Studies of intrinsic properties of gamma ray bursts detected by the HETE-II satellite

Abstract

Analysis of HETE-II data is discussed with the aim of understanding the intrinsic properties of gamma-ray bursts (GRBs). A technique is developed that allows the simultaneous estimation of source and background counts during a burst with coded aperture instruments such as the Wide-field X-ray Monitor (WXM) on HETE-II. A closely related photon-by-photon statistical bootstrap analysis is then described that can be used to compute the non-Gaussian error distribution of GRB temporal statistics. This is applied to the T90 and To.45 duration measures. The distribution of T90 has been extensively studied since the availability of the BATSE gamma-ray burst catalogs and is widely believed to be log-normally distributed. It is shown that intrinsically, GRBs may in fact have a much narrower distribution, and the wide log-normal may primarily be due to measurement artifacts. Computation of the TO.45 parameter enables the inference of redshifts through the recently discovered Liso-Epk-TO.45 relation. This in turn allows the compilation of a flux-limited sample of bursts with redshifts that is free of the observational selection effects inherent in spectroscopic catalogs. This analysis is performed for 31 WXM bursts and redshift-corrected distributions of T90 and T0.45 are computed.

(cont.) It is shown for the first time that the distribution of T0.45 can be modeled by an exponential distribution. The redshifts calculated through the Liso-Epk-TO.45 relation are also used to calculate the implied isotropic luminosities. The normalized luminosity function and redshift distribution of gamma-ray bursts are derived using the non-parametric methods of Lynden-Bell and Efron & Petrosian. The results imply strong evidence for luminosity evolution with redshift and are consistent with prior studies based on BATSE bursts. Concordance cosmology (QA = 0.7, QM = 0.3, Ho=70 km s-1 Mpc-1) is assumed throughout the analysis. Effects of log-normal errors in the redshifts are estimated using Monte-Carlo methods. Results indicate that a fraction close to 10% of GRBs are to be expected at high redshifts (> 5) in consonance with theoretical predictions of high-redshift Swift detections.