1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[apj]{emulateapj}
3: %
4: \usepackage{apjfonts,mathptmx}
5:
6:
7: \def\cxo {\emph{Chandra}}
8:
9: %
10: \begin{document}
11: \journalinfo{The Astrophysical Journal, in press}
12: \submitted{The Astrophysical Journal, in press}
13:
14: \title{\emph{Chandra} astrometry sets a tight upper limit to the\\
15: proper motion of SGR\,1900+14}
16:
17: \author{A.~De~Luca,\altaffilmark{1,2,4} P.~A.~Caraveo,\altaffilmark{1} P.~Esposito,\altaffilmark{1,3,4} and K.~Hurley\altaffilmark{5}}
18:
19: \altaffiltext{1}{INAF - Istituto di Astrofisica Spaziale e Fisica Cosmica Milano, via E.~Bassini 15, 20133 Milano, Italy}
20: \altaffiltext{2}{IUSS - Istituto Universitario di Studi Superiori, viale Lungo Ticino Sforza, 56, 27100 Pavia, Italy}
21: \altaffiltext{3}{Universit\`a degli Studi di Pavia, Dipartimento di Fisica Nucleare e Teorica, via A.~Bassi 6, 27100 Pavia, Italy}
22: \altaffiltext{4}{INFN - Istituto Nazionale di Fisica Nucleare, sezione di Pavia, via A.~Bassi 6, 27100 Pavia, Italy}
23: \altaffiltext{5}{Space Sciences Laboratory, University of California, 7 Gauss Way Berkeley, CA 94720-7450 USA}
24: \email{deluca@iasf-milano.inaf.it}
25:
26: \shortauthors{A.~De~Luca et al.}
27: \shorttitle{\emph{Chandra} proper motion of SGR\,1900+14}
28:
29:
30: % \date{}
31:
32:
33: \begin{abstract}
34: The soft gamma-ray repeater (SGR) SGR\,1900+14 lies a few arcminutes
35: outside the edge of the shell supernova remnant (SNR) G42.8+0.6. A
36: physical association between the two systems has been proposed - for
37: this and other SGR-SNR pairs - based on the expectation of high space
38: velocities for SGRs in the framework of the magnetar model. The large
39: angular separation between the SGR and the SNR center, coupled with the
40: young age of the system, suggest a test of the association with a proper
41: motion measurement.
42: %%% If the
43: %%% association is correct, a proper motion of at least $0.11$ arcsec yr$^{-1}$
44: %%% is expected for the neutron star.
45: We used a set of three \cxo/ACIS
46: observations of the field spanning $\sim5$ years to perform accurate
47: relative astrometry in order to measure the possible angular displacement
48: of the SGR as a function of time. Our investigation sets a $3\sigma$ upper
49: limit of 70 mas yr$^{-1}$ to the overall proper motion of the SGR. Such a
50: value argues against an association of SGR\,1900+14 with G42.8+0.6 and adds
51: further support to the mounting evidence for an origin of the SGR within a
52: nearby, compact cluster of massive stars.
53: \end{abstract}
54:
55: \keywords{X-rays: individual (SGR\,1900+14) -- ISM: individual (G42.8+0.6) -- X-rays: stars -- stars: neutron -- supernova remnants.}
56:
57: \section{Introduction}
58:
59: Soft gamma-ray repeaters
60: \citep[SGRs - see e.g. ][for reviews]{hurley00sgr,woods06,mereghetti08}
61: are a handful of four (plus a few candidates) sources of short
62: bursts of soft gamma rays. SGRs were originally thought to be a peculiar
63: subclass of Gamma-Ray Bursts (GRBs).
64: %- standing out
65: %among GRBs because of repeated bursting and softer spectra.
66: Contrary to the behavior of classical GRBs, SGRs produce series of bursts
67: over various time scales and the events are characterized by a soft spectral
68: shape. SGRs have a very rich high energy phenomenology. Apart from the flaring
69: activity, characterized by the emission of multiple, short ($\sim$0.1 s)
70: bursts with peak luminosities of $\sim$$10^{41}$ erg~s$^{-1}$, sometimes
71: culminating in dramatic, very energetic \emph{giant flares}
72: %- the most intense bursts of
73: %gamma rays of cosmic origin observed from Earth -
74: with peak luminosity exceeding $10^{47}$ erg~s$^{-1}$
75: \citep[as was the case of the 2004 December 27 event from SGR\,1806--20 -][]{hurley05short},
76: %Only one giant flare per source has been recorded so far (refs).
77: %On the other side,
78: SGRs display persistent emission from 0.1 to hundreds of keV, with pulsations
79: in the 5--8 s range, and with significant variability in flux, spectral shape,
80: pulse shape and pulsed fraction; they also exhibit glitches and an
81: irregular period derivative.
82:
83: This phenomenology has been interpreted within the framework of the
84: \emph{magnetar} model. SGRs are believed to be young, isolated neutron stars
85: (INSs) endowed with an ultra-high magnetic field (\mbox{$B\sim10^{15}$ G}),
86: which is thought to be the energy reservoir for all high energy emissions
87: \citep{duncan92,thompson95,thompson96}. It is now commonly accepted that
88: Anomalous X-ray Pulsars (AXPs), a peculiar class of X-ray pulsars
89: \citep[see ][for reviews]{woods06,mereghetti08}, could also be magnetars,
90: in view of a close similarity of several aspects of the high energy
91: phenomenology.\\
92: \indent The identification of SGRs with young INSs is on a rather firm basis.
93: However, in the past, an important supporting argument has been their possible
94: association with supernova remnants (SNRs). This provided a link between SGR'
95: formation and supernova explosions, and also set an upper limit to their age,
96: in view of the short lives of SNRs. Indeed, a possible SNR association had been
97: claimed for all SGRs \citep{hurley00sgr}. The validity of this association was
98: later reconsidered as new multiwavelength observations became available
99: \citep[see ][and references therein]{gaensler01}.
100: In one case (SGR\,1806--20) the mere existence of a nearby SNR was ruled out.
101: For the remaining cases, the SGR positions, close to the edges or even outside
102: their SNRs, cast doubt on the associations, based on two main considerations:
103: (i) the significant chance alignement probability; (ii) the need for large
104: spatial velocities (well in excess of 1000 km s$^{-1}$).
105: %The latter point is not
106: %unexpected for neutron stars in general, nor for magnetars in
107: %particular (refs).
108: Although such large velocities are not unheard of in the neutron star family
109: \citep{hobbs05}, AXPs do not show any evidence for them. This would suggest
110: that AXPs and SGRs are different classes of INSs, originating in intrinsically
111: different supernova explosion processes, in contrast to the robust evidence
112: for a very close relationship between SGR and AXP families. In view of this,
113: all SGR-SNR associations have been reconsidered, and a possible association of
114: two SGRs with nearby, massive star clusters has been proposed (SGR\,1806--20
115: by \citealt{fuchs99}; SGR\,1900+14 by \citealt{vrba00}).
116: %%> In any case, no {\em direct} evidences proving or ruling out the
117: %%> genetical link between a SGR and a nearby SNR have been obtained yet.
118:
119: %%> A conclusive measure could be the detection of a proper motion
120: %%> for a SGR consistent in direction and modulus with the
121: %%> observed offset with respect to the SNR's center and with
122: %%> a reasonable age estimate for the two systems.
123: %%> Such an observation is now within reach by performing relative
124: %%> astrometry on multi-epoch observations collected with
125: %%> the Chandra X-ray observatory. The superb angular resolution
126: %%> and the stability of aspect reconstruction of Chandra instruments
127: %%> allow to measure with high accuracy the relative displacement
128: %%> of an X-ray source with respect to a set of reference sources,
129: %%> which has been proved by Motch et al.(2007,2008).
130:
131: %%> We will study here the case of SGR\,1900+14.
132:
133: \section{SGR\,1900+14 and SNR G42.8+0.6}
134: SGR\,1900+14 lies outside the rim of G42.8+0.6, a shell-type, $\sim$$10^4$
135: year old radio SNR located at a distance of $3\div9$ kpc \citep{mlrh01}. An
136: association between the two systems was originally proposed by
137: \citet{vasisht94} and \citet{hurley96}.
138: %%Significant uncertainties affecting the age and the distance
139: %%of both the SGR and the SNR prevented from drawing firm
140: %%conclusions about their association.
141: SGR\,1900+14 lies towards a rather complicated region of the Galaxy, and
142: \citet{gaensler01} estimated a chance probability for the alignment of a
143: SNR to be as high as 4\%. \citet{kaplan02sgr} even increased this estimate
144: to $\sim$23\%. \citet{lorimer00} detected a young ($\tau_c \sim 38$ kyr)
145: radio pulsar about 2 arcmin away from the SGR. Such a pulsar could also be
146: plausibly associated with G42.8+0.6. Moreover, \citet{vrba00} discovered
147: a massive star cluster very close ($\sim$12 arcsec) to the position of the
148: SGR. Such observations weakened the case for an association between SGR\,1900+14
149: and G42.8+0.6. Very recently, \citet{wachter08} discovered an infrared ring
150: surrounding SGR\,1900+14 with \emph{Spitzer}. This structure was interpreted
151: as the rim of a dust cavity produced by past giant flares from the SGR,
152: heated by a nearby star cluster. This would point to a possible association
153: of SGR\,1900+14 with the star cluster, although a large difference in
154: reddening towards the SGR and the cluster remains to be explained
155:
156: In this work, we directly probe the association of SGR\,1900+14 and the SNR
157: G42.8+0.6 through a proper motion measurement. Such a possibility had been
158: suggested by \citet{hurley02}.
159: The angular separation between the SGR and the center of the SNR is $\sim$18
160: arcmin \citep{hurley02}.
161: %The proper motion depends
162: %on the age of the system.
163: Adopting a value of \mbox{$10^4$ yr}
164: %%% as the maximum age allowed by the magnetar model for the SGR
165: as a conservative estimate for the age of the system
166: \citep{thompson00}, the SGR-SNR association would
167: require a proper motion of at least 0.11 arcsec yr$^{-1}$.
168: %(where $t_{10}$ is the age in units of 10 kyr) is expected.
169: Such a proper motion can be measured using multi-epoch observations with the
170: \cxo\ X-ray Observatory. The superb angular resolution of its optics and
171: the stability of aspect reconstruction with its imaging detectors allow
172: measurements of tiny angular displacements through accurate relative
173: astrometry \citep[see e.g.][]{motch07,motch08}.
174:
175:
176: \section{\emph{Chandra} observations and data analysis}
177: \label{data}
178:
179: SGR\,1900+14 has been observed three times by \cxo\ between 2001 and 2006
180: using the Advanced CCD Imaging Spectrometer (ACIS). The first observation
181: was carried out in response to an AO-2 proposal whose goal was to obtain a
182: baseline measurement for proper motion studies. A log of the available data
183: is given in Table~\ref{cxodata}.
184: %%%%%%%%%%%%%%%%%%%TABLE
185: \begin{deluxetable}{lcccc}
186: % \tabletypesize{\scriptsize}
187: \tablecolumns{1}
188: \tablewidth{0pt}
189: \tablecaption{\emph{Chandra} observations of the field of SGR\,1900+14.\label{cxodata}}
190: \tablehead{
191: \colhead{Date} & \colhead{MJD} & \colhead{Obs. ID} & \colhead{Instrument} & \colhead{Exposure time}
192: }
193: \startdata
194: 2001 Jun 17 & 52,077 & 1954 & ACIS/I & 30.1 ks\\
195: 2002 Mar 11 & 52,344 & 3449 & ACIS/S & 2.7 ks \\
196: %2002, November 2$^{nd}$ & ACIS & 29.7 \\
197: 2006 Jun 4 & 53,890 & 6731 & ACIS/I & 25.0 ks\enddata
198: \tablecomments{All observations were performed using the Timed Exposure mode and the Faint event telemetry format.}
199: \end{deluxetable}
200: %%%%%%%%%%%%%%%%%%%TABLE
201: We retrieved event files from the \cxo\ X-ray Center Data Archive. All the
202: datasets have gone through
203: `reprocessing III'\footnote{See http://cxc.harvard.edu/ciao/repro\_iii.html.}
204: with updated software and calibration. According to the \cxo\ X-ray Center
205: guidelines,\footnote{See http://cxc.harvard.edu/ciao/threads/createL2/.}
206: no further reprocessing is required and archival `level 2' data were adopted
207: as a starting point. We used the \cxo\ Interactive Analysis of Observation
208: software (CIAO version 3.3) for our analysis.
209:
210: No significant background flares affected the observations. We removed pixel
211: randomization (we checked \emph{a posteriori} that fully consistent results
212: are obtained by using standard pixel position randomization) and we
213: generated images of the field by selecting photons in the 0.3--8 keV energy
214: range, binning the CCD pixel size by a factor of 2. The target was imaged
215: close to the aimpoint on the ACIS/I3 detector for Observations 1 and 3 and
216: on the ACIS/S3 detector for Observation 2.
217:
218: Source detection was done using the WAVDETECT task, with wavelet
219: scales ranging from 1 to 16 pixels, spaced by a factor $\sqrt{2}$. A
220: detection threshold of $10^{-5}$ was selected in order to avoid missing
221: faint sources. Cross-correlation of the source lists produced for each
222: observation (adopting a maximum source distance of 3 arcsec) allowed us to
223: reject spurious detections (about $\sim$70 per field in Observation 1 and
224: Observation 3). We identified a set of 31 common sources in Observation 1
225: and Observation 3 (excluding the target), while only six such sources
226: were found in Observation 2, which was significantly shorter and had a
227: smaller field of view.
228:
229: The probability of a chance alignment of two false detections is estimated
230: to be of order $\leq$0.1\%. Thus, we may safely assume that all of the
231: selected common sources are real.
232:
233: The uncertainty affecting the source's positions in the reference frame of
234: each image depends on the source signal-to-noise, as well as on the distance
235: from the aimpoint (because of point-spread function [PSF] degradation as a
236: function of off-axis angle). The position of the target, the brightest source
237: in the field, located close to the aimpoint, was determined with a $1\sigma$
238: error of 0.02 pixels per coordinate, while the position of a typical faint
239: source located several arcmin off-axis is affected by an uncertainty of order
240: 1 pixel per coordinate.
241:
242: \section{Relative Astrometry}
243:
244: Relative astrometry relies on accurate image superposition, based on a grid
245: of good reference sources. The positions of the sources selected in
246: Section~\ref{data}, together with their uncertainties, were used to compute
247: the best transformation needed to superimpose our multi-epoch images. We
248: took Observation 3 as a reference. To register the frames, we used a simple
249: rotation and translation. We found a strong dependence of the residuals on
250: the positions of the reference sources as a function of the distance to the
251: aimpoint. Superimposing Observation 1 to Observation 3, these residuals are
252: of order 0.2 pixels per coordinate within 4 arcmin from the aimpoint (12
253: reference sources); residuals grow to $\sim$0.6 pixels per coordinate between
254: 4 and 6 arcmin off-axis (8 reference sources); using 10 sources at off-axis
255: distances larger than 6 arcmin, the residuals are of order 1.3 pixels per
256: coordinate. This is most likely due to the degradation of the PSF with
257: off-axis angle, which hampers an accurate localization of the sources.
258:
259: We decided to use only the inner portion of the field (we checked
260: {\em a posteriori} that no different astrometric solutions are obtained
261: using the entire sample of reference sources). After excluding a source
262: deviating at more than $3\sigma$ with respect to the root mean square
263: (r.m.s.), we obtained a very good superposition using 11 sources, yielding
264: a $1\sigma$ error on the frame registration as small as 50 mas per coordinate.
265: The residuals of the reference source positions are $\sim$100 mas per
266: coordinate. To be conservative, this value was assumed as the $1\sigma$
267: uncertainty affecting our frame registration. We note that the best fit
268: roto-translation implies a frame registration with a similar uncertainty
269: (i.e. no significant transformation is required). We repeated the same
270: exercise to superimpose Observation 2 to Observation 3, which yielded similar,
271: consistent results, although based on a smaller sample of reference sources.
272: We then computed the target position in the reference frame of Observation 3,
273: in order to evaluate its possible displacement over the $\sim$5 year interval
274: spanned by the observations.
275:
276: We found no significant displacement in either coordinate. Accounting for the
277: uncertainty in the target position in each image, as well as for the
278: uncertainty involved in the frame superposition, a linear fit to the observed
279: relative positions sets an upper limit to the proper motion of SGR\,1900+14 of
280: 17 mas yr$^{-1}$ per coordinate. The $3\sigma$ upper limit to the overall
281: proper motion of the source in the plane of the sky is 70 mas yr$^{-1}$.
282:
283: \section{Absolute astrometry: X-ray versus radio positions.}
284: A precise localization of SGR\,1900+14 was obtained on 1998 September 10,
285: thanks to Very Large Array (VLA) observations of the source after the giant
286: flare of August 27 \citep{hurley99}. The accurate radio position was
287: $\rm RA = 19^h 07^m 14\fs33$, $\rm Dec.= +09\degr 19\arcmin 20\farcs1$
288: (J2000) with a $1\sigma$ uncertainty of $0\farcs15$ per coordinate
289: \citep{frail99}. We will take advantage of \cxo\ accurate absolute astrometry
290: to compare the X-ray position of the target with the 1998 radio position.
291:
292: \cxo\ absolute localization accuracy for on-axis sources has been carefully
293: evaluated by the calibration team. A typical radial uncertainty of
294: $\sim$$0\farcs4$ at 68\% confidence level affects ACIS/I
295: positions\footnote{See http://cxc.harvard.edu/cal/ASPECT/celmon/.},
296: while for ACIS/S observations this uncertainty is $\sim$$0\farcs2$.
297: In order to assess (and possibly improve) the absolute astrometry of the
298: \cxo\ dataset on SGR\,1900+14, we have cross-correlated the `good' source
299: list obtained in Section~\ref{data} with sources in the Two-Micron All-Sky Survey
300: \citep[2MASS,][]{skrutskie06} catalog, which has an astrometric accuracy of order
301: $0\farcs1$.\footnote{See http://spider.ipac.caltech.edu/staff/hlm/2mass/overv/overv.html.}
302:
303: For Observation~1 and Observation~2, we found 5 coincidences in 2MASS
304: within $0\farcs8$ of the X-ray position. Increasing the correlation radius
305: up to $2\farcs7$ yields no further match. This suggests that the 5 2MASS
306: sources are very likely to be the infrared counterparts of the corresponding
307: X-ray sources. Infrared colors and X-ray to infrared flux ratio suggest such
308: sources to be late stars of K--M spectral class. We used the 5 source
309: positions to register the \cxo\ images on the accurate 2MASS reference frame
310: by fitting a roto-translation, which yielded a r.m.s. of $\sim$200 mas per
311: coordinate.
312: %Such a value was assumed as the $1\sigma$ uncertainty affecting
313: %absolute coordinates of the corrected X-ray frame.
314: We note that
315: the \cxo-to-2MASS superposition did not require a significant transformation
316: (i.e., the corrections are of the same order of the residuals). We repeated
317: the same operation using Observation~3, where 2 of the above sources were
318: found and were used to adjust the astrometry by fitting a simple translation,
319: with r.m.s. residual of $\sim$300 mas per coordinate.
320: We computed the overall uncertainty in the absolute position of SGR\,1900+14
321: by summing the target localization accuracy on Chandra images, the
322: r.m.s. of the Chandra-2MASS frame superposition, and the 2MASS absolute
323: astrometric accuracy, in quadrature.
324:
325: The resulting positions of the target are given in Table \ref{abspos}. These
326: positions are (as expected) fully consistent with the accurate radio one
327: computed in 1998. With a simple linear fit, we estimate that absolute
328: astrometry sets a $3\sigma$ upper limit to the proper motion of SGR\,1900+14
329: of 100 mas yr$^{-1}$ per coordinate. Although such a limit is slightly less
330: stringent than the one obtained through relative astrometry, this is an
331: important consistency check of our result.
332: \begin{deluxetable}{lllc}
333: % \tabletypesize{\scriptsize}
334: \tablecolumns{1}
335: \tablewidth{0pt}
336: \tablecaption{Multi-epoch absolute positions (J2000) of SGR\,1900+14.\label{abspos}}
337: \tablehead{
338: \colhead{RA (error)} & \colhead{Dec. (error)} & \colhead{Date} & \colhead{Instrument}
339: }
340: \startdata
341: $19^{\rm h}07^{\rm m}14\fs33$ ($0\fs15$) & $+09\degr19\arcmin20\farcs1$ ($0\farcs15$) & 1998 Sep 10 & VLA \\
342: $19^{\rm h}07^{\rm m}14\fs33$ ($0\fs21$) & $+09\degr 19\arcmin 19\farcs6$ ($0\farcs21$) & 2001 Jun 17 & ACIS/I \\
343: $19^{\rm h}07^{\rm m}14\fs33$ ($0\fs31$) & $+09\degr 19\arcmin 19\farcs8$ ($0\farcs31$) & 2002 Mar 11 & ACIS/S \\
344: $19^{\rm h}07^{\rm m}14\fs31$ ($0\fs21$) & $+09\degr 19\arcmin 19\farcs8$ ($0\farcs21$) & 2006 Jun 4 & ACIS/I\enddata
345: \tablecomments{Errors are at 68\% confidence level.}
346: \end{deluxetable}
347:
348: \section{Conclusions}
349: Relative astrometry using a set of three \cxo\ observations spanning
350: 5 years yields no evidence for any angular displacement of SGR\,1900+14
351: in the plane of the sky.
352:
353: Our $3\sigma$ upper limit of 70 mas yr$^{-1}$
354: on the SGR proper motion could still be consistent
355: with a physical association of the SGR to G42.8+0.6. Indeed, the $\sim18'$
356: angular separation between the SGR and the SNR center would require, for an
357: association to hold,
358: $18' \, / \, (70\, mas \,yr^{-1})\,\sim\,15500$ yr as a $3\sigma$ {\em lower} limit
359: to the SGR/SNR age. This would point to a system significantly older than usually assumed.
360: The age of G42.8+0.6 is rather uncertain, in view of the poorly constrained
361: distance to the SNR \citep{mlrh01}, and could therefore fit within such a
362: picture.
363: On the other hand, estimating the true age of SGR 1900+14 is very difficult.
364: The characteristic age $\tau_c \, = \, P/ 2\dot{P}$ of the SGR, derived under standard magneto-dipole
365: braking assumptions, is as low as $\sim1300$ yr \citep{woods06}.
366: Such a value should be treated with caution, since the spin-down rate of
367: SGR 1900+14 has been observed to undergo significant variations \citep[but remaining
368: very high, in the $6-20\times10^{-11}$ s s$^{-1}$ range, throughout 20 years
369: of observations,][]{woods06}. The observed $\dot{P}$ changes
370: did not show any obvious correlation with variability in the SGR persistent
371: emission, nor with bursting activity, and ultimately, the physical mechanisms driving the peculiar
372: SGR spin-down evolution are not understood.
373:
374: In any case, it has been argued \citep{kouveliotou99,thompson00}
375: that $\tau_c$ could underestimate the true age of the SGR.
376: For instance, an additional
377: torque due to a charged particle wind -- leading to a different
378: long-term evolution wrt. pure magnetodipole torque (which implies $\dot{P}\propto
379: P^{-1}$) -- could play an important
380: role. \citet{thompson00} calculated the braking due to such a wind torque
381: (resulting in $\dot{P}\propto P$) and
382: estimated the true age of the SGR could be of $\sim4000$ years. Even such a
383: revised age seems uncomfortably low to be consistent with a $3\sigma$ lower
384: limit of 15500 yr. An additional hypothesis would be required, namely that we are
385: observing SGR 1900+14 in a transient phase of accelerated
386: spin-down \citep{kouveliotou99,thompson00}, lasting a fraction $\epsilon$
387: of the SGR true age $\tau$. Setting $\tau \, = \, \epsilon^{-1} \, P/ \dot{P}$,
388: our lower limit to the SGR age translates to a $3\sigma$ upper limit to
389: $\epsilon$ of $\sim0.15$. Unless such a behaviour to be unique
390: to SGR 1900+14, this would imply that SGRs spend a large majority of
391: their life in a ``slowly braking'' regime (which ignores the
392: braking due to their expected, very high dipole fields). Furthermore,
393: slowly braking SGRs should outnumber, by a factor $\epsilon^{-1}$,
394: SGRs with accelerated braking such as SGR 1900+14 \citep[][proposed to
395: identify such
396: slowly braking SGRs with Anomalous X-ray Pulsars; however
397: AXPs do not display, on average, a significantly slower spin-down than SGRs]{thompson00}.
398:
399: We believe that it is not worth further speculation to consider these issues.
400: In view of the above difficulties in associating SGR 1900+14 and
401: G42.8+0.6 (taking into account our upper limit to the SGR proper motion),
402: and of the
403: high chance alignement proability for the two systems,
404: Occam's razor argues against any physical link between them.
405:
406:
407:
408:
409:
410: %%% The
411: %%% $3\sigma$ upper limit to the angular displacement of the source rules out
412: %%% the minimum value of 0.11 arcsec yr$^{-1}$ required by the association of
413: %%% the SGR with SNR G42.8+0.6. The lack of a measurable proper motion allows
414: %%% us to exclude any physical link
415: %%% between the soft gamma-ray repeater and the
416: %%% supernova remnant, which appear to be (nearly) co-aligned on the plane of the
417: %%% sky by chance, as suspected by \citet{gaensler01} and \citet{kkf02}.
418:
419: Such a conclusion adds further support to the association of SGR\,1900+14 with the
420: nearby cluster of massive stars discovered by \citet{vrba00}. Morphological
421: evidence for a physical interaction of the gas surrounding the SGR and the
422: cluster's stars \citep{wachter08} argue against a simple chance alignment
423: for the two systems, suggesting that SGR\,1900+14
424: originated within the cluster.
425:
426: Of course, this leaves open the question of the fate of the remnant left by
427: the supernova in which the SGR originated. Dense gas and dust clouds
428: associated with the massive star cluster could hide the SNR emission.
429: If the birthplace of the SGR lies within the star cluster, a rather
430: tight upper limit to its space velocity may be set. The angular separation
431: of $\sim$12 arcsec \citep{vrba00} between the position of the SGR and the
432: center of the cluster implies a projected velocity of
433: $86\,d_{15}\,\tau_{10}^{-1}$ km s$^{-1}$ for the neutron star, where
434: $d_{15}$ is the distance in units of 15 kpc and $\tau_{10}$ is the age in
435: units of 10 kyr. Unless the SGR velocity is almost aligned along the line
436: of sight, this estimate is somewhat at odds with one of the basic expectations
437: of the magnetar model. As discussed by \citet{duncan92}, suppression of
438: convection due to the high magnetic field should give rise to anisotropies
439: in the core-collapse process, resulting in a very high recoil velocity
440: for the newborn neutron star (\mbox{$\sim$1000 km s$^{-1}$}). We note that to date, no
441: evidence for a high space velocity of any magnetar candidate has been found.
442:
443: In this context, a physical association of SGR\,1806--20 with a nearby cluster
444: of massive stars \citep{fuchs99} indirectly becomes more robust. The same is
445: true for other magnetar candidates, namely CXO\,J164710.2--455216
446: \citep{muno06} and possibly 1E\,1048.1--5937 \citep{gaensler05_1e1048},
447: although a revised distance estimate by \citet{durant06}
448: argues against the latter association. This mounting
449: evidence that a sizeable fraction of the magnetar family had very massive
450: progenitors raises several questions concerning magnetar origin and, more
451: generally, neutron star formation in core-collapse supernovae. Is there any
452: link between progenitor mass and the generation of high magnetic fields in
453: their compact remnants? What channels lead to the formation of young neutron
454: stars as diverse as magnetars
455: \citep[B-field$\ \, \sim10^{14}$ G, see e.g. ][]{mereghetti08}, energetic
456: `standard' radio pulsars (B-field$\ \, \sim10^{12}$ G) and Central Compact
457: Objects \citep[B-field$\ \, \leq10^{11}$ G, see e.g. ][]{deluca08}?
458: What is the maximum mass for a star to generate a neutron star?
459:
460: Focusing on the case of SGR\,1900+14, more detailed investigations of the
461: compact star cluster would be extremely useful, in order to assess the number
462: and class of its components and thus to estimate its age. Although this is a
463: rather difficult task, in view of the large reddening and the crowded region,
464: it would be very rewarding since it would set a lower limit to the mass of
465: the progenitor of the soft gamma-ray repeater.
466:
467: \acknowledgements
468: Chandra data analysis is supported by INAF-ASI contract n.I/023/05/0.
469: KH is grateful for support under the \cxo\ \mbox{AO-2} guest investigator program,
470: Smithsonian grant \mbox{GO1-2053X}. This research has made use of data obtained from the Chandra Data Archive
471: and software provided by the \cxo\ X-ray Center (CXC) in the application
472: packages CIAO, ChIPS, and Sherpa. This publication makes use of data products from the Two-Micron All-Sky
473: Survey, which is a joint project of the University of Massachusetts and
474: the Infrared Processing and Analysis Center/California Institute of
475: Technology, funded by the National Aeronautics and Space Administration
476: and the National Science Foundation.
477:
478:
479: \bibliographystyle{apj}
480: \bibliography{biblio}
481:
482: \end{document}
483: