The SAPP pipeline for the determination of stellar abundances and atmospheric parameters of stars in the core program of the PLATO mission

Abstract

We introduce the SAPP (Stellar Abundances and atmospheric Parameters Pipeline), the prototype of the code that will be used to determine parameters of stars observed within the core program of the PLATO space mission. The pipeline is based on the Bayesian inference and provides effective temperature, surface gravity, metallicity, chemical abundances, and luminosity. The code in its more general version has a much wider range of potential applications. It can also provide masses, ages, and radii of stars and can be used with stellar types not targeted by the PLATO core program, such as red giants. We validate the code on a set of 27 benchmark stars that includes 19 FGK-type dwarfs, 6 GK-type subgiants, and 2 red giants. Our results suggest that combining various observables is the optimal approach, as this allows the degeneracies between different parameters to be broken and yields more accurate values of stellar parameters and more realistic uncertainties. For the PLATO core sample, we obtain a typical uncertainty of 27(syst.) ± 37 (stat.) K for T eff , 0.00 ± 0.01 dex for log g, 0.02 ± 0.02 dex for metallicity [Fe/H], −0.01 ± 0.03 R for radii, −0.01 ± 0.05 M for stellar masses, and −0.14 ± 0.63 Gyr for ages. We also show that the best results are obtained by combining the ν max scaling relation with stellar spectra. This resolves the notorious problem of degeneracies, which is particularly important for F-type stars.

Non-LTE radiative transfer with Turbospectrum

“…We adopt the reference stellar parameters for these stars from Jofré et al (2018), with a few updates as described in Gent et al (2022). The T eff estimates for the majority of the stars are based on angular diameters determined from interferometric data collected with several different facilities, such as Center for High Angular Resolution Astronomy (CHARA) or the Very Large Telescope Interferometer (VLTI), together with bolometric flux estimates.…”

“…, and O'Brian et al (1991) (15)Blackwell et al (1982a) (16) average ofBlackwell et al (1982b) and O'Brian et al (1991) (17) average ofBard et al (1991) and O'Brian et al (1991) (18) average of Blackwell et al (1982a) and O'Brian et al (1991) Article number, page 10 of 16 Jeffrey M. Gerber et al: Non-LTE radiative transfer with Turbospectrum a The stellar parameters come from Jofré et al (2018), with a few updates as described inGent et al (2022). We also include the literature [Fe/H] value fromJofré et al (2018) that was derived using 1D NLTE, as a reference.…”

Non-LTE Radiative Transfer with Turbospectrum

“…We adopt the reference stellar parameters for these stars from Jofré et al (2018), with a few updates as described in Gent et al (2022). The T eff estimates for the majority of the stars are based on angular diameters determined from interferometric data collected with several different facilities, such as CHARA or the VLTI, together with bolometric flux estimates.…”

“…a The stellar parameters come fromJofré et al (2018), with a few updates as described inGent et al (2022). We also include the literature [Fe/H] value fromJofré et al (2018) that was derived using 1D NLTE as a reference.…”

An Interpretable Machine-learning Framework for Modeling High-resolution Spectroscopic Data*

“…The overall accuracy hinges on the accuracy of the stellar benchmark labels. The mechanics of this approach could be adapted to fit within existing pipelines such as SAPP (Gent et al 2022).…”