Magnetic flare model of γ-ray bursts

Abstract

The thermal synchrotron (TS) interpretation of γ-ray burst continuum spectra¹ has recently gained support from the analysis of an expanding database^2–5. However, this interpretation requires an emission region which is hot (kT∼0.2–1mc²), optically very thin (nh ≲ 10²¹ cm⁻²) with a highly super-Eddington flux F ∼ 10³⁰ erg cm⁻² s⁻¹ (»F_Edd = 10²⁵ M/M_⊙) for a 10-km neutron star. This picture is similar to that first proposed for the 5 March 1979 event^6–8. In addition there are hints that the emission layer is very dense (n_e ⩽ 10²⁴-10²⁶ cm⁻³) and thin (h ⩽ 10⁻³ cm). For example, events which show simultaneous redshifted 511-keV annihilation lines and low energy self-absorption (see Fig. 1) allow us to estimate a lower limit in the range 10²³-10²⁶ cm⁻¹ (ref. 9) to the pair density, provided that the annihilation region coincides with the synchrotron source. To maintain the pair population close to maxwellian, the collision excitation rate into higher Landau levels⁹ (that is, the pitch-angle scattering rate) must exceed the cyclotron decay rates. Coulomb scattering between pairs and protons is too slow. Even if collective processes whose rates are close to the electron plasma frequency, or scattering by heavy ions (for example, Z = 26, rates ∝ Z²), are invoked, a particle density n ⩾ 10²⁵ cm⁻³ is still needed. Both arguments point towards a dense but thin emitting sheet with h ≲ 10⁻³–10⁻⁵ cm. The severe energetics and persistence of such a hot, dense, thin emitting layer prompted us to consider a picture in which the emission regions lie at the foot points of reconnecting magnetic loops, powered by downward impinging electromagnetic waves. We now examine the structure of the emitting sheet, and the generation, propagation and coupling of the electromagnetic energy fluxes to the surface layer, and show that a flare-like model can account for most of the general features of γ-ray bursts.