Other ``normal'' Seyfert nuclei

While MCG-6-30-15 has been subjected to particularly detailed study, there have also been intensive studies of other Seyfert galaxies. Given the limited collecting area of ASCA, multi-day observations of bright AGN were required in order to obtain reasonable signal-to-noise in the relativistic iron line. Prior to the existence of many such long datasets, Paul Nandra and co-workers showed that the relativistic iron line is present in the ``average'' ASCA spectrum of a sample of two dozen Seyfert-1 galaxies [226]. However, in the past 5 years, deep observations of Seyfert-1 galaxies with ASCA and, more recently, Chandra and XMM-Newton have been performed. Here we summarize the results of these studies.

In addition to MCG-6-30-15, relativistically broad iron lines with very high signal-to-noise have been found by ASCA in the Seyfert galaxies NGC 3516 and NGC 4151. In NGC 3516, the red-wing of the iron line appears to track changes in the continuum flux, as expected by the simple reprocessing model. The blue-wing of the iron line, however, displays higher amplitude variability that is uncorrelated with the continuum flux, suggesting more complex changes in the pattern of fluorescence across the disk [227,228]. More interestingly, the ASCA data suggested the presence of an absorption line at $\sim 5.9{\rm\thinspace keV}$ (absorbing flux from the broad iron emission line). If real, this feature may arise from redshifted resonance absorption by ionized iron in tenuous plasma above the disk, with the energy shift due to either gravitational redshift [229] or infall of the material [228]. If the latter interpretation is true, this may be a rare detection of material actually in the process of accreting into the black hole. Recent simultaneous Chandra grating and XMM-Newton observations of NGC 3516 have also revealed fascinating substructure within the broad line profile in the form of narrow emission spikes at 5.6keV and 6.2keV [230]. These may be the first indications of the expected transient non-axisymmetric illumination.

The high quality ASCA data on NGC 4151 also reveals interesting variability and complexity within the line profile [231,232,233]. During the 1995-May ASCA observation, the iron line profile displayed significant variability, especially in the strength of the red-wing, despite relatively small changes in the observed continuum [233]. On the other hand, during the 2000-May observation, the opposite conclusion appears to hold, with the continuum undergoing significantly larger changes than the iron line [234]. Thus, it appears that the iron line in a given AGN can change its variability behavior, although the non-uniformity of the data analysis methods employed in these (independent) studies prevents one from making a straightforward conclusion.

In comparison with the very broad accretion disks iron lines in MCG-6-30-15, NGC 3516 and NGC 4151, the well-known Seyfert galaxy NGC 5548 has developed a reputation for displaying a rather narrow iron line. Fitting accretion disk models to ASCA data for this object suggested that the line emitting region of the disk was truncated at an inner radius of about $r=10M$ [235,236]. However, we now know that this source displays a composite iron line [237]. The Chandra-HETG has revealed a ``narrow'' core to the line (with a velocity dispersion of $4500^{+3500}_{-2600}\hbox{${\rm\thinspace km}{\rm\thinspace s}^{-1}\,$}$) originating via fluorescence of material a substantial distance from the black hole. The large equivalent width of this narrow core ($130^{+60}_{-50}{\rm\thinspace eV}$) suggests that this material subtends a large part of the sky as seen by the primary X-ray source. Once this component is accounted for in medium-resolution spectra, the Chandra data are entirely consistent with a relativistic accretion disk extending down to the radius of marginal stability [237]. However, the last word has yet to be spoken about NGC 5548. Very recently (in fact, after the initial submission of this review), there was a convincing non-detection of the broad line in NGC 5548 by XMM-Newton [238]. In order to make sense of results from these different observatories, we are forced to conclude that broad iron lines, in at least some objects, are transitory.

Figure 16: Model line profiles for a narrow iron line superposed on an iron line from an accretion disk viewed at $46^\circ$ (solid line), compared with a line from a disk viewed at an inclination of $20^\circ$. These two cases would be difficult to distinguish in data of limited resolution and signal-to-noise. Figure from [239].

However, our experiences with NGC 5548 does illustrates a point of some general practical importance -- at medium spectral resolution, narrow (i.e., low-velocity) iron line components can be blended together with the relativistic line profile and, unless account for, produce significant systematic errors in measurements of, for example, the disk inclination, emissivity profile and inner emitting radius. For example, it has been claimed on the basis of ASCA measured iron line profiles that the rather absorbed ``intermediate'' Seyfert galaxies are viewed face-on [240], a result that runs counter to our current understanding of AGN geometry (see Fig. 16). However, this result is likely to be an artifact of blending between relativistic iron line profiles and narrow lines -- including narrow lines with reasonable strengths in the spectral models shows that the data for these systems are consistent with fairly edge-on disks [239]. Chandra has made a major contribution in this respect -- high resolution grating spectra with the Chandra/HETG can unambiguously identify the narrow iron line component, thereby allowing it to be subtracted from the medium resolution spectra of the relativistic line profiles. This is another sense in which MCG-6-30-15 is a particularly good system for studying relativistic line profiles -- Chandra/HETG spectra show that it has, at most, a very weak narrow line component [241] (which, interestingly, suggests that the larger-scale environment of this SMBH is very ``clean'' with little evidence for an obscuring torus).

Figure 17: Iron line equivalent width $W_{\rm K\alpha}$ against the relative reflection normalization ${\cal R}$ for the joint ASCA/RXTE campaign on NGC 5548 [236]. The dotted line is the proportionality relationship expected, assuming reflection from a planar slab of cold gas with cosmic abundances. The solid line is the best fit proportionality relationship. The data show an obvious anti-correlation, possibly arguing for flux-correlated changes in the ionization state of the disk.

NGC 5548 was also the first source for which a particular mystery associated with spectral variability was discovered. Jim Chiang and collaborators performed contemporaneous observations of NGC 5548 with ASCA and RXTE, thereby allowing details of the iron line and the reflection continuum to be constrained simultaneously [236]. As the continuum source varied, it was found that the iron line flux was fairly constant or, stated another way, the equivalent width was inversely proportional to the continuum flux. This is partially explained by the narrow component to the iron emission line noted above, which occurs in material light weeks or more from the black hole. The light travel time eliminates any line variability over the course of the month-long ASCA/RXTE monitoring campaign. However, at the same time, it was found that the continuum reflection fraction as measured by RXTE was positively correlated with the continuum flux. In other words, the relative strength of the fluorescent iron line appeared to be anti-correlated with the relative strength of the reflection continuum (see Fig.17). Since the iron line and reflection continuum are different facets of the same phenomenon (i.e. X-ray reprocessing in the disk atmosphere), this result is most puzzling. In principle, unmodeled flux correlated changes in the ionization structure of the accretion disk atmosphere might decouple the iron line and apparent reflection continuum strength in this way [203], although the success of detailed ionized disk models in explaining this observation has yet to be demonstrated. Very similar behavior is also found in RXTE studies of MCG-6-30-15 by Julia Lee and collaborators [242].

To summarize the above discussion, iron lines from relativistic accretion disks appear to be generic features in the X-ray spectra of normal Seyfert-1 galaxies, although NGC 5548 and the intermediate Seyferts demonstrate that some care might be needed to extract the relativistic line profile from the data. In one source, MCG-6-30-15, the strength of the relativistic effects argues strongly for a rapidly-rotating black hole. Furthermore, the data suggest that the central regions of the accretion disk may be extracting the black hole spin energy. However, substantial uncertainties remain. In particular, we do not possess even a phenomenological characterization of iron line variability that can predict how the line changes as a function of continuum flux or other source properties. Line variability is of particular interest since it undoubtedly is the key to substantial understanding of accretion disk physics.

The lack of a simple correlation between the iron line strength and the observed X-ray continuum has lead some researchers to throw out the whole picture of a relaticistically-broadened fluorescence iron line, leading them to favour rather complex and seemingly contrived X-ray continua as an explanation for the observed X-ray spectrum (e.g., [234]). However, in the opinion of one of us (CSR), this is an extreme over-reaction to the failure of a particularly simple-minded realization of the fluorescent iron line model. A realistic MHD turbulent disk with have a ``surface'' that, at a local level, may be highly complex both in terms of geometry and ionization state. For example, it is easy to image that the (turbulent) surface is comprised of filaments and/or sheets which have a variety of ionization states. The resulting response of the iron line flux from such a complex surface to a change in the irradiating coronal flux (even assuming no direct thermal/dynamical coupling between the corona and the disk surface) is very unlikely to be linear.

The on-going operation of XMM-Newton promises to make significant progress in the near future. In addition to MCG-6-30-15 and NGC 3516, there are a small number of other normal Seyferts for which XMM-Newton data has already been reported [243,244,245,238]. On the basis of these early studies, it has been suggested that the ASCA studies may have over-estimated the generality of relativistic iron lines in the spectra of Seyfert galaxies. However, robust conclusions must await a more systematic analysis of current and future XMM-Newton datasets, part of which would include a reconciliation with the older yet still valid data from ASCA and RXTE.