AGNs and galaxy interactions

0.34 MB·English

by  M. Sol Alonso

Mon.Not.R.Astron.Soc.000,000–000 (0000) Printed4February2008 (MNLATEXstylefilev2.2)
AGNs and galaxy interactions
M. Sol Alonso 1,2, Diego G. Lambas 1,4, Patricia Tissera 1,3 and Georgina Coldwell 1,4
1 Consejo Nacional de Investigaciones Cient´ıficasy T´ecnicas.
2 Complejo Astron´omico El Leoncito, Argentina
3 Institutode Astronom´ıa y F´ısica del Espacio, Argentina.
4 ObservatorioAstron´omico de la Universidad Nacional de Co´rdoba, Argentina.
7
0
0
2
4February2008
n
a
J
ABSTRACT
8

1 We performastatisticalanalysisofAGNhostcharacteristicsandnuclearactivity
v for AGNs in pairs and without companions. Our study concerns a sample of AGNs
2 derivedfromtheSDSS-DR4databyKauffmannetal(2003)andpairgalaxiesobtained
9 from the same data set by Alonso et al. (2006). An eye-ball classification of images
1 of 1607 close pairs (r < 25 kpc h−1, ∆V < 350 km s−1) according to the evidence
p
1 of interaction through distorted morphologies and tidal features provides us with a
0
more confident assessment of galaxy interactions from this sample. We notice that,
7
at a given luminosity or stellar mass content, the fraction of AGNs is larger for pair
0
galaxies exhibiting evidence for strong interaction and tidal features which also show
/
h sings of strong star formation activity. Nevertheless, this process accounts only for a
p ∼10%increaseofthefractionofAGNs.Asinpreviousworks,wefindAGNhoststobe
- redder and with a larger concentration morphological index than non-AGN galaxies.
o
This effect does not depend whether AGN hosts are in pairs or in isolation. The
r
t OIII luminosity of AGNs with strong interaction features is found to be significantly
s
a larger than that of other AGNs, either in pairs or in isolation. Estimations of the
: accretion rate, L[OIII]/M , show that AGNs in merging pairs are actively feeding
v BH
their black holes, regardless of their stellar masses. We also find that the luminosity
i
X of the companion galaxy seems to be a key parameter in the determination of the
r black hole activity. At a given host luminosity, both the OIII luminosity and the
a L[OIII]/M are significantly largerin AGNs with a brightcompanion(M <−20)
BH r
than otherwise.
Key words: cosmology:theory - galaxies: formation - galaxies: evolution.
1 INTRODUCTION iesshowedthatthestarburstsarefueledbygasinflowspro-
ducedbythetidaltorquesgeneratingduringtheencounters.
Galaxy-galaxy interactions can be an effective mechanism
Theefficiencyofthismechanismsdependsontheparticular
to regulate the star formation activity in galaxies. Studies
internal characteristics of galaxies and their gas reservoir.
of galaxies in pairs in larger surveys such as the 2dFGRS
and the SDSS showed with high statistical signal that star
formation is enhanced by a factor of two in close systems Besides feeding star formation, these gas inflows could
(Lambas et al. 2003; Alonso et al 2004). The trend of in- also feed a central black hole and AGN activity (Sanders
creasing star formation for decreasing relative velocity and et al. 1988). The co-evolution of galaxies and black holes
projected separation was first showed by Barton & Geller is now widely accepted although many details on how this
(1998). This trend was confirmed to be in place indepen- coexistence works are still understudied (Heckmann et al.
dentlyofthecharacteristicsofenvironmentbyAlonsoetal. 2004). Toomre & Toomre (1972) suggested that collisional
(2006). disruptionandgasdissipationcouldfeedthenuclearactivity
Thephysicsbehindthetriggering ofstarformation ac- of galaxies. Other ideas about the fueling of active galaxy
tivityduringgalaxy-galaxyinteractionshavebeenexplained nucleusarethediskofbarredgalaxiesgeneratedbyinternal
by theoretical (Martinet 1995 and references there in) and instabilities which can lead thegas tothecentreofgalaxies
numerical analysis (e.g. Toomre & Toomre 1978; Barnes & (Schwartz1981,Shlosman,Begelman&Frank1990)andthe
Hernquist1992,1996;Mihos&Hernquist1996).Thesestud- presence of more that one supermassive black hole which
2 M. Sol Alonso et al.
seems likely in the nucleus of merger remnants (Begelman, 6670 squaredegreesof skyimaged infivewave-bands(u,g,
Blanford and Rees 1980). r, i and z) containing photometric parameters of 180 mil-
Certainly,alltheseprocessesmayprovidecluestosolve lion unique objects. The main galaxy sample is essentially
the AGN fueling problem. However, AGN-galaxy interac- a magnitude limited spectroscopic sample (Petrosian mag-
tion paradigm has been the issue of several studies which nitude) rlim<17.77, most of galaxies span a redshift range
aim at determining, for example, the frequency of AGNs 0<z<0.25 with amean redshift at z =0.1 (Strausset al.
in a sample of interacting galaxies, the proximity effects of 2002).
a close companion on the AGN power, etc. Dahari (1985), The SDSS-DR4 has more than 500000 galaxy spectra
in a sample of 167 systems, found that there is an excess andincludesdifferentgalaxypropertiessuchasmagnitudes,
of Seyferts among spiral interacting galaxies and also that star formation indicator, concentration index parameters,
Seyfertactivityandtheexistenceofexternaltidalforcesare etc. From SDSS-DR4, K03 have constructed a catalogue
interrelated. However, Seyfert nuclei were not found in in- with a subset of 88178 narrow emission line galaxies that
teractingellipticals.Fromstatisticalstudyofcompanionsof areclassified as AGN,within redshift range 0.02 <z< 0.3.
Seyfert galaxies Fuentes-Williams & Stocke (1988) derived Consideringtheunifiedmodel(Antonucci1993),AGNs
that Seyfert galaxies have a marginal excess of compara- can be separated into two categories: Type 1 AGN where
blesize companions and a clear excess of faint companions. the black hole and the associated continuum and broad
However,Keel(1996) foundthat thepresenceof AGNdoes emission-line region is viewed directly and Type 2 AGNs
not seem to be related with the type of collisional interac- whereonlythenarrowlineregioncanbeobservedduetothe
tion. obscuringmedium.TheBaldwin,Phillips&Terlevich(BPT,
Conversely Schmitt (2001) analysed a sample of dif- 1981)line-ratiodiagramallowustodistinguishtype2AGNs
ferent types of active galaxies such as LINERS, transition from normalstarforminggalaxies byconsideringtheinten-
galaxies, absorption-line galaxies, Seyfert and HII galaxies. sityratiosoftwopairsofrelativelystrongemissionlines.The
This authorshowed that galaxies of different activity types sample of Type 2 AGN of K03 was selected taking into ac-
havesimilarpercentageofcompanionsthangalaxiesofsimi- counttherelationbetweenspectrallines,[OIII]λ5007,Hβ,
larmorphologicaltypeswithnonuclearactivity.Thisresult [NII]λ6583 and Hα luminosities where an AGN is defined
suggests that interactions between galaxies arenot aneces- if
saryconditiontotriggerthenuclearactivityinAGNs.Sim-
log([OIII]/Hβ)>0.61/(log([NII/Hα])−0.05)+1.3 (1)
ilarresultswerefoundbyKelm,Focardi&Zitelli(2004) by
studying a sample of Seyfert in UZC-Compact Groups and Wethen use 88000 Type2 AGNsin K03 catalog. The
by Coldwell & Lambas (2006) who analyzed the percent- luminosity of [OIII]λ5007 emission line will be used as a
age of close companions of SDSS quasars within the range tracer of AGN activity.
rp <100 kpch−1 and ∆V <350km s−1.
An interesting approach to study the role of interac-
2.1 SSDS-DR4 galaxy pair catalog
tions on the feeding of both star formation and black holes
is to study AGN activity in galaxy pairs. In fact, Sortchi- In our previous works (Lambas et al., 2003; Alonso et al.
Bergmann et al. (2001) found a correlation between active 2004;2006),weselectedgalaxiesinpairsadoptingprojected
galactic nucleus and star formation activity in interacting relative distance and relative velocity thresholds: rp < 100
galaxies by analyzing 35 Seyfert 2 nuclei. The galaxy spec- kpc h−1 and ∆V < 350 km s−1. We found that these lim-
tral information available in large surveys provides unique its are adequate to define galaxy pairs with enhanced star
information and high statistical numbers to study on more formation activity. In these works we also detected a clear
robust basis the relation between these processes. In this correlation between star formation activity and the prox-
paper, we combine the galaxy pair sample constructed by imity to a close neighbour supporting the physical scenario
Alonso et al. (2006) from the Sloan Digital Sky Survey wheretidaltorquesgeneratedduringtheinteractionstrigger
(SDSS-DR4) with a sample of 88000 narrow-line AGN se- gas inflows, feeding the star formation activity.
lected by Kauffmann et al (2003, hereafter K03) from the Forthepurposeofanalysingtheeffectsofcloseinterac-
SDSS-DR4,which is hithertothelargest sample of AGNs. tions, which are known to produce thestrongest effects, we
In section 2, webrieflydescribetheAGNandpaircat- selectedclosepairgalaxiesbyrequiringrp <25kpch−1and
alogs. Section 3 analyses the properties of hosts of Type-2 ∆V < 350 km s−1. With this further restriction to closer
AGN galaxies. Section 4 provides a comparative statistical relative separations, the effects of interactions are largely
analysisofAGNhostsinpairsandwithoutcompanionswith increased as shown by Lambas et al. (2003). Galaxy pairs
the aim of unveiling the role of interactions in triggering are only selected within a redshift range 0.01<z <0.10 in
AGNactivity.Insection5wegiveourmainconclusionsand order to avoid strong incompleteness at larger distances as
discussion of theresults. well as significant contributions from peculiar velocities at
low redshift. The final pair catalog in the SDSS-DR4 com-
prises 1607 close galaxy pairs. Following the procedure ex-
plained by Lambas et al. (2003), we constructed a control
2 SDSS-DR4 DATA
sample by selecting those galaxies without a close compan-
The SDSS (York et al. 2000) in fiveoptical bands will map ion within rp < 100 kpc h−1 and ∆V < 350 km s−1. The
one-quarter of the entire sky and perform a redshift sur- controlsampleisalsorequiredtohavethesameredshiftand
vey of galaxies, quasars and stars. The fourth data release, magnitudedistribution as our close galaxy pair catalog.
DR4,isthemajorcataloguewhichprovidesimages,spectra, As discussed in Alonso et al. (2006), the effects of in-
and redshifts for download. The imaging portion comprises completeness or aperture (e.g see also Balogh et al.2004ab)
AGNs and galaxy interactions 3
donotintroduceimportantbiasinthegalaxypaircatalogs.
Regarding incompleteness, by combining the spectroscopic 0.25
andphotometricsurveys,Alonsoetal.(2006)estimatedthat 0.2 (a) (b)
the spectroscopic catalog has only an incompleteness of ≈ 0.2
0.15
9.5%. Hence,neithereffect isexpectedto haveastrong im-
0.15
pact in our close galaxy pair catalog.
0.1
0.1
0.05 0.05
3 THE HOSTS OF TYPE 2 AGNS IN GALAXY
0
PAIRS 0
-23 -22 -21 -20 -19 -18 9 10 11 12
Asafirststep,weidentifyAGNhostsinourclosepaircata-
log(i.e.pairswithrp <25kpch−1 and∆V <350kms−1) 0.2
bycross-correlating itwiththeAGNcatalog. Wefound498 0.2 (c) (d)
galaxypairswithonememberexhibitingAGNactivity.This
0.15
represent ≈ 30% of the total close pair catalogue. We also 0.15
detectedthat108closepairshavebothmembersclassifiedas 0.1
0.1
AGNs(≈7%).Then,thefinalSDSS-DR4AGN-paircatalog
comprises 606 pairs. 0.05 0.05
Forthepurposeofproperlyassessingthesignificanceof
theresultsobtainedfromAGNgalaxiesinpairs,wedefined 0 0
a sample of isolated AGN galaxies from the SDSS-DR4 by 1 2 3 1.5 2 2.5 3 3.5 4
(u-r) C = (r90/r50)
selecting those galaxies without a companion galaxy within
rp <100 kpc h−1 and ∆V <350 km s−1, within the same
redshiftrangeoftheAGNclosepaircatalog.Theprocedure Figure1.(a)r-bandabsolutemagnitude,(b)Massinstars,(c)
u-rcolourand (d)Concentration parameter distributions.Solid
followed to construct the sample of isolated AGN galaxies,
linescorrespondtoAGNsinpairs,dashedlinestoAGNswithout
assuresthatitwillhavethesameselectioneffectsthanAGN
companions, and dotted lines to a control sample of non-AGN
galaxy pair sample, and consequently, it can be used to es-
galaxieswithasimilarluminositydistributionthanAGNs.
timate the actual difference between AGN galaxies in pairs
and isolated AGNs, unveiling the effect of the interactions.
A similar reasoning motivated the definition of a non-AGN
InFig.2weshowtheBPTdiagnosticdiagram,plotting
galaxy sample with the same luminosity distribution than
theratio [OIII]λ5007/Hβ versusratio [NII]λ6583/Hα for
galaxies in theAGN-paircatalog.
AGN pair galaxies and AGNs without a near companion.
Thedistributionofabsolutemagnitudein r-band(Mr) As it can be appreciated, there is not an evident difference
andthestellarmass(M∗),ofAGNsinpairs,isolatedAGNs,
between isolated and AGNsin pairs.
and a control sample of non-AGN galaxies with a similar
Hence, colours, morphology as measured by C, or lu-
luminosity distribution than the hosts of AGNs are shown
minosity, do not provide evidence that interactions could
in Fig.1. It can be seen that the luminosity distribution is
play a role in the triggering or regulation of the AGN ac-
similarinthethreesamples,butthefractionoflargestellar
tivity. AGN hosts show similar distributions regardless of
mass objects among AGN hosts is larger than that of non-
thepresenceofaclosecompanion.However,takingintoac-
AGN hosts: the percentage of galaxies with log(M∗) > 11
count previous results such as those of Strochi-Bergmann
is 15.8% for AGNs in pairs, 19.7% for isolated AGNs and
et al. (2001), a clear signal of correlation between theAGN
9.8% for non-AGN hosts. In fact, there are very few AGNs
activityandinteractions couldbedetectedbyrestrictingto
in galaxies with M∗ < 1010 M⊙. As it is well know, the activestarformingsystems.Inourpreviousstudiesofgalax-
majorityoflowmassgalaxieshaveyoungstellarpopulation
ies in pairs we found that there is an important fraction of
and a significant fraction are currently experiencing strong
galaxies that, although being close in projected space, do
star formation activity.
notexhibitstrongstarformationactivity.Itisveryunlikely
InFig.1c,weplotthe(u−r)distributionsforthesame that at these close projected distances (rp < 25 kpc h−1
samples. As it can be appreciated from this figure colour and ∆V <350 km s−1) pairs could be strongly affected by
distribution of the control sample of non-AGN galaxies is
interlopers (Alonso et al. 2004; Perez et a. 2005). However,
clearly bimodal as expected (e.g. Baldry et al. 2003). How-
it is possible that,owing to particular characteristics of the
everAGNs,regardless ofbeinginpairsorinisolation, have
orbital parameters or internal structure, these pairs do not
(u−r)coloursconsistentwitharedhostpopulation.Fig.1d
experience strong tidal torques that can drive gas inflows.
shows the concentration parameter distribution ,C, defined
Motivated by these facts, in the next Section we discuss a
as the ratio between the radii enclosing 90 % and 50% of
new classification for close galaxy pairs based on a visual
the galaxy light in the Petrosian r-band. Recall that Strat-
morphological analysis of their members.
eva et al. (2001) found that galaxies with C > 2.6 are
mostly early-type galaxies, whereas spirals and irregulars
have 2.0 < C < 2.6. From the distributions of Fig.1d, it
3.1 Classification of pair galaxies
canbeappreciatedthatAGNsresidepreferentially inbulge
dominatedgalaxiesincontrasttothecontrolsampleofnon- ByusingthephotometricSDSS-DR4,weclassifiedallgalax-
AGN galaxies which show the expected C bimodality. ies in the close pair catalogue taking into account the
4 M. Sol Alonso et al.
1.5
1
0.5
0
-0.5
-1
-1.5 -1 -0.5 0 0.5
Figure2.BPTdiagnosticdiagram.Thefilledcirclescorrespond
toAGN pairgalaxiesand filledtrianglesto AGNswithoutcom-
panions. The dashed curve represents the demarcation between
starburstgalaxiesandType2AGNsdefinedbyKauffmannetal.
(2003).
eye-ball detection of features characteristics of interactions.
Three categories havebeen defined:
• Pairs with evidenceof an ongoing merging process (m:
Figure3.Examplesofgalaxyimagesinclosepairswithdifferent
merging pairs) classification:evidenceofanongoingmergingprocess,m,(upper
• Pairswithsignsoftidalinteractionsbutnotnecessarily panels), pair with a sign of tidal interaction but not necessarily
merging (t:tidal pairs) merging ,t, (medium panels) and pair galaxies showing no evi-
• Pairsshowingnoevidenceofdistortedmorphologies(n: denceofdistortedmorphologies,n,(lowerpanels).
non-interacting pairs).
Fig.3 shows images of typical examples of close pair
galaxiesforthethreedifferentvisualclassification: m,tand
galaxy pairs (m) extracted at random from our total sam-
n.Hereafterwewillcarryouttheanalysisbydistinguishing
ple of 383 m-pairs (with typical projected separation rp <
them according to these categories. Note that although the 12 kpch−1).We calculated thetotal magnitudeof thepair
m category referes to systems in advance stages of interac-
which acording to our previous analysis should be within
tions, the two systems could be still individualized by the
0.02 magnitudes compared to SDSS measurements in the
SDSSsurvey.
case that deblending is accurately working. We find, how-
ever, that the deblending procedure in SDSS introduces a
larger uncertainty since by adding the quoted luminosities
3.1.1 Analysis of deblending for Merging Pair Galaxies
in SDSS, the typical difference raises to 0.12 rms. In Fig.4
In order to analyse the effects of how reliable SDSS photo- weshowr-bandmagnitudderivedfromSExtractorroutines
metric deblendingprocedure works in merging galaxy pairs versusr-band magnitud extracted from SDSS,for a sample
(m) for which we expect the largest uncertainties, we have ofisolatedgalaxiesandforasubsampleof50mergingpairs.
carried out suitable tests to determine independent magni- Thissimpleandrobusttestgivesusefulestimatesoftheun-
tudesfor thissample. certaintiesintroducedbythedeblendingprocedureadopted
Forthispurposewehavestudieddirectlygalaxyimages in SDSS which we argue are about 0.12 mag. We also no-
extractedfrom SDSSandwehaveusedSExtractorroutines ticethelack ofsystematicsinSDSSdeblendingprocedures,
to derive galaxy magnitudes. We have used first, a set of since the difference between our Sextractor measurements
isolated galaxies which serve as a control sample for test- and quoted SDSS magnitudes have an approximately zero
ing the photometry. For this sample, our photometry gives mean.
r-band magnitudes with 0.02 rms difference compared to In sum, this analysis indicates that SDSS magnitudes
those quoted in SDSS, indicating the consistency of our re- canbesuitablyusedinouranalysisinthefollowingsections
sultswith SDSSmagnitudeswhennodeblendingprocedure provided a rms scatter of 0.12 magnitudes is not seriously
is required. The second sample is composed of 50 merging affecting theresults.
AGNs and galaxy interactions 5
Table 1.Percentages ofAGNs
Categories TotalPairs m t n Total Control
Numberofclosepairs 1607 383 688 536 14359
%ofAGNs 32% 31% 32% 28% 23%
(a)
(b)
Figure 4. r-band magnitud derived from SExtractor routines Figure5.FractionofAGNsinmerging-pairgalaxies(solidlines),
versus r-band magnitud extracted from SDSS, for a sample of intidalpairs(dotted lines),innon-interactingpairs(dot-dashed
isolatedgalaxies (open circles)and forasubsample ofclose pair lines) and in galaxies without companions (dashed lines), as a
galaxies(triangles). functionofMr (a)andM∗ (b).
3.1.2 Frequency of AGNs gionwherethemaincontributiontotheopacitycomesfrom
Forthethreecategoriesdefinedabove,wecalculatetheper- ionized metals. The break index Dn(4000) (Kauffmann et
al. 2002) is defined as the ratio of the average flux density
centage of pairs with AGNs. As we can notice from Table
1, we found that ≈ 30 % of close galaxy pairs host AGNs, by using narrow continuum bands (3850-3950 ˚A and 4000-
regardless of the category they belong to, and that isolated 4100 ˚A). The Dn(4000) indicator is suitably correlated to
the mean age of the stellar population in a galaxy and can
galaxies without companions have a lower AGN frequency
be used to estimate the star formation rate per unit stellar
(sin23%).
mass,SFR/M∗,(Brinchmannetal.2004).Themajorityof
SinceAGNphenomenonisastrongfunctionofhostlu-
starformationtakesplacepreferentiallyingalaxieswithlow
minosity, we analysed the fraction of AGNs in each of the
paircategoriesandinthecontrolsample,fordifferentstellar Dn(4000)values,forinstance,only12%and2%ofthetotal
mass content and luminosity intervals. In Fig.5, we display SFRdensityareassociated togalaxieswithDn(4000)>1.8
thefractionofAGNsasafunctionofr-bandluminosityand and Dn(4000)>2.0 respectively.
Wehaveanalysed thebehaviourof thisparameteras a
stellar mass The results indicate that the fraction of host
function of stellar mass and luminosity for AGNs in pairs
with AGN activity increases with luminosity in a similar
and in isolation. The resultsare given in Fig.6 where it can
fashion for AGNs in pairs or in isolation. However, we no-
beseenthatAGNsinmergingpairshavesignificantlylower
tice that, at a given mean luminosity or mean stellar mass,
mergingpairstendtohaveahigherfrequencyofAGNsthan values of Dn(4000). This finding suggests that these galax-
ieshaveexperienced morerecent episodes of star formation
pairs in the two other categories or in the control sample.
Nevertheless, this excess is approximately ≈10%. as expected since they are effectively in an ongoing merger
event.
The trends shown in Fig.5 and Fig.6 support the sce-
nario where interactions could drive both star formation
3.1.3 Stellar population ages in AGN hosts
and AGN activity via the triggering of gas inflows to the
The strongest discontinuity occurring at 4000 ˚A in the op- central region. The fact that we are actually detecting the
tical spectrum of a galaxy arises of the accumulation of a strongest signal for merging pairs could be indicating that
large number of spectral lines in a narrow wavelength re- fusions could be more efficient than tidal torques produced
6 M. Sol Alonso et al.
strumental resolution of SDSS spectra is σ∗ ≈ 60 to 70 km
s−1. The aperture corrections to thestellar velocity disper-
(a)
sions are small in early-type galaxies in SDSS (Bernardi et
al. 2003), so we do not apply such corrections.
InFig.7a,weshowthedistributionofL[OIII]forAGNs
in galaxies in merging pairs, tidal-pairs, non-interacting
pairs and without a close companions. As it can be clearly
appreciated, AGNsin pairs with evidence of ongoing merg-
ing processes (m) show larger luminosities L[OIII] than
AGNs in isolation or in the other two pair categories. In
fact,while62.4%ofAGNsinmergingpairshaveL[OIII]>
106.5L⊙,only43.3% ofisolatedAGNsshowsuchapowerful
(b) activity. Note that the other two pair categories have AGN
luminosity distributions similar to that of isolated AGNs.
In Fig.7b , we show the distribution of black holes
masses, MBH,forAGNsgroupedsimilarly toFig.7a.Itcan
be appreciated in this figure that although there is a trend
for MBH to be systematically larger for AGNs in merging
pairs, the difference is not as significant as the signal de-
tected for L[OIII],being only of less than 0.15 dex.
It is interesting to further investigate the strength of
AGNs as a function of host luminosity and stellar mass.
Figure 6. Dn(4000) as a function of (a) Mr and (b) M∗. The Therefore, we calculated the mean L[OIII] as a function
different types of lines indicate AGNs in merging-pair galaxies
of r-band luminosity and stellar mass content of the corre-
(solid), AGNs in tidal-pairs (dotted), AGNs in non-interacting
spondinghostgalaxies.Theserelationshavebeenestimated
pairs (dot-dashed) and AGNs in galaxies without companions
for AGNs in the three pair categories (m, t, n) and for iso-
(dashed).
lated AGNs. From the results shown in Fig.8a, it can be
clearly appreciated that, in general, most luminous hosts
duringtheorbital decay to feed both AGNand star forma- show the highest L[OIII]. We can also see that AGNs in
tion activity. merging pairs show higher L[OIII], regardless of the lumi-
nosityandstellarmassofthehostgalaxy,indicatingthatthe
blackholeactivityisstrongerforAGNsinadvancestagesof
interactions.Infact,forhostgalaxieswithM∗ >1010.5M⊙,
4 PROPERTIES OF AGNS IN INTERACTING
themeanL[OIII]ofmergingpairsislargerbyhalfanorder
GALAXIES: O[III] LUMINOSITY AND
ofmagnitudethanthemeanL[OIII]oftherestofthesam-
BLACK HOLE MASS
ples.AsimilarestimationforMBH confirmsthetrendfound
As a tracer of the AGN activity, we focus here on the lu- inFig.7b.Black holemasses areslightly larger forAGNsin
minosity of the [OIII]λ5007 line, calculated by Kauffmann merging pairs than for the rest of thecategories.
et al. (2003) in SDSS-DR4 AGN galaxies. Although this The ratio of [OIII]luminosity to black hole mass (R =
line can be excited by massive stars as well as by AGNs, log(L[OIII]/MBH))providesausefulmeasureoftheaccre-
itisknowntoberelativelyweakinmetal-rich,star-forming tion rate onto a black hole (Heckman et al. 2004). In our
galaxies. The [OIII] line also has the advantage of being calculations we have not converted the L[OIII] to volume-
strongandeasilydetectedinmostgalaxies.Itshouldbeno- averaged luminosities but by using the relations displayed
ticed that the narrow-line emission is likely to be affected in Fig. 2 of Heckman et al. (2004) it is possible to infer an
by dust within the host galaxy (Kauffmann 2003). Thus it accretion time for AGNs in pairs. In Fig.9 we show R as a
is important to correct the L[OIII] luminosities for the ef- functionofhostluminosity(a),stellarmass(b)andBHmass
fects of extinction. For AGNs in SDSS, Kauffmann et al. (c), for AGNs in the three defined pair categories (m, t, n)
(2003)measuredtheextinctionusingtheBalmerdecrement andinisolatedhosts.Inagreementwiththepreviousresults,
findingthatthebestapproximationforadustcorrectionto we found that black holes in smaller or fainter systems are
L[OIII]isbasedontheratioHα/Hβ.Wewillconsideredthe more active than black holes in larger or brighter systems.
luminosity L[OIII]=106.5L⊙ as alimit between weak and Heckmanetal.(2004)concludesthatthemostrapidlygrow-
powerfulAGNs.AweakL[OIII]emission indicates thatthe ingblackholesarethosewithMBH <few×107M⊙ withan
black hole is not rapidly growing (Heckman et al. 2004). implied growth time ∼ twice the age of the universe. From
Weestimatedblackholesmasses,MBH,forbothAGNs Fig.9c we can appreciate that MBH < few×107M⊙ have
with and without companion, by using the observed corre- significantly higher accretion rates than larger ones.
lation between MBH and the bulge velocity dispersion σ∗ Therelations shown inFig.9 indicatethat,although in
(Tremaine et al. 2002) generalthesmallestblackholesaretheonesthatexhibitthe
highest rates of accretion, AGNs in merging pairs host the
logMBH =8.13+4.02log(σ∗/200) (2)
most rapidly growing black holes. In fact, AGNs in merg-
Werestrict thisanalysistoAGNgalaxies withσ∗ >70 ing pairs have the most powerful black holes and the ones
km s−1, corresponding to logMBH > 6.3, because the in- that are growing faster, besides hosting stellar populations
AGNs and galaxy interactions 7
0.06
(a)
0.04
0.02
0
5 6 7 8 9
log (Lum(OIII))
0.25
(b)
0.2
0.15
0.1
0.05
0
6 7 8 9
log (MassBH)
Figure 7. Distributions of L[OIII] (a) and Black Hole Mass (b)
in AGNs merging-pair galaxies (solid lines), AGNs tidal-pairs
(dottedlines),AGNsnon-interactingpairs(dot-dashedlines)and Figure 8. L[OIII] (a,b) and MBH (c,d) as a function of Mr
AGNsgalaxieswithoutcompanions (dashedlines). andM∗ forAGNsinmergingpairs(solidlines),tidalpairs(dot-
ted lines), non-interacting pairs (dot-dashed lines)and inAGNs
galaxieswithoutclosecompanions (dashedlines).
which show clear signal of being importantely rejuvenated
by recent starbursts.
Finally,wehavealsoinvestigatedtheroleplayedbythe Ourmain results are:
AGN companion in powering the AGN activity. This was 1- The fraction of AGNs as a function of r-band lumi-
accomplishedbycalculating,foreachAGNinpairs,theOIII nosity or stellar mass is larger for merging pairs. However,
luminosity and L[OIII]/MBH as a function of its r-band this is a relative small effect accounting for an increase of
luminosity and stellar mass content, considering separately less ∼10% dueto tidal interactions.
pairswithabrightandafaintcompanion.Forthatpurpose, 2-AGNhostsareredderandwithalargerconcentration
we adopted Mr = −20 as the magnitude threshold. The morphological index than non-AGN galaxies. This effect is
results shown in Fig.10 clearly indicate that AGN activity similar for AGNhosts in pairs or in isolation.
in hosts with bright companions is significantly enhanced 3- We find that the locus of AGNs with and without
with respect to that of AGNswith faint companions, being close companions in the BPT diagram are consistent with
boththeAGNpowerandaccretionratesignificantlylarger. each other.
This finding provides evidences that nuclear activity is not 4-Atagivenhostr-bandluminosityorstellarmass,we
onlyaffectedbythepresenceofaveryclosecompanion but findtheAGNOIIIluminositytobeenhancedforAGNswith
that the relative mass (or luminosity) is also an important strong interaction features (i.e. merging pairs). We found
issue to take intoaccount. thatwhile62%ofAGNsinmergingpairshaveLum[OIII]>
106.5 thispercentagedecreasesto43%inthecaseofisolated
AGNs.
5- Estimations of the mean accretion rates onto the
5 SUMMARY AND CONCLUSIONS
black holesof AGNsin pairs in thethreedefinedcategories
Wehaveperformed astatistical analysisofboth,host char- and in isolation indicate that AGNs in merging pairs have
acteristics and nuclear activity, of AGNs in pairs and with- moreactiveblackholesthanotherAGNs,atagivenr-band
out close companions. Our study is based on the sample of luminosity or stellar mass content. We estimated than 27%
AGNs derived from the SDSS-DR4 release by Kauffmann of AGNs in merging pairs have Lum[OIII] > 106.5 and
et al. (2003) and the pair galaxies obtained from the same MBH < 3×107M⊙ and can be considered as being feed-
data release. We have complemented the SDSS-DR4 data ing their black holes with high efficiency (Heckmann et al.
with the addition of a eye-ball classification of images of 2003). In the case of isolated AGNs (selected to have the
1607 close pairs (rp <25 kpc h−1, ∆V < 350 km s−1) same,thispercentagedecreasesto21%.Ontheotherhand,
according to the evidence of interaction through distorted all AGNs in merging pairs have Dn4000 < 1.8 which indi-
morphologies and tidal features. Also, we have performed cates that they also have active star formation formation.
a photometric analysis to assess the reliability of the SDSS Interestinglywealsofoundthat17% ofisolated AGNshave
magnitude deblending procedure in merging pairs, finding powerful emission, MBH < 3×107M⊙ and have also ex-
negligible systematic effects. perienced important star formation activity in recent past
8 M. Sol Alonso et al.
Figure10.MeanLum[OIII](a,c)andR=L[OIII]/MBH (b,
d) as a function of Mr and M∗ for AGNs in pairs (solid lines),
AGN galaxy with a bright pair companion (Mr < −20.0) (dot-
ted line), and AGN with a faint pair companion (Mr > −20.0)
(dashedline).
American European Network for Astrophysicsand Cosmol-
ogy and Conicet from Argentina.
Funding for the creation and distribution of the SDSS
Archive has been provided by the Alfred P. Sloan Founda-
tion, the Participating Institutions, the National Aeronau-
tics and Space Administration, the National Science Foun-
dation, the U.S. Department of Energy, theJapanese Mon-
Figure 9. Mean R = log(L[OIII]/MBH) as a function of M∗ bukagakusho,and the Max Planck Society. The SDSS Web
(a),Mr (b)andMBH (c)forAGNsinmergingpairs(solidlines), site is http://www.sdss.org/.
tidalpairs(dottedlines),non-interactingpairs(dot-dashedlines)
The SDSS is managed by the Astrophysical Research
andAGNsgalaxieswithoutclosecompanions (dashedlines).
Consortium (ARC) for the Participating Institutions. The
Participating Institutions are The University of Chicago,
Fermilab,theInstituteforAdvancedStudy,theJapan Par-
with Dn4000 < 1.8. These trends open the possibility that
ticipationGroup,TheJohnsHopkinsUniversity,theKorean
isolatedAGNswithactiveblackholeshaveexperiencedare-
ScientistGroup,LosAlamosNationalLaboratory,theMax-
cent merger. It is also possible that secular evolution could
Planck-Institute for Astronomy (MPIA), the Max-Planck-
beresponsible offeeding theblack holeand thestar forma-
Institute for Astrophysics (MPA), New Mexico State Uni-
tion activity if a gaseous disk is present. More detail data
versity,UniversityofPittsburgh,UniversityofPortsmouth,
are needed to draw a conclusion on this point.
Princeton University,theUnited States NavalObservatory,
6- We find that the luminosity (or mass in stars) of
and theUniversity of Washington.
the companion galaxy is a key parameter to determine the
activityoftheblackhole.BoththeOIIIluminosityandthe
accretion rates L[OIII]/MBH for AGNs with bright close
companions are significantly larger than for those of AGNs REFERENCES
with faint neighbours.
Alonso,M.S.,Tissera,P.B.,Coldwell,G.&Lambas,D.G.,2004,
MNRAS,352,1088.
Alonso,M.S.,Lambas,D.G.,Tissera,P.B.&Coldwell,G.,2006,
MNRAS,367,1029.
ACKNOWLEDGEMENTS Antonucci,R.,1993,ARA&A,31,473.
Baldry,I.K.;Glazebrook,K.;Brinkmann,J.;Ivezic,Z.;Lupton,
Wethank GuinevereKauffmann for useful comments. R.H.;Nichol,R.C.;Szalay, A.S.,2004,ApJ,600,681.
This work was partially supported by the European Baldwin,J.A.,Phillips,M.M.&Terlevich,R.,1981,PASP,93,
Union’s ALFA-II program, through LENAC, the Latin 5.
AGNs and galaxy interactions 9
Balogh M., Baldry I. K., Nichol R., MillerC., Bower R., Glaze-
brookK.,2004,ApJLetters,615,101.
BaloghM.,etal.,2004, MNRAS348,1355.
Barnes,J.E.&Hernquist,L.,1992, ARA&A,30,705.
Barnes,J.E.&Hernquist,L.,1996, ApJ,471,115.
Barrow, J. D., Bhavsar, S. P., & Sonoda, B. H., 1984, MNRAS,
210,19p.
Begelman,M.C.,Blandford,R.D.,&Rees,M.J.,1980,Nature,
287,307.
Bernardi,M.etal.,2003,AJ,125,1817.
Brinchmann, J., Charlot, S., White, S. D. M., Tremonti, C.,
Kauffmann, G., Heckman, T. & Brinckmann, J., MNRAS,
2004,351,1151.
Coldwell,G.V.&Lambas,D.G.,2006, MNRAS,inpress.
Gunn,J.E.,1979,inActiveGalacticNuclei,ed.C.Hazard&S.
Milton(Cambridge:CambridgeUniv.Press),213.
Dahari,O.,1985, ApJS,57,643.
Fuentes-Williams,T.&Stocke, J.T.,1988,AJ,96,1235.
Heckman, T. M.; Kauffmann, G.; Brinchmann, J.; Charlot, S.;
Tremonti,C.;White,S.D.M.,2004,ApJ,613,109.
Kauffmann,G.,etal.,2003,MNRAS,341,33.
Kauffmann,G.,etal.,2003,MNRAS,346,1055.
Keel,W.C.1996,AJ,111,696.
Kelm,B.,Focardi,P.&Zitelli,V.,2004,A&A,418.25.
Lambas,D.G.,Tissera,P.B.,Alonso,M.S.&Coldwell,G.,2003,
MNRAS,346,1189.
Martinet,L.,1995, FCPh,15,341.
MihosJ.C.&HernquistL.,1996, ApJ,464,641.
P´erez M. J., Tissera P., Lambas D. G., Scannapieco C., 2006,
A&A,449,23.
Sanders,D.B.,Soifer,B.T.,Elias,J.H.,Madore,B.F.,Matthews,
K.,Neugebauer,G.&Scoville,N.,1988, ApJ,325,74.
Schmitt,H.R.,2001,AJ,122,2243.
Schwartz, M.P.,1981,ApJ,247,77.
Shlosman, I., Begelman, M. C. & Frank, J., 1990, Nature, 345,
679.
Storchi-Bergmann, T.; Gonzlez Delgado, R. M.; Schmitt, H. R.;
CidFernandes, R.;Heckman, T.,2001, ApJ,559,147.
Strateva, I.etal.,2001,AJ,122,1861.
Strauss,M.etal.,2002,AJ,124,1810.
Toomre,A.&Toomre,J.,1972,AJ,178,623.
Tremaine,S.etal.,2002,ApJ,574,740.
YorkD.O.etal.,2000, ApJ,120,1579.