The Merger History, AGN and Dwarf Galaxies of Hickson Compact Group 59

38 Pages·2012·1.08 MB·English

https://ntrs.nasa.gov/search.jsp?R=20120012974 2019-01-12T19:13:03+00:00Z
The merger history, AGN and dwarf galaxies of
Hickson Compact Group 59
I. S. Konstantopoulos1, S. C. Gallagher2, K. Fedotov2,3, P. R. Durrell4, P. Tzanavaris5,6,7,
1
1 A. R. Hill2, A. I. Zabludoff8, M. L. Maier9, D. M. Elmegreen10, J. C. Charlton1,
0 K. E. Johnson11,12, W. N. Brandt1, L. M. Walker11, M. Eracleous1, A. Maybhate13,
2
C. Gronwall1, J. English14, A. E. Hornschemeier5, J. S. Mulchaey15
t
c
O

3

ABSTRACT
]
O
Compactgroupgalaxiesoftenappearunaffectedbytheirunusuallydenseenvironment. Closer
C examination can, however, reveal the subtle, cumulative effects of multiple galaxy interactions.
. Hickson Compact Group (HCG) 59 is an excellent example of this situation. We present a
h
p photometric study of this group in the optical (HST), infrared (Spitzer) and X-ray (Chandra)
- regimesaimedatcharacterizingthestarformationandnuclearactivityinitsconstituentgalaxies
o
andintra-groupmedium. Weassociatefivedwarfgalaxieswiththegroupandupdatethevelocity
r
t dispersion, leading to an increase in the dynamical mass of the group of up to a factor of 10 (to
s
a 2.8×1013 M(cid:2)), andasubsequentrevisionofits evolutionarystage. Starformationisproceeding
[ at a level consistent with the morphological types of the four main galaxies, of which two are

star-forming and the other two quiescent. Unlike in some other compact groups, star-forming
1
v complexes across HCG 59 closely follow mass-radius scaling relations typical of nearby galaxies.
1 Incontrast,the ancientglobularcluster populationsingalaxiesHCG59AandB showintriguing
0 irregularities, and two extragalactic H ii regions are found just west of B. We age-date a faint
5
stellarstreamintheintra-groupmediumat∼1Gyrtoexaminerecentinteractions. Wedetecta
0
likely low-luminosity AGN in HCG 59A by its ∼1040 erg s−1 X-ray emission; the active nucleus
.
0
rather than star formation can account for the UV+IR SED. We discuss the implications of our
1
findings in the context of galaxy evolution in dense environments.
1
1
Subjectheadings: galaxies: clusters: individual(HCG59)—galaxies: starclusters—galaxies: evolution
:
v — galaxies: interactions — galaxies: active— galaxies: dwarf — galaxies: fundamental parameters
i
X
1. Introduction
r 1DepartmentofAstronomy&Astrophysics,ThePenn-
a
sylvania State University, University Park, PA 16802; Compactgalaxygroupspopulate the highden-
[email protected] sitytailofthegalaxynumberdensitydistribution.
2Department of Physics & Astronomy, The University
ofWesternOntario,London,ONN6A3K7,Canada
3Herzberg Institute of Astrophysics, Victoria, LaSerena,Chile
BCV9E2E7,Canada 10Department of Physics & Astronomy, Vassar College,
4Department of Physics & Astronomy, Youngstown Poughkeepsie, NY12604
StateUniversity,Youngstown, OH44555 11Department of Astronomy, University of Virginia,
5Laboratory for X-ray Astrophysics, NASA Goddard Charlottesville,VA22904
SpaceFlightCenter,Greenbelt,MD20771 12National Radio Astronomy Observatory, Char-
6Department of Physics and Astronomy, The Johns lottesville,VA22903
HopkinsUniversity,Baltimore,MD21218 13Space Telescope Science Institute, Baltimore,
7NASAPost-doctoral ProgramFellow MD21218
8Steward Observatory, University of Arizona, Tucson, 14UniversityofManitoba,Winnipeg,MN,Canada
AZ85721 15CarnegieObservatories,Pasadena,CA 91101
9Gemini Observatory, Casilla603, Colina el Pino S/N,
1
The systems catalogued by Hickson (1982, Hick- to this investigation by quantifying the gas rich-
sonCompactGroups,orHCGs)exhibitsome fea- ness of twelve groups with the relation of Hi-
tures, such as dynamical and evolutionary states, to-dynamical mass, log(MHi)/log(Mdyn). This
elliptical fractions and X-ray properties of the gave rise to the hypothesis of an alternate, two-
intra-groupmedium(IGM)similartogalaxyclus- prongedevolutionarydiagramforHCGs,whichwe
ters, their massive, more populous counterparts. explored in Konstantopoulos et al. (2010). In one
In contrast to the well-studied cluster galaxies, path, the galaxies have strong interactions before
however,the specific effects of the compact group exhaustingtheircoldgasreservoirsforstarforma-
environment on the evolution of its galaxies are tion, in the other, gas is processed by star forma-
not yet clear. tion within individual galaxies prior to late-stage
HCGs are defined through criteria of isola- dry mergers.
tion and surface brightness1 that give rise to self- Furthermore, the mid-IR colors of HCG galax-
gravitating, dense groupings of a few (typically ies show an interesting bimodal distribution that
four) main members. Because of their masses, distinguishesstar-formingfromquiescentsystems.
these galaxies orbit around the group barycen- Walker et al.(2010)interpretthisstatisticallysig-
ter rather sluggishly, with velocity dispersions nificantgapasevidenceforacceleratedgalaxyevo-
on the order of σCG ∼ 250 kms−1 (Tago et al. lution in the compact group environment. Their
2008; Cox 2000), cf. galaxy cluster dispersions similar mid-IR color distributions relate HCGs to
of σcluster ∼ 750 kms−1 (Binggeli et al. 1987; the infall regions of clusters and set them apart
The & White 1986; Cox 2000). This trait makes fromanyothergalaxysamplecompared,interact-
HCGs valuable laboratories for galaxy evolution: ing or quiescent. This theme was expanded by
the low velocity dispersions force some galax- Tzanavaris et al.(2010)whofoundthisgapappar-
ies into strong, prolonged interactions while oth- ent also in the distribution of specific star forma-
ers appear undisturbed but are apparently un- tion rates for HCG galaxies. These observations
dergoing enhanced secular evolution. That is to together point to compact groups as local exam-
say, this latter population is affected by gravita- ples of the plausible building blocks of clusters in
tional interplay with their neighbors, but evolve the early universe.
more subtly, without obvious, strong interactions Inaddition, HCGs, whichareisolatedby selec-
(Konstantopoulos et al. 2010). tion, could potentially help explain the evolution-
Relating the various observational characteris- ary history of some field ellipticals. For example,
tics of compact groups to those of clusters is im- Rubin et al. (1990) originally proposed (see also
portantforunderstandingwhethertheyconstitute Gallagher et al. 2008) that HCG 31 will evolve
theirownclass,oriftheyaresimplymini-clusters. into a single, field elliptical through a wet merger
Perhaps more appropriately, structures like com- (onewheregasisstillavailableduringtheinterac-
pact groups may be considered plausible building tion). ‘Fossilgroups’,theprobableultimatefateof
blocks of clusters at higher z (e.g. Fujita & Goto isolated groupings, were examined by Jones et al.
2004; Rudick et al. 2006). Revealing past inves- (2003), who defined a criterion of diffuse X-ray
tigations of HCGs as a class have focussed on emission in excess of 0.5 × 1042h−2 erg s−1 for
70
gas content. Their members are typically defi- suchaclassification. Thisarisesfromthe process-
cient in H i gas when compared to galaxies of ing of a group’s IGM during a merger (or series
similar morphological types and masses (e.g. the of mergers), but the low total mass of most local
sampleofisolatedgalaxiesinHaynes & Giovanelli compactgroupssuggeststheirpotentialwelllacks
1984). Verdes-Montenegro et al. (2001) proposed the depth required to heat the IGM to X-ray de-
an evolutionary sequence based on mapping the tectable levels (Mulchaey & Zabludoff 1998). Us-
spatial distribution of Hi across a large sample of ing multiple mergers as a vehicle toward a fossil
HCGs. Johnson et al.(2007,hereafterJ07)added groupend-statemapsonepathofgalaxyevolution
fromthe ‘blue cloud’ ofstar-formingdisk galaxies
1 θN ≥ 3 θG, i.e. a circular area defined by three galaxy- tothe‘redsequence’ofquiescentbulge-dominated
mean-radii about the group is devoid of galaxies of com- galaxies (Bell et al. 2004).
parable brightness. A group surface brightness of μ <
26.0magdefines galaxydensity. Fossilgroupformationmay provideananalogy
2
toclustercentersorsub-clumpswherethebuildup onciled. Table 1 summarizes some of the general
ofcDgalaxiesoccurs. Ifthisturnsouttobevalid, characteristicsofthefourgalaxies,whileTable??
thestudyofcompactgroupscouldalsohelpillumi- presents some derived and literature values of the
nate morphological transformations in the inner- mass content and nuclear identifications in the
most cores of clusters. Exploring these different four galaxies.
scenariosmayprovefruitfulforourunderstanding This paper is organized in the following way:
ofgalaxyevolutionandthebuildupofstellarmass Section 2 presents the optical, IR, and X-ray
in the universe. Making meaningful progress in datasetsusedthroughoutthiswork. Section3pro-
this area requires detailed multi-wavelength stud- videsafullaccountoftheyoungandoldstarclus-
iesinordertomapthe rangeofphysicalprocesses ter populations, which we use as our prime diag-
affecting galaxies that are found in these environ- nosticsofcurrentstarformationandancientinter-
ments,determinetheirhistories,andprojecttheir actions. In Section 4 we discuss the main findings
evolution. of this work. Finally, in Section 5 we summarize
A consistent treatment of a large sample of the work presented and offer ties to previous and
HCGs is therefore in order. In this work we future work in this series.
continue the series of Gallagher et al. (2010) and
Konstantopoulos et al. (2010) and providea com- 2. Observations
prehensive, multi-wavelength study of HCG 59.
2.1. HST optical imaging
We will look at the current state of the group
through its star formation and nuclear activity;
The analysis presented in this paper is based
investigate its past through the star cluster pop- largely on HST-ACS/WFC multi-band data. Im-
ulations; try to unravel the history of mergers in ages were taken in the F435W, F606W and
the group; examine its dwarf galaxy system; and F814W bandsintwopointingstocoverallknown
place it in the context of HCGs in general.
giantgroupmembers. Wewillrefertothesefilters
The core of HCG 59 consists of four giant asB435, V606, I814 (andthe setasBVI) to denote
galaxies, a typical number for HCGs in general. the closest matches in the Johnson photometric
The group lies at a distance of 60 Mpc, based system. The notation does not, however, imply
on a recession velocity of vR = 4047 kms−1 a conversion between the two systems. The ob-
(Hickson et al. 1992, corrected to the reference servations were executed on 2007 February 24, as
frame defined by the 3K Microwave Background) part of GO program 10787 (PI: Jane Charlton).
and H0 = 73 kms−1Mpc−1. Three of the galax- Theexposuretimeswere1710,1230and1065sec-
ies, A (type Sa), B (E0), and C (Sc), have onds in the BVI bands respectively. Three equal
seemingly undisturbed morphologies, and the sub-exposures were taken with each filter with
fourth(D,Im)isanunusuallylargeirregularwith a three-point dither pattern (sub-pixel dither-
a normal, peaked light profile. The total stel- ing). Images were reduced ‘on the fly’ to pro-
lar mass of the group is MT∗OT =3.14×1010 M(cid:2) duce combined, geometrically corrected, cosmic-
(fromthe2MASSKs-bandluminosities;Tzanavaris et al. ray cleaned images. For the analysis of point
2010),whiletheHimassofMHi =3.09×109 M(cid:2) sources,weusedthe standardHST pipeline prod-
is comparable to the value expected for the mor- ucts with a nominal pixel scale of 0(cid:4).(cid:4)05 per pixel.
phologicaltypesandstellarmassesofthemember For analysis of the extended sources, we ran
galaxies, according to Verdes-Montenegro et al. MultiDrizzle (Fruchter & Sosey 2009) with the
(2001). This is therefore a somewhat gas-rich pixel scale set to 0(cid:4).(cid:4)03 per pixel to improve the
compact group, given that HCGs typically con- spatialresolution. Theabsoluteimageastrometry
tain only about a third of the Hi expected. On was checked with the world coordinate system of
the other hand, the J07 scheme classifies the Hi the Two Micron All-Sky Survey catalog (2MASS;
content of the galaxy group as a Type II, i.e. in- Skrutskie et al. 2006) by identifying four unsatu-
termediateingascontent,accordingtoitsratioof rated point sources in common; the averageoffset
gas-to-dynamical mass of log(MHi)/log(Mdyn) = was∼0.01(cid:4)(cid:4)inRAandDec. Thefourmaingalaxy
0.81±0.05. These classificationsarebasedondif- I814-bandlightprofileswerefitwithS´ersicprofiles
ferent criteria and the disparity can thus be rec- using GALFIT (S´ersic 1968; Peng et al. 2010a);
3
Table 1: Basic information on HCG 59 main members
Identifier Coordinatesa Type m vR Referencesb
(J2000) H89c (mag) (kms−1)
A: IC 0737 11:48:27.55+12:43:38.7 Sa 14.82 (B) 4109 [1], [2], [3]
.
B: IC 0736 11:48:20.08+12:42:59.5 E0d 15.60 (B) 4004 [4], [2], [5]
C: KUG 1145+129 11:48:32.44+12:42:19.5 Sc 15.90 (g) 4394 [4], [2]
D: KUG 1145+130 11:48:30.64+12:43:47.8 Im 16.00 (g) 3635 [4], [2]
aCoordinates are the centroids from fitting the HST I814-band images of each galaxy with S´ersic profiles. See § 2.1 for more
details.
b[1]:Evansetal.(2010);[2]:deVaucouleursetal.(1991);[3]:Hicksonetal.(1992);[4]:Yorketal.(2000);[5]:Falcoetal.(1999).
cHicksonetal.(1989)
dTheRC3designationfor59Bis‘S0?’,andsothereissomeuncertaintyastoitsclassification.
Fig. 1.— HST BVI color-composite imaging of HCG 59. The red colors of galaxies A and B imply little
star formation, while C and D show some signatures of nebular emission in green. Also visible is the newly
catalogued dwarf galaxy I (Section 4.1).
4
Table 2: Masses and star formation rates of HCG 59 galaxies
ID M∗a MH2b SFRc sSFRd Nucleus e
(×109 M(cid:2)) (M(cid:2) yr−1) (×10−11 yr−1)
A 17.40 10.2 4.99±0.67f 28.66 C/AGN
B 8.29 <8.8 0.02±0.01 0.19 C
C 3.03 <7.2 0.16±0.03 5.15 Hii
D 2.67 <6.6 0.48±0.04 18.15 Hii
a,c,dStellar masses, star formation rates (SFR) and specific SFRs (sSFR) aredrawn from Tzanavarisetal. (2010). The published
stellarmasseswereoffbyafactorof7.4;thesevalueshavebeencorrectedforthiserror.
bFromtheCOobservations ofVerdes-Montenegroetal.(1998)
eFrom Mart´ınezetal. (2010); ‘C’ stands for ‘composite’, i.e. one that falls at the Hii/AGN overlap region, as defined in
Kewleyetal.(2006);seeSections 4.4,4.5fordiscussiononthesedesignations.
fThisvalueisheavilyaffectedbytheAGNingalaxyA,aswillbeelaboratedinSection2.6.
the best-fitting centroidpositions aregivenin Ta- routine in IRAF2: χ values below 3.0; a sharp-
ble 1. ness in the range [−2.0,2.0]; and a photometric
We used the images, presented in Figure 1, to error less than 0.3 mag. Aperture corrections are
characterize the optical morphology of the galax- first measured between 3 and 10 pixels and then
ies, and to detect and photometer star clusters added to the Sirianni et al. (2005) corrections to
and cluster complexes. All reported magnitudes infinity. Finally, foreground (Galactic) extinction
are in the Vega magnitude system. In Section 3, with E(B−V)=0.037is accountedfor using the
we present the analysis of these two scales (clus- standard Galactic extinction law (a correction of
ters and complexes) of the star formation hierar- AV ∼0.12 mag; Schlegel et al. 1998).
chy and also distinguish between young massive In order to fortify the selection against stars,
clusters (YMCs) and globular clusters (GCs). we apply a conservative absolute magnitude cut
at MV < −9 mag, which produces the high-
2.2. Optical point source photometry confidencesample. Wedo,however,definealarger
sample by relaxing the magnitude cut and apply-
Wefollowthesamerationaleappliedinourpre-
ing stricter PSF-fitting criteria to detect globular
viouswork(Gallagher et al.2010;Konstantopoulos et al.
clusters, which are expected to be fainter and
2010)andusestarclusterstoinferthestarforma-
point-like. In order to minimize contamination
tion activity and history in each of the HCG 59
from marginally resolved sources such as com-
galaxies. At the adopted distance to HCG 59 of
pactbackgroundgalaxies,wefollowRejkuba et al.
60 Mpc, we expect some contamination by super-
(2005) and apply hyperbolic filters (starting nar-
giantstars,whichcanhaveabsoluteV-bandmag-
row for bright sources and widening for fainter
nitudes as bright as −8.5 (Efremov et al. 1986).
sources) with a maximum cut-off at the above
Atthisdistance,oneACSpixelmeasures∼13pc,
mentioned criteria of magnitude error, χ and
(cf. the average star cluster radius of ∼4 pc; e.g.
sharpness. We will refer to this as the extended
Scheepmaker et al. 2007), meaning that clusters
sample.
areatmostmarginallyresolvedandcanbeconsid-
The application of these criteria assigns 240
eredpointsourcesforthepurposesofselectionand
brightstar cluster candidates (SCCs) to the high-
photometry. We select clusters using the method
confidencesampleand948totheextendedsample.
described in Gallagher et al. (2010); in brief, we
Specifically, the numbers of detected SSCs (ex-
performtheinitialselectiononmedian-dividedim-
ages,requireselectioninallthreebands,andfilter
the resulting catalog using point spread function 2 IRAF is distributed by the National Optical Astronomy
Observatories, which are operated by the Association of
(PSF) photometry. Our PSF filtering applied the
Universities for Research in Astronomy, Inc., under coop-
following criteria from the output of the ALLSTAR
erativeagreement withtheNationalScienceFoundation.
5
tended sample numbers in parentheses) in galax- all 3 filters. Assuming a Gaussian GC luminosity
ies A through D are 7 (29), 77 (213), 13 (63) and function with a peak at MV = −7.4±0.2 and a
65 (217), with a further 78 (426) objects coinci- dispersion σ =1.2±0.2, our faint-end cutoff then
dentwithwhatwouldbetheintra-groupmedium. samples 39±8% of the entire GCLF.
Wewillprovideafullanalysisoftheseclusterpop- We have selected GC candidates (GCCs) ac-
ulations in Section 3.1. cording to the color-space distribution of Milky
Inordertotestthecompletenessofthefinallist Way GCs. We de-reddened the colors of globular
of SCCs, we used ADDSTAR to add 3000 artificial clusters from the Harris (1996) catalog by their
stars to the image (over the entire field, includ- listed E(B−V) values, and then defined a paral-
ingthe galaxies)inthe apparentmagnituderange lelogrambased on the intrinsic (B−V)−(V −I)
24–28,i.e.absolutemagnitudesof(−9.89,−5.89). colordistributionoftheMWGCs. Thisparallelo-
Because the final catalogue only contains sources gram was then converted to the ACS filters using
detected in all three filters, we include this effect the ‘synthetic’ transformations in Sirianni et al.
by calculating completeness fractions based only (2005). The color selection region adopted here
on artificialstars detected in all three bands (e.g. is 0.10magwiderin(V −I)thanthatusedinthe
Da Rocha et al. 2002). The limiting magnitudes analysis of HCG 7 (Konstantopoulos et al. 2010),
for the 90% and 50% recovery rates are (26.56, but still does not exceed the boundaries of the
27.25), (26.51, 27.19) and (26.47, 27.15) in the Harris (1996) MW GCs. To quantify, 95 of 97
B435 ,V606 andI814 bandsrespectively(afterpho- MW GCs from the Harris (1996) catalog (those
tometriccorrectionsareapplied). Forthedistance with BVI information) lie within this box. All
modulus used of 33.89 mag, m = 26.5 mag corre- point sources in the HCG 59 fields with 1σ er-
sponds to M (cid:4)−7.4 mag. rorbarsthatoverlapourcolorselectionregionare
Our assessment of the state of star formation consideredGC candidates,andareplottedinFig-
in HCG 59 is not limited to star clusters. Star ure 2.
cluster complexes represent a larger scale of star Due to the close (projected) proximity of the
formation,astheopticallyblendedconcentrations galaxies in the group, it is likely that the halo
of gas, stars, and dust that make up small star- GCsineachsystemwillappearsuperposed. Inan
forming regions,and likely include groups of clus- attempt to quantify the GCCs in each galaxy, we
ters. In contrast to star clusters, these can be re- usetherelationshipbetweenthegalacticmassand
solvedtoevengreaterdistancesthanstudiedhere, the radial extent of the GC systems in galaxies of
as the fractal distribution of gas about a galaxy Rhode et al.(2007). Tocomputetheexpectedsize
gives rise to such structures at all scales (e.g. as of each halo, we have adopted the mass-to-light
demonstrated for M33 by Bastian et al. 2007). conversions in that work, although we stress the
general conclusions we reach are not dependent
2.3. Globular Cluster Candidate Selection onthe detailed size of any givenhalo. As the pre-
dicted masses of all of the group galaxies are just
Globular clusters are also selected from the
HST images. Since the process is tuned to the belowthelowestmassgalaxiesintheRhode et al.
(2007) sample, we adopted a radial extent of 15
colordistributionsfoundinHCG59,weprovidea
kpc (or 56(cid:4)(cid:4) at the assumed distance to HCG 59)
full accountbelow. As contaminationfrom super-
for each of galaxies A, B and C. Galaxy D is a
giants is not a problem for objects with GC-like
lower luminosity system, and as seen in Figure 2,
colors, we adopt a fainter magnitude limit to se-
GCCs in this object already lie within the pro-
lect old GC candidates than we used for SCCs.
jected halo of GCCs in galaxy A. Discussion of
We have chosen a cutoff at V606 = 26, which cor-
responds to MV ∼ −7.7 at our adopted distance the individual GC systems in each group galaxy
follows in Section 3.3.
modulus for HCG 59, or slightly more luminous
than the expected peak in the globular cluster lu-
2.4. BackgroundandForegroundContam-
minosity function (GCLF) at MV ∼ −7.4 (e.g. ination
Ashman & Zepf 1998; Harris 2001). The major-
ity of GC candidates brighter than this limit lie Contamination in our color-selected sample
above the 90% photometric completeness level in of GC candidates is expected from a variety of
6
Fig. 2.— Positions of detected globular cluster candidates, marked on HST I814 imaging. We have also
markedtheexpectedGChaloeswithblueboundaries. ‘OuterB’referstothelocationofapossibleexcessof
GCs, discussed in the text. This is symmetric to the ‘A-B bridge’ (with respect to B; see Fig. 4) and could
be a tidally induced redistribution of the GC population (see Sections 3.3 and 4.4).
7
sources, including foreground Milky Way halo 2006; Newberg et al. 2007; Yanny et al. 2009);
stars, reddened young clusters, and unresolved the presence of such streams is not accounted
background galaxies. With the small number of for in traditional Milky Way star count mod-
GC candidates present in some of the HCG 59 els. To investigate the impact of Sgr Stream
galaxies (discussed below), contamination can be stellar populations, in Figure 3 we have overlaid
significant. 12 Gyr isochrones of Marigo et al. (2008) with a
Predictions from Milky Way star count mod- range of metallicities expected for the Sgr lead-
els(theBesan¸conmodelofRobin et al.2003)sug- ing arm, [M/H] ∼ −1 ± 0.5 (e.g. Chou et al.
gestthatonly3–4foregroundMilkyWaystarswill 2007; Yanny et al. 2009), at distances between
appear in the magnitude and color ranges for ex- 26 and 36 kpc onto color-magnitude diagrams
pectedGCsineachofourACSfields. Determining of the point sources in both of our ACS fields
the contamination from younger, reddened clus- (assuming d ∼ 31 kpc, with a spread of ±5
tersismoredifficult,particularlyinthecentralre- kpc; Newberg et al.2007;Niederste-Ostholt et al.
gionsofthelate-typegalaxiesCandD,wheresuch 2010;Correnti et al.2010). Fromthis,weseethat
objects could be present. Unresolved background stars just below the Sgr Stream main sequence
galaxies are not likely contributing in any signifi- turnoff do have colors and magnitudes similar to
cant way to the numbers of objects in our fields; that of the brighter (V606 <25) GC candidates in
analyses of the background objects (with GC-like our study, making some contamination likely.
colors) in HCG 7 (Konstantopoulos et al. 2010) ToestimatethetotalcontaminationinourGCC
showed that the predicted foreground Milky Way sample, we assume those GC candidates that lie
star counts were similar to the observed putative faroutside the GCsystemhalos(asshowninFig-
backgroundcontamination,leavinglittle roomfor ure 2) are instead contaminating sources. The
background galaxies to contribute significantly. one exception to this is a region (called ‘outer B’
We also consider the Pirzkal et al. (2005) anal- in Figure 2) that lies outside the GC system of
ysis of stars in the Hubble Ultra-Deep Field galaxy B, opposite to the direction of galaxy A.
(HUDF). Within the range of colors shown in We will return to this feature below. There are a
Figure 8, they found the main contaminant of totalof18objectsin8.4arcmin2,orabackground
‘void sky’ to be M-stars, however, with a V606- surface density Σback = 2.3±0.5 arcmin−2. This
I814 of ∼ 2.0, they are too red to be considered is much higher than the predicted surface den-
in our analysis. All Main Sequence stars detected sity of MW halo stars from the Besan¸con model
in the HUDF are too bright to be mistaken for (ΣMW ∼ 0.4 arcmin−2), indicating that Sgr lead-
star clusters by our detection algorithm. In fact, ing arm stars are the dominant foregroundsource
the only class of stellar object that can be found of contamination in our sample.
in the color-space occupied by our cluster can- Of course, for the above analysis we are mak-
didates is white dwarfs, which Pirzkal et al. find ing the assumption that these contaminating ob-
to have a density of 1.1±0.3×10−2 pc−3. The jects are not bona-fide ‘intra-group’ GCs that lie
maximum Galactic volume covered by our two faroutsidethemaingalaxiesofthegroup. Totest
pointings is a cube of ∼ 8000 px on a side, or this, we compare the V606 luminosity function for
∼0.24pc−3,assumingascaleheightof400pc(the the background source sample with the luminos-
maximum height quoted by Pirzkal et al.). Such ity function ofthe largeGC candidate population
a volume might be expected to host ∼2.6×10−3 surrounding galaxy B. We show this in the right
white dwarfs. Therefore, we consider this poten- panel of Figure 3. The extrahalo luminosity func-
tial source of contamination to be negligible. tion does not show a sharp rise with increasing
The dominant source of contamination to our magnitude as expected of a GC luminosity func-
GCCsamplewillactuallybe starsfromthe Sagit- tion and exhibited by the GCCs in galaxy B. Al-
tarius dwarf galaxy. Our ACS fields (at l ∼254◦, thoughadefinitivecomparisonisnotpossiblewith
b ∼ +69◦) are superposed on a rather dense part so few objects in the IGM area, this is consistent
of the tidally induced (and bifurcated) leading with these IGM objects being contaminants and
arm (stream ‘A’ from Belokurov et al. 2006) of not a part of a diffuse population of intragroup
Sgr (e.g., Majewski et al. 2003; Belokurov et al. GCs. Thus we adopt the surface density above
8
Fig. 3.—DiagnosticdiagramsrelatedtotheGCpopulation. Ontheleftweshowcolor-magnitudediagrams
of all detected GCCs. The solid and dashed lines show Marigo et al. (2008) evolutionary tracks for 12 Gyr-
oldmainsequencestars,withZ =−0.6and−1.3respectively. The linesaredouble tobracketthe distances
appropriate for the Sagittarius dwarf, with its leading spiral arm superposed along this line of sight. Their
distribution in the CMD dispels our worries of significant contamination. The right panel shows the lumi-
nosityfunctionofglobularclusterscandidates(filledcircles)ingalaxyBandsourcesconsidered‘background’
(open circles and dashed lines). The clear discrepancy supports the quality of our GC selection. The LF of
‘outer B’ sources is designated with a solid line.
9
as that of the ‘background’ in the analyses that attempt to determine whether these are HCG 59
follow. members through a phase-space analysis.
The Hickson (1982) naming convention assigns
2.5. Las Campanas wide-field imaging:
letters in order of brightness. Since our imaging
lowsurface brightnesslightand dwarf
does not cover all five dwarf candidates, we used
galaxies
the SDSS r-band photometryto consistently clas-
We extend the coverage of the HST ob- sify thegalaxiesasHCG59FthroughJ.We have
servations through wide-field imaging with the omitted the letter E, as it was assigned in the
LasCampanasObservatory(LCO)2.5-metertele- original catalog to a background galaxy. We at-
scope. We took B- and R-band images of a 25(cid:4) temptedtomeasurestellarmassesforthesegalax-
diameter around the group with the Wide Field ies using 2MASS K-band images (Skrutskie et al.
Reimaging CCD Camera (WFCCD). The data 2006),however,theyarebelowthedetectionlimit
wereobtainedon2007 July 07aspartofanimag- of that survey. Table 3 summarizes all of the in-
ing campaign that covers all 12 HCGs in the J07 formationpresentedinthissection: measuredand
sample. The B and R filter exposure times were SDSS photometry, radial velocities, galaxy mor-
300 s and 600 s, respectively. phologies and projected distances from the group
barycenter. The latter two properties will be dis-
Theseimagesallowforthedetectionoflowsur-
cussed in Section 4.1
face brightness features, such as the signatures of
past interactions, over a large area. We present
2.6. Spitzer observations: infrared spec-
this analysis in Figure 4, where we stacked the B
tral energy distributions
and R images and applied a Gaussian smoothing
filter to the result. This image shows only fea- The optical imaging was complemented by
tures and R-band contours that register at least Spitzer imaging in the mid-infrared (IRAC 3.6–
3σ above the background. We find two faint fea- 8 μm and MIPS 24 μm observations) presented
tures, a ‘bridge’ that appears to connect galaxies in J07 and shown in Figure 5. In addition to
A and B and an arc extending from B toward a the Rayleigh-Jeans tail of stellar photospheric
compact structure to its north-west (we will later emission, the IRAC bands probe the presence
refer to this as the ‘B-I arc’). There is another of hot dust and polycyclic aromatic hydrocar-
compact, extended source in the space between bons (PAHs), while the 24 μm observations trace
galaxies C and D. cooler thermal dust emission. The dust and PAH
The original purpose of the LCO observing emission are both stimulated by star formation
programwas to prepare a sample of dwarf galaxy activity. The harder spectra of active galactic nu-
candidates for spectroscopic follow-up. Though clei typically destroy PAH molecules while heat-
our redshift survey has yet to cover HCG 59, it is ing dust to hotter temperatures than found in
covered by the Sloan Digital Sky Survey (SDSS; galaxies with star formation alone. At low AGN
York et al. 2000): a spectroscopic search sweep- luminosities, the IR SEDs are often ambiguous
ing a radius of 30 arcminutes around the nom- (particularly in the presence of star formation;
inal center of the group (the geometric center e.g., Gallagher et al. 2008).
of the region enveloping the four known mem- The Spitzer images were combined with JHKS
bers) yields seven spectra with redshifts in the observations from 2MASS (Skrutskie et al. 2006)
range 0.01–0.02: galaxies C and D and five com- to plot the IR spectral energy distribution (SED)
pact galaxies. We therefore consider the mem- of each galaxy (following J07), presented in the
bership of SDSS galaxies J114817.89+124333.1 frequency-spaceplotofFigure6. Wehaveusedthe
and J114813.50+123919.2, which are covered by Silva et al. (1998) templates for galaxies of vari-
our wide-field imaging and J114930.72+124037.5, ous morphological types. These map the SED of
J114940.11+122338.6 and J114912.21+123753.8 different galaxies as the sum of starlight and gas
whichlieatprojecteddistancesgreaterthan13ar- and dust emission from star formation and inter-
cminutesfromthe groupcenter. The firstofthese stellar cirrus. We calculate the spectral index of
galaxies is also present in the HST imaging, but theSEDwithintheIRACbandsthroughasimple
lies partly in the ACS chip gap. In Section 4.1 we power-lawfit. ThiswasdefinedbyGallagher et al.
10