1: %\documentclass[12pt,preprint]{aastex}

2: \documentclass[english]{emulateapj}

3: 

4: %\usepackage[T1]{fontenc}

5: %\setcounter{tocdepth}{3}

6: \usepackage{graphicx}

7: \usepackage{amssymb}

8: \usepackage{natbib}

9: \usepackage{amsmath}

10: %\usepackage{babel}

11: 

12: 

13: %= Used defined commands==================================

14: \makeatletter

15: %% Because html converters don't know tabularnewline

16: \providecommand{\tabularnewline}{\\}

17: %% A simple dot to overcome graphicx limitations

18: \newcommand{\lyxdot}{.}

19: 

20: 

21: %% Bold symbol macro for standard LaTeX 

22: \providecommand{\boldsymbol}[1]{\mbox{\boldmath $#1$}}

23: \newcommand{\kmps}{\mathrm{km~s^{-1}}}

24: \newcommand{\Kelvin}{\mathrm{K}}

25: \newcommand{\Msun}{\mathrm{M_{\sun}}}

26: \newcommand{\Rsun}{\mathrm{R_{\sun}}}

27: \newcommand{\Lsun}{L_{\sun}}

28: \newcommand{\MsunPerYear}{\mathrm{M_{\sun}\,yr^{-1}}}

29: \newcommand{\micon}{\mu m}

30: 

31: \makeatother

32: %=========================================================

33: 

34: \slugcomment{}

35: \shorttitle{}

36: \shortauthors{3-D simulations of Precessing Outflows from AGN}

37: 

38: 

39: 

40: \begin{document}

41: 

42: \title{Three-Dimensional Simulations of Inflows Irradiated by a Precessing

43: Accretion Disk in Active Galactic Nuclei: Formation of Outflows}

44: 

45: 

46: \author{Ryuichi Kurosawa and Daniel Proga}

47: 

48: 

49: \affil{Department of Physics and Astronomy, University of Nevada Las Vegas,

50: Box~454002, 4505~Maryland Pkwy, Las Vegas, NV 891541-4002}

51: 

52: 

53: \email{\{rk,dproga\}@physics.unlv.edu}

54: 

55: \begin{abstract}

56: We present three-dimensional (3-D) hydrodynamical simulations of gas

57: flows in the vicinity of an active galactic nucleus (AGN) powered

58: by a precessing accretion disk. We consider the effects of the radiation

59: force from such a disk on its environment on a relatively large scale

60: (up to $\sim10$~pc). We implicitly include the precessing disk by

61: forcing the disk radiation field to precess around a symmetry axis

62: with a given period ($P$) and a tilt angle ($\Theta$). We study time

63: evolution of the flows irradiated by the disk, and investigate basic

64: dependencies of the flow morphology, mass flux,

65: angular momentum on different combinations of $\Theta$ and $P$. As

66: this is our first attempt to model such 3-D gas flows, we consider

67: a simplest form of radiation force i.e., force due to electron scattering,

68: and neglect the forces due to line and dust scattering/absorption.

69: Further, the gas is assumed to be nearly isothermal. 

70: We find the gas flow settles into a configuration with two components,

71: (1) an equatorial inflow and (2) a bipolar inflow/outflow with the

72: outflow leaving the system along the poles (the directions of disk

73: normals). However, the flow does not always reach a steady state.

74: We find that

75: the maximum outflow velocity and the kinetic outflow power at the

76: outer boundary can be reduced significantly with increasing $\Theta$.

77: We also find that of the mass inflow rate across the inner boundary

78: does not change significantly with increasing $\Theta$. The amount

79: of the density-weighted mean specific angular momentum deposited to

80: the environment by the precessing disk increases as $P$ approaches

81: to the gas free-fall time ($t_{\mathrm{ff}}$), and then decreases

82: as $P$ becomes much larger than $t_{\mathrm{ff}}$. Generally, the

83: characteristics of the flows are closely related to a combination

84: of $P$ and $\Theta$, but not to $P$ and $\Theta$ individually. Our

85: models exhibit helical structures in the weakly collimated outflows.

86: Although on different scales, the model reproduces the Z- or S- shaped

87: density morphology of gas outflows which are often seen in radio observations

88: of AGNs.

89: \end{abstract}

90: 

91: \keywords{accretion, accretion -- disks -- galaxies: jets -- galaxies: kinematics

92: and dynamics-- methods: numerical -- hydrodynamics }

93: 

94: 

95: \section{Introduction}

96: 

97: \label{sec:Introduction}

98: 

99: Powered by accretion of matter onto a super massive ($10^{6}$--$10^{10}\,\Msun$)

100: black hole (SMBH), Active Galactic Nuclei (AGN) release large amount

101: of energy (e.g., \citealt{Lynden-Bell:1969}) as electromagnetic radiation

102: ($10^{10}$--$10^{14}\Lsun$) over a wide range of wavelengths, from

103: the X-ray to the radio. The very central location of AGN in their

104: host galaxies indicates that the radiation from AGN can play an important

105: role in determining the physical characteristics (e.g., the ionization

106: structure, the gas dynamics, and the density distribution) of their

107: surrounding environment in different scales i.e., from the scale of

108: AGN itself to a lager galactic scale, and even to an inter-galactic

109: scale (e.g., \citealt{Quilis:2001}; \citealt{DallaVecchia:2004};

110: \citealt{McNamara:2005}; \citealt{Zanni:2005}; \citealt{Fabian:2006};

111: \citealt{Vernaleo:2006}). The feedback process of AGN in the form

112: of mass or energy outflows, in turn, is one of key elements in

113: galaxy formation/evolutionary models (e.g., \citealt{ciotti:1997},

114: \citeyear{ciotti:2001}, \citeyear{ciotti:2007}; \citealt{Silk:1998}; \citealt{king:2003};

115: \citealt{Begelman:2005}; \citealt{Hopkins:2005}; \citealt{Murray:2005}; \citealt{Sazonov:2005};

116: \citealt{Silk:2005};

117: \citealt{Springel:2005}; \citealt{Brighenti:2006}; \citealt{Fabian:2006b}; \citealt{Fontanot:2006};

118: \citealt{Thacker:2006}; \citealt{Wang:2006}, \citealt{Tremonti:2007}). 

119: 

120: Although the AGN outflows can be driven by magnetocentrifugal force

121: (e.g., \citealt{Blandford:1982}; \citealt{Emmering:1992};

122: \citealt{Konigl:1994}; \citealt{Bottorff:1997}) and thermal pressure

123: (e.g., \citealt{Weymann:1982}; \citealt{Begelman:1991};

124: \citealt{Everett:2007}), it is the radiation force from the luminous

125: accretion disk that is most likely the dominant force driving winds

126: capable of explaining the blueshifted absorption line features often

127: seen in the UV and optical spectra of AGN

128: (e.g.,~\citealt{Shlosman:1985}; \citealt{Murray:1995b}; \citealt{Proga:2000};

129: \citealt{Proga:2004}). In reality, these three forces may interplay and

130: contribute to the dynamics of the outflows in AGN in somewhat

131: different degrees.

132: 

133: Another complication in the outflow gas dynamics is the presence of

134: dust. The radiation pressure on dust can drive dust outflows,

135: and their dynamics is likely to be coupled with the gas dynamics (e.g.,

136: \citealt{Phinney:1989}; \citealt{Pier:1992}; \citealt{Emmering:1992};

137: \citealt{Laor:1993}; \citealt{Konigl:1994}; \citealt{Murray:2005}).

138: The AGN environment on relatively large scales ($10^{2}-10^{3}$~pc)

139: is known to be a mixture of gas and dust (e.g.~\citealt{Antonucci:1984};

140: \citealt{Miller:1990}; \citealt{Awaki:1991}; \citealt{Blanco:1990};

141: \citealt{Krolik:1999}); however, in much smaller scales ($<\sim10$~pc)

142: one does not expect much dust to be present because the temperature

143: of the environment is high ($>10^{4}\,\Kelvin$). Concentrating on

144: only the gas component, the dynamics of the outflows in smaller scales

145: was studied by e.g..~\citet{Arav:1994}, \citet{Proga:2000} in

146: 1-D and 2-D, respectively. 

147: 

148: Radio observations show that a significant fraction of extended extragalactic

149: sources display bending or twisting jets from their host galaxies.

150: For example, \citet{Florido:1990} found that $\sim11$\% of their

151: sample (368 objects) show anti-symmetrically bending jets (S-shaped

152: or Z-shaped morphology) while $\sim9$\% show the symmetrically bending

153: jets (U-shaped morphology). Similarly \citet{Hutchings:1988} studied

154: the morphology of the radio lobes from 128 quasars (with $z<1$),

155: and found that $30$\% of the sample show a sign of bending jets.

156: The bending and misalignment of jets are also observed in parsec scales

157: in compact radio sources (e.g.~\citealt{Linfield:1981}; \citealt{Appl:1996};

158: \citealt{Zensus:1997}). Examples of the radio maps displaying the

159: S- or Z-shaped morphology of jets can be found in e.g.~\citet{Condon:1984},

160: \citet{Hunstead:1984}, and \citet{Tremblay:2006}. 

161: 

162: Using the data available in literature, \citet{Lu:2005} compiled

163: the list (see their Tab.~1) of 41 known extragalactic radio sources

164: which show an evidence of jet precession, along with their jet precession

165: periods ($P$) and the half-opening angle ($\psi$) of jet precession

166: cones. According to this list, a large fraction (67\%) of system has

167: rather small half-opening angles, i.e., $\psi<\sim15^{\circ}$. A

168: large scatter in the precession periods are found in their sample;

169: however, most of the precession periods are found in between $10^{4}$

170: and $10^{6}$~yr (see also \citealt{Roos:1988}). Note that the precession

171: periods are usually too long to be determined directly by variability

172: observations. Typically the precession periods are found by fitting

173: the radio map with a kinematic jet model (e.g., \citealt{Gower:1982};

174: \citealt{Veilleux:1993}). Interestingly, \citet{Appl:1996} showed

175: that a typical precession period of tilted massive torus around SMBH

176: is $\sim10^{6}$~yr.

177: 

178: The S- and Z-shaped morphology seen in the observations mentioned

179: above can naturally explained by precessing jets. Further, the precessing

180: of jets can occur if the underlying accretion disk is tilted

181: (or warped) with respect to the symmetry plane. There are at least

182: five known mechanisms that can causes warping and precession of in

183: accretion disks (1)~the Bardeen-Petterson effect (\citealt{bardeen:1975};

184: see also \citealt{Schreier:1972}; \citealt{Nelson:2000}; \citealt{Fragile:2005};

185: \citealt{King:2005}), (2)~tidal interactions in binary BH system

186: (e.g., \citealt{Roos:1988}; \citealt{Sillanpaa:1988}; \citealt{katz:1997};

187: \citealt{Romero:2000}; \citealt{Caproni:2004}), (3)~radiation-driven

188: instability (e.g., \citealt{Petterson:1977}; \citealt{Pringle:1996};

189: \citealt{Maloney:1996}; \citealt{Armitage:1997}), (4)~magnetically-driven

190: instability (\citealt{Aly:1980}; \citealt{Lai:2003}), and (5)~Disk-ISM

191: interaction (e.g., \citealt{Quillen:1999}). Using a small sample

192: of AGN, \citet{Caproni:2006b} examined whether mechanisms (1)--(4)

193: are capable of explaining the observed precession periods. Similarly

194: \citet{Tremblay:2006} searched for a possible cause of disk precession

195: and warping of the FR~I radio source 3C~449 using mechanisms~(2),

196: (3) and (4) above. In general, it is very difficult to determine the exact cause of disk/jet

197: precession for a given AGN system because of large uncertainties in model

198: parameters and observed precession periods (which are also often model

199: dependent). 

200: 

201: \citet{Kochanek:1990} presented a hydrodynamical simulation of jet

202: propagation along the surface of an axisymmetric hollow/cone to approximate

203: a jet with fast precession; however, intrinsically non-axisymmetric

204: nature of the dynamics of jet precession requires the problem to be

205: solved/simulated in 3-D. Hydrodynamical simulations of extragalactic

206: radio sources with precessing jets in full 3-D have been performed

207: by e.g., \citet{Cox:1991}, \citet{Hardee:1992}, \citet{Hardee:1994},

208: Typically, in these models, the jets are driven at the origin by a

209: small-amplitude precession to break the symmetry and excite helical

210: modes of the Kelvin-Helmholtz instability. Careful stability/instability

211: analysis of such simulations has been presented by \citet{Hardee:1995}.

212: The effect of magnetic field has been also investigated by e.g., \citet{Hardee:1995b}

213: while the effect of optically thin radiative cooling on the Kelvin-Helmholtz

214: instability has been investigated by e.g., \citet{Xu:2000}. 

215: Precession of relativistic jets in 3-D with or without magnetic field

216: has been also studied (e.g., \citealt{Hardee:2001};

217: \citealt{Hughes:2002}; \citealt{Aloy:2003}; \citealt{Mizuno:2007}).

218: On much larger scales, \citet{Sternberg:2007} studied the

219: effect of precessing massive slow jets onto the intergalactic medium

220: (IGM) in a galaxy cluster, and found such jets can inflate a fat bubble

221: in the IGM. 

222: In the models mentioned above, jets themselves are injected on small

223: scales, and the jet propagations are studied. However, it is also

224: possible to model a self-consistent production of a jet and its

225: subsequent propagation. For example, a jet can be produced from an

226: infalling matter by radiation pressure due to

227: a luminous accretion disk (e.g., see \citealt{Proga:2007};

228: \citealt{Proga:2007b}, for axisymmetric cases).

229: 

230: Regardless of the exact cause of disk/jet precession,

231: the observations (e.g.~\citealt{Florido:1990}; \citealt{Hutchings:1988})

232: suggest that a significant fraction of AGN contain warped or precessing

233: disks. One might expect the details of the radiative feedback

234: processes in such systems are different from the ones predicted by

235: axi-symmetry models (e.g.~\citealt{Proga:2000}; \citealt{Proga:2007};

236: \citealt{Proga:2007b}). If they differ, then by 

237: how much? In this paper, we explore the effects of disk precession

238: on the gas dynamics in the AGN environment by simulating the outflows

239: driven by the radiation force from a luminous precessing accretion

240: disk around a SMBH. Specifically, we will examine how the mass-accretion

241: rate, the outflow powers (kinetic and thermal), the morphology of

242: the flows, and the specific angular momentum of the gas are affected

243: by the presence of a precessing disk and its radiation field. This

244: is our first step toward a full extension of the axisymmetric radiation-driven

245: wind model of \citet{Proga:2007} to a full 3-D model.

246: 

247: In the following section, we describe our method and model assumptions,

248: and we give the results of our 3-D hydrodynamical simulations in \S~\ref{sec:Results}.

249: Our conclusions are summarized in \S~\ref{sec:Conclusions}.

250: 

251: 

252: \section{Method}

253: 

254: Our basic model configuration is shown in Figure~\ref{fig:model-config}.

255: The model geometry and the assumptions of the SMBH and the disk are

256: very similar to those in \citet{Proga:2007}. In

257: Figure~\ref{fig:model-config}, a SMBH with its mass

258: $M_{\mathrm{BH}}$ and its Schwarzschild radius $r_{\mathrm{S}}=2GM_{\mathrm{BH}}/c^{2}$

259: is placed at the center of the cartesian coordinate system ($x$,

260: $y$, $z$). The X-ray emitting corona regions is defined as a sphere

261: with its radius $r_{*}$, as shown in the figure. The

262: geometrically-thin and optically-thick flat accretion disk (e.g.,

263: \citealt{shakura:1973})

264: is placed \emph{near} the $x$-$y$ plane.

265: In case for an axisymmetric model, the $z$-axis in the figure becomes

266: the symmetry axis, and the accretion disk is on the $x$-$y$ plane.

267: To simulate the disk precession, we assume that the angular momentum

268: ($\boldsymbol{J}_{\mathrm{D}}$) of the accretion disk is tilted from

269: the $z$-axis by an angle $\Theta$. In other words,

270: the accretion disk is assumed to be tilted by $\Theta$ from the

271: $x$-$y$ plane. Further, the accretion disk hence its 

272: angular momentum $\boldsymbol{J}_{\mathrm{D}}$ is assumed to precess

273: around the $z$-axis with the precession period $P$. The 3-D hydrodynamic

274: simulations will be performed in the spherical coordinate system ($r$,

275: $\theta$, $\phi$), and in between the inner boundary $r_{\mathrm{i}}$

276: and the outer boundary $r_{\mathrm{o}}$. The poles of the spherical

277: coordinate system coincides with the $z$-axis. 

278: The radiation forces, from

279: the corona region (the sphere with its radius $r_{*}$) and the accretion

280: disk, acting on the gas located at a location ($p$) in the field

281: are assumed to be only in radial direction. The magnitude of the radiation

282: force due to the corona is assumed to be a function of radius $r$,

283: but that due to the accretion disk is assumed to be a function of

284: $r$ and the angle ($\theta'$) between the disk angular momentum

285: $\boldsymbol{J}_{\mathrm{D}}$ and the position vector $\boldsymbol{r}$

286: as shown in the figure. The point-source like approximation for the

287: disk radiation pressure at $P$ is valid when $r_{\mathrm{D}}\ll r_{\mathrm{i}}$.

288: In the following, we will describe our radiation hydrodynamics, our

289: implementation of the continuum radiation sources (the corona and

290: disk), the model parameters and assumptions. 

291: 

292: %====================================

293: 

294: 

295: 

296: %

297: \begin{figure*}

298: \begin{center}

299: 

300: \includegraphics[width=0.95\textwidth]{f1.eps}

301: 

302: \end{center}

303: 

304: 

305: \caption{Basic model configuration. A super massive blackhole (BH) with its

306: Schwarzschild radius $r_{S}$ is located at the center of the cartesian

307: coordinate system ($x$, $y$, $z$) where the $y$-axis is perpendicular

308: to and into the page. The normal vector or the angular momentum ($\boldsymbol{J}_{\mathrm{D}}$)

309: of the accretion disk, spanning from its inner radius $r_{*}$ to

310: its outer radius $r_{D}$, is misaligned with the $z$-axis by a tilt

311: angle $\Theta$ i.e., the accretion disk is tilted by $\Theta$ from

312: the $x$-$y$ plane. The accretion disk hence

313: its angular momentum $\boldsymbol{J}_{\mathrm{D}}$ is assumed to

314: precess around the $z$-axis with the precession period $P$. The

315: 3-D hydrodynamic simulations are performed in the spherical coordinate

316: system ($r$, $\theta$, $\phi$). The simulations are performed in the

317: radial range between the inner boundary

318: $r_{\mathrm{i}}$ and the outer boundary $r_{\mathrm{o}}$. The

319: radiation pressure from the central BH on a point $p$ with its 

320: position vector $\boldsymbol{r}$ is in radial direction, and is a function

321: of $r$. Whereas the radiation pressure from the accretion disk is assumed to be a function

322: of $r$ and $\theta'$ where the latter is the angle between $\boldsymbol{J}_{\mathrm{D}}$

323: and $\boldsymbol{r}$ (see Secs.~\ref{sub:Continuum-Radiation-Source}

324: and \ref{sub:Precessing-Disk} for details). The point-source like

325: approximation for the disk radiation pressure at $P$ is valid when

326: $r_{\mathrm{D}}\ll r_{\mathrm{i}}$. Note that the figure is not to

327: scale. }

328: 

329: \label{fig:model-config}

330: \end{figure*}

331: 

332: 

333: %====================================

334: 

335: 

336: \subsection{Hydrodynamics}

337: 

338: \label{sub:Hydrodynamics}

339: 

340: We employ 3-D hydrodynamical simulations of the outflow from and accretion

341: onto a central part of AGN, using the ZEUS-MP code \citep[c.f.,][]{Hayes:2006}

342: which is a massively parallel MPI-implemented version of the ZEUS-3D

343: code (c.f., \citealt{Hardee:1992}; \citealt{Clarke:1996}). The ZEUS-MP

344: is a Eulerian hydrodynamics code which uses the method of finite differencing

345: on a staggered mesh with a second-order-accurate, monotonic advection

346: scheme \citep{Hayes:2006}. To compute the structure and evolution

347: of a flow irradiated by a strong continuum radiation of AGN, we solve

348: the following set of HD equations: 

349: \begin{eqnarray}

350:   \frac{D\rho}{Dt}+\rho\,\boldsymbol{\nabla}\cdot\boldsymbol{v} 

351:   & = & 0, 

352:   \label{eq:hydro01}

353: \end{eqnarray}

354: 

355: \begin{equation}

356:   \rho\frac{D\boldsymbol{v}}{Dt}=-\boldsymbol{\nabla}P+\rho\,\boldsymbol{g}+\rho\,\boldsymbol{g}_{\mathrm{rad}},

357: \label{eq:hydro02}

358: \end{equation}

359: 

360: \begin{equation}

361:   \rho\frac{D}{Dt}\left(\frac{u}{\rho}\right)=-P\,\boldsymbol{\nabla}\cdot\boldsymbol{v}+\rho\,\mathcal{C},

362:   \label{eq:hydro03}

363: \end{equation}

364:  where $\rho$, $u$, $P$ and $\boldsymbol{v}$ are the mass density,

365: energy density, pressure, and the velocity of gas respectively. Also,

366: $\boldsymbol{g}$ is the gravitational force per unit mass. The Lagrangian/co-moving

367: derivative is defined as $D/Dt\equiv\partial/\partial t+\boldsymbol{v}\cdot\boldsymbol{\nabla}$.

368: We have introduced two new components to the ZEUS-MP in order to treat

369: the gas dynamics more appropriate for the gas flow in and around AGN.

370: The first is the acceleration due to radiative force per unit mass

371: ($\boldsymbol{g}_{\mathrm{rad}}$) in equation~(\ref{eq:hydro02}),

372: and the second is the the effect of radiative cooling (and heating)

373: simply as the net cooling rate ($\mathcal{C}$) in equation~(\ref{eq:hydro03}).

374: As this is our first 3-D simulations with this code, we consider a

375: simplest case i.e., $\mathcal{C}=0$ , but $g_{\mathrm{rad}}\neq0$.

376: We also use $\gamma=1.01$ in the equation of state $P=\left(\gamma-1\right)u$

377: where $\gamma$ is the adiabatic index. In the following, our implementation

378: of $\boldsymbol{g}_{\mathrm{rad}}$ will be described. 

379: 

380: 

381: \subsubsection{Radiation Force}

382: 

383: To evaluate the radiative acceleration due to line absorption/scattering,

384: we follow the method in \citet{Proga:2000} who applied the modified

385: Castor, Abbott \& Klein (CAK) approximation \citep{Castor:1975}.

386: Their model works under the assumption of the Sobolev approximation

387: (e.g., \citealt{Sobolev:1957}; \citealt{Castor:1970}; \citealt{Lucy:1971});

388: hence, the following conditions are assumed to be valid: (1)~presence

389: of large velocity gradient in the gas flow, and (2)~the intrinsic

390: line width is negligible compared to the Doppler broadening of a line.

391: Following \citet{Proga:2000}, the radiative acceleration of a unit

392: mass at a point $\boldsymbol{r}$ can be written as 

393: \begin{equation}

394:   \boldsymbol{g}_{\mathrm{rad}}=\oint_{\Omega}\left[1+\mathcal{M}\right]\left[\frac{\sigma_{e}I\left(\boldsymbol{r},\hat{\boldsymbol{n}}\right)}{c}\right]\hat{\boldsymbol{n}}\,

395:  d\Omega 

396: \label{eq:rad_force}

397: \end{equation}

398:  where $I$ is the frequency-integrated continuum intensity in the

399: direction $\hat{\boldsymbol{n}}$, and $\Omega$ is the solid angle

400: subtended by the source of continuum radiation. Also, $\sigma_{e}$

401: is the electron scattering cross section. The force multiplier $\mathcal{M}$

402: is a function of optical depth parameter $\tau$ which is similar

403: to the Sobolev optical depth (c.f.~\citealt{Rybicki:1978}), and

404: can be written as \begin{equation}

405: \tau=\frac{\sigma_{e}\,\rho\, v_{\mathrm{th}}}{\left|Q\right|}\label{eq:tau_parameter}\end{equation}

406: where $Q=dv_{l}/dl$ is the directional derivative of the velocity

407: field in direction $\hat{\boldsymbol{n}}$, $dl$ is the line element

408: in the same direction, and $v_{\mathrm{th}}$ is the thermal velocity

409: of the gas. Further equation~(\ref{eq:rad_force}) can be simplified

410: greatly when the continuum radiation source is approximated as a point,

411: i.e.,~when $r\gg r_{\mathrm{c}}$ where $r_{\mathrm{c}}$ is the

412: radius of the radiation source. In our case, we consider the accretion

413: disk which emits most of the radiation from the innermost part, between

414: $r_{*}$ and $r_{\mathrm{D}}$ in Figure~\ref{fig:model-config};

415: hence, the condition $r\gg r_{\mathrm{c}}$ is satisfied. Using this

416: approximation, the radiative acceleration $\boldsymbol{g}_{\mathrm{rad}}$

417: will be radial only, and be a function of radial position and polar

418: angle (if the contribution from the disk luminosity is included),

419: i.e. $\boldsymbol{g}_{\mathrm{rad}}=g_{\mathrm{rad}}\left(r,\theta\right)\hat{\boldsymbol{r}}$.

420: This simplification is very useful for our purposes as it reduces

421: the computational time significantly hence it enables us to perform

422: large-scale 3-D simulations. Unlike \citet{Proga:2000}, we consider

423: the case in which the radiative acceleration is dominated by the continuum

424: process, i.e. $\mathcal{M}=0$ in equation~(\ref{eq:rad_force})

425: in this paper since we initially intend to investigate the basic characteristics

426: of the impact of the disk precession that do not depend on the details

427: of the radiation force model. The models with $\mathcal{M}\neq0$

428: in equation~(\ref{eq:rad_force}) and $\mathcal{C}\neq0$ will be

429: presented in a forthcoming paper. 

430: 

431: 

432: \subsection{Continuum Radiation Source }

433: 

434: \label{sub:Continuum-Radiation-Source}

435: 

436: As mentioned earlier, we consider two different continuum radiation

437: sources in our models: (1)~the accretion disk, and (2)~the central

438: spherical corona. Since the geometry of the central engine in AGN

439: is not well understood, we assume that is consist of a spherically

440: shaped corona with its radius $r_{*}$ and the innermost part of the

441: accretion disk (c.f.~Fig.~\ref{fig:model-config}). The disk is

442: assumed to be flat, Keplerian, geometrically-thin and optically thick.

443: The disk radiation is assumed to be dominated by the radiation from

444: the disk radius between $r_{*}$ and $r_{\mathrm{D}}$ where $r_{*}=3\, r_{\mathrm{S}}$

445: and $r_{*}<r_{\mathrm{D}}\ll r_{\mathrm{i}}$ (c.f., Fig.~\ref{fig:model-config}).

446: Note that the exact size of $r_{\mathrm{D}}$ does not matter as

447: long as it satisfies this condition in order for the point-source

448: approximation mentioned in \S.~\ref{sub:Hydrodynamics} to be valid. 

449: 

450: In terms of the disk mass-accretion rate ($\dot{M}_{\mathrm{D}}$),

451: the mass of the BH ($M_{\mathrm{BH}}$) and the Schwarzschild radius

452: ($r_{\mathrm{S}}$), the total luminosity ($L$) of the system can

453: be written as 

454: \begin{align}

455: L & =\eta\dot{M}_{\mathrm{D}}c^{2}\label{eq:total-luminosity}\\

456:   & =\frac{2\eta GM_{\mathrm{BH}}\dot{M}_{\mathrm{D}}}{r_{\mathrm{S}}}\label{eq:total-luminosity2}

457: \end{align}

458:  where $\eta$ is the rest mass conversion efficiency (e.g., \citealt{shakura:1973}).

459: Following \citet{Proga:2007} and \citet{Proga:2007b}, we simply

460: assume the system essentially radiates only in the UV and the X-ray

461: bands. The total luminosity of the system $L$ is then the sum of

462: the UV luminosity $L_{\mathrm{UV}}$ and the X-ray luminosity $L_{\mathrm{X}}$

463: i.e., $L=L_{\mathrm{UV}}+L_{\mathrm{X}}$. Further, we assume that

464: the disk only radiates in the UV and the central corona in the X-ray.

465: The ratio of the disk luminosity ($L_{\mathrm{D}}$) to the total

466: luminosity is parametrized as $f_{\mathrm{D}}=L_{\mathrm{D}}/L$,

467: and that of the corona luminosity ($L_{*}$) to the total luminosity

468: as $f_{*}=L_{*}/L$. Consequently, $f_{\mathrm{D}}+f_{*}=1$.

469: 

470: In the point-source approximation limit, the radiation flux from the

471: central X-ray corona region can be written as 

472: \begin{equation}

473:   \mathcal{F}_{*}=\frac{L_{*}}{4\pi r^{2}}

474:   \label{eq:corona-flux}

475: \end{equation}

476:  where $r$ is the radial distance from the center (by neglecting

477: the source size). Here we neglect the geometrical obscuration of the

478: corona emission by the accretion disk and vice versa. On the other

479: hand, the disk radiation depends on the polar angle $\theta$ because

480: of the source geometry. Again following \citet{Proga:2007} and

481: \citet{Proga:2007b} (see also \citealt{proga:1998}), the disk intensity

482: $I_{\mathrm{D}}$ is assumed to be radial and $I_{\mathrm{D}}\propto\left|\cos\theta'\right|$.

483: This follows that the disk radiation flux at the distance $r$ from

484: the center can be written as 

485: \begin{equation}

486: \mathcal{F}_{\mathrm{D}}=2\,\left|\cos\theta'\right|\,\frac{L_{\mathrm{D}}}{4\pi

487:   r^{2}}

488: \label{eq:disk-flux}

489: \end{equation}

490:  where $\theta'$ is the angle between the disk normal and the position

491: vector $\boldsymbol{r}$ (c.f., Fig.~\ref{fig:model-config}). The leading term $2$ in this expression

492: comes from the normalization of the polar angle dependency. Finally

493: by using eqs.~(\ref{eq:rad_force}), (\ref{eq:corona-flux}) and (\ref{eq:disk-flux}),

494: the radiative acceleration term in equation~(\ref{eq:hydro03}) can

495: be written as 

496: \begin{equation}

497:    \boldsymbol{g}_{\mathrm{rad}}=\frac{\sigma_{e}L}{4\pi r^{2}c}\left\{

498: f_{*}+2\,\left|\cos\theta'\right|\, f_{\mathrm{D}}\right\}

499: \,\boldsymbol{\hat{r}}\, .

500: \label{eq:rad-force-final}

501: \end{equation}

502: 

503: 

504: 

505: \subsection{Precessing Disk}

506: 

507: \label{sub:Precessing-Disk}

508: 

509: As we noted before, here we do not model the precession of the accretion

510: disk itself, but rather manually force the precession. We do not specify

511: the cause of the precession either. We simply assume that the disk

512: precession exists, and investigate its consequence to the AGN environment.

513: The UV emitting portion of the disk spans from $r_{*}$ to $r_{\mathrm{D}}$

514: (c.f., Fig.~\ref{fig:model-config}). We assume that $r_{\mathrm{D}}\ll r_{\mathrm{i}}$

515: where the $r_{\mathrm{i}}$ is the inner radius of the computational

516: domain of the hydrodynamic simulations. This means that the disk itself

517: is not in the computational domain. The effect of the precessing disk

518: is included as precessing radiation field in the hydrodynamics of

519: the gas (through eq.~{[}\ref{eq:hydro02}]). 

520: 

521: We assume that the disk is tilted from the $x$-$y$ plane (in the

522: cartesian coordinate system) by $\Theta$ as in Figure~\ref{fig:model-config}.

523: Equivalently, the disk angular momentum $\boldsymbol{J}_{\mathrm{D}}$

524: (assuming a flat uniform Keplerian disk) deviates from the $z$-axis

525: by $\Theta$. Further, we assume that $\boldsymbol{J}_{\mathrm{D}}$

526: precesses around the $z$-axis with precession period $P$. With these

527: assumptions, the components of the $\boldsymbol{J}_{\mathrm{D}}$

528: in the cartesian coordinate system can be written as 

529: \begin{align}

530:   J_{\mathrm{D}x} & =J_{D}\sin\Theta\,\cos\left(\frac{2\pi t}{P}\right),\label{eq:J_D-x-component}\\

531:   J_{\mathrm{D}y} & =J_{D}\sin\Theta\,\sin\left(\frac{2\pi t}{P}\right),\label{eq:J_D-y-component}\\

532:   J_{\mathrm{D}z} & =J_{D}\cos\Theta\label{eq:J_D-z-component}

533: \end{align}

534: where $t$ is the time measured from the beginning of hydrodynamic

535: simulations. Here we set $\boldsymbol{J}_{\mathrm{D}}$ to be on the

536: $x-z$ plane (as shown in Fig.~\ref{fig:model-config}) at $t=0$.

537: By setting $\Theta=0$, the model reduced to an asymmetric case as

538: in \citet{Proga:2007}. To compute the radiative acceleration as expressed

539: in equation~(\ref{eq:rad-force-final}), one requires the angle between

540: $\boldsymbol{J}_{\mathrm{D}}$ and the position vector $\boldsymbol{r}$

541: at which the set of the HD equations (eqs.~{[}\ref{eq:hydro01}],

542: {[}\ref{eq:hydro02}] and {[}\ref{eq:hydro03}]) are solve. This can

543: be obtained simply by finding the inner product of $\boldsymbol{J}_{\mathrm{D}}$

544: and $\boldsymbol{r}$. 

545: 

546: 

547: \subsection{Model Setup}

548: 

549: \label{sub:Model-Setup}

550: 

551: In all models presented here, the following ranges of the coordinates

552: are adopted: $r_{\mathrm{i}}\leq r\leq r_{\mathrm{o}}$, $0\leq\theta\leq\pi$

553: and $0\leq\phi<2\pi$ where $r_{\mathrm{i}}=500\, r_{*}$ and $r_{\mathrm{o}}=2.5\times10^{5}\, r_{*}$.

554: The radius of the central and spherical X-ray corona region $r_{*}$

555: coincides with the inner radius of the the accretion disk (Fig.~\ref{fig:model-config}).

556: In our simulations, the polar and azimuthal angle ranges are divided

557: into 128 and 64 zones, and are equally spaced. In the $r$ direction,

558: the gird is divided into 128 zones in which the zone size ratio is

559: fixed at $\Delta r_{k+1}/\Delta r_{k}=1.04$. 

560: 

561: For the initial conditions, the density and the temperature of gas

562: are set uniformly i.e., $\rho=\rho_{o}$ and $T=T_{o}$ everywhere

563: in the computational domain where $\rho_{o}=1.0\times10^{-21}\,\mathrm{g\, cm^{-3}}$

564: and $T_{o}=2\times10^{7}\,\Kelvin$ through out this paper. The initial

565: velocity of the gas is simply set to zero everywhere. 

566: 

567: At the inner and outer boundaries, we apply the outflow (free-to-outflow)

568: boundary conditions, in which the field values are extrapolated beyond

569: the boundaries using the values of \emph{the ghost zones} residing

570: outside of normal computational zones (see \citealt{Stone:1992} for

571: more details). At the outer boundary, all HD quantities (except the

572: radial velocity) are fixed constant, to their initial values (e.g., $T=T_{o}$

573: and $\rho=\rho_{o}$), during the the evolution of each model. The

574: radial velocity components are allowed to float. \citet{Proga:2007} applied

575: these conditions to represent a steady flow condition at the outer

576: boundary. They found that this technique leads to a solution that

577: relaxes to a steady state in both spherical and non-spherical accretion

578: with an outflow (see also \citealt{Proga:2003b}). This imitates the

579: condition in which a continuous supply of gas is available at the

580: outer boundary. 

581: 

582: 

583: \section{Results}

584: 

585: \label{sec:Results}

586: 

587: 

588: \subsection{Reference Values}

589: 

590: \label{sub:Reference-Values}

591: 

592: %====================================

593: 

594: 

595: 

596: %

597: \begin{table*}

598: 

599: \caption{\label{tab:Model-Summary}Model Summary}

600: 

601: \scriptsize

602: 

603: \begin{center}

604: 

605: \begin{tabular}{cccccccccc}

606: \hline 

607: \hline&

608: $\Theta$&

609: $P$&

610: $\dot{M}_{\mathrm{in}}\left(r_{\mathrm{o}}\right)$&

611: $\dot{M}_{\mathrm{net}}\left(r_{\mathrm{i}}\right)$&

612: $\dot{M}_{\mathrm{out}}\left(r_{\mathrm{o}}\right)$&

613: $v_{r}^{\mathrm{max}}\left(r_{\mathrm{o}}\right)$&

614: $P_{k}\left(r_{\mathrm{o}}\right)$&

615: $P_{\mathrm{th}}\left(r_{o}\right)$&

616: $j_{\rho}$\tabularnewline

617: Model&

618: $\left(\,^{\circ}\right)$&

619: $\left(\mathrm{yr}\right)$&

620: $\left(10^{25}\mathrm{\, g\, s^{-1}}\right)$&

621: $\left(10^{25}\,\mathrm{g\, s^{-1}}\right)$&

622: $\left(10^{25}\,\mathrm{g\, s^{-1}}\right)$&

623: $\left(\mathrm{km\, s^{-1}}\right)$&

624: $\left(10^{40}\,\mathrm{erg\, s^{-1}}\right)$&

625: $\left(10^{40}\,\mathrm{erg\, s^{-1}}\right)$&

626: $\left(j_{0}\right)$\tabularnewline

627: \hline

628: I&

629: $0$&

630: $\infty$&

631: $-2.2$&

632: $-0.6$&

633: $1.6$&

634: $1500$&

635: $3.0$&

636: $260$&

637: $0$\tabularnewline

638: II&

639: $5$&

640: $1.6\times10^{4}$&

641: $-2.3$&

642: $-0.6$&

643: $1.7$&

644: $640$&

645: $1.4$&

646: $290$&

647: $0.2$\tabularnewline

648: III&

649: $15$&

650: $1.6\times10^{4}$&

651: $-2.2$&

652: $-0.4$&

653: $1.8$&

654: $620$&

655: $1.1$&

656: $310$&

657: $0.2$\tabularnewline

658: IV&

659: $5$&

660: $1.6\times10^{5}$&

661: $-2.2$&

662: $-0.6$&

663: $1.6$&

664: $1900$&

665: $3.0$&

666: $260$&

667: $0.05$\tabularnewline

668: \hline

669: \end{tabular}

670: 

671: \end{center}

672: 

673: 

674: \tablecomments{The model output values are averaged over the last $2\times10^{12}\,\mathrm{s}$

675: of the hydrodynamic simulations. }

676: 

677: \normalsize

678: \end{table*}

679: 

680: 

681: %====================================

682: 

683: We consider four different cases which have different combinations

684: of the disk tilt angle ($\Theta$) and the disk precession period ($P$),

685: as summarized in Table~\ref{tab:Model-Summary}. The following parameters

686: are common to all the models presented here, and are exactly the same

687: as in \citet{Proga:2007}. We assume that the central BH is non-rotating

688: and has mass $M_{\mathrm{BH}}=10^{8}\, M_{\odot}$. The size of the

689: disk inner radius is assumed to be $r_{*}=3r_{s}=8.8\times10^{13}\,\mathrm{cm}$

690: (c.f.~Sec.~\ref{sub:Model-Setup}). The mass accretion rate ($\dot{M}_{a}$)

691: onto the central SMBH and the rest mass conversion efficiency ($\eta$)

692: are assumed to be $1\times10^{26}\,\mathrm{g\, s^{-1}}$ and $0.0833$,

693: respectively. With these parameters, the corresponding accretion luminosity

694: of the system is $L=7.5\times10^{45}\,\mathrm{erg\, s^{-1}=2\times10^{12}\,\Lsun}$.

695: Equivalently, the system has the Eddington number $\Gamma=0.6$ where

696: $\Gamma\equiv L/L_{\mathrm{Edd}}$ and $L_{\mathrm{Edd}}$ is the

697: Eddington luminosity of the Schwarzschild BH i.e., $4\pi cGM_{\mathrm{BH}}/\sigma_{e}$.

698: The fractions of the luminosity in the UV ($f_{\mathrm{UV}}$) and

699: that in the X-ray ($f_{\mathrm{X}}$) are fixed at $0.95$ and $0.05$

700: respectively, as in \citet{Proga:2007} (see their Run~C). 

701: 

702: Important reference physical quantities relevant to our systems are

703: as follows. The Compton radius, $R_{C}\equiv GM_{\mathrm{BH}}\mu\, m_{p}/kT_{C}$,

704: is $8\times10^{18}\,\mathrm{cm}$ or equivalently $9\times10^{4}\, r_{*}$

705: where $T_{C}$, $\mu$ and $m_{p}$ are the Compton temperature, the

706: mean molecular weight of gas and the proton mass, respectively. Here

707: we assume that the gas temperature at infinity is $T_{\infty}=T_{C}=2\times10^{7}\,\Kelvin$;

708: hence, the corresponding speed of sound at infinity is $c_{\infty}=(\gamma kT_{C}/\mu m_{p})^{1/2}=4\times10^{7}\,\mathrm{cm\, s^{-1}}$

709: where $\gamma$ is the adiabatic index. In this paper, $\gamma=1.01$

710: (almost isothermal) is adopted to imitate a gas in Compton equilibrium

711: with the radiation field. The corresponding Bondi radius \citep{Bondi:1952}

712: is $R_{B}=GM_{\mathrm{BH}}/c_{\infty}^{2}=4.8\times10^{18}\,\mathrm{cm}$

713: while its relation to the Compton radius is $R_{B}=\gamma^{-1}R_{C}$.

714: The Bondi accretion rate (for the isothermal flow) is $\dot{M}_{B}=3.3\times10^{25}\,\mathrm{g\, s^{-1}}=0.52\,\MsunPerYear$.

715: The corresponding free-fall time ($t_{\mathrm{ff}}$) of gas from

716: the Bondi radius to the inner boundary is $2.1\times10^{11}\,\mathrm{sec}=7.0\times10^{3}\,\mathrm{yr}$

717: which is about $2.3$ times smaller than the precession period used

718: for Models~II and III, and about $23$ times smaller than that of

719: Model~IV (c.f.~Tab.~\ref{tab:Model-Summary}). 

720: 

721: 

722: \subsection{Comparison of axisymmetric models in 2-D and 3-D}

723: 

724: \label{sub:Comparison-2D-3D}

725: 

726: Before we proceed to the main precession disk models, we briefly compare

727: our axisymmetric model (Model~I) with the axisymmetric models presented

728: earlier by \citet{Proga:2007} who used very similar model parameters

729: as in our Model~I. The main differences here are in the treatment

730: of the radiation force and that in the radiative heating/cooling.

731: As mentioned earlier, we set the force multiplier $\mathcal{M}=0$

732: (in eq.~{[}\ref{eq:rad_force}]) and the net cooling rate $\mathcal{C}=0$

733: (in eq.~{[}\ref{eq:hydro03}]) while \citet{Proga:2007} used non-zero

734: values of those two terms. In our Model~I, the adiabatic index is

735: set to $\gamma=1.01$ (essentially isothermal), but their models use

736: $\gamma=5/3$. However, \citet{Proga:2007} found that their Run~A

737: is nearly isothermal despite $\gamma=5/3$ was used (see their Fig.~1).

738: Another important difference is the numerical codes used. \citet{Proga:2007}

739: used the ZEUS-2D code \citep{Stone:1992}. 

740: 

741: Overall geometry of the flow in Model~I (Figs.~\ref{fig:Density-and-velocity}

742: and \ref{fig:3-d-plots}) is similar to those in \citet{Proga:2007}.

743: The matter accretes onto the central BH near the equatorial plane,

744: and strong outflows occur in polar direction. The collimation of our

745: model is relatively weak compared to their Run~C which uses exactly

746: the same disk and corona luminosities as in our Model~I. The wider

747: bipolar outflow pattern seen here resembles that of their Run~A which

748: has the highest X-ray heating. The difference and the resemblance

749: seen here are caused by the following two key factors: (1) nearly

750: isothermal equation of state and (2) no radiative cooling ($\mathcal{C}=0$)

751: in our model . These condition keep the temperature of gas warm everywhere

752: in the computational domain, and the temperature is essentially that

753: set at the outer boundary ($T_{\infty}=T_{C}=2\times10^{7}\,\Kelvin$).

754: This will result in a very similar situation as in Run~A of \citet{Proga:2007}

755: in which the gas temperature is also relatively high because of the

756: high X-ray heating and cooling. The high temperature hence the ionization

757: state of the gas makes the line force in their model very inefficient,

758: resulting in the situation in which the gas is almost entirely driven

759: by the continuum process (electron scattering) and thermal effects

760: just as in our Model~I. 

761: 

762: Although not shown here, we have also checked the internal consistency

763: of the ZEUS-MP (3-D) code by running the axisymmetric models (Model~I)

764: in both 2-D and 3-D modes. We find that the results from the both

765: runs agree with each other in all aspects e.g., inflow and outflow

766: geometry, density distribution, velocity, mass accretion and outflow

767: rates. 

768: 

769: %====================================

770: 

771: 

772: 

773: %

774: \begin{figure*}

775: \begin{center}

776: 

777: \begin{tabular}{cc}

778: \includegraphics[clip,width=0.52\textwidth]{f2a.eps}&

779: \hspace{-0.8 cm}\includegraphics[clip,width=0.52\textwidth]{f2b.eps}\tabularnewline

780: \includegraphics[clip,width=0.52\textwidth]{f2c.eps}&

781: \hspace{-0.8 cm}\includegraphics[clip,width=0.52\textwidth]{f2d.eps}\tabularnewline

782: \end{tabular}

783: 

784: \end{center}

785: 

786: 

787: \caption{Comparison of the density and velocity fields on the $x$-$z$ plane

788: from Models~I (upper-left) II (upper-right), III (lower-left) and

789: IV (lower-right). The density maps shown in the background are given

790: in logarithmic scale (base 10) and in cgs units. The contours of the

791: Mach number are overlaid along with the arrows which indicate the

792: directions of the velocity on the x-$z$ plane. The units of both

793: $x$ and $z$ axes are in pc. The time slices of each models are chosen

794: such that the density and velocity fields are representative of each

795: model. }

796: 

797: \label{fig:Density-and-velocity}

798: \end{figure*}

799: 

800: 

801: %====================================

802: 

803: %====================================

804: 

805: 

806: 

807: %

808: \begin{figure*}

809: \begin{center}

810: 

811: \begin{tabular}{cc}

812: \includegraphics[clip,width=0.48\textwidth]{f3a.eps}&

813: \includegraphics[clip,width=0.48\textwidth]{f3b.eps}\tabularnewline

814: \includegraphics[clip,width=0.48\textwidth]{f3c.eps}&

815: \includegraphics[clip,width=0.48\textwidth]{f3d.eps}\tabularnewline

816: \end{tabular}

817: 

818: \end{center}

819: 

820: 

821: \caption{Two-level iso-density (left panels) and the corresponding sonic (right

822: panels) surfaces in 3-D for Models~I (upper panels) and II (lower

823: panels). The density levels used here are $\log\,\rho=-20.5$ (blue)

824: and $-21$ (red) where $\rho$ is in $\mathrm{g\, cm^{-3}}$. The

825: time slices of the simulation data used here are as in Fig.~\ref{fig:Density-and-velocity}.

826: The sizes of the plotting boxes are $14.2$~pc in all directions

827: ($x$, $y$ and $z$). }

828: 

829: \label{fig:3-d-plots}

830: \end{figure*}

831: 

832: 

833: %====================================

834: 

835: %====================================

836: 

837: 

838: 

839: %

840: \begin{figure*}

841: \begin{center}

842: 

843: \begin{tabular}{cc}

844: \includegraphics[clip,width=0.48\textwidth]{f4a.eps}&

845: \includegraphics[clip,width=0.48\textwidth]{f4b.eps}\tabularnewline

846: \includegraphics[clip,width=0.48\textwidth]{f4c.eps}&

847: \includegraphics[clip,width=0.48\textwidth]{f4d.eps}\tabularnewline

848: \end{tabular}

849: 

850: \end{center}

851: 

852: 

853: \caption{As in Fig.~\ref{fig:3-d-plots}, but for Models~III and IV. }

854: 

855: \label{fig:3-d-plots-2}

856: \end{figure*}

857: 

858: 

859: %====================================

860: 

861: 

862: \subsection{Dependency on the disk tilt angle $\Theta$}

863: 

864: \label{sub:Dependency-on-beta}

865: 

866: We now examine the model dependency on disk tilt angle ($\Theta$)

867: while keeping all other parameters fixed. The results from Models~I,

868: II and III (c.f., ~Tab~\ref{tab:Model-Summary}), which use $\Theta=0^{\circ}$,

869: $5^{\circ}$ and $15^{\circ}$ respectively, are compared for this

870: purpose. Note that the observations suggest that a large fraction

871: of AGN have rather small i.e., $\Theta<\sim15^{o}$ (e.g.,~\citealt{Lu:2005}). 

872: 

873: Figure~\ref{fig:Density-and-velocity} shows the slices of the density

874: and velocity fields (on the $y=0$ plane) from snapshots of our four

875: simulations. The snapshots are chosen at the time when the models

876: reached a (semi-)steady state for Models~I and II. As we will see

877: later, the flow never reaches steady state in Model~III; therefore,

878: we chose the snapshot of the model at the time when the flow pattern

879: is a typical of a whole simulation time sequence. While accretion occurs

880: mainly on the equatorial plane ($z=0$) for Model~I, it occurs in

881: a inclined plane with a pitch angle (a angle between the equatorial

882: plane and the accretion plane) similar to the disk inclination angle,

883: for Models~II and III. In the precessing disk models (II and III),

884: the deviation from the axisymmetric is clearly seen in both density

885: distribution of gas and the shapes of the Mach number contours. Corresponding

886: 3-D density and Mach number contour surfaces of these models are also

887: shown in Figs.~\ref{fig:3-d-plots} and \ref{fig:3-d-plots-2}. The

888: morphology of the density distribution seen in Figure~\ref{fig:Density-and-velocity}

889: resemble that of the Z-shaped (for Model~II) and the S-shaped (for

890: Model~III) radio jets (e.g.~\citealt{Condon:1984}; \citealt{Hunstead:1984};

891: \citealt{Tremblay:2006}) although in different scales. Obviously,

892: the difference between the Z- and S- shapes are simply due to the

893: difference in the viewing angles. Unlike the MHD precessing jet models,

894: the bending structures of the density distributions seen here are

895: shaped by the geometry of the sonic surfaces. When accreting material

896: from the outer boundary encounters the relatively low density but

897: high-speed outflowing gas launched by the radiation force from the

898: inner part, the gas becomes compressed, and forms higher density regions.

899: The flows in the bending density structure itself are rather complex

900: (especially in Model~III), but the direction of the flow becomes

901: outward (in radial direction) as they approach the sonic surface (excluding

902: the one shaped like a disk formed by the \emph{accreting} gas in the

903: inner region). Relatively large curvatures of the flows seen in both

904: density and the Mach number contours of Models~II and III can be

905: also understood from the fact the precession period used in these

906: models ($P=16000\,\mathrm{yr}$) is comparable to the gas free fall

907: time ($t_{\mathrm{ff}}=7000\,\mathrm{yr}$, c.f., \S~\ref{sub:Reference-Values}).

908: The curvatures or the {}``twists'' of the weakly collimated bipolar

909: flows can be clearly seen in the 3-D representation of these models

910: in Figs.~\ref{fig:3-d-plots} and \ref{fig:3-d-plots-2}. 

911: 

912: %====================================

913: 

914: 

915: 

916: %

917: \begin{figure*}

918: \begin{center}

919: 

920: \begin{tabular}{cc}

921: \includegraphics[clip,width=0.45\textwidth]{f5a.eps}&

922: \hspace{0.5cm}\includegraphics[clip,width=0.45\textwidth]{f5b.eps}\vspace{0.25cm}\tabularnewline

923: \includegraphics[clip,width=0.45\textwidth]{f5c.eps}&

924: \hspace{0.5cm}\includegraphics[clip,width=0.45\textwidth]{f5d.eps}\tabularnewline

925: \end{tabular}

926: 

927: \end{center}

928: 

929: 

930: \caption{Comparison of mass flux and energy flux as a function of radius for

931: Models~I (upper-left), II (upper-right), III (lower-left) and IV

932: (lower-right). The plot for each model is subdivided into two panels:

933: top (mass flux) and bottom (energy flux). In the mass flux plots,

934: the inflow (dash: $\dot{M}_{\mathrm{in}}$), outflow (solid: $\dot{M}_{\mathrm{o}}$)

935: and net (dot: $\dot{M}_{\mathrm{net}}$) mass fluxes, as defined in

936: equation~(\ref{eq:mdot2}), are separately plotted as a function

937: of radial distance from the center. The absolute values of $\dot{M}_{\mathrm{in}}$

938: and $\dot{M}_{\mathrm{net}}$ are plotted here since they are negative

939: at all radii. In the energy flux plots, the kinetic energy (solid)

940: and the thermal energy (dot), defined in eqs.~(\ref{eq:power-kinetic})

941: and (\ref{eq:power-thermal}), are shown. Note that the times slices

942: of the model simulations used here to computed the fluxes are same

943: as those in Figure~\ref{fig:Density-and-velocity}. }

944: 

945: \label{fig:mass-engery-flux}

946: \end{figure*}

947: 

948: 

949: %====================================

950: 

951: We compute the mass fluxes as a function of radius for a quantitative 

952: comparison of the characteristics of the flows in the models. Following

953: \citet{Proga:2007}, the net mass flux ($\dot{M}_{\mathrm{net}}$),

954: the inflow mass flux ($\dot{M}_{\mathrm{in}}$) and the outflow mass

955: flux ($\dot{M}_{\mathrm{out}}$) can be computed from 

956: \begin{align}

957:   \dot{M}\left(r\right) & =\oint_{s}\rho\,\boldsymbol{v\,\cdot}d\boldsymbol{a}\label{eq:mdot}\\

958:                         & =r^{2}\oint_{4\pi}\rho v_{r}\, d\Omega\label{eq:mdot2}

959: \end{align}

960: where $v_{r}$ is the radial component of velocity $\boldsymbol{v}$.

961: In the equation above, $\dot{M}=\dot{M}_{\mathrm{net}}$ if all $v_{r}$

962: are included. Similarly, $\dot{M}=\dot{M}_{\mathrm{in}}$ for $v_{r}<0$

963: and $\dot{M}=\dot{M}_{\mathrm{out}}$ for $v_{r}>0$. Also, $d\boldsymbol{a}=\boldsymbol{\hat{r}}\, r^{2}\sin\theta\, d\theta\, d\phi$

964: and $d\Omega=\sin\theta\, d\theta\, d\phi$. Similarly we further

965: define the \emph{outflow} power in the form of kinetic energy ($P_{k}$)

966: and that in the thermal energy ($P_{\mathrm{th}}$) as functions of

967: radius i.e., 

968: \begin{align}

969:   P_{k}\left(r\right) & =r^{2}\oint_{4\pi}\rho v_{r}^{3}\,

970:   d\Omega

971:   \label{eq:power-kinetic}

972: \end{align}

973: and

974: \begin{align}

975:   P_{\mathrm{th}}\left(r\right) & =r^{2}\oint_{4\pi}uv_{r}\,

976:   d\Omega\,.

977:   \label{eq:power-thermal}

978: \end{align}

979: where $v_{r}>0$. 

980: 

981: The resulting mass fluxes and the outflow powers of the models are

982: summarized in Figure~\ref{fig:mass-engery-flux}. In all cases, the

983: mass inflow flux exceeds the mass outflow rate at all radii. For Models~I

984: and II, the net mass fluxes ($\dot{M}_{\mathrm{net}}$) are almost

985: constant at all radii, indicating that the flows in these models are

986: steady. Despite the presence of the disk precession in Model~II,

987: the flow becomes steady. The density distribution and the velocity

988: field become almost constant in the coordinate system co-rotating

989: with the disk precession period. On the other hand, $\dot{M}_{\mathrm{net}}$

990: for Model~III does not remain constant as $r$ becomes larger ($r>10^{18}\,\mathrm{cm}$)

991: because of the unsteady nature of the flow (c.f., Figs.~\ref{fig:Density-and-velocity}

992: and \ref{fig:3-d-plots-2}). As the disk tilt angle $\Theta$ increases,

993: the direction of the outflows, which are normally in polar directions

994: ($\pm z$ directions) with an absence of the disk tilt, moves toward

995: the equatorial plane (the $x$-$z$ plane) where the flow is predominantly

996: inward. This opposite flows makes it harder for the outflowing gas

997: to reach the outer boundary. Further, since the disk is precessing,

998: the direction of the outflow is constantly changing. This results

999: in continuous collisions between the inflowing and outflowing gas

1000: especially for a larger $\Theta$ model. The net mass fluxes at the

1001: inner boundary $\dot{M}_{\mathrm{net}}\left(r_{\mathrm{i}}\right)$

1002: are $-0.6$, $-0.6$ and $-0.4\times10^{25}\,\mathrm{g\, s^{-1}}$

1003: (or equivalently $-0.10$, $-0.1$0 and $-0.06$~$\MsunPerYear$)

1004: for Models~I, II and III respectively (Tab.~\ref{tab:Model-Summary}),

1005: indicating the net mass flux inward (negative signs indicate inflow)

1006: decreases slightly, but not significantly as the disk tilt angle $\Theta$

1007: increases. 

1008: 

1009: The ratios of the total mass outflow flux to the total mass inflow

1010: at the outer boundary ($\mu=\left|\dot{M}_{\mathrm{out}}/\dot{M}_{\mathrm{in}}\right|$)

1011: are $0.7$3, $0.74$, $0.82$  for Models I, II, and III (see also

1012: Tab.~\ref{tab:Model-Summary}). These values indicates that the high

1013: efficiency of the outflow by the radiation pressure even for a modest

1014: Eddington number used here i.e., $\Gamma=0.6$. This conversion efficiency

1015: $\mu$ (from the outflow to inflow) is about the same for Models~I

1016: and II, but it slightly ($\sim12$\%) increases for Model~III which

1017: has the highest disk tilt angle. Overall characteristics of the mass-flux

1018: curves as a function radius for Models~I and II are also very similar

1019: to each other. The curves for Model~III are also similar to those

1020: of Models~I and II; however, they differ in the outer radii ($r<\sim10^{18}\mathrm{cm}$),

1021: mainly because of the unsteady nature of the flow in this model. 

1022: 

1023: The maximum speed of the outflow in the radial direction

1024: $v_{r}^{max}\left(r_{\mathrm{\mathrm{o}}}\right)$ 

1025: decreases as $\Theta$ increases (Tab.~\ref{tab:Model-Summary}). The

1026: reduction in the speed is very significant ($\Delta v_{r}=-860\,\kmps$)

1027: as $\Theta$ increase from $0^{\circ}$ to $5^{\circ}$ while the change

1028: is relatively small ($\Delta v_{r}=-20\,\kmps$) as $\Theta$ changes

1029: from $5^{\circ}$ to $15^{\circ}$. 

1030: 

1031: Figure~\ref{fig:mass-engery-flux} also shows the outflow powers

1032: ($P_{k}$ and $P_{\mathrm{th}}$) of the models as a function of radius,

1033: as defined in eqs.~(\ref{eq:power-kinetic}) and (\ref{eq:power-thermal}).

1034: As for the mass flux curves in the same figure, the dependency of

1035: the energy flux curves on radius for Models~I, II and III are very

1036: similar to each others. A small but noticeable deviations of the curves

1037: for Model~III from those for Models~I and II are seen at the large

1038: radii ($r>3\times10^{18}\,\mathrm{cm}$). The figure shows that in

1039: all three models, the outflow power is dominated by thermal process

1040: ($P_{\mathrm{th}}\approx10-100\, P_{k}$ at all radii). This can be

1041: explained by the high temperatures of the gas ($T\approx T_{C}=2.0\times10^{7}\,\Kelvin$

1042: ) in the computational domains caused by the (almost) isothermal equation

1043: of state and the temperature ($2.0\times10^{7}\,\Kelvin$) fixed at

1044: the outer boundary. The kinetic powers or the radiation forces are

1045: not as significant as the pressure gradient force in these models;

1046: however, their importance cannot be ignored since they {}``shape''

1047: the geometry of the outflow as they strongly depend on the polar angle

1048: position of a point in the computational field. We also note that

1049: as $\Theta$ increases, the kinetic power at the outer boundary $P_{k}\left(r_{\mathrm{o}}\right)$

1050: decreases significantly e.g.~$P_{k}\left(r_{\mathrm{o}}\right)$

1051: of Model~III is three times smaller than that of the axisymmetric

1052: model, Model~I (see Table~\ref{tab:Model-Summary}).

1053: 

1054: Next we examine the degree of non-axisymmetry in Model~I, II and

1055: III, and seek for any obvious dependency on $\Theta$. For this purpose,

1056: we compute the center of mass (CM) of the gas on the planes perpendicular

1057: to the $z$-axis (as a function of $z$) i.e., $x_{c}\left(z\right)$

1058: and $y_{c}\left(z\right)$ which are defined as 

1059: \begin{equation}

1060:   x_{c}\left(z\right)=\frac{\int_{-r_{\mathrm{o}}}^{r_{\mathrm{o}}}\int_{-a}^{a}\int_{-\infty}^{\infty}x\,\delta\left(z-z'\right)\rho\left(x,y,z'\right)dz'\, dx\, dy}{m\left(z\right)}

1061:   \label{eq:cm-x-comp}

1062: \end{equation}

1063: and 

1064: \begin{equation}

1065:   y_{c}\left(z\right)=\frac{\int_{-r_{\mathrm{o}}}^{r_{\mathrm{o}}}\int_{-a}^{a}\int_{-\infty}^{\infty}y\,\delta\left(z-z'\right)\rho\left(x,y,z'\right)dz'\, dx\, dy}{m\left(z\right)}

1066:   \label{eq:cm-y-comp}

1067: \end{equation}

1068: where 

1069: \begin{equation}

1070:   m\left(z\right)=\int_{-r_{\mathrm{o}}}^{r_{\mathrm{o}}}\int_{-a}^{a}\int_{-\infty}^{\infty}\delta\left(z-z'\right)\rho\left(x,y,z'\right)dz'\, dx\, dy\,\,,

1071:   \label{eq:mass-on-plane}

1072: \end{equation}

1073: and $a\left(y,z\right)=$ $\left(r_{\mathrm{o}}^{2}-z^{2}-y^{2}\right)^{1/2}$.

1074: The results are shown in Figure~\ref{fig:Center-of-mass}. As expected,

1075: the CM position remains constant and on the $z$-axis ($x_{c}\left(z\right)=0$

1076: and $y_{c}\left(z\right)=0$) for Model~I, as this is an axisymmetric

1077: model. For both Models~II and III, the maximum amount of deviations

1078: for each component of the CM ($\left|x_{c}\right|$and $\left|y_{c}\right|$)

1079: is about $0.3$~pc which is relatively small compared to the outer

1080: boundary radius ($r_{\mathrm{o}}=7.1\,\mathrm{pc}$). The $x_{c}$

1081: and $y_{c}$ curves are anti-symmetric about the $z=0$ position since

1082: our model accretion disk hence the radiation force is symmetric about

1083: the origin of the coordinate system. The plot also shows that the

1084: positions of the maxima and minima in the $x_{c}$ and $y_{c}$ curves

1085: do not coincide, but they are rather shifted in both $+z$ and $-z$

1086: directions. This clearly demonstrates a helical or twisting nature

1087: of the flows, as one can also simply see it in the 3-D density and

1088: Mach number contour plots in Figs.~\ref{fig:3-d-plots} and \ref{fig:3-d-plots-2}. 

1089: 

1090: To summarize, as the tilt angle of the disk precession $\Theta$ increases,

1091: reductions of the maximum outflow velocity ($v_{r}$) and the kinetic

1092: outflow power ($P_{k}$) at the outer boundary $r_{\mathrm{o}}$ occur,

1093: as a consequence of the stronger interactions between the outflowing

1094: and inflowing gas of as $\Theta$ increases. The net mass inflow flux

1095: ($\dot{M}_{\mathrm{net}}$) at the inner boundary does not strongly

1096: depend on $\Theta$. The thermal outflow energy power dominates the

1097: kinetic outflow power in our models here because of the high temperature

1098: of set at the outer boundary and because the gas is (almost) isothermal.

1099: The flows of Models~II and III show helical structures; however,

1100: the radius of the helices (base on the CM positions along the z-axis)

1101: does not change greatly as $\Theta$ increases from $5^{\circ}$ to

1102: $15^{\circ}$. 

1103: 

1104: %====================================

1105: 

1106: 

1107: 

1108: %

1109: \begin{figure}[h]

1110: \begin{center}

1111: 

1112: \includegraphics[clip,width=0.45\textwidth]{f6.eps}

1113: 

1114: \end{center}

1115: 

1116: 

1117: \caption{Positions of the center of mass of the gas on the planes perpendicular

1118: to the $z$-axis for Models I, II, III and IV from the top bottom.

1119: The locations of the center of mass, $x_{c}$ (sold) and $y_{c}$

1120: (dash) as defined in eqs.~(\ref{eq:cm-x-comp}) and (\ref{eq:cm-y-comp}),

1121: are computed as a function of $z$. All models except Model~I show

1122: a clear sign of deviations from axisymmetry; however the displacement

1123: of the center of mass remains fairly small ($\left|x_{c}\right|<\sim0.3\,\mathrm{pc}$

1124: and $\left|y_{c}\right|<\sim0.3\,\mathrm{pc}$), at all $z$ locations,

1125: compared to the size of outermost radius $r_{\mathrm{o}}$ ($7.1$~pc)

1126: of the computational domain. The patterns in $x_{c}$ and $y_{c}$

1127: curves for Models~II and III indicate that the flow density structures

1128: are helical.}

1129: 

1130: \label{fig:Center-of-mass}

1131: \end{figure}

1132: 

1133: 

1134: %====================================

1135: 

1136: 

1137: \subsection{Dependency on the disk precession period $P$}

1138: 

1139: \label{sub:Dependency-on-P}

1140: 

1141: We now examine the dependency of the model on the disk precession

1142: period ($P$). We vary the value of $P$ while fixing the disk tilt

1143: angle to $\Theta=5^{\circ}$. For this purposed, we compare Models~I,

1144: II and IV as summarized in Table.~\ref{tab:Model-Summary}. The precession

1145: periods $P$ are $\infty$, $1.6\times10^{4}$ and $1.6\times10^{5}$~yr

1146: respectively for Models~I, II and IV. In the units of the free-fall

1147: time ($t_{\mathrm{ff}}=7.0\times10^{3}$~yr ) from the Bondi radius

1148: (\S~\ref{sub:Reference-Values}), they are $\infty$, $2.3$ and

1149: $23$ respectively. Note that the observations suggest that typical

1150: values of jet precession period are $P=10^{4}$--$10^{6}$~yr (c.f.,

1151: Tab.1 in \citealt{Lu:2005}). 

1152: 

1153: Figure~\ref{fig:Density-and-velocity} shows that similarities between

1154: Model~IV and Model~I in their morphology of the density distribution

1155: and Mach number surfaces. At a given time, the flow in Model~IV is almost axisymmetric,

1156: and the symmetry axis is tilted also by $\Theta=5^{\circ}$ from the

1157: $z$-axis. This is caused by the relatively long precession period

1158: for Model~IV compared to the dynamical time scale or the gas free-fall

1159: time scale $t_{\mathrm{ff}}$. The curvature or helical motion of

1160: the gas is not significant, and it does not greatly affect the overall

1161: morphology of the flow, except for the outermost part of the flow

1162: where the flow is slightly turbulent due to the shear of the slowly

1163: precessing flow and the outer boundary. This can be clearly seen in

1164: the 3-D plots Figure~\ref{fig:3-d-plots-2}. As the precession period

1165: $P$ becomes shorter and comparable to $t_{\mathrm{ff}}$ (as in Model~II),

1166: the flow shows more curvature and the helical structures. 

1167: 

1168: The mass flux curves (c.f., eq.~{[}\ref{eq:mdot2}]) for Models~IV

1169: in Figure~\ref{fig:mass-engery-flux} also show that nature of the flows between

1170: Models~I and IV are very similar to each other. Overall characteristics

1171: of the curves are also similar to that of Model~II. In fact, the

1172: net mass flux $\dot{M}_{\mathrm{net}},$ the inflow mass flux $\dot{M}_{\mathrm{in}}$

1173: and the outflow mass flux $\dot{M}_{\mathrm{out}}$ at the outer boundary

1174: of Model~IV are identical to those of Model~I (see Table~\ref{tab:Model-Summary}).

1175: Also note that Models~I, II and IV all have same $\dot{M}_{\mathrm{in}}$

1176: value at the inner boundary, i.e., the mass inflow rate across the

1177: inner boundary in insensitive to the change in the precession period

1178: for $\Theta=5^{\circ}$. 

1179: 

1180: The similarity between and Models~I and IV can be also seen in the

1181: outflow powers, $P_{k}$ and $P_{\mathrm{th}}$. Figure~\ref{fig:mass-engery-flux}

1182: shows $P_{k}$ and $P_{\mathrm{th}}$ as a function of radius for

1183: Model~IV are almost identical to those of Model~I. The $P_{k}$

1184: and $P_{\mathrm{th}}$ values at outer boundary are indeed identical

1185: (Table~\ref{tab:Model-Summary}). A slight increase in the maximum

1186: outflow velocity at the outer boundary $v_{r}^{\mathrm{max}}(r_{\mathrm{o}})$

1187: is observed for Model~IV, compared to Model~I. The kinetic outflow

1188: power $P_{k}$ and the maximum outflow velocity at the outer boundary

1189: $v_{r}^{\mathrm{max}}(r_{\mathrm{o}})$ decreases as the precession

1190: period become comparable to $t_{\mathrm{ff}}$. 

1191: 

1192: The CM positions  $x_{c}$ and $y_{c}$ as a function of $z$ (see

1193: \S~\ref{sub:Dependency-on-beta}) for Model~IV is shown in Figure~\ref{fig:Center-of-mass}.

1194: Compared to Model~II, the maximum displacement of the CM is about

1195: 40 times smaller in Model~IV i.e.~$\left|x_{c}\right|<\sim0.075\,\mathrm{pc}$

1196: and $\left|y_{c}\right|<\sim0.075\,\mathrm{pc}$. The $x_{c}$ curve

1197: for Model~IV shows a rather complex pattern compared to that in Model~II.

1198: This and the visual inspection of the density and the Mach number

1199: contour surfaces in Figure~\ref{fig:3-d-plots-2}, indicates that

1200: the bipolar outflow flows are slightly twisted, but does not have

1201: clear helical structure. 

1202: 

1203: 

1204: \subsection{Time Evolution of Mass Accretion/Outflow Rates and Angular Momentum }

1205: 

1206: \label{sub:Time-dependet-behavours}

1207: 

1208: %====================================

1209: 

1210: 

1211: 

1212: %

1213: \begin{figure*}

1214: \begin{center}

1215: 

1216: \begin{tabular}{cc}

1217: \includegraphics[clip,width=0.45\textwidth]{f7a.eps}&

1218: \hspace{0.5cm}\includegraphics[clip,width=0.45\textwidth]{f7b.eps}\tabularnewline

1219: \includegraphics[clip,width=0.45\textwidth]{f7c.eps}&

1220: \hspace{0.5cm}\includegraphics[clip,width=0.45\textwidth]{f7d.eps}\tabularnewline

1221: \end{tabular}

1222: 

1223: \end{center}

1224: 

1225: 

1226: \caption{The mass flow rates across the outer boundary and the density-weighted

1227: mean specific angular momentum $j_{\rho}$ (c.f.~eqs.~[\ref{eq:mdot2}]

1228: and [\ref{eq:def_rho_mean_j}]) plotted as a function of time for Models~I

1229: (upper-left), II (upper-right), III (lower-left), and IV (lower-right).

1230: For each model, the plot is subdivided into two panels: mass-inflow/outflow

1231: rates (top panels) and $j_{\rho}$ (bottom panels). In the top panels,

1232: the mass-inflow rate at the outer boundary (solid), and the mass-outflow

1233: rate at the outer boundary (dash) are shown separately. The values

1234: of $j_{\rho}$ are in the units of $j_{0}$ which is defined as the

1235: specific angular momentum of the gas in Keplerian orbit at the inner

1236: boundary ($r=r_{\mathrm{in}}$). Note that the precession period used

1237: here are $\infty$, $5.0\times10^{11}\,\mathrm{s}$, $5.0\times10^{11}\,\mathrm{s}$

1238: and $5.0\times10^{12}\,\mathrm{s}$ for Models I, II, III and IV respectively.}

1239: 

1240: \label{fig:mdot-j-evolution}

1241: \end{figure*}

1242: 

1243: 

1244: %====================================

1245: 

1246: Next, we examine the variability or steadiness of the flows in each

1247: model by monitoring the mass fluxes at the outer boundary as in equation~(\ref{eq:mdot2})

1248: and the angular momentum of the system as a function of time. For

1249: the latter, we compute the density-weighted mean specific angular

1250: momentum $\boldsymbol{j}_{\rho}$ of the systems defined as: 

1251: \begin{equation}

1252:   \boldsymbol{j}_{\rho}=\frac{\int_{V}\rho\,\left(\boldsymbol{r}\times\boldsymbol{v}\right)\, dV}{\int_{V}\rho\, dV}

1253:   \label{eq:def_rho_mean_j}

1254: \end{equation}

1255: where the denominator is simply the total mass of the gas in the computational

1256: domain. Note that the radiation force (eq.~[\ref{eq:rad-force-final}])

1257: and the gravitational force are in radial direction only. Consequently,

1258: they do not exert torque onto the system; hence, they do not contribute

1259: to the change in the angular momentum of the system directly. The

1260: system can gain the angular momentum in the following way. In our

1261: models, the strength of the

1262: disk radiation field depends on the angle measured from the disk

1263: normal (c.f., eq.~[\ref{eq:disk-flux}]). 

1264: This causes gas pressure gradients in azimuthal direction, and  

1265: contributes to the angular momentum of gas locally,  forming

1266: vorticity. The precession of radiation field hence the precessing

1267: outflow will cause the gas with preferred sign of vorticity to escape

1268: from the outer boundary, resulting a change in the net angular momentum

1269: of the gas in the computational domain. 

1270: 

1271: Figure~\ref{fig:mdot-j-evolution} shows $\dot{M}_{\mathrm{in}}\left(r_{o}\right)$,

1272: $\dot{M}_{\mathrm{out}}\left(r_{o}\right)$ and $j_{\rho}$ for Models~I,

1273: II, III and IV as a function of time. For Models I, II and IV, both

1274: the mass fluxes and $j_{\rho}$ reach to asymptotic values by $t\approx7\times10^{12}\,\mathrm{s}$. Small

1275: oscillations of $j_{\rho}$ around the asymptotic values are seen

1276: for Models~II and IV. On the other hand, the mass fluxes of Model~III

1277: has much larger amplitudes of the oscillations around an asymptotic

1278: value. By visual inspections, their oscillations do not seem have

1279: a clear periodicity associated with them. We performed the Lomb-Scargle

1280: periodgram analysis (e.g.~\citealt{Horne:1986}; \citealt{Press:1992})

1281: on the $\dot{M}_{\mathrm{in}}\left(r_{o}\right)$, $\dot{M}_{\mathrm{out}}\left(r_{o}\right)$

1282: and $j_{\rho}$ curves for Model~III. Only $\dot{M}_{\mathrm{out}}\left(r_{o}\right)$

1283: shows a relatively strong signal at $P_{\mathrm{LS}}=1.36\times10^{12}\,\mathrm{s}$

1284: which is about $2.7$ times longer than the precession period of Model~III.

1285: On the other hand $\dot{M}_{\mathrm{in}}\left(r_{o}\right)$ and $j_{\rho}$

1286: curves do not have any obvious period associated with them, but they

1287: are rather stochastic. 

1288: 

1289: As mentioned in \S~\ref{sub:Dependency-on-beta}, as the disk tilt

1290: angle $\Theta$ increases the direction of the outflows, which normally

1291: exists in polar directions with an absence of the disk tilt, moves

1292: toward the equatorial plane (the x-$z$ plane) where the flow is predominantly

1293: inward. In addition, the precession of the disk causes the direction

1294: of the outflow to change constantly; hence, causing constant creation

1295: of the shock between the inflowing and the outflowing gas. This leads

1296: into a very unstable flow of the gas at all time for a model with

1297: a larger $\Theta$ e.g., Model~III with $\Theta=15^{\circ}$. The flow,

1298: of course, can be stabilized if the precession period is increased

1299: to a value much larger than the free-fall time $t_{\mathrm{ff}}$. 

1300: 

1301: The amount of the (density-weighted) mean specific angular momentum

1302: deposited to the gas by the precessing disk (measured by $j_{\rho}$)

1303: is largest in Models~II and III (see Table~\ref{tab:Model-Summary}),

1304: and that in Model~IV is about $4$ times less than those of Models~II

1305: and III. For all models, a time-averaged value (by using the last

1306: $2\times10^{12}\,\mathrm{s}$ of the simulation) of $j_{\rho}$ is

1307: used in Table~\ref{tab:Model-Summary}. It seems that the faster

1308: the disk precesses, the lager the amount of angular momentum transferred

1309: to the environment; however, this trend does not continues, as we

1310: increase the disk precession speed even faster. Although not shown

1311: here, a model with exactly the same set of parameters as in Model~II,

1312: but with $P=1600\,\mathrm{yr}$ (10 times faster rotation), showed

1313: that the the value of $j_{\rho}$ decrease to $\sim0.01$, which is

1314: even smaller than Model~IV (with $P=160000\,\mathrm{yr}$). This

1315: indicates that the amount of angular momentum deposited to the gas

1316: depends on how close the precession period to the dynamical time scale

1317: of the flow. 

1318: 

1319: In principle, it is possible to model the change in the angular momentum

1320: of the accretion disk itself through the transfer of angular momentum

1321: from the environment, we ignored this effect for simplicity (and this

1322: is also our model limitation). To model the interaction of the disk

1323: angular momentum and the angular momentum of the surrounding gas properly,

1324: we need to model the dynamics of the gas in the accretion disk itself

1325: as well as the dynamic of the gas which is much larger scale as in

1326: our models here. This is computationally challenging with our current

1327: code since we have to resolve the length scale of the innermost part

1328: of the accretion disk ($\sim10^{-5}$~pc) to the large scale outflow/inflow

1329: gas ($\sim10$~pc). 

1330: 

1331: 

1332: \section{Conclusions}

1333: 

1334: \label{sec:Conclusions}

1335: 

1336: We have studied the dynamics of the gas under the influences of the

1337: gravity of SMBH and the radiation force from the luminous accretion

1338: disk around the SMBH. The rotational axis of the disk was assumed

1339: to be tilted with respect to the symmetry axis with a given angle

1340: $\Theta$ and a precession period $P$ (c.f., Figure~\ref{fig:model-config}).

1341: We have investigated the dependency of the flow morphology, mass accretion/outflow

1342: rates, angular momentum of the flows for different combinations of

1343: $\Theta$ and $P$. This is a natural extension of similar but more

1344: comprehensive 2-D radiation hydrodynamics models of AGN outflow models

1345: by ~\citet{Proga:2000}, \citet{Proga:2007}. As this is our first

1346: attempt for modeling such gas dynamics in full 3-D, we have used a

1347: reduced set of physical models described in \citet{Proga:2007} i.e.,

1348: the radiation force due to line and dust scattering/absorption, and

1349: the radiative cooling/heating are omitted. In the following, we summarize

1350: our main findings through this investigation. 

1351: 

1352: (1)~Our assumption of the adiabatic index ($\gamma=1.01$) keeps

1353: the mean temperature of the gas in the computational domain relatively

1354: high ($\sim2\times10^{7}\,\Kelvin$) which is essentially determined

1355: by the outer boundary condition. For our axisymmetric model (Model~I:

1356: Figs.~\ref{fig:Density-and-velocity} and \ref{fig:3-d-plots}),

1357: this results in the flow morphology very similar to the model with

1358: a relatively high X-ray heating (see Run~A in \citealt{Proga:2007})

1359: in which the line force is inefficient because of the high gas temperature

1360: and hence the high ionization state of the gas. 

1361: 

1362: (2)~Although in different scales, we were able to reproduced the

1363: Z- or S- shaped density morphology of the gas outflows (Fig.~\ref{fig:Density-and-velocity})

1364: which are often seen in the radio observations of AGN (e.g.~\citealt{Florido:1990};

1365: \citealt{Hutchings:1988}). The bending structure seen here are shaped

1366: by the shape of the sonic surfaces. When accreting material from the

1367: outer boundary encounters the relatively low density but high speed

1368: outflowing gas launched by the radiation force from the inner part,

1369: the gas becomes compressed, and forms higher density regions. 

1370: 

1371: (3)~As the tilt angle of the disk precession $\Theta$ increases,

1372: the reduction of the maximum outflow velocity ($v_{r}$) and the kinetic

1373: outflow power $P_{k}$ at the outer boundary $r_{\mathrm{o}}$ decrease

1374: as a consequence of the stronger interactions between the outflowing

1375: and inflowing gas (Tab.~\ref{tab:Model-Summary}). The net mass inflow

1376: rate ($\dot{M}_{\mathrm{net}}$) at the inner boundary does not change

1377: significantly with increasing $\Theta$. 

1378: 

1379: (4)~A relatively high efficiency of the outflow ($\mu=\dot{M}_{\mathrm{out}}/\dot{M}_{\mathrm{in}}$)

1380: by the radiation pressure were observed in our models ($70$--$80$~\%;

1381: see also Tab.~\ref{tab:Model-Summary}) for a Eddington number ($\Gamma$)

1382: of $0.6$ here. The conversion efficiency $\mu$ (from the outflow

1383: to inflow) is about the same for Models~I and II, but it slightly

1384: ($\sim12$\%) increased for Model~III which has the highest disk

1385: tilt angle. 

1386: 

1387: (5)~The thermal outflow energy power dominates the kinetic outflow

1388: power (Fig.~\ref{fig:mass-engery-flux}) in the models presented

1389: here because of the high temperature of the flow (as mentioned above). 

1390: 

1391: (6)~The flows of Models~II and III show helical structures (c.f.,

1392: Figs.~\ref{fig:3-d-plots} and \ref{fig:3-d-plots-2}); however,

1393: the radius of the helices does not change as $\Theta$ increases from

1394: $5^{\circ}$ to $15^{\circ}$, based on the locations of the center

1395: of mass (Figure~\ref{fig:Center-of-mass}) of the planes perpendicular

1396: to the symmetry axis (the $z$-axis in Fig.~\ref{fig:model-config}).

1397: We leave for a future investigation to test whether these trends continue

1398: as $\Theta$ becomes larger than $15^{\circ}$. 

1399: 

1400: (7)~The characteristics of the flows are closely related to a combination

1401: of $P$ and $\Theta$, but not to $P$ and $\Theta$ individually. Even

1402: with a relatively large disk tilt angle $\Theta$, if the precession

1403: period is much larger than the dynamical time scale of a system, the

1404: flow geometry obviously becomes almost axisymmetric (c.f.~Model~IV

1405: in Figs.~\ref{fig:Density-and-velocity} and \ref{fig:3-d-plots-2}). 

1406: 

1407: (8)~The gas dynamics of a model with a relative large disk tilt angle

1408: ($\Theta=15^{\circ}$) with a precession period comparable to the gas

1409: free-fall time ($t_{\mathrm{ff}}$) of the system (e.g.,~Model~III)

1410: does not reach a steady state because the outflows driven by the luminous

1411: accretion disk constantly collides with the inflowing/accreting gas

1412: as the disk precesses hence as the outflow direction changes. 

1413: 

1414: (9)~The amount of the density-weight mean specific angular momentum

1415: ($j_{\rho}$) deposited by the precessing disk is largest for Models~II

1416: and III (Tab.~\ref{tab:Model-Summary} and Fig.~\ref{fig:mdot-j-evolution})

1417: which have the precession period comparable to $t_{\mathrm{ff}}$. 

1418: 

1419: The models represented here are mainly for exploratory purpose --

1420: to examine the basic model dependencies on $\Theta$ and $P$ -- with

1421: a relatively simple set of physics but in full 3-D. In the follow-up

1422: paper, we will improve our model by including the physics omitted

1423: here (the line scattering/absorption, dust scattering/absorption,

1424: and the radiative cooling/heating) as in the 2-D models of e.g.,~\citet{Proga:2007}. 

1425: 

1426: 

1427: \acknowledgements{}

1428: 

1429: This work was supported by NASA through grant HST-AR-10680 from the

1430: Space Telescope Science Institute, which is operated by the Association

1431: of Universities for Research in Astronomy, Inc., under NASA contract

1432: NAS5-26555. This work was also supported by the National Center for

1433: Supercomputing Applications under AST070036N and utilized the Xeon

1434: Linux Cluster, Tungsten. Authors are grateful for original developer

1435: of ZEUS-MP for making the code publicly available. We thank Prof.~Jim

1436: Pringle and Monika Mo\'{s}cibrodzka for the critical reading of the

1437: manuscript, and comments. We also thank Agnieszka Janiuk for helpful

1438: discussion and support. 

1439: 

1440: %\bibliographystyle{apj}

1441: %\bibliography{local}

1442: 

1443: \begin{thebibliography}

1444: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1445: 

1446: \bibitem[{Aloy {et~al.}(2003)Aloy, {Mart{\'{\i}}}, {G{\'o}mez}, {Agudo},

1447:   {M{\"u}ller}, \& {Ib{\'a}{\~n}ez}}]{Aloy:2003}

1448: Aloy, M.-{\'A}., {Mart{\'{\i}}}, J.-M., {G{\'o}mez}, J.-L., {Agudo}, I.,

1449:   {M{\"u}ller}, E., \& {Ib{\'a}{\~n}ez}, J.-M. 2003, \apjl, 585, L109

1450: 

1451: \bibitem[{Aly(1980)}]{Aly:1980}

1452: Aly, J.~J. 1980, \aap, 86, 192

1453: 

1454: \bibitem[{Antonucci(1984)}]{Antonucci:1984}

1455: Antonucci, R.~R.~J. 1984, \apj, 278, 499

1456: 

1457: \bibitem[{Appl {et~al.}(1996)Appl, Sol, \& Vicente}]{Appl:1996}

1458: Appl, S., Sol, H., \& Vicente, L. 1996, \aap, 310, 419

1459: 

1460: \bibitem[{Arav {et~al.}(1994)Arav, Li, \& Begelman}]{Arav:1994}

1461: Arav, N., Li, Z.-Y., \& Begelman, M.~C. 1994, \apj, 432, 62

1462: 

1463: \bibitem[{Armitage \& Pringle(1997)}]{Armitage:1997}

1464: Armitage, P.~J. \& Pringle, J.~E. 1997, \apjl, 488, L47

1465: 

1466: \bibitem[{Awaki {et~al.}(1991)Awaki, Koyama, {Inoue}, \&

1467:   {Halpern}}]{Awaki:1991}

1468: Awaki, H., Koyama, K., {Inoue}, H., \& {Halpern}, J.~P. 1991, \pasj, 43, 195

1469: 

1470: \bibitem[{Bardeen \& Petterson(1975)}]{bardeen:1975}

1471: Bardeen, J.~M. \& Petterson, J.~A. 1975, \apjl, 195, L65

1472: 

1473: \bibitem[{Begelman {et~al.}(1991)Begelman, de~Kool, \& Sikora}]{Begelman:1991}

1474: Begelman, M., de~Kool, M., \& Sikora, M. 1991, \apj, 382, 416

1475: 

1476: \bibitem[{{Begelman} \& {Nath}(2005)}]{Begelman:2005}

1477: {Begelman}, M.~C. \& {Nath}, B.~B. 2005, \mnras, 361, 1387

1478: 

1479: \bibitem[{Blanco {et~al.}(1990)Blanco, Ward, \& Wright}]{Blanco:1990}

1480: Blanco, P.~R., Ward, M.~J., \& Wright, G.~S. 1990, \mnras, 242, 4P

1481: 

1482: \bibitem[{Blandford \& Payne(1982)}]{Blandford:1982}

1483: Blandford, R.~D. \& Payne, D.~G. 1982, \mnras, 199, 883

1484: 

1485: \bibitem[{Bondi(1952)}]{Bondi:1952}

1486: Bondi, H. 1952, \mnras, 112, 195

1487: 

1488: \bibitem[{Bottorff {et~al.}(1997)Bottorff, {Korista}, {Shlosman}, \&

1489:   {Blandford}}]{Bottorff:1997}

1490: Bottorff, M., {Korista}, K.~T., {Shlosman}, I., \& {Blandford}, R.~D. 1997,

1491:   \apj, 479, 200

1492: 

1493: \bibitem[{Brighenti \& Mathews(2006)}]{Brighenti:2006}

1494: Brighenti, F. \& Mathews, W.~G. 2006, \apj, 643, 120

1495: 

1496: \bibitem[{Caproni {et~al.}(2006)Caproni, {Livio}, {Abraham}, \& {Mosquera

1497:   Cuesta}}]{Caproni:2006b}

1498: Caproni, A., {Livio}, M., {Abraham}, Z., \& {Mosquera Cuesta}, H.~J. 2006,

1499:   \apj, 653, 112

1500: 

1501: \bibitem[{Caproni {et~al.}(2004)Caproni, {Mosquera Cuesta}, \&

1502:   {Abraham}}]{Caproni:2004}

1503: Caproni, A., {Mosquera Cuesta}, H.~J., \& {Abraham}, Z. 2004, \apjl, 616, L99

1504: 

1505: \bibitem[{Castor(1970)}]{Castor:1970}

1506: Castor, J.~I. 1970, \mnras, 149, 111

1507: 

1508: \bibitem[{Castor {et~al.}(1975)Castor, Abbott, \& Klein}]{Castor:1975}

1509: Castor, J.~I., Abbott, D.~C., \& Klein, R.~I. 1975, \apj, 195, 157

1510: 

1511: \bibitem[{Ciotti \& Ostriker(1997)}]{ciotti:1997}

1512: Ciotti, L. \& Ostriker, J.~P. 1997, \apjl, 487, L105

1513: 

1514: \bibitem[{Ciotti \& Ostriker(2001)}]{ciotti:2001}

1515: ---. 2001, \apj, 551, 131

1516: 

1517: \bibitem[{Ciotti \& Ostriker(2007)}]{ciotti:2007}

1518: ---. 2007, \apj, 665, 1038

1519: 

1520: \bibitem[{Clarke(1996)}]{Clarke:1996}

1521: Clarke, D.~A. 1996, \apj, 457, 291

1522: 

1523: \bibitem[{Condon \& Mitchell(1984)}]{Condon:1984}

1524: Condon, J.~J. \& Mitchell, K.~J. 1984, \apj, 276, 472

1525: 

1526: \bibitem[{Cox {et~al.}(1991)Cox, Gull, \& Scheuer}]{Cox:1991}

1527: Cox, C.~I., Gull, S.~F., \& Scheuer, P.~A.~G. 1991, \mnras, 252, 558

1528: 

1529: \bibitem[{Dalla~Vecchia {et~al.}(2004)Dalla~Vecchia, {Bower}, {Theuns},

1530:   {Balogh}, {Mazzotta}, \& {Frenk}}]{DallaVecchia:2004}

1531: Dalla~Vecchia, C., {Bower}, R.~G., {Theuns}, T., {Balogh}, M.~L., {Mazzotta},

1532:   P., \& {Frenk}, C.~S. 2004, \mnras, 355, 995

1533: 

1534: \bibitem[{Emmering {et~al.}(1992)Emmering, {Blandford}, \&

1535:   {Shlosman}}]{Emmering:1992}

1536: Emmering, R.~T., {Blandford}, R.~D., \& {Shlosman}, I. 1992, \apj, 385, 460

1537: 

1538: \bibitem[{Everett \& Murray(2007)}]{Everett:2007}

1539: Everett, J.~E. \& Murray, N. 2007, \apj, 656, 93

1540: 

1541: \bibitem[{{Fabian} {et~al.}(2006){Fabian}, {Celotti}, \&

1542:   {Erlund}}]{Fabian:2006b}

1543: {Fabian}, A.~C., {Celotti}, A., \& {Erlund}, M.~C. 2006, \mnras, 373, L16

1544: 

1545: \bibitem[{Fabian {et~al.}(2006)Fabian, {Sanders}, {Taylor}, {Allen},

1546:   {Crawford}, {Johnstone}, \& {Iwasawa}}]{Fabian:2006}

1547: Fabian, A.~C., {Sanders}, J.~S., {Taylor}, G.~B., {Allen}, S.~W., {Crawford},

1548:   C.~S., {Johnstone}, R.~M., \& {Iwasawa}, K. 2006, \mnras, 366, 417

1549: 

1550: \bibitem[{Florido {et~al.}(1990)Florido, Battaner, \&

1551:   Sanchez-Saavedra}]{Florido:1990}

1552: Florido, E., Battaner, E., \& Sanchez-Saavedra, M.~L. 1990, \apss, 164, 131

1553: 

1554: \bibitem[{Fontanot {et~al.}(2006)Fontanot, {Monaco}, {Cristiani}, \&

1555:   {Tozzi}}]{Fontanot:2006}

1556: Fontanot, F., {Monaco}, P., {Cristiani}, S., \& {Tozzi}, P. 2006, \mnras, 373,

1557:   1173

1558: 

1559: \bibitem[{Fragile \& Anninos(2005)}]{Fragile:2005}

1560: Fragile, P.~C. \& Anninos, P. 2005, \apj, 623, 347

1561: 

1562: \bibitem[{Gower {et~al.}(1982)Gower, {Gregory}, {Unruh}, \&

1563:   {Hutchings}}]{Gower:1982}

1564: Gower, A.~C., {Gregory}, P.~C., {Unruh}, W.~G., \& {Hutchings}, J.~B. 1982,

1565:   \apj, 262, 478

1566: 

1567: \bibitem[{Hardee \& Clarke(1992)}]{Hardee:1992}

1568: Hardee, P.~E. \& Clarke, D.~A. 1992, \apjl, 400, L9

1569: 

1570: \bibitem[{Hardee \& Clarke(1995)}]{Hardee:1995b}

1571: ---. 1995, \apjl, 451, L25

1572: 

1573: \bibitem[{Hardee {et~al.}(1995)Hardee, Clarke, \& Howell}]{Hardee:1995}

1574: Hardee, P.~E., Clarke, D.~A., \& Howell, D.~A. 1995, \apj, 441, 644

1575: 

1576: \bibitem[{Hardee {et~al.}(1994)Hardee, Cooper, \& Clarke}]{Hardee:1994}

1577: Hardee, P.~E., Cooper, M.~A., \& Clarke, D.~A. 1994, \apj, 424, 126

1578: 

1579: \bibitem[{Hardee {et~al.}(2001)Hardee, {Hughes}, {Rosen}, \&

1580:   {Gomez}}]{Hardee:2001}

1581: Hardee, P.~E., {Hughes}, P.~A., {Rosen}, A., \& {Gomez}, E.~A. 2001, \apj, 555,

1582:   744

1583: 

1584: \bibitem[{Hayes {et~al.}(2006)Hayes, {Norman}, {Fiedler}, {Bordner}, {Li},

1585:   {Clark}, {ud-Doula}, \& {Mac Low}}]{Hayes:2006}

1586: Hayes, J.~C., {Norman}, M.~L., {Fiedler}, R.~A., {Bordner}, J.~O., {Li}, P.~S.,

1587:   {Clark}, S.~E., {ud-Doula}, A., \& {Mac Low}, M.-M. 2006, \apjs, 165, 188

1588: 

1589: \bibitem[{Hopkins {et~al.}(2005)Hopkins, {Hernquist}, {Cox}, {Di Matteo},

1590:   {Martini}, {Robertson}, \& {Springel}}]{Hopkins:2005}

1591: Hopkins, P.~F., {Hernquist}, L., {Cox}, T.~J., {Di Matteo}, T., {Martini}, P.,

1592:   {Robertson}, B., \& {Springel}, V. 2005, \apj, 630, 705

1593: 

1594: \bibitem[{Horne \& Baliunas(1986)}]{Horne:1986}

1595: Horne, J.~H. \& Baliunas, S.~L. 1986, \apj, 302, 757

1596: 

1597: \bibitem[{Hughes {et~al.}(2002)Hughes, {Miller}, \& {Duncan}}]{Hughes:2002}

1598: Hughes, P.~A., {Miller}, M.~A., \& {Duncan}, G.~C. 2002, \apj, 572, 713

1599: 

1600: \bibitem[{Hunstead {et~al.}(1984)Hunstead, {Murdoch}, {Condon}, \&

1601:   {Phillips}}]{Hunstead:1984}

1602: Hunstead, R.~W., {Murdoch}, H.~S., {Condon}, J.~J., \& {Phillips}, M.~M. 1984,

1603:   \mnras, 207, 55

1604: 

1605: \bibitem[{Hutchings {et~al.}(1988)Hutchings, Price, \& Gower}]{Hutchings:1988}

1606: Hutchings, J.~B., Price, R., \& Gower, A.~C. 1988, \apj, 329, 122

1607: 

1608: \bibitem[{Katz(1997)}]{katz:1997}

1609: Katz, J.~I. 1997, \apj, 478, 527

1610: 

1611: \bibitem[{King(2003)}]{king:2003}

1612: King, A. 2003, \apjl, 596, L27

1613: 

1614: \bibitem[{King {et~al.}(2005)King, {Lubow}, {Ogilvie}, \&

1615:   {Pringle}}]{King:2005}

1616: King, A.~R., {Lubow}, S.~H., {Ogilvie}, G.~I., \& {Pringle}, J.~E. 2005,

1617:   \mnras, 363, 49

1618: 

1619: \bibitem[{Kochanek \& Hawley(1990)}]{Kochanek:1990}

1620: Kochanek, C.~S. \& Hawley, J.~F. 1990, \apj, 350, 561

1621: 

1622: \bibitem[{K\"{o}nigl \& Kartje(1994)}]{Konigl:1994}

1623: K\"{o}nigl, A. \& Kartje, J.~F. 1994, \apj, 434, 446

1624: 

1625: \bibitem[{Krolik(1999)}]{Krolik:1999}

1626: Krolik, J.~H. 1999, {Active Galactic Nuclei: From the Central Black Hole to the

1627:   Galactic Environment} (Princeton: Princeton Univ. Press)

1628: 

1629: \bibitem[{Lai(2003)}]{Lai:2003}

1630: Lai, D. 2003, \apjl, 591, L119

1631: 

1632: \bibitem[{Laor \& Draine(1993)}]{Laor:1993}

1633: Laor, A. \& Draine, B.~T. 1993, \apj, 402, 441

1634: 

1635: \bibitem[{Linfield(1981)}]{Linfield:1981}

1636: Linfield, R. 1981, \apj, 250, 464

1637: 

1638: \bibitem[{Lu \& Zhou(2005)}]{Lu:2005}

1639: Lu, J.-F. \& Zhou, B.-Y. 2005, \apjl, 635, L17

1640: 

1641: \bibitem[{Lucy(1971)}]{Lucy:1971}

1642: Lucy, L.~B. 1971, \apj, 163, 95

1643: 

1644: \bibitem[{Lynden-Bell(1969)}]{Lynden-Bell:1969}

1645: Lynden-Bell, D. 1969, \nat, 223, 690

1646: 

1647: \bibitem[{Maloney {et~al.}(1996)Maloney, {Begelman}, \&

1648:   {Pringle}}]{Maloney:1996}

1649: Maloney, P.~R., {Begelman}, M.~C., \& {Pringle}, J.~E. 1996, \apj, 472, 582

1650: 

1651: \bibitem[{McNamara {et~al.}(2005)McNamara, {Nulsen}, {Wise}, {Rafferty},

1652:   {Carilli}, {Sarazin}, \& {Blanton}}]{McNamara:2005}

1653: McNamara, B.~R., {Nulsen}, P.~E.~J., {Wise}, M.~W., {Rafferty}, D.~A.,

1654:   {Carilli}, C., {Sarazin}, C.~L., \& {Blanton}, E.~L. 2005, \nat, 433, 45

1655: 

1656: \bibitem[{Miller \& Goodrich(1990)}]{Miller:1990}

1657: Miller, J.~S. \& Goodrich, R.~W. 1990, \apj, 355, 456

1658: 

1659: \bibitem[{Mizuno {et~al.}(2007)Mizuno, Hardee, \& Nishikawa}]{Mizuno:2007}

1660: Mizuno, Y., Hardee, P., \& Nishikawa, K.-I. 2007, \apj, 662, 835

1661: 

1662: \bibitem[{Murray {et~al.}(1995)Murray, {Chiang}, {Grossman}, \&

1663:   {Voit}}]{Murray:1995b}

1664: Murray, N., {Chiang}, J., {Grossman}, S.~A., \& {Voit}, G.~M. 1995, \apj, 451,

1665:   498

1666: 

1667: \bibitem[{Murray {et~al.}(2005)Murray, Quataert, \& Thompson}]{Murray:2005}

1668: Murray, N., Quataert, E., \& Thompson, T.~A. 2005, \apj, 618, 569

1669: 

1670: \bibitem[{Nelson \& Papaloizou(2000)}]{Nelson:2000}

1671: Nelson, R.~P. \& Papaloizou, J.~C.~B. 2000, \mnras, 315, 570

1672: 

1673: \bibitem[{Petterson(1977)}]{Petterson:1977}

1674: Petterson, J.~A. 1977, \apj, 216, 827

1675: 

1676: \bibitem[{Phinney(1989)}]{Phinney:1989}

1677: Phinney, E.~S. 1989, in {Theory of Accretion Disks}, ed. F.~Meyer (NATO ASI

1678:   Ser. C, 290; Dordrecht: Kluwer), 457

1679: 

1680: \bibitem[{Pier \& Krolik(1992)}]{Pier:1992}

1681: Pier, E.~A. \& Krolik, J.~H. 1992, \apjl, 399, L23

1682: 

1683: \bibitem[{Press {et~al.}(1992)Press, {Teukolsky}, {Vetterling}, \&

1684:   {Flannery}}]{Press:1992}

1685: Press, W.~H., {Teukolsky}, S.~A., {Vetterling}, W.~T., \& {Flannery}, B.~P.

1686:   1992, {Numerical recipes in FORTRAN. The art of scientific computing}

1687:   (Cambridge: University Press, 1992, 2nd ed.)

1688: 

1689: \bibitem[{Pringle(1996)}]{Pringle:1996}

1690: Pringle, J.~E. 1996, \mnras, 281, 357

1691: 

1692: \bibitem[{Proga(2007)}]{Proga:2007}

1693: Proga, D. 2007, \apj, 661, 693

1694: 

1695: \bibitem[{Proga \& Begelman(2003)}]{Proga:2003b}

1696: Proga, D. \& Begelman, M.~C. 2003, \apj, 592, 767

1697: 

1698: \bibitem[{Proga \& Kallman(2004)}]{Proga:2004}

1699: Proga, D. \& Kallman, T.~R. 2004, \apj, 616, 688

1700: 				   

1701: \bibitem[{Proga {et~al.}(2007)Proga, Ostriker, \&

1702:     Kurosawa}]{Proga:2007b}

1703: Proga, D., Ostriker, S.~P., \& Kurosawa, R., 2007, preprint (astro-ph/0708.4037)

1704: 

1705: \bibitem[{Proga {et~al.}(1998)Proga, Stone, \& Drew}]{proga:1998}

1706: Proga, D., Stone, J.~M., \& Drew, J.~E. 1998, \mnras, 295, 595

1707: 

1708: \bibitem[{Proga {et~al.}(2000)Proga, Stone, \& Kallman}]{Proga:2000}

1709: Proga, D., Stone, J.~M., \& Kallman, T.~R. 2000, \apj, 543, 686

1710: 

1711: \bibitem[{Quilis {et~al.}(2001)Quilis, Bower, \& Balogh}]{Quilis:2001}

1712: Quilis, V., Bower, R.~G., \& Balogh, M.~L. 2001, \mnras, 328, 1091

1713: 

1714: \bibitem[{Quillen \& Bower(1999)}]{Quillen:1999}

1715: Quillen, A.~C. \& Bower, G.~A. 1999, \apj, 522, 718

1716: 

1717: \bibitem[{Romero {et~al.}(2000)Romero, {Chajet}, {Abraham}, \&

1718:   {Fan}}]{Romero:2000}

1719: Romero, G.~E., {Chajet}, L., {Abraham}, Z., \& {Fan}, J.~H. 2000, \aap, 360, 57

1720: 

1721: \bibitem[{Roos(1988)}]{Roos:1988}

1722: Roos, N. 1988, \apj, 334, 95

1723: 

1724: \bibitem[{Rybicki \& Hummer(1978)}]{Rybicki:1978}

1725: Rybicki, G.~B. \& Hummer, D.~G. 1978, \apj, 219, 654

1726: 

1727: \bibitem[{Sazonov {et~al.}(2005)Sazonov, {Ostriker}, {Ciotti}, \&

1728:   {Sunyaev}}]{Sazonov:2005}

1729: Sazonov, S.~Y., {Ostriker}, J.~P., {Ciotti}, L., \& {Sunyaev}, R.~A. 2005,

1730:   \mnras, 358, 168

1731: 

1732: \bibitem[{Schreier {et~al.}(1972)Schreier, {Giacconi}, {Gursky}, {Kellogg}, \&

1733:   {Tananbaum}}]{Schreier:1972}

1734: Schreier, E., {Giacconi}, R., {Gursky}, H., {Kellogg}, E., \& {Tananbaum}, H.

1735:   1972, \apjl, 178, L71

1736: 

1737: \bibitem[{Shakura \& Sunyaev(1973)}]{shakura:1973}

1738: Shakura, N.~I. \& Sunyaev, R.~A. 1973, \aap, 24, 337

1739: 

1740: \bibitem[{Shlosman {et~al.}(1985)Shlosman, Vitello, \&

1741:     Shaviv}]{Shlosman:1985}

1742: Shlosman, I., Vitello, P.~A., \& Shaviv, G. 1985, \apj, 294, 96

1743: 

1744: 

1745: \bibitem[{{Silk}(2005)}]{Silk:2005}

1746: {Silk}, J. 2005, \mnras, 364, 1337

1747: 

1748: \bibitem[{{Silk} \& {Rees}(1998)}]{Silk:1998}

1749: {Silk}, J. \& {Rees}, M.~J. 1998, \aap, 331, L1

1750: 

1751: 

1752: \bibitem[{Sillanp\"{a}\"{a} {et~al.}(1988)Sillanp\"{a}\"{a}, Haarala,

1753:   {Valtonen}, {Sundelius}, \& {Byrd}}]{Sillanpaa:1988}

1754: Sillanp\"{a}\"{a}, A., Haarala, S., {Valtonen}, M.~J., {Sundelius}, B., \&

1755:   {Byrd}, G.~G. 1988, \apj, 325, 628

1756: 

1757: \bibitem[{Sobolev(1957)}]{Sobolev:1957}

1758: Sobolev, V.~V. 1957, Soviet Astronomy, 1, 678

1759: 

1760: \bibitem[{Springel {et~al.}(2005)Springel, Di~Matteo, \&

1761:   Hernquist}]{Springel:2005}

1762: Springel, V., Di~Matteo, T., \& Hernquist, L. 2005, \apjl, 620, L79

1763: 

1764: \bibitem[{Sternberg \& Soker(2007)}]{Sternberg:2007}

1765: Sternberg, A. \& Soker, N. 2007, preprint (astro-ph/0708.0932)

1766: 

1767: \bibitem[{Stone \& Norman(1992)}]{Stone:1992}

1768: Stone, J.~M. \& Norman, M.~L. 1992, \apjs, 80, 753

1769: 

1770: \bibitem[{{Thacker} {et~al.}(2006){Thacker}, {Scannapieco}, \&

1771:   {Couchman}}]{Thacker:2006}

1772: {Thacker}, R.~J., {Scannapieco}, E., \& {Couchman}, H.~M.~P. 2006, \apj, 653,

1773:   86

1774: 

1775: \bibitem[{Tremblay {et~al.}(2006)Tremblay, {Quillen}, {Floyd}, {Noel-Storr},

1776:   {Baum}, {Axon}, {O'Dea}, {Chiaberge}, {Macchetto}, {Sparks}, {Miley},

1777:   {Capetti}, {Madrid}, \& {Perlman}}]{Tremblay:2006}

1778: Tremblay, G.~R., {Quillen}, A.~C., {Floyd}, D.~J.~E., {Noel-Storr}, J., {Baum},

1779:   S.~A., {Axon}, D., {O'Dea}, C.~P., {Chiaberge}, M., {Macchetto}, F.~D.,

1780:   {Sparks}, W.~B., {Miley}, G.~K., {Capetti}, A., {Madrid}, J.~P., \&

1781:   {Perlman}, E. 2006, \apj, 643, 101

1782: 

1783: \bibitem[{Tremonti {et~al.}(2007)Tremonti, Moustakas, \&

1784:   Diamond-Stanic}]{Tremonti:2007}

1785: Tremonti, C.~A., Moustakas, J., \& Diamond-Stanic, A.~M. 2007, \apj,

1786:   663, L77

1787: 

1788: \bibitem[{Veilleux {et~al.}(1993)Veilleux, Tully, \&

1789:   Bland-Hawthorn}]{Veilleux:1993}

1790: Veilleux, S., Tully, R.~B., \& Bland-Hawthorn, J. 1993, \aj, 105, 1318

1791: 

1792: \bibitem[{Vernaleo \& Reynolds(2006)}]{Vernaleo:2006}

1793: Vernaleo, J.~C. \& Reynolds, C.~S. 2006, \apj, 645, 83

1794: 

1795: \bibitem[{Wang {et~al.}(2006)Wang, Chen, \& Hu}]{Wang:2006}

1796: Wang, J.-M., Chen, Y.-M., \& Hu, C. 2006, \apjl, 637, L85

1797: 

1798: \bibitem[{Weymann {et~al.}(1982)Weymann, Scott, Schiano, \&

1799:   {Christiansen}}]{Weymann:1982}

1800: Weymann, R.~J., Scott, J.~S., Schiano, A.~V.~R., \& {Christiansen}, W.~A. 1982,

1801:   \apj, 262, 497

1802: 

1803: \bibitem[{Xu {et~al.}(2000)Xu, Hardee, \& Stone}]{Xu:2000}

1804: Xu, J., Hardee, P.~E., \& Stone, J.~M. 2000, \apj, 543, 161

1805: 

1806: \bibitem[{Zanni {et~al.}(2005)Zanni, {Murante}, {Bodo}, {Massaglia}, {Rossi},

1807:   \& {Ferrari}}]{Zanni:2005}

1808: Zanni, C., {Murante}, G., {Bodo}, G., {Massaglia}, S., {Rossi}, P., \&

1809:   {Ferrari}, A. 2005, \aap, 429, 399

1810: 

1811: \bibitem[{Zensus(1997)}]{Zensus:1997}

1812: Zensus, J.~A. 1997, \araa, 35, 607

1813: 

1814: \end{thebibliography}

1815: 

1816: 

1817: 

1818: 

1819: \end{document}

1820: 

1821: % LocalWords:  Las rk dproga pc collimated AGNs Lynden Quilis DallaVecchia Laor

1822: % LocalWords:  Zanni Vernaleo ciotti Sazonov Springel Brighenti Fontanot Konigl

1823: % LocalWords:  Tremonti Blandford Emmering Bottorff Weymann Begelman Shlosman

1824: % LocalWords:  Phinney Antonucci Awaki Blanco Krolik Arav Florido Hutchings BH

1825: % LocalWords:  Linfield Appl Zensus Condon Hunstead Tremblay Roos Gower bardeen

1826: % LocalWords:  Veilleux Petterson Schreier Sillanpaa katz Caproni Pringle Aly

1827: % LocalWords:  Maloney Armitage Lai Quillen Kochanek axisymmetric Hardee Xu IGM

1828: % LocalWords:  Aloy Mizuno Sternberg Schwarzschild cartesian shakura Eulerian

1829: % LocalWords:  advection Dt CAK Rybicki th dv dl Keplerian proga eqs eq cGM kT

1830: % LocalWords:  precesses cccccccccc Bondi iso axisymmetry dx dy maxima minima

1831: % LocalWords:  dV Lomb Scargle periodgram NAS AST Xeon cibrodzka Agnieszka et

1832: % LocalWords:  Janiuk natexlab al mez Agudo ller Ib ez Koyama Inoue Halpern de

1833: % LocalWords:  Kool Sikora Korista Livio Mosquera Cuesta Ostriker Scheuer Dalla

1834: % LocalWords:  Vecchia Theuns Balogh Mazzotta Frenk Johnstone Iwasawa Battaner

1835: % LocalWords:  Saavedra Cristiani Tozzi Anninos Unruh Rosen Fiedler Bordner ud

1836: % LocalWords:  Doula Hernquist Baliunas Lubow Ogilvie Hawley nigl Kartje Draine

1837: % LocalWords:  Zhou Nulsen Carilli Sarazin Blanton Nishikawa Chiang Grossman Hu

1838: % LocalWords:  Voit Quataert Papaloizou ASI Ser Dordrecht Kluwer Teukolsky Bodo

1839: % LocalWords:  Vetterling Flannery Kallman astro Sunyaev Giacconi Gursky Shaviv

1840: % LocalWords:  Tananbaum Vitello Sillanp Haarala Valtonen Sundelius Soker Storr

1841: % LocalWords:  O'Dea Chiaberge Macchetto Miley Capetti Perlman Moustakas Stanic

1842: % LocalWords:  ArXiv Schiano Christiansen Murante Massaglia Rossi

1843: