2020MNRAS.494..108W

High-resolution spectroscopy (R ≥ 20 000) is currently the only known method to constrain the orbital solution and atmospheric properties of non-transiting hot Jupiters. It does so by resolving the spectral features of the planet into a forest of spectral lines and directly observing its Doppler shift while orbiting the host star. In this study, we analyse VLT/CRIRES (R=100 000) L-band observations of the non-transiting giant planet HD 179949 b centred around 3.5 µm. We observe a weak (3.0σ, or S/N = 4.8) spectral signature of H₂O in absorption contained within the radial velocity of the planet at superior-conjunction, with a mild dependence on the choice of line list used for the modelling. Combining this data with previous observations in the K band, we measure a detection significance of 8.4 σ for an atmosphere that is most consistent with a shallow lapse-rate, solar C/O ratio, and with CO and H₂O being the only major sources of opacity in this wavelength range. As the two sets of data were taken 3 yr apart, this points to the absence of strong radial-velocity anomalies due, e.g. to variability in atmospheric circulation. We measure a projected orbital velocity for the planet of K_P = (145.2 ± 2.0) km s^–1 (1σ) and improve the error bars on this parameter by ∼70 per cent. However, we only marginally tighten constraints on orbital inclination (66.2^+3.7_–3.1 deg) and planet mass (0.963^+0.036_–0.031 Jupiter masses), due to the dominant uncertainties of stellar mass and semimajor axis. Follow ups of radial-velocity planets are thus crucial to fully enable their accurate characterization via high-resolution spectroscopy.