An exoplanet is an astronomical object that orbits one or more stars different from the Sun and has a mass lower than that necessary to produce the thermonuclear fusion of deuterium. Among the variety of techniques implemented to discover these objects, one of the most reliable is the radial velocity (RV) method. This technique is based on measuring the Doppler effect experienced by the spectral lines of a star, searching for signals caused by the gravitational field of a planet. The search of low-mass exoplanets through this technique requires high-resolution spectrographs such as HARPS or CARMENES, that achieve an RV precision around 1 m/s, or ESPRESSO, which provides an improved RV precision up to 10 cm/s. The RV precision will determine the minimum amplitude of the planetary signals that can be detected by the instrument, and therefore, the mass of the possible planetary companions.
The wavelength calibration of the spectrographs is key to take advantage of the resolution and stability of these instruments and thus be able to maximize the precision limit. Currently, the Laser Frequency Comb (LFC) is considered one of the most accurate calibration systems, providing a short-term repeatability of 1 cm/s through a regular pattern of spectral lines referred to an atomic clock.
Regardless of the RV precision achieved, the detection of planetary signals is hampered by the short-term and long-term stellar activity signals. The former are caused by the stellar rotation along with the presence of spots and plages in the stellar surface, while the latter are related to the magnetic cycle of the star. The modeling and subtraction of these signals are fundamental to detect and characterize correctly the planetary signals, especially in M dwarfs, in which the periodicity of the short-term activity signals is compatible with the habitable zone of these stars, where liquid water could potentially exist on the surface of a planet. To distinguish which is the origin of the RV signals detected, different chromospheric activity indicators are used, along with relations between the stellar rotation and other properties of the star.
In this thesis, we develop a calibration pipeline for the LFC in order to achieve the most precise wavelength solution possible. We compare the LFC results with those obtained by a ThAr lamp (the most commonly used calibration instrument) using more than nine thousand HARPS spectra from the LFC commissioning and previous testing campaigns. The pipeline includes an RV calculation module that allows the user to self-build Cross-Correlation Functions (CCFs) using different stellar masks. We perform an RV accuracy study comparing our results with those reported by NASA's Jet Propulsion Laboratory (JPL), and finally broaden the module's availability to other spectrographs.
We contribute to the HADES (HArps-n red Dwarf Exoplanet Survey) and RoPES (Rocky Planets in Equatorial Stars) programs, which are focused on the search and characterization of rocky exoplanets with the ultimate goal of detecting Earth-like planets in the habitability zone. The first program is based on a sample of 79 M dwarfs observed with HARPS-N, while the second one contains a sample of 30 G- and K-type stars observed with HARPS-N and HARPS with the aim of taking advantage of the LFC features in this last spectrograph. We conducted a stellar activity study of the sample from both programs. The most prolifi c is the one carried out on Barnard's Star (the closest single star to the Solar system) based on a 15 yr dataset coming from eight different spectrographs and four photometric sources. We computed different activity indicators to characterize the rotation of the star (including its differential rotation) and its long-term magnetic cycle, both being key elements for the discovery of the first planetary companion detected around this star. The analyses of the remaining
stars served to cross-correlate the detection of 33 new rotation signals and 18 new cycle signals within the M-type star sample.
The detailed stellar activity analysis of the star sample considered in this thesis led to the discovery of a super-Earth around the M-dwarf GJ 740 using HARPS-N and CARMENES data. The planet has a minimum mass of
2.96 +0.50-0.48 Me, it is located at 0.029 +0.001-0.001 AU from its parent star, and presumably has a rocky composition. Additionally, we contributed to the detection of eight other new exoplanets around seven different stars from our sample: GJ 625 b, GJ 3942 b, Gl 15Ac, Gl 686 b, Gl 49 b, GJ 685 b, HD176986 b, and
HD176986 c.
The extremely high RV precision provided by the new-generation spectrograph ESPRESSO has expanded the limits to reach planetary signals with lower amplitudes. We combine ESPRESSO spectroscopy with K2 photometry to characterize the two-planet system orbiting the K2-38 star, in one of the first published works within the ESPRESSO GTO. We fi nd that K2-38 b is an iron-rich super-Earth with a size of 1.54+-0.14 Re and a mass of 7.3 +1.1-1.0 Me (and therefore, one of the greatest densities reported to date) while K2-38 c is a rocky sub-Neptune with an H2 envelope that has a 2.29+-0.26 Re and a mass of 8.3 +1.3 Me. Each planet is located on each side of the radius valley (a region that lacks detected planets because of photoevaporation effects) due to the different irradiation levels and evaporation processes that they are exposed to, along with core-powered mass-loss mechanisms.