Author: Houjun Lu
Abstract: Traditionally gamma-ray bursts (GRBs) are classified in the t90-hardness ratio two-dimensional plane into two classes: long/soft vs. short/hard GRBs. Afterglow and host galaxy observations suggest that these two categories may correspond to two different types of progenitors: those related to massive stars (Type II) and those related to compact stars (Type I). However, recent observations suggest that the duration classification scheme does not always match the physical classification scheme. Multiwavelength criteria are needed to diagnose the physical category of a particular GRB, but some important information (e.g. supernova signature, host galaxy information) are not available during the prompt phase. In this paper, we suggest to add the ``amplitude'' of GRB prompt emission as the third dimension into the consideration of GRB classification with the prompt γ-ray data. We define three new parameters with the prompt emission data. The first parameter is defined as f = Fp/FB, where Fp and FB are the peak flux and the background flux, respectively. The second parameter is defined by feff = F'p/FB, where F'p the peak flux of the a pseudo-GRB, which is generated by scaling the original lightcurve down until the measured t90 above the background is just less than 2 s. This is an "effective amplitude" of a real long GRB that is observed as a "short" GRB with its "tip-of-iceberg" above the background. The third parameter is defined as feff,z = F'p,z/FB, where F'p,z is the peak flux of another pseudo-GRB, which is generated by progressively moving a GRB to a higher redshift z, until its measured "rest frame" duration t90/(1+z) is just below 2 s. We systematically derive these parameters for the Swift GRBs detected between December 2004 to December 2011. We find that most short GRBs are likely not "tip-of-iceberg" of long GRBs. However, one needs to be cautious if a short GRB has f < 2 , since it could be either intrinsically short or intrinsically long with the extended emission episodes buried beneath the background. We also find that most long GRBs would show up as rest-frame short GRBs if they have a high enough redshift. This interprets the interesting observational fact that the three highest-z GRBs (080913, 090423, 090429B) all have a rest frame duration shorter than 2 s, but are all likely Type II GRBs.