Journal. Astronomy & Astrophysics
Abstract. Transmission spectroscopy can be used to constrain the properties of exoplanetary atmospheres. During a transit, the light blocked from the atmosphere of the planet leaves an imprint in the light coming from the star. This has been shown for many exoplanets with different techniques, with both photometry and spectroscopy. We aim at testing chromatic line-profile tomography as a new tool to investigate exoplanetary atmospheres. The signal imprinted on the cross-correlation function (CCF) by a planet transiting its star is dependent on the planet-to-star radius ratio. We want to verify if the precision reachable on the CCF obtained from a subset of the spectral orders of the HARPS spectrograph is enough to discriminate the radius of a planet at different wavelengths. Methods. We analyze HARPS archival data of three transits of HD 189733b. We divide the HARPS spectral range in 7 broadbands, calculating for each band the ratio between the area of the out-of-transit CCF and the area of the signal imprinted by the planet on it during the full part of the transit. We take into account the effect of the limb darkening using the theoretical coefficients of a linear law. Averaging the results of three different transits allows us to obtain a good quality broadband transmission spectrum of HD 189733b, with a precision greater than that of the chromatic RM effect. Results. We proved chromatic line-profile tomography to be an interesting way to reveal broadband transmission spectra of exoplanets: our analysis of the atmosphere of HD 189733b is in agreement with other ground- and space-based observations. The independent analysis of different transits puts in evidence the probability that stellar activity plays a role in the extracted transmission spectrum. Care has thus to be taken when claiming for Rayleigh scattering in the atmosphere of exoplanets orbiting active stars using only one transit.