CNN emulator for decaying magnetic field fast-cooling synchrotron spectra is trained on synthetic data and used in Bayesian fits to GRB 231020A, favoring the decaying-field model over the standard version.
Low-energy Spectra of Gamma-ray Bursts from Cooling Electrons
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
The low-energy spectra of gamma-ray bursts' (GRBs) prompt emission are closely related to the energy distribution of electrons, which is further regulated by their cooling processes. We develop a numerical code to calculate the evolution of the electron distribution with given initial parameters, in which three cooling processes (i.e., adiabatic, synchrotron and inverse Compton cooling) and the effect of decaying magnetic field are coherently considered. A sequence of results are presented by exploring the plausible parameter space for both the fireball and the Poynting-flux-dominated regime. Different cooling patterns for the electrons can be identified and they are featured by a specific dominant cooling mechanism. Our results show that the hardening of the low-energy spectra can be attributed to the dominance of synchrotron self-Compton cooling within the internal shock model, or to decaying synchrotron cooling within the Poynting-flux-dominated jet scenario. These two mechanisms can be distinguished by observing the hard low-energy spectra of isolated short pulses in some GRBs. The dominance of adiabatic cooling can also lead to hard low-energy spectra when the ejecta is moving at an extreme relativistic speed. The information from the time-resolved low-energy spectra can help to probe the physical characteristics of the GRB ejecta via our numerical results.
fields
astro-ph.HE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Modeling Gamma-Ray Burst Spectra with Convolutional Neural Networks: Fast-Cooling Synchrotron Emission in a Decaying Magnetic Field
CNN emulator for decaying magnetic field fast-cooling synchrotron spectra is trained on synthetic data and used in Bayesian fits to GRB 231020A, favoring the decaying-field model over the standard version.