On the Numerical Treatment and Dependence of Thermohaline Mixing in Red Giants

TL;DR

This paper investigates the numerical challenges in modeling thermohaline mixing in red giants, revealing that lithium abundance predictions are highly sensitive to model resolution and numerical schemes, which questions previous results.

Contribution

It demonstrates the extreme sensitivity of lithium predictions to numerical details in stellar models and provides guidelines for achieving converged solutions in thermohaline mixing simulations.

Findings

Lithium destruction varies by orders of magnitude across models.

Numerical resolution critically affects the predicted mixing outcomes.

Previous literature results on lithium in red giants may be unreliable.

Abstract

In recent years much interest has been shown in the process of thermohaline mixing in red giants. In low and intermediate mass stars this mechanism first activates at the position of the bump in the luminosity function, and has been identified as a likely candidate for driving the slow mixing inferred to occur in these stars. One particularly important consequence of this process, which is driven by a molecular weight inversion, is the destruction of lithium. We show that the degree of lithium destruction, or in some cases production, is extremely sensitive to the numerical details of the stellar models. Within the standard 1D diffusion approximation to thermohaline mixing, we find that dfferent evolution codes, with their default numerical schemes, can produce lithium abundances that differ from one another by many orders of magnitude. This disagreement is worse for faster mixing. We perform experiments with four independent stellar evolution codes, and derive conditions for the spatial and temporal resolution required for a converged numerical solution. The results are extremely sensitive to the timesteps used. We find that predicted lithium abundances published in the literature until now should be treated with caution.