0705.0702/ms.tex
1: \pdfoutput=1
2: \documentclass{emulateapj}
3: \pdfoutput=1
4: 
5: \usepackage{color}
6: 
7: \newcommand{\av}{\ensuremath{A_{V}}}
8: \newcommand{\kt}{\ensuremath{k_{\rm{B}}T}}
9: \newcommand{\lx}{\ensuremath{L_{\rm{X}}}}
10: \newcommand{\lbol}{\ensuremath{L_{\rm{bol}}}}
11: \newcommand{\emm}{\ensuremath{EM}}
12: \newcommand{\fx}{\ensuremath{F_{\rm{X}}}}
13: \newcommand{\nh}{\ensuremath{N_{\rm H}}}
14: \newcommand{\ri}{\textit{R}}
15: \newcommand{\ji}{\textit{J}}
16: \newcommand{\hi}{\textit{H}}
17: \newcommand{\ki}{\textit{K}}
18: \newcommand{\ksi}{\textit{K\ensuremath{_{\rm{s}}}}}
19: 
20: \slugcomment{}
21: \shorttitle{High-Resolution X-ray Imaging of RCW\,49}
22: \shortauthors{Tsujimoto et al.}
23: 
24: \begin{document}
25: \title{An X-ray Imaging Study of the Stellar Population in RCW\,49}
26: \author{M.~Tsujimoto\altaffilmark{1,2,3}, E.~D.~Feigelson\altaffilmark{1},
27: L.~K.~Townsley\altaffilmark{1}, P.~S.~Broos\altaffilmark{1},
28: K.~V.~Getman\altaffilmark{1}, J.~Wang\altaffilmark{1}, G.~P.~Garmire\altaffilmark{1},
29: D.~Baba\altaffilmark{4}, T.~Nagayama\altaffilmark{5}, M.~Tamura\altaffilmark{6}, \&
30: E.~B.~Churchwell\altaffilmark{7}}
31: 
32: \altaffiltext{1}{Department of Astronomy \& Astrophysics, Pennsylvania State University,
33: 525 Davey Laboratory, University Park, PA 16802.}
34: \altaffiltext{2}{Department of Physics, Rikkyo University, 3-34-1, Nishi-Ikebukuro,
35: Toshima, Tokyo 171-8501, Japan.}
36: \altaffiltext{3}{Chandra Fellow. E-mail: \texttt{tsujimot@astro.psu.edu}.}
37: \altaffiltext{4}{Department of Physics, Graduate School of Science, Nagoya University,
38: Furo-cho, Chikusa, Nagoya, 464-8602 Japan.}
39: \altaffiltext{5}{Department of Astronomy, Graduate School of Science, Kyoto University,
40: Kitashirakawa-Oiwake-cho, Sakyo, Kyoto, 606-8502 Japan.}
41: \altaffiltext{6}{National Astronomical Observatory of Japan, 2-21-1, Osawa, Mitaka,
42: Tokyo 181-8588, Japan.}
43: \altaffiltext{7}{Department of Astronomy, University of Wisconsin, 475 North Charter
44: Street, Madison, WI 53706.}
45: 
46: 
47: \begin{abstract}
48:  We present the results of a high-resolution X-ray imaging study of the stellar
49:  population in the Galactic massive star-forming region RCW\,49 and its central OB
50:  association Westerlund 2. We obtained a $\sim$40~ks X-ray image of a
51:  $\sim$17\arcmin$\times$17\arcmin\ field using the \textit{Chandra X-ray Observatory}
52:  and deep near-infrared (NIR) images using the Infrared Survey Facility in a concentric
53:  $\sim$8\farcm3$\times$8\farcm3 region. We detected 468 X-ray sources and identified
54:  optical, NIR, and \textit{Spitzer Space Telescope} mid-infrared (MIR) counterparts for
55:  379 of them. The unprecedented spatial resolution and sensitivity of the X-ray image,
56:  enhanced by optical and infrared imaging data, yielded the following results: (1) The
57:  central OB association Westerlund 2 is resolved for the first time in the X-ray
58:  band. X-ray emission is detected from all spectroscopically-identified early-type stars
59:  in this region. (2) Most ($\sim$86\%) X-ray sources with optical or infrared
60:  identifications are cluster members in comparison with a control field in the Galactic
61:  Plane. (3) A loose constraint (2--5~kpc) for the distance to RCW\,49 is derived from
62:  the mean X-ray luminosity of T Tauri stars. (4) The cluster X-ray population consists
63:  of low-mass pre--main-sequence and early-type stars as obtained from X-ray and NIR
64:  photometry. About 30 new OB star candidates are identified. (5) We estimate a cluster
65:  radius of 6\arcmin\--7\arcmin\ based on the X-ray surface number density profiles. (6)
66:  A large fraction ($\sim$90\%) of cluster members are identified individually using
67:  complimentary X-ray and MIR excess emission. (7) The brightest five X-ray sources, two
68:  Wolf-Rayet stars and three O stars, have hard thermal spectra.
69: \end{abstract}
70: 
71: \keywords{X-rays: stars --- infrared: stars --- ISM: \ion{H}{2} regions: individual (RCW\,49)
72: --- open clusters and associations: individual (Westerlund 2) --- stars: Wolf-Rayet ---
73: stars: pre--main-sequence}
74: 
75: 
76: \section{INTRODUCTION}
77: \ion{H}{2} regions, which are recognized by diffuse hydrogen emission ionized by UV photons
78: from stars earlier than B2, are the sites of massive star formation. Strong radiation
79: and winds from early-type stars in massive star-forming regions (SFRs) have effects that
80: cannot be understood by a simple extrapolation of low-mass SFRs. X-ray studies of
81: massive SFRs, which are potentially important to detect hot gas as well as stellar
82: constituents, had lagged behind those of low-mass regions in the past due to their
83: greater distance, source confusion, and obscuration. Both high spatial resolution and
84: hard X-ray sensitivity are required to overcome these difficulties.
85: 
86: The \textit{Chandra X-ray Observatory} is capable of conducting high-resolution and
87: high-sensitivity imaging observations in the 0.5--8.0~keV band. Its $\sim$1\arcsec\
88: spatial resolution and sensitivity are sufficient to resolve scales $\sim$10$^{17}$~cm
89: (a typical size of ultra-compact \ion{H}{2} regions; \citealt{churchwell02}) and to detect
90: X-ray emission from embedded late-type pre--main-sequence sources at a distance of
91: several kiloparsecs, typical of well-studied Galactic massive SFRs.
92: 
93: Along with the \textit{Chandra} Orion Ultradeep Project \citep{getman05}, which led to a
94: series of X-ray imaging studies of the nearest ($D\lesssim$~0.5~kpc) rich SFR, we have seen a
95: surge of results from X-ray studies of massive SFRs at larger distances. Galactic
96: massive SFRs with X-ray imaging surveys of their stellar contents include: NGC\,3603
97: \citep{moffat02}, W\,3 \citep{hofner02}, the Carina nebula
98: \citep{evans03,evans04,sanchawala06}, NGC\,6530 \citep{damiani04}, the Trifid Nebula
99: \citep{rho04}, S\,106 \citep{giardino04}, Ara OB1 \citep{skinner05}, Cepheus B
100: \citep{getman06}, RCW\,38 \citep{wolk06}, the Arches and Quintuplet clusters
101: \citep{dqwang06}, NGC\,2362 \citep{delgado06}, NGC\,2264 \citep{flaccomio06}, Westerlund
102: 1 \citep{skinner06}, IC\,1396N \citep{getman07a}, NGC\,6357 \citep{jwang07a}, the
103: Rosette nebula \citep{townsley03,jwang07b}, M\,16 \citep{linsky07}, Cygnus OB2
104: \citep{albacete07}, and M\,17 \citep{townsley03,broos07}. Early results are reviewed in
105: \citet{townsley06b} and \citet{feigelson07}.
106: 
107: \medskip
108: 
109: This paper describes the results of the \textit{Chandra} imaging study of the stellar
110: contents in a southern massive SFR RCW\,49 and its central OB association Westerlund
111: 2. The region was imaged in the mid-infrared (MIR) band as a part of the Galactic Legacy
112: Infrared Mid-Plane Survey Extraordinaire program (GLIMPSE; \citealt{benjamin03}) using
113: the \textit{Spitzer Space Telescope} \citep{werner04}. We also obtained deep
114: near-infrared (NIR) images of this region using the Infrared Survey Facility. Many
115: previous \textit{Chandra} studies were combined with NIR imaging data, but none was
116: presented with MIR imaging data, which is a growing tool for the stellar census studies
117: in massive SFRs with the launch of \textit{Spitzer}. We demonstrate the ability of
118: high-resolution X-ray imaging to identify and understand the nature of the stellar
119: population in massive SFRs with a particular emphasis on the combination with NIR and
120: MIR imaging.
121: 
122: The outline of this paper is as follows. In \S~\ref{sect:s2}, we give a brief review of
123: RCW\,49 and Westerlund 2. In \S~\ref{sect:s3}, we present our X-ray and NIR observations
124: of this region. In \S~\ref{sect:s4}, we describe our X-ray and NIR data reduction and
125: the identification of X-ray sources with optical, NIR, and MIR sources using our deep
126: NIR data as well as archived data. In the data reduction, special care is taken in
127: source extraction, photometry, and cross-correlation in crowded regions. In total 468
128: X-ray sources are detected and 379 of them are identified by optical and infrared
129: sources.
130: 
131: In \S~\ref{sect:s5}, we discuss how the X-ray data are relevant to discriminate
132: intrinsic RCW\,49 members, identify new sources, and infer their nature. Seven
133: multi-faceted and loosely-related results are discussed in each subsection. In the first
134: subsection (\S~\ref{sect:s5-1}), we show that the central OB association is resolved in
135: the X-ray image and all the massive members in this region are detected in the X-ray. In
136: \S~\ref{sect:s5-2}, we argue that most of the X-ray sources with optical or infrared
137: counterparts are intrinsic to RCW\,49 by a comparison study with the X-ray population of
138: a control field in the Galactic Plane. In \S~\ref{sect:s5-3}, we obtain a loose
139: constraint on the distance to RCW\,49 using the mean X-ray luminosity of T Tauri stars
140: in the mass range of 2.0--2.7~$M_{\odot}$. In \S~\ref{sect:s5-4}, we evaluate the nature
141: of X-ray sources using X-ray and infrared photometric data. We employ the X-ray versus
142: NIR flux and luminosity plots in addition to the conventional NIR color-magnitude
143: diagram. We argue that the majority of X-ray sources are low-mass pre--main-sequence
144: sources and early-type main sequence stars. Some early-type stars are found to be harder
145: and more luminous in the X-ray than others with a similar NIR brightness. In
146: \S~\ref{sect:s5-5}, we study the spatial distribution of X-ray-identified cluster
147: members. Because of the low level of contamination of X-ray samples, we can identify
148: cluster members far from the central OB association. We find a possible excess
149: population at $\sim$4\arcmin\ and the cluster boundary at 6\arcmin--7\arcmin\ from the
150: cluster center. In \S~\ref{sect:s5-6}, we examine the fraction of RCW\,49 members
151: individually selected using X-ray emission and find that about half of the RCW\,49
152: sources brighter than \ksi$\sim$14~mag are identified by X-rays. The rate increases to
153: $\sim$90\% when combined with MIR excess emission, another indicator of cluster
154: membership complementary to the X-ray emission signature. In the last subsection
155: (\S~\ref{sect:s5-7}), we discuss the X-ray spectra of the five brightest X-ray
156: sources. We summarize these results in \S~\ref{sect:s6}.
157: 
158: 
159: \section{RCW\,49 and Westerlund 2}\label{sect:s2}
160: RCW\,49 was identified as an optically-visible \ion{H}{2} region by \citet{rogers60}. It
161: is also known as Kes\,10 \citep{kesteven68}, G\,284.3--00.3, and Gum\,29
162: \citep{gum55}. The visual extinction toward the region is \av\,$=$\,5--6~mag
163: \citep{carraro04,rauw07,ascenso07}. The stellar constituents in the region were studied
164: by imaging studies in the optical \citep{moffat75,moffat91,piatti98,carraro04,rauw07},
165: infrared \citep{churchwell04,whitney04,uzpen05,ascenso07}, and X-ray
166: \citep{goldwurm87,belloni94} bands. At least a dozen OB stars comprise the central OB
167: association called Westerlund 2 (Wd\,2; \citealt{westerlund60}). The association is also
168: known as NGC\,3247 \citep{dreyer88}, ESO\,127-18, and C\,1024-576.
169: 
170: \citet{moffat91} identified six O6--7 stars based on low-resolution optical
171: spectroscopy. \citet{rauw07} added six more O-type stars and gave a consistent
172: classification based on medium-resolution optical spectroscopy. These stars were
173: reclassified toward earlier spectral classes ranging between O3 and O6.5 \citep{rauw07}.
174: 
175: Three more early-type stars are found beyond the cluster core. One is a star with a
176: spectral type of O4\,V(f) or O5\,V(f) \citep{uzpen05,rauw07}, which we refer to as
177: MSP\,18 \citep{moffat91}. Another is a binary Wolf-Rayet star WR\,20a (WN6$+$WN6;
178: \citealt{vanderhucht01,rauw05}), which is the most massive binary in the Galaxy with a
179: well-determined mass \citep{rauw05}. Another Wolf-Rayet star WR\,20b
180: \citep{vanderhucht01} lies $\sim$3\farcm7 from Wd\,2 \citep{shara91}. These early-type
181: stars in and around Wd\,2 are probably the cause of exotic high energy phenomena
182: discovered in this region, such as extended hard X-ray emission \citep{townsley04} found
183: in the same \textit{Chandra} data set presented in this paper and TeV gamma-ray emission
184: \citep{aharonian07} found by the High Energy Stereoscopic System telescope.
185: 
186: A much larger number of lower-mass members are anticipated, many of which are at the
187: pre--main-sequence stage that can be identified either by infrared excess or X-ray
188: emission, arising respectively from circumstellar disks and envelopes and from enhanced
189: coronal magnetic activity. Some of the low-mass members were detected by mid-infrared
190: excess emission using the \textit{Spitzer} data \citep{whitney04}. In the X-ray band,
191: the two previous studies did not have sufficient spatial resolution and sensitivity to
192: distinguish individual stars. In these studies, respectively using \textit{Einstein} and
193: \textit{ROSAT}, only one \citep{goldwurm87} and seven \citep{belloni94} X-ray sources
194: were detected in the \textit{Chandra} studied field presented here. Most of these
195: sources are isolated early-type stars and unresolved stars in Wd\,2 (1E 1022.1--5730 in
196: \citealt{goldwurm87} and source number 24 in \citealt{belloni94}).
197: 
198: The ionized gas content at $\sim$10$^{4}$~K in RCW\,49 was studied by a radio continuum
199: imaging study at 0.843, 1.38, and 2.38~GHz \citep{whiteoak97}. Two shells appear in the
200: images, one around the complex of Wd\,2 and WR\,20a with a diameter of $\sim$7\farcm3
201: (radio ring A) and the other around WR\,20b with a diameter of $\sim$4\farcm1 (radio
202: ring B). Enhanced emission was found at the position where these two shells overlap. A
203: similar global structure was found in the dust content. The MIR image taken by the
204: Infrared Array Camera (IRAC) onboard \textit{Spitzer} traces the dust distribution via
205: polycyclic aromatic hydrocarbon emission at its three ([3.6], [5.8], and [8.0] $\mu$m)
206: of four bands \citep{churchwell04}. Moreover, the \textit{Spitzer} image reveals a
207: network of highly structured filaments, pillars, and shocks at arcsecond scales, some of
208: which are suggestive of instabilities.
209: 
210: Both gas and dust distributions are dominated by the massive stars in Wd\,2 and the two
211: Wolf-Rayet stars via their strong stellar winds and UV radiation that sculpt the
212: surrounding interstellar medium. These effects not only suppress further star formation,
213: but also trigger the second generation of star formation. From the spatial distribution
214: of MIR excess sources, \citet{whitney04} argued that star formation is taking place
215: preferentially in $\sim$4\arcmin\ around Wd\,2. The radius corresponds to the
216: outer boundary of the radio ring A where swept-up material is accumulated, suggesting
217: that these young sources are a consequence of induced star formation.
218: 
219: \medskip
220: 
221: The distance to RCW\,49 is highly controversial, ranging from 2 to 8~kpc in the
222: literature (\citealt{churchwell04} and references therein). The kinematic distance
223: determination gives a poor constraint because the region is in the tangential direction
224: of the Saggitarius-Carina arm. The debate on the distance continues in two most recent
225: papers; \citet{rauw07} argued for 8.0$\pm$1.4~kpc based on optical spectro-photometric
226: study of bright stars in Wd\,2, while \citet{ascenso07} proposed 2.8~kpc based on NIR
227: magnitudes and colors of RCW\,49 sources on the Henyey track. With the present X-ray
228: data, we can also loosely constrain the distance using the mass-stratified mean X-ray
229: luminosity (\S~\ref{sect:s5-3}). This gives an estimate of 2--5~kpc. Despite intensive
230: study, the distance to this cluster remains an open question. We adopt 4.2~kpc as used
231: by \citet{churchwell04}, which is roughly a geometric mean of the smallest and the
232: largest distance estimates, to avoid the worst case when the question is settled.
233: 
234: The association of Wolf-Rayet stars and an O4--O5\,V(f) star with Wd\,2 suggests that
235: the star cluster is no older than a few Myrs. No cluster member appears to be located
236: away from the zero-age--main-sequence isochrone curve in the optical color-magnitude
237: diagram, from which \citet{carraro04} estimated an age of $\lesssim$
238: 2~Myr. \citet{ascenso07} similarly derived the cluster age of $\sim$2~Myr by fitting the
239: color-magnitude distribution of NIR sources with theoretical calculations.
240: 
241: 
242: \section{OBSERVATIONS}\label{sect:s3}
243: We carried out an X-ray observation of RCW\,49 using the Advanced CCD Imaging
244: Spectrometer (ACIS; \citealt{garmire03}) onboard the \textit{Chandra X-ray Observatory}
245: \citep{weisskopf02} from 2003 August 23 UT 18:20 to August 24 UT 4:54. Four imaging
246: array (ACIS-I) chips covered a 17\arcmin$\times$17\arcmin\ field centered at (R.\,A.,
247: decl.)$=$(10$^{\rm h}$24$^{\rm m}$00\fs5, --57\arcdeg45\arcmin18\arcsec) in the equinox
248: J2000.0 for a 36.7~ks exposure. ACIS-I covers the 0.5--8.0~keV energy band with a
249: spectral resolution of $\sim$150~eV at 6~keV and a point spread function (PSF) radius of
250: $\sim$0\farcs5 within $\sim$2\arcmin\ of the on-axis position, degrading to
251: $\sim$6\arcsec\ at a 10\arcmin\ off-axis angle. The data were taken with the very faint
252: telemetry mode and the timed exposure CCD operation with a frame time of 3.2~s.
253: 
254: We conducted near-infrared (NIR) observations on 2004 December 25 and 28 using the
255: Simultaneous three-color InfraRed Imager for Unbiased Surveys (SIRIUS;
256: \citealt{nagayama03}) mounted on the Cassegrain focus of the Infrared Survey Facility
257: (IRSF) 1.4~m telescope at the South African Astronomical Observatory. SIRIUS is a NIR
258: imager capable of obtaining simultaneous images in the \ji, \hi, and \ksi\ bands using
259: two dichroic mirrors. The instrument is equipped with three HAWAII arrays of
260: 1024$\times$1024 pixels. The pixel scale of 0\farcs45 is an excellent match with the
261: on-axis spatial resolution of \textit{Chandra}. With ten-point dithering, we covered
262: 8\farcm3$\times$8\farcm3 fields at two positions, one aimed at RCW\,49 (10$^{\rm
263: h}$24$^{\rm m}$01\fs9, --57\arcdeg45\arcmin31\arcsec) and the other at a control region
264: (10$^{\rm h}$27$^{\rm m}$13\fs5, --58\arcdeg01\arcmin26\arcsec) free from very bright
265: sources displaced by $\sim$30\arcmin\ from RCW\,49. Both fields have the same Galactic
266: latitude.
267: 
268: Each frame was exposed for 30~s and the total exposure time was $\sim$30~minutes equally
269: for the two regions. Additional short (5~s) exposure frames were also taken for the two
270: regions to extend the dynamic range of the data set to brighter sources. The observing
271: conditions were photometric with a seeing of $\lesssim$ 1\farcs0 on both nights.
272: 
273: Figures~\ref{fg:f1} (a) and (b) respectively show the ACIS and SIRIUS images of the study
274: field. The SIRIUS image has a wider coverage by $\sim$2.9 times than the image
275: presented in \citet{ascenso07}.
276: 
277: \begin{figure*}[hbtp]
278:  \figurenum{1}
279:  \epsscale{0.5}
280:  \plotone{f1a.jpg}
281:  \plotone{f1b.jpg}
282:  \caption{(a) Smoothed X-ray image. Red, green, and blue indicate the intensity in the
283:  0.5--1.7, 1.7--3.8, and 3.8--8.0~keV bands, respectively. The fields of view of the
284:  four ACIS chips and SIRIUS are shown by solid and dashed squares. Free-style lines
285:  taken from \citet{churchwell04} indicate the global features in the radio image
286:  \citep{whiteoak97}. (b) NIR image. Red, green, and blue indicate the intensity in the
287:  \ksi, \hi, and \ji\ bands, respectively. The positions of some landmark sources are
288:  given in small squares in both images and labeled in panel (a).}\label{fg:f1}
289: \end{figure*}
290: 
291: 
292: 
293: \section{DATA REDUCTION AND ANALYSIS}\label{sect:s4}
294: \subsection{X-ray Data}
295: \subsubsection{Reduction \& Source Extraction}\label{sect:s4-1-1}
296: We reprocessed the raw data distributed by the \textit{Chandra} X-ray Center to obtain
297: an X-ray event list. A background rejection algorithm specific to data taken with the
298: very faint mode was applied. Events were further cleaned by removing cosmic-ray
299: afterglow and applying filters based on the event grades, status, and good-time
300: intervals \citep{townsley03}. Charge transfer inefficiency was corrected
301: \citep{townsley02} and the positions of all events were improved using a sub-pixel
302: repositioning technique \citep{tsunemi01}.
303: 
304: We detected sources using the wavdetect algorithm in the CIAO package\footnote{See
305: http://asc.harvard.edu/ciao/ for detail.} independently for the soft (0.5--2.0~keV),
306: hard (2.0--8.0~keV), and full (0.5--8.0~keV) band images. The software failed to detect
307: sources in crowded regions, where we manually added candidate sources with the aid of
308: smoothed images in different kernel sizes and PSF-deconvolved images obtained using a
309: maximum likelihood technique \citep{lucy74,townsley06}. In total, 556 X-ray source
310: candidates were identified including 60 manually added sources. We took a stance to add
311: sources manually in a liberal manner here, and we examined their validity in a rigorous
312: manner later.
313: 
314: For all the candidate sources, we used ACIS Extract\footnote{See the ACIS Extract User's
315: Guide at http://www.astro.psu.edu/xray/docs/TARA/ae\_users\_guide.html for detail. A
316: complete description of the procedure using the ACIS Extract package can be found in
317: \citet{getman05}.} version 3.107 for automated systematic source and background event
318: extractions. The source positions were refined by correlating the event distribution
319: with the PSF or by using the centroid of the events. Signals were then extracted from a
320: 90\% encircled energy polygon of 1.5~keV X-rays to derive source counts in 0.5--8.0~keV
321: ($C_{\rm{src}}$). When two extraction regions overlap, the regions were reduced. A
322: masked dataset was created by removing the events in a circle with a $\sim$99\%
323: encircled energy radius around all sources, from which background events were extracted
324: locally around each source to derive their background and net counts, $C_{\rm{bkg}}$ and
325: $C_{\rm{net}} = C_{\rm{src}} - C_{\rm{bkg}} \times A_{\rm{src}}/A_{\rm{bkg}}$, where
326: $A_{\rm{src}}$ and $A_{\rm{bkg}}$ are the integrals of the masked exposure map values
327: within the source and background extraction regions.
328: 
329: Each source candidate was tested for validity using the photometric significance (PS)
330: and probability of no source ($P_{\rm{B}}$) statistics extracted by ACIS Extract. PS is
331: the commonly used photometric signal-to-noise ratio given by
332: \begin{equation}
333:  \mathrm{PS} = \frac{C_{\mathrm{net}}}{\Delta C_{\mathrm{net}}},
334: \end{equation}
335: where the approximation by \citet{gehrels86} was employed to calculate the uncertainty
336: $\Delta C_{\rm{net}}$. $P_{\rm{B}}$ is the probability for the null hypothesis that all
337: detected counts or more are explained by fluctuations of the background and is defined
338: by
339: \begin{equation}
340:  P_{\mathrm{B}} = 1 - \sum\limits_{n=0}^{C_{\mathrm{src}}-1} P(C_{\mathrm{bkg}} \times A_{\mathrm{src}}/A_{\mathrm{bkg}}, n),
341: \end{equation}
342: where $P(\lambda,n)$ indicates a Poisson probability distribution function of a mean
343: $\lambda$ to have $n$ counts.
344: 
345: We recognize 468 source candidates with $\mathrm{PS}$~$\ge$~1.0 and
346: $P_{\rm{B}}$~$\le$~1.0$\times$10$^{-2}$ to be valid X-ray sources
347: (Table~\ref{tb:t1}). None of them is bright enough to cause pile-up. Almost all the
348: sources are newly detected here.
349: 
350: 
351: \subsubsection{Photometry \& Time Variability}
352: For all valid sources, ACIS Extract was used for recalculating their source and
353: background counts, producing instrumental responses, and constructing spectra and light
354: curves for systematic photometry and spectroscopy analyses. The photometry results are
355: compiled in Table~\ref{tb:t1}. Columns (1) and (2) are the sequential number and the
356: source name. Columns (3)--(6) give position information (R.\,A. and decl.\, in the
357: equinox J2000.0, their uncertainty, and the off-axis angle). Columns (7)--(9) show
358: counts in 0.5--8.0~keV; $C_{\rm{net}}$, $\Delta C_{\rm{net}}$, and
359: $C^{\prime}_{\rm{bkg}} = C_{\rm{bkg}} \times A_{\rm{src}}/A_{\rm{bkg}}$, whereas column
360: (10) shows the net counts in 2.0--8.0~keV ($C_{\rm{net,hard}}$). Column (11) indicates
361: the PSF fraction of source extraction, in which smaller numbers than the default value
362: ($\sim$0.9) indicate that the extraction region was reduced to avoid
363: overlapping. Columns (12) and (13) are for PS and $P_{\rm{B}}$. Columns (14) and (15)
364: are sets of flags to indicate extraction anomalies and source variability. Column (16)
365: gives the effective exposure time, the time needed to accumulate the observed counts if
366: the source were at the aim point. Column (17) lists the median energy of the extracted
367: X-ray events corrected for the background. Median energy is proposed to be a better
368: quantity to characterize X-ray spectral hardness and to give a reliable estimate of
369: extinction than the conventionally-used hardness ratio
370: \citep{hong04,feigelson05}. Column (18) is the photometric flux discussed in
371: \S~\ref{sect:s4-1-3}.
372: 
373: 
374: X-ray variability was examined for 377 bright sources off the chip gaps using a
375: Kolmogorov-Smirnov test. The null hypothesis probability ($P_{\rm{KS}}$) that the flux
376: is constant was tested and sources were placed into three classes; (a) no evidence for
377: variability ($P_{\rm{KS}} > 5 \times 10^{-2}$), (b) possibly variable ($5 \times 10^{-3}
378: < P_{\rm{KS}} \le 5 \times 10^{-2}$), and (c) definitely variable ($P_{\rm{KS}} \le 5
379: \times 10^{-3}$). Seven sources are in the ``definitely variable'' class, showing
380: flare-like light curves.
381: 
382: \subsubsection{Spectroscopy \& Flux Estimates of Faint Sources}\label{sect:s4-1-3}
383: Table~\ref{tb:t2} gives the results from the X-ray spectral fits obtained with ACIS
384: Extract, where 230 spectra with PS $\ge$ 2 were fitted by an optically-thin thermal
385: plasma (APEC) model \citep{smith01} with interstellar absorption. The best-fit values of
386: plasma temperature (\kt), the volume emission measure (\emm), and the extinction column
387: density (\nh) were derived. The photometric and spectroscopic tables (Tables~\ref{tb:t1}
388: and \ref{tb:t2}) are in a format similar to those found in associated papers
389: \citep{getman06,townsley06,getman07a,jwang07a,broos07,jwang07b}.
390: 
391: The X-ray flux can be derived by the spectral fits for most of the sources in
392: Table~\ref{tb:t2}. For the remaining sources, which are mostly faint, we estimated their
393: X-ray flux using broad-band photometry: the mean energy times the net count rate divided
394: by the mean effective area \citep{tsujimoto05}. The consistency between the two flux
395: estimates was checked for the brightest 173 sources (Fig.~\ref{fg:f2}), for which both
396: values are available. They are overall in good agreement with a systematic offset
397: (20\%). We hereafter use the spectroscopic flux for the 173 bright sources and the
398: photometric flux corrected for the offset for the remaining sources.
399: 
400: \begin{figure}[hbtp]
401:  \figurenum{2}
402:  \epsscale{1.0}
403:  \plotone{f2.jpg}
404:  \caption{Consistency between the two X-ray flux estimates in the 0.5--8.0~keV band. The
405:  values derived from spectral fits (abscissa) are plotted against those from broad-band
406:  photometry (ordinate) for 173 sources with successful thermal fits in
407:  Table~\ref{tb:t2}. The dashed-and-dotted line indicates $y=x$, while the dashed line
408:  indicates the best-fit model of $y=1.20x$.}\label{fg:f2}
409: \end{figure}
410: 
411: \subsection{NIR Data}
412: All SIRIUS frames were reduced following standard procedures using IRAF, including
413: dark-current correction, sky subtraction, and flat-fielding. Dithered images were
414: registered to construct combined images in the three bands (\ji, \hi, and \ksi). Sources
415: were extracted and their photometry was measured using the daofind and daophot tasks in
416: IRAF.
417: 
418: In order to cope with source crowding in the SIRIUS images, we used the following
419: procedure \citep{stetson87}. First, we constructed the average PSF from dozens of bright
420: sources across each image. The profile of each extracted source was fitted with the
421: average PSF and the source was masked from the image if the fit was successful. The
422: procedure was repeated until most sources were masked. We then reconstructed the PSF
423: from the masked image, which we consider is freer from contamination than the original
424: one. By repeating the procedure, we separated closely spaced sources and deblended their
425: overlapping signals.
426: 
427: We extracted 10,540 and 9,768 sources in RCW\,49 and the control fields, respectively,
428: at levels larger than 3 $\sigma$ of the sky background noise in all three bands. We
429: derived their positions and magnitudes in the \ji, \hi, and \ksi\ bands in the
430: longer-exposure images. We replaced magnitudes with those derived from the
431: shorter-exposure images when they were saturated at levels brighter than 11.0 (\ji),
432: 11.5 (\hi), and 12.0 (\ksi) mag. The 3 $\sigma$ detection limits of the longer-exposure
433: images in the RCW\,49 field are $\sim$20.5 (\ji), $\sim$19.8 (\hi), and $\sim$18.4
434: (\ksi) mag, whereas the 10 $\sigma$ limits are $\sim$19.0 (\ji), $\sim$17.8 (\hi), and
435: $\sim$16.8 (\ksi) mag. These limits were $\sim$0.5~mag shallower than those measured in
436: the control field because of the larger source confusion. All the detected sources are
437: used for the X-ray counterpart search, while only those with magnitude uncertainty less
438: than 0.1~mag are used for photometric analysis.
439: 
440: 
441: \subsection{Optical and Infrared Identifications of X-ray Sources}
442: We identified the X-ray sources in the optical, NIR, and MIR bands using the Naval
443: Observatory Merged Astrometric Dataset (NOMAD)\footnote{See
444: http://www.nofs.navy.mil/nomad/ for details.}, the Two-Micron All-Sky Survey (2MASS;
445: \citealt{skrutskie97}), SIRIUS, and GLIMPSE. GLIMPSE is a Legacy Program of the
446: \textit{Spitzer Space Telescope} to survey a $\sim$220 deg$^{2}$ region of the Galactic
447: Plane at four MIR bands ([3.6], [4.5], [5.8], and [8.0] $\mu$m) with a
448: 1\farcs6--1\farcs9 resolution using IRAC \citep{fazio04}. A 1\fdg7$\times$0\fdg7 region
449: encompassing RCW\,49 was mapped with ten 1.2~s exposures in the initial phase of the
450: program \citep{churchwell04}.
451: 
452: First, the ACIS image was shifted to match the absolute astrometry of 2MASS using the
453: closest ACIS--2MASS pairs. The residual displacements of the pairs in the R.\,A. and
454: decl. directions are plotted in Figure~\ref{fg:f3}, in which the rms (1~$\sigma$) of the
455: displacements is $\sim$0\farcs56 (Fig.~\ref{fg:f3}; the inner dotted circle). We
456: recognized 216 ACIS--2MASS pairs within 2 $\sigma$ displacement (Fig.~\ref{fg:f3}; the
457: outer dotted circle) to be physical counterparts. We similarly found NOMAD, SIRIUS, and
458: IRAC counterparts within 2 $\sigma$ of ACIS sources after a correction of the systematic
459: offset, where 1 $\sigma$ is $\sim$0\farcs63, $\sim$0\farcs38, and $\sim$0\farcs59,
460: respectively. Consequently, 230, 299, and 156 ACIS sources were found to have NOMAD,
461: SIRIUS, and IRAC counterparts. Both the numbers of false positives (unrelated pairs
462: identified as counterparts) and false negatives (physical pairs unidentified as
463: counterparts) are estimated to be $<$10.
464: 
465: \begin{figure}[hbtp]
466:  \figurenum{3}
467:  \epsscale{1.0}
468:  \plotone{f3.jpg}
469:  \caption{Astrometric consistency between ACIS and 2MASS after correcting for a
470:  systematic offset. The differences of R.\,A. and decl. between the ACIS and 2MASS
471:  positions ($\Delta$R.\,A. and $\Delta$Decl.) are plotted for all the closest pairs. The
472:  dotted circles are drawn at 1 and 2 $\sigma$ radius, whereas the solid square
473:  represents the ACIS pixel size of 0\farcs492$\times$0\farcs492.}\label{fg:f3}
474: \end{figure}
475: 
476: The resultant identifications and the optical (\ri), NIR (\ji, \hi, and \ksi), and MIR
477: ([3.8], [4.5], [5.8], and [8.0]) photometry are summarized in
478: Table~\ref{tb:t3}. The \ri-band magnitude is from NOMAD, which is a compilation work of
479: several catalogues. NIR magnitudes are either from 2MASS or SIRIUS, which is indicated
480: by the NIR flag column.
481: 
482: 
483: In total, 379 of the 468 ACIS sources have either NOMAD, 2MASS, SIRIUS, or IRAC
484: counterparts. Except for massive stars in and around Wd\,2, almost all the X-ray sources
485: are poorly studied in other wavelengths. Spectroscopic classifications are available for
486: five sources outside of Wd\,2 \citep{moffat91,rauw07}: HD\,90273 (O\,7), HD\,302752 (A),
487: MSP\,91 (G0\,V--III), MSP\,158 (G0\,III), and MSP\,218 (A0\,III), among which only HD\,90273
488: was detected in X-rays. The identifications with sources in earlier work are also given
489: in Table~\ref{tb:t3}.
490: 
491: 
492: 
493: \section{DISCUSSION}\label{sect:s5}
494: \subsection{Early-type stars in Westerlund 2 and Beyond}\label{sect:s5-1}
495: Figure~\ref{fg:f4} shows a close-up view of Wd\,2 in the SIRIUS and ACIS images. The
496: \textit{Chandra} observation of RCW\,49 resolved X-ray sources in the central OB
497: association. A dozen early-type stars from O7 to O3 are present in the cluster core ;
498: MSP\,151, 157, 167, 171, 175, 182, 183, 188, 199, 203, and 263 \citep{rauw07}. In
499: addition, three early-type stars, an O4\,V(f) or O5\,V(f) star (MSP\,18;
500: \citealt{uzpen05,rauw07}) and two Wolf-Rayet stars (WR\,20a $=$ MSP\,240 and WR\,20b),
501: lie outside the cluster core.
502: 
503: \begin{figure*}[hbtp]
504:  \figurenum{4}
505:  \epsscale{0.5}
506:  \plotone{f4a.jpg}
507:  \plotone{f4b.jpg}
508:  \caption{Close-up views of Wd\,2. (a) SIRIUS \ksi-band image. The positions of ACIS
509:  (crosses), SIRIUS (pluses), IRAC (diamonds), and 2MASS (squares) sources are shown.
510:  O-type stars in \citet{rauw07} are also shown by thick squares with their names (the
511:  prefix ``MSP'' is omitted). (b) \textit{Chandra} image. The color coding is the same
512:  as in Figure~\ref{fg:f1} (a). The positions of Chandra sources are shown by
513:  crosses with their sequential numbers (Table~\ref{tb:t1}).}\label{fg:f4}
514: \end{figure*}
515: 
516: We detected X-ray emission from all of these early-type sources. The identifications are
517: given in Table~\ref{tb:t3}. The results illustrate clearly that high-resolution X-ray
518: imaging observations, even with a short exposure of 30--40~ks, are powerful enough to
519: resolve and identify the crowded massive members in OB associations at a few kiloparsecs
520: away, as was found by \citet{jwang07a} in NGC\,6357, \citet{broos07} in M\,17, and
521: similar studies.
522: 
523: \subsection{Galactic \& Extra-Galactic Contaminants}\label{sect:s5-2}
524: The population of X-ray sources detected in ACIS images of Galactic massive SFRs is a
525: superposition of the young cluster members, field stars in the Galactic Plane, and
526: extragalactic background. In order to examine how many X-ray sources toward RCW\,49 are
527: intrinsic to the cluster, we compare the RCW\,49 population with that in a control field
528: of the Galactic Plane \citep{ebisawa05}.
529: 
530: The control field, the ``Scutum region'' at ($l$, $b$) $\sim$ (28\fdg5, 0\fdg0), is an
531: ideal place for our comparison study. First, the region was covered by X-ray and NIR
532: imaging studies with a sensitivity similar to RCW\,49. Second, it does not harbor a
533: young stellar cluster and the X-ray population there is a typical Galactic Plane
534: population superposed on obscured extragalactic contaminants. Third, the extinction in
535: the two regions is similar. The typical \nh\ in RCW\,49 is derived to be
536: 10$^{22}$--10$^{22.5}$~cm$^{-2}$ from Figure~\ref{fg:f5}, which is consistent with the
537: \textit{ROSAT} \citep{belloni94} and NIR extinction \citep{ascenso07} measurements. The
538: Scutum region shows a bimodal \nh\ distribution unlike RCW\,49 (Fig.~\ref{fg:f5}), but
539: their median values are the same (10$^{22.2}$~cm$^{-2}$). In any case, the extinction
540: correction is not very important in the hard X-ray band, where the opacity is less than
541: one.
542: 
543: \begin{figure}[hbtp]
544:  \figurenum{5}
545:  \epsscale{1.0}
546:  \plotone{f5.jpg}
547:  \caption{Histograms of the \nh\ values of X-ray sources derived by spectral fits
548:  (Table~\ref{tb:t2}). The solid and dashed histograms are for the RCW\,49 and the Scutum
549:  regions, respectively.}\label{fg:f5}
550: \end{figure}
551: 
552: We use the radial profile of X-ray surface number density (SND) in RCW\,49 to argue that
553: a large fraction of our X-ray sources consists of the intrinsic RCW\,49
554: population. First, the radial SND profile of all X-ray sources (Fig.~\ref{fg:f6}a) shows
555: a centrally peaked shape. The peak at R.\,A.$=$10:24:01.6 and decl.$=$--57:45:31, which
556: was derived as the mean position of all X-ray sources, is consistent with the position
557: of Wd\,2. Second, the SND of RCW\,49 has an overpopulation compared to the Scutum region
558: (Fig.~\ref{fg:f6}a; \textit{dotted histogram}) out to $\sim$6\arcmin--7\arcmin. Within
559: 5\arcmin\ of the cluster center where we have SIRIUS coverage, the number of X-ray
560: sources is 374 while the projected number of the Scutum region is $\sim$69, indicating
561: that $\sim$82\% of the X-ray sources are intrinsic to RCW\,49.
562: 
563: \begin{figure}[hbtp]
564:  \figurenum{6}
565:  \epsscale{1.0}
566:  \plotone{f6.jpg}
567:  \caption{Radial SND profile of (a) all X-ray sources, (b) sources with optical, NIR, or
568:  MIR counterparts, and (c) those without counterparts. Solid and dashed histograms
569:  indicate the data from RCW\,49 and the Scutum region. The center of the profiles is the
570:  mean position of all X-ray sources in RCW\,49 and the same detector position for the
571:  Scutum region. The uncertainty is estimated by $\sqrt{N}$, where $N$ is the number of
572:  sources in each bin.}\label{fg:f6}
573: \end{figure}
574: 
575: Our claim that the X-ray population suffers little contamination by non--cluster-members
576: is more evident for those with stellar (optical, NIR, and MIR) identifications
577: (Fig.~\ref{fg:f6}b). We have 330 identified X-ray sources within 5\arcmin\ of the Wd\,2
578: center, while the projected number in the Scutum region is $\sim$47, which accounts only
579: for $\sim$14\% of the RCW\,49 count.
580: 
581: This overpopulation holds in the entire X-ray flux range of the \textit{Chandra}
582: study. In the comparison of the hard-band $\log{N}-\log{S}$ relations
583: (Fig.~\ref{fg:f7}), we find the total and identified X-ray populations of the RCW\,49
584: region (\textit{thick histograms}) are well above those of the Scutum region
585: (\textit{broken histograms}).
586: 
587: \begin{figure}[hbtp]
588:  \figurenum{7}
589:  \epsscale{1.0}
590:  \plotone{f7.jpg}
591:  \caption{$\log{N}-\log{S}$ plots of RCW\,49 (solid) and the Scutum region
592:  (broken). Hard-band absorption-uncorrected flux was used. Histograms are constructed
593:  for all sources (black) and those with stellar counterparts (red),
594:  respectively.}\label{fg:f7}
595: \end{figure}
596: 
597: 
598: \subsection{Distance Constraint from X-ray Data}\label{sect:s5-3}
599: The distance to RCW\,49 is controversial despite intensive studies across many
600: wavelengths (\S~\ref{sect:s2}). In the present study, we obtained a rich X-ray sample
601: and learned that most of the identified X-ray sources are intrinsic to RCW\,49. It is
602: therefore worthwhile to attempt to constrain the distance with the new X-ray data set.
603: 
604: We use a method proposed in \citet{getman07b} to constrain the distance based on X-ray
605: data. In this method, a constraint comes from the dependence of X-ray luminosity on the
606: mass of late-type pre--main-sequence sources. The relation between these two stellar
607: quantities is established consistently in several young star clusters, including Cepheus
608: B \citep{getman06}, the Orion Nebula Cluster (\citealp[Table~6]{getman06};
609: \citealt{preibisch05}), the Taurus Molecular Cloud \citep {guedel07}, and CG\,12
610: \citep{getman07b}. Note that the X-ray luminosity versus mass relation is not degenerate
611: with respect to distance because pre--main-sequence X-ray luminosity and mass depend
612: differently on the assumed distance.
613: 
614: First, we estimate the stellar mass of X-ray sources, for which NIR photometry is
615: available (Table~\ref{tb:t3}). We construct a \ji/(\ji--\hi) color-magnitude diagram
616: assuming various distances from 0.5 to 8~kpc and derive the mass by comparing the
617: dereddened \ji-band magnitudes to the \citet{siess00} model. Then, we extract the X-ray
618: sources with an estimated mass of 2.0--2.7~$M_{\odot}$ and derive the median value of
619: their logarithmic hard-band absorption-corrected X-ray luminosity ($\log L_{\rm{h,c}}$
620: in Table~\ref{tb:t2}) in order to minimize the extinction effect. The mass range is
621: limited by the degeneracy in the mass estimate above $\sim$2.7~$M_{\odot}$ and by our
622: X-ray sensitivity below $\sim$2.0~$M_{\odot}$.
623: 
624: Figure~\ref{fg:f8} shows the median X-ray luminosity at each assumed distance, in which
625: the estimated mean X-ray luminosity increases monotonically as the distance
626: increases. We compare the median X-ray luminosity in the same mass range of the Orion
627: Nebula Cluster \citep{getman06}, for which the distance is reliably measured. We found
628: that the distance is constrained to be 2--5~kpc, which is consistent with the claim by
629: \citep{ascenso07} but not with \citep{rauw07}.
630: 
631: \begin{figure}[hbtp]
632:  \figurenum{8}
633:  \epsscale{1.0}
634:  \plotone{f8.jpg}
635:  \caption{Median X-ray luminosity versus assumed distance of RCW\,49 sources. The
636:  logarithmic median value of hard-band absorption-corrected X-ray luminosities in the
637:  mass ranges of 2.0--2.7~$M_{\odot}$ is plotted at each assumed distance. The error bars
638:  show the median absolute dispersion (MAD) divided by the number of samples (7--46
639:  sources) at each bin, which represent the uncertainty of median \lx\ determination. The
640:  dotted and dashed lines indicates the median \lx\ and its MAD divided by the number of
641:  samples from the Orion Nebula Cluster in the same mass range, respectively
642:  \citep{getman06}.}\label{fg:f8}
643: \end{figure}
644: 
645: 
646: \subsection{Nature of X-ray Cluster Population}\label{sect:s5-4}
647: Sources of different natures have different X-ray and NIR brightness and colors. In
648: massive SFRs, where the mass range of stellar constituents is wider than low-mass SFRs,
649: the brightness is of prime importance for estimating the nature. We employ X-ray versus
650: NIR brightness plots (Figs.~\ref{fg:f9}a and \ref{fg:f9}b) to estimate the nature of
651: X-ray cluster members of RCW\,49, aided by the conventional NIR color-magnitude diagram
652: (Fig.~\ref{fg:f11}a).
653: 
654: Figure~\ref{fg:f9} (a) shows the observed \ksi-band magnitude ($F_{K\rm{s}}$) versus
655: hard-band X-ray flux ($F_{\rm{h}}$) of all the X-ray sources with \ksi-band
656: identification. Figure~\ref{fg:f9} (b) is restricted to a sample with X-ray spectral
657: fits, showing the extinction-corrected absolute \ksi-band luminosity ($L_{K\rm{s}}$)
658: versus hard-band X-ray luminosity ($L_{\rm{h,c}}$). The constant flux ratio
659: lines between the \ksi-band and the hard X-ray band (\textit{dotted-and-dashed lines} in
660: Figure~\ref{fg:f9}a and b) are a proxy for the canonical X-ray versus bolometric
661: luminosity, which is independent from our distance estimate.
662: 
663: \begin{figure*}[hbtp]
664:  \figurenum{9}
665:  \epsscale{0.5}
666:  \plotone{f9a.jpg}
667:  \plotone{f9b.jpg}
668:  \caption{Scatter plots of X-ray and NIR brightness of NIR-identified X-ray sources. (a)
669:  Hard-band X-ray flux ($F_{\rm{h}}$) versus \ksi-band flux ($F_{K\rm{s}}$) or magnitude
670:  ($m_{K\rm{s}}$) for 339 X-ray sources with \ksi-band counterparts. (b)
671:  Extinction-corrected hard-band X-ray luminosity ($L_{\rm{h,c}}$) versus \ksi-band
672:  luminosity ($L_{K\rm{s}}$) or absolute magnitude ($M_{K\rm{s}}$) for 124 X-ray sources
673:  with \nh\ estimates in Table~\ref{tb:t2}. Sources are NIR-loud toward the top left and
674:  X-ray-loud toward the bottom right of the figures. Wolf-Rayet stars are shown by large
675:  diamonds. Sources brighter than B2\,V in NIR are shown by upward triangles for bright
676:  (more than 50 counts) and hard (best-fit temperature larger than 2~keV), downward
677:  triangles for bright and soft, and squares for faint (less than 50 counts)
678:  sources. Sources fainter than B2\,V in NIR are indicated by circles. Those with X-ray
679:  variability are marked with filled symbols. Spectroscopically-identified early-type
680:  stars are labeled. The prefix ``MSP'' is omitted for \citet{moffat91} sources. ACIS\,29
681:  indicates the source number 29 in Tables~\ref{tb:t1}, \ref{tb:t2}, and \ref{tb:t3}. The
682:  dashed-and-dotted lines indicate the iso-$x/y$ lines from 10$^{-7}$ to 10$^{2}$ (the
683:  logarithmic values are given beside each line).} \label{fg:f9}
684: \end{figure*}
685: 
686: It is easily recognized that the sources form several groups in both plots. The most
687: populous group centered at ($F_{\rm{h}}$, $F_{K\rm{s}}$) $\sim$ ($10^{-14.8}$,
688: $10^{-12.7}$)~ergs~s$^{-1}$~cm$^{-2}$ in Figure~\ref{fg:f9} (a) and ($L_{\rm{h.c}}$,
689: $L_{K\rm{s}}$) $\sim$ ($10^{31}$, $10^{33.5}$)~ergs~s$^{-1}$ in Figure~\ref{fg:f9} (b)
690: are 1.0--3.0~$M_{\odot}$ low-mass pre--main-sequence sources in RCW\,49 for the
691: following reasons. Their \ksi-band brightness (13--16~mag) is consistent with the NIR
692: brightness of pre--main-sequence sources in this mass range. Their hard-band X-ray
693: luminosities 10$^{30}$--10$^{31.5}$~ergs~s$^{-1}$ and flaring are typical of this class
694: of stars in near-by regions (e.g., \citealt{wolk06,getman06}). The ratio of X-ray to NIR
695: brightness of 10$^{-1}$--10$^{-3}$ is consistent with the fractional X-ray to bolometric
696: luminosity of 10$^{-3}$--10$^{-5}$ common among these sources. The \ksi-band bolometric
697: corrections for A0\,V, F0\,V, and G0\,V stars are --3.75, --2.85, and --2.25~mag
698: \citep{tokunaga00,drilling00}.
699: 
700: A number of sources deviate in the upward and rightward directions of the primary group
701: in Figure~\ref{fg:f9}. Among them, two Wolf-Rayet stars occupy the top right end of all
702: sources with ($F_{\rm{h}}$, $F_{K\rm{s}}$) $\sim$ ($10^{-12.8}$,
703: $10^{-10.2}$)~ergs~s$^{-1}$~cm$^{-2}$ in Figure~\ref{fg:f9} (a) and ($L_{\rm{h,c}}$,
704: $L_{K\rm{s}}$) $\sim$ ($10^{33}$, $10^{36}$)~ergs~s$^{-1}$ in Figure~\ref{fg:f9} (b).
705: 
706: The sources in between the late-type pre--main-sequence and Wolf-Rayet star groups, both
707: in X-ray and NIR brightness, are early-type main-sequence stars. They have a large
708: dispersion in the horizontal direction in Figure~\ref{fg:f9}, indicating that sources
709: with a similar NIR brightness can differ significantly in hard X-ray brightness. We
710: first extract early-type star candidates based on the NIR color-magnitude diagram
711: (Fig.~\ref{fg:f10}a). X-ray sources with dereddened \ki-band magnitude brighter than a
712: B2\,V star are considered to be early-type candidates and are labeled ``ET'' in
713: Table~\ref{tb:t3}. A total of 43 candidate stars include 13 spectroscopically-identified
714: O-type stars from preceding studies \citep{moffat91,rauw07}. These early-type candidates
715: are plotted by different symbols based on their X-ray spectral hardness.
716: 
717: \begin{figure*}[hbtp]
718:  \figurenum{10}
719:  \epsscale{0.5}
720:  \plotone{f10a.jpg}
721:  \plotone{f10b.jpg}
722:  \caption{NIR diagrams of RCW\,49. SIRIUS sources with X-ray and MIR excess emission are
723:  respectively shown by blue filled and red open circles. X-ray sources with 2MASS but
724:  without SIRIUS photometry are plotted by blue filled triangles. 2MASS and SIRIUS colors
725:  are transformed into the California Institute of Technology color system to compare
726:  with the theoretical calculations using the formulae given respectively by
727:  \citet{carpenter01} and \citet{nakajima05}. (a) Color-magnitude diagram. A 2~Myr
728:  isochrone curve of pre--main-sequence stars (thick solid curve) is from \citet[$M <
729:  1.4~M_{\odot}$]{baraffe98} and \citet[$3.0~M_{\odot} > M >
730:  1.4~M_{\odot}$]{siess00}. The reddening lines of 1, 2, and 3~$M_{\odot}$ sources are
731:  shown with solid lines. A color-magnitude curve of main sequence stars (thick dashed
732:  curve) is from \citet{tokunaga00,drilling00}. The reddening lines of O9\,V and B2\,V
733:  stars are shown with dashed lines. The SIRIUS \ki-band magnitude limit is indicated by
734:  the dotted line. Spectroscopically-identified early-type stars are labeled. The prefix
735:  ``MSP'' is omitted for \citet{moffat91} sources. (b) Color-color diagram. The intrinsic
736:  colors of dwarfs and giants (thick curves) are from \citet{tokunaga00}, the classical T
737:  Tauri star locus (thick line) is from \citet{meyer97}, and the intrinsic colors of
738:  Herbig Ae Be stars (dotted curve) are from \citet{lada92}.}\label{fg:f10}
739: \end{figure*}
740: 
741: The upward and downward triangle sources, which respectively indicate hard and soft
742: X-ray emission by X-ray spectroscopy, occupy different regions naturally reflecting the
743: fact that harder O stars are brighter in the hard X-ray band. The two groups appear to
744: form two separate clustering in Figure~\ref{fg:f9} and have different X-ray to NIR
745: brightness ratio. These suggests that they have different X-ray production
746: mechanisms. Eight spectroscopically-identified sources (HD\,90273, MSP\,151, 171, 182,
747: 183, 199, 203, and 263) are in the soft O-star group with $L_{\rm{h,c}}$/$L_{K\rm{s}}
748: \sim 10^{-4}$, while three (MSP\,18, 167, and 188) are in the hard O-star group with
749: $L_{\rm{h,c}}$/$L_{K\rm{s}} = 10^{-2} - 10^{-3}$. Source number 29 (ACIS 29) may belong
750: either to the hard O-star or Wolf-Rayet star group, but the nature of this source is
751: unknown without any spectroscopic studies. We note that the source has infrared excess
752: emission (Table~\ref{tb:t3}) and is an off-cluster source at $\sim$4\farcm5 away from
753: the center, which is similar to WR\,20a and MSP\,18.
754: 
755: The soft O stars share common characteristics with traditional X-ray-emitting O stars
756: \citep{berghoefer97} for having soft and constant X-ray emission due to internal shocks
757: in their stellar winds \citep{lucy80}. The hard O stars are among the growing class of
758: new-type O stars with hot thermal emission (\S~\ref{sect:s5-7-2}). The three hard O
759: stars (MSP\,18, 167, and 188) are bright enough for detailed X-ray spectroscopy, and we
760: discuss them in \S~\ref{sect:s5-7-2}.
761: 
762: 
763: 
764: \subsection{Spatial Distribution of X-ray Cluster Members}\label{sect:s5-5}
765: We can individually identify cluster members based on their X-ray identification with
766: a low level of contamination. Unlike optical and infrared imaging, the X-ray detections
767: are not affected by non-uniform and intense diffuse emission. These two features of
768: X-ray imaging studies are particularly important outside of the cluster core, where the
769: optical and NIR SNDs of cluster members fall to a comparable level to those of
770: irrelevant sources. Using the X-ray SND profiles, we obtain a sensible estimate of the
771: spatial scale and distribution of cluster members away from the cluster core.
772: 
773: In the X-ray SND profile (Fig.~\ref{fg:f6}a), we see two interesting features. One is
774: that the profile shows an overall decline, reaching the background level at
775: 6\arcmin--7\arcmin. The X-ray declining profile is smoother and more extended than the
776: NIR profile constructed from 2MASS \citep[Fig.~6]{ascenso07}. We obtained a larger
777: estimate of the cluster size by twice than the NIR estimate of 4\farcm1\
778: \citep{ascenso07}. Our estimated radius is $\sim$8~pc at the assumed distance, which is
779: far larger than the Orion Nebula Cluster and somewhat larger than NGC\,6357
780: \citep{jwang07a}.
781: 
782: The other feature is a deviation from the monotonic trend at $\sim$4\arcmin. A similar
783: excess can also be found in the identified and unidentified X-ray source profile
784: (Figs.~\ref{fg:f6}b and c). A $\sim$4\arcmin\ radius partial ring around Wd\,2 is the
785: most conspicuous feature both in the radio and MIR images (Fig.~\ref{fg:f1}a;
786: \citealt{whiteoak97,churchwell04}), where gas and dust are compressed by winds and
787: radiation from massive sources in the central OB association. The possible peak at
788: $\sim$4\arcmin\ in the X-ray profiles, although the significance is low, suggests that a
789: second generation of triggered star formation is present in this ring. A deeper X-ray
790: image would confirm this possibility and identify members of the triggered population.
791: 
792: We identified thirty new OB star candidates based on the NIR photometry of X-ray sources
793: (\S~\ref{sect:s5-4}). It is interesting to note that many of them are located outside of
794: the cluster core. In addition to the three off-core early-type stars (WR\,20a, 20b, and
795: MSP\,18 at $\sim$0\farcm6, $\sim$3\farcm7, $\sim$1\farcm0\ from the Wd\,2 center,
796: respectively; Fig.~\ref{fg:f1}a) considered to be physically associated with RCW\,49, we
797: have 14 early-type candidates beyond 3$\arcmin$ of the cluster center. One of them
798: (HD\,90273) is a spectroscopically-identified O star \citep{cannon19}. This implies
799: either that these are run-away O stars as suggested for MSP\,18 \citep{uzpen05} or that
800: the cluster was formed without primordial mass segregation. We note that the OB stars in
801: the Rosette Nebula Cluster similarly do not exhibit mass segregation
802: \citep{jwang07b}. Follow-up studies are mandatory to confirm the early-type nature of
803: these off-core candidates and to understand the cause of their large spatial
804: distribution.
805: 
806: 
807: \subsection{Discrimination of Intrinsic Cluster Members}\label{sect:s5-6}
808: We detected 10,540 and 9,768 sources in the SIRIUS images on and off of the RCW\,49
809: region (\S~\ref{sect:s3}; hereafter called the object and the control fields). This
810: indicates that the NIR sample is seriously contaminated by field stars. Individual
811: cluster members, most of which are low-mass pre--main-sequence sources, are
812: traditionally identified by infrared excess emission from their circumstellar disks and
813: envelopes. MIR excess studies are more sensitive than NIR excess studies
814: \citep{haisch01}. \citet{whitney04} identified 256 cluster members in the
815: \textit{Chandra} field from the MIR excess signature based on the MIR color-color
816: diagrams using \textit{Spitzer}/IRAC. However, the infrared excess is only sensitive to
817: classical T Tauri stars and protostars with rich circumstellar matter. Weak-line T Tauri
818: stars are usually missed. On the contrary, X-ray emission in the $\gtrsim$1~keV band
819: does not significantly depend on whether pre--main-sequence sources have disks or not
820: \citep{feigelson07}. We now use the X-ray emission as a signature of cluster membership
821: and examine how the \textit{Chandra} data complement the \textit{Spitzer} data.
822: 
823: For this purpose, we construct three \ksi-band luminosity functions (KLFs). One is the
824: intrinsic KLF of RCW\,49 obtained by subtracting the KLF of the control field from that
825: of the object field. The other two are the KLFs of SIRIUS sources with X-ray and MIR
826: excess emission signatures. We examine what fraction of the first KLF is recovered by
827: each and the combination of the latter two KLFs.
828: 
829: Figure \ref{fg:f11} (a) shows the KLFs in the object and the control fields
830: (\textit{solid and dashed histograms}), while the difference of the two is plotted by a
831: solid histogram in Figure \ref{fg:f11} (b). Sources in the two fields have slightly
832: different characteristic extinction, for which we compensate by shifting the object KLF
833: brightward by 0.1~mag based on the following argument. The difference between the mean
834: NIR colors in the object and control fields represents the difference in the
835: extinction. In order to avoid the effect of excess emission seen among
836: pre--main-sequence sources, we use only the (\ji--\hi) color of sources without any
837: evidence of infrared excess. The mean (\ji--\hi) values in the object and control fields
838: are 0.94 and 0.83 mag with $\sim$0.01~mag uncertainty. The difference can be converted
839: to $\Delta A_{K} \sim 0.1$~mag \citep{tokunaga00,benjamin03}.
840: 
841: \begin{figure}[hbtp]
842:  \figurenum{11}
843:  \epsscale{1.0}
844:  \plotone{f11.jpg}
845:  \caption{(a) KLFs of the object and control fields from SIRIUS (solid and dashed
846:  histograms). The KLF of the object field is shifted leftward by 0.1~mag to compensate
847:  for extra reddening relative to the control field. (b) Intrinsic KLF of RCW\,49 (solid
848:  histogram) derived as the subtraction of the KLFs of the object and control fields. The
849:  Poisson uncertainty is given by error bars. The KLFs of SIRIUS sources with X-ray and
850:  MIR excess (ME) identifications are given by triangles and diamonds, respectively. The
851:  KLFs of SIRIUS sources with either of and both of the two identifications (the union
852:  and intersection) are given by dotted and dashed histograms,
853:  respectively.}\label{fg:f11}
854: \end{figure}
855: 
856: First, we compare the intrinsic KLF of RCW\,49 with each of the KLFs of sources with
857: X-ray and MIR excess emission, which are respectively shown with triangles and diamonds
858: in Figure~\ref{fg:f11} (b). Each of the X-ray and MIR-excess KLFs represent part of the
859: intrinsic KLF. The recovery rate drops at \ksi~$\gtrsim$~14~mag because the
860: \textit{Chandra} and \textit{Spitzer} sensitivities do not match that of SIRIUS. Also,
861: the \textit{Spitzer} image suffers confusion in the most crowded part of the
862: cluster. The recovery of cluster members by the X-ray KLF is also expected to drop in
863: the magnitude range for late B and A type sources, because some of them are
864: intrinsically X-ray faint unlike earlier and later type sources
865: \citep{preibisch05}. These sources would appear at $\sim$12~mag at the assumed
866: distance. The smaller X-ray recovery rate at 12.0--12.5~mag may be attributable to this
867: effect (Fig.~\ref{fg:f11}b).
868: 
869: We nonetheless do not attempt to correct statistically for these types of incompleteness
870: because we are interested in how many of the intrinsic cluster members are
871: \textit{individually} identified by the current data set. The total counts of sources in
872: the intrinsic KLF at \ki$<$14.0 and $<$14.5~mag are 221 and 358. Respectively, 125
873: (57\%) and 179 (50\%) of them are recovered in the X-ray KLF, while 84 (38\%) and 98
874: (27\%) are in the MIR excess KLF. Each of the X-ray and MIR excess KLFs represent about
875: 30--60\% of the intrinsic RCW\,49 KLF.
876: 
877: Second, we compare the intrinsic KLF with the combination of X-ray and MIR excess
878: KLFs. The KLFs of the union and the intersection sets of X-ray and MIR excess sources
879: are shown by dotted and dashed histograms in Figure~\ref{fg:f11} (b). The union set
880: recovers 90\% (200) and 74\% (266) of the intrinsic cluster members at \ki$<$14.0 and
881: $<$14.5~mag, respectively. The intersection set is very small. These indicate that X-ray
882: and MIR excess emission selections operate in a complementary fashion in identifying
883: cluster members and that the combination of the two recovers a large fraction of the
884: members. The complementarity probably comes from the fact that richer circumstellar
885: matter around pre--main-sequence sources increases the MIR excess detection rate due to
886: stronger emission and decreases the X-ray detection rate due to larger extinction. This
887: is best examined in the NIR color-color diagram in Figure~\ref{fg:f10} (b). SIRIUS
888: sources with X-ray and MIR excess signatures (blue and red dots) are segregated in the
889: plot. The X-ray emission preferentially detects sources with no or smaller NIR excess.
890: 
891: \subsection{X-ray Emission from Massive Stars}\label{sect:s5-7}
892: We present the results of detailed X-ray spectroscopy for five exceptionally bright
893: X-ray sources with more than 200 counts. Two of them are Wolf-Rayet stars (WR\,20a and
894: 20b) and the rest are O-type stars (MSP\,18, 167, and 188). All these sources are
895: located at the top right end of X-ray and NIR brightness plots (Fig.~\ref{fg:f9}).
896: 
897: \subsubsection{Wolf-Rayet stars}\label{sect:s5-7-1}
898: The X-ray spectra of WR\,20a and WR\,20b are shown in Figure~\ref{fg:f12} (a) and (b),
899: respectively. Both of them are characterized by emission lines from highly ionized ions
900: such as Mg, Si, S, Ar, and Ca, and a hard tail up to $\sim$5~keV. The \ion{Si}{13}
901: K$\alpha$ emission as well as the hard tail indicate that the emission is more likely
902: from a plasma of multiple temperatures than from that of a uniform temperature. We
903: present the results of a two-temperature plasma fit in Table~\ref{tb:t4}. In the
904: fitting, the abundances of conspicuous lines (Si and S) were thawed to obtain the
905: best-fit values. The abundances of other elements were fixed to be 1 solar.  It is known
906: that hydrogen is heavily depleted in Wolf-Rayet stars (e.g.,
907: \citealt{vanderhucht86,morris00}), but we used the cosmic abundance \citep{anders89} as
908: we do not know the level of hydrogen depletion for these stars. Once it is known, the
909: abundance values presented here should be decreased accordingly.
910: 
911: \begin{figure*}[hbtp]
912:  \figurenum{12}
913:  \epsscale{1.0}
914:  \plotone{f12.jpg}
915:  \caption{X-ray spectra of bright sources. The source identifications and the models
916:  employed to fit the spectra are given in Table~\ref{tb:t4}. Grouped data with errors
917:  are shown in the upper panels over-plotted with the best-fit model convolved with
918:  mirror and detector responses (solid line). The energy of the K$\alpha$ lines of some
919:  representative highly ionized ions are shown. The lower panels show the residuals of
920:  the fit.}\label{fg:f12}
921: \end{figure*}
922: 
923: WR\,20a is an eclipsing binary of two Wolf-Rayet stars of spectral type WN6ha
924: \citep{bonanos04,rauw05}. Using the best-fit orbital solution by \citet{bonanos04}, the
925: \textit{Chandra} observation covers the orbital phase from 0.36 to 0.48. Because of the
926: uncertainty in the orbital period determination of 3.686 $\pm$ 0.01 days, the
927: \textit{Chandra} phase range has uncertainty from 0.18--0.30 to 0.55--0.67. We compare
928: the X-ray light curves and the eclipsing optical light curve (Fig.~\ref{fg:f13}) with the
929: aim to obtain a hint in the X-ray production mechanism of this source. One artificial
930: effect is introduced by the fact that the source was observed in a gap between ACIS I2
931: and I3. The source oscillated across the gap at a period of 1000~s due to the dithering
932: motion of the telescope. About 60\% of the photons were lost in the gap. We binned the
933: light curves with a binning size of 1000~s to cancel this effect.
934: 
935: \begin{figure}[hbtp]
936:  \figurenum{13}
937:  \epsscale{1.0}
938:  \plotone{f13.jpg}
939:  \caption{Light curves of the binary system WR\,20a. (a) \textit{I}-band magnitude of
940:  one orbit \citep{bonanos04}. (b) X-ray count rate in the 0.5--8.0~keV band and (c) the
941:  hardness ratio (HR) in the phase 0.36 to 0.48 (the range is indicated by two dotted
942:  lines in a; the uncertainty is about $\pm$0.2). We used the ephemeris by
943:  \citet{bonanos04} with an orbital period of 3.686 days and a Julian date origin of
944:  2,453,124.569 day. HR here is defined as (H$-$S)/(H$+$S) where H and S are the count
945:  rates in the hard (2.0--8.0~keV) and the soft (0.5--2.0~keV) bands, respectively.}\label{fg:f13}
946: \end{figure}
947: 
948: No apparent variation was found in either the average X-ray count rate
949: (Fig.~\ref{fg:f13}b) or the hardness ratio (Fig.~\ref{fg:f13}c). In contrast, a clear
950: decline is seen in the \textit{I}-band magnitude (Fig.~\ref{fg:f13}a) during the phase
951: of about 0.3--0.7. The \textit{Chandra} phase coverage is uncertain from 0.18--0.30 to
952: 0.55--0.67. When its lower limit is adopted, the \textit{Chandra} observation was
953: conducted out of eclipses. When other values are adopted, it was conducted during an
954: optical eclipse. In this case, the lack of the eclipse in the X-ray light curves
955: indicates that the X-ray emitting region is far enough from the stellar surface so as
956: not affected by the photospheric eclipsing. From the orbital solution by
957: \citet{bonanos04}, the radii of the two stars are $\sim$19$R_{\odot}$ and $a\sin{(i)}$
958: is $\sim$53$R_{\odot}$, where $a$ is the semi-major axis of the orbit,
959: $i=$74.5$^{\circ}$ is the inclination angle, and $R_{\odot}$ is the radius of the
960: Sun. The system is observed close to the edge-on in the line of sight. At the middle of
961: the eclipse, the two stars are separated by $2a\cos{(i)} \sim 29 R_{\odot}$. Therefore,
962: the X-ray emitting region is extended to at least a few stellar radii.
963: 
964: 
965: \subsubsection{O-type Stars}\label{sect:s5-7-2}
966: At least two out of three bright O stars show emission line features of highly ionized
967: Mg, Si and S, indicating their thermal plasma origin. We fit all three spectra with
968: a thin-thermal plasma model with variable abundance of elements with
969: conspicuous features. Table~\ref{tb:t4} gives the results. All of them are characterized
970: by a hard spectrum with the best-fit plasma temperature of $\gtrsim$1.5~keV and luminous
971: X-ray brightness of 2--8$\times$10$^{32}$~ergs~s$^{-1}$.
972: 
973: Such hard and luminous X-ray-emitting O stars have been reported in some other massive
974: SFRs \citep{albacete03,rho04,skinner05,broos07,wolk06}. The X-ray properties of these
975: sources are quite different from the classical view of main-sequence O-type stars
976: characterized by soft thermal spectra of $\lesssim$1~keV and the X-ray to bolometric
977: luminosity ratio of $\sim$10$^{-7}$ \citep{berghoefer97}. Hard X-ray--emitting O stars
978: in massive SFRs should be a new distinct class of X-ray sources with a different X-ray
979: production mechanism.
980: 
981: Several interpretations have been proposed for such hard X-ray emission from O
982: stars. One is magnetically-confined wind shocks
983: \citep{babel97,schulz03,gagne05,stelzer05}, where magnetized winds produce hard X-ray
984: emission in large-scale shocks near the stellar surface. Another is similar to the X-ray
985: emission in WR binaries; i. e., winds from each component of a close binary of two
986: O-type stars collide with each other to make hard X-rays
987: \citep{zhekov00,skinner07}. The hard X-rays might be from the magnetic reconnections
988: between the two binary components \citep{schulz06}. The binary nature of the three
989: O-type stars in RCW\,49 is not confirmed yet, but these sources might be
990: ``X-ray-spectroscopic binaries'' if this interpretation is correct. Yet another idea is
991: that inverse Compton emission of UV photons from these O stars accounts for their hard
992: spectra \citep{albacete03}. The X-ray production mechanism of such curious O stars is
993: not yet certain, but enriching the sample using plots like Figure~\ref{fg:f9} may lead
994: to a better understanding of the mechanism.
995: 
996: 
997: \section{Summary}\label{sect:s6}
998: We observed the Galactic massive star-forming region RCW\,49 and its central OB
999: association Westerlund 2 using \textit{Chandra}/ACIS, IRSF/SIRIUS, and
1000: \textit{Spitzer}/IRAC. The main results are summarized as follows.
1001: 
1002: \begin{enumerate}
1003:  \item 468 X-ray sources were detected in a $\sim$40~ks ACIS exposure covering a
1004:  $\sim$17\arcmin$\times$17\arcmin\ field. Previous X-ray studies detected less than 10
1005:  sources in the studied field and almost all the X-ray sources reported here are new
1006:  detections.  The results of X-ray photometric, timing, and spectroscopic analyses are
1007:  presented and summarized in Tables~\ref{tb:t1} and \ref{tb:t2}.
1008: 
1009:  \item We obtained SIRIUS images in the \ji, \hi, and \ksi\ bands of a concentric
1010:  $\sim$8\farcm3$\times$8\farcm3 field and extracted 10,540 sources at a 10~$\sigma$
1011:  detection limits of $\sim$19.0 (\ji), $\sim$17.8 (\hi), and $\sim$16.8 (\ksi) mag.
1012: 
1013:  \item 379 X-ray sources were identified with optical, NIR, and MIR counterparts using
1014:  NOMAD, 2MASS, SIRIUS, and GLIMPSE data. The results are given in Table~\ref{tb:t3}.
1015: 
1016:  \item The central OB association Westerlund 2 was resolved in the X-ray image for the
1017:  first time. X-ray emission was detected from all of the spectroscopically-identified
1018:  early-type stars consisting of two Wolf Rayet stars and a dozen O type stars.
1019: 
1020:  \item About 86\% of the X-ray sources with optical or infrared identifications were
1021:  found to be cluster members by comparing with the X-ray population in a control field
1022:  in the Galactic Plane. The X-ray overpopulation against the control field is seen in
1023:  all measured flux ranges.
1024: 
1025:  \item We compared the intrinsic \ki-band luminosity function of NIR sources in RCW\,49
1026:  with the KLFs of sources with X-ray or MIR excess identification. We found that
1027:  30--60\% of the cluster members at \ksi$<$14~mag are individually identified using the
1028:  two observational signatures of cluster membership. When X-ray and MIR excess
1029:  signatures are combined, about 90\% of cluster sources are individually located. The
1030:  two indicators of the cluster membership work in a complementary fashion, in which the
1031:  former excels in detecting weak-line T Tauri stars with smaller attenuation while the
1032:  latter predominantly detects classical T Tauri stars or protostars with richer
1033:  circumstellar matter.
1034: 
1035:  \item A loose constraint on the distance to RCW\,49 was derived to be 2--5~kpc based on
1036:  the mean hard-band X-ray luminosity of T Tauri stars in the mass range of
1037:  2.0--2.7~$M_{\odot}$.
1038: 
1039:  \item The cluster X-ray population consists of low-mass pre--main-sequence sources and
1040:  early-type sources based on X-ray and NIR photometry. Late-type pre--main-sequence
1041:  sources, Wolf-Rayet stars, and early-type main sequence stars occupy different regions
1042:  of X-ray versus NIR brightness plots. Thirty new OB candidates were identified based on
1043:  the NIR color-magnitude diagram, several times more than the number of
1044:  optically-identified OB stars.
1045: 
1046:  \item Using the X-ray surface number density radial profiles, we gave an estimate of
1047:  the cluster radius to be 6\arcmin--7\arcmin, or $\sim$8~pc at the assumed distance. A
1048:  possible excess is seen at $\sim$4\arcmin\ of the center, suggesting that secondary
1049:  star formation is taking place at the ring. Fourteen OB star candidates are found
1050:  outside of Wd\,2, suggesting an absence of mass segregation.
1051: 
1052:  \item Detailed X-ray spectroscopy of the two Wolf-Rayet stars was presented. Both
1053:  sources show two-temperature thermal plasmas with characteristic temperatures of
1054:  0.5--1.0 and 4--6~keV and 1--8~keV luminosities of 2$\times$10$^{33}$~ergs~s$^{-1}$. No
1055:  X-ray flux or hardness variation was found from WR\,20a, an eclipsing binary,
1056:  suggesting that the X-ray emission arises far from stellar surfaces.
1057: 
1058:  \item Detailed X-ray spectroscopy of three luminous O-type stars were presented. They
1059:  are characterized by thermal spectra with a temperature of $\gtrsim$1.5~keV. These are
1060:  among the growing samples of hard X-ray emitting O stars recently found in some massive
1061:  star-forming regions. Early-type stars appears to have two sub groups, hard and soft,
1062:  based on the X-ray spectroscopy.
1063: \end{enumerate}
1064: 
1065: 
1066: \acknowledgments
1067: 
1068: The authors express gratitude to Barbara Whitney for providing the \textit{Spitzer}
1069: data. M.\,T. thanks for the hospitality at the Academia Sinica Institute of Astronomy
1070: and Astrophysics in Taipei and Theoretical Institute of Advanced Research in
1071: Astrophysics in Hsinchu, Taiwan during the course of this work. We wish to thank the
1072: staff of the South African Astronomical Observatory for their kind support during our
1073: observation.
1074: 
1075: This work is financially supported by the Japan Society for the Promotion of Science,
1076: the NASA contract NAS8-38252, and the Chandra Contract SV4-74018 issued by the
1077: \textit{Chandra} X-ray Center on behalf of NASA under Contract NAS8-08060. The
1078: IRSF/SIRIUS project was financially supported by the Sumitomo Foundation and the
1079: Grant-in-Aids for Scientific Research Numbers 10147207 and 10147214 of Ministry of
1080: Education, Culture, Sports, Science, and Technology of Japan.
1081: 
1082: This work is based in part on observations made with the \textit{Spitzer Space
1083: Telescope} operated by Jet Propulsion Laboratory under a contract with NASA and data
1084: products from the Two Micron All Sky Survey, which is a joint project of the University
1085: of Massachusetts and the Infrared Processing and Analysis Center/California Institute of
1086: Technology, funded by the National Aeronautics and Space Administration and the National
1087: Science Foundation.
1088: 
1089: We also made use of the SIMBAD database, operated at CDS, Strasbourg, France. IRAF is
1090: distributed by the National Optical Astronomy Observatories, which are operated by the
1091: Association of Universities for Research in Astronomy, Inc., under cooperative agreement
1092: with the National Science Foundation.
1093: 
1094: Facilities: \facility{CXO(ACIS)}, \facility{IRSF(SIRIUS)}, \facility{SST(IRAC)}
1095: 
1096: \begin{thebibliography}{}
1097:  \bibitem[Ascenso et al.(2007)]{ascenso07} Ascenso, J., Alves, J., Beletsky, Y., \&
1098: 			       Lago, M.~T.~V.~T.\ 2007, \aap, in press
1099:  \bibitem[Aharonian et al.(2007)]{aharonian07} Aharonian, F. et al\ 2007, \aap, in press
1100:  \bibitem[Albacete Colombo et al.(2003)]{albacete03} Albacete Colombo, J.~F.,
1101: 			       M{\'e}ndez, M., \& Morrell, N.~I.\ 2003, \mnras, 346, 704
1102:  \bibitem[Albacete Colombo et al.(2007)]{albacete07} Albacete Colombo, J.~F., Flaccomio,
1103: 			       E., Micela, G., Sciortino, S., \& Damiani, F.\ 2007,
1104: 			       \aap, 464, 211 
1105:  \bibitem[Anders \& Grevesse(1989)]{anders89} Anders, E., \& Grevesse, N.\ 1989, \gca,
1106: 			       53, 197 
1107:  \bibitem[Babel \& Montmerle(1997)]{babel97} Babel, J., \& Montmerle, T.\ 1997, \apjl,
1108: 			       485, L29
1109:  \bibitem[Baraffe et al.(1998)]{baraffe98} Baraffe, I., Chabrier, G., Allard, F., \&
1110: 			       Hauschildt, P.~H.\ 1998, \aap, 337, 403
1111:  \bibitem[Bauer et al.(2004)]{bauer04} Bauer, F.~E., et al.\ 2004, Advances in Space
1112: 			       Research, 34, 2555
1113:  \bibitem[Belloni \& Mereghetti(1994)]{belloni94} Belloni, T.~\& Mereghetti, S.\ 1994,
1114: 			       \aap, 286, 935 
1115:  \bibitem[Benjamin et al.(2003)]{benjamin03} Benjamin, R.~A., et al.\ 2003, \pasp, 115, 953 
1116:  \bibitem[Berghoefer et al.(1997)]{berghoefer97} Berghoefer, T.~W., Schmitt,
1117: 			       J.~H.~M.~M., Danner, R., \& Cassinelli, J.~P.\ 1997,
1118: 			       \aap, 322, 167
1119:  \bibitem[Bonanos et al.(2004)]{bonanos04} Bonanos, A.~Z., et al.\ 2004, \apjl, 611, L33 
1120:  \bibitem[Broos et al.(2007)]{broos07} Broos, P.~S., Feigelson, E.~D., Townsley, L.~K.,
1121:  Getman, K.~V., Wang, J., Garmire, G.~P., Jiang, Z., \& Tsuboi, Y.\ 2007, \apj, in press
1122:  \bibitem[Cannon \& Pickering(1919)]{cannon19} Cannon, A.~J., \& Pickering, E.~C.\ 1919,
1123: 			       Annals of Harvard College Observatory, 94, 1
1124:  \bibitem[Carpenter(2001)]{carpenter01} Carpenter, J.~M.\ 2001, \aj, 121, 2851 
1125:  \bibitem[Carraro \& Munari(2004)]{carraro04} Carraro, G.~\&  Munari, U.\ 2004, \mnras,
1126: 			       347, 625
1127:  \bibitem[Churchwell(2002)]{churchwell02} Churchwell, E.\ 2002, \araa, 40, 27 
1128:  \bibitem[Churchwell et al.(2004)]{churchwell04} Churchwell, E., et al.\ 2004,
1129: 			       \apjs, 154, 322 
1130:  \bibitem[Damiani et al.(2004)]{damiani04} Damiani, F., Flaccomio, E., Micela, G.,
1131: 			       Sciortino, S., Harnden, F.~R., Jr., \& Murray, S.~S.\
1132: 			       2004, \apj, 608, 781 
1133:  \bibitem[Delgado et al.(2006)]{delgado06} Delgado, A.~J., Gonz{\'a}lez-Mart{\'{\i}}n,
1134: 			       O., Alfaro, E.~J., \& Yun, J.\ 2006, \apj, 646, 269
1135:  \bibitem[Dreyer(1888)]{dreyer88} Dreyer, J.~L.~E.\ 1888, \memras, 49, 1 
1136:  \bibitem[Drilling \& Landolt(2000)]{drilling00} Drilling,~J.~S. \& Landolt,~A.~U.\
1137:                                2000, in Allen's Astrophysical Quantities, ed. A.~N.~Cox
1138:                                (New York: Springer-Verlag), 381
1139:  \bibitem[Ebisawa et al.(2005)]{ebisawa05} Ebisawa, K., et al.\ 2005, \apj, 635, 214 
1140:  \bibitem[Evans et al.(2003)]{evans03} Evans, N.~R., Seward, F.~D., Krauss, M.~I.,
1141: 			       Isobe, T., Nichols, J., Schlegel, E.~M., \& Wolk, S.~J.\
1142: 			       2003, \apj, 589, 509
1143:  \bibitem[Evans et al.(2004)]{evans04} Evans, N.~R., Schlegel, E.~M., Waldron, W.~L.,
1144: 			       Seward, F.~D., Krauss, M.~I., Nichols, J., \& Wolk,
1145: 			       S.~J.\ 2004, \apj, 612, 1065
1146:  \bibitem[Fazio et al.(2004)]{fazio04} Fazio, G.~G., et al.\ 2004, \apjs, 154, 10 
1147:  \bibitem[Feigelson et al.(2005)]{feigelson05} Feigelson, E.~D., et al.\ 2005, \apjs,
1148: 			       160, 379
1149:  \bibitem[Feigelson et al.(2007)]{feigelson07} Feigelson, E., Townsley, L., G{\"u}del,
1150: 			       M., \& Stassun, K.\ 2007, Protostars and Planets V, 313,
1151: 			       ed. B.~Reipurth, D.~Jewitt, \& K.~Keil (Tuscon, Univ of
1152: 			       Arizona Press), 313
1153:  \bibitem[Flaccomio et al.(2006)]{flaccomio06} Flaccomio, E., Micela, G., \& Sciortino,
1154: 			       S.\ 2006, \aap, 455, 903 
1155:  \bibitem[Garmire et al.(2003)]{garmire03} Garmire, G.~P., Bautz, M.~W., Ford, P.~G.,
1156: 			       Nousek, J.~A., \& Ricker, G.~R.\ 2003, \procspie, 4851, 28
1157:  \bibitem[Gagn{\'e} et al.(2005)]{gagne05} Gagn{\'e}, M., Oksala, M.~E., Cohen, D.~H.,
1158:                                  Tonnesen, S.~K., ud-Doula, A., Owocki, S.~P., Townsend,
1159:                                  R.~H.~D., \& MacFarlane, J.~J.\ 2005, \apj, 628, 986
1160:  \bibitem[Gehrels(1986)]{gehrels86} Gehrels, N.\ 1986, \apj, 303, 336 
1161:  \bibitem[Getman et al.(2005)]{getman05} Getman, K.~V., et al.\ 2005, \apjs, 160, 319
1162:  \bibitem[Getman et al.(2006)]{getman06} Getman, K.~V., Feigelson, E.~D., Townsley, L.,
1163: 			       Broos, P., Garmire, G., \& Tsujimoto, M.\ 2006, \apjs,
1164: 			       163, 306 
1165:  \bibitem[Getman et al.(2007a)]{getman07a} Getman, K.~V., Feigelson, E.~D., Garmire, G.,
1166: 			       Broos, P., \& Wang, J.\ 2007, \apj, 654, 316
1167:  \bibitem[Getman et al.(2007b)]{getman07b} Getman, K.~V., et al. in prep.
1168:  \bibitem[Giardino et al.(2004)]{giardino04} Giardino, G., Favata, F., \& Micela, G.\
1169: 			       2004, \aap, 424, 965
1170:  \bibitem[Goldwurm et al.(1987)]{goldwurm87} Goldwurm, A., Caraveo, P.~A., \& Bignami,
1171: 			       G.~F.\ 1987, \apj, 322, 349
1172:  \bibitem[G\"{u}del et al.(2007)]{guedel07} G\"{u}del, M., et al.\ 2007, \aap, in press
1173:  \bibitem[Gum(1955)]{gum55} Gum, C.~S.\ 1955, \memras, 67, 155 
1174:  \bibitem[Haisch et al.(2001)]{haisch01} Haisch, K.~E., Lada, E.~A., Pi{\~n}a, R.~K.,
1175:                                Telesco, C.~M., \& Lada, C.~J.\ 2001, \aj, 121, 1512 
1176:  \bibitem[Hofner et al.(2002)]{hofner02} Hofner, P., Delgado, H., Whitney, B.,
1177: 			       Churchwell, E., \& Linz, H.\ 2002, \apjl, 579, L95
1178:  \bibitem[Homeier \& Alves(2005)]{homeier05} Homeier, N.~L., \& Alves, J.\ 2005, \aap, 430, 481
1179:  \bibitem[Hong et al.(2004)]{hong04} Hong, J., Schlegel, E.~M., \& Grindlay, J.~E.\
1180: 			       2004, \apj, 614, 508
1181:  \bibitem[Kesteven(1968)]{kesteven68} Kesteven, M.~J.~L.\ 1968, Australian Journal of
1182: 			       Physics, 21, 369
1183:  \bibitem[Lada \& Adams(1992)]{lada92} Lada, C.~J., \& Adams, F.~C.\ 1992, \apj, 393, 278 
1184:  \bibitem[Linsky et al.(2007)]{linsky07} Linsky, J.~L., Gagn{\'e}, M., Mytyk, A.,
1185: 			       McCaughrean, M., \& Andersen, M.\ 2007, \apj, 654, 347
1186:  \bibitem[Lucy(1974)]{lucy74} Lucy, L.~B.\ 1974, \aj, 79, 745 
1187:  \bibitem[Lucy \& White(1980)]{lucy80} Lucy, L. B., \& White, R. 1980, \apj, 241, 300
1188:  \bibitem[Meyer, Calvet, \& Hillenbrand(1997)]{meyer97} Meyer, M. R., Calvet, N.,
1189:                            \& Hillenbrand, L. A. 1997, \aj, 114, 288
1190:  \bibitem[Moffat \& Vogt(1975)]{moffat75} Moffat, A.~F.~J., \& Vogt, N.\ 1975, \aaps,
1191: 			       20, 125
1192:  \bibitem[Moffat et al.(1991)]{moffat91} Moffat, A.~F.~J., Shara, M.~M., \&
1193: 			       Potter, M.\ 1991, \aj, 102, 642 
1194:  \bibitem[Moffat et al.(2002)]{moffat02} Moffat, A.~F.~J., et al.\ 2002, \apj, 573, 191
1195:  \bibitem[Morris et al.(2000)]{morris00} Morris, P.~W., van der Hucht, K.~A., Crowther,
1196: 			       P.~A., Hillier, D.~J., Dessart, L., Williams,  P.~M., \&
1197: 			       Willis, A.~J.\ 2000, \aap, 353, 624
1198:  \bibitem[Muno et al.(2006)]{muno06} Muno, M.~P., et al.\ 2006, \apjl, 636, L41 
1199:  \bibitem[Nagayama et al.(2003)]{nagayama03} Nagayama, T., et al.\ 2003, \procspie,
1200: 			       4841, 459
1201:  \bibitem[Nakajima et al.(2005)]{nakajima05} Nakajima, Y., et al.\ 2005, \aj, 129, 776 
1202:  \bibitem[Piatti et al.(1998)]{piatti98} Piatti, A.~E., Bica, E., \& Claria, J.~J.\
1203: 			       1998, \aaps, 127, 423 
1204:  \bibitem[Preibisch et al.(2005)]{preibisch05} Preibisch, T., et al.\ 2005, \apjs, 160, 401
1205:  \bibitem[Rauw et al.(2005)]{rauw05} Rauw, G., et al.\ 2005, \aap, 432, 985 
1206:  \bibitem[Rauw et al.(2007)]{rauw07} Rauw, G., Manfroid, J., Gosset, E., Naz{\'e}, Y.,
1207: 			       Sana, H., De Becker, M., Foellmi, C., \& Moffat,
1208: 			       A.~F.~J.\ 2007, \aap, 463, 981 
1209:  \bibitem[Rho et al.(2004)]{rho04} Rho, J., Ram{\'{\i}}rez, S.~V., Corcoran, M.~F.,
1210: 			       Hamaguchi, K., \& Lefloch, B.\ 2004, \apj, 607, 904
1211:  \bibitem[Rogers et al.(1960)]{rogers60} Rodgers, A.~W., Campbell, C.~T., \& Whiteoak,
1212: 			       J.~B. 1960, MNRAS, 121, 103
1213:  \bibitem[Sanchawala et al.(2007)]{sanchawala06} Sanchawala, K., Chen, W.-P., Lee,
1214: 			       H.-T., Chu, Y.-H., Nakajima, Y., Tamura, M., Baba, D., \&
1215: 			       Sato, S.\ 2007, \apj, 656, 462 
1216:  \bibitem[Schulz et al.(2003)]{schulz03} Schulz, N.~S., Canizares, C., Huenemoerder, D.,
1217:                                  \& Tibbets, K.\ 2003, \apj, 595, 365 
1218:  \bibitem[Schulz et al.(2006)]{schulz06} Schulz, N.~S., Testa, P., Huenemoerder, D.~P.,
1219: 			       Ishibashi, K., \& Canizares, C.~R.\ 2006, \apj, 653, 636 
1220:  \bibitem[Seward et al.(1979)]{seward79} Seward, F.~D., Forman, W.~R., Giacconi, R.,
1221: 			       Griffiths, R.~E., Harnden, F.~R., Jr., Jones, C., \& Pye,
1222: 			       J.~P.\ 1979, \apjl, 234, L55 
1223:  \bibitem[Shara et al.(1991)]{shara91} Shara, M.~M., Smith, L.~F., Potter, M., \&
1224: 			       Moffat, A.~F.~J.\ 1991, \aj, 102, 716
1225:  \bibitem[Siess et al.(2000)]{siess00} Siess, L., Dufour, E., \& Forestini, M.\ 2000,
1226: 			       \aap, 358, 593
1227:  \bibitem[Skinner et al.(2003)]{skinner03} Skinner, S., Gagn{\'e}, M., \& Belzer, E.\
1228: 			       2003, \apj, 598, 37 5
1229:  \bibitem[Skinner et al.(2005)]{skinner05} Skinner, S.~L., Zhekov, S.~A., Palla, F., \&
1230: 			       Barbosa, C.~L.~D.~R.\ 2005, \mnras, 361, 191
1231:  \bibitem[Skinner et al.(2006)]{skinner06} Skinner, S.~L., Simmons, A.~E., Zhekov,
1232: 			       S.~A., Teodoro, M., Damineli, A., \& Palla, F.\ 2006,
1233: 			       \apjl, 639, L35 
1234:  \bibitem[Skinner et al.(2007)]{skinner07} Skinner, S.~L., Simmons, A.~E., Audard, M.,
1235: 			       \& Guedel, M.\ 2007, \apj, in press
1236:  \bibitem[Skrutskie et al.(1997)]{skrutskie97} Skrutskie, M.~F., et al.\ 1997, ASSL
1237:                                Vol.~210: The Impact of Large Scale Near-IR Sky Surveys,
1238:                                25 
1239:  \bibitem[Smith et al.(2001)]{smith01} Smith, R.~K., Brickhouse, N.~S., Liedahl, D.~A.,
1240: 			       \& Raymond, J.~C.\ 2001, \apjl, 556, L91 
1241:  \bibitem[Stelzer et al.(2005)]{stelzer05} Stelzer, B., Flaccomio, E., Montmerle, T.,
1242:                                  Micela, G., Sciortino, S., Favata, F., Preibisch, T.,
1243:                                  \& Feigelson, E.~D.\ 2005, \apjs, 160, 557 
1244:  \bibitem[Stetson(1987)]{stetson87} Stetson, P.~B.\ 1987, \pasp, 99, 191 
1245:  \bibitem[Tokunaga(2000)]{tokunaga00} Tokunaga, A. T. 2000, in Allen's Astrophysical
1246: 			       Quantities, ed. A. N. Cox (4th ed.; New York: Springer),
1247: 			       143
1248:  \bibitem[Townsley et al.(2002)]{townsley02} Townsley, L.~K., Broos, P.~S., Nousek,
1249: 			       J.~A., \& Garmire, G.~P.\ 2002, Nuclear Instruments and
1250: 			       Methods in Physics Research A, 486, 751 
1251:  \bibitem[Townsley et al.(2003)]{townsley03} Townsley, L.~K., Feigelson, E.~D.,
1252:                                  Montmerle, T., Broos, P.~S., Chu, Y.-H., \& Garmire,
1253:                                  G.~P.\ 2003, \apj, 593, 874 
1254:  \bibitem[Townsley et al.(2004)]{townsley04} Townsley, L.~K., Feigelson, E.~D,
1255: 			       Montmerle, T., Broos, P.~S, Chu, Y.-H., Garmire, G.~P.,
1256: 			       \& Getman, K.~V.\ 2004, Proc. X-Ray and Radio
1257: 			       Connections (astro-ph/0406349)
1258:  \bibitem[Townsley et al.(2006a)]{townsley06} Townsley, L.~K., Broos, P.~S., Feigelson,
1259: 			       E.~D., Garmire, G.~P., \& Getman, K.~V.\ 2006, \aj, 131,
1260: 			       2164 
1261:  \bibitem[Townsley(2006b)]{townsley06b} Townsley, L.~K.\ 2006, Proc. Massive Stars:
1262: 			       From Pop III and GRBs to the Milky Way (astro-ph/0608173)
1263:  \bibitem[Tsujimoto et al.(2005)]{tsujimoto05} Tsujimoto, M., Kobayashi, N., \& Tsuboi,
1264: 			       Y.\ 2005, \aj, 130, 2212
1265:  \bibitem[Tsunemi et al.(2001)]{tsunemi01} Tsunemi, H., Mori, K., Miyata, E., Baluta,
1266: 			       C., Burrows, D.~N., Garmire, G.~P., \& Chartas,
1267: 			       G.\  2001, \apj, 554, 496 
1268:  \bibitem[Uzpen et al.(2005)]{uzpen05} Uzpen, B., et al.\ 2005, \apj, 629, 512 
1269:  \bibitem[van der Hucht et al.(1986)]{vanderhucht86} van der Hucht, K.~A., Cassinelli,
1270: 			       J.~P., \& Williams, P.~M.\ 1986, \aap, 168, 111
1271:  \bibitem[van der Hucht(2001)]{vanderhucht01} van der Hucht, K.~A.\ 2001, New Astronomy
1272: 			       Review, 45, 135
1273:  \bibitem[Wang et al.(2006)]{dqwang06} Wang, Q.~D., Dong, H., \& Lang, C.\ 2006, \mnras, 371, 38
1274:  \bibitem[Wang et al.(2007a)]{jwang07a} Wang, J., Townsley, L.~K., Feigelson, E.~D.,
1275: 			       Getman, K.~V., Broos, P.~S., Garmire, G.~P., \&
1276: 			       Tsujimoto, M.\ 2007, \apjs, 168, 100 
1277:  \bibitem[Wang et al.(2007b)]{jwang07b} Wang, J., et al. in prep.
1278:  \bibitem[Weisskopf et al.(2002)]{weisskopf02} Weisskopf, M.~C., Brinkman, B.,
1279: 			       Canizares, C., Garmire, G., Murray, S., \& Van
1280: 			       Speybroeck, L.~P.\ 2002, \pasp, 114, 1
1281:  \bibitem[Werner et al.(2004)]{werner04} Werner, M.~W., et al.\ 2004, \apjs, 154, 1 
1282:  \bibitem[Westerlund(1960)]{westerlund60} Westerlund, B.\ 1960, Arkiv for Astronomi, 2, 419 
1283:  \bibitem[Whiteoak \& Uchida(1997)]{whiteoak97} Whiteoak, J.~B.~Z., \& Uchida, K.~I.\
1284: 			       1997, \aap, 317, 563 
1285:  \bibitem[Whitney et al.(2004)]{whitney04} Whitney, B.~A., et al.\ 2004, \apjs, 154, 315 
1286:  \bibitem[Wolk et al.(2006)]{wolk06} Wolk, S.~J., Spitzbart, B.~D., Bourke, T.~L., \&
1287: 			       Alves, J.\ 2006, \aj, 132, 1100 
1288:  \bibitem[Zhekov \& Skinner(2000)]{zhekov00} Zhekov, S.~A., \& Skinner, S.~L.\ 2000,
1289: 			       \apj, 538, 808
1290: \end{thebibliography}
1291: 
1292: \clearpage
1293: 
1294: \begin{turnpage}
1295: \input{t1_stab}
1296: \end{turnpage}
1297: 
1298: \clearpage
1299: 
1300: \begin{turnpage}
1301: \input{t2_stab}
1302: \end{turnpage}
1303: 
1304: \clearpage
1305: 
1306: \begin{turnpage}
1307: \input{t3_stab}
1308: \end{turnpage}
1309: 
1310: \clearpage
1311: 
1312: \begin{turnpage}
1313: \begin{deluxetable}{ccccccccccccc}
1314:  \centering
1315:  \tabletypesize{\tiny}
1316:  \tablewidth{0pt}
1317:  \tablecolumns{13}
1318:  \tablecaption{Detailed X-ray Spectroscopy of Bright Sources\label{tb:t4}}
1319:  \tablehead{
1320:  \colhead{X-ray\tablenotemark{a}} & 
1321:  \colhead{Name} & 
1322:  \colhead{Model\tablenotemark{b}} & 
1323:  \colhead{Label\tablenotemark{c}} &
1324:  \colhead{\nh\tablenotemark{d}} & 
1325:  \colhead{\kt$_{1}$\tablenotemark{b,d}} &
1326:  \colhead{\kt$_{2}$\tablenotemark{b,d}} &
1327:  \colhead{\emm$_{1}$/\emm$_{2}$} &
1328:  \colhead{\fx\tablenotemark{e}} &
1329:  \colhead{\lx\tablenotemark{e}} &
1330:  \colhead{Si\tablenotemark{d}} & 
1331:  \colhead{S\tablenotemark{d}} & 
1332:  \colhead{$\chi^{2}$/d.o.f.}\\
1333:  \colhead{ID} &
1334:  &
1335:  &
1336:  &
1337:  \colhead{(10$^{22}$~cm$^{-2}$)} &
1338:  \colhead{(keV)} &
1339:  \colhead{(keV)} &
1340:     &
1341:  \colhead{(ergs s$^{-1}$~cm$^{-2}$)} &
1342:  \colhead{(ergs s$^{-1}$)} &
1343:  \colhead{($Z_{\odot}$)} &
1344:  \colhead{($Z_{\odot}$)} &
1345:  }
1346:  \startdata
1347:  132 & WR\,20a & 1T & \nodata & 2.0 (1.1--2.4) & 2.0 (1.7--3.3) & \nodata & \nodata & 3.7$\times$10$^{-13}$ & 1.5$\times$10$^{33}$ & 0.7 (--2.8) & 1.7 (1.0--5.1) & 40.8/27 \\
1348:  &         & 2T & (a) & 2.1 (0.9--3.0) & 4.2 (2.2--) & 0.8 (0.08--1.1) & 0.6 & 3.9$\times$10$^{-13}$ & 2.2$\times$10$^{33}$ & 1.2 (0.3--4.1) & 4.0 (1.5--) & 25.3/24 \\
1349:  396 & WR\,20b & 1T & \nodata & 2.5 (1.9--3.4) & 3.0 (1.9--6.4) & \nodata & \nodata & 1.9$\times$10$^{-13}$ & 7.6$\times$10$^{32}$ & 1.0 (--5.3) & 1.6 (--6.4) & 33.6/18 \\
1350:  &         & 2T & (b) & 3.9 (2.2--5.8) & 5.5 (2.8--) & 0.48 (0.30--0.72) & 0.1 & 2.2$\times$10$^{-13}$ & 1.8$\times$10$^{33}$ & 1.2 (0.3--5.0) & 3.4 (0.69--) & 20.3/17 \\
1351:  \hline
1352:  192 & MSP\,188 & 1T & (c) & 0.31 (0.11-0.56) & 3.6 (2.6--6.2) & \nodata & \nodata & 1.2$\times$10$^{-13}$ & 2.9$\times$10$^{32}$ & 6.2 (2.4--11) & 2.1 (0.0--8.0) & 41.7/10 \\
1353:  217 & MSP\,167 & 1T & (d) & 1.4 (0.2--2.8) & 1.4 (0.9--4.7) & \nodata & \nodata & 6.1$\times$10$^{-14}$ & 4.9$\times$10$^{32}$ & 0.9 (0.2--3.1) & 0.5 (--4.7) & 12.5/9 \\
1354:  229 & MSP\,18  & 1T & (e) & 0.63 (0.30--0.99) & 4.1 (2.6--8.9) & \nodata & \nodata & 2.7$\times$10$^{-13}$ & 7.2$\times$10$^{32}$ & \nodata & \nodata & 17.0/18 \\
1355: \enddata
1356:  \tablenotetext{a}{Source numbers follow Table~\ref{tb:t1}.}
1357:  \tablenotetext{b}{``1T'' and ``2T'' indicate the fits by an absorbed one-temperature
1358:  and two-temperature thin-thermal plasma (\texttt{vapec}) model, respectively. The
1359:  amount of extinction (\nh), the plasma temperature (\kt$_{1}$, \kt$_{2}$), X-ray flux
1360:  (\fx) and luminosity (\lx), the abundance of Si and S relative to the cosmic values,
1361:  and the goodness of the fit ($\chi^{2}$/d.o.f.) are derived. The cosmic abundance
1362:  \citep{anders89} is used. For the 2T fits, \kt$_{1}$ and \kt$_{2}$ represent the
1363:  higher and lower temperature component. The \emm\ ratio of the two components is
1364:  given as \emm$_{1}$/\emm$_{2}$.}
1365:  \tablenotetext{c}{The label of figures for the spectra and the best-fit models in
1366:  Figure~\ref{fg:f11}.}
1367:  \tablenotetext{d}{Uncertainties of a 90\% confidence range are given in parentheses.}
1368:  \tablenotetext{e}{Values in the 1.0--8.0~keV range. A distance of 4.2~kpc is assumed
1369:  to derive \lx.}
1370: \end{deluxetable}
1371: \end{turnpage}
1372: 
1373: \end{document}
1374: