Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/~golim/papers/meurer2003.pdf Дата изменения: Wed May 5 17:48:35 2004 Дата индексирования: Tue Oct 2 11:18:43 2012 Кодировка: Поисковые слова: galaxy cluster

The Astrophysical Journal, 599:L83­L86, 2003 December 20
2003. The American Astronomical Society. All rights reser ved. Printed in U.S.A.

THE DISCOVERY OF GLOBULAR CLUSTERS IN THE PROTOSPIRAL GALAXY NGC 2915: IMPLICATIONS FOR HIERARCHICAL GALAXY EVOLUTION Gerhardt R. Meurer,1 J. P. Blakeslee,1 M. Sirianni,1 H. C. Ford,1 G. D. Illingworth,2 N. Benitez,1 M. Clampin,3 ґ F. Menanteau,1 H. D. Tran,4 R. A. Kimble,3 G. F. Hartig,5 D. R. Ardila,1 F. Bartko,6 R. J. Bouwens,1 T. J. Broadhurst,7 R. A. Brown,5 C. J. Burrows,5 E. S. Cheng,3 N. J. G. Cross,1 P. D. Feldman,1 D. A. Golimowski,1 C. Gronwall,8 L. Infante,9 J. E. Krist,5 M. P. Lesser,10 A. R. Martel,1 G. K. Miley,11 M. Postman,5 P. Rosati,12 W. B. Sparks,5 Z. I. Tsvetanov,1 R. L. White,1,5 and W. Zheng1
Received 2003 October 7; accepted 2003 November 12; published 2003 December 4

ABSTRACT We have discovered three globular clusters beyond the Holmberg radius in Hubble Space Telescope Advanced Camera for Sur veys images of the gas-rich dark matter­dominated blue compact dwarf galaxy NGC 2915. The clusters, all of which start to resolve into stars, have MV 606 p 8.9 to 9.8 mag, significantly brighter than the peak of the luminosity function of Milky Way globular clusters. Their colors suggest a metallicity [Fe/H] 1.9 dex, typical of metal-poor Galactic globular clusters. The specific frequency of clusters is at a minimum normal, compared to spiral galaxies. However, since only a small portion of the system has been sur veyed, it is more likely that the luminosity and mass normalized cluster content is higher, like that seen in elliptical galaxies and galaxy clusters. This suggests that NGC 2915 resembles a key phase in the early hierarchical assembly of galaxies--the epoch when much of the old stellar population has formed but little of the stellar disk. Depending on the subsequent interaction histor y, such systems could go on to build up larger elliptical galaxies, evolve into normal spirals, or in rare circumstances remain suspended in their development to become systems like NGC 2915. Subject headings: galaxies: evolution -- galaxies: halos -- galaxies: individual (NGC 2915) -- galaxies: star clusters
1. INTRODUCTION

All galaxies with massive old stellar populations are thought to contain globular clusters (GCs). They are particularly noticeable where the old stellar population is dominant such as in elliptical (E), dwarf elliptical (dE), and central dominant (cD), as well as spiral galaxies with prominent bulges. Dwarf spheroidal galaxies generally do not contain GCs, probably because they have insufficient mass to make their formation likely. The exceptions are the most massive dwarf spheroidals Fornax (Hodge 1961) and Sagittarius (Ibata, Gilmore, & Ir win 1994), which each have at least four GCs. Disk galaxies dominated by Population I stars contain fewer GCs per unit luminosity, presumably because of star formation in the disk after the formation of the Population II component. Galaxies with a high MH i /L B ratio have yet to form much of their bar yonic mass into stars. They typically are blue and not considered likely hosts for populous GC systems.
Department of Physics and Astronomy, The Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218; meurer@pha.jhu.edu. 2 UCO/Lick Obser vator y, University of California at Santa Cruz, Santa Cruz, CA 95064. 3 NASA Goddard Space Flight Center, Greenbelt, MD 20771. 4 W. M. Keck Obser vator y, 65-1120 Mamalahoa Highway, Kamuela, HI 96743. 5 Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218. 6 Bartko Science and Technology, P.O. Box 670, Mead, CO 80542-0670. 7 Racah Institute of Physics, Hebrew University, Jerusalem 91904, Israel. 8 Department of Astronomy and Astrophysics, Pennsylvania State University, 525 Davey Lab, University Park, PA 16802. 9 Departmento de Astronomia y Astrofisica, Pontificia Universidad Catolica ґ ґ ґ de Chile, Casilla 306, Santiago 22, Chile. 10 Steward Obser vator y, University of Arizona, 933 North Cherr y Avenue, Tucson, AZ 85721. 11 Leiden Obser vator y, Postbus 9513, 2300 RA Leiden, Netherlands. 12 European Southern Obser vator y, Karl-Schwarzschild-Strasse 2, D-85748 Garching, Germany.
1

NGC 2915 is an extreme gas-rich galaxy having MH i /L B p 1.7 M, /L B, , (Meurer et al. 1996, hereafter M96). Its regularly rotating H i disk extends to over 5 times beyond the readily detectable optical emission providing an excellent dynamical tracer for the mass distribution; not coincidentally NGC 2915 has one of the highest known mass-to-light ratios in a single galaxy (M96). Furthermore, while its optical morphology is that of a blue compact dwarf (Meurer, Mackie, & Carignan 1994, hereafter MMC94), its H i disk clearly shows spiral arms that are not apparent in the optical. In this Letter, we report the discover y of three luminous GCs found in Hubble Space Telescope Advanced Camera for Sur veys (ACS; Ford et al. 2003) images of NGC 2915 that were obtained in order to look for a stellar heating source for the H i disk. That issue will be discussed in a separate article (G. R. Meurer et al. 2003, in preparation, hereafter M03).
2. DATA AND ANALYSIS

ACS Wide-Field Camera images were obtained of a field cens tered at R.A. p 09h25m36.48, decl. p 76 35 52.4 (J2000.0). The images cover projected radii of 45 ­257 , whereas the Holmberg radius RHo p 114 . We obtained 2, 2, and 4 images for a total exposure of 2480, 2600, and 5220 s in the filters F475W (g475), F606W (V606), and F814W (I814), respectively. The basic processing of the images was done using the CALACS pipeline (Hack 1999). We used the program Apsis (Blakeslee et al. 2002) to align and combine the images incorporating geometric correction and rejection of cosmic rays and hot pixels. Here we present photometr y in the natural system of the filters, with zero points selected so that Vega would have a magnitude of 0.0 in all bands. In order to compare our observations with previous work, we convert the previous work to this system, as needed, using the calibrations of M. Sirianni et al. (2003, in preparation). The most important correction is to L83


L84
TABLE 1 NGC 2915 Properties Quantity E(B V ) ...... D ............. Mg ............ M ........... * MT ............ MV ............ LB ............. MT /LB ........ Value 0.28 0.04 4.1 0.3 7.4 # 108 3.2 # 108 2.1 # 1010 ­16.42 3.4 # 108 62 Units mag Mpc M, M, M, mag LB, , solar Description

MEURER ET AL.

Vol. 599

Foreground extinction Distance ISM mass Mass in stars Total dynamical mass Absolute magnitude V band B-band luminosity Mass-to-light ratio

the V606 photometr y, since the F606W filter straddles the wavelength of traditional V and R filters.
3. RESULTS

Table 1 presents adopted global properties for NGC 2915. The foreground extinction, E( B V ), is from the Schlegel, Finkbeiner, & Davis (1998) extinction maps. It is significantly larger than E( B V ) p 0.15 0.05 estimated by MMC94 but consistent with the position of the field star red giant branch (RGB; M03). Extinction-corrected photometr y employing the Cardelli, Clayton, & Mathis (1989) extinction cur ve is denoted with a "0" subscript. The distance, D, was derived from the field star RGB tip (M03). It is consistent with but improves on previous estimates D p 5.3 1.3 Mpc (MMC94) and D p 3.8 0.5 Mpc (Karachentsev et al. 2003). The remaining quantities in Table 1 were derived from MMC94 and M96 after correcting to the new E( B V ) and D. As shown in Figure 1, the three sources are clearly GCs whose brightest stars are resolved. Table 2 compiles the properties of the clusters. The photometric quantities were measured using a circular aperture having a radius of r p 3 , with the local sky subtracted using an annulus having radii of 5 and 7 . 5. The cluster size r1/ 2 is the circular aperture radius encompassing half the V606 light as measured from cur ves of growth. Compared to Galactic GCs, these clusters are large and luminous but not abnormally so. Only 16% of the clusters in the Harris (1996) database13 have V606 luminosities brighter than G3; only three clusters are more luminous than G1. The clusters' r1/ 2 ranges from about 5 to 9 pc, placing them in the upper quartile of Galactic GCs that have r1/ 2 ranging from 0.3 to 24.7 pc (Harris 1996). The clusters are noticeably elongated with ellipticity e { 1 b/a similar to the canonical flattened Galactic GCs M22
13

and q Cen (e p 0.14 and 0.17, respectively; Harris 1996). The combination of high luminosity and appreciable flattening is also seen in the cluster M31-G1 (Pritchet & van den Bergh 1984; Meylan et al. 2001). The (g475 V606 ) 0 and (V606 I814 ) 0 colors of the clusters are compared to Milky Way GCs (Harris 1996) in Figure 2. Their colors are virtually identical, implying similar metallicities, assuming that they are old and nearly coeval. We derive their metallicity by fitting the metallicity-color relationship from the Harris database after converting the colors to (V606 I814 ) 0. We employed an unweighted least-squares fit with an iterative 2.5 j rejection resulting in [Fe/ H] p 5.37 5.36(V606 I814 ) 0 with a dispersion of 0.29 dex. The metallicity for the three clusters is then [Fe/ H] p 1.9 0.4 dex, consistent with low metallicity Galactic GCs. Our stellar population analysis, in progress (M03), indicates that the stars at the outskirts of the clusters have very similar I814, 0 versus (V606 I814 ) 0 color-magnitude diagrams, dominated by a narrow and blue RGB. This is also consistent with low [Fe/H], if the clusters are old.
4. DISCUSSION

4.1. Cluster Formation Efficiency Because of the blue core and gas-rich nature of NGC 2915, we had not expected to find GCs in our images. However, in retrospect this discover y should not have been surprising. GCs have previously been discovered around morphologically simЁ ilar systems; Ostlin, Bergvall, & Ronnback (1998) find a conЁ siderable population of at least 65 old GCs about the blue compact (albeit not dwarf) galaxy ESO 338-IG04. In addition, NGC 2915 is not a dwarf system in terms of mass, and optical imaging shows it to be dominated by an older stellar population for R 1 0.5 kpc (MMC94). These facts suggest that old GCs might have been expected. Is the number of clusters found anomalous? To address this we estimate the total number of clusters in the system. Unfortunately, our images sample only a small fraction of the galaxy. If the GCs are distributed spherically out to RH i p 10 (where the H i distribution ends; M96), then we have surveyed only 3.5% of the available area. Here we consider two cases for the possible distribution of old clusters. Case 1 assumes that we are lucky and have managed to obser ve all the clusters in NGC 2915. While this is unlikely, it provides a strict lower limit to the cluster formation efficiency. The more likely case 2 is that the GC system is like that of more luminous systems--having a spherically symmetric power-law radial dis-

See http://physwww.physics.mcmaster.ca/harris/mwgc.dat.

Fig. 1.--Three color images of the three GCs, from left to right: G1, G2, and G3. The images are shown on the same logarithmic stretch, with F814W, F606W, and F475W representing red, green, and blue, respectively. Each image is 8 on a side and has the same orientation. The large arrow of the compass in the middle panel points to north; the small arrow points to east.


No. 2, 2003

GLOBULAR CLUSTERS IN NGC 2915
TABLE 2 Cluster Properties ID G1 ...... G2 ...... G3 ...... R.A.a (J2000.0) 09 25 56.273 09 25 27.445 09 25 41.954 Decl.a (J2000.0) 76 35 14.53 76 36 31.55 76 36 33.70 R (kpc) 3.0 3.3 2.4 MV606 9.82 9.04 8.92 (g V606)0 (V606 I814)0 0.66 0.65 0.64 r1/2 (pc) 8.8 4.6 5.9 e 0.11 0.16 0.13

L85

475

0.62 0.57 0.59

Note.--Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. a The absolute accuracy of the positions 0 . 2 in each coordinate is set by the astrometric calibration that employed 86 stars in the Hubble Space Telescope Guide Star Catalog version 2 (http://www-gsss.stsci.edu/ gsc/gsc2/GSC2home.htm). The relative accuracy of the positions is less than 0 . 05 in each coordinate.

tribution in number per unit area--N( R) R a (Harris 1991). We assume that this extends between R p 0.5 Kpc and RH i, where the minimum R insures that the predictions are finite and was chosen to correspond to the size of the central starforming population (MMC94). For 2.5 ! a ! 0.5, we estimate that the total number of clusters in the system is 7­12 times higher than found on our images. For case 2 we adopt a correction factor of 9. Hence there are at least three GCs in the system (case 1), with a more likely number being 27 (case 2). In principle, we should correct the total GC estimate for the finite luminosity sampling of the images. However, SExtractor (Bertin & Arnouts 1996) catalogs of our images show that we can detect extended objects down to V606 27, which corresponds to an MV 606 2 for sources in NGC 2915. This is well below the peak of the GC luminosity function MV 7.5 (Secker 1992; revised to agree with the Hipparcos RR Lyrae zero point, e.g., Carretta et al. 2000). Hence we apply no luminosity sampling correction. The fact that we have found only GCs much brighter than this peak is somewhat puzzling.

We consider three measures of cluster formation efficiency. First, the specific frequency, SN p NT10 0.4(MV 15), is the V luminosity normalized cluster content (Harris & van den Bergh 1981). SN is typically around 1 for spiral galaxies and increases toward earlier galaxy types, with E galaxies having SN 4. In the center of galaxy clusters SN ranges from 2.5 to 12.5 (Blakeslee, Tonr y, & Metzger 1997; Blakeslee 1999). We consider two measures of the mass normalized contribution: hGC, the number of clusters per 109 M, in dynamical mass (Blakeslee et al. 1997), and e cl, the fractional bar yonic (gas and stars) mass in GCs (McLaughlin 1999). As done by McLaughlin (1999), we assume that each cluster has an average mass AM cl S p 2.4 # 10 5 M, when calculating e cl. Blakeslee et al. (1997) find an average hGC p 0.7 with a scatter of 30%, while McLaughlin (1999) computes an average e cl p 0.0026 with a 20% uncertainty. Table 3 tabulates SN, hGC, and e cl for NGC 2915 under the above two cases. Relevant quantities used for these calculations are given in Table 1. Table 3 also lists literature values and uncertainties of the efficiencies for "normal" systems. The uncertainty of our estimates is large: for a Poissonian distribution yielding a count of 3, the 95% confidence limits on the cluster formation efficiencies are 0.3 and 2.3 times the estimated value. We find that at a minimum (case 1) SN is close to normal for a spiral galaxy, while hGC and e cl are low compared to normal GC systems. In the more likely case 2, SN is ver y high compared to normal gas-rich systems. hGC and e cl are also somewhat high compared to literature values but in reasonable agreement considering the Poissonian uncertainty of our measurements. While we cannot rule out the possibility that NGC 2915 is like a normal spiral galaxy in terms of SN, it is more likely that it has a high luminosity normalized cluster content, whereas the mass normalized content is closer to normal. 4.2. A Missing Link in Galaxy Evolution? It is interesting to interpret these results within the framework of hierarchical evolution through comparison with other systems. Blakeslee (1999) has discussed how the SN of the young Milky Way must have been fairly high, similar to the values for cluster E galaxies, after the formation of the Galactic halo but before stellar disk formation. The increase in luminosity from later star formation in the disk would cause a decrease

Fig. 2.--Color-color plots of the NGC 2915 GCs ( pink filled diamonds) compared to Galactic GCs ( plus signs; from Harris 1996) and "simple stellar population" models from Bruzual & Charlot (2003; colored connected dots). All data have been corrected for Galactic extinction. The red line segment shows the E(B V ) p 0.1 reddening vector (Cardelli et al. 1999). Population models having ages of 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1.0, 2.0, 5.0, 10, 15, and 20 Gyr are plotted in four constant metallicity model sequences having [Fe/H] p 2.25 (black), 1.65 ( purple), 0.33 (green), and 0.09 (orange). All models employ the standard stellar evolution and initial mass function prescriptions from Bruzual & Charlot (2003).

TABLE 3 GC System Properties Quantity SN ....... hGC ...... ecl ....... Case 1 0.81 0.14 0.00067 Case 2 7.3 1.3 0.0061 Literature 0­12.5 0.7 0.2 0.0026 0.0005


L86

MEURER ET AL.

Vol. 599

in the Galactic SN to its present low value, while leaving the number of GCs per unit mass unchanged. In NGC 2915 we have an example of a present-day galaxy with an old stellar component, including GCs, but a spiral disk that is still mainly in the form of gas. In this sense, NGC 2915, like the centers of rich galaxy clusters, has an elevated value of SN compared to typical spirals because of a lower efficiency for converting gas into stars. Clusters typically have 2­5 times as much mass in gas as in stars (Arnaud et al. 1992), similar to what we see in NGC 2915. The difference, however, is that the gas in clusters now resides in the hot intracluster medium, while the gas in NGC 2915 still retains the potential to be converted into stars. To a large extent, the fate of the gaseous disk (and the future evolution of SN in NGC 2915) must depend on the surrounding environment. Simulations show that when a disk galaxy enters the environment of a rich galaxy cluster, much of the gas in its disk and halo will be removed by the combination of tidal and ram pressure stripping (Abadi, Moore, & Bower 1999; Bekki, Couch, & Shioya 2002; Bekki et al. 2001; Gnedin 2003). For example, if a system similar to the Galaxy were to fall into a rich cluster, it would have its gas disk truncated down to a radius of a few kiloparsecs within a few tens of megayears by ram pressure stripping alone (Abadi et al. 1999). Indeed, models of present-day evolution of galaxy clusters invoke mildly truncated star formation in field galaxies accreting onto the cluster as the cause of the Butcher-Oemler effect (Balogh, Navarro, & Morris 2000; Kodama & Bower 2001). It is likely that in their early evolution clusters were assembled from building blocks similar to NGC 2915 that had their interstellar matter stripped from them by these mechanisms. The stripped gas was then virialized to become the cluster X-ray halo. In this scenario galaxy clusters have high SN because they formed out of subclumps that had already efficiently formed star clusters but that never had the chance to form disks. In a less hostile environment, a building block similar to NGC 2915 could go on to form a normal spiral galaxy. If the

H i disk of NGC 2915 were to be perturbed enough to efficiently form stars in a fairly quiescent fashion, this would result in an additional 7 # 10 8 M, of stars forming with no additional GCs. Assuming a M/L V 1 M, /L V, , for the additional stars, then the system would evolve toward SN 2, a fairly normal value for spiral galaxies, while hGC and e cl remain fixed at their normal values. The SN in NGC 2915 is then anomalously high because the formation of its disk has not proceeded, presumably because of a lack of external perturbations (M96). If so, then we may expect other galaxies with high MH i /L B ratios and extended H i disks to also have a significant GC population, especially if they have a strong old population. While this scenario seems compelling, we caution that it is possible that NGC 2915 only superficially resembles the building block that we describe. We have not proven that the number of clusters is anomalously high. Furthermore, we cannot yet rule out the possibility that the clusters are of higher metallicity and younger than Galactic GCs. This is shown in Figure 2, where we overplot Bruzual & Charlot (2003) population models on the two-color diagram illustrating the strong agemetallicity degeneracy for the filters used here. If the clusters are not old, they may represent the remnants of a starburst occurring as recently as a few gigayears ago. We are undertaking additional obser vations (imaging and spectroscopic) to get a more accurate census of star clusters in the NGC 2915 system and determine their nature. The results of these studies should determine whether NGC 2915 is just a gas-rich galaxy with a peculiar star formation histor y or whether it truly presents us with a rare local view of how galaxies looked in the epoch of cluster assembly. ACS was developed under NASA contract NAS 5-32864, and this research has been supported by NASA grant NAG57697. We thank the technical and administrative support staff of the ACS science team for their committed work on the ACS project.

REFERENCES Abadi, M. G., Moore, B., & Bower, R. G. 1999, MNRAS, 308, 947 Arnaud, M., Rothenflug, R., Boulade, O., Vigroux, L., & Vangioni-Flam, E. 1992, A&A, 254, 49 Balogh, M. L., Navarro, J. F., & Moore, S. L. 2000, ApJ, 540, 113 Bekki, K., Couch, W. J., Drinkwater, M. J., & Gregg, M. D. 2001, ApJ, 557, L39 Bekki, K., Couch, W. J., & Shioya, Y. 2002, ApJ, 577, 651 Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393 Blakeslee, J. P. 1999, AJ, 118, 1506 Blakeslee, J. P., Anderson, K. R., Meurer, G. R., Benitez, N., & Magee, D. 2002, in ASP Conf. Ser. 295, Astronomical Data Analysis Software and Systems XII, ed. H. E. Payne, R. I. Jedrzejewski, & R. N. Hooks (San Francisco: ASP), 257 Blakeslee, J. P., Tonr y, J. L., & Metzger, M. R. 1997, AJ, 114, 482 Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000 Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 Carretta, E., Gratton, R. G., Clementini, G., & Fusi Pecci, F. 2000, ApJ, 533, 215 Ford, H. C., et al. 2003, Proc. SPIE, 4854, 81 Gnedin, O. Y. 2003, ApJ, 589, 752 Hack, W. J. 1999, CALACS Operation and Implementation, ISR ACS-99-03 (Baltimore: STScI) Harris, W. E. 1991, ARA&A, 29, 543 ------. 1996, AJ, 112, 1487 Harris, W. E., & van den Bergh, S. 1981, AJ, 86, 1627 Hodge, P. W. 1961, AJ, 66, 83 Ibata, R., Gilmore, G., & Ir win, M. 1994, Nature, 370, 194 Karachentsev, I. D., et al. 2003, A&A, 398, 479 Kodama, T., & Bower, R.G. 2001, MNRAS, 321, 18 McLaughlin, D. E. 1999, AJ, 117, 2398 Meurer, G. R., Carignan, C., Beaulieu, S., & Freeman, K. C. 1996, AJ, 111, 1551 (M96) Meurer, G. R., Mackie, G., & Carignan, C. 1994, AJ, 107, 2021 (MMC94) Meylan, G., Sarajedeni, A., Jablonka, P., Djorgovski, S. G., Bridges, T., & Rich, R. M. 2001, AJ, 122, 830 Ё Ostlin, G., Bergvall, N., & Ronnback, J. 1998, A&A, 335, 85 Ё Pritchet, C. J., & van den Bergh, S. 1984, PASP, 96, 804 Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525 Secker, J. 1992, AJ, 104, 1472