A Point Source Excess in Abell 1185: Intergalactic Globular Clusters?

6 Pages·0.35 MB·English

by  Andres Jordan

Accepted forpublicationinAJ
PreprinttypesetusingLATEXstyleemulateapjv.11/12/01
A POINT SOURCE EXCESS IN ABELL 1185: INTERGALACTIC GLOBULAR CLUSTERS?1,2
Andr´es Jorda´n3
DepartmentofPhysicsandAstronomy,RutgersUniversity,Piscataway,NJ08854
[email protected]
Michael J. West
DepartmentofPhysics&Astronomy,UniversityofHawaii,Hilo,HI96720
[email protected]
Patrick Coˆt´e
DepartmentofPhysicsandAstronomy,RutgersUniversity,Piscataway,NJ08854
[email protected]
and
3
0
0 Ronald O. Marzke
2
DepartmentofPhysics&Astronomy,SanFranciscoStateUniversity,SanFrancisco,CA94132
n [email protected]
Accepted for publication inAJ
a
J
ABSTRACT
8
Deep imaging with WFPC2 and the Hubble Space Telescope has been used to searchfor a population

1 of intergalactic globular clusters (GCs) belonging to Abell 1185, a richness class I cluster at cz = 9800
v km s−1. The field is noteworthy in that contains no bright galaxies and yet is centered on the peak
2 of the cluster’s X-ray surface brightness distribution. We detect a total of 99 point sources in this
2 field to a limiting magnitude of I ≃ 26. An identical analysis of the Hubble Deep Field North,
F814W
1 which serves as our control field, reveals a total of 12 objects in the same magnitude range. We discuss
1 possible explanations for this discrepancy, and conclude that the point source excess is likely due to the
0
presence of GCs within A1185. The number and spatial distribution of these GCs are consistent with
3
theirbeingintergalacticinnature,althoughwecannotruleoutthepossibilitythatsomeoftheGCsmay
0
be associated with neighboring bright galaxies. Deeper imaging with the Advanced Camera for Surveys
/
h may resolve this ambiguity.
p
Subject headings: Galaxies: clusters: individual (Abell 1185) – galaxies: evolution – galaxies: star
-
o clusters
r
t
s 1. introduction tions such as individual red giant branch stars (Ferguson,
a
Tanvir & von Hippel 1998;Durrell et al. 2002),planetary
: Although most of the stars residing in galaxy clusters
v nebulae(Theuns&Warren1997;M´endezetal.1997Feld-
are contained within distinct galaxies, it has been known
i meier, Ciardullo & Jacoby 1998; Feldmeier et al. 2002;
X since the study of Zwicky (1951) that clusters often have
Okamura et al. 2002), type Ia supernovae (Gal-Yam
r diffuse stellarcomponentsthatareintergalacticinnature.
et al. 2002) and even HII regions (Gerhard et al. 2002).
a The existence of this low surface brightness starlight has
These studies typically find the intergalactic component
been firmly established by numerous subsequent studies
to comprise 10 – 20% of the total cluster luminosity.
(e.g., Thuan & Kormendy 1977; Melnick, White & Hoes-
Globularclusters(GCs)representanadditional,andpo-
sel 1977; V´ılchez-G´omez, Pell´o & Sanahuja 1994; Bern-
tentially powerful, tracer of this intergalactic component.
stein et al. 1995) but, due to the difficulties inherent in
Indeed, the possibility that galaxy clusters may contain a
the analysis of extended, low surface brightness features,
population of intergalactic GCs (IGCs) has been consid-
thereremainsasurprisinglevelofdisagreementonitsbasic
ered many times (van den Bergh 1958; Muzzio, Mart´ınez
properties including spatial distribution, luminosity and
& Rabolli 1984; White 1987; West et al. 1995). Such an
color (see the excellent review in Durrell et al. 2002).
IGCpopulationmighteitherformin situ,assuggestedby
Our ability to characterize the properties of this inter-
West (1993) and Cen (2001), or arise through the disrup-
galactic stellar component has improved dramatically in
tion or tidal stripping of cluster galaxies. Ample observa-
recent years, thanks in large part to advances in instru-
tionalevidenceshowsthattheselatterprocessesmustplay
mentation and to the targetted analysis of tracer popula-
aroleintheevolutionofgalaxyclusters(e.g.,Weil,Bland-
1 Based on observations with the NASA/ESA Hubble Space Tele- Hawthorn, & Malin 1997; Gregg & West 1998), as pre-
scope obtained at the Space Telescope Science Institute, which is
dicted by numerical simulations of cluster evolution (see,
operated by the Association of Universities for Research in Astron-
omy,Inc.,underNASAcontract NAS5-26555 e.g., Dubinski 1999 and references therein).
2 Based in part on data obtained at the W. M. Keck Observatory, If,assuggestedbytheabovestudies,10–20%oftheto-
which is operated as a scientific partnership among the California talstellarluminosityinclusterstakestheformofadiffuse
Institute of Technology, the University of California, and NASA, intergalactic component, then it is possible to estimate
and was made possiblebythe generous financial support of the W.
the total number of IGCs expected within a given clus-
M.KeckFoundation.
3 ClaudioAnguitaFellow ter. Early-type galaxies in rich cluster environments have
1
2 JORDA´N ET AL.
GC specific frequencies (Harris & van den Bergh 1981) of abovethelocalbackground. TheFWHMofpointsources,
S ≃ 4 (Harris 1991). A cluster such as Virgo, with a requiredtoproperlyclassifythe detections,waschosenon
N
total luminosity of LV ≃ 1.5×1012 LV,⊙ inside a radius the basis of artificial star tests (see below) to be 0′.′16.
of 6◦ (Sandage, Binggeli & Tammann 1985)4 should then Tocarryoutprofile-fittingphotometry,andtoaidinset-
contain ≃ 104 GCs associated with its intergalactic com- ting the parameters for SExtractor, we constructed point
ponent. spreadfunctions(PSFs)forourframesasfollows. Weused
However,theunambiguousdetectionofIGCshasproven DAOPHOT (Stetson 1987) to create artificial WFPC2
difficult, with conflicting claims on whether or not such frames containing 121 stars positioned in a 11×11grid on
IGCs mightalreadyhavebeen detected in nearbyclusters theimage. PSFsfortheF814Wfilterwerekindlyprovided
(e.g.,Westetal.1995;Harris,Harris&McLaughlin1998; by P.B. Stetson. Each artificial frame was then shifted,
Cˆot´e et al. 2001; Hilker 2001; Mar´ın-Franch & Aparicio creating a series of images that had the same dither pat-
2002). Much of the confusion in this debate stems from tern as the actual images for both A1185 and HDF-N.
thefactthat,ifIGCsfollowthesamedistributionwithina Each set of images was combined into a final frame using
given cluster as the other baryoniccomponents (i.e.,clus- the same procedures that were used for the actual obser-
ter galaxiesand X-rayemitting gas),they should be most vations. The final images created in this way thus con-
apparent at, or near, the dynamical center of the cluster. tain point sources constructed from the single-frame PSF
Since most galaxy clusters have a giant elliptical galaxy but have been subjected to the same DITHER process as
in close projection to their center, the direct detection of the actual data. DAOPHOT was used to determine PSFs
IGCs has been hampered by the difficulties involved in which vary quadratically with position in the frames for
disentangling the hypothesizedpopulationof intergalactic both the A1185 and HDF-N frames.
GCs from those intrinsic to the central galaxy. SExtractor assigns each detected object a classification
The richgalaxyclusterAbell 1185ispeculiar inthis re- parameter,CLASS, which ranges from≃ 0 for galaxiesto
spect as its brightest cluster galaxy is offset by ∼ 3′, or ≃ 1 for point sources. CLASS is obtained using a neural
about 120 kpc, from the centroid of its X-ray gas distri- networktrainedonsimulatedimages(fordetailsseeBertin
bution (Mahdavi et al. 1996). Assuming that the X-ray & Arnouts 1996). We consider point source candidates to
centroid marks the dynamical center of the cluster mass have CLASS > 0.8. An optimized value for the FWHM
distribution, and is not the result of an ongoing merger, parameter was determined by carrying out artificial star
then the detection of IGCs in Abell 1185 promises to be tests on the A1185 and HDF-N images, adding 25 stars
more straightforwardthan in most clusters. at a time (so as not to alter the crowding conditions of
In this paper, we present deep HST observations of a the frame) until a total of 5000 stars were added to each
field centered on the peak of the X-ray distribution of chip. TheFWHMquotedabovewasfoundtobethemini-
A1185. These observations were designed to test the IGC mumvaluethatensuresthatthebulkoftheartificialstars
hypothesis by searching for the expected point source ex- brighterthanourdetectionlimitareclassifiedcorrectlyas
cesscomparedtoablankfield. Asshownbelow,theobser- point sources.
vations, which were designed to sample the brightest part Figure 1 shows the distribution of CLASS parameters
of the GC luminosity function at the distance of A1185, for artificial stars with I ≤ 26 mag recovered from
F814W
do indeed contain a point source excess that starts at the theA1185andHDF-Nfields. Themajority(85%)ofthese
expected magnitude for an old GC population. objects have CLASS ≥ 0.8. The fraction of misclassified
artificialstarsdependssensitivelyonthemagnitudecutoff,
2. observations and analysis fallingto∼8%forI ≤25.5and∼5%forI ≤
F814W F814W
WeusedWFPC2onHST toobtainF814W(I)imaging 24.5. Based on these artificial star tests, we find the 50%
ofA1185aspartofprogramGO-8164. Thetotalexposure completeness level lies at IF814W ∼ 26 mag for A1185
time was 13000 s, divided equally among ten, dithered and IF814W ∼ 26.2 mag for HDF-N. As Figure 1 makes
images. The raw data were processed with the standard clear, the distributions for the two fields are similar, as is
STScIpipelineusingthebestavailablecalibrationfiles. As necessary if we are to compare directly the counts for the
a comparison field, we used a subset of the Hubble Deep two fields.
Field North(HDF-N) observations(Williams etal.1996), PSF-fitting photometry was then carried out with
chosen to have a total exposure time identical to that of DAOPHOT, taking the sky from an annulus with inner
the A1185 field. Given the enhanced noise and smaller and outer radii of 5 and 10 pixels. Instrumental magni-
field of view of the PC chip compared to the WF chips, tudesweretransformedintotheVEGAMAGsystemusing
we limit our analysis to the latter. zero-pointstaken fromthe HST WFPC2 Data Handbook
Cosmic ray masks and the shifts required to register (Baggett et al. 2002). A correctionfor foreground extinc-
the images were obtained using the DITHER package in tionwasperformedusingthereddeningcurvesofCardelli,
IRAF.5 The images were combined by applying integer Clayton & Mathis (1989), with the value of E(B − V)
shifts and averaging to produce the final images. SEx- taken from the DIRBE maps of Schlegel, Finkbeiner &
tractor (Bertin & Arnouts 1996) was used to detect and Davis(1998). Afinalcorrectionof0.1mag wasappliedto
classify the sources present on the field, setting the de- correctfromaPSFradiusof0′.′5tooneofinfiniteaperture
tection threshold to 5 connected pixels having counts 2σ (Holtzman et al. 1995).
Those objects having a DAOPHOT χ statistic6 greater
4 Foranadopted Virgodistanceof16Mpc(Ferrareseetal.1996)
than 1.75 were discarded and any point sources brighter
5 IRAF is distributed by the National Optical Astronomy Obser-
vatories, which are operated by the Association of Universities for 6 χ is a robust estimate of the ratio of the observed to expected
ResearchinAstronomy,Inc.,undercooperative agreementwiththe scatter aboutthemodelprofile(Stetson &Harris1988).
NationalScienceFoundation.
Intergalactic Globular Clusters in Abell 1185 3
than 3σ above the expected turnover magnitude of the counts for the HDF-N field are in good agreement with
globular cluster luminosity function at the distance of those of Elson, Santiago & Gilmore (1996) who find 14
A1185 were omitted from the analysis. The threshold on point sources in the same magnitude interval. In what
χ was determined on the basis of the artificial star tests. follows,we examine a number ofpossible explanations for
We assumed M =−8.4 mag and σ =1.4 mag for the this point source excess.
I,TO
absolute magnitude of the turnover and the dispersion of
3.1. Galactic Field Stars
the Gaussian representation of the GC luminosity func-
tion, respectively (Harris 2002). The distance to A1185 is We first consider the possibility that the A1185 field
calculated directly fromits redshift, z =0.0325,assuming containsalargernumberGalacticfieldstarsthandoesthe
H0 = 72 km s−1 Mpc−1 (Freedman et al. 2001). This HDF-N field. However, this explanation appears unlikely
procedure results in the rejection of all objects brighter given the size of the excess (i.e., nearly an order of mag-
than IF814W = 23.1 mag and eliminates 9 objects in the nitude) and the roughly similar Galactic latitudes of the
A1185fieldand10inHDF-N.Forcomparison,6−7galac- two fields: b = 67◦.8 for A1185 and b = 54◦.8 for the HDF-
tic stars with I .23 are predicted by the Bahcall-Soneira N field. Indeed, according to the Bahcall-Soneira Galaxy
Galaxy model (Bahcall & Soneira 1980; Bahcall 1986) in model(Bahcall& Soneira1980;Bahcall1986),bothfields
these fields are predicted to contain only 4-6 stars in the magnitude
As a final check on the point-source nature of the re- range 23 . I . 26. We conclude that Galactic field stars
maining detections, we examined the radial profile of the are unlikely to be the origin of the observed excess.
differencebetweentheobservedprofileofeachsourcewith
3.2. Background Galaxy Cluster
the PSF profile expected at that position in the field.
While a trend between the median difference and radius Couldtheexcessbedue toadistantbackgroundgalaxy
might then be expected if the sources were extended no cluster which happens to fall in the A1185 field? The
such trend was detected, supporting the SExtractor clas- low surface density of high-redshift clusters (e.g. Post-
sifications. manetal.2002)suggeststhatthisexplanationisunlikely.
More importantly, we expect almost all background giant
3. the nature of the excess
ellipticals to be resolved on our WF frames: i.e., the an-
Our final catalog consists of 99 point sources in the gular diameter distance turns around at z ∼ 1.5 in such
A1185field. Bycontrast,wefind12sourcesintheHDF-N a way that, for z & 1, 1′′ corresponds to ≃ 6 kpc. For
field,whichservesasourcontrol. Theluminosityfunctions instance, Lyman break galaxies at z ∼ 3, though physi-
for both fields are shown in Figure 2. For A1185 we also cally compact, are still clearly resolved in deep WFPC2
show the expected luminosityfunction for GCs atthe dis- images, with half-light radii >∼ 0′.′2 (Steidel et al. 1996).
tance of A1185, multiplied by the completeness function A definitive test should be possible with multi-color data
and normalized to the observed number of sources. Our (and particularly IR colors) for the point source popula-
tion in A1185.
M I (mag)
−12 −11 −10 −9
0
12 2
A1185
HDFN
5 A1185
1
10
N 10
8
5
pdf
6 0
0
2
4 5 HDF−N
1
2 N 10
5
0
0
0.0 0.2 0.4 0.6 0.8 1.0
23 24 25 26 27
CLASS
I F814W (mag)
Fig. 1.— Probability density function (PDF) for the SExtrac-
tor CLASS parameter, recovered fromthe artificial star tests. The Fig. 2.— (Top Panel) Luminosity function of point sources de-
solidlineshowsthePDFforA1185whilethedashedlineindicates tected inA1185. Thesolidlineistheexpected luminosityfunction
that for HDF-N. Both distributions were obtained with a normal of GCs at the distance of A1185, multiplied by the completeness
kerneldensityestimate. Thedottedverticallineshowsouradopted function and normalized to the observed counts. (Bottom Panel)
thresholdforpointsourcedetections. Luminosityfunctionofpointsourcesdetected intheHDF-Nfield.
4 JORDA´N ET AL.
3.3. Nuclei of dE,N Galaxies and assuming that the surface density of dwarfs in Virgo
has the approximate form e−r/α with α ≃ 1◦.7 (Ferguson
A third possibility is that our point source catalog in-
&Sandage1989)7,theexpecteddEsurfacedensityatthe
cludes a population of dE,N galaxies. At the distance of
center of A1185 is roughly 5500 deg−2. Thus, we expect
A1185, the bright nuclei of such galaxies would be un-
our WFPC2 field to contain . 8 dE and dE,N galaxies.
resolved. As Binggeli & Cameron (1991) have shown, the
Giventhatthesegalaxieswouldtypicallycontainonly≃5
nucleiofdE,NgalaxiesmimicthebrightendoftheGClu-
clusterseach(e.g.,Lotzetal.2001),andthatourphotom-
minosity function in the Virgo cluster, so it is conceivable
etry samples only the brightest16%of the GC luminosity
that our sample may include some dE,N galaxies, partic-
function at the distance of A1185, we expect only . 8
ularly since our WFPC2 field is locatednear the center of
point sources to be GCs associated with dwarfgalaxies in
A1185 where the galaxy surface density is expected to be
A1185.
high. Therearetwoseparateissuesthatmustbeaddressed
to test the plausibility of this explanation: (1) the prob-
3.5. Faint dE galaxies
ability that dE,N galaxies would be classified incorrectly
by SExtractor as point sources; and (2) overall number We have also explored the possibility that the point
of dE,N galaxies expected in our A1185 field. This latter source catalog in A1185 has been contaminated by faint
point will be examined in detail below. dE galaxies with small exponential scale-lengths. This
To address the first issue, we carried out simulations in possibility has been investigated by adding simulated dE
which artificial dE,N galaxies were added to our frames. galaxies to the WFPC2 frames, with the galaxies mod-
These frames were then run through the same reduction eled as pure exponential disks having scale-lengths of
procedures as the actual observations and, at the end of r = 0.1 kpc. This choice of scale-length corresponds to
e
each run, the classifications returned by SExtractor were the low-end of the values found for Local Group dwarfs
recorded. The dE,N galaxies were simulated using the (Mateo 1998). It is these highly concentrated dwarfs that
data presented in Table 1 of Lotz et al. (2001) which are cause for the greatest concern, as they are the ones
gives total magnitudes, exponential length scales, nuclei most likely to be misclassified at the distance of A1185.
V-band magnitudes and V −I colors for a sample of 27 We added 150 artificial dE galaxies having absolute mag-
dE,N galaxies in the Virgo and Fornax clusters that have nitudesM =−11.7, −11.2, 10.7, −10.2, −9.7and−9.2
I
been observed with HST. To simulate the dE,N galaxies, (1200 in total) to each WF chip and analyzed the images
we first used DAOPHOT to add point sources with the in the same way as the actual data. Only for M =−10.7
I
tabulated I magnitudes after shifting to the distance of and−10.2wereany dwarfsmisclassified,andthen only in
A1185. The task MKOBJECTS in IRAF was then used ∼1%ofthecases. Combiningthisresultwiththefactthat
to add an exponential disk having the appropriate expo- weexpect∼8dEgalaxiesinthis magnituderange(deter-
nential scale, ellipticity and total magnitude (minus the mined from an extrapolation of the fitted Schechter func-
contribution of the nucleus). The colors of the exponen- tion for the dE population in Virgo; Sandageet al. 1985),
tial disks were taken to be (V − I) = 0.65 − 0.022M we find that the number of compact dE galaxies in our
V
and (B−V) = 0.41−0.022M ; these relations were ob- point source catalog is completely negligible.
V
tainedbycombininga[Fe/H]–M relationfordwarfgalax-
V
ies (Coˆt´e et al. 2000) with the color–metallicity relations 3.6. A1185 Globular Clusters
presentedinBarmbyetal.(2000). EachgalaxyintheLotz
Thecalculationsabovesuggestthatthenumberofdwarf
et al.(2001)sample was simulated50 times per WF chip,
galaxies and Galactic field stars expected in the A1185
with the galaxies being simulated in groups of 9 so as not
field fall far short of that needed to explain the point
to overcrowd the fields. These simulations revealed that
source excess. We conclude that the point source excess
all dE,Ngalaxieswereclassifiedcorrectlyas non-stellarin
is probably the result of bona fide GCs in A1185, but
all trials. We therefore conclude that the nuclei of dE,N
is it plausible that the sources are GC associated with
galaxies cannot be the origin of the point source excess in
nearby ellipiticals rather with the cluster as a whole? As
A1185.
Figure 3 shows, a few of the sources are close in projec-
tion to neighboring galaxies, but the bulk of the popula-
3.4. GCs associated with Dwarf Galaxies
tionisnotobviouslyassociatedwithgalaxies. Toexamine
A related possibility is that the point sources detected this issue more quantitatively, we consider the GC sys-
in the A1185 field represent bona fide GCs, but ones that tems of the six bright galaxies that are nearest to the
are associated with dwarf galaxies that may be present WFPC2 field: NGC 3550 (BCG), NGC 3552, CGCG
in A1185. We examine this possibility by calculating the 155-081, 2MASSXi J1110398+284224, MCG+05-27-004
expected number of dE and dE,N galaxies in the A1185 NED01 and MCG+05-27-004 NED02. These six objects
field. Since the velocity dispersions of Virgo and A1185 are marked by crosses in Figure 3, and labelled according
are similar (Binggeli, Tammann & Sandage 1987; Mah- to the above order.
davietal.1996),theirdwarfpopulationscanbecompared ToestimatethenumberofGCsthatareassociatedwith
directly. Based on the Virgo dE+dE,N luminosity func- these galaxies and fall in the WFPC2 field, it is first nec-
tion published by Sandage, Binggeli & Tammann (1985), essary to assume a radial density profile for their GC sys-
expect ∼ 1000 dE and dE,N galaxies in Virgo out to a tems. On theoretical grounds, the profiles of non-central
distance of 5◦. We have adopted a cutoff at the faint end galaxiesare expected to be tidally truncated by the mean
of the luminosity function of MB =−11.5 as there are no 7 Thisvalue of αis appropriate fordE + dE,N fainter than MB ∼
known GC systems belonging to dwarfs fainter than that. −13. Brighterdwarfsexhibitingenerallargerαandthiswoulddrive
Given that A1185 is ∼ 9 times more distant than Virgo, downtheestimatednumberofdwarfsinourfield.
Intergalactic Globular Clusters in Abell 1185 5
cluster field(e.g.,Merritt1984),butprecisevaluesfor the GCs associated with neighboring galaxies, if the galaxies
truncation radii of cluster galaxies have not been deter- are tidally truncated as described above.
mined observationally, nor has the process been probed However,thisestimatedependsrathersensitivelyonthe
adequately in numerical simulations. As a first estimate, assumedrecipefortidallimitation. Wehavethereforecal-
we use the theoretical predictions for the tidal radius put culated the expected contribution when this effect is ne-
forward in Merritt (1984), assuming α = β = 1.5 in his glected entirely. Since we are now interested in the spa-
equation (6a). The cluster velocity dispersion is taken to tial distribution of GCs at very large projected radii, the
be σ = 736 km s−1 (Mahdavi et al. 1996) and the ve- power-law representations of the radial density is inade-
v
locity dispersion for each galaxy is estimated using the quate (Rhode & Zepf 2000). We instead assume that all
Faber-Jackson relation (Faber & Jackson 1976). For the galaxies have GC surface density profiles similar to the
radialdensityprofilesinsidethetidalradius,weusepower r1/4 profile of M49. Beyond r = 100 kpc, the profiles are
laws of the form rα with α given by set to zero, consistent with the results of Rhode & Zepf
α=−0.29M −8.0 (1) (2000). This is a rather conservative approach, as the
V
fainter galaxies will certainly have less extended profiles
(Kaisler et al. 1996). Absolute magnitudes, M , were
V (see Eq. 1). Following the same procedure as described
measured for the A1185 galaxies by running SExtractor
above,wethenfindanexpectedcontributionof∼62GCs,
on a 30s V-band exposure taken with the Low-Resolution
still lowerthan the observednumber of point sources,but
ImagingSpectrometer (Okeet al.1995)atthe W.M Keck
nowwithinafactorof∼50%oftheactualexcess. Wecau-
Observatory on 19 March, 1999. To estimate the total
tion,however,thatthepreviousestimateisprobablymore
number of GCs belonging to each galaxy, we assume a
realistic,astheclustergalaxieshavealmostcertainlybeen
constant GC specific frequency of S = 4 (Harris 1991).
N tidallytruncatedatsome level. Indeed,dynamicalmodel-
Integrating the various GC density profiles over the WF
ingoftheGCsystemofM87—thecentralellipticalinthe
chips, and taking into account the fact that we observe
Virgo cluster — has shown the velocity dispersion profile
only the brightest ∼ 16% of the GC luminosity function,
ofits GC systemto risebeyondr ∼20kpc, as expectedif
wefindatotalcontributionof∼15GCsfromtheGCsys-
the GCs are orbiting in the potential welldefined by both
temsofthesegalaxies.8 Ofthe15expectedGCs,13would
the galaxy and its parent cluster (Coˆt´e et al. 2001). At a
be on WF4. Although this detector is indeed observed to
projected distance of 100 kpc, the gravitational potential
containlargestnumber of pointsources(39), the factor of
in which the GCs orbit is dominated not by the central
sixdiscrepancybetweentheobservedandpredictedcounts
galaxy, but by the cluster itself.
suggestthatthepointsourceexcessisunlikelytobedueto
Could a population of IGCs be responsible for the ob-
8 The galaxy 2MASSXi J1110398+284224 lies partly on the WF4, served point source excess? West et al. (1995) proposed
so we neglect any GCs falling within 10′′ of its center since these a phenomenological model in which the high specific fre-
GCswouldgoundetected inourreductionpipeline. quencies of some centrally dominant galaxies (e.g., Harris
1991) are explained by the presence of a population of
IGCs. In this scenario, putative populations of IGCs pro-
videan“excess”GCpopulationthatelevatestheobserved
1 specificfrequenciesaboveauniversalvalueofSN =4. Fol-
lowing West et al. (1995) and Blakeslee et al. (1997), we
assume that the surface density of IGCs is given by
Σ =Σ T /(1+r2/r2)
IGC 0 X c
where where r and T are the cluster’s core radius and
c X
X-ray temperature, respectively. To find the constant of
proportionality in this relation, we performed a linear fit
to the data presented in Figure 13 of Blakeslee, Tonry
3
& Metzger (1997). Taking into account that the area of
their detector was 22 arcmin2, the predicted IGC surface
density is
Σ =33T /(1+r2/r2) arcmin−2
IGC X c
Using a value of T = 3.9 keV for A1185 (Jones & For-
X
4 man 1999) and recalling that we are sensitive to only the
6
brightest∼16%oftheGCluminosityfunction,weexpect
our WFPC2 field to contain ∼ 110 IGCs. The predicted
5 2 numberofIGCsisthusoftherightorderneededtoexplain
the point source excess in A1185.
4. conclusions
Fig. 3.— Digitized sky survey image of A1185, showing the
WFPC2fieldanddetected pointsources(circles). Northisupand Having considered the expected contributions of many
East is to the left inthis image, which measures 6.′0×6.′0. Crosses different sources, it seems clear that the observed point
denote thesixgalaxieswhoseglobularclustersystemsaremodeled source excess is best explained by a population of GCs in
in § 3. The numbering is as described in the text. The thin lines
A1185. Whether these GCs are intergalacticin nature, or
shows contours of constant X-raysurface brightness based on Ein-
steinIPCobservations. whether they are associated with neighboring galaxies, is
6 JORDA´N ET AL.
anopenquestion. Thedifficultyindecidingbetweenthese Binggeli,B.,Tammann,G.A.,&Sandage, A.1987,AJ,94,251
scenarios lies in the unknown extent to which the cluster Binggeli,B.&Cameron,L.M.1991,A&A,252,27
Blakeslee,J.P.,Tonry,J.L.,&Metzger,M.R.1997,AJ,114,482
meanfieldhasimposedatidallimitationontheindividual
Cardelli,J.A.,Clayton, G.C.,&Mathis,J.S.1989,ApJ,345,245
cluster galaxies. If tidal truncation has been less effective Cen,R.2001, ApJ,560,592
than assumed in the above analysis, then many, and per- Coˆt´e, P., Marzke, R.O., West, M.J., & Minniti, D. 2000, ApJ, 533,
869
haps most, of the observed point sources observed might Coˆt´e,P.,McLaughlin,D.E.,Hanes,D.A.,Bridges,T.J.,Geisler,D.,
be explained by GCs associated with individual galaxies. Merritt,D.,Hesser,J.E.,Harris,G.L.H.,&Lee,M.G.2001,ApJ,
559,828
Ontheotherhand,iftheclustergalaxieshavebeentidally
Durrell, P.R., Ciardullo, R., Feldmeier, J.J.; Jacoby, G.H., &
limitedasprescribedinMerritt(1984),theIGChypothesis Sigurdsson,S.2002,ApJ,570,119
emergesasthemostplausibleexplanationfortheobserved Dubinski, J. 1999, in ASP Conf. Ser. 182, Galaxy Dynamics, ed.
D.R.Merritt,J.A.Sellwood,&M.Valluri(SanFrancisco: ASP),
excess.
491
Based on the evidence above, it seems very likely Elson,R.A.W.,Santiago,B.X.,&GilmoreG.F.,NewAstronomy,1,
that the point sources observed in our WFCP2 field are 1
Fabian, A.C.,Nulsen, P.E.J.,& Canizares, C.R.1984, Nature, 310,
bonafide GCs residing in A1185. Nevertheless, this asser- 733
tion can be tested with deeper imaging of this field. Any Feldmeier,J.J.,Ciardullo,R.,&Jacoby, G.H.1998,ApJ,503,109
Feldmeier, J.J., Mihos, J.C., Morrison, H.L., Rodney, S.A., &
population of GCs, whether or not it is intergalactic in
Harding,P.2002,ApJ,575,779
nature, should be characterized a near-Gaussianluminos- Ferguson,H.C.,&Sandage, A.1989,ApJ,346,L53
ity function. Atthe distanceofA1185,the GCluminosity Ferguson, H.C., Tanvir, N.R., & von Hippel, T. 1998, Nature, 391,
461
function is expected to show a turnover at ITO ≃ 27, Ferrarese,L.,etal.1996,ApJ,464,568
which is well within the imaging capabilities of HST and Fleming,D.E.B.,Harris,W.E.,Pritchet, C.J.,&Hanes,D.A.1995,
AJ,109,1044
the Advanced Camera for Surveys. While the luminosity
Freedman,W.L.etal.2001,ApJ,553,47
function alone would would not allow us to distinguish Gal-Yam,A.,Maoz,D.,Guhathakurta,P.,&Filippenko,A.V.2002,
between GCs belonging to individuals galaxies and those AJ,inpress(astro-ph/0211334).
Gerhard, O., Arnaboldi, M., Freeman, K.C., & Okamura, S. 2002,
associated with the cluster as a whole, it may be possible
ApJ,inpress(astro-ph/0211341)
to discriminate between these alternatives by examining Gregg,M.D.,&West,M.J.1998,Nature,396,549
the spatial distribution of an expanded sample of sources. Harris,W.E.,&vandenBergh,S.1981,AJ,86,1627
Harris,W.E.1986,AJ,91,822
Additional clues to the nature of these GCs, including Harris,W.E.1991,ARA&A,29,543
ages and metallicities, will be possible once optical and Harris,W.E.inStarClusters,editedbyL.LabhardtandB.Binggeli
(Berlin: Springer),223
infrared colors are in hand.
Harris, W.E., Harris, G.L.H., & McLaughlin, D.E. 1998, AJ, 115,
1801
Hilker,M.2001,preprint(astro-ph/0210466)
WewouldliketothankP.B.StetsonforprovidingPSFs
Holtzman,J.A.,Burrows,C.J.,Casertano,S.,Hester,J.J.,Trauger,
for the F814Wfilter andananonymousrefereefor helpful J.T.,Watson, A.M.,&Worthey, G.1995,PASP,107,1065
comments. SupportforprogramGO8164wasprovidedby Jones,C.,&Forman,W.1999,ApJ,511,65
Kaisler,D.,Harris,W.E.,Crabtree,D.R.,&Richer,H.B.1996,AJ,
NASA through a grant from the Space Telescope Science
111,2224
Institute,whichisoperatedbytheAssociationofUniversi- Lotz, J.M., Telford, R., Ferguson, H.C., Miller,B.W., Stiavelli, M.,
tiesforResearchinAstronomy,Inc.,underNASAcontract &Mack,J.2001,ApJ,552,572
Mahdavi, A., Geller, M.J., Fabricant, D.G., Kurtz, M.J., Postman,
NAS5-26555. Additional support for this work was pro- M,&McLean, B,1996,AJ,111,64
videdbytheNationalScienceFoundationfromgrantAST- Mar´ın-Franch,A.&Aparicio,A.2002,ApJ,568,174
Mateo,M.1998,ARA&A,36,435
0205960andthroughagrantfromthe AssociationofUni-
McLaughlin,D.E.,&Pudritz,R.E.1996,ApJ,469,194
versitiesforResearchinAstronomy,Inc.,underNSFcoop- Melnick,J.,White,S.D.M.,&Hoessel,J.1977,MNRAS,180,207
erative agreement AST-9613615 and by Fundaci´on Andes Merritt,D.1984, ApJ,276,26
M´endez, R.H., Guerrero, M.A., Freeman, K.C., Arnaboldi, M.,
under project No.C-13442. This research has made use Kudritzki,R.P.,Hopp,U.,Capaccioli,M.,&Ford,H.1997,ApJ,
of the NASA/IPAC ExtragalacticDatabase (NED) which 491,L23
Muzzio,J.C.,Mart´ınez,R.E.,&Rabolli,M.1984, ApJ,285,7
is operated by the Jet Propulsion Laboratory, California
Okamura,S.,etal.2002, PASJ,inpress(astro-ph/0211352)
Institute of Technology, under contract with the National Oke,J.B.,Cohen,J.G.,Carr,M.,Cromer,J.,Dingizian,A.,Harris,
Aeronautics and Space Administration. This work was F.H., Labrecque, S., Lucinio, R., Schaal, W., Epps, H., & Miller,
J.1995,PASP,107,375
basedinpartonobservationsobtainedattheW.M.Keck
Postman,M.,Lauer,T.R.,Oegerle,W.,&Donahue, M.2002,ApJ,
Observatory, which is operated jointly by the California 579,93
Institute of Technology and the University of California. Rhode,K.L.,&Zepf,S.E.2001,AJ,121,210
Sandage, A.,Binggeli,B.,&Tammann,G.A.1985,AJ,90,1759
We are grateful to the W. M. Keck Foundation for their Schlegel,D.J.,Finkbeiner,D.P.,&Davis,M.1998,ApJ,500,525
vision and generosity. Steidel,C.C.,Giavalisco,M.,Dickinson,M.,&Adelberger,K.1996,
AJ,112,352
Stetson, P.B.1987, PASP,99,191
Stetson, P.B.,&Harris,W.E.1988, AJ,96,909
Theuns,T.,&Warren,S.J.1997, MNRAS,284,L11
Thuan,T.X.,&Kormendy,J.1977, PASP,89,466
REFERENCES vandenBergh,S.1958,Obs.,78,85
vandenBergh,S.1984,PASP,96,236
Baggett,S.,etal.2002,inHSTWFPC2DataHandbook,v.4.0,ed. V´ılchez-Go´mez,R.,Pello´,R.,&Sanahuja,B.1994,A&A,283,37
B.Mobasher,Baltimore,STScI Weil,M.L.,Bland-Hawthorn,J.,&Malin,D.F.1997,ApJ,490,664
Bahcall,J.N.,&Soneira,R.M.1980, ApJS,44,73 West,M.J.1993,MNRAS,265,755
Bahcall,J.N.1986,ARA&A,24,577 West,M.J.,Coˆt´e,P.,Jones,C.,Forman,W.,&Marzke,R.O.1995,
Barmby,P.,Huchra,J.P.,Brodie,J.P.,Forbes,D.A.,Schroder,L.L., ApJ,453,L77
&Grillmair,C.J.2000, AJ,119,727 White,R.E.1987,MNRAS,227,185
Bertin,E.&Arnouts,S.1996,A&AS,117,393 Williams,R.E.etal.1996,AJ,112,1335
Bernstein, G.M., Nichol, R.C., Tyson, J.A., Ulmer, M.P., & Zwicky,F.1951,PASP,63,61
Wittman,D.1995, AJ,110,1507