Grid Search in Stellar Parameters: a software for spectrum analysis of single stars and binary systems

TL;DR

This paper introduces GSSP, a high-performance, parallel software tool for analyzing medium- to high-resolution stellar spectra to accurately determine atmospheric parameters and chemical compositions of single and binary stars.

Contribution

The paper presents GSSP, a novel, efficient, and parallelized software for stellar spectrum analysis, capable of handling large data sets from space missions.

Findings

GSSP successfully recovers known stellar parameters from simulated data.

The software's accuracy depends on the signal-to-noise ratio of spectra.

Application to real Kepler stars shows good agreement with previous studies.

Abstract

The currently operating space missions, as well as those that will be launched in the near future, (will) deliver high-quality data for millions of stellar objects. Since the majority of stellar astrophysical applications still (at least partly) rely on spectroscopic data, an efficient tool for the analysis of medium- to high-resolution spectroscopy is needed. We aim at developing an efficient software package for the analysis of medium- to high-resolution spectroscopy of single stars and those in binary systems. The major requirements are that the code has a high performance, represents the state-of-the-art analysis tool, and provides accurate determinations of atmospheric parameters and chemical compositions for different types of stars. We use the method of atmosphere models and spectrum synthesis, which is one of the most commonly used approaches for the analysis of stellar spectra. Our Grid Search in Stellar Parameters (GSSP) code makes use of the OpenMPI implementation, which makes it possible to run in parallel mode. The method is first tested on the simulated data and is then applied to the spectra of real stellar objects. The majority of test runs on the simulated data were successful in the sense that we could recover the initially assumed sets of atmospheric parameters. We experimentally find the limits in signal-to-noise ratios of the input spectra, below which the final set of parameters gets significantly affected by the noise. Application of the GSSP package to the spectra of three Kepler stars, KIC11285625, KIC6352430, and KIC4931738, was also largely successful. We found an overall agreement of the final sets of the fundamental parameters with the original studies. For KIC6352430, we found that dependence of the light dilution factor on wavelength cannot be ignored, as it has significant impact on the determination of the atmospheric parameters of this binary system.