Astron. Astrophys. 336, 855-877 (1998)

3. Results and discussion - the point sources

3.1. The bright point sources - IC 4329A, IC 4329 & H1-P3

The three bright sources visible in the IC 4329A field are all especially interesting. H1-P3 is bright, shows marginal evidence for extension in the PSPC (though the fact that no evidence for extension is seen in the HRI data implies that the source is truly unresolved) and appears associated with a quite bright (B magnitude = 16.5) star-like object, some distant. H5-P6 is also very bright, appears to be significantly extended (both in the PSPC and HRI data), and is associated with the giant elliptical galaxy IC 4329. Finally, H11-P8 is extremely bright, showing a great deal of structure, and is undoubtedly due to the Seyfert galaxy, IC 4329A.

It is worth noting again here that an analysis of the PSPC data (in conjunction with COMPTON GRO observations ) has already been published by Madejski et al. (1995) (M95). They deal however, almost exclusively with the spectral properties of IC4329A (plus those of IC 4329 and H1-P3, or as they call it, S3), and so, in the discussion that follows, many of the results we present have not been addressed by M95, and are new, though we do compare the results of our spectral analysis with those of M95.

As in M95, spectra of all three bright objects (IC 4329A, IC 4329 and H1-P3) were analysed. Spectra were extracted for all three objects from within circles of (for IC 4329A and IC 4329) and (for H1-P3) at the position of each source (we note that an extraction radius of for IC 4329A, as used in M95, is likely to be too large, given that this is approximately the distance between IC 4329A and IC 4329). Background spectra were extracted as follows: for IC 4329A, from an annulus to from IC 4329A, thus avoiding contamination from any other bright features; for IC 4329, from a radius circle situated equidistant, on the opposite side of IC 4329A, thus ensuring that contamination from the very bright central source could be removed; and for H1-P3, from a to annulus centred on H1-P3, again avoiding any bright features.

The three background-subtracted spectra, once corrected for exposure and vignetting effects, were fitted with standard spectral models (thermal bremsstrahlung, power law, blackbody and Raymond & Smith (1977) hot plasma models). A number of extra, more complex models have been attempted as regards the IC 4329A spectrum, as in M95, and the results of all the best fits are given below in Table 4 as follows: Source (column 1), spectral model (whether PL - power law plus absorption, BB - blackbody plus absorption, PL/E - power law plus absorption and an edge, RS Raymond & Smith hot plasma plus absorption) (column 2), fitted (column 3), fitted spectral index , where (column 4), fitted temperature (kT, in keV) (column 5) (Note here that in the case of the PL/E model, this column gives the edge energy in keV). The next columns give the metallicity (solar, where an `F' indicates that the value has been frozen) (column 6), the reduced (column 7), and three values of the (0.1-2.4 keV) luminosity (columns 8-10). Two values of luminosity as calculated using the PSPC results are given; one (column 8) gives the `intrinsic' luminosity of the source (i.e. correcting for the total ), the second (column 9) gives an `emitted' luminosity (i.e. correcting merely for the Galactic ). The final luminosity column (column 10) gives the intrinsic (0.1-2.4 keV) luminosity, using the count rate observed with the HRI, and calculating the fluxes, assuming identical spectral models as inferred from the PSPC data. All luminosities are calculated for an assumed distance of 64 Mpc (which is almost certainly incorrect in the case of H1-P3, as discussed below).

Table 4. Results of the best model fits to the IC 4329A, IC 4329 and H1-P3 spectra (see text). Models are: PL (power law plus absorption), BB (blackbody plus absorption), PL/E (power law plus absorption and an edge), RS (Raymond & Smith hot plasma plus absorption). In the case of the PL/E fit, the temperature kT refers to the temperature of the edge. Three (0.1-2.4 keV) luminosities are tabulated. One, the intrinsic PSPC luminosity of the source, two, the Galactic -corrected (i.e. emitted) PSPC luminosity (Galactic  cm^-2), and lastly, the intrinsic HRI luminosity, using the HRI count rates in conjunction with the models suggested by the PSPC data.

Although no thermal model (whether a Raymond & Smith hot plasma model or a thermal bremsstrahlung model) is able to fit the IC 4329A data adequately, a simple power law model does gives quite an acceptable fit, the fitted parameters agreeing well with M95 and with Rush et al. (1996). However, as in M95, close inspection of the residuals does suggest an edge-like feature at around 0.7 keV. Incorporating this edge into the model does improve the fit (the data and residuals are shown in Fig. 7), and we are able to reproduce the best-fit results of M95 very accurately. Firstly, a photon index of 1.730.33 is suggested, consistent with M95, with the Ginga data (Piro et al. 1990; Fiore et al. 1992), and with the ASCA data (Cappi et al. 1996). Secondly, the edge feature at 0.720.07 keV is found at exactly the same energy as in M95. As M95 suggest, the energy of this edge is inconsistent with that expected if there were a neutral absorber present, and this strongly suggests the presence of an ionized absorber (OVI , OVII ). Further modelling, to address the question of the true nature of this absorber, is possible. However, because of the modest spectral resolution of the PSPC, one cannot distinguish between different models, i.e. between an ionized absorber model, a partial covering by neutral material model, and a high column cold absorber model (see M95 for a detailed discussion).

Fig. 7. IC 4329A spectrum with the best-fit power-law plus absorption edge model (see Table 4). The pulse height spectrum of the total X-ray emission is indicated by crosses, and the fit, by the solid line.

It is worth noting that the inferred , 27.9 cm^-2, is substantially larger than the Galactic in the direction of IC 4329A (4.4 cm^-2; Dickey & Lockman 1990), indicating the presence of a large intrinsic absorption. This is not too surprising given the edge-on nature of the galaxy. In their study of the soft X-ray properties of Seyfert galaxies in the ROSAT All-Sky Survey, Rush et al. (1996) also find a very significant excess in the best fit .

The intrinsic (0.1-2.4 keV) luminosity of IC 4329A, 6.3 erg s^-1, is very large, within the top 10% or so of the Seyferts within the Rush et al. (1996) All-Sky Survey sample. It is an extremely luminous galaxy, and an extremely luminous Seyfert galaxy as well.

The IC 4329 spectrum on the other hand, is only fitted adequately well by a thermal (Raymond & Smith hot plasma) spectrum. The best fit, while keeping the metallicity frozen at solar, results in a well-constrained, 1.080.06 keV spectrum, absorbed by a column of cm^-2, a column consistent with (though on the low side of) the Galactic value. This result is entirely consistent with M95. Fitting of the spectrum while letting the metallicity optimize, gives a column entirely consistent with the Galactic value and a low (0.40.2 solar) metallicity, though the fitted temperature is less well constrained than in the frozen-metallicity case. The intrinsic (0.1-2.4 keV) luminosity of IC 4329, 7.9 erg s^-1, is somewhat higher than average when compared to optically similar systems (Fabbiano et al. 1992). The fitted temperature is entirely consistent with that of ellipticals, ASCA observations resulting in temperatures for several early-type galaxies of between 0.7 and 1.2 keV (Matsushita et al. 1994; Rangarajan et al. 1995). Similarly, the low fitted metallicity appears to be consistent with ellipticals, high-resolution studies with ASCA revealing abundances solar in several cases (Loewenstein et al. 1994; Matsushita et al. 1994).

The H1-P3 spectrum is best fit with a power law model of photon index 2.350.33, absorbed by a column of cm^-2, consistent with that out of our own Galaxy. It is almost certainly due to a background quasar, given the facts that it is unresolved in the HRI data, it appears coincident with a quite bright (B mag = 16.5) star-like object, and it has a spectrum consistent with that of quasars (power law with photon indices in the range 2.20.2; Branduardi-Raymont et al. 1994; Roche et al. 1995).

Finally, note that, in Table 3, it appeared that, in the cases of these bright sources, the inferred HRI and PSPC fluxes did not agree particularly well. This could have been attributable to time-variability or the assumption of the wrong spectral model. We have already seen however (Figs. 4-6) that none of these sources appear to be particularly time-variable in either the HRI or the PSPC, and usage of the correct spectral model, as has been done here, has only really aided the situation in the case of IC 4329, and then, only slightly. In the case of H1-P3, it is possible that the source has varied between the HRI and PSPC observations. The fact that the object is very likely to be a QSO or background AGN adds some credence to this. Also bare in mind that the quality of fit to the H1-P3 PSPC spectra is not excellent, the reduced being only 1.4. In the case of IC 4329A and IC 4329, the situation is rather intriguing. Both sources appear extended however, and this will lead to a reduction in the calculated HRI count rates, compared to the PSPC count rates. Furthermore, IC 4329A is extremely bright, accentuating the above effect. Finally, as may be the case (Fig. 2), and is discussed in detail later, if there were a large amount of low-surface brightness, diffuse emission in the vicinity of these two galaxies, this could very well lead to a reduction in the HRI count rates compared to the PSPC count rates, the HRI being relatively far less sensitive to this type of emission than the PSPC.

3.2. The secondary point sources

Moving on to the remaining point sources, many interesting results have been obtained. Feature P4, for instance, appears elongated in the east-west direction in the PSPC image (Fig. 3). The HRI is able to resolve this feature into two separate, equally bright sources (H2 and H4), the more western of which (H2) appears coincident with a bright stellar-like object, with a B magnitude of 10.7. The apparent optical counterpart to H4 (not seen in the APM finding charts of Irwin et al. (1994)) is much fainter (B=14.7, see Fig. 8 (left)). What is rather striking though, is that, what appears to be a `twin' of H2/4-P4 can be seen on the opposite side of IC 4329A, at an extremely similar projected distance from the bright central galaxy. This source, H16-P12, appears coincident with a rather faint (B=17.6) stellar-like object, less than west of HRI position (Fig.8_(right)). The positioning of these two sources with respect to the central bright galaxy, IC 4329A, is both unusual and intriguing.

Fig. 8. Contours of HRI emission overlayed onto optical digitized sky survey images for the H2/H4 field (left) and the H16 field (right). The X-ray image has been smoothed with a Gaussian filter of FWHM, and the contours correspond to 0.0625 and 0.1875 cts arcsec-2.

On July 6, 1997, during an observing campaign at the 2.2m ESO/MPG telescope at La Silla observatory, we obtained spectra of these optical candidates using the EFOSC2 spectrograph with grism #4, a -wide long slit and the LORAL CCD. This gave a dispersion of 2 Å per pixel, a spectral coverage of 4100-7500 Å, and a spectral resolution of 12 Å FWHM. The seeing was typically 1.5". The data were reduced according to the procedure given in Pietsch et al. (1998).

Through these optical observations, we have established that all three sources have nothing at all to do with IC 4329A, and are in fact Galactic foreground or background objects. Sources H2 and H4 to the south-west, when compared with Jacoby et al.'s (1984) library of stellar spectra, appear to be foreground stars of types G3 V and M5 V respectively. Furthermore, the north-eastern source looks to be a background quasar with a redshift of . The HRI count rates measured are consistent with the X-ray fluxes expected from these source classes.

All of the remaining non-bright sources appear to have very close () optical counterparts, the brightest of which is that associated with H17-P14 (B=14.7).

© European Southern Observatory (ESO) 1998

Online publication: July 27, 1998
helpdesk.link@springer.de