X-ray imaging of the ionisation cones in NGC 5252
Key Words.:Galaxies:active – Galaxies:individual:NGC 5252 – Galaxies:Seyfert – X-ray:galaxies
Context:The physical conditions of the gas forming the narrow line regions (NLR) in active galactic nuclei (AGN) have been extensively studied in the optical band. Recently, detailed X-ray studies have shown how the emission in the 0.1-2 keV band detected in Seyfert 2 galaxies is associated to gas lying close to or associated with the NLR.
Aims:We take advantage of the spectacular extension (15”) of the NLR in the type II Seyfert galaxy NGC 5252 and of the complementary characteristics of – and to investigate the physical conditions of the gas in this galaxy.
Methods:The X-ray data from – are used to define the spectral properties of the ionising nuclear source. The data are used to trace the spatial characteristics of the soft X-ray emission. This information is then compared to the optical HST characteristics of the NLR in NGC 5252.
Results:The X-ray spectrum of the nucleus of NGC 5252 is intrinsically flat (1.4-1.5), and absorbed by neutral gas with a column density N10 cm. Below 1 keV a soft excess is detected. The high-resolution spectrum obtained with the – RGS shows the presence, in the 0.2-1.5 keV range, of emission lines which strongly indicate that the soft X-ray component is essentially due to ionised gas. Moreover, the soft X-ray emission is spatially resolved around the nucleus and well overlaps the images obtained in narrow optical bands centered around the [OIII] emission line at 5007. The [OIII]/soft-X flux ratios along the ionisation cones is basically constant. This indicates that the electron density does not significantly deviates from the r law (constant ionisation parameter) moving outward from the nucleus.
Conclusions: This result combined with previous optical studies suggest two plausible but different scenarios in the reconstruction of the last 30000 years history of the central AGN. The most promising one is that the source is indeed a “quasar relic” with steady and inefficient energy release from the accretion of matter onto the central super-massive black-hole. This scenario is suggested also by the flat nuclear X-ray spectrum that suggests an advection dominate accretion flow (ADAF) like emission mechanism.
Obscuration of the nuclear emission in type II AGN allows the study of soft X-ray spectral components, which are normally outshone by the direct component in type I unobscured objects. It has been well known since the early day of X-ray spectroscopy that excess emission above the extrapolation of the absorbed nuclear radiation is present in almost all bright Seyfert 2s (Turner et al. 1997). This excess appears smooth when measured with instruments with moderate energy resolutions such as CCD. However, high-resolution (grating) measurements with Chandra and XMM-Newton revealed that this excess is generally due to a blending of strong recombination lines from He- and H-like transitions of elements from Carbon to Nitrogen (Sako et al. 2000, Sambruna et al. 2001, Kinkhabwala et al. 2002, Armentrout et al. 2007). X-ray spectral diagnostics (Kinkhabwala et al. 2002, Guainazzi & Bianchi 2006) and a close morphological coincidence between the soft X-rays and the [OIII] in Extended Narrow Line Regions (ENLR; Bianchi et al. 2006, Bianchi et al. 2010) strongly indicate that the gas is photoionised by the AGN, with an important role played by resonant scattering.
In this context, NGC 5252 represents an extraordinary laboratory to study the feedback between the AGN output and circumnuclear gas on kpc scale, thanks to its spectacular ionisation cones (Tadhunter & Tsvetanov 1989).
NGC 5252 is classified as Seyfert 1.9 (ost ()) S0 (dev ()) nearby (z=0.023,) galaxy (N=2.1410 cm, Dickey & Lockman, 1990). Small radio jets (r4”) have been detected and found to be aligned with the ionisation cones (wilson94 ()). Nonetheless, the host galaxy luminosity (M-22. capetti ()), mass (M2.410M, marc ()) and the mass of the central super-massive black-hole (M10M, capetti ()) are more typical of quasar than Seyfert galaxies. These pieces of evidence led capetti () to speculate that NGC 5252 is most probably to be considered a QSO relic. This view is in agreement with ”downsizing” scenarios about the evolution of super-massive black-hole (SMBH) in cosmic times. Accordingly with these scenarios, most massive SMBHs formed and evolved earlier than lower mass ones.
Ionisation cones are one of the strongest argument in favour of the Seyfert unification scenarios (Antonucci 1993). For this reason, NGC 5252 is also an important laboratory to test AGN geometrical models. From a diferent point of view, AGN activity has been recognized, since a while as a key component of the SMBH host galaxy co-evolution and AGN feedback is likely to self-regulate or be responsible of the observed properties (Menci et al. 2004). The very existence of ionization cones witness that feedback/winds are or were active and thus these sources are ideal laboratories for feedback.
X-ray measurements allows to directly link the properties of the gas emitting optical lines with the intrinsic AGN power, which in type II AGN can be truly measured only at energies larger than the soft photoelectric cut-off due to the AGN obscuring matter. Furthermore, the morphological coincidence between X-rays and optical emission in ENLR (Bianchi et al. 2006) points to a fundamental physical link between the two wavebands. They need to be studied simultaneously in order to derive the correct energy budget in the ionisation cones. Prompted by these motivations, we have performed deep X-ray observations of NGC 5252 at the highest spatial and spectral resolution currently available with Chandra and XMM-Newton. The results of these observations are the subject of this paper.
2 The nuclear spectrum
NGC5252 was observed by XMM-Newton on 2003, July 18th, with the EPIC CCDs (MOS and pn in full window, see Tab. 1) as the prime instrument for a total duration of 67 ks. The Observation Data Files (ODFs) were reduced and analysed using the latest Science Analysis System (SAS) software package (gabriel03 ()) with associated latest calibration files. We used the epproc and emproc tasks to generate event files and remove dead and hot pixels. Several time intervals with a high background rate were identified in the single events light curve at energy 10 keV and were removed, yielding a net exposure of 50 ks for the MOS and 38 ks for the pn. Pile-up is negligible in this source, according to the epatplot SAS task outcome. Patterns 12 and 4 were considered for MOS and pn, respectively. Source counts were extracted from a circular region with radius 50, thus encompassing a large fraction of the optically-defined galaxy. Background was estimated using both blank-sky files and locally from a offset source-free region. Light curves in the soft (0.5-2 keV) and hard (2-10 keV) energy bands were first investigated. We found no significant flux nor spectral variations, thus considered the time averaged spectrum.
The best-fit spectrum is shown in Fig. 1.
If the nuclear emission is modeled with a simple absorbed power-law we obtain an extremely flat photon index [=1.050.10 and =(2.20.1) cm, the reported errors are, here and hereafter, at 90% confidence level], plus a soft component emerging at energies below E1 keV. It can be parametrised with a scattered power-law with a steep photon index =3.00.2. There is also evidence for one (or few) soft emission lines in addition to the soft continuum, a Fe K line with E6.440.05 keV and EW=5025 eV, and an absorption edge at E=7.00.1 keV and optical depth =0.310.05. A significantly better fit is obtained in case of a power-law plus a thermal component, namely in (mewe85 ()), is used to model the soft X-ray band of NGC 5252 (/d.o.f.=1436/1410 in the first case while /d.o.f.=1329/1410 with ). In this case, the temperature of the plasma is 0.17 keV. The investigation on the true nature of this soft X-ray component will be the subject of the next Sections. Here it is important to note that, whatever the fitting of the data below 1 keV, the best-fit model for the nuclear emission of NGC 5252 above 1 keV is typical in shape, but flatter (1.4-1.5) than normally found from Seyfert 2 galaxies (1.5-2.5, Turner & Pounds 1989; turner97 (); risaliti02 (); cappi06 (); Dadina 2008). It is however consistent with the previous ASCA measurement (1-1.5, cappi96 ()), confirming the need in this source for a more complex absorption (either multiple, ionised or both) in order to recover a steeper canonical photon index. With the above model we measure, for the soft (0.5-2 keV) component, a flux of 3.5 10erg cms corresponding to a luminosity of 4.110 erg/s and, for the hard (2-10 keV) component, a flux of 8.9 10erg cms corresponding to a (unabsorbed) luminosity of 1.210 erg/s. This is consistent, within a few tens percent, with previous ASCA and BeppoSAX values (Cappi et al. 1996; Dadina 2007). It is worth noting here that, from IR diagnostics, we expect that star forming activity should contribute to less than 1% to the total soft X-ray emission (Cappi et al. 1996).
3 High-resolution spectroscopy of the AGN environment
The Reflection Grating Spectrometer (RGS; derherder01 ()) on board XMM-Newton observed NGC 5252 simultaneously to the EPIC cameras (Tab. 1, Fig. 1). It produces high-resolution (first order resolution 600-1700 km s) spectra in the 6–35 Å (0.35–2 keV) range. Its 2.5 diameter slit fully encompasses the ionisation cones and the host galaxy. The RGS spectrum represents therefore only the average conditions of the soft X-ray emitting gas across the nucleus and the cone.
count-rate in the 0.2-10 keV band; count-rate in the 0.4-1.2 keV
RGS data were reduced starting from the Observation Data Files with SASv6.5 (gabriel03 ()), and using the latest calibration files. The SAS meta-task rgsproc was used to generate source and background spectra, assuming as a reference coordinate coincident with the optical nucleus of NGC 5252. Background spectra were generated using both blank field maps - accumulated across the whole mission - and a “local” background accumulated during the observation. The former, based on a model of the estimated background on the basis of the count rates detected in the most external of the camera CCDs, overestimates the intrinsic background level during the observation. We have therefore employed the “local” background hereafter. A correction factor to the count background spectrum has been applied to take into account the size of the extraction region, which corresponds to the area of the RGS active CCDs outside the 98% percentage point of the line spread function in the cross-dispersion direction.
We simultaneously fit the spectra of the two cameras
following the procedure outlined in Guainazzi & Bianchi
|(eV)||(10 erg s)||(km s)||(km s)|
|Ovii He- (f)||110|
|Ovii He- (i)||0.9|
|Ovii He- (r)||1.4|
None of them is a Radiative Recombination Continuum (RRC). A (admittedly loose) constrain on the width of the Gaussian profile can be obtained on the Ovii He- triplet only: km s (8.2 eV).
Diagnostic parameters involving the intensity of the Ovii He- triplets can be in principle used to pinpoint the physical process responsible for the bulk of the X-ray emission in high-resolution spectra. The detection of the forbidden () component only allows to set lower limits on the standard triplet diagnostics (gabriel69 (); porquet00 ()): , (where R is the ratio between forbidden and intercobination lines and depends on the electron density, while G is the ratio between intercombination plus forbidden lines and the resonance line, gabriel69 (); porquet00 ()).
These limits, although fully consistent with photoionised plasmas, do not rule out collisional ionisation. Guainazzi & Bianchi (2007) proposed a criterion to discriminate, on a statistical basis, between AGN- and starburst-powered sources based on the location of the source in an empirical observable plane: integrated luminosity of the He- and H-like Oxygen lines, , against the intensity ratio between the and the Oviii Ly-. In NGC 5252 , and erg s. These values put NGC 5252 in the plane locus preferentially occupied by photoionised (AGN) sources (guainazzi09 ()). We estimated also the flux density associated to the continuum, using a line-free energy range between 586 and 606 eV: erg s.
4 X-ray imaging of the ionisation cones
Chandra observed NGC 5252 on August 11, 2003 with the ACIS-S detector in standard VFAINT configuration. Data reduction was performed with CIAO version 3.3 and associated CALDB version 3.2. “Level 1” events were corrected for bad pixels, gain spatial dependency, and charge transfer inefficiency via acis_process_events.
Although the correction for read-out streaks was applied as well, some out-of-time events remain in the final cleaned event list, and were removed by applying a 2 pixels (1)-wide tilted rectangular box around the streak.
ACIS-S images in the 2 around the optical core of NGC 5252 are shown in Fig. 2 in the 0.2–1 keV and in the 1–10 keV energy bands. The soft band clearly shows extended emission in the North-South direction on both sides of the nucleus. On the contrary, the hard band image is point-like. We extracted a radial profile of the latter, and compared it with the expected instrumental Point Spread Function (PSF) of a source with the same spectral energy distribution as the NGC 5252 nucleus. The two profiles are perfectly consistent up to 30 off-axis (see Fig. 3).
In order to characterize the spectral behavior of the diffuse emission, we have extracted spectra from four regions, identified in the soft image (Fig. 2): the nucleus (N), a S-E source about 3.2 from the nucleus (SENS), and the South (SS) and North (NS) diffuse Spots. Background spectra were generated from a large circle 57 wide around the galaxy core, once a inner circle, as well as 5 circles around each serendipitous point sources were removed. Alternative choices of the background regions do not substantially change the results presented in this section. Source spectra were rebinned in order to over-sample the intrinsic instrumental energy resolution by a factor 3, and to have at least 25 background-subtracted counts in each spectral bin. The latter criterion ensures the applicability of the statistics.
For all spectra, we have employed a baseline model that include a thermal emission component from collisionally excited plasma (mekal in Xspec; mewe85 ()). This choice was done for simplicity and the only information obtained using mekal is the flux of the thermal component. This is particularly true for the SS and NS regions where the complexity of the mekal model is well above the quality of the data. Moreover, a photoelectrically-absorbed power-law was always included in the data. The physical meaning of the latter is different depending on the region where the spectrum was extracted. For the nuclear region, the non-thermal component represents the contribution of the active nucleus; for the other regions, the integrated contribution of hard galactic sources such as, for example, X-ray binaries, cataclysmic variables or supernova remnants. We therefore refrain from attributing a physical meaning to the power-law spectral indeces in the latter case. The spectra and corresponding best-fits are shown in Fig. 4. A summary of the spectral results is presented in Tab. 3. The Chandra data confirm that the nuclear spectrum
derived using to fit the spatially resolved data
in units of erg s cm
in units of erg s
is remarkably flat, and is seen through a substantial column density ( cm).
The background subtraction for the spectrum of SENS could be contaminated by the spilling of the nuclear emission. The encircled energy fraction at a distance equal to that between sources N and SENS is 97.5%. However, subtracting a properly rescaled nuclear spectrum to the SENS spectrum yields negative counts above 2 keV. In order to have an independent estimate of the spectral behavior of SENS, we have extracted images 10 around the nuclear region in narrow energy bands (Fig. 5): 200-400 eV,
500-600 eV, 600-700 eV and 700-1000 eV. In a line-dominated plasma, the above energy ranges correspond to bands dominated by Cvi and Cv K, Ovii He-, Oviii Ly-, and Fe-L transitions, respectively. Each image was normalized to the peak of the nuclear emission in that energy band. The soft X-ray SENS spectrum is comparatively dominated by Oxygen transitions, with little contribution in either the Carbon or the Iron band.
5 Comparing soft X-ray and [OIII] morphologies
NGC 5252 was observed in the [OIII] band with the WFPC2 on-board HST on 1995, July 23, using the linear ramp filter FR533N. The data were downloaded from MAST (multi-mission archive at STScI). The images were processed through the standard OTFR (on-the-fly reprocessing) calibration pipeline which performs analog-to-digital conversion, bad pixel masking, bias and dark subtraction, flat field correction and photometric calibration. The cosmic rays rejection was performed combining the two images that are usually taken for this scope. Geometric distortion was corrected using the multi drizzle script (koekemoer02 ()).
The relative Chandra-HST astrometry is clearly a fundamental issue for this work. Chandra has a nominal position accuracy of 0.6 while the absolute astrometry of HST is accurate to 1-2. Fortunately, to align the two astrometric solutions, we could use a point-like source detected both in the WFPC2 and Chandra fields. This source was previously detected at radio wavelengths (wilson94 ()) and is most probably associated to a background quasar (Tsvetanov et al. 1996). Moreover, as a second reference point, we used the brightest emission peak in the nuclear region of NGC 5252 itself.
Images were calibrated in flux using a constant flux conversion factor of , corresponding to the flux producing a count rate of 1 s in the filter band. The above value is appropriate for the instrumental configuration employed during the NGC 5252 exposure, as indicated by the PHOTFLAM keyword in the image file.
The HST image is shown in Fig. 6.
Tsvetanov et al. (1996), Morse et al. (1998) & Capetti et al. (2005) discussed it in details. We refer the reader to these papers for an extensive discussion on the properties of the optical emission. Their main outcomes can be summarized as follows:
the surface brightness is dominated by the unresolved nucleus
a half-ring structure is apparent S-E of the nucleus at a maximum projected distance of 1.5 kpc. It is probably associated with the near side of an inclined gas disk, whose far side is obscured by the host galaxy dust (morse98 ());
the large scale ionisation cone is traced by thin shells of enhanced emission at either side of the nucleus, well aligned along a P.A.110 at distances between 5 and 11 kpc. Fainter co-aligned structures at scales as large as 20 kpc are detected as well in the O[iii] images; however, we will not discuss these latter structures, as they are beyond the region where X-ray emission associated with NGC 5252 is detected;
there is no evidence of radial motions. The measured velocities of the different structures are fully explained by the rotations of the two disks [nonetheless Acosta-Pulido et al. (1996) claimed the detection of radial motions describing the kinematic properties of the [OIII] emission arcs].
In Fig. 7 we present
the superposition between the soft X-rays iso-intensity contours to the [OIII] image. The HST image spatial resolution has been degraded with a 8 pixel wavelet kernel to match the resolution of the Chandra optics. Regions of enhanced X-ray emission exhibit a remarkable coincidence with the morphology of the optical narrow-band image.
We have calculated the ratio between the [OIII] band and the 0.5-2 keV flux (Fig. 8) for the regions specified, after splitting region SS into two sub-regions divided by a E-W line at (regions “SSNorth” - SSN - and “SSSouth” - SSS - respectively). The ratio exhibit a dynamical range smaller than a factor 2 over distances ranging from less than 100 pc to 1.5 kpc with a slight tendency to decrease with the distance from the nucleus(r). This last effect, however, is most probably an observational artifact due to the decreasing in surface brightness of the arcs moving away from the nucleus coupled with the sensitivity limits of to extended sources. At a first glance, the [OIII]/Soft-X ratio profile as a function of the distance from the nucleus seems to suggest that the electron density follow a r relation since the number of ionising photons and of the overall average ionisation state of the nuclear species remain almost constant. This result is in agreement with Bianchi et al. (2006), who assumed, however, a very simplified geometry of the emitting gas. A more detailed investigation on this topic in NGC 5252 is hampered by the quality of the data.
The soft X-ray emission of NGC 5252 is clearly
extended and ACIS images demonstrate that
the spectacular ionisazion cones observed in [OIII] have counterparts
in the 0.1-1 keV band.
The cumulative soft X-ray spectrum observed by – is described by
a soft power-law (3). The ACIS images suggest that this is probably due
to a blend of emission lines that mimics such steep power-law as demonstrated
in other type II Seyferts like NGC 1068, Circinus galaxy and Mrk 3 (kin02 (), brink02 (),
ogle03 (), sam01 (), sako00 (), b05 (), pp05 ()).
This scenario is supported also by the detection in the RGS high resolution
spectrum of three emission
lines, of CV, OVII and OVIII, probably due to photoionised gas. This is consistent also by previous
optical studies that excluded
collisional ionisation along the cones of NGC 5252 (Tsvetanov et al. 1996).
Moreover, the presence of in situ ionisation sources due to shocks formed by
large scale outflows interacting with the interstellar matter has been
excluded (morse98 ()).
Thus the source of ionising photons is most probably the nucleus. Under this
we can use the imaging of the arcs to study the physical condition
of the gas along the ionisation cones. In particular,
the constant of [OIII]/(0.5-2 keV) flux ratio along the ionisation cones
within the inner
Optical spectroscopic studies (Acosta-Pulido et al. 1996) suggest that the
radial dependency of the ionization parameter
On the contrary, having U constant and r would imply that L has remained constant during the last 5000 (30000) years. This is consistent with the “quasar-relic” scenario proposed by capetti (). These authors suggested that the nucleus of NGC 5252 is indeed the “relic” of a nucleus that already experienced the activity phase in the past and that now persists in an almost quiescent phase. This is suggested by the high mass of the SMBH (M=10M, Capetti et al. 2005) that indicates that the nucleus has already accreted in the past, the low Eddington ratio (L10L, assuming the bolometric correction from Marconi et al. (2004), L(1/22)L), and the early type (S0) morphology of the AGN host galaxy. In literature it is also reported that the optical emission line ratios in the inner 30” are typical of LINERS (gon ()), thus suggesting that a low efficiency engine is acting at the nucleus of the source. It is worth noting that also the detection of two counterotating disks suggests that NGC 5252 is a ”quasar-relic”. These disks are tracers of a major merging event that occurred, most probably, more than 10 years ago, since the stellar disk of NGC 5252 is undisturbed. If the merging event triggered a phase of AGN activity (see Jogge 2006, and references therein for a discussion on this topic), we can expect that it lasted few/some 10 years (mar (); ste (); jac (); gon2 ()) after which the source has persisted in a quiescent state.
Finally, it is worth noting that the spectrum of the nucleus hosted by NGC 5252 is confirmed to be quite flat (Cappi et al. 1996). As shown, if modeled with a simple absorbed power-law its photon index points to a very hard spectrum (1). The low EW (50 eV) of the neutral (E6.4 keV) iron line is consistent with what expected if the FeK line is produced via transmission in the observed column (N210 cm, Makishima 1986) thus excluding a reflection dominated spectrum. To reconcile, at least marginally, the hardness of the NGC 5252 nuclear spectrum, with previous results for Seyfert galaxies (1.5-2.5; Turner & Pounds, 1989; Nandra & Pounds, 1994; Smith & Done, 1996; Dadina 2008), we must invoke complex absorption models involving partial covering of the source and/or the presence of ionised absorbers along the line of sight. In this case the spectral index becomes 1.4-1.5. It is interesting to note that the flat photon index may be a further clue suggesting that the X-rays may be produced in an ADAF, Narayan & Yi, 1994) as expected in a “quasar-relic”.
Acknowledgements.This paper is based on observations obtained with XMM-Newton, an ESA science mission with instruments and contributions directly funded by ESA Member States and the USA (NASA). MD greatfully acknowledge Barbara De Marco for the helpful discussions. MD, MC and GM greatfully acknowledge ASI financial support under contract I/23/05/0. CV greatfully acknowledge ASI financial support under contract I/088/06/0.
- offprints: M. Dadina, e-mail: firstname.lastname@example.org
- This paper discusses a sample of 69 RGS spectra of type 1.5, 1.8, 1.9 and 2 Seyfert galaxies. The observation of NGC 5252 discussed in this paper belongs to this sample as well.
- The outer arcs and filaments (tad ()) are most probably too weak to be detected in X-rays. Considering the extension of the outer [OIII] arcs, the minimum detectable flux between 0.1-1 keV is F510erg s cm while, assuming a constant [OIII]/soft X-ray ratio, the expected flux should be 10 times lower.
- U=, where L is the source’s luminosity, is the density of the ionized gas, is the distance between the source of ionizing photons and the ionized matter.
- Acosta-Pulido, J. A., Vila-Vilaro, B., Perez-Fournon, I., Wilson, A. S., & Tsvetanov, Z. I. 1996, ApJ, 464, 177
- Antonucci, R. 1993, ARA&A, 31, 473
- Baskin, A., & Laor, A. 2005, MNRAS, 358, 1043
- Bianchi, S., Chiaberge, M., Evans, D. A., Guainazzi, M., Baldi, R. D., Matt, G., & Piconcelli, E. 2010, arXiv:1002.0800
- Bianchi, S., Guainazzi, M., & Chiaberge, M. 2006, A&A, 448, 499
- Bianchi, S., Miniutti, G., Fabian, A. C., & Iwasawa, K. 2005, MNRAS, 360, 380
- Brinkman, A. C., Kaastra, J. S., van der Meer, R. L. J., et al. 2002, A&A, 396, 761
- Capetti, A., Marconi, A., Macchetto, D., & Axon, D. 2005, A&A, 431, 465
- Cappi M., Mihara T., Matsuoka M.., Hayashida K., Weaver K.A., Otani C., 1996, ApJ, 458, 149
- Cappi M., Panessa F., Bassani L., et al., 2006, A&A, 446, 459
- Dadina, M. 2007, A&A, 461, 1209
- Dadina M., 2008, A&A, 485, 417
- der Herder J., Brinkman A.C., Kahn S.M., et al., 2001, A&A, 365, L7
- de Vaucouleurs, G., de Vaucouleurs, A., Corwin, H. G., Jr., Buta, R. J., Pasturel, G., & FouqueÂ´, P. 1991, Third Reference Catalogue of Bright Galaxies (New York: Springer) Devereux, N. A., Ford,
- Dickey, J. M., & Lockman, F. J. 1990, ARA&A, 28, 215
- Gabriel A.H., Jordan C., 1969, MNRAS, 145, 241
- Gabriel C., Denby M., Fyfe D. J., Hoar J., Ibarra A., 2003, in ASP Conf. Ser., Vol. 314 Astronomical Data Analysis Software and Systems XIII, eds. F. Ochsenbein, M. Allen, & D. Egret (San Francisco: ASP), 759
- Goncalves, A. C., Veron, P., & Veron-Cetty, M.-P. 1998, A&A, 333, 877
- Gonçalves, T. S., Steidel, C. C., & Pettini, M. 2008, ApJ, 676, 816
- Guainazzi M., Bianchi S., 2006, MNRAS, 374, 1290
- Guainazzi M., Risaliti G., Nucita A., et al., 2009, A&A, 505, 589
- Jakobsen, P., Jansen, R. A., Wagner, S., & Reimers, D. 2003, A&A, 397, 891
- Jogee, S. 2006, Physics of Active Galactic Nuclei at all Scales, 693, 143
- Kinkhabwala A., Sako M., Behar E., et al., 2002, ApJ, 575, 732
- Koekemoer, A. M., Fruchter, A. S., Hook, R. N., & Hack, W. 2002, in The 2002 HST Calibration Workshop, ed. S. Arribas, A. Koekemoer, & B. Whitmore, Baltimore, MD: Space Telescope Science Institute, 339
- Makishima, K. 1986, The Physics of Accretion onto Compact Objects, 266, 249
- Marconi, A., & Hunt, L. K. 2003, ApJ, 589, L21
- Marconi A., Risaliti G., Gilli R., Hunt L.K., Maiolino R., Salvati M., 2004, MNRAS, 351, 169
- Martini, P., & Weinberg, D. H. 2001, ApJ, 547, 12
- McLure, R. J., & Dunlop, J. S. 2001, MNRAS, 327, 199
- Menci, N., Fiore, F., Perola, G. C., & Cavaliere, A. 2004, ApJ, 606, 58
- Mewe R., Gronenschild E.H.B.M., van der Oord G.H.J., 1985, A&AS, 62, 197
- Morse J.A., Cecil G., Wilson A.S., Tsvetanov Z.I. 1998, ApJ, 505, 159
- Nandra, K., & Pounds, K. A. 1994, MNRAS, 268, 405
- Narayan, R., & Yi, I. 1994, ApJ, 428, L13
- Ogle, P. M., Brookings, T., Canizares, C. R., Lee, J. C., & Marshall, H. L. 2003, A&A, 402, 849
- Osterbrock, D. E., & Martel, A. 1993, ApJ, 414, 552
- Porquet D., Dubau J., 2000, A&AS, 143, 495
- Pounds, K. A. & Page, K. L. 2005, MNRAS, 360, 1123
- Risaliti G., 2002, A&A, 386, 379
- Sako, M., Kahn, S. M., Paerels, F., & Liedahl, D. A. 2000, ApJ, 543, L115
- Sambruna, R. M., Brandt, W. N., Chartas, G., et al. 2001, ApJ, 546, L9
- Smith, D. A., & Done, C. 1996, MNRAS, 280, 355
- Steidel, C. C., Hunt, M. P., Shapley, A. E., Adelberger, K. L., Pettini, M., Dickinson, M., & Giavalisco, M. 2002, ApJ, 576, 653
- Tadhunter, C., & Tsvetanov, Z. 1989, Nature, 341, 422
- Tsvetanov, Z. I., Morse, J. A., Wilson, A. S., & Cecil, G. 1996, ApJ, 458, 172
- Turner T.J., George I.M., Nandra K., Mushotzky R.F., 1997, ApJS 113, 23
- Turner, T. J., & Pounds, K. A. 1989, MNRAS, 240, 833
- Wilson, A. S., & Tsvetanov, Z. I. 1994, AJ, 107, 1227