Fast spectral fitting of hard X-ray bremsstrahlung from truncated power-law electron spectra

¹ Department of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ, UK e-mail: john@astro.gla.ac.uk
² Astronomický ústav AV ČR, v.v.i., Fričova 298, 25165 Ondřejov, Czech Republic
³ CNR-INFM LAMIA, via Dodecaneso 33, 16146 Genova, Italy e-mail: michele.piana@univr.it
⁴ Dipartimento di Informatica, Università di Verona, Ca' Vignal 2, Strada le Grazie 15, 37134 Verona, Italy

Received: 1 February 2008
Accepted: 31 May 2008

Abstract

Context. Hard X-ray bremsstrahlung continuum spectra, such as from solar flares, are commonly described in terms of power-law fits, either to the photon spectra themselves or to the electron spectra responsible for them. In applications various approximate relations between electron and photon spectral indices are often used for energies both above and below electron low-energy cutoffs.

Aims. We examine the form of the exact relationships in various situations, and for various cross-sections, showing that empirical relations sometimes used can be highly misleading especially at energies below the low-energy cutoff, and consider how to improve fitting procedures.

Methods. We obtain expressions for photon spectra from single, double and truncated power-law electron spectra for a variety of cross-sections and for the thin and thick target models and simple analytic expressions for the non-relativistic Bethe-Heitler case.

Results. We show that below the low-energy cutoff Kramers and other constant spectral index forms commonly used are very poor approximations to accurate results, but that our analytical forms are a good match; and that above a low-energy cutoff, the Kramers and non-relativistic Bethe-Heitler results match reasonably well with results for up to energies around 100 keV.

Conclusions. Analytical forms of the non-relativistic Bethe-Heitler photon spectra from general power-law electron spectra are good match to exact results for both thin and thick targets and they enable much faster spectral fitting than evaluation of the full spectral integrations.