1: %%
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 2004 January 9
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17: %%
18:
19: \documentclass[12pt,preprint]{aastex}
20: %\usepackage{rotate}
21: %\usepackage{deluxetable}
22: %% manuscript produces a one-column, double-spaced document:
23:
24: %% \documentclass[manuscript]{aastex}
25:
26: %% preprint2 produces a double-column, single-spaced document:
27: %
28: %
29: % V0.3 - 15/6/2006 S.M.
30: % V0.4 - spedita ai coautori
31: % V0.5 - 7/7/2006 S.M.+P.E. - nuove figure e commenti DG, SZ, MF, RT
32: % V0.6 - 13/7 SM+PE figure spettri - tolto hardening
33: % V0.7 - 15/7 Silvia
34: % V0.8 - 17/7 small changes - Abstract, summary - SM
35: % V0.10 - 23/7 SM + link sivia + ultimi commenti DG/MF - submitted 24/7
36: % V1.00 - 16/8 SM taken from ApJ web server to make changes after referee
37:
38: %% \documentclass[preprint2]{aastex}
39:
40: %% Sometimes a paper's abstract is too long to fit on the
41: %% title page in preprint2 mode. When that is the case,
42: %% use the longabstract style option.
43:
44: %% \documentclass[preprint2,longabstract]{aastex}
45:
46: %% If you want to create your own macros, you can do so
47: %% using \newcommand. Your macros should appear before
48: %% the \begin{document} command.
49: %%
50: %% If you are submitting to a journal that translates manuscripts
51: %% into SGML, you need to follow certain guidelines when preparing
52: %% your macros. See the AASTeX v5.x Author Guide
53: %% for information.
54:
55: \newcommand{\vdag}{(v)^\dagger}
56: \newcommand{\myemail}{sandro@iasf-milano.inaf.it}
57:
58: %% You can insert a short comment on the title page using the command below.
59:
60: % \slugcomment{Submitted to The Astrophysical Journal on ..../2006}
61:
62: %% If you wish, you may supply running head information, although
63: %% this information may be modified by the editorial offices.
64: %% The left head contains a list of authors,
65: %% usually a maximum of three (otherwise use et al.). The right
66: %% head is a modified title of up to roughly 44 characters.
67: %% Running heads will not print in the manuscript style.
68:
69: \shorttitle{XMM-Newton observations of SGR 1900$+$14}
70:
71: \shortauthors{Mereghetti et al.}
72:
73: %% This is the end of the preamble. Indicate the beginning of the
74: %% paper itself with \begin{document}.
75:
76: \def \src {SGR\thinspace1900$+$14}
77: \def\approxgt{\mathrel{\hbox{\rlap{\lower.55ex \hbox {$\sim$}}
78: \kern-.3em \raise.4ex \hbox{$>$}}}}
79: \def\approxlt{\mathrel{\hbox{\rlap{\lower.55ex \hbox {$\sim$}}
80: \kern-.3em \raise.4ex \hbox{$<$}}}}
81: \def \xmm {XMM-Newton }
82: \def \sax {Beppo-SAX }
83: \def\pdot {\dot P}
84: \def\flux {\mbox{erg cm$^{-2}$ s$^{-1}$}}
85: \def\lum {\mbox{erg s$^{-1}$}}
86: \def\nh {$N_{\rm H}$}
87:
88:
89: \begin{document}
90:
91: %% LaTeX will automatically break titles if they run longer than
92: %% one line. However, you may use \\ to force a line break if
93: %% you desire.
94:
95:
96: \title{The first XMM-Newton observations of the Soft Gamma-ray
97: Repeater SGR 1900$+$14 \footnote{Based on obervations obtained with XMM-Newton,
98: an ESA science mission with instruments and contributions
99: directly funded by ESA Member States and NASA.}}
100:
101:
102:
103:
104: %% Use \author, \affil, and the \and command to format
105: %% author and affiliation information.
106: %% Note that \email has replaced the old \authoremail command
107: %% from AASTeX v4.0. You can use \email to mark an email address
108: %% anywhere in the paper, not just in the front matter.
109: %% As in the title, use \\ to force line breaks.
110:
111: \author{S. Mereghetti, P. Esposito\altaffilmark{2}, A. Tiengo}
112: \affil{INAF - IASF Milano, via Bassini 15, I-20133 Milano, Italy}
113: % \email{sandro@mi.iasf.cnr.it}
114:
115:
116: \author{S. Zane}
117: \affil{Mullard Space Science Laboratory, University College
118: London, \\ Holmbury St. Mary, Dorking Surrey, RH5 6NT, United Kingdom}
119:
120:
121: \author{R. Turolla}
122: \affil{Universit\`a di Padova, Dipartimento di Fisica, via Marzolo
123: 8, I-35131 Padova, Italy}
124:
125: \author{L. Stella, G.L. Israel}
126: \affil{INAF - Osservatorio Astronomico di Roma,
127: via Frascati 33, \\ I-00040 Monteporzio Catone, Italy}
128:
129:
130: \author{ D. G\"{o}tz}
131: \affil{CEA Saclay, DSM/DAPNIA/Service d'astrophysique, F-91191 Gif-sur-Yvette, France}
132:
133: \author{M. Feroci}
134: \affil{INAF - IASF Roma, v. Fosso del Cavaliere 100, I-00133 Roma,
135: Italy }
136:
137:
138: %% Notice that each of these authors has alternate affiliations, which
139: %% are identified by the \altaffilmark after each name. Specify alternate
140: %% affiliation information with \altaffiltext, with one command per each
141: %% affiliation.
142:
143: \altaffiltext{2}{Universit\`a di Pavia, Dipartimento di Fisica Nucleare e Teorica and INFN-Pavia, via Bassi 6, I-27100 Pavia, Italy}
144: %\altaffiltext{2}{SRON - National Institute for Space Research,
145: %Sorbonnelaan, 2, 3584 CA, Utrecht, The Netherlands }
146:
147: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
148: %% style, abstract will output a Received/Accepted line after the
149: %% title and affiliation information. No date will appear since the author
150: %% does not have this information. The dates will be filled in by the
151: %% editorial office after submission.
152:
153: \begin{abstract}
154: A $\sim$50 ks XMM-Newton observation of \src\ has been carried
155: out in September 2005, after almost three years during which no
156: bursts were detected from this soft gamma-ray repeater. The 0.8-10
157: keV spectrum was well fit by a power law plus blackbody model with
158: photon index $\Gamma$=1.9$\pm$0.1, temperature kT=0.47$\pm$0.02
159: keV and N$_H$ = $(2.12\pm0.08)\times10^{22}$ cm$^{-2}$, similar to
160: previous observations of this source. The flux was
161: $\sim5\times10^{-12}$ \flux , a factor 2 dimmer than the typical
162: value and the smallest ever seen from \src . The long term fading
163: of the persistent emission has been interrupted by the recent
164: burst reactivation of the source. A target of opportunity
165: XMM-Newton observation performed in April 2006 showed a flux
166: $\sim$15\% higher. This variation was not accompanied by
167: significant changes in the spectrum, pulsed fraction and light
168: curve profile. We searched for emission and absorption lines in
169: the spectra of the two observations, with negative results and
170: setting tight upper limits of 50--200 eV (3$\sigma$), depending on
171: the assumed line energy and width, on the equivalent width of
172: lines in the 1-9 keV range.
173: \end{abstract}
174:
175: %% Keywords should appear after the \end{abstract} command. The uncommented
176: %% example has been keyed in ApJ style. See the instructions to authors
177: %% for the journal to which you are submitting your paper to determine
178: %% what keyword punctuation is appropriate.
179:
180: %% Authors who wish to have the most important objects in their paper
181: %% linked in the electronic edition to a data center may do so in the
182: %% subject header. Objects should be in the appropriate "individual"
183: %% headers (e.g. quasars: individual, stars: individual, etc.) with the
184: %% additional provision that the total number of headers, including each
185: %% individual object, not exceed six. The \objectname{} macro, and its
186: %% alias \object{}, is used to mark each object. The macro takes the object
187: %% name as its primary argument. This name will appear in the paper
188: %% and serve as the link's anchor in the electronic edition if the name
189: %% is recognized by the data centers. The macro also takes an optional
190: %% argument in parentheses in cases where the data center identification
191: %% differs from what is to be printed in the paper.
192:
193: \keywords{stars: individual (SGR 1900+14) -- stars: neutron}
194:
195: %% From the front matter, we move on to the body of the paper.
196: %% In the first two sections, notice the use of the natbib \citep
197: %% and \citet commands to identify citations. The citations are
198: %% tied to the reference list via symbolic KEYs. The KEY corresponds
199: %% to the KEY in the \bibitem in the reference list below. We have
200: %% chosen the first three characters of the first author's name plus
201: %% the last two numeral of the year of publication as our KEY for
202: %% each reference.
203:
204: \section{Introduction}
205:
206: The few X-/gamma-ray sources known as Soft Gamma-ray Repeaters
207: (SGRs) are generally believed to provide the most convincing
208: evidence for the existence of magnetars, i.e. neutron stars with
209: magnetic fields well above the quantum critical value $B_{c} =
210: \frac{m_{e}^{2}c^{3}}{e\hbar}=4.4\times10^{13}$~G \citep{duncan92,
211: paczynski92}.
212:
213: The first sources of this class were discovered in the seventies
214: as transient emitters of short (tens of ms) bursts of hard X-rays
215: \citep{mazets79,cline80,laros86}. Only in more recent years it was
216: possible to identify and study in detail their counterparts in the
217: classical 1--10~keV X-ray band. This led to the discovery of
218: periodic pulsations and secular spin-down in SGR 1806--20 (P=7.5
219: s, \citealt{kouveliotou98}) and \src\ (P=5.2 s,
220: \citealt{hurley99,kouveliotou99}) thus confirming the neutron star
221: nature of this class of sources. Occasionally, SGRs emit very
222: energetic giant flares, during which up to a few 10$^{46}$ ergs
223: are released in a few tenths of a second. The extreme properties
224: of these events, of which only three, each one for a different
225: source, have been observed to date, are the main motivation for
226: the magnetar interpretation \citep{thompson95,thompson96}.
227:
228: Among the four confirmed SGRs, SGR 1806--20 and \src\, offer the
229: best prospects for detailed spectral and timing studies in the
230: soft X-ray band (E$<$10 keV). Both these sources, despite
231: alternating periods of bursting activity with intervals of
232: quiescence lasting months or years, remained at flux levels of the
233: order of 10$^{-11}$ \flux. The other Galactic soft repeater, SGR
234: 1627--41, had a similar X-ray flux when it was discovered as a
235: bursting source in 1998 \citep{woods99c}, but since then its
236: luminosity decreased by a factor $\sim$25 and it is now a rather
237: faint source \citep{mereghetti06}. SGR 0526--66 has a relatively
238: high luminosity ($\sim2\times10^{35}$ \lum), but being at the
239: Large Magellanic Cloud distance, its observed flux is only
240: $\sim$10$^{-12}$ \flux\ \citep{kulkarni03}. Furthermore, its study
241: with low spatial resolution instruments is complicated by the
242: presence of diffuse emission from the surrounding supernova
243: remnant N49.
244:
245: We started in 2003 a long--term monitoring program to study the
246: time evolution of the spectral properties of SGR 1806--20 using
247: the XMM-Newton X-ray satellite. Thanks to its imaging capability
248: and large effective area we could obtain spectra of a much better
249: quality than those available from previous satellites. The
250: emission of the 27 December 2004 giant flare was also a fortunate
251: occurrence, since we could observe how the source properties
252: evolved in the two years leading to the flare and how they changed
253: after this dramatic event \citep{mte05,tiengo05}.
254:
255: A similar observational campaign could not be carried out for
256: \src. In fact this source is located in a sky region that, until
257: recently, was not accessible to XMM-Newton observations due to
258: technical constraints in the satellite pointing. Thus the first
259: observation of \src\ with this satellite could be obtained only
260: in September 2005, during a long period of inactivity (the last
261: bursts before the XMM-Newton observation were reported in November
262: 2002, \citealt{hurley02}). \src\ became active again in March
263: 2006: bursts were observed by Swift \citep{palmer06} and
264: Konus-Wind \citep{golenetskii06}. We therefore requested a target
265: of opportunity XMM-Newton observation, that was carried out on 1
266: April 2006.
267:
268:
269:
270:
271:
272: \section{Observations}
273:
274:
275: We present the results obtained with the EPIC instrument,
276: consisting of two MOS and one pn cameras
277: \citep{turner01,struder01}. In all the observations the pn was
278: operated in Full Frame mode and the MOSs in Large Window mode
279: (time resolution: 73.4 ms and 0.9 s respectively). Both the pn and
280: the MOSs mounted the medium thickness filter. All the data were
281: processed using version 6.5.0 of the \emph{XMM-Newton Science
282: Analysis System} and the most recent calibration files (last
283: update on 2005 December 14). Response matrices and effective area
284: files were generated ad-hoc with the SAS tasks \emph{rmfgen} and
285: \emph{arfgen}; spectral fits were performed using the XSPEC
286: v11.3 software \citep{arnaud96}.
287:
288: The first observation of \src\ was divided in two parts, starting
289: on 2005 September 20 01:44 UT and 22 01:36 UT, respectively.
290: Since there was no evidence for variations in the flux and
291: spectrum of the source from September 20 to 22, we added the two
292: data sets and analyzed them together. After filtering for particle
293: induced flares we obtained a net exposure time of 38.9 ks in the
294: pn camera, and of 47.4 ks in the two MOSs.
295:
296: The second observation started on 1 April 2006 and lasted $\sim$22
297: ks, yielding net exposure times of 12.7 ks in the pn camera and of
298: 15.7 ks in the two MOSs.
299:
300:
301: The 0.8--10 keV image obtained with the pn camera in September
302: 2005 is shown in Fig.\ref{ds9}. \src\
303: %, with a count rate of $0.615\pm0.004$ counts s$^{-1}$,
304: is the brightest source at the
305: center of the field. Several other objects, detected here for the
306: first time are visible. As expected for such a low Galactic
307: latitude field, many of them can be associated with foreground
308: stars based on their soft spectrum and positional coincidence with
309: bright optical counterparts. A relatively bright spatially
310: resolved source is also visible $\sim5'$ to the West of \src , but
311: it is very likely unrelated to the SGR. Its spectrum is well
312: described by an optically thin plasma emission model (MEKAL in
313: XSPEC) with temperature \mbox{$kT=7^{+3}_{-2}$ keV} and a high
314: absorption of $N_{\rm H}=(3.6^{+1.0}_{-0.7})\times10^{22}$
315: cm$^{-2}$. This spectrum and the spatial extension of about one
316: arcminute are consistent with emission from a cluster of galaxies
317: at redshift $z\sim0.6$ and with a 2--10 keV luminosity of
318: $\sim$$2\times10^{44}$ \lum. Its coordinates are R.A.=19$^{h}$
319: 06$^{m}$ 53$^{s}$.7, Dec.=+09$^{\circ}$ 20$'$ 47$''$ (J2000).
320: % 19:06:53.704 e +09:20:47.52
321:
322:
323: \section{Timing and spectral results}
324:
325: Except for the periodic pulsations, \src\ did not show flux
326: variability within the two observations, but it was about 15\%
327: brighter in April 2006, after the burst reactivation. We searched
328: for the presence of bursts in both observations, by a careful
329: analysis of light curves binned with different time resolution,
330: but none could be found. With a standard folding analysis of the
331: Solar system barycentered light curves, we measured a spin period
332: of \mbox{$5.198346\pm0.000003$ s} in September 2005 and
333: \mbox{$5.19987\pm0.00007$ s} in April 2006. In Figure \ref{fol} we
334: show the background subtracted pulse profiles in three different
335: energy ranges. The pulsed fractions (values reported in the
336: corresponding figures) have been computed by fitting a sinusoid to
337: the light curves. There is no evidence for changes in the pulsed
338: fractions and light curve shapes between the two observations. The
339: two period measurements correspond to a spin-down rate of
340: (9.2$\pm$0.4)$\times10^{-11}$ s s$^{-1}$.
341:
342:
343: We extracted spectra for \src\ by selecting source counts with
344: patterns 0--4 for the pn camera and 0--12 for the MOS cameras from
345: circles of 40\arcsec\ radius. The background spectra were
346: extracted from composite regions located on the same chip as the
347: source. The spectra were rebinned to have at least 30 counts in
348: each bin and to oversample the instrumental energy resolution by a
349: factor three. Fits were performed in the energy range 0.8--12 keV,
350: since the source is heavily absorbed and only few counts are
351: detected at lower energies.
352:
353: In Fig.~\ref{spec1} and \ref{spec2} we show the spectrum obtained
354: with the pn camera in the September 2005 observation, fitted with
355: a power law and with a power law plus black body model,
356: respectively. The latter clearly provides a better fit, as it can
357: be seen from the residuals shown in the lower panels of the
358: figures. Similar results were obtained using the spectra from the
359: MOS. We therefore performed simultaneous fits of the spectra from
360: the three cameras, obtaining photon index $\Gamma$=1.9$\pm$0.1,
361: blackbody temperature kT=0.47$\pm$0.02 keV, and absorption N$_H$ =
362: $(2.12\pm0.08)\times10^{22}$ cm$^{-2}$. An acceptable fit could
363: also be obtained with the sum of two blackbodies with temperatures
364: of 0.53 and 1.9 keV (see Table~\ref{fits}).
365:
366: The second observation gave entirely consistent spectral
367: parameters, except for a statistically significant variation in
368: the flux. The background subtracted count rates (0.8-10 keV) measured with the
369: pn camera were $0.615\pm0.004$ counts s$^{-1}$ in September 2005
370: and $0.720\pm0.008$ counts s$^{-1}$ in April 2006.
371: Indeed the April 2006 data are well described simply
372: rescaling in normalization (by $\sim$15\%) the best fit spectra of
373: the September 2005 observation.
374:
375:
376: For both observations we performed phase-resolved spectroscopy
377: extracting the spectra for different selections of phase
378: intervals. No significant variations with phase were detected, all
379: the spectra being consistent with the model and parameters of the
380: phase-averaged spectrum, simply rescaled in normalization.
381:
382: No evidence for emission or absorption lines was found by
383: inspecting the residuals from the best fit models. We computed
384: upper limits on the lines equivalent widths as a function of the
385: assumed line energy and width. This was done by adding Gaussian
386: components to the model and computing the allowed range in their
387: normalization. The most constraining results were obtained for the
388: September 2005 observation. They are summarized in Figure
389: \ref{lines}, where the top panel refers to the phase averaged
390: spectrum and the other ones to the spectra of the pulse maximum
391: (phase from 0.25 to 0.75 of Fig.~\ref{fol}) and minimum.
392:
393:
394:
395: \section{Discussion}
396:
397: Previous spectral studies of the persistent X-ray emission of
398: \src, carried out with ASCA \citep{hurley99}, BeppoSAX (e.g.,
399: \citealt{woods99b}, \citealt{esposito06} and references therein)
400: and Chandra \citep{kouveliotou01}, showed that a blackbody plus
401: power law model often provides a better fit than a single power
402: law. The blackbody temperature was always of $\sim$0.4-0.5 keV and
403: the power law photon index $\Gamma\sim$2 (except for the only
404: BeppoSAX observation carried out before the 1999 giant flare, that
405: had a harder spectrum with $\Gamma$=1.1). The XMM-Newton best fit
406: parameters are in agreement with these values, but the flux of
407: $\sim$$4.8\times10^{-12}$ \flux ~measured in our September 2005
408: observation is the lowest ever detected from \src. A $\sim$30\%
409: decrease of the persistent emission, compared to the
410: ``historical'' level of $\sim$$10^{-11}$ \flux, had already been
411: noticed in the last BeppoSAX observation \citep{esposito06}, that
412: was carried out in April 2002, six months earlier than the last
413: bursts reported before the recent reactivation. This is
414: illustrated in Fig.~\ref{multi}, where we have plotted the long
415: term evolution of the pulse period, X--ray flux and bursting rate
416: of \src.
417:
418:
419:
420: The long term fading experienced by \src\ in 2002-2005 might be
421: related to the apparent decrease in the bursting activity in this
422: period and can be compared to that of SGR 1627--41. SGR~1627--41
423: experienced a short period of bursting activity in June-July 1998
424: and, during the following $\sim$2 years, its 2-10 keV flux
425: decreased with time as a power law
426: F(t)$\propto$(t$-$t$_0$)$^{-0.6}$, with t$_0$ indicating the time
427: of the outburst \citep{mereghetti06}. As suggested by
428: \citet{kouveliotou03}, this behavior is likely due to the fact
429: that, during outbursts, a substantial amount of energy is
430: deposited in the deep crustal layers ($\sim500-600$ m in depth)
431: due to shear dissipation and magnetic reconnection. Heat is then
432: transported inwards (because the conductivity increases at larger
433: densities) and later gradually transferred to the surface.
434: \citet{lyubarsky02} computed the surface cooling evolution, in
435: plane parallel approximation, by assuming a constant magnetic
436: field perpendicular to the surface and by solving numerically the
437: heat flow equation. They found that in a time scale of a few days
438: the deep crustal layers are cooled by inward heat flow, and that
439: 80\% of the deposited energy is transferred to the core and
440: re-radiated over longer timescales as surface X-ray emission.
441: Quite independently on the details of the initial energy
442: deposition, this model gives a cooling luminosity that scales in
443: time as $t^{-0.7}$, in agreement to what has been observed for SGR
444: 1627--41 \citep{mereghetti06}.
445:
446: An alternative scenario to explain the``afterglows'' following
447: magnetars outbursts is that surface heating is caused by the
448: currents flowing in an azimuthally twisted magnetosphere
449: \citep{tlk02,gotthelf05}. The basic idea is that the toroidal
450: component of the internal magnetic field stresses the crust,
451: inducing a deformation and causing the external field to acquire
452: an azimuthal component. In this case a current density in excess
453: of the Goldreich-Julian current (which is expected for a simple
454: dipolar field) is required to thread the magnetosphere. As the
455: twist angle grows, the bursting activity is expected to increase
456: and larger returning currents heat the star surface producing more
457: thermal photons. By assuming a simple cylindrically symmetric and
458: self-similar magnetosphere, \citet{tlk02} derived an upper limit
459: for the luminosity of the returning currents, $L_X^{rc}\simeq
460: 10^{35} B^p_{14} \Delta \phi$~erg s$^{-1}$, where $\Delta \phi$ is
461: the twist angle and $B^p_{14}$ is the polar value of the magnetic
462: field in unit of $10^{14}$~G. In this model the luminosity decay
463: is dictated by the time evolution of the current (and that of the
464: consequent, almost instantaneous surface heating), but no detailed
465: computations have been performed so far.
466:
467:
468: The luminosity decay shown by \src\ has been much smaller than
469: that of SGR 1627--41, since the flux of the former source only
470: faded by a factor $\sim$2 in three years. By fitting the observed
471: decay with a power law gives $F(t) \propto (t-t_0)^{-0.17}$, where
472: we have taken as t$_0$ the time of the intermediate flare of 18
473: April 2001 \citep{feroci03}. The flatter slope may be an
474: indication that a mechanism of the second kind (i.e. surface
475: heating by returning currents) is at work. On the other hand, the
476: one-dimensional model computed by \citet{lyubarsky02} assumed that
477: the internal magnetic field is essentially radial. There is now
478: increasing theoretical and observational evidence that strong
479: poloidal and toroidal components can be present in the neutron
480: star crustal magnetic field (see e.g. \citealt{geppert06} and
481: references therein). This affects dramatically the heat transfer,
482: that becomes strongly anisotropic. A strong magnetic field
483: channels the heat flow along its field lines and, in the presence
484: of large meridional components, can produce large inhomogeneities
485: in the surface temperature distribution. Moreover, toroidal
486: fields substantially limit the radial conductivity (heat
487: blanketing) forcing energy to be transferred into narrow regions
488: along the polar axis. Although no detailed computations are
489: available, we may argue that, by assuming that the initial energy
490: deposition per unit volume is the same, crustal fields with large
491: poloidal and toroidal components might produce flatter power law
492: luminosity decays, due to a combination of a smaller emitting
493: surface area and of the lower efficiency of the radial
494: conductivity in establishing a substantial thermal gradient
495: between the core and the surface (the latter being proportional to
496: the flux of heat outward).
497:
498:
499: We found no evidence for emission or absorption lines in the X-ray
500: spectra. The upper limits obtained in the longer observation of
501: September 2005 are the most constraining ever obtained for this
502: source in the ~1-10 keV energy range. An emission line at 6.4 keV
503: was possibly detected with the PCA instrument on RXTE in August
504: 1998 \citep{strohmayer00}. This line, visible only for the first
505: 0.3 s of a particularly long and hard burst, had an equivalent
506: width of $\sim$400 eV and was interpreted as Fe fluorescence from
507: relatively cool material possibly ejected during the giant flare
508: that occurred two days before its detection. Thus it is not
509: surprising that we do not find evidence for the same feature in
510: the spectrum of the persistent emission.
511:
512: In models involving ultra-magnetized neutron stars, proton
513: cyclotron features are expected to lie in the X-ray range, for
514: surface magnetic fields strengths of $\sim 10^{14}-10^{15}$~G.
515: Detailed calculations of the spectrum emerging from the
516: atmosphere of a magnetar in quiescence have confirmed this basic
517: expectation \citep{zane01,ho01}. Model spectra exhibit a strong
518: absorption line at the proton cyclotron resonance, $E_{c,p}\simeq
519: 0.63 z_G(B/10^{14}\, {\rm G})$ keV, where $z_G$, typically in the
520: 0.70--0.85 range, is the gravitational red-shift at the neutron
521: star surface. However, no evidence for persistent cyclotron
522: features have been reported to date in SGRs, despite some features
523: have been possibly detected during bursts (see e.g.
524: \citealt{strohmayer00,ibrahim03}).
525:
526: Indeed some reasons have also been proposed to explain the absence
527: of cyclotron lines in magnetars, besides the obvious possibility
528: that they lie outside the sampled energy range. First, it must be
529: noticed that the atmospheric models available so far only account
530: for a single temperature and a single value of magnetic field
531: strength and inclination in the atmosphere, and no source of
532: heating besides the standard core cooling is taken into account.
533: Again, the lack of a standard atmosphere in active magnetars, and
534: the fact that their magnetic field topology and surface
535: temperature distribution are likely to be complex, makes the non
536: detection of proton cyclotron features in the persistent emission
537: not surprising.
538:
539: Moreover, in the model discussed by \citet{tlk02}, magnetars have
540: highly twisted magnetospheres that can support current flows.
541: These, in turn, can substantially distort the thermal emission
542: from the neutron star surface. The presence of charged particles
543: ($e^-$ and ions) produces a large resonant scattering depth and
544: the resonant frequency depends on the local value of the magnetic
545: field. If the source flux at the cyclotron resonance does not
546: exceed the luminosity of the returning currents, the distributions
547: of both electrons and ions are spatially extended, in which case
548: repeated scatterings could lead to the formation of a hard tail,
549: typically observed below $\sim 10$~keV, instead of a narrow line.
550: Another implication of this model is that the twisted
551: magnetospheres can act as a source of gamma rays, either through
552: bremsstrahlung from a thin turbulent layer of the star's surface
553: heated to kT$\sim$100 keV by magnetospheric currents or through
554: synchrotron emission from pairs produced at a height of $\sim100$
555: km above the neutron star \citep{thompson05}. Indeed a hard
556: X--ray tail extending to 100 keV has been recently discovered in
557: \src\ with the INTEGRAL satellite \cite{gotz06}. A different
558: explanation for the absence of lines involves vacuum polarization
559: effects. It has been calculated that in strongly magnetized
560: atmospheres this effect can significantly reduce the equivalent
561: width of cyclotron lines, thus making difficult their detection
562: \citep{ho03}.
563:
564:
565:
566:
567:
568: \section{Conclusions}
569:
570: Thanks to the high sensitivity of the EPIC instrument on
571: XMM-Newton we have obtained the first high quality spectra of the
572: persistent X--ray emission from \src , setting tight limits on the
573: presence of emission and absorption lines. In September 2005 the
574: source was found at a luminosity level of 1.3$\times10^{35}$ \lum\
575: (for d=15 kpc), a factor two smaller than the typical value
576: observed in the past, and in line with the trend of luminosity
577: decrease already observed in the latest BeppoSAX observations
578: performed in April 2002. The target of opportunity XMM-Newton
579: observation of April 2006 showed that the decreasing luminosity
580: trend in \src\ has been interrupted by the recent onset of bursts
581: emission.
582: %However, this reactivation did not induce dramatic
583: %changes in the spectral and timing properties of the persistent
584: %X--ray flux. No significant variations were seen in the pulse
585: %profiles, pulsed fractions, and spectral shape.
586: However, the moderate flux increase was not associated with
587: significant changes in the X--ray spectral and timing properties,
588: probably because the source is, up to now, only moderately active.
589: Future observations with XMM-Newton will be essential to monitor
590: the spectral and flux variations for this source, possibly in
591: connection with its renewed bursting activity, as it has been
592: successfully done for its twin source SGR 1806--20.
593:
594:
595:
596:
597: We thank N.Schartel and the staff of the XMM-Newton Science Operation
598: Center for performing the target of opportunity observation.
599: This work has been partially supported by the Italian Space Agency
600: and INAF through contract ASI/INAF I/023/05/0
601: and by the MIUR under grant PRIN 2004-023189.
602:
603:
604:
605: \bibliographystyle{aa}
606: \bibliography{biblio}
607:
608: %\end{references}
609: \clearpage
610:
611:
612: \begin{deluxetable}{cccccccccc}
613: \rotate
614: \tabletypesize{\scriptsize}
615: \tablecolumns{2}
616: \tablewidth{0pc}
617: \tablecaption{Summary of the spectral results in
618: the 0.8--12 keV energy range\label{fits}}
619: \tablehead{\colhead{Model\tablenotemark{a}} &
620: \colhead{Observation} & \colhead{$N_{\rm H}$} &
621: \colhead{$\Gamma$} & \colhead{$k_B T_1$} &
622: \colhead{$R_{bb\,1}$\tablenotemark{b}} & \colhead{$k_B T_2$} &
623: \colhead{$R_{bb\,2}$\tablenotemark{b}} & \colhead{Flux\tablenotemark{c}} & \colhead{$\chi^{2}_{r}$ (d.o.f.)}\\
624: \colhead{} & \colhead{} & \colhead{($10^{22}$ $\rm cm^{-2}$)} & \colhead{} &
625: \colhead{(keV)} & \colhead{(km)} & \colhead{(keV)} & \colhead{(km)}&
626: \colhead{($10^{-12}$ erg cm$^{-2}$ s$^{-1}$)} & \colhead{}}
627: \startdata
628: PL & A & $2.57\pm0.05$ & $2.84\pm0.04$ & \nodata & \nodata & \nodata & \nodata & $4.85\pm0.07$ & 1.65 (467) \\
629: \nodata & B & $2.71\pm0.08$ & $2.81\pm0.06$ & \nodata & \nodata & \nodata & \nodata & $5.6\pm0.1$ & 1.21 (323) \\
630: BB+BB & A & $1.82\pm0.06$ & \nodata & $0.53^{+0.01}_{-0.02}$ & $3.7^{+0.3}_{-0.2}$ & $1.9\pm0.1$ & $0.22\pm0.02$ & $4.6\pm0.1$ & 1.32 (465) \\
631: \nodata & B & $2.0^{+0.1}_{-0.2}$ & \nodata & $0.53^{+0.03}_{-0.02}$ & $3.9\pm0.4$ & $1.9^{+0.2}_{-0.1}$ & $0.23^{+0.04}_{-0.03}$ & $5.3\pm0.2$ & 1.04 (321) \\
632: PL+BB & A & $2.12\pm0.08$ & $1.9\pm0.1$ & $0.47\pm0.02$ & $4.0^{+0.4}_{-0.3}$ & \nodata & \nodata & $4.8\pm0.2$ & 1.24 (465) \\
633: \nodata & B & $2.3^{+0.1}_{-0.2}$ & $1.9\pm0.2$ & $0.47\pm0.03$ & $4.2\pm0.5$ & \nodata & \nodata & $5.5\pm0.4$ & 1.00 (321) \\
634: \enddata
635: \tablenotetext{a}{Errors are quoted at the 90\% confidence level
636: for a single parameter.} \tablenotetext{b}{Radius at infinity
637: assuming a distance of 15 kpc.} \tablenotetext{c}{Flux in the
638: 2--10 keV range, corrected for the absorption. The flux errors take into
639: account the whole range of uncertainties in the spectral parameters.}
640: \end{deluxetable}
641:
642:
643:
644:
645: \clearpage
646: \begin{figure}
647: \includegraphics[angle=00,width=15cm]{f1.ps}
648: \caption{\label{ds9} EPIC-pn image of the \src\ field in the
649: 0.8--10 keV energy range. North is to the top, East to the left.}
650: \end{figure}
651:
652: \clearpage
653:
654: \begin{figure}
655: \includegraphics[angle=90,width=15cm]{f2.ps}
656: \caption{\label{fol}Folded light curves in the total (0.8-10 keV),
657: soft (0.8-4 keV), and hard (4-10 keV) energy range for the two
658: observations. The background has been subtracted. The
659: corresponding pulsed fraction is indicated on each panel
660: (1$\sigma$ errors).
661: }
662: \end{figure}
663:
664: \clearpage
665:
666:
667: \begin{figure}
668: \includegraphics[angle=270,width=15cm]{f3.ps}
669: \caption{\label{spec1} EPIC pn spectrum of \src\ from the
670: September 2005 observation. Top: data and best fit power law
671: model. Bottom: residuals from the best fit model in units of
672: standard deviations.}
673: \end{figure}
674:
675:
676: \begin{figure}
677: \includegraphics[angle=270,width=15cm]{f4.ps}
678: \caption{\label{spec2} EPIC pn spectrum of \src\ from the
679: September 2005 observation. Top: data and best fit power law plus
680: blackbody model. Bottom: residuals from the best fit model in
681: units of standard deviations.}
682: \end{figure}
683:
684:
685: \clearpage
686:
687: \begin{figure}
688: \includegraphics[angle=90,width=13cm]{f5.ps}
689: \caption{\label{lines} Upper limits (at 3 $\sigma$) on spectral
690: features in the 2005 pn data of \src\ . The top panel refers to
691: the phase-averaged spectrum and the two lower panels to the
692: spectra at the pulse minimum and maximum.
693: }
694: \end{figure}
695: \clearpage
696:
697:
698: \begin{figure}
699: \includegraphics[angle=0,width=15cm]{f6.ps}
700: \caption{\label{multi} Long term evolution of the pulse period
701: (top panel), X--ray flux (middle panel) and bursting rate observed with
702: the IPN (bottom
703: panel) of SGR 1900+14. The vertical dashed lines indicate the
704: times of the 27 August 1998 giant flare and of the 18 April 2001
705: intermediate flare. }
706: \end{figure}
707: \clearpage
708:
709: \end{document}
710: