Binary neutron star merger simulations with different initial orbital frequency and equation of state

We present results from three-dimensional general relativistic simulations of binary neutron star coalescences and mergers using public codes. We considered equal mass models where the baryon mass of the two Neutron Stars (NS) is $1.4M_{\odot}$, described by four different equations of state (EOS) for the cold nuclear matter (APR4, SLy, H4, and MS1; all parametrized as piecewise polytropes). We started the simulations from four different initial interbinary distances ($40, 44.3, 50$, and $60$ km), including up to the last 16 orbits before merger. That allows to show the effects on the gravitational wave phase evolution, radiated energy and angular momentum due to: the use of different EOSs, the orbital eccentricity present in the initial data and the initial separation (in the simulation) between the two stars. Our results show that eccentricity has a major role in the discrepancy between numerical and analytical waveforms until the very last few orbits, where "tidal" effects and missing high-order post-Newtonian coefficients also play a significant role. We test different methods for extrapolating the gravitational wave signal extracted at finite radii to null infinity. We show that an effective procedure for integrating the Newman-Penrose $\psi_4$ signal to obtain the gravitational wave strain $h$ is to apply a simple high-pass digital filter to $h$ after a time domain integration, where only the two physical motivated integration constants are introduced. That should be preferred to the more common procedures of introducing additional integration constants, integrating in the frequency domain or filtering $\psi_4$ before integration.