F. Merging neutron stars
Neutron star binary mergers [86, 276] or neutron star - black hole binary mergers  (hereafter called mergers) also produce a black hole - accretion disk system and are candidates for the inner engines of GRBs, specifically of short GRBs. These mergers take place because of the decay of the binary orbits due to gravitational radiation emission as was beautifully demonstrated in the famous binary pulsar PSR 1913+16 .
These mergers take place at a rate of 10-6 events per year per galaxy [278, 300, 414]. This is the rate of merger of binaries of the type of PSR 1913+16 whose life time is or order of several 108 years. Various population synthesis calculations suggest that there is also another population of short lived binaries [18, 297, 410, 411]. These binaries form with very close orbits and hence with short lifetimes of the order of 105 yrs. Even thought the overall rate of such mergers could be comparable to those of the PSR 1913+16 type one cannot expect to catch a binary in our galaxy in such a stage. Similarly unlike the long lived mergers that may be kicked from their host galaxy within their long life time [44, 276] this short lived population remains within the galaxy when they merge .
Earlier simulations of mergers focused on the gravitational radiation from this system. Davies et al.  begun a series of numerical simulation of neutron star merger that focused on GRB related aspects [11, 349, 350, 351]. Using a SPH scheme they followed NS mergers under different assumptions (Newtonian with ad hoc addition of gravitational radiation back reaction or Post Newtonian), with different equations of state (adiabatic or realistic) and with different initial spin axis and mass rations and different estimates of the effects of neutrino cooling. A parallel set of simulations was carried out by Janka and Ruffert , Ruffert and Janka [354, 355, 356], Ruffert et al.  who used particle in cell methods. Both kinds of simulations yield comparable results. The merger results in a black hole - accretion disk system. The mass of the accretion disk is of order 0.1M and it depends, of course somewhat on the orientation of the spins and the relative masses of the two neutron stars.
A merger releases ~ 5 × 1053 ergs but most of this energy is in the form of low energy neutrinos and gravitational waves. Still there is enough energy available to power a GRB but is not clear how the GRB is produced. A central question is, of course, how does a merger generate the relativistic wind required to power a GRB. Eichler et al.  suggested that about one thousandth of these neutrinos annihilate and produce pairs that in turn produce gamma-rays via e+ e- . This idea was criticized on several grounds by different authors the main problem is that it does not produce enough energy. For example Jaroszynksi  pointed out that a large fraction of the neutrinos will be swallowed by the black hole that forms. An alternative source of energy within the merger model is the accretion power of a disk that forms around the black hole. This brings us back to the canonical black hole - accretion disk scenario.