astro-ph0507152/ms.tex
1: %\documentstyle[emulateapj]{article}
2: %\documentstyle[aas2pp4]{article}
3: %\documentstyle[aaspp]{article}
4: %\documentstyle[apjpt]{article}
5: %\documentstyle[12pt]{article}
6: \documentclass[12pt,preprint]{aastex}
7: 
8: \begin{document}
9: 
10: 
11: %\documentclass[12pt,preprint]{aastex}
12: %\usepackage{natbib}
13: %\citestyle{aa}
14: 
15: %\received{}
16: %\accepted{}
17: %\journalid{}{}
18: %\articleid{}{}
19: 
20: 
21: 
22: \title{On how much X-ray and UV radiation processes are coupled 
23: in accretion disks: AGN case}
24: 
25: \author{Daniel Proga,\altaffilmark{1}}
26: 
27: \affil{$^1$ Princeton University Observatory, Peyton Hall, Princeton, NJ 08544.e-mail: dproga@astro.princeton.edu}
28: 
29: \def\LSUN{\rm L_{\odot}}
30: \def\MSUN{\rm M_{\odot}}
31: \def\RSUN{\rm R_{\odot}} 
32: \def\MSUNYR{\rm M_{\odot}\,yr^{-1}}
33: \def\MSUNS{\rm M_{\odot}\,s^{-1}}
34: \def\MDOT{\dot{M}}
35: 
36: \newbox\grsign \setbox\grsign=\hbox{$>$} \newdimen\grdimen \grdimen=\ht\grsign
37: \newbox\simlessbox \newbox\simgreatbox
38: \setbox\simgreatbox=\hbox{\raise.5ex\hbox{$>$}\llap
39:      {\lower.5ex\hbox{$\sim$}}}\ht1=\grdimen\dp1=0pt
40: \setbox\simlessbox=\hbox{\raise.5ex\hbox{$<$}\llap
41:      {\lower.5ex\hbox{$\sim$}}}\ht2=\grdimen\dp2=0pt
42: \def\simgreat{\mathrel{\copy\simgreatbox}}
43: \def\simless{\mathrel{\copy\simlessbox}}
44: 
45: 
46: \begin{abstract}
47: Within the standard accretion disk theory for active galactic nuclei (AGN), 
48: the observed X-rays are often modeled by Compton up-scattering 
49: of ultraviolet (UV) disk photons inside a hot disk corona.
50: Here, we point out that for many AGN, radiation pressure due 
51: to the very same UV disk photons can drive a flow from the disk into 
52: the corona and couple the processes producing X-rays and UV photons.
53: This coupling could lead to quenching of the disk corona 
54: because the regions above the UV disk will be too dense, too opaque, 
55: and consequently too cold. We discuss various consequences of 
56: this new type of the X-ray/UV coupling on the dynamical and 
57: radiative properties of AGN. 
58: \end{abstract}
59: 
60: 
61: \keywords{accretion, accretion disks  -- galaxies: active: nuclei --
62: methods: numerical -- radiation mechanisms}
63: 
64: \section{Introduction}
65: High X-ray fluxes observed in active galactic nuclei (AGN) 
66: are a serious challenge for the standard accretion disk theory.
67: They are also a serious challenge for any AGN outflow model
68: because no matter what is the physics of
69: X-ray production, one has to deal with the so-called overionization 
70: problem:  how outflows absorbing the ultraviolet (UV) radiation avoid 
71: full photoionization due to strong  X-ray radiation.
72: 
73: In most pictures depicting the centers of AGN, the X-rays are 
74: produced in a small region, referred to as the central engine, whereas 
75: the outflows are produced outside of the central engine. 
76: Therefore, the X-ray production mechanisms
77: are usually considered to operate separately from the outflow production 
78: mechanisms.
79: In this paper, we discuss new dynamical and radiative consequences of 
80: UV  photons being emitted into the central engine.
81: In particular, we argue that radiation pressure due to UV lines 
82: (line force) couples the X-ray and UV radiation processes 
83: by driving the disk material into the regime above the disk
84: where X-rays are to be produced.
85: 
86: 
87: \section{Disk Corona and Disk Outflow: ``a tale of two merging cities''} 
88: The spectral energy distribution (SED) of active galactic nuclei (AGN) 
89: is very broad. It spans the wavelength range from radio to hard X-rays.
90: Most of the AGN luminosity, $L$ is in 
91: the optical--UV--IR regime but some significant fraction is in the X-ray band. 
92: It is commonly accepted that AGN are powered by accretion of matter 
93: onto a supermassive black hole (BH). AGN are typically very luminous 
94: (0.001~$L_{Edd}\simless L \simless 1~L_{Edd}$, where $L_{Edd}$ 
95: is the Eddington limit).
96: For most AGN, the accretion flow
97: is thought to form an optically thick, geometrically thin Keplerian
98: disk. In the standard picture, this disk radiates thermally mostly
99: in the optical--UV regime (Shakura \& Sunyaev 1973).
100: 
101: \subsection{Disk Corona}
102: Generally, the optically thick disk model can account for
103: the optical--UV radiation but it does not account for the spectral shape
104: and high flux observed in X-rays. 
105: The most commonly accepted model for the production of X-rays is 
106: multiple Compton up-scattering (Comptonization) of UV photons 
107: (e.g., Sunyaev \& Titarchuck 1980) off hot electrons in a disk corona
108: (e.g., Walter \& Courvoisier 1992; Haardt \& Maraschi 1991, 1993;
109: Sobolewska et al. 2004).
110: The derived X-ray spectrum depends on 
111: the temperature and optical
112: depth of the scattering electrons.
113: Despite recent advances in observations and modeling 
114: of X-rays, the geometry and radiation processes responsible for X-ray
115: emission remain poorly constrained. 
116: Two types of geometries are being considered for an accretion disk
117: and Comptonizing corona: (1) ``slab'' or ``sandwich'' geometries 
118: (e.g., the top and bottom panels of  Fig.~1 in Reynolds \& Nowak 2003, 
119: hereafter RN) and
120: (2) ``sphere+disk geometries'' (e.g., the middle two panels  
121: of  Fig.~1 in RN). 
122: 
123: Here we focus on exploring the first class of geometries where
124: the hot corona is thought to be located immediately above the disk.
125: The disk emits soft thermal photons which provide 
126: the main source of cooling for the hot electrons in the corona.
127: It is possible that at the same time, the hard photons produced by
128: Comptonization are an important source of heating for the reflecting
129: matter which reprocesses them into soft photons.
130: Thus there is a radiative coupling between the X-ray and UV radiation.
131: Many models account for this coupling and assume 
132: that a significant fraction of the gravitational power
133: is dissipated in the hot tenuous layers  
134: (e.g., Haardt \& Maraschi 1993; Svensson \& Zdziarski 1994). 
135: Although the basic idea behind
136: a hot disk corona is  straightforward, the physical model of the corona
137: remains one of the biggest challenges in the field. Phenomenologically,
138: one can imagine that the corona is heated by dissipation of
139: the accretion power via magnetic processes (e.g., Galeev et al. 1979; 
140: Field \& Rogers 1993).
141: 
142: Direct studies of these complex multidimensional, time-dependent  
143: processes by means of numerical magnetohydrodynamical (MHD) simulations 
144: have begun. For example, Miller \& Stone (2000) found that in 
145: a stratified disk, a  MHD turbulence driven by the magnetorotational 
146: instability (MRI, Balbus \& Hawley 1998) is capable of driving 
147: magnetogravitational modes of the Parker instability.
148: However, it has not been demonstrated yet that magnetic buoyancy 
149: can supply the disk corona with sufficient power to explain 
150: the observed X-ray emission. In fact, several numerical simulations indicate 
151: that local dissipation not magnetic buoyancy
152: is the primary saturation mechanism of the MRI (e.g., Brandenburg et al. 1995; 
153: Stone et al. 1996; Miller \& Stone 2000, Turner 2004).
154: 
155: \subsection{Disk Wind}
156: Viable models for AGN must also account for
157: mass outflows which is another important aspect of activity in galactic nuclei.
158: The most relevant to this paper are the outflows that can be 
159: inferred from spectral features observed in the UV and X-ray bands
160: (i.e., we do not discuss AGN jets).
161: Broad absorption lines (BALs) in QSOs are the best example
162: of spectral features revealing the existence of
163: such outflows. These lines are almost always blueshifted relative to 
164: the emission-line rest frame, indicating the presence of outflows from 
165: the active nucleus, with velocities as large as 0.2~c
166: (e.g., Turnshek 1998). The apparent X-ray properties can be affected
167: by outflows, too. For example, 
168: the relative strength of the soft X-ray flux anti-correlates 
169: with the C~IV absorption equivalent width for QSOs 
170: (e.g., Brandt, Laor \& Wills 2000). 
171: 
172: There has been considerable time and effort spent to understand AGN outflows
173: (e.g., Arav, Shlosman \& Weymann 1997; Crenshaw, Kraemer \& George 2002, 
174: and references therein). In particular, many theoretical models have been 
175: proposed to explain outflows in AGNs 
176: (e.g., Crenshaw et al. 2002).  
177: One of most plausible scenarios for AGN outflows is a wind from an
178: accretion disk around a black hole where
179: the line force drives a wind from a disk by the 
180: local disk radiation at radii where the disk radiation is mostly in the UV
181: (e.g., Murray et al. 1995,  MCGV hereafter; 
182: Proga, Stone, \& Kallman 2000, PSK hereafter; 
183: Proga \& Kallman 2004, PK hereafter).
184: In this scenario, AGN outflows are a natural consequence of the standard 
185: accretion disk theory because the theory predicts high enough radiative flux 
186: and gas opacity in the UV regime, for the line force to drive an outflow. 
187: 
188: Numerical simulations of the radiation driven disk winds
189: illustrate why and how a wind from an
190: accretion disk can account for AGN outflows. In particular, the
191: simulations have been essential in 
192: studying the robustness of the radiation launching and
193: acceleration of the wind (e.g., PSK, PK). 
194: For example, PK considered 
195: relatively unfavorable conditions for line driving (LD) as they took into 
196: account the central engine radiation as a source of ionizing photons 
197: but neglected its contribution to the radiation force. Additionally,
198: they accounted for the attenuation of the X-ray radiation by computing
199: the X-ray optical depth in the radial direction assuming that only electron 
200: scattering contributes to the opacity. The main result of the simulations 
201: is that the disk atmosphere can 'shield' itself  from external X-rays
202: so that the local disk radiation can launch gas off the disk
203: photosphere.  
204: 
205: \subsection{The Role of the Failed Wind}
206: 
207: LD can change the flow near the accretion disk 
208: in many ways. A powerful wind is just the most dramatic of them 
209: for which fairly strict requirements must be met. 
210: Using some physical arguments as well as 
211: numerical simulations one can show that line-driven disk winds 
212: are produced only when the effective luminosity of the disk (i.e. 
213: the luminosity of the disk, $L_D$ times the total line opacity 
214: in the optically thin case, $M_{max}$) exceeds the Eddington limit 
215: (e.g., Proga, Stone \& Drew 1998, PSD hereafter).
216: For the BH mass $M_{BH}=10^8~\MSUN$ of a typical quasar, PK found that 
217: for $L_D > 0.3 L_{Edd}$ a strong disk wind develops 
218: whereas for $L_D \simless 0.3 L_{Edd}$ there is no disk wind. 
219: For a less luminous disk or stronger ionizing radiation that reduces 
220: $M_{max}$, or both, the line force can still lift material off the disk 
221: but it fails to accelerate the flow to escape velocity. 
222: Such a failed disk wind has been
223: found in simulations with and without X-ray ionization (see PSD~98 for 
224: no X-ray cases, runs 1 and 6 there, and PSK and PK  for X-ray cases). 
225: 
226: For  X-ray cases, a large fraction of the failed wind is not fully ionized
227: and its temperature is comparable the disk effective temperature, $T_D$
228: (see Fig.~1). 
229: LD can then change the vertical structure of an accretion disk 
230: by dynamically increasing the disk scale height. Therefore, a failed wind 
231: solution can be referred to as a puffed-up disk. 
232: 
233: The disk wind solution is very sensitive to
234: $M_{BH}$: for a fixed ratio, $L_D/L_{Edd}$
235: (e.g., 0.5) it is easier to produce a wind for
236: $M_{\rm BH} \simgreat 10^7~\MSUN$ than for
237: $M_{\rm BH} \simless 10^7~\MSUN$ (PK). 
238: Thus less dramatic but very important changes in the density distribution
239: above the disk occur for a broad range of AGN luminosities and BH masses
240: (e.g., $L_D M_{max} < L_{Edd}$).
241: We note that LD can change the density profiles 
242: even inside the disk. For example, dense and massive clumps
243: of gas lifted by radiation can fall back on the disk (e.g., Fig. 1)
244: and perturb the disk structure and radiative properties.
245: 
246: 
247: To illustrate properties of the puffed-up disk/failed wind, we
248: show results from one of PK's simulations for an AGN with $M_{BH}=10^8~\MSUN$
249: (we refer a reader to PK for a description of the calculations).
250: PK present detail results from only one of the simulations
251: where $M_{BH}=10^8~\MSUN$ and $L_D=0.5~L_{Edd}$.
252: Here, we present and discuss results for the same $M_{BH}$ but
253: for $L_D=0.3~L_{Edd}$. For these parameters,
254: the disk radiation launches the flow off the disk but fails
255: to accelerate it to escape velocity (Fig.~1).
256: 
257: The top panel of Fig.~1 shows very clearly a high density
258: flow above the disk, i.e., the line force can maintain gas with
259: the density as high as $10^{-14}~{\rm g~cm^{-3}}$ as high as
260: $40\%$ of radius along the equator.
261: We note that in the radiation dominated regime,
262: the standard disk theory predicts
263: the disk scale height, $H_D=3 L/L_{Edd}~r_\ast=0.9 r_\ast$ 
264: (where $r_\ast\equiv 6 GM_{BH}/c^2$ is the inner disk radius and
265: unit of the length scale: $r'=r/r_\ast$ and $z'=z/r_\ast$).
266: The density distribution in the vertical
267: direction  is much broader compared not only to the standard disk model
268: but also compared to the X-ray heated disk corona. For example,
269: for the gas temperature $T=8\times10^8$~K, the scale height of a disk corona 
270: in hydrostatic equilibrium (HSE), 
271: $H_C/r_\ast=[k T r^3/(\mu m_p GM_{BH})]^{1/2}=0.7$ at $r'=20$.
272: The puffed-up disk is far from HSE and displays unsteady 
273: behavior: the flow is complex with a few filaments 
274: and various knots and clumps of gas moving both upwards and
275: downwards.  The direction and speed of motion at any one position is apt
276: to change unpredictably with time. 
277: There are two main reasons for this behavior: 
278: (i) the gravity and driving flux differently scale with $z'$ 
279: (e.g., PSD) and (ii) the overionization of the innermost
280: flow by the central engine radiation (compare the left and right panels).
281: 
282: PK's line-driven wind model takes into account X-ray radiation 
283: from a point source
284: located at the origin of the coordinate system. Thus it does not
285: account for the X-ray emission as envisioned in the disk corona scenario
286: described in \S2.1. However, we can estimate some of the effects of
287: X-rays emitted immediately above the UV disk.
288: For example, we can estimate the photoionization parameter, 
289: $\xi\equiv 4 \pi F_x/n$ inside the puffed-up disk ($F_x$ is the 
290: X-ray flux and $n$ is the number  density).
291: We assume that at a give point ($r', z'$) above the disk,
292: the ionizing flux  is comparable to the flux emitted by the 
293: disk at ($r',0$), i.e., $F_x(r',z')\approx \sigma T^4_D(r')$ (where $T_D$
294: is the effective disk temperature at $r'$).
295: For $r'=20$ this yields $T_D=20, 000$~K  and 
296: $F_X=2\times 10^{13}~{\rm erg~s^{-1}~cm^{-2}~s^{-1}}$.
297: For the density of $10^{-13}~{\rm g~cm^{-3}}$, typical for the region
298: above the disk at $z'\approx 4$, $\xi=5000$. 
299: This high $\xi$ implies that at this location, the gas would be 
300: fully ionized. However, for a few of reasons, 
301: this does not mean that the energy released above the disk
302: must suppress formation of the failed wind.
303: 
304: Firstly, the failed wind is opaque as illustrated in the middle panel 
305: of Fig.~1 which plots an estimate of the electron scattering optical depth
306: as a function of position over the unit length scale, 
307: $\tau\equiv \rho r_\ast \sigma_e$. 
308: Thus the size of a region fully ionized by 
309: locally dissipated energy will be small compared
310: to the size of the failed wind. In other words, the failed wind
311: can likely shield itself from locally produced X-rays as it does
312: from external X-rays. 
313: Secondly, as we eluded to above,
314: the presence of dense and cold gas above the disk creates unfavorable
315: conditions for transport and liberation of the energy 
316: above the disk in the first place. 
317: Contrary to the standard disk corona
318: scenario where the magnetized bubble carrying energy expands almost
319: freely once outside the disk, here
320: the bubble must rise through an extended, high density, dynamic region.
321: Generally, in a case of a disk with a failed disk wind,
322: the strength of a magnetic field needed to dominate over
323: the gas  and radiation pressure 
324: outside the disk must be orders of magnitude higher than in a case
325: of a bare disk.
326: Finally, the critical assumption about the X-ray production
327: is the fact that the hot gas is  tenuous
328: so that the main cooling mechanism is inverse Compton emission.
329: However, for the gas density 
330: in the failed wind of $>10^{-14}~{\rm g~cm^{-3}}$ 
331: the bremsstrahlung losses will exceed the Compton losses (e.g., eq. 1 in
332: Haardt \& Maraschi 1993).  Therefore,
333: even if the hot electrons were produced
334: above the disk, they will cool very efficiently by bremsstrahlung
335: instead of inverse Compton emission and the coronal X-ray flux
336: would be suppressed. 
337: \section{Summary and Discussion}
338: 
339: In the standard accretion disk theory for AGN, strong X-ray radiation
340: is accounted for by allowing the gravitational energy to be dissipated 
341: above the disk. This energy dissipation is to lead to formation of a hot disk 
342: corona located immediately above the disk and  inside this 
343: corona, UV disk photons are to be up-scattered to X-rays.  
344: We point out here, that for a broad range of AGN properties, radiation 
345: pressure due to the very same UV disk photons can drive a flow from 
346: the disk into the corona.
347: 
348: LD as a mechanism producing disk winds has been studied extensively, but LD
349: as a mechanism changing the density distribution above and inside the disk 
350: has not been given its due.
351: Therefore, we comment mostly on some qualitative  effects
352: that would be very important and relevant to AGN and other accretion disk
353: systems such as X-ray binaries (XBs). In particular, we use some physical 
354: arguments, quote and present results from simulations of line-driven disk 
355: winds to explore possible coupling between the X-ray and UV production 
356: processes.
357: 
358: The base for our discussion are simulations of line-driven flows from 
359: accretion disks studied by PK. The simulations show that 
360: the flows are opaque and can shield themselves not only from external X-rays 
361: but also from X-rays produced locally (i.e., in the region just above 
362: the disk where the flows are driven into). These opaque flows can lead to 
363: a suppression of energy dissipation above the disk. If so, the corona and 
364: X-rays would have to be produced interior to the UV emitting part of the disk.
365: This means that the UV and X-ray emitting regions do not overlap
366: and argues against a two-phase accretion disk model (i.e.,
367: against the slab/sandwich type of geometries). If this is true then
368: we have an inverse overionization problem: UV driven flow suppresses
369: X-ray production. If however, line-driven flows were suppressed 
370: by locally emitted X-rays, then the AGN wind would have to be launched from 
371: the disk exterior to the UV disk where coronal activity is negligible. 
372: It is possible that the intermediate situation occurs where X-rays are 
373: produced above the disk and the disk material is launched off the disk 
374: at different location/time. 
375: Such a double activity above the disk would lead to formation of a two-phase 
376: corona where high temperature plasma bubbles coexist with cooler 
377: and denser clumps.
378: 
379: The wind quenching of the hot corona raises many questions about the geometry, 
380: dynamics and energetics of AGN. For example, if the X-rays are quenched by 
381: the failed wind, what happens to the energy pumped into the gas via 
382: magnetic fields? Could this energy contribute to the UV continuum? Will 
383: the resulted UV continuum drive a flow different than that predicted by 
384: current models? 
385: 
386: To address these and other issues, one would need to perform global 
387: simulations of a turbulent MHD accretion disk 
388: with line driven flows in the radiation-dominated regime.
389: We anticipate that this will be possible soon
390: as local simulations of these type without LD
391: have been already preformed (e.g., Turner 2004; Hirose et al. in preparation).
392: 
393: There are many observational implications of the fact that the radiation
394: driven wind or the  failed wind can quench the X-ray production.
395: For example, one would expect an inverse correlation between X-ray fluxes
396: and winds. Perhaps, the observed anti-correlation between
397: the relative strength of the soft X-ray flux 
398: and the C~IV absorption equivalent width for QSOs 
399: (e.g., Brandt et al. 2000)
400: is due to the wind quenching rather than obscuration. The wind quenching
401: can also play a role in solving the overionization problem in QSOs.
402: The expected inverse correlation between X-ray fluxes
403: and winds is consistent with the fact that
404: BALs are not observed in X-ray sources such as Seyfert galaxies.
405: 
406: The wind quenching can also affect the disk reflected spectrum, 
407: big blue bump and iron lines.
408: The observed spectra are complex convolutions of the primary  
409: and reprocessed photons because some fraction of the X-rays are always 
410: intercepted by the optically thick matter and reprocessed before escaping.
411: Thus, to understand those spectra, it is necessary to compute very carefully 
412: the effects of reprocessing as many physical processes play a role (e.g., 
413: Ross \& Fabian 1993; {\.Z}ycki et al. 1994; Nayakshin et al. 2000). 
414: Future calculations of X-ray spectra should include 
415: the effects of line-driven flows because the flows change the key parameters
416: determining the spectra, i.e., the temperature and optical
417: depth of the scattering electrons.
418: These effects have at their core a coupling between X-rays 
419: and UV photons and therefore can help to better constrain the AGN models.
420: 
421: We note that the X-ray/UV coupling is unlikely to operate in the solar corona.
422: For this reasons among others, solar and AGN coronal activities may 
423: significantly differ.
424: However, this coupling or its variant may operate 
425: in XBs. For AGN and XBs, the same 
426: two types  of disk/corona geometries are being considered 
427: (e.g., Zdziarski \& Gierli{\'n}ski 2004 and Fig. 14 there).
428: Except for smallest radii, the bound-free and 
429: line opacities and radiation flux
430: are likely high enough for radiation pressure to 
431: lift dense material off the disk in some spectral/luminosity states of XBs. 
432: Therefore, one can expect the sphere-disk geometry in these systems because
433: of quenching of the disk corona.
434: In fact, this geometry is inferred from analysis of the photon index of 
435: Comptonization spectrum and frequencies of quasi periodic oscillations 
436: observed in some XBs
437: (e.g., Titarchuk \& Fiorito 2004; Titarchuk \& Shaposhnikov 2005).
438: It appears then that we identified a physical mechanism that 
439: can determine a geometry of an accretion disk and corona in 
440: a wide range of accreting systems.
441: 
442: We thank T. Kallman, G. Richards, A. R{\'o}{\.z}a{\'n}ska,
443: A. Siemiginowska, and J.Stone for useful discussions. 
444: We acknowledge support from NASA under LTSA grant NAG5-11736
445: and support provided by NASA through grants  HST-AR-09947.01-A
446: and  HST-AR-10305.05-A from the Space Telescope Science Institute, 
447: which is operated by the Association of Universities for Research 
448: in Astronomy, Inc., under NASA contract NAS5-26555.
449: 
450: 
451: \clearpage
452: 
453: \begin{thebibliography}{}
454: 
455: \bibitem[]{435}  Arav N.,  Shlosman I., \&  Weymann. R. ed. 1997, in ASP Conf. 
456: Ser. 128,  Mass Ejection from Active Galactic Nuclei, (San Francisco: ASP)
457: 
458: \bibitem[]{438} Crenshaw, D.M., Kraemer, S.B., \& George, I.M. ed. 2002,
459: in ASP Conf. Ser. 255, Mass Outflow in Active Galactic Nuclei: 
460: New Perspectives, (San Francisco: ASP)
461: 
462: \bibitem[]{442} Balbus, S. A., \& Hawley, J. F. 1998, Rev. Mod. Phys., 70, 1 
463: 
464: \bibitem[]{444} Brandenburg, A., Nordlund, A., Stein, R. F., \& Torkelsson, U. 1995, ApJ, 446, 741
465: 
466: \bibitem[]{446} Brandt, W.N., Laor, A., \& Wills, B. J. 2000, ApJ, 528, 637
467: 
468: \bibitem[]{448} Field, G. B., \& Rogers, R. D. 1993, ApJ, 403, 94
469: 
470: \bibitem[]{450} Galeev, A. A., Rosner, R., \& Vaiana, G. S. 1979, ApJ, 229, 318
471: 
472: \bibitem[]{452} Haardt, F., \& Maraschi, L.,  1991, ApJ, 380, L51
473: 
474: \bibitem[]{454} Haardt, F., \& Maraschi, L.,  1993, ApJ, 413, 507
475: 
476: \bibitem[]{456} Miller, K., \& Stone, J.M. 2000, ApJ, 534, 398 
477: 
478: \bibitem[]{458} Murray, N., Chiang, J., Grossman, S. A., \& Voit, G. M. 1995, ApJ, 451, 498 
479: (MCGV)
480: 
481: \bibitem[]{461} Nayakshin, S., Kazanas, D, \& Kallman, T.R. 2000, ApJ, 537, 833
482: 
483: 
484: \bibitem[]{464} Proga, D., \& Kallman, T.R. 2004, ApJ, 616, 688 (PK)
485: 
486: \bibitem[]{466} Proga, D., Stone J.M., \& Drew J.E. 1998, MNRAS, 295, 595 (PSD)
487: 
488: \bibitem[]{765} Proga, D., Stone, J.M., \& Kallman, T.R. 2000, ApJ, 543, 686 
489: (PSK)
490: 
491: \bibitem[]{} Reynolds, C. S., \& Nowak, M. A. 2003, Phys. Rep., 377, 389
492: 
493: \bibitem[]{471} Ross, R. R. \& Fabian, A. C. 1993, MNRAS, 261, 74
494: 
495: \bibitem[]{473}  Shakura N.I., \& Sunyaev R.A. 1973 A\&A, 24, 337
496: 
497: \bibitem[Sobolewska et al.(2004)]{2004ApJ...608...80S} Sobolewska, M.~A., 
498: Siemiginowska, A., {\.Z}ycki, P.~T.\ 2004, \apj, 608, 80 
499: 
500: \bibitem[]{477} Stone, J. M., Hawley, J. F., Gammie, C. F., \& Balbus, S. A. 
501: 1996, ApJ, 463, 656 
502: 
503: \bibitem[]{480} Sunyaev, R., \& Titarchuck, L.G. 1980, A\&A, 86, 121
504: 
505: \bibitem[]{482} Svensson, R., \& Zdziarski, A.A. 1994, ApJ 436, 599
506: 
507: \bibitem[]{} Titarchuk, L. G., \& Fiorito, R. 2004, ApJ, 612, 988
508: 
509: \bibitem[]{} 	Titarchuk, L. \& Shaposhnikov, N. 2005, ApJ, 626, 298
510: 
511: \bibitem[]{484} Turner, N.J. 2004, ApJ, 605, L45
512: 
513: \bibitem[]{486} Turnshek, D. A. 1988, in QSO Absorption Lines: 
514: Probing the Universe, ed. J. C. Blades, D. A. Turnshek, \& C. A. Norman
515: (Cambridge: Cambridge Univ. Press), 17
516: 
517: \bibitem[]{490} Walter, R., \& Courvoisier, T. J.-L. 1992, A\&A, 2389, 157
518: 
519: \bibitem[]{} Zdziarski, A. A., \& Gierli{\'n}ski, M. 2004, Prog. Theor. Phys. Suppl., 155, 99
520: 
521: \bibitem[]{492} {\.Z}ycki, P. T., Krolik, J. H., Zdziarski, A. A., \& Kallman, T. R. 1994, ApJ, 437, 597
522: \end{thebibliography}
523: 
524: \clearpage
525: 
526: \begin{figure}
527: \begin{picture}(180,580)
528: \put(180,410){\special{psfile=f1a.ps angle =90
529: hoffset=130 voffset=-15 hscale=32 vscale=32}}
530: \put(180,210){\special{psfile=f1b.ps angle =90
531: hoffset=130 voffset=-15 hscale=32 vscale=32}}
532: \put(180,10){\special{psfile=f1c.ps angle =90
533: hoffset=130 voffset=-15 hscale=32 vscale=32}}
534: 
535: \end{picture}
536: \caption{\tiny
537: {\it From top to bottom:}
538: Maps of logarithmic density (top panel), optical depth
539: (middle panel), and gas temperature
540: (bottom panel) of the AGN failed disk wind, described in the text. 
541: The density and temperature maps are overplotted with
542: the direction of the poloidal velocity field. 
543: In making this figure, 
544: we used the density and optical depth floors of $10^{-20}~{\rm g~cm^{-3}}$
545: and $10^{-2}$, respectively. 
546: In all three panels, the disk rotational axis is along the left hand vertical
547: frame, while the disk photosphere is along the lower horizontal frame.
548: Note the difference in the range along
549: the $r'$ and $'z'$ axises.}
550: \end{figure}
551: 
552: \end{document}
553: 
554: