AMUSE-Field II. Nucleation of early-type galaxies in the field vs. cluster environment
Vivienne F. Baldassare, Elena Gallo, Brendan P. Miller, Richard M. Plotkin, Tommaso Treu, Monica Valluri, Jong-Hak Woo
DD RAFT VERSION N OVEMBER
11, 2018
Preprint typeset using L A TEX style emulateapj v. 5/2/11
AMUSE-FIELD II. NUCLEATION OF EARLY-TYPE GALAXIES IN THE FIELD VS. CLUSTER ENVIRONMENT V IVIENNE
F. B
ALDASSARE , E LENA G ALLO , B RENDAN
P. M
ILLER , R
ICHARD
M. P
LOTKIN , T OMMASO T REU , M ONICA V ALLURI , J ONG -H AK W OO Draft version November 11, 2018
ABSTRACTThe optical light profiles of nearby early type galaxies are known to exhibit a smooth transition from nu-clear light deficits to nuclear light excesses with decreasing galaxy mass, with as much as 80 per cent of thegalaxies with stellar masses below M (cid:12) hosting a massive nuclear star cluster. At the same time, whileall massive galaxies are thought to harbor nuclear super-massive black holes (SMBHs), observational evidencefor SMBHs is slim at the low end of the mass function. Here, we explore the environmental dependence ofthe nucleation fraction by comparing two homogeneous samples of nearby field vs. cluster early type galax-ies with uniform Hubble Space Telescope (HST) coverage. Existing
Chandra X-ray Telescope data for bothsamples yield complementary information on low-level accretion onto nuclear SMBHs. Specifically, we reporton dual-band (F475W & F850LP) Advanced Camera for Surveys (ACS) imaging data for 28 out of the 103field early type galaxies that compose the AMUSE-Field Chandra survey, and compare our results against thecompanion HST and Chandra surveys for a sample of 100 Virgo cluster early types (ACS Virgo Cluster andAMUSE-Virgo surveys, respectively). We model the two-dimensional light profiles of the field targets to iden-tify and characterize NSCs, and find a field nucleation fraction of +17% − (at the 1 σ level), consistent withthe measured Virgo nucleation fraction across a comparable mass distribution ( +17% − ). Coupled with theChandra result that SMBH activity is higher for the field, our findings indicate that, since the last epoch of starformation, the funneling of gas to the nuclear regions has been inhibited more effectively for Virgo galaxies,arguably via ram pressure stripping. INTRODUCTIONThe assembly and merging history of supermassive blackholes (SMBHs) at the centers of massive galaxies appears toproceed in close connection with – and possibly even regu-late – the growth of their host galactic bulges. Perhaps themost well known incarnation of this is the “ M BH - σ ” relation,or the correlation between the mass M BH of a SMBH andthe velocity dispersion σ of the bulge (Ferrarese & Merritt2000, Gebhardt et al. 2000, Tremaine et al. 2002, G¨ultekinet al. 2009, McConnell & Ma 2013; see also Woo et al. 2010,2013 for M- σ relation for active galactic nuclei). Tight scalingrelations have also been claimed between SMBH mass andbulge mass/luminosity (Marconi & Hunt 2003, H¨aring & Rix2004). While the number of reliable dynamical measurementsfor black hole mass has grown by a factor of about 5 overthe last decade or so, it remains unclear whether these powerlaw scaling relations break down at the highest and lowestmasses (Lauer et al. 2007, Greene et al. 2010, Kormendy &Ho 2013), and whether classical vs. pseudo-bulges lead to dif-ferent scaling relations (see e.g. Jiang et al. 2011, Kormendyet al. 2011). Related to the above issues, the black hole massfunction itself is largely unconstrained at low masses (Greene& Ho 2007, Kelly et al. 2010), and its determination may inturn be biased by the assumption of a single power-law rela-tion with log-normal scatter (Kelly & Merloni 2012).Based on detailed morphological analysis of a large sample Department of Astronomy, University of Michigan, Ann Arbor, MI48109 Physics and Astronomy Department, Macalester College, Saint Paul,MN 55105 Physics Department, University of California, Santa Barbara, CA93106 Astronomy Program, Department of Physics and Astronomy, SeoulNational University, Seoul, Republic of Korea of nearby early type galaxies spanning over four decades instellar mass ( M (cid:63) ), Ferrarese et al. (2006a) proposed that, forgalaxies with M (cid:63) less than a few times M (cid:12) , compactstellar nuclei – with half-light radii between 2-5 pc and about20 times brighter than typical globular clusters (B¨oker et al.2004) – may take over from SMBHs as the dominant form ofmass aggregation in galactic nuclei (a similar result has beenreported for a comparably large sample of spiral galaxies byRossa et al. 2006).Most relevant to this paper, detailed work on the occurrencerate and properties of these “nuclear star clusters” (NSCs;B¨oker et al. 2004; Walcher et al. 2005; Carollo et al. 1998;Matthews et al. 1999; Balcells et al. 2003) was carried outwithin the Advanced Camera for Surveys Virgo Cluster Sur-vey (ACS VCS; Cˆot´e et al. 2004), consisting of dual-band Hubble Space Telescope (HST) ACS observations of 100early-type galaxies in the Virgo Cluster. Later augmented bythe Fornax Cluster Survey (Jord´an et al. 2007), the ACS VCSshowed that between 66% and 82% of early-type galaxieswith absolute B-magnitude M B < − host nuclear star clus-ters, a much larger percentage than had been thought basedon ground observations (Cˆot´e et al. 2006, Turner et al. 2012).Cˆot´e et al. (2007) confirmed that NSCs preferentially residein galaxies with − . < M B < − mag, and found thatthere is a smooth transition from nuclear light deficit to nu-clear light excess with decreasing galaxy luminosity. Also us-ing the ACS VCS data, Ferrarese et al. (2006a) found that themasses of NSCs correlate with the virial masses of their hostspheroidal galaxies, and that this relation extends from thescaling relation between SMBH masses and the bulge massesof their host galaxies, possibly indicating a common growthmechanism for NSCs and SMBHs, and perhaps a shared for-mation mechanism. A common scaling relation for NSCs andSMBHs was also found independently by Wehner & Har- a r X i v : . [ a s t r o - ph . GA ] J un ris (2006). These results would be consistent with a sce-nario whereby SMBHs are the dominant – perhaps sole –mode of nuclear mass aggregation at the center of bright mas-sive galaxies, becoming progressively less common down themass function and disappearing entirely at the faint end, to bereplaced by NSCs (the existence of a common scaling rela-tion for the SMBHs and the NSCs has been by challenged byGraham 2012 and Leigh et al. 2012).Over the last few years, an increasing number of galaxieshosting both NSCs and SMBHs have been identified (Gra-ham & Spitler 2009; Neumayer & Walcher 2012; Gallo et al.2010). In particular, Seth et al. (2008a) took a somewhat dif-ferent approach from previous studies by searching for activegalactic nuclei (AGN) in 176 galaxies that were previouslyknown to contain NSCs. Based on their analysis, at least 10%of their sample – spanning a wide range in masses and Hub-ble types – hosts both NSCs and SMBHs, strongly suggestingthat galaxies harboring NSCs “have AGN fractions consistentwith the population of galaxies as a whole”. In order to bet-ter understand the connection between these objects, as wellas their respective formation mechanisms, it is desirable toundertake systematic studies that characterize both NSCs andSMBH activity over a sample that is unbiased with respect tonuclear properties.With the goal of delivering the first unbiased census oflow-level SMBH activity in the local universe, the AMUSE-Virgo (AGN Multi-wavelength Survey of Early-Type Galax-ies; Gallo et al. 2008, Gallo et al. 2010, Miller et al. 2012b,Leipski et al. 2012 ) survey acquired Chandra X-ray Telescope
ACIS (Advanced CCD Imaging Spectrometer) observationsfor all 100 early-type galaxies targeted by the ACS VCS. Asthe survey was tailored to probe down to Eddington scaledX-ray luminosities as low as log( L X / L Edd ) (cid:39) − , it providesa relatively inexpensive (i.e., compared to dynamical studies)way to identify SMBHs in formally “inactive” nearby galac-tic nuclei. While, in the absence of a NSC, the issue of po-tential contamination to the nuclear X-ray signal from brightlow mass X-ray binaries (LMXBs) can be addressed quan-titatively based on the known shape and normalization of theX-ray luminosity function of LMXBs (Gilfanov 2004), thepresence of a NSC demands a more conservative treatment(see Gallo et al. 2010 for a detailed discussion). For the VCSsample, the combined Chandra and HST data indicate that be-tween 24% and 34% of the targeted galaxies host a bona fide X-ray active SMBH. The fraction of hybrid nuclei, hostingboth a SMBH and a NSC is estimated between 0.3% and 7%for M (cid:63) below M (cid:12) and to be lower than 32% above it (atthe 95% confidence level; Gallo et al. 2010).Born as an extension of AMUSE-Virgo, the AMUSE-FieldChandra survey (Miller et al. 2012a, Miller et al. 2012b,Plotkin et al. 2014 on ULXs, Miller et al. 2014, submitted)was designed to deliver the first measurement of low-levelSMBH activity in a field environment. AMUSE-Field tar-geted a volume-limited sample of 103 nearby field early typesspanning over three orders of magnitude in host stellar mass;analogous to the Virgo sample, the field galaxies were se-lected based solely on optical properties as classified by Hy-perLeda (Paturel et al. 2003), to create a sample that is un-biased with respect to nuclear properties. For the field sam-ple, 45% ±
7% of targets were found to host an X-ray activeSMBH (Miller et al. 2012a). However, this measurement re- The sample is comprised of early type galaxies only, ensuring negligiblecontamination from high mass X-ray binaries. lies on the assumption that the fraction of field objects hostinga NSC is the same as that found for the AMUSE-Virgo targets.In order to properly compare the incidence of SMBH activityas well as stellar nucleation as a function of host stellar massfor the field sample as well as for Virgo, we acquired dual-band HST/ACS observations for a subsample of the AMUSE-Field galaxies.Combined, the Chandra/ACIS and HST/ACS data of theVirgo and field samples provide uniform multi-wavelength in-formation on the frequency of SMBHs and NSCs in the localuniverse, across the mass spectrum and across environment.In this work, we report on the analysis of the ACS obser-vations of the field targets. We model the galaxies’ surfacebrightness profiles to determine what fraction host NSCs, andcompare our results to those from the Virgo cluster. This pa-per is organized as follows. Section 2 describes our sample,data reduction, and analysis. Section 3 presents our resultson the fraction of nucleated (as in hosting a NSC), early-typefield galaxies, and our comparison to the fraction found for theVirgo Cluster. Section 4 discusses the implications of theseresults in the context of NSC formation models. DATA ANALYSISThe full AMUSE-Field sample is comprised of 103 early-type galaxies. We refer the reader to Section 2 of Miller et al.(2012a) for a detailed description of the selection criteria. OurHST program aimed at acquiring dual band images for thegalaxies with detected nuclear X-ray emission from Chandra(52 out 103 objects). Out of those, 8 already had archival HSTdata in both the F475W and F850LP filters. The remaining 44targets were approved for a Snapshot survey in Cycle 19 (PI:Gallo, ID 12951), and 17 of them were eventually observed,with 340.0 s and 340.0 + . (cid:48)(cid:48) . (cid:48)(cid:48) g and z bands defined in Table 7 Sloan g and z bands are roughly equivalent to HST F475W and F850LPbands, respectively. of Appendix A of Bell et al. (2003). In calculating masses,magnitude corrections were made for the slight deviations be-tween the HST F475W & F850LP and Sloan g & z filters (EricBell, private communication). Excluding the three aforemen-tioned non-early type galaxies and two galaxies (NGC 1370and NGC 3073) for which heavy dust contamination aroundthe nucleus prevented us from doing accurate photometry, ouranalysis is based on a sample of 23 early-type, field galaxies.Properties of individual galaxies are listed in Table 1.For the purpose of a meaningful comparison to the VirgoCluster nucleation fraction, we follow the definition of nucle-ation given by Cˆot´e et al. (2006), whereby a nucleated galaxyis one whose inner light profile lies systematically above thatof a S´ersic profile fit to the galaxy excluding the inner 0 . (cid:48)(cid:48) . (cid:48)(cid:48) (cid:48)(cid:48) , i.e. Dullo & Gra-ham 2013). While GALFIT does not allow the position angleand isophotal ellipticity to radially vary within a component,we did not constrain components to have the same positionangle and ellipticity. Example GALFIT models and residualsare shown in Figure 1.The presence of a NSC is signaled by a light excess with re-spect to the extrapolation of the best-fit galaxy model withinthe inner 0 . (cid:48)(cid:48)
5. When present, such excess was fit with an addi-tional S´ersic component. A good fit was chosen to have resid-uals between the data and model consistently between − + RESULTS3.1.
Nucleation Fraction
Following Cˆot´e et al. (2007), we adopt the ∆ . parameterto quantify the degree of nucleation in our targets. ∆ . isdefined as log L g /L s where L g is the total luminosity of thebest-fit galaxy model inside a break radius R b , and L s is theluminosity of just the outer S´ersic component in this region.Break radius R b is equal to . R e , which is where deviations A S´ersic profile is defined as I S´ersic ( R ) = I e exp (cid:40) − b n (cid:34)(cid:18) RR e (cid:19) /n − (cid:35)(cid:41) (1)where R e is the effective radius which encloses half of the model’s light, I e isthe intensity at the effective radius, n is the S´ersic index, and b n is a constantthat is dependent on n . from the outer S´ersic profile tend to occur. Negative ∆ . values indicate nuclear light deficits; these “cored” galax-ies tend to be very luminous ellipticals (Graham & Guzm´an2003). A positive ∆ . value indicates a nuclear light excess,i.e. a NSC. ∆ . is plotted against absolute B-band magnitude (fromMiller et al. 2012a) in Figure 6 for 22 of the 23 objects inour analysis; we exclude NGC 4697 from this calculation be-cause its nuclear disk precluded us from calculating an accu-rate ∆ . value. For objects with nuclear light deficits, wecalculate L g based on the flux inside R b , as measured by EL-LIPSE, as opposed to the best fit model, and use the outerS´ersic component to calculate L s . The Spearman rank for thisrelation, calculated using the ∆ . values measured from theg band data, is 0.36 with a p-value of 0.11, indicating a posi-tive correlation between magnitude and ∆ . at the 89% con-fidence level. Unlike Cˆot´e et al. (2007), who find a consistenttrend from nuclear light deficit to excess with a clearly definedtransition region between − < M B < − . mag, we donot observe such a sharp transition region for our sample. Wecheck that the same qualitative results hold if the analysis isperformed in the z band.Next, we compare our results for early-type field galaxiesto those for the 100 early-type galaxies that compose the ACSVCS (Cˆot´e et al. 2004), in order to test whether the nucle-ation fraction has an environmental dependence. In order toproperly account for the different mass distributions acrossthe two samples, we use the procedure outlined in Section2.2 of Miller et al. (2012b) to match the mass distributions,thus controlling for stellar mass (the mass distribution of thefield sample is biased toward high stellar masses, since ob-servations targeted objects with X-ray detections). In brief,we represent the field and Virgo M (cid:63) distributions as a sum ofGaussian functions, then use a weighting function (equivalentto the ratio of the field to Virgo Gaussian representations) todraw subsamples from the Virgo sample that have the samemass distribution as the field sample. Figure 7 shows the two M (cid:63) distributions and their Gaussian representations, alongwith the weighting function that is used to draw subsamplesfrom the Virgo cluster sample. We draw 500 such subsamplesof 23 Virgo galaxies each, and find that they contain, on av-erage, 7.16 nucleated objects, corresponding to +17% − ofgalaxies (error given at 1 σ confidence level, Gehrels 1986).For the field, we found six out of 23 objects to be nucleated,corresponding to +11% − (errors given at the 1 σ confidencelevel). Poisson statistics shows that for an expected value ofsix nucleated galaxies, there is a 15% chance of finding eightor more nucleated objects in a sample of 23. This argues forno statistically significant difference in the nucleation frac-tions of the field and Virgo samples.It is important to note that our fitting procedure differs fromthat of the ACS VCS; GALFIT performs 2-D modeling, whilethe ACS VCS fits directly to the 1-D surface brightness pro-file. Additionally, the ACS VCS only fitted either a S´ersicor core-S´ersic profile to the galaxy surface brightness profile,while we sometimes include an extra component to the outerregions of the galaxy. In order to explore whether this differ-ence in method would result in differing classifications, andif so, whether it would affect the soundness of our nucleationfraction comparison, we tested our fitting procedure on eightACS VCS objects. For seven out of eight objects, the classi-fications were consistent between methods. One object, VCC2095, which was classified by the ACS VCS as nucleated, was NGC 1331F850 LP
NGC 5831F850 LP
NGC 3379F850 LP
NGC 1331model
NGC 5831model
NGC 3379model
NGC 1331residual
NGC 5831residual
NGC 3379residual F IG . 1.— HST/ACS F850LP images, GALFIT 2-D models, and GALFIT residuals for three AMUSE-Field targets. NGC 1331 (left column) was fit with adouble-S´ersic profile; NGC 5831 (center column) was fit with a single S´ersic profile; NGC 3379 was fit with a S´ersic profile and an additional outer, low-S´ersicindex component. Scaling is the same within each image/model/residual set. found not to be nucleated based on our GALFIT modeling.However, we also failed to identify a NSC when fitting to the1-D surface brightness profile, as done in Cˆot´e et al. (2006).We believe this particular discrepancy to be due to the ACSVCS supplementing the light profile fitting-based classifica-tions with by-eye classifications. To summarize, while theremay be slight inconsistencies between classification methods,we expect the effect on the nucleation fraction to be smallcompared to the 1- σ error bars, and not to affect the final con-clusion that the nucleation fraction is consistent across envi-ronment. 3.2. Nucleation and X-ray Emission
As discussed in Section 1, the X-ray luminosity threshold ofthe Chandra AMUSE surveys demands a careful assessmentof the possible contamination of the nuclear X-ray signal frombright LMXBs, as opposed to low-luminosity SMBHs. WhileMiller et al. (2012a) assumed that the same level of nucle-ation (and thus contamination) measured in Virgo applied tothe field galaxies, the ACS observations presented here en-able us to verify this assumption by directly measuring thenucleation fraction. While, in the absence of a NSC, the X-ray Luminosity Function (XLF) – and thus expected number – of LMXBs within the Chandra point spread function sim-ply scales within the enclosed stellar mass (Gallo et al. 2008,2010, and references therein), the presence of a NSC likelyimplies an enhanced LMXB contribution to the nuclear X-raysignal, as discussed below. In order to quantify this effect, fol-lowing Gallo et al. (2010) we adopt the functional shape of theXLF for LMXBs in globular clusters as estimated by Sivakoffet al. (2007) . We expect this to be a conservative estimateof the probability of contamination for NSCs, as illustrated bythe following discussion.Pooley et al. (2003) finds that the number of LMXBs in adense stellar environments scales as the stellar encounter rate, Γ , to the 0.74 power; in turn, Γ ∝ ρ r c ν , where ρ is the cen-tral density, r c is the system core radius, and ν is the cen-tral velocity dispersion. While the NSCs have higher centraldensities compared to globulars (NSCs are about an order ofmagnitude more massive than the typical Milky Way globu-lar cluster (Walcher et al. 2005), the smaller core radii andthe observed higher velocity dispersions for the NSCs sug- Number of LMXBs ( L X > . × erg / s ) ∝ . g − z ) r − . , cor M . , where r h , cor is the half light radius in pc, and M is the stellar massin units of M (cid:12) . F IG . 2.— For each AMUSE-Field object with HST coverage, we present a 1-D light profile, best fit GALFIT model profile (projected to a 1-D surface brightnessprofile), residuals, and color gradient. The semi-major axis returned by ELLIPSE is plotted on the x-axis. The top panel of each plot gives the light profile and thebest fit model profile (a single or double S´ersic profile, either with or without an outer disk component). Open circles represent the z band data, and open squaresrepresent the g band data. The best fit to the z band data is given by the solid, red line, while the best fit to the g band data is given by the blue, long dashed line.If multiple components are necessary to fit a light profile, the individual components are shown in short dashed lines. In the middle panel is the residual betweenthe data and best fit profile (dashed, blue line for g band; solid, red line for z band). The bottom panel is the (g-z) color gradient. gest that globular clusters would have higher encounter rates,and thus a larger number of LMXBs (see Walcher et al. 2005for velocity dispersions of NSCs; see Harris 1996 for veloc-ity dispersions of Milky Way globular clusters, see Capuzzo-Dolcetta & Miocchi 2008 for a comparison between the NSCand globular cluster profiles).Table 2 lists probabilities of contamination for objects thatcontain NSCs and have nuclear X-ray detections. For ref-erence, we point out that NGC 3384, which has the high-est expected number of LMXBs of all the hybrid nuclei, isknown to host a SMBH based on dynamical evidence (Geb-hardt et al. 2003). The listed probabilities have been incor-porated in an accompanying paper (Miller et al. , submitted)where the Virgo and field samples combined are used to pro- vide the first measurement of the SMBH occupation fractionin the local universe. DISCUSSION AND CONCLUSIONSDetailed morphological analysis of a sample of 100 earlytype galaxies in the Virgo cluster has shown evidence thatthe majority of galaxies with stellar mass below M (cid:12) host NSCs (e.g. Ferrarese et al. 2006a). In this paper, weexplored the possibility of an environmental dependence ofthe nucleation fraction using dual-band (F475W & F850LP)HST/ACS images of a sample of 28 field early type galaxiesout of the 103 galaxies that compose the AMUSE-Field X-ray survey (Miller et al. 2012a). After controlling for differ-ent stellar mass distributions, we found there to be no statis- F IG . 3.— Light profiles for AMUSE-Field objects with HST coverage. See Figure 2 for description. tically significant difference between the nucleation fractionfor field ( +17% − ) and Virgo ( +17% − ) early-type galax-ies (from the ACS VCS survey). Here, we discuss our resultsin the context of NSC formation theories, and compare themeasured nucleation fractions with the active fractions as es-timated from X-ray diagnostics.The mechanism behind the mass assembly and evolutionof NSCs is still uncertain, but two prevalent formation theo-ries have emerged. The dissipationless model suggests thatNSCs form from the infall of globular clusters through dy-namical friction (Tremaine et al. 1975), while the dissipative model posits that NSCs form by gas accumulating at the cen-ter of a galaxy (via mergers, i.e. Mihos & Hernquist 1994, ortransport of gas in a disk, i.e. Milosavljevi´c 2004) and form-ing stars in situ . Additionally, both processes may contributeto nucleus formation, and evidence for recurring episodes ofstar formation and/or distinct kinematics (Rossa et al. 2006;Seth et al. 2008b; Paudel et al. 2011; Seth et al. 2010) – par- ticularly in late-type galaxies – implies that, even if globularcluster infall forms the bulk of the mass in the NSC, some gasneeds to accrete to the center and form stars at later times.Simulations show that in-falling globular clusters can formNSCs with global properties that generally match those of ob-served NSCs, and can reproduce scaling relations observedbetween NSCs and their host galaxies (Capuzzo-Dolcetta &Miocchi 2008, Antonini et al. 2012, Antonini 2013, Agar-wal & Milosavljevi´c 2011, Gnedin et al. 2013). This for-mation scenario is consistent with the observed radial distri-bution of the number of globular clusters near galaxy cen-ters being flatter than that of the spheroidal stellar compo-nent with the same age and metallicity (Capuzzo-Dolcetta &Mastrobuono-Battisti 2009), indicative of inner depletion ofglobular clusters via dynamical processes. There is also evi-dence, particularly in late-type galaxies, for more recent gasaccretion, from observations of multiple stellar populations(e.g. Walcher et al. 2006, Rossa et al. 2006) and kinematic F IG . 4.— Light profiles for AMUSE-Field objects with HST coverage. See Figure 2 for description. features that cannot be reproduced solely by globular clusterinfall (Hartmann et al. 2011). Overall, while globular clusterinfall is able to reproduce the bulk properties of NSCs (massesand radii), subsequent injection of gas seems to be necessaryto explain the full spectrum of observed properties.Further information on the nature of nucleation comes fromcomparing different environments. For example, Turner et al.(2012) explored the nucleation fraction and NSC propertiesbetween the Virgo and Fornax early-type cluster members(through the ACS VCS and ACS Fornax Cluster Survey,FCS), and found them to be consistent with each other. Theysuggest that this agreement in nucleation fraction between thetwo different cluster environments – with Fornax being sub-stantially smaller, colder and denser than its Northern coun-terpart – indicates that factors related to large scale environ-mental properties do not have large effects on the formationof NSCs in early-type galaxies. Although our sample size issmaller than those of both the VCS and FCS, our results con- firm and strengthen this hypothesis, as we find no statisticallysignificant difference in the nucleation fraction between thefield and the clusters’ samples. In addition, early-type galax-ies in both cluster environments as well as the field demon-strate a trend in which their nuclei move from having nuclearlight deficits to nuclear light excesses as galaxy luminositydecreases.In Figure 8, we explore whether environmental differencesare reflected in the colors of field and Virgo NSCs. As dis-cussed by Turner et al. (2012), there is a tendency for VirgoNSCs (displayed as red filled circles) to become redder withincreasing host galaxy mass/luminosity (see the caption ofFigure 8 for a quantitative analysis). According to single stel-lar population models (Kotulla et al. 2009), the extremely redcolors of the more massive Virgo NSCs (with ( g-z ) as highas 1.7) can only be achieved in less than a Hubble time forsuper-solar metallicities (see Figure 9). Thus, Turner et al.(2012) interpret these results as being suggestive of two main F IG . 5.— Light profiles for AMUSE-Field objects with HST coverage. See Figure 2 for description. channels of growth for NSCs, with low mass nuclei being pri-marily assembled via globular cluster infall, and higher massnuclei growing further through subsequent accretion, merg-ers, and/or tidal torques involving metal-enriched gas. Oursample size is too small to ascertain whether field NSCs (dis-played as blue triangles in Figure 8) show the same reddeningwith host stellar mass. If the higher mass Virgo nuclei do red-den as a result of growth through the aforementioned mecha-nisms, we might expect that NSCs in the field (where eventslike mergers and tidal torques are less common) show a moreconstant ( g-z ) color as a function of stellar mass.At the same time, highly sub-Eddington SMBH activity, asindicated by X-ray observations, changes with environment.The higher incidence of nuclear X-ray activity in the fieldsample (50% ±
7% versus 32% ± ± ∼ Gyr (Figure 9). This is comparable to the crossing time fora massive galaxy cluster, making it likely that gas could havebeen stripped between the last episode of star formation andthe present day, decreasing the amount of gas available to fun-nel to the nucleus.Complementary questions about the interplay betweenSMBHs, NSCs and globular clusters can potentially be F IG . 6.— ∆ . = log( L g /L s ) (as measured using g band data) versus absolute B-band magnitude. Positive values of ∆ . indicate there is a nuclear lightexcess, while negative values indicate a nuclear light deficit. Dashed vertical lines bound the transition region from light deficit to excess identified by Cˆot´e et al.(2007).F IG . 7.— Histograms of stellar mass distributions for the ACS VCS sample and Field sample with HST coverage. We represent these histograms as the sumof Gaussians. Also shown is the weighting function (arbitrary normalization) used to draw mass distribution matched subsamples from the Virgo Cluster; theweighting function is equivalent to the ratio of the Field to Virgo Gaussian representations. addressed by exploring their relative spatial distributionwithin clusters. E.g., Peng et al. (2008) presents evidencefor an environmental dependence of the specific frequency of globular clusters. They find that Virgo dwarf galaxieswith high specific frequencies tend to reside within 1 Mpc ofM87, implying that globular cluster formation may be biasedtoward denser environments. Although Binggeli et al. (1987)found that nucleated galaxies in Virgo are more centrallyconcentrated than non-nucleated galaxies, Cˆot´e et al. (2006)did not replicate this trend, and attribute this finding to thesurface brightness limit of Binggeli et al. (1987). However, Defined as the number of globular clusters normalized to a galaxy lumi-nosity of M V = − (Harris & van den Bergh 1981) Cˆot´e et al. (2006) used only galaxies with B T ≥ . totest this result, corresponding to 40 nucleated galaxies andjust 4 non-nucleated galaxies. Whether nucleated galaxiesare more centrally concentrated than non-nucleated galaxies,whether this aligns with the concentration of galaxies withhigher frequencies of globular clusters, and whether thesecorrelate at all with nuclear SMBH accretion, warrantsfurther investigation.V.F.B. is supported by the National Science FoundationGraduate Research Fellowship Program grant DGE 1256260.This work was supported in part by the National ScienceFoundation under grant no. NSF PHY11-25915 and byNASA through grant HST-GO-12591 from the Space Tele-0 -15 -16 -17 -18 -19 -20 . . . . . . . Absolute B mag ( g - z ) Virgo NSCs
Field NSCs F IG . 8.— (g-z) color of NSC versus host galaxy absolute B-band magnitude. Red circles correspond to Virgo Cluster galaxies; blue diamonds to Field galaxies.Shown are the results of a linear regression analysis to test the presence of a relation of the form ( g − z ) − . a + b ( M b ). A statistically significant correlationis found for the Virgo NSCs (with best-fit slope β = − . ± . , at 1 σ ), while no significant correlation is found for the field.F IG . 9.— A simple model for the evolution of the (g-z) color of a single stellar population with stellar mass M (cid:63) = M (cid:12) . Plotted for comparison are themedian colors of the Virgo and Field NSCs in galaxies with log ( M (cid:63) / M (cid:12) ) > . . scope Science Institute, which is operated by the Associationof Universities for Research in Astronomy, Inc., under NASAcontract NAS 5-26555. J.H.W. acknowledges support by theNational Research Foundation of Korea (NRF) grant fundedby the Korea government (MEST; No. 2012-006087). V.F.B.thanks Oleg Gnedin for useful suggestions.1 REFERENCESAgarwal, M., & Milosavljevi´c, M. 2011, ApJ, 729, 35Antonini, F. 2013, ApJ, 763, 62Antonini, F., Capuzzo-Dolcetta, R., Mastrobuono-Battisti, A., & Merritt, D.2012, ApJ, 750, 111Balcells, M., Graham, A. W., Dom´ınguez-Palmero, L., & Peletier, R. F.2003, ApJ, 582, L79Bell, E. F., McIntosh, D. H., Katz, N., & Weinberg, M. D. 2003, ApJS, 149,289Binggeli, B., Tammann, G. A., & Sandage, A. 1987, AJ, 94, 251B¨oker, T., Sarzi, M., McLaughlin, D. E., et al. 2004, AJ, 127, 105Capuzzo-Dolcetta, R., & Mastrobuono-Battisti, A. 2009, A&A, 507, 183Capuzzo-Dolcetta, R., & Miocchi, P. 2008, ApJ, 681, 1136Carollo, C. M., Stiavelli, M., & Mack, J. 1998, AJ, 116, 68Cˆot´e, P., Blakeslee, J. P., Ferrarese, L., et al. 2004, ApJS, 153, 223Cˆot´e, P., Piatek, S., Ferrarese, L., et al. 2006, ApJS, 165, 57Cˆot´e, P., Ferrarese, L., Jord´an, A., et al. 2007, ApJ, 671, 1456Dullo, B. T., & Graham, A. W. 2013, ApJ, 768, 36Ferrarese, L., & Merritt, D. 2000, ApJ, 539, L9Ferrarese, L., Cˆot´e, P., Bonta, E. D., et al. 2006a, ApJ, 644, L21Ferrarese, L., Cˆot´e, P., Jord´an, A., et al. 2006b, ApJ, 164, 334Gallo, E., Treu, T., Jacob, J., et al. 2008, ApJ, 680, 154Gallo, E., Treu, T., Marshall, P., et al. 2010, ApJ, 714Gebhardt, K., Bender, R., Bower, G., et al. 2000, ApJ, 539, L13Gebhardt, K., Richstone, D., Tremaine, S., et al. 2003, ApJ, 583, 92Gehrels, N. 1986, ApJ, 303, 336Gnedin, O. Y., Ostriker, J. P., & Tremaine, S. 2013, ArXiv e-printsGraham, A. W. 2012, MNRAS, 422, 1586Graham, A. W., & Guzm´an, R. 2003, AJ, 125, 2936Graham, A. W., & Spitler, L. R. 2009, MNRAS, 397, 2148Greene, J. E., & Ho, L. C. 2007, ApJ, 667, 131Greene, J. E., Peng, C. Y., Kim, M., et al. 2010, ApJ, 721, 26G¨ultekin, K., Richstone, D. O., Gebhardt, K., et al. 2009, ApJ, 698, 198H¨aring, N., & Rix, H.-W. 2004, ApJ, 604, L89Harris, W. E. 1996, AJ, 112, 1487Harris, W. E., & van den Bergh, S. 1981, AJ, 86, 1627Hartmann, M., Debattista, V. P., Seth, A., Cappellari, M., & Quinn, T. R.2011, MNRAS, 418, 2697Jiang, Y.-F., Greene, J. E., & Ho, L. C. 2011, ApJ, 737, L45Jord´an, A., Blakeslee, J. P., Cˆot´e, P., et al. 2007, ApJS, 169, 213Kelly, B. C. 2007, ApJ, 665, 1489 Kelly, B. C., & Merloni, A. 2012, Advances in Astronomy, 2012Kelly, B. C., Vestergaard, M., Fan, X., et al. 2010, ApJ, 719, 1315Kormendy, J., Bender, R., & Cornell, M. E. 2011, Nature, 469, 374Kormendy, J., & Ho, L. C. 2013, ARA&A, 51, 511Kotulla, R., Fritze, U., Weilbacher, P., & Anders, P. 2009, MNRAS, 396, 462Lauer, T. R., Tremaine, S., Richstone, D., & Faber, S. M. 2007, ApJ, 670,249Leigh, N., B¨oker, T., & Knigge, C. 2012, MNRAS, 424, 2130Leipski, C., Gallo, E., Treu, T., et al. 2012, ApJ, 744, 152Marconi, A., & Hunt, L. K. 2003, ApJ, 589, L21Matthews, L. D., Gallagher, III, J. S., Krist, J. E., et al. 1999, AJ, 118, 208McConnell, N. J., & Ma, C.-P. 2013, ApJ, 764, 184Mei, S., Blakeslee, J. P., Cˆot´e, P., et al. 2007, ApJ, 655, 144Mihos, J. C., & Hernquist, L. 1994, ApJ, 437, L47Miller, B., Gallo, E., Treu, T., & Woo, J.-H. 2012a, ApJ, 747—. 2012b, ApJ, 745LMilosavljevi´c, M. 2004, ApJ, 605, L13Neumayer, N., & Walcher, C. J. 2012, Advances in Astronomy, 2012Paturel, G., Petit, C., Prugniel, P., et al. 2003, A&A, 412, 45Paudel, S., Lisker, T., & Kuntschner, H. 2011, MNRAS, 413, 1764Peng, C. Y., Ho, L. C., Impey, C. D., & Rix, H.-W. 2002, AJ, 124, 266Peng, E. W., Jord´an, A., Cˆot´e, P., et al. 2008, ApJ, 681, 197Plotkin, R. M., Gallo, E., Miller, B. P., et al. 2014, ApJ, 780, 6Pooley, D., Lewin, W. H. G., Anderson, S. F., et al. 2003, ApJ, 591, L131Rossa, J., van der Marel, R. P., B¨oker, T., et al. 2006, AJ, 132, 1074Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525Seth, A., Ag¨ueros, M., Lee, D., & Basu-Zych, A. 2008a, ApJ, 678, 116Seth, A. C., Blum, R. D., Bastian, N., Caldwell, N., & Debattista, V. P.2008b, ApJ, 687, 997Seth, A. C., Cappellari, M., Neumayer, N., et al. 2010, ApJ, 714, 713Sivakoff, G., Jord´an, A., Sarazin, C., et al. 2007, ApJ, 660, 1246Tremaine, S., Gebhardt, K., Bender, R., et al. 2002, ApJ, 574, 740Tremaine, S. D., Ostriker, J. P., & Spitzer, Jr., L. 1975, ApJ, 196, 407Turner, M., Cote, P., Ferrarese, L., et al. 2012, ApJSWalcher, C. J., B¨oker, T., Charlot, S., et al. 2006, ApJ, 649, 692Walcher, C. J., van der Marel, R. P., McLaughlin, D., et al. 2005, ApJ, 618,237Wehner, E. H., & Harris, W. E. 2006, ApJ, 644, L17Woo, J.-H., Schulze, A., Park, D., et al. 2013, ApJ, 772, 49Woo, J.-H., Treu, T., Barth, A. J., et al. 2010, ApJ, 716, 269
APPENDIXADDITIONAL INFORMATION ON INDIVIDUAL OBJECTS
ESO 540-014
This object was misclassified as an early-type galaxy based on observations from the ground. Our observations reveal clumpsof star formation and an irregular morphology with no clear nucleus. This object may be considered a dE/dIrr object, as definedin Ferrarese et al. 2006. This object was excluded from analysis.
NGC 0855
This object is highly irregular, with many clumps of star formation extending in a bar across the galaxy. There is also a greatdeal of dust contamination. This object may be considered a dE/dIrr object, as defined in Ferrarese et al. 2006. This object wasexcluded from analysis.
NGC 1172
NGC 1172 is elliptical, with dust lanes stretching across the galaxy and diffuse dust throughout. NGC 1172 is nucleated and iswell fit by a double S´ersic profile with the addition of low ( n ∼ . ) S´ersic index component to fit the outer regions. NGC 1331
NGC 1331 is an elliptical galaxy with a nuclear star cluster. This object’s light profile is well fit by a double-S´ersic profile.The nucleus is slightly bluer than the outer regions of the galaxy. In fitting with GALFIT, the S´ersic index of the nucleus had tobe held fixed.
NGC 1340
NGC 1340 is nucleated and is well fit by a double S´ersic profile.
NGC 1370
NGC 1370 has an extremely dusty torus. Dust dominates a significant portion of the galaxy in g band, and is visible around thenucleus in z band. This object was excluded from analysis.2
TABLE 1G
ALAXY P ROPERTIES
Object Name R.A. Decl. Distance M g M ∗ ( g − z ) AB Nucleated?(deg) (deg) (Mpc) (mags) ( M (cid:12) ) (mags)NGC 4125 182.025000 65.174167 23.7 -21.17 11.2 1.37 nNGC 3585 168.321250 -26.754722 19.9 -20.94 11.1 1.32 nNGC 4036 180.361667 61.895833 24.2 -20.7 11.0 1.41 nNGC 4291 185.075833 75.370833 32.2 -20.67 10.9 1.23 nNGC 1340 52.082083 -31.068056 20.6 -20.45 10.9 1.32 yNGC 4278 185.028333 29.280833 18.5 -20.40 10.8 1.39 nNGC 5831 226.029167 1.220000 26.9 -20.25 10.6 1.23 nNGC 4697 192.149583 -5.800833 12.2 -20.11 10.6 1.37 nNGC 3115 151.308333 -7.718611 9.7 -20.07 10.7 1.47 nNGC 5582 215.179583 39.693611 28.2 -20.02 10.6 1.14 nNGC 3379 161.956667 12.581667 11.1 -19.83 10.7 1.46 nNGC 1439 56.208333 -21.920556 26.4 -19.83 10.6 1.35 nNGC 5845 226.503333 1.633889 32.7 -19.82 10.7 1.42 nNGC 1426 55.704583 -22.108333 23.3 -19.66 10.6 1.37 yNGC 4648 190.435000 74.420833 25.4 a -19.63 10.4 1.27 nNGC 3384 162.070417 12.629167 9.2 -19.24 10.4 1.37 yNGC 1172 45.400000 -14.836667 22.0 -19.09 10.3 1.30 yNGC 3377 161.926250 13.985833 10.2 -18.94 10.3 1.31 nUGC 07767 188.885000 73.674722 27.5 -18.57 10.0 1.21 nNGC 1331 51.617917 -21.355278 22.9 -18.14 9.8 1.06 yNGC 4121 181.985833 65.113889 24.8 a -18.08 9.8 1.11 nNGC 2970 145.879583 31.976944 25.9 a -17.86 9.6 0.96 yPGC 056821 240.697917 19.787222 27.0 a -17.18 9.5 1.23 nNGC 3265 157.778333 28.796667 23.0 a – – – –NGC 3073 150.217083 55.618889 33.4 – – – –NGC 1370 53.810833 -20.373611 13.2 a – – – –NGC 0855 33.514583 27.877222 12.96 – – – –ESO 540-014 10.298750 -21.131667 22.4 a – – – – Notes.
R.A. and Decl. are taken directly from the HyperLeda database. Distances are calculated from the redshift-independent mod0 distance modulus inHyperLeda ( a For objects lacking a mod0 distance modulus, the modz redshift-based distance modulus was used). M g , M (cid:63) , and color are determined asdescribed in Section 2.TABLE 2N UCLEAR S TAR C LUSTER P ROPERTIES
Host Galaxy r h log( M NSC ) ( g − z ) M NSC /M gal N( > L X ) 1- P X (pc) ( M (cid:12) ) (mag) (%)(1) (2) (3) (4) (5) (6) (7)NGC 3384 8.0 7.8 1.46 .003 0.37 30.9NGC 1340 131.8 8.9 1.32 .010 0.10 9.5NGC 1426 57.6 8.6 0.81 .010 0.42 34.3NGC 1172 26.1 8.3 1.05 .010 0.16 14.8NGC 2970 11.3 7.5 0.67 .008 0.01 1.0NGC 1331 16.2 7.3 0.69 .003 0.02 1.9 Notes.
Column 2: Half light radius. Column 3: Mass of NSC. Column 4: (g-z) color of NSC. Column 5: Fraction of host galaxy stellar mass contained in theNSC. Column 6: For objects with nuclear X-ray detections, number of expected LMXBs with log( L X ) > . within the Chandra PSF. Column 7: Forgalaxies with nuclear X-ray detections, probability that the Chandra PSF is contaminated by an LMXB of log( L X ) > . . NGC 1426
NGC 1426 is nucleated and is best fit by a double-S´ersic profile.
NGC 1439
NGC1439 has a dusty disk around its nucleus extending about .5”. The majority of the galaxy is well fit by a single S´ersicprofile, but an extra component is necessary to fit the region past 20”.
NGC 2970
NGC 2970 has a faint spiral structure, perhaps from a merger. It is the bluest object in our sample with a ( g − z ) color of 0.96.NGC 2970 has a nuclear star cluster and is well fit by a double-S´ersic profile. NGC 3073
NGC 3073 suffers from dust contamination throughout the galaxy. It is particularly problematic in g band, but the dust is alsovisible in z band. This object was excluded from analysis.3
NGC 3115
NGC 3115 is a highly edge-on S0 galaxy. The light profile was fit by a S´ersic profile plus an exponential disk.
NGC 3265
NGC 3265 appears to have very diffuse spiral arms and is highly contaminated by dust. This object was excluded from analysis.
NGC 3585
NGC 3585 appears to have a disk. The light profile was fit by a S´ersic profile plus an outer exponential disk component.
NGC 3377
NGC 3377 has an extremely depleted core, which is also very blue. Excluding the central region, we can fit the light profile ofthis galaxy with a S´ersic profile plus an outer disk component. NGC 3377 has a dust lane as well as diffuse dust contamination inthe g band. There is a steep color gradient from the nucleus to the outer regions of the galaxy, with the g − z color getting redderwith increasing radius. NGC 3379
NGC 3379 had a chip gap across the nucleus, but we were able to correct for it. There appears to be a disk in the central 2” butthe chip gap covers part of it. This galaxy is well fit by a S´ersic profile plus an outer disk-like component.
NGC 3384
NGC 3384 required a multiple component fit for the outer galaxy, and was found to have a nuclear star cluster. This galaxy hasbeen referenced in the literature as having both a NSC and SMBH (Graham & Spitler 2009).
NGC 4036
NGC 4036 suffers from significant dust contamination, including some contamination surrounding the nucleus. The lightprofile was fit by a S´ersic component plus an outer exponential disk component.
NGC 4121
NGC 4121 is well fit by a S´ersic component for most of the galaxy plus an outer exponential disk component.
NGC 4125
NGC 4125 suffers from diffuse dust contamination across most of the galaxy and across the nucleus. The light profile is wellfit by a single S´ersic component.
NGC 4278
NGC 4278 is well fit by a single S´ersic component.
NGC 4291
NGC 4291 has a depleted core, but the outer regions of the galaxy are well fit by a single S´ersic profile.
NGC 4648
The light profile of NGC 4648 is well fit by a S´ersic component for most of the galaxy plus an outer exponential disk component.
NGC 4697
NGC 4697 has a disk in the center which complicates light profile fitting. The disk extends from 0.4” to 4.0”, correspondingto a dip in the light profile visible in this region. While NGC 4697’s light profile can be fit with a double S´ersic profile, we arenot convinced it is truly nucleated as opposed to appearing to have a central light excess with respect to the nuclear disk.
NGC 5582
The light profile of NGC 5582 is well fit by a single S´ersic profile.
NGC 5831
The light profile of NGC 5831 is well fit by a single S´ersic profile.
NGC 5845
NGC 5845 has a dusty disk in the center. Its light profile is well fit by a single S´ersic profile.
PGC 056821
The light profile of PGC 056821 is well fit by a single S´ersic profile. The images of PGC 056821 taken in the F850LP filtersuffer from streaks of scattered light from a nearby star. Some of these streaks cut across the galaxy. These were corrected for bymasking them, creating a model using the IRAF ELLIPSE and BMODEL tasks, and filling in those regions with the model.