Parsec-scale X-ray Flows in High-mass Star-forming Regions
ABSTRACT The Chandra X-ray Observatory is providing remarkable new views of massive star-forming regions, revealing all stages in the life cycle of high-mass stars and their effects on their surroundings. We present a Chandra tour of several high-mass star-forming regions, highlighting physical processes that characterize the life of a cluster of high-mass stars, from deeply-embedded cores too young to have established an HII region to superbubbles so large that they shape our views of galaxies. Along the way we see that X-ray observations reveal hundreds of stellar sources powering great HII region complexes, suffused by both hard and soft diffuse X-ray structures caused by fast O-star winds thermalized in wind-wind collisions or by termination shocks against the surrounding media. Finally, we examine the effects of the deaths of high-mass stars that remained close to their birthplaces, exploding as supernovae within the superbubbles that these clusters created. We present new X-ray results on W51 IRS2E and 30 Doradus and we introduce new data on Trumpler 14 in Carina and the W3 HII region complexes W3 Main and W3(OH). Comment: 6 pages, 3 figures, to appear in the proceedings of IAU Symposium 227,"Massive Star Birth - A Crossroads of Astrophysics," eds. R. Cesaroni, E. Churchwell, M. Felli, and C.M. Walmsley
- SourceAvailable from: Milton Virgilio Fernandes[Show abstract] [Hide abstract]
ABSTRACT: Stellar clusters are potential acceleration sites of very-high-energy (VHE, E > 100GeV) particles since they host supernova remnants (SNRs) and pulsar wind nebulae (PWNe). Additionally, in stellar clusters, particles can also be accelerated e.g. at the boundaries of wind-blown bubbles, in colliding wind zones in massive binary systems or in the framework of collective wind or wind/supernova(SN) ejecta scenarios. Motivated by the detection of VHE gamma-ray emission towards Westerlund 2 and assuming similar particle acceleration mechanisms at work, Westerlund 1 is an even more promising target for VHE gamma-ray observations given that massive star content and distance are more favorable for detectable VHE gamma-ray emission compared to Westerlund 2. Here, H.E.S.S. observations of massive stellar clusters in general with special emphasis on the most massive stellar cluster in the galaxy, Westerlund 1 are summarized. Comment: Invited talk at HEPIMS Workshop (Jaen 2009); 10 pages, 4 figures, to appear in PASP06/2009;
- [Show abstract] [Hide abstract]
ABSTRACT: We present an X-ray tour of diffuse emission in the 30 Doradus star-forming complex in the Large Magellanic Cloud using high spatial resolution X-ray images and spatially resolved spectra obtained with the Advanced CCD Imaging Spectrometer on board the Chandra X-Ray Observatory. The dominant X-ray feature of the 30 Doradus nebula is the intricate network of diffuse emission generated by interacting stellar winds and supernovae working together to create vast superbubbles filled with hot plasma. We construct maps of the region showing variations in plasma temperature (T = 3-9 million degrees), absorption [NH = (1-6) × 1021 cm-2], and absorption-corrected X-ray surface brightness [SX = (3-126) × 1031 ergs s-1 pc-2]. Enhanced images reveal the pulsar wind nebula in the composite supernova remnant N157B, and the Chandra data show spectral evolution from nonthermal synchrotron emission in the N157B core to a thermal plasma in its outer regions. In a companion paper we show that R136, the central massive star cluster, is resolved at the arcsecond level into almost 100 X-ray sources. Through X-ray studies of 30 Doradus the complete life cycle of such a massive stellar cluster can be revealed.The Astronomical Journal 12/2007; 131(4):2140. · 4.97 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: We report here results from a Chandra ACIS observation of the stellar populations in and around the M17 H II region. The field reveals 886 sources with observed X-ray luminosities (uncorrected for absorption) between ~ 29.3 ergs s-1 < log LX < 32.8 ergs s-1, 771 of which have stellar counterparts in infrared images. In addition to comprehensive tables of X-ray source properties, several results are presented:The Astrophysical Journal Supplement Series 12/2008; 169(2):353. · 16.24 Impact Factor
Massive Star Birth: A Crossroads of Astrophysics
Proceedings IAU Symposium No. 227, 2005
R. Cesaroni, E. Churchwell, M. Felli, & C.M. Walmsley, eds.
c ? 2005 International Astronomical Union
Parsec-scale X-ray Flows in High-mass
L. K. Townsley, P. S. Broos, E. D. Feigelson, and G. P. Garmire
Department of Astronomy & Astrophysics, 525 Davey Laboratory, Pennsylvania State
University, University Park, PA 16802, USA
email: townsley, patb, edf, email@example.com
Abstract. The Chandra X-ray Observatory is providing remarkable new views of massive star-
forming regions, revealing all stages in the life cycle of high-mass stars and their effects on their
surroundings. We present a Chandra tour of several high-mass star-forming regions, highlighting
physical processes that characterize the life of a cluster of high-mass stars, from deeply-embedded
cores too young to have established an HII region to superbubbles so large that they shape our
views of galaxies. Along the way we see that X-ray observations reveal hundreds of stellar sources
powering great HII region complexes, suffused by both hard and soft diffuse X-ray structures
caused by fast O-star winds thermalized in wind-wind collisions or by termination shocks against
the surrounding media. Finally, we examine the effects of the deaths of high-mass stars that
remained close to their birthplaces, exploding as supernovae within the superbubbles that these
clusters created. We present new X-ray results on W51 IRS2E and 30 Doradus and we introduce
new data on Trumpler 14 in Carina and the W3 HII region complexes W3 Main and W3(OH).
Most stars are born in massive star-forming regions (MSFRs); the most massive stars
live out their short lives in this environment and eventually transform it when they
explode as supernovae. In the meantime, they have a profound influence on their natal
neighborhood, generating HII regions and wind-blown bubbles and often triggering new
generations of stars to form in the surrounding molecular clouds. The kinetic power of a
massive O-star’s winds injected into its stellar neighborhood over its lifetime equals that
input in its supernova explosion; essentially from the moment they are born high-mass
stars cause changes in their environment on parsec scales.
X-ray observations probe different energetic components of MSFRs than traditional
optical and IR studies. Stars of virtually all masses and stages emit X-rays in their
youth, although the mechanisms for X-ray emission vary with stellar mass. For OB
stars excavating an HII region within their nascent molecular cloud, diffuse X-rays may
be generated as fast winds shock the surrounding media (Weaver et al. 1977); we have
recently discovered such diffuse emission with X-ray observations of M 17 and the Rosette
Nebula (Townsley et al. 2003). X-ray studies also detect the presence of past supernovae
through the shocks in their extended remnants.
Chandra and its Advanced CCD Imaging Spectrometer (ACIS) camera give us the sen-
sitivity, spatial resolution, and broad bandpass to detect diffuse X-ray emission generated
by these high-mass stars and to separate it from the hundreds of pre-main sequence X-
ray-emitting stars seen in these fields. Chandra routinely penetrates heavy obscuration
(AV > 100 mag) with little source confusion or contamination from unrelated objects
to reveal the young stellar populations in MSFRs. Before the Chandra era, the relative
X-ray contributions of high-mass and low-mass stars, OB winds, and supernova remnant
shocks in these regions were largely unknown.
2Townsley et al.
Through the Chandra General Observer and Guaranteed Time programs, we are pur-
suing a multi-year study of MSFRs, cataloguing and characterizing the point source
populations (e.g. Getman et al. 2005) as well as searching for diffuse emission. Last
year we reviewed our Chandra observations of M17, RCW49, and W51A (Townsley et
al. 2004). We show other examples of our program in this contribution, presenting new
results for W3, W51A IRS2E, Trumpler 14 in Carina, and 30 Doradus in the LMC.
W3 is an obscured complex of high-mass stars, H II regions, and associated molecular
clouds situated 2.3 kpc from the Sun, part of a vast star formation complex also containing
the W4 superbubble, the massive stellar clusters IC 1805, IC 1795, and NGC 896, and
several unnamed IR clusters (Carpenter et al. 2000). It is bordered to the west by HB3,
a very large, evolved supernova remnant (SNR), which is clearly seen in the ROSAT
All-Sky Survey image and was observed with Einstein (Leahy et al. 1985). Radio studies
(Routledge et al. 1991) suggest that the SNR shock has not yet reached the W3 H II
regions, but it is influencing the distribution of CO in the W3 molecular cloud. Lada et al.
(1978) and Thronson et al. (1985) argued that star formation in W3 is being induced by
the expansion of W4, which is sweeping up molecular gas into a high-density layer, within
which stars are forming. Oey et al. (2005) propose that the young (3–5 Myr) OB cluster
IC 1795, triggered to form by W4, is blowing its own second-generation superbubble at
the molecular cloud interface, triggering in turn the W3 MSFR.
Figure 1. Left: An ACIS-I mosaic of W3 (red = 0.5–2 keV, green = 2–8 keV), including three
pointings on W3 Main totaling 78 ks and a single 72-ks pointing on W3(OH). Each ACIS-I
image is 17?×17?, or ∼ 11 pc on a side. Right: A zoomed image of W3 Main, showing the large
number of young, embedded stars revealed by this observation.
A 40-ksec Chandra observation of W3 Main using the ACIS imaging array (ACIS-
I) revealed the ionizing sources for many of its HII regions and over 200 point sources
(Hofner et al. 2002). We have recently obtained more Chandra/ACIS-I observations of
W3 Main and the adjacent field W3(OH) (Figure 1), showing a rich stellar population
around W3 Main, the older cluster IC 1795 (perhaps exhibiting soft diffuse emission), and
several small embedded clusters in and around W3(OH). The W3(OH) field is noticeably
lacking in sources compared to the W3 Main field. Although partly an obscuration effect,
this also illustrates the intrinsic difference in the size of these clusters; hard X-rays are
largely unaffected by the obscuring material so the 2–8 keV (green) component in Figure 1
reflects an intrinsic difference in cluster size between W3(OH) and W3 Main.
Parsec-scale X-ray Flows3
3. W51A and the Enigmatic Source IRS2E
W51 is one of the most massive star-forming complexes in the Galaxy but is difficult
to observe because of its distance (∼ 7 kpc) and high obscuration. Our 72-ksec Chan-
dra/ACIS observation of W51A (Townsley et al. 2004) detected many of the known radio
HII regions (Mehringer 1994) as diffuse X-ray sources. We also see ∼ 450 point sources,
revealing the highest-mass and youngest inhabitants fueling the HII regions and just
emerging from their dusty natal cocoons.
Buried in one of the youngest and richest high-mass complexes, G49.5-0.4, we have
discovered an enigmatic hard X-ray source at the center of an embedded high-mass
stellar cluster (Figure 2). CXOW51 J192340.1+143105 is spatially coincident with a
deeply-embedded mid-IR source (Kraemer et al. 2001) known as IRS2 East (IRS2E). It
is surrounded by powerful masers and ultra-compact HII regions yet has no associated
radio HII region itself. Evidence for infall is seen in this region (Sollins et al. 2004).
NH = 3x1023 cm-2
(AV = 150 mag)
kT = 18 keV
LX,corr (2.7--8 keV) = 2x1033 ergs/s
W51 IRS2E, 1 apec + gaussian at 6.5 keV
Figure 2. Left: A binned ACIS image of the W51 IRS2 complex, 8??× 12??, with a J2000
coordinate grid. The extraction region for IRS2E, containing 90% of the 1.5 keV point spread
function, is shown, as is the cometary HII region W51d. The locations of the UCHII region
W51d2 and the strong SiO maser are noted. Right: ACIS spectrum of W51 IRS2E: the upper
panel shows the source spectrum and model fit; the lower panel shows the fit residuals.
IRS2E emits most of its X-ray photons in a broad 6.5-keV line probably due to fluo-
rescent iron. Although common in AGN, this type of X-ray spectrum is highly unusual
for stellar sources, as it requires an embedded source with substantial emission above
the iron absorption edge (7.1 keV). Such a hard X-ray spectrum could be generated by
colliding winds in a massive cluster (Cant´ o et al. 2000), but the X-ray emission of IRS2E
varies by a factor of two in 40 ksec, ruling out the cluster explanation. We suspect that
it is a colliding wind binary, perhaps a younger version of Eta Carinae (Corcoran et al.
2004) or HD 5980, a luminous blue variable in the SMC (Naz´ e et al. 2002). The absence
of an HII region around this source suggests that it is very young.
4. Trumpler 14 in Carina
The Carina complex, at a distance of ∼2.8 kpc (Tapia et al. 2003), is a remarkably
rich star-forming region at the edge of a giant molecular cloud (GMC), containing 8 open
clusters with at least 64 O stars, 2 Wolf-Rayet stars, and the luminous blue variable Eta
Carinae (Feinstein 1995). ISO discovery of 22µm grains in the bright radio HII region
Carina I may imply that a supernova occurred in this region (Chan & Onaka 2000);
the presence of WR stars also may indicate past supernovae, although no well-defined
remnant has ever been seen.
Tr 14 is an extremely rich, young (∼1 My), compact OB cluster near the center of the
4 Townsley et al.
Carina complex, containing at least 30 O and early B stars (V´ azquez et al. 1996). Tr 14
is probably at the same distance as its neighboring, equally rich cluster Trumpler 16
but is thought to be younger (Walborn 1995). These two clusters contain the highest
concentration of O3 stars known in the Galaxy; their ionizing flux and winds may be
fueling a bipolar superbubble (Smith et al. 2000).
An Einstein X-ray study of the Carina star-forming complex was performed by Seward
& Chlebowski (1982). They detected ∼30 point sources, mostly individual high-mass
stars and the collective emission from unresolved cluster cores. They also detected diffuse
emission pervading the entire region and speculated that it may be due to O star winds.
Based on experience with Chandra, we now know that thousands of the lower-mass stars
in these young clusters were likely to be contributing to the diffuse flux seen in the
Einstein data. A major goal of our Chandra observation was to resolve out a significant
fraction of this point source emission so a better determination of the spatial and spectral
characteristics of the diffuse component can be made.
Figure 3. Left: A smoothed image (red = 0.5–2 keV, blue = 2–8 keV) of the 57-ksec ACIS
observation of Tr 14 in Carina, with the cluster imaged on the ACIS-I array (17?×17?, or ∼ 14 pc
on a side at D = 2.8 kpc) and bright diffuse emission seen in the off-axis S2 and S3 CCDs (each
8.5?×8.5?). We find ∼ 1600 point sources on the I array plus extensive diffuse emission across the
whole field. Right: Smoothed soft-band image (red = 0.5–0.7 keV, green = 0.7–1.1 keV, blue =
1.1–2.3 keV) of the 21-ksec ACIS observation of 30 Doradus (from Townsley et al. 2005a), with
30 Dor Main imaged on the I array (covering ∼ 250 pc on a side at D = 50 kpc); the large shell
30 Dor C, the Honeycomb SNR, and SN1987A are seen in the off-axis CCDs S3 and S4. Bright,
soft diffuse emission dominates the entire field.
The aimpoint of our 57-ksec ACIS observation of Tr 14 (Figure 3 left) was the central
star in the cluster, HD 93129AB, a very early-type (O2I–O3.5V) binary (Walborn et al.
2002), with the two components separated by ∼ 1??. Since these are resolved in the ACIS
data, we can see that the two components have very different spectra; HD 93129B shows
a typical O-star X-ray spectrum (kT = 0.5 keV, or T ∼ 6 MK), while HD 93129A shows
a similar soft component (kT = 0.6 keV) but also exhibits a much harder component,
with kT = 3.0 keV (T ∼ 35 MK), and is ten times brighter in X-rays than HD 93129B.
This hard spectrum and high X-ray luminosity are indicative of a colliding-wind binary
(Portegies Zwart et al. 2002); in fact HD 93129A was recently discovered to be a spectro-
scopic binary (Nelan et al. 2004). Additionally, while the O3V star HD 93128 is soft and
fainter in X-rays, we find that the O3V star HD 93250 in Tr 14 shows a two-component
spectrum and X-ray luminosity almost identical to HD 93129A. XMM-Newton observa-
tions also show a hard spectral component for HD 93250 (Albacete Colombo et al. 2003);
this source is very likely a colliding-wind binary as well.
The diffuse emission in Tr 14 is quite soft and shows abundances typical of OB wind
termination shocks. We also see soft, bright diffuse emission in the off-axis CCDs of the
Parsec-scale X-ray Flows5
ACIS array, far from any of the Carina massive stellar clusters. Spectral fits to this diffuse
emission require abundances of O, Ne, Si, and Fe to be more than twice the solar value;
this is evidence that the emission may be from an old “cavity” supernova remnant that
exploded inside the Carina superbubble, as suggested by Chu et al. (1993).
5. 30 Doradus
Early in the Chandra mission, we obtained a ∼ 21 ksec observation of the most lumi-
nous Giant Extragalactic HII Region and “starburst cluster” in the Local Group, 30 Do-
radus in the Large Magellanic Cloud. The ACIS pointing was centered on the young,
dense OB cluster R136, a testbed for understanding recent and ongoing star formation
in the 30 Dor complex. The presence of evolved supergiants ∼ 25 Myr old and embedded
massive protostars shows that 30 Dor is the product of multiple epochs of star formation,
including a new generation of embedded stars currently forming, possibly as a result of
triggered collapse from the effects of R136 (Brandner et al. 2001). Supernovae pervade
the region but may go undetected due to age and environment (Chu & Mac Low 1990).
Nearby are two msec pulsars and SN1987A. 30 Dor produced at least five plasma-filled
superbubbles with ∼100-pc scales (Wang & Helfand 1991), likely products of strong OB
winds and multiple supernovae.
The right panel of Figure 3 shows a smoothed ACIS image from our 21-ksec observation
of 30 Dor, including the off-axis CCDs S3 and S4 as well as the main 30 Dor nebula on
the ACIS-I array. We see a bright concentration of X-rays associated with the R136 star
cluster, the bright SNR N157B to the southwest, a number of new widely-distributed
compact X-ray sources, and diffuse structures associated with the superbubbles produced
by the collective effects of massive stellar winds and their past supernova events (Townsley
et al. 2005a). Some of these are center-filled while others are edge-brightened, indicating
a complicated mix of viewing angles and perhaps filling factors.
Our spectral analysis of the superbubbles reveals a range of absorptions (NH = 1–
6×1021cm−2), plasma temperatures (T = 3–9×106K), and abundance variations. We
find ∼ 100 sources associated with the central massive cluster R136 (Townsley et al.
2005b); some bright, hard X-ray point sources in the field are likely colliding-wind binaries
(Portegies Zwart et al. 2002). Comparing the X-ray data to visual and IR images, we
find that hot plasma fills the shells outlined by ionized gas and warm dust.
Interactions between powerful O star winds and the ISM lead to parsec-scale soft X-ray
emission as predicted by Weaver et al. (1977) and others, but with much fainter X-ray
luminosities; this hot plasma may pervade the Galactic plane but is hard to detect due to
obscuration. Wind-wind interactions lead to harder X-rays; this emission may provide a
way to determine close binarity in massive stars or to detect embedded massive clusters.
The 104K Str¨ omgren Sphere that defines classical HII regions is really a Str¨ omgren Shell
filled with 106K plasma in many MSFRs. Only a small portion of the wind energy and
mass appears in the observed diffuse X-ray plasma, though; it could be dissipated via
turbulence, mass-loading, or fissures into the ISM (Townsley et al. 2003). We see bright,
soft diffuse X-rays in some regions; enhanced metallicity and luminosity compared to
wind-generated emission and the presence of these structures in regions that also contain
evolved stars implies that these X-ray features are the remains of cavity SNRs.
Chandra chronicles the life cycle of massive stars through studies of MSFRs. In W3
and W51A we see massive embedded protostars; winds from main sequence O stars fill