I discuss three-dimensional SPH simulations of neutron star black hole ( BH ) encounters .
The calculations are performed using a nuclear equation of state and a multi-flavor neutrino treatment , general relativistic effects are mimicked using the Paczynski-Wiita pseudo-potential and gravitational radiation reaction forces . Most of the explored mass range ( 14 to 20 M _ { \odot } ) has not been considered before in numerical simulations .
The neutron star is always disrupted during the first approach after most of its mass has been transferred directly into the hole .
In none of the analyzed cases episodic mass transfer is found .
For the lower end of the mass range ( M _ { BH } \leq 16 M _ { \odot } ) an accretion disk of moderate density ( \rho \sim 10 ^ { 10 } g cm ^ { -3 } ) and temperature ( T < 2.5 MeV ) forms around the hole ; the rest of the material forms a rapidly expanding tidal tail , up to 0.2 M _ { \odot } of which are unbound .
For higher mass black holes ( M _ { BH } \geq 18 M _ { \odot } ) almost the complete neutron star disappears in the hole without forming any accretion disk .
In these cases a small fraction of the star ( between 0.01 and 0.08 M _ { \odot } ) is spun up by gravitational torques and dynamically ejected . None of the investigated systems of this study yields conditions that are promising to launch a GRB .
While we can not completely exclude that a subset of neutron star black hole binaries , maybe black holes with low masses and very large initial spins , can produce a GRB , this seems to happen -if at all- only in a restricted region of the available parameter space .
I argue that the difficulty to form promising disks together with the absence of any observed neutron star black hole binary may mean that they are insignificant as central engines of the observed , short-hard GRBs and that the vast majority of the latter ones is caused by double neutron star coalescences .