Merging black hole binaries with the SEVN code

Abstract: Studying the formation and evolution of black hole binaries (BHBs) is essential for the interpretation of current and forthcoming gravitational wave (GW) detections. We investigate the statistics of BHBs that form from isolated binaries, by means of a new version of the SEVN population-synthesis code. SEVN integrates stellar evolution by interpolation over a grid of stellar evolution tracks. We upgraded SEVN to include binary stellar evolution processes and we used it to evolve a sample of $1.5\times{}10^8$ binary systems, with metallicity in the range $\left[10^{-4};4\times 10^{-2}\right]$. From our simulations, we find that the mass distribution of black holes (BHs) in double compact-object binaries is remarkably similar to the one obtained considering only single stellar evolution. The maximum BH mass we obtain is $\sim 30$, $45$ and $55\, \mathrm{M}\odot$ at metallicity $Z=2\times 10^{-2}$, $6\times 10^{-3}$, and $10^{-4}$, respectively. A few massive single BHs may also form ($\lesssim 0.1\%$ of the total number of BHs), with mass up to $\sim 65$, $90$ and $145\, \mathrm{M}\odot$ at $Z=2\times 10^{-2}$, $6\times 10^{-3}$, and $10^{-4}$, respectively. These BHs fall in the mass gap predicted from pair-instability supernovae. We also show that the most massive BHBs are unlikely to merge within a Hubble time. In our simulations, merging BHs like GW151226 and GW170608, form at all metallicities, the high-mass systems (like GW150914, GW170814 and GW170104) originate from metal poor ($Z\lesssim{}6\times 10^{-3}$) progenitors, whereas GW170729-like systems are hard to form, even at $Z = 10^{-4}$. The BHB merger rate in the local Universe obtained from our simulations is $\sim 90 \mathrm{Gpc}^{-3}\mathrm{yr}^{-1}$, consistent with the rate inferred from LIGO-Virgo data.