Formation of Double Neutron Star systems as implied by observations

Double Neutron Stars (DNS) have to survive two supernovae and still remain bound. This sets strong limits on the nature of the second collapse in these systems. We consider the masses and orbital parameters of the DNS population and constrain the two distributions of mass ejection and kick velocities directly from observations with no a-priori assumptions regarding evolutionary models and/or the types of the supernovae involved. We show that there is strong evidence for two distinct types of supernovae in these systems, where the second collapse in the majority of the observed systems involved small mass ejection ($\Delta M\lesssim 0.5M_{\odot}$) and a corresponding low-kick velocity ($v_{k}\lesssim 30 $km\,s$^{-1}$). This formation scenario is compatible, for example, with an electron capture supernova. Only a minority of the systems have formed via the standard SN scenario involving larger mass ejection of $\sim 2.2 M_{\odot}$ and kick velocities of up to $400$km\,s$^{-1}$. Due to the typically small kicks in most DNS (which are reflected by rather low proper motion), we predict that most of these systems reside close to the galactic disc. In particular, this implies that more NS-NS mergers occur close to the galactic plane. This may have non-trivial implications to the estimated merger rates of DNS and to the rate of LIGO / VIRGO detections.