Problems in the analysis of gamma-ray burst spectra

Spectral measurements by the Solar Maximum Mission have been used to confirm the cyclotron lines in gamma-ray bursts reported from the Konus experiment. We present ISEE-3 data for the same burst (GB800419) during the same period of time which show no line. We discuss various problems in the analysis of scintillator spectra and point out that unfolded spectra are not necessarily unique and that the position of a data point in a deconvolved spectrum may vary depending on the assumed overall shape of the spectrum. As a result, if the analysis assumes a soft spectrum (such as optically thin thermal bremsstrahlung) an absorption feature might appear, whereas a harder spectrum (such as a Comptonized blackbody) would not require the feature. Since the continuum shape probably changes during the duration of a typical burst, the nonuniqueness of the spectral unfolding, combined with the assumption that the continuum is optically thin thermal bremsstrahlung, could give rise to spurious absorption features which vary on a time scale of seconds. Despite these problems, there is still some evidence for narrow spectral lines in the range 45 to 65 keV but not for the broad lines reported from the Konus experiment. Unfortunately, the range 45 to 65 keV is the most difficult spectral region to unfold