Neutron capture elements in the early Universe

Sammanfattning: Metal-poor (MP) stars in the Milky Way (the Galaxy) and its satellite galaxies open a window into the earliest times in the history of the Universe, probing the chemistry of the earliest times. Recent galaxy formation simulations predict that the oldest MP stars are those on tightly bound orbits in the inner regions of a galaxy (Tumlinson 2010). Applying this to the Milky Way, MP stars in the bulge would have a higher probability to have formed at higher redshifts compared to MP halo stars. In this work we have measured the abundances of Sr, Y, Ba and Eu in 48 MP stars from the inner regions of the Galaxy. This is the first time neutron capture elements have been studied for a significant sample of stars in the bulge. Our results show that the bulge sample generally behaves similarly to the halo stars, with some differences. We find an increasing scatter in the abundances of Sr and Ba with decreasing metallicity as in halo stars, this is not the case for Y and Eu where our stars show a constant scatter scale. As reported in François et al. (2007) for their sample, we find an anticorrelation between [Y/Sr] and [Sr/H], this is not seen in other halo samples, more stars at the lowest metallicities are needed to confirm this behavior. Our stars are on average r-enhanced compared to the halo stars with ∼ 50% of the stars having [Eu/Fe]> 0.6, favoring r-process production sites with high yields such as neutron star mergers and magnetorotationally driven supernovae. We find one star with exceptionally high [Sr/Ba]=2.04 and [Y/Ba]= 1.72 that is slightly carbon enhanced with [C/Fe]=0.65. This star could be in a binary system, only a handful of stars from the literature show similar values. Similar to halo MP stars, we find an anticorrelation between the abundances of the light s-process (ls) elements Sr and Y and the heavy s-process (hs) element Ba. This behavior is not explained using the yields from the main r-process only. We compare our results to Galactic chemical evolution models (one for the halo and another for an old bulge population) that include massive fast-rotating stars (spinstars) as a production site for ls elements at the earliest times along with the main r-process from magnetorotationally driven supernovae. The results from this comparison show that the halo model better matches our data than the bulge model. However, more work on the spinstar models is needed to better predict the trends from our stars. The source of ls elements in the early Universe remains unknown for the time being.