Titans metal-poor reference stars. Ba abund.

We assessed the performance of one-dimensional under local thermodynamic equilibrium (1D LTE) synthesis, 1D non-LTE synthesis, and abundance corrections based on three-dimensional model atmospheres under non-LTE (3D non-LTE) to model Ba lines. Alongside barium abundances and its isotopic ratios, we determined Eu and other neutron-capture element abundances to validate the method in the TITANSmetal-poor benchmark stars.

The observational results are compared with the predictions of stochastic Galactic chemical evolution models that account for the inhomogeneous mixing in the early times. We find that 1D LTE and 3D non-LTE Ba abundance determinations are equivalent, whereas the 1D non-LTE approach leads to systematic underestimations. These underestimations bias isotopic fractions towards higher r-process contributions.

The inferred s- and r-process fractions demonstrate that [Ba/Eu] alone is an ambiguous tracer for ordinary stars within the -0.8<=[Ba/Eu]<=0dex range. However, we identify an observational pattern in the [Ba/Eu]-[Ba/Fe] plane that can distinguish, at a given [Ba/Fe], the dominant neutron capture process. The comparison of our set of models, both for the proto-Milky Way halo and for the Gaia-Enceladus galaxy is used to put constraints on the production of Ba at low metallicity, especially evaluating the role of rotating massive stars.

The method developed here can be applied with confidence to both ordinary stars and peculiar stars enhanced in barium.