astro-ph0502428/ms.tex
1: \documentstyle[12pt,psfig,draft,lscape,array,amssymb]{nature-ejb}
2: 
3: \textwidth16cm\textheight24.2cm\voffset-2cm\hoffset+1.5cm
4: \voffset-0.35cm
5: 
6: \def\simlt{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
7: \def\simgt{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
8: 
9: \def\ale{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
10: \def\age{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
11: 
12: \def\nodata{---}
13: 
14: \def\ra#1#2#3{#1$^{\rm h}$#2$^{\rm m}$#3$^{\rm s}$}
15: \def\dec#1#2#3{$#1^\circ#2'#3''$}
16: 
17: \newcommand{\sgr}{SGR\,1806$-$20}
18: 
19: \begin{document}
20: 
21: \title{\bf Discovery of a Radio Source following the 27 \\ December 2004 Giant Flare from SGR\,1806$-$20}
22: 
23: \author{ 
24: P.~B.~Cameron\affiliation[1]{\scriptsize Division of Physics, Mathematics and Astronomy,
25:   105-24, California Institute of Technology, Pasadena, CA 91125, USA},
26: P.~Chandra\affiliation[2]{\scriptsize Tata Institute of Fundamental Research, Mumbai 400 005,
27:   India}$^,$\affiliation[3]{\scriptsize Joint Astronomy Programme, 
28:   Indian Institute of Science, Bangalore 560 012, India},
29: A.~Ray\affiliationmark[2],
30: S.~R.~Kulkarni\affiliationmark[1],
31: D.~A.~Frail\affiliation[4]{\scriptsize National Radio Astronomy Observatory, Socorro, NM 87801, USA},
32: M.~H.~Wieringa\affiliation[5]{\scriptsize Australia Telescope National Facility, CSIRO,
33: P.O. Box 76, Epping NSW 1710, Australia},
34: E.~Nakar\affiliation[6]{\scriptsize Theoretical Astrophysics 130-33, California Institute of
35: Technology, Pasadena, CA\,91125, USA},
36: E.~S.~Phinney\affiliationmark[6],
37: Atsushi Miyazaki\affiliation[7]{\scriptsize Shanghai Astronomical Observatory 80 Nandan Road 
38: Shanghai 200030, China},
39: Masato Tsuboi\affiliation[8]{\scriptsize Nobeyama Radio Observatory,National Astronomical 
40: Observatory of Japan Minamisaku, Nagano 384-1305, Japan},
41: Sachiko Okumura\affiliationmark[8],
42: N.~Kawai\affiliation[9]{\scriptsize Department of Physics Tokyo Institute of Technology 
43: Ookayama 2-12-1, Meguro-ku,Tokyo 152-8551, Japan},
44: K.~M.~Menten\affiliation[10]{\scriptsize Max Planck Institut f\"ur Radioastronomie, Auf dem H\"ugel
45: 69; 53121 Bonn, Germany} \&\
46: F.~Bertoldi\affiliation[11]{\scriptsize University of Bonn, Auf dem H\"uegel 71, 53121 Bonn, Germany}
47: }
48: \date{\today}{}
49: \headertitle{\sgr\ Fading Radio Source}
50: \mainauthor{Cameron et al.}
51: 
52: \summary{
53: Over a decade ago it was established that the remarkable high energy
54: transients, known as soft gamma-ray repeaters (SGRs), are a Galactic
55: population\cite{kf93,mtk+94} and originate from neutron stars with
56: intense ($\simlt 10^{15}\,$G) magnetic fields (``magnetars''\cite{wt04}).
57: On 27 December 2004 a giant flare\cite{GCN2920} (fluence\cite{GCN2936}
58: $\age 0.3\,$erg cm$^{-2}$) was detected from \sgr.  Here we report
59: the discovery of a fading radio counterpart.  We began a monitoring
60: program from 0.2\,GHz to 250\,GHz and obtained a high resolution
61: 21-cm radio spectrum which traces the intervening interstellar
62: neutral Hydrogen clouds.  Analysis of the spectrum yields the first
63: direct distance measurement of \sgr. The source is located at a 
64: distance greater than 6.4\,kpc and we argue that it is nearer than 9.8\,kpc.  
65: If true, our distance estimate lowers the total energy of the explosion 
66: and relaxes the demands on theoretical models. 
67: The energetics and the rapid decay of the
68: radio source are not compatible with the afterglow model that is
69: usually invoked for gamma-ray bursts. 
70: Instead we suggest that the rapidly decaying radio emission arises from the debris
71: ejected during the explosion.
72: }
73: 
74: \maketitle
75: 
76: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
77: 
78: On 3 January 2005 we observed \sgr\
79: with the Very Large Array (VLA) and identified and promptly
80: reported\cite{GCN2928} a new radio source at right ascension
81: $\alpha_{J2000}=$\ra{18}{08}{39.34} and declination
82: $\delta_{J2000}=$\dec{-20}{24}{39.7} (with an uncertainty of $\pm
83: 0.1$'' in each coordinate) coincident with the quiescent X-ray
84: counterpart\cite{kfk+02}. In Table~\ref{tab:fluxes} we report the
85: results of a subsequent monitoring program undertaken with the VLA, the
86: Giant Metrewave Radio Telescope (GMRT), the Australia Telescope
87: Compact Array (ATCA), the Nobeyama Millimeter Array (NMA) and the
88: Institut de Radioastronomie Millim$\acute{\rm e}$trique (IRAM) 30m Telescope. 
89: 
90: The radio source decays in all frequency bands, but the behaviour
91: is complex (Figure~\ref{fig:LightCurves}).  At each band we model
92: the flux by a power law, $S_\nu(t)\propto t^\alpha$, but
93: allow for changes in the temporal indices $\alpha$ (``breaks'') at two epochs. These
94: breaks are clearly seen in our highest signal-to-noise ratio data.
95: Following the first break (9 days, postburst) the light curve steepens 
96: to $-4 \simlt \alpha \simlt -3$.
97: The radio source\cite{fkb99} from SGR\,1900+14 following the 27 August 1998 giant flare\cite{fhd+01}
98: showed a similar rapid decay at 8\,GHz.
99: Subsequently around day 14 the light curve flattens to
100: $\alpha\sim -1$.  At any given epoch, the radio spectrum can be
101: modeled by a power law, $S_\nu \propto \nu^{\beta}$.  The spectral
102: index, $\beta$, steepens with time, changing from $\sim -0.7$ to $-0.9$ (see
103: Figure~\ref{fig:LightCurves} and Supplemental Information).  
104: 
105: We confirm claims that the source is resolved\cite{GCN2933} by an independent
106: analysis.  We find that it is elongated with a major axis $\theta\sim
107: 77$\,milliarcsecond (mas) and an axial ratio of 2:1 (Table~\ref{tab:size}).
108: We considered four expansion models, $\theta\propto t^{s}$ with 
109: unconstrained $s$ and  plausible expansion models ($s=0$, 2/5, and 1). 
110: The best fit model corresponds to no expansion ($s=0.04\pm 0.15$). However,
111: due to the limited range of our observations we prefer not to model the dynamics of the
112: explosion.
113: 
114: We took advantage of the brightness of the radio source and obtained
115: a high resolution spectrum (Figure~\ref{fig:hone}, second panel) centered around
116: the 21-cm line of atomic Hydrogen (HI).   Intervening interstellar
117: clouds appear as absorption features in the spectrum. These clouds
118: are expected to participate in the rotation of the Galaxy and the
119: absorption features allow us to infer ``kinematic'' distance estimates.
120: Such estimates have several complications. First, in the inner Galaxy the
121: radial velocity curve is double-valued (see Figure~\ref{fig:hone}, third panel)
122: leading to a ``near'' distance estimate
123: (d$_l$) and a ``far'' distance estimate (d$_u$) for each velocity. Second, in
124: some directions, there are features with non-circular motion e.g.
125: the ``3-kpc expanding arm'' and the ``$-30$ km\,s$^{-1}$ spiral
126: arm''\cite{ce04}.  Finally, in the innermost part of the Galaxy
127: there is a deficit of cold gas\cite{kjbd02}.
128: 
129: Significant HI absorption toward \sgr\ is seen over the velocity
130: range $-$20 km\,s$^{-1}$ to +85 km\,s$^{-1}$ (Figure~\ref{fig:hone}, second panel).  
131: There is also a weak (2.5-$\sigma$) absorption feature coincident in velocity with a
132: clearly detected $^{12}$CO(1-0) emission feature identified\cite{cwd+97} as MC94 (Figure~\ref{fig:hone}, first panel).
133: Adopting a simple Galactic rotation curve with a circular velocity
134: $\Theta_\circ=220$ km\,s$^{-1}$ and a Galactic center distance
135: R$_\circ$=8.5\,kpc, the near distance to \sgr\ (for V$_{LSR}$=95 
136: km\,s$^{-1}$) is d$_l$=6.4\,kpc.
137: 
138: The two HI emission clouds seen at velocities above 100 km\,s$^{-1}$
139: toward \sgr\ (Figure~\ref{fig:hone}, first panel), with no corresponding HI absorption, may be used to
140: infer an upper limit to the distance provided that we can be
141: reasonably certain that cold neutral gas exists at these velocities.
142: The HI absorption spectrum toward the nearby ($\Delta\theta=0.77^\circ$)
143: extragalactic source J1811$-$2055 shows a strong and broad absorption
144: feature between 110 and 130 km\,s$^{-1}$ (Figure~\ref{fig:hone}, fourth panel).
145: The only HI emission in this direction\cite{hb97} above 60 
146: km\,s$^{-1}$ corresponds to an HI cloud at this same velocity. This
147: feature can be traced in absorption toward several other
148: extragalactic radio sources in this direction \cite{gd89}, suggesting
149: that cold gas at $\sim$120 km\,s$^{-1}$ is widespread.  
150: Adopting the same Galactic rotation curve as above, the
151: absorbing cloud at +120 km\,s$^{-1}$ can either be located at 7\,kpc
152: or 9.8\,kpc (see Figure~\ref{fig:hone}, third panel).
153: We thus suggest an upper limit to the distance, d$_u$=9.8\,kpc.  
154: 
155: Our new distance estimate is smaller than previous (indirect) values\cite{fnk04,ce04} 
156: of 12 to 15\,kpc. Accepting our estimate has several important implications.
157: It results in a reduction of the total energy released ($\propto d^2$) as
158: well as the rate of such events in nearby galaxies\cite{ngp+05} ($\propto d^3$), 
159: and calls into question the association\cite{fmc+99} of \sgr\ 
160: with a star cluster along the same line-of-sight.
161: Therefore, claims that magnetars originate from more massive stars than normal
162: neutron stars\cite{gmo+05} may be called into question.
163: 
164: Next we consider the energetics of the material giving rise to 
165: the radio emission.
166: As with many other radio sources, the power law spectrum can be
167: attributed to energetic electrons with a power law energy distribution
168: ($dN/d\gamma \propto \gamma^{-p}$; here $\gamma$ is the Lorentz
169: factor of electrons and we measure $p = 2.24 \pm 0.04$ on day 7, 
170: which as is a typical value for strong shocks) 
171: which gyrate in a magnetic field and emit synchrotron radiation.  
172: We apply the minimum energy formulation for synchrotron sources\cite{Pac70,sr77}
173: to the radio spectrum (from 0.2\,GHz to 100\,GHz) of 3 January 2005 and find the
174: energy of the radio emitting source and the associated magnetic
175: field strength are $U_{min} \sim  10^{43} d_{10}^{17/7}
176: \theta_{75}^{9/7}$ erg and $B_{min} \sim 13 d_{10}^{-2/7}
177: \theta_{75}^{-6/7}$ mG; here, the distance is 10$d_{10}$\,kpc and the
178: angular diameter is 75$\theta_{75}$\,mas (Table~\ref{tab:size}).
179: 
180: Evidently the amount of energy released in the $\gamma$-ray flare,
181: $E_{\gamma,{\rm iso}}\simgt 4\times 10^{45}\,$erg\,s$^{-1}$ (assuming
182: unbeamed, isotropic emission),
183: substantially exceeds $U_{\rm min}$. 
184: In contrast, the ratio $U_{\rm min}/E_{\gamma,{\rm iso}}$ is unity for 
185: GRBs and as a result the lower energy and
186: longer duration emission is correctly regarded as arising from the
187: shock of the circumburst medium (the ``afterglow'' model).  Thus,
188: based solely on energetics, there is no {\em prima facie} reason to
189: suggest that the radio source is the afterglow of the $\gamma$-ray
190: flare. 
191: 
192: Furthermore, as discussed above, the radio emission decays quite
193: rapidly 9 days after the burst.
194: Such a rapid decay is incompatible
195: with the afterglow model (in the non-relativistic limit)  for which
196: we expect\cite{fwk00} $\alpha=3\beta+0.6$.  We conclude (in contrast
197: to refs.~\pcite{cw03}, \pcite{nps05} and \pcite{wwf+05}) that  the
198: radio emission must be powered by something very different from
199: that which produced the $\gamma$-ray emission. 
200: 
201: In summary, the radio emission can be described by two components:
202: a rapidly decaying component and a slowly decaying component.
203: The latter
204: becomes detectable when the former has decayed significantly.  The
205: rapid decay is phenomenologically similar to that seen from accreting
206: Galactic sources (e.g. ref.  \pcite{hrh+00}) -- the so-called
207: ``plasmon'' model framework in which the radio emission arises from
208: a ball of electrons and magnetic field which are initially shocked and
209: then cool down by expansion. 
210: We make the specific suggestion that
211: the radio emission up until about 2 weeks is a result of the shocking
212: of the debris given off in the explosion (the ``reverse shock'').
213: In this framework the slowly decaying component is the emission
214: arising from the forward shock as the ejecta slams into the circumburst
215: medium. A requirement of this suggestion is that the energy inferred
216: in the slowly decaying component should be comparable to $U_{\rm min}$.
217: Separately, we note that the comparable ratios $U_{min}/E_{\gamma,{\rm iso}}$
218: and the temporal and spectral similarities of 
219: the giant flares from \sgr\ and SGR\,1900$+$14 suggest a common mechanism for launching
220: these flares and similar circumstellar environments.
221: 
222: 
223: Regardless of the suggestions and speculations, it is clear that radio afterglow is telling us something
224: entirely different from that revealed by the $\gamma$-ray emission.
225: If our suggestion of a reverse shock origin is correct then radio
226: observations allow us to probe the ejecta. Taken together
227: it appears that rapid and intense radio monitoring of
228: such flares will be highly fruitful in the future.
229: 
230: \bibliographystyle{nature-pap}
231: \bibliography{journals,magnetar}
232: 
233: \bigskip
234: \noindent
235: {\bf Supplementary Information} accompanies the paper on
236: {\bf www.nature.com/nature.}
237: 
238: \begin{acknowledge}
239: ATCA is funded by the Commonwealth of Australia for operation as a
240: National Facility managed by CSIRO.  We thank K. Newton-McGee and B.
241: Gaensler for scheduling and performing observations with the ATCA.
242: GMRT is run by the National Centre for Radio Astrophysics-Tata
243: Institute of Fundamental Research, India.  We thank the GMRT staff and
244: in particular C. H. Ishwara-Chandra and D. V. Lal for help with
245: observations and analysis.  The VLA is a facility of the National
246: Science Foundation operated under cooperative agreement by Associated
247: Universities, Inc.  NMA is a branch of the National Astronomical
248: Observatory, National Institutes of Natural Sciences, Japan.  IRAM is
249: supported by INSU/CNRS (France), MPG (Germany) and IGN (Spain).  We
250: thank A.  Weiss from IRAM for help with the observations.  We
251: gratefully acknowledge S. Corbel, S. S.  Eikenberry and R. Sari for
252: useful discussions.  Our work is supported in part by the NSF and
253: NASA.
254: 
255: \bigskip
256: \noindent
257: {\bf Correspondence} and requests for materials should be addressed to P.B.C.
258: (pbc@astro.caltech.edu).
259: 
260: 
261: 
262: \end{acknowledge}
263: 
264: \clearpage
265: \vspace{0.5cm}
266: \centerline{\bf FIGURES \& CAPTIONS}
267: \vspace{0.5cm}
268: 
269: \noindent
270: {\bf Table~\ref{tab:fluxes}}: {\small 
271: Flux density measurements of the transient radio counterpart to
272: \sgr\ from the VLA, GMRT, NMA, and ATCA as a function of frequency
273: and time.  The reported errors are 1-$\sigma$.  In addition to these
274: measurements, we obtained IRAM-30m observations on 8 and 9 January
275: 2005 using MAMBO-2 at 250\,GHz which show no detection with a value
276: of $0.57 \pm 0.46$~mJy at the position of the radio source.  Finally,
277: we detect linearly polarized emission from the source at the 1.5\%
278: to 2.5\% level.  See the Supplemental Information for observational
279: details.\\ $^a$ The epoch of the flare, $t_0$, was 27.90 December
280: 2004.\\ $^b$ ATCA observations in this column have a frequency of
281: 8.6\,GHz.\\ $^c$ These values represent 2$\sigma$ upper limits.\\
282: $^d$ The frequency is 1.06\,GHz for the 16.37 January 2005 and 4.01
283: February 2005 GMRT observations.\\ 
284: }
285: 
286: 
287: \noindent
288: {\bf Figure~\ref{fig:LightCurves}}: {\small Broadband temporal
289: behavior of the transient radio source coincident with \sgr.  The
290: abscissa indicates days elapsed since the giant flare on 27.90
291: December 2004.  The displayed flux density measurements (denoted
292: with symbols) were obtained in six frequency bands with the VLA,
293: GMRT, and ATCA (Table~\ref{tab:fluxes}).  The error bars denote
294: 1-$\sigma$ uncertainties.  With the exception of the 6.1\,GHz data
295: (which is insufficiently sampled at early and late times and is not
296: shown), the light curves with $\nu > 1\,$GHz are best fit by power-law
297: models (shown as lines, $S_\nu \propto t^{\alpha_i}$) with two
298: breaks at $t_1 \sim 9$ days and $t_2 \sim 15$ days (see
299: the Supplemental Information for exact values).  The
300: temporal index varies chromatically in the time before and after
301: the first break (denoted by regions A and B respectively).  
302: The exponent value ranges from $-2 \simlt \alpha_A \simlt -1$ and $-4
303: \simlt \alpha_B \simlt -3$; here the subscript identifies the region
304: of interest.  After day $\sim 15$ (region C) the source decay
305: flattens to $\alpha_C \sim -0.9$ at these frequencies which persists
306: until day 51.
307: Region B, the period of steep light curve decline, is shaded gray. 
308: The light curves with $\nu < 1\,$GHz do not show these temporal breaks or late time flattening
309: Apparently a single power law decay model 
310: with $\alpha=-1.57 \pm 0.11$ (0.24\,GHz)
311: and $\alpha=-1.80 \pm 0.08$ (0.61\,GHz)
312: provides a good statistical description of the data.
313: 
314: Our substantial frequency coverage (over three decades) allows
315: an excellent characterization of the spectrum.  The spectrum
316: is consistent with a single power law slope ($S_\nu \propto \nu^\beta$)
317: at all epochs.  On day 7, before the first temporal break, we find $\beta =
318: -0.62 \pm 0.02$.  The spectrum steepens to a value of $\beta = -0.76
319: \pm 0.05$ (day 15), reaching $\beta = -0.9 \pm 0.1$ (days 21--51).\\
320: }
321: 
322: 
323: \noindent
324: {\bf Table~\ref{tab:size}}: {\small Source size measurements and
325:    95\% confidence limits of the radio source as measured with the
326:    VLA at 8.46\,GHz.  The source is clearly resolved at all VLA
327:    epochs. The best constraints on the source size come from the
328:    observations which occurred closest to the transit of the
329:    source on Jan 3rd and Jan 6th.  The result is a source with size
330:    $\theta \sim 77$ mas with an axial ratio of $\sim 0.5$ and a
331:    position angle (PA) of 60 degrees (measured clockwise from the
332:    North).  The flux centroid did not change position within the
333:    limits of our astrometric accuracy ($\pm$ 100 mas). 
334: 
335:    The best fit model is consistent with no expansion, 
336:    $s = 0.04 \pm 0.15$ ($\theta(t) \propto t^s$) with a $\chi^2 = 5.3$ 
337:    with 4 degrees of freedom.  
338:    Sedov-Taylor ($s=2/5$) and free expansion ($s=1$) model were also fit, and
339:    yield $\chi^2 = 20.6$ and $\chi^2 = 91$, respectively. These fits
340:    have  five 
341:    degrees of freedom.
342:    See the Supplemental Information for the details of the source size
343:    measurements.\\
344: }
345: 
346: \noindent
347: {\bf Figure~\ref{fig:hone}}: {\small Cold atomic and molecular hydrogen spectra 	
348:    toward \sgr. These spectra were used to derive a distance estimate for \sgr.
349:    {\em (Top Panel)}. HI emission
350:    (upper curve, thick line) in the direction of \sgr, determined
351:    by averaging two adjacent spectra taken by Hartmann \&
352:    Burton\cite{hb97} at $l,b$=(10.0$^\circ$,0.0$^\circ$) and
353:    $l,b$=(10.0$^\circ$,$-0.5^\circ$). The lower curve (thin line)
354:    is the $^{12}$CO(1$-$0) spectrum (from ref.~\pcite{ce04}).
355:    For display purposes the brightness temperature  has been
356:    scaled up by a factor of 11.4.
357:    {\em (Second Panel).}
358:    The HI absorption spectrum taken toward \sgr.
359:    The two horizontal bars illustrate the radial velocity
360:    measurements\cite{fnk04,ce04} of the nearby star LBV 1806$-$20
361:    (36$\pm$10\,km\,s$^{-1}$ and 10 $\pm$20\,km\,s$^{-1}$).  
362:    The absorption spectra were
363:    made with the Very Large Array on 4 January 2005, using a 1.56
364:    MHz bandwidth in both hands of polarization centered 50 
365:    km\,s$^{-1}$ with respect to the local standard of rest.  The bandwidth
366:    was divided into 256 channels each 6.1 kHz in width, or a velocity
367:    resolution of 1.3 km\,s$^{-1}$ covering a velocity range of $-$115
368:    km\,s$^{-1}$ to $+$215 km\,s$^{-1}$.
369:    {\em (Third Panel).} The distance as a function of radial velocity adopting a
370:    simple Galactic rotation curve with a circular velocity
371:    $\Theta_\circ=220$ km\,s$^{-1}$ and a galactic center distance
372:    R$_\circ$=8.5\,kpc.  
373:    {\em (Bottom Panel).} The HI absorption spectrum
374:    of nearby extragalactic source J1811$-$2055 at
375:    $l,b$=(9.8$^\circ$,$-1.0^\circ$).  \\
376: 
377: The lower limit to the distance is firmly established by the 95\,km\,s$^{-1}$
378: absorption feature from MC94 (see Text). 
379: An upper limit to the distance of the SGR is suggested by the {\it
380: absence} of strong absorption at +120 km s$^{-1}$, seen toward
381: J1811$-$2055 and several other extragalactic radio sources in this
382: direction\cite{gd89}. One could argue that the
383: 120\,km\,s$^{-1}$ cold cloud is small and not present along the
384: line-of-sight to the \sgr.  However, this hypothesis also requires
385: the absence of any other cloud between 95\,km\,s$^{-1}$ (distance
386: of 6.4\,kpc) and 86\,km\,s$^{-1}$ (distance of 10.6\,kpc).
387: The mean absorption coefficient drops in the inner Galaxy ($R \simlt 4.5$)
388: kpc, giving a mean free path between clouds of 2.3 kpc\cite{gd89}.
389: The distance interval from 6.4\,kpc to 10.6\,kpc
390: corresponds to $\sim 1.8$ mean free paths. So, the probability of finding no
391: clouds in this gap is 16\%. 
392: Thus our upper limit of 9.8\,kpc is not a certainty but quite likely.
393: }
394: 
395: \begin{landscape}
396: \begin{table}
397: \begin{center}
398: \setlength{\extrarowheight}{-0.090in}
399: \begin{tabular}{>{\scriptsize}l >{\scriptsize}l >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c}
400: \hline
401: \hline
402: & & $\Delta t^a$&S$_{0.24}$&S$_{0.610}$&S$_{1.46}$&S$_{2.4}$&S$_{4.86}$&S$_{6.1}$&S$_{8.46}^b$&S$_{102}$\\
403: Epoch & Telescope & (days)&  (mJy)& (mJy)& (mJy)& (mJy)& (mJy)& (mJy)& (mJy)&(mJy)\\
404: \hline
405: 3.84 Jan 2005 & VLA & 6.94 & \nodata       & \nodata      & 178 $\pm$ 4 & \nodata    & 79 $\pm$ 2    & \nodata   & 53 $\pm$ 1    & \nodata\\
406: 4.17 Jan 2005 & NMA & 7.27 & \nodata       & \nodata      & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & 16.3 $\pm$ 5.6\\
407: 4.41 Jan 2005 & GMRT& 7.51 & 466 $\pm$ 28  & 224 $\pm$ 13 & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & \nodata  \\
408: 4.59 Jan 2005 & VLA & 7.69 & \nodata       & \nodata      & 161 $\pm$ 4 & \nodata & 66 $\pm$ 2    & \nodata   & 44 $\pm$ 1    & \nodata \\
409: 5.26 Jan 2005 & ATCA& 8.36 & \nodata       & \nodata      & 127 $\pm$ 3 & 80 $\pm$ 2 & \nodata       &  \nodata  & \nodata       & \nodata\\
410: 5.66 Jan 2005 & VLA & 8.76 & \nodata       & \nodata      & \nodata     & \nodata    & 55 $\pm$ 1    & \nodata   & 33 $\pm$ 1    & \nodata \\
411: 5.85 Jan 2005 & ATCA& 8.93 & \nodata       & \nodata      & 113 $\pm$ 3 & 63 $\pm$ 2 & 53  $\pm$ 2   & \nodata   & 30 $\pm$ 1    & \nodata \\
412: 6.26 Jan 2005 & ATCA& 9.36 & \nodata       & \nodata      & 96 $\pm$ 3  & 73 $\pm$ 2 & 45 $\pm$ 2    &  \nodata  & 23 $\pm$ 1    & \nodata \\ 
413: 6.38 Jan 2005 & GMRT& 9.48 & 462 $\pm$ 29  & 142 $\pm$ 8  & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & \nodata \\
414: 6.77 Jan 2005 & VLA & 9.87 & \nodata       & \nodata      & 93 $\pm$ 2  & \nodata    & 38 $\pm$ 1    & \nodata   & 23.5 $\pm$ 0.5& \nodata \\
415: 6.77 Jan 2005 & ATCA& 9.87 & \nodata       & \nodata      & 85 $\pm$ 3  & 67 $\pm$ 2 &  40 $\pm$ 1   & 32 $\pm$ 1& \nodata       & \nodata \\
416: 7.20 Jan 2005 & ATCA& 10.30& \nodata       & \nodata      & 88 $\pm$ 2  & 55 $\pm$ 1 & \nodata       &  \nodata  & \nodata       & \nodata \\ 
417: 7.25 Jan 2005 & GMRT& 10.35& 231 $\pm$ 20  & 125 $\pm$ 9  & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & \nodata\\
418: 7.90 Jan 2005 & VLA & 11.00& \nodata       & \nodata      & 71 $\pm$ 2  & \nodata    & 26 $\pm$ 1    & \nodata   & 16.5 $\pm$ 0.5& \nodata\\
419: 8.19 Jan 2005 & ATCA& 11.29& \nodata       & \nodata      & 67 $\pm$ 3  & 38 $\pm$ 2 & 24 $\pm$ 1    & 20 $\pm$ 1& \nodata       & \nodata\\
420: 8.24 Jan 2005 & GMRT& 11.34& 250 $\pm$ 17  & 104 $\pm$ 8  & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & \nodata\\ 
421: 9.06 Jan 2005 & ATCA& 12.16& \nodata       & \nodata      & 42 $\pm$ 2  & 32 $\pm$ 1.5& 21 $\pm$ 1   & \nodata   & 11.4 $\pm$ 1  & \nodata\\ 
422: 9.26 Jan 2005 & GMRT& 12.36& 176 $\pm$ 20  & 86 $\pm$ 7   & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & \nodata\\
423: 10.07 Jan 2005& ATCA& 13.16& \nodata       & \nodata      & 32 $\pm$ 2  & 24  $\pm$ 1& 17  $\pm$ 1   & \nodata   & 10 $\pm$ 1    & \nodata\\
424: 10.16 Jan 2005& GMRT& 13.26& 155 $\pm$ 17  & 82 $\pm$ 7   &\nodata      & \nodata    & \nodata       & \nodata   & \nodata       & \nodata\\
425: 10.60 Jan 2005& VLA & 13.70& \nodata       & \nodata      & \nodata     & \nodata    & \nodata       & \nodata   & 8.7 $\pm$ 0.4 & \nodata\\
426: 12.00 Jan 2005& NMA & 15.10& \nodata       & \nodata      & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & 7.16$^c$\\
427: 12.04 Jan 2005& ATCA& 15.14& \nodata       & \nodata      & 24 $\pm$ 1.5& 16 $\pm$ 1 & 12 $\pm$ 1    & 9.3 $\pm$ 1& 7.3 $\pm$ 1  & \nodata\\
428: 13.00 Jan 2005& NMA & 16.10& \nodata       & \nodata      & \nodata     & \nodata    & \nodata       & \nodata   & \nodata       & 5.50$^c$\\
429: 14.04 Jan 2005& ATCA& 17.14& \nodata       & \nodata      & 23 $\pm$ 1  & 15 $\pm$ 1 & 9.7 $\pm$ 1   & 7.3 $\pm$ 1& 5.5 $\pm$ 1  & \nodata\\
430: 16.25 Jan 2005& GMRT& 19.35& 96 $\pm$ 23   & 31 $\pm$ 5   &\nodata      & \nodata    & \nodata       & \nodata    & \nodata        & \nodata\\
431: 16.37 Jan 2005& GMRT& 19.47& \nodata       & \nodata      & 20 $\pm$ 2$^d$& \nodata  & \nodata       & \nodata    & \nodata        & \nodata\\
432: 18.01 Jan 2005& ATCA& 21.11& \nodata       & \nodata     & 24 $\pm$ 1.5 & 17 $\pm$ 1 & 6.2 $\pm$ 1   & 4.7 $\pm$ 1& 3.8$\pm$ 1 & \nodata\\
433: 20.10 Jan 2005& ATCA& 23.20& \nodata       & \nodata     & 19 $\pm$ 1.5 & 10 $\pm$ 1.5& 5  $\pm$ 1   & \nodata    & 3.2$\pm$ 1 & \nodata\\
434: 22.07 Jan 2005& ATCA& 25.17& \nodata       & \nodata     & 18 $\pm$ 1   & 11 $\pm$ 1 & 5 $\pm$ 1     & 4.3$\pm$ 1 & 2.0$\pm$ 1 & \nodata\\
435: 23.84 Jan 2005& ATCA& 26.94& \nodata       & \nodata     & 12 $\pm$ 1   & 7.9 $\pm$ 1& 5.6 $\pm$ 1   & 3.7$\pm$ 1 & 3.6$\pm$ 1 & \nodata\\
436: 24.85 Jan 2005& ATCA& 27.95& \nodata       & \nodata     & 12 $\pm$ 1   & 11 $\pm$ 1 & 4.2 $\pm$ 1   & 5.1$\pm$ 1 & 3.6$\pm$ 1 & \nodata\\
437: 26.26 Jan 2005& GMRT& 29.36& 104 $\pm$ 31  & 19 $\pm$ 6   &\nodata      & \nodata    & \nodata       & \nodata  & \nodata        & \nodata\\
438: 4.01  Feb 2005& GMRT& 38.14& \nodata       & \nodata      & 10$^{c,d}$  & \nodata    & \nodata       & \nodata  & \nodata        & \nodata\\
439: 16.87 Feb 2005& ATCA& 50.97& \nodata       & \nodata      & 10 $\pm$ 1  & 6.3 $\pm$ 1& 3.3 $\pm$ 0.4 & \nodata  & 2.1 $\pm$ 0.3  & \nodata\\
440: 24.01 Feb 2005& GMRT& 58.11& 28.2 $\pm$ 9  &6.6 $\pm$ 1.4 & \nodata     & \nodata    & \nodata       & \nodata  & \nodata     & \nodata\\
441: \hline
442: \end{tabular}
443: \end{center}
444: \caption[]{}
445: \label{tab:fluxes}
446: \end{table}
447: \end{landscape}
448: 
449: \begin{figure}
450: \centerline{\psfig{file=figure1.ps,width=6in,angle=0}}
451: \bigskip
452: \bigskip
453: \caption[]{}
454: \label{fig:LightCurves}
455: \end{figure}
456: 
457: 
458: 
459: \begin{table}[ht]
460: \begin{center}
461: \setlength{\extrarowheight}{-0.090in}
462: \begin{tabular}{>{\scriptsize}l >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c >{\scriptsize}c}
463: \hline
464: \hline
465: Epoch&Beam&Beam PA&Fit Major Axis&Axial Ratio&Fit PA\\
466: & (mas)&(degrees)& (mas)& & (degrees)\\
467: \hline
468: 03 Jan 2005 & $349 \times 170$  &  11.8   & 78.2$^{+3.0}_{-2.9}$ & 0.34$^{+0.21}_{-0.34}$ & 54.6$^{+6.7}_{-6.0}$\\
469: 04 Jan 2005 & $593 \times 173$  & $-40.3$  & 72.4$^{+14.5}_{-48}$ & 0.00$^{+0.90}_{-0}$       & 69$^{+20}_{-67}$\\
470: 05 Jan 2005 & $397 \times 168$  & $-25.7$  & 55$^{+18}_{-10}$     & 0.66$^{+0.34}_{-0.66}$   & 74$^{+90}_{-90}$\\
471: 06 Jan 2005 & $329 \times 178$  & $-16.5$  & 75.7$^{+3.0}_{-3.0}$ & 0.48$^{+0.18}_{-0.33}$ & 51.8$^{+9.0}_{-8.8}$\\
472: 07 Jan 2005 & $532 \times 178$  &  40.4  & 78$^{+26}_{-18}$     & 0.60$^{+0.4}_{-0.60}$       & 69$^{+90}_{-90}$\\
473: 10 Jan 2005 & $560 \times 161$  & $-38.8$  & 112$^{+30}_{-42}$    & 0.34$^{+0.33}_{-0.34}$ & 18$^{+28}_{-18}$\\
474: \hline
475: \end{tabular}
476: \end{center}
477: \caption[]{}
478: \label{tab:size}
479: \end{table}
480: 
481: 
482: \begin{figure}
483: \centerline{\psfig{file=figure2.ps,width=4.5in,angle=0}}
484: \bigskip
485: \bigskip
486: \caption[]{}
487: \label{fig:hone}
488: \end{figure}
489: 
490: \clearpage
491: \vspace{0.5cm}
492: \centerline{\bf SUPPLEMENTAL INFORMATION}
493: \vspace{0.5cm}
494: 
495: {\bf Observational Details}: {\small 
496: For the VLA observations we used the standard continuum mode with 2 $\times$ 50\,MHz bands, with
497: the exception of the 1.46\,GHz observation of 4 January 2005, which was taken in spectral line mode with 8 channels of 
498: width 3.1\,MHz. We used the extragalactic source 3C\,286 (J1331+305) for flux calibration, 
499: while the phase was monitored with J1820-254, J1751$-$253, and J1811$-$209. 
500: The listed flux densities and  uncertainties were measured from the resulting maps by fitting an 
501: elliptical Gaussian model to the afterglow emission.  
502: The GMRT observations were performed in dual frequency mode with 16\,MHz bandwidth divided into a
503: total of 128 frequency channels for 610\,MHz observations, 
504: and 6\,MHz of bandwidth divided into 64 channels for the 235\,MHz observations.
505: The observations at 1060 MHz at GMRT were carried out with a
506: bandwidth of 32 MHz.
507: 3C\,48 (J0137+331) and 3C\,286 were used as flux calibrators, and J1822$-$096 was used as the phase calibrator. 
508: These sources were also used for bandpass calibration.
509: We obtained the flux densities of the source by fitting a Gaussian with a background level plus a slope and
510: removed the contribution from a nearby weak source.
511: Because of the high density of radio sources in the galactic plane
512: in which \sgr\ is located, the antenna system noise temperature
513: has notable contributions from the sky within the telescope beam.
514: This reduces the signal to noise ratio and an appropriate correction must be
515: made to the observed flux (especially at low frequencies), since the
516: flux calibrators which establish the flux scale lie well outside the
517: galactic plane and are in an environment of less sky temperature.
518: We applied a $T_{sys}$ correction factor for 3C\,48 of 3.88 and 1.93, and for 3C\,286 of 3.87 and 1.8, for the
519: 235\,MHz and 610\,MHz, respectively.
520: Both the VLA and GMRT data were reduced and analyzed using the Astronomical Image Processing System.
521: The ATCA observations were performed in snapshot mode with 100MHz of effective bandwidth.
522: The amplitude calibrator was J1934$-$638, whereas J1711$-$251, J1817$-$254, and 
523: J1811$-$209 where used as phase calibrators.
524: The last of these was observed in a rapid (3 minute) cycle mode to compensate for its poor phase stability.
525: The flux densities were determined by performing a local parabolic fit to the peak closest to the known position of the source.
526: The NMA observations were performed at 102\,GHz in D-configuration (the most compact configuration) on 4 January 2005, and in
527: AB configuration (longest baseline configuration) on 12, 13 January 2005. 
528: We used NRAO530 for the phase calibration, and assumed it to have a flux density of 2.3\,Jy.\\
529: }
530: 
531: {\bf Details of Source Size Measurements}: {\small   The source sizes were measured by modeling the calibrated
532:    visibilities with the model-fitting procedure in DIFMAP.  This
533:    procedure employs the Levenberg-Marquardt non-linear least
534:    squares minimization technique while fitting a 6 parameter elliptical Gaussian to
535:    the visibilities.  
536:    The errors were determined with DIFWRAP using the following scheme:
537:    the source size parameters were stepped in small increments around
538:    their best-fitted value to form a grid of values. At each grid point
539:    the source size parameters were held fixed while the other model
540:    parameters were allowed to 'relax' with 4 model-fitting rounds.  
541:    The 95\% confidence limits were determined by those models that had a
542:    $\Delta\chi^2 < 12.8$ as measured from the best-fit total $\chi^2$
543:    (Press, W.~H., Teukolsky, S.~A., Vetterling, W.~T. and Flannery, B.~P.
544:    {\it Numerical Recipies in C. The art of scientific computing}.
545:    Cambridge: University Press, 2nd ed. 1992).  
546:    As a check we used phase only
547:    self-calibration as well as phase and amplitude self-calibration,
548:    both of which give consistent source size measurements.  We also
549:    used 30 second time-averaged data sets (to reduce the number of
550:    degrees of freedom), and found the best fit model parameters agreed
551:    to within the errors.\\
552: }
553: 
554: 
555: \begin{table}[ht]
556: \begin{center}
557: \setlength{\extrarowheight}{-0.090in}
558: \begin{tabular}{>{\scriptsize}l >{\scriptsize}c >{\scriptsize}c
559: >{\scriptsize}c
560: >{\scriptsize}c >{\scriptsize}c}
561: \hline
562: \hline
563: Frequency&$\alpha_A$&$t_1$&$\alpha_B$&$t_2$&$\alpha_C$\\
564: (GHz)& &(days)& &(days)& \\
565: \hline
566: 0.240& $-1.7 \pm 0.1$ & \nodata & \nodata & \nodata & \nodata\\
567: 0.61 & $-1.9 \pm 0.1$ &  \nodata & \nodata & \nodata & \nodata\\
568: 1.4 & $-2.0\pm 0.2$ & $10.7\pm 0.3$ & $-4.1\pm 0.3$ & $13.8\pm 0.2$ & $-0.85 \pm 0.2$\\
569: 2.4 & $-0.95\pm 0.3$ & $9.8\pm 0.2$ & $-3.5\pm 0.2$ & $13.8\pm 0.5$ & $-0.95\pm 0.2$\\
570: 4.9 & $-1.55\pm 0.15$ & $8.8\pm 0.2$ & $-3.1\pm 0.2$ & $18.6\pm 3.0$ & $-0.65\pm 0.3$\\
571: 6.1 & $-2.3 \pm 0.1$ & \nodata & \nodata &\nodata  &\nodata \\
572: 8.5 & $-2.00\pm 0.15$ & $8.1\pm 0.3$ & $-2.8\pm 0.24$ & $18.0\pm 3.0$ & $-0.64\pm0.4$\\
573: \hline
574: \end{tabular}
575: \end{center}
576: %\caption[]{}
577: \label{tab:breaks}
578: \end{table}
579: 
580: {\bf Table~\ref{tab:breaks}}: {\small Summary of temporal indices and breaks in the light curve of \sgr\ in 7 frequency bands.
581: The fits represent minimums in $\chi^2$ subject to the conditions that the two power-law slopes are continuous at 
582: the break point and disagree by more than 1-$\sigma$.
583: The first and second break points are denoted by $t_1$ and $t_2$, respectively.
584: The temporal decay indices ($S_\nu \propto t^{\alpha_i}$) 
585: are $\alpha_A$ for $t<t_1$, $\alpha_B$ for $t_1<t<t_2$, and $\alpha_C$ for $t>t_2$.
586: }
587: 
588: \end{document}
589: