2018A&A...620A..54S

Aims. In an effort to improve spectroscopic methods of stellar parameters determination, we implemented non-solar-scaled opacities in a simultaneous derivation of fundamental parameters and abundances. We wanted to compare the results with the usual solar-scaled method using a sample of solar-type and evolved stars.
Methods. We carried out a high-precision determination of stellar parameters and abundances by applying non-solar-scaled opacities and model atmospheres. Our sample is composed of 20 stars, including main sequence and evolved objects. The stellar parameters were determined by imposing ionization and excitation equilibrium of Fe lines, with an updated version of the FUNDPAR program, together with plane-parallel ATLAS12 model atmospheres and the MOOG code. Opacities for an arbitrary composition and v_micro were calculated through the opacity sampling (OS) method. We used solar-scaled models in the first step, and then continued the process, but scaled to the abundance values found in the previous step (i.e. non-solar-scaled). The process finishes when the stellar parameters of one step are the same as in the previous step, i.e. we use a doubly iterated method.
Results. We obtained a small difference in stellar parameters derived with non-solar-scaled opacities compared to classical solar-scaled models. The differences in T_eff, logg, and [Fe/H] amount to 26K, 0.05dex, and 0.020dex for the stars in our sample. These differences can be considered the first estimation of the error due to the use of classical solar-scaled opacities to derive stellar parameters with solar-type and evolved stars. We note that some chemical species could also show an individual variation greater than those of the [Fe/H] (up to ∼0.03dex) and varying from one species to another, obtaining a chemical pattern difference between the two methods. This means that condensation temperature T_c trends could also present a variation. We include an example showing that using non-solar-scaled opacities, the solution found with the classical solar-scaled method indeed cannot always verify the excitation and ionization balance conditions required for a model atmosphere. We discuss in the text the significance of the differences obtained when using solar-scaled versus non-solar-scaled methods.
Conclusions. We consider that the use of the non-solar-scaled opacities is not mandatory in every statistical study with large samples of stars. However, for those high-precision works whose results depend on the mutual comparison of different chemical species (such as the analysis of condensation temperature T_c trends), we consider its application to be worthwhile. To date, this is probably one of the most precise spectroscopic methods for stellar parameter derivation.