Fundamental parameters for low-mass stars in eclipsing binariesthe mass-radius relation

  1. Ramón Iglesias Marzoa
Supervised by:
  1. María Jesús Arévalo Morales
  2. Mercedes López Morales

Defence university: Universidad de La Laguna

Year of defence: 2019

  1. Astrofísica

Type: Thesis

Teseo: 597613


This work is inscribed in the framework of an observational long-term project to measure reliable fundamental parameters of low-mass stars (M<1 Msun). The final goal of this project is to obtain precise values of masses and radii for stars in this mass range in order to compare them with those predicted by the theoretical models. This set of parameters can only be obtained simultaneously from radial velocity and light curve observations of double-lined detached eclipsing binary stars with low-mass components. Previous studies of low-mass stars in detached eclipsing binaries showed that, for a given mass, the radii are larger than those predicted by stellar models by up to 10-15%. Most of these studies are based on the analysis of short period eclipsing binaries (P<3 days), in which the components are tidally synchronized and the orbit is circularized. The most feasible explanation for this discrepancy is the "magnetic disruption" hypothesis. In this hypothesis, the stellar magnetic fields are strengthened by the high rotational velocities of the tidally locked components. As a consequence, the stellar activity is also enhanced, and, typically, these low-mass stars show extensive spots covering a noticeable fraction of the stellar surfaces, among other features. As the stellar spots are regions of lower temperature, they irradiate less flux, specially at shorter wavelengths. To compensate the flux blocked by the spotted surface, the stellar radius must be larger than in an unspotted star. The study of these discrepancies is limited by the low number of known detached eclipsing systems with low-mass components, mainly due to the low intrinsic luminosity of these objects. This situation has improved with the launch of wide-field photometric time-series surveys, which provide new low-mass eclipsing systems candidates. Our long-term observational project consists in the spectroscopic and photometric follow-up, and the characterization of a number of such systems, in order to accurately measure their masses and radii. In this thesis I present the results obtained for two of these low-mass eclipsing binaries: T-Cyg1-12664 and NSVS 10653195. At the beginning of this work, we realized the necessity to fit simultaneously in a reliable way all the orbital parameters of spectroscopic binaries, from which a significant fraction are also eclipsing binaries. The usual procedure is to fix the orbital period, which is typically measured from a periodogram or from light curves, if they are available; and then to fit by eye an initial set of parameters to use them as a starting point for other fitting procedures. The fact that the function to fit has six or seven parameters, depending whether the system is single-lined or double-lined, complicates the fit. Hence, we developed a robust code to fit simultaneously all the spectroscopic orbital parameters using a technique known as adaptive simulated annealing (ASA). This code is called "rvfit" and is designed to be used without any a priory knowledge of the values of the orbital parameters, which makes this code suitable to be used in radial velocities (RV) surveys. The code assumes a star-star or star-planet system in keplerian orbits, so it is not suitable for multi-planetary systems. To demonstrate the performance of the code, we fitted several systems, including double-line and single-line spectroscopic binaries and exoplanet systems. As a test, we also fitted the radial velocity curve of Kepler 78b, the first Earth-like transiting exoplanet discovered in the Kepler mission data. The eclipsing binary T-Cyg1-12664 was discovered in the T-Cyg1 field of the TrES survey, and turned out to also be included in the Kepler Mission field of view. This means that there is photometric high-precission continuous monitoring of the light curve of this system over several years with a cadence of ~27 min. This long-term monitoring allows to study the evolution of spots and search for period changes. Our ground-based differential photometry overlaps with the Kepler observations and this allows to accurately characterize the system. T-Cyg1-12664 was first analyzed by Devor (2008), and later by Çakirli et al (2013), but the physical parameters obtained by the later authors differ from values expected for low-mass stars. In this last paper, the authors assumed a circular orbit, but we demonstrate that, instead, the orbit is slightly elliptical. Our results show that the system of T-Cyg1-12664 is composed of an oversized G6V primary star and a M3V secondary near the full-convection limit, which agrees well with models. The low-mass eclipsing binary NSVS 10653195 was listed in the NSVS survey low-mass eclipsing binaries candidates, and its light curve was first analyzed by Coughlin (2007). Other light curves were later analyzed by Wolf et al. (2010), Zhang et al. (2014), and Zhang et al. (2015). All these publications fitted only light curves, and therefore, the masses of the components and the scale of the system (the semi-major orbital axis, a) remained unknown. Our study performs, for the first time, a joint analysis of radial velocities and optical and infrared light-curves, and provides accurate masses and radii for the components. By the time we fitted this system, the Gaia Data Release 2 (DR2) was available and the distance to this system was know with precission. This binary system is composed of similar K6V and K7V stars, and their physical dimensions confirm the radius inflation scenario. Finally, I present in the Appendixes the results of a photometric calibration of several low-mass eclipsing binary systems in our observational program, in order to obtain reliable out-of-eclipse photometric colors. Some of these results are used in two of the paper published. The photometric colors are key to measure the mean effective temperatures of the binary. The Appendixes also contain copies of the three papers published with the results of Chapters 2, 3, and 4.