The Astrophysical Journal, 611:270–293, 2004 August 10 A # 2004. The American Astronomical Society. All rights reserved. Printed in U.S.A. INFRARED SURFACE BRIGHTNESS FLUCTUATIONS OF MAGELLANIC STAR CLUSTERS1 Rosa A. Gonza´lez Centro de Radioastronomı´a y Astrofı´sica, Universidad Nacional Autonoma de Me´xico, Campus Morelia, Michoaca´n CP 58190, Mexico; [email protected] Michael C. Liu Institute for Astronomy, University of Hawaii, 2680 Woodlawn Drive, Honolulu, HI 96822; [email protected] and Gustavo Bruzual A. Centro de Investigaciones de Astronomı´a, Apartado Postal 264, Me´rida 5101-A, Venezuela; [email protected] Received 2003 November 27; accepted 2004 April 16 ABSTRACT We present surface brightness fluctuations (SBFs) in the near-IR for 191 Magellanic star clusters available in the Second Incremental and All Sky Data releases of the Two Micron All Sky Survey (2MASS) and compare them with SBFs of Fornax Cluster galaxies and with predictions from stellar population models as well. We also construct color-magnitude diagrams (CMDs) for these clusters using the 2MASS Point Source Catalog (PSC). Our goals are twofold. The first is to provide an empirical calibration of near-IR SBFs, given that existing stellar population synthesis models are particularly discrepant in the near-IR. Second, whereas most previous SBF studies have focused on old, metal-rich populations, this is the first application to a system with such a wide range of ages (106 to more than 1010 yr, i.e., 4 orders of magnitude), at the same time that the clusters have a very narrow range of metallicities (Z 0:0006 0:01, i.e., 1 order of magnitude only). Since stellar population synthesis models predict a more complex sensitivity of SBFs to metallicity and age in the near-IR than in the optical, this analysis offers a unique way of disentangling the effects of age and metallicity. We find a satisfactory agreement between models and data. We also confirm that near-IR fluctuations and fluctuation colors are mostly driven by age in the Magellanic cluster populations and that in this respect they constitute a sequence in which the Fornax Cluster galaxies fit adequately. Fluctuations are powered by red supergiants with high-mass pre- cursors in young populations and by intermediate-mass stars populating the asymptotic giant branch in intermediate- age populations. For old populations, the trend with age of both fluctuation magnitudes and colors can be explained straightforwardly by evolution in the structure and morphology of the red giant branch. Moreover, fluctuation colors display a tendency to redden with age that can be fitted by a straight line. For the star clusters only, (H¯ À K¯ s) ¼ (0:21Æ 0:03) log (age) À (1:29 Æ 0:22); once galaxies are included, (H¯ À K¯ s) ¼ (0:20 Æ 0:02) log (age) À (1:25 Æ 0:16). Finally, we use for the first time a Poissonian approach to establish the error bars of fluctuation measurements, instead of the customary Monte Carlo simulations. Subject headings: astronomical data bases: miscellaneous — galaxies: distances and redshifts — galaxies: star clusters — Magellanic Clouds — stars: AGB and post-AGB Online material: machine-readable table 1. INTRODUCTION measurements are expressed in m¯ and M¯ , which are, respec- tively, the apparent and absolute magnitudes of L¯. While the mean surface brightness of a galaxy is inde- SBF magnitudes, however, depend on not only galaxy pendent of distance, the variance about the mean decreases distances but also the age and metallicity of stars. Therefore, with distance; i.e., given the same angular resolution, more SBFs also offer a unique possibility to investigate unresolved distant galaxies appear smoother. This is the principle behind stellar populations. For example, as a luminosity-weighted surface brightness fluctuation (SBF) measurements (Tonry & mean, M¯ is much more sensitive to giant stars than integrated Schneider 1988; Blakeslee et al. 2001), one of the most colors (Worthey 1993a; Ajhar & Tonry 1994). For the same powerful methods to determine cosmological distances (e.g., reason, M¯ is relatively insensitive to differences in the initial Tonry et al. 1997; Liu & Graham 2001; Jensen et al. 2003). SBFs arise from Poisson fluctuations in the number of stars mass function (IMF) for intermediate-age and old systems. We engaged in this work with the aim of providing an within a resolution element, and they are measured through empirical calibration of near-IR SBFs, specifically for the the observed ratio of the variance to the mean surface brightness of a galaxy, that is, the ratio (denoted L¯)ofthe study of unresolved stellar populations. The near-IR is very favorable for SBF measurements, from the point of view of second to the first moment of the stellar luminosity function, improved signal (the light of intermediate-age and old pop- scaled by the inverse of 4d 2,whered is the distance. SBF ulations is dominated by the asymptotic giant branch [AGB] and the red giant branch [RGB]), reduced dust extinction, and, 1 This research has made use of the NASA / IPAC Infrared Science Archive, which is operated by the Jet Propulsion Laboratory, California In- last but not least, the model prediction that near-IR SBFs stitute of Technology, under contract with the National Aeronautics and Space might help break the age-metallicity degeneracy (Worthey Administration. 1993b). However, there is the very important disadvantage 270 SURFACE BRIGHTNESS FLUCTUATIONS 271 that existing stellar population synthesis models are particu- photometry of 61 rich star clusters in the MC; it constitutes a larly discrepant in the near-IR spectral region (Charlot et al. smooth, one-dimensional sequence of increasing age and de- 1996; Liu et al. 2000; Blakeslee et al. 2001). The disagreement creasing metallicity. Elson & Fall (1985) assigned SWB is as high as 0.2 mag in (V À K ), compared to 0.05 mag classes to 147 more clusters using UBV photometry, assuming in (B À V ). The ill-determined contribution of AGB stars to a low and uniform reddening (EBÀV 0:1 Æ 0:1) toward and the integrated light may be the most important source of this in the clouds, which is valid in general for clusters older than a problem (Ferraro et al. 1995). Such an uncertainty is bound few times 107 yr (Charlot & Fall 2000). The Elson & Fall to compromise the calibration of M¯ . An empirical calibration (1985) classification is parameterized by s,wheres ¼ (5:75 Æ of near-IR SBFs is therefore essential. 0:26)SWB class þ (9:54 Æ 1:45). We have grouped the clus- ters in superclusters according to their s-parameter, rather than 2. OUR STRATEGY their SWB class, as shown in Table 1. Nevertheless, we have In theory, a good starting point to assess the impact of assigned ages and metallicities to the superclusters from stellar population variations on SBFs and to empirically cal- Cohen (1982), by virtue of their SWB types.2 ibrate M¯ would be to derive fluctuations for a number of In order to compare the results obtained for star clusters ¯ ¯ simple stellar populations with known distances. This is not a with those for galaxies, we use MKs and MF160W derived for a new idea. Ajhar & Tonry (1994) attempted to calibrate the sample of Fornax Cluster galaxies by, respectively, Liu et al. SBF zero point with V and I photometry of Galactic globular (2002) and Jensen et al. (2003). Likewise, we have taken the clusters. Their derived M¯ I, however, does not constrain M¯ I for J, H,andKs integrated fluxes and colors of the Fornax gal- galaxies, since the range of ages of Galactic globular clusters axies directly from the 2MASS Second Incremental Release is small, and in general their metallicities do not overlap with Extended Source Catalog (XSC), via the GATOR catalog Web those of spiral bulges and early-type galaxies. About five query page. Finally, ages and abundances have been adopted members (Harris 1996) of the inner disk subgroup of the from Kuntschner (1998). Parameter values for the Fornax Galactic globular clusters have metallicities in the right range; Cluster galaxies are all presented in Table 2. aside from the fact that ages and distances of those clusters are We had to convert M¯ F160W of the galaxies to M¯ H .From generally not very well determined yet, their number is so the photometric transformations between the Hubble Space small that their analysis will be dominated by stochastic Telescope (HST ) NICMOS Camera 2 filters and the CIT/CTIO effects (see x 4). Finally, Ajhar & Tonry (1994) found that system, published by Stephens et al. (2000) for cool giants optical data alone are inadequate to decouple the effects of age of near-solar metallicity, we get and metallicity reliably. A study of the clusters of the Magellanic Clouds (MC) mH ¼ mF160W þ(0:080 Æ 0:069)À(0:243 Æ 0:046)ðÞmJ ÀmK : would seem to constitute a better course of action from the point of view of their relevance to early-type galaxies, for two ð1Þ reasons: the clusters have very well known distances, and they span a much wider range of ages (106 to 1010 yr) than the However, the transformation coefficients for fluctuation mag- Galactic globular clusters (all with 1010 yr). The oldest MC nitudes may be different from these, since the spectrum of the clusters are as old or older than elliptical galaxies and spiral fluctuations is not the same as the spectra of the stars used to bulges. On the other hand, it is true that their metallicity is low derive the transformation. Buzzoni (1993), for transformations (Z 0:0006 0:01), but their slow chemical enrichment his- of SBF magnitudes between Johnson R, I and Cousins R, I, tory means that clusters with ages between a few million years and Blakeslee et al.