Kiloparsec-scale Spatial Offsets in Double-peaked Narrow-line Active Galactic Nuclei. I. Markers for Selection of Compelling Dual Active Galactic Nucleus Candidates

Abstract

Merger-remnant galaxies with kpc-scale separation dual active galactic nuclei
(AGNs) should be widespread as a consequence of galaxy mergers and triggered
gas accretion onto supermassive black holes, yet very few dual AGNs have been
observed. Galaxies with double-peaked narrow AGN emission lines in the Sloan
Digital Sky Survey are plausible dual AGN candidates, but their double-peaked
profiles could also be the result of gas kinematics or AGN-driven outflows and
jets on small or large scales. To help distinguish between these scenarios, we
have obtained spatial profiles of the AGN emission via follow-up long-slit
spectroscopy of 81 double-peaked narrow-line AGNs in SDSS at 0.03 < z < 0.36
using Lick, Palomar, and MMT Observatories. We find that all 81 systems exhibit
double AGN emission components with ~kpc projected spatial separations on the
sky, which suggests that they are produced by kpc-scale dual AGNs or kpc-scale
outflows, jets, or rotating gaseous disks. In addition, we find that the
subsample (58%) of the objects with spatially compact emission components may
be preferentially produced by dual AGNs, while the subsample (42%) with
spatially extended emission components may be preferentially produced by AGN
outflows. We also find that for 32% of the sample the two AGN emission
components are preferentially aligned with the host galaxy major axis, as
expected for dual AGNs orbiting in the host galaxy potential. Our results both
narrow the list of possible physical mechanisms producing the double AGN
components, and suggest several observational criteria for selecting the most
promising dual AGN candidates from the full sample of double-peaked narrow-line
AGNs. Using these criteria, we determine the 17 most compelling dual AGN
candidates in our sample.

Full text

Page 1

arXiv:1111.2862v1 [astro-ph.CO] 11 Nov 2011
Submitted for Publication in ApJ
Preprint typeset using LATEX style emulateapj v. 03/07/07
KILOPARSEC-SCALE SPATIAL OFFSETS IN DOUBLE-PEAKED NARROW-LINE AGNS. I. MARKERS FOR
SELECTION OF COMPELLING DUAL AGN CANDIDATES
Julia M. Comerford1,9, Brian F. Gerke2,10, Daniel Stern3, Michael C. Cooper4,11, Benjamin J. Weiner5,
Jeffrey A. Newman6, Fiona Harrison7, Kristin Madsen7, and R. Scott Barrows8
1Astronomy Department, University of Texas at Austin, Austin, TX 78712
2Kavli Institute for Particle Astrophysics and Cosmology, M/S 29, Stanford Linear Accelerator Center,
2575 Sand Hill Road, Menlo Park, CA 94725
3Jet Propulsion Laboratory, California Institute of Technology, MS 169-221, 4800 Oak Grove Drive, Pasadena, CA 91109
4Center for Galaxy Evolution, Department of Physics and Astronomy, University of California, Irvine, 4129 Frederick Reines Hall,
Irvine, CA 92697
5Steward Observatory, University of Arizona, 933 N. Cherry Avenue, Tucson, AZ 85721
6Pittsburgh Particle Physics, Astrophysics, and Cosmology Center, Department of Physics and Astronomy, University of Pittsburgh,
Pittsburgh, PA 15260
7Space Radiation Laboratory, California Institute of Technology, MS 105-24, Pasadena, CA 91125 and
8Arkansas Center for Space and Planetary Sciences, University of Arkansas, Fayetteville, AR 72701
Submitted for Publication in ApJ
ABSTRACT
Merger-remnant galaxies with kpc-scale separation dual active galactic nuclei (AGNs) should be
widespread as a consequence of galaxy mergers and triggered gas accretion onto supermassive black
holes, yet very few dual AGNs have been observed. Galaxies with double-peaked narrow AGN emission
lines in the Sloan Digital Sky Survey are plausible dual AGN candidates, but their double-peaked pro-
files could also be the result of gas kinematics or AGN-driven outflows and jets on small or large scales.
To help distinguish between these scenarios, we have obtained spatial profiles of the AGN emission
via follow-up longslit spectroscopy of 81 double-peaked narrow-line AGNs in SDSS at 0.03 < z < 0.36
using Lick, Palomar, and MMT Observatories. We find that all 81 systems exhibit double AGN emis-
sion components with ∼kpc projected spatial separations on the sky (0.2 h−1
kpc; median ∆x = 1.1 h−1
70kpc), which suggests that they are produced by kpc-scale dual AGNs or
kpc-scale outflows, jets, or rotating gaseous disks. Further, the objects split into two subpopulations
based on the spatial extent of the double emission components and the correlation between projected
spatial separations and line-of-sight velocity separations. These results suggest that the subsample
(58+5
−6%) of the objects with spatially-compact emission components may be preferentially produced
by dual AGNs, while the subsample (42+6
−5%) with spatially-extended emission components may be
preferentially produced by AGN outflows. We also find that for 32+8
emission components are preferentially aligned with the host galaxy major axis, as expected for dual
AGNs orbiting in the host galaxy potential. Our results both narrow the list of possible physical
mechanisms producing the double AGN components, and suggest several observational criteria for
selecting the most promising dual AGN candidates from the full sample of double-peaked narrow-line
AGNs. Using these criteria, we determine the 17 most compelling dual AGN candidates in our sample.
Subject headings: galaxies: active – galaxies: interactions – galaxies: nuclei
70kpc < ∆x < 5.5 h−1
70
−6% of the sample the two AGN
1. INTRODUCTION
It is now established that most bulge-dominated galax-
ies host central supermassive black holes (SMBHs) and
that galaxies frequently merge with one another.
a natural consequence there must exist a subpopula-
tion of galaxies that harbor two SMBHs, which are ob-
servable if they are accreting gas as active galactic nu-
clei (AGNs). The kpc-scale separation dual AGNs are
an intermediate evolutionary stage between the galaxy
pairs separated by tens of kpc used as proxies for the
merger rate (e.g., Lin et al. 2008; Ellison et al. 2008;
de Ravel et al. 2011) and the sub-pc separation binary
SMBHs (e.g., Eracleous et al. 2011) that are expected to
produce gravitational waves upon coalescence. Although
there are thousands of known galaxy pairs with tens of
As
9NSF Astronomy and Astrophysics Postdoctoral Fellow
10Current address: Lawrence Berkeley National Lab, 1 Cyclotron
Road, MS 90-4000, Berkeley, CA 94720
11Hubble Fellow
kpc separations, only a handful of kpc-scale dual AGNs
(Komossa et al. 2003; Hudson et al. 2006; Bianchi et al.
2008; Koss et al. 2011; Fu et al. 2011c), one 100-pc-scale
AGN pair (Fabbiano et al. 2011), and one 10-pc-scale
AGN pair (Rodriguez et al. 2006) have been confirmed.
The small number of confirmed dual AGNs is in con-
flict with expectations based on the galaxy merger rate
and the triggering of gas accretion onto SMBHs dur-
ing mergers (e.g., Springel et al. 2005; Hopkins et al.
2005; Van Wassenhove et al. 2011).
eries of individual dual AGN systems continue (e.g.,
Comerford et al. 2009b; Barrows et al. 2011), a system-
atic survey of dual AGNs is now necessary to close the
gap between the observed and expected numbers.
Such a systematic survey of dual AGNs would shed
light on drivers of galaxy evolution such as galaxy merg-
ers and SMBH growth. As direct observational tracers
of galaxy mergers, dual AGNs can be used to estimate
the galaxy merger rate. Further, the number of dual
SMBHs powering AGNs can set constraints on the acti-
Although discov-

Page 2

2 Comerford et al.
vation probability of SMBHs in merging galaxies, which
is a key parameter in models of the growth of SMBHs by
gas accretion during galaxy mergers.
With the advent of large spectroscopic surveys of
galaxies, it is now possible to conduct a systematic cen-
sus of dual AGNs. Two dual AGN candidates were
identified in the DEEP2 Galaxy Redshift Survey based
on their longslit spectra, which revealed two compo-
nents of AGN-fueled [O III] λ5007 emission separated by
∼kpc spatially on the sky and a few hundred km s−1in
line-of-sight velocity (Gerke et al. 2007; Comerford et al.
2009a).The velocity separations manifest as double-
peaked emission lines in the one-dimensional spectra.
However, the extreme rarity of double-peaked AGNs de-
mands a larger survey for statistical studies.
Recently, 340 unique AGNs with double-peaked
[O III] λ5007 were identified in the Sloan Digital Sky
Survey (SDSS) at 0.01 < z < 0.7 (Wang et al. 2009;
Liu et al. 2010b; Smith et al. 2010). These are plausi-
ble dual AGN candidates, but velocity offsets in AGN
emission lines can also be produced by narrow-line region
(NLR) effects such as small-scale (∼
ics and outflows (e.g., Das et al. 2005, 2006; Smith et al.
2011) or larger-scale AGN outflows (e.g., Crenshaw et al.
2010; Fischer et al. 2011).
carry insufficient spatial information, spatially resolved
follow-up slit spectroscopy is required to determine both
whether the double-peaked emission lines come from spa-
tially distinct regions and what the physical scales of the
regions are. These observations will help constrain which
physical mechanisms are producing the double-peaked
AGN lines and which objects are good candidates for
dual AGNs.
We have obtained follow-up longslit observations of 81
(or one-fourth) of the double-peaked narrow-line AGNs
in SDSS using a combination of the Kast Spectrograph
at Lick Observatory, the Double Spectrograph at Palo-
mar Observatory, and the Blue Channel Spectrograph at
MMT Observatory. Since the orientation of the double
emission components on the sky is unknown, we observe
each object at two position angles, generally separated
by 90◦, so that we can determine the full spatial separa-
tion of the two emission components on the sky as well
as their position angle. We use this information, as well
as the spatial extent of the emission and existing multi-
wavelength observations of the host galaxies, to constrain
what mechanisms (gas rotation, AGN outflows and jets,
or dual AGNs) on what scales (∼
kpc) produce the line profiles, and to identify which ob-
jects are the best candidates for dual AGNs.
We assume a Hubble constant H0= 70 km s−1Mpc−1,
Ωm= 0.3, and ΩΛ= 0.7 throughout, and all distances
are given in physical (not comoving) units.
<100 pc) gas kinemat-
Since SDSS fiber spectra
<100 pc, ∼kpc, or ∼10
2. THE SAMPLE AND OBSERVATIONS
2.1. The Sample
We selected targets from the 340 unique active galaxies
in SDSS with double-peaked [O III] λ5007 emission lines
presented in the catalogs of Wang et al. (2009), Liu et al.
(2010b), and Smith et al. (2010) (we note that one of the
objects in Smith et al. 2010 was originally identified in
Xu & Komossa 2009). These AGNs were selected using
standard emission line diagnostics (Baldwin et al. 1981;
Kewley et al. 2001). Wang et al. (2009) select 87 double-
peaked Type 2 AGNs that have blueshifted and red-
shifted components of [O III] λ5007, relative to the host
galaxy redshifts, that have a wavelength difference ≥ 1˚ A
and a flux ratio between the two components of 0.3 – 3;
Liu et al. (2010b) select 167 double-peaked Type 2 AGNs
by eye, through visual identification of systems that
have similar double-peaked profiles in both [O III] λ5007
and [O III] λ4959; Smith et al. (2010) select 86 double-
peaked Type 1 AGNs and 62 double-peaked Type 2
AGNs through visual selection of systems with double
peaks in both [O III] λ5007 and [O III] λ4959. After ac-
counting for the overlaps between these samples, there
are 340 unique double-peaked AGNs. We drew our tar-
gets from these 340 objects.
Since our aim was to use follow-up longslit spec-
troscopy of these objects to measure ∼kpc spatial off-
sets between the [O III] λ5007 emission components, we
selected targets as double-peaked AGNs that had 1) air-
masses below 1.5 during our observing runs, 2) red-
shifts where∼
<kpc spatial offsets were resolvable with
the spectrograph pixel scale, and 3) sufficiently bright
r-band magnitudes to minimize exposure times and max-
imize the number of objects observed in this campaign.
The latter two criteria are spectrograph- and telescope-
dependent, and we selected targets for each telescope
based on the spectrograph pixel size and the aperture
of the telescope as described below. We made no selec-
tions on any other galaxy properties.
We obtained follow-up slit spectroscopy for a sample of
81 double-peaked AGNs at 0.03 < z < 0.36 and 14.5 <
r < 18.5.
2.2. Observations
We obtained longslit spectroscopy of the 81 double-
peaked AGNs using the Kast Spectrograph at the Lick
3-m telescope (pixel size 0.′′78), the Double Spectrograph
(DBSP) on the Palomar 5-m Hale telescope (pixel size
0.′′47 for the red detector and 0.′′39 for the blue detector),
and the Blue Channel Spectrograph on the MMT 6.5-m
telescope (pixel size 0.′′29). We optimized the aperture
size and pixel size of each telescope-spectrograph combi-
nation to observe the lower redshift and brighter objects
with Lick/Kast, the medium redshift and brightness ob-
jects with Palomar/DBSP, and the higher redshift and
fainter objects with MMT/Blue Channel. We note that
since we resolve each double-peaked AGN’s two emission
components in velocity space, it is possible to measure
their positional centroids to much better than the seeing
limit and that the limiting angular scale is instead the
pixel scale.
We observed subsamples of our target list as follows:
16 objects at 0.04 < z < 0.16 (median z = 0.08, where
1′′corresponds to 1.5 h−1
70kpc) and 14.5 < r < 17.5
with Lick/Kast; 16 objects at 0.06 < z < 0.15 (me-
dian z = 0.11, where 1′′corresponds to 2.0 h−1
and 15.0 < r < 18.2 with Palomar/DBSP; and 49 ob-
jects at 0.03 < z < 0.36 (median z = 0.15, where 1′′
corresponds to 2.6 h−1
70kpc) and 14.5 < r < 18.5 with
MMT/Blue Channel. Table 1 summarizes the observa-
tions. We used a 1200 lines mm−1grating with each tele-
scope/spectrograph, centered such that the wavelength
range spanned Hβ and [O III] for the redshift range of
70kpc)

Page 3

Kiloparsec-scale Spatial Offsets in Double-peaked Narrow-line AGNs3
TABLE 1
Summary of Observations
SDSS Designation Telescope/InstrumentObservation Date (UT)
θ1 (◦)
θ2 (◦)Exposure Time (s)
SDSS J000249.07+004504.8
SDSS J000656.85+154847.9
SDSS J000911.58−003654.7
SDSS J011659.59−102539.1
Lick/Kast
MMT/Blue Channel
MMT/Blue Channel
MMT/Blue Channel
2009 August 16 / 2009 August 17
2010 November 5
2010 November 5
2010 November 4
69.4
44.4
67.0
28.8
158.0
134.4
157.0
118.8
3600
1080
1080
1080
Note. — We observed each object at two position angles, θ1 and θ2 (given in degrees east of north), and the exposure time
given is for each position angle.
(This table is available in its entirety in a machine-readable form in the online journal. A portion is shown here for guidance
regarding its form and content.)
TABLE 2
Summary of Measurements
SDSS
Name
RedshiftSpec.
Type
∆v
∆x
(′′)
∆x
70kpc)
Phys.
Extent
θsky
(◦)
∆θ
(◦)
e
Other Obs.Reference
(km s−1)(h−1
0002+0045
0006+1548
0009−0036
0116−1025
0.087
0.125
0.073
0.150
2
2
2
2
511 ± 3
359 ± 3
304 ± 5
288 ± 5
0.58 ± 0.05
0.37 ± 0.02
0.25 ± 0.01
1.02 ± 0.06
0.95 ± 0.08
0.84 ± 0.05
0.35 ± 0.01
2.68 ± 0.16
c
e
c
c
63.9 ± 5.0
172.6 ± 3.9
56.2 ± 1.1
117.0 ± 2.6
46.3
51.8
79.1
1.8
0.106
0.040
0.188
0.151
NIR/VLBA/AO
AO
NIR
NIR
1, 2, 4, 5, 6
2, 6
1, 4
1, 4
Note. — Column 3 shows the spectral type of the AGNs. Column 4 shows the velocity offset between the [O III] λ5007 peaks in the SDSS
spectrum. Column 5 shows the angular projected spatial offset between the two [O III] λ5007 emission features on the sky, as measured from
our slit spectroscopy. Column 6 shows the physical projected spatial offset. Column 7 shows whether the double AGN emission features in
the two-dimensional spectrum appear compact (c) or extended (e) in physical extent. Column 8 shows the position angle between the two
[O III] λ5007 emission features on the sky, measured from our slit spectroscopy in degrees east of north. Column 9 shows the absolute value
of the difference between the measured position angle and the isophotal position angle of the major axis of the object from SDSS r-band
photometry (with the same error as in Column 8). Column 10 shows the ellipticity of the object from SDSS photometry. Column 11 lists
the other observations of the object from near infrared (NIR), Very Long Baseline Array (VLBA), and adaptive optics (AO). References: (1)
Liu et al. (2010b), (2) Wang et al. (2009), (3) Smith et al. (2010), (4) Shen et al. (2011), (5) Tingay & Wayth (2011), (6) Fu et al. (2011b).
(This table is available in its entirety in a machine-readable form in the online journal. A portion is shown here for guidance regarding its form
and content.)
each night’s subsample of targets.
We observed each target twice, with the slit at two
different position angles, in order to determine the orien-
tation of the AGN emission components on the plane of
the sky. We typically made observations at the isophotal
position angle of the major axis of the object in SDSS r-
band photometry and the corresponding orthogonal po-
sition angle. However, if the SDSS imaging showed a
companion near our target we aligned the position angle
to include that companion rather than using the position
angle of the galaxy, and at times mechanical constraints
on the rotation of the telescope prevented us from ob-
serving at a second position angle that was orthogonal
to the first.
The data were reduced following standard procedures
in IRAF and IDL.
3. ANALYSIS OF SDSS DATA
3.1. Velocity Measurements
In this section we present measurements of the veloc-
ity offsets between the redshifted and blueshifted compo-
nents of the double-peaked [O III] λ5007 emission lines
in the SDSS spectra. Although velocity offsets have been
measured in the literature (Wang et al. 2009; Liu et al.
2010b; Smith et al. 2010), cases where different authors
measured the velocity offset for the same object often
conflict. Consequently, for completeness and consistency,
we measure all of the velocity offsets from the SDSS spec-
tra here (Table 2).
For each galaxy,we fit two Gaussians to the
continuum-subtracted [O III] λ5007 emission line profile,
which is the emission line with the highest signal-to-noise
ratio in our observations. We measure the line-of-sight
velocity difference based on the wavelengths of the peaks
of the best-fit Guassians, and the error in the velocity
difference is derived from the errors in the peak wave-
lengths of the best-fit Guassians added in quadrature.
3.2. Ellipticities
Using the r-band Stokes parameters Qr and Ur from
the SDSS photometric pipeline (Stoughton et al. 2002),
we compute the ellipticity e =?Q2
in our sample.
r+ U2
rfor each object
4. LONGSLIT SPECTRAL ANALYSIS
4.1. Spatial Separation Measurements
For each two-dimensional longslit spectrum, we deter-
mine the projected spatial separation between the two
[O III] λ5007 emission features by measuring the spatial
centroid of each emission component individually. We
measure the spatial centroid of an emission component
by first centering a 2˚ A (rest-frame) wide window on the
emission. Next, at each spatial position we sum the flux,
weighted by the inverse variance, over all wavelengths
within the window. We place a 10 pixel window around
the emission, and define the window’s center to be the
spatial position where the summed flux is maximum. We
then fit a quadratic to the summed flux to locate a peak,
define a narrow window centered on the peak flux, and

Page 4

4 Comerford et al.
Fig.1.— Segments of the two-dimensional longslit spectra
and one-dimensional SDSS spectra of five example double-peaked
AGNs that exhibit two spatially-compact emission components.
Each object was observed at two roughly orthogonal position an-
gles, as labeled in the two-dimensional spectra. The spectra are
shifted to the rest frame of the host galaxy, and the dotted ver-
tical lines denote the rest wavelengths of Hβ, [O III] λ4959, and
[O III] λ5007. Night-sky emission features have been subtracted
from the two-dimensional spectra.
Fig.2.— As Figure 1, but for three example double-peaked
AGNs that exhibit spatially-extended emission.
compute the line centroid within this window. We derive
the error on this spatial centroid by repeatedly adding
noise to the spectrum drawn from a Gaussian with the
variance of the pixels in the window and redoing all cen-
troid measurements.
For a given target, we measure the projected spatial
separation between the two [O III] λ5007 emission com-
ponents at both position angles observed, and then com-
bine the two spatial separations to yield the spatial offset
between the two emission components on the sky and the
position angle of their offset. If observations at position
angles θ1and θ2yield spatial separation measurements of

Page 5

Kiloparsec-scale Spatial Offsets in Double-peaked Narrow-line AGNs5
x1and x2, respectively, then the true position angle θsky
of the two emission components on the sky is determined
through the numerical solution of
x1cos(θsky− θ2) = x2cos(θsky− θ1).(1)
With θskydetermined above, the spatial separation of
the two emission components on the sky is then ∆x =
x1/cos(θsky−θ1) or equivalently, ∆x = x2/cos(θsky−θ2).
4.2. Spatial Extent of the Emission
We also classify by eye whether the AGN emission com-
ponents in the longslit spectra are spatially “compact”
or “extended”. We find that each emission feature dis-
plays two components, and if both components appear
spatially distinct and compact at both position angles
observed we label the object “compact”. If one or both
of the components appear spatially extended or diffuse
at one or both position angles, we label the object “ex-
tended”.This extended structure is expected for gas
kinematics, outflows, or jets, though we note it could
also be produced by dual AGNs where one or both of
the AGNs also has visible emission from gas kinematics,
outflows, or jets.
We find 47 spatially-compact objects and 34 spatially-
extended objects, corresponding to 58+5
the sample, respectively. Figure 1 shows examples of the
spatially-compact object spectra, while Figure 2 shows
examples of the spatially-extended object spectra.
Sections 5 and 6 address other observations of our sam-
ple and some objects that we find have unusual longslit
spectra; for results for the complete sample skip to Sec-
tion 7.
−6% and 42+6
−5% of
5. MULTIWAVELENGTH OBSERVATIONS FROM
THE LITERATURE
In an effort to classify the nature of the double-peaked
[O III] λ5007 emission lines in SDSS spectra, some au-
thors have conducted follow-up observations of subsam-
ples of these objects in optical, near infrared (NIR), ra-
dio, and X-ray. These observations aimed to test the dual
AGN hypothesis by resolving systems with two AGNs
or two stellar components, where the stellar components
are presumed to be remnants of the progenitor galaxies
from the merger that produced the dual AGNs. Here we
summarize these observations and their overlap with our
sample (see also Table 2).
The follow-up optical slit spectroscopy and NIR
imaging of 31 Type 2 Seyferts with double-peaked
[O III] λ5007 (Liu et al. 2010a; Shen et al. 2011) in-
cludes 14 objects in our sample. None of our objects
exhibit two stellar components in the NIR imaging. How-
ever, in three of our objects we measure that the emis-
sion components in the longslit spectra are separated by
less than the ∼ 0.′′4 spatial resolution of the NIR imag-
ing, such that stellar nuclei coincident with the emission
components would not be spatially resolved in the NIR
imaging. In the remaining 11 objects, stellar nuclei that
are spatially-coincident with the double emission com-
ponents could be resolved in the NIR imaging, so either
they have no secondarystellar component at these spatial
separations, or the secondary stellar nucleus is obscured
or too faint to be visible in the NIR imaging.
Follow-up adaptive optics (AO) imaging strived to re-
solve smaller-separation stellar components in 55 galax-
ies with double-peaked [O III] λ5007 (Fu et al. 2011a;
Rosario et al. 2011; Fu et al. 2011b). In addition to ob-
taining integral field unit (IFU) spectroscopy of 42 galax-
ies with double-peaked [O III] λ5007, Fu et al. (2011b)
also analyzed high-resolution ACS/WFC2 archival HST
images of six additional double-peaked galaxies. These
IFU, AO, and HST observations include 38 objects in our
sample, of which seven show double stellar components
in AO.
Although position angles between the stellar compo-
nents are not reported in the literature (Fu et al. 2011a;
Rosario et al. 2011; Fu et al. 2011b), we estimate the po-
sition angles for these seven objects based on our best
estimates of the bright centroid of each component in
the published images. In Section 7.4.2 we compare these
position angle estimates, as well as the reported spatial
separations, to the position angles and spatial separa-
tions measured from our longslit spectroscopy.
Further, Very Long Baseline Array (VLBA) observa-
tions of 11 Type 2 Seyferts that have double-peaked
[O III] λ5007 as well as Faint Images of the Radio Sky
at Twenty-cm (FIRST) detections reveal compact radio
emission in only two objects (Tingay & Wayth 2011).
One (SDSS J151709.21+335324.7) of these two galaxies
is part of our sample, and this galaxy also has a radio
jet that suggests a radio jet-driven outflow produces the
double-peaked [O III] λ5007 lines (Rosario et al. 2010).
None of the VLBA observations show double radio cores,
although Expanded Very Large Array observations of an-
other double-peaked AGN (not in our sample) revealed
two radio cores that confirmed it as a dual AGN system
(Fu et al. 2011c).We also find no significant correla-
tion between FIRST detections and other properties of
double-peaked AGNs that are measured here, such as
line-of-sight velocity separation and projected physical
separation.
Finally, Chandra observations of the double-peaked
AGN in SDSS J171544.05+600835.7 show two X-ray
sources with separation ∆x
orientation θsky = 147◦, which are consistent with
the separation and orientation measured between the
double [O III] λ5007 sources in the longslit spectra
(Comerford et al. 2011). While this is strong evidence
for dual AGNs, additional follow-up observations are re-
quired for confirmation of dual AGNs in this system.
=1.9 h−1
70
kpc and
6. UNUSUAL SPECTRA
Three of our targets have particularly unusual and
noteworthy spectra, which we discuss briefly here. Fig-
ure 3 shows the spectra and Figure 4 shows the corre-
sponding SDSS images of these galaxies.
SDSS J080740.99+390015.2
Although this galaxy was classified as a double-peaked
AGN in Wang et al. (2009), its spectra display triple-
peaked narrow lines and single-peaked broad lines (Fig-
ure 3, top), a line profile that could be explained by
outflows or jets that affect the NLR only. We exclude
this galaxy from the analysis in this paper, as we desire
a homogeneous sample of double-peaked AGNs.