Galaxy Pairs in the 2dFGRS II. Effects of interactions on star formation in groups and clusters

0.46 MB·English

by  M. Sol Alonso

Mon.Not.R.Astron.Soc.000,1–9(0000) Printed2February2008 (MNLATEXstylefilev2.2)
Galaxy Pairs in the 2dFGRS II. Effects of interactions on
star formation in groups and clusters
M. Sol Alonso 1,2, Patricia B. Tissera 1,3, Georgina Coldwell 1,4 and Diego G. Lambas 1,4
1 Consejo Nacional de Investigaciones Cient´ıficasy T´ecnicas.
4 2 Complejo Astron´omico El Leoncito, Argentina
0 3 Institutode Astronom´ıa y F´ısica del Espacio, Argentina.
0 4 ObservatorioAstron´omico de la Universidad Nacional de Co´rdoba, Argentina.
2

n
a
J 2February2008

2
2
ABSTRACT

1
v We analyse the effects of galaxy interactions on star formation in groups and
5 clusters of galaxies with virial masses in the range 1013−1015M⊙. We find a trend
5
for galaxy-galaxy interactions to be less efficient in triggering star formation in high
4
density regions in comparison with galaxies with no close companion. However, we
1
obtain the same relative projected distance and relative radial velocity thresholds
0
for the triggering of significant star formation activity (r ∼ 25h−1 kpc and ∆V ∼
4 p
0 100km s−1)asfoundinthefield.Thus,the natureofstarformationdrivenbygalaxy
/ interactions is nearly independent of environment, although there is a general lower
h
level of star formation activity in massive systems. The above results reflect, on one
p
hand, the local nature of star formation induced by tidal interactions and, on the
-
o other,the roleplayedbytheinternalpropertiesofgalaxies.Byusinga2dFGRSmock
r catalog we estimate the contamination by spurious pairs finding that the statistics
t
s are cleary dominated by real pairs. We also found the behaviour of the trends to be
a
robust against the use of more restrictive relative velocity thresholds. We obtain a
:
v similar radial and relative velocity distribution of the pairs with respect to the group
i centerscomparedtothoseofothertypicalgroupmembers,indicatingthatgalaxypairs
X
have no particular location and dynamics within groups. We find that galaxy pairs
r in rich groups are systematically redder and with a lower present-day star formation
a
activitythanothergroupmembers.Ahigherefficiencyofgalaxy-galaxyinteractionsin
dense regionsin the past, or the fact that early-type galaxies,the majorityin groups,
are dominated by a central stellar spheroid being more stable against tidal torques,
could provide suitable explanations to this observed trend.
Key words: cosmology:theory - galaxies:formation - galaxies:evolution - galaxies:
1 INTRODUCTION galaxy pairs in the 100K 2dF public release, Lambas et al.
(2003, hereafter PaperI) showed that galaxies with a close
Different observations have shown that galaxy interactions
companion have significant larger SF than isolated galax-
are powerful mechanisms to trigger star formation (e.g.,
ies with similar luminosity and redshift distribution only if
Kennicutt et al 1998; Yee & Ellingson 1995). It has been they are at a projected distance rp <25h−1 kpcand radial
also found that both the rates of interactions and mergers velocity ∆V < 100 kms−1. These authors also found that
(LeFevreet al. 2000; Patton et al. 2002) andthecomoving
more luminous galaxies in pairs are the most affected ones
star formation rate of the Universe increase at least up to
byinteractionswhencomparedtoisolatedgalaxieswiththe
z ≈2 (Stanway et al. 2003). Barton et al. (2000) and Lam-
same luminosity, expect for similar luminosity pairs where
bas et al. (2003) have carried out statistical analysis of the
both members haveenhanced SF.
starformation inpairsinthefieldfindingthatproximityin
radial velocity andprojected distancecould belinkedtoan Numericalstudiesofgalaxycollisionsprobethatgasin-
increaseofthestarformation(SF)activity.Asimilarconclu- flowscan betriggered duringgalaxy interactions producing
sionisreachedbyPostman&Geller(1984)fromtheanalysis a sharp increase of the star formation activity even before
ofgalaxieswithinandaroundvoids.Byanalyzing1258field the actual fusion of the galaxies, depending on the orbital
2 Alonso et al.
parameters and the internal dynamical properties of the of the Hα line. Consequently, we restrict to galaxy pairs
systems (e.g., Mihos 1992; Mihos & Hernquist 1994, 1996; at z ≤ 0.1 in order to prevent the results from strong bi-
Barnes & Hernquist 1996). It is well known that galaxies ases. The 2dFGRS spectra are classified by a parameter, η,
dominated by a central stellar component are more stable whichisalinearcombination ofthefirstandsecondprinci-
against tidaltorques.Inthiscase,tidalinteractionsareless palcomponentswhichisolatestherelativestrengthofemis-
efficient in inducing star formation during the orbital de- sion and absorption lines present in each galaxy spectrum.
cayphase(Binney&Tremaine1987).Byusingcosmological Physically,η is related tothespecificstarformation ratein
simulations,Tisseraetal.(2002) foundthattheresponseof a galaxy, given thecorrelation with the equivalent width of
a galactic system to tidal interactions varies along it evo- Hαfound in emission lines galaxies (Bland-Hawthorn et al.
lutionary history, expected to be stronger at early stages 2002). Galaxies with low star-formation rates have typical
of evolution when the systems have internal properties not valuesη<−1.4andactivelystarformingsystemsη>−1.4
suitable for providingstability. (Madgwick et al. 2002). We study star formation induced
On the other hand, it is well known that galaxies in bytidalinteractionsestimatingthestellarbirthrateparam-
groups and clusters have significantly reduced star forma- eter, b=SFR/<SFR> which indicates the present level
tion with respect to those in the field and that, the star ofstarformationactivityofagalaxyrelatedtoitsmeanpast
formation activity depends on the distance to the centre history.Following PaperI,throughoutthisworkweusethe
(Mart´ınez et al. 2002; Dom´ınguez et al. 2002). However, it linear correlation between b and η, b = 0.25η+1.06, as an
is still uncertain how relevant the global environment is in estimateof thestarformation activity in2dFGRSgalaxies.
the regulation of the star formation in galaxies. Loveday, InPaperI,weanalysedasampleof1853galaxypairsin
Tresse&Maddox(1999)foundthatgalaxieswithprominent the 100 K release of the 2dF galaxy redshift survey defined
emission-linesdisplayweakerclusteringthanmorequiescent by a projected distance (rp = 100h−1 kpc and a relative
galaxies. Tegmark & Bromley (1999) also found that early radial velocity (∆V = 350kms−1. (We adopt in this paper
spectraltypesaremorestronglyclusteredthanlatespectral H0=100hkms−1Mpc−1).Theselimitsprovedtobereliable
types.Besides,Carteretal.(2001)suggest thatthetrigger- onestoselectinteractingpairswithenhancedstarformation
ing of star formation occurs on a smaller spatial scale and activity. By applying the same selection criteria, we identi-
whether a galaxy forms stars or not is strongly correlated fiedatotalof9174pairsinthe2dFGRS.Inordertoanalyse
with the surrounding galaxy density averaged over a scale in detail the properties of galaxy interactions in high den-
of a few Mpc. Lewis et al. (2002) confirmed this last result sityenvironmentsweconstructedacatalogofpairsingroups
by studying the environmental dependence of galaxy star- by cross-correlating the total galaxy pairs catalog with the
formation rates near clusters, finding that it is insensitive 2dFGRS group catalog obtained by Merch´an & Zandivarez
to the global large-scale structure in which the galaxy is (2004, in preparation). These authors identified groups by
embedded. The authors also obtained that the distribution using a slightly modified version of the group finding algo-
of star-formation rates is correlated with both the distance rithmdevelopedbyHuchra&Gellerwithaminimumnum-
from the cluster centre and the local projected density (see berof4members,anouternumberdensityenhancementof
also Dom´ınguez et al. 2002). 80 and a linking radial cutoff of 200 kms−1. The sample in
Takingintoaccounttheseresults,inthisworkwefocus thecatalogcomprises6076groupsspanningovertheredshift
on the analysis of galaxy-galaxy interactions within groups range of 0.003≤z ≤0.25 with a mean redshift z ≃0.1. As
andclusterswiththeaimatassessingifthisphysicalmecha- aresult ofthiscross-correlation we obtain asampleof 4658
nismplaysasignificantroleinstarformationtriggering.For galaxies pairs in groups.
this purpose we constructed the hitherto largest sample of Despite of the fact that the 2dF public catalog is
interacting pairs in groups and clusters from the 2dFGRS. not complete, we argue that galaxy pairs searching is not
By means of spectroscopy and colour analysis, we explore severely affected by completeness effects. This is based on
thedependenceofthestarformation ingalaxy pairsonrel- thefactthatalthoughtheminimumfiberseparationfor2dF
ative projected separation, radial velocity and groupcentric spectroscopy is approximately 25 arcsec, the survey strat-
distance. egy was to repeat the measurements in each field with new
fiberpositions inordertoachievethehighestcompleteness.
Thus, from this point of view thereis nobias against small
angularseparationswhichwouldintroducespuriousresults,
2 DATA AND ANALYSIS specially at higher redshifts. Therefore, the inclusion of a
pair in our catalog dependsmostly on the inclusion of each
2.1 Pairs in groups in the 2dFGRS
memberinthesurvey,whichwererandomlyselectedwithin
The 2dF Galaxy Redshift Survey comprises over 220000 the target of each field. We argue that there are not signif-
spectra of galaxies located in two contiguos declination icant selection effects on the pair sample which could bias
strips (Colles et al 2001). The spectral properties of 2dF- ourstatistical results on star formation activity.
GRS galaxies are characterised using the principal compo- Following the procedure outlined of Paper I, we focus
nent analysis (PCA) described by Madgwick et al. (2002). our attention on the effects of interactions on star forma-
This analysis makes use of the spectral information in the tionbycomparingwithasuitablecontrolsamplewhichdif-
rest-frame wavelength range 3700˚A to 6650˚A, thereby in- fers from the pair catalog only on the fact that galaxies in
cluding all the major optical diagnostic between OII and groups in the latter have a close companion. Using Monte
Hα line. For galaxies with z > 0.15, the relation between Carlo algorithm we select for each galaxy pair, two other
the derived star formation rates and the spectral classifica- members of the 2DFGRS group catalog. Therefore, in this
tion can be affected by sky absortion bands contamination paper, the control sample corresponds to 9316 galaxies in
Galaxy Pairs in the 2dFGRS II. Effects of interactions on star formation in groups and clusters 3
groups and clusters which do not have a companion within
rp < 100h−1 kpc and ∆V < 350km/s. We stress the fact
thatthiscomparisonsampleofgalaxiesingroupssharesthe
same environment and has the same redshift distributions
0.3
than thesample of galaxy pairs in groups.
2.1.1 Testing the effects of spurious galaxy pairs
Theselectionofgalaxypairsbyusingprojectedvelocitydif-
0.2
ferences (∆V) and projected separation (rp) has the draw-
backthatspuriouspairscanbeincluded.Theuseofcut-offs
for both variables helps to diminsh the problem, although
they do not solve it complete. In particular, the effects of
spurious pairs are expected to be stronger in high density
regions (Mamon 1986; 1987). 0.1
In order to assess the effects of spurious pairs in our
observationalanalysis,wehaveusedthe2dFGRSmockcat-
alog constructed by Merch´an and Zandivarez (2002) from
a gravitacional numerical simulation of the concordance Λ
cold dark matter universe (Ωm = 0.3,Ωγ = 0.7, H = 0
70kms−1Mpc−3 and σ8 =0.9). The authors performed this 0 0.5 1
simulationsbyusingtheHYDRAN-bodycodedevelopedby D
Couchman, Thomas & Pearce (1995) with 1283 particles in
a cubiccomovingvolume of 180 h−1 Mpcper side, starting
Figure 1. Distribution of normalised groupcentric distances D
at z=50. for close galaxy pairs restricted to rp < 25h−1 kpc (solid lines)
¿From the 2dFGRS mock catalog a mock galaxy pair andthecontrol sample(dashedline).
catalogwasconstructedbyapplyingthesameobservational
cut-offs defined in PaperI: rp < 100 h−1 kpc and ∆V <
350kms−1 .Similary,aclosemockpaircatalogwasobtained sion (σ), respectively (D=RD/RVir and v=V/σ).We re-
by requiring: rp < 25 h−1 kpc and ∆V < 100kms−1. The stricttheanalysistogroupswithmorethan10membersfor
orbital parameters such as major semi-axis, eccentricities, thepurposeofavoidinglargeuncertaintiesinthedetermina-
boundedenergy,etc.,wereestimatedforbothmockpaircat- tion of group centre, mean velocity and velocity dispersion
alogsbyassumingatwo-bodyproblemscenario.Weadopted owing to small number statistics. Since close pairs are the
e < 1 and negative bounded energy to distinguish between most likely tohavetidally enhancedstarformation activity
real and spurious pairs. we further restrict this analysis to close pairs rp < 25 h−1
Wefoundthatforthecompletemockpaircatalog,71% kpc. The resulting distributions of close pairs and of other
were real pairs while for the close catalog, the percentage membergalaxies areshown in Fig.1 and Fig. 2, from where
waslarger:79%.Wealsoimpossedtheconditionofanextra it can be appreciated that close pairs and the other group
neighbour within rp < 400 h−1 kpc and ∆V < 500 kms−1 membersare similarly concentrated and haveacomparable
in order to segregate pairs according to environmemt. For relativevelocity distribution with respect to thehost group
galaxy pairs in dense regions, we found that 73% and 79% centre.
were real pairs in all and close pair mock catalogs, respec-
tively. ¿From these estimations, we conclude that although
spurious pairs are present, real binary systems cleary dom- 2.3 Comparison of the star formation in galaxy
inate the statistics. In agreement with previous work, we pairs and in the control sample
found that the contamination is larger in denser regions,
We have also computed the mean star formation birthrate
however, the differences with that of low density environ-
parameter b = SFR/ < SFR > for different bins of nor-
ment are not significant, at least, when our cut-off criteria
malizedgroupcentricdistanceD forthegroupgalaxypairs,
are adopted. Nevertheless, we have included estimations of
theclose pairs and thecontrol sample.
the effects of spurious pairs along the paper that can help
The results from Fig.3 show clearly that the star for-
to furtherassess theirimpact on theresults.
mation in galaxy pairs and in the control sample strongly
increases for larger groupcentric distance, approaching the
meanvalueforfieldgalaxies intheoutskirts.Thesimilarity
2.2 Position and velocity distribution of galaxy
of thesetrendsin both samples shows, on average, that the
pairs in groups
environment has the same effects in all group members. A
We firstly investigate whether galaxy pairs have a partic- similarbehaviourisfoundfortheclosepairsalthoughtthey
ular radial location in groups with respect to the control havelarger star formation activity,as expected.
sample.Forthispurpose,wehaveanalysed thedistribution In order to improve our understanding of the star for-
ofprojectedradialdistance,RD,andrelativevelocity,V ,of mationpropertiesofgalaxyinpairs,wehavecalculated the
pairswithrespecttothehostgroupcentrenormalisedtothe fraction of strong star forming galaxies (f⋆ =N(b>¯bcon)),
group virial radius (RVir) and group mean velocity disper- where ¯bcon is the mean birth rate parameter of the corre-
4 Alonso et al.
spondingcontrol sample(seealso PaperI).Inthesmall box
ofFig. 3weshowthecorrespondingf⋆ forthegalaxypair,
the close pair and the control samples as a function of the
galactrocentric distance. This figure agrees with the trend
0.3
found for the mean b, indicating that in the central regions
thestarformationactivityisweakerforpair,closepairsand
control galaxies.
0.2
2.4 Star Formation as a function of orbital
parameters
We analyse the dependence of star formation on pair rela-
tiveprojected separation rp andrelativeradialvelocity ∆V
0.1 estimating mean values < b > as a function of rp and ∆V
for our sample of galaxy pairs in groups.
TheresultsareshowninFig.4andFig.5.Itcanbeseen
that,asitoccursinpairsinthefield(PaperI),atrp <25h−1
kpcthestarformationactivityissignificantlyenhancedover
thecontrolsample.Similarly,pairswithsmaller relativera-
0 dial velocity differences have larger mean b values, consis-
0 1 2 3 tent with the results found for field pairs. This behaviour
v
indicatesthatthephysicsof starformation inducedbypair
interactions operates in a similar fashion in high density
Figure 2.Distributionofnormalisedradialvelocitiesv forclose environments, although with a lower general level of star
galaxy pairs restricted to rp < 25h−1 kpc (solid lines) and the formation activity.
controlsample(dashedline).
Given the reduced star formation activity of pairs in
the central regions of groups (Fig. 3), we have carried out
a similar analysis adopting the restriction RD/RVir < 0.5.
Theresultsforthissubsampleofcentrally located pairs are
shown in Fig. 4and Fig. 5 (dashed lines) from which it can
beseen thesimilar behaviourofthecentrally located pairs,
22
albeit with an overall weaker star formation activity.
We detect a trend for the mean star formation birth
rate parameter b of close galaxy pairs to be more densely
populated by early type galaxies with respect to the corre-
sponding mean values (¯b) of thecontrol samples from pairs
in the field (< b > /¯b = 1.43 ±0.14) to pairs in the
11..55 field
densest regions (< b > /¯bgroups = 1.09±0.01). This trend
indicates that tidal torques generated by interactions could
belessefficientinpairsystems(withsimilar orbitalparam-
eters) inhigh densityregions. Thefact thatdensest regions
aremorepopulatedbyearlytypegalaxies suggeststhatthe
different response could due to differences in the dynamics
11
of the galaxies in pairs and the available gas reservoir to
form stars in these systems.
We should also take into account the possibility that
spurious galaxy pairs which are more probable in denser
regions, can contribute to diminish the signal, producing
00 00..22 00..44 00..66 00..88 11 theobservedtrend.Inordertoassesstheeffectsofspurious
DD pairs,weusestricterdifferencevelocitycut-offs,similarlyto
theanalysiswecarriedoutinPaperI.AswecanseefromFig.
Figure3.Meanbparameterversusnormalisedgroupcentricdis- 6, as pairs with larger velocity differences are excluded, the
tance (D) for galaxy pairs (solid line), close pairs (long dashed meanbvaluesbehavesimilartoclosepairs.Fromthisfigure,
line)andthecontrolsample(dashedline)ingroups.Thedotted wesee that spurious pairs could contaminate theresults by
linecorrespondstothemeanbvalueobtained inPaperIforfield reducingthestarformation activitysignal but,theireffects
galaxies. The small box shows the fraction f⋆ of galaxies with are not significant enough to change the trends. The small
b > ¯b in the galaxy pair (solid line), in close galaxy pairs (long
boxinFig.6showsthepointsdistributionofbparameterasa
dashedline)andinthecontrol sample(dashedline).
functionofprojecteddistanceforallgalaxiesinpairs.¿From
thisfigureitisclearthatatsmallprojectedseparationsthere
arefew pointswith low starformations activity.Hence,the
enhacementoftheSFactivityforveryclosepairsseemsnot
Galaxy Pairs in the 2dFGRS II. Effects of interactions on star formation in groups and clusters 5
1.2 1.1
1.1
1
1
0.9
0.9
0.8
0.8
0 20 40 60 80 100 0 100 200 300
Figure4.Meanbparameterversusprojecteddistanceforgalaxy Figure5.Meanbparameterversusrelativevelocityforgalaxies
pairsingroups(solidline)andpairsrestrictedtoD<0.5(dashed ininteracting pairsin groups (solid line)and galaxies with D<
line).Thesmallboxcorrespondtothefractionf⋆ofgalaxieswith 0.5(dashed line).Thesmallboxshows the fractionf⋆ of strong
b>¯b. starforminggalaxiesinthesample.
tobesignificantlybiasedbyinterpolerswhichcanartificially
diminish it.
¿From thisanalysis, we findno indications for an envi- 1.2
ronmental dependence of the relative separations and rela-
tive velocity thresholds for star formation induced by tidal
torques. However, our results might indicate a decrease in
the efficiency of tidal torques to induce SF as we move to 1.1
denser environments(see also Section 2.6).
2.5 Pair bj−R colours in groups 1
Star formation in galaxies can also be studied by analysing
integrated colours which reflect the relative fraction of old
andnewstellarpopulations.Weacknowledgethatmetallic-
0.9
ity could introduce complications in this direct interpreta-
tion, therefore this analysis is complementary of the previ-
ous one, serving to deepen our understanding on star for-
mation and galaxy evolution. By providing an estimate of
0.8
the fraction of new stars, galaxy colours reflect a different
timescale than emission lines. For instance, in the case of
anongoingburstofstarformation prominentemission lines 0 20 40 60 80 100
are expected although the integrated colour is not likely to
beaffected.
Therefore, one may ask whether the distribution of Figure6.Meanbparameterversusprojecteddistanceforgalaxy
colour indexes of galaxies in pairs at a given spectral type pairs in groups with ∆V < 350 kms−1 (solid line) and pairs
differs from that of galaxies in the control sample. For this restrictedto∆V <100kms−1(dottedline).Thesmallboxshows
purpose,we haveanalysed thedistribution of bj−R colour the points distribution of b parameter versus projected distance
forpairsingroups.
indexes available in the 2dFGRS data release. In Fig.7 we
show the histograms corresponding to bj − R colour dis-
tributions of pairs, close pairs and control sample galax-
ies It can be appreciated in this figure that the complete
sample of galaxies pairs and galaxies in the control sample
6 Alonso et al.
0.25
0.2 1.2
0.15
1.1
0.1
0.05
1
0
0.5 1 1.5 0 0.2 0.4 0.6 0.8 1
D
Figure 7. bj−R colour distributions of galaxies in pairs (solid Figure 8. bj−R colour versus distance to the group centre for
thickline)andincontrol sample(solidthinlines).We have also galaxypairs(solidline)andthecontrolsample(dashedline).The
includedthedistributionforthesubsampleofclosegalaxypairs: smallboxshowstothefractionf⋆⋆ ofgalaxiesbj−R<bj−¯ R
rp<25kpcand∆V <100kms−1 (dashedthickline).
on local galaxy density and groupcentric radius while low
have a remarkably similar distribution of bj −R colour in- virial mass groups show no significant trends.
dexes,butclosepairsgalaxiesrestrictedtorp <25kpcand In this subsection, and in order to have a sufficiently
∆V <100 km/s havea distribution shifted tothe blue. large number of pairs, we have modified the 10 minimum
In Fig. 8 we display the mean bj −R colours in bins number member condition to 4 members. These enlarged
of groupcentric distance for galaxy pairs in groups and for groupsampleisdividedintoapoorandrichgroupsubsam-
the corresponding thecontrol sample. This plot shows that ples at MVir=1013.5M⊙.
galaxypairsingroupshavesimilarcoloursthanthosewith- InFig.9weshowthat,asexpected,galaxypairsinrich
out a close companion independently of the distance to the groups show a weaker star formation activity than galaxy
groupcentre.Thisresultisinagreementwiththatshownin pairsinthetotalsample.Weseethat,inhighmassgroups,
Fig. 3 where it was found a similar behavior for the mean thereisnodependenceofthemeanbparameteronrp.Con-
birth rate parameter <b>. versely, galaxy pairs in poor groups have a higher star for-
Similarly to the estimation of the fraction of actively mation activity than pairs in the total sample and show a
starforminggalaxies,wehavecalculatedthefraction f⋆⋆ of significant dependence of mean b parameter on rp. A simi-
galaxy in pairs with bj −R <bj−¯ Rcon where bj−¯ Rcon is lar behaviour appears for b vs ∆V as shown Fig. 10. This
themeancolourofthecorrespondingcontrolsample.Inthe weakdependenceofbonorbitalparametersinthehighmass
small box of Fig. 8 we show the f⋆⋆ for galaxies in pairs group subsample may be associated to a high efficiency of
and the corresponding control sample as a function of the starformationintheseregionsinthepastwhichleaveslittle
galactrocentric distance. fuel for new bursts. Also, taking into account the correla-
tionbetweenthespectraltypeηandmorphologyshowedby
Madgwicketal.(2002),thecentralregionsofclusterswould
bemostly populated byearly typegalaxies dominated bya
2.6 Results restricted to rich and poor group
central stellar spheroid which can provide stability to the
subsamples
systemagainsttidaltorques(e.g.,Binney&Tremaine1987;
Importantobservationalevidenceshowsthatdenseenviron- Mihos&Hernquist1996;Tisseraetal.2002).Hence,inthese
ments can affect many galaxy properties, in particular the systems tidal field could not be that effective to drive gas
star formation rate (Dressler et. al 1985). In studying this inflows and trigger starbursts even if there were gas avail-
point,Dom´ınguez etal. (2002) foundacleardistinction be- able.
tweenhighandlowvirialmassgroups(adoptingamasslimit In Fig 11, we show that mean bj −R colours of pairs
of 1013.5M⊙) in the 2dF release 100 K Group Galaxy Cat- in the rich group subsample are redder than those in the
alog constructed by Merch´an & Zandivarez (2002). These total group sample. It can be also seen a significant decline
authorsfoundthatmassivegroupshaveasignificantdepen- of the mean bj −R colours at small projected separations
denceoftherelativefractionoflowstarformationsgalaxies (rp <25h−1 kpc)forthethreesamples.InFig.12itcanbe
Galaxy Pairs in the 2dFGRS II. Effects of interactions on star formation in groups and clusters 7
1.6
1.6
1.4
1.4
1.2
1.2
1
1
0.8 0.8
0 20 40 60 80 100 0 100 200 300
Figure9.Meanbparameterversusprojecteddistanceforgalaxy Figure10.Meanbparameterversusrelativevelocityforgalaxy
pairsinallgroups(solidline)andintherichandpoorsubsamples pairsinallgroups(solidline)andintherichandpoorsubsamples
(dottedanddashedlines,respectively).Thesmallboxshowsthe (dottedanddashedlines,respectively).Thesmallboxshowsthe
fraction f⋆ of galaxies with b > ¯b in the different subsamples fraction f⋆ of galaxies with b > ¯b in the different subsamples
(same code line). Fractions have been estimated with respect to (same code line). Fractions have been estimated with respect to
thecorresponding¯bvalueofeachsubsample thecorresponding¯bvalueofeachsubsample.
appreciatedasimilarbehaviourforthemeanbj−Rcolours
as a function of pair relative velocity ∆V. 1.4
Thebparametersderivedfromthespectraltypeindexη
haveno resolution for low star forming galaxies (Madgwick
et al. 2002), the majority in high mass groups. Thus, we
computed the fractions of star forming galaxies with b >¯b
which shows similar trends for rich and poor group sub-
samples. Fig. 11 shows that closer pairs are systematically
1.2
redinbothsub-samples,ifinterloperswereamajorproblem
we would not see the decreasing trends with a projected
separations (rp). Fig. 13 shows the relation between mean
bj−Rcoloursforgalaxiesinpairsandthehostgroupvirial
mass. Also we display the results for the control sample.
Wecan seeherethatgalaxy pairsaresystematically redder
than those of the control sample. These trends agree with 1
thoseshowninFig.14,fromwhereitcanbeconcludedthat
galaxiesinpairsinrichgroupshaveapresentdaylowerstar
formation activity than galaxies without close companions.
Although effects by spurious pairs should be always taken 0 20 40 60 80 100
intoaccount,wearguethisresultisrobustagainst spurious
pairsonthebasisofpreviousdiscussions(Sections2..1.1and
2.4) and the fact that a further restriction to thedifference Figure11.Meanbj−Rgalaxycoloursversusprojecteddistance
velocity cut-off yields similar relations albeit blueshifted. forgalaxypairsinallgroups(solidline)andintherichandpoor
subsamples (dotted and dashed lines, respectively). The small
The results of confronting the analysis of the pair and
control catalogs, suggest that galaxy interactions in rich box shows the fraction f⋆⋆ of galaxies with bj −R < bj−¯ R in
allsubsamples.Fractionshavebeenestimatedwithrespecttothe
groups were more efficient in triggering star formation so correspondingbj−¯ Rofeachsubsample.
thatcurrently,thesegalaxiesshowlessstarformationactiv-
ity and have reddercolours than other group members.
8 Alonso et al.
1.4
1.4
1.2
1
1.3 (a)
0.8
1.2
0.9
0.8
1.1
0.7
0.6
(b)
1
0.5
0 100 200 300
Figure 12. Mean bj−R galaxy colours versus relative velocity Figure 14. Mean b parameter versus group virial masses for
forgalaxypairsinallgroups(solidline)andintherichandpoor galaxypairs(solidline)andcontrolsample(dashedline)ingroups
subsamples(dottedanddashedlines,respectively).Thesmallbox oflowandhighmass(aandbrespectively).Thedottedlinerepre-
shows the fraction f⋆⋆ of galaxies with bj −R < bj−¯ R in the sentsthemeanbparameterofisolatedgalaxiesinthefield(paper
differentsubsamples.Fractionshavebeenestimatedwithrespect I).
tothecorrespondingbj−¯ Rofeachsubsample.
3 CONCLUSIONS
1.2 Wehaveanalysed thepropertiesof pairsofgalaxies inhigh
(a)
density environments corresponding to groups and clusters
ofgalaxies with virial masses 1013−1015M⊙.Westress the
1.1 fact that the analysis discussed in this paper is based on
the comparison between two galaxy samples which differed
betweeneachotheronlyonthefactthatoneofsamplecom-
1
prisesgalaxieswithclosecompanions,andtheother,donot.
Hence, we are always estimating the statistical differences
0.9 inthepropertiesofgalaxiesintroducedbythepresenceofa
companion.
The main conclusions can besummarized as follow:
1.3
(b) • We find that the radial and relative velocity distribu-
tionsofclosepairs,normalisedtothegroupvirialradiusand
1.25
group mean velocity dispersion, with respect to the group
1.2 centersaresimilartothoseofgroupmemberswithnocom-
panions indicating a no particular location for galaxy pairs
1.15
within thegroups.
1.1 • The general lower efficiency of star formation in dense
environmentsisaccompaniedbyalowerenhancementofthe
starformationinducedbyinteractionsassociatedtogalaxies
of early-morphology which mostly populate rich systems.
• Similarrelativeprojectedseparationandrelativeveloc-
Figure13.Meanbj−Rgalaxycoloursversusgroupvirialmasses itythresholdsrp ∼25h−1 kpcand∆V ∼100kms−1 within
for galaxy pairs (solid line) and control sample (dashed line) in which star formation is significantly enhanced with respect
groupsoflowandhighmass(aandbrespectively). togalaxies without aclose companion, is obtained for pairs
in the field and in modest mass groups.
• Wefindthatstarformationstronglyincreasesforlarger
groupcentricdistanceforbothpairs andthecontrolsample
in a similar fashion.
Galaxy Pairs in the 2dFGRS II. Effects of interactions on star formation in groups and clusters 9
• Thedistributionof bj−Rcolour indexesof galaxies in MadgwickD.S.etal.,2002a, MNRAS,333,133(M02).
closepairsarebluerthanofgalaxiesingroupswithnoclose MadgwickD.S.,Somerville,R.,Lahav,O.,Ellis,R.,2002b,MN-
companions. RAS,submitted.
• Forthesubsampleofpairsinhighmassgroups,wefind Mamon,G.A.,1986,ApJ,307,426.
Mamon,G.A.,1987,ApJ,321,622.
aweakerdependenceofthestarformationbirthparameterb
Mart´ınez,H.J.,Zandivarez,A.,Dom´ınguez,M.,Mercha´n,M.E.,
onrelativeprojected distanceandrelativevelocity.Inthese
Lambas,D.G.,2002,MNRAS,333,31.
massive systems,galaxies in pairs aresystematically redder
Mercha´n,M.,Zandivarez,A.,2002,MNRAS,335,216.
and with a lower present-day star formation activity than
Mihos, J. C., Richstone, D. O., Bothun, G. D., 1992, ApJ, 400,
galaxies in thecontrol sample. 153.
Galaxies in groups, dominated by a central stellar Mihos,J.C.,Hernquist,L.,1994,ApJ,425,13.
spheroid would be more stable against tidal torques gen- Mihos,J.C.,Hernquist,L.,1996,ApJ,464,641.
erated by interactions (Binney & Tremaine 1987; Mihos & Patton, D. R., Pritchet, C. J., Yee, H. K. C., Ellingson, E. &
Hernquist 1996; Tissera et al. 2002) which would not be Carlberg,G.,1997,AJ,475,29.
veryeffectiveindrivinggasinflowsandtriggeringstarbursts, Patton, D.R.,etal.,2002,ApJ,565,208.
Postman,M.&Geller,M.J.,1984,ApJ,281,95.
evenifthereweresufficientgasavailable.Also,wemaycon-
Stanway, E. R., Bunker, A. J., McMahon, R. G., Ellis,
siderthepossibilitythatgalaxyinteractionshavebeenvery
R.S., Treum T., McCarthy, P. J., ApJ, submitted (astro-
efficient in stimulating star formation in the past leaving
ph/0308124)
little fuel for new bursts.
Tegmark,M.&Bromley,B.C.,1999, ApJ,518,L69.
The analysis of a mock catalog of galaxy pairs derived TisseraP.B.,2000,ApJ,534,636.
from numerical simulationns to mimic the 2dFGRS catalog Tissera P.B., Dom´ınguez-Tenreiro R., Scannapieco C., Sa´iz A.,
showed that the statistics are not dominated by spurious 2002,MNRAS,333,327.
pairs. This finding is supported by the fact that the results ToomreA.,ToomreJ.1972, ApJ,178,623.
are robust against stricter velocity differencecut-offs. Yee,H.K.C.&Ellingson,E.,1995, ApJ,445,37.
4 ACKNOWLEDGMENTS
We acknowledge useful discussions with Manuel Merch´an
and Michael Balogh. WethankstheanonymousReferee for
a detail revision that helped to improve this paper. This
workwaspartiallysupportedbytheConsejoNacionaldeIn-
vestigaciones Cient´ıficas y T´ecnicas, Agencia de Promoci´on
de Ciencia y Tecnolog´ıa, Fundaci´on Antorchas and Secre-
tar´ıa de Ciencia y T´ecnica de la Universidad Nacional de
C´ordoba.
REFERENCES
Barnes,J.,Hernquist,1996,ApJ,471,115.
BartonE.J.,GellerM.J.,KenyonS.J.,2000, ApJ,530,660.
Barrow, J. D., Bhavsar, S. P., & Sonoda, B. H., 1984, MNRAS,
210,19p.
Baugh,C.M.,Cole,S.&Frenk,C.S.,1996,MNRAS,283,1361.
Bland-Hawthorn,J.,2002dshg.conf,373.
Binney, J., Tremaine, S., 1987, Galactic Dynamics (Princeton
UniversityPress.
Carter, B. J., Fabricant, D. G., Geller, M. J., Kurzt, M. J. &
MacLean,B.,2001,AJ,556,606.
Colless, M. M., et al. (the 2dFGRS Team), 2001, MNRAS, 328,
1039.
Dom´ınguez, M.J.,Zandivarez, A.A.,Mart´ınez, H.J.,Mercha´n,
M.E.,Muriel,H.,Lambas,D.G.,2002,MNRAS,335,825.
Donzelli,C.J.,Pastoriza,M.G.,1997,ApJS,111,181.
DresslerA.,1980,ApJ,236,351.
Dressler A., Thompson, I. B., Shectman, S. A., 1985, ApJ, 288,
481.
Grogin,N.A.&Geller,M.J.,2000, AJ,119,32.
KennicuttR.Jr.,1998,ApJ,498,541
Lambas, D. G., Tissera, P. B., Alonso, M. S. Coldwell, G. 2003,
MNRAS,346,1189.
Larson,R.B.,Tinsley,B.M.,1978, ApJ,219,46.
LeFevre,O.etal.,2000,MNRAS,311,565.
Lewis,I.,etal.(the2dFGRSTeam),2002,MNRAS,334,673.
Loveday, J.,Tresse,L.,&Maddox,S.,1999,MNRAS,310,281.