La photométrie des transits permet de détecter des planètes extrasolaires en mesurant leur rayon. Dans cet objectif, le télescope spatial CoRoT (Convection, Rotation et Transit planétaires), lancé en décembre 2006, est doté d'un photomètre de haute précision permettant à la fois l'étude de la structure interne des étoiles par astérosismologie et la détection de planètes par la méthode des transits. Pour cette thèse, j'ai développé différents outils informatiques permettant la détection, l'analyse détaillée de transits dans les courbes de lumière de CoRoT. J'ai appliqué ces outils aux quelques 12 000 étoiles observées durant chacune des dix premières campagnes d'observation. La collaboration des équipes de détection et de suivi au sol par d'autres méthodes d'observation a permis, à ce jour, la découverte de quinze planètes et deux naines brunes.Il est possible de contraindre les modèles de ces exoplanètes grâce à la connaissance des paramètres (masse, rayon): gazeuses, de glace, telluriques, ou d'autres types mixtes.La recherche de planètes rocheuses en particulier est un objectif motivé tant par la rareté des détections de ces objets jusqu'à présent (liée à leur faibles rayon et masse), que par la grande variété potentielle de leur nature.Ainsi, la découverte de CoRoT-7b, la première exoplanète compatible avec un modèle rocheux et dont le rayon (1,6 rayon terrestre) et la masse (environ 7 masses terrestres) ont pu être mesurés, a permis d'élaborer un modèle physique auquel j'ai contribué.J'ai étudié la possibilité d'observer cette planète très chaude en proche infrarouge avec le JWST, au cours de son orbite, afin d'estimer le contraste de température entre les faces jour et nuit. Ceci doit permettre de confirmer / invalider notre modèle qui suppose l'absence d'une atmosphère suffisamment dense pour redistribuer la chaleur à la surface de CoRoT-7b. / The transit photometry makes it possible to detect exoplanets by measuring their radii. Pursuing this goal, the space telescope CoRoT (Convection, Rotation and planetary Transits), launched in December 2006, is equiped with a high-precision photometer allowing both planet detection by transit photometry, and stellar physics studies (asteroseismology).For this PhD thesis, I have developed various computing tools for the detection and detailed analysis of the transits in CoRoT light curves. I have applied these tools to almost 12 000 stars observed during each of the first ten campaigns of observation. The collaboration between the detection and ground based follow-up teams led to the discovery, up to now, of fifteen planets and two brown dwarfs.It is possible to constrain the physical natures of these exoplanets thanks to the knowledge of the masses and radii: they can be gazeous, icy, rocky or with an mixed nature.The search for rocky planets in particular, is a goal motivated by their singular nature, and both by the paucity of detections of these objects (due to their low masses and radii), Thus, the discovery of CoRoT-7b -- the first exoplanet compatible with a rocky model and whith measured radius (1.6 Earth radius) and mass (around 7 Earth masses) -- allowed us to develope a physical model to which I contributed :I studied the possibility of observing this very hot planet in the near infrared range with JWST, during its orbit, to estimate the temperature contrast between the day and the night faces. This should allow to confirm / invalidate our model with atmosphere dense enough to redistribute heat at the surface of CoRoT-7b.

The search for extrasolar planets had been ongoing for many years when Mayorand Queloz discovered 51 Pegasi b in 1995. It was a giant gas planet similar to Jupiter, but with a larger radius and of only half of Jupiter’s mass. Theso called Hot Jupiter was observed to orbit its host star 7 times closer thanMercury is orbiting the Sun. Theoretical models at the time stated that gasgiants could not form in such a short distance to the host star. Thus, thisdiscovery was completely unexpected. It was the beginning of a new field ofresearch where the diversity of exoplanets is the most remarkable discovery, challenging theoretical models. Thanks to the Kepler space telescope and anew generation of space missions such as TESS, thousands of exoplanets havebeen discovered and thousands of planet candidates await confirmation. In this thesis I have studied all confirmed exoplanets to this date, which havebeen discovered by the radial velocity and/or the transit method. The planetparameters and their stellar hosts are available on NASA’s Exoplanet Archive.For all planets < 100 M⊕, I have assessed and updated the parameters for eachplanet in particular when several solutions exist. There are several types ofplanets, but the focus of this work are small planets which come in two sizes: Rocky super-Earths, and the slightly larger and lower density sub-Neptune. Different types of planets have different radii and mass ranges, which togetherwith composition and interior structure are separating the types from each other. These mass and radii ranges are however not universally defined, and in thisreport the super-Earth and sub-Neptune ranges are discussed together with their typical characteristics. The radii and mass ranges of the two different classes of small planets are overlapping and are often difficult to classify. In particularfor planets in between 2 R⊕ and 3 R⊕, there is an ambiguity of structure and composition. This report will also investigate how planet properties depend on the stellarhost properties and on the orbital distances to the stars. One of my mainresults is that sub-Neptunes are common orbiting host stars with low metallicity, in contrast to super-Earths which are common orbiting host stars with highmetallicity. Other parameters, such as stellar effective temperature, seem to have no influence on planet properties. Super-Earth’s are found at a wide range of orbital distances while the sub-Neptunes cluster in a narrow range of orbital distances to their host star. Sub-Neptunes have an atmosphere, and are orbiting at distances where the atmosphere does not evaporate from intense host star radiation. If an atmospheree vaporates, only the rocky core of the planet is left. Thus, some super-Earths might have been sub-Neptunes that have lost their atmospheres. My second main result is that planets with characteristics of sub-Neptunes (with respect to density and interior structure) of 10 M⊕ to 15 M⊕ have radiibetween 2 R⊕ and 4.5 R⊕. Sub-Neptunes in the upper mass limit, between 15M⊕ to 17 M⊕, have radii from 2.6 R⊕ to 7.5 R⊕. And finally, my third result is the relation between planet density and equilibrium temperature. The density of all planets with masses < 15 M⊕ is Earth-like for equilibrium temperatures > 1400 K. For lower equilibrium temperatures corresponding to longer orbital periods, or lower-mass and cooler stars, planetswith masses < 15 M⊕ have a larger spread in densities. However, it never fallsbelow a diagonal linear trend in the density against equilibrium temperature diagram described by ρ = 2.6 × log10(Teq) − 7.46.

Plus d’un millier d’exoplanètes ont été découvertes depuis une dizaine d’années. Plus incroyable encore, nous pouvons maintenant caractériser les atmosphères de ces mondes lointains. Des spectres de Jupiter-chauds tels que HD 189733b et HD 209458b et de planètes similaires à Neptune telles que GJ1214b sont déjà disponibles et ceux de planètes plus petites le seront bientôt. La plupart des observations caractérisent l’état moyen de l’atmosphère. Pour les cas les plus favorables, l’observation des courbes de phase et la technique de cartographie par éclipse secondaire permettent d’obtenir une résolution en longitude et en latitude. Les planètes les plus proches de leurs étoiles sont aussi les plus faciles à observer. Ces mondes chauds sont radicalement différents des exemples que nous avons dans le système solaire. Modéliser correctement leurs atmosphères est un défi à relever pour comprendre les observations présentes et à venir. Durant cette thèse, j’ai développé des modèles de différente complexité pour comprendre les interactions entre la structure thermique, la composition, la circulation atmosphérique et l’évolution à long terme des exoplanètes irradiées. La forte luminosité de leur étoile hôte détermine le climat de ces planètes. Elle engendre une circulation atmosphérique qui maintient l’atmosphère dans un état de déséquilibre thermique et chimique, affectant son évolution. Avec les futurs instruments de nombreuses autres planètes vont être découvertes et caractérisées. Nos modèles seront testés sur une large diversité de planètes, ouvrant les portes de la climatologie aux exoplanètes. / More than a thousand exoplanets have been discovered over the last decade. Perhaps more excitingly, probing their atmospheres has become possible. We now have spectra of hot Jupiters like HD 189733b and HD 209458b, of Neptune-like planets like GJ1214b and even smaller planets are within reach. Most exoplanet atmospheric observations are averaged spatially, often over a hemi- sphere (during secondary eclipse) or over the limb of the planet (during transit). For favorable targets, longitudinal and latitudinal resolution can also be obtained with phase curve and secondary eclipse mapping techniques respectively. The closer the planet orbits to its star, the easier it is to observe. These hot planets strongly differ from the examples we have in our solar-system. Proper models of their atmospheres are challenging yet necessary to understand current and future observations. In this thesis, I use a hierarchy of atmospheric models to understand the interactions between the thermal structure, the composition, the atmospheric circulation and the long-term evolution of irradiated planets. In these planets, the large stellar irradiation dominates the energy budget of the atmosphere. It powers a strong atmospheric circulation that transports heat and material around the planet, driving the atmosphere out of thermal and chemical equilibrium and affecting its long-term evolution. Future instruments (Gaia, SPIRou, CHEOPS, TESS, PLATO etc) will discover many more planets that the next generation of telescopes (GMT, TMT, E-ELT or JWST) will characterize with an unprecedented accuracy. Models will be tested on a large sample of planets, extending the study of climates to exoplanets.

This report describes simulations of observations with the near-infrared high-resolution spectrometer SIMPLE that is proposed to the ESO telescope E-ELT. We simulate M4 and G2 stars with transiting Earth-like planets and the goal is to distinguish spectral features originating from the atmosphere of the exoplanet. Noise levels of different magnitudes are added to the simulations and the minimal signal-to-noise required for detection of the atmosphere is estimated. Our conclusion is that detection of atmospheric features looks promising using this setup.