Star formation history and X-ray binary populations: the case of the Small Magellanic Cloud

Abstract

Using Chandra, XMM-Newton, and optical photometric catalogs we study the young X-ray binary (XRB) populations of the Small Magellanic Cloud. We find that the Be/X-ray binaries (Be-XRBs) are observed in regions with star formation rate bursts ~25-60 Myr ago. The similarity of this age with the age of maximum occurrence of the Be phenomenon (~40 Myr) indicates that the presence of a circumstellar decretion disk plays a significant role in the number of observed XRBs in the 10-100 Myr age range. We also find that regions with strong but more recent star formation (e.g., the Wing) are deficient in Be-XRBs. By correlating the number of observed Be-XRBs with the formation rate of their parent populations, we measure a Be-XRB production rate of ~1 system per 3 x 10^(-3) M$_{\odot}$/yr. Finally, we use the strong localization of the Be-XRB systems in order to set limits on the kicks imparted on the neutron star during the supernova explosion. Comment: 15 pages, 2 figures, published in ApJ Letters

Full text

Page 1

arXiv:1006.2381v1 [astro-ph.HE] 11 Jun 2010
Star formation history and X-ray binary populations: the case of
the Small Magellanic Cloud
V. Antoniou1,2,3, A. Zezas1,2,4, D. Hatzidimitriou4,5, V. Kalogera6
ABSTRACT
Using Chandra, XMM-Newton and optical photometric catalogs we study
the young X-ray binary (XRB) populations of the Small Magellanic Cloud. We
find that the Be/X-ray binaries (Be-XRBs) are observed in regions with star
formation rate bursts ∼25-60 Myr ago. The similarity of this age with the age
of maximum occurrence of the Be phenomenon (∼40 Myr) indicates that the
presence of a circumstellar decretion disk plays a significant role in the number
of observed XRBs in the 10-100 Myr age range. We also find that regions with
strong but more recent star formation (e.g., the Wing) are deficient in Be-XRBs.
By correlating the number of observed Be-XRBs with the formation rate of their
parent populations, we measure a Be-XRB production rate of ∼1 system per
3×10−3M⊙/yr. Finally, we use the strong localization of the Be-XRB systems in
order to set limits on the kicks imparted on the neutron star during the supernova
explosion.
Subject headings: Magellanic Clouds—pulsars: general—stars: early-type—stars:
emission-line, Be—stars: formation—X-rays: binaries
1Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA; vanto-
niou@head.cfa.harvard.edu
2Physics Department, University of Crete, P.O. Box 2208, GR-710 03, Heraklion, Crete, Greece
3Department of Physics and Astronomy, Iowa State University, 12 Physics Hall, Ames, IA 50011, USA
4IESL/Foundation for Research and Technology-Hellas, P.O. Box 1527, GR-711 10 Heraklion, Crete,
Greece
5Section of Astrophysics, Astronomy and Mechanics, Department of Physics, University of Athens,
Panepistimiopolis, GR-157 84 Zografos, Athens, Greece
6Department of Physics and Astronomy, Northwestern University, 2145 Sheridan Road, Evanston, IL
60208, USA

Page 2

– 2 –
1. Introduction
Nearby star-forming galaxies offer a unique environment to study the young (<100
Myr) X-ray binary (XRB) populations. One of the best cases is the Small Magellanic Cloud
(SMC), which at ∼60 kpc is our second nearest star-forming galaxy (Hilditch et al. 2005).
Its proximity, well mapped extinction (Zaritsky et al. 2002), moderate Galactic foreground
absorption (NH≃ 6 × 1020cm−2; Dickey & Lockman 1990), small line-of-sight depth of its
young, central stellar populations (<10 kpc; Crowl et al. 2001; Harries et al. 2003), and its
well-determined recent star formation history (SFH; Harris & Zaritsky 2004 [HZ04]) make
the SMC the ideal environment for directly studying the link between XRB populations and
star formation (SF). Furthermore, the wealth of multi-wavelength data allows us to clarify
the X-ray sources and obtain an even more precise picture of their population.
Several studies have compared the number of Be/X-ray binaries (Be-XRBs) in the Mag-
ellanic Clouds and the Galaxy (e.g., Majid et al. 2004, Haberl & Pietsch 2004, Coe et
al. 2005), concluding that the SMC hosts an unusually large number of these systems. There
is only one identified supergiant XRB located in the SMC Wing (SMC X-1; Webster et
al. 1972) in a population of ∼100 High-Mass XRBs (HMXBs; e.g., Liu et al. 2005, Antoniou
et al. 2009b [Paper II]). However, only few of those HMXBs have determined spectral types
(e.g., out of the 92 listed in Liu et al. 2005, 53 are cited as Be-XRBs but only 19 have
been confirmed spectroscopically). Later works by Antoniou et al. (2009a [Paper I]), Haberl
et al. (2008), McBride et al. (2008), Shtykovskiy & Gilfanov (2005), Coe et al. (2005) and
others have increased the number of known Be-XRBs to 67 to date. Nevertheless, this over-
abundance can be partly explained by the enhanced SMC SFH ∼40 Myr ago (e.g., Majid et
al. 2004, Shtykovskiy & Gilfanov 2007 [SG07]). However, Antoniou et al. (2009b) show that
even after accounting for the difference in the star formation rate (SFR) between the SMC
and the Galaxy, the SMC hosts ∼1.5 times more Be-XRBs than the Galaxy down to a lim-
iting luminosity of LX≥ 1034erg s−1. This residual excess can be explained by the different
metallicity of these galaxies, as justified by population synthesis models (Dray 2006) and re-
cent observations of Be stars (e.g., Wisniewski & Bjorkman 2006, Martayan et al. 2007). The
work of Antoniou et al. (2009b) also indicated spatial variations of the Be-XRB populations
within the SMC Bar, which could be evidence for small supernova (SN) kicks.
The SMC Bar hosts stellar populations with ages <100 Myr [HZ04] and the vast majority
of the SMC pulsars (Galache et al. 2008). [SG07] found that the age distribution of the
HMXBs peaks at ∼20-50 Myr after the SF event, while McSwain and Gies (2005) observed
a strong evolution in the fraction of Be stars with age up to 100 Myr, with a maximum at
∼25-80 Myr. These results motivated us to investigate the connection between the spatially
resolved SFH in and around the SMC Bar and the number and spatial distribution of the

Page 3

– 3 –
XRBs. In this study, we use the results from our Chandra survey of the central, most actively
star forming, SMC Bar (A. Zezas et al. 2010, in preparation; Papers I, II), and data from
our XMM-Newton survey of the outer SMC regions which host young and intermediate age
stellar populations (∼10-500 Myr; [HZ04]).
2.X-ray observations and data analysis
Using the ACIS-I detector on board Chandra we observed five fields in the central part of
the SMC (the so called SMC Bar), with typical exposure times of 8-12 ks. These observations
yielded a total of 158 sources, down to a limiting luminosity of ∼ 4×1033erg s−1(0.5-7.0 keV
band), reaching the luminosity range of quiescent HMXBs (typically LX∼ 1033− 1035erg
s−1; van Paradijs & McClintock 1995). The analysis of the data, the source-list and their
X-ray luminosity functions are presented in A. Zezas et al. (2010, in preparation), while their
optical counterparts and resulting classification are given in Papers I, II.
Our XMM-Newton survey consists of five observations in the outer SMC Bar, performed
with the three EPIC (MOS1, MOS2, and PN) detectors in full frame mode. One of these
fields was affected by high background flares and it is not included in this work. The data were
analyzed with the XMM-Newton Science Analysis System (SAS) version 7.0.0. After pro-
cessing the raw data with the epchain and emchain tasks, we filtered any bad columns/pixels
and high background flares (excluding times when the total count rate deviated more than
3σ from the mean), resulting in 5-18 ks net exposures. We only kept events of patterns 0-4
for the PN and 0-12 for the MOS detectors. Source detection was performed simultaneously
in five energy bands (0.2-0.5, 0.5-1.0, 1.0-2.0, 2.0-4.5, and 4.5-12.0 keV), and the three EPIC
detectors with the maximum likelihood method (threshold set to 7) of the edetect chain
task. The detected sources were visually inspected for spurious detections. We detected 186
sources down to a limiting luminosity of ∼ 3.5×1033erg s−1(0.2-12 keV), out of which 4-8
sources are expected to be spurious based on the calibration of Watson et al. (2009).
In Table 1, we give the ID and the coordinates of the X-ray fields, along with the
properties of the dominant SF event in each field (see §3).
2.1. X-ray source classification
New HMXBs and candidate Be-XRBs are identified based on their X-ray and optical
properties. Hardness ratios between the soft (0.5-1.0 keV), medium (1.0-2.0 keV), and hard
(2.0-4.5 keV) bands were used as an initial measurement of their X-ray spectral properties.

Page 4

– 4 –
A hard X-ray spectrum or hardness ratio (equivalent to a photon index of Γ ∼1) is indicative
of a pulsar binary (e.g.,Haberl et al. 2008). For the identification of the optical counterparts
of the XMM-Newton sources we followed the analysis of Antoniou et al. (2009b). We cross-
correlated their coordinates with the OGLE-II (Udalski et al. 1998) and MCPS (Zaritsky et
al. 2002) catalogs, and searched for optical matches in a 5′′radius around each X-ray source
(which includes the boresight error of XMM-Newton; e.g.,Brusa et al. 2007). Given the small
number of X-ray sources with independently known optical counterparts, we cannot correct
these observations for boresight errors. Based on the position of these counterparts on the
V,B − V color-magnitude diagram, we identify sources with early OB-type counterparts,
while from hardness ratio or spectral analysis we identified those hard X-ray sources (Γ ∼1),
strongly suggesting they are XRB pulsars. Although Monte-Carlo simulations indicate a
significant number of spurious sources in these fields, the identification of a hard X-ray
source with an early-type counterpart suggests that this is a true match.
We find that 15 XMM-Newton sources have O- or B-type counterparts, while only 8
of those are hard X-ray sources, suggesting they are HMXBs. Since all but one of the
confirmed SMC HMXBs are Be-XRBs, they are also candidate Be-XRBs. Their properties
are presented in Table 2. The X-ray luminosity is derived assuming a power-law spectrum
of Γ = 1 and H I column density equal to 4.81, 6.63, and 4.51 × 1020cm−2for fields 1, 2,
and 3, respectively (based on Dickey & Lockman 1990). The X-ray spectra of two sources
with >200 counts were fitted with an absorbed power law, resulting in a photon index of
Γ = 0.65 ± 0.04 and 0.97 ± 0.25, and a column density of NH= (3.31 ± 0.02) × 1020cm−2
and (0.30±0.15)×1022cm−2, for sources 2-1 (by simultaneously fitting its MOS1 and MOS2
spectra) and 3-1 (from its PN spectrum), respectively. Source 2-1 in particular is a known
Be-XRB pulsar with a period of 169.3 s (Lochner et al. 1998) associated with emission-
line object [MA93]623 (Meyssonnier & Azzopardi 1993; 3.3′′away), source 3-1 remained
unclassified in Sasaki et al. 2000 (ROSAT HRI src ID 11), while source 3-3 is the only one
not included in the pipeline EPIC detection list.
If we include to the above sources the confirmed and candidate Be-XRBs that lie within
the Chandra and XMM-Newton fields (from this work and those mentioned in §1), we have
a total of 54 (39) and 11 (2) HMXBs (Be-XRBs), respectively. For Chandra fields this is the
sum of unique Be-XRBs, i.e., sources detected in two overlapping fields are counted once.
3.SFH and XRB populations
The recent SFH in our Chandra and XMM-Newton fields is derived by averaging the
spatially resolved SFH of the MCPS regions (∼ 12′× 12′; [HZ04]) encompassed by them.

Page 5

– 5 –
We find that:
(i) For the Chandra fields, the most recent major burst peaked ∼42 Myr ago, and it had a
duration of ∼40 Myr. Moreover, there were older SF episodes (∼0.4 Gyr ago) with lower
intensity but longer duration, besides a more recent episode (∼7 Myr) observed only in
Chandra field 4.
(ii) For XMM-Newton field 3, the most recent major burst occurred ∼67 Myr ago. We also
observed two fields with very young populations (most recent major burst at ∼11 and 17
Myr ago for fields 1 and 2, respectively). XMM-Newton field 2 had an additional intense
burst ∼67 Myr ago (Table 1).
In order to investigate the link between stellar and XRB populations, we calculate the
average SFH for the MCPS regions (∼ 12′× 12′; [HZ04]) that host one or more Be-XRBs
(candidate and confirmed) detected in our Chandra and XMM-Newton surveys (39 and 2,
respectively; see §2.1). The SFH in each region is weighted by the encompassed number of
Be-XRBs, and the error bars are derived based on the upper and lower limits of [HZ04]. We
repeat this exercise for the 15 MCPS regions without any known Be-XRB in our surveys.
The two SFHs are presented in Figure 1. The SFH of the Be-XRBs (black points) is strongly
peaked at ∼42 Myr, while fields without any Be-XRB (gray points) have minimal SFR at
this age. This underscores the difference between the fields with and without Be-XRBs, and
suggests a clear connection between an SF event and the observed Be-XRBs.
Following the above comparison, we also construct the SFH of the MCPS regions hosting
one or more known X-ray pulsars within any of our fields1(Figure 1; black points), and
for those that do not host such sources (gray points). A large fraction of these pulsars
(∼60%; Liu et al. 2005) also appears in the Be-XRBs sample, since the vast majority of
their companions are Be stars. This link is reinforced by the fact that all the counterparts of
these X-ray sources lie on the region of the color-magnitude diagram consistent with main-
sequence stars of age ∼40 Myr (Paper II). For completeness we present both, since the pulsar
and the Be-XRB samples are selected on the basis of their timing and optical properties,
respectively. In total, in the MCPS regions that overlap with the Chandra fields lie 30 X-ray
pulsars, while in the XMM-Newton fields only 2 (sum of unique sources as in Section 2.1).
As expected, the pattern in their SFH is very similar to that of Be-XRBs. For regions rich in
X-ray pulsars the SF peaks at ∼42 Myr, while for regions without pulsars there is no peak
at this age.
The average SFH of the MCPS regions with and without any Be-XRBs detected in the
1Based on the on-line census of Malcolm Coe (http://www.astro.soton.ac.uk/∼mjc/smc/ as of 2009 June
18).