The Chemical Homogeneity of Single-Lined Spectroscopic Binaries in Open Clusters

Abstract

Using SDSS-V DR19 Milky Way Mapper APOGEE data, we measure the impact that close binarity has on surface chemistry across the Hertzsprung-Russell diagram in a broad set of abundances by studying single-lined spectroscopic binaries (SB1s) in open clusters. We derive binary membership and orbital parameters for 103 SB1s by analysing APOGEE radial velocities with The Joker and UltraNest. We perform a detailed abundance analysis with BACCHUS to derive abundances in fourteen chemical species: Si, Fe, C, N, O, Na, Mg, Al, Ca, Ti, Cr, Ni, Ce, and Nd. Leveraging the assumptions of chemical homogeneity in open clusters, we compare the surface abundances of SB1s to non-binary stars at similar evolutionary states. We find that a subset of binaries with significant UV excess have a $Δ$[C/N] that is 0.2--0.5 dex higher than expected, resulting in overestimated [C/N]-based ages for those stars. This points to pollution from an evolved companion and has implications for [C/N]-based age studies of the broader Milky Way. At the population level, we find that SB1s in our sample can be treated as statistically chemically homogeneous with their single-star counterparts, and we find no connection between orbital separation and chemical enrichment or depletion. We show that at separations up to ~5 pc, co-eval stars can be considered chemically homogeneous with one another within current abundance precisions, regardless of multiplicity.

Figures (10)

• Figure 1: Top: The orbital fit for SDSS 66658896, a short-period binary, as determined by UltraNest. We show the best fit orbital solution compared to the observed data points in the top left in MJD and orbital phase across the top row. In the bottom row, we show the UltraNest final live-point sampling for the orbital period, histograms of the sampling counts for period and eccentricity, and the residual between the observed data and model across orbital phase. Bottom: The orbital fit for SDSS 66661885, a long-period binary, as determined by UltraNest. The ordering of information in panels is the same as in the top plot.
• Figure 2: Top: A comparison between the orbital solutions for unimodal stars analysed by the Joker, UltraNest, and WOCS. It is clear that while both codes exhibit reasonable agreement in period, there is significantly less agreement in eccentricity. This is likely due to the different priors used in the two codes. It is also clear that the Joker/UltraNest parameters broadly agree with those derived independently by WOCS. Bottom: The period-eccentricity distribution of all binaries in our sample with confident orbital solutions, as solved for by UltraNest.
• Figure 3: Top Left: A comparison of the interpolated MIST isochrone for Messier 67 (NGC 2682) and the observed stars in $(BP-RP)$ and $G$. Top Right: Here we show the effective temperatures and surface gravities used, as compared to the published values, for M67. Bottom: We show the three Si lines used in this study to derive broadening: 16680.8Å, 16828.2Å, and 16094.8Å, with the observed spectrum of a typical star in black. We show a synthetic spectrum generated using an interpolated MARCS stellar atmosphere at that star's derived log(g), $T_{\rm eff}$, [Fe/H], and microturbulence in red. While there are slight differences between the observed and synthetic spectrum on the left wing in the rightmost plot, this does not affect the synthetic fit to the line core---and therefore the derived convolution.
• Figure 4: Here we compare the derived values of $v \sin{i}$ to the published Astra values. We find strong agreement between the two samples at values below $\sim$30 km s$^{-1}$ ($\sim$85% of the sample).
• Figure 5: Here we show comparisons between the observed spectrum (black) and synthetic spectrum (salmon) for one of the stars in our sample, with our chosen absorption features shown as dashed lines (§\ref{['sec:line_selection']}). We also show cutouts of two regions that contain C and Ce features to demonstrate the quality of our spectral synthesis. The shaded band shows a typical uncertainty in synthetic fit when considering the uncertainty in all the atmospheric parameters.