2015A&A...579A..94G

Heavy-elements, i.e. those beyond the iron peak, mostly form via two neutron capture processes: the slow (s-) and the rapid (r-) process. Metal-poor stars should contain fewer isotopes that form via the s-process, according to currently accepted theory. It has been shown in several investigations that theory and observation do not agree well, raising questions on the validity of either the methodology or the theory. We analyse the metal-poor star HD140283, for which we have a high quality spectrum. We test whether a three-dimensional (3D) local thermodynamic equilibrium (LTE) stellar atmosphere and spectrum synthesis code permits a more reliable analysis of the iron abundance and barium isotope ratio than a one-dimensional (1D) LTE analysis. Using 3D hydrodynamical model atmospheres, we examine 91 iron lines of varying strength and formation depth. This provides us with the star's rotational speed. With this, we model the barium isotope ratio by exploiting the hyperfine structure of the singly ionised 4554Å resonance line, and study the impact of the uncertainties in the stellar parameters. The star's rotational speed was found to be 1.65±0.05km/s. Barium isotopes under the 3D paradigm show a dominant r-process signature as 77±6±17% (f_odd=0.38±0.02±0.06) of barium isotopes form via the r-process, where errors represent the assigned random and systematic errors, respectively. We find that 3D LTE fits reproduce iron line profiles better than those in 1D, but do not provide a unique abundance (within the uncertainties). However, we demonstrate that the isotopic ratio is robust against this shortcoming. Our barium isotope result agrees well with currently accepted theory regarding the formation of the heavy-elements during the early Galaxy. The improved fit to the asymmetric iron line profiles suggests that the current state of 3D LTE modelling provides excellent simulations of fluid flows. However, the abundances they provide are not yet self-consistent. This may improve with non-LTE considerations and higher resolution models.