Quark deconfinement and the duration of short gamma-ray bursts

We propose a model for short duration gamma-ray bursts (sGRBs) based on the formation of a quark star after the merger of two neutron stars. We assume that the sGRB central engine is a proto-magnetar, which has been previously invoked to explain the plateaulike X-ray emission observed following both long and short GRBs. Here, we show that: (i) a few milliseconds after the merger it is possible to form a stable and massive star made in part of quarks; (ii) during the early cooling phase of the incompletely formed quark star, the flux of baryons ablated from the surface by neutrinos is large and it does not allow the outflow to achieve a bulk Lorentz factor high enough to produce a GRB; (iii) after the quark burning front reaches the stellar surface, baryon ablation ceases and the jet becomes too baryon poor to produce a GRB; (iv) however, between these two phases a GRB can be produced over the finite time scale required for the baryon pollution to cease; a characteristic time scale of the order of 0.1 s naturally results from the time the conversion front needs to cover the distance between the rotational pole and the latitude of the last closed magnetic field line; (v) we predict a correlation between the luminosity of the sGRB and its duration, consistent with the data; (vi) our model also predicts a delay of the order of ten seconds between the time of the merger event and the sGRB, allowing for the possibility of precursor emission and implying that the jet will encounter the dense cocoon formed immediately after the merger.