Age and mass constraints for a young massive cluster in M31 based on spectral-energy-distribution fitting
Jun Ma, Song Wang, Zhenyu Wu, Zhou Fan, Yanbin Yang, Tianmeng Zhang, Jianghua Wu
aa r X i v : . [ a s t r o - ph . C O ] J a n AJ, in press
Preprint typeset using L A TEX style emulateapj v. 11/10/09
AGE AND MASS CONSTRAINTS FOR A YOUNG MASSIVE CLUSTER IN M31 BASED ONSPECTRAL-ENERGY-DISTRIBUTION FITTING
Jun Ma,
Song Wang,
Zhenyu Wu, Zhou Fan, Yanbin Yang, Tianmeng Zhang, Jianghua Wu, Xu Zhou, Zhaoji Jiang, and Jiansheng Chen AJ, in press
ABSTRACTVDB0-B195D is a massive, blue star cluster in M31. It was observed as part of the Beijing-Arizona-Taiwan-Connecticut (BATC) Multicolor Sky Survey using 15 intermediate-band filters covering awavelength range of 3000–10,000 ˚A. Based on aperture photometry, we obtain its spectral-energy dis-tribution (SED) as defined by the 15 BATC filters. We apply previously established relations betweenthe BATC intermediate-band and the Johnson-Cousins
U BV RI broad-band systems to convert ourBATC photometry to the standard system. A detailed comparison shows that our newly derived
V RI magnitudes are fully consistent with previous results, while our new B magnitude agrees to within 2 σ .In addition, we determine the cluster’s age and mass by comparing its SED (from 3000 to 20,000˚A,comprising photometric data in the 15 BATC intermediate bands, optical broad-band BV RI , and2MASS near-infrared
JHK s data) with theoretical stellar population synthesis models, resulting inage and mass determinations of 60 . ± . . − . × M ⊙ , respectively. This age andmass confirms previous suggestions that VDB0-B195D is a young massive cluster in M31. Subject headings: galaxies: individual (M31) – galaxies: star clusters – galaxies: stellar content INTRODUCTION
Young massive star clusters (YMCs) are among themain objects resulting from violent star-forming episodestriggered by galaxy collisions, mergers, and close en-counters (see de Grijs & Parmentier 2007, and referencestherein). They are also referred to as ‘young populousclusters,’ a term first coined by Hodge (1961), who usedit to describe 23 clusters containing bright, blue starsin the Large Magellanic Cloud. In Hodge (1961), the‘young’ aspect is demonstrated by the fact that all clus-ters have main sequences that extend to absolute magni-tudes brighter than M V = 0, while ‘populous’ describestheir richness (stellar membership). However, YMCs arealso observed in quiescent galaxies (Larsen & Richtler1999) and in the disks of isolated spirals, although highercluster-formation efficiencies are associated with environ-ments exhibiting high star-formation rates (see Larsen2004; Cao & Wu 2007, and references therein). It hasbecome clear that, in many ways, YMCs resemble youngversions of the old globular clusters (GCs) associatedwith all large galaxies (see Larsen et al. 2004, and ref-erences therein). YMCs are seemingly absent in theMilky Way; possibly the best example of a GalacticYMC is Westerlund 1, a heavily reddened cluster withan age and mass of 4–5 Myr (Crowther et al. 2006) and M cl ∼ M ⊙ (Clark et al. 2005), respectively.Since the pioneering work of Tinsley (1968, 1972) andSearle (1973), evolutionary population synthesis model-ing has become a powerful tool to interpret integrated National Astronomical Observatories, Chinese Academy ofSciences, Beijing 100012, P. R. China;[email protected] Key Laboratory of Optical Astronomy, National Astronomi-cal Observatories, Chinese Academy of Sciences, Beijing 100012,P. R. China Graduate University, Chinese Academy of Sciences, Beijing100039, P. R. China spectrophotometric observations of galaxies and theircomponents, such as star clusters (e.g., Anders et al.2004). The evolution of star clusters is usually modeledby means of the simple stellar population (SSP) approx-imation. An SSP is defined as a single generation ofcoeval stars formed from the same progenitor molecularcloud (thus implying a single metallicity), and governedby a given stellar initial mass function (IMF).Age and metallicity are two basic star cluster parame-ters. The most direct method to determine a cluster’sage is by employing main-sequence photometry, sincethe absolute magnitude of the main-sequence turnoff ispredominantly affected by age (see Puzia et al. 2002,and references therein). However, until recently (cf.Perina et al. 2009), this method was only applied tothe star clusters in the Milky Way and its satellites(e.g., Rich et al. 2001), although Brown et al. (2004) es-timated the age of an M31 GC using extremely deep im-ages observed with the
Hubble Space Telescope (HST) ’sAdvanced Camera for Surveys. Generally, the ages ofextragalactic star clusters are determined by compar-ing their observed spectral-energy distributions (SEDs)and/or spectroscopy with the predictions of SSP mod-els (Williams & Hodge 2001a,b; de Grijs et al. 2003a,b,c;Bik et al. 2003; Jiang et al. 2003; Beasley et al. 2004;Puzia et al. 2005; Ma et al. 2006; Fan et al. 2006;Ma et al. 2007, 2009; Caldwell et al. 2009; Wang et al.2010). Nevertheless, SSP models assume that clusterIMFs are fully populated, i.e., that clusters contain in-finite numbers of stars with a continuous distribution ofstellar masses, and that all evolutionary stages are wellsampled. Real clusters, however, contain a finite numberof stars. Therefore, a disagreement between the observedcluster colors and theoretical colors derived from SSPmodels may become apparent (see Piskunov et al. 2009;Popescu & Hanson 2010, and references therein). Otherlimitations inherent to SSP models arise from our poor Ma et al.understanding of some advanced stellar evolutionarystages, such as the supergiant and the asymptotic-giant-branch (AGB) phases (see Bruzual & Charlot 2003, andreferences therein).Located at a distance of 785 ±
25 kpc, correspond-ing to a distance modulus of ( m − M ) = 24 . ± .
07 mag (McConnachie et al. 2005), M31 is the near-est and largest spiral galaxy in the Local Groupof galaxies. It has been the subject of many GCstudies and surveys, dating back to the early studyof Hubble (1932). Based on previous publications(Hubble 1932; Seyfert & Nassau 1945; Hiltner 1958;Mayall & Eggen 1953; Kron & Mayall 1960), Vete˘snik(1962) compiled the first large M31 GC catalog, con-taining
U BV photometric data of approximately 300GC candidates. Over the past decades, several ma-jor catalogs of M31 GCs and GC candidates havebeen published, including major efforts by the Bolognagroup (Battistini et al. 1980, 1987, 1993), Barmby et al.(2000), Galleti et al. (2004, 2005, 2006, 2007), Kim et al.(2007), Caldwell et al. (2009), and Peacock et al. (2010).Following on from the first extensive spectroscopic surveyof M31 GCs by van den Bergh (1969), a significant num-ber of authors (e.g., Huchra et al. 1982; Huchra et al.1991; Dubath & Grillmair 1997; Federici et al. 1993;Jablonka et al. 1998; Barmby et al. 2000; Perrett et al.2002; Galleti et al. 2006; Lee et al. 2008, and referencestherein) have studied their spatial, kinematic, and chem-ical (metallicity) properties.M31 is known to host a large number of youngstar clusters (e.g., Fusi Pecci et al. 2005; Caldwell et al.2009; Wang et al. 2010, and references therein).Fusi Pecci et al. (2005) presented a comprehensive studyof 67 very blue star clusters, which they referred to as‘blue luminous compact clusters’ (BLCCs). Since theyare quite bright ( − . ≤ M V ≤ − . < HST ’s Wide Field and Planetary Camera-2 (WFPC2).They obtained the reddening values, ages, and metallic-ities of their sample clusters by comparing the observedcolor-magnitude diagrams (CMDs) and luminosity func-tions with theoretical models.VDB0-B195D was first detected by van den Bergh(1969). Its color is extremely blue (e.g., U − B = − . U = 14 .
66 mag; van den Bergh 1969).As a consequence, van den Bergh (1969) asserted thatVDB0-B195D is the brightest open cluster in M31. Hedetermined an integrated stellar spectral type equiva-lent to A0, which implies that the cluster contains mas-sive stars. In addition, VDB0-B195D is particularly ex-tended and most previous photometric studies did notinclude the full extent of the object’s light distribution(see for details Perina et al. 2009). We will provide anoverview of previous studies that included the clusterin § BV RI andnear-infrared
JHK s filters from the Two Micron All Sky Survey (2MASS) taken from Perina et al. (2009), we ob-tained the SED of VDB0-B195D in 22 filters, coveringthe wavelengh range from 3000 to 20,000 ˚A.In this paper, we describe the details of the observa-tions and our approach to the data reduction in §
2. In §
3, we determine the age and mass of VDB0-B195D bycomparing observational SEDs with population synthesismodels. We discuss the implications of our results andprovide a summary in § OPTICAL AND NEAR-INFRARED OBSERVATIONS OFTHE YMC VDB0-B195D
Historical overview
VDB0-B195D was first given the designation ‘0’ (i.e.,VDB0), the brightest open cluster in M31, in the catalogof van den Bergh (1969). Battistini et al. (1987) identi-fied VDB0-B195D independently and called it B195D. InBattistini et al. (1987), B195D was given a low level ofconfidence (class D) of being a genuine cluster (classesA and B were assigned very high and high levels ofconfidence, respectively). It was only recently indepen-dently confirmed to be a single object. Caldwell et al.(2009) presented a new catalog containing 670 likely starclusters, stars, possible stars, and galaxies in the fieldof M31, all with updated high-quality coordinates ac-curate to 0 . ′′ , based on images from either the LocalGroup Galaxies Survey (LGGS) (Massey 2006) or theDigitized Sky Survey (DSS). They use the designationVDB0-B195D, associated with α = 00 h m s .
43 and δ = +40 ◦ ′ ′′ . HST /WFPC2 imaging survey of youngmassive GCs in M31. They initially selected VDB0-B195D as two YMCs in M31, but their WFPC2 imagesshowed unequivocally that these two sample objects are,in fact, the same cluster. In addition, the
HST imagesclearly confirmed that VDB0-B195D is a real cluster.However, it is difficult to establish whether it is moresimilar to ordinary open clusters, similar to those in thedisk of the Milky Way, than to YMCs that may evolveto become disk GCs (see for details Perina et al. 2009).Spectral observations of VDB0-B195D were obtainedby van den Bergh (1969)—yielding classification spec-tra and the object’s radial velocity—and Perrett et al.(2002), who used them for determination of its radialvelocity and metallicity.
Archival images of the BATC Multicolor SkySurvey
Observations of the YMC VDB0-B195D were obtainedwith the BATC 60/90cm Schmidt telescope located atthe XingLong station of the National Astronomical Ob-servatory of China (NAOC). This telescope is equippedwith 15 intermediate-band filters covering the opticalwavelength range from 3000 to 10,000 ˚A. The filter sys-tem was specifically designed to avoid contamination bythe brightest and most variable night-sky emission lines.Descriptions of the BATC photometric system can befound in Fan et al. (1996). Before February 2006, a FordAerospace 2k ×
2k thick CCD camera was installed, witha pixel size of 15 µ m and a field of view of 58 ′ × ′ , yield-ing a resolution of 1 . ′′ pixel − . Since February 2006, a young massive star cluster in M31 3new E2V 4k ×
4k thinned CCD with a pixel size of 12 µ m has been in operation, featuring a resolution of 1 . ′′ pixel − . The blue quantum efficiency of the new, thinnedCCD is 92.2% at 4000 ˚A, which is much higher than forthe old, thick device (see for details Fan et al. 2009). Afield including VDB0-B195D in the a – c filters was ob-served with the thinned CCD, and in d – p bands with thethick CCD. Fig. 1 shows a finding chart of VDB0-B195Din the BATC g band (centered at 5795 ˚A), obtained withthe NAOC 60/90cm Schmidt telescope. We adopt anaperture with a radius of 15 ′′ (shown in Fig. 1) for theintegrated photometry discussed in this paper.The BATC survey team obtained 61 images of VDB0-B195D in 15 BATC filters between January 2004 andNovember 2006. Fan et al. (2009) performed the datareduction of these images, which formed part of theirM31-7 field. Table 1 contains the observation log, includ-ing the BATC filter names, the central wavelength andbandwidth of each filter, the number of images observedthrough each filter, and the total observing time per fil-ter. Multiple images through the same filter were com-bined to improve image quality (i.e., increase the signal-to-noise ratio and remove spurious signal). Intermediate-band photometry of VDB0-B195D
We determined the intermediate-band magnitudes ofVDB0-B195D on the combined images using a standardaperture photometry approach, i.e., the phot routine in daophot (Stetson 1987). Calibration of the magnitudezero level in the BATC photometric system is similarto that of the spectrophotometric AB magnitude sys-tem. For flux calibration, the Oke-Gunn (Oke & Gunn1983) primary flux standard stars HD 19445, HD 84937,BD +26 ◦ ◦ ∼ ′′ . Inspection ensured that this apertureis adequate for photometry, i.e., VDB0-B195D does notshow any obvious signal beyond this radius. In addi-tion, this aperture is nearly the same as that adoptedby Perina et al. (2009) to determine the cluster’s pho-tometry in the BV RI bands, based on the M31 imagingsurvey of Massey (2006) (see § r ≈ ′′ for integrated photome-try, i.e., r = 9 pixels for the 2k ×
2k thick CCD camera,and r = 12 pixels for the 4k ×
4k thinned CCD camera.VDB0-B195D is projected onto the disk of M31, wherethe background is bright and fluctuates, potentially asa function of distance from the cluster center. To avoidcontamination from background fluctuations, we adoptedannuli for background subtraction spanning between 10and 15 pixels for the 2k ×
2k thick CCD camera, and from13 to 20 pixels for the 4 ×
4k thinned CCD camera, bothcorresponding to ∼ ′′ . While these annuli are spa-tially as close as possible to the region dominated by cluster light (so that any differences in background fluxare minimized), they are wide enough to average out anyexpected background fluctuations. The calibrated pho-tometry of VDB0-B195D in 15 filters is summarized incolumn (6) of Table 1, in conjunction with the 1 σ magni-tude uncertainties, which include uncertainties from thecalibration errors of both the M31-1 field standard stars(see for details Fan et al. 2009; Jiang et al. 2003) and‘the secondary standard stars’ in common between theM31-1 and M31-7 fields used for calculation of the meanmagnitude offsets between the standard and instrumen-tal magnitudes (see for details Fan et al. 2009), as wellas those resulting from our daophot application. Optical broad-band and near-infrared 2MASSphotometry of VDB0-B195D
Four independent sets of photometric data exist forVDB0-B195D. van den Bergh (1969) obtained
U BV photometry using observations of the 200-inch Hale tele-scope, Battistini et al. (1987) performed
U BV R pho-tometry based on photographic plates observed with the152 cm Ritchey-Chr´etien f /8 telescope of the Univer-sity of Bologna in Loiano, King & Lupton (1991) ob-tained U BV photometry for VDB0-B195D using obser-vations with the University of Hawaii’s 2.2 m telescopeon Mauna Kea using the f /10 secondary and coronene-coated 584 ×
416 GEC CCD, and Sharov et al. (1995)performed
U BV photometry based on photo-electric ob-servations with the 2.6 m Shain telescope of the CrimeanAstrophysical Observatory. In addition, in the RevisedBologna Catalogue (RBC) of M31 GCs published byGalleti et al. (2004), the photometric data of VDB0-B195D in optical bands are based on Battistini et al.(1987) and Sharov et al. (1995), and transformed to thereference system of Barmby et al. (2000) by applying off-sets derived from objects in common between the rele-vant catalog and the data set of Barmby et al. (2000).In the RBC, VDB0-B195D was regarded as two objects.We list these photometric data in Table 2 for comparison.Note that, in the latest RBC incarnation (version 3.5, up-dated on 27 March 2008), VDB0-B195D is included as asingle object.Galleti et al. (2004) also determined 2MASS
JHK s photometric magnitudes for VDB0-B195D (transformedto the CIT photometric system; Elias et al. 1982, 1983),which we have included in Table 3. In addition,Perina et al. (2009) realized that VDB0-B195D is a par-ticularly extended object and that it is possible thatthe photometry of Sharov et al. (1995) (compiled in theRBC) was obtained with apertures that were not largeenough to include all of its flux. Therefore, they redeter-mined its photometric values in the BV RI bands basedon the M31 imaging survey of Massey (2006) using anaperture with r = 14 . ′′ , which are also listed in Table 3.From a comparison of the values in Tables 2 and 3, itis clear that the magnitudes of van den Bergh (1969)are brighter, while the results of the three other ref-erences are consistent. The magnitudes determined byPerina et al. (2009) are much brighter, however, becauseof their careful inclusion of all of the cluster’s flux. Tocompare our photometric results with previously pub-lished values, we transformed the magnitudes of VDB0-B195D in the BATC intermediate bands to broad-band U BV RI -equivalent photometry based on the relation- Ma et al.
NE VDB0-B195D
Fig. 1.—
Image of VDB0-B195D in the BATC g band, obtained with the NAOC 60/90cm Schmidt telescope. VDB0-B195D is circledusing an aperture with a radius of 15 ′′ . The field of view of the image is 11 ′ × ′ . ships obtained by Zhou et al. (2003). These are alsolisted in Table 3, and the uncertainties include thoseoriginating from the transformation based on the rela-tionships of Zhou et al. (2003) and their calibration er-rors (column 5 of their Table 3). In Fig. 2, we show theresult of the comparison. In general, the other photomet-ric data are fainter than ours and those of Perina et al.(2009). Fig. 2 and Table 3 show that our new V RI mag-nitudes agree with the results of Perina et al. (2009), andthat the B magnitude obtained in this paper is 0.32 magbrighter than that of Perina et al. (2009). Consideringthe photometric errors of both Perina et al. (2009) andour current study, these two B -band photometric resultsare consistent within 2 σ . In addition, we should keepin mind that, although the V RI magnitudes obtained inthis paper are consistent with the results of Perina et al.(2009) within 1 σ , the disagreement in B magnitudes atthis level is understandable. This is caused by the factthat the original photometry in the present paper was ob- tained in the proprietary BATC filters and transformedto the U BV RI system using transformation equations.Zhou et al. (2003) determined these conversions basedon the broad-band
U BV RI magnitudes of 48 stars fromLandolt (1983, 1992) and Galad´ı-Enr´ıquez et al. (2000)in the Landolt SA95 field, and their photometric datain the 15 BATC intermediate-band filters. In addition,the central wavelengths and bandwidths of the BATCand
U BV RI systems differ. In fact, a similar significantdisagreement of B -band photometric data for some M31GCs was reported by Wang et al. (2010), citing similararguments. STELLAR POPULATION OF VDB0-B195D
Stellar populations and synthetic photometry
To determine the age and mass of VDB0-B195D,we compared its SED with theoretical stellar popula-tion synthesis models. The SED consists of photo-metric data in the 15 BATC intermediate bands ob- young massive star cluster in M31 5
Fig. 2.—
Comparison of photometric data from different sources with new determinations in this paper for VDB0-B195D. The datapoints shown as black dots are from Perina et al. (2009). tained in this paper and optical broad-band
BV RI and 2MASS near-infrared
JHK s data from Perina et al.(2009), listed in Table 3. We used the galev SSP models (e.g., Kurth et al. 1999; Schulz et al. 2002;Anders & Fritze-v. Alvensleben 2003) for our compar-isons. The galev
SSPs are based on the Padova stellarisochrones, with the most recent versions using the up-dated Bertelli et al. (1994) isochrones (which include thethermally pulsing asymptotic giant-branch phase), and aSalpeter (1955) stellar IMF with lower- and upper-masslimits of 0.10 and between 50 and 70 M ⊙ , respectively,depending on metallicity. The full set of models spansthe wavelength range from 91˚A to 160 µ m. These mod-els cover ages from 4 × to 1 . × yr, with an ageresolution of 4 Myr for ages up to 2.35 Gyr, and 20 Myrfor greater ages. The galev SSP models include fiveinitial metallicities, Z = 0 . , . , . , .
02 (solarmetallicity), and 0.05.Since our observational data consist of integrated lumi-nosities through the set of BATC filters, we convolved the galev
SSP SEDs with the BATC intermediate-, opti-cal broad-band
BV RI , and 2MASS filter-response curvesto obtain synthetic optical and near-infrared photometryfor comparison. The synthetic i th filter magnitude canbe computed as m = − . R ν F ν ϕ i ( ν )d ν R ν ϕ i ( ν )d ν − . , (1)where F ν is the theoretical SED and ϕ i the responsecurve of the i th filter of the BATC, BV RI , and 2MASSphotometric systems. Here, F ν varies with age and metallicity. Since the observed magnitudes in the BV RI and 2MASS photometric systems are given in the Vegasystem, we transformed them to the AB system for ourfits.
Reddening and metallicity of VDB0-B195D
To obtain the intrinsic SED of VDB0-B195D, its pho-tometry must be dereddened. To date, only Perina et al.(2009) obtained reddening values for VDB0-B195D.They compared the observed CMD with theoreticalisochrones and determined E ( B − V ) = 0 . ± .
03 mag.Caldwell et al. (2009) were unable to derive the cluster’sreddening value because of the presence of a foregroundfield star, so they adopted E ( B − V ) = 0 . ± .
17 mag(external rms error), equivalent to the mean reddeningof the young clusters in M31. In this paper, we thereforeadopt the reddening value from Perina et al. (2009).In addition, cluster SEDs are affected by age andmetallicity effects. Therefore, we can only accuratelyconstrain a cluster’s age if the metallicity is known.Perina et al. (2009) found that the CMD of VDB0-B195D, based on their
HST /WFPC2 observations, isbest reproduced by the solar-metallicity models ofGirardi et al. (2002). We therefore adopt solar metal-licity for VDB0-B195D.
The ‘lowest-luminosity-limit’ test
The lowest-luminosity limit (LLL;Cervi˜no & Luridiana 2004) implies that it is mean-ingless to compare a cluster with population synthesismodels to obtain its age and mass if its integrated Ma et al.luminosity is lower than the luminosity of the mostluminous star included in the model for the relevantage. The LLL method states that clusters fainter thanthis limit cannot be analyzed using standard proceduressuch as χ minimization of the observed values withrespect to the mean SSP models (see also Barker et al.2008). Below the LLL, cluster ages and masses cannotbe obtained self-consistently. To take into account theeffects on the integrated luminosities of statisticallysampling the stellar IMF (e.g., Cervi˜no et al. 2000,2002; Cervi˜no & Luridiana 2004), we used the theoret-ical Padova isochrones at http://stev.oapd.inaf.it/cmd(CMD2.2). This interactive Web interface providesisochrones for a number of photometric systems, in-cluding optical broad-band, 2MASS, and the BATCdata used here. We obtained the solar-metallicity( Z = 0 . § § m − M ) = 24 .
47 mag (785 kpc) forM31 (McConnachie et al. 2005). The upper luminositylimit has been corrected for extinction, based on a red-dening value of E ( B − V ) = 0 .
20 mag. The interstellarextinction curve, A λ , is taken from Cardelli et al. (1989), R V = A V /E ( B − V ) = 3 . Fit results
In the previous section, the LLL test proves that theluminosity of VDB0-195D is higher than the luminosityof its brightest star expected for a given cluster age, i.e.,that using SSP models is not completely meaningless. Inaddition, the bright absolute magnitude of VDB0-195Dallows us to consider a possibility that the cluster is mas-sive enough and IMF sampling effects should not stronglyimpact the fitting results. So we will determine the clus-ter’s age and mass estimates based on direct compar-isons with SSP mean values in this section. However, weshould keep in mind that this approach is a compromise.In fact, the fitting results (Fig. 4 and Table 5) showprobable problem even for relative massive clusters.We use a χ minimization test to determine which galev SSP models are most compatible with the ob-served SEDs, χ = X i =1 [ m intr ν i − m mod ν i ( t )] σ i , (2)where m mod ν i ( t ) is the integrated magnitude in the i th filter of a theoretical SSP at age t (for solar metallicity), m intr ν i is the intrinsic, integrated magnitude, and σ i is the magnitude uncertainty, defined as σ i = σ ,i + σ ,i + ( R λ i ∗ σ red ) + σ ,i . (3)Here, σ obs ,i is the observational uncertainty from col-umn (6) of Table 1 and column (2) of Table 3, σ mod ,i is the uncertainty associated with the model itself, σ red is the uncertainty in the reddening value, and R λ i = A λ i /E ( B − V ), where A λ i is taken from Cardelli et al.(1989), R V = A V /E ( B − V ) = 3 .
1, and σ md ,i is theuncertainty in the distance modulus, for the i th filter.Charlot et al. (1996) estimated the uncertainty associ-ated with the term σ mod ,i by comparing the colors ob-tained from different stellar evolutionary tracks and spec-tral libraries. Following Ma et al. (2007, 2009), we adopt σ mod ,i = 0 .
05 mag.Perina et al. (2009) pointed out that VDB0-B195D isa particularly extended object and that the photometricmeasurements of van den Bergh (1969), Battistini et al.(1987), King & Lupton (1991), and Sharov et al. (1995)did not include all of its flux. Therefore, we adopt thephotometry of Perina et al. (2009) to fit the observedSED with theoretical SSPs for our age determination.The fit yielding the minimum χ value ( χ (min)) wasadopted as the best fit and we adopted the correspond-ing age value, 60 . ± . . ± . k – p and IJHK s photometry; using onlythe blue part of the cluster’s SED ( B, a – e , where anyeffects caused by stochasticity may be smaller) yieldsan age of 72 . ± . χ /ν < χ (min) /ν + 1,the resulting age is within the 68.3% probability range;here, ν = 21 is the number of free parameters, i.e., thenumber of observational data points minus the numberof parameters used in the theoretical model. Therefore,the accepted age range is derived from those fits thathave χ (min) /ν < χ /ν < χ (min) /ν + 1. The bestreduced- χ —defined as χ ν (min) = χ (min) /ν —and ageare listed in Table 4. The best fit to the SED of VDB0-B195D is shown in Fig. 4, where we display the intrinsiccluster SED (symbols with error bars), as well as theintegrated SED (open circles) and spectrum of the best-fitting model. From Fig. 4, we note that the observa-tional data in the b , d , o , and p BATC filters and in the K s band do not match the best-fitting model very well(the difference is approximately 0.3 mag). Photometricuncertainties in these filters may cause some differences,although this might not be the main reason for the dis-crepancy. As we know, observational star clusters’ SEDsare affected by age, metallicity and reddening. If thereddening value and metallicity adopted in this paperare not problematic, discrepancy between our observa-tions and the best-fitting model may reflect the difficultyin achieving an appropriate (but formal) fit of an SED ofa single, real cluster by SSP models. However, as we willsee below, the reddening value adopted in this paper maybe bigger than the actual reddening of VDB0-B195D. Inaddition, the differences between the photometric dataand the model in Fig. 4 show a somewhat systematicbehavior with wavelength: in bluer passbands the clus-ter seems to be more luminous than predicted by themodel, while in redder passbands it is fainter than thecorresponding model predictions. A blue excess and red young massive star cluster in M31 7 Fig. 3.—
Lowest-luminosity limit for the filters used in this paper. The curves indicate the luminosities of the most luminous star on eachisochrone for the relevant passband. The light-gray area shows the absolute magnitudes of VDB0-B195D based on a reddening value of E ( B − V ) = 0 .
20 mag (Perina et al. 2009). We used a distance modulus of ( m − M ) = 24 .
47 mag (785 kpc) for M31 (McConnachie et al.2005) to calculate the absolute magnitudes.
Ma et al.deficiency in the observed SED with respect to the modelpredictions may indicate a shortage of red giants (RGs),which can occur when the cluster is either younger orless massive (or both) than the corresponding best-fittingmodel suggests. In other words, IMF discreteness mayplay a role: due to a relatively longer main-sequence(MS) phase and shorter RG phase, a random young clus-ter is typically bluer than predicted by SSP models. Atthe same time, we find that the reddening value adoptedaffects the fitting result greatly. In fact, the best fit to theSED of VDB0-B195D improves a great deal when adopt-ing a smaller reddening value such as E ( B − V ) = 0 . χ ν (min) = 0 .
73; the resulting age (64 . ± . . ± . E ( B − V ) = 0 . galev models include absolute magnitudes (in the Vegasystem) in 77 filters for SSPs of 10 M ⊙ , including 66 fil-ters of the HST , Johnson
U BV RI (see for details Landolt1983), Cousins RI (see for details Landolt 1983), and JHK (Bessell & Brett 1988) systems. The difference be-tween the intrinsic absolute magnitudes and those givenby the model provides a direct measurement of the clus-ter mass, in units of 10 M ⊙ . However, we should keepin mind that this is only correct for cluster masses above10 M ⊙ . We estimated the mass of VDB0-B195D usingmagnitudes in all of the BV RI and
JHK s bands. There-fore, we transformed the 2MASS JHK s magnitudes tothe photometric system of Bessell & Brett (1988) usingthe equations given by Carpenter (2001). The resultingmass determinations for VDB0-B195D are listed in Ta-ble 5 with their 1 σ uncertainties including contributionsfrom uncertainties in extinction and distance modulus.From Table 5, we see that the mass of VDB0-B195D ob-tained based on the magnitudes in different filters is verydifferent. (The highest mass obtained, based on the B -band magnitude, is 0 . × M ⊙ more massive thanthat obtained using the K s magnitude.) In addition, themass estimates differ systematically with filters. Pro-vided that VDB0-B195D is massive enough to be fittedby SSP models, a systematic trend of masses based ondifferent passbands may indicate a problem with redden-ing value adopted for the cluster. If the actual reddeningis smaller than the adopted value, the actual luminositywould be overestimated. This effect is small in redderfilters but strong in bluer filters. As discussed in age es-timation, a smaller reddening value can improve the fit-ting result greatly. In fact, a smaller reddening value canreduce the mass discrepancies based on the magnitudesin different filters. When we adopted E ( B − V ) = 0 . σ . We listthese estimates in Table 7. From Table 5, we know thatthe mass of VDB0-B195D obtained in paper is between(1 . − . × M ⊙ when the reddening value is adoptedto be ( B − V ) = 0 . SUMMARY AND DISCUSSION
VDB0-B195D was previously shown to be a massivecluster based on
HST /WFPC2 observations. Its color isextremely blue and it is very bright, particularly in bluebands. In addition, VDB0-B195D is an extended object,and most previous photometric measurements did notinclude its full flux distribution (see Perina et al. 2009, for details).In this paper, we obtained the cluster’s SED in the15 BATC intermediate-band filters. We subsequentlydetermined its age and mass by comparing our multi-color photometry with theoretical stellar population syn-thesis models. Our multicolor photometric data con-sist of 15 intermediate-band filters obtained in this pa-per, and broad-band
BV RI and 2MASS
JHK s fromPerina et al. (2009), covering a wavelength range from3000 to 20,000 ˚A. Our results show that VDB0-B195D isa genuine YMC in M31.To understand the real nature of the BLCCs,Perina et al. (2009, 2010) performed an HST imag-ing survey of 20 BLCCs in M31’s disk. As a testcase, Perina et al. (2009) presented details of the data-reduction pipeline that will be applied to all survey dataand describe its application to VDB0-B195D. They es-timated the object’s age, by comparison of the observedCMD with theoretical isochrones from Girardi et al.(2002), at ≃
25 Myr. In addition, they constrained re-alistic upper and lower limits to the cluster’s age, in-dependent of the adopted metallicity, within the rela-tively narrow range from 12 to 63 Myr. Using Maraston’sSSP models of solar metallicity (Maraston 1998, 2005),Salpeter (1955) and Kroupa (2001) IMFs, and photo-metric values in the V and 2MASS J , H , and K s bands,Perina et al. (2009) concluded that the mass of VDB0-B195D is > . × M ⊙ , with their best estimates inthe range ≃ (4 − × M ⊙ .Caldwell et al. (2009) presented an updated catalog of1300 objects in M31, including spectroscopic and imag-ing surveys, based on images from either the LGGS orthe DSS and spectra taken with the Hectospec fiber po-sitioner and spectrograph on the 6.5 m MMT. They de-rived ages and reddening values for 140 young clustersby comparing their observed spectra with model spectrafrom the Starburst99 SSP suite (Leitherer et al. 1999).The results show that these clusters are less than 2 Gyrold, while most have ages between 10 and 10 yr (the ageof VDB0-B195D they derive is log age / yr = 7 . V -band photometry andmodel mass-to-light ratios (Leitherer et al. 1999) corre-sponding to the derived spectroscopic ages. This resultedin masses ranging from 2 . × to 1 . × M ⊙ . Themass of VDB0-B195D obtained by Caldwell et al. (2009)is log M cl /M ⊙ = 5 . σ . The age and mass obtained in this paperconfirms that VDB0-B195D is genuinely a YMC in M31.As we know, SSP models describe a very special case ofa continuous distribution of stellar mass (or light) alongisochrones. This is well approximated by clusters withmasses larger than 10 M ⊙ . Also, for cluster masses ofabout 10 M ⊙ , SSP models can probably still be appliedsince a systematic difference between SSP models and ob- young massive star cluster in M31 9 Fig. 4.—
Best-fitting, integrated theoretical galev
SEDs compared to the intrinsic SED of VDB0-B195D. The photometric measurementsare shown as symbols with error bars (vertical for uncertainties and horizontal for the approximate wavelength coverage of each filter).Open circles represent the calculated magnitudes of the model SED for each filter. We used a distance modulus of ( m − M ) = 24 .
47 mag(785 kpc) for M31 (McConnachie et al. 2005) to calculate the absolute magnitudes. servations should, on average, be smaller than 0.05 magfor clusters older than 10 Myr (see Fig. 3 in Piskunovet al. 2009). However, from the results of this paper,we may conclude that, probably, a formal fitting of SSPmodels to observed SEDs cannot be used without cau-tion even for relatively massive (or apparently massive)clusters, and it is highly doubtful that this approach canbe applied in a routine work providing accurate clusterparameters. The relative accuracy of 10% for age and20% found for the mass of VDB0-B195D seems to berather formal and not very confident. In addition, obser-vational star clusters’ SEDs are affected by reddening, aneffect that is also difficult to separate from the combinedeffects of age and metallicity (Calzetti 1997; Vazdekis etal. 1997; Origlia et al. 1999). Only the metallicity andreddening are derived accurately (and, ideally, indepen-dently), these degeneracies are largely (if not entirely)reduced, and ages can then also be estimated accuratelybased on a comparison of multicolor photometry span-ning a significant wavelength range (de Grijs et al. 2003b; Anders et al. 2004) with theoretical stellar populationsynthesis models. It is true that the discrepancy betweenour observations and the best-fitting model is great, andthe mass of VDB0-B195D obtained based on the magni-tudes in different filters is very different. However, whenwe adopt a smaller reddening value, the results improvegreatly. So, we conclude that the actual reddening valueof VDB0-B195D may be smaller than E ( B − V ) = 0 . REFERENCESAnders, P., & Fritze-v. Alvensleben, U. 2003, A&A, 401, 1063Anders, P., Bissantz, N., Fritze-v. Alvensleben, U., & de Grijs, R.2004, MNRAS, 347, 196Barker, S., de Grijs, R., & Cervi˜no, M. 2008, A&A, 484, 711Barmby, P., Huchra, J., Brodie, J., Forbes, D., Schroder, L., &Grillmair, C. 2000, AJ, 119, 727Battistini, P., B`onoli, F., Braccesi, A., Fusi Pecci, F., Malagnini,M. L., & Marano, B. 1980, A&AS, 42, 357Battistini, P., B`onoli F., Braccesi, A., Federici, L., Fusi Pecci, F.,Marano, B., & B¨orngen, F. 1987, A&AS, 67, 447 Battistini, P., B`onoli, F., Casavecchia, M., Ciotti, L., Federici, L.,& Fusi Pecci F. 1993, A&A, 272, 77Beasley, M. A., et al. 2004, AJ, 128, 1623Bertelli, G., Bressan, A., Chiosi, C., Fagotto, F., & Nasi, E. 1994,A&AS, 106, 275Bessell, M. S., & Brett, J. M. 1988, PASP, 100, 1134Bik, A., Lamers, H. J. G. L. M., Bastian, N., Panagia, N., &Romaniello, M. 2003, A&A, 397, 473B`onoli, F., Delpino, F., Federici, L., & Fusi Pecci, F. 1987, A&A,185, 25
Brown, T. M., et al. 2004, ApJ, 613, L125Bruzual, A. G., & Charlot, S. 2003, MNRAS, 344, 1000Caldwell, N., Harding, P., Morrison, H., Rose, J. A., Schiavon, R.,& Kriessler, J. 2009, AJ, 137, 94Calzetti, D. 1997, AJ, 113, 162Cao, C., & Wu, H. 2007, AJ, 133, 1710Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345,245Carpenter, J. M. 2001, AJ, 121, 2851Cervi˜no, M., Luridiana, V., & Castander, F. J. 2000, A&A, 360,L5Cervi˜no, M., Valls-Gabaud, D., Luridiana, V., & Mas-Hesse,J. M. 2002, A&A, 381, 51Cervi˜no, M., & Luridiana, V. 2004, A&A, 413, 145Charlot, S., Worthey, G., & Bressan, A. 1996, ApJ, 457, 625Clark, J. S., Negueruela, I., Crowther, P. A., & Goodwin, S. P.2005, A&A, 434, 949Crowther, P. A., Hadfield, L. J., Clark, J. S., Negueruela, I., &Vacca, W. D. 2006, MNRAS, 372, 1407Dubath, P., & Grillmair, C. J. 1997, A&A, 321, 379de Grijs, R., & Parmentier, G. 2007, ChJAA, 7, 155de Grijs, R., Bastian, N., & Lamers, H. J. G. L. M. 2003a,MNRAS, 340, 197de Grijs, R., Fritze-v. Alvensleben, U., Anders, P., Gallagher, J.S., Bastian, N., Taylor, V. A., & Windhorst, R. A. 2003b,MNRAS, 342, 259de Grijs, R., Anders, P., Lynds, R., Bastian, N., Lamers, H. J. G.L. M., & O’Neill, E. J. Jr. 2003c, MNRAS, 343, 1285Elias, J. H., Frogel, J. A., Matthews, K., & Neugebauer, G. 1982,AJ, 87, 1029Elias, J. H., Frogel, J. A., Hyland, A. R., & Jones, T. J. 1983, AJ,88, 1027Fan, X., et al. 1996, AJ, 112, 628Fan, Z., Ma, J., de Grijs, R., Yang Y., & Zhou X. 2006, MNRAS,371, 1648Fan, Z., Ma, J., & Zhou X. 2009, RAA, 9, 993Federici, L., Bonoli, F., Ciotti, L., Fusi Pecci, F., Marano, B.,Lipovetsky, V. A., Neizvestny, S. I., & Spassova, N. 1993, A&A,274, 87Fusi Pecci, F., Bellazzini, M., Buzzoni, A., De Simone, E.,Federici, L., & Galleti, S. 2005, AJ, 130, 554Galad´ı-Enr´ıquez, D., Trullols, E., & Jordi, C. 2000, A&AS, 146,169Galleti, S., Federici, L., Bellazzini, M., Fusi Pecci, F., & Macrina,S. 2004, A&A, 426, 917Galleti, S., Bellazzini, M., Federici, L., & Fusi Pecci, F. 2005,A&A, 436, 535Galleti, S., Federici, L., Bellazzini, M., Buzzoni, A., & Fusi Pecci,F. 2006, A&A, 456, 985Galleti, S., Bellazzini, M., Federici, L., Buzzoni, A., & Fusi Pecci,F. 2007, A&A, 471, 127Girardi, L., et al. 2002, A&A, 391, 195Hiltner, W. A. 1958, ApJ, 128, 9Hodge, P. W. 1961, ApJ, 133, 413Hubble, E. P. 1932, ApJ, 76, 44Huchra, J., Stauffer, J., & van Speybroeck, L. 1982, ApJ, 259, L57Huchra, J. P., Brodie, J. P., & Kent, S. M. 1991, ApJ, 370, 495Jablonka, P., Bica, E., Bonatto, C., Bridges, T. J., Langlois, M.,& Carter, D. 1998, A&A, 335, 867Jiang, L., Ma, J., Zhou, X., Chen, J., Wu, H., & Jiang Z. 2003,AJ, 125, 727Kim, S., et al. 2007, AJ, 134, 706King, J. R., & Lupton, R. H. 1991, ASP Conf. Ser. 13:, Theformation and evolution of star clusters, 13, 575Kron, G. E., & Mayall, N. U. 1960, AJ, 65, 581 Kroupa, P. 2001, MNRAS, 322, 231 AJ, 65, 581Kurth, O. M., Fritze-v. Alvensleben, U., & Fricke, K. J. 1999,A&AS, 138, 19Landolt, A. U. 1983, AJ, 88, 439Landolt, A. U. 1992, AJ, 104, 340Larsen, S. S., & Richtler, T. 1999, A&A, 345, 59Larsen, S. S. 2004, ASP Conf. Ser. 322: The Formation andEvolution of Massive Young Star Clusters, 322, 19Larsen, S. S., Brodie, J. P., & Hunter, D. A. 2004, AJ, 128, 2295Lee, M. G., Hwang, H. S., Kim, S. C., Park, H. S., Geisler, D.,Sarajedini, A., & Harris, W. E. 2008, ApJ, 674, 886Leitherer, C., et al. 1999, ApJS, 123, 3Ma, J., de Grijs, R., Yang, Y., Zhou, X., Chen, J., Jiang, Z., Wu,Z., & Wu, J. 2006, MNRAS, 368, 1443Ma, J., et al. 2007, ApJ, 659, 359Ma, J., et al. 2009, AJ, 137, 4884McConnachie, A. W., Irwin, M. J., Ferguson, A. M. N., Ibata, R.A., Lewis, G. F., & Tanvir, N. 2005, MNRAS, 356, 979Maraston, C. 1998, MNRAS, 300, 872Maraston, C. 2005, MNRAS, 362, 799Marigo, P., Girardi, L., Bressan, A., Groenewegen, M. A. T.,Silva,L., & Granato, G. L. 2008, A&A, 482, 883Massey, P., Olsen, K. A. G., Hodge, P. W., Strong, S, B., Jacoby,G. H., Schlingman, W., & Smith, R. C. 2006, AJ, 131, 2478Mayall, N. U., & Eggen, O. J. 1953, PASP, 65, 24Origlia, L., Goldader, J. D., Leitherer, C., Schaerer, D., & Oliva,E. 1999, ApJ, 514, 96Oke, J. B., & Gunn, J. E. 1983, ApJ, 266, 713Peacock, M., et al. 2010, MNRAS, 402, 803Perina, S., et al. 2009, A&A, 494, 933Perina, S., et al. 2010, A&A, 511, 23Perrett, K. M., Bridges, T. J., Hanes, D. A., Irwin, M. J., Brodie,J. P., Carter, D., Huchra, J. P., & Watson, F. G. 2002, AJ, 123,2490Piskunov, A. E., Kharchenko, N. V., Schilbach, E., R¨oser, S.,Scholz, R. D., & Zinnecker, H. 2009, A&A, 507L, 5Popescu, B., & Hanson, M. M. 2010, ApJ, 713L, 21Puzia, T. H., Zepf, S. E., Kissler-Patig, M., Hilker, M., Minniti,D., & Goudfrooij, P. 2002, A&A, 391, 453Puzia, T. H., Perrett, K. M., & Bridges, T. J. 2005, A&A, 434,909Racine, R. 1991, AJ, 101, 865Rich, R. M., Shara, M. M., & Zurek, D. 2001, AJ, 122, 842Salpeter, E. E. 1955, ApJ, 121, 161Schulz, J., Fritze-v. Alvensleben, U., M¨oller, C. S., & Fricke, K. J.2002, A&A, 392, 1Searle, L., Sargent, W. L. W., & Bagnuolo, W. G. 1973, ApJ,179, 427Seyfert, C. K., & Nassau, J. J. 1945, ApJ, 102, 377Sharov, A. S., Lyutyi, V. M., & Esipov, V. F. 1995, SvAL, 21, 240Stetson, P. B. 1987, PASP, 99, 191Tinsley, B. M. 1968, ApJ, 151, 547Tinsley, B. M. 1972, ApJ, 178, 319Vazdekis, A., Peletier, R. F., Beckman, J. E., & Casuso, E. 1997,ApJS, 111, 203van den Bergh, S. 1969, ApJS, 19, 45Vete˘snik, M. 1962, Bull. Astron. Inst. Czech., 13, 182Wang, S., Fan, Z., Ma, J., de Grijs, R., & Zhou, X. 2010, AJ, 139,1438Williams, B. F., & Hodge, P. W. 2001a, ApJ, 548, 190Williams, B. F., & Hodge, P. W. 2001b, ApJ, 559, 851Yan, H. J., et al. 2000, PASP, 112, 691Zheng, Z. Y., et al. 1999, AJ, 117, 2757Zhou, X., et al. 2003, A&A, 379, 361 young massive star cluster in M31 11
TABLE 1BATC photometry of the M31 YMC VDB0-B195D.
Filter Central wavelength Bandwidth Number of images Exposure time Magnitude(˚A) (˚A) (hours) a . ± . b . ± . c . ± . d . ± . e . ± . f . ± . g . ± . h . ± . i . ± . j . ± . k . ± . m . ± . n . ± . o . ± . p . ± . TABLE 2Comparison of broad-band photometry of VDB0-B195D.
Filter Mag a Mag b Mag c Mag d Mag e Mag f U . ± .
012 14 . ± .
01 15.110 15.140 B . ± .
010 15 . ± .
01 15.410 15.510 V . ± .
013 15 . ± .
01 15.190 15.280 R a van den Bergh (1969); b Battistini et al. (1987); c King & Lupton (1991), uncertainties are the medianuncertainties in the mean for all sample cluster measurements; d Sharov et al. (1995); e Photometry from Galleti et al. (2004), based onBattistini et al. (1987); f Photometry from Galleti et al. (2004), based on Sharov et al. (1995).
TABLE 3Recently determined photometry for VDB0-B195D.
Filter Mag a Mag b Mag c U . ± . B . ± .
09 14 . ± . V . ± .
05 14 . ± . R . ± .
11 14 . ± . I . ± .
11 14 . ± . J . ± .
07 13 . ± . H . ± .
12 13 . ± . K s . ± .
15 12 . ± . a Perina et al. (2009); b Galleti et al. (2004); c This paper.
TABLE 4Age estimate of VDB0-B195D based on the the galev models.
Age log (Age) χ ν (min)(Myr) [yr] (per degree of freedom)60 . ± . . ± .
05 2.2
TABLE 5Mass estimates (and uncertainties) of VDB0-B195D based on the galev models.
B V R I J H K s Mass (10 M ⊙ )1 . ± .
18 1 . ± .
13 1 . ± .
17 1 . ± .
16 1 . ± .
13 1 . ± .
16 1 . ± . TABLE 6Comparison of age and mass estimates of VDB0-B195D.
Age Age log (Age) log (Age) Mass Mass log (Mass) log (Mass) (Myr) (Myr) [yr] [yr] (10 M ⊙ ) (10 M ⊙ ) [ M ⊙ ] [ M ⊙ ]25 60 . ± . . ± .
05 4 − . − . . . − . Perina et al. (2009); This paper; Caldwell et al. (2009).
TABLE 7Mass estimates (and uncertainties) of VDB0-B195D based on the galev models with E ( B − V ) = 0 . . B V R I J H K s Mass (10 M ⊙ )1 . ± .
12 1 . ± .
10 1 . ± .
14 1 . ± .
14 1 . ± .
12 1 . ± .
16 1 ..
16 1 .. ± ..