1: \documentclass[apj]{emulateapj}
2: %\documentclass[12pt,preprint]{aastex}
3: %\usepackage{psfig}
4: \begin{document}
5: \title{The Compact X-ray Source 1E 1547.0-5408 and the Radio Shell
6: G327.24-0.13: A New Proposed Association between a Candidate Magnetar
7: and a Candidate Supernova Remnant}
8: \author{Joseph D. Gelfand}
9: \affil{Harvard-Smithsonian Center for Astrophysics}
10: \affil{60 Garden Street Cambridge, MA 02138}
11: \email{jgelfand@cfa.harvard.edu}
12: \and
13: \author{B. M. Gaensler\altaffilmark{1}}
14: \affil{Harvard-Smithsonian Center for Astrophysics, Cambridge, MA
15: 02138}
16: \and
17: \affil{School of Physics A29, The University of Sydney, NSW 2006, Australia}
18: \email{bgaensler@usyd.edu.au}
19:
20: \altaffiltext{1}{Alfred P. Sloan Research Fellow,
21: Australian Research Council Federation Fellow}
22:
23: \begin{abstract}
24: We present X-ray, infrared and radio observations of the field
25: centered on X-ray source 1E~1547.0--5408 in the Galactic Plane.
26: Analysis of a new {\it Chandra} observation of this source shows it is
27: unresolved at arc-second resolution, and analysis of a new {\it XMM}
28: observation shows that its X-ray spectrum is best described by an
29: absorbed power-law and blackbody model. A comparison of the X-ray
30: flux observed from 1E~1547.0--5408 between 1980 and 2006 by {\it
31: Einstein}, {\it ASCA}, {\it XMM}, and {\it Chandra} reveals that its
32: absorbed 0.5--10~keV X-ray flux decreased significantly during this
33: period, from $\sim2\times10^{-12}$~ergs~cm$^{-2}$~s$^{-1}$ to
34: $\sim3\times10^{-13}$~ergs~cm$^{-2}$. No pulsations in the X-ray
35: emission from 1E~1547.0--5408 were observed during the most recent
36: {\it XMM} observation, which allows us to put a 5$\sigma$ confidence
37: upper limit of 14\% for the 0.5--10~keV peak-to-peak pulsed fraction
38: (for sinusoidal pulses with periods slower than 1.8s). A
39: near-infrared observation of this field shows a source with magnitude
40: $K_s = 15.9\pm0.2$ near the position of 1E~1547.0--5408, but the
41: implied X-ray to infrared flux ratio indicates the infrared emission
42: is most likely from an unrelated field source, allowing us to limit
43: the IR magnitude of any counterpart to 1E~1547.0--5408 to $\ga
44: 17.5$. Archival radio observations reveal that 1E~1547.0--5408 sits at
45: the center of a faint, small ($4\arcmin$ diameter) radio shell,
46: G327.24--0.13, which is possibly a previously unidentified supernova
47: remnant. The X-ray properties of 1E~1547.0--5408 suggest that this
48: source is a magnetar --- a young neutron star whose X-ray emission is
49: powered by the decay of its extremely strong magnetic field, $B \sim
50: 10^{14-15}$~G. The spatial coincidence between this source and
51: G327.24--0.13 suggests that 1E~1547.0--5408 is associated with a young
52: supernova remnant, supporting a neutron star interpretation.
53: Additional observations are needed to confirm the nature of both
54: 1E~1547.0--5408 and G327.24--0.13, and to determine whether these
55: sources are physically associated. If confirmed, this pair will be an
56: important addition to the small number of known associations between
57: magnetars and supernova remnants.
58: \end{abstract}
59: \keywords{stars: neutron, ISM: supernova remnants, ISM: individual
60: (G327.24-0.13), X-rays: stars, X-rays: individual (1E 1547.0-5408)}
61:
62: \section{Introduction}
63: \label{intro}
64: Massive stars ($M \ga 8 M_{\odot}$) end their lives in supernovae,
65: often forming neutron stars. For many decades, it was thought that
66: these neutron stars had short initial spin periods ($P_0 \ll 1~{\rm
67: s}$), dipole surface magnetic fields with strengths $B \sim
68: 10^{12}~{\rm G}$, and were most often observed as radio pulsars.
69: While this describes the majority of known neutron stars, several new
70: classes of neutron stars have since been discovered. Most notable of
71: these are Anomalous X-ray Pulsars (AXPs) and Soft Gamma Repeaters
72: (SGRs), which have spin periods longer than most normal radio pulsars
73: ($P\sim5-12$~s) and high period-derivatives ($\dot{P}\sim10^{-11}$
74: s/s) \citep{woods04}. Due to the strong surface dipole magnetic
75: fields inferred from these timing properties ($B \sim 10^{14-15}~{\rm
76: G}$), these sources are believed to be ``magnetars'', neutron stars
77: whose X-ray emission is powered by the decay of these extremely strong
78: field \citep{duncan92,thompson95,thompson96}.
79:
80: Currently there are only 13 confirmed magnetars. As a result, it is
81: difficult to determine the spatial distribution, birth-rate, and
82: active lifetime of these sources, information vital to understanding
83: the relationship between magnetars and other products of core collapse
84: supernovae. This uncertainty is further exacerbated by the presence
85: of at least one transient magnetar, XTE~J1810-197 \citep{ibrahim04}.
86: A powerful way of determining the relationship between magnetars and
87: other classes of neutron stars is to use environmental information to
88: constrain their ages and progenitors of these populations. This is
89: easiest done for those neutron stars associated with SNRs, since
90: observations of SNRs allow an independent estimates of distances and
91: ages, the densities of the surrounding medium, and the explosion
92: energies of the progenitor supernovae. To date, there are only two
93: secure magnetar/SNR associations, 1E~2259+586 in SNR CTB 109
94: \citep{fahlman81} and 1E~1841-045 in SNR Kes~73 \citep{vasisht97},
95: %and AX~J1845--0258 in SNR~G29.6+0.1 \citep{gaensler99},
96: though SGR~0526-66 is possibly associated with Large Magellanic Cloud
97: SNR N49 \citep{marsden96,gaensler01}. Based on the ages of these SNRs
98: (not including SNR~N49), as well as the offset between these magnetars
99: and the centers of their SNRs, \citet{gaensler01} concluded that
100: magnetars are young ($<10000$~yr) neutron stars with low projected
101: space velocities ($<$500~km~s$^{-1}$). This assertion that magnetars
102: have a low spatial velocity is also supported by recent proper motion
103: measurements of XTE~J1810-197 \citep{helfand07}. Additionally, an
104: analysis of the X-ray emission from these SNRs implies that the level
105: of energy injection at early times, from both supernova ejecta and
106: from the magnetar spin down, was $\sim10^{51}$~ergs, constraining the
107: initial spin period of these magnetars to $P_0 \ga 5$~ms
108: \citep{vink06}. While much has been learned from detailed studies of
109: the few associations between magnetars and SNRs, much more can be
110: gained by identifying additional examples of such associations.
111:
112: In this paper, we propose a new potential magnetar/SNR association,
113: between the X-ray source 1E~1547.0-5408 and the Galactic radio shell
114: G327.24-0.13. In \S\ref{1e1547}, we present both X-ray (\S\ref{xray})
115: and near-infrared (near-IR) (\S\ref{nir}) observations of the field
116: around 1E~1547.0-5408. In \S\ref{radio}, we present archival radio
117: observations which show that 1E~1547.0-5408 is located at the center
118: of G327.24-0.13. In \S\ref{conclusions}, we argue the compact nature
119: of 1E~1547.0-5408, its lack of a bright near-IR counterpart, and its
120: location of 1E~1547.0-5408 source at the center of G327.24-0.13
121: implies this source is a neutron star, and that the X-ray spectrum and
122: variability of 1E~1547.0-5408 strongly suggest that it is a magnetar.
123:
124: \section{Observations of 1E~1547.0-5408}
125: \label{1e1547}
126:
127: \subsection{X-ray Data}
128: \label{xray}
129:
130: X-ray source 1E~1547.0-5408 was discovered by {\it Einstein} during a
131: search for X-ray counterparts of unidentified $\gamma$-ray sources
132: \citep{lamb81}, in this case the $\gamma$-ray source 1CG327--0
133: detected by {\it Cos-B} \citep{hermsen77}. This detection was
134: confirmed by the {\it ASCA} Galactic plane survey, in which the source
135: AX~J155052-5418 was detected at the same position \citep{sugizaki01}.
136: Recently, this field has been observed twice with {\it XMM--Newton},
137: (an archival observation in 2004 and a new observation in 2006), and
138: again with the {\it Chandra X-ray Observatory} in 2006. A summary of
139: these observations is given in Table \ref{obssum}. Both {\it XMM}
140: datasets were processed using the standard tasks given in {\sc
141: xmm-sas~v.6.5.0}, while the {\it Chandra} data were analyzed using
142: {\sc ciao~v.3.4}. All datasets were then filtered using the standard
143: energy and quality criteria.
144:
145: As shown in Fig.~\ref{xrayimg}, due to the improved spatial resolution
146: of {\it XMM} and {\it Chandra}, we detect two point sources in the
147: overlapping position error circles of 1E~1547.0-5408 and
148: AX~J155052-5418 -- a bright source that we designate
149: CXOU~J155054.1-541824 (=XMMU~J155054.3--541825) and a much fainter
150: source we designate XMMU~J155053.7-541925. This second source is
151: detected in both {\it XMM} observations but not the {\it Chandra}
152: observation because it fell in a chip gap. For both sources, their
153: extents are consistent with the point spread function of the telescope
154: at that position. The most accurate position for
155: CXOU~J155054.1-541824 comes from the {\it Chandra} observation, which
156: gives $\alpha_{\rm J2000}=$15:50:54.11 and $\delta_{\rm
157: J2000}=$--54:18:23.8. This position has not been registered to an
158: external reference frame, due to the lack of field X-ray sources with
159: counterparts at other wavelengths. Therefore, the error in this
160: position is dominated by the pointing accuracy of {\it Chandra} -- a
161: typical 99\% confidence radius of $\sim0\farcs8$. As shown in Table
162: \ref{src2cnt}, the observed count rates of both CXOU~J155054.1-541824
163: and XMMU~J155053.7-541925 declined between the 2004 and 2006 {\it XMM}
164: observations -- in the case of CXOU~J155054.1-541824, the count rate
165: declined by $\sim30\%$, while for XMMU~J155053.7-541925 the count rate
166: declined by $\sim85\%$.
167:
168: We have quantified the apparent variation of the X-ray flux from
169: CXOU~J155054.1-541824 via a more detailed spectral analysis of the
170: observations listed in Table \ref{obssum}. Here, and in the
171: subsequent discussion, we assume that 1E~1547.0-5408 and
172: AX~J155052-5418 correspond to the same X-ray source, and that this
173: source is a blend of CXOU~J155054.1-541824, XMMU~J155053.7-541925, and
174: other adjacent field sources that can be seen at the high angular
175: resolution of {\it XMM} and {\it Chandra}, but not with {\it ASCA} and
176: {\it Einstein}. Since CXOU~J155054.1-541824 substantially dominates
177: the emission, we here on use 1E~1547.0-5408 to indicate this main
178: source, except with a distinction with fainter, adjacent sources needs
179: to be made. For the {\it Einstein} observation, we used the procedure
180: defined by \citet{mcgarry05} to determine the flux of 1E~1547.0-5408.
181: To do this, we assume the X-ray spectrum of 1E~1547.0-5408 is well
182: modeled by an absorbed power law assuming a hydrogen column density
183: $N_H=4.2\times10^{22}$~cm$^{-2}$ and a photon index $\Gamma=4.7$, as
184: derived below from spectral fits to the {\it ASCA}, {\it XMM}, and
185: {\it Chandra} observations. To determine the X-ray flux of
186: 1E~1547.0-5408 as measured by {\it ASCA}, {\it XMM}, and {\it
187: Chandra}, we simultaneously fit the observed spectrum of this source
188: to an absorbed power law model using {\sc Xspec} v.11.3.1. We chose
189: an absorbed power law model since this model provides a good fit to
190: each of these spectra individually. For the {\it XMM} and {\it
191: Chandra} observations, spectral regions were chosen to minimize
192: contamination from XMMU~J155053.7-541925 and other field sources. For
193: the {\it ASCA} observation, this was not possible due to the poor
194: point spread function of this instrument. Holding the $N_H$,
195: $\Gamma$, and normalization constant between these different
196: observations, which implies a constant flux, resulted in a poor fit
197: (reduced $\chi^2=2.19$). Allowing the normalizations to vary
198: independently, but holding $N_H$ and $\Gamma$ fixed, lowered the
199: reduced $\chi^2$ to 1.26, a substantial improvement. As shown in
200: Table \ref{xrayfluxes}, the results from this analysis imply that
201: 1E~1547.0-5408 is a variable X-ray source.
202:
203: To determine what physical model best describes the X-ray spectrum of
204: 1E~1547.0-5408, we used {\sc Xspec} v.11.3.1 to jointly fit the
205: spectrum of this source measured during the 2006 August {\it XMM}
206: observation by the {\sc Mos1}, {\sc Mos2}, and {\sc pn} instruments to
207: a number of different models, including a power law, bremsstrahlung,
208: power-law plus blackbody, and two blackbodies, all attenuated by
209: interstellar absorption. During this observation, between 0.5 and
210: 10~keV a total of 1005$\pm$33 counts above the background were
211: collected from this source by the {\sc Mos1} detector, 1062$\pm$34 by
212: the {\sc Mos2} detector, and 2351$\pm$51 by the {\sc pn} detector.
213: The resultant X-ray spectrum is shown in Fig. \ref{src1spec}. As
214: shown in Table \ref{xrayspec}, an absorbed power-law plus blackbody
215: model produces the best fit. According to the f-test, the decrease in
216: $\chi^2$ for the best two-component model (absorbed power-law plus
217: blackbody) over the best one-component model (absorbed bremsstrahlung)
218: is statistically significant at 99.9\% confidence, so the use of
219: second component is justified. Therefore, we conclude that either an
220: absorbed power-law plus blackbody or an absorbed two blackbodies
221: (these two models produced statistically indistinguishable fits)
222: provides the best description of the X-ray spectrum of 1E~1547.0-5408.
223:
224: We also searched for pulsed X-ray emission from 1E~1547.0-5408 using
225: the data collected during the 2006 {\it XMM} observation using the
226: $Z_n^2$ test \citep{buccheri83}. To do so, we first extracted events
227: from two circular regions centered around the position
228: CXOU~J155054.1-541824, one $9\farcs5$ in radius and the other
229: $26\farcs5$ in radius, from the filtered {\sc Mos1}, {\sc Mos2}, and
230: {\sc pn} datasets. We then barycentered the arrival times of these
231: photons to the solar system reference frame. This search was
232: conducted on the {\sc Mos1} (0.9s time resolution), {\sc Mos2} (0.9s
233: time resolution), and {\sc pn} (73.4~ms time resolution) datasets
234: individually, as well as on them jointly. To allow for the
235: possibility that the pulsed fraction might have a strong energy
236: dependence, we applied various energy cuts to the event lists derived
237: from both spatial regions. For all searches the minimum frequency was
238: $2\times10^{-5}$~Hz, the maximum frequency was 0.6~Hz for the {\sc
239: Mos} and joint datasets and 6.8~Hz for the {\sc pn} data, and the
240: frequency step was $2\times10^{-6}$~Hz, oversampling the Nyquist
241: frequency by a factor of five. For each combination of spatial
242: region, detectors, and energy range, we searched for periods summing
243: up to a maximum harmonic $n=1,2,...,10$. In none of these different
244: combinations did we detect a statistically significant signal.
245: Between 0.5 and 10~keV, the most sensitive dataset for a sinusoidal
246: ($n=1$) pulse profile comes from combining the three detectors and
247: using the large spatial region. In this dataset, we have 5305 photons
248: in 52201 independent trials. Using the equations in \citet{leahy83},
249: we are able to place a 5$\sigma$ upper limit on the peak-to-peak pulse
250: fraction $f_{\rm pulse}$, defined as \citep{patel03}:
251: \begin{eqnarray}
252: \label{pulsefrac}
253: f_{\rm pulse} & = & \frac{N_{\rm max}-N_{\rm min}}{N_{\rm max}+N_{\rm min}}
254: \end{eqnarray}
255: where $N_{\rm max}$ and $N_{\rm min}$ are the maximum and minimum
256: number of counts in the pulse profile, of $f_{\rm pulse}<14\%$.
257:
258: \subsection{Near-IR Observation of 1E~1547.0-5408}
259: \label{nir}
260:
261: Using the position of CXOU~J155054.1-541824 as measured by {\it
262: Chandra}, we searched for a near-IR counterpart in a three minute
263: $K_s$ ($\lambda=2.15\mu$m) observation of this field taken on 13 June
264: 2006 with the Persson's Auxiliary Nasmyth Infrared Camera ({\sc
265: panic}) instrument on the 6.5m Baade Magellan telescope at the Las
266: Campanas Observatory in Chile. This image was sky-subtracted using
267: the standard procedures in the {\sc iraf} software package, and
268: registered using the position of 2MASS sources in the field. We used
269: the {\sc iraf} task {\tt daofind} to identify objects in this field
270: and measure their instrumental magnitudes, and used the 144 2MASS
271: stars in this field with one and only counterpart within 1\arcsec~of
272: their position to determine the conversion between instrumental and
273: astronomical magnitudes. To search for a counterpart, we used the
274: 99\% error ($\sim0\farcs8$) of the {\it Chandra} position quoted in
275: \S\ref{xray}. As shown in Fig.~\ref{nirimg}, only one source was
276: detected inside this region, and this source has an observed magnitude
277: of $K_s=15.9\pm0.2$. Using the source magnitude distribution of
278: objects detected in our $K_s$ observation, we determine that our image
279: is complete to sources with a $K_s$ magnitude $\la 17.5$.
280:
281: \section{The Galactic radio shell G327.24-0.13}
282: \label{radio}
283:
284: 1E~1547.0-5408 falls within the field of view of multiple recent
285: southern hemisphere radio surveys: the Molonglo Galactic Plane survey
286: (MGPS, 843~MHz; \citealt{green99}), the Sydney University Molonglo Sky
287: Survey (SUMSS, 843~MHz; \citealt{bock99,green02}) and in both the test
288: region \citep{gaensler01b} and survey region \citep{sgps} of the
289: Southern Galactic Plane Survey (SGPS, 1.4~GHz). All four surveys
290: detected a faint $\sim4^{\prime}$ diameter shell centered on
291: CXOU~J155054.1-541824, which we designate G327.24-0.13. In both the
292: MGPS and SUMSS (Figure~\ref{radimg}) images, there is some evidence
293: for enhanced emission from the center of G327.24-0.13. At both
294: frequencies, the flux was determined by subtracting the observed flux
295: from G327.24-0.13 by estimates for the diffuse Galactic background at
296: this position obtained using nearby regions, and the error is
297: dominated by the uncertainty in the background. At 843~MHz, the flux
298: density of G327.24-0.13 -- excluding any interior emission -- is
299: $0.5\pm0.1$~Jy while at 1.4~GHz, the flux density is $0.3\pm0.1$~Jy.
300: These fluxes correspond to a radio spectral index between 843~MHz and
301: 1.4~GHz of $\alpha=-0.9\pm0.6$ ($S_{\nu}\propto\nu^\alpha$).
302:
303: \section{Discussion}
304: \label{conclusions}
305:
306: In this section, we use the observational results presented in
307: \S\ref{1e1547} and \S\ref{radio} to determine the nature of
308: 1E~1547.0-5408 and G327.24-0.13. Based on the results of the {\it
309: XMM} and {\it Chandra} observations, we believe 1E~1547.0-5408 is a
310: blend of compact X-ray sources dominated by the X-ray emission of
311: CXOU~J155054.1-541824. Therefore, we conclude that
312: XMMU~J155053.7-541925 and the other sources detected by {\it XMM}
313: and/or {\it Chandra} are negligible contributors to the X-ray
314: properties of 1E~1547.0-5408 as measured by {\it Einstein} and {\it
315: ASCA}. We also believe that these sources are unrelated to
316: CXOU~J155054.1-541824.
317:
318: Since CXOU~J155054.1-541824 is unresolved by {\it Chandra},
319: 1E~1547.0-5408 is most likely either an active galactic nucleus (AGN),
320: a neutron star (either a rotation-powered pulsar, a magnetar, or a
321: compact central object (CCO)), a non-degenerate star, a X-ray binary,
322: or a cataclysmic variable (CV). A major clue into the nature of
323: 1E~1547.0-5408 is its possible association with a near-IR source, as
324: discussed in \S\ref{nir}.
325:
326: If 1E~1547.0-5408 is associated with this near-IR source, it is most
327: likely either a non-degenerate star in the Milky Way or a AGN. If a
328: non-degenerate star, then the absorbed bremsstrahlung is the most
329: realistic description of the X-ray emission. If an AGN, only the
330: absorbed bremsstrahlung fit to the observed spectrum gives parameters
331: similar to that observed from other AGN. To determine if either
332: identification is reasonable, we compared the X-ray and IR fluxes of
333: this source to those of known stars and AGN, similar to the approach
334: used by \citet{kaplan04}. For stars, we used data obtained by the
335: {\it Chandra} Orion Ultradeep Project \citep{getman05}, and for AGN,
336: we used the X-ray \citep{kenter05} and near-IR \citep{jannuzi04} data
337: from the XBo\"{o}tes survey. For both surveys, we only used the
338: observed 2--7~keV flux since this quantity is less sensitive to
339: interstellar absorption and choice of X-ray spectral model than the
340: 0.5--2~keV flux. To correct for the difference in $N_H$ observed
341: towards 1E~1547.0-5408 from the Galactic value of $N_H$ towards
342: sources in the Bo\"otes field ($N_H=10^{20}$~cm$^{-2}$;
343: \citeauthor{kenter05} \citeyear{kenter05}) and the value observed
344: towards sources in the Orion field ($N_H = 10^{21}-10^{23}$~cm$^{-2}$;
345: \citeauthor{getman05} \citeyear{getman05}), we determined the 2--7~keV
346: X-ray and near-IR fluxes of 1E~1547.0-5408 over this range of $N_H$.
347: As shown in Figure \ref{fluxratimg}, the IR and X-ray properties of
348: 1E~1547.0-5408 are inconsistent with both populations. Therefore, we
349: conclude that 1E~1547.0-5408 is not associated with this near-IR
350: source. To compute the probability of their association being a
351: coincidence, we shifted the RA and DEC of 1E~1547.0-5408 by a random
352: amount between $\pm0\farcs8-10\arcsec$, and determined if there is a
353: near-IR source within $0\farcs8$ of its adjusted position. This
354: analysis implies a false coincidence rate of $\sim$14\%, implying that
355: there is a reasonable possibility that the near-IR object is an
356: unrelated field source. As a result, we adopt a $K_s$ magnitude of
357: 17.5 as an upper limit on $K_s$ magnitude 1E~1547.0--5408.
358:
359: As a result, we are left with the possibility that this X-ray source
360: is a neutron star, X-ray binary, or a CV. The location of
361: 1E~1547.0-5408 in the center of SNR candidate G327.24-0.13 strongly
362: implies that this source is a neutron star. This identification is
363: supported by the fact that the bremsstrahlung and power-law fits to
364: the X-ray spectrum of 1E~1547.0-5408 shown in Table \ref{xrayspec} are
365: inconsistent with the spectra expected from X-ray binaries (a modified
366: blackbody with kT$\sim$1-2~keV; \citeauthor{white88}
367: \citeyear{white88}) and CVs (bremsstrahlung emission with kT$>1$~keV;
368: \citeauthor{eracleous91} \citeyear{eracleous91}). The X-ray spectrum
369: of 1E~1547.0-5408 is also inconsistent with that observed from
370: rotation-powered pulsars (a power law with $\Gamma=1-2$;
371: \citeauthor{cheng04} \citeyear{cheng04}). However, the parameters for
372: the absorbed power law, the absorbed power law and blackbody, and the
373: absorbed two blackbody models are similar to that observed from both
374: magnetars \citep{mereghetti01} and CCOs \citep{pavlov04}. As shown in
375: Table \ref{xrayirdata}, the observed X-ray variability of
376: 1E~1547.0-5408 described in \S\ref{xray} is similar to that observed
377: from several magnetars. CCOs, on the other hand, appear to be steady
378: X-ray sources, and therefore we conclude the 1E~1547.0-5408 is most
379: likely a magnetar. One crucial difficulty with this interpretation is
380: our failure to detect pulsed X-ray emission from this source.
381: Magnetars typically have high pulsed fractions in the X-ray regime,
382: for example before its recent outburst CXOU~J164710.2-455216 had a
383: pulsed fraction of $\sim50\%$ \citep{muno06}, much higher than the
384: 5$\sigma$ upper limit derived in \S\ref{xray} of 14\% on the 0.5--10~keV
385: peak-to-peak pulsed fraction from 1E~1547.0-5408 for a sinusoidal
386: pulse profile, the characteristic pulse profile for magnetars.
387: However, this upper limit is higher than the pulsed fraction of at
388: least one magnetar, 4U~0142+61, which has a 0.5--7~keV peak-to-peak
389: pulsed fraction of $\approx7\%$ \citep{patel03,gohler05}.
390:
391: If this identification is correct, we can compare the X-ray properties
392: of 1E~1547.0-5408 to those of other magnetars. For the power-law plus
393: blackbody model of the spectrum, the 2--10 keV unabsorbed flux of
394: 1E~1547.0-5408 ranges from
395: $\sim(0.5-3)\times10^{-12}$~ergs~cm$^{-2}$~s$^{-1}$. For a distance
396: $d=4d_4$~kpc (it will be argued below that $d_4 \approx 1$ is a
397: reasonable distance estimate), this translates to a luminosity range
398: of $\sim(0.9-7)d_4\times10^{33}$~ergs~s$^{-1}$. Comparing this
399: value to those given in the SGR/AXP Online Catalog\footnote{Available
400: online at {\tt
401: http://www.physics.mcgill.ca/$\sim$pulsar/magnetar/main.html}.}, this
402: falls within the X-ray luminosity range spanned by confirmed
403: magnetars. With this information, we can also compare the near-IR and
404: X-ray properties of 1E~1547.0-5408 to that of other magnetars, whose
405: properties are listed in Table \ref{xrayirdata}. As shown in Figure
406: \ref{fluxratimg}, the range of X-ray fluxes observed from
407: 1E~1547.0--5408, as well as the upper limit on the near-IR flux is
408: consistent with that observed from confirmed magnetars.
409:
410: The radio spectral index of G327.24-0.13 derived in \S\ref{radio} is
411: consistent with a non-thermal origin, implying that this source is
412: either a stellar wind bubble or a SNR. The lack of any mid-IR
413: counterpart to G327.24-0.13 in the {\it Spitzer} {\sc glimpse} survey
414: \citep{glimpse} gives support to the latter interpretation. If
415: 1E~1547.0-5408 is a magnetar and G327.24-0.13 is the SNR created by
416: its progenitor, then this system becomes a new addition to the handful
417: of associations between magnetars and SNRs. Assuming that the two
418: sources are indeed associated, we can obtain initial constraints on
419: the possible distance to 1E~1547.0--5408/G327.24-0.13 as follows.
420: Recent evidence has suggested that magnetar progenitors are especially
421: massive stars \citep{gaensler05, muno06}, and therefore expect
422: significant star--formation activity in their vicinity. Indeed, in
423: this case there are two nearby ($\la0\fdg5$ away) thermal radio
424: sources, G326.96+0.03 and G327.99-0.09, both associated with a large
425: star-forming complex in the Scutum-Crux spiral arm. H\,{\sc i}\
426: absorption and hydrogen recombination line measurements indicate that
427: the distance to G326.96+0.03 and G327.99-0.09 both fall within the
428: range $d=3.7-4.3$~kpc \citep{caswell87,sgpstest}. By associating
429: 1E~1547.0-5408/G327.24-0.13 with this region, we argue that $d_4
430: \approx 1$. If so, the observed size of G327.24-0.13 ($\sim4\arcmin$)
431: implies a diameter of $\sim5d_4$~pc, making G327.24-0.13 one of the
432: smallest, and therefore probably youngest, known SNRs.
433:
434: As mentioned in \S\ref{xray}, 1E~1547.0-5408 was originally discovered
435: in a search for the X-ray counterpart of unidentified $\gamma$-ray
436: source 1CG327--0. Quiescent emission from magnetars has been detected
437: at energies as high as $\sim$100~keV by the {\it Integral} satellite
438: \citep{gotz06}, but not in the $>100$~MeV range detected by {\it
439: Cos-B} from 1CG327--0 \citep{hermsen77}. Though magnetars have been
440: known to be the source of intense $\gamma$-ray flares
441: \citep{hurley05}, these events are rare (only three have been detected
442: in the past 30 years), and the maximum photon energy detected from
443: these flares ($\sim1$~MeV; \citeauthor{hurley05} \citeyear{hurley05})
444: is significantly less than the photon energies detected from
445: 1CG327--0. As a result, we do not believe that 1E~1547.0-5408 is the
446: X-ray counterpart of 1CG327--0. Another possibility is that 1CG327--0
447: is associated with G327.24-0.13, since high-energy $\gamma$-ray
448: emission has been detected from some SNRs. However, the non-detection
449: of 1CG327--0 by {\it EGRET} implies this source is variable, making
450: any association of 1CG327--0 with the SNR candidate G327.24--0.13
451: unlikely. Future {\it GLAST} observations of this region should
452: determine if 1CG327--0 is a real $\gamma$-ray source and, if so, if it
453: is affiliated with 1E~1547.0-5408 or G327.24-0.13.
454:
455: To summarize, in this paper we presented X-ray, near-IR, and radio
456: observations of 1E~1547.0-5408 and of the field around it. A
457: consistent explanation of these observation is that 1E~1547.0-5408 is
458: a magnetar possibly associated with SNR candidate
459: G327.24--0.13. Deeper X-ray observations of 1E~1547.0-5408 are needed
460: to confirm its identification as a magnetar by detecting X-ray
461: pulsations from this source, deep near-IR observations of this source
462: are needed to discover its near-IR counterpart, and additional radio
463: observations of G327.24--0.13 are required to determine if it is a
464: SNR. The identification of 1E~1547.0-5408 as a magnetar candidate
465: illustrates the importance of follow-up observations of other bright,
466: unidentified X-ray sources in the Galactic Plane for understanding the
467: X-ray population of the Milky Way and for discovering new members of
468: exotic classes of neutron stars.
469:
470: \acknowledgements J.D.G and B.M.G. are supported by NASA through LTSA
471: grant NAG5-13032. B.M.G. is also supported by an Alfred P. Sloan
472: Research Fellowship. J.D.G. would like to thank Julia Bryant for
473: generously carrying out the Magellan observation as well as the
474: initial IR data reduction, Ingyin Zaw, Jenny Greene, Craig Heinke,
475: Wynn Ho, Jae Sub Hong, Xavier Koenig, Dave Monet, Dan Padnaude, and
476: Pat Slane for useful discussions, Ryan Hickox for providing the X-ray
477: and IR emission of sources in the Bo\"otes survey, and David Kaplan,
478: Manual Torres, and Maryam Modjaz for help analyzing the Magellan data.
479: The Australia Telescope is funded by the Commonwealth of Australia for
480: operation as a National Facility managed by CSIRO. The MOST is
481: operated by the University of Sydney with support from the Australian
482: Research Council and the Science Foundation for Physics within the
483: University of Sydney. This research has used the resources of the
484: High Energy Astrophysics Science Archive Research Center (HEASARC).
485: The Two Micron All Sky Survey is a joint project of the University of
486: Massachusetts and the Infrared Processing and Analysis
487: Center/California Institute of Technology.
488:
489: \bibliography{ms}
490: \bibliographystyle{apj}
491:
492: \clearpage
493: %% Tables
494:
495: \begin{table}
496: \begin{center}
497: \scriptsize
498: \caption{Properties of the new and archival X--ray observations of
499: 1E~1547.0-5408 presented in this paper \label{obssum}}
500: \begin{tabular}{cccc}
501: \hline
502: \hline
503: {\sc Date} & {\sc Telescope} & {\sc Instruments} & {\sc Exposure Time} [ks] \\
504: \hline
505: 1980 March 12 & {\it Einstein} & HRI & 3.6 \\
506: 1998 February 24 & {\it ASCA} & GIS & 6.7 \\
507: 2004 February 8 & {\it XMM--Newton} & {\sc Mos1}, {\sc Mos2}, {\sc
508: pn} & 8.7, 8.7, 6.2 \\
509: 2006 July 1 & {\it Chandra} & {\sc Acis-I} & 9.5 \\
510: 2006 August 22 & {\it XMM--Newton} & {\sc Mos1}, {\sc Mos2}, {\sc pn}
511: & 45.1, 45.1, 38.6 \\
512: \hline
513: \hline
514: \end{tabular}
515: \tablecomments{For the 2004 February 8 {\it XMM} observations, the
516: {\sc Mos1} detector was operated in Large Window Mode using the
517: Medium filter, the {\sc Mos2} detector was operated in Full Window
518: Mode using the Thin filter, and the {\sc pn} detector was operated in
519: Full Window Mode using the Medium filter. For 2006 July 1 {\it
520: Chandra} observation, the {\sc Acis-I} instrument was operated in
521: Faint mode. For the 2006 August 22 {\it XMM} observation, the {\sc
522: Mos1} and {\sc Mos2} detectors were operated in Small Window Mode,
523: and the {\sc pn} detector in Full Window Mode, all using the Medium
524: Filter. The definitions of the Window Modes on {\it XMM} can be
525: found at {\tt
526: http://xmm.esac.esa.int/external/xmm\_user\_support/documentation/uhb/node28.html},
527: and the telemetry modes of {\it Chandra} are defined in {\tt
528: http://cxc.harvard.edu/proposer/POG/html/ACIS.html\#tth\_sEc6.13.1}.
529: For all observations, the quoted exposure time accounts for
530: dead-time and the removal of background flares.}
531: \end{center}
532: \end{table}
533:
534:
535: \begin{table}
536: \begin{center}
537: \caption{0.5--10~keV background subtracted count rate of
538: CXOU~J155054.1-541824 and XMMU~J155053.7-54192 measured during the
539: 2004 and 2006 {\it XMM} observations \label{src2cnt} }
540: \begin{tabular}{ccccc}
541: \hline \hline & \multicolumn{2}{c}{CXOU~J155054.1-541824} &
542: \multicolumn{2}{c}{XMMU~J155053.7-541925} \\
543: \hline
544: {\it Detector} & {\it 2004 February} & {\it 2006 August} &
545: {\it 2004 February} & {\it 2006 August} \\
546: \hline
547: {\sc Mos1} (counts ks$^{-1}$) & $33\pm2$ & $22.7\pm0.8$ & $4.2\pm0.8$
548: & $0.5\pm0.2$ \\
549: {\sc Mos2} (counts ks$^{-1}$) & $29\pm2$ & $23.7\pm0.8$ & $2.6\pm0.7$
550: & $0.4\pm0.2$ \\
551: {\sc pn} (counts ks$^{-1}$) & $85\pm4$ & $62\pm1$ & $\cdots$ & $2.0\pm0.3$ \\
552: \hline
553: \hline
554: \end{tabular}
555: \end{center}
556: \tablecomments{The same source and background regions were used for
557: both observations. In the 2004 {\it XMM} observation,
558: XMMU~J155053.7-541925 fell in a chip-gap in the {\sc pn} detector
559: which is why no count-rate is given for this detector. Errors
560: denote the 1$\sigma$ confidence interval.}
561: \end{table}
562:
563: \begin{table}
564: \begin{center}
565: \caption{The absorbed 0.5--10~keV flux of 1E~1547.0-5408 as
566: measured by the observations listed in Table \ref{obssum}
567: \label{xrayfluxes}}
568: \begin{tabular}{ccc}
569: \hline
570: \hline
571: {\sc Date} & {\sc Observatory} & {\sc Absorbed Flux} [$\times 10^{-12}$ ergs
572: cm$^{-2}$ s$^{-1}$] \\
573: \hline
574: 1980 March 12 & {\it Einstein} & $1.9_{-0.5}^{+1.0}$ \\
575: 1998 February 24 & {\it ASCA} & $2.1^{+0.3}_{-0.3}$ \\
576: 2004 February 8 & {\it XMM} & $0.45_{-0.03}^{+0.02}$ \\
577: 2006 July 1 & {\it Chandra} & $0.30_{-0.04}^{+0.04}$ \\
578: 2006 August 22 & {\it XMM} & $0.31_{-0.03}^{+0.01}$ \\
579: \hline
580: \hline
581: \end{tabular}
582: \end{center}
583: \tablecomments{The quoted errors denote the 90\% confidence interval.
584: The {\it ASCA}, {\it XMM}, and {\it Chandra} fluxes are from a joint
585: absorbed power law fit to the spectra, where the hydrogen column
586: density ($N_H$) and photon index ($\Gamma$) were held fixed for
587: these observations, but the normalizations were allowed to vary.
588: This fit produced $N_H=4.2\pm0.2\times10^{22}$cm$^{-2}$ and
589: $\Gamma=4.7\pm-0.2$ (errors denote 90\% confidence level of the
590: parameters), with a reduced $\chi^2=1.26$ for 263 degrees of
591: freedom. For the {\it Einstein} flux, we determined the 90\%
592: interval in the observed count rate, and then converted this count
593: rate to a flux using the same method described by
594: \citet{mcgarry05}.}
595: \end{table}
596:
597: \begin{table}
598: \begin{center}
599: \caption{Spectral Fits to 2006 August {\it XMM} observation of
600: CXOU~J155054.1-541824 \label{xrayspec}}
601: \footnotesize
602: \begin{tabular}{ccccc}
603: \hline
604: \hline
605: {\sc Parameter} & \multicolumn{4}{c}{\sc Value} \\
606: \hline
607: \hline
608: Model & {\bf phabs * pow} & {\bf phabs * bremss} & {\bf phabs *
609: (pow+bb)} & {\bf phabs*(bb+bb)} \\
610: \hline
611: N$_{H}$ [cm$^{-2}$] & $4.3^{+0.3}_{-0.2}\times10^{22}$ &
612: $3.0\pm0.2\times10^{22}$ & $3.1^{+0.7}_{-0.8}\times10^{22}$ &
613: $2.5_{-0.2}^{+0.3}\times10^{22}$ \\
614: $\Gamma_1$/$kT_1 [kT] $ & $4.8\pm0.2$ & $1.1\pm0.1$ &
615: $3.7^{+0.8}_{-2.0}$/$0.43^{+0.03}_{-0.04}$ & $0.45_{-0.06}^{+0.04}$ \\
616: $kT_2$ [kT] & $\cdots$ & $\cdots$ & $\cdots$ & $1.1^{+1.4}_{-0.3}$ \\
617: Absorbed Flux [ergs~cm$^{-2}$~s$^{-1}$] & $3.1_{-0.3}^{+0.1}\times10^{-13}$ &
618: $3.0_{-0.6}^{+0.3}\times10^{-13}$ & $3.1_{-1.2}^{+0.3}\times10^{-13}$
619: & $3.1_{-0.9}^{+0.1}\times10^{-13}$ \\
620: Unabsorbed Flux [ergs~cm$^{-2}$~s$^{-1}$] &
621: $2.8_{-1.9}^{+0.1}\times10^{-11}$ & $2.3_{-0.5}^{+0.1}\times10^{-12}$
622: & $3.6_{-2.4}^{+0.4} \times10^{-12}$ & $1.0_{-0.2}^{+0.01}\times10^{-12}$ \\
623: $\chi^2$/d.o.f (reduced $\chi^2$) & $147.9/126$ (1.17) & $144.7/126$
624: (1.14) & $134.4/124$ (1.08) & $135.3/124$ (1.09) \\
625: \hline
626: \hline
627: \end{tabular}
628: \end{center}
629: \tablecomments{The results from an absorbed power-law ({\bf phabs *
630: pow}), an absorbed bremsstrahlung ({\bf phabs * bremss}), an
631: absorbed power-law plus blackbody ({\bf phabs * (pow+bb)}), and an
632: absorbed blackbody plus blackbody ({\bf phabs * (bb+bb)}) fit to the
633: {\sc Mos1}, {\sc Mos2}, and {\sc pn} spectra of
634: CXOU~J155054.1-541824 obtained during the 2006 {\it XMM}
635: observation. The fits were done between 0.5--10~keV. The {\sc
636: Mos1} and {\sc Mos2} channels were binned such that there were a
637: minimum of 25 counts per bin, and the {\sc pn} channels were binned
638: such that there were a minimum of 50 counts per bin. Errors
639: indicate 90\% confidence intervals for each quantity, and ``d.o.f''
640: stands for degrees of freedom. Both the absorbed and unabsorbed flux
641: were calculated between 0.5 and 10~keV.}
642: \end{table}
643:
644: \begin{table}
645: \begin{center}
646: \tiny
647: \caption{The 2--7~keV Absorbed X-ray Flux and $K_s$ Flux of
648: 1E~1547.0-5408 and Some Magnetars \label{xrayirdata}}
649: \begin{tabular}{cccc}
650: \hline
651: \hline
652: Source & X-ray Flux [ergs cm$^{-2}$ s$^{-1}$] & $K_s$ Flux [ergs
653: cm$^{-2}$ s$^{-1}$] & References \\
654: \hline
655: 1E~1547.0-5408 & $(0.3-1.7)\times10^{-12}$ & $\la1\times10^{-14}$ & $\cdots$ \\
656: \hline
657: SGR 1900+14 & $(0.4-1.9)\times10^{-11}$ & $\la5\times10^{-16}$ &
658: \citet{kaplan02,esposito07}\\
659: SGR 1627-41 & $(0.1-1.9)\times10^{-12}$ & $\la1\times10^{-15}$ &
660: \citet{mereghetti06,wachter04}\\
661: SGR 1806-20 & $(0.7-1.6)\times10^{-11}$ & $(0.2-2.1)\times10^{-15}$ &
662: \citet{rea05b,woods07} \\
663: & & & \citet{kosugi05,israel05} \\
664: SGR 0526-66 & $(4.1-4.4)\times10^{-13}$ & $\la3\times10^{-16}$ &
665: \citet{klose04,kulkarni03}\\
666: XTE J1810-197 & $(0.002-2.8)\times10^{-11}$ & $(3.2-5.4)\times10^{-16}$ &
667: \citet{halpern05,gotthelf05}\\
668: & & & \citet{rea04} \\
669: 1E 1048.1-5937 & $(4.2-9.7)\times10^{-12}$ & $(0.3-2.0)\times10^{-15}$ &
670: \citet{tiengo05} \\
671: & & & \citet{wang02,durant05} \\
672: %1E 2259+586 & & $2.3-7.7\times10^{-16}$ & \\
673: % & & & \citet{hulleman01,tam04} \\
674: 4U 0142+61 & $(5.6-6.7)\times10^{-11}$ & $(0.6-1.5)\times10^{-15}$ &
675: \citet{gohler05,patel03} \\
676: & & & \citet{durant06} \\
677: CXOU J164710.2-455216 & $(0.001-1.7)\times10^{-11}$ &
678: $\la4\times10^{-16}$ & \citet{muno07,wang06} \\
679: 1RXS J170849.0-400910 & $(2.6-2.8)\times10^{-11}$ & $(2.2-3.2)\times10^{-15}$ &
680: \citet{campana07,durant06b} \\
681: \hline
682: \hline
683: \end{tabular}
684: \end{center}
685: \tablecomments{Only confirmed magnetars which had sufficient X-ray
686: spectral information available in the literature are included in
687: this table. The X-ray fluxes quoted in this table are the absorbed
688: flux, and no corrections have been made to either the X-ray flux or
689: $K_s$ fluxes for the different values of $N_H$ measured towards
690: these sources. The X-ray flux range of 1E~1547.0-5408 was
691: calculated using the same procedure described in the caption of
692: Table \ref{xrayfluxes}.}
693: \end{table}
694:
695: \clearpage
696: %% Figures
697: \begin{figure}
698: \begin{center}
699: \includegraphics[angle=0,scale=0.45]{f1a.eps}
700: \includegraphics[angle=0,scale=0.45]{f1b.eps}
701: \end{center}
702: \figcaption{X-ray images of the field around 1E~1547.0-5408 obtained
703: during the 2004 {\it XMM} observation ({\it top}) and 2006 {\it
704: Chandra} {\sc Acis--I} observation ({\it bottom}). The {\it XMM}
705: image combines data collected by the {\sc Mos1}, {\sc Mos2}, and
706: {\sc pn} instruments. Both images have been smoothed by a
707: 5\arcsec~Gaussian. The white circle indicates the position and
708: positional uncertainty of 1E~1547.0-5408 as measured by {\it
709: Einstein}, the black circle indicates the position and positional
710: uncertainty of 1E~1547.0-5408 and AX~J155052-5418, while the arrows
711: indicate the position of
712: CXOU~J155054.1-541824=XMMU~J155054.3--541825 as well as unrelated
713: field sources XMMU~J155053.7-541925 and
714: CXOU~J155107.7--541925=XMMU~J155108.1--541926. \label{xrayimg}}
715: \end{figure}
716:
717: \begin{figure}
718: \begin{center}
719: \includegraphics[angle=0,scale=1.0]{f2.eps}
720: \end{center}
721: \figcaption{The X-ray spectrum of CXOU~J155054.1-541824, as measured
722: by the {\sc Mos1}, {\sc Mos2}, and {\sc pn} detectors in the 2006
723: {\it XMM} observation. The {\sc Mos1} and {\sc Mos2} channels were
724: binned such that there was a minimum of 25 counts per bin, and the
725: {\sc pn} channels were binned such that there was a minimum of 50
726: counts per bin. The lines indicates the fit of the absorbed power
727: law plus blackbody model given in Table~\ref{xrayspec}, and the
728: bottom panel indicates the number of standard deviations the data
729: deviates from the model. \label{src1spec}}
730: \end{figure}
731:
732: \begin{figure}
733: \begin{center}
734: \includegraphics[angle=-90,scale=0.75]{f3.eps}
735: \end{center}
736: \figcaption{ $K_s$ image ($\lambda=2.15\mu$m) of the field around
737: CXOU~J155054.1-541824 (linear greyscale), overlaid with the 99\%
738: positional error circle of CXOU~J155054.1-541824 as measured by {\it
739: Chandra} (black circle) and the location of near-IR sources in this
740: field as determined by the {\it IRAF} package {\tt daofind} (white
741: circles). \label{nirimg}}
742: \end{figure}
743:
744: \begin{figure}
745: \begin{center}
746: \includegraphics[angle=270,scale=0.7]{f4.eps}
747: \end{center}
748: \figcaption{843~MHz image of G327.24-0.13 from the SUMSS Survey
749: \citep{bock99}, overlaid with the {\it Chandra} position of
750: CXOU~J155054.1-541824 (black cross; significantly larger than the
751: 99\% positional uncertainty of $\sim0\farcs8$). The synthesized
752: beam of the radio observation is shown in the lower-left corner
753: (black ellipse), and has dimensions
754: $53\arcsec\times43\arcsec$. \label{radimg}}
755: \end{figure}
756:
757: \begin{figure}
758: \begin{center}
759: \includegraphics[angle=0,scale=1.0]{f5.eps}
760: \end{center}
761: \figcaption{Observed IR Flux ($F_{K_s}$) versus Absorbed 2--7~keV
762: X-ray ($F_{X,2-7}$) flux of 1E~1547.0-5408, sources in the XBo\"otes
763: survey (filled circles) believed to be dominated by Active Galactic
764: Nuclei, sources in the {\it Chandra} Orion Ultradeep Project (open
765: stars) believed to be non-degenerate stars, and selected magnetars
766: -- both magnetars with detected near-IR counterparts (shaded
767: rectangles) and without (upper limits). If 1E~1547.0-5408 is
768: associated with the near-IR source within its positional error
769: circle shown in Figure \ref{nirimg}, the observed location of
770: 1E~1547.0-5408 is the filled square, while the locations of
771: 1E~1547.0-5408 for different values of $N_H$ are shown by the black
772: line, with the gray rectangle denoting the 1$\sigma$ error in
773: $F_{K_s}$. The units of $N_H$ are cm$^{-2}$, and the location of
774: 1E~1547.0-5408 for $0<N_H<10^{21}$~cm$^{-2}$ is indistinguishable
775: from the location for $N_H=10^{21}$~cm$^{-2}$ on this plot. The
776: sources in the XBo\"otes survey are gridded into columns due to the
777: small number of X-ray photons detected from most of these sources.
778: The X-ray and IR data for the magnetars, along with those for
779: 1E~1547.0-5408 (assuming that it is not detected in our near-IR
780: observation, and thus only have an upper limit on its near-IR
781: flux), are given in Table \ref{xrayirdata}.
782: \label{fluxratimg}}
783: \end{figure}
784:
785: \end{document}
786: