The gamma-ray burst ( GRB ) 060218 has \sim 10 ^ { 5 } times lower luminosity than typical long GRBs , and is associated with a supernova ( SN ) .
The radio afterglow displays no jet break , so that this burst might arise from a mildly-relativistic spherical outflow produced by the SN shock sweeping the stellar surface .
Since this model is energetically difficult , we propose that the radio afterglow is produced by a non-relativistic phase of an initially collimated outflow ( jet ) .
Our jet model is supported by the detection of optical linear polarization in the SN component .
We also show analytically that the jet can penetrate a progenitor star .
We analyzed the observational data of the prompt emission of this burst and obtained a smooth power-law light curve which might last longer than 10 ^ { 6 } s. This behavior contrasts with the long intermittent activities with the X-ray flares of typical GRBs , implying that the central engine of this burst is different from those of typical GRBs .
This argument is consistent with the analysis of the SN component of this burst , which suggests that the progenitor star was less massive and collapsed to a neutron star instead of a black hole .
The collimation-corrected event rate of such low-luminosity GRBs is estimated to be \sim 10 times higher than that of typical long GRBs , and they might form a different GRB population : low-luminosity GRBs are produced by mildly-relativistic jets from neutron stars at the collapses of massive stars , while typical long GRBs by highly-relativistic jets from black holes .
We suggest that the central engine of GRB 060218 is a pulsar ( or a magnetar ) with the initial rotation period P _ { 0 } \sim 10 ms and the magnetic field B \sim 10 ^ { 16 } G. A giant flare from the magnetar might be observed in future .