Download AGN host galaxies at redshift z ≈ 0.7: peculiar or not?

c ESO 2008
Astronomy & Astrophysics manuscript no. agn
December 20, 2008
AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
A. Böhm1 , L. Wisotzki1 , E. F. Bell2 , K. Jahnke2 , C. Wolf3 , D. Bacon4 , M. Balogh5 , F. D. Barazza6 , M. Barden7 ,
J. A. R. Caldwell8 , M. E. Gray9 , B. Häußler9 , C. Heymans10 , S. Jogee11 , E. van Kampen12 , K. Lane9 ,
D. H. McIntosh13 , K. Meisenheimer2 , C. Y. Peng14 , S. F. Sánchez15 , A. N. Taylor10 , and X. Zheng16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Astrophysikalisches Institut Potsdam, An der Sternwarte 16, D-14482 Potsdam, Germany
Max-Plack-Institut für Astronomie, Königstuhl 17, D-69117 Heidelberg, Germany
Department of Astrophysics, Denys Wilkinson Building, University of Oxford, Keble Road, Oxford, OX1 3RH, UK
Institute of Cosmology and Gravitation, University of Portsmouth, Hampshire Terrace, Portsmouth, PO1 2EG, UK
Department of Physics and Astronomy, University Of Waterloo, Waterloo, Ontario, N2L 3G1, Canada
Laboratoire d’Astrophysique, École Polytechnique Fédérale de Lausanne (EPFL), Observatoire, CH-1290 Sauverny, Switzerland
Institute for Astro- and Particle Physics, University of Innsbruck, Technikerstr. 25/8, A-6020 Innsbruck, Austria
University of Texas, McDonald Observatory, Fort Davis, TX 79734, USA
School of Physics and Astronomy, The University of Nottingham, University Park, Nottingham NG7 2RD, UK
Institute for Astronomy, University of Edinburgh, Blackford Hill, Edinburgh, EH9 3HJ, UK
Department of Astronomy, University of Texas at Austin, 1 University Station, C1400 Austin, TX 78712-0259, USA
European Southern Observatory, Karl-Schwarzschild-Strasse 2, D-85748 Garching bei München, Germany
Department of Astronomy, University of Massachusetts, 710 North Pleasant Street, Amherst, MA 01003, USA
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA
Centro Hispano Aleman de Calar Alto, C/Jesus Durban Remon 2-2, E-04004 Almeria, Spain
Purple Mountain Observatory, National Astronomical Observatories, Chinese Academy of Sciences, Nanjing 210008, PR China
Received ???; accepted ???
ABSTRACT
Aims. We perform a quantitative comparison between the morphologies of the host galaxies of Active Galactic Nuclei (AGN) and
quiescent galaxies in the intermediate redshift regime (z ≈ 0.7). The imaging data are taken from the the large HST/ACS mosaics of
the GEMS and STAGES surveys. Our main aim is to test whether the AGN host galaxy population shows an increased fraction of
morphological perturbations with respect to non-AGN. Such perturbations could be signatures of gravitational interactions which are
suspected as one of the triggering mechanisms of nuclear activity.
Methods. Using images of non-AGN galaxies and stars we create synthetic AGN images to investigate the impact of an optical nucleus
(typical for type-1 AGN) on the morphological analysis of AGN host galaxies. Galaxy morphologies are described using asymmetry
index A, concentration index C, Gini coefficient G and the M20 index. A sample of ∼ 200 synthetic AGN is matched to 21 real AGN
in terms of redshift, host brightness and host–to–nucleus ratios to ensure a reliable comparison between active and quiescent galaxies.
Results. The optical nuclei strongly affect the quantitative morphological properties of the underlying host galaxy even if a careful
multi-component modelling of the AGN images is performed. After accounting for the effect of the nucleus, active and inactive
galaxies show similar distributions in C, A, G and M20 . The morphologies of the distant AGN host population are of intermediate
type, they are clearly distinct from undisturbed early-types or galaxies undergoing strong gravitational interactions.
Conclusions. Our findings seem to argue against major mergers as the main physical driver of AGN activity at intermediate redshifts.
The signatures of gravitational interactions might be too weak to be detected by the commonly used descriptors like CAS at a
significant rate, potentially due to a delay between the time of strongest morphological perturbation and the onset of the AGN phase.
Key words. galaxies: active – galaxies: interactions – galaxies: evolution
1. Introduction
For a long time, Active Galactic Nuclei (AGN) were considered
a very special class of objects, characterised by an accreting supermassive black hole (SMBH) in their center. Only during the
last decade, a SMBH has been found in practically all galaxies
that, firstly, are nearby enough to determine the stellar kinematics in the center at high spatial resolution and, secondly, have a
significant bulge component. It also became evident that the formation and evolution of galaxies and SMBHs are tightly linked.
E.g. the mass of the SMBH is observed to be correlated with
the luminosity (e.g. Marconi & Hunt 2003), velocity dispersion (e.g. Gebhardt et al. 2000; Ferrarese & Merritt 2000) and
Send offprint requests to: [email protected]
mass (e.g. Häring & Rix 2004) of the spheroidal component of
a galaxy.
However, the main mechanism that turns a quiescent SMBH
into an AGN is still under debate. Candidates for this are secular processes like bar-driven gas inflows, gravitional interactions or, as the extreme case, galaxy mergers. Is is likely that the
main triggering mechanism depends on the considered luminosity regime and the cosmic epoch.
One possible approach to identify this process is to compare
the morphological properties of AGN host galaxies and quiescent (non-AGN) galaxies. To this end, some previous studies
have been utilizing descriptors like CAS (Conselice et al. 2000).
Grogin et al. (2005), e.g., found no differences between active
and non-active galaxies at z < 1.3 in terms of their asymmetry
2
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
index A in the Chandra Deep Fields North and South. Similarly,
Pierce et al. (2007) used the Gini coefficent G and M20 index
(Lotz et al. 2004) to quantify the morphologies of X-ray selected
AGN at 0.2 < z < 1.2 from the DEEP2 survey and concluded
that the majority of the AGN reside in non-peculiar early-type
hosts. These studies could be taken as evidence that signs of
gravitational interactions in AGN at intermediate redshifts are
scarce. However, it should be noted that these analyses did not
include a decomposition of the AGN images into the nucleus
and host components. Although an X-ray selection yields a high
fraction of absorbed or so-called “optically dull” AGN, it is not
clear to which extent optical nuclei had an impact on these results.
Performing a decomposition into the contributions of host
and nucleus, Sanchez et al. (2004) found that 12 out of 15 (type1) AGN at 0.5 < z < 1.1 from the GEMS survey reside in
early-type galaxies, with roughly one third of the hosts showing signs of gravitational interactions. Moreover, the rest-frame
U − V colors of the majority of the early-type hosts were significantly bluer than those of non-AGN early-types, indicating
recent or ongoing star formation. In fact, the same has been
found for X-ray selected AGN in the Extended Chandra Deep
Field South (Böhm et al. 2007, Silverman et al. 2008, Wisotzki
et al., in prep.). These findings could be understood in a scenario
where some mechanism triggers star formation and, with some
delay, AGN activity, which eventually suppresses star formation
via feedback processes (e.g. Schawinski et al. 2007).
A few studies have previously investigated the impact of an
optical nucleus on the analysis of the underlying host galaxy
(Sanchez et al. 2004, Simmons & Urry 2008, Kim et al. 2008,
Gabor et al. 2008). In this paper, we will for the first time use real
galaxy images for the simulation of type-1 AGN images (i.e. objects with prominent optical nuclei) and use these for a morphological comparison between quiescent and active galaxies at intermediate redshifts. A detailed discussion of our approach and
results with respect to earlier work will be given in Sect. 5.
The paper is organized as follows: in Sect. 2, we briefly introduce the utilized surveys, Sect. 3 describes the selection of
the AGN sample. The construction of the non-AGN comparison
sample and the simulated AGN images is outlined in Sect. 4. We
perform the morphological analysis of both samples and discuss
our results in Sect. 5; Sect. 6 will give a short summary.
Througout the paper, we assume a flat concordance cosmology with ΩΛ = 0.7, Ωm = 0.3 and H0 = 70 km s−1 Mpc−1 .
Unless otherwise stated, all given magnitudes are in the AB system.
2. Data sets
Our morphological analysis relies on HST/ACS imaging data
taken from the GEMS survey (Rix et al. 2004) and the STAGES
survey (Gray et al. 2008).
GEMS is a two band, F606W and F850LP (similar to the
V− and z-bands), survey consisting of 78 ACS pointings. This
survey covers a continuous field of 28′ × 28′ in the extended
Chandra Deep Field South. Details on the data reduction are
given in Caldwell et al. (2008). The STAGES survey spans a
quarter square degree centered on the A901/2 supercluster at
z = 0.165. Imaged in F606W, the mosaic of 80 ACS tiles forms
the second largest deep image taken by HST. Details on the data
reduction can be found in Gray et al. (2008).
In brief, the data reduction comprised the standard steps
(bias subtraction, flat-fielding, flux calibration), drizzling of the
data from the original 0.05 arcsec/pixel to a final sampling of
0.03 arcsec/pixel, background estimation and variance determination for each pixel. For both data sets, the individual exposures
were multi-drizzled with a Gaussian kernel. Note that the initial
drizzling of the GEMS data was based on a square kernel. In
terms of the detectability of a faint host galaxy of an AGN with
a bright central nucleus, the data produced with a Gaussian kernel have a significantly higher sensitivity. For the purpose of the
analysis presented here, we will use only the F606W images of
GEMS data, to allow a combination with the STAGES tiles as a
basis for a homogeneous sample.
Both the STAGES and GEMS areas are covered by
COMBO-17 (Wolf et al. 2003), a photometric redshift survey
based on imaging in 12 medium- and 5 broad-band filters with
the Wide Field Imager at the MPG/ESO 2.2m telescope on La
Silla, Chile. COMBO-17 incorporates the estimated redshifts
and spectral energy distribution (SED) classifications of several
104 objects. For galaxies brighter than R = 24 (Vega), the error
on the photometric redshifts is typically σz ∼ 0.02(1 + z). For
non-AGN galaxies, stellar mass estimates M∗ are also available
(Wolf et al. 2004, Borch et al. 2006).
3. AGN sample
3.1. Local Point Spread Function
The analysis of the underlying host galaxy in AGN images is a
challanging task, particularly in cases of low flux contrasts between the host and nucleus. The key issue for an accurate removal of the central point source and a determination of the host
properties is a good knowledge of the Point Spread Function
(PSF). Its shape is depending on the used filter, CCD position,
observation date and spectral shape of the source. Tests on the
GEMS data showed that the temporal variations are negligible
compared to the spatial variations (see Sanchez et al. 2004).
Due to the variations of the PSF shape across the CCD (with
pixel–to–pixel flux variations up to 20%), it is mandatory to construct individual, local PSFs for each AGN. These local PSFs
were determined by normalizing and averaging the nearest 35
stars from all tiles (only using stars from GEMS for GEMS AGN
and from STAGES for STAGES AGN) around a given AGN position on the CCD. Very faint (V > 24) or saturated stars were
rejected. The typical maximum distance between the AGN and
the stars that were included was ∼ 38 arcsec in case of GEMS
and ∼ 24 arcsec for the STAGES data.
A small fraction of the STAGES imaging (8 tiles out of 80)
was performed ∼ 6 months later than the main part, which was
spread over only three weeks of observations. The tiles with delayed imaging (no. 29, 46, 75 . . . 80; see Heymans et al. 2008
for details) show significant deviations of their PSFs from those
of the data taken earlier and these tiles were therefore neglected
in our analysis. To construct local PSFs just based on these tiles
was no option, it would have required to include stars from very
large angular distances or to use local PSFs with a much lower
S /N than for majority of the AGN sample.
3.2. Definition of AGN sample
Within the COMBO-17 catalogue, there are 278 objects for
which ACS imaging from STAGES or GEMS is available and
which are classfied as AGN. We restrict our analysis to AGN
in the redshift range 0.5 < z < 1.1. The lower limit here been
chosen simply due to the lack of AGN identified with COMBO17 at lower redshifts. The upper limit is mainly motivated by the
shift of the wavelengths probed by the F606W filter into the rest-
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
3
frame UV which strongly affects the morphological appearance
and would limit the comparability to objects at lower redshifts.
Moreover, there is a dearth of quiescent galaxies in COMBO-17
at redshifts z > 1.1. Because of the SED-based selection technique, the vast majority of the COMBO-17 AGNs are type I,
i.e., they show broad emission lines and an optical nucleus in direct images. We introduced a cut at R < 24 Vega mag for SED
and redshift reliability. In the redshift range 0.5 < z < 1.1, this
results in a subsample of 33 AGN. Finally, we rejected objects
which either were located close to the edges of the tiles or were
observed during the delayed STAGES observations (see last section). The cleaned sample comprised 28 AGN with robust photometric redshifts and good ACS imaging data at 0.5 < z < 1.1.
As a final constraint, we were checking which of the AGN
have a detectable host galaxy. To this end, the local PSF of each
AGN was normalized to the AGN’s flux within an annulus of 2
pixels radius and then this scaled PSF was subtracted from the
AGN image. By definition, this approach is an over-subtraction
of the nucleus because the normalization to the central flux includes the contribution of the host galaxy. In other words, a
scaled PSF subtraction is a way to give a conservative lower
limit on the host-to-nucleus flux ratio. Only objects showing, after scaled PSF subraction, a flux residual of more than 10% of
the total flux, were included in our final sample. This resulted in
21 AGN at a median redshift of hzi = 0.71.
Some of the derived host galaxy parameters are listed in table 1. table to be included
3.3. Decomposition of AGN images
The basis for the selection of the comparison sample was the
COMBO-17 catalogue of the GEMS field. For the pre-selection,
the following constraints were used:
To accurately determine the H/N ratios and the structural properties of the AGN host galaxies, we utilized the GALFIT package
(Peng et al. 2002). GALFIT allows multi-component fitting of
galaxy images in a χ2 minimization approach. For the decomposition of the AGN images, the individually constructed local
PSF was used as a model for the unresolved nucleus of the AGN,
and a Sérsic profile (Sérsic 1968) as a model for the host galaxy.
This model can be used to approximate various kinds of surface
brightness profiles with a Sérsic index n = 1 being equivalent to
an exponential disk and n = 4 to a de Vaucouleurs profile.
The two-component GALFIT modelling introduces a total of
10 free parameters. For the Sérsic component, these were the position xs , ys , total brightness ms , effective radius reff , Sérsic index
n, position angle θ and axis ratio q. For the nucleus model, only
position xn , yn and brightness mn were free parameters. The user
has to define initial guesses for all parameters which GALFIT
uses to start the fitting process. To determine these initial guess
values of ms , xs , ys , reff , θ, mn , xn , yn , Source Extractor (Bertin
& Arnouts 1996) was used. The brightnesses ms and mn were
set according to the Source Extractor output MAG AUTO (Kron
magnitudes) on the sum of both components, with the flux ratio
between host and nucleus taken from the conservative estimate
via scaled PSF subtraction as outlined above. The initial guess
value of the Sérsic index was n = 2.5, representing an intermediate type between an n = 1 exponential disk and a de Vaucouleurs
profile. To perform the fits, thumbnails of size 256 × 256 pixel2 ,
corresponding to 7.68 × 7.68 arcsec2 , were created. Neighboring
objects and bad pixels were masked.
GALFIT allows to set constraint intervals on the free fitting
parameters. The relative positions ∆x = xs − xn and ∆y = ys − yn
were required to be within 5 pixels, the Sérsic index was constrained to 0.25 ≤ n ≤ 8, and the effective radius to 0.1 arcsec
≤ reff ≤ 3 arcsec. All fits were visually checked, in particular
for the fit residuals. For a few objects, it was necessary to set
additional constraints e.g. on the position of the components to
acquire the best fitting results.
4. Comparison sample
As outlined in the introduction, previous studies that were utilizing morphological descriptors like CAS or G/M20 were not
taking into account the impact of the optical nucleus on the
quantitative AGN host galaxy properties. To do so, we have constructed a special kind of comparison sample in three main steps.
Firstly, we defined a set of non-AGN galaxies that is matched
to the redshifts and magnitudes of our sample of 21 AGN hosts.
Secondly, we have used these quiescent galaxies to create synthetic AGN images that cover the same host–to–nucleus ratios as
the real AGN (section 4.3). Thirdly, the synthetic AGN images
have been decomposed into the host and nucleus components
using GALFIT in the same way as the real AGN. After subtracting — for each individual object — the model nucleus from the
AGN image, the recovered host galaxies of the synthetic AGN
have been combined with the host galaxies of the real AGN and
analysed in CAS and G/M20 space (this part will be described in
section 5).
4.1. Selection
1. Apparent Vega magnitude R < 24. In this magnitude regime,
we can rely on photometric redshifts accurate to within 34% and robust SED classifications from COMBO-17 (Wolf
et al. 2003).
2. Source classified as a “galaxy” on the basis of its SED (but
not as an AGN)
3. No X-ray counterpart in the point source catalogue of the
Extended Chandra Deep Field South (Lehmer et al. 2005).
This criterion was applied to de-select absorbed or “optically
dull” AGN.
4. Sérsic index n > 1.5 (derived on the F850LP images, i.e. zband, by Häussler et al. 2007). This constraint was set to
reject bulgeless galaxies. According to the relation between
bulge mass and SMBH mass, bulgeless host galaxies are not
expected to be present in the AGN sample.
3536 galaxies with available GEMS data passed these criteria. In the next step, 10 quiescent counterparts were selected for
each of the 21 AGN hosts within a bin ∆z/z ≤ 0.025 of the AGN
redshift and within ∆mV ≤ 0.05 mag of the total brightness of the
host as determined with the GALFIT decomposition. When less
than 10 non-AGN were found in the catalogue within these limits, they were slightly relaxed and the search repeated iteratively
until 10 counterparts could be selected. After visual inspection
of all 210 objects in the resulting sample, seven galaxies had to
be rejected because they were unresolved. The final non-AGN
sample hence contained 203 galaxies.
The COMBO photometric redshift distributions of the 21
AGN hosts and the 203 quiescent galaxies are shown in Fig. 1.
To allow a direct comparison between the two samples, the
z-distribution of the AGN has been normalized to a total of
203 objects. The median photometric redshift of the AGN sample is hzi = 0.71, that of the quiescent comparison sample is
hzi = 0.70. Fig. 2 displays the distributions in total apparent
4
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
galaxies, objects with tidal tails or minor merger candidates;
3. interactions of “class II”, showing evidence for violent gravitational interactions, with the vast majority of them being major
merger candidates. This established a two-dimensional classification scheme. Our quiescent comparison sample comprises:
1. 40 early-type galaxies, of which 29 are undisturbed, 7 fall
into category “interaction class I”, and 4 are assigned “interaction class II”
2. 94 disk galaxies: 51 undisturbed, 30 interactions of class I,
13 interactions of class II
3. 47 late-types/irregulars: 17 undisturbed, 14 interactions of
class I, 16 interactions of class II
Fig. 1. Comparison between the redshifts of the AGN (dashed) and the
comparison sample of quiescent galaxies (solid). The histograms are
normalized to the same total number of objects.
22 of the non-AGN galaxies were too faint to perform a reliable classification. Examples are shown in Fig. 3.
At the redshifts considered here, even based on high quality ACS imaging data, it is a complicated matter to visually
distinguish between undisturbed irregular galaxies, where the
morphological appearance results from small-scale kinematical
anomalies, and mildly disturbed irregular galaxies which we refer to as “class I” interactions. We cannot rule out an overlap
between these two types due to misclassifications. We are also
aware that it is challenging to assign a morphological type to a
galaxy undergoing a major merger, during which the initial appearance of an object will be strongly affected or even destroyed.
This is reflected in the relatively high fraction of interactions of
class II among the late-type/irregular galaxies (34% compared to
respective fractions of 10% and 13% for early-types and disks).
Within a given morphological class, cases of strong morphological perturbations can however be identified robustly and these
are the particularly interesting objects to test the interactionactivity scenario.
4.3. Synthetic AGN images
Fig. 2. Distribution of the AGN host galaxy magnitudes as derived
with GALFIT (dashed) compared to the quiescent comparison sample
(solid). The histograms are normalized to the same total number of objects.
V magnitude of the AGN hosts (from GALFIT) and the quiescent comparison sample (Source Extractor MAG AUTO values). Again, both samples’ distributions are normalized to the
same number of objects for the sake of comparability. The median total magnitudes of the two samples are hmv i = 23.05
(AGN) and hmv i = 23.33 (comparison sample), respectively.
4.2. Visual classification
To have some kind of calibration at hand for the quantitative
analysis in CAS and G/M20 space, we performed a visual classification of all galaxies in the quiescent comparison sample.
Our classification scheme considers the morphological appearance of a galaxy in a Hubble-like scheme with three different
types: 1. early-type (ellipticals and lenticulars); 2. disk+bulge
galaxies (which we in the following will refer to as “disks”),
frequently with spiral structures, partly showing a bar; 3. latetypes/irregulars. In addition, we roughly determined the degree
of gravitational interactions in three categories: 1. undisturbed
objects; 2. galaxies with signs of mild gravitations interactions
(which we define as interactions of “class I”), e.g. lopsided
The modifications of the quiescent galaxies’ ACS images described in the following constitute the backbone of our analysis.
To ensure a direct comparability of the morphologies of AGN
hosts and non-AGN galaxies, the same kind of residuals that
occur after the subtraction of the nucleus in the central part of
an AGN host, were introduced in the quiescent galaxy images.
This was achieved by adding a synthetic nucleus to each quiescent galaxy, constructing their local PSFs and finally decomposing them into a host and nucleus component with GALFIT. For
the synthetic nuclei, we decided to use real stellar images from
GEMS, not model profiles. This way, it is ensured that the deviations between the PSF shapes of a nucleus and the corresponding
local PSF will be statistically the same as in the case of the real
AGN, and the residuals of the nucleus in simulated AGN and real
AGN images — after subtracting the best-fit model nucleus —
will have the same noise properties.
To pre-select a sample of stars that could be used as synthetic
nuclei, the GEMS catalogue of stars was constrained to the Vband magnitude range of the nuclei of the real AGN: 19.6 <
V < 25.3. Since the PSF shape also depends on the SED of
the source, and the nuclei of AGN have very blue colors, only
blue stars were selected, using upper limits (B − V) < 0.6 and
(V − R) < 0.6 (both Vega). For the four faintest stars with V >
23.7, these limits had to be relaxed due to the lack of very blue
stars in that magnitude range. Five stars were rejected because
they showed cosmics or bad pixels close to the central peak. The
final list for the creation of synthetic nuclei consisted of 58 stars
with median colors of hB − Vi ≈ 0.44 and hV − Ri ≈ 0.43 (both
Vega), respectively.
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
5
Fig. 3. Examples of galaxies from the non-AGN comparison sample. Each row displays six examples of the three different morphological types
that were denoted via visual classification. From top to bottom row: early-type galaxies; disk galaxies; late-type disks / irregulars. From left to
right, the galaxies are ordered according to their degree of morphological peculiarity: non-interacting (two leftmost columns), interactions of “class
I”, e.g. lopsided galaxies or tidal tails (two middle columns), interactions of “class II”, i.e. evidence of violent gravitational interactions and major
mergers (two rightmost columns). Thumbs will be resized to same intrinsic scale
For each of the quiescent galaxies, a star was picked from
this list such that the flux ratio between galaxy and star was the
same as the H/N ratio of the AGN to which the quiescent galaxy
initially was assigned in the construction of the comparison sample. Small normalizations of the order of a few percent were applied to the stellar fluxes to achieve exactly the same flux ratios
between a galaxy and the star on the one hand and AGN host and
nucleus on the other. After normalization, the quiescent galaxy
(synthetic “host”) and stellar image (synthetic “nucleus”) were
simply co-added with the star positioned at the galaxy’s central
coordinates according to the GALFIT modelling on the inital
image. The variance frames of the two components were also
combined to be used in the following GALFIT decompostion.
The local PSF for each synthetic AGN was created based
on the CCD position of the star that was used as a synthetic
nucleus (with that star being excluded from the local PSF construction catalogue of stars). The 35 nearest stars around a given
synthetic nucleus’ CCD position were normalized and averaged.
Local PSFs for the synthetic AGN hence have the same quality as those of the real AGN. Finally, using GALFIT, all synthetic AGN were decomposed into nucleus and Sérsic host with
the same configuration as described in the case of the real AGN
(section 3.3). The main steps of the creation and decompostion
of the synthetic AGN are illustrated in Fig. 4.
The host–to–nucleus ratios H/N of the real AGN and the
synthetic AGN are compared to each other in Fig. 5. As in the
case of the redshifts and host magnitudes, the distributions of the
two samples differ only slightly, with median values of hH/Ni =
0.60 for the simulated AGN and hH/Ni = 0.63 in case of the real
AGN.
Fig. 4. Illustration of the AGN simulation procedure. Upper left: original quiescent galaxy (COMBO-17 #938) at z = 0.52, apparent brightness V = 23.27. Upper middle: star of brightness V = 21.46 used as a
synthetic nucleus. Upper right: simulated AGN (co-added galaxy and
star) with a host-to-nucleus ratio of H/N = 0.18. Lower left: GALFIT
model of the host galaxy (from fitting a nucleus + host model to the
simulated AGN). Lower middle: recovered host galaxy after subtracting the GALFIT model of the nucleus from the simulated AGN image.
Lower right: residual after subtracting the combined galaxy + nucleus
model from the simulated AGN image. Each of the images is ∼ 2 arcsec
on the side.
4.4. On the stability of the AGN decomposition
Our synthetic AGN images offer the possibility to test the
GALFIT recovery of the main host galaxy parameters like total
brightness, effective radius and Sérsic index for a broad range
of H/N ratios. A similar test has been performed by Sanchez
6
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. 5. Distribution of the host–to–nucleus ratios H/N of the real AGN
(dashed) compared to the simulated AGN (solid). The histograms are
normalized to the same total number of objects.
et al. (2004), however in that study no real galaxy images but
GALFIT model profiles and hence perfectly symmetric host
galaxies have been used. Even though studying the robustness
of AGN host galaxy parameters derived with GALFIT is not the
raison d’étre of our analysis, it is an interesting byproduct.
The plots in Fig. 6 compare the initial host parameters (with
suffix “ini”), as derived on the pure galaxy images prior to the
addition of a synthetic nucleus, and the recovered host parameters (suffix “sim”) from simultaneous fitting of galaxy and and
additional synthetic nucleus.
In Fig. 6a, we show the total brightness vhost,ini of the quiescent galaxies (single Sérsic modelling on the original images,
before the addition of synthetic nuclei) plotted against the deviations ∆vhost = vhost,ini − vhost,sim between the initial brightness vhost,ini and the brightness vhost,sim determined from twocomponent GALFIT modelling after the addition of the synthetic
nucleus. The values of ∆vhost basically are distributed around
zero with a scatter of ∼ 0.25 mag. The scatter is mostly is independent of the galaxy magnitude except for objects brighter
than vhost,ini ≈ 21.3. It is surprising that most of these galaxies
have large positive values of ∆vhost , meaning that the fluxes of
these objects are overestimated when they are fitted simultaneously with a synthetic nucleus. It is quite obvious that this occurs
because the fitting algorithm attributes part of the nucleus’ flux
to the host, because the Sérsic indices of most of these galaxies are overestimated, too (leading to large values of the fraction
nini /nsim between Sérsic index in the original image and the simulated AGN, depicted by filled red circles).
Fig. 6b shows the deviations between original host magnitude and recovered host magnitude as a function of input
host–to–nucleus flux ratio H/Nini . The scatter in ∆vhost strongly
increases towards low values of H/Nini , which is expected.
Moreover, in the regime H/Nini < 1, i.e. in cases where the nucleus outshines the host, the recovered host magnitudes vhost,sim
tend to be too bright. This trend indicates that part of the bright
nucleus’ flux is attributed to the host component by the fitting algorithm. A similar results has been found by Kim et al. (2008).
These authors argued that mismatches between the nucleus’ profile and the input local PSF cause this systematic effect.
Fig. 6c shows the input effective radii reff,ini determined on
the original images plotted against the deviations between input
effective radius reff,ini and recovered effective radius reff,sim . This
parameter is recovered to within less than 30% for 3/4 of the
objects. Fig. 6d shows the deviations between input and recovered effective radius as a function of input host–to–nucleus ratio
H/Nini . Like the recovered host brightnesses, the recovered effective radii have a larger scatter for low host–to–nucleus ratios.
Fig. 6e displays the ratio ∆n = log(nini /nsim ) between original and recovered Sérsic index as a function of the input host–
to–nucleus ratio. It is evident that n is a parameter that is much
harder to determine than the brightness or effective radius of the
host galaxy. At small host–to–nucleus ratios, the scatter in ∆n
is approx. 0.4 dex. The recovered Sérsic indices also are on average too small. This systematic discrepancy is further demonstrated in Fig. 6f, where the ratios between initial and recovered
Sérsic index are plotted as a function of input total host magnitude vhost,ini . Towards fainter host magnitudes, the Sérsic indices are systematically underestimated in a decomposition into
nucleus and host component. For the distribution in host–to–
nucleus ratio considered here (the median for our AGN sample
is hH/Ni = 0.63), this effect starts to kick in at host magnitudes
fainter than vhost,ini ≈ 23. Due to this systematic effect, it is hard
to determine the morphological type of an AGN host straightforwardly from the best-fit parameters. A similar trend has been
found by Sanchez et al. (2004) on the basis of simulated (and
thus perfectly symmetric) host galaxies and nuclei. Those simulations utilized single component host galaxies (Sérsic profiles
with either n = 1 or n = 4). In contrast to this, many of the
host galaxies in our sample consist of two or more components
(bulge, disk, potential bar, etc.).
Fig. 7a shows the host magnitude deviations ∆vhost plotted
as a function of the Sérsic index deviations ∆n. Not surprisingly,
there is an anti-correlation of these two parameters. A too faint
model profile for the host galaxy (∆vhost < 0) is accompanied by
an underestimated Sérsic index, because a fraction of the host
galaxies’ flux is assigned to the model nucleus. Since this effects
occurs at small galactocentric radii, the host model profile becomes too shallow (vice versa in cases of an overestimated host
flux). A false host model flux directly translates into a wrong
value of the model host–to–nucleus ratio (see Fig. 7b, displaying ∆H/N as a function of the host magnitude deviations ∆vhost ).
Fig. 7c shows the fraction between the input host–to–nucleus
ratio (H/N)ini and the recovered ratio (H/N)sim , plotted as a function of (H/N)ini . Very slightly underestimated H/N for low input
values of H/N, very slightly overrestimated H/N for large input
host–to–nucleus ratios.
In Fig. 7d, the ratio between the input and recovered H/N
ratio is plotted against original total host magnitude vhost,ini . In
general, the H/N ratio is overestimated for very bright hosts —
which is a consequence of the frequently overestimated host
fluxes in this regime (cf. Fig. 6a).
The comparisons between single-component light profile
modelling of quiescent galaxies and two-component modelling
of these galaxies with added synthetic nuclei allows to give a
systematic error for the decomposition of AGN images using
GALFIT. Ignoring extreme ouliers by taking only the best 95%
of the fits into account, host fluxes, H/N ratios, effective radii
and Sérsic indices are recovered to within fractional errors of
σvhost =0.23, σH/N =0.34, and σreff =0.5, σn =0.89, respectively.
5. Morphological analysis and discussion
Prior to the following analysis, the best-fit model nuclei from the
GALFIT decomposition were subtracted from all synthetic and
real AGN images.
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
7
Fig. 6. GALFIT recovery of host galaxy parameters for the simulated AGN sample. (a): deviations between input host magnitude vhost,ini and
recovered host magnitude as a function of input host magnitude. (b): deviations between original host magnitude and recovered host magnitude
vhost,sim as a function of input host–to–nucleus flux ratio H/Nini (c): deviations between input effective radius reff,ini and recovered effective radius
reff,sim as a function of input effective radius (d): deviations between input effective radius and recovered effective radius as a function of input
host–to–nucleus flux ratio (e): deviations between input Sérsic index nini and recovered Sérsic index nsim as a function of input host–to–nucleus
flux ratio (f): deviations between input Sérsic index nini and recovered Sérsic index nsim as a function of input host magnitude.
To define a galaxies’ outer contours, the segmentation maps
created with the Source Extractor package have been used. In
others words, these maps determined which pixel of an image
belongs to which galaxy. All these maps have been visually inspected. In a few cases, e.g. when the deblending of multiple
galaxies was not satifying, Source Extractor was re-run with dif-
ferent input parameters until the segmentation maps were accurate. In particular, major merger candidates were not deblended
into the individual sub-components but analyzed as a single object. In the SExtractor configuration used here, the flux at the
“egdes” of a given galaxy corresponds to the 1 σ sky noise level.
8
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. 7. GALFIT recovery of host galaxy parameters for the simulated AGN sample. (a): deviations between input host magnitude vhost,ini and
recovered host magnitude vhost,sim as a function of the deviations between input Sérsic index nini and recovered Sérsic index nsim (b): deviations
between input host–to–nucleus ratio H/Nini and the recovered host–to–nucleus ratio H/Nsim as a function of the deviations between original host
magnitude and recovered host magnitude (c): deviations between input host–to–nucleus ratio H/Nini and the recovered host–to–nucleus ratio
H/Nsim as a function of input host–to–nucleus flux ratio (d): deviations between input host–to–nucleus ratio H/Nini and the recovered host–to–
nucleus ratio H/Nsim as a function of input host magnitude.
5.1. Definition of morphological descriptors
To quantify morphology and peculiarity, we here rely on two parameter spaces which have been frequently used in the literature,
in particular for distant galaxies. These are, on the one hand,
the concentration index C and asymmetry index A (Conselice et
al. 2000) and, on the other, the Gini coefficient G and M20 index
(Lotz et al. 2004).
The asymmetry parameter A is measured by subtracting a
galaxy image rotated by 180◦ from the original image:
X |I(i, j) − I180 (i, j)| X |B(k, l) − B180 (k, l)|
A=
−
.
(1)
|I(i, j)|
|B(k, l)|
i, j
k,l
Here, I is the original image, I180 is the image rotated by 180◦
about the galaxy’s center, and B and B180 are the background and
rotated background. This latter contribution to A due to the background noise is estimated from blank sky regions in the direct
vicinity of the object under scrutiny. The first term is summed
over all pixels within the 1σ isophotes of the galaxy. The central
pixel is determined by shifting the galaxy on a 50 × 50 pixel grid
and finding the minimum of A.
There are various definitions of the concentration index C in
the literature which only have slight differences to each other.
We adopt C as the ratio between the flux within an elliptical area
G1 and the flux within another elliptical area G2 , with these areas
having the same axis ratio and position angle, but the major axis
of G2 being 0.3 times the size of the major axis of G1 (Abraham
et al. 1996):
P
i, j∈G2 Ii j
C=P
.
(2)
k,l∈G1 Ikl
Here, Ii j is the flux at the pixel (i, j) within the inner isophotal
region G2 and Ikl is the flux at the pixel (k, l) within the larger
isophotal region G1 .
The Gini coefficient originally has been developed for applications in social sciences. When computed for galaxy images,
it is a measure of the concentration of an object’s light profile,
similar to C, but insensitive to the definition of a “central” pixel.
We adopt the Gini coefficient as given in Lotz et al. (2004):
n
G=
X
1
(2i − n − 1)| fi |
| ¯f |n(n − 1) i
(3)
where | fi | are the absolute flux values of the i-th pixel (ranked in
increasing order) and | ¯f | is the mean of the absolute flux values
of all n pixels assigned to an object.
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
The M20 index is based on the total second-order moment
Mtot , which is the flux in each pixel fi multiplied by the squared
distance to the center of the galaxy, summed over all pixels of a
given object:
Mtot =
n
X
i
Mi =
n
X
fi · ((xi − xc )2 + (yi − yc )2 )
(4)
i
where xc , yc is the galaxy’s center. The center is computed by
finding xc , yc such that Mtot is minimized. M20 then is defined as
the normalized second order moment of the brightest 20% of the
galaxy’s flux. To compute M20 , the pixels are rank-ordered by
flux, and Mi is summed over the brightest pixels until the sum of
the brightest pixels equals 20% of the total galaxy flux, followed
by a normalization to Mtot (Lotz et al. 2004):
!
P
X
Mi
M20 = log i
while
fi < 0.2 ftot .
(5)
Mtot
i
Here, ftot is the total flux of a galaxy and fi are the fluxes for
each pixel i, ordered according to their flux values such that f1
is the brightest pixel, f2 the second brightest etc. M20 traces the
spatial distribution of bright features like off-center star forming
regions.
5.2. Quantitative morphological comparison of AGN hosts
and quiescent galaxies
Our visual classifications of the comparison sample will be utilized as a calibration in the CAS and G/M20 parameter spaces.
Since these classifications have been performed prior to the inclusion of a synthetic optical nucleus, they are much more robust
than it would be feasible in the case of the nucleus-subtracted
host galaxy images of the real AGN that are affected by the nucleus’ residuals in the center of the images (cf. Fig. 6).
In Fig. 8 (left-hand plots), we show the distribution of the
quiescent galaxies in terms of C vs. A. Only the 181 objects
for which a visual classification was feasible are shown. The
top, middle and bottom graphs show the morphological types
(early-types, disks, late-type/Irr galaxies) individually. Each of
these three sub-classes is further divided into undisturbed objects
(small open symbols), interactions “class I”, i.e. midly disturbed
morphologies (large open symbols) and interactions “class II”,
i.e. galaxies with signatures of violent gravitational interactions.
From undisturbed early-types to undisturbed disks and further
to undisturbed late-type galaxies, concentration C on average decreases, while asymmetry A increases. The distinguishability between undisturbed and interacting galaxies appears to decrease
from top to bottom, i.e. from early-type galaxies to late-types
and irregulars. A two-dimensional KS test yields a probability of P = 6 · 10−5 that both sub-samples are drawn from the
same distribution (the small numbers not allow to sub-divide
interacting early-type galaxies into class I and class II here).
Comparing undisturbed disks and disks of interaction class I,
we find Pu,int1 = 0.001; undisturbed disks compared to disks of
interaction class II yield Pu,int2 = 6 · 10−6 . Undisturbed and interaction class I late/Irr galaxies do not satisfy the null hypothesis
(Pu,int1 = 0.38), only undisturbed and interaction class II late/Irr
galaxies are well-separated with Pu,int2 = 5 · 10−5 .
In G/M20 space (right-hand plots of Fig. 8 ), the separation
between the sub-classes seems to be slightly worse than in CAS
space. This is justified by the KS tests. Undisturbed and interaction class I/II early-types have a probability P = 0.017 to be
drawn from the same parent distribution. Undisturbed disks and
9
disks of interaction class I yield Pu,int1 = 0.077, for undisturbed
disks and disks of interaction class II we find Pu,int2 = 0.001.
Like for CAS, undisturbed and interaction class I late/Irr galaxies are hardly distinguished (Pu,int1 = 0.11), undisturbed and interaction class II late/Irr objects show a slightly better separation with Pu,int2 = 0.04. One additional aspect we would like to
point out is that, on the basis of our quisecent sample, we do not
find a separate area characterized by high G and low M20 values
that is populated mainly by major merger candidates (cf. Lotz et
al. 2004).
The distributions in CAS and G/M20 are completely different in the presence of an optical nucleus. In Fig. 9, we show the
of simulated AGN (quiescent galaxies plus synthetic nucleus)
and the real AGN including their optical nuclei. Note that the
displayed ranges in C, G and M20 are slightly different from
Fig. 8. Compared to the “pure” quiescent galaxies, the synthetic
AGN have moved to higher values for the concentration index
and Gini coefficient, which is expected after the addition of a
nuclear point source. The M20 indices become smaller, indicating the loss of “information” because the combined light profile of host+nucleus is smoother than that of the host alone. The
mean asymmetry of all types (median value hAisim = 0.30) is
almost unchanged with respect to the original galaxy images
(hAiorig = 0.27). This is somewhat surprising because the nuclear point sources should make the combined light profile less
elliptical. However, the synthetic nucleus is usually not perfectly
centered on the pixel grid, and since the asymmetry index is minimized only on an integer pixel shift basis, even an almost symmetric light profile can yield an intermediate value for A. One
other net effect of the nuclei is that the distribution in A became
narrower.
The distinguishability between different morphological subclasses (for the simulated AGN hosts) is, when compared to
the initial distributions in Fig. 8, strongly degraded. The various
types are mixed with a high level of “noise”. Repeating the KS
test to quantify how well undisturbed and interacting galaxies are
separated, we find the following in CAS space: for early-types
P = 0.031 (both interaction classes merged due to the small
number of objects); for disks Pu,int1 = 0.066 and Pu,int2 = 0.008;
for late-types/irregulars Pu,int1 = 0.15 and Pu,int2 = 0.018. With
the same test in G/M20 , the respective probabilities for the null
hypothesis are P = 0.005 for early-types; Pu,int1 = 0.14 and
Pu,int2 = 0.14 for disks; Pu,int1 = 0.34 and Pu,int2 = 0.25 for
late-types/irregulars.
Simulated and real AGN cover similar regions in the parameter spaces. Before we perform a quantitative comparison, we
will go one step further. We stress again that previous analyses
of AGN hosts that utilized CAS or G/M20 have not gone beyond
the situation illustrated in Fig. 9.
For the next computation of C, A, G and M20 , the best-fit
nuclei from two-component modelling with GALFIT were subtracted from all (simulated and real) AGN. The resulting distributions are shown in Fig. 10. In C/A space (left plot), the visually derived morphological types are better separated than in
the presence of a nucleus (cf. Fig. 9). However, overlap between
the individual types of the simulated hosts is clearly worse than
based on the original images (cf. Fig. 8). Again we will first
use a KS test to quantify the distinction between undisturbed
and disturbed galaxies. In C/A space, the result is for early-types
P = 2.4 · 10−4 ; for disks Pu,int1 = 0.65 and Pu,int2 = 0.018;
for late-types/irregulars Pu,int1 = 0.22 and Pu,int2 = 0.039. In
G/M20 space, the respective probabilities for the null hypothesis are P = 1.3 · 10−3 for early-types; Pu,int1 = 0.64 and
10
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. 8. Distribution of the quiescent galaxies in concentration index C versus asymmetry A (left) and in Gini coefficient G versus M20 (right). Top
row: early-type galaxies, middle row: disks, bottom row: late-type and irregular galaxies. Small symbols depict undisturbed galaxies, large open
symbols show objects of interaction class I, large filled symbols are galaxies of interaction class II.
Pu,int2 = 0.083 for disks; Pu,int1 = 0.54 and Pu,int2 = 0.16 for
late-types/irregulars.
Some objects feature asymmetry indices above unity. This
indicates strong additional noise in the center of the images that
is not related to the sky background but residuals from the nucleus subtraction. As will be discussed further below, this effect
particularly comes into play for objects with low H/N ratios.
We will use the KS test now to determine which subsample of the quiescent galaxies (more precisely, the nucleus-
subtracted, synthetic AGN hosts) is most similar to the (real)
AGN hosts. Comparing the AGN to undisturbed early-types,
early-types of interaction class I+II, undisturbed disks, disks of
interaction class I, disks of interaction class II, late-types of interaction class I, late-types of interaction class II, we find in
CAS space: Pe,undist = 0.011, Pe,int1/2 = 0.041, Pd,undist = 0.38,
Pd,int1 = 0.18, Pd,int2 = 0.022, Pl,undist = 0.17, Pl,int1 = 0.009,
Pl,int2 = 0.001. Undisturbed disks hence show the highest probability, followed by disks of interaction class I. The same com-
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
11
Fig. 9. Distribution of the simulated AGN (quiescent galaxies with added synthetic nucleus) and real AGN host galaxies, in concentration index
C versus asymmetry A (left) and in Gini coefficient G versus M20 (right). Note that the axes partly show different ranges with respect to Fig. 8.
parison in G/M20 space yields Pe,undist = 0.001, Pe,int1/2 = 0.43,
Pd,undist = 0.24, Pd,int1 = 0.39, Pd,int2 = 0.11, Pl,undist = 0.046,
Pl,int1 = 0.007, Pl,int2 = 0.001. Here, interacting early-types
show the highest probability, followed by disks of interaction
class I.
The impact of an optical nucleus on the location of a galaxy
in C/A and G/M20 space in further illustrated in Fig. 11. Each
line in these plots connects the initial position of a quiescent
galaxy in the respective morphological parameter space to its location after simulating and subtracting a nucleus (indicated by a
small circle). The impact of the nucleus mostly is independent
from the visually determined degree of peculiarity. Only in the
CAS system, strongly interacting disk galaxies and, in particular,
strongly interacting late-types/irregulars are less sensitive to an
optical nucleus than undisturbed or weakly interacting galaxies.
This is not a surprise since these objects show pronounced, extended morphological peculiarities. More importantly, based on
these trends we argue that strong morphological disturbances
of real AGN hosts — if present — should be detectable in CAS
space with our data.
12
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. 10. Distribution of the simulated and real AGN host galaxies, after the subtraction of the best-fit model nucleus, in concentration index C
versus asymmetry A (left) and in Gini coefficient G versus M20 (right)
One parameter that is of major importance for the impact of
the optical nucleus is, of course, the contrast between the host
and the central point source. This is demonstrated in Fig. 12.
For high H/N ratios, A keeps relatively constant, C is mostly
becoming smaller. Towards smaller H/N values, there is no systematic trend in C anymore, but A is increased due to the artifical noise from the residuals. In G/M20 , low-H/N objects show
rather random shifts, which towards intermediate values of host–
to–nucleus constrast become more systematic in the sense that
they are diagonal (with G decreasing and M20 increasing or vice
versa). For high H/N, there is a general trend that G mostly is
decreasing with respect to its initial value and M20 is increasing.
As a test, we checked whether the S /N in the descriptor
spaces could be enhanced by masking the innermost parts of
the galaxies which, after the subtraction of the nucleus, are affected by residuals. For the analysis in C/A space, we alternatively used the definition of C given in Conselice et al. (2000):
C = 5 log(R20 /R80 ), where R20 and R80 are the radii enclosing
20% and 80% of an object’s total flux. Since this concept is not
based on fixed radii, it is more reliable than Eq. 2 for usage on
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
13
Fig. 11. Illustration of the impact of the optical nucleus on the quantitative morphologies of the simulated AGN galaxies in concentration index C
versus asymmetry A (left) and in Gini coefficient G versus M20 (right). Each vector depicts where a given galaxy was located originally, i.e. before
the inclusion of a synthetic nucleus, and after the addition and subtraction of that nucleus (indicated by a circle). For reasons of clarity, only a
fraction of the respective sub-samples is shown.
galaxy images that include a mask at small radii. We compared
the results for masks of various diameters (2, 3 or 4 pixels), but
the sensitivity distinction between the various sub-types was on
average lower than on the basis of Eq. 2 without any masking
of the inner parts of the objects. We give as an example the results of the KS test using a mask of 2 pixel radius. In C/A space,
the probability for that disturbed and undisturbed early-types are
drawn from the same distribution is P = 0.003; for disks we
find Pu,int1 = 0.19 and Pu,int2 = 0.01; for late-types/irregulars
Pu,int1 = 0.41 and Pu,int2 = 0.27. In G/M20 space, the respective
probabilities for the null hypothesis are P = 0.04 for early-types;
Pu,int1 = 0.79 and Pu,int2 = 0.062 for disks; Pu,int1 = 0.1 and
Pu,int2 = 0.27 for late-types/irregulars.
14
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. 12. Illustration of the impact of the optical nucleus on the quantitative morphologies of the simulated AGN galaxies in concentration index C
versus asymmetry A (left) and in Gini coefficient G versus M20 (right). Each vector depicts where a given galaxy was located originally, i.e. before
the inclusion of a synthetic nucleus, and after the addition and subtraction of that nucleus (indicated by a circle). For reasons of clarity, only
a quarter of the sample is shown. Line types denote three different regimes in host–to–nucleus ratio, each holding roughly the same number of
objects: H/N < 0.35 (dotted), 0.35 < H/N < 1 (thin solid) and H/N > 1 (thick solid).
5.3. Discussion
Our morphological comparison does not yield evidence that the
majority of intermediate-z AGN hosts are galaxies that are undergoing major mergers. This is at variance with straightforward
expectation in the framework of a merger-AGN connection. We
want to address three possible explanations in the following:
1. The morphological descriptors like CAS or G/M20 might
miss a significant fraction of strong gravitational interactions on
single-orbit ACS data of intermediate-redshift galaxies;
2. there might be a significant delay between a major merger and
the onset of the AGN phase, which in turn could lower the probability to detect merger signatures in a sample of AGN hosts, or
3. violent gravitational interactions might not be the main triggering mechanism of nuclear activity at the redshift regime considered here, which corresponds to look-back times between ∼ 5
and ∼ 8 Gyr.
Regarding the first of these possibilities, our analysis shows
that, at least in the CAS system, major mergers should be detectable at a high success rate, even in the presence of an optical nucleus (cf. Fig. 10). Recall that this statement refers to major merger candidates that have been visually identified in nonAGN galaxies. Our test thus goes beyond using real AGN hosts
to compare the visually derived merger fraction to that found
with morphological descriptors. However, even if major mergers are triggering the nuclear activity for a significant fraction of
our sample, it is not necassarily the case that the morphological imprint of the merger is still detectable by the time the AGN
phase has started. The start of the SMBH accretion and the time
of the strongest morphological perturbations (hence the best “detectability” of the merger) might not occur simultaneously.
Observational evidence for a delay between a merger and
the onset of an AGN phase has been found e.g. by Schawinski
et al. (2007) and Bennert et al. (2008). Synthetic observations to
estimate the time during which a major merger could be identified have been performed by Conselice (2006; using the CAS
space) and Lotz et al. (2008; based on CAS and G/M20 ). The involved time scales depend on a large number of parameters, like
the mass ratios, gas and dust content, or the geometry and sense
of rotation of the merging galaxies. The reported times of the
strongest perturbations, in terms of CAS or G/M20 , range from a
few 100 Myr to & 1 Gyr after the first encounter. However, these
synthetic simulations did not include the impact of an AGN’s
optical nucleus on the image analysis, nor did they attempt to
estimate the time when the nuclear accretion starts.
Concerning the third hypothesis given above, it can be argued that, at least at higher redshifts and luminosities, there is
both observational evidence and numerical simulations which
backup a connection between major mergers and AGN phases.
Ultraluminous infrared galaxies (ULIRGS), e.g., which are
mostly found in merging systems (e.g. Sanders & Mirabel 1996),
show — with respect to the overall population of luminous
galaxies — a strongly enhanced probability to contain an AGN
(Canalizo & Stockton 2001). Numerical simulations have shown
that major mergers can efficiently transport gas to the central
parts of the merger remnant to feed the central SMBH: e.g. Di
Matteo et al. (2005), Hopkins et al. (2008). At later stages, the
feedback by the AGN can terminate the star formation; this scenario might be a necessity to produce massive, red and dead
ellipticals observed in the low-z universe (e.g. Khalatyan et
al. 2008).
In the low-redshift universe, on the other hand, studies utilizing very large samples drawn from the Sloan Digital Sky Survey
did not find evidence for an increaseed fraction of interacting
galaxies among AGN hosts (e.g. Kauffmann et al. 2003, Li et
al. 2008).
Our results would be in compliance with weak gravitational
interactions (fly-bys, minor mergers) playing an important role
in AGN triggering at the cosmics epoch under scrutiny. Such
processes can induce the transport of gas with low angular momentum to the central parts of the host to feed the SMBH (ref);
the overall host morphology would not be strongly affected. The
comparison between simulated and real AGN in CAS and G/M20
space showed that a significant fraction of the hosts could be described as midly disturbed morphologies. To quantify this fraction is hard (use the new blind visual classifications here).
Moreover, our results show that weak interactions of AGN hosts
are much harder to detect than strong ones, i.e. the fraction of
midly disturbed host morphologies might be underestimated.
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
6. Summary
to be included . . .
Acknowledgements. AB acknowledges funding by the Deutsches Zentrum für
Luft- und Raumfahrt (50 OR 0404); EFB and KJ are supported by the Emmy
Noether Programme of the Deutsche Forschungsgemeinschaft; CW by a PPARC
Advanced Fellowship; MB and EvK acknowledge support by the Austrian
Science Foundation FWF under grant P18416; MEG by an Anne McLaren
Research Fellowship; CH by an European Commission Programme 6th framework Marie Cure Outgoing International Fellowship under contract MOIF-CT2006-21891, and a CITA National fellowship; SJ by NASA under LTSA Grant
NAG5-13063 and NSF under AST-0607748; DHM by NASA under LTSA Grant
NAG5-13102; CYP by STScI and NRC-HIA Fellowship programs; SFS by
the Spanish MEC grants AYA2005-09413-C02-02 and the PAI of the Junta de
Andalucı́a as research group FQM322. Support for STAGES was provided by
NASA through GO-10395 from STScI operated by AURA under NAS5-26555.
References
Abraham, R.G., Tanvir, N.R., Santiago, B.X., et al. 1996, MNRAS, 279, L47
Bennert, N, Canalizo, G., Jungwiert, B., Stockton, A., Schweizer, F., Peng, C.Y.,
Lacy, M. 2008, ApJ 677, 846
Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393
Böhm, A., Wisotzki, L., & GEMS team 2007, ASP Conf. Ser. 379, Metcalfe, N.
& Shanks, T. (eds.), 185
Borch, A., Meisenheimer, K., Bell, E.F., et al. 2006, A&A, 453, 869
Caldwell, J.A.R., McIntosh, D.H., Rix, H.-W., et al. 2008, ApJS, 174, 136
Canalizo, G., & Stockton, A. 2001, ApJ, 555, 719
Conselice, C.J., Bershady, M.A., & Jangren, A. 2000, ApJ, 529, 886
Conselice, C.J. 2006, ApJ, 638, 686
Di Matteo, T., Springel, V., & Hernquist, L. 2005, Nature, 433, 604
Ferrarese, L., & Merritt, D. 2000, ApJL, 539, 9
Gabor, J.M., Impey, K., Jahnke, K., et al. 2008, ApJ, accepted (arXiv:0809.0309)
Gebhardt K., Bender, R., Bower, G., et al. 2000, ApJL, 539, 13
Grogin, N.A., Conselice, C.J., Chatzichristou, E., et al. 2005, ApJ, 627, 97
Gray, M.E., et al., MNRAS, in prep.
Häring, N., & Rix, H.-W. 2004, ApJ, 604, 89
Häussler, B., McIntosh, D.H., Barden, M., et al. 2007, ApJS, 172, 615
Heymans, C., Gray, M.E., Peng, C.Y., et al. 2008, MNRAS, 385, 1431
Hopkins, P.F., Hernquist, L., Cox, T.J., & Keres, D. 2008, ApJS, 175, 356
Jahnke, K., Sanchez, S.F., Wisotzki, L., et al. 2004, ApJ, 614, 568
Kauffmann, G., Heckman, T.M., Tremonti, C., et al. 2003, MNRAS, 346, 1055
Khalatyan, A., Cattaneo, A., Schramm, M., Gottlöber, S., Steinmetz, M., &
Wisotzki, L. 2008, MNRAS, 387, 13
Kim, M., Ho, L.C., Peng, C.Y., Barth, A.J., Im, M. 2008, ApJS, accepted
(arXiv:0807.1334)
Lehmer, B., Brandt, W.N., Alexander, D.M., et al. 2005, ApJS, 161, 21
Li, C., Kauffmann, G., Heckman, T.M., White, S. D. M., Jing, Y.P. 2008, MRAS,
385, 1915
Lotz, J., Primack, J., & Madau, P. 2004, AJ, 128, 163
Lotz, J., Jonsson, P., Cox, T.J., & Primack, J. 2008, MNRAS, accepted
(arXiV:0805.1246)
Marconi, A., & Hunt, L.K. 2003, ApJ, 589, 21
Peng, C.Y., Ho, L.C., Impey, C.D., & Rix, H.-W. 2002, AJ, 124, 266
Pierce, C.M., Lotz, J.M., Laird, E.S., et al. 2007, ApJ, 660, 19
Rix, H.-W., Barden, M., Beckwith, S.V.W., et al. 2004, ApJS, 152, 163
Sanchez, S.F., Jahnke, K., Wisotzki, L., et al. 2004, ApJ, 614, 586
Schawinski, K., Thomas, D., Sarzi, M., et al. 2007, MNRAS, 382, 1415
Sérsic, J.L., 1968, Atlas de Galaxias Australes (Córdoba: Obs. Astron. Uni.
Nacional Cordoba),1
Silverman, J.D., Mainieri, V., Lehmer, B.D., et al. 2008, ApJ, 675, 1025
Simmons, B.D., & Urry, C.M. 2008, ApJ, 683, 644
Wolf, C., Meisenheimer, K., Rix, H.-W., Borch, A., Dye, S., & Kleinheinrich,
M. 2003, A&A, 401, 73
Wolf, C., Meisenheimer, K., Kleinheinrich, M., et al. 2004, A&A, 421, 913
Appendix A: Additional plots
15
16
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. A.1. Distribution of the simulated and real AGN host galaxies, after the subtraction of the best-fit model nucleus, in concentration index C
versus asymmetry A (left) and in Gini coefficient G versus M20 (right)
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. A.2. Same as Fig. 10; using Knud’s classifications on the basis of the original images of the quiescent galaxies.
17
18
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. A.3. Same as Fig. 10; using Eric’s classifications on the basis of the original images of the quiescent galaxies.
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
19
Fig. A.4. Distribution of the quiescent galaxies in concentration index C versus asymmetry A (left) and in Gini coefficient G versus M20 (right).
Top row: early-type galaxies, middle row: disks, bottom row: late-type and irregular galaxies. Small symbols depict undisturbed galaxies, large
open symbols show objects of interaction class I, large filled symbols are galaxies of interaction class II.
20
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. A.5. Same as Fig. 8; using Knud’s classifications on the basis of the original images of the quiescent galaxies.
A. Böhm et al.: AGN host galaxies at redshift z ≈ 0.7: peculiar or not?
Fig. A.6. Same as Fig. 8; using Eric’s classifications on the basis of the original images of the quiescent galaxies.
21