1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3: \usepackage{apjfonts}
4: \usepackage{graphicx}
5: \usepackage{amsmath}
6: \usepackage{amssymb}
7: \usepackage{amsfonts}
8: \usepackage{natbib} % This is required to use the apj.bst
9: \citestyle{aa} % and be able to use the AASTEX 5.x
10: % citations (\citep, \citet, etc.) - bph
11: \bibliographystyle{apj}
12:
13: \shorttitle{Transient Black-Hole LMXB in Cen A}
14: \shortauthors{SIVAKOFF ET AL.}
15:
16: \slugcomment{Accepted for publication in ApJ Letters}
17:
18: \begin{document}
19:
20: \title{A Transient Black-Hole Low-Mass X-Ray Binary Candidate in Centaurus A}
21:
22:
23: \author{
24: G.~R.~Sivakoff\altaffilmark{1},
25: R.~P.~Kraft\altaffilmark{2},
26: A.~Jord\'{a}n\altaffilmark{2,3},
27: A.~M.~Juett\altaffilmark{4},
28: D.~A.~Evans\altaffilmark{2},
29: W.~R.~Forman\altaffilmark{2},
30: M.~J.~Hardcastle\altaffilmark{5},
31: C.~L.~Sarazin\altaffilmark{6},
32: M.~Birkinshaw\altaffilmark{7,2},
33: N.~J.~Brassington\altaffilmark{2},
34: J.~H.~Croston\altaffilmark{5},
35: W.~E.~Harris\altaffilmark{8},
36: C.~Jones\altaffilmark{2},
37: S.~S.~Murray\altaffilmark{2},
38: S.~Raychaudhury\altaffilmark{9,2},
39: K.~A.~Woodley\altaffilmark{8},
40: D.~M.~Worrall\altaffilmark{7,2}
41: }
42: \altaffiltext{1}{%
43: Department of Astronomy,
44: The Ohio State University,
45: 4055 McPherson Laboratory
46: 140 W. 18th Avenue, Columbus, OH 43210-1173, USA;
47: sivakoff@astronomy.ohio-state.edu%
48: }
49: \altaffiltext{2}{%
50: Harvard-Smithsonian Center for Astrophysics,
51: 60 Garden Street,
52: MS-67, Cambridge, MA 02138, USA%;
53: }
54: \altaffiltext{3}{%
55: Clay Fellow%
56: }
57: \altaffiltext{4}{%
58: NASA Postdoctoral Fellow,
59: Laboratory for X-ray Astrophysics,
60: NASA Goddard Space Flight Center,
61: Greenbelt, MD 20771, USA%;
62: }
63: \altaffiltext{5}{%
64: School of Physics, Astronomy, and Mathematics,
65: University of Hertfordshire,
66: Hatfield AL10 9AB, UK%;
67: }
68: \altaffiltext{6}{%
69: Department of Astronomy,
70: University of Virginia,
71: P. O. Box 400325,
72: Charlottesville, VA 22904-4325, USA%;
73: }
74: \altaffiltext{7}{%
75: Department of Physics,
76: University of Bristol,
77: Tyndall Avenue,
78: Bristol BS8 ITL, UK%;
79: }
80: \altaffiltext{8}{%
81: Department of Physics and Astronomy,
82: McMaster University,
83: Hamilton, ON L8S 4M1, Canada%;
84: }
85: \altaffiltext{9}{%
86: School of Physics and Astronomy,
87: University of Birmingham,
88: Edgbaston, Birmingham B15 2TT, UK%;
89: }
90:
91:
92: \begin{abstract}
93:
94:
95: We report the discovery of a bright transient X-ray source, CXOU
96: J132518.2$-$430304, towards Centaurus A (Cen A) using six new {\it
97: Chandra X-Ray Observatory} observations in 2007 March--May. Between 2003
98: and 2007, its flux has increased by a factor of $>770$.
99: The source is likely a low-mass X-ray binary in Cen A with unabsorbed
100: 0.3--10 keV band luminosities of $(2$--$3) \times 10^{39} {\rm \, erg
101: \, s}^{-1}$ and a transition from the steep-power law state to the thermal
102: state during our observations. CXOU J132518.2$-$430304 is the most luminous
103: X-ray source in an early-type galaxy with extensive timing information that
104: reveals transience and a spectral state transition. Combined with its
105: luminosity, these properties make this source one of the strongest candidates to
106: date for containing a stellar-mass black hole in an early-type galaxy. Unless
107: this outburst lasts many years, the rate of luminous transients in Cen A is
108: anomalously high compared to other early-type galaxies.
109: \end{abstract}
110: \keywords{
111: binaries: close ---
112: galaxies: elliptical and lenticular, cD ---
113: galaxies: individual (Centaurus A, NGC 5128) ---
114: X-rays: binaries
115: }
116:
117: \section{Introduction}
118:
119: Over the 40 years of X-ray astronomy, the detailed study of Galactic X-ray
120: binaries (XRBs) has placed strong constraints on theories of X-ray binary
121: evolution and accretion. X-ray variability studies have been important in
122: understanding disk accretion and the interplay between different emission
123: components, e.g., jet, corona, and disk.
124:
125: All Galactic low mass X-ray binaries (LMXBs), which have companions of
126: $\lesssim 1 M_{\odot}$, that are confirmed BHs are also transient systems
127: with X-ray luminosities that vary by orders of magnitude \citep{RM2006}.
128: While most have outbursts ranging from weeks to months, one unusual transient,
129: GRS~1915$+$105 has been continuously emitting since its outburst began 15 years
130: ago. The transient behavior of BH (and some neutron star) LMXBs has been
131: attributed to a disk instability first identified in cataclysmic variables
132: \citep[CVs;][]{P1996}. While this theory correctly identifies which XRBs are
133: likely transients, it has yet to accurately predict the outburst lengths and
134: recurrence times, which are necessary to determine an X-ray duty cycle
135: \citep[for a review, see][]{K2006}.
136:
137: The Galactic population of BH LMXBs has been studied in great detail, but the
138: small sample size, $\sim$190 Galactic LMXBs of which $\sim10$--25\% are
139: confirmed or candidate BH systems \citep{LPH2006}, and non-uniform observational
140: sampling limit the ability of Galactic studies to measure the duty cycle. With
141: the {\em Chandra X-ray Observatory}, we can now study the tens to hundreds of
142: identified LMXBs in individual nearby galaxies
143: \citep[e.g.,][]{SIB2000,SJS+2007}. With multi-epoch observations of early-type
144: galaxies, we can measure the variability of extragalactic LMXBs. Interestingly,
145: the most luminous extragalactic LMXBs are persistent on timescales of several
146: years \citep{Ir2006}.
147:
148: Centaurus A (NGC~5128, Cen A), at $3.7{\rm \, Mpc}$ \citep[the average of 5 distance
149: indicators, see \S~6 in][]{FMS+2007}, is the nearest radio galaxy and the nearest optically
150: luminous \citep[$M_B = -21.1$;][]{DHM+1979} early-type galaxy. Prior to 2007,
151: {\it Chandra} had targeted Cen A with the ACIS detectors four times, 1999
152: December 5, 2000 May 17, 2002 September 3, and 2003 September 14 (Observations
153: \dataset[ADS/Sa.CXO#obs/00316]{316},
154: \dataset[ADS/Sa.CXO#obs/00962]{962},
155: \dataset[ADS/Sa.CXO#obs/02978]{2978},
156: and
157: \dataset[ADS/Sa.CXO#obs/03965]{3965}).
158: In 2007, {\it Chandra} performed six further deep observations ($\sim 100 {\rm
159: \, ks}$ each) of Cen A to study the X-ray properties of its jet, lobes, and XRBs
160: \citep{HKS+2007,JSM+2007,WBK+2007,KHS+2008s}.
161: As the nearest
162: luminous early-type galaxy, Cen A is one of the premier targets for studying
163: extragalactic LMXB variability. In this analysis, we report results on the most
164: luminous XRB candidate in Cen A, CXOU J132518.2$-$430304.
165: The errors on our spectral fits refer to
166: single-parameter $ 90\% $ confidence intervals and all other errors refer to $1
167: \sigma $ confidence intervals. All fluxes and luminosities are in the
168: 0.3--10 keV band.
169:
170: \section{Observations and Data Reduction}
171: \label{sec:obs}
172:
173: \begin{deluxetable}{llcccc}
174: \tabletypesize{\footnotesize}
175: \tablewidth{0pt}
176: \tablecaption{Chandra Observations of CXOU J132518.2$-$430304 in Cen A\label{tab:obs}}
177: \tablehead{
178: \colhead{Obs.} &
179: \colhead{Date} &
180: \colhead{Exposure} &
181: \colhead{OAA} &
182: \colhead{Net Counts} &
183: \colhead{Net Rate}\\
184: \colhead{(\#)}&
185: \colhead{(2007)} &
186: \colhead{(${\rm s}$)} &
187: \colhead{($\arcmin$)} &
188: \colhead{(${\rm cnt}$)} &
189: \colhead{($10^{-2} {\rm \, cnt \, s}^{-1}$)}
190: }
191: \startdata
192: \dataset[ADS/Sa.CXO#obs/07797]{7797} & March 22 & 96888 & 0.9 & 6471 & $6.679\pm0.084$ \\
193: \dataset[ADS/Sa.CXO#obs/07798]{7798} & March 27 & 90839 & 7.3 & 7942 & $8.742\pm0.100$ \\
194: \dataset[ADS/Sa.CXO#obs/07799]{7799} & March 30 & 94783 & 7.3 & 6981 & $7.365\pm0.090$ \\
195: \dataset[ADS/Sa.CXO#obs/07800]{7800} & April 17 & 90843 & 7.4 & 6259 & $6.890\pm0.089$ \\
196: \dataset[ADS/Sa.CXO#obs/08489]{8489} & May 8 & 93936 & 3.3 & 8987 & $9.567\pm0.102$ \\
197: \dataset[ADS/Sa.CXO#obs/08490]{8490} & May 30 & 94425 & 0.5 & 7908 & $8.375\pm0.095$
198: \enddata
199: \tablecomments{Observed counts are in the 0.5--$7 {\rm \, keV}$ band.
200: The source in observations 7797, 8489, and 8490 is at small off-axis angles
201: (OAAs), and their rates are underestimated due to event pileup.}
202: \end{deluxetable}
203: \begin{figure*}
204: \center
205: \includegraphics[height=170pt]{f1a.eps}\includegraphics[height=170pt]{f1b.eps}
206: \caption{
207: {\it Left:} {\it Chandra} 0.3--2 keV image of Cen A from Observation 7797, with
208: a logarithmic intensity scale and Gaussian smoothing (FWHM=3 pixels).
209: The image is limited to soft energies
210: to minimize the readout streak from the central AGN.
211: The $\sim 30\arcsec \times 30\arcsec$ box indicates
212: the location of CXOU J132518.2$-$430304.
213: {\it Right:}
214: {\it Chandra} 0.3--10 keV subimages
215: ($30\arcsec \times 30\arcsec$)
216: of CXOU J132518.2$-$430304 with the source extraction
217: region indicated for all {\it Chandra} observations except
218: 0962 and 2978, both of which show no source.
219: The logarithmic intensity scale
220: has been changed.
221: The source clearly went into outburst between
222: Observations 3965 and 7797 and has remained bright for at least
223: 70 days. The source regions due to the varying PSFs are overlaid.
224: \label{fig:image}}
225: \end{figure*}
226:
227: New {\it Chandra} observations of Cen~A were taken
228: in 2007 (Table~\ref{tab:obs}) and contained no high-background periods.
229: Our analysis includes only events with ASCA grades of 0, 2, 3, 4, and 6. Photon
230: energies were determined using the gain file acisD2000-01-29gain\_ctiN0001.fits,
231: correcting for time dependence of the gain, charge-transfer inefficiency, and
232: quantum efficiency degradation. We excluded bad pixels, bad columns, and columns
233: adjacent to bad columns or chip node boundaries.
234: The absolute astrometry is accurate to $\pm0\farcs1$ \citep{HKS+2007}.
235: We used {\sc ciao 3.4}%
236: \footnote{See \url{http://asc.harvard.edu/ciao/}.}
237: with {\sc caldb 3.3.0.1} and NASA HEASARC's {\sc ftools 6.2}%
238: \footnote{See
239: \url{http://heasarc.gsfc.nasa.gov/docs/software/lheasoft/}%
240: \label{ftn:heasoft}.} for data reduction and analysis.
241: We used ACIS Extract 3.131%
242: \footnote{See \url{http://www.astro.psu.edu/xray/docs/TARA/%
243: ae\_users\_guide.html}} to refine the position of CXOU J132518.2$-$430304
244: by correlating the 0.5--$7 {\rm \, keV}$ photons near the {\sc wavdetect}
245: coordinates against the count weighted, combined X-ray PSF (at 1.497 keV) of the
246: observations. For each observation
247: we created source regions corresponding to a polygon
248: encircling 90\% of the X-ray PSF and
249: local circular background regions that began just beyond the region encircling
250: 97\% of the X-ray PSF and had three times the source area. These regions were
251: used to extract spectra and response files for
252: spectral fitting using {\sc xspec 11.3.}\footnotemark[\ref{ftn:heasoft}]
253:
254: \section{Properties of CXOU J132518.2$-$430304}
255:
256: CXOU J132518.2$-$430304 is located $2\farcm6$ SW of the AGN in Cen A. We
257: estimate that its position of R.A.\ $= 13^{\rm h} 25^{\rm m} 18\fs24$ and Dec.\
258: $= -43\arcdeg03\arcmin04\farcs5$ (J2000) is accurate to $0\farcs2$.
259:
260: In each observation, CXOU J132518.2$-$430304 was the brightest non-nuclear
261: source.
262: At this position, there is no source in any
263: of the four previous observations of Cen A (see Figure~\ref{fig:image}).
264: In those four observations, we find a combined $3\sigma$ upper limit of 14.9 net
265: counts. The spectral models fitted to the new observations (Table~{\ref{tab:spec}}) give
266: PSF-corrected, absorbed fluxes of $(8.5$--$11.5) \times 10^{-13} {\rm \, erg \,
267: cm}^{-2} {\rm \, s}^{-1}$. The same models correspond to fluxes of $\lesssim 1.1
268: \times 10^{-15} {\rm \, erg \, cm}^{-2} {\rm \, s}^{-1}$ in the old
269: observations. The flux has increased by a factor of $>770$.
270:
271: CXOU J132518.2$-$430304 is a transient source, whose outburst began after 2003
272: September 14. Since it is detected in all new observations, the outburst
273: duration is at least 70 days.
274:
275: We fitted the $0.5$--$10 {\rm \, keV}$
276: spectra of CXOU J132518.2$-$430304, grouping spectral channels
277: to have $\ge 25$
278: counts. Since the spectra of Galactic XRBs are often fitted with the combination
279: of a power-law and a multi-color disk blackbody model \citep{RM2006}, we adopted
280: this model attenuated by
281: a fixed Galactic absorption term
282: \citep[$N_H= 8.41 \times 10^{20} {\rm \, cm}^{2}$;][]{DL1990}
283: and a variable local absorption term $N_{\rm H, local}$.
284: The power-law photon index
285: $\Gamma$, the inner temperature of the disk $kT_{\rm diskbb}$, and the
286: normalizations of the two components were also variable parameters.
287: Given the high count rates, multiple
288: photons may be associated with an event, distorting the measured spectrum.
289: Therefore, we convolved the physical model with the pileup model of
290: \citet{D2001}.
291: Its key parameters are the grade-morphing
292: parameter $\alpha$ and the number of independent regions $N_{\rm reg,p}$ over
293: which pileup is calculated. We assumed $\alpha=0.5$ for all fits, which was
294: consistent with fits where $\alpha$ was allowed to vary. With their off-axis
295: PSF, we determined that $N_{\rm reg,p}=5$ for Observations 7798--7800 and pileup
296: is minimal ($\lesssim 4\%$). For the nearly on-axis Observations 7797, 8489, and
297: 8490, $N_{\rm reg,p}=1$ and pileup is a concern ($15$--$23\%$).
298:
299: We first attempted to jointly fit the spectra, requiring a fixed physical
300: spectral shape, but allowing the normalizations to vary between
301: observations. This fit is unacceptable with $\chi^2 = 1248.4$ for 999 degrees of
302: freedom (dof), suggesting that there is spectral variability
303: between the six observations. We next allowed each
304: observation to have its own spectral shape. We present the spectra in
305: Figure~{\ref{fig:spec}} and summarize those fits in Table~{\ref{tab:spec}}.
306: Added together, the $\chi^2 = 992.1$ for 979 dof is an acceptable fit
307: that is significantly
308: preferred over models with only an absorbed power-law ($\chi^2 = 1239.9$ for 991 dof) or absorbed disk blackbody ($\chi^2 = 1416.1$ for 991 dof).
309: Using error-weighted averages, it is clear that the first four observations are
310: dominated, $L_{\rm unabs,pow}/L_{\rm unabs,tot} = 0.77\pm0.05$, by a power-law
311: component, $\Gamma = 2.29\pm0.08$, attenuated by a $N_{\rm H, local}$ of
312: $(2.3\pm0.5)
313: \times 10^{21} {\rm \, cm}^{-2}$.
314: The sub-dominant disk component has a temperature that is unconstrained in
315: observations 7797 and 7800, but that observations 7798 and 7799 find to be $k
316: T_{\rm diskbb} = 0.9 \pm 0.2$~keV. In the last two observations, a disk
317: blackbody component, $kT_{\rm diskbb} = 1.01\pm0.03 {\rm \, keV}$, dominates,
318: $L_{\rm unabs,pow}/L_{\rm unabs,tot} = 0.10\pm0.06$, over a poorly constrained
319: power-law component. In addition, the $N_{\rm H, local}$ is also lower,
320: $(0.5\pm0.3) \times 10^{21} {\rm \, cm}^{-2}$.
321: In Observation 8489, a very steep
322: power-law component with high absorption leads to poor constraints on
323: column density and $L_{\rm unabs,pow}/L_{\rm unabs,tot}$.
324:
325: \begin{deluxetable*}{llllllllll}
326: \tabletypesize{\footnotesize}
327: \tablewidth{0pt}
328: \tablecaption{X-ray Spectral Fits to Individual Observations of
329: CXOU J132518.2$-$430304\label{tab:spec}}
330: \tablehead{
331: \colhead{Obs.} &
332: \colhead{$L_{\rm abs,tot}$} &
333: \colhead{$N_{H{\rm ,local}}$} &
334: \colhead{$L_{\rm unabs,tot}$} &
335: \colhead{$\Gamma$} &
336: \colhead{$L_{\rm unabs,pow}$} &
337: \colhead{$kT_{\rm diskbb}$} &
338: \colhead{$L_{\rm unabs,diskbb}$} &
339: \colhead{$L_{\rm unabs,pow}/L_{\rm unabs,tot}$} &
340: \colhead{$\chi^2/{\rm dof}$}\\
341: &
342: \colhead{($10^{39} {\rm \, erg \, s}^{-1}$)} &
343: \colhead{($10^{21} {\rm \, cm}^{-2}$)} &
344: \colhead{($10^{39} {\rm \, erg \, s}^{-1}$)} &
345: &
346: \colhead{($10^{39} {\rm \, erg \, s}^{-1}$)} &
347: \colhead{(keV)}&
348: \colhead{($10^{39} {\rm \, erg \, s}^{-1}$)} &
349: &
350: }
351: \startdata
352: 7797 & $1.49^{+0.05}_{-0.05}$ & $2.8^{+1.2}_{-1.3}$ & $2.76^{+1.12}_{-0.64}$ & $2.34^{+0.66}_{-0.44}$ & $2.63^{+0.57}_{-1.03}$ & $1.23 \, [{\rm unc}] $ & $0.13 [<0.68] $ & $0.95 [>0.75] $ & 140.2/151\\
353: 7798 & $1.69^{+0.06}_{-0.05}$ & $2.3^{+1.2}_{-1.0}$ & $2.91^{+1.14}_{-0.51}$ & $2.45^{+0.67}_{-0.37}$ & $2.20^{+0.79}_{-0.57}$ & $1.20^{+0.28}_{-0.31}$ & $0.71^{+0.39}_{-0.08}$ & $0.76^{+0.05}_{-0.08}$ & 159.4/168\\
354: 7799 & $1.50^{+0.04}_{-0.04}$ & $2.4^{+1.4}_{-0.6}$ & $3.42^{+1.65}_{-0.90}$ & $2.29^{+0.11}_{-0.08}$ & $2.60^{+0.31}_{-0.17}$ & $0.07 \, [{\rm unc}] $ & $0.81 [<2.15] $ & $0.76 [>0.57] $ & 142.6/156\\
355: 7800 & $1.39^{+0.04}_{-0.05}$ & $1.5^{+1.3}_{-1.1}$ & $1.99^{+0.59}_{-0.31}$ & $2.04^{+0.38}_{-0.57}$ & $1.50^{+0.92}_{-0.67}$ & $0.66^{+0.33}_{-0.12}$ & $0.49^{+0.36}_{-0.33}$ & $0.75^{+0.18}_{-0.26}$ & 163.1/144\\
356: 8489 & $1.88^{+0.02}_{-0.02}$ & $0.5^{+3.3}_{-0.3}$ & $2.20^{+3.69}_{-0.06}$ & $2.25 \, [{\rm unc}] $ & $0.05 [<3.75] $ & $0.99^{+0.04}_{-0.06}$ & $2.15^{+0.02}_{-0.25}$ & $0.02 [<0.64] $ & 186.4/186\\
357: 8490 & $1.80^{+0.00}_{-0.09}$ & $0.5^{+0.3}_{-0.3}$ & $2.04^{+0.05}_{-0.11}$ & $-3.00 [<0.80] $ & $0.20^{+0.05}_{-0.12}$ & $1.05^{+0.04}_{-0.06}$ & $1.84^{+0.03}_{-0.12}$ & $0.10^{+0.03}_{-0.06}$ & 200.4/174
358: \enddata
359: \tablecomments{PSF-corrected luminosities in
360: the 0.3--10 keV band are calculated assuming a distance of $3.7 {\rm \, Mpc}$.
361: Errors indicate 90\% confidence intervals. Brackets indicate parameter is
362: unconstrained in at least one direction.
363: An additional Galactic
364: $N_H = 8.41 \times 10^{20} {\rm \, cm}^{-2}$ is also included.}
365: \end{deluxetable*}
366:
367: We used the Rayleigh statistic to search observations for periodic signals
368: with frequencies between $10^{-5} {\rm \, Hz}$ and $10^{-1} {\rm \, Hz}$,
369: testing every $10^{-5} {\rm \,
370: Hz}$. We only found periodic signals
371: related to known instrumental effects.
372: We looked for additional count-rate variability within each
373: observation using the Kolmogorov-Smirnov (K-S) test. Initial tests on the
374: 0.3--$10.0 {\rm \, keV}$ band indicated 99.96\% significant variability in
375: Observation 7797 and 7800. The total flux variations are of order 10\% and 20\%,
376: respectively. Additional K-S tests on the 0.3--$1.0 {\rm \, keV}$ (soft),
377: 1.0--$2.0 {\rm \, keV}$, and 2.0--$10.0 {\rm \, keV}$ (hard) bands show that the
378: variability in Observation 7797 (a $\sim 5 \times 10^{3} {\rm \, s}$ flare and a
379: lower rate over $\sim 5 \times 10^{4} {\rm \, s}$) and
380: 7800 (a $\sim 8 \times 10^{3} {\rm \, s}$ dip and a
381: higher rate over $\sim 5 \times 10^{4} {\rm \, s}$)
382: primarily originates in the hard and soft bands, respectively.
383:
384: \begin{figure}
385: \center
386: \includegraphics[width=0.85\linewidth,clip=true]{f2.eps}
387: \caption{
388: Spectra of CXOU J132518.2$-$430304 (0.5--$10.0 {\rm \, keV}$) with the best-fit
389: spectral model for each of the new observations. All spectra have been rescaled
390: for visibility. The spectral fit parameters are listed in
391: Table~{\ref{tab:spec}}. A spectral transition from the steep power-law state to
392: the thermal dominant state appears to occur between Observations 7800 and 8489.
393: \label{fig:spec}}
394: \end{figure}
395:
396: A {\it Hubble Advanced Camera for Surveys} observation
397: \dataset[ads/Sa.HST#J8Z012010]{(J8Z012010)}
398: indicates that any counterpart has $m_{f606w} > 24.9 {\rm \, (AB)}$. From
399: {\it Hubble WFPC2} observations
400: \dataset[ads/Sa.HST#U3LBA101M,ads/Sa.HST#U3LBA102M,ads/Sa.HST#U3LBA103M,
401: ads/Sa.HST#U3LBA104M,ads/Sa.HST#U3LBA105M,ads/Sa.HST#U3LBA106M]{(U3LBA101M -- U3LBA106M)},
402: the de-reddened galaxy color $(V-I)_0$ at the position of CXOU
403: J132518.2$-$430304 is $1.13\pm0.04$, which is consistent with K giants
404: \citep{P1998}. Since O/B stars would significantly alter the measured $(V-I)$
405: color, we rule out their presence and the possibility that CXOU
406: J132518.2$-$430304 is a high-mass X-ray binary. A 2007 June Very Large Array
407: observation places a $3\sigma$, 8.4 GHz upper limit of $0.24 {\rm \, mJy}$
408: (Goodger et al.\, in preparation).
409:
410: \section{Discussion}
411:
412: CXOU J132518.2$-$430304 is a transient source with an outburst duration of $> 70
413: {\rm \, days}$. Among transient sources, its large X-ray to optical flux ratio
414: of $\log(F_X/F_{\rm opt}) \gtrsim 3.5$ is characteristic of only strongly
415: absorbed AGN, CVs, and XRBs. The local column density of $N_{\rm H,
416: local} \sim 2 \times 10^{21} {\rm \, cm}^{-2}$ is inconsistent with the source
417: being a strongly absorbed AGN. If the source is Galactic in nature, its X-ray
418: luminosity is $\lesssim 4 \times 10^{35} {\rm \, erg \, s}^{-1}$ and any
419: potential companion must be a K dwarf or later in the outskirts of the Galaxy.
420: Since the spectra are poorly fit by thermal bremsstrahlung models, we rule out
421: that CXOU J132518.2$-$430304 is a luminous Galactic CV. While the source could
422: be a Galactic LMXB, the field density of such sources is extremely low and
423: transient LMXBs are typically more luminous in outburst. Thus, it is unlikely
424: that CXOU J132518.2$-$430304 is a Galactic LMXB. Thus, we conclude that CXOU
425: J132518.2$-$430304 is most likely an XRB in Cen A. Our measured color of the
426: galaxy at the position of this source rules out the possibility of it being a
427: high-mass XRB.
428:
429: Assuming CXOU J132518.2$-$430304 is a transient LMXB in Cen A, its 0.3--10.0 keV
430: unabsorbed luminosities are $(2$--$3) \times 10^{39} {\rm \, erg \,
431: s}^{-1}$, which is above the lower limit used to define an ultraluminous X-ray
432: source in the {\it Chandra} band \citep{IBA2004}. Although a handful of Galactic
433: LMXBs with BHs also have similar peak luminosities \citep{JN2004}, CXOU
434: J132518.2$-$430304 appears to be one of the most luminous extragalactic
435: transient LMXBs discovered to date. The proximity of Cen A allows us to study
436: this LMXB in greater detail than in any other early-type galaxy.
437:
438: Since its average outburst luminosity is $\sim 17$ times the Eddington
439: limit for a solar mass object accreting ionized hydrogen, CXOU
440: J132518.2$-$430304 is more likely to be a BH-XRB than a neutron star XRB. If we
441: assume that the compact object mass is less than 30 $M_\odot$, CXOU
442: J132518.2$-$430304 is accreting at an Eddington efficiency of $\gtrsim 0.5$.
443: This source appears to be a very efficiently accreting BH-XRB.
444: Active BH-XRBs are often divided into three states, a thermal state
445: (high/soft state), a hard state (low/hard state), and the steep power-law
446: state (very high state) \citep{RM2006}. The steep power-law state is
447: characterized by a $\Gamma > 2.4$ power-law dominating over a $\sim 1 {\rm \,
448: keV}$ disk blackbody component, which is well matched to the spectra of CXOU
449: J132518.2$-$430304 in the first four observations; typically, the power-law
450: component in the hard state has $1.4 < \Gamma < 2.1$. In two of these
451: observations, we saw some evidence for intraobservation variability. The mixture
452: of bands that vary points to $\sim 10^{4} {\rm \, s}$ timescale variability in
453: both the disk and power-law components. The last two observations
454: are clearly best fit by the thermal state, with disk temperatures that are
455: consistent with Galactic BH-LMXBs \citep[e.g.,][]{MR2005}. Thus, we conclude
456: that CXOU J132518.2$-$430304 has likely undergone a state transition from the
457: steep-power law state to the thermal dominant state.
458: The accretion column density also changes during the transition; however, to our
459: knowledge no such change has been observed in Galactic BH-LMXBs. It is unclear
460: if the local accretion column density is actually changing or if the power-law
461: model does not sufficiently capture the still unknown physics of the steep
462: power-law state at the lower energies observed by {\it Chandra}.
463:
464: Many Galactic BH-LMXBs are relatively short-term transients with infrequent
465: outbursts with durations of about a month \citep{MR2005}. One notable exception
466: is the recurring transient GX~339$-$4, which outbursts every few years. In
467: \citet{KKF+2001}, a luminous transient source, CXOU J132519.9$-$430317
468: ($L_X \sim 2 \times 10^{39} {\rm \, erg \, s} ^{-1}$),
469: was detected 21\arcsec\ ESE of CXOU J132518.2$-$430304
470: in Observation
471: 316, but was undetected five months later. The transient source 1RXH
472: J132519.8$-$430312, which has been associated with CXOU J132519.9$-$430317, was
473: luminous over 10 days in 1995, but undetected in other ROSAT observations
474: spanning an 8 year timescale from 1990--1998 \citep{SDE2000}. Our deep ACS
475: image reveals that the counterpart found by \citet{GFS+2006} is an
476: aggregate of several stars consistent with being red giants in Cen A.
477: CXOU
478: J132519.9$-$430317, which appears to be a recurring transient with one outburst
479: of at least 10 days and a second outburst with an upper limit of 835 days
480: \citep{SDE2000,KKF+2001}, appears similar to GX~339$-4$.
481: Some BH-XRB transients, e.g., XTE~J1550$-$564 and GRO~J1655$-$40, have
482: intermediate-duration outbursts lasting months, while GRS~1915$+$105 has
483: remained bright since its outburst approximately 15 years ago \citep{MR2005}.
484: In Cen A, CXOU J132518.2$-$430304 is in outburst for at least 70 days
485: with roughly constant luminosity,
486: which could be consistent with outburst timescales of months or years.
487:
488: Luminous ($L_X > 8\times 10^{38} {\rm \, erg \, s} ^{-1}$) LMXBs in other
489: early-type Galaxies tend to be persistent sources. In \citet{Ir2006}, none of
490: the 18 (15) luminous LMXBs attributed to NGC 1399 (M87) were transient sources.
491: This study used two (six) epochs spanning 3.3 (5.3 yr). This places a 95\% lower
492: limit of 50 yr for the outburst duration of such sources. On the other hand, the
493: two luminous non-nuclear sources in Cen A have both been identified as
494: transients. Since the current data do not strongly constrain the duration of
495: either transient, we first assumed that the previously known transient in Cen A
496: undergoes $100 {\rm \, d}$ outbursts and that CXOU J132518.2$-$430304 is
497: undergoing a long-duration (${\rm yrs}$) outburst akin to GRS~1915$+$105. In
498: that case, the rate of $100 {\rm \, d}$ outburst transients in Cen A is
499: $0.6^{+1.5}_{-0.5} {\rm \, yr}^{-1}$.
500: Since both NGC 1399 and M87 have absolute $K_s$-band magnitude $\approx$1.5 mags
501: brighter than Cen A \citep{ TDB+2001,SCS+2006}, they each have approximately 4
502: times the stellar mass of Cen A. By scaling the number of $100 {\rm \, d}$
503: outburst transients with stellar mass and considering when the galaxies were
504: observed, we find the rates are consistent with no such transients being found
505: by the
506: \citet{Ir2006} study of NGC 1399 and M87.
507:
508: On the other hand, CXOU J132518.2$-$430304 may have an outburst duration much
509: shorter than decades. If we assume both Cen A sources are transients with $100
510: {\rm \, d}$ outbursts, the rate of such transients is $1.3^{+1.7}_{-0.8} {\rm \,
511: yr}^{-1}$, and we would predict the \citet{Ir2006} study should have found
512: $8.4^{+11.1}_{-\phn5.4}$ such transients in NGC~1399 and M87. The lower limit is
513: only consistent with no detected transients at the 5\% level.
514:
515: As a more model independent comparison of the rates of transients, we calculate
516: the fraction of sources with outburst durations shorter than years. If the
517: outburst of CXOU J132518.2$-$430304 lasts for decades, the difference in the
518: fractions between Cen A ($0.5\pm0.4$) and NGC 1399 and M87 ($<0.04$) are
519: different only at the $80.9\%$ level. However, if the outburst is much shorter,
520: the fraction in Cen A is $>0.40$, while the fraction in NGC 1399 and M87 is
521: $<0.04$; these fractions are different at the $98.5\%$ level of confidence.
522:
523: Extensive X-ray observations of the bulge of M31 have revealed $\sim$60 transient
524: sources \citep{TPC2006,WNG+2006,VG2007} over $\sim 4.6 {\rm \, yr}$, all
525: with $L_X < 4\times 10^{38} {\rm \, erg \, s} ^{-1}$. Since the
526: $M_{K_s}$ of Cen A is about 0.3 brighter than M31, one
527: expects to find $4.5^{+5.9}_{-2.8}$ luminous transients in M31, if both Cen
528: A transients undergo $100 {\rm \, d}$ outbursts. Thus, it is not
529: surprising that no such luminous transients in M31 have been observed.
530:
531: New observations of Cen A
532: are needed to determine if CXOU J132518.2$-$430304 is undergoing
533: an outburst of months or years and if the rate of luminous
534: transients in Cen A is anomalously high compared with other early-type
535: galaxies.
536:
537:
538: \acknowledgments
539: This work was supported by NASA through {\it Chandra} award
540: GO7-8105X from the CXO, which is operated by the SAO
541: under NASA contract NAS8-03060, and {\it HST} award GO-10597 from
542: STScI, which is operated by AURA under NASA contract NAS5-26555.
543:
544: {\it Facilities:} \facility{CXO (ACIS)}, \facility{HST (ACS/WFC,WFPC2)}, \facility{VLA}
545:
546: %%% BIBLIOGRAPHY
547:
548: \bibliography{ms}
549:
550: \end{document}
551:
552: