The Molecular Gas Content of z<0.1 Radio Galaxies: Linking the AGN Accretion Mode to Host Galaxy Properties

0.43 MB·English

by  V. Smolcic

PreprinttypesetusingLATEXstyleemulateapjv.11/10/09
THE MOLECULAR GAS CONTENT OF Z <0.1 RADIO GALAXIES:
LINKING THE AGN ACCRETION MODE TO HOST GALAXY PROPERTIES
V. Smolcˇic´1,2,3 and D. A. Riechers4,5
ABSTRACT
One of the main achievements in modern cosmology is the so-called‘unified model’, which success-
fully describesmostclassesof activegalacticnuclei (AGN) within a single physicalscheme. However,
1
there is a particular class of radio-luminous AGN that presently cannot be explained within this
1
framework – the ‘low-excitation’ radio AGN (LERAGN). Recently, a scenario has been put forward
0
which predicts that LERAGN, and their regular ‘high-excitation’ radio AGN (HERAGN) counter-
2
parts represent different (red sequence vs. green valley) phases of galaxy evolution. These different

n evolutionarystatesarealsoexpectedtobereflectedintheirhostgalaxyproperties,inparticulartheir
a cold gas content. To test this, here we present CO(1→0) observations towarda sample of 11 of these
J systems conducted with CARMA. Combining our observationswith literature data, we derive molec-

7 ular gas masses (or upper limits) for a complete, representative, sample of 21 z < 0.1 radio AGN.
2 Our results yield that HERAGN on average have a factor of ∼ 7 higher gas masses than LERAGN.
We also infer younger stellar ages, lower stellar, halo, and central supermassive black masses, as well

] as higher black hole accretion efficiencies in HERAGN relative to LERAGN. These findings support
O
the ideathathigh-andlow-excitationradioAGNformtwophysicallydistinctpopulationsofgalaxies
C that reflect different stages of massive galaxy build-up.
h. Subject headings: galaxies: fundamental parameters – galaxies: active, evolution – cosmology: obser-
p vations – radio continuum: galaxies
-
o
r 1. INTRODUCTION tent with these two classes of radio AGN dividing in a
t
s Over the past two decades a standard model of AGN stellar mass vs. color plane in such a way that LER-
a AGN occupy the red sequence and HERAGN inhabit
has emerged. In this ‘unified’ model efficient disk accre-
[
the so called “green valley”, a sparsely populated region
tion of cold matter on the central supermassive black
2 hole (BH) provides the radiation field that photoion- between the blue-cloud and the red-sequence (Smolˇci´c
v izes emission-line regions. However, there is a certain 2009).
5 fraction of AGN identified by radio observations that The stellar mass vs. color plane can be interpreted
3 posesachallengeto the unifiedmodel,the so-calledlow- as a time-sequence for galaxy evolution. Galaxies
9 are thought to evolve from an initial star-formation-
excitation radio AGN (hereafter: LERAGN). The main
4 dominated state with blue optical colors into the most
difference between high-excitation radio AGN (HER-
. massivered-and-deadgalaxiesthroughatransitionphase
1 AGN) and these LERAGN is that the latter do not ex-
reflected in the green valley (Bell et al. 2004a, 2004b;
0 hibit strongemissionlines in their opticalspectra (Jack-
Borch et al. 2006; Faber et al. 2007; Brown et
1 son & Rawlings 1997;Evans et al. 2006).
al. 2007). In recent years it has been suggested
1 Recently, Hardcastle et al. (2006) have suggested that
: high- and low-excitation radio AGN may represent a that radio outflows from AGN likely play a crucial
v role in this massive galaxy build-up (Croton et al. 2006;
principal separator between populations fundamentally
Xi different in their black hole accretion mechanisms (see Bower et al. 2006; Sijacki & Springel 2006; Sijacki et al.
2007; Fanidakis et al. 2010). In this context the radio-
r also Evans et al. 2006; Allen et al. 2006; Kewley et al.
a 2006). They developed a model in which central super- AGN feedback (often called the “radio” or “mainte-
nance” mode), which is thought to limit stellar mass
massive black holes of HERAGN accrete in a standard
growth in already massive galaxies, is expected to occur
(radiatively efficient) way from the cold phase of the in-
only in LERAGN (Smolˇci´c 2009).
tragalactic medium (IGM), while those of LERAGN are
Furthermore,ithasbeenshownthatthecosmicevolu-
poweredin a radiatively inefficient manner by Bondi ac-
tion of the space density of various types of radio AGN
cretionofthe hotIGM.Smolˇci´c(2009)showedthatlow-
issignificantlydifferent(e.g.Peacocketal.1985,Willott
andhighexcitationradioAGNexhibitnotonlysystemic
etal.2001;Smolcicetal.2009). Basedonastudy ofthe
differences in their black hole masses and accretion rate
evolution of the radio AGN luminosity function out to
properties, but also in their host galaxy properties,such
z =1.3,Smolˇci´cetal.(2009)haveshownthatthecomov-
as stellar masses and stellar populations. This is consis-
ingspacedensityoflow-luminosityradioAGN (predom-
1EuropeanSouthernObservatory,Karl-Schwarzschild-Strasse inantly LERAGN) only modestly declines since z =1.3,
2,D-85748Garchingb. Mu¨nchen,Germany while that of powerful AGN (predominantly HERAGN)
2ArgelanderInstitutforAstronomy,AufdemHu¨gel71,Bonn, dramaticallydiminishesoverthesamecosmictimeinter-
D-53121,Germany
3ESOALMACOFUNDFellow val. ThissuggeststhatLERAGNandHERAGNnotonly
4California Institute of Technology, MC249-17, 1200 East representphysicallydistinctgalaxypopulations,butalso
CaliforniaBoulevard,Pasadena,CA91125,USA populations in different stages of massive galaxy build-
5HubbleFellow
2 Smolˇci´c & Riechers
up. If this is the case, the molecular gas masses and available. The total calibration is estimated to be ac-
fractions in low- and high- excitation radio AGN are ex- curate to 15%. Data reduction was performed using the
pected to directly reflect this trend. MIRIAD package. The CO(1→0) spectra are shown in
We here investigate this idea by observing Fig. 2.
CO(J=1→0) emission of a carefully selected, rep-
resentative sample of nearby (z < 0.1) HE- and 3. RESULTS
LERAGNwith CARMA. We adopta ΛCDM cosmology 3.1. CO Data
with H0 =70, ΩM =0.3, ΩΛ =0.7. CO(1→0)hasbeendetectedin4(3C33,3C321,3C403,
2. DATA andNGC6109)outofthe11galaxiesinourCARMA-CO
2.1. Sample sample (see Fig. 2). To parameterize the emission lines
detectedinthesefourgalaxieswefitGaussianprofilesto
We here utilize a sample of 21 Type 2 AGN atz <0.1 the line and underlying continuum emission (see ⁀tab:co
that have been observed in X-rays (with Chandra or
for line/continuum properties).
XMM-Newton) by Evans et al. (2006). 18 out of the 21
Two of our four CO detected sources (3C321 and
AGN have been drawn from the 3CRR survey, adding 3C403) have recently been detected in the CO(1→0)
3 more sources (3C 403, 3C405 and Cen A) for com-
transitionby Flaqueretal.(2010)usingthe IRAM30m
pleteness (see Evans et al. 2006 for details). The sam-
telescope. The line parameters reported by Flaquer et
ple properties are summarized in ⁀tab:physprops (see al. are in good agreement with ours.
also Tab. 1 in Evans et al. 2006). We separate our 3σ upper limits for CO(1→0) non-detections are de-
AGN into LERAGN (i.e. LINERs) and HERAGN (i.e. terminedbyassumingalinewidthof300 kms−1,corre-
Seyferts) using standard diagnostic tools based on opti-
sponding to the averagewidth of the detected lines. We
calemissionlinefluxratioswherepossible(seeFig.1and furthercomplement⁀tab:co with datafromliteraturefor
⁀tab:physprops ; Kauffmann et al. 2003; Kewley et al. the8sourceswithalreadyexistingCO(1→0)detections,
2001, 2006; Smolˇci´c 2009; Buttiglione et al. 2009). For and the 2 sources (3C 388 and 3C405) that had to be
this we make use of the emission line fluxes extracted excluded from our sample.
from high resolution spectroscopy of 3CR sources pre-
sented in Buttiglione et al. (2009, 2010, 2011, see also 3.2. Ancillary Data
Tab. 1 in Smolˇci´c 2009). In cases where the relevant
Wesummarizethephysicalpropertiesofthe21sources
emission line fluxes are not available, we make use of
in our low- and high-excitation radio AGN sample in
the galaxy type information available in the NASA Ex-
tragalactic Database6 to separate the sources into LE- ⁀tab:physprops. We adoptthe 178MHz luminosities,ac-
and HE-RAGN. The sample contains 9 HERAGN and cretion efficiencies, and black hole masses from Evans et
12 LERAGN. al. (2006). The stellar ages of our sources, taken from
Buttiglione et al. (2009), were derived by fitting stellar
2.2. CO(1→0) observations and data reduction population synthesis models to the (Hα portion of the
sources’)opticalspectra. Combiningthestellarageswith
At the time of observations, 8 out of the 21 Type 2
2MASS K-band luminosities (where available) we com-
AGN in our sample have already been detected in
CO(1→0). Thus, we observed the CO(1→0) transition putedthestellarmassesofoursourcesfollowingDroryet
al. (2004). Drory et al. have parameterized the mass-to-
linetowardtheremaining13AGNusingthetheCARMA
lightratioinK-bandasafunctionofstellarage(seetheir
(Combined Array for Research in Millimeter-wave As-
Fig.1)usingsimplestellarpopulationmodels(Maraston
tronomy) Interferometer. Observations were performed
1998), and a Salpeter initial mass function. The total
during Summer/2009 and Spring/2010 for about 4 to
systematic uncertainty of such a derived mass-to-light
15 hours per source (⁀tab:obs ). All targets were ob- ratio is estimated to be ∼ 25−30%. Lastly, from the
served under good to excellent weather conditions at
CO(1→0)luminositiesinferredforoursources(see⁀tab:co
3mm with 15 antennas (corresponding to 105 baselines)
) we estimated the molecular (H ) mass using a conver-
in the two most compact, E and D configurations (2009 2
sion factor of α = M /L′ = 1.5 (K km s−1 pc2)−1
and 2010, respectively). Data on two objects (3C 388 H2 CO
(Evans et al. 2005).
and 3C 405) had to be discarded due to technical prob-
Wefindsystematicdifferencesintheaverageblackhole
lems, and are excluded in the following. The receivers
were tuned to the redshifted CO(J=1→0) line frequen- and host galaxy properties of the low- and high- exci-
tation sources (i.e. LINERs and Seyferts, respectively)
cies (ν =115.2712GHz; see ⁀tab:obs for exact observ-
rest in our sample (⁀tab:physprops ). This is illustrated in
ing frequencies), centering them in the upper sideband.
Fig. 3, where we also indicate the average properties
Three bands with 15 channels of 31.25MHz width each
of our low- and high-excitation radio AGN, computed
were utilized. The bands were overlapped by 2 channels
using the ASURV statistical package and assuming log-
to improve calibration of the correlated dataset, leading
toaneffectivebandwidthof1281.25MHz(3500 kms−1) normal distributions in luminosity and mass. The aver-
per sideband. Phase calibration was performed by ob- age properties are specifically given in⁀tab:averageprops
serving bright nearby radio quasars every 15minutes. .7 Compared to LERAGN, HERAGN on average
Bandpass calibration was performed once per track on
bright quasars. Fluxes were bootstrapped relative to 7 ItshouldbekeptinmindthatinASURVthereisanimplicit
assumptionthatthecensoreddatafollowasimilardistributionto
planets, or monitored radio quasars if no planet was
thatofthemeasuredpopulation. Ifthisisnotthecase,”average”
values calculated by ASURV willbe generallybiased upwards (as
6 http://nedwww.ipac.caltech.edu our upper limits typically lie towards the bottom-end of the dis-
CO in Radio Galaxies 3
TABLE 1
Physical propertiesof the z<0.1sample of radioAGN
name redshift type L178−MHz L0.5−10keV/LEDD stellarage M∗ MBH MH2
[W/Hz/srad] [Gyr] [M⊙] [M⊙] [M⊙]
3C31 0.017 Seyfert 9.08×1023 <2.0×10−4 3 2.4×1011 7.8×107 (5.1±0.4)×108
3C33 0.060 Seyfert 3.95×1025 1.6×10−3 5 1.3×1011 4.8×108 (3.75±1.5)×108
3C98 0.030 Seyfert 8.75×1024 3.7×10−4 2 7.9×1010 1.7×108 <7.8×107
3C321 0.097 Seyfert 2.6×1025 – 13 7.0×1011 – (3.3±0.6)×109
3C403 0.059 Seyfert 3.5×1025 3.3×10−3 5 2.4×1011 2.6×108 (6.6±1.6)×108
3C449 0.017 Seyfert 6.51×1023 <7.0×10−3 3 2.4×1010 5.1×107 (1.1±0.2)×108
3C452 0.081 Seyfert 7.54×1025 3.3×10−3 13 4.5×1011 3.5×108 8.1×108,c
CenA 0.0008 Sey2a 5.4×1023 3.0×10−5 – – 2.0×108 1.4×108
3C405 0.0565 Sey2a 4.90×1027 8.5×10−4 – – 2.5×109 <3.3×108
3C66B 0.022 LINER 2.21×1024 <4.4×10−5 3 3.0×1010 6.9×108 <7.8×107
3C84 0.018 LINER 3.74×1024 <9.2×10−6 – – 1.9×109 (2.14±0.02)×109
3C264 0.022 LINER 2.20×1024 <1.8×10−5 13 4.4×1011 7.1×108 (9.3±1.8)×107
3C272.1 0.004 LINER 3.1×1022 <8.5×10−7 13 2.9×1011 1.5×109 (9.3±3.2)×105,c
3C274 0.004 LINER 3.4×1024 <4.3×10−7 2 1.5×1011 2.4×109 (1.65±0.15)×107
3C296 0.025 LINER 1.43×1024 <1.2×10−5 13 1.1×1012 1.3×109 <5.7×107
3C338 0.032 LINER 8.63×1024 <2.0×10−5 13 1.3×1012 1.7×109 3×107
3C388 0.091 LINER 4.29×1025 – 9 – – <1.2×109,b
3C465 0.030 LINER 6.41×1024 <2.2×10−4 13 1.0×1012 2.1×109 <1.95×107
NGC1265 0.027 LERGa 3.39×1024 <6.8×10−6 13 1.5×1010 1.0×109 <5.7×107
NGC6109 0.0296 LERGa 1.86×1024 – – – – (1.3±0.3)×108
NGC6251 0.0244 LERGa 1.20×1024 <2.0×10−4 – – 6.0×108 <7.5×107
a BasedonNED;LERGabbreviates “low-excitationradiogalaxy”
b AdoptedfromSaripallietal.(2007), andscaledtothecosmologyusedhere.
c Notconsideredinourstatisticalanalysis(see⁀tab:co fordetails).
Thefirst,secondandthirdcolumnsdenotethesource,itsredshiftandAGNtype. Thelastwasinferredeitherviaopticaldiagnosticdiagrams
(seeFig.1),oradoptedfromNED.Thefourthcolumnshowstheradiocontinuumluminosity,adoptedfromEvansetal.(2006). Thefifth
column,alsoadoptedfromEvansetal.(2006),representstheaccretionefficiency(inEddingtonunits)derivedfromX-rayobservations of
thecoresoftheAGN(theupperlimitsareobtainedassumingNH =1024 atomscm−2;seeEvansetal.2006fordetails). Thesixthcolumn
showsthestellarageofthesourcebasedonfittingstellarpopulationsynthesismodelstotheopticalspectraofthesources(encompassing
the Hα portion of the spectrum; see Tab. 5 in Buttiglione et al. 2009). The seventh column shows the stellar mass derived using the
2MASSK-bandluminosityandstellarage(where available)followingDroryetal.(2004; seetext fordetails). Thesecond tolastcolumn
showstheblackholemass,adoptedfromEvansetal.(2006),andthelastcolumnreportsthemoleculargasmassobtainedfromCO(1→0)
observations (see ⁀tab:co ) using a conversion factor of α = MH2/L′CO = 1.5 (K km s−1 pc2)−1, and assuming H0 = 70, ΩM = 0.3,
ΩΛ=0.7. ThehorizontallinesseparateSeyferts(i.e.HERAGN;top)andLINERs(i.e.LERAGN;bottom).
TABLE 2
Summaryof observationswith the CARMAinterferometer.
Source RA(J2000) DEC(J2000) configuration obs.freq. On-source beam rms/channel
[GHz] time[hr] [arcsec] [mJy]
3C296 141652.94 +104826.50 D&E 112.460 14.1 6.4”×5.5” 2.1
3C321 153143.45 +240419.10 E 105.079 4.0 9.8”×7.2” 3.8
3C33 010852.86 +132013.80 D&E 108.746 14.8 8.7”×6.3” 1.8
3C403 195215.80 +023024.47 E 108.849 5.7 9.0”×6.9” 4.7
3C452 224548.77 +394115.70 D&E 106.634 15.6 6.8”×5.1” 1.8
3C465 233829.52 +270155.90 E 111.914 8.9 9.3”×6.5” 3.0
3C66B 022311.41 +425931.38 E 112.790 4.4 8.7”×6.3” 5.1
3C98 035854.43 +102603.00 E 111.914 4.8 9.3”×6.7” 3.7
3C83.1B(NGC1265) 031815.86 +415127.80 D&E 112.241 13.5 4.5”×3.7” 2.1
NGC6109 161740.54 +350015.10 D&E 111.957 10.6 5.0”×4.0” 2.7
NGC6251 163231.97 +823216.40 E 112.526 6.53 9.0”×8.2” 3.5
4 Smolˇci´c & Riechers
Fig.1.— Spectroscopic diagnostic diagrams for our 3CRR (filled dots) sources that separate AGN into LINERs and Seyferts. The
emission line fluxes have been taken from Buttiglione et al. (2009; see their Tab. 1). The regions (separated following Kauffmannetal.
2003; Kewleyetal. 2001, 2006) that are occupied by star forming(SF), composite (Comp.) and AGN (Seyfert and LINER) galaxies are
indicatedinthepanels. Each3CRRgalaxyisalsolabeled.
TABLE 3
Molecular gasproperties
Source z Scont zCO ∆vFWHM ICO(1→0) L′CO
[mJy] [kms−1] [Jy kms−1] [K kms−1 pc2]
3C33 0.060∗ 31.8±0.4 0.060 400±200 1.5±0.6 (2.5±1.0)×108
3C66B 0.022∗ 113.6±0.8 – – <2.4 <5.2×107
3C83.1B(NGC1265) 0.027∗ 40.5±0.4 – – <1.2 <3.8×107
3C98 0.030∗ 8.0±0.4 – – <1.3 <5.2×107
3C296 0.025∗ 144.4±0.4 – – <1.4 <3.8×107
3C321 0.097∗ 9.1±0.6 0.097 320±70 5.0±0.9 (2.2±0.4)×109
3C403 0.059∗ 4.6±0.5 0.059 350±100 2.8±0.7 (4.4±1.1)×108
3C452+ 0.081∗ 44.7±0.6 – – – <5.4×108
3C465 0.030∗ 20.2±0.1 – – <0.3 <1.3×107
NGC6109 0.0296∗∗ 17.7±0.3 0.0301 230±50 2.2±0.4 (8.8±1.7)×107
NGC6251 0.0244∗∗ 624.9±0.6 – – <1.9 <5.0×107
CenAe 0.0008∗∗ – – – – 9.4×107
3C272.1(M84)b,c 0.004∗ – 0.0028 200 1.8±0.6 (6.2±2.1)×105
3C274(M87)b 0.004∗ – 0.0035 200 20±2 (1.1±0.1)×107
3C31a 0.017∗ – 0.0169 450 27±2 (3.4±0.2)×108
3C449b 0.017∗ – 0.0169 500 6±1 (7.6±1.3)×107
3C84(NGC1275)a 0.018∗ – 0.0176 200 104±1 (1.43±0.01)×109
3C264b 0.022∗ – 0.02 200 3.5±0.7 (6.2±1.2)×107
3C338d 0.032∗ – 0.030 – 0.495 2.0×107
3C405a 0.0565∗∗ – – – <1.5 <2.2×108
3C388 0.091 – – – – –
Thecolumnsshow thesource, itsredshift,the observedcontinuum fluxdensity(Scont), theredshiftbased ontheCO(1→0)emissionline
(zCO),thelinewidthathalf-maximum(vFWHM),theCOlineintensity(ICO(1→0)),andluminosity(L′CO;seeeq.4inEvansetal.2005).
ForsourcesinwhichtheCOlinewasnotdetected wereport3σ upperlimits,computed assuming∆vFWHM=300 kms−1.
∗ adopted fromButtiglioneetal.(2009)
∗∗ adopted fromEvans etal.(2006)
+ Duetothestrongcontributionofcomplex,steeplyslopedmmcontinuum emissionfromextended radiojetstothemmemissionofthis
source, the continuum was fitted over only 33 channels, where the jet contribution is estimated to be small after deconvolution. Due to
thisuncertainty, however,wedonotconsiderthissourceinourstatisticalanalysis.
a zCO,∆vFWHM andICO(1→0) areadoptedfromEvansetal.2005. L′CO wascomputedusingthecosmologyadoptedhere.
b zCO,∆vFWHM andICO(1→0) aretaken fromFlaquer etal.(2010). Giventhattheir observations wereconducted withtheIRAM30m
telescope, wetake1K=4.95Jy,andcomputeL′ usingthecosmologyadoptedhere.
c tentative detection CO
d ICO(1→0) adopted fromLeonetal.(2001).
e adoptedfromEckartetal.(1990), andscaledtothecosmologyusedhere.
CO in Radio Galaxies 5
Fig. 2.—CO(1→0)spectra(histograms)ofthe11radiogalaxiesinoursampleobservedwithCARMA.Thedataareshownataresolution
of31.25MHz(83–89 kms−1). TheblackcurvesareGaussianfitstothelineprofilesandrest-frame2.6mmcontinuum emission.
6 Smolˇci´c & Riechers
TABLE 4
Averageproperties of the z<0.1 high-andlow-excitation radioAGN
AGN L178−MHz stellarage M∗ MBH MH2
type [W/Hz/srad] [Gyr] [M⊙] [M⊙] [M⊙]
HERAGN (7.2±4.9)×1024 6.3±1.6 (1.7±0.7)×1011 (2.5±1.0)×108 (2.9±1.2)×108
LERAGN (2.5±1.1)×1024 10.2±1.4 (2.4±1.4)×1011 (1.3±0.2)×109 (4.3±1.9)×107,∗
∗ The given limit was computed excluding the tentative CO detection in 3C 272.1 (see ⁀tab:co ). Including the gas mass for this source
yieldsanaverageof(1.8±1.5)×107 M⊙.
Fig.3.— Distribution of physical properties of high- and low-excitation radio AGN (HEARGN and LERAGN, resp.), drawn from
⁀tab:physprops (excluding the tentative COdetection in3C 272.1). Average values (given in⁀tab:averageprops ) areindicated by vertical
dot-dashedlines
CO in Radio Galaxies 7
have a factor of ∼ 3 higher radio continuum luminosi-
ties, significantly higher accretion efficiencies, but about
an order of magnitude lower mass central black holes.
Furthermore, their host galaxies have about a factor of
1.5 younger stellar populations and stellar masses, but
about a factor of ∼ 7 higher molecular gas masses. As
discussed in the next section, this is consistent with the
idea that high- and low- excitation radio AGN form two
physicallydistinctpopulationsofgalaxiesthatreflectdif-
ferent phases of massive galaxy formation.
4. DISCUSSIONANDSUMMARY
Our main result is that HERAGN have systemati-
cally higher molecular gas masses (a factor of ∼ 7; see
⁀tab:averageprops ), compared to LERAGN. Flaquer et
al. (2010) have found a similar trend by dividing their
sample (∼ 50 radio AGN observed with the IRAM 30m
telescope,partiallyoverlappingwithoursample)intoFR
class I and II objects. They find that the molecular gas
mass in FR IIs is a factor of ∼ 4 higher than that in
FRIs. TheFRclasscanbe takentoroughlycorrespond
to the low- and high- excitation classification.8 Flaquer
et al. (2010) have, however, concluded that the system-
aticdifferencestheyfindarelikelyaresultofaMalmquist Fig. 4.— CO vs. FIR luminosity for our local AGN sources
bias, i.e. simply due to a systematically higher redshift detectedwithIRAS.ThelinesrepresenttheL′CO−LFIRcorrelation
derivedbyRiechersetal.(2006).
of their FR-II sources. Although our HERAGN lie on
average at a slightly higher redshift, compared to our radio AGN sample, are in excellent agreement with the
LERAGN(0.046vs.0.030,resp.) in the following wear- systematic differences in various properties of high- and
gue that the systematic differences we find in molecular low- excitation radio AGN, both on pc- and kpc- galaxy
gas mass are not due to a Malmquist bias. scales (see Sec. 1 and⁀tab:physprops ).
Mori´c et al. (2010) have shown that the redshift dis-
We find that, on average, HERAGN have lower stel-
tributions of carefully selected samples of radio-selected
lar masses and stellar ages compared to LERAGN (see
LINERs and Seyferts are approximately the same (see
⁀tab:averageprops ; see also Smolˇci´c 2009). This is con-
their Fig. 6). This eliminates Malmquist bias from their
sistent with HERAGN and LERAGN being green val-
results. They find that the detection fractionin the FIR
ley and redsequence sources,respectively. Furthermore,
is significantly lowerfor LINERs than forSeyferts (6.5%
we show that HERAGN have on average higher radio
vs. 22%, resp.) in their sample. Assuming that the
luminosities than LERAGN, consistent with the results
star formation law parameterized by L′ (as a proxy
CO presented in Kauffmann et al. (2008). Kauffmann et al.
for total gas mass) and L (as a proxy for star forma-
FIR have shown that the fraction of radio AGN with strong
tionrate;e.g.,Kennicutt1998;Solomon&VandenBout
emission lines in their spectra significantly rises beyond
2005;Bigieletal.2008),onaverage,correctlyrepresents ∼ 1025 W Hz−1. In general, the comparison of the
the star formation properties of these samples (as con-
black hole and host galaxy properties inferred for our
firmedbytheCO/FIRluminositiesoftheIRASdetected
21z <0.1AGNwithmuchlargersamplesofradioAGN
sourcesanalyzedhere;seeFig.4),the loweraverageFIR
(Kauffmann et al. 2008;Smolˇci´c2009)suggests that our
luminosityinlow-excitationsources(i.e.LINER)implies
AGNsampleisrepresentativeofhigh-andlow-excitation
lower gas masses than in high-excitation (i.e. Seyfert)
radio AGN in the nearby universe.
types of galaxies. A similar result is obtained based on
From the average stellar masses that we infer for our
average (optically derived) star formation rates9, sug-
high- and low excitation sources we extrapolate that
g3elsotwinegrtthhaatntihnosSeeyinferLtIsNiEnRasraerdeshbiyft-ambaotuctheadfascatmorploef. they occupy ∼ 3 × 1013 M⊙ and ∼ 5 × 1014 M⊙ ha-
los, respectively (e.g. Behroozi et al. 2010; Moster et al.
These findings suggest that the systematic differences
2010). Compared to the systematic molecular gas mass
in molecular gas mass in high- and low-excitation radio
difference, this yields an even more dramatic discrep-
AGN are physical, and not due to Malmquist bias.
ancy of more than 2 orders of magnitude in the average
The systematically higher molecular gas masses that
molecular gas fractions in HE- (∼10−5) and LE-RAGN
wefindinHERAGN,relativetoLERAGNinourz <0.1
(∼9×10−8). Thediscrepancyremainssignificant(about
anorderofmagnitude)ifthe averagegas-to-stellarmass
8 Almost all FR I - low power - radio galaxies are LERAGN,
fraction(whichcan be interpretedas starformationeffi-
while optical hosts of FR IIs, which are typically more powerful
than FR Is (Fanaroff & Riley 1974; Owen 1993; Ledlow & Owen ciency) is considered.
1996), usually have strong emission lines. Note however that the On small scales, the average black hole accretion effi-
correspondencebetweentheFRclassandthepresenceofemission
ciencies in HE- and LE-RAGN suggest different super-
linesisnotone-to-one.
9Mori´cetal.(2010)derivedstarformationratesforeachgalaxy massive black-hole accretion mechanisms (standard disk
intheirsampleviastellarpopulationsynthesismodelfittingtothe accretion of cold gas in HERAGN vs. Bondi accretion
SDSSphotometryofthehostgalaxy(seealsoSmolˇci´cetal.2008). of hot gas in LERAGN; see Evans et al. 2006; Hard-
8 Smolˇci´c & Riechers
castle et al. 2006). Furthermore, the higher black hole sightfuldiscussions. Theresearchleadingtotheseresults
masses in LERAGN suggest a later evolution stage of hasreceivedfundingfromtheEuropeanUnion’sSeventh
their host galaxies,comparedto that of HERAGN. This Framework programme under grant agreement 229517.
is further strengthened by the higher stellar masses in DRacknowledgessupportfromNASAthroughanaward
LERAGN, as well as older stellar ages, and less mas- issued by JPL/Caltech, and from NASA through Hub-
sive gas reservoirs. In the blue-to-red galaxy formation ble Fellowship grant HST-HF-51235.01 awarded by the
picture, blue gas rich galaxies are thought to transform Space Telescope Science Institute, which is operated by
intoread-and-deadgas-poorgalaxies,the stellarpopula- the Association of Universities for Research in Astron-
tions in the host galaxies of HERAGN are expected to omy,Inc., forNASA, under contractNAS 5-26555. Sup-
be youngerandhavelowermasses,while theirmolecular portforCARMAconstructionwasderivedfromtheGor-
gas reservoirs– fueling further stellar mass growth– are don and Betty Moore Foundation, the Kenneth T. and
expected to be higher than those in LERAGN. This is Eileen L. Norris Foundation, the James S. McDonnell
in very good agreement with the results presented here. Foundation, the Associates of the California Institute of
Thus, in summary, our results strengthen the idea that Technology,the UniversityofChicago,the states ofCal-
low- and high-excitationradio AGN form two physically ifornia,Illinois, and Maryland,and the NationalScience
distinct galaxy populations that reflect different stages Foundation. Ongoing CARMA development and oper-
of massive galaxy formation. ations are supported by the National Science Founda-
tionunderacooperativeagreement,andbytheCARMA
partner universities.
The authors thank F. BertoldiandK.Knudsenfor in-
REFERENCES
Allen,S.W.,Dunn,R.J.H.,Fabian,A.C.,Taylor,G.B.,& Fanidakis,N.,Baugh,C.M.,Benson,A.J.,Bower,R.G.,Cole,
Reynolds,C.S.2006,MNRAS,372,21 S.,Done,C.,&Frenk,C.S.2010,MNRAS,1547
Baldwin,J.A.,Phillips,M.M.,&Terlevich,R.1981,PASP,93,5 Ocan˜aFlaquer,B.,Leon,S.,Combes,F.,&Lim,J.2010,A&A,
Behroozi,P.S.,Conroy,C.,&Wechsler,R.H.2010,ApJ,717, 518,A9
379 Hardcastle,M.J.,Evans,D.A.,&Croston,J.H.2006,MNRAS,
Bell,E.F.,etal.2004a,ApJ,608,752 370,1893
Bell,E.F.,etal.2004b,ApJ,600,L11 Jackson,N.,&Rawlings,S.1997, MNRAS,286,241
Borch,A.,etal.2006, A&A,453,869 Kauffmann,G.etal., 2003,MNRAS,341,33
Bower,R.G.,Benson,A.J.,Malbon,R.,Helly,J.C.,Frenk, Kauffmann,G.,Heckman,T.M.,Best,P.N.2008, MNRAS,384,
C.S.,Baugh,C.M.,Cole,S.,Lacey, C.G.2006, MNRAS,370, 953
645 Kewley,L.J.,Dopita,M.A.,Sutherland,R.S.,Heisler,C.A.,
Brown,M.J.I.,Dey,A.,Jannuzi, B.T.,Brand,K.,Benson, Trevena,J.2001,ApJ,556,121
A.J.,Brodwin,M.,Croton,D.J.,Eisenhardt,P.R.2007,ApJ, Kewley,L.J.,Groves,B.,Kauffmann,G.,Heckman,T., 2006,
654,858 MNRAS,372,961
Buttiglione,S.,Capetti,A.,Celotti,A.,Axon,D.J.,Chiaberge, Leon,S.,Lim,J.,Combes,F.,&van-Trung,D.2001,QSOHosts
M.,DuccioMacchetto, F.,&Sparks,W.B.2009, andTheirEnvironments,185
arXiv:0901.1764 Mori´c,I.,Smolˇci´c,V.,Kimball,A.,Riechers,D.A.,Ivezi´c,Zˇ.,&
Buttiglione,S.,Capetti,A.,Celotti,A.,Axon,D.J.,Chiaberge, Scoville,N.2010,ApJ,724,779
M.,Macchetto, F.D.,&Sparks,W.B.2010,A&A,509,A6 Moster,B.P.,Somerville,R.S.,Maulbetsch, C.,vandenBosch,
Buttiglione,S.etal.,2011,A&A,525,A28 F.C.,Maccio`,A.V.,Naab,T.,&Oser,L.2010,ApJ,710,903
Croton,D.J.,etal.2006,MNRAS,365,11 Peacock, J.A.1985,MNRAS,217,601
Downes,D.,&Solomon,P.M.1998, ApJ,507,615 Riechers,D.A.,etal.2006,ApJ,650,604
Drory,N.,Bender,R.,Feulner,G.,Hopp,U.,Maraston,C., Saripalli,L.,&Mack,K.-H.2007, MNRAS,376,1385
Snigula,J.,&Hill,G.J.2004,ApJ,608,742 Sijacki,D.,&Springel,V.2006,MNRAS,366,397
Eckart,A.,Cameron,M.,Rothermel,H.,Wild,W.,Zinnecker,H., Sijacki,D.,Springel,V.,diMatteo, T.,&Hernquist,L.2007,
Rydbeck, G.,Olberg,M.,&Wiklind,T.1990,ApJ,363,451 MNRAS,380,877
Evans,A.S.,Mazzarella,J.M.,Surace,J.A.,Frayer,D.T., Smolˇci´c,V.,etal.2008,ApJS,177,14
Iwasawa, K.,&Sanders,D.B.2005, ApJS,159,197 Smolˇci´c,V.2009,ApJ,699,L43
Evans,D.A.,Worrall,D.M.,Hardcastle,M.J.,Kraft,R.P.,& Smolˇci´c,V.,etal.2009,ApJ,696,24
Birkinshaw,M.2006,ApJ,642,96 SolomonP.M.,VandenBoutP.A.,2005, ARA&A,43,677
Faber,S.M.,etal.2007, ApJ,665,265 Willott,C.J.,Rawlings,S.,Blundell,K.M.,Lacy,M.,&Eales,
S.A.2001, MNRAS,322,536