4. Spectral analysis
Data were selected in the energy ranges 0.1-4 keV, 1.6-10 keV, 8-40 keV and 15-40 keV, respectively for the LECS, MECS, HPGSPC and PDS, where the instrument responses are well determined and there are sufficient counts. All spectra have been rebinned to at least 20 counts for energy channel, in order to ensure the applicability of test in the spectral fits.
Exploiting the BeppoSAXspectral capability we were able to obtain the simultaneous broad band spectrum (0.1-200 keV) of GS 1843+00. The source shows a very hard spectrum strongly absorbed at lower energies. No deviation from a smooth continuum is observed. This can be seen in Fig. 6 in which the Crab ratio, upper panel, and the ratio times the functional form of the Crab (a featureless power-law with in this energy range), lower panel, are reported. To extract more physical information we fitted the phase averaged spectra obtained from the four co-aligned instruments simultaneously. The conventional model used to describe the spectrum of X-ray pulsar (Pravdo et al. 1978; White et al. 1983) is an absorbed power law with exponential cut-off at higher energies, i.e. a photon spectrum of the form
where E is the photon energy, is the power-law photon index, is the absorbing column and is the photoelectric absorption cross sections due to cold matter (Morrison & McCammon 1983). The high-energy cut-off is modeled by the function of the form:
where is the cut-off energy and is the e-folding energy.
Using this model, we obtained a of 1.08 for 477 degrees of freedom (dof). The best-fit parameters are summarized in Table 1. The spectrum together with the best-fit model are shown in the upper panel of Fig. 7. Fit residuals in terms of are reported in the lower panel and its show no clear evidence of any absorption or emission features. Normalization factors, between the instruments, were left free in the fits. Setting the MECS as reference, the relative normalizations are 0.83 for the LECS, 1.02 for the HPGSPC and 0.79 for the PDS. These values are in good agreement with the ones obtained from the intercalibration analysis of the four Narrow Field Instruments (Fiore et al. 1999). The inclusion in the model of a gaussian line gives a marginal improvement in fit quality (at less then 90 confidence level) for a fluorescent line at 6.4 keV with a flux level of photons cm-2 s-1.
Table 1. Spectral Parameters for the GS 1843+00 broad band fit. All quoted uncertainties are at 90% confidence for a single parameter ()
© European Southern Observatory (ESO) 2000
Online publication: June 5, 2000