Neutron star mergers have been long considered as promising sites of heavy r -process nucleosynthesis .
We overview observational evidence supporting this scenario including : the total amount of r -process elements in the Galaxy , extreme metal poor stars , geological radioactive elemental abundances , dwarf galaxies , and short gamma-ray bursts ( sGRBs ) .
Recently , the advanced LIGO and Virgo observatories discovered a gravitational-wave signal of a neutron star merger , GW170817 , as well as accompanying multi-wavelength electromagnetic ( EM ) counterparts .
The ultra-violet , optical , and near infrared observations point to r -process elements that have been synthesized in the merger ejecta .
The rate and ejected mass inferred from GW170817 and the EM counterparts are consistent with other observations .
We find however that , within simple one zone chemical evolution models ( based on merger rates with reasonable delay time distributions as expected from evolutionary models , or from observations of sGRBs ) , it is difficult to reconcile the current observations of the europium abundance history of Galactic stars for [ Fe/H ] \gtrsim - 1 .
This implies that to account for the role of mergers in the Galactic chemical evolution , we need a Galactic model with multiple populations that have different spatial distributions and/or varying formation rates .