4. XRBs IN ELLIPTICAL AND S0 GALAXIES

As discussed in the 1995 chapter (Fabbiano 1995), XRBs could not be directly detected in E and S0 galaxies with pre-Chandra telescopes, because of the distance of these galaxies and the limited angular resolution of the telescopes. The presence of XRBs in E and S0 galaxies was predicted by Trinchieri & Fabbiano (1985), based on an analogy with the bulge of M31, for which such a population could be detected (Van Speybroeck et al. 1979; see also Fabbiano, Trinchieri & Van Speybroeck 1987). This early claim was reinforced by differences in the average spectral properties of E and S0 galaxies with different X-ray-to-optical luminosity ratios, that suggested a baseline X-ray faint XRB emission (Kim, Fabbiano & Trinchieri 1992; Fabbiano, Kim & Trinchieri 1994), and by the ASCA discovery of a hard spectral component in virtually all E and S0 galaxies (Matsushita et al. 1994), which, however, could also have been due, at least in part, to accreting massive nuclear black holes (Allen, Di Matteo & Fabian, 2000).

The Chandra images (Fig. 11) leave no doubt about the presence of rich populations of point-like sources in E and S0 galaxies. Published results, of which the first one is the paper on NGC 4697 by Sarazin, Irwin & Bregman (2000), include point-source detections in a number of galaxies. These source populations have been detected with varying low-luminosity detection thresholds (a function of galaxy distance and observing time). While most of the detected sources have luminosities in the 10³⁷ - 10³⁹ ergs s^-1 range, some were detected at luminosities above 10³⁹ ergs s^-1, in the Ultra-Luminous-X-ray (ULX) source range (see Section 3). A representative summary (limited to papers published or in press as of May 2003) in given in Table 1.

Figure 11. Chandra ACIS image of the Virgo elliptical NGC 4365, using archival data. The white ellipse is the D25 isophote, from de Vaucouleurs et al. (1991).

The X-ray colors or co-added spectra of these sources are consistent with those of LMXBs (see above references, and Irwin, Athey & Bregman 2002); however, a variety of spectral properties have been reported in some cases, similar to the spectral variety of Galactic and Local Group XRBs, including a few instances of very soft and supersoft (i. e., all photons below ~ 1 keV) sources (e.g., NGC 4697, Sarazin, Irwin & Bregman 2000; M84, Finoguenov & Jones 2002; NGC 1399, Angelini, Loewenstein & Mushotzky 2001; NGC 1316, Kim & Fabbiano 2003). The overall spatial distribution of these sources follows that of the stellar light, but there are exceptions, such as in NGC 720, where the most luminous sources follow arcs (Jeltema et al. 2003), NGC 4261 and NGC 4697, where the X-ray source distributions are highly asymmetric (Zezas et al. 2003), and NGC 4472, where the X-ray source distribution may be more consistent with that of Globular Clusters (GCs) than of the general field stellar light (Kundu, Maccarone & Zepf 2002; Maccarone, Kundu & Zepf 2003). No firm conclusion on the origin and evolution of these sources exists. Given the old stellar population of the parent galaxies, and the life-times of LMXBs, it has ben suggested that these sources may be outbursting transients (Piro & Bildsten 2002). Alternatively, more recent formation and evolutions in GCs may result in steady sources (Maccarone, Kundu & Zepf 2003). With the exception of NGC 5128, which is near enough to allow detection of sources in the 10³⁶ ergs s^-1 luminosity range, and for which multiple observations demonstrate widespread source variability (Kraft et al. 2001), the Chandra observations performed so far typically only give a single snapshot of the most luminous part of the XRB population in a given galaxy. In NGC 5128, a comparison of the two Chandra observations reveals at least five transients (sources that disappear with a dimming factor of at least 10), supporting the Piro & Bildsten scenario.

Chandra observations of highly significant asymmetries in the spatial distribution of X-ray sources in otherwise regular old elliptical galaxies (Zezas et al. 2003) may suggest rejuvenation of the stellar population of these galaxies. In NGC 4261, the most significant example, all the detected sources are luminous, above the Eddington limit for a neutron star accretor. If the X-ray sources were standard LMXBs belonging to the dominant old stellar population, we would expect their spatial distribution to be consistent (within statistics) with that of the stellar light. However this is not so, as indicated by Kolmogorov-Smirnov tests and Bayesian block analysis. On the basis of simulations of galaxy interactions (Hernquist & Spergel 1992; Mihos & Hernquist 1996), this result suggests that the luminous XRBs may belong to a younger stellar component, related to the rejuvenating fall-back of material in tidal tails onto a relaxed merger remnants.