Thèse de doctorat
Résumé : About half of the stars in our Galaxy are born with a companion forming so-called binary systems. In these systems, the two stellar components are gravitationally bound and orbit each other during their evolution, which can be strongly affected by interaction with the companion, especially when one of the stars evolves to giant dimensions. Binary interactions also affect the evolution of the period and the eccentricity of the orbit. Many aspects of interaction physics in binaries are not understood yet, and investigating the products that result from interacting systems is crucial to unravelthe physical mechanisms involved. These statements are applicable to binaries of all masses, but this thesis focuses on low- and intermediate-mass systems, with components with initial masses between one and five times the mass of the Sun (Msun).Among the prototypical families of post-interaction binary systems in this mass range, one must mention Barium (Ba) stars. Ba stars are main-sequence or giant stars which show an enhancement of chemical elements that should not yet be overabundant at these evolutionary stages. Currently it is widely accepted that these chemicals were transferred from a more evolved companion during a phase of mass transfer in the system, and that this companion evolved into a cool white dwarf, which is too dim to be directly detected. However, the motion of the Ba star and its white-dwarf companion orbiting each other is detectable, and understanding the observed orbital properties of Ba-star systems, as well as the stellar properties of the Ba star and its polluter, is a key to the system’s interaction history. Moreover, identifying the mechanisms that impact the formation of Ba stars will teach us about binary evolution and interaction physics.This thesis presents observational and theoretical efforts made to contribute to a better understanding of the properties of Ba stars. The beginning of this work was well timed with the first data release of the Gaia mission, and this thesis exploits the synergy between Gaia data, of unprecedented quality, high-resolution spectroscopy, long-term binary monitoring programmes, and state-of-the-art stellar and binary evolution models. In the first part, we focus on the development of a methodology to accurately locate Ba stars on the Hertzsprung-Russell diagram and determine their evolutionary status and their masses. Ba giants have masses from 1 to 5 Msun, with the distribution peaking around 2 Msun. Accurate and precise stellar masses are critical in astrophysics, but very difficult to measure, and the mass distribution of Ba stars presented here brings new observational constraints that were inaccessible before the Gaia era and this work.Giant Ba stars have been intensively investigated since their discovery, but main-sequence Ba stars are much less sampled and studied. Since the polluting mass-transfer is expected to happen when the Ba star is still on the main-sequence, the properties of these systems are important to have a complete picture of the formation and evolution of Ba stars. The second part of this thesis presents the largest systematic study of the stellar and orbital properties of main-sequence Ba stars and a thorough comparison of these properties to those of Ba giants. Main-sequence Ba stars show orbital periods between 200 and 11 000 days and eccentricities over the whole range from 0 to > 0.8, similar to Ba giants. However, the main-sequence Ba stars studied in this thesis cover a much narrower mass range than their giant counterparts, from 0.9 to 1.6 Msun. Thedistributions of masses, periods and eccentricities that resulted from this analysis provide strong constraints to theoretical studies.The comparison between main-sequence and giant Ba stars allowed us to investigate their evolution between these two phases. A second stage of binary interaction, this time between the main-sequence Ba star and its white-dwarf companion, also takes place in some systems, affecting the distribution of orbits observed among Ba giants. Our binary evolution models suggest that main-sequence Ba stars with M < 2.0 Msun in systems with periods below 500 to 1000 days (depending on mass and metallicity) will merge with their companions and never become core-He burning giants.Despite our contribution to describe the properties of Ba stars, further research is necessary to entirely understand their formation and evolution. From the observational side, identifying more massive main-sequence Ba stars is necessary to complete the full evolutionary picture. Moreover, not much is known about the white-dwarf companions, and hence about their progenitors, responsible for the contamination of the Ba-star envelopes. This information can constrain the initial conditions that lead to the formation of a Ba-star system. Of course, further theoretical research is also necessary, and the observational constraints presented here will contribute to the investigation of interaction mechanisms in low- and intermediate-mass binary systems.Astronomy is a field in evolution itself, and these are exceptional times for stellar and binary evolution research. New telescopes, instruments, surveys, and theories keep coming, providing new pieces to solve this and many other puzzles. As an example, after the third Gaia data release, we will have new crucial information opening different perspectives to study Ba stars through the individual masses of the unseen white-dwarfs companions. Low- and intermediate-mass stars are very common, and they are major contributors of carbon, nitrogen, many heavy elements, and dust to the host Galaxy, and since at least half of them evolve in binary systems, interaction physics are essential to describe not only stellar evolution, but also the chemical evolution of the Galaxy.
Près de la moitié des étoiles de notre Galaxie naissent avec un compagnon et sont amenées à interagir durant leur évolution via des processus encore mal compris. En étudiant des systèmes d’étoiles doubles qui ont subit des interactions, nous pourrons découvrir les mécanismes physiques impliqués. Dans cette thèse, cette stratégie est appliquée pour étudier une certaine classe de systèmes binaires dont font partie les étoiles à baryum (Ba). Les étoiles Ba sont des étoiles géantes ou de la séquence principale (Ba naines) qui sont caractérisées par une sur-abondance en éléments chimiques, tel le baryum, qui ne peut provenir de l'étoile elle-même. Cette pollution chimique résulte d'un transfert de matière antérieur entre l'étoile initialement la plus massive et son compagnon. L'étoile à l’origine de la pollution est maintenant au stage naine blanche. L'analyse des paramètres orbitaux des étoiles à Ba et des propriétés des composantes stellaires, est une clé pour comprendre l’histoire de ces interactions.Cette thèse présente les efforts observationnels et théoriques apportés pour atteindre ce but. Cette thèse exploite la synergie entre les données du satellite Gaïa mise à disposition au début de ce travail, la spectroscopie haute résolution, les programmes de surveillance binaire à long terme et les modèles d'évolution stellaire et binaire de pointe.Dans la première partie, nous décrivons la méthodologie utilisée pour localiser avec précision grâce aux données Gaïa, les étoiles à Ba dans le diagramme de Hertzsprung-Russell et ainsi déterminer leur statut évolutif et leur masse. Nous montrons notamment que les géantes ont des masses variants entre 1 et 5 Msun avec une distribution piquée à 2 Msun.La deuxième partie présente la plus grande étude systématique des propriétés stellaires et orbitales des étoiles à Ba naine et une comparaison approfondie de ces propriétés avec celles des étoiles géantes. Les étoiles Ba de la séquence principale montrent des périodes orbitales comprises entre 200 et 11 000 jours et des excentricités sur toute la plage de 0 à 0.8, similaires aux Ba géantes. Cependant, les étoiles Ba naine couvrent une gamme de masse plus étroite que leurs homologues géantes, entre 0.9 à 1.6 Msun.D'un point de vue évolutif, les étoiles à Ba géante sont les descendantes des étoiles à Ba naine. Nous avons effectué la modélisation de ces systèmes avec le code d'évolution stellaire BINSTAR et comparé les prédictions de nos modèles avec les observations obtenues. Les simulations permettent de rendre compte de la majorité de propriétés orbitales des géantes mais suggèrent que les étoiles à Ba naine de masse M < 2.0 Msun et avec des périodes inférieures à 500 jours (pour une composition solaire) fusionneront avec leur compagnon et ne deviendront jamais des géantes.