Molecular Gas Kinematics and Star Formation Properties of the Strongly-Lensed Quasar Host Galaxy RXS J1131-1231

2.9 MB·

by  T. K. Daisy Leung

ACCEPTEDTOTHEAPJ
PreprinttypesetusingLATEXstyleemulateapjv.5/2/11
MOLECULARGASKINEMATICSANDSTARFORMATIONPROPERTIESOFTHESTRONGLY-LENSED
QUASARHOSTGALAXYRXSJ1131 1231
−
T.K.DAISYLEUNG,DOMINIKA.RIECHERS,ANDRICCARDOPAVESI
DepartmentofAstronomy,SpaceSciencesBuilding,CornellUniversity,Ithaca,NY14853,USA;[email protected]
AcceptedtotheApJ
ABSTRACT
We report observations of CO(J=2 1) and CO(J=3 2) line emission towards the quadruply-
→ →
7 lensed quasar RXSJ1131 1231 at z=0.654 obtained using the Plateau de Bure Interferometer (PdBI)
−
1 and the Combined Array for Research in Millimeter-wave Astronomy (CARMA). Our lens modeling
0 shows that the asymmetry in the double-horned CO(J=2 1) line profile is mainly a result of differ-
2 ential lensing, where the magnification factor varies from→3 to 9 across different kinematic compo-
∼ ∼
n nents. Theintrinsicallysymmetriclineprofileandasmoothsource-planevelocitygradientsuggestthat
a the host galaxy is an extended rotating disk, with a CO size of R 6kpc and a dynamical mass of
CO
J M 8 1010M . We also find a secondary CO-emitting source ne∼ar RXSJ1131 1231 whose loca-
6 tiodnyni∼sco×nsistentw(cid:12)iththeoptically-faintcompanionreportedinpreviousstudies. Th−elensing-corrected
2 molecular gas masses are M =(1.4 0.3) 1010M and (2.0 0.1) 109M for RXSJ1131 1231
gas (cid:12) (cid:12)
± × ± × −
andthecompanion,respectively. Wefindalensing-correctedstellarmassofM =(3 1) 1010M and
] ∗ (cid:12)
± ×
A a star formation rate of SFR =(120 63)M yr−1, corresponding to a specific SFR and star forma-
FIR (cid:12)
±
G tion efficiency comparable to z 1 disk galaxies not hosting quasars. The implied gas mass fraction of
∼
18 4% is consistent with the previously-observed cosmic decline since z 2. We thus find no evi-
h. ∼dence±forquenchingofstarformationinRXSJ1131 1231. Thisagreeswith∼ourfindingofanelevated
p M /M ratioof>0.27+0.11%comparedtothelo−calvalue,suggestingthatthebulkofitsblackhole
- BH bulge −0.08
o massislargelyinplacewhileitsstellarbulgeisstillassembling.
r Subject headings: infrared: galaxies – galaxies: high-redshift – galaxies: ISM – galaxies: evolution –
t
s quasars: individual(RXSJ1131 1231)–radiolines: ISM
a −
[

1 1. INTRODUCTION ies. Attemptstoextendthisrelationouttohigherredshifts,
v Many recent studies of galaxy evolution have been fo- beyond the peak epoch of star formation and AGN activ-
0 ity, have been made in recent years. These studies find
cused on investigating the interplay between star forma-
3 thathigh-zAGNhostgalaxiesdonotappeartofollowthe
tionandactivegalacticnucleus(AGN)activityacrosscos-
8 same M M relation as nearby spheroidal galaxies.
mic epochs (e.g., Di Matteo et al. 2005; Alexander et al. BH bulge
7 −
(e.g., Walteretal.2004;Borysetal.2005;McLureetal.
0 2005;Hopkinsetal.2006;Coppinetal.2008;Pageetal.
2006;Pengetal.2006;Riechersetal.2008;Coppinetal.
. 2012; Simpson et al. 2012; Lamastra et al. 2013). It is
1 2008; Alexander et al. 2008). Yet, the M M rela-
currently not well-understood when and how the super- BH bulge
0 −
tion remains poorly-constrained at intermediate redshifts
7 massive black holes (SMBHs) and stellar populations of
due to the difficulty in separating the stellar component
1 present-day massive galaxies were assembled, but it is
contributingtotheopticalemissionfromthatofthebright
: clearthattheco-movingstarformationrateandtheblack
v AGN.Thisstemsfromthelimitedresolvingpower,which
hole accretion rate densities both increased substantially
Xi sincez> 3andreachedtheirclimaxatz 2,followedby restricts the dynamic range that can be achieved at posi-
∼ tionsneartheAGN.Stronggravitationallensingprovides
r a rapid decline toward z 0 (e.g., Hopkins & Beacom
a 2006; Madau & Dickinso∼n 2014). A leading explanation themagnificationnecessarytospatiallyseparatetheAGN
emission from the host galaxy stellar emission, signifi-
for this decline is the decrease in molecular gas content
cantly improving the dynamic range that can be achieved
andstarformationefficiency(e.g.,Erbetal.2006;Carilli
instudiesoftheirhostgalaxieswithcurrentinstruments.
& Walter 2013; Walter et al. 2014), but direct molecular
The quasar RXS J113151.62 123158 (hereafter
gasmeasurementsatintermediateredshift(0.2 < z < 1)
−
RXJ1131) at z =0.658 (Sluse et al. 2003, hereafter
that could confirm this explanation remain largely lim- s,QSO
S03) is a particularly well-suited source for studying
ited to spatially unresolved CO observations of a modest
the evolution of molecular gas properties in quasar host
sampleof 30ultra-luminousinfraredgalaxies(ULIRGs;
∼ galaxies and the connection between SMBHs and their
Combesetal.2011,2013).
host galaxies at intermediate redshift given its unique
Meanwhile, empirical scaling relations such as the
lensing configuration. The stellar emission in the host
M M relation (e.g., Magorrian et al. 1998; Häring
BH− bulge galaxyofRXJ1131islensedintoanEinsteinringof1.(cid:48)(cid:48)83
& Rix 2004) have been established locally, suggesting a
in radius that is clearly separated from the quadruply
co-evalgrowthbetweenlocalSMBHsandtheirhostgalax-
imaged quasar emission (Claeskens et al. 2006, hereafter
2 Leung,Riechers&Pavesi
C06). The foreground lens is an elliptical galaxy at terpretationofthissystemandtheconclusionofthispaper
z =0.295 (S03). Reis et al. (2014) report a high spin donotrelyonthisquantity.
L
parameter of a 0.9 for the moderate-mass black hole The MIRIAD package was used to calibrate the visibil-
in RXJ1131 (M∼ =8 107M ; Sluse et al. 2012), with ity data. The calibrated visibility data were imaged and
BH (cid:12)
×
an intrinsic bolometric luminosity of L =1.3 1045 deconvolved using the CLEAN algorithm with “natural”
bol,X
ergss−1 (Pooleyetal.2007). × weighting. This yields a synthesized clean beam size of
In this paper, we present CO(J=2 1) and 3.(cid:48)(cid:48)2 1.(cid:48)(cid:48)9atapositionangle(PA)of8◦forthelowerside-
CO(J=3 2) line observations and broad→band pho- band×image cube. The final rms noise is σ = 13.3 mJy
tometry s→panning rest-frame UV to radio wavelengths beam−1 over a channel width of 25MHz. An rms noise
towards RXJ1131 to study the properties of its molecular of σ=0.83mJybeam−1 is reached by averaging over the
gas, dust and stellar populations. In §2, we outline line-freechannelsinbothsidebands.
details of the observations and of our data reduction
2.2. PdBICO(J=2 1)
procedures. In§3,wereportresultsfortheCO(J=2 1)
→
andCO(J=3 2)emissionandbroadbandphotomet→ryin Observations of the CO(J=2 1) rotational line
RXJ1131. In§→4,wepresentlensmodelinganddynamical (ν =230.53800 GHz) towards R→XJ1131 were carried
rest
modeling of the CO(J=2 1) data and spectral energy outusingtheIRAMPlateaudeBureInterferometer(PdBI;
distribution (SED) modelin→g of the photometric data. In Program ID: S14BX; PI: D. Riechers). Based on the
§5, we discuss the ISM properties of the host galaxy of CARMA CO(J=3 2) line redshift of z =0.655,
CO(3−2)
→
RXJ1131 and compare them to other galaxy populations theCO(J=2 1)lineisredshiftedtoν =139.256GHz.
obs
→
at low and high redshift. Finally, we summarize the Two observing runs were carried out on 2014 Decem-
main results and present our conclusions in §6. We use ber 06 and 2015 February 05 under good weather con-
a concordance ΛCDM cosmology throughout this paper, ditions in the C and D array configurations, respectively.
with parameters from the WMAP9 results: H = 69.32 This resulted in 3.75 hours of cumulative six antenna-
0
kms−1 Mpc−1,Ω =0.29,andΩ =0.71(Hinshawetal. equivalent on-source time after discarding unusable visi-
M Λ
2013). bility data. The 2 mm receivers were used to cover the
redshifted CO(J=2 1) line and the underlying contin-
2. OBSERVATIONS →
uumemission, employingacorrelatorsetupthatprovides
2.1. CARMACO(J=3 2) aneffectivebandwidthof3.6GHz(dualpolarization)and
→
Observations of the CO(J=3 2) rotational a native spectral resolution of 1.95MHz ( 4.2kms−1).
line (ν =345.79599GHz) toward→s RXJ1131 at The nearby quasars B1127 145 and B11∼24 186 were
rest − −
z =0.658 were carried out with the Combined observedevery22minutesforpointing,secondaryampli-
s,QSO
Array for Research in Millimeter-wave Astronomy tude,andphasecalibration,andB1055+018wasobserved
(CARMA; Program ID: cf0098; PI: D. Riechers) in the as the bandpass calibrator for both tracks. MWC349 and
Darrayconfiguration. Thelinefrequencyisredshiftedto 3C279wereobservedasprimaryfluxcalibratorsfortheC
ν =209.10443GHzatthequasarredshift. Observations andDarrayobservations,respectively,yieldingcalibration
obs
werecarriedouton2014February02underpoor1.5mm accuracybetterthan15%.
weather conditions and on 2014 February 17 under good The GILDAS packagewasusedtocalibrateandanalyze
1.5mm weather conditions. This resulted in a total the visibility data. The calibrated visibility data were im-
on-sourcetimeof2.94hoursafterflaggingpoorvisibility aged and deconvolved using the CLEAN algorithm with
data. The correlator setup provides a bandwidth of 3.75 “natural”weighting. Thisyieldsasynthesizedcleanbeam
GHz in each sideband and a spectral resolution of 12.5 size of 4.(cid:48)(cid:48)44 1.(cid:48)(cid:48)95 (PA = 13◦). The final rms noise
MHz ( 17.9 kms−1). The line was placed in the lower is σ=1.45mJy×beam−1 over 10 MHz (21.5kms−1). The
GsidHezb.an∼dThweitrhadtihoeqluoacsaalrsosJc1i1ll2a7tor1t8u9ne(dfirtsot tνrLaOck∼) 2an1d6 acgoinntginuouvmeri3m.1a6gGeHatzνocofntli∼ne1-3fr9eGeHbaznidswpirdotdhu.ceTdhibsyyaievledrs-
3C273 (second track) were obser−ved every 15 minutes anrmsnoiseof0.082mJybeam−1.
for pointing, amplitude, and phase calibration. Mars
was observed as the primary absolute flux calibrator and 2.3. KarlG.JanskyVeryLargeArray(Archival)
3C279 was observed as the bandpass calibrator for both Our analysis also uses archival data of the 4.885GHz
tracks. radio continuum obtained with the Karl G. Jansky Very
Given that the phase calibrator used for the first track Large Array (VLA; Program ID: AW741; PI: Wucknitz).
wasfaintandwasobservedunderpoorweatherconditions Observations were carried out on 2008 December 29 un-
andthatthephasecalibratorusedforthesecondtrackwas derexcellentweatherconditionsintheAarrayconfigura-
far from our target source, the phase calibration is sub- tion for a total of 7 hours on-source time. The C-band
par, with an rms scatter 50◦ over a baseline length of receiverswereused∼withacontinuummodesetup,provid-
∼
135m. Wethusconservativelyestimateacalibrationac- ingabandwidthof50MHzforthetwoIFbandswithfull
∼
curacyof 40%basedonthefluxscaleuncertainties,the polarization. ThenearbyradioquasarJ1130 149wasob-
∼ −
gainvariationsovertime,andthephasescatteronthecal- servedevery10minutesforpointing,amplitude,andphase
ibrateddata. WethereforetreatthederivedCO(J=3 2) calibration. J1331+305wasobservedastheprimaryflux
→
lineintensitywithcautionandensurethatourphysicalin- calibrator, and J0319+415 was observed as the bandpass
MolecularGasKinematicsoftheStrongly-LensedQuasarHostGalaxyRXJ1131 1231 3
−
calibrator, yielding 10% calibration accuracy. We used
AIPStocalibratethe∼visibilitydata. Thecalibratedvisibil- TABLE1
OBSERVEDPROPERTIESOFRXJ1131ANDITSCOMPANION
ity data were imaged and deconvolved using the CLEAN
algorithm with robust=0, which was chosen to obtain a Parameter Unit Value
higherresolutionimagegivenhighSNR.Thisyieldsasyn-
z 0.6541±0.0002
thesizedcleanbeamsizeof0.(cid:48)(cid:48)49 0.(cid:48)(cid:48)35(PA=0.18◦)and CO(2−1)
afinalrmsnoiseofσ =13µJybe×am−1. ICO(2−1) Jykms−1 20.3±0.6
S a Jykms−1beam−1 8.12±0.30
CO(2−1)
2.4. HST(Archival) FWHMCO(2−1) b kms−1 179±9,255±28
FWHM c kms−1 220±72
CO(2−1)
WeobtainedanHSTimagetakenwiththeACSusingthe
I Jykms−1 35.7±6.9
CO(3−2)
F555W filter (V-band) from the Hubble Legacy Archive.
aPeakfluxdensityintheintensitymap.
The details of the observations can be found in C06. We
bFromfittingadoubleGaussiantotheobservedCO(J=2→1)spectrum
apply an astrometric correction to the optical image by (Figure1).
adoptingtheVLA5GHzmapasthereferencecoordinate cFromfittingadoubleGaussianwithacommonFWHMtothede-lensed
frame. We shift the latter to the east by 0.(cid:48)(cid:48)5963 in R.A. CO(J=2→1)spectrum(Figure5).
and+0.(cid:48)(cid:48)8372inDec.,whichisconsistentwiththetypical
astrometricprecision(1(cid:48)(cid:48) 2(cid:48)(cid:48))ofimagesfromtheHubble
Legacy Archive1. This −astrometric correction is critical
to avoid artificial spatial offsets between different emit-
ting regions and to carry out our lens modeling, in which
the absolute position of the foreground lensing galaxy is
basedonitscoordinatesinthehigh-resolutionopticalim-
age. TheVLAimageiscalibratedusingawell-monitored
phase calibrator, with absolute positional accuracy of 2
∼
mas. For this reason, the absolute alignment between the
VLA image and other interferometric images reported in
this paper are expected to have an astrometric precision
better than 0.(cid:48)(cid:48)1, modulo uncertainties related to the SNR FIG.1.— Continuum-subtracted spectrum (histogram) of CO(J=2→1)
andphaseinstability. emission towards RXJ1131, with a spectral resolution of 22kms−1. The
solidblacklineshowsadouble-Gaussianfittothelineprofile. Thevelocity
3. RESULTS scaleiswithrespecttoz=0.6541, whichcorrespondstothelinecenterof
RXJ1131basedonthede-lensedlineprofile(Figure5).Adetaileddiscussion
3.1. CO(J=2 1)Emission ofthiseffectispresentedin§4.1.2.
→
We detect CO(J=2 1) line emission towards the tegrated line emission in the image plane yields
background source RX→J1131 in the PdBI data at (cid:38)27σ 5.(cid:48)(cid:48)1 0.(cid:48)(cid:48)7 3.(cid:48)(cid:48)7 0.(cid:48)(cid:48)7, which is consistent with that ob-
± × ±
significance. Based on this measurement, we refine the tained by visibility-plane fitting within the uncertainties.
redshiftofRXJ1131toz =0.6541 0.00022. Theemis- SincethespatialdistributionoftheobservedCOemission
CO
sion is spatially and kinematically re±solved with a highly isunlikelytobefullydescribedbyasimpleGaussianand
asymmetric double-horned line profile as shown in Fig- appears to be a superposition of at least two components
ure 1. Fitting a double Gaussian results in peak flux (top left panel of Figure 2), we also fit two Gaussians to
densities of 75.3 2.6 mJy and 24.0 2.0 mJy, and a theintensitymap. Thisyieldsdeconvolvedsourcesizesof
FWHM of 179 ±9 kms−1 and 255 ±28 kms−1 for the 3.(cid:48)(cid:48)8 0.(cid:48)(cid:48)4 1.(cid:48)(cid:48)9 0.(cid:48)(cid:48)4 and 3.(cid:48)(cid:48)6 0.(cid:48)(cid:48)3 1.(cid:48)(cid:48)5 0.(cid:48)(cid:48)3, sepa-
twocomponents±,respectively. Thepe±aksareseparatedby rated±by ×2.(cid:48)(cid:48)2in±RAand 1.(cid:48)(cid:48)7±inDec×. The±deconvolved
∼ ∼
∆v =400 12kms−1. The total integrated line flux is source sizes of both models suggest that the gravitation-
sep
20.3 0.6Jy±kms−1. allylensedCOemissionismoreextendedthantheoptical
We±constructazerothordermomentmap,red/bluechan- “Einstein ring”, which has a diameter of 3.(cid:48)(cid:48)6 (i.e., the
∼
nel maps, and first and second moment maps, as shown “Einsteinring”formedbyCOemissionislikelytohavea
in Figure 2, using the uv-continuum subtracted data cube largerdiametercomparedtotheopticalone). Thisiscon-
over a velocity range of ∆v 750kms−1. The higher- sistent with the centroid position of the redshifted emis-
order moment maps are prod∼uced using unbinned chan- sion,whichisalongthequasararcseenintheopticalim-
nel maps with 3σ clipping. The peak flux density is age,andtheblueshiftedemission,whichisoffsetfurtherto
8.12 0.30 Jykms−1 beam−1 in the intensity map. Ob- theSE(toprightofFigure2). Therefore,theCO-emitting
serve±d properties of the CO(J=2 1) emission line are region in RXJ1131 is likely to be more extended than its
summarizedinTable1. → stellarandquasaremission.
The deconvolved source size FWHM obtained from We also place an upper limit on HNC(J=2 1) line
→
fitting a single two-dimensional Gaussian to the in- emissionintheforegroundgalaxyatz 0.295. Assuming
a typical line width of 300kms−1, thi∼s corresponds to a
1http://hla.stsci.edu/hla_faq.html 3σ limitof0.35Jykms−1 beam−1.
2 This redshift is derived by fitting a double-Gaussian to the de-lensed
spectrum (Figure 5) instead of the observed spectrum (Figure 1) to avoid 3.2. CO(J=3 2)Emission
biasesinourredshiftdeterminationduetodifferentiallensing(see§4.1.2). →
4 Leung,Riechers&Pavesi
FIG.2.—Topleft:overlayofthevelocity-integratedCO(J=2→1)emissiononanarchivalHSTV-band(F555W)image.Topright:sameastopleft,except
thecontoursarecolor-codedtorepresentthered-andblueshiftedemission,whichareextractedbyintegratingoverv∈[−19,357]kms−1and∈[−395,−19]
kms−1,respectively. Thecontoursinbothtoppanelsstartat3σandincrementinstepsof±3σ,whereσ=0.3mJybeam−1forthetopleftpanel,andσ=0.4
mJybeam−1(red)and0.5mJybeam−1(blue)forthetoprightpanel. Thecrossesdenotethelocationoftheforegroundgalaxyatz=0.295. Contoursforthe
first(bottomleft)andsecond(bottomright)momentmapsoftheCO(J=2→1)lineemissionareshowninstepsof50kms−1,and100kms−1,respectively.
Thesynthesisbeamsizeis4.(cid:48)(cid:48)4×2.(cid:48)(cid:48)0,atPA=13◦.
We detect CO(J=3 2) line emission towards infertheCO(J=2 1)lineintensity.
RXJ1131 in the CARMA→data at (cid:38)5σ significance. The Assuming that t→he spatial extent of CO(J=2 1) and
→
CO(J=3 2) spectrum appears to be consistent with a CO(J=3 2) is similar and therefore the emission is
→ →
double-peaked profile, as shown in Figure 3, where we magnified by the same amount, the measured line in-
over-plotspectraoftheCO(J=2 1)andCO(J=3 2) tensities correspond to a brightness temperature ratio
→ →
lines. We extract the peak fluxes and their corresponding of r =T /T =0.78 0.37. The
32 CO(J=3→2) CO(J=2→1)
±
uncertaintiesfortheblueandredwingindependently. We quoted error bar is derived by adding the uncertainties
find a peak line flux of 5.13 1.43 Jykms−1 beam−1 for associated with the CO line intensities and those from
thebluewing,indicatinga(cid:38)±3σ detectionforthiscompo- absolute flux calibrations in quadrature. This brightness
nentalone, andapeaklinefluxof11.45 1.99Jykms−1 temperature ratio is consistent with thermalized excita-
beam−1 fortheredwing,indicatinga 6±σ detection. We tionwithintheuncertainties,ascommonlyobservedinnu-
measure a line intensity of 35.7 6.9∼Jykms−1 (Table 1) clear regions of nearby ULIRGs and high-z quasars (e.g.,
by summing up fluxes over the±FWZI linewidth used to Weiß et al. 2007; Riechers et al. 2011b; Carilli & Walter
MolecularGasKinematicsoftheStrongly-LensedQuasarHostGalaxyRXJ1131 1231 5
−
2013), but also with the lower excitation seen in normal
star-forming disks (e.g., Dannerbauer et al. 2009; Carilli
&Walter2013;Daddietal.2015).
FIG.3.—CARMACO(J=3→2)lineprofile(solid)withoutcontinuum
subtractionisover-plottedonthecontinuum-subtractedPdBICO(J=2→1)
line profile (dashed). The velocity scale is the same as in Figure 1. The
spectralresolutionforCO(J=3→2)andCO(J=2→1)is36kms−1 and
22kms−1,respectively.
3.3. ContinuumEmission
No 1.4mm continuum emission is detected at the po-
sition of CO(J=3 2) down to a 3σ limit of 2.49mJy
beam−1. This is c→onsistent with the spectrum shown in
Figure3.
We detect 2.2mm continuum emission at an inte-
grated flux density of 1.2 0.2 mJy, with a peak flux of
S =799 88µJybeam−1±centered on the lensing galaxy
ν
±
(Figure 4). Slightly extended emission along the lensing
arc is also detected. This suggests that we detect emis-
sion in both the foreground and the background galaxy
and that the emission is marginally resolved along its
major axis. We subtract a point source model in the
visibility-planetoremovetheunresolvedpartoftheemis- FIG.4.—Top:overlayofthePdBI2mmcontinuumemissionontheopti-
calimage.Contoursstartandincrementinstepsof±3σ,whereσ2mm=0.082
sion, which we here assume to be dominated by the fore- mJybeam−1.Bottom:overlayoftheVLA5GHzcontinuumemissiononthe
groundgalaxy. Theemissionintheresidualmapcoincides opticalimage.Contourscorrespondto±2nσ,whereσ5GHz=13µJybeam−1
spatially with the lensing arc. We measure a flux den- andnisanintegerrunningfrom2to5. Radioemissiontowardsthefore-
groundradiocoreisdetectedat(cid:38)57σsignificance.Thesynthesisbeamsize
sity of Sν=0.39 0.08mJy for this residual component. is4.(cid:48)(cid:48)4×2.(cid:48)(cid:48)0,atPA=13◦forthePdBIobservations(top),and0.(cid:48)(cid:48)5×0.(cid:48)(cid:48)4
Thisfluxdensity±isconsistentwiththedifferencebetween (PA=0.18◦)fortheVLAobservations(bottom).
the integrated and the peak flux density measured in the
original continuum map ( 0.4 mJy). We therefore adopt resolved emission from the jets and the core of the fore-
S =0.39 0.12 mJy as t∼he best estimate for the 2mm ground elliptical galaxy as well as emission toward the
ν
continuum±fluxofthebackgroundgalaxy(RXJ1131). We backgroundquasar. Multiplepeaksareseenalongthearc
here quote a conservative error bar, which is derived by with their centroids coincident with the optical emission
adding the uncertainty associated with the flux density fromthequasar. Weextractthefluxdensitiesforthelens-
of the point-source model (δS =0.088mJy) with that of ing arc and the radio core in Table 2. We find a spectral
ν
the peak flux in the residual map (0.08mJy) in quadra- index of α2mm= 0.02 0.07 for the foreground galaxy
6cm − ±
ture. We caution that this does not account for the sys- and α2mm= 0.35 0.21 for the background galaxy by
tematicuncertaintiesofthede-blendingprocedure,where fitting6pcmower−-laws (±S να) to their continuum fluxes at
ν
∝
we have assigned 100% of the point source flux to the 5GHzand2mm. Thespectralslopederivedfortheback-
foreground galaxy. We report the peak flux in the orig- groundsourceisflatterthanthetypicalslopeofpuresyn-
inal map (S =799 88µJybeam−1) for the foreground chrotron emission (α 0.7; e.g., Andreani et al. 1993).
ν
± ∼−
galaxy, which is the best estimate possible at the resolu- This likely suggests that at least a fraction of the ob-
tion of our observations, but we acknowledge that a non- served 2mm emission arises from thermal dust emission.
negligiblecontributionfromthebackgroundsourcetothe This spectral slope would be even shallower if the back-
peakfluxcannotberuledout. ground source contributes to the unresolved fraction of
The VLA C-band continuum image in Figure 4 shows the 2mm flux. In this case, the 2mm flux of the fore-
6 Leung,Riechers&Pavesi
ground galaxy would be lower than the value reported Monte Carlo simulations, each of 500 iterations. We in-
hereandleadtoaslopesteeperthanα2mm 0.02,which corporate the decomposed 24µm data in our SED fitting
6cm ∼−
is flatter than that typical of elliptical galaxies. Assum- to provide some constraints on the Wien tail beyond the
ing a spectral slope of α 0.7 to account for syn- dust peak of the SED of RXJ1131. Details of the SED
∼−
chrotron radiation in RXJ1131, we expect a flux density modelingarepresentedin§4.5.
of S =0.122 0.004 mJy at 2mm. The flux excess of Extraction of the Herschel/SPIRE photometry at 250,
2mm
±
S2mm=0.27 0.08mJythereforelikelyarisesduetother- 350,and500µmwascarriedoutusing SUSSEXTRACTOR
±
maldustemission. within the Herschel Interactive Processing Environment
(HIPE;Ott2010)onLevel2mapsobtainedfromtheHer-
3.4. Photometry schelScienceArchive. Thesemapswereprocessedbythe
Wecompilemid-IR(MIR)tofar-IRbroadbandphotom- SPIRE pipeline version 13.0 within HIPE. The SUSSEX-
etry from various catalogs available on the NASA/IPAC TRACTOR task estimates the flux density from an image
convolved with a kernel derived from the SPIRE beam.
Infrared Science Archive (IRSA) in Table 2 with aper-
ture corrections when warranted. These data were ob- The flux densities measured by SUSSEXTRACTOR were
confirmed by using the Timeline Fitter, which performs
tained using the Cerro Tololo Inter-American Observa-
photometrybyfittinga2DellipticalGaussiantotheLevel
tory(CTIO)fortheTwoMicronAllSkySurvey(2MASS;
Skrutskieetal.2006), theWide-fieldInfraredSurveyEx- 1 data at the source position given by the output of SUS-
plorer(WISE;Wrightetal.2010),theInfraredAstronomi- SEXTRACTOR. Thefluxesobtainedfrombothmethodsare
consistentwithintheuncertainties.
calSatellite(IRAS;Neugebaueretal.1984),andtheMulti-
band Imaging Photometer (MIPS; Rieke et al. 2004) and
4. ANALYSIS
Mid-infrared Infrared Array Camera (IRAC; Fazio et al.
2004)on the SpitzerSpace Telescope. We retrievePBCD 4.1. LensModeling
(level 2) Spitzer/IRAC images from the Spitzer Heritage AttheangularresolutionoftheCO(J=2 1)data,the
Archive and perform aperture photometry on the channel source is resolved (cid:38)2 resolution elements.→Given the ex-
1imagetoextractthefluxdensityat3.6µmsinceitisnot tentofthelensedemission(seeFigure2),thisimpliesthat
availablefromtheIRSAarchive. we do not resolve structures (e.g. knots and arcs) of the
The emission in the IRAC images is slightly extended. lensed emission in our CO(J=2 1) data. Nevertheless,
We thus use an HST image ( 0.(cid:48)(cid:48)07 resolution) to deter- the high spectral resolution of th→ese data provides kine-
∼
mine the origin of their centroids, all of which are found maticinformationonspatialscalessmallerthanthebeam
to be centered at the position corresponding to the lensed (see Figure 2). Hence, we reconstruct the intrinsic line
emission from the background galaxy. To recover the profileandsource-planevelocitystructurebycarryingout
diffuse background emission, we subtract a point source aparametriclensmodelingoverdifferentchannelslicesof
model centered on the lensing galaxy, using the average theinterferometricdatausingourlensingcodeUVMCMC-
FWHMfoundbyfittingaGaussianprofiletoseveralfield FIT(Bussmannetal.2015a;seeBussmannetal.2015bfor
starswiththeIMEXAMroutineofIRAF.Weperformaper- detailsofthecode). Ourapproachfollowsasimilarstrat-
ture photometry on the residual image to obtain decom- egy as Riechers et al. (2008), who reconstruct a source-
posed flux measurements of the background galaxy. The plane velocity gradient and constrain the gas dynamics in
photometry for the foreground galaxy is then obtained by thez>4quasarhostgalaxyofPSSJ2322+1944,whichis
subtractingthebackgroundemissionfromtheobservedto- also lensed into an Einstein ring configuration. To ensure
tal flux. The resulting photometry in Table 2 is obtained adequate SNRs for lens modeling, we bin the frequency
after performing an aperture correction described in the channelsbyafactoroffivetoproducesevenindependent
IRAC Instrument Handbook3 to correct for the fact that ∆v 105kms−1 channels (dashed line in Figure 5) that
theimagingwascalibratedusinga12(cid:48)(cid:48) aperture,whichis cove∼rthefulllinewidthof 750kms−1.
largerthantheaperture(5.(cid:48)(cid:48)8)weusedtoperformaperture We model the lens mas∼s distribution using a singular
photometry. isothermal ellipsoid (SIE) profile, which is described by
We fit a power-law spectrum to the decomposed IRAC fivefreeparameters: thepositionaloffsetinR.A.andDec.
photometrytodisentanglethebackgroundandforeground relative to an arbitrary chosen fixed coordinate in the im-
emission from the total flux observed in the MIPS 24µm age, the Einstein radius, the axial ratio, and the position
band. Thespectralindicescorrespondingtothebest-fitting angle. Positionaloffsetbetweentheforegroundgalaxyand
curvesareα= 1.8andα= 0.85forthelensinggalaxy thepre-definedcoordinateisinitializedusingtheVLAra-
and RXJ1131,−respectively. −The latter is consistent with diocontinuummap. Weimposeauniformpriorof 0.(cid:48)(cid:48)05
±
the mean 3.6 8µm spectral slope of α= 1.07 0.53 in both ∆R.A. and ∆Dec., motivated by the astrometry
− − ±
found for unobscured AGN (Stern et al. 2005). An ex- uncertainties in the VLA image as well as the uncertain-
trapolationofthefitto24µmyields33.96 0.01mJyand ties provided by previous SIE lens model (C06). We ini-
±
25.19 0.03mJyfortheforegroundgalaxyandRXJ1131, tialize the Einstein radius based on the model parameters
±
respectively. Theuncertaintiesarethestandarddeviations reportedbyC06andimposeauniformpriorusing 3σof
±
of the extrapolated fluxes obtained from two independent theiruncertainties. Thesourcesaremodeledusingellipti-
calGaussianprofiles,whichareparameterizedbysixfree
3http://irsa.ipac.caltech.edu/data/SPITZER/docs/irac/iracinstrumenthandbook/ parameters: thepositionaloffsetinR.A.andDec. relative
MolecularGasKinematicsoftheStrongly-LensedQuasarHostGalaxyRXJ1131 1231 7
−
TABLE2 TABLE3
PHOTOMETRYDATA LENSPARAMETERSCONSTRAINEDBY
MODELSOFSEVENVELOCITYCHANNELS
Wavelength Frequency FluxDensity Instrument
(µm) (GHz) (mJy) Parameters Medianvalues
Combined/Unresolved OffsetinRA ((cid:48)(cid:48)) 0.004±0.027
OffsetinDec ((cid:48)(cid:48)) 0.003±0.027
1.25 239834 1.009±0.090 CTIO/J-Band AxialRatio 0.56±0.16
1.65 181692 1.448±0.121 CTIO/H-Band PositionAngle (deg) 103±22
2.17 138153 2.064±0.160 CTIO/Ks-Band EinsteinRadiusa ((cid:48)(cid:48)) 1.833±0.002
3.4 88174.2 7.027±0.142 WISE/W1
3.6 83275.7 5.618±0.002 Spitzer/IRAC NOTE.—Parametersdescribingtheforeground
lensareobtainedbasedonthemedianinthepre-
4.5 66620.5 7.803±0.002 Spitzer/IRAC
liminarymodels(seetextfordetails). Allangu-
4.6 65172.3 8.872±0.163 WISE/W2 lar offsets are with respect to α=11h31m51.s44,
5.8 51688.4 10.720±0.005 Spitzer/IRAC δ=−12◦31(cid:48)58.(cid:48)(cid:48)3(J2000).
8.0 37474.1 14.470±0.004 Spitzer/IRAC
12 24982.7 21.960±0.425 WISE/W3 a This corresponds to mass of
12 24982.7 <400 IRAS M(θ<θE)=(7.42±0.02)×1011M(cid:12) within
22 13626.9 55.110±1.878 WISE/W4 theEinsteinradius.
24 12491.4 70.204±0.026 Spitzer/MIPS
25 11991.7 <500 IRAS
60 4996.54 <600 IRAS
100 2997.92 <1000 IRAS
250 1199.17 289.4±9.6 Herschel/SPIRE
350 856.55 168.2±8.6 Herschel/SPIRE
500 599.585 56.8±8.8 Herschel/SPIRE
1387.93 216 <2.49 CARMA
2152.82 139.256 1.23±0.22 PdBI Observed De-lensed
(RXJ only)
ForegroundLensingGalaxy(deblendedbands)
0.555 540167 0.056±0.006 HST-ACS/V-Band
0.814 368295 0.238±0.013 HST-ACS/I-Band
1.6 187370 0.539±0.041 HST-NICMOS(NIC2)/H-Band
3.6 83275.7 0.585±0.003a Spitzer/IRAC
4.5 66620.5 1.794±0.003a Spitzer/IRAC
5.8 51688.4 3.163±0.006a Spitzer/IRAC
8.0 37474.1 4.589±0.006a Spitzer/IRAC
2152.82 139.256 0.799±0.082 PdBI
61414 4.8815 0.866±0.027 VLA
BackgroundGalaxyRXJ1131(deblendedbands)
0.555 540167 0.009±0.004b HST-ACS/V-Band
0.814 368295 0.041±0.005b HST-ACS/I-Band
1.6 187370 0.133±0.004b HST-NICMOS(NIC2)/H-Band
3.6 83275.7 5.034±0.002 Spitzer/IRAC De-lensed De-lensed
4.5 66620.5 6.009±0.002 Spitzer/IRAC (RXJ1131) (Companion)
5.8 51688.4 7.557±0.003 Spitzer/IRAC
8.0 37474.1 9.881±0.004 Spitzer/IRAC
2152.82 139.256 0.39±0.12c PdBI
61414 4.8815 1.273±0.042 VLA
REFERENCES.—TheHSTphotometryisadoptedfromC06.
NOTE. —TheIRACphotometryforchannel1(3.6µm)isextracteddirectlyfromthe
imageandfromtheSpitzerHeritageArchiveforchannels2−4(4.5,5.8,and8.0µm). The
fluxuncertaintiesquotedforradioandmmobservations(PdBI,CARMA,andVLA)donot
includethosefromabsolutefluxcalibration.Allupperlimitsare3σ.
a FluxobtainedusingaperturephotometryaftersubtractingtheemissionofRXJ1131from
thetotalemission.
bAcontributionfromthequasarhasbeenremoved(seeC06),andthusthefluxdensitycor-
respondstothehostgalaxyonly.
cFluxextractedfromtheresidualmapaftersubtractingapoint-sourcemodel.ForSEDmod-
eling,weuseS2mm=0.27±0.08mJytoexcludesynchrotronemission(see§3.3).
to the lens, the intrinsic flux density, the effective radius,
theaxialratio,andthepositionangle. Thepositionofeach
source is allowed to vary between 1.(cid:48)(cid:48)5 (i.e., within the FIG.5.— Top: the full resolution CO(J=2→1) spectrum (yel-
Einsteinradius)andtheeffectiverad±iusisallowedtovary low histogram) and the binned spectrum (dashed line) with the seven
from0.(cid:48)(cid:48)01 2(cid:48)(cid:48). ∆v∼105kms−1channelsusedforlensmodeling.Thelightbluehistogram
showsthe“intrinsic”lineprofileofRXJ1131aftersubtractingacontribution
−
OurcodeusesanMarkovChainMonteCarlo(MCMC) fromitscompaniongalaxyandcorrectingforlensingusingthemagnification
approach to sample the posterior probability distribution factorµL asannotatedbythehorizontalbarshownaboveeachrespective
modelchannel. Bottom: the“intrinsic”lineprofileofRXJ1131(lightblue)
function (PDF) of the model parameters. In each model,
andofitscompanion(darkblue).They-axesareshownonalogscale.
we require a target acceptance rate of 0.25 0.5 and
∼ −
checkforchainconvergencebyinspectingtraceplotsand
8 Leung,Riechers&Pavesi
byrequiringthatthesamplesareobtainedbeyondatleast
an autocorrelation time. We thus employ 50,000 sam- TABLE4
∼ MAGNIFICATIONFACTORSOFVARIOUSKINEMATIC
ples as the initial “burn-in” phase to stabilize the Markov
COMPONENTSINCO(J=2→1)
chains (which we then discard) and use the final 5,000
∼
steps, sampled by 128 walkers, to identify the posterior. VelocityRange(kms−1) Source1µL Source2µL
Here,weidentifythebest-fitmodelandthequoteduncer-
346−260 6.7±2.5 7.2±5.6
taintiesusingthemedianandthe68%confidenceintervals 238−153 7.6±1.6
inthemarginalPDFs. 131−45 8.7±2.0
Wefirstobtain apreliminarylensmodelforeach chan- 24−-62 4.1±0.9
nel slice independently, where their lens parameters are -84−-170 4.2±0.6
-191−-277 4.3±2.4
allowedtovaryandareinitializedaccordingtotheafore-
-300−-385 3.1±0.9
mentioned way. We obtain the final model by repeating
the modeling over each slice but fixing their lens param- weightedaverage 4.4
median 5.5
eters to the overall median in the preliminary models, as
listed in Table 3. This ensures that all models share the NOTE. —Firstcolumncorrespondstotherest-framevelocity
rangestakenfromthecenterofanunbinnedchannel(seeFigure5).
samelensprofile. ThemagnificationfactorsinTable4are Eachrowcorrespondstoa(binned)channelsliceusedforlensmod-
determined by taking the ratio between the image plane eling.Source1isRXJ1131andsource2isitscompanion.
fluxandthesourceplanefluxofeachmodel. magnification factors listed in Table 4, which to first or-
Our model parameters in Table 3, describing the der takes into account the effect of differential lensing.
mass distribution of the lensing galaxy, are consis- This yields I =2.93 0.70 Jykms−1, where
CO(J=2→1)
tent (within the uncertainties) with that of the SIE the uncertainty includes those o±n the magnification fac-
model presented by C06. We find a mass of tors. Adopting the same brightness temperature ratio and
M(θ < θE)=(7.47±0.02)×1011M(cid:12) withintheEinstein αCO as used for the companion, this corresponds to a gas
radius. mass of M =(1.38 0.33) 1010 M , which implies a
gas (cid:12)
± ×
gasmassratioof 7:1betweenRXJ1131anditscompan-
4.1.1. InterpretationoftheSource-planeMorphology ∼
ion.
The reconstructed source locations, as represented by The spatial resolution of the data in hand is a few arc-
magenta ellipses in Figure 6, demonstrate an intrinsic ve- sec,whichimpliesthatdespitethehighSNRandspectral
locity gradient across the source plane, which is consis- resolution, constraints on the intrinsic sizes of the lensed
tent with a kinematically-ordered disk-like galaxy. Ad- galaxies are modest, and thus the magnification factors
ditional support to the disk conjecture can be found in may be under-predicted (see e.g., Bussmann et al. 2015b;
thedouble-hornedlineprofile(Figure1)andtheobserved Dyeetal.2015;Rybaketal.2015).
(imageplane)velocityfield(Figure2). Furthermore,C06
also find that the reconstructed source plane emission in 4.1.2. SpatialExtentandDifferentialLensing
optical-NIRisbest-reproducedusingan=1Sersicprofile. In the image-plane integrated line map shown in Fig-
WethusinterpretRXJ1131asadiskgalaxy. ure 2, the redshifted component is cospatial with the Ein-
A better fit is found for the lens model of the red-most stein ring that is seen in the optical image, with most of
channelifweaddasecondsourcecomponent(seetopleft itsapparentfluxoriginatingalongthelensingarc,whereas
panelinFigure6). Thisisconsistentwithpreviousresults centroid of the blueshifted emission is offset to the SE of
reported by Brewer & Lewis (2008, hereafter B08), who thelensingarc. ThissuggeststhattheCO-emittingregion
find an optically faint companion (component F in their inRXJ1131isextended. Tofurtherillustratethis,weshow
paper) 2.4kpcinprojectionfromtheAGNhostgalaxyin thechannelmapsof21.5kms−1 widthandaspatialspec-
∼
V-band, and with C06, who find evidence for an interact- tra map of 1.(cid:48)(cid:48)5 resolution in Figure 7 and Figure 8, re-
inggalaxynearRXJ1131. Spatially,theredvelocitycom- spectively. Thesefiguresshowthatredshiftedemissionis
ponent of the CO emission also consistent with this com- presenttothewest, peakingtowardthelensingarc(black
ponentF.ItisthereforelikelythatwedetectCO(J=2 1) crossesinFigure7),andshiftstotheeastwithdecreasing
→
emissioninacompaniongalaxy. velocity (blue wing). This is consistent with the source
We decompose the total line flux into two components: plane positions in our models and is suggestive of an ex-
one from RXJ1131 and the other from its companion. tendedCOemittingregion.
Sincethecompanionisonlydetectedinthered-mostchan- Previous studies of RXJ1131 find evidence for differ-
nel, wederiveitsintrinsicgasmassusingthebest-fitflux ential lensing across the HSTV-, I-, and H-bands, where
densitiesandmagnificationfactorsobtainedfromthemod- the magnification factor varies from 10.9 to 7.8 (C06).
elsofthischannel. Assumingabrightnesstemperaturera- Thisindicatesthattheemissionfromdifferentstellarpop-
tio of r21=1 between CO(J=2 1) and CO(J=1 0) ulations within the host galaxy have various spatial ex-
→ →
lines and a CO luminosity-to-H2 mass conversion fac- tents and positions with respect to the caustic. The best-
tor of αCO =0.8M(cid:12) (K km pc2)−1, we find a molecu- fit lens models obtained here for different CO channels
lar gas mass of M =(1.92 0.09) 109 M . For the showthatdifferentiallensingalsoplaysanimportantrole
gas (cid:12)
± ×
molecular gas mass in RXJ1131, we derive its intrinsic in the observed CO(J=2 1) emission, with a magnifi-
→
line flux over the FWZI linewidth using the respective cation factor (µ ) that varies from 8.7 to 3.1 across dif-
L
MolecularGasKinematicsoftheStrongly-LensedQuasarHostGalaxyRXJ1131 1231 9
−
FIG.6.—Best-fitlensmodelsofthePdBICO(J=2→1)dataindifferentvelocitychannels,aslistedinTable4. Top: eachpanelcorrespondstoachannel
mapofwidth107.5kms−1centeredontheindicatedvelocity.Theobservedemission(redcontours)isover-plottedatopthebest-fitmodel(grayscale).Bottom:
residualimagesobtainedbytakingaFouriertransformaftersubtractingthebest-fitmodelfromthedataintheuv-domain. Inallpanels,thelocationofthe
foregroundlensinggalaxyisindicatedbyablackdotanditscriticalcurveistracedbytheorangesolidline;locationsandmorphologies(half-lightradii)of
thereconstructedsourcesarerepresentedbymagentaellipses;thecausticcurvesarerepresentedascyanlines. Contoursstartat±3σandincrementinsteps
of3×2nσ,wherenisapositiveinteger. ThebeamofthePdBIobservationsisshowninthebottomrightcorner. Thereconstructedsource-planepositions,as
representedbythemagentaellipses,demonstrateanintrinsicvelocitygradientoftheCO(J=2→1)emissioninRXJ1131. Thebest-fitmodelofthered-most
channel(topleftpanel)containstwosourcecomponents—RXJ1131anditscompaniongalaxy.
10 Leung,Riechers&Pavesi
FIG.7.— ChannelmapsofthePdBICO(J=2→1)dataat22kms−1resolution. Blackcrossesindicatethepositionsofthelensedknots(AGNemission,
whichcorrespondtocomponentsABCDinC06). Thewhite-filledstarindicatesthepositionoftheforegroundlensinggalaxy(componentGinC06). Central
velocitiesareshownatthetopofeachmap.Contoursstartandincrementinstepsof±3σ.Thebeamisdenotedinthebottomrightpanel.