1: % EDITOR: COMMENT OUT THESE Three LINES TO PRINT OUT THE MANUSCRIPT

2: \documentclass{article}

3: \usepackage{emulateapj}

4: %\usepackage{times}

5: % EDITOR: UNCOMMENT  THIS LINE TO PRINT OUT THE MANUSCRIPT

6: %\documentclass{aastex}

7: 

8: %\usepackage{aaspp4}

9: \usepackage{psfig}

10: %\setlength{\topmargin}{1.5cm}

11: 

12: \def\lc{l_{\rm c}}

13: \def\lbb{l_{\rm bb}}

14: \def\tautr{\tau_{\rm tr}}

15: \def\tauc{\tau_{\rm c}}

16: \def\taue{\tau_{\rm e}}

17: \def\gta{\ga}

18: \def\lta{\la}

19: \def\Tbb{T_{\rm bb}}

20: \def\ledd{L_{\rm Edd}}

21: \def\taut{\tau_{\rm T}}

22: \def\tauff{\tau_{\rm ff}}

23: \def\msun{{\,M_\odot}}

24: \def\lsun{{\,L_\odot}}

25: \def\sw{Schwarzschild~}

26: \newcommand\fe{Fe K$\alpha$~}

27: \newcommand\rozanska{R\'o$\dot{\rm z}$a\'nska }

28: \newcommand\zycki{$\dot{\rm Z}$ycki }

29: \newcommand\tl{t_{\rm l}}

30: \newcommand\tff{\tau_{\rm ff}}

31: \newcommand\fx{F_{\rm x}}

32: \newcommand\tx{\tau_{\rm x}}

33: \newcommand\thydro{t_{\rm h}}

34: \newcommand\dm{\dot{m}}

35: \newcommand\fdisk{F_{\rm d}}

36: \newcommand\tcomp{T_{\rm comp}}

37: \newcommand\pram{P_{\rm ram}}

38: \newcommand\prad{P_{\rm rad}}

39: \newcommand\ptot{P_{\rm tot}}

40: \newcommand\pg{P_{\rm gas}}

41: \newcommand\lnet{\Lambda_{\rm net}}

42: \newcommand\taueff{\tau_{\rm eff}}

43: \newcommand\teff{T_{\rm eff}}

44: \newcommand\pmin{P_{\rm min}}

45: \newcommand\pstar{P_*}

46: \newcommand\tstar{T_*}

47: \newcommand\thot{\tau_{\rm h}}

48: \newcommand\tskin{\tau_{\rm s}}

49: \newcommand\fbb{F_{\rm BB}}

50: \newcommand\ibb{I_{\rm BB}}

51: \def\mean#1{\langle #1 \rangle}

52: %

53: % Reference macros

54: %

55: % To generate reference to a paper in Ap.J. volume 300, p.123 

56: % write \apj{Claus, S. 1990}{300}{123}

57: %

58: %\def\refindent{\par\noindent\hangindent=3pc\hangafter=1 }

59: %\def\aa#1#2#3{\refindent#1, A\&A, #2, #3}

60: %\def\aasup#1#2#3{\refindent#1, A\&AS, #2, #3}

61: %\def\aj#1#2#3{\refindent#1, AJ, #2, #3}

62: %\def\apj#1#2#3{\refindent#1, {\it ApJ}, {#2}, #3.}

63: %\def\apjlett#1#2#3{\refindent#1, {\it ApJL}, {#2}, #3.}

64: %\def\apjsup#1#2#3{\refindent#1, ApJS, #2, #3}

65: %\def\araa#1#2#3{\refindent#1, ARA\&A, #2, #3}

66: %\def\baas#1#2#3{\refindent#1, BAAS, #2, #3}

67: %\def\icarus#1#2#3{\refindent#1, Icarus, #2, #3}

68: %\def\mnras#1#2#3{\refindent#1, {\it MNRAS}, {#2}, #3.}

69: %\def\nature#1#2#3{\refindent#1, {\it Nature}, {#2}, #3.}

70: %\def\pasj#1#2#3{\refindent#1, PASJ, #2, #3}

71: %\def\pasp#1#2#3{\refindent#1, PASP, #2, #3}

72: %\def\qjras#1#2#3{\refindent#1, QJRAS, #2, #3}

73: %\def\science#1#2#3{\refindent#1, Science, #2, #3}

74: %\def\sov#1#2#3{\refindent#1, Soviet Astr., #2, #3}

75: %\def\sovlett#1#2#3{\refindent#1, Soviet Astr.\ Lett., #2, #3}

76: %\def\refpaper#1#2#3#4{\refindent#1, #2, #3, #4}

77: %\def\refbook#1{\refindent#1}

78: \def\degs{$^\circ$}

79: \def\biggldb{\biggl[\!\!\biggl[}

80: \def\biggrdb{\biggr]\!\!\biggr]}

81: \def\um{{\,\mu\rm m}}

82: \def\cm{{\rm\,cm}}

83: \def\km{{\rm\,km}}

84: \def\au{{\rm\,AU}}

85: \def\pc{{\rm\,pc}}

86: \def\kpc{{\rm\,kpc}}

87: \def\mpc{{\rm\,Mpc}}

88: \def\sec{{\rm\,s}}

89: \def\yr{{\rm\,yr}}

90: \def\gm{{\rm\,g}}

91: \def\kms{{\rm\,km\,s^{-1}}}

92: \def\kelvin{{\rm\,K}}

93: \def\erg{{\rm\,erg}}

94: \def\ev{{\rm\,eV}}

95: \def\hz{{\rm\,Hz}}

96: \def\>{$>$}

97: \def\<{$<$}

98: \def\bsl{$\backslash$}

99: \def\simlt{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}

100: \def\simgt{\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}}

101: \def\sqr#1#2{{\vcenter{\hrule height.#2pt

102:       \hbox{\vrule width.#2pt height#1pt \kern#1pt

103:          \vrule width.#2pt}

104:       \hrule height.#2pt}}}

105: \def\square{\mathchoice\sqr34\sqr34\sqr{2.1}3\sqr{1.5}3}

106: %def\ledd{1}

107: 

108: 

109: \begin{document}

110: 

111: \title{Narrow Moving Fe K$\alpha$ lines from magnetic flares in AGN.}

112: 

113: \author{Sergei Nayakshin\altaffilmark{1} \& Demosthenes Kazanas}

114: 

115: \affil{NASA/GSFC, LHEA, Code 661, Greenbelt, MD, 20771}

116: \altaffiltext{1}{Also Universities Space Research Association}

117: 

118: \begin{abstract}

119: We point out that luminous magnetic flares, thought to occur in

120: standard AGN accretion disks, cannot be located much higher than few

121: pressure scale heights above the disk. Using this fact, we estimate

122: the fraction of the disk surface illuminated by a typical flare. This

123: fraction turns out to be very small for geometrically thin disks,

124: which implies that the instantaneous Fe K$\alpha$ emission line from a

125: specific magnetic flare is {\em narrow}. The line is red- or

126: blue-shifted depending on the position of the observer relative to the

127: flare and sweeps across the line band with time. We present several

128: examples of theoretical time-resolved line profiles from such flares

129: for \sw geometry.  The observations of such moving features with

130: future X-ray telescopes will present a powerful test of the accretion

131: disk geometry and may also test General Relativity in the strong field

132: limit.

133: \end{abstract}

134: 

135: \keywords{accretion, accretion disks ---radiative transfer ---

136: line: formation --- X-rays: general}

137: %--- radiation mechanisms: non-thermal}

138: 

139: 

140: \section{Introduction}\label{sect:intro}

141: 

142: There is much current debate about the structure of the inner part of

143: the accretion flow in Active Galactic Nuclei (AGN). One of the

144: possibilities often discussed is the standard (geometrically thin,

145: optically thick) accretion disk in which the X-ray activity is

146: provided by magnetic flares above the disk (e.g., Galeev, Rosner \&

147: Vaiana 1979; Haardt, Maraschi \& Ghisellini 1994). The main difficulty

148: of this model is its intrinsic physical complexity, which does not

149: naturally relate the X-ray emission to the {\em global} disk

150: parameters, such as the accretion rate in the disk, black hole mass

151: $M$, etc. However the model is attractive because of the physical

152: parallel with the Solar magnetic flares, and the fact that it

153: naturally explains the broad \fe lines observed in a number of AGN --

154: see the recent review by Fabian et al. (2000). Any robust

155: observational prediction that can be used to test the model is thus of

156: a substantial value.

157: 

158: Herein we show that time-resolved \fe line profiles of the magnetic

159: flare model should consist of relatively narrow features due to the

160: limited extent (in radius and height) of the active regions.  As they

161: circle the black hole, the corresponding \fe line features should

162: sweep across $\sim 4-8$ keV energy band. This pattern is unique to the

163: magnetic flare model (in particular, the line width is much narrower

164: than the profiles of the lamppost-like models calculated by Reynolds

165: et al. [1999], Young \& Reynolds [2000] and Ruszkowski [2000]), and it

166: could be used in the future to verify or falsify the basic premises of

167: this model.

168: 

169: 

170: \section{The flare height and the active region size}

171: 

172: The size of a single magnetic reconnecting region, $\Delta R$, and its

173: height above an accretion disk are of theoretical and observational

174: import, yet there is no clear discussion of this issue in the

175: literature. If one assumes that magnetic field pressure in the flux

176: tube greatly exceeds the gas pressure there, then its equilibrium

177: structure should be dictated by a force-free equilibrium.  Parker

178: (1979, \S 8.4) computes the force-free equilibrium for a flux tube not

179: bounded by any external pressure. He shows that, mathematically

180: speaking, the magnetic field fills all the available (empty)

181: space. However, one can check that most of the magnetic {\em energy}

182: is confined within a volume of roughly same linear size as the

183: separation between the footpoints of the tube, i.e., in close

184: proximity to the footpoints.  Further, it is generally believed that

185: the size of a magnetic flux tube inside the disk should be no larger

186: than the size of the largest turbulent cell, which is $\sim H$, the

187: disk pressure height scale (e.g., Galeev et al.  1979). These

188: considerations thus require that both the size and the height of a

189: magnetic flare be $\Delta R \sim H$.

190: 

191: At the same time, one could argue that the magnetic field need not

192: completely dominate the energy (and pressure) content of the flux tube

193: above the disk, and hence maybe the gas pressure expands the flux tube

194: much above $H$. While this is in principle possible dynamically (e.g.,

195: Romanova et al. 1998), one can show that the corona has to be

196: magnetically dominated or the energy balance conditions will require

197: unrealistically large size for the emitting region (e.g., Merloni \&

198: Fabian 2001). These authors derive also a limit on the size of a

199: magnetic flare (see their equation 8), which however depends on the

200: average number of magnetic flares, $N$; this is not a well determined

201: quantity since the usual variability arguments (e.g., Haardt et

202: al. 1994) may not be applied if a single flare in the {\em light

203: curve} is produced by a multitude of reconnecting flux tubes (e.g. as

204: in the flare avalanche model of Poutanen \& Fabian 1999; also see

205: below).

206: 

207: We now show that one can derive an even tighter limit on the flux tube

208: size. The magnetic flux, $\Psi$, is a constant along a flux tube

209: (e.g., Parker 1979). Since any flux tube is anchored in the disk

210: mid-plane, the maximum of the magnetic flux is $\Psi \sim B_e H^2$

211: where $B_e$ is the equipartition magnetic field in the disk mid-plane,

212: and $H^2$ is the order of magnitude estimate of the maximum tube cross

213: section within the disk. The magnetic pressure in the loop above the

214: disk is thus

215: \begin{equation}

216: {B^2\over 8\pi}  \sim \beta \lambda^4 P_d \left({H\over \Delta

217: R}\right)^4\;,

218: \label{mp}

219: \end{equation}

220: where we introduced parameter $\beta \equiv B_{\rm max}^2/B_e^2\leq

221: 1$; $B_{\rm max}$ and $\lambda^2 H^2$ are the actual magnetic field

222: and the tube's cross section in the mid-plane ($\lambda < 1$), and

223: $P_d$ is the total disk pressure there.

224: 

225: The constraint on the size of the  active region comes from the demand 

226: that the mechanism responsible for the formation of the resulting

227: spectrum be the Comptonization of seed photons by an optically thin, 

228: mildly relativistic plasma (e.g. Poutanen \& Svensson 1996), a 

229: process known to yield power law spectra in agreement with those 

230: observed in AGN.  

231: %

232: %The observed X-ray spectra of accreting black holes are power-laws in

233: %the zeroth order approximation. The most robust way to produce such

234: %spectra is Comptonization of seed photons in an optically thin mildly

235: %relativistic plasma (e.g., Poutanen \& Svensson 1996).  

236: In order for

237: Comptonization to be the dominant emission mechanism, the compactness

238: parameter, $l$, of the emission region must be greater than about 

239: $\sim 0.1$ (Fabian 1994; Nayakshin 1998, \S 2). To estimate the value 

240: of $l$ we assume that the luminosity of a flare, $L_1$, is given by

241: $(4\pi/3) (\Delta R)^3 (B^2/8\pi) \, t_r^{-1}$, where $t_r$ is the

242: reconnection time scale, here parameterized as $b \, \Delta R/c$ with

243: $b > 1$. With this assumption, the X-ray flux from the active region 

244: is $\fx \sim (B^2/8\pi)\,c/3b$, which with the use of Eq.(\ref{mp}) 

245: yields for the compactness

246: %

247: \begin{equation}

248: l \equiv {\sigma_T \fx \Delta R\over m_e c^3}= \beta \lambda^{4} {P_d

249: \sigma_T H\over 3 b m_e c^2}\, \left[{H\over \Delta R}\right]^3\;.

250: \label{es1}

251: \end{equation}

252: It is a simple matter to show, using equations of Svensson \&

253: Zdziarski (1994; SZ94), that for both radiation- and gas-dominated

254: disks,

255: \begin{equation}

256: {P_d \sigma_T H\over 3 m_e c^2}\, = {1\over 6\sqrt{2}}\,

257: (m_p/m_e)\alpha^{-1} \eta^{-1} r^{-3/2} \dm J(r)\;,

258: \label{es2}

259: \end{equation}

260: where $\alpha$ is the viscosity parameter; $\dm$ is the dimensionless

261: accretion rate (for $\dm=1$ the disk luminosity equals that of

262: Eddington $\ledd$); $\eta\simeq 0.06$ is the efficiency of accretion;

263: $J(r) \equiv 1 - \sqrt{3/r}$, and $r = R c^2/2 GM$.  The function

264: $J(r)/r^{3/2}$ has a maximum at $r=16/3$, with the value $\simeq

265: 0.02$. Therefore, for any radius in the disk,

266: \begin{equation}

267: l \leq 73\, \beta \lambda^{4} {\dm\over \alpha b} \,\left[{H\over \Delta

268: R}\right]^3\;.

269: \label{es3}

270: \end{equation}

271: The corresponding limit on the size of

272: the active region is

273: \begin{equation}

274: {\Delta R\over H} \leq 4.2 \left[{l\over

275: 0.1}\right]^{-1/3}\left[{\beta\over \alpha_2 b_1}\right]^{1/3}

276: \lambda^{4/3} \left({\dm \over 0.01}\right)^{1/3}\;.

277: \label{es4}

278: \end{equation}

279: where $\alpha_2\equiv

280: \alpha/100$ and $b_1\equiv b/10$.

281: Thus, the requirement of sufficiently large compactness for an active

282: region limits its height to no more than several times the disk scale

283: height $H$. This limit can be even tighter if one demands larger

284: values for $l$ (e.g. $l\simeq 10$ was assumed by Poutanen \& Fabian

285: 1999).

286: 

287: Let us now compare $H$ with the radius, $R$. Using equation (7) of

288: SZ94, one finds

289: \begin{equation}

290: \frac{H}{R} = \frac{1}{16}\, \frac{\dm}{0.01} \frac{J(r)}{r} \;.

291: \label{hr}

292: \end{equation}

293: This is a tiny number: for $r=5$, i.e., for radii around which the

294: radiation flux in the Shakura-Sunyaev disk reaches maximum, $H/R

295: \simeq 0.003$ for $\dm = 0.01$.  Note that our considerations do {\em

296: not} preclude the presence of larger size magnetic loops, but

297: they show that these loops will have very weak magnetic fields and

298: hence are unlikely to be responsible for the X-ray radiation that we

299: observe from AGN.

300: 

301: 

302: 

303: Let us now consider the area of the disk illuminated by a flare.  A

304: source that is a height $\Delta R$ above the disk will illuminate area

305: $\sim \pi (\Delta R)^2$ immediately below the flare, and the rest of

306: the disk will receive a negligible amount of X-ray illumination. This

307: area, $\pi (\Delta R)^2$, is a very small fraction of the disk full

308: area, according to Eq. (\ref{hr}). This fact has an important

309: observational implication: the instantaneous \fe line profile from a

310: single magnetic flare should be {\em narrow} (see below). With an

311: estimate of the extent $\Delta R$ of a magnetic flare 

312: (Eq. \ref{es4}) one can now estimate its luminosity as a fraction 

313: of the total disk luminosity $L = \dm \ledd$ (at $r = 6$):

314: %

315: \begin{equation}

316: \frac{L_1}{L}\simeq  1.4\times 10^{-2} \frac{\beta

317: \lambda^4}{\alpha_2 b_1} \left(\frac{H}{\Delta R}\right)^2

318: \frac{\dm}{0.01}\;.

319: \label{l1}

320: \end{equation}

321: %

322: Because both $\lambda$ and $\beta$ must be smaller than unity by

323: definition, the above relation shows that if the X-ray luminosity is a

324: good fraction of the bolometric luminosity, then one needs $N$ (the

325: number of such X-ray emitting regions) $\sim$ from few tens up to few

326: thousand at any one time. If flares occurred completely randomly,

327: independently of each other, then one would expect an RMS variability

328: amplitude $\sim 1/\sqrt{N}$. This latter number seems to be too small

329: given that it is not atypical for AGN to exhibit variations in X-ray

330: flux by factors of 2 or so.

331: 

332: To account for the observed variability one needs to invoke avalanches

333: of magnetic flares, i.e., require that each observed X-ray flare

334: consists of many correlated (in time and space) individual active

335: regions (e.g., Poutanen \& Fabian 1999). Physically, it is likely that

336: magnetic flux tubes rise above the disk but do not immediately

337: reconnect, settling down into a quasi-static equilibrium

338: state. However, with time, more and more magnetic flux tubes arise

339: from the disk, covering the disk surface. Additionally, the tubes that

340: are already above the disk are sheared by differential rotation of the

341: foot-points and hence some of them will be taken out of equilibrium

342: (i.e., the reconnection will start).  If some region of the disk is

343: particularly closely packed with flux tubes, then it is possible that

344: the active magnetic flares will affect its neighbors, setting them off

345: as in a chain reaction, producing the flare avalanche and leading to a

346: flare-like event in the observer's light curve.

347: 

348: Note that for this scenario to be plausible, the magnetic flares

349: taking part in the avalanche must be near each other, otherwise it is

350: hard to see how they can interact over distances greater than their

351: own size, $\Delta R$. Therefore, from now on, we will accept that the

352: active flares contributing to a particular observer's flare in the

353: light curve (which can be as large as $\sim 50$\% of the average X-ray

354: luminosity of an AGN) closely pack a disk region of area $A$.  Assume

355: that the active region produces a fraction $\zeta$ of the bolometric

356: luminosity, $L$. The maximum X-ray flux produced by flares in the

357: active region is $\fx \sim c P_d b^{-1}$. The area of the disk covered

358: with magnetic flares responsible for the luminosity $\zeta L$ is thus

359: $A\sim \zeta L/ \fx$. We can compare this area with the effective area

360: of the inner accretion disk $\pi R^2$ if we note that $L \sim \fdisk

361: \pi R^2$, where $\fdisk = (9/8) \alpha c_s P_d$ is the Shakura-Sunyaev

362: disk flux, and $c_s$ is the mid-plane sound speed:

363: \begin{equation}

364: \frac{A}{\pi R^2}\sim \zeta \alpha b \frac{c_s}{c} \simeq 10^{-4}

365: \alpha \zeta b\;,

366: \label{ratio}

367: \end{equation}

368: where we used $T\sim 10^5$ to estimate $c_s$. The combination of the

369: remaining factors in the equation above is probably less than unity,

370: and so clearly $A\ll R^2$.  If we assume that the active region has

371: about same dimensions in the radial ($\delta R$) and azimuthal (R

372: $\delta \phi$) directions, then we conclude that $\delta R/R \sim

373: \delta \phi/2\pi \sim 0.01$.

374: 

375: \section{Time resolved \fe line profile}

376: 

377: An immediate implication of the discussion of the previous section is

378: the prediction that the instantaneous \fe line profile as seen by an 

379: observer at

380: infinity should be narrow. Indeed, for a non-rotating black hole, and

381: neglecting photon ray bending due to strong gravitational fields for a

382: moment, the observer sees \fe line photons of energy $E(R, \phi)$ from

383: a point source at $(R, \phi)$

384: \begin{equation}

385: E(R,\phi) = E_0 (1- 2/r)^{1/2} \left[\gamma (1 - v

386: \cos\alpha_0)\right]^{-1}\;,

387: \label{e0}

388: \end{equation}

389: where $v$ is the orbital velocity in units of speed of light at radius

390: $r$, $\alpha_0$ is the angle this velocity makes with the direction to

391: the observer, $\gamma \equiv (1-v^2)^{-1/2}$, and

392: $E_0$ is the photon rest frame energy.  Approximately, the width of

393: the line profile will be

394: \begin{equation}

395: \delta E = |\frac{\partial E}{\partial R} \delta R | + |\frac{\partial

396: E}{\partial \phi} \delta \phi |\;.

397: \label{eshift}

398: \end{equation}

399: For a pole-on observer the second term vanishes and $\delta E \sim

400: 0.1 E_0 (\delta R/R) \sim 6$ eV (!) for the chosen values of $\delta R,

401: \delta \phi$, much

402: smaller than the red or blue-shift of the line centroid itself.

403: $\delta E$ generally increases with increasing inclination angle. The

404: maximum of the second term in eq. (\ref{eshift}) is for an equatorial

405: observer when $\alpha = \pi/2$ or $3 \pi/2$: max$|\partial E/\partial

406: \phi| = E(R,\phi) v\simeq E_0 (r-2)^{-1/2}$, which yields $\delta E

407: \simeq 130 {\rm eV} (\delta \phi/0.01\pi)$. Hence the width of \fe

408: line profile from a \sw disk flare should vary between as little as

409: $\sim 6$ eV to $\sim 100$ eV, with a ``reasonable'' value of $\sim 30$

410: eV for Seyfert 1 Galaxies that are thought to be nearly pole-on.  Note

411: that the contrast of these features to the continuum flux is

412: larger (by a factor of $\sim 10$) for nearly face-on than for nearly

413: edge-on disks.

414: 

415: 

416: 

417: Let us now calculate the time-resolved \fe line profile from a single

418: rotating spot. We will assume that its radial size $\delta R$ is

419: negligible, and $\delta \phi/(2\pi) = 0.01$.  Note that the time delay

420: between the start of a magnetic flare and the \fe line emission from

421: the flare is very small, i.e., it is $\sim H/c$, and so we can neglect

422: it.  For this first study, we will assume that the \fe line emissivity is

423: isotropic in the rest frame and is also constant over the active region's

424: life time. Finally, we will consider the case of a flare lasting

425: exactly one rotation for a non-rotating black hole at a radius $R = 12

426: GM/c^2$.

427: 

428: In polar spherical coordinates with the black hole at the center and

429: the disk treated as a plane at $\theta = \pi/2$, we place the observer

430: at $r_o = 10^3$, azimuthal angle $\phi = 0$ and polar angle $\theta =

431: \theta_o$. The active region turns on at $\phi = - \pi$ and turns off

432: at $\phi = \pi$ by assumption. We consider three different 

433: values of the observer

434: inclination angles: $\theta_o = 4$, $27$ and $63$ degrees. To

435: calculate the time dependent line profiles, we isotropically (in the

436: rest frame of the active region) emit a large number of photons and

437: then trace their trajectories until they reach radius $r_o$. The

438: photon ray tracing is performed using appendices A3 and A4 of Reynolds

439: et al. (1999). Only the photons that arrived within angle $\theta =

440: \theta_o \pm 2^{\circ}$ and $\phi = \pm 0.005\times 2\pi$ are

441: recorded. Figure 1 shows the resulting \fe line trajectory in the

442: energy-time diagram for the three chosen angles. One can easily derive

443: an analytical expression for these tracks neglecting photon ray

444: bending (e.g., eq. \ref{eshift} with $\alpha$ calculated as a function

445: of photon arrival time). This expression agrees with the curves shown

446: in Figure 1 quite well (because for $r=12$ and not too large

447: $\theta_o$, the ray bending is relatively weak).

448: 

449: When the observer is nearly pole-on, there is almost no Doppler boost,

450: so the photon line trajectory in Figure 1 is a straight line . At

451: larger angles, line photons can be either red or blue shifted

452: depending on the azimuthal separation of the source and the

453: observer. Finally, one should note that the larger the inclination

454: angle, the shorter (in time) the blue-shifted section of the S-shaped

455: trajectory is. This is simply due to the fact that the source moves

456: towards the observer for $-\pi < \phi < 0$ and then it moves away from

457: the observer for $0 < \phi < \pi$.  Also note that similar

458: trajectories can be noticed in some of the response functions

459: calculated by Ruszkowski (2000; see his Figures 5,7 \& 9).  The latter

460: are however much broader because the X-ray source is located much

461: higher above the disk.

462: 

463: \begin{figure*}[t]

464: \centerline{\psfig{file=fig.eps,width=.4\textwidth,angle=90}}

465: \caption{\fe line trajectories in energy-time space for three

466: different viewing angles: $\theta = 4$, 27 and $63^\circ$. The flare

467: is located at radius $R = 12 GM/c^2$. The larger the viewing angle,

468: the larger the spread between maximum and minimum energy observed.}

469: \label{fig:temp}

470: \end{figure*}

471: 

472: 

473: 

474: \section{Discussion}

475: 

476: In this Letter we showed that the area illuminated by X-rays from an

477: active magnetic flare is small compared to the total disk area.  The

478: time-resolved \fe line profile from this active region should then be

479: a narrow feature sweeping with the rotation period across the entire

480: range of red(blue)-shift associated with the given radius and

481: inclination angle.  We calculated several examples of such line

482: profiles for a non-rotating black hole. We now discuss the

483: implications of our results assuming that such a narrow moving \fe

484: line trajectory can be observed in the future with observatories such

485: as XMM-Newton or (more realistically) Constellation-X.

486: 

487: (1) The trajectories of the \fe feature in the energy-time diagram are

488: very different from those obtained by Reynolds et al. (1999), Young \&

489: Reynolds (2000) and Ruszkowski (2000) for the lamp-post model or

490: highly elevated flares.  Indeed, these latter profiles are

491: considerably broader -- i.e., $\delta E\sim 1$ keV with the red- and

492: blue-shifted wings of the line appearing simultaneously. This is due

493: to the fact that a highly elevated X-ray source illuminates a large

494: fraction of the disk area. In contrast, we found $\delta E \simlt 100$

495: eV with the line being either red- or blue-shifted at any particular

496: time. Future observations of sufficient energy and time resolution

497: should be able to confirm these results thereby setting limits on the

498: size and height of X-ray emitting regions in accretion disks.

499: 

500: (2) The much more restricted extent of the X-ray emitting region

501: argued for in this note compared to that of e.g. the lamp-post model

502: presents a much more unequivocal probe of the underlying geometry. In

503: the lamp-post case, one of the complications is that even if the

504: accretion disk proper does not exist at radii smaller than the last

505: stable orbit, there is still gas there.  This gas is thought to be

506: free-falling into the black hole.  As shown by Reynolds \& Begelman

507: (1997), the vertical Thomson depth of this material can be

508: significant, thus suggesting that it will also produce a fluorescent

509: \fe line emission if it is illuminated by X-rays. And indeed, if the

510: X-ray source is located on the height $\simgt$~few~$GM/c^2$ above the

511: disk, then this region also contributes to the observed line profile

512: (Reynolds \& Begelman 1997). Young, Ross \& Fabian (1998) argued that

513: a detailed account of disk photo-ionization can still constrain the

514: location of the innermost radius of the disk in the particular case of

515: the well known AGN~~ MCG-6-30-15. For time-dependent reflection, the

516: distinction between rotating and non-rotating holes is in a red-ward

517: moving bump in the line profile, as shown by Reynolds et al. (1999)

518: and Young \& Reynolds (2000). However, recently, it was realized that

519: thermal ionization instability plays a crucial role in the

520: photo-ionized reflection (Nayakshin, Kazanas \& Kallman 2000).  

521: Recent modeling of time-dependent reflection that included thermal

522: ionization instability (Nayakshin \& Kazanas 2001) shows that the

523: physics of the problem is far richer than it has been thought based

524: on constant density models. We therefore feel that distinguishing

525: between rotating and non-rotating black holes in the lamp-post

526: geometry can be non trivial in practice.

527: 

528: In contrast, magnetic flares can only occur where there is a

529: continuous energy production in the underlying disk, that is at radii

530: greater than the last stable orbit. Since the flares illuminate only

531: the disk immediately below them, the region inwards of the last stable

532: orbit receives a negligible amount of X-rays and hence emits no line.

533: Observation of a \fe line shifted by an amount only consistent with

534: $R < 6 GM/c^2$, would strongly argue for a rotating black hole, then.

535: 

536: 

537: (3) \fe line trajectories similar to those in Figure 1 could be used to

538: constrain circular orbits of test particles in the gravitational

539: field, as well as photon ray bending. In principle, this information

540: can be used to test General Relativity in the strong field limit.

541: Note that, as long as the line profile is sufficiently narrow, this

542: procedure would require knowledge of only the \fe line energy versus time

543: and not the amplitude of the line emission; hence complications due to

544: anisotropy of the magnetic flare emissivity or photo-ionization

545: physics may be neglected (this is not so for the lamppost case because

546: it may be hard to define the peak of the line if the profile is broad

547: and is overlayed on the top of a continuum emission).

548: 

549: 

550: 

551: The real \fe line profiles should be produced by many magnetic flares

552: (see eq. \ref{l1}). If they occur randomly, at completely independent

553: locations, then it will be perhaps impossible to distinguish

554: individual line ``trails'' if of sufficiently large number. However, 

555: as we discussed in \S 2, this would also preclude any substantial

556: variability in the X-ray continuum flux, while one often

557: observes variations of up to $\sim 50$ \% in the continuum flux (e.g.,

558: Edelson et al. 2000). Therefore, we believe that observed large

559: amplitude excursions of the AGN X-ray flux can only be accounted by

560: magnetic flares if the latter are not independent of each other, i.e.,

561: if they take part in flare avalanches (Poutanen \& Fabian

562: 1999). Further, it is difficult to see how small scale flares can

563: affect each other unless they are in physical proximity. Therefore, 

564: the line profile from a magnetic flare avalanche should also be narrow

565: and hence the points made above hold true. On the other hand, the

566: active region (the avalanche region) may be sheared by the

567: differential rotation in the disk (although magnetic field can also be

568: strong enough throughout the active region to inhibit differential

569: rotation [E. Vishniac, private communication]). Thus, the region may

570: become larger in extent and the point-source approximation will be

571: invalid.  Whether this is the case or not could be determined directly

572: from time-resolved \fe line spectroscopy.

573: 

574: 

575: %The observations of the features we propose in this paper

576: %could be performed, in the absence of a direct observation

577: %a narrow feature moving across the line width, through a 

578: %cross-correlation analysis of the flux in fairly narrow

579: %bins within the observed line profile. By varying the width 

580: %of the energy bin, a search for a peak in the cross-correlation 

581: %function with another bin within the profile can provide (if 

582: %detected), not only the position and extent of the X-ray emitting 

583: %region but also of inclination angle of the accretion disk. 

584: %Clearly, emission by flares at the smaller radii, while 

585: %potentially more luminous, cover a broader range of energies

586: %at much shorter time scales and are therefore less easily

587: %detectable. 

588: 

589: In summary, we conclude that instantaneous \fe line profiles from

590: magnetic flares and highly elevated X-ray sources, such as a lamppost,

591: are very much different (e.g., compare Fig. 1 to Figures in Reynolds

592: et al. 1999), and hence time-resolved \fe line observations have an

593: enormous potential for constraining accretion disk theories.  Given

594: this, Astrophysics community should spare no effort in developing

595: future space observatories such as Constellation-X.

596: 

597: 

598: 

599: 

600: 

601: 

602: 

603: 

604: 

605: 

606: 

607: 

608: 

609: 

610: 

611: 

612: 

613: 

614: 

615: 

616: 

617: \begin{thebibliography}{}

618: 

619: %\bibitem[]{} Blandford, R.D., \& McKee, C.F., 1982, \apj, 255, 419

620: 

621: %\bibitem[]{} Blandford, R.D., \& Begelman, M.C. 1999, \mnras, F3, L1

622: 

623: %\bibitem[]{} Done, C., \& Nayakshin, S. 2001, to be submitted to ApJ.

624: 

625: 

626: \bibitem[]{} Edelson, R., et al. 2000, \apj, 534, 180

627: 

628: \bibitem[]{} Fabian, A.C. 1994, ApJS, 92, 555

629: 

630: \bibitem[]{} Fabian, A.C., Iwasawa, K., Reynolds, C.S., \& Young,

631: A.J. 2000, PASP accepted (astro-ph/0004366)

632: 

633: \bibitem[]{} Galeev, A. A., Rosner, R., \& Vaiana, G. S., 1979, \apj,

634: 229, 318

635: 

636: \bibitem[Haardt, Maraschi and Ghiselli (1994)]{h94} Haardt F.,

637: Maraschi, L., \& Ghisellini, G. 1994, \apjl, 432, L95

638: 

639: %\bibitem[]{} Ichimaru, S. 1977, \apj, 214, 840

640: 

641: \bibitem[]{} Merloni, A. \& Fabian, A.C., 2001, to appear in \mnras

642: 

643: %\bibitem[]{} Narayan, R. \& Yi. I. 1984, \apj, 428, L13

644: 

645: \bibitem[]{} Nayakshin S. 1998, PhD thesis, University of Arizona

646: 

647: %\bibitem[]{} Nayakshin, S., \& Kallman, T. 2001, to appear in ApJ

648: 

649: \bibitem[]{} Nayakshin, S., Kazanas, D., \& Kallman, T. 2000, ApJ,

650: 537, 833

651: 

652: 

653: \bibitem[]{} Nayakshin, S., \& Kazanas, D. 2001, to be submitted to ApJ

654: 

655: \bibitem[]{} Parker, E.N. 1979, Cosmical Magnetic Fields, Clarendon

656: Press, Oxford.

657: 

658: \bibitem[]{} Poutanen, J. \& Svensson, R. 1996, \apj, 470, 249

659: 

660: \bibitem[]{} Poutanen, J. \& Fabian, A.C. 1999, \mnras, 306, L31

661: 

662: %\bibitem[]{} Rees, M.J., Begelman, M.C., Blandford, R.D., \& Phinney,

663: %E.S. 1982, Nature, 295, 17

664: 

665: \bibitem[Reynolds and Begelman (1997)]{rb97} Reynolds, C.S.,

666: \& Begelman, M.C. 1997, \apj, 488, 109

667: 

668: \bibitem[]{} Reynolds, C.S., Young,

669: A.J., Begelman, M.C., \& Fabian, A.C. 1999, \apj, 514, 164

670: 

671: \bibitem[]{} Romanova, M. M. Ustyugova, G. V., Koldoba, A. V.,

672: Chechetkin, V. M., \& Lovelace, R. V. E 1998, \apj, 500, 703

673: 

674: \bibitem[]{} Ruszkowski, M. 2000, \mnras, 315, 1

675: 

676: %\bibitem[Shakura \& Sunyaev 1973] {sun73} Shakura, N.I., \& Sunyaev,

677: %R.A. 1973, A\&A, 24, 337

678: 

679: %\bibitem[]{} Svensson, R. 1996, A\&A Supl., 120, 475

680: 

681: \bibitem[]{} Svensson, R. \& Zdziarski, A. A. 1994, \apj, 436, 599

682: 

683: \bibitem[]{} Young, A.J. \& Reynolds, C.S., \apj, 2000, 529, 101

684: 

685: \end{thebibliography}{}

686: 

687: 

688: 

689: \end{document}

690: 

691: \begin{figure*}

692: \begin{minipage}[t]{.3\textwidth}

693: \centerline{\psfig{file=3.6.epsi,width=\textwidth,angle=90}}

694: \end{minipage}\hfill

695: \begin{minipage}[t]{.3\textwidth}

696: {\psfig{file=27.epsi,width=\textwidth,angle=90}}

697: \end{minipage}\hfill

698: \begin{minipage}[t]{.3\textwidth}

699: {\psfig{file=54.epsi,width=\textwidth,angle=90}}

700: \end{minipage}

701: \caption{Photon line ``tracks'' -- photon energy versus time for three

702: different inclination angles. From left to right, the observer is

703: located at $\theta_o= 4$, 27 and 54$^\circ$. See text for

704: details. Note that the gaps in the line trajectory are due to finite

705: number of photons traced.}

706: \label{fig:1}

707: \end{figure*}

708: 

709: 

710: 

711: 

712: 

713: These new physical effects in \fe line reverberation problem

714: (discussed here and in Nayakshin \& Kazanas 2001) should not be viewed

715: as complicating the cause for actual observing time-resolved \fe line

716: profiles. On the contrary, as long as we have a clear understanding of

717: these effects, we can always use them to our advantage to increase the

718: amount of physics learned from observations. In addition, e

719: 

720: 

721: 

722: In retrospective, it is interesting to note that \fe line

723: reverberation turns out to be quite different from the optical-UV line

724: reverberation formulated by Blandford \& McKee (1982). In the latter

725: case, the proper motion of the continuum source is irrelevant, since

726: the reverberating region is much greater in size than the inner

727: accretion disk. As our calculations show, for time-resolved \fe line

728: profiles from magnetic flares, however, the motion of the flare itself

729: cannot be neglected.  In addition, the physics of reprocessing of the

730: incident X-rays into \fe line photons produces an important relaxation

731: time-scale that is long but comparable with the light crossing time of

732: the reverberation region (Nayakshin \& Kazanas 2001).  [Note that for

733: magnetic flares, this time scale is considerably shorter than it is

734: for the lamppost.] There is no similar relaxation time-scale for the

735: optical/UV line reverberation (Blandford \& McKee 1982).

736: 

737: 

738: (2) Although we did not show any line trajectories for the Kerr

739: metric, clearly, they will be quite distinct from those of a

740: non-rotating black hole. Indeed, 

741: 

742: 

743: (3) The conclusions about geometry of the inner disk will be much more

744: robust than those obtained with conventional time-averaged line

745: profiles. The latter can be very broad, and often the distinction

746: between different disk configurations is in the far wings of the

747: line. The exact shape of these wings, in our opinion, is less reliable

748: observationally, because they are typically weak, and one needs to

749: assume that the instrument response matrix and the illuminating

750: continuum shape are perfectly known.

751: 

752: (4) Finally, such line profiles may be used to test the theory of

753: General Relativity in the strong limit, since the energy-time \fe line

754: trajectories similar to Fig. (1) can be used to constrain circular

755: orbits of test particles in the gravitational field, as well as photon

756: ray bending.  Note that this would require knowledge of only \fe line

757: energy versus time and not the amplitude of the \fe line emission, and

758: hence complications due to anisotropy of the magnetic flare emissivity

759: or photo-ionization physics may be neglected (this is not so for the

760: lamppost case -- see Nayakshin \& Kazanas 2001).

761: 

762: 

763: 

764: Though not exhibited herein,

765: the trajectories in the energy-time diagram corresponding

766: to the Kerr metric are substantially broader and of shorter

767: period than those presented in Figure 1. Observations 

768: of moving \fe features consistent with such values of the parameters

769: would provide  support for associating the Kerr metric 

770: with the compact object and may even be able to serve as 

771: tests of General Relativity in the strong field limit.

772: 

773: 

774: 

775: In order to compute time-resolved \fe line profiles from a

776: rotating active region, one should first compute the local reflected

777: spectrum from photo-ionized gas. For magnetic flares, the problem is

778: complicated by the existence of winds and has not yet been solved

779: completely self-consistently (e.g., Nayakshin \& Kallman 2001, \S 4.1;

780: also Done \& Nayakshin 2001). In addition, one would need to know

781: whether the luminosity of the active region changes with time as it

782: rotates.