Carbon monoxide fluorescence from Titan's atmosphere

Abstract

We report on the discovery of emissions due to carbon monoxide from Titan's atmosphere, from mid-infrared observations with the ISAAC spectrometer at the Very Large Telescope and covering the 4.50–4.85 μm range. We detected about 45 emission lines coinciding with CO ro-vibrational lines, including CO(1–0) (P18 to R11) and CO(2–1) (P11 to R11). We show that these emissions cannot be generated thermally but occur in non-LTE conditions, due to radiative de-excitation from the v = 1 and v = 2 CO levels after excitation at 4.7 and 2.3 μm by solar radiation. A complete fluorescence model is then developed, allowing to compute the state populations of the two most abundant CO isotopes and N 2(1). It includes absorption by CO and CH 4, and vibrational–thermal and vibrational–vibrational collisional exchanges with CO, N 2, CH 4, and H 2. Emerging radiances at the top of the atmosphere are evaluated with a line-by-line code and compared to observations. Contribution functions show that the CO emissions sound Titan's stratosphere: while the (1–0) lines generally probe two layers, located respectively at 100–250 km and 300–550 km, the (2–1) lines are sensitive to the intermediate layer at 150–300 km. A sensitivity study is performed to establish the effect of the main model parameters (temperature profile, collisional scenario, and CO stratospheric abundance) on the results. Models reproduce the essential structure of the observed emissions. The (1–0) fundamental band is generally well fit with a nominal CO mixing ratio of 32 ppm—as inferred in the troposphere from observations at 4.80–5.10 μm (Lellouch et al., 2003, Icarus 162, 126–143). However, this band is only weakly dependent on the CO abundance, and given temperature and collisional scenario uncertainties, it constrains the CO stratospheric mixing ratio only to within a factor of ∼3. In addition, the nominal model with 32 ppm CO underestimates the first hot (2–1) transition by approximately a factor of 2. This discrepancy can be resolved by a combined adjustment of collisional rates and an increased CO stratospheric ratio of 60 ppm, consistent with the determination of Gurwell and Muhleman (2000, Icarus 145, 653–656). In contrast, the CO vertical profile suggested by Hidayat et al. (1998, Icarus 133, 109–133), strongly depleted above 200 km, cannot match the data for any realistic collisional scenario, and is therefore not supported by our results.