0% found this document useful (0 votes)
99 views6 pages

Aleksandar M. Diamond-Stanic, John Moustakas, Christy A. Tremonti, Alison L. Coil, Ryan C. Hickox, Aday R. Robaina, Gregory H. Rudnick, & Paul H. Sell

This document summarizes a study that discovered compact, obscured star formation in galaxies at z ~ 0.6 that exhibit outflows with velocities greater than 1000 km/s. The study estimates star formation rate surface densities approaching 3000 M☉/yr/kpc^2, comparable to the Eddington limit from radiation pressure on dust grains. It argues that feedback from a compact starburst in the form of radiation pressure and supernova ram pressure is sufficient to produce the high-velocity outflows, without needing feedback from an active galactic nucleus. The study uses Hubble Space Telescope imaging to measure galaxy sizes and estimates star formation rates and stellar masses from infrared data to argue that the outflows could be driven by extremely dense starbursts

Uploaded by

jimjorge
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
99 views6 pages

Aleksandar M. Diamond-Stanic, John Moustakas, Christy A. Tremonti, Alison L. Coil, Ryan C. Hickox, Aday R. Robaina, Gregory H. Rudnick, & Paul H. Sell

This document summarizes a study that discovered compact, obscured star formation in galaxies at z ~ 0.6 that exhibit outflows with velocities greater than 1000 km/s. The study estimates star formation rate surface densities approaching 3000 M☉/yr/kpc^2, comparable to the Eddington limit from radiation pressure on dust grains. It argues that feedback from a compact starburst in the form of radiation pressure and supernova ram pressure is sufficient to produce the high-velocity outflows, without needing feedback from an active galactic nucleus. The study uses Hubble Space Telescope imaging to measure galaxy sizes and estimates star formation rates and stellar masses from infrared data to argue that the outflows could be driven by extremely dense starbursts

Uploaded by

jimjorge
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Submitted to ApJ Letters, 2012 May 9

Preprint typeset using LATEX style emulateapj v. 5/2/11

HIGH-VELOCITY OUTFLOWS WITHOUT AGN FEEDBACK:


EDDINGTON-LIMITED STAR FORMATION IN COMPACT MASSIVE GALAXIES
Aleksandar M. Diamond-Stanic1,2 , John Moustakas1 , Christy A. Tremonti3 , Alison L. Coil1 , Ryan C. Hickox4 ,
Aday R. Robaina5 , Gregory H. Rudnick6 , & Paul H. Sell3

arXiv:1205.2368v1 [astro-ph.CO] 10 May 2012

Submitted to ApJ Letters, 2012 May 9

ABSTRACT
We present the discovery of compact, obscured star formation in galaxies at z 0.6 that exhibit
& 1000 km s1 outflows. Using optical morphologies from the Hubble Space Telescope and infrared
photometry from the Wide-field Infrared Survey Explorer, we estimate star formation rate (SFR)
surface densities that approach SFR 3000 M yr1 kpc2 , comparable to the Eddington limit from
radiation pressure on dust grains. We argue that feedback associated with a compact starburst in the
form of radiation pressure from massive stars and ram pressure from supernovae and stellar winds is
sufficient to produce the high-velocity outflows we observe, without the need to invoke feedback from
an active galactic nucleus.
Subject headings: galaxies: evolution galaxies: kinematics and dynamics galaxies: ISM
galaxies: starburst
1. INTRODUCTION

The central regions of elliptical galaxies are thought


to form in compact starbursts (Kormendy et al. 2009;
Hopkins et al. 2009). Feedback associated with such
starbursts can produce outflows driven by thermal energy
from supernova explosions (Chevalier & Clegg 1985),
stellar winds (Leitherer et al. 1992), and momentum input from both supernova ram pressure and radiation
pressure on dust grains (Murray et al. 2005). It has been
argued that such feedback imposes a limit on the maximum star-formation rate (SFR) surface density (SFR )
for starbursts (Lehnert & Heckman 1996; Meurer et al.
1997; Murray et al. 2005; Thompson et al. 2005) and the
maximum stellar surface density for elliptical galaxies
and star clusters (Hopkins et al. 2010).
Galactic winds are ubiquitous in star-forming galaxies
at all redshifts and generally exhibit outflow velocities
in the 100500 km s1 range, which can be attributed
to the stellar processes described above (Heckman et al.
2000; Shapley et al. 2003; Martin 2005; Rupke et al.
2005; Weiner et al. 2009; Rubin et al. 2010). Outflows
with significantly higher velocities (|v| > 1000 km s1 )
were discovered by Tremonti et al. (2007) in a sample
of massive (M 1011 M ) post-starburst galaxies at
z 0.6, and it was suggested that a more energetic
source such as feedback from an accreting supermassive
black hole (Silk & Rees 1998; Di Matteo et al. 2005) may
be responsible for launching the winds (see Fabian 2012,
for a recent review).
However, it also plausible that feedback from a com1 Center for Astrophysics and Space Sciences, University of
California, San Diego, La Jolla, CA 92093, USA
2 Center for Galaxy Evolution Fellow; [email protected]
3 Department of Astronomy,
University of WisconsinMadison, Madison, WI 53706, USA
4 Department of Physics and Astronomy, Dartmouth College,
Hanover, NH 03755, USA
5 Institut de Ci
encies del Cosmos, University of Barcelona,
08028 Barcelona, Spain
6 Department of Physics and Astronomy, University of
Kansas, Lawrence, KS 66045, USA

pact starburst could expel gas with such large velocities.


Indeed, there is evidence for a positive correlation between outflow velocity and starburst luminosity (Martin
2005; Rupke et al. 2005; Tremonti et al. 2007), albeit
with significant scatter. Furthermore, Heckman et al.
(2011) recently found outflows with maximum velocities
reaching 1500 km s1 in a sample of local starbursts with
compact nuclei, and argued that such velocities could be
explained by a wind launched from r0 100 pc and
driven by feedback from massive stars and supernovae.
In this Letter, we measure sizes and SFRs for a sample of massive galaxies at z 0.6 that exhibit &
1000 km s1 outflows, expanding on the initial study by
Tremonti et al. (2007). We seek to test whether the energetic outflows in these galaxies could have been driven
by feedback from starbursts with very large SFR surface
densities. Our analysis combines galaxy sizes measured
with the Hubble Space Telescope (HST) with SFRs and
stellar masses estimated from Wide-field Infrared Survey
Explorer (WISE), Spitzer Space Telescope, Sloan Digital Sky Survey (SDSS), and Galaxy Evolution Explorer
(GALEX) photometry.
2. ANALYSIS

2.1. Morphologies and Sizes


We observed 29 galaxies with HST (programs 12019
and 12272) selected from a parent sample of 103
galaxies at 0.35 < z < 1.0 with post-starburst spectral features: B or A-star dominated stellar continua
and moderately weak nebular emission (EW([O ii]) <
20
A; see C. Tremonti, et al., in preparation for more
details). The galaxies targeted with HST were those
with the youngest derived post-burst ages, tburst .
300 Myr. Therefore this sample has bluer U B colors
and stronger emission lines than typical post-starburst
samples (Coil et al. 2011). Our subsequent UVIR SED
analysis (see Section 2.2) reveals significant ongoing star
formation (SFR > 50 M yr1 ) in the 14/29 galaxies
with WISE 22 m detections, calling into question the
post-starburst nature of roughly half of the HST sample.

Diamond-Stanic et al.

Fig. 1. Left: Observed UVIR SEDs (rest = 0.115 m)


for three galaxies with the largest SFR surface densities (SFR
3000 M yr1 kpc2 ). The top two SEDs are offset by 5 and
10 magnitudes, respectively. We show stellar population fits to
the rest = 0.13 m emission (black solid line) and three templates for dust emission (M82 starburst, blue dashed line; Arp 220
starburst, green dot-dashed line; obscured quasar, red dotted line;
Polletta et al. 2007) scaled to match the obs = 4.6 m band.
The starburst templates provide reasonable fits to the obs = 12
and 22 m WISE photometry, while the quasar template does not.
Right: HST/WFC3 F814W images (probing rest 5000
A) showing that these galaxies are dominated by a compact nucleus.

Using the F814W filter on the WFC3/UVIS channel,


which has 0.04 pixels, we obtained 4 10 min exposures
in a single orbit for each galaxy. The dithered images
were processed with MultiDrizzle7 to produce science
mosaics with 0.02 pixels. For each galaxy, we use GALFIT (Peng et al. 2002) to model the two-dimensional surface brightness profile with a single Sersic component
(characterized by Sersic index n and effective radius re ),
using stars in the images to construct the model pointspread function (PSF). In cases where the best-fit model
returns n > 4, we also fit an n = 4 de Vaucouleurs model,
yielding a larger re value (due to the covariance between
n and re ); we use these these larger effective radii in our
analysis.
In this Letter, we highlight the galaxies with the smallest re and largest SFR values because such extreme
starbursts could conceivably produce the high-velocity
outflows we observe (see Section 3). We show HST images for the three highest SFR galaxies in Figure 1. In
all three cases, the single-component GALFIT model accounts for > 85% of the total flux. The residuals show
diffuse emission that is consistent with these systems being late-stage galaxy mergers, although we defer a detailed study of the merger stage to future work.
For the most compact galaxy (J0905+5759, re =
0.013 or 100 pc), we also show the observed onedimensional surface brightness profile in Figure 2. We
compare to the profiles of six stars in the same image,
the best-fit de Vaucouleurs model, and a de Vaucouleurs
model with re = 0.04 (the native WFC3/UVIS pixel
size, which corresponds to a physical scale of 290 pc).
This comparison illustrates that this galaxy, while only
7

http://stsdas.stsci.edu/multidrizzle/

Fig. 2. One-dimensional surface brightness profile for


J0905+5759, which has the smallest effective radius in the sample.
The observed profile is shown as the solid black line, and the profiles of six stars in the same image are shown in blue, normalized to
the same central surface brightness. The best-fit de Vaucouleurs
profile with re = 100 pc is shown as a dashed red line, and for
comparison a broader profile with re = 0.04 = 290 pc is shown as
the dotted green line. This galaxy is quite compact, but is more
extended than a point source.

marginally resolved with an re that is 20% of the image


FWHM, is clearly more extended than a point source.
For such a compact source, there is uncertainty in our
re measurement given uncertainties in the model PSF.
To quantify this, we used TinyTim to generate a model
PSF that is narrower than the stars in the image (convolving with a FWHM = 0.04 Gaussian, whereas the
image FWHM is 0.074 ) and found that this increased
the re in the GALFIT model by a factor of two. We also
fit a two-component PSF+Sersic model, but found that
the Sersic component dominates the fit, yielding a similar re . Furthermore, the spectrum of the galaxy shows
no evidence for an AGN contribution to the optical continuum (see Figure 3), so there is no clear motivation
for including an unresolved, point-source component in
the model. We conclude that our re estimate is accurate
within a factor of two.
2.2. Star Formation Rates and Stellar Masses
We gathered photometry from the WISE All-Sky Release, the SDSS Seventh Data Release, and GALEX General Release 6. We also obtained 5 30 sec dithered
exposures at 3.6 m and 4.5 m for all sources with
the Infrared Array Camera on the Spitzer Space Telescope as part of GO program 60145. We used the post
basic calibrated data to perform aperture photometry
on all sources and point-source photometry on sources
in crowded fields. We show spectral energy distributions
(SEDs) for the three highest SFR galaxies in Figure 1.
We estimate IR-based SFRs for the 25/29 galaxies with WISE 12 or 22 m detections by fitting
Chary & Elbaz (2001) templates to their 12 and 22 m
fluxes. For the 14/25 galaxies with 22 m detections,
this yields SFRs that agree with those obtained from
the Rujopakarn et al. (2012) method based on 24 m
luminosity with a scatter of 0.05 dex. Several authors

Compact Starbursts Driving High-Velocity Outflows

Keck/HIRES (J0905+5759). The spectra are dominated


by light from a young (t < 50 Myr) stellar population.
We highlight the interstellar medium Mg ii 2796, 2803
absorption lines, which are used to measure outflow
velocities. At low spectral resolution (e.g., the top
right panel of Figure 3) it is not possible to determine the intrinsic shape of the absorption line profile and therefore the covering factor of the outflowing
gas. However, the Keck/HIRES spectrum of J0905+5759
(FWHM 8 km s1 ) reveals that the gas covers the entire continuum source near the velocity centroid (v =
2470 km s1 ) indicating a galaxy-wide outflow.

for the three


Fig. 3. Spectra covering rest = 25005200 A
galaxies shown in Figure 1. For clarity the top two spectra are
offset by +10 and +20 units. These spectra are dominated by
the light of a young stellar population but have relatively weak
nebular emission lines ([O ii] 3727, H 4861, [O iii] 5007) and
strong Mg ii 2796, 2803 absorption arising from the interstellar
medium. The absense of broad Mg ii or H emission lines rules
out a significant contribution to the optical continuum light from
an AGN. The panels on the right highlight the region around the
Mg ii doublet in both wavelength and velocity space to illustrate
the outflow kinematics. The spectrum in the bottom panel has
sufficient spectral resolution (FWHM 8 km s1 ) to resolve the
intrinsic shape of the absorption-line profile, revealing that the gas
near the centroid velocity (v = 2470 km s1 ) covers the entire
galaxy.

have shown that the shape of the IR SED for starforming galaxies depends on SFR (Rujopakarn et al.
2011; Elbaz et al. 2011), with more compact starbursts
having larger total-IR (81000 m) to mid-IR (824 m)
ratios, characteristic of the most luminous galaxies in the
local universe (Rieke et al. 2009). If we used the most luminous local templates for the 8/25 sources with SFRs
in the ULIRG regime (SFRIR > 100 M yr1 ), we would
obtain SFRs that are larger by 0.5 dex than the values
we adopt for this paper.
We also estimate SFRs and stellar masses based on
stellar population fits to the rest = 0.13 m SEDs using the method of Moustakas et al. (2011). For the 14/25
galaxies with SFRIR > 50 M yr1 , there is agreement
between these UV-based SFR estimates and SFRIR with
a scatter of 0.32 dex. For an SMC dust law, we find a median attenuation of AV = 0.4 mag. The observed H luminosities, uncorrected for dust extinction, are typically
factors of 1020 smaller than expected from the UV and
IR SFRs. This can be reconciled by either strong differential dust attenuation (i.e., AV 23 mag for the lineemitting region), escaping ionizing photons from matterbounded H ii regions, or a recently quenched starburst
(t > 5 Myr) with a small ratio of ionizing ( < 912
A)
to non-ionizing UV photons.
2.3. Outflow Kinematics and Covering Factors
We present rest = 25005200
A spectroscopy for
three high-SFR sources in Figure 3 based on data
from MMT/Blue Channel and SDSS (J1506+5402),
Magellan/MagE (J1341-0321), and Keck/LRIS and

3. DISCUSSION
The compact sizes (re 100 pc) and large SFRs
(SFR 200 M ) for the three galaxies described above
imply extremely large SFR surface densities (SFR
3000 M yr1 kpc2 ). To place these galaxies in context, we plot SFR versus stellar mass for the 25/29
galaxies detected by WISE in Figure 4. We include comparison samples of 105 star-forming galaxies at 0.5 <
z < 1.5 from Wuyts et al. (2011) and gas-rich mergers
at z < 0.3 including 32 ULIRGs from Veilleux et al.
(2006), five Lyman break analogs with dominant central
objects from Overzier et al. (2009), and the local compact starburst Arp 220 (Scoville et al. 1997; Kennicutt
1998; Rodrguez Zaurn et al. 2008). We also mark
the empirical threshold for launching winds (SFR
0.1 M yr1 kpc2 , Heckman 2002), the 90th percentile
limit for the surface brightness of starbursts over a wide
range in redshift measured using UV, H, far-IR, and radio continuum emission (SFR 25 M yr1 kpc2 for
a Chabrier IMF, Meurer et al. 1997), and the theoretical
limit for a starburst limited by feedback from radiation
pressure (SFR 3000 M yr1 kpc2 , Murray et al.
2005; Thompson et al. 2005; Hopkins et al. 2010). The
most luminous, compact starbursts in our sample exhibit
SFR surface densities that reach the Eddington limit,
suggesting that their growth is being regulated by momentum input from massive stars.

3.1. Constraints on Ongoing AGN Activity


While the SEDs (Figure 1) and optical spectra (Figure 3) for our sample indicate that their bolometric output is dominated by star formation, it is worthwhile to
consider the level of ongoing AGN activity and its effect on our measurements. Among the high-SFR galaxies, the strongest case for detectable AGN emission can
be made for J1506+5402, which has the most luminous
[O iii] 5007 line in the HST sample. This source also
has a weak [Ne v] 3426 emission line, which is associated with AGN activity (Gilli et al. 2010). It was observed with the Chandra X-ray Observatory (proposal ID
11700896) and had four detected counts, corresponding
to a 0.58.0 keV X-ray luminosity of 1042.7 erg s1 (see
P. Sell et al., in preparation for more details on the 12/29
galaxies with Chandra observations). Using the relationship between 210 keV X-ray and 12.3 m mid-IR luminosity for AGNs from Gandhi et al. (2009), one would expect a source with LX 5 1042 erg s1 to have LMIR
71042 erg s1 , which is 400 smaller than the observed
luminosity for J1506+5402 (LMIR 3 1045 erg s1 ).
Based on its [O iii] luminosity (1042.1 erg s1 , which may

Diamond-Stanic et al.

Fig. 4. SFR surface densities and stellar masses for the HSTWISE sample described in this paper (black circles, symbol size proportional
to outflow velocity), along with samples of z < 0.3 gas-rich mergers (orange squares) and z 1 star-forming galaxies (shown with 68%,
95%, and 99.7% contours; see Section 3). We mark the empirical threshold for launching winds (dotted line, SFR 0.1 M yr1 kpc2 ;
Heckman 2002), the 90th-percentile starburst intensity limit from Meurer et al. (1997) (dashed line, SFR 25 M yr1 kpc2 ), and the
Eddington limit from radiation pressure on dust grains (solid line, SFR 3000 M yr1 kpc2 ; Murray et al. 2005; Thompson et al.
2005; Hopkins et al. 2010). The representative error bar in the top-right portion of the plot corresponds to uncertainties of 0.3 dex in SFR
and 0.2 dex in stellar mass. Our HSTWISE sample overlaps with the region characterized by gas-rich mergers, and extends to very large
SFR surface densities near the Eddington limit, suggesting growth that is limited by momentum injection from massive stars.

also include a contribution from star formation) and the


calibration for type 1 AGNs from Heckman et al. (2005),
one would expect an intrinsic 210 keV luminosity of
LX 5 1043 erg s1 , suggesting absorption by a factor of 10 in the X-rays. However, even if the X-ray
attenuation were a factor of 100, which is typical for local Compton-thick AGNs (Diamond-Stanic et al. 2009),
the expected mid-IR AGN contribution would only be
relevant at the .30% level. We conclude that the bolometric output of the galaxies in our sample is dominated
by star formation and that our results regarding large
SFR values are not affected by AGN contamination.
3.2. The Outflow Launching Mechanism
Are the high-velocity outflows observed in these galaxies produced by a compact starburst? Murray et al.
(2011) argued that massive star clusters with large

gas surface densities are the ideal launching point for


galactic-scale outflows driven by radiation pressure, and
that the outflow velocity should scale with escape velocity of the most massive star clusters in a galaxy. For
our sample, if one assumes that the spatial extent of
the stellar mass is comparable to that of the rest-frame
V -band light (see Section 3.3), then half of the stellar
mass (M 1011 M ) will be within the effective radius
(re 100 pc). Such a compact stellar population would
have an escape velocity comparable to the & 1000 km s1
outflow velocities we observe:

p
2GM /r

1/2
1/2 
M
r
km s1 (1)
= 2100
1011 M
200 pc

vesc =

Compact Starbursts Driving High-Velocity Outflows


This argument, combined with the fact that we observe
galaxies with significant dust-obscured star formation
and SFR values near the Eddington limit, suggests that
momentum input from massive stars in the form of radiation pressure is a viable mechanism for launching these
outflows.
In addition to radiation pressure, we also expect significant momentum flux from stellar winds and supernovae.
For example, a starburst with SFR 200 M yr1 would
have radiation pressure Lbol /c 3 1035 dyne and ram
pressure p 510 1035 dyne from stellar winds and
supernovae (Leitherer et al. 1992, 1999; Veilleux et al.
2005). Furthermore, Heckman et al. (2011) noted that
such a large momentum injection (p 1035 dyne) from a
small initial radius r0 100 pc could accelerate a cloud
with column density NH 1021 cm2 to a terminal velocity v 1800 km s1 . Thus, the energetics of compact starbursts are sufficient to produce the high-velocity
outflows we observe, and it is plausible that both radiation pressure on dust grains and supernova ram pressure
contribute to driving the winds.
3.3. Placing These Galaxies in Context
It is clear from Figure 4 that the high-SFR galaxies in
our sample constitute a rare population, suggesting that
they represent an unusual or short-lived phase. Mergers
of gas-rich galaxies are a viable mechanism for producing compact starbursts (Mihos & Hernquist 1996), and
such gas-rich major mergers are rare at z 0.6 due
to the decline in both the gas fraction (Tacconi et al.
2010) and the fractional major merger rate (Lotz et al.
2011) of galaxies since z 2. Furthermore, our highSFR galaxies are caught in a particular time interval
where there is both vigorous star formation and strong
feedback. The length of this phase may be set by the
gas consumption timescale or the timescale for feedback
to suppress subsequent star formation. Based on the
KennicuttSchmidt (K-S) relation (Kennicutt 1998), a
compact starburst with re 100 pc, SFR 200 M ,
and SFR 3000 M yr1 kpc2 would have a gas surface density of gas 1011 M kpc2 corresponding to
Mgas 3 109 M inside 100 pc, which would be consumed on a timescale 20 Myr. This scenario could be
tested with CO observations of molecular gas masses and
kinematics for these extreme galaxies.
Considering our full HSTWISE sample in Figure 4, we
find values of SFR spanning four orders of magnitude.
This could be explained as an evolutionary sequence
where the high-SFR galaxies represent the peak of the
starburst when the high-velocity outflows are launched,
while the lower SFR galaxies represent a subsequent,
post-starburst phase. It is interesting to note that all
12/25 galaxies above the SFR 25 M yr1 kpc2 limit
from Meurer et al. (1997) exhibit outflows (with median
centroid velocity v = 1500 km s1 ), while all 7/25

galaxies without detected outflows (the smallest black


circles in Figure 4) have SFR < 20 M yr1 kpc2 . If a
compact starburst is a requirement for the production of
high-velocity outflows, it may be that the sources without outflows have not gone through such a phase (see A.
Robaina, et al., in preparation for a discussion of the relationship between galaxy morphology, outflow velocity,
and stellar population age in this sample).
Finally, we consider the implications of our results for
models of massive galaxy formation. Simulations of major galaxy mergers with large gas fractions (fgas 50%)
can produce M 1011 M remnants with re 1 kpc
(Wuyts et al. 2010), but our sample includes galaxies of
similar mass that are smaller in the rest-frame V band
by almost an order of magnitude. If the mass in these
galaxies follows their V -band light, it would be extremely
challenging for them to grow in size from re 0.1 kpc
to re 5 kpc to reach the local sizemass relation
(Shen et al. 2003) in the t 6 Gyr since z = 0.6. For
comparison the compact, quiescent galaxies observed at
z 2 (Trujillo et al. 2007; van Dokkum et al. 2008) have
t 10 Gyr to grow by a factor of 5. In this context, it
is worth noting that the half-light radius (ignoring dust
attenuation) at rest-frame V band can be a factor of 510
smaller than the half-mass radius for a gas-rich merger
at final coalescence near the peak of starburst activity
(Wuyts et al. 2010). One could test for such size discrepancies and probe the radial dependence of the massto-light ratio for these galaxies by measuring sizes at restframe near-IR wavelengths.
4. SUMMARY
We have measured large SFR surface densities for
galaxies that exhibit & 1000 km s1 outflows. The
largest SFR values are comparable to the Eddington
limit from radiation pressure on dust grains, and such
compact starbursts are expected to have substantial momentum input from massive stars and supernovae. Highvelocity outflows have been previously interpreted as a
signpost of AGN feedback, but given that feedback from
a compact starburst is capable of producing such a signature and is clearly present in this sample, we conclude
that the outflows we observe are driven by star formation.

We acknowledge useful discussions with and assistance from James Aird, Brandon Kelly, Dusan Keres,
David Law, Alexander Mendez, Kate Rubin, Art Wolfe,
and Stijn Wuyts. AMD acknowledges support from
the Southern California Center for Galaxy Evolution, a
multi-campus research program funded by the University of California Office of Research. Support for HSTGO-12272 was provided by NASA through a grant from
STScI. Support for Spitzer-GO-60145 was provided by
contract 1419615 from JPL/Caltech. This paper includes
data obtained at the W.M. Keck Observatory.

REFERENCES
Chary, R., & Elbaz, D. 2001, ApJ, 556, 562
Chevalier, R. A., & Clegg, A. W. 1985, Nature, 317, 44
Coil, A. L., Weiner, B. J., Holz, D. E., et al. 2011, ApJ, 743, 46
Diamond-Stanic, A. M., Rieke, G. H., & Rigby, J. R. 2009, ApJ,
698, 623

Di Matteo, T., Springel, V., & Hernquist, L. 2005, Nature, 433,


604
Elbaz, D., Dickinson, M., Hwang, H. S., et al. 2011, A&A, 533,
A119
Fabian, A. C. 2012, arXiv:1204.4114
Gandhi, P., Horst, H., Smette, A., et al. 2009, A&A, 502, 457

Diamond-Stanic et al.

Gilli, R., Vignali, C., Mignoli, M., et al. 2010, A&A, 519, A92
Heckman, T. M., Lehnert, M. D., Strickland, D. K., & Armus, L.
2000, ApJS, 129, 493
Heckman, T. M. 2002, Extragalactic Gas at Low Redshift, 254,
292
Heckman, T. M., Ptak, A., Hornschemeier, A., & Kauffmann, G.
2005, ApJ, 634, 161
Heckman, T. M., et al. 2011, ApJ, 730, 5
Hopkins, P. F., Cox, T. J., Dutta, S. N., et al. 2009, ApJS, 181,
135
Hopkins, P. F., Murray, N., Quataert, E., & Thompson, T. A.
2010, MNRAS, 401, L19
Kennicutt, R. C., Jr. 1998, ApJ, 498, 541
Kormendy, J., Fisher, D. B., Cornell, M. E., & Bender, R. 2009,
ApJS, 182, 216
Lehnert, M. D., & Heckman, T. M. 1996, ApJ, 472, 546
Leitherer, C., Robert, C., & Drissen, L. 1992, ApJ, 401, 596
Leitherer, C., Schaerer, D., Goldader, J. D., et al. 1999, ApJS,
123, 3
Lotz, J. M., Jonsson, P., Cox, T. J., et al. 2011, ApJ, 742, 103
Martin, C. L. 2005, ApJ, 621, 227
Meurer, G. R., Heckman, T. M., Lehnert, M. D., Leitherer, C., &
Lowenthal, J. 1997, AJ, 114, 54
Mihos, J. C., & Hernquist, L. 1996, ApJ, 464, 641
Moustakas, J., Zaritsky, D., Brown, M., et al. 2011,
arXiv:1112.3300
Murray, N., Quataert, E., & Thompson, T. A. 2005, ApJ, 618, 569
Murray, N., M
enard, B., & Thompson, T. A. 2011, ApJ, 735, 66
Overzier, R. A., Heckman, T. M., Tremonti, C., et al. 2009, ApJ,
706, 203
Peng, C. Y., Ho, L. C., Impey, C. D., & Rix, H.-W. 2002, AJ,
124, 266
Polletta, M., Tajer, M., Maraschi, L., et al. 2007, ApJ, 663, 81

Rodrguez Zaurn, J., Tadhunter, C. N., & Gonz


alez Delgado,
R. M. 2008, MNRAS, 384, 875
Rieke, G. H., Alonso-Herrero, A., Weiner, B. J., et al. 2009, ApJ,
692, 556
Rubin, K. H. R., Weiner, B. J., Koo, D. C., Martin, C. L.,
Prochaska, J. X., Coil, A. L., & Newman, J. A. 2010, ApJ, 719,
1503
Rujopakarn, W., Rieke, G. H., Eisenstein, D. J., & Juneau, S.
2011, ApJ, 726, 93
Rujopakarn, W., Rieke, G. H., Weiner, B. J., et al. 2011,
arXiv:1107.2921
Rupke, D. S., Veilleux, S., & Sanders, D. B. 2005, ApJS, 160, 115
Scoville, N. Z., Yun, M. S., & Bryant, P. M. 1997, ApJ, 484, 702
Shapley, A. E., Steidel, C. C., Pettini, M., & Adelberger, K. L.
2003, ApJ, 588, 65
Shen, S., Mo, H. J., White, S. D. M., et al. 2003, MNRAS, 343,
978
Silk, J., & Rees, M. J. 1998, A&A, 331, L1
Tacconi, L. J., Genzel, R., Neri, R., et al. 2010, Nature, 463, 781
Thompson, T. A., Quataert, E., & Murray, N. 2005, ApJ, 630, 167
Tremonti, C. A., Moustakas, J., & Diamond-Stanic, A. M. 2007,
ApJ, 663, L77
Trujillo, I., Conselice, C. J., Bundy, K., et al. 2007, MNRAS, 382,
109
van Dokkum, P. G., et al. 2008, ApJ, 677, L5
Veilleux, S., Cecil, G., & Bland-Hawthorn, J. 2005, ARA&A, 43,
769
Veilleux, S., Kim, D.-C., Peng, C. Y., et al. 2006, ApJ, 643, 707
Weiner, B. J., et al. 2009, ApJ, 692, 187
Wuyts, S., Cox, T. J., Hayward, C. C., et al. 2010, ApJ, 722, 1666
Wuyts, S., F
orster Schreiber, N. M., van der Wel, A., et al. 2011,
ApJ, 742, 96

You might also like