A Comprehensive Analysis of Fermi Gamma-Ray Burst Data. I. Spectral Components and Their Possible Physical Origins of LAT/GBM GRBs

Abstract

We present a systematic analysis of the spectral and temporal properties of 17 GRBs co-detected by GBM and LAT on board the Fermi satellite by May 2010. We performed a time-resolved spectral analysis of all the bursts with the finest temporal resolution allowed by statistics, in order to avoid temporal smearing of different spectral components. We found that the time-resolved spectra of 14 out of 17 GRBs are best modeled with the Band function over the entire Fermi spectral range, which may suggest a common origin for emissions detected by LAT and GBM. GRB 090902B and GRB 090510 require the superposition between an MeV component and an extra power law component, with the former having a sharp cutoff above E_p. For GRB 090902B, this MeV component becomes progressively narrower as the time bin gets smaller, and can be fit with a Planck function as the time bin becomes small enough. In general, we speculate that phenomenologically there may be three elemental spectral components : (I) a Band-function component (e.g. in GRB 080916C) that extends in a wide energy range and does not narrow with reducing time bins, which may be of the non-thermal origin; (II) a quasi-thermal component (e.g. in GRB 090902B) with the spectra progressively narrowing with reducing time bins; and (III) another non-thermal power law component extending to high energies. The spectra of different bursts may be decomposed into one or more of these elemental components. We compare this sample with the BATSE sample and investigate some correlations among spectral parameters. We discuss the physical implications of the data analysis results for GRB prompt emission, including jet compositions (matter-dominated vs. Poynting-flux-dominated outflow), emission sites (internal shock, external shock or photosphere), as well as radiation mechanisms (synchrotron, synchrotron self-Compton, or thermal Compton upscattering).