Stellar evolution models with entropy-calibrated mixing-length parameter: application to red giants

TL;DR

This paper introduces a new method for calibrating the mixing-length parameter in stellar evolution models using entropy from 2D and 3D simulations, improving the accuracy of red giant predictions.

Contribution

It integrates entropy-based calibration of the mixing-length parameter into stellar evolution codes, replacing traditional solar-calibrated approaches for better modeling of evolved stars.

Findings

Red giant branch positions align better with observations.

Models show improved radius and temperature estimates.

Calibration method is applicable across various stellar types.

Abstract

We present evolutionary models for solar-like stars with an improved treatment of convection that results in a more accurate estimate of the radius and effective temperature. This is achieved by improving the calibration of the mixing-length parameter, which sets the length scale in the 1D convection model implemented in the stellar evolution code. Our calibration relies on the results of 2D and 3D radiation hydrodynamics simulations of convection to specify the value of the adiabatic specific entropy at the bottom of the convective envelope in stars as a function of their effective temperature, surface gravity and metallicity. For the first time, this calibration is fully integrated within the flow of a stellar evolution code, with the mixing-length parameter being continuously updated at run-time. This approach replaces the more common, but questionable, procedure of calibrating the length scale parameter on the Sun, and then applying the solar-calibrated value in modeling other stars, regardless of their mass, composition and evolutionary status. The internal consistency of our current implementation makes it suitable for application to evolved stars, in particular to red giants. We show that the entropy calibrated models yield a revised position of the red giant branch that is in better agreement with observational constraints than that of standard models.