Impact of the convective mixing-length parameter α on stellar metallicity

Context. Mixing-length theory is used to treat stellar convection. As a simulation in one-dimensional stellar atmospheres models, the mixing-length parameter α is calibrated from the Sun and then applied to other stars. However, there is no strong evidence to suggest that α should be the same for stars of different evolutionary stages.
Aims: We evaluate the impact of the α value on the metallicity of different types of stars and investigate the correlation between the metallicity discrepancy (Δ[Fe/H]) and stellar parameters (T_eff, log g).
Methods: We selected ten well-studied field stars and one open cluster of nine members for which high-resolution and high signal-to-noise spectra are available. The model atmospheres were calculated with the code MAFAGS-OS. We derived iron abundances from Fe I and Fe II lines both under local thermodynamic equilibrium and non-LTE conditions using a spectrum synthesis method. After deriving [Fe/H] for each line, we calculated Δ[Fe/H] with two different α values, fixed solar-calibrated α, and α obtained for each star individually. Finally, we investigated the correlation between Δ[Fe/H] caused by revised α with stellar parameters.
Results: For FGK dwarf stars, the Δ[Fe/H] caused by the α correction is less than 0.02 dex, while for turn-off and giant stars, the Δ[Fe/H] values are no more than 0.03 dex, which are lower than typical uncertainties in metallicity. For main-sequence stars, Δ[Fe/H] versus T_eff and Δ[Fe/H] versus log g are well fit by linear relations.