Double neutron stars: merger rates revisited

We revisit double neutron star (DNS) formation in the classical binary evolution scenario in light of the recent LIGO/Virgo DNS detection (GW170817). The observationally estimated Galactic DNS merger rate of $R_{\rm MW}=21^{+28}_{-14}$ Myr$^{-1}$, based on 3 Galactic DNS systems, fully supports our standard input physics model with $R_{\rm MW} =24$ Myr$^{-1}$. This estimate for the Galaxy translates in a non-trivial way (due to cosmological evolution of progenitor stars in chemically evolving Universe) into a local ($z\approx0$) DNS merger rate density of $R_{\rm local}=48$ Gpc$^{-3}$yr$^{-1}$, which {\em is not} consistent with the current LIGO/Virgo DNS merger rate estimate ($1540^{+3200}_{-1220}$ Gpc$^{-3}$yr$^{-1}$). Within our study of the parameter space we find solutions that allow for DNS merger rates as high as $R_{\rm local} \approx 600^{+600}_{-300}$ Gpc$^{-3}$yr$^{-1}$ which are thus consistent with the LIGO/Virgo estimate. However, our corresponding BH-BH merger rates for the models with high DNS merger rates exceed the current LIGO/Virgo estimate of local BH-BH merger rate ($12$-$213$ Gpc$^{-3}$yr$^{-1}$). Apart from being particularly sensitive to the common envelope treatment, DNS merger rates are rather robust against variations of several of the key factors probed in our study (e.g. mass transfer, angular momentum loss, natal kicks). This might suggest that either common envelope development/survival works differently for DNS ($\sim$ 10-20 Msun stars) than for BH-BH ($\sim$ 40-100 Msun stars) progenitors, or high BH natal kicks are needed to meet observational constraints for both types of binaries. Note that our conclusion is based on a limited number of (21) evolutionary models and is valid only within this particular DNS and BH-BH isolated binary formation scenario.