2009ApJ...700.1349T

We report extensive radial velocity measurements of the two giant components of the detached, 104 day period binary system of Capella. Our highly accurate three-dimensional orbital solution based on all existing spectroscopic and astrometric observations including our own yields much improved masses for the primary and secondary of 2.466±0.018 M_☉ and 2.443±0.013 M _☉, with relative errors of only 0.7% and 0.5%, respectively. The mass ratio is considerably closer to unity than previously believed, which has an impact on assessing the evolutionary state of the system. Improved values are presented also for the radii (11.87±0.56 R_☉ and 8.75±0.32 R_☉), effective temperatures (4920±70 K and 5680±70 K), and luminosities (79.5±4.8 L_☉ and 72.1±3.6 L_☉). The distance is determined to be 13.042±0.028 pc, based on the accurate orbital parallax. The projected rotational velocities and individual rotation periods are also known. Capella is unique among evolved stars in that, in addition to all of the above, the chemical composition is known as well. This includes the overall metallicity [m/H], the carbon isotope ratio ¹²C/¹³C for the primary, and the lithium abundance and C-to-N ratios for both components. We present new or revised values for some of these. The latter three quantities are sensitive diagnostics of evolution, and change drastically for giants as a result of the deepening of the convective envelope during the first dredge-up. The secondary is crossing the Hertzprung gap, while the primary is believed to be in the longer-lived core helium-burning phase. Previous studies using only the masses, temperatures, and luminosities have found good agreement with stellar evolution models placing the primary in the clump. Here, we compare all of the constraints simultaneously against three sets of current models. We find that they are unable to match all of the observations for both components at the same time, and at a single age, for any evolutionary state of the primary. This shows the great importance of chemical information for assessing the evolutionary state of giant stars. A comparison with models of tidal evolution yields similarly disappointing results, when tested against the fact that the orbit is circular, the primary is rotating synchronously, the secondary ∼12 times faster than synchronous, and the spin axes are apparently aligned with the axis of the orbit. When confronted in detail, our understanding of the advanced stages of stellar evolution is thus still very incomplete.