A Search for Gamma-ray Burst Subgroups in the SWIFT and RHESSI Databases

arXiv:0912.3945v1 [astro-ph.HE] 19 Dec 2009 A Search for Gamma-ray Burst Subgroups in the SWIFT and RHESSI Databases Jakub ˇRípa∗, David Huja∗, Attila Mészáros∗, René Hudec†, Wojtek Hajdas∗∗and Claudia Wigger∗∗,‡ ∗Astronomical Institute of the Charles University, V Holešoviˇckách 2, Prague, Czech Republic †Astronomical Institute, Academy of Sciences of the Czech Republic, Ondˇrejov, Czech Republic ∗∗Paul Scherrer Institut, Villigen, Switzerland ‡Kantonsschule Wohlen, Switzerland Abstract. A sample of 286 gamma-ray bursts (GRBs) detected by the Swift satellite and 358 GRBs detected by the RHESSI satellite are studied statistically. Previously published articles, based on the BATSE GRB Catalog, claimed the existence of an intermediate subgroup of GRBs with respect to duration. We use the statistical χ2 test and the F-test to compare the number of GRB subgroups in our databases with the earlier BATSE results. Similarly to the BATSE database, the short and long subgroups are well detected in the Swift and RHESSI data. However, contrary to the BATSE data, we have not found a statistically significant intermediate subgroup in either Swift or RHESSI data. Keywords: gamma-ray astrophysics, gamma-ray bursts PACS: 01.30.Cc, 95.55.Ka, 95.85.Pw, 98.70.Rz INTRODUCTION By studying the duration distribution of Gamma-Ray Bursts (GRBs), it was originally found (results from BATSE and Konus-Wind instruments [1, 2]) that there exist two subclasses; the short one with typical durations of less than 2 s and the long one with GRB lasting more than 2 s. However, some articles point to existence of three subclasses of GRBs in the BATSE database with respect to their durations [3, 4]. Works [9] and [10] using a multivariate analysis reveals at least three groups of GRBs in the BATSE data. The article [6] says that the third subclass (with intermediate duration), observed by BATSE, is a bias caused by an instrumental effect. Therefore, we decided to investigate durations and number of groups of GRBs in the RHESSI and Swift databases. RHESSI AND SWIFT INSTRUMENTS The Ramaty High Energy Solar Spectroscopic Imager (RHESSI) is a NASA Small Explorer satellite designed to study hard X-rays and gamma-rays from solar flares (see at: http://hesperia.gsfc.nasa.gov/hessi). It consists mainly of an imaging tube and a spectrometer. The spectrometer consists of nine germanium detectors (7.1 cm diameter and 8.5 cm height). They are lightly shielded only, thus making RHESSI also very useful to detect non solar photons from any direction. The energy range for GRB detection extends from about 50 keV up to 17 MeV depending on the direction. The energy and time resolutions are: ∆E = 3 keV (at 1000 keV), ∆t = 1 µs. An effective area reaches up to 150 cm2 at 200 keV. With a field of view of about half of the sky, RHESSI observes about one or two gamma-ray bursts per week. The Swift satellite is described at http://swift.gsfc.nasa.gov/docs/swift. DATA SAMPLES Two samples are used. The first is the set of 358 GRBs observed by the RHESSI satellite and covers the period from February 2002 to April 2008 (see: http://grb.web.psi.ch). We have used the SSW program under IDL and authors’ routines to derive count light-curves in the energy range 25 - 1500 keV. FIGURE 1. Distribution of the GRB RHESSI durations with 2 log-normal (left) and 3 log-normal (right) fits. The second data-sample is the set of 286 GRBs detected by the Swift satellite. The sample covers the period from November 2004 to December 2007 (see at: http://swift.gsfc.nasa.gov/docs/swift/archive/grb_table). The Swift energy range is 15 - 150 keV. Here we present a comparison of the statistical analysis of these two samples and discuss our results with similar analyses of the BATSE Catalog done by others ([3]-[5]). DURATION DISTRIBUTION The duration distributions of GRBs observed by RHESSI and SWIFT are shown in Figure 1 and Figure 2, respectively. As duration, we use T90 i.e. the time interval during which the cumulative counts increase from 5% to 95% above a background. We followed the method done in [3] and fitted the data by one, two and three log-normal functions. We used the χ2 test to evaluate these fits, the measurement errors being the square root of the number of GRBs per bin. In the case of one log-normal fit for both samples we obtained the goodness of fit < 0.01 %. Therefore, the assumption that there is only one subclass can be rejected. The parameters of the other fits, together with the χ2 values and goodness-of-fits, are noted in Table 1. The question is whether the improvement in χ2, when adding a third log-normal function, is statistically significant. To answer this question, we used the F-test as described in [7]. In the case of the RHESSI data, the obtained critical value F0 implies the probability P(F>F0) = 8.6 % that the improvement is just a statistical fluctuation. This probability is still too high to doubtlessly reject the hypothesis that the improvement in χ2 is being accidentally.

This content is AI-processed based on open access ArXiv data.