1:
2: %\documentclass[12pt,preprint]{aastex}
3: \documentclass{emulateapj}
4:
5: %\usepackage{sidecap}
6:
7: \shorttitle{Variations in the spectral slope of Sgr~A* during a NIR flare}
8: \shortauthors{Gillessen et al.}
9:
10: \begin{document}
11:
12: \title{Variations in the spectral slope of Sgr~A* during a NIR flare}
13:
14: \author{S.~Gillessen$^1$, F.~Eisenhauer$^1$, E.~Quataert$^2$, R.~Genzel$^{1,\,3}$,
15: T.~Paumard$^1$, S.~Trippe$^1$,
16: T.~Ott$^1$, R.~Abuter$^1$,
17: A.~Eckart$^4$, P.~O.~Lagage$^5$, M.~D.~Lehnert$^1$, L.~J.~Tacconi$^1$, F.~Martins$^1$}
18: \affil{$^1\,$Max-Planck-Institut f\"ur extraterrestrische Physik, 85748 Garching, Germany}
19: \affil{$^2\,$Astronomy Department, University of California, Berkeley, CA 94720, USA}
20: \affil{$^3\,$Physics Department, University of California, Berkeley, CA 94720, USA}
21: \affil{$^4\,$1. Physikalisches Institut, Universit\"at zu K\"oln, 50937 K\"oln, Germany}
22: \affil{$^5\,$UMR 7158, CEA-CNRS-Universit\'e Paris 7, DSM/DAPNIA/Service d'Astrophysique, CEA/Saclay, France}
23:
24: \begin{abstract}
25: We have observed a bright flare of Sgr~A* in the near infrared
26: with the adaptive optics assisted integral field
27: spectrometer SINFONI\footnote{This work is based on observations
28: collected at the European Southern Observatory, Paranal, Chile. Program ID: 075.B-0547(B)}.
29: Within the uncertainties, the observed spectrum is featureless and can be described
30: by a power law. Our data suggest that the spectral index
31: is correlated with the instantaneous flux and that both
32: quantities
33: experience significant changes within less than one hour.
34: We argue that the near infrared flares from Sgr~A* are due to synchrotron emission of
35: transiently heated electrons, the emission being
36: affected by orbital dynamics and synchrotron cooling, both
37: acting on timescales of $\approx 20$ minutes.
38: \end{abstract}
39:
40: \keywords{blackhole physics --- Galaxy: center --- infrared: stars --- techniques: spectroscopic}
41:
42: \section{Introduction}
43:
44: The detection of stellar orbits \citep{sch02,eis05,ghe03,ghe05} close to Sgr~A* has proven that
45: the Galactic Center (GC) hosts a massive black hole (MBH) with a mass of $(3.6\pm 0.2) \times 10^6\,
46: \mathrm{M}_{\odot}$. Sgr~A* appears rather
47: dim in all wavelengths, which is
48: explained by accretion flow models
49: \citep{nar95,qua99b,mel00,mel01}.
50: In the near infrared (NIR) it was detected after diffraction limited
51: observations at the 8-m class telescopes had become possible \citep{gen03,ghe04}.
52: Usually the emission is not detectable.
53: However, every few hours Sgr~A* flares in the NIR, reaching up to
54: $K\approx 15\,$mag. A first flare spectrum was obtained by \cite{eis05},
55: showing a featureless, red spectrum ($\nu S_{\nu}\sim\nu^\beta$ with
56: $\beta \approx-3$).
57:
58: \section{Observations and data reduction}
59: \label{datared}
60: We observed the GC on 2005 June 18 from
61: 2:40 to 7:15 UT with SINFONI \citep{eis03, bon04}, an adaptive optics (AO) assisted integral
62: field spectrometer which is mounted at the Cassegrain focus
63: of ESO-VLT Yepun (UT4).
64: The field of view was 0.8''$\times$0.8'' for individual exposures, mapped
65: onto 64$\times$32 spatial pixels. We observed in K-band with
66: a spectral resolution of FWHM
67: $0.5\,$nm. The first
68: 12 integrations lasted $5\,$min each. During those we noticed
69: NIR activity of Sgr~A* and
70: we switched to 4 minute exposures. We followed Sgr~A* for another 32 exposures.
71: In total we interleaved nine integrations on
72: a specifically chosen off field (712'' W, 406'' N of Sgr~A*).
73: The seeing was $\approx\,$0.5'' and the optical coherence time
74: $\approx 3\,$ms, some short-time deteriorations excluded. The AO
75: was locked on the closest optical guide star ($m_R=14.65$,
76: 10.8'' E, 18.8'' N of Sgr~A*),
77: yielding a spatial resolution of $\approx 80\,$mas FWHM, close to the
78: diffraction limit of UT4 in K-band ($\approx 60\,$mas).
79:
80: Our detection triggered immediate follow-up observations with VISIR,
81: a mid-infrared (MIR) instrument
82: mounted at ESO-VLT Melipal (UT3). From 5:25 UT onwards
83: VISIR was pointing to the
84: GC. At the position of Sgr~A* no significant flux was seen. A conservative upper limit
85: of 40$\,$mJy (not dereddened) at 8.59$\,\mu$m is reported (Lagage et al., in prep).
86:
87:
88: \begin{figure}
89: \epsscale{1}
90: \plotone{f1.eps}
91: \caption{The Galactic Center as seen
92: with SINFONI. Channel maps from
93: the bright and the dim sample obtained
94: by taking the mean in spectral dimension
95: from $2.05\,\mu$m to
96: $2.15\,\mu$m and from $2.25\,\mu$m to
97: $2.35\,\mu$m. All images are scaled the same way
98: and use an identical color map.
99: Sgr~A* is brighter in the
100: longer wavelength maps, indicating that it is
101: is redder than the field stars. Furthermore,
102: the effect is more pronounced in the dim sample,
103: meaning that Sgr~A* is redder therein
104: than in the bright sample.}
105: \label{adb}
106: \end{figure}
107:
108: The reduction of the SINFONI data followed the standard
109: steps:
110: From all source data we subtracted the respective sky frames to correct for instrumental
111: and atmospheric background. We applied
112: flatfielding, bad pixel correction, a search for cosmic ray hits,
113: and a correction for the optical distortions of SINFONI.
114: We calibrated the wavelength dimension with line emission lamps and tuned
115: on the atmospheric OH-lines of the raw frames.
116: Finally we assembled the data into cubes
117: with a spatial grid of 12.5$\,$mas/pix.
118:
119: \section{Analysis}
120:
121: \begin{figure}
122: \epsscale{.80}
123: \plotone{f2.eps}
124: \caption{Lightcurve and variation of the spectral power law index
125: $\beta$ during the flare.
126: Time is counted from 2:40 UT.
127: Top: Flux ratio flare/S2.
128: Thin dots are exposures
129: affected by bad seeing.
130: Middle/top: $\beta$ using the small apertures background.
131: Middle/bottom: $\beta$ using the off state subtraction background.
132: Bottom: $\beta$ using the constant subtraction background.}
133: \label{lightcurve}
134: \end{figure}
135:
136: \subsection{Flux determination}
137:
138: For all 44 cubes we extracted a
139: collapsed image (median in spectral dimension) of
140: a rectangular region (0.25''$\times$0.5'') centered on Sgr~A*
141: and containing the three S-stars S2, S13, and S17.
142: We determined the flux of Sgr~A*
143: from a fit with five Gaussians to each of these images. Four
144: Gaussians with a common width describe the four sources.
145: The fifth Gaussian (with a width $3.5\times$ wider, typical for the halo
146: from the imperfect AO correction) accounts
147: for the halo of the brightest star S2 ($K \approx 14\,$mag). The halos of the weaker
148: sources (all $K < 15\,$mag) could be neglected for the flux measurement.
149: We fixed the positions of all sources
150: (known a-priori from a combined cube) and
151: the amplitude ratios for the stars. Five parameters were left free:
152: An overall amplitude, the background,
153: the width, the flux ratio halo/S2, and $\cal{F}$, the
154: flux ratio Sgr~A*/S2. This procedure disentangles real variability from
155: variations in the background, the Strehl ratio, and the seeing.
156:
157: As a crosscheck we determined $\cal{F}$ in a second way for all images;
158: for both Sgr~A* and S2 we measured the flux difference
159: between a signal region centered on source and a reasonable,
160: symmetric background region.
161: The such determined values agreed very well
162: with the fits. For further analysis we
163: used the fitted ratios and included the difference
164: between the two estimates in the errors.
165:
166: \subsection{Choice of background}
167:
168: The value of the spectral power law index $\beta$
169: crucially depends on the background subtraction. Subtracting too much light
170: would artificially make the signal look redder than it is. Hence, a
171: reasonable choice for the background region excludes the nearby sources S2 and
172: S17. Furthermore, the background flux is
173: varying spatially. Actually Sgr~A* lies close to
174: a saddle point in the background light distribution, caused by S2 and S17 (see Fig.~\ref{adb}).
175: In the East-West direction
176: the background has a maximum close to Sgr~A* and in the North-South direction
177: a minimum.
178: A proper estimate of the background can be achieved in two ways:
179: a) working with small enough, symmetric apertures, and b)
180: subtracting from the signal an off state obtained at the position of Sgr~A* from cubes in
181: which no emission is seen. We used the first method as well as two variants
182: of the second.\\
183: {\bf Small apertures:}
184: The local background at a position $\vec{x}$ can be estimated by averaging over a small,
185: symmetric region centered on $\vec{x}$.
186: Given the background geometry we have chosen a ring with inner radius $3\,$pix
187: and outer radius $7\,$pix. The circular symmetry was only broken since we explicitely excluded those pixels
188: with a distance to S17 and S2 smaller than 3 and 7 pixels respectively.
189: Unfavorable of this method is that the local background is only approximated, since a
190: sufficiently large region has to be declared as signal region.\\
191: {\bf Off state subtraction:}
192: The local background can be extracted from cubes in
193: which no signal is seen. Since the seeing conditions change from
194: cube to cube, one still has to correct for the varying amount of stray light in the
195: signal region. We estimate this variation by measuring the difference
196: spectrum between signal and off cube in a stray light region.
197: The latter must not contain any field stars and should be
198: as far away from the nearby sources as Sgr~A*.
199: We used two stray light regions to the left and to the
200: right of Sgr~A*, between 5 and 10 pixels away.
201: The disadvantage of this method is that one needs a suitable off state.
202: The latter point is critical for our data.
203: Even though Sgr~A* has not been detected directly in
204: the first three cubes, the light at its position appears
205: redder than the local
206: background\footnote{Inspecting older (non-flare) SINFONI data
207: we found cases similar to the new data and other cases in which the light
208: was identical to the local background emission. This is consistent with
209: the L-band observations by \cite{ghe04,ghe05b}.}.
210: Assuming that this
211: light is due to a very dim, red state of Sgr~A*, we would subtract
212: too much red light and artificially make the flare too blue.
213: In this sense, this method yields an upper limit for $\beta$.\\
214: {\bf Constant subtraction:}
215: The off state method can be varied to obtain a lower limit for $\beta$.
216: Assuming that the true off state spectrum has the color of S2, one can
217: demand that it is flat after division by S2. In our data, the
218: S2 divided off state
219: spectrum is rising towards longer wavelengths. Hence, in this third
220: method we estimate the background
221: at blue wavelengths and use that constant as background for
222: all spectral bins.
223:
224: \begin{figure*}
225: \epsscale{1}
226: \begin{center}
227: \includegraphics[width=5.5cm]{f3a.eps}
228: \includegraphics[width=5.5cm]{f3b.eps}
229: \includegraphics[width=5.5cm]{f3c.eps}
230: \end{center}
231: \caption{Correlation plots between the flare flux and the spectral index $\beta$.
232: Points with error bars represent the flare, blue dots are $\beta_{S17}$.
233: Thick black dots mark the points with an error $\Delta \beta < 1$, red triangles
234: mark the data with good seeing (FWHM $<75\,$ mas). The line is a fit to the thick black dots.
235: Open circles denote the data from \cite{eis05} - near 2$\,$mJy - and \cite{ghe05b}
236: - near 7$\,$mJy.
237: Left: Small apertures method; middle: off state subtraction method; right: constant subtraction method.
238: }
239: \label{corr}
240: \end{figure*}
241:
242: \subsection{Determination of spectral power law index}
243:
244: We obtained spectra as the median of all pixels inside a
245: disk with radius $3\,$pix centered on source minus the median spectrum of the pixels
246: in the selected off region.
247: In order to correct for the interstellar extinction
248: we then divided by the S2-spectrum (obtained in the same way as the signal).
249: Next the temperature of S2 is corrected by multiplying
250: by the value of $\nu S_{\nu}$
251: of a blackbody with $T\approx 25000\,$K. After binning the data
252: into 60 spectral channels it is finally fit with a power law $\sim \nu^\beta$.
253: With this definition, red emission has $\beta<0$. The error on $\beta$
254: is obtained as the
255: square sum of the formal fit error and the standard deviation in a sample of
256: 20 estimates for $\beta$ obtained by varying the on and off region selection.
257:
258: \section{Results}
259: We observed a strong
260: (flux density up to $8\,$mJy
261: or $\nu L_\nu \approx 10^{35}\,$ergs/s), long
262: (more than 3 hours) flare which
263: showed significant brightness variations
264: on timescales as short as 10 minutes (Fig.~\ref{lightcurve}, top).
265: While the data is
266: not optimal for a periodicity search (poorer sampling
267: than our previous imaging data), it is worth noting that the
268: highest peak in the periodogram (significance of $\approx 2\,\sigma$)
269: lies at a period of $\approx 18\,$min.
270: This is also the timescale found by \cite{gen03}, who identify the
271: quasi-periodicity with the orbital time close to the
272: last stable orbit (LSO) of the MBH.
273:
274: We divided the data into three groups: a) the cubes of the
275: first peak near $t=50\,$min ("preflare"), b) the cubes at $t>100\,$min
276: with $\cal{F}\,$$<$ 0.25 ("dim state"), and c) the cubes at $t>100\,$min
277: with $\cal{F}\,$$>$ 0.25 ("bright state"). For the three sets
278: we created
279: combined cubes in which we determined
280: $\beta$, using
281: all three background estimates.
282: In all cases we obtained the correct
283: spectral index $2.9 \pm 0.5$ for S17 (a star with a spectrum similar
284: to S2 but a flux comparable to Sgr~A*). For Sgr~A* we get:\\
285:
286: \begin{tabular}{r|ccc}
287: $\beta\,\,$&preflare&dim state&bright state\\
288: \hline
289: off state subtr.& $-1.4\pm 0.4$ & $-0.7\pm 0.4$ & $+0.4\pm 0.2$\\
290: small apertures& $-1.8\pm 0.3$ & $-1.7\pm 0.4$ & $-0.1\pm 0.3$ \\
291: constant subtr.& $-3.4\pm 0.4$ & $-2.3\pm 0.3$ & $-0.3\pm 0.2$
292: \end{tabular} \\
293:
294: The values obtained from the small apertures lie
295: between the values from the other two methods, consistent
296: with the idea that they yield upper and lower limits.
297: The absolute values vary systematically according to the chosen background
298: method. However, independent from that, it is clear that the preflare is
299: redder than the dim state which in turn is redder than the bright state.
300: An obvious question then is whether
301: flux and $\beta$ are directly correlated.
302:
303: Hence, we applied the spectral analysis to the individual cubes.
304: We kept the data in which Sgr~A* is detected,
305: the error $\Delta \beta < 1.5$ and the spectral
306: index for S17 does not deviate more than $1.5$ from
307: the expected value. The resulting
308: spectral indices appear to be correlated with the flux
309: (Fig.~\ref{lightcurve}, \ref{corr}).
310: The values match the results in \cite{eis05}
311: and \cite{ghe05b}.
312: Bright flares are indeed bluer than weak flares, as suspected
313: by \cite{ghe05b} and consistent with the earlier
314: multi-band observations of \cite{gen03}. Our key new result is that
315: this even holds
316: within a single event.
317:
318: For all three background methods
319: it is clear that the main event was preceded by a weak, red event.
320: For the small apertures and the
321: constant subtraction method instantaneous flux and
322: spectral index are correlated. For the off state subtraction method one could instead group the
323: data into a preflare at the beginning and a bluer, brighter
324: main event.
325:
326: We checked our data for contamination effects. If stray light
327: would affect the flare signal, $\beta$
328: should be correlated with the seeing (measured by the width
329: from the multiple fits). Since we did not find such a correlation,
330: we exclude significant contamination.
331:
332: \section{Interpretation}
333:
334: Theoretical models predict that the mm-IR emission from Sgr~A* is
335: synchrotron emission from relativistic electrons close to the LSO
336: \citep{liu01,qua03,liu04,liu06}.
337: Radiatively inefficient accretion
338: flow (RIAF) models with a thermal electron population ($T_\mathrm{e}\approx 10^{11}\,$K)
339: produce the observed peak in the submm but fail to produce
340: enough flux at 2$\,\mu$m. The
341: NIR emission requires transiently heated or
342: accelerated electrons as proposed by \cite{mar01}.
343:
344: A conservative interpretation of our data is suggested by
345: Fig.~\ref{corr}, middle. There was a weak, red event before
346: and possibly independent from the
347: main flare which was a much bluer event.
348: Plausibly the preflare
349: is then due
350: to the high-energetic tail of the submm peak \citep{gen03}. The main flare
351: requires nevertheless a population of heated electrons.
352:
353: A more progressive interpretation follows from
354: the correlation between flux and $\beta$ (Fig.~\ref{corr}, left \&
355: right). It suggests that
356: the NIR variability is caused by the combination of
357: transient heating with subsequent cooling and orbital dynamics of relativistic
358: electrons. In the following subsections we will exploit this idea.
359:
360: \subsection{Synchrotron emission}
361:
362: In the absence of continued energy injection, synchrotron cooling will suppress the
363: high energy part of the electron distribution function. This results in a
364: strong cutoff in
365: the NIR spectrum with a cutoff frequency decreasing with time. At a fixed
366: band one expects that the light becomes redder as the flux decreases.
367: This can cause the correlation between
368: flux and $\beta$. The observed flare
369: apparently needs several heating and cooling events.
370:
371: The synchrotron cooling time is comparable to the observed timescale
372: of decaying flanks as in Fig.~\ref{lightcurve} (top) or \cite{gen03}.
373: In a RIAF model \citep{yua03} the magnetic field $B$ is related to the
374: accretion rate $\dot{M}$. For disk models with
375: $\dot{M} \approx 10^{-8}\,M_{\odot}/$yr
376: \citep{ago00,qua00} one has $B\approx 30\,$G at a
377: radial distance of $3.5\,R_\mathrm{S}$ (Schwarzschild radii).
378: The cooling time for electrons emitting at $\lambda = 2 \,
379: \lambda_2 \,\mu$m for $B = 30 \, B_{30} \,$G is
380: \begin{equation}
381: t_{\rm cool} \approx 8 \, B_{30}^{-3/2} \, \lambda_2^{1/2} \, {\rm min} \,\,.
382: \label{cool}
383: \end{equation}
384: This is similar to the orbital timescale, making
385: it difficult to disentangle flux
386: variations due to heating and cooling from
387: dynamical effects due to the orbital motion.
388:
389: \subsection{Orbital dynamics}
390: \label{dyn}
391:
392: The timescale of $\approx 20$ minutes for the observed variations
393: suggests orbital motion close to the LSO as one possible
394: cause. Any radiation produced propagates through strongly curved space-time.
395: Beaming, multiple images, and Doppler shifts
396: have to be considered \citep{hol97,hol99}.
397: Recent progress has been made in simulating these effects
398: when observing a spatially limited emission region orbiting the MBH
399: \citep{pau05, bro05a, bro05b, bro05c}.
400:
401: That the emission region is small compared to the accretion disk
402: can be deduced from X-ray observations.
403: \cite{eck04,eck05} report similar timescales for X-ray- and IR-flares.
404: This excludes that the X-ray emission is
405: synchrotron light as one
406: would expect cooling times $<1\,$min
407: (eq.~\ref{cool}). The X-ray emission is naturally explained by IC scattering of
408: the transiently heated electrons that have a
409: $\gamma$-factor of $\gamma\approx 10^3$.
410: The IC process scatters the seed photon field up according to
411: $\nu_\mathrm{IC}\approx \gamma^2 \,\nu_\mathrm{seed}$. Given
412: $\nu_\mathrm{IC} \approx 10^{18}\,$Hz for X-rays
413: the seed photon frequency is $\nu_\mathrm{seed}\approx 10^{12}\,$Hz - the
414: submm regime. The X-ray intensity for the synchrotron-self-Compton case is
415: given by
416: \begin{equation}
417: X/IR = L_\mathrm{SSC}/L_\mathrm{sync} \approx
418: 10^{-2} L_{35} \left(R/R_\mathrm{S}\right)^{-2} B_{30}^{-2}\,\,,
419: \label{l37}
420: \end{equation}
421: where $L_{35}$ is the seed photon field luminosity in units of $10^{35}\,$ergs/s
422: and $R$ the radius of the emission region.
423: The highest value of $L_{35}$ compatible with the MIR- and
424: submm-data is reached for a flat spectrum ($\beta=0$) of the
425: heated electrons. Thus,
426: $L_{35} \lesssim 1$. Via equation~(\ref{l37}) we have
427: $R/R_\mathrm{S} \approx 0.1 \sqrt{L_{35}/(X/IR)}$.
428: From the observed ratios $X/IR \approx [0.1\,-\,1]$ we derive
429: $R < 0.3\,R_\mathrm{S}$. That means, the emission region
430: has to be a small spot.
431:
432: A plausible dynamical model
433: is discussed in Paumard et al. (in prep).
434: It reproduces typical lightcurves
435: as observed by \cite{gen03} or in Fig.~\ref{lightcurve}.
436: Due to the Doppler effect
437: the observed light corresponds to different rest frame
438: frequencies depending on the orbital phase. If the source spectrum
439: is curved,
440: flux and spectral index appear correlated. Following this
441: interpretation,
442: the emission during the brightest part
443: originates from a rest-frequency with larger $\beta$
444: than the dimmer state emitted at shorter
445: wavelengths. Such a concavely curved spectrum is
446: naturally expected from the synchrotron models.
447:
448: \begin{thebibliography}{}
449: \bibitem[Agol(2000)]{ago00} Agol, E. 2000, \apj, 538, L121
450: \bibitem[Bonnet et al.(2004)]{bon04} Bonnet, H. et al. 2004, ESO Messenger, 117, 17
451: \bibitem[Broderick \& Loeb(2005a)]{bro05a} Broderick, A. \& Loeb, A. 2005, MNRAS, 363, 353
452: \bibitem[Broderick \& Loeb(2005b)]{bro05b} Broderick, A. \& Loeb, A., \apj, 636, L109
453: \bibitem[Broderick \& Loeb(2005c)]{bro05c} Broderick, A. \& Loeb, A., MNRAS, submitted (astro-ph/0509237)
454: \bibitem[Eckart et al.(2004)]{eck04} Eckart, A. et al. 2004, A\&A, 427, 1
455: \bibitem[Eckart et al.(2005)]{eck05} Eckart, A. et al., A\&A, in press (astro-ph/0512440)
456: \bibitem[Eisenhauer et al.(2003)]{eis03} Eisenhauer, F. et al. 2003, Proc. SPIE, 4841, 1548
457: \bibitem[Eisenhauer et al.(2005)]{eis05} Eisenhauer, F. et al. 2005, \apj, 628, 246
458: \bibitem[Genzel et al.(2003)]{gen03} Genzel, R. et al. 2003, Nature, 425, 934
459: \bibitem[Ghez et al.(2003)]{ghe03} Ghez, A. et al. 2003, \apj, 586, L127
460: \bibitem[Ghez et al.(2004)]{ghe04} Ghez, A. et al. 2004, \apj, 601, L159
461: \bibitem[Ghez et al.(2005a)]{ghe05} Ghez, A., Salim, S., Hornstein, S., Tanner, A.,
462: Lu, J., Morris, M. \& Becklin, E. 2005, \apj, 620, 744
463: \bibitem[Ghez et al.(2005b)]{ghe05b} Ghez, A. et al. 2005, \apj, 635, in press
464: \bibitem[Hollywood et al.(1997)]{hol97} Hollywood, J. \& Melia, F. 1997, ApJS, 112, 423
465: \bibitem[Hollywood et al.(1999)]{hol99} Hollywood, J., Melia, F.,
466: Close, L., McCarthy, D., Dekeyser, T. 1999, \apj, 448,
467: \bibitem[Liu \& Melia(2001)]{liu01} Liu, S. \& Melia, F. 2001, \apj, 561, L77
468: \bibitem[Liu et al.(2004)]{liu04} Liu, S., Petrosian, V. \& Melia, F. 2004, \apj, 611, L101
469: \bibitem[Liu et al.(2006)]{liu06} Liu, S., Melia, F. \& Petrosian, V. 2006, \apj, 636, L798
470: \bibitem[Markoff et al.(2001)]{mar01} Markoff, S., Falcke, H., Yuan, F. \& Biermann, P. 2001, A\&A, 379, L13
471: \bibitem[Melia et al.(2000)]{mel00} Melia, F., Liu, S. \& Coker, R. 2000, ApJ, 545, L117
472: \bibitem[Melia et al.(2001)]{mel01} Melia, F., Liu, S. \& Coker, R. 2001, ApJ, 553, 146
473: \bibitem[Narayan et al.(1995)]{nar95} Narayan, R., Yi, I.,
474: \& Mahadevan, R.\ 1995, \nat, 374, 623
475: \bibitem[Paumard et al.(2005)]{pau05} Paumard, T. et al. 2005, AN, 326, 568
476: \bibitem[Quataert \& Narayan(1999)]{qua99b} Quataert, E. \& Narayan, R. 1999, \apj, 520, 298
477: \bibitem[Quataert \& Gruzinov(2000)]{qua00} Quataert, E. \& Gruzinov, A. 2000, \apj, 539, 809
478: \bibitem[Quataert(2003)]{qua03} Quataert, E. 2003, ANS, 324, 435
479: \bibitem[Sch\"odel et al.(2002)]{sch02} Sch\"odel, R. et al. 2002, Nature, 419, 694
480: \bibitem[Yuan et al.(2003)]{yua03} Yuan, F., Quataert, E. \& Narayan, R. 2003 \apj, 598, 301
481: \end{thebibliography}
482:
483: \end{document}
484:
485: