Fundamental effective temperature measurements for eclipsing binary stars

Abstract

Modern high-precision spectroscopy and photometry has made it possible to directly measure masses and radii for stars in detached eclipsing binaries (DEBs) to 0.5% or better. These stars, due to their wide orbits, are suitable for testing models of single stars. However, effective temperature (Teff) estimates for FGK-type stars are lagging behind, with Jofre et al. (2019) noting that most Teff estimates are no more accurate than 50 K. This is an important issue to address, as inaccurate Teff estimates limit the calibration of stellar models using DEBs. A consistent Teff scale for stars with a range of masses and ages is essential to avoid spurious trends in population studies for exoplanet host stars and Galactic archaeology. This thesis aims to address this problem by developing a new method to measure the fundamental, i.e. direct, effective temperatures for stars in DEBs. The new method is based on the Stefan-Boltzmann law. We use a Bayesian approach to obtain the integrated bolometric uxes for the two stars from observed magnitudes, colours, and ux ratios. Angular diameters are obtained from measurements of the stellar radii using TESS light curves and radial velocities measured from high-resolution spectroscopy, and parallax from the Gaia satellite. Fundamental effective temperatures have been measured for five FGK-type stars in three DEBs: the F7V+K0V binary AI Phoenicis, the F5V+F6V binary CPD-54 810, and the primary component of the F+M binary with a low-mass component EBLM J0113+31. The results significantly improve on the accuracies of existing Teff estimates: better than ±0:4% for AI Phoenicis and ±0:7% for CPD-54 810 and EBLM J0113+31 A. The choice of model SED has no impact on the derived effective temperatures. This work provides the basis for building a large sample of well-studied FGKtype stars with very accurate and precise Teff measurements. Such a sample has a wide applicability in the rest of astrophysics, as it can be used for testing and calibrating stellar models and large-scale spectroscopic surveys.