Astron. Astrophys. 346, 407-414 (1999)

5. Comparison with ASCA and ROSAT results

Nandra et al. (1997) detected with ASCA a weak ( =  eV) and narrow iron line, whose best-fit parameters are inconsistent with the ones corresponding to the "baseline" BeppoSAX best-fit model. We have therefore re-analyzed the ASCA observation to check the robustness of the iron line variability. ASCA (Tanaka et al. 1994) payload include a pair of Charged-coupled devices (SIS0 and SIS1, 0.57-9 keV) and a pair of gas scintillation proportional counters (GIS2 and GIS3, 0.7-10 keV). It observed NGC 4593 on January 9, 1994. Spectra have been extracted from the screened event files, which have been filtered according to standard criteria, using extraction radii of 2.30´, 3´ and 6´ for the SIS1, SIS0 and GIS, respectively. Net exposure times amount to 30 and 33 ks for the SIS and GIS, respectively. The spectra have been rebinned in order to have at least 20 counts per energy channel. Background spectra have been extracted from blank sky pointing event files, using the same extraction region in detector coordinates as the source. The same BeppoSAX "baseline" model has been applied to the spectra of all detectors simultaneously, except for the fact that ASCA is capable to resolve the OVII and OVIII absorption edges. The parameter , which cannot be constrained by ASCA given the limited bandwidth of its instruments, has been forced to be comprised in the 90% confidence level range determined by BeppoSAX data, i.e. : 0.6-1.6. The time-averaged 2-10 keV flux in the ASCA observation is very close to that measured by BeppoSAX ( erg s^-1 cm^-2). No significant deviation exists between the ASCA and BeppoSAX best-fit parameters, with the exception of a slight steepening of the intrinsic power-law (, cf. Table 1). In particular, the iron emission line best-fit parameters are consistent with the one obtained from the BeppoSAX data analysis within the statistical uncertainties and there is no evidence for a much weaker line in the ASCA data. It is worth noticing that no soft excess is required. The ASCA spectra and residuals against the best-fit are shown in Fig. 9.

Fig. 9. Spectra and best-fit model (upper panel ) and residuals in units of standard deviations (lower panel ) when the modified BeppoSAX "baseline" model is applied to the January 1994 NGC 4593 ASCA observation. Only SIS1 and GIS3 data are shown for clarity

Walter & Fink (1993) report the measure of a very steep () intrinsic spectral index with ROSAT. However, their analysis does not take into account the possible contribution of a ionized absorber. We therefore reanalyzed the data of a pointed NGC 4593 ROSAT/PSPC observation, to compare BeppoSAX and ROSAT findings. The ROSAT observation was performed July 14 1992. The average spectrum was extracted, using a circular region about 2´ radius around the source centroid. The background spectrum was extracted from a surrounding annulus of radii 43 and 75, after removing a 2´ circular area around a weak serendipitous source at , . Total exposure time is 1261 s. Publicly available response matrices, appropriate for the epoch of the observation, were employed and spectral fit were performed in the energy range 0.1-2 keV, where these matrices are best calibrated. A simple power-law model with photoelectric absorption is an adequate description of the spectrum ( dof), with best-fit parameters:  cm^-2, , 0.1-2 keV flux  erg cm^-2 s^-1. The intrinsic spectral index derived by ROSAT is indeed steeper than measured by BeppoSAX or ASCA, but by an amount much smaller than reported by Walter & Fink (1993). The soft excess above a power-law is only 20% in flux against the 270% reported by these authors. The different energy band-passes in which the spectra are taken, or slight residuals misalignment in the cross calibration between ROSAT detectors and other missions (cf. Yaqoob et al. 1994; Iwasawa et al. 1998; Iwasawa et al. 1999) may easily account for this effect. Interestingly enough, no absorption edge is required by the ROSAT data despite the better effective area of the PSPC in comparison to the LECS (the energy resolution is comparable at 0.5 keV). The upper limit on the optical depth of a 0.77 keV absorption edge is 0.26. This might be suggestive of a long-term positive correlation between flux and warm absorber features, since the 0.1-2 keV flux is 60% in the ROSAT observation than in the BeppoSAX one. This aspect may be worth further investigation by future X-ray monitoring programs.

© European Southern Observatory (ESO) 1999

Online publication: May 21, 1999
helpdesk.link@springer.de