Mixing due to internal gravity waves can explain the CNO surface abundances of B-type detached eclipsing binaries and single stars

TL;DR

This paper investigates whether internal gravity wave (IGW)–induced envelope mixing can explain the observed surface nitrogen trends in single and binary B-type stars. Using MESA, the authors simulate main-sequence tracks with IGW-driven mixing near the convective core boundary and compare the predicted surface nitrogen to measurements from detached eclipsing binaries and single B stars, constraining the required mixing strength to $log(D_{ m env}/\mathrm{cm^{2}\,s^{-1}})\approx 5$--$6$. They find that binaries, often fast-rotating due to tidal effects, show no nitrogen enrichment, while slowly rotating or evolved single B stars can be nitrogen-enhanced, consistent with rotation-modulated IGW efficiency. The results suggest a nuanced interplay between IGWs and rotational mixing, motivating targeted asteroseismic studies to further test wave-driven mixing as the origin of the observed abundance patterns and to refine models of massive-star interiors.

Abstract

Observations of double-lined spectroscopic eclipsing binaries are ideal to study stellar evolution. They have tight model-independent constraints on their masses and radii. With the addition of spectroscopically determined effective temperatures and surface abundances, they can be used to calibrate and improve models. Here we determine whether the observed trends of surface nitrogen abundance in single and binary stars can be explained by wave-induced mixing occurring in the stellar envelope. We use MESA to run the simulations. We compare the outcome of the models to observations of the surface nitrogen abundance for samples of detached eclipsing binary systems and of single B-type stars. From this we determine the amount of wave-induced mixing required to bring the model predictions in agreement with the observations. We find nitrogen to be enriched at the surface of theoretical models with wave-induced mixing provided that we use levels above log(Denv)=5-6 at the convective core boundary. A prominent observation is that the B-type components of detached eclipsing binaries do not show any nitrogen surface enhancement, which can be explained by their relatively fast rotation enforced by the tidal forces in the systems. The slowly rotating or evolved stars among the sample of single B stars do reveal a nitrogen enhancement. Our findings on the difference between single B stars and B-type components of detached binary systems can potentially be explained by internal wave-induced mixing profiles based on recent 2-dimensional hydrodynamical simulations of rotating B stars. Such wave-induced mixing decreases with increasing rotation and may act in combination with additional rotational mixing. Our findings motivate future asteroseismic studies in samples of single B stars and pulsating eclipsing binaries with B-type components as optimal laboratories to further test our interpretations.

Figures (8)

• Figure 1: Shape of the overall internal mixing profile in our stellar models. The green zone on the left is the fully mixed convective core, the dark blue zone indicates the convective overshoot zone, and the light blue area is the radiative envelope where mixing takes place through IGWs. The size of the overshoot zone is determined by $\alpha_{ov}$, and the starting height of the envelope mixing is determined by log(D$_{\rm env}$/cm$^{2}$s$^{-1}$). This figure is inspired by Figure 3 of Pedersen2021.
• Figure 2: Kiel diagram for the components of the binary systems from Andrew2020 along with the stellar evolution tracks of models with masses from 5-20 M$_\odot$. For 5 and 20 M$_\odot$ the tracks with the maximum amount of IGW envelope mixing, log(D$_{\rm env}$/cm$^{2}$s$^{-1}$)=5 and 6, are shown with the dashed-dotted and dashed lines, respectively. None of the models in this figure include the effects of rotational mixing. The binaries discussed in Section \ref{['discrepancy']} are indicated by the black symbols (U Oph), red symbols (V1034 Sco), and blue symbols (V380 Cyg). The stellar parameters of the binaries are reported in Table \ref{['BinaryData']}.
• Figure 3: Theoretical boundaries of N/C and N/O based on changes in the CNO-ratio following the description given by Martins2015CNO for the elemental mixtures adopted by Brott2011 (grey dashed line), Nieva2012 (black dashed-dotted line), and Asplund2009 (light grey dotted line). The position of the binary star from Andrew2020 are shown with their respective error bars.
• Figure 4: Tracks in the Kiel-diagram of 10 M$_\odot$ (right) and 17 M$_\odot$ (left) models, with the surface nitrogen compared to the initial surface abundance on the colour-scale, mind that a lighter colour indicates a larger amount of nitrogen on the surface. The top panel shows models with different levels of initial rotational velocity (blue scale), where left models of either set have an an initial rotational velocity of 15% of the critical rotational velocity and the models more to the right have initial rotational velocity 40% of the critical rotational velocity. These models do not include mixing based on IGWs. The bottom panel shows the models with different levels of envelope mixing based on IGWs and do not include the effects of rotational mixing. The tracks that branch off to lower log(g) are those with log(D$_{\rm env}$/cm$^{2}$s$^{-1}$)=5 for the 10 M$_\odot$ model and log(D$_{\rm env}$/cm$^{2}$s$^{-1}$)=6 for the 17 M$_\odot$ model. The other two tracks have log(D$_{\rm env}$/cm$^{2}$s$^{-1}$)=0.
• Figure 5: Observed values of N/C and N/O for single B-type stars from the literature (see text for references). The theoretical boundaries drawn by the lines are the same as in Fig. \ref{['TheoreticallySpeaking']}. The black stars indicate binary systems from Andrew2020, Mayer2013, Andrew2014B, and Andrew2016. The point in the lower right corner shows representative errors for all stars.