Low-Luminosity Gamma-Ray Bursts as a Distinct GRB Population:A Firmer Case from Multiple Criteria Constraints

The intriguing observations of Swift/BAT X-ray flash XRF 060218 and the BATSE-BeppoSAX gamma-ray burst GRB 980425, both with much lower luminosity and redshift compared to other observed bursts, naturally lead to the question of how these low-luminosity (LL) bursts are related to high-luminosity (HL) bursts. Incorporating the constraints from both the flux-limited samples observed with CGRO/BATSE and Swift/BAT and the redshift-known GRB sample, we investigate the luminosity function for both LL- and HL-GRBs through simulations. Our multiple criteria, including the log N - log P distributions from the flux-limited GRB sample, the redshift and luminosity distributions of the redshift-known sample, and the detection ratio of HL- and LL- GRBs with Swift/BAT, provide a set of stringent constraints to the luminosity function. Assuming that the GRB rate follows the star formation rate, our simulations show that a simple power law or a broken power law model of luminosity function fail to reproduce the observations, and a new component is required. This component can be modeled with a broken power, which is characterized by a sharp increase of the burst number at around L < 10^47 erg s^-1}. The lack of detection of moderate-luminosity GRBs at redshift ~0.3 indicates that this feature is not due to observational biases. The inferred local rate, rho_0, of LL-GRBs from our model is ~ 200 Gpc^-3 yr^-1 at ~ 10^47 erg s^-1, much larger than that of HL-GRBs. These results imply that LL-GRBs could be a separate GRB population from HL-GRBs. The recent discovery of a local X-ray transient 080109/SN 2008D would strengthen our conclusion, if the observed non-thermal emission has a similar origin as the prompt emission of most GRBs and XRFs.