Because of their proximity, nearby spiral galaxies is where the early work on extra-galactic XRB populations begun (see Fabbiano 1995). For the same reason, these are the galaxies where the deepest samples of sources have been acquired with Chandra and XMM-Newton. Here we will discuss first the recent work done on M31, which, not surprisingly, is the galaxy that has been studied in most detail. We will then review the results on M81, M83, and M101, to provide examples of the XRB populations in a wider variety of spirals. We conclude this section with a summary of the work on actively star-forming galaxies. We note that this field is evolving rapidly, with an increasing number of galaxies being surveyed and with the sensitivity limit being pushed to fainter fluxes, with ever deeper Chandra observations.
Being at a distance of only ~ 700 kpc, M31 (NGC 224, the Andromeda nebula) is the spiral (Sb) galaxy closest to us. M31 has been observed by virtually all the X-ray observatories since Uhuru, the first X-ray satellite (for a history of the X-ray observations of galaxies, see Fabbiano & Kessler 2001). Starting with the Einstein Observatory and following on with ROSAT, M31 has been the prime target for systematic studies of a population of extragalactic XRBs, and for comparisons with our own Galactic XRBs (e.g., Long & Van Speybroeck 1983; Trinchieri & Fabbiano 1991; Primini, Forman & Jones 1993; Supper et al. 1997, 2001). Chandra and XMM-Newton observations, both by themselves and in combination, are providing new insight on the characteristics of the XRB population of M31. With its subarcsecond resolution, Chandra is unique in resolving dense source regions, such as the circum-nuclear region of M31, and detecting faint sources (Garcia et al. 2000). Given the proximity of M31 and the relatively low density of luminous XRBs, XMM-Newton provides valuable data on the XRB population of this galaxy, if one excludes the centermost crowded core (Shirey et al. 2001).
Source variability and counterparts - Multiple observations of the same fields with these two observatories (and comparison with previous observations) have confirmed the general source variability characteristic of XRBs. XMM-Newton work, following the first statement of source variability (Osborne et al. 2001), includes detailed studies of interesting luminous sources. Trudolyubov, Borozdin & Priedhorsky (2001) report the discovery of three transient sources, with maximum X-ray emission in the 1037 ergs s-1 range: a candidate low-mass black-hole binary, a source with a long (> 1 year) outburst, and a supersoft transient. Trudolyubov et al. (2002b) report an 83% modulation with a 2.78 hr period in the X-ray source associated with the globular cluster (GC) Bo 158. Comparison with earlier XMM-Newton observations and with the ROSAT PSPC data, allows these authors to conclude that the modulation is anticorrelated with the source flux, suggesting perhaps a larger less obscured emission region in high state. This source resembles Galactic `dip' XRBs, and could be an accreting neutron star. Its period suggest a highly compact system (separation ~ 1011 cm).
Widespread source variability is evident from Chandra observations, both from a ks HRC study of the bulge (Kaaret 2002), from a set of eight Chandra ACIS observations of the central 17' × 17' taken between 1999 and 2001 (Kong et al. 2002), and from a 2.5 years 17 epochs survey with the Chandra HRC (Williams et al. 2003), which also includes the data from Kaaret (2002).
Kong et al. find 204 sources, including nine supersoft sources,
with a detection limit of ~ 2 × 1035 ergs s-1.
This detection limit is 5 times fainter than that of the
ROSAT HRI catalog
(Primini, Forman &
Jones 1993),
which lists only 77 sources in the surveyed area.
They report 22 globular cluster (GC) identifications, 2 supernova
remnants, and 9 planetary nebulae associations.
By comparing the different individual data sets,
they establish that 50% of the sources vary on timescales of months, and
13 are transients. The spectra of the most luminous sources can be
fitted with power-laws with
~ 1.8,
and, of these, 12 show coordinated flux and spectral variability. Two
sources exhibit harder spectra with increasing count rate, reminiscent of
Galactic Z sources (e.g.
Hasinger & van der
Klis 1989).
All these characteristics point to an XRB population similar to that of
the Milky Way. The HRC survey
(Williams et al.
2003)
reports fluxes and light curves for 173 sources,
and finds variability in 25% of the sources; 17 of these sources are
transients, and two of these are identified with variable HST
WFPC2 U band counterparts. One of these two sources is
also a transient in the optical and has global properties suggesting a
~ 10 M
black hole X-ray nova with a period
9 days.
Williams et al.
(2003)
determine that at any given time there are
1.9 ± 1.3 active X-ray transients in M31, and from here
they infer that the ratio of neutron star to black hole LMXBs in M31 is ~ 1, comparable to that in the Galaxy.
Globular Cluster (GC) sources - The recent X-ray populations studies of M31 with Chandra and XMM-Newton demonstrate the importance of large area surveys of the entire galaxian system. A targeted study of GCs with three Chandra fields at large galactocentric radii (Di Stefano et al. 2002) revives the old suggestion (Long & Van Speybroeck 1983) that the M31 GC sources are more X-ray luminous than Galactic GC sources. This hypothesis had been dismissed with the ROSAT M31 survey (Supper et al. 1997), which however covered only the central 34' of M31. Di Stefano et al. (2002) find that in their fields the most luminous sources are associated with GCs. They detect 28 GCs sources, 15 of which are new detections: 1/3 of these sources have LX(0.5 - 7 keV) > 1037 ergs s-1; 1/10 of the sources have LX(0.5 - 7 keV) > 1038 ergs s-1. The X-ray luminosity function (XLF) of the M31 GC sources differs from the Galactic GC XLF, by both having a larger number of sources, and by extending a decade higher in X-ray luminosity (the most luminous M31 GC is Bo 375 with LX > 2 × 1038 ergs s-1; compare with Milky Way GCs, that emit less than 1037 ergs s-1).
Supersoft sources (SSS) -
SSS are very soft X-ray sources, with most of the emission below 1 keV,
and spectra that can be fitted with black body temperatures of
100 eV (see Chapter by
Kahabka in this volume).
SSS were first discovered in M31 with ROSAT
(Supper et al.
1997).
As noted above,
Kong et al.
(2002)
reported nine SSS in their Chandra observations of M31.
Recent work by
Di Stefano et al.
(2003)
reports 33 SSSs in the same fields surveyed for GCs by
Di Stefano et al.
(2002),
of which only two were known since the ROSAT times.
Two SSSs are identified with symbiotic stars and two with supernova
remnants, but the bulk are likely to be
supersoft XRBs. These sources are highly variable, and may be classified
in two spectral groups:
sources with kT 100 eV, and
other sources with harder emission, up to kT ~ 300 eV.
Sixteen of them (on average the most luminous) cluster in the bulge,
others are found in both the disk and the halo of M31.
Di Stefano et al.
(2003)
point out that some of these sources are detected
with luminosities well below
1037 ergs s-1, the luminosity of a 0.6
M white
dwarf steadily burning hydrogen, and are therefore likely
to be lower mass white dwarfs or luminous cataclysmic variables.
The bulge - The XLFs of the global core population [Kaaret 2002 (Chandra HRC); Kong et al. 2002 (Chandra ACIS); Trudolyubov et al. 2002a (XMM-Newton)] all are in general agreement with each other and with the Einstein (Trinchieri & Fabbiano 1991) and ROSAT studies (Primini, Forman & Jones 1993). However, because of the resolution and sensitivity of Chandra, both Kong et al. (2002) and Kaaret (2002) can can look at the bulge source population in greater detail than ever before.
Kong et al. divide the detected sources in three groups, based on their galactocentric position: inner bulge (2` × 2'), outer bulge (8' × 8', excluding the inner bulge sources), and disk (17' × 17', excluding the two bulge regions). When considering the entire bulge population, these authors find a general low luminosity break of the XLF at ~ 2 × 1037 ergs s-1, in agreement with Trudolyubov et al. (2002a). However, they also find that the break appears to shift to lower luminosities with decreasing galactocentric radius, going from 0.18 ± 0.08 × 1037 ergs s-1 in the inner bulge to 2.10 ± 0.39 × 1037 ergs s-1 in the outermost `disk' region. They note that if the breaks mark episodes of star formation, the more recent of these events must have occurred at larger radii. The slopes of the XLFs also vary (0.67 ± 0.08 in the center, 1.86 ± 0.40 in the outermost region), but this trend is the opposite of that expected from progressively young populations, where more luminous, short lived sources, may be found (see e.g. Kilgard et al. 2002; Zezas & Fabbiano 2002; Section 2.4). Kong et al. suggest that the XRB populations of the central regions of M31 may instead all be old (see Trudolyubov et al. 2002a), with the shifts of the break resulting from the inclusion of new classes of fainter sources in the inner regions, rather than from a disappearance of the most luminous sources.
Kaaret (2002) contributes to the debate on the nature of the inner bulge sources by investigating their spatial distribution. He shows that the the number of X-ray sources detected in the centermost regions of the bulge (< 100") is in excess of what would be expected on the basis of the radial distribution of the optical surface brightness, and suggests that this result may be consistent with a GC origin for the LMXBs.
X-ray source populations in different galaxian fields - With the increased rate of papers on M31, resulting from the XMM-Newton and Chandra surveys of this galaxy, we are now realizing that the X-ray source population of M31 is more varied than previously thought, and that there are correlations between the properties of the X-ray sources and those of the stellar field to which they belong.
In contrast with previous reports (e.g.
Trinchieri & Fabbiano
1991;
Kong et al.
2002),
Trudolyubov et al.
(2002a),
by using a larger definition for the radius of the bulge (15'), with
XMM-Newton observations conclude that,
although the XLFs of bulge and disk sources have a similar cumulative
slope (-1.3), disk sources are all fainter than
LX < 2 × 1037 ergs s-1,
while bulge sources can have luminosities as high as
LX ~ 1038 ergs s-1.
They suggest that the most luminous sources are associated with the
older stellar population, as in the Milky Way
(Grimm, Gilfanov &
Sunyaev 2002).
However, the fields studied by
Trudolyubov et al.
(2002a)
do not include the areas surveyed
by Di Stefano et
al. (2002),
where the most luminous GC sources are found
(see Fig. 1).
Figure 1. Regions of M31 observed with
Chandra and XMM-Newton. Dots are detected Chandra
sources; yellow crosses and blue diamonds identify supernova remnants
and OB associations in the field (not X-ray sources), respectively (from
Kong et al.
2003).
A Chandra ACIS study of XLFs from different regions of M31
(Fig. 1;
Kong et al.
2003),
uses a follow-up of the
Di Stefano et al.
(2002)
survey. The results (Fig. 2) show that the
sources in the central 17' × 17' region are overall more luminous
than those from the outer fields (as noticed by
Trudolyubov et al.
2002a),
but only if one removes the GC population, which appears to have a
relatively more numerous high luminosity component than the central
sources. The slopes of the XLFs of the external fields also vary, and
there is an indication that these differences are related to variations
in the stellar populations of the different fields: Field 1, which has
the steepest slope (cumulative -1.7+0.34-0.15) and
also the lowest density of X-ray sources,
does not appear to have a large young
population of stars; Field 2, with the largest X-ray source population
and the flattest XLF slope (cumulative -0.9) is in the region with the
youngest stellar population. This slope is the closest to that
(0.63 ± 0.13) derived by
Grimm, Gilfanov &
Sunyaev (2002)
for the high mass X-ray binaries (HMXBs) in the Galaxy; Field 3, with
an intermediate XLF slope instead does not appear to cover a large
stellar population.
The overall integrated slope is instead similar to that found by Grimm
et al. for the Galactic low-mass X-ray binary (LMXB) population,
suggesting that these sources dominate the X-ray emission of M31.
Figure 2. Cumulative XLFs and best-fit
power-laws from different fields of M31
(Kong et al.
2003).
Williams et al. (2003), using the Chandra HRC survey of M31,
distinguish between a roughly radially symmetric
bulge population (within a 7' radius) and a field population, outside
this inner region. They report different XLFs for bulge and disk
sources, with a flatter broken power-law representing well the disk
distribution. Their survey has a wider (although shallower) coverage of
the entire M31 galaxy, than the
Trudolyubov et al.
(2002a)
work, and covers also the southern half of the disk, where the
X-ray sources are significantly more luminous than in the northern disk,
surveyed with XMM-Newton by Trudolyubov et al..
The Trudolyubov et
al. (2002a),
Kong et al.
(2003), and
Williams et al .
(2003)
papers are illuminating in demonstrating the variability of the XLF in
different regions, and in pointing out
how a good spatial sampling and supporting multi-wavelength information, are
needed to get a complete picture of the XRB population of M31.
