The role of neutron star mergers in the chemical evolution of the Galactic halo

Aims. We explore the problem of the site of production of Eu. We use also the information present in the observed spread in the Eu abundances in the early Galaxy, not only its average trend. Moreover, we extend to other heavy elements (Ba, Sr, Rb, Zr) our investigations to provide additional constraints to our results. Methods. We adopt a stochastic chemical evolution model taking into account inhomogeneous mixing. The adopted yields of Eu from neutron star mergers (NSM) and from core-collapse supernovae (SNII) are those that are able to explain the average [Eu/Fe]-[Fe/H] trend observed for solar neighborhood stars, in the framework of a well-tested homogeneous model for the chemical evolution of the MilkyWay. Rb, Sr, Zr, and Ba are produced by both the s- and r-process. The s-process contribution by spinstars is the same as in our previous papers. Results. NSM that merge in less than 10 Myr or NSM combined with a source of r-process generated by massive stars can explain the spread of [Eu/Fe] in the Galactic halo. The combination of r-process production by NSM and s-process production by spinstars is able to reproduce the available observational data for Sr, Zr and Ba. We also show the first predictions for Rb in the Galactic halo. Conclusions. We confirm previous results that either NSM with very short time scale or both NSM and at least a fraction of SNII should have contributed to the synthesis of Eu in the Galaxy. The r-process production by NSM - complemented by an s-process production by spinstars - provide results compatible with our previous findings based on other r-process sites. We critically discuss the weak and strong points of both NSM and SNII scenarios for producing Eu and eventually suggest that the best solution is probably a mixed one in which both sources produce Eu. In fact, this scenario better reproduces the scatter observed in all the studied elements. [abridged]