Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 340, L19-L22 (1998)

Previous Section Next Section Title Page Table of Contents

3. The average X-ray spectra

Spectral analysis was performed separately on the average TOO1 and TOO2 NFI spectra. Spectra were selected in the energy ranges 0.3-10 keV, 1.8-10 keV, 5-20 keV and 15-50 keV for the LECS, MECS, HPGSPC and PDS, respectively. For TOO2 only an upper limit from the PDS in the 15-50 keV energy range was obtained, since the source was much fainter [FORMULA]10 keV than in TOO1. Factors were included in the spectral fitting to allow for known normalization differences between the instruments. Uncertainties and upper limits are quoted at 90% confidence throughout. Fit results are listed in Tables 1 and 2. No simple model, e.g. absorbed power-law, broken or exponentially cut-off power-law, thermal bremsstrahlung, or multi-temperature disk blackbody (Mitsuda et al. 1984) plus power-law, gives a satisfactory fit to either observation. This last model has been successfully fit to the spectra of many soft X-ray transients (e.g., Tanaka & Lewin 1995). Inspection of the residuals reveals the presence of strong emission lines in the spectra. Including such features in the models brings a significant, albeit insufficient, improvement in fit quality. The description of the lines is given in Table 2. All fitted lines are narrow and unresolved by the LECS and MECS. However a blend of narrow lines cannot be excluded. If allowed to vary, the Gaussian widths, [FORMULA], remain small compared to the instrument resolution and the fit statistics do not change significantly. Therefore [FORMULA] was fixed at 0.1 keV.


Table 1. Results of fits to the BeppoSAX NFI spectra. Model code: PL = power-law; CO PL = cut-off power-law; BKNPL = broken power-law; DBBPL = disk blackbody + power-law; 2BRMS = double bremsstrahlung; 2MEKAL = double MEKAL. Except for the 2MEKAL model, emission lines fixed at the energies given in Table 2 are included. The cut-off and break energies are listed for the CO PL and BKNPL models, respectively. The metal abundance "Fe/He" with respect to solar is given for the 2MEKAL model. [FORMULA] is in units of [FORMULA] atom [FORMULA], kT and E are in keV


Table 2. Two bremsstrahlung and narrow Gaussian emission lines model fits to XTE J0421+560. The O VIII and Ne X features in TOO2 are assumed to be blue-shifted from their rest energies of 0.65 and 1.02 keV (see text)

The best fit for TOO1 is obtained using a cut-off power-law with narrow Gaussian emission features. The fit is formally unacceptable with a [FORMULA] of 1.23, for 718 degrees of freedom (dof), but models the overall shape of the 1-20 keV spectrum reasonably well. Next best is a model consisting of two thermal bremsstrahlung components plus emission lines. The observed TOO1 fluxes [FORMULA] and [FORMULA] are 0.3 and [FORMULA] ergs cm-2 s-1. At a distance of 1 kpc these correspond to luminosities of [FORMULA] and [FORMULA] erg s-1.

The X-ray spectrum of XTE J0421+560 changed dramatically between TOO1 and TOO2 (see Fig. 1) with the appearance of strong soft emission at [FORMULA]1 keV. All the models listed in Table 1 show a reduction in [FORMULA] of at least a factor [FORMULA]1.8 between TOO1 and TOO2. Such a change may result from obscuration by material in an expanding shell. The best description of the TOO2 data is achieved with a double bremsstrahlung model including narrow Gaussian emission lines. The observed TOO2 fluxes [FORMULA] and [FORMULA] are 1.7 and 0.05 [FORMULA] ergs cm-2 s-1. The double MEKAL model, which fits the TOO1 spectrum reasonably well, provides a very poor fit to TOO2.

[FIGURE] Fig. 1. Top : Deconvolved photon spectra of XTE J0421+560 during TOO1 and TOO2 using two bremsstrahlung components plus emission lines (see Table 2). Bottom : data to model ratios, where the line normalizations are set equal to zero in the models. The inset shows the TOO2 model ratio near 1 keV where there is strong soft emission

The double bremsstrahlung plus narrow emission lines model gives a reasonable and "simple" parameterization of both spectra and was therefore chosen to compare TOO1 and TOO2. Table 2 shows that the continuum temperatures decreased significantly between TOO1 and TOO2. Features at 1.9, 2.5 and 6.7 keV, identified with He-like K[FORMULA] emission from Si, S, and Fe, are observed in both spectra. There are no large changes in their equivalent widths, EW, or mean energies between the observations. In TOO1 a feature is present at 0.99 keV with an EW of 163 eV. This may be identified with K[FORMULA] emission from Ne X and/or a number of Fe-L transitions. In TOO2 intense features are present at 0.74 and 1.15 keV with EWs of 1420 and 635 eV, respectively. There are no prominent emission lines with energies close to 0.74 keV. However, if the 1.15 keV feature is interpreted as the Doppler shifted Ne X /Fe-L complex (observed at 0.99 keV in TOO1), and if the 0.74 keV feature is also Doppler shifted by the same amount, its rest energy is 0.63 keV - close to the energy of the prominent O VIII line at 0.65 keV. We therefore tentatively identify the TOO2 0.74 keV feature with blue-shifted O VIII emission. The upper limit to a Gaussian emission feature at 0.65 keV during TOO1 is 36 eV. Finally, the TOO1 fits improve significantly (at 99% confidence using the F-statistic) when a Ca XIX line is included at 3.9 keV. However, the EW is small ([FORMULA] eV) and this line is not required in TOO2.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: November 9, 1998