Retired galaxies: not to be forgotten in the quest of the star formation -- AGN connection

2.1 MB·English

by  G. Stasińska

Mon.Not.R.Astron.Soc.000,1–17(2014) Printed19January 2015 (MNLATEXstylefilev2.2)
Retired galaxies: not to be forgotten in the quest of the star
formation – AGN connection
5
1 G. Stasińska1⋆, M. V. Costa Duarte1,2, N. Vale Asari3, R. Cid Fernandes3, L. Sodré Jr.2
0
2 1LUTH, Observatoire de Paris, CNRS, Université Paris Diderot; Place Jules Janssen 92190 Meudon, France
2Departamento de Astronomia, Instituto de Astronomia, Geofísica e CiênciasAtmosféricas, Universidade de São Paulo, São Paulo SP, Brazil
n 3Departamento de Física - CFM - Universidade Federal de Santa Catarina, Florianópolis, SC, Brazil
a
J

5
1

Accepted ....Received;inoriginalform
]
A
G
ABSTRACT
.
h
p We propose a fresh look at the Main Galaxy Sample of the Sloan Digital Sky
- Surveybypackingthe galaxiesinstellarmass andredshiftbins.We showhowimpor-
o
tantitis to considerthe emission-lineequivalentwidths, inadditiontothe commonly
r
t used emission-line ratios, to properly identify retired galaxies (i.e. galaxies that have
s
stopped forming stars and are ionized by their old stellar populations) and not mis-
a
[ take them for galaxies with low-level nuclear activity. We find that the proportion
of star-forming galaxies decreases with decreasing redshift in each mass bin, while

1 that of retired galaxiesincreases.Galaxies with M⋆ >1011.5M⊙ have formed alltheir
v
stars at redshift larger than 0.4. The population of AGN hosts is never dominant
12 for galaxy masses larger than 1010M⊙. We warn about the effects of stacking galaxy
spectra to discuss galaxy properties. We estimate the lifetimes of active galactic nu-
8
clei(AGN)relyingentirelyondemographicarguments—i.e.withoutanyassumption
3
0 on the AGN radiative properties. We find upper-limit lifetimes of about 1–5 Gyr for
. detectable AGN in galaxies with masses between 1010–1012M⊙. The lifetimes of the
1 AGN-dominated phases are a few 108 yr. Finally, we compare the star-formationhis-
0
tories ofstar-forming,AGN andretiredgalaxiesas obtained by the spectralsynthesis
5
code starlight. Once the AGN is turned on it inhibits star formation for the next
1
: ∼ 0.1 Gyr in galaxies with masses around 1010M⊙, ∼ 1 Gyr in galaxies with masses
v around 1011M⊙.
i
X
Key words: galaxies: evolution – galaxies: statistics – galaxies: stellar content –
r galaxies : active.
a
1 INTRODUCTION wasfoundthatclassicalSeyfertgalaxiesconstituteabout1%
offieldgalaxies withabsolute magnitudesMB smaller than
It is nowadays considered that episodes of nuclear activity −20, and that the percentage of AGN hosts increases with
are of prime importance in the evolution of galaxies and luminosity, reaching 20% for galaxies with MB below −21
in the building up of the present-day Universe, due to the (usingH0=100kms−1).Sincethen,theSloanDigitalSur-
stronginterplaybelievedtoexistbetweentheactivegalactic
vey (SDSS, York et al. 2000) has provided a spectroscopic
nuclei (AGN) and star formation (SF) (see Fabian 2012 for
database of nearly one million galaxies which allows one to
a detailed review). Observational evidence of this interplay study the properties of galaxies up to a redshift z ∼ 0.4.
requires a large and complete sample of galaxies with ad-
Many studies have used SDSS data to address the nature
equate spectra and well defined criteria to classify galaxies
ofAGNhostsandtherelation betweennuclearactivityand
into appropriate categories. One of the pioneering studies
star formation (e.g. Kauffmann et al. 2003 – referred to as
in this direction is that of Huchra & Burg (1992), based
K03in thispaper–; Schawinskietal. 2007; Leeet al.2007;
on the CfA redshift survey (Huchra et al. 1983) which is a
Denget al. 2012; Lamassa & Heckman 2013). Somestudies
magnitude-limited spectroscopic survey of 2500 galaxies. It
have complemented optical data from the SDSS with data
atotherwavelengths(e.g.Kauffmannetal.2007fortheUV;
⋆ E-mail:[email protected]
2 Stasińska et al.
Constantinetal.2009forX-rays;Rosarioetal.2013forthe another important merit. It allows one to compare in each
mid-infrared). bin spectra encompassing regions of similar physical sizes
Most of these studies adopt the classification used by andtostudyapertureeffectsbycomparingbinsofdifferent
K03 based on the [O iii] λ5007/Hβ vs [N ii] λ6584/Hα redshifts.Thisisimportantsincethe3′′diameterSDSSfiber
emission-line ratio diagram introduced by Baldwin et al. doesnotcovertheentiregalaxies,especiallyatlowredshifts.
(1981,commonlyreferredtoastheBPTdiagram),inwhich Aperture effects are expected to be important particularly
galaxies are divided into star-forming (those below the di- for spirals, for which thecentral zones contain an old bulge
visory line introduced by K03), “pure” AGN (those lying and – possibly – an active nucleus while the outer zones
abovetheKewley etal. 2001 –K01– divisoryline),and the correspond to the star-forming disk. A full assessment of
“transition” or “composite” ones lying in-between. As em- apertureeffectsrequiresintegralfieldspectroscopyofrepre-
phasized by Cid Fernandes et al. (2010, 2011) this classifi- sentativesamplesofgalaxies.Afewpreliminarystudieshave
cation leaves aside a large number of galaxies that cannot been published (Brinchman et al. 2004; Kewley, Jansen &
be classified in the BPT diagram because they lack some Geller 2005) butinvestigations based on integral field spec-
of the required emission lines in their spectra, most often troscopy are just starting (Gerssen et al. 2012; Mast et al.
Hβ.Inaddition,it doesnot account forthepresenceof“re- 2013; Iglesias-Páramo et al. 2013). As will become evident
tired” galaxies (following the nomenclature introduced by in the course of the paper, for most aspects of our study
Stasińska et al. 2008). These systems have stopped form- we need not separate aperture affects and evolution in the
ing stars and are ionized by their old stellar populations interpretation of our results.
(namelythehotlow-massevolvedstars,orHOLMES),with- This paper is organized as follows. In Sect. 2, we de-
out a detectable contribution of an AGN. Finally, the K03 fine our master sample of galaxies, explain the preliminary
line to separate star-forming galaxies from AGN hosts has treatment applied to their spectra and divide galaxies in
been drawn empirically without physical justification. Con- mass and redshift bins. In Sect. 3 we discuss the BPT and
sequently,it is uncertain in the region of the BPT diagram WHAN emission line diagnostics including aperture effects
wheretheSFandAGNwingsmerge.Stasińskaetal.(2006, anddefineBPTandWHANsubsamples.InSect.4,wepro-
hereafter S06) showed that many galaxies below the K03 vide a detailed census of the galaxy emission-line spectral
line do in fact contain an AGN, and proposed a divisory types in our master sample. This allows us to proceed, in
line based on photoionization models to separate pure SF Sect. 5, to a direct estimation of AGN lifetimes. In Sect. 6,
galaxies from galaxies containinga detectable AGN.Itis to weproceed onestep furtherand analyze thestar formation
be noted that the new photoionization models of Dopita et histories of the different categories of galaxies in mass and
al.(2013)areinagreementwiththeS06onesasregardsthe redshift bins, in an attempt to better understandthe inter-
identification of pureSF galaxies. play between the star-forming, AGN and retired phases of
Cid Fernandes et al. (2011) proposed a different classi- galaxy evolution in the past 3–4 Gyr. The main outcomes
ficationofgalaxieswhichtakesintoaccounttheexistenceof of thethis study are summarised in Sect.7.
retired galaxies by considering the equivalent width of the Threeappendicescompletethispaper:AppendixAdis-
Hα line (in their WHAN diagram). In this paper, we show cusses selection effects in the SDSS Main Galaxy sample,
that the WHAN diagram leads to a much more compre- AppendixBwarnsabouttheeffectsofstackinggalaxyspec-
hensiveandreliable view ofgalaxy evolution thanthecom- tra to discuss galaxy properties, and Appendix C provides
monly used BPT diagram – although it also shares some of some additional material.
thedrawbacks of the BPT diagram. ThroughoutthepaperweconsideraΛCDMcosmology
Toproperly addresstheproblem ofthestarformation- with H0=70 km s−1 Mpc−1, Ωm =0.30, and ΩΛ =0.70.
AGN connection, it is important not only to use safe diag-
nostics of thenature of galaxies, but also todivide galaxies
accordingtotheirmasses,sinceitknownthatmassisafun-
damentalparameterforthepaceofgalaxyevolution(Cowie 2 THE DATABASE
et al. 1996; Heavens et al. 2004; Cid Fernandes et al. 2007;
2.1 Sample selection
Asari et al. 2007; Jimenez et al. 2007; Haines et al. 2007).
Anotheraspectnottobeneglectedisthatpastepisodes We consider the 7th Data Release of the Sloan Digital Sky
of nuclear activity cannot be detected – one can only rec- Survey (SDSS/DR7, Abazajian et al. 2009) which covers
ognize those that take place at the time when a galaxy is 9380 sq.deg. in the sky and presents fiber-fed spectroscopy
observed. Splitting the observational data in redshift slices in the wavelength range 3800–9200 Å with mean spectral
therefore can provide some clues on the past activity of resolution of λ/∆λ ∼1800 for nearly one million galaxies.
galaxies. Other factors such as morphology or environment Since we are interested in the relative populations of dif-
also play a role in the evolution of galaxies and have been ferent kinds of galaxies, we need a well-defined subsample.
addressed byvarious authors(e.g. Haineset al. 2007; Bam- Wetherefore restrict ourselves totheMain Galaxy Sample,
ford et al. 2009; Lietzen et al. 2011; Schawinski et al. 2010; which isacomplete flux-limitedsample of galaxies down to
2012, 2014), and will need to be reexamined in future pa- a magnitude mr = 17.77 (Strauss et al. 2002). Note that,
pers in the context of the present work. Here, we focus on because we work with mass-redshift bins, we do not need
just the following aspects: the correct assignation of galaxy a volume-limited sample, which would severely restrict the
spectral classes implying a proper identification of the re- usable redshift range. This issue will be further discussed
tired galaxies, and the importance of presenting the results below.
of any analysis in binsof mass and redshift. We select galaxies with z-band covering factor larger
Dividing the observational data into redshift bins has than 20% to reduce aperture effects in the determination
Retired galaxies and the star formation – AGN connection 3
of the stellar masses. We further impose a cut in redshift masses of the galaxies are much smaller than the width of
(z > 0.002) to garantee that luminosity distances are not themass bins.
dominated by peculiar motions (e.g. Ekholm et al. 2001).
Thisreducesoursampletoatotalof574,473objects.Addi-
tionalrestrictions havetobemade,dependingontheprop-
3 EMISSION-LINE DIAGNOSTICS OF THE
erties underdiscussion, and will be explicited below.
GALAXIES
3.1 Some generalities
2.2 Analysis of the data
As mentioned in the introduction, the vast majority of
Our work on the galaxy spectra makes use of the spectral
emission-line diagnostics of the ionization of galaxies are
synthesiscodestarlight(CidFernandesetal.2005). This
basedontheBPTdiagram wherepureSFgalaxiesarecon-
is a inverse stellar population synthesis code which decom-
sidered to lie below the K03 line and pure AGN above the
posestheobservedspectraincontributionsfromsimplestel-
K01line,thespacebetweenthetwolinesbeingoccupiedby
lar populations (SSPs) of given age and metallicity. The li-
so-called composite objects. Nevertheless, signs of AGN ac-
braryofSSPsisbasedontheBruzual&Charlot(2003)evo-
tivitycan bedetected well belowtheK03line, asshown by
lutionarymodelsofgalaxies,adoptingaChabrier(2003)ini-
S06.Ontheotherhand,“pure” AGNsliewellabovetheK01
tialmassfunction(IMF),“Padova1994” evolutionarytracks
line.NotethatKewleyetal.(2013) nowdefineanotherline
(Bertelli et al. 1994) and the STELIB spectral library (Le
to separate star-forming galaxies from galaxies containing
Borgne et al. 2003), as explained in detail in Asari et al.
an AGN, which actually is veryclose to theone of S06.
(2007).Spectralregionscontainingbadpixels,emissionlines
As shown by Stasińska et al. (2008), the BPT diagram
and the Na D doublet are not considered for the fits. The
ishowevernotabletodistinguishbetweengalaxiescontain-
stellar extinctionis anoutcomeof thestarlightfitting.It
ing a weak AGN and retired galaxies, i.e. galaxies ionized
is computed using a Cardelli, Clayton & Mathis (1989) ex-
bytheirHOLMES(whichcanalso explaintheionization of
tinctionlawwithRV =3.1.Thetotalmassesofstarspresent theextraplanar gas in spiral galaxies, see Flores-Fajardo et
inthegalaxies,M⋆,areobtainedaftercorrectingforthefrac- al.2011).Theexistenceofsuchgalaxieshassincethenbeen
tionofthegalaxyz bandluminosityoutsidetheSDSSfiber
evidencedbyintegral-fieldspectroscopyofgalaxies,showing
as explained in Cid Fernandes et al. (2005). The intensities
thattheextendedemissionseeninlow-ionizationgalaxiesis
of the emission lines were measured after subtracting the
inconsistent with a central point source for the ionization,
modelled stellar spectrum from the observed one. More de-
ruling out nuclear activity as the dominant source for the
tails can be found in Mateus et al. (2006), Stasińska et al.
emission (Sarzi et al. 2010; Kehrig et al. 2012; Papaderos
(2006), Asari et al. (2007) and Cid Fernandes et al. (2010).
et al. 2013; Singh et al. 2013). Cid Fernandes et al. (2011)
All the data used in this paper can be retrieved from the
proposed to use the equivalent width of Hα, EW(Hα), as a
starlight database1. function of [N ii]/Hα (the WHAN diagram) to distinguish
We generally did not make any restriction concerning
between galaxies containing an AGN and retired galaxies.
the signal-to-noise (S/N) ratio in the continuum since we
Thisistheseconddiagnosticdiagram wewillconsiderhere.
are mainly interested in the stellar masses, which are very
Anotheradvantage of using the WHAN diagram instead of
robustoutputsfromstarlight:asshowninCidFernandes
the BPT is that it requires only 2 lines, while the BPT
et al. (2005), the uncertainty in M⋆ is of 0.1 dex for a S/N diagram requires 4 lines. We must note, however, that nei-
ratioof5at4000Å,andobjectswithsmallerS/Nrepresent
ther the BPT diagram nor the WHAN diagram are able to
only 1.5 per thousand of the objects in our sample (3 per
identifyAGNsinlow-metallicitygalaxies(S06;Grovesetal.
thousand of the ones with redshift z > 0.2). Concerning
2006).
the line intensities, we may have to restrict the sample by
The BPT diagram can also be used to rank the metal-
imposing criteria related tothe quality of emission lines, as licitiesinthestar-formingwing,sincethe[Oiii]/[Nii]ratio
detailed below.
has been shown since Alloin et al. (1979) to be a metallic-
ityindicator.The[Nii]/Hαratiohasalsobeenshowntobe
stronglycorrelatedwithmetallicity(Storchi-Bergmannetal.
2.3 Mass and redshift bins
1994;VanZeeetal.1998)ingiantHiiregions,andcanthus
In order to convey a synthetic view of the realm of SDSS serve to rank metallicities in the SF region of the WHAN
galaxies we divide our samples of galaxies in bins of mass diagram. In the AGN/retired zone, however, its interpre-
and redshift. We consider mass bins of 0.5 dex for log M⋆1 tation is not straightforward since [N ii]/Hα both depends
between8.5and12.5. Concerningredshift,weconsiderbins on the heating power of the main ionizing source (AGN or
of∆z =0.05untilz=0.40.Theseredshiftlimitscorrespond HOLMES) and on thenitrogen enrichment.
to sampled galactic radii of 1.5 to 8.1 kpc for the adopted
cosmology.Thewidthsoftheadoptedmass-redshiftbinsare
a compromise between i) obtaining a significant number of 3.2 Aperture effects on the BPT and WHAN
objectsineachbintorobustlydefinemeanpropertiesandii) diagrams
definingasufficientnumberofbinsinthemassandredshift
An important thing to realize is that both the BPT and
ranges considered. Note that the errors in the individual
theWHANdiagramsaresensitivetoapertureeffects.More
explicitly, the galaxy classification inferred from these di-
1 http://casjobs.starlight.ufsc.br/casjobs/ agrams depends on the relative proportions of bulge (or
1 ThroughoutthepaperM⋆isexpressedinunitsofsolarmasses. spheroid) and disk sampled by theSDSSfiber.
4 Stasińska et al.
tinycoverage total coverage
case BPT WHAN BPT WHAN
a oldbulge+SFdisk L R C/SF AGN/SF
b AGN+SFdisk AGN AGN C/SF AGN/SF
c AGN+SFbulge+SFdisk C AGN C AGN/SF
d SFbulge+SFdisk SF SF SF SF
e oldbulge+olddisk/spheroid L R L R
Table1.ClassificationofgalaxiesusingtheBPTandWHANdiagrams,fortwoextremefibercoveringfactors.“C”standsfor“composite,
“L” stands for“LINER”,“R” standsfor“retired”.
The contingency table presented in Table 1 schemati-
cally shows the diagnoses that would be obtained from the
BPT diagram in its classical use (since Kewley et al. 2006)
and from the WHAN diagram for the five possible galaxy
configurations and two extreme fiber covering factors. We
notethatincasea,boththeBPTandtheWHANdiagram
may interpret the emission-line pattern as being a conse-
quenceofthepresenceofanAGN(since“composite” means
AGN+star-forming) whilethegalaxy containsnoAGNat
all! TheBPT, in addition,findsthreemoreinstances ofthe
presence of an AGN (in terms of LINER) while the galaxy
contains no AGN at all. These are cases a and b for a tiny
fibercoverage, and case e for a total fibercoverage.
3.3 The BPT and WHAN subsamples
From our parent sample, we definetwo subsamples: a sam-
ple based on the WHAN diagram (sample W), and a sub-
sample based on the BPT diagram (sample B). Sample W
is the least restrictive of the two. Since within the WHAN
diagram there is a continuity between emission-line galax-
ies and lineless galaxies in the sense that lineless galaxies
would befound at vanishingvaluesof EW(Hα),herewedo
notimpose acondition on S/Nonthelineintensities.This,
of course, results in a somewhat uncertain classification of
galaxieswithlowS/Nbutshouldnotbiastheglobalpicture.
Ontheotherhand,weremovegalaxieswheredefectsinthe
spectra at the wavelengths of Hα and [N ii] do not allow
even a coarse evaluation of the ratio of their intensities. To
do this, we do not consider the criterion of Cid Fernandes
et al. (2010), since we judge it to be too severe. Instead,
we measure the nebular velocity dispersion σgas for the Hα
and [N ii] lines and count the pixels within ±1σgas of the
peakofeachline.Wethenremovefromoursamplegalaxies
withmorethan25% badpixelsoneitheremission line.The
total numberof galaxies that were removed with this crite-
rionis64,802,leaving509,671galaxiesinsampleW.Sample
Bcontains 217,391 galaxies, which are selected by applying
to sample W the requirement that the S/N must be larger
than3forallthefourBPTdiagnosticlines:[Oiii]5007,Hβ,
[N ii]6584 and Hα.
Retired galaxies and the star formation – AGN connection 5
0.4 0.4
sampleB sampleW
Hβ 1 3 1 1 α)3.0 3 4 2 5 12
gOIII[]/−01 217391 012Nlog 1 2 18 4 gEWH(01..05 509671 012Nlog 2 12 160 323
0.3 lo −2 0 0.3 lo −2 0
log[NII]/Hα log[NII]/Hα
1 5 76 252 9 1 16 166 2581 470
4 14 177 1962 1108 7 4 19 324 4915 15361 394
0.2 0.2
z z
1 23 285 4892 12326 1244 3 1 26 343 8802 50135 16209 72
1 28 660 7627 33262 19648 1058 1 29 684 9685 66980 83825 9009 3
0.1 0.1
70 1459 12251 35543 40702 13541 432 73 1498 13283 51400 88248 38533 1868
2822 6659 7686 5798 3774 834 4 3007 7666 10741 10656 8789 2043 7
0.0 0.0
9 10 11 12 9 10 11 12
logM⋆[M⊙] logM⋆[M⊙]
Figure 1. BPT diagrams for sample B (i.e., a subset of Figure 2. WHAN diagrams for sample W (i.e., only re-
sampleWofobjectswithgoodS/NinallfourBPTlines) moving galaxies with bad pixels in [N ii]6594 or Hα) in
in mass-redshift bins. The number of objects contained mass-redshift bins. The number of objects contained in
ineachpanelisindicated.Thecolorscaleisinlogarithm each panel is indicated. The color scale is in logarithm
of the number of objects per pixel, as indicated in the of the number of objects per pixel, as indicated in the
color bar in the inset which shows the entire sample B color bar in the inset which shows the entire sample W
in the BPT diagram. The curves are the S06 delimita- in the WHAN diagram. The continuous grey lines indi-
tions between pure SF galaxies and galaxies containing catethedelimitationsbetweenSFgalaxies,galaxieswith
an AGN. strong and weak AGNs, and the dashed grey linemarks
thelimitforretiredgalaxiesaccordingtoCidFernandes
et al. (2011).
4 A DETAILED CENSUS OF GALAXY
EMISSION-LINE SPECTRAL TYPES FROM
3.4 The BPT and WHAN diagrams in mass and THE SDSS DATA
redshift bins
Cid Fernandes et al. (2010) discussed the proportions of
Fig. 1 show the BPT diagram for sample B in the (M⋆,z) purestar-forminggalaxies/AGNhosts/retiredgalaxiesfora
bins we have defined. The total number of objects in each volume-limitedsampleofSDSSgalaxies with emission lines
bin is indicated. The curves represent the S06 line to dis- without paying special attention to their masses and red-
tinguishpureSFgalaxiesfromAGNhosts.Asexpectedthe shifts. Here, we extend this discussion by evaluating these
lowest mass binsarepopulatedonlyforthelowest redshifts proportions in mass and redshift bins.
(because the sample is flux limited), and the highest mass The first thing to consider in such an endeavour is the
bins have very few objects below z = 0.1 (because of the questionofcompleteness. Ifobservationalselection removes
decrease of the galaxy mass function towards high masses, somekindofgalaxiesfromoursample,wemusttakethisinto
see e.g. Panter et al. 2004). We clearly see how – at any accountintheoverallcensus.Byworkingin(M⋆,z)bins,we
redshift–therightwingbecomesmoreandmorepopulated are less prone to discriminations against certain categories
with increasing stellar mass, leaving virtually no object in ofobjects(definedbytheircoloursorbytheiremission-line
theSF wing at thehighest masses. properies). In Appendix A we discuss in detail the bias ex-
Fig. 2 shows the WHAN diagram for sample W in the pected in each (M⋆,z) bin. It turns out that some bins are
(M⋆,z) bins. The grey lines indicate the delimitations be- definitelyfreefromcolourbias.Atthesmallestredshifts,the
tween the four categories of emission-line galaxies accord- binswithlogM⋆ >10arecomplete.Inthenextredshiftin-
ing to Cid Fernandes et al. (2011). In this diagram we also terval, bins with log M⋆ > 10.5 are complete, and so on.
see how the SF zone becomes gradually depopulated with All the subpanels where the sample is complete are flagged
increasing stellar mass. But here we understand that this withayellowbackgroundinFig.A1.AsshowninAppendix
change essentially benefits the category of retired galaxies A, adjacent bins are not much affected by colour bias. As
and not to that of AGN hosts. redshift increases, the limiting magnitude mr of the Main
6 Stasińska et al.
Galaxy Sample will reduce the number of complete mass strong and weak AGN in a single category) and retired
bins by playing against galaxies of smaller masses. What is galaxies (with or without emission lines) as a function of
problematicforourstudyisthattherecouldbeacolourdis- z in thedifferentmassbinsinoursample, forbothsamples
crimination against some kind of galaxies. A general prop- B and W.
erty of a surveylimited bymagnitude in ther band is that Let usfirst focus on thedifferences between Fig. 3 and
the galaxies that are more easily missed are red ones, since the top panel of Fig. 4. Both sets of plots show galaxies in
blue galaxies with same masses are more luminous, as seen massandredshiftsbinsforsampleB,theformerclassifying
inAppendixA.Thus,atthelowestredshift,thelogM⋆ <9 galaxies based on their position on the BPT and the latter
bin could be missing an important number of red galaxies. ontheWHANdiagram.Formasses10.5<logM⋆ <12,the
However, the selection bias will become less important for BPT classes tricks us into finding an increase in AGN ac-
9<logM⋆ <9.5,andwillprobablyvanishforlogM⋆ >10. tivityatsmallerredshifts.Whenclassifying thesame setof
In the next redshift bin, the selection bias will start being galaxies using the WHAN this trend disappears. By telling
noticeable for log M⋆ <10.5, an so on. apart AGN hosts from retired galaxies, the staggering con-
clusionisthatagreatpartofwhatisattributedtoAGNbe-
haviourontheBPTisinfactjustduetogalaxyretirement.
4.1 BPT-based demography Given thedispersion of galaxy typeswithin amass-redshift
bin, one needs to be extremely careful when working with
Before turning to the WHAN diagnostic diagram, it is in-
averagedgalaxypropertiesorstackedspectra(seeAppendix
structive to first carry out a demographic study using the
B), even when binning in stellar mass and redshift.
BPTdiagraminthewaypromotedbyK03,sincethisisstill
The bottom panel of Fig. 4 shows galaxies as classified
themostpopularwaytoseparategalaxiesintostar-forming
on the WHAN diagram for sample W, which is much more
and AGN hosts.
inclusive than sample B.
Figure 3 shows the fraction of galaxies of different
The general behaviour is that, as redshift decreases, in
emission-line spectral types for sample B as a function
each mass bin the proportion of SF galaxies decreases, the
of redshift for our differentmass bins,as obtained from the
proportionofretiredgalaxiesincreaseswhiletheproportion
canonical use of the BPT diagram, i.e. using the K03 and
ofAGNhostsrathertendstodecrease(atleastforlogM⋆ >
K01lines todistinguish between SF,“composite” andAGN
10.5). As was the case in Fig. 3, we see that in the Main
galaxies. In each panel we indicate the total number of ob-
Galaxy Sample SF galaxies dominate the whole population
jectsrepresented.Thelargersymbolsrepresent(M⋆,z)bins
of galaxies at any redshift for log M⋆ <10.5. On the other
that are judged free of bias (see Appendix A), and adja-
hand,retiredgalaxiesalwaysdominatethewholepopulation
centbinsareprobablylittleaffectedbybias.Weonly plot
for log M⋆ > 11.5. In the intermediate mass bins, retired
bins with a minimumof 5 galaxies.Weseethatforlog
galaxies dominate at the lowest redshifts. As was the case
M⋆ <9allthegalaxiesareofSFtype;AGNhostsarefound
fortheBPTdiagram,theredshiftbehaviourisqualitatively
only for masses larger than that. Until log M⋆ = 10, SF
what is expected from aperture effects alone. For example,
types still constitute the dominant population of galaxies.
at intermediate masses the increasing proportion of retired
However, as noted above, the Main Galaxy Sample may be
galaxiesasredshiftdecreasesisexpectedasthefibresamples
missingapopulationoflow-massredgalaxies,whoseimpor-
decreasingportionsofthestar-formingdisk.Thisbehaviour
tance is difficult to assess. Among this population, galaxies
can also be dueto galaxy evolution.
containing an AGN but not experiencing present-day star
Inspiteofthisuncertaintyin theinterpretation,in ab-
formation could perhaps exist. At higher values of M⋆ the
senceofanycolourbias,Fig.4wouldnicelydepictthedown-
tendency starts reversing and for log M⋆ > 11.5 we find
sizing paradigm in the local Universe. The vast majority of
thatthepopulationofgalaxiesisvastlydominatedbyAGN
galaxies with log M⋆ > 11.5 have formed all their stars at
hosts. We also note that the proportion of SF galaxies al-
redshiftlargerthan0.4.Galaxieswith10.5<logM⋆ <11.5
ways tends to decrease with decreasing redshift while the
graduallystopformingstarsbetweenz=0.4andthepresent
proportion of AGNs increases. Qualitatively, this is what is
time.GalaxieswithlogM⋆ <10.5stillformstarspresently.
expectedfrom mereapertureeffectswherethecontribution
However, as explained in Appendix A, we expect the Main
of the old bulge with respect to the star-forming disk in-
Galaxy Sample tomiss red galaxies for masses below acer-
creases as redshift decreases.
tainthresholdmasstoanextentthatwearenotabletoeval-
uate, but which is certainly more important for the lowest
masses. This means that the redshift evolution of the spec-
4.2 WHAN-based demography
traltypesofgalaxieswithmassesbelow1010.5M⊙cannotbe
Ascommentedbefore,theBPTdiagramdemandsfourlines obtainedfromtheSDSSMainGalaxySample(althoughwe
to be observed with good S/N, which selects against weak- will see in Sect. 5 that our spectral-type census is probably
line objects and leaves aside about half of the galaxies of correct down to1010M⊙).
our master sample. More importantly, the BPT diagram is In the mass and redshift bins that we consider free of
not able todistinguish between hosts of weak AGNand re- colour bias, thepopulation of AGN hosts is neverpredomi-
tired galaxies since both categories display similar line ra- nant(exceptatthehighestreshiftofthe11<logM⋆ <11.5
tios. Emission-line equivalent widths need to be considered bin where they outnumber the retired galaxies by a tiny
as well. margin). Note that we reach such a conclusion in spite of
Now we classify the galaxies according to the WHAN the fact that our definition of AGN host is less restrictive
diagram,whichovercomestheselimitationsoftheBPT.Fig- thantheonewhichisgenerallyused(i.e.thatofK03).Also
ure4showsthefractionofSFgalaxies,AGNhosts(putting noteworthy is the fact that we do not see an increase in
Retired galaxies and the star formation – AGN connection 7
Figure 3.Fractionofgalaxiesofdifferentemission-linespectral typesasafunctionofredshiftforthedifferent massbins.Thespectral
types areobtained fromthe canonical useof theBPT diagram. Bluetriangles: SF; green squares: AGN; cyan circles:composite. Large
symbolscorrespond to(M⋆,z)binsthatarejudgeddevoidofcolourbias(seetextandAppendixA).
Figure 4.Fraction ofgalaxiesofdifferentemission-linespectral types (bluetriangles: SF;greensquares :AGN(strong andweak); red
circles:retired(withorwithoutemissionlines)asafunctionofredshiftforthedifferentmassbinsasobtainedfromtheWHANdiagram.
Largesymbolscorrespondto(M⋆,z)binsthatarejudgeddevoidofcolourbias.Top: WHANclassificationofsampleB,whichimposesa
S/Nlimitto[Oiii]5007,Hβ,[Nii]6584,andHα.Bottom: WHANclassificationofsampleW,whichonlyexcludesgalaxieswhosespectra
around[Nii]6584orHαisunreliable.
the proportion of AGN hosts with decreasing redshift, as 5 AGN LIFETIMES
was the case with the BPT diagram. Indeed, this increase
AGNlifetimeshavebeenestimatedbyanumberofmethods
is fake, and only due to retired galaxies being mistaken for
basedonblackholedemographicsorAGNradiativeproper-
AGN hosts.
tieswhichledtoquitedispersedresults(Martini2004).Part
of this dispersion may be due to the fact that the different
studies do not actually refer to the same objects (e.g. high
or low level of activity,massive or less massive galaxies).
We have found that the standard use of the BPT dia- From the census of emission-line spectral types pre-
gram and our use of the WHAN diagram provide very dif- sented above, we can estimate the AGN lifetimes in each
ferent panoramas of the demographics of galaxy emission- mass bin (and even in each redshift bin) by dividing the
linespectraltypesacross thepast4Gyr(which isthelook- number of AGN hosts in a given bin by the total number
back time corresponding to z =0.4). One advantage of the of galaxies in that bin and multiplying it by the age of the
WHAN classification is that it does not set aside a large Universeatthecorrespondingredshift.ThisisshowninFig.
number of the galaxies. The other one is that it does not 5. We find lifetimes of the order of 1-5 Gyr for log M⋆ be-
attributeto nuclearactivity theemission-line ratios seen in tween 10 and 12. Statistics for lower masses are affected by
weak-line objects. In the remaining of the paper, we there- thecolourbiasdiscussedinAppendixAsothelifetimes are
foreonlyconsiderthegalaxyspectralclassesthatwedefined notreliable:sinceweprobablymissretired/redgalaxiesata
with the help of the WHAN diagram (i.e. SF, AGN hosts higherproportion in thebiased bins, thefractions of AGNs
and retired) and only analyze sample W. and SF in such bins are probably upper limits. At higher
8 Stasińska et al.
Figure 5. AGN lifetimes computed from the census of AGN hosts in the WHAN diagram as a function of redshift, for different mass
bins. Top: considering all galaxies with a detectable AGN; Bottom: considering AGN-dominated galaxies. As in Fig. 3, large symbols
correspond to(M⋆,z)binsthatarejudgeddevoidofcolourbias.
redshift,theAGNpopulationisalsocontaminatedbybulge It is also well-known that “red and dead” galaxies are more
+ disk systems in disguise (case a of Table 1). This means massivethangalaxiesformingstarspresently(Heavensetal.
that the AGN lifetimes we show are also upper limits in 2004).However,atagivenmassandredshift,somegalaxies
thesecases. Inthe12–12.5 mass range,thelifetimes appear have already stopped forming stars while others still form
to be shorter, about 0.3 Gyr. We can do the same, now re- them, some galaxies host an AGN while others do not. It
stricting ourselves to those objects where the AGN heavily hasrepeatedlybeenshownbydetailedstudiesofnearbyob-
contributes to the emission lines (i.e. galaxies which have jects that nuclear activity in Seyfert galaxies is linked with
EW(Hα) > 3Å and are above the K01 line; for simplicity somelevelofstar-formingactivity(Shlosman1990;Storchi-
wewillrefertothemAGN-dominatedgalaxies,althoughthis Bergman et al. 1996; Raiman & Storchi Bergmann 2000;
term is somewhat inaccurate since stellar ionization might Veilleux 2001; González Delgado et al. 2009). In order to
be preponderant in a number of cases). This is done in the furtherdiscuss therelation between star formation and nu-
bottom panel of Fig. 5. Now the lifetime appears to be of clear activity, here we study the star-formation histories of
theorderofafew108 yr.ThisimpliesthatstrongAGNare thevarious categories of galaxies considered.
short-lived with respect to weak AGN,or that thephase of
strong nuclear activity is shorter by a factor of 3–10 than
theperiod where nuclear activity can be detected. 6.1 Determination of the star-formation histories
Note that for 10<logM⋆ <12 we do not see any sig-
nificant variation of AGN lifetimes with respect to redshift The star-formation histories can be obtained from the
even in the redshift domain where absence of colour bias is starlight analysis of the galaxy spectra, which provides a
notguaranteed.Thiscanperhapsbetakenasan indication descriptionofgalaxiescontentintermsofsimplestellarpop-
that the colour bias is not very strong for those bins. Nei- ulations of different ages. The specific star-formation rate,
ther do we see any strong variation of the estimated AGN SSFR(t), which measures the mass converted into stars at
lifetimes with respect to M⋆. time t with respect to the total mass converted into stars,
Itmustbenotedthattheobtainedlifetimesarestrongly is computed for each galaxy by smoothing the stellar pop-
relatedtothecriterionusedtodefinetheactivity,asindeed ulation age distribution, as explained in Asari et al. (2007,
illustratedbythetopandbottompanelsofFig.5,andcare theirEq.6)withthefollowingmodification.Foragalaxyat
must be taken when they are used in other contexts. agivenredshift,partofthelight(hencestellarmass)maybe
attributed by starlight to stellar populations older than
the age of the Universe at that redshift. In order to cor-
rect for that, we find thelookback time tmax corresponding
6 STAR FORMATION HISTORIES
to each galaxy redshift and consider that the SSFR(t) at
Wehavejustseenintheprevioussectionthatgalaxieshost- t=tmax is the sum of the contributions of all stellar popu-
ing AGN are, on average, more massive than SF galaxies2. lations older than tmax.
Notethatthetimeresolutionofstellarpopulationanal-
ysesdecreaseswithlookbacktime,sothatitisnotpossible,
2 This was already found byK03, even though their AGN sam- for any given galaxy, to detect short episodes of star for-
ple was strongly contaminated by retired galaxies, which are on mation that happened long ago. Besides, the history of the
average moremassivethanSFones. activityingalaxiesisnotrecordedintheiremission lines.A
Retired galaxies and the star formation – AGN connection 9
statistical way to circumvent these limitations is to follow
a population of galaxies through different redshifts. This
allows us not only to detect short episodes of star forma-
tion which are nowadays too old to be discerned, but also
todirectly measuretheemission-line properties inthispast
epoch. Our dataset allows us to peep into the starburst-
AGNconnection in thelast 4Gyrof thelife of thegalaxies
(with thecaveat that welook at increasingly large portions
of galaxies as redshift increases).
6.2 Star formation histories of the various galaxy
spectral types
Figure 6 shows the variations of the median SSFRs (thick
blue curves) and 16th and 84th percentiles (thin curves) as
afunctionoflookbacktimeineachofourmass-redshiftbins
for our three galaxy spectral types. The top panel concerns
SFgalaxies, themiddlepanelconcernsAGNhosts, andthe
bottompanelconcernsretiredgalaxies.Onlybinscontaining
atleast50objectsareshown,toallowforarobustrepresen-
tation of the star-formation history in the considered bins.
For convenience, in each (M⋆,z) subpanel, we consider the
lookbacktimewithrespecttothegalaxyredshift,andnota
cosmological lookbacktimetakingintoaccounttheredshift
ofthegalaxy.Theinsetindicatesthescaleonthe(logSSFR,
logt)planeusedtorepresentthestar-formationhistoriesin
each bin. In each panel, the cyan background points rep-
resent all the galaxies from our sample that belong to the
same spectral class.
Wenowturntothediscussionofthestar-formationhis-
tories.ConcerningSFgalaxies,thefirstthingtonoticefrom
Fig.6isthat,atanyredshift,theSSFRchangesfromslightly
decreasingwithtimeforthehighestmasses(logM⋆ >11.5)
through roughly flat for intermediate masses to increasing
with time for the smaller masses. We also see that, for a
given mass bin, the ratio of recent to past star-formation
rate increases steadily with z for all mass bins. There is,
of course, a certain dispersion in all those relations, as seen
fromthecurvesindicatingthepercentiles,butthedescribed
tendencies are unambiguous.
To interpret the observed trends, we first must ask
whether,inagivenmassbin,thegalaxiesinthepanelswith
higher values of z are the predecessors of the galaxies with
lowerz.Thisisnotnecessarilytrue,becauseofmassgrowth,
which, in the(log M⋆, z) plane, tends to move the galaxies
towardstherightasredshiftdecreases.Thiseffect,however,
isingeneralnotcrucialforthegalaxiesofoursample,given
thewidthofourredshiftbins.Indeed,foratypicalSSFRof
10−10 yr−1 assuggestedbythetoppanelofFig.6,themass
of stars in the galaxy doubles during a time interval of 10
Gyr, while thewidth of our mass bins is 0.5 dex.There are
episodes where the SSFR is significantly higher than 10−10
yr−1, but, as seen in Fig. 6, they are short, so they do not
contribute much to the stellar mass growth. Thus, roughly,
inagivenM⋆ bin,galaxiesathigherredshiftscouldbecon-
sideredastheprogenitorsofgalaxies seenatlowest redshift
(but,ofcourse,theymaychangeclassduringtheirevolution,
forexampleanepisodicstarformationmaytakeplaceoran Figure6.Star-formationhistories,i.e.SSFRvslookbacktimein
AGN may appear). In a given mass range, aperture effects thegalaxiesredshiftframein(M⋆,z)bins.Top:SFgalaxies;mid-
qualitatively explain the observed tendencies of the SSFR dle: AGN hosts; bottom: retired galaxies. Thick curves: median
with redshift, since as redshift increases increasing portions SSFR; thin curves: 16th and 84th percentiles. SSFRs are shown
ofthestarformingdiskiscoveredbytheSDSSfiber.Inad- onlyforbinswithatleast50objects.Thecyanbackgroundpoints
showthe(M⋆,z)diagramfortherelevantsubsampleofgalaxies.
10 Stasińska et al.
dition, we must take into account the existence of the bias
againstredgalaxieswhich,inagivenmassbin,likelygrows
withincreasingredshift(seeAppendixA).Thus,theappar-
ent redshift evolution of the SSFR in a given mass bin is
perhaps simply due to these two causes and not an evolu-
tionary effect.
ThemiddlepanelofFig.6isanalogoustothetoppanel
but is for AGN hosts. Note that, because of our decision to
show results only for subpanels populated by at least 50
galaxies, some panels have disappeared while a few others
haveappearedwithrespecttotheSFcase.Comparedtothe
star-formation histories of SF galaxies, those of AGN hosts
show similar tendencies, with, however, the SSFRs being
lower at small lookback times than those of SF galaxies in
thesame (M⋆,z) bin. Wewill return to thispoint below.
ThebottompanelofFig.6isanalogoustothetoppanel
but concerns galaxies appearing as retired instead of star-
forming.HerewewitnessaverystrongdecreaseoftheSSFR
with time in all the subpanels, with a rather small disper-
sion3.
Figure 7. Star-formation histories in mass-redshift bins for all
6.3 Comparison of the star-formation histories of
the galaxies. They are represented in blue for SF galaxies, in
the three galaxy spectral classes: a clue to the
greenforAGNhosts,andinredforretiredgalaxies.
star formation – AGN connection?
Galaxies that are now appearing as retired must have been
Another interpretation of Fig. 7 could be as follows.
seen as star-forming in the past, perhaps not continuously,
The population of AGN hosts could actually be composed
butatleastepisodically.Thustheirprogenitors–oratleast
of retired galaxies and of massive SF galaxies, resulting in
some of them – are actually in thetop (or middle) panel of
median star-formation histories of AGN hosts intermediate
Fig. 6. It is therefore interesting to compare in more detail
between SF and retired galaxies. In such a circumstance
thestar-formation histories of ourthree classes of objects.
Fig. 7 would not necessarily imply theexistenceof a causal
Fig. 7 displays the median star-formation histories of
link between the AGN phenomenon and recent star forma-
SF galaxies (blue curves), AGN hosts (green) and retired
tion. One would then expect a dichotomy in the distribu-
galaxies (red)in(M⋆,z)bins.Onecanseethat,atanyred- tion of thestar-formation rates of AGN hosts. Fig. 8 shows
shift,thestar-formationhistoriesasobtainedbystarlight
thedistribution of thestar-formation rates at t=24.5 Myr
for SF galaxies and AGN hosts are roughly similar. The
(which Asari et al. 2007 found that best matches the Hα-
largest differencesoccurat smallredshift andsmall masses.
calibratedSFR)forthethreegalaxytypes.Thedistribution
Anotherthing to note is that the SSFRsof retired galaxies
forAGNhostsisnotbimodal,whichrulesdownthissecond
areidenticaltothoseofSFandAGNgalaxiesatthelargest
interpretation.
lookback times diverging at lookback times of one to a few
The overall quenching factor (as guessed from the
Gyr.
SSFR(AGN)/SSFR(SF)ratioinagivenbin)decreaseswith
The curves in Fig. 7 can be read as follows. Judging
increasing M⋆. At low M⋆, AGN hosts have specific star-
from the point where the green and blue curves split apart
formationratesnearly10timeslowerthanSFgalaxies,while
in the figure, the nuclear activity kick-off in SF galaxies
at high M⋆ this differencereduces to a factor of 2 or less.
reduces the level of star formation during the next 0.1–1
Fig. 9 is analogous to Fig. 7 but here we only consider
Gyr with respect to SF galaxies in the same (M⋆,z) bin. AGN hosts that are dominated by the AGN (i.e. galaxies
In low mass galaxies, M⋆ ∼ 109−10.5M⊙, once the AGN is whichhaveEW(Hα)>3Åand areabovetheK01lineinthe
turned on it inhibits star formation for the next ∼ 1 Gyr.
BPT diagram.) Qualitatively, the behaviour is the same as
In high mass galaxies this delay seems shorter: ∼ 0.1 Gyr
inFig.7,exceptthatthequenchingfactorislarger(compare
for M⋆ ∼1011M⊙ and perhaps even less for higher masses, top and bottom of Fig. 10).
as can be read from Fig. 10 (top), which shows a close-up
of Fig. 7. All these AGNs keep forming stars, albeit at a
reduced rate.
7 SUMMARY
In thispaper, we have addressed anew thetopic of the star
3 Webelievethattheupturnatsmalllookbacktimesseenatthe formation–AGN4connectioninthecontextofglobalgalaxy
smallestredshiftsisanartefactduetotheincompletetreatmentof evolution by studyingseveral aspects.
horizontalbranchand/orbluestragglersintheBC03evolutionary
synthesis models used in the starlight analysis. As previously
shownbyOcvirk(2010)andGonzálezDelgado&CidFernandes 4 Note that our working definition of an AGN host is that the
(2010),thelackofsucholdandhotphasestendstoproducefake galaxy contains a detectable AGN, but the AGN itself may be
burstsinafullspectral fittinganalysis. weak,inparticularwithrespect totheionizationduetomassive