Arunav Kundu

A black hole in a globular cluster

Globular star clusters contain thousands to millions of old stars packed within a region only tens of light years across. Their high stellar densities make it very probable that their member stars will interact or collide. There has accordingly been considerable debate about whether black holes should exist in these star clusters. Some theoretical work suggests that dynamical processes in the densest inner regions of globular clusters may lead to the formation of black holes of ∼1,000 solar masses. Other numerical simulations instead predict that stellar interactions will eject most or all of the black holes that form in globular clusters. Here we report the X-ray signature of an accreting black hole in a globular cluster associated with the giant elliptical galaxy NGC 4472 (in the Virgo cluster). This object has an X-ray luminosity of about 4 × 1039 erg s-1, which rules out any object other than a black hole in such an old stellar population. The X-ray luminosity varies by a factor of seven in a few hours, which excludes the possibility that the object is several neutron stars superposed.

Low-Mass X-Ray Binaries and Globular Clusters in Early-Type Galaxies

A high fraction of the low-mass X-ray binaries (LMXBs) in early-type galaxies are associated with globular clusters (GCs). Here we discuss the correlations between LMXBs and GCs in a sample of four early-type galaxies with X-ray source lists determined from Chandra observations. There is some evidence that the fraction of LMXBs associated with GCs (fX-GC) increases along the Hubble sequence from spiral bulges (or spheroids) to S0s to Es to cDs. On the other hand, the fraction of GCs that contain X-ray sources appears to be roughly constant at fGC-X ~ 4%. There is a strong tendency for the X-ray sources to be associated with the optically more luminous GCs. However, this correlation is consistent with a constant probability of finding an LMXB per unit optical luminosity; that is, it seems to result primarily from the larger number of stars in optically luminous GCs. The probability of finding a bright LMXB per unit optical luminosity in the GCs is about 1.5 ? 10-7 LMXBs per L?,I for LX 1 ? 1038 ergs s-1 (0.3-10 keV) and rises to about 2.0 ? 10-7 LMXBs per L?,I at lower X-ray luminosities, LX 3 ? 1037 ergs s-1. This frequency appears to be roughly constant for different galaxies, including the bulges of the Milky Way and M31. There is a tendency for the X-ray sources to be found preferentially in redder GCs, which is independent of optical luminosity correlation. This seems to indicate that the evolution of X-ray binaries in a GC is affected by either the metallicity or the age of the GC, with younger and/or more metal rich GCs having more LMXBs. There is no strong difference in the X-ray luminosities of GC and non-GC LMXBs. There is a weak tendency for the brightest LMXBs, whose luminosities exceed the Eddington luminosity for a 1.4 M? neutron star, to avoid GCs. That may indicate that black hole X-ray binaries are somewhat less likely to be found in GCs, as seems to be true in our Galaxy. On the other hand, there are some luminous LMXBs associated with GCs. There is no clear evidence that the X-ray spectra or variability of GC and non-GC X-ray sources differ. We also find no evidence for a difference in the spatial distribution of GC and non-GC LMXBs. Many of these results are similar to those found in NGC 1399 and NGC 4472 by Angelini et al. and Kundu et al., respectively.

The low-mass X-ray binary-globular cluster connection in NGC 4472

We have analyzed the low-mass X-ray binary (LMXB) candidates in a Chandra observation of the giant elliptical galaxy NGC 4472. In a region observed by the Hubble Space Telescope (HST), approximately 40% of the bright (LX>~1037 ergs s-1) LMXBs are associated with optically identified globular clusters (GC). This is significantly higher than the fraction of bright LMXBs in Galactic GCs and confirms that GCs are the dominant sites of LMXB formation in early-type galaxies. The ~4% of NGC 4472 GCs hosting bright LMXBs, on the other hand, is remarkably similar to the fraction of GCs with LMXBs in every other galaxy. Although statistical tests suggest that the luminosity of a cluster is an important driver of LMXB formation in GCs, this appears largely to be a consequence of the greater number of stars in bright clusters. The metallicity of GCs is a strong determinant of LMXB specific frequency, with metal-rich clusters about 3 times more likely to host LMXBs than metal-poor ones. There are weaker dependences on the size of a GC and its distance from the center of the galaxy. The X-ray luminosity does not depend significantly on the properties of the host GC. Furthermore, the spatial distribution and X-ray luminosity function of LMXBs within and outside GCs are indistinguishable. The X-ray luminosity function of both GC-LMXBs and non-GC-LMXBs reveal a break at ~3×1038 ergs s-1, strongly suggesting that the brightest LMXBs are black hole accretors. Based on observations made with the NASA/ESA Hubble Space Telescope, obtained at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-26555, and on observations made with the Chandra X-Ray Observatory.

Wide Field Planetary Camera 2 Imaging of the Globular Cluster System of the S0 Galaxy NGC 3115

Using the Wide Field Planetary Camera 2 on board the Hubble Space Telescope we have identified 144 globular cluster candidates in the inner region of the S0 galaxy NGC 3115. The color distribution of the globular clusters is bimodal, with a blue peak at V-I = 0.96 mag and a red peak at V-I = 1.17 mag. There are roughly equal number of blue and red clusters. The mean [Fe/H] of the blue population is -1.36 dex, while the red population has a mean [Fe/H] of -0.37 dex. Following Ashman & Zepf, we interpret the bimodality as evidence of multiple epochs of cluster formation in the metal enrichment history of the galaxy. Assuming that the blue peak corresponds to a primordial, 15 Gyr old, metal-poor population of globular clusters, we derive an age of 11 ± 3 Gyr for the younger, red clusters. We also find that the age and metallicities of the globular clusters are consistent with the red giant stars in the field, suggesting that the two systems formed coevally. The spatial distribution of both the red and the blue clusters are flattened with major and minor axes roughly colinear with the galaxy light profile. While the red clusters follow the thick disk component of the galaxy, the blue cluster distribution is less flattened and mimics the bulge/halo light distribution. We find a local specific frequency of 1.3 ± 0.1 in the central region of the galaxy, which is lower than the global value of 2.3 ± 0.5 reported in the literature, implying that the specific frequency increases with galactocentric distance. We have also observed the globular cluster luminosity function 2 mag deeper than the turnover luminosity and derived a distance modulus of m-M = 29.8 ± 0.3 mag to NGC 3115. The average half-light radii of the clusters is ≈2 pc, and the red clusters are ≈20% smaller than the blue ones.

THE FUELING OF NUCLEAR ACTIVITY. I. A NEAR-INFRARED IMAGING SURVEY OF SEYFERT AND NORMAL GALAXIES

We present near-infrared, modified K-band images of a sample of Seyfert and control galaxies. Our sample includes all known Seyfert galaxies in the Revised Shapely-Ames Catalog and Sandage & Tammanns 1987 extension to the RSA with recessional velocities less than 5000 km s-1 and logarithmic axial ratios less than 0.2 (excluding the two early-type Seyfert galaxies Centaurus A and Perseus A). A control sample of normal galaxies, matched to the Seyfert sample in Hubble type, redshift, inclination, and blue luminosity, has also been observed. To quantify the incidence of bars in both samples, elliptical fits to the isophotes of each galaxy have been performed. In agreement with earlier studies, we find that many galaxies classified as unbarred in the optical, display evidence for bars in the near-infrared.

From globular clusters to tidal dwarfs: Structure formation in the tidal tails of merging galaxies

Using V and I images obtained with the Wide Field Planetary Camera 2 (WFPC2) of the Hubble Space Telescope, we investigate compact stellar structures within tidal tails. Six regions of tidal debris in the four classic Toomre sequence mergers: NGC 4038/39 (Antennae), NGC 3256, NGC 3921, and NGC 7252 (Atoms for Peace) have been studied in order to explore how the star formation depends on the local and global physical conditions. These mergers sample a range of stages in the evolutionary sequence and tails with and without embedded tidal dwarf galaxies. The six tails are found to contain a variety of stellar structures, with sizes ranging from those of globular clusters up to those of dwarf galaxies. From V and I WFPC2 images, we measure the luminosities and colors of the star clusters. NGC 3256 is found to have a large population of blue clusters (0.2 V-I 0.9), particularly in its western tail, similar to those found in the inner region of the merger. In contrast, NGC 4038/39 has no clusters in the observed region of the tail, only less luminous point sources likely to be individual stars. NGC 3921 and NGC 7252 have small populations of clusters along their tails. A significant cluster population is clearly associated with the prominent tidal dwarf candidates in the eastern and western tails of NGC 7252. The cluster-rich western tail of NGC 3256 is not distinguished from the others by its dynamical age or by its total H I mass. However, the mergers that have few clusters in the tail all have tidal dwarf galaxies, while NGC 3256 does not have prominent tidal dwarfs. We speculate that star formation in tidal tails may manifest itself either in small structures like clusters along the tail or in large structures such as dwarf galaxies, but not in both. Also, NGC 3256 has the highest star formation rate of the four mergers studied, which may contribute to the high number of star clusters in its tidal tails.

The M31 globular cluster system: ugriz and K-band photometry and structural parameters

We present an updated catalogue of M31 globular clusters (GCs) based on images from the Wide Field Camera (WFCAM) on the United Kingdom Infrared Telescope and from the Sloan Digital Sky Survey (SDSS). Our catalogue includes new, self-consistent ugriz and K-band photometry of these clusters. We discuss the difficulty of obtaining accurate photometry of clusters projected against M31 due to small-scale background structure in the galaxy. We consider the effect of this on the accuracy of our photometry and provide realistic photometric error estimates. We investigate possible contamination in the current M31 GC catalogues using the excellent spatial resolution of these WFCAM images combined with the SDSS multicolour photometry. We identify a large population of clusters with very blue colours. Most of these have recently been proposed by other works as young clusters. We distinguish between these, and old clusters, in the final classifications. Our final catalogue includes 416 old clusters, 156 young clusters and 373 candidate clusters. We also investigate the structure of M31s old GCs using previously published King model fits to these WFCAM images. We demonstrate that the structure and colours of M31s old GC system are similar to those of the Milky Way. One GC (B383) is found to be significantly brighter in previous observations than observed here. We investigate all of the previous photometry of this GC and suggest that this variability appears to be genuine and short lived. We propose that the large increase in its luminosity may have been due to a classical nova in the GC at the time of the previous observations in 1989.

Probing the Formation of Low-Mass X-Ray Binaries in Globular Clusters and the Field*

We present an analysis of LMXBs and GCs in five early-type galaxies using Chandra X-ray and HST optical data. Of the 186 LMXBs within the optical fields of view, 71 are in GCs, confirming that LMXBs are formed particularly efficiently in clusters due to dynamical interactions. However, there is no statistically significant correlation between the distance of a cluster from the center of its host galaxy and its LMXB hosting probability. LMXBs are preferentially found in luminous and metal-rich GCs. Metal-rich clusters are 3.4 times more likely to host LMXBs than metal-poor ones. This is slightly higher than that measured in other surveys, likely because of larger contamination of the GC sample in previous ground-based data sets and the inclusion of galaxies with intermediate-age clusters in others. Intriguingly, the LMXBs in NGC 1399 are preferentially in the reddest clusters of the metal-rich GC subsystem. This indicates that the red peak of the bimodal GC color distribution itself encompasses clusters with a range of enrichment histories. The strength of this effect varies from galaxy to galaxy, possibly indicating differences in their metal enrichment histories. Field LMXBs in our program galaxies are more concentrated toward the center of their host galaxies than GC LMXBs. This suggests that a majority of field LMXBs are formed in situ and are not a population that has escaped from current GCs. This is consistent with previous specific frequency-based studies. The brightest X-ray sources in GCs appear to be preferentially associated with luminous, metal-rich clusters. We show that it is probable that some of these clusters host multiple bright LMXBs, while the probability is much lower for metal-poor GCs. If this interpretation is correct, our study implies that LMXBs in more metal-rich cluster systems should reveal a longer high-luminosity X-ray tail and show less X-ray variability than metal-poor cluster populations.

An Explanation for Metallicity Effects on X-Ray Binary Properties

We show that irradiation-induced stellar winds can explain two important metallicity effects in X-ray binaries: the higher numbers and softer spectra of X-ray binaries in metal-rich globular clusters (GCs) compared to those in metal-poor ones. As has been previously noted by Iben, Tutukov, and Fedorova, the winds should be stronger at lower metallicity because of less efficient line cooling. This speeds up the evolution of low-mass X-ray binaries (LMXBs) in metal-poor clusters and hence reduces their number. These winds can also provide extra material near the accreting object, which can create an intrinsic absorber that hardens the X-ray spectra of the metal-poor cluster systems relative to the metal-rich ones, as suggested by observations. We outline some additional observational predictions of this model.

Hot Populations in M87 Globular Clusters

To explore the production of UV-bright stars in old, metal-rich populations like those in elliptical galaxies, we have obtained Hubble Space Telescope (HST) Space Telescope Imaging Spectrograph far- and near-UV photometry of globular clusters (GCs) in four fields in the giant elliptical (gE) galaxy M87. To a limit of mFUV ~ 25 we detect a total of 66 GCs in common with the deep HST optical-band study of Kundu et al. Despite strong overlap in V- and I-band properties, the M87 GCs have UV-optical properties that are distinct from clusters in the Milky Way and in M31. M87 clusters, especially metal-poor ones, produce larger hot horizontal-branch populations than do Milky Way analogs. In color plots including the near-UV band, the M87 clusters appear to represent an extension of the Milky Way sequence. Cluster mass is probably not a factor in these distinctions. The most metal-rich M87 GCs in our sample are near solar metallicity and overlap the local E galaxy sample in estimated Mg2 line indices. Nonetheless, the clusters produce much more UV light at a given Mg2, being up to 1 mag bluer than any gE galaxy in (FUV - V) color. The M87 GCs do not appear to represent a transition between Milky Way-type clusters and E galaxies. The differences are in the correct sense if the clusters are significantly older than the E galaxies. Comparisons with Galactic open clusters indicate that the hot stars lie on the extreme horizontal branch, rather than being blue stragglers, and that the extreme horizontal branch becomes well populated for ages 5 Gyr. Existing model grids for clusters do not match the observations well, due to poorly understood giant branch mass loss or perhaps high helium abundances. We find that 41 of our UV detections have no optical-band counterparts. Most appear to be UV-bright background galaxies seen through M87. Eleven near-UV variable sources detected at only one epoch in the central field are probably classical novae. Two recurrent variable sources have no obvious explanation but could be related to activity in the relativistic jet.