Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 342, L41-L44 (1999)

Previous Section Next Section Title Page Table of Contents

3. Spectral analysis

LECS and MECS data were rebinned in order to sample the energy resolution of the detector with an accuracy proportional to the count rate: one channel for LECS and 5 channels for MECS. Spectral data from LECS (0.5-4.5 keV), MECS (2-10 keV), and PDS (13-150 keV) have been fitted simultaneously. Normalization constants have been introduced to allow for known differences in the absolute cross-calibration between the detectors. The values of the two constants have been allowed to vary in a [FORMULA]10% interval around the suggested (Cusumano et al. 1998, Dal Fiume private communication) values (C[FORMULA]/C[FORMULA]=0.65, C[FORMULA]/C[FORMULA]=0.90). The best fit values turned out to be within [FORMULA]5% of the suggested ones. The spectral analysis has been performed by means of the XSPEC 10.0 package, and using the instrument response matrices released by the BeppoSAX Science Data Centre in September 1997. All the quoted errors correspond to 90% confidence intervals for one interesting parameter ([FORMULA] of 2.71). Source plus background light curves did not indicate significant flux variability. Therefore the data from the whole observation were summed together for the spectral analysis.


Table 1. Spectral fits results. The Fe edge energy was fixed at 7.1 keV. Note: [FORMULA] PDS data are not included.

All the models used in what follows contain an additional term to allow for the absorption of X-rays due to our Galaxy that in the direction of NGC 2110 amounts to [FORMULA] cm-2 (Elvis et al. 1989). The energy values of the emission line(s) and absorption edge(s) are given in the reference system of the emitting source, unless otherwise stated.

3.1. The 0.5-10 keV spectrum

The LECS-MECS spectrum has been fitted with a simple absorbed power law model with the addition of a narrow gaussian line to account for FeK[FORMULA] emission. The fit is satisfactory ([FORMULA]), and results in a flat spectrum with photon index [FORMULA] and N[FORMULA] cm-2. The line feature is centered at [FORMULA] keV, with an equivalent width (EW) of 176[FORMULA] eV. If the line width is allowed to vary, the additional parameter gives only a negligible statistical improvement in the fit ([FORMULA]), and in any case, the upper limit obtained, [FORMULA] keV, is consistent with the energy resolution of the MECS, which at 6.4 keV is [FORMULA]7% (FWHM, Boella et al. 1997b). The line width was then frozen to zero for the subsequent analysis. The introduction of an absorption edge consistent with neutral Fe ([FORMULA] keV, [FORMULA]) turns out in a significant improvement of the fit ([FORMULA], corresponding to [FORMULA]95% confidence), and gives a first hint for the presence of an additional absorber. The LECS+MECS 2-10 keV data on NGC 2110 confirm basically the previous results obtained by GINGA (Hayashi et al. 1996), BBXRT (Weaver et al. 1995), and ASCA (Hayashi et al. 1996; Turner et al. 1997), and allow a better measurement of the optical depth of the Fe edge detected by ASCA. A second hint for the presence of a complex absorber is given by the fact that the observed line intensity is too high to be produced by transmission through the measured absobing column (Leahy & Creighton 1993). The consistency is reached only if we add a second absorber responsible for the observed FeK edge: assuming the Fe cross section given by Leahy & Creighton (1993), the measured optical depth [FORMULA] corresponds to an equivalent hydrogen column density [FORMULA] cm-2, which is then consistent with the measured Fe line EW.

[FIGURE] Fig. 1. Confidence (68%, 90%, and 99%) contour plot of the 2-10 keV and 13-150 keV photon indices for a double power law model. The two indices are different at [FORMULA]99% confidence.

3.2. The Seyfert 1 nucleus in the PDS spectrum

The most important result of the present work is that the high energy spectrum (13-150 keV) is well fitted ([FORMULA]) by a simple power law model with [FORMULA]. This is the first evidence for the presence of a steep, X-ray spectrum in NGC 2110, previously classified as a "flat-spectrum" source (Smith & Done 1996). In the effort of verifying, and quantifying the significance of the spectral steepening above 10 keV, a broken power law model was used. For simplicity the knee of the broken power law was frozen at 10 keV, but a knee with free energy did not affect the results significantly. The model results in a good ([FORMULA]) fit to the data and the two photon indices are [FORMULA] and [FORMULA]. Fig. 2 shows clearly that the two indices are different at [FORMULA]99% confidence, and that [FORMULA] is consistent with the canonical Seyfert 1 X-ray spectrum slope (Nandra & Pounds 1994). The observed [FORMULA] remains significant also if we introduce the systematics due to cross-calibration between MECS and PDS which are of the order of 3%, or [FORMULA] (Cusumano et al. 1998).

[FIGURE] Fig. 2. Broad band spectrum with residuals for the dual absorber plus FeK line model (see text).

3.3. The broad-band 0.5-150 keV spectrum

Given the 13-150 keV slope, the X-ray spectrum below 10 keV must be interpreted as the result of some kind of reprocessing of the primary emission. In what follows we have tried to model the broad-band BeppoSAX spectrum of NGC 2110 in order to reconcile the observed 2-10 keV flatness with the steepening observed at [FORMULA] keV. As a first attempt we considered the hypothesis that the observed spectrum might include a reflection component that would produce a flatter spectrum below 20-30 keV and a steepening at higher energies. The reflector(s), in the unified model scenario, could be identified with the accretion disk, the Broad Line Region (BLR), or the circumnuclear torus. The reflection component obtained with this model (PEXRAV in XSPEC) is low ([FORMULA] and [FORMULA] for [FORMULA][FORMULA]10000 and 50 keV respectively) and is therefore not consistent with the Fe line EW (which requires a reflection component [FORMULA]). Moreover the spectral index remains flat ([FORMULA]). On the other hand the measured absorption column density cannot explain the observed Fe line EW and FeK edge optical depth, unless a large, [FORMULA], Fe overabundance is introduced in the system.

A further physical situation that can, in an AGN environment, harden and therefore conceal an intrinsic steep spectrum is the presence of a complex absorber. We applied this model and the result (the spectrum and residuals are shown in Fig. 2) was very satisfying ([FORMULA]), with the slope increased to [FORMULA], while the additional partially covering column density was [FORMULA] cm-2, for a covering fraction [FORMULA]. Contours of [FORMULA] versus [FORMULA] are reported in Fig. 3. The best fit value of [FORMULA] is consistent with the optical depth inferred from the additional absorption edge (see section 3.1), and, as a confirmation of this result, the addition of a FeK edge is not required by this model. The partial covering absorbing column is therefore consistent both with the observed depth of the FeK edge, and also with the measured Fe line EW.

[FIGURE] Fig. 3. Confidence contours for the covering fraction vs equivalent Hydrogen column density of the partially covering material in the dual absorber model.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: February 23, 1999