Classification of Swift's gamma-ray bursts

Context. Two classes of gamma-ray bursts have been identified in the BATSE catalogs characterized by durations shorter and longer than about 2 seconds. There are, however, some indications for the existence of a third class. Swift satellite detectors have different spectral sensitivity than pre-Swift ones for gamma-ray bursts. Therefore we reanalyze the durations and their distribution and also the classification of GRBs. Aims. We analyze the bursts duration distribution, published in The First BAT Catalog, whether it contains two, three or more groups. Methods. Using The First BAT Catalog the maximum likelihood estimation was used to analyze the duration distribution of GRBs. Results. The three log-normal fit is significantly (99.54% probability) better than the two for the duration distribution. Monte-Carlo simulations also confirm this probability (99.2%). Similarly, in previous results we found that the fourth component is not needed. The relative frequencies of the distribution of the groups are 7% short 35% intermediate and 58% long. Conclusions. Similarly to the BATSE data, three components are needed to explain the BAT GRBs’ duration distribution. Although the relative frequencies of the groups are different than in the BATSE GRB sample, the difference in the instrument spectral sensitivities can explain this bias. This means theoretical models may be needed to explain three different type of gamma-ray bursts.