Stellar activity in exoplanet hosts

Abstract

[cat] Actualment, la major part dels esforços per la cerca i caracterització d’exoplanetes de tipus terrestre es centren en aquells que orbiten estrelles de baixa massa. Algunes de les característiques importants de l’estructura i els processos d’aquest tipus d’estrelles són encara poc coneguts, i per tant és important fer-ne un estudi acurat com un dels següents passos en el camp de l’exoplanetologia. En concret, el senyal produïit pels fenòmens d’activitat estel•lar, degut a la presència de taques i fàcules, introdueix variacions en les mesures fotomètriques i espectroscòpiques amb una periodicitat modulada per la rotació de l’estrella. En aquesta tesi s’ha realitzat un estudi dels fenòmens d’activitat a la fotosfera d’estrelles de baixa massa a través de diverses tècniques que ens han permès modelitzar o bé simular les relacions entre les propietats estel•lars i les dades observables. La simulació de mostres estadístiques d’estrelles basada en les relacions conegudes entre l’activitat estel•lar i la rotació ha permés implementar una tècnica per estimar la inclinació de l’eix de les estrelles a partir de mesures espectroscòpiques, resultant en un catàleg amb les millors candidates per una cerca de planetes amb trànsits. Per a la caracterització dels fenòmens relacionats amb l’activitat i un estudi acurat dels seus efectes sobre les mesures i cerques d’exoplanetes, s’ha aprofitat la disponibilitat de dades fotomètriques de Kepler juntament amb tècniques de modelització de taques, centrant-nos particularment en l’estudi de LHS 6343 A per obtenir informació d’activitat en la seva superfície. A més, s’ha dissenyat un simulador dels efectes d’activitat a la fotosfera basat en tècniques d’integració de superfície i models d’atmosfera Phoenix, que permet obtenir sèries temporals de dades sintètiques i estudiar els efectes de les regions actives sobre les mesures de trànsits de planetes. En particular, s’ha modelat el cas de HD189733. Els resultats d’aquest treball conclouen que és essencial una correcta modelització del senyal de l’activitat estel•lar en el camp de l’exoplanetologia, i es proporcionen algunes eines i estratègies per tal de caracteritzar i corregir aquests efectes i obtenir-ne informació astrofísica.

[eng] Most of the efforts on the search and characterization of Earth-like exoplanets are currently focused on low mass stars. Some important properties related to the structure and processes in this type of stars are still unknown, so a careful characterization is essential as one of the next steps in exoplanet sciences. The characterization of stellar activity in low mass stars was carried out through several techniques that allowed us both to model and to simulate the relationships between the observational data and the stellar properties. Several empyrical relations for low mass stars allow to find correlations between certain activity indicators and the rotation period. These have permitted us to generate synthetic samples of stars with stochastic distributions of stellar and geometric properties allowing to estimate the inclination of the rotation axis from the distribution in the activity-vsini diagram. The methodology was applied to a sample of 1200 observed low mass stars and the best candidates for a targeted transit search were selected. Spot modelling techniques allow to obtain physical information about the stellar surface from time series photometric and spectroscopic data. In this work we analyse Kepler photometry of LHS 6343 A, an M-dwarf being eclipsed by a brown dwarf companion every 12.718 days, and showing photometric oscillations with the same periodicity and a phase lag of 100º from the eclipses. The accurate modeling of the Kepler data allowed to explain these oscillations with the presence of active regions appearing at a fixed longitude, thus suggesting a possible magnetic connection between both components. On the other hand, we also studied an alternative explanation for the photometric oscillations in LHS 6343 A in terms of the Doppler beaming effect, showing that this could be the main cause of the observed oscillations. Stellar activity effects are responsible for the noise observed at different amplitude and timescales on time series data. Such noise represents one of the main limitations for exoplanetary sciences. In order to characterize it, we designed a methodology to simulate the photosphere of an active rotating star through the integration of small surface elements from Phoenix atmosphere models. This allows to characterize the signal produced by activity and further study its relationship with the stellar properties, as well as the possible effects produced on exoplanet measurements. The methodology allowed us to present several strategies in order to correct or reduce the effects of spots on the photometry of exoplanet transits, as these may induce significant variations on the measurement of the planetary radius. We focused on a comprehensive analysis of HD 189733, a K5 star hosting a giant planet, which has simultaneous photometric (MOST) and spectroscopic (SOPHIE) data available. An accurate surface map was obtained using the methodology above, accurately reproducing the light curve and radial velocity observations. Such map was used in order to study the effects of activity on the exoplanet transits. We showed that the effects of spot-crossing events are significant even for mid-infrared wavelengths. Moreover, the chromatic effects of spots not occulted by the planet show a signal with a wavelength dependence and amplitude that are very similar to the signature of the atmosphere of a planet dom- inated by dust. The radial velocity theoretical curve is in agreement with the observations up to the typical instrumental systematics of SOPHIE. The results from this work conclude that it is essential to correctly model the stellar activity signals for exoplanetary sciences, and we provide some tools and strategies to characterize and reduce such effects and extract astrophysical information.