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

2: %\documentclass{emulateapj}

3: %\usepackage{apjfonts}

4: %\usepackage{graphicx}

5: %\usepackage[usenames,dvips]{color}

6: %\documentclass[manuscript]{aastex}

7: 

8: %    For ApJ emulation mode use one of the following two:

9: %

10: %\documentclass[12pt]{aastex}             % two column (default)

11: %\documentclass[12pt,onecolumn]{aastex}   % single column

12: %                  plus:

13: %\usepackage{natbib,emulateapj5,psfig}

14: 

15: \def\msun{{~M}_{\odot}}

16: \def\ergs{{\,erg\,s^{-1}}}

17: \def\medd{{\dot M_{Edd}}}

18: \def\mdot{{\dot M}}

19: \def\be{\begin{equation}}

20: \def\ee{\end{equation}}

21: 

22: \begin{document}

23: 

24: \shorttitle{Outflow influences on ADAF} \shortauthors{F. G., Xie;

25: F., Yuan}

26: 

27: \title{The influences of outflow on the dynamics of inflow}

28: 

29: \author{Fu-Guo Xie,\altaffilmark{1,}\altaffilmark{2,}\altaffilmark{3}

30:  Feng Yuan\altaffilmark{1,}\altaffilmark{2}}

31: 

32: \altaffiltext{1}{Shanghai Astronomical Observatory, 80 Nandan Road,

33: Shanghai 200030, China} \altaffiltext{2}{Joint Institute for Galaxy

34: and Cosmology (JOINGC) of SHAO and USTC} \altaffiltext{3}{Graduate

35: School of the Chinese Academy of Sciences, Beijing 100039, China;

36: fgxie@shao.ac.cn}

37: 

38: \begin{abstract}

39: 

40: Both numerical simulations and observations indicate that in an

41: advection-dominated accretion flow most of the accretion material

42: supplied at the outer boundary will not reach the inner boundary.

43: Rather, they are lost via outflow. Previously, the influence of

44: outflow on the dynamics of inflow is taken into account only by

45: adopting a radius-dependent mass accretion rate $\dot{M}=\dot{M}_0

46: (r/r_{\rm out})^s$ with $s>0$. In this paper, based on a 1.5 dimensional

47: description to the accretion flow, we investigate this problem in

48: more detail by considering the interchange of mass, radial and

49: azimuthal momentum, and the energy between the outflow and inflow.

50: The physical quantities of the outflow is parameterized based on our

51: current understandings to the properties of outflow mainly from

52: numerical simulations of accretion flows. Our results indicate that

53: under reasonable assumptions to the properties of outflow, the main

54: influence of outflow has been properly included by adopting

55: $\dot{M}=\dot{M}_0 (r/r_{\rm out})^s$.

56: 

57: \end{abstract}

58: 

59: \keywords{accretion, accretion disks --- black hole physics ---

60: hydrodynamics --- ISM: jets and outflows}

61: 

62: \section{Introduction}

63: 

64: There are now strong observational evidences for the existence of

65: outflow in accretion flow system. One of the best examples comes

66: from Sgr A*, the supermassive black hole located at our Galactic

67: center. The accretion flow in this source is likely in the form of

68: the advection-dominated accretion flow (ADAF, or radiatively

69: inefficient accretion flow; Yuan, Quataert \& Narayan 2003). On one

70: hand, from the observational results from $Chandra$ combined with

71: the Bondi accretion theory we can calculate the value of the mass

72: accretion rate at the outer boundary---the Bondi radius. On the

73: other hand, radio polarization observations constrain the accretion

74: rate at the innermost region of the accretion flow nearly two orders

75: of magnitude lower than that determined at the Bondi radius (e.g.,

76: Marrone et al. 2006). This implies that about $99\%$ of the material

77: available at the Bondi radius will not finally enter into the black

78: hole horizon, rather, they must be lost in the form of outflow.

79: Outflow seems to exist also in more luminous sources whose accretion

80: mode is different from the ADAF. For example, the blueshifted

81: absorption lines, which indicates the existence of outflowing

82: materials, have been detected in the X-ray spectrum of some Seyfert

83: 1 sources (e.g., NGC 3783: Kaspi et al. 2001) and more spectacularly

84: in quasars (e.g., PG 1115+80: Chartas, Brandt \& Gallagher 2003).

85: The existence of outflow has been paid more and more attention

86: recently in the field of galaxy formation because of its feedback

87: effect in the coevolution of galaxy and the central active galactic

88: nuclei (e.g., Silk \& Rees 1998; Granato et al. 2004; Springel et

89: al. 2005).

90: 

91: Many work has been done on the origin and dynamics of outflow

92: (e.g., Xu \& Chen 1997; Blandford \& Begelman 2004;

93: Xue \& Wang 2005) in the frame of self-similar hydrodynamical

94: solution. Magnetic field, especially its

95: poloidal component, may presumably serve as the most promising

96: mechanism on producing outflow, as proposed by, e.g., Blandford \&

97: Payne (1982). This has been confirmed in non-radiative

98: magnetohydrodynamical (MHD) numerical simulations of accretion flows

99: (e.g., Stone \& Pringle 2001; Igumenshchev et al. 2003; Vlahakis \&

100: K$\ddot{o}$nigl 2003; McKinney 2006). Radiation pressure could be

101: another important mechanism in luminous accretion disk (Proga 2003).

102: But even in the absence of magnetic field and strong radiation,

103: outflow is likely present in ADAFs. This is first proposed

104: from analytical argument that the Bernoulli parameter of an ADAF

105: is large or even positive because of the small radiative energy

106: loss. This implies that the gas is inclined to escape once they are

107: perturbed (Narayan \& Yi 1994; Blandford \&

108: Begelman 1999). This suggestion was later confirmed by numerical

109: simulation by Stone, Pringle \& Begelman (1999).

110: 

111: Since the outflow is likely very strong (Misra \& Taam 2001), they

112: may provide an additional important sink of angular momentum and energy.

113: So it is important to investigate its dynamical influence to

114: inflow. This problem has been investigated by Kuncic \& Bicknell (2007)

115: in the context of the standard thin disk. In this paper we focus on ADAFs.

116: Blandford \& Begelman (1999, hereafter BB99) examined this

117: question through a one-dimensional self-similar approach. A

118: phenomenological way was adopted in which they parameterized the

119: rate at which mass, angular momentum and energy are extracted

120: through outflow, regardless the mechanism of the formation of

121: outflow.

122: 

123: BB99 gives a quite general description, covering a broad kind of

124: outflow including Poynting flux whose mass flux is zero while energy

125: flux is not. It is based on self-similar assumption. For the purpose

126: of application and comparison with observation, however, we need to

127: discuss it based on global solutions, because self-similar solution

128: is too simplified to be used to calculate the emitted spectrum.

129: Quataert \& Narayan (1999) presented the first effort on this

130: aspect. They calculate the global solution of inflow when strong

131: outflow is present by using a radius-dependent mass accretion rate,

132: $\dot{M}\propto r^s$ while keeping all other equations describing

133: inflow such as the momentum and energy equations unchanged. This is

134: roughly equivalent to assuming that the specific angular momentum

135: and energy of outflow is identical to the inflow at the same radius

136: where outflow is launched (see e.g., eqs.(\ref{mass})-(\ref{ion})).

137: This approach is subsequently adopted in almost all following works

138: (e.g., Yuan, Quataert \& Narayan 2003). In this paper, we refer to

139: this treatment as standard treatment.

140: 

141:  While the approximation of Quataert

142: \& Narayan (1999; see also Yuan, Quataert \& Narayan 2003) may

143: capture the most important influence of outflow to inflow, it is not

144: obvious in what degree we can use this approximation or how good

145: this approximation is when we compare the theoretical prediction

146: such as the spectrum to observations. This is the aim of the present

147: paper. More specifically, by considering the conservations of fluxes

148: of mass, momentum, and energy of the combined inflow/outflow system,

149: we focus on the influence of outflow on the dynamics of inflow. We

150: will use a ``1.5 dimension'' description of the accretion flow,

151: which means the height-integrated equations will be used instead of

152: a fully two-dimensional description, but the conservation equations

153: take into account the outflow in the vertical direction as well. The

154: paper is organized as follows: in \S 2 we present basic equations

155: for our model and discuss the main properties of outflow. The

156: calculation results are presented in \S 3. The last section is

157: devoted to a summary.

158: 

159: \section{Accretion Model with Outflow }

160: 

161: \subsection{Basic Equations}

162: 

163: We adopt a cylinder coordinate ($r, \phi, z$) to describe a steady

164: axisymmetric ($\partial /\partial t= \partial / \partial\phi = 0 $)

165: accretion flow. The Paczy\'nski \& Wiita potential (Paczy\'nski \&

166: Wiita 1980) $\psi = -GM_{BH}/(r-r_g)$ is adopted to mimic the

167: geometry of a Schwarzschild black hole, with $M_{BH}$ is the mass of

168: black hole and $r_g \equiv 2 G M_{BH}/c^2$ is the Schwarzschild

169: radius of the black hole. As shown in Fig. \ref{scheme}, we divide

170: the whole accretion flow at each radius into two parts, i.e., inflow

171: and outflow. For the inflow, we assume a hydrostatic balance in

172: vertical direction ($\upsilon_z = 0$ for the inflow) and assume all

173: quantities such as the radial and azimuthal velocity ($\upsilon_r$

174: and $\upsilon_\phi$), ions and electron temperature ($T_i, T_e$) and

175: the sound speed ($c_s$) are only functions of radius $r$. Such an

176: isothermal assumption in the vertical direction results in a density

177: distribution of $\rho(r,z) = \rho(r,0) \exp(-z^2/2 H^2)$ in the

178: inflow, where $H = c_s/\Omega_K$ is the vertical scale height of

179: inflow\footnote{This kind of vertical density structure is based on

180: a first-order expand of Paczy\'nski \& Wiita potential, which is not

181: exact far away from the equatorial plane (Gu \& Lu 2007).}. We set

182: $z=H$ as the surface where an outflow launches. The vertical

183: gradients of above quantities are absorbed by their discontinuity

184: between inflow and outflow at this surface, except that the density

185: distribution is continuous and the density of outflow at $z=H$ is

186: then $e^{-1/2} \rho(r,0)$. Note that the vertical velocity of

187: outflow $\upsilon_{z,w}$, will ``compress'' the inflow because of

188: momentum conservation, thus the vertical scale height may be

189: smaller. We neglect this effect here.

190: 

191: From compressible Navier-Stokes equations, we can write the

192: equations of the conservations of mass, momentum, and energy for the inflow as

193: follows (see Appendix A for details):

194: \begin{equation}

195: \label{mass} \frac{d \mdot(r)}{d r} = \eta_1 4 \pi r \rho

196: \upsilon_{z,w}

197: \end{equation}

198: \begin{equation}

199: \label{radial} \upsilon_r \frac{d \upsilon_r}{d r} + \eta_1

200: \upsilon_{z,w} \frac{\upsilon_{r,w}-\upsilon_r}{H} = r

201: (\Omega^2-\Omega_k^2) -\frac{1}{\rho}\frac{d P}{d r} -\frac{1}{2}

202: \frac{d c_s^2}{d r}

203: \end{equation}

204: \begin{equation}

205: \label{angular} \rho r \upsilon_r  \frac{d}{d r}(r^2 \Omega)+ \eta_1

206: r^2 \rho \upsilon_{z,w}

207: \frac{\upsilon_{\phi,w}-\upsilon_\phi}{H}=\frac{1}{H}\frac{d}{d r}

208: (r^2 H \tau_{\phi r})

209: \end{equation}

210: \begin{equation}

211: \label{electron} \rho \upsilon_r(\frac{d \epsilon_e}{d

212: r}-\frac{p_e}{\rho^2}\frac{d \rho}{d r}) + \eta_1 \rho

213: \upsilon_{z,w} \frac{\epsilon_{e,w}-\epsilon_e}{H}= \delta q^+

214: +q_{ie} - q^-

215: \end{equation}

216: \begin{equation}

217: \label{ion} \rho \upsilon_r(\frac{d \epsilon_i}{d

218: r}-\frac{p_i}{\rho^2}\frac{d \rho}{d r}) + \eta_1 \rho

219: \upsilon_{z,w} \frac{\epsilon_{i,w}-\epsilon_i}{H}= (1-\delta) q^+ -

220: q_{ie}

221: \end{equation}

222: Here all quantities have their usual meanings. The specific internal

223: energy of electrons and ions are $\epsilon_e, \epsilon_i$,

224: respectively. The pressure $P$ is the sum of gas and magnetic

225: pressure $P=P_{gas} + P_{mag}$. The inflow's accretion rate is

226: defined as $\dot M(r) \equiv -4 \pi r \rho \upsilon_r H$ and $\eta_1

227: \equiv \rho_w/\overline{\rho}$ is the density ratio of outflow and

228: inflow (Appendix A). Parameter $\delta$ describes the fraction of

229: the turbulent energy dissipation rate $q^+$ ($\equiv \tau_{\phi r} r

230: d \Omega/d r$) that heats electrons directly. Energy transfers from

231: ions to electrons through Coulomb collisions at a volume rate

232: $q_{ie}$, and radiative cooling rate is denoted by $q^-$. The

233: quantities with subscript {\it w} denote the quantities of

234: wind/outflow just away from the launching surface $z = H$. We take

235: the $\alpha$ viscosity description for the stress tensor $\tau_{\phi

236: r}$ (Shakura \& Sunyaev 1973):

237: \begin{equation}

238: \label{vis} \tau_{\phi r} = - \alpha P

239: \end{equation}

240: where $\alpha$ is the dimensionless viscosity parameter. Other

241: stress tensor components are neglected for simplicity, except that

242: the $\phi z$ component is considered by taking into account the

243: angular momentum exchange between inflow and outflow at $z=H$.

244: 

245: Obviously, it is impossible to directly solve the eqs.

246: (\ref{mass})-(\ref{ion}). We therefore introduce the following

247: parameters ``$\xi$'' to evaluate the radial, azimuthal, and vertical

248: velocity, and the ion and electron temperatures of outflow in terms

249: of inflow, \be \upsilon_{r,w}=\xi_r \upsilon_{ff},\ee \be

250: \upsilon_{\phi,w}=\xi_\phi \upsilon_{\phi}, \ee \be

251: \upsilon_{z,w}=\xi_z c_s,\ee \be T_{i,w}=\xi_{T_i} T_i, \ee \be

252: T_{e,w}=\xi_{T_e} T_e.\ee Here $\upsilon_{ff}$ are the free-fall

253: velocity, $\upsilon_{\phi}$, and $c_s$ are azimuthal velocity and

254: the sound speed of the inflow, respectively. We assume that these

255: parameters are independent of radius. While this assumption is

256: simple, we think it can capture the main physics of the influence of

257: outflow in a reasonable way. Specifically, this simple assumption

258: does not mean all quantities are a power-law function of radius as

259: the usual ``power-law'' assumption  of the mass flux of inflow. If

260: we know the values of these parameters, we will be able to get the

261: global solution of eqs. (\ref{mass})-(\ref{ion}).

262: 

263: \subsection{Outflow's Properties}

264: 

265: We now estimate the properties of outflow. Generally all these

266: quantities should be a function of $z$. Here we consider the

267: properties of outflow when they are just launched or detached from inflow. All

268: their subsequent evolution should be due to outflow itself and does

269: not affect the inflow any longer.

270: 

271: The first quantity is the strength of the outflow, or the mass lost

272: rate. BB99 assume $\mdot \propto r^{s}$ with $0 \leq s < 1$. This

273: ensures that the mass accretion rate decreases while the released

274: energy increases with accretion (BB99). The strength of outflow in

275: our notion is mainly governed by $\xi_z$. The above range

276: corresponds to $0 \leq \xi_z < - \upsilon_r/\eta_1 c_s \approx 0.2$

277: (ref. eq. (\ref{ratio}), but note $s(r)$ now is a function of $r$).

278: 

279: We next consider the value of $\xi_{\phi}$. The vertical

280: distribution of angular momentum of the accretion flow is

281: complicated. Two-dimensional self-similar analysis on ADAF based on

282: hydrodynamics shows that the specific angular momentum of outflow is

283: lower than inflow (Narayan \& Yi 1995; Xu \& Chen 1997; Blandford \&

284: Begelman 2004). This result is confirmed later by numerical

285: simulations (e.g., Stone, Pringle \& Begelman 1999). However, any

286: magnetic coupling between inflow and outflow will likely lead to

287: transportation of angular momentum from the former to latter (Spruit

288: 1996; Stone \& Pringle 2001; BB99; Blandford \& Begelman 2004).

289: Therefore in this paper we explore $\xi_{\phi}$ in a range around

290: unity, $0.8<\xi_{\phi}<1.2$.

291: 

292: Hydrodynamical and MHD simulations also reveal that the specific

293: internal energy or the temperature of the gas increases from the

294: equator to higher altitude (e.g., Stone, Pringle \& Begelman 1999;

295: De Villiers et al. 2005; Beckwith, Hawley \& Krolik 2008).

296: %the vertical convection and the small-scale diffusions are

297: %effective, we may argue outflowing gas can only have slightly larger

298: %internal energy than that of the corresponding inflow; on the other

299: %hand, if these effects are compressed (e.g., convection is highly

300: %compressed by magnetic field in MHD simulations; Stone \& Pringle

301: %2001), outflow may have significant larger internal energy. Under

302: %these considerations,

303: One underlying reason may be that the gas with higher internal

304: energy may escape more easily. We therefore consider $\xi_{T_i}=1,

305: 1.5$ and $\xi_{T_e} =1, 1.5$.

306: 

307:  It is highly unclear about the radial velocity of outflow when they

308: are just launched although we somehow know how they will be

309: accelerated later. But we speculate that it should be positive, and

310: should not be larger than a fraction of the local Keplerian or

311: free-fall velocity $\upsilon_{ff}$. Fortunately, although it

312: may be important for the dynamics of outflow itself, we find that

313: the value of $\xi_r$ has minor effect on the inflow. We simply set

314: $\xi_r \equiv 0.2$ in our calculations.

315: 

316: \section{Results}

317: 

318: For our specific model, we adopt the black hole mass $M_{BH} \equiv

319: 4.0\times 10^6 \msun$, accretion rate at the outer boundary $r_{out}

320: = 10^4 r_s$ is $\mdot=1.1 \times 10^{-5} \medd$, where $\medd = {\rm

321: 10 L_{Edd}}/c^2$, is the Eddington accretion rate. The values for

322: other parameters are $\alpha = 0.1, \beta \equiv P_{gas}/P_{total} =

323: 0.9, \delta = 0.3$. These parameters are close to those in Yuan,

324: Quataert \& Narayan (2003) to model the supermassive black hole in

325: our Galactic center. There they assume $\mdot\propto r^s$ with

326: $s=0.27$ being a constant. Under our notation, we have

327: \begin{equation}

328: \label{ratio} s(r) = \frac{d \ln \mdot(r)}{d \ln r} = \eta_1 \xi_z

329: \frac{\upsilon_k}{- \upsilon_r},

330: \end{equation}

331: where $\upsilon_k$ is the Keplerian velocity. We would like to note

332: that $s(r)$ now is not a constant as in Yuan, Quataert \& Narayan (2003)

333: (or Quataert \& Narayan 1999). The slope of $\mdot(r)$ now

334: is steeper at large radius while flatter at small radius, because of

335: the quicker increase of $v_r$ compared to $\upsilon_k$. This is

336: shown in Fig. \ref{mdot}, where we adjust the parameter $\xi_z$ so

337: that the accretion rates at $r_{out}$ and horizon are the same as

338: the case of $\mdot\propto r^s$ with $s=0.48$.

339: 

340: We first investigate the effect of $\xi_z$. Fig. \ref{xiz} shows the

341: effects of various $\xi_z$ on inflow, with other outflow parameters

342: fixed at $\xi_\phi = \xi_{T_e}= \xi_{T_i} = 1.0$ and $\upsilon_{r,w}

343: = \upsilon_r$. The four plots show the Mach number, profiles of

344: density, temperature, and specific angular momentum. The dotted,

345: dashed, and long-dashed lines correspond to $\xi_z = 0.01, 0.05$ and

346: $0.15$, respectively. As the outflow becomes stronger ($\xi_z$

347: increases), the gas density decreases while the ion temperature

348: decreases. This is very similar to the case of the standard

349: treatment (with increasing $s$). The decrease of ions temperature is

350: because when more and more accretion material is lost via the

351: outflow, the density profile becomes flatter thus the compression

352: work which is an important heating mechanism for ions becomes

353: weaker. Different from the ions, the electrons temperature has no

354: obvious relation with the strength of outflow. This is because

355: different from ions the compression work in the electron energy

356: equation is about one order of magnitude smaller due to the lower

357: electron temperature (ref. eq. \ref{electron}).

358: 

359: We mentioned before that the value of the radial velocity of outflow

360: or equivalently $\xi_r$, has minor effect on the dynamics of inflow.

361: The ``kick back'' force due to the discrepancy of the radial

362: velocity between the inflow and outflow is manifested by the second

363: term in eq. (\ref{radial}). Since $\upsilon_{z,w} = \xi_z c_s,

364: H=c_s/\Omega_k$ and $\upsilon_r \sim \alpha \upsilon_k$, this term

365: is roughly $\alpha \xi_z (\ll 1)$ times of the gravitational force

366: thus can be neglected. So we simply fix $\xi_r = 0.2$ in this paper.

367: 

368: We now check how good the standard treatment is. For this purpose,

369: we first get the global solution with the standard treatment with

370: $\mdot=2\times 10^{-5} (r/r_{out})^{0.25}$ and $r_{out}=10^4r_g$.

371: We then get the global solution of eqs. (\ref{mass}) - (\ref{ion}) for

372: various sets of outflow parameters of $\xi_{\phi}=0.8, 1.0, 1.2$,

373: $\xi_{T_i}=1.0$ (referred to Case A) and $1.5$ (referred to as Case

374: B), and $\xi_{T_e}=1.5$. For each set of these parameters, we adjust

375: the value of $\xi_z$ so that the mass accretion rates at $r_{out}$

376: and black hole horizon are equal to the values in the above standard

377: treatment. By doing this, we want to focus on the influence on inflow of

378: the transportation of angular momentum and internal energy between

379: inflow and outflow, which is neglected in the standard treatment. Note the

380: profile of accretion rate in this case is similar to Fig. 2.

381: 

382: Fig. \ref{mratedyn1} shows the comparison of Case A with the

383: standard treatment. We can see that our models intend to have lower

384: densities compared to the standard treatment, although the accretion

385: rates at the outer and inner boundaries are the same. We can easily

386: understand this by looking at bottom-left panel of Fig.

387: \ref{mratedyn1}. Our solutions have higher ion temperature and lower

388: electron temperature at the inner region of the inflow. The higher

389: ion temperature is because the density profile in the inner region

390: is steeper thus the compression heating is stronger. The lower

391: electron temperature is because $\xi_{T_e}=1.5>1$ which implies that

392: some internal energy is transferred into the outflow from inflow

393: (ref. eq. \ref{electron}).

394: %In addition, the electron temperature decreases with increasing $\xi_{\phi}$.

395: %This is because a larger $\xi_{\phi}$ causes a quicker decrease of the angular

396: %momentum in the inner region, as shown in the figure. This then

397: %result in a weaker viscous heating rate in the electron energy equation.

398: 

399: Fig. \ref{mratedyn2} shows the dynamical influences of outflow for

400: Case B. Compared to Case A, the ion temperature is lower. This is

401: obviously because $\xi_{T_i}$ is larger, $\xi_{T_i}=1.5$, so some

402: internal energy is transferred from inflow to outflow. The lower ion

403: temperature results in a smaller $H$. This, combined with the

404: smaller radial velocity, make the density of inflow higher compared

405: to Case A and almost identical to that of the standard treatment, as

406: shown in the figure.

407: %Since the density profile is the same with the

408: %standard treatment while $\xi_{T_i}=1.5>1$, the ion temperature of

409: %the inflow is lower than the standard treatment.

410: We also see from

411: the figure that both the value of the specific angular momentum and

412: its slope are higher compared to the standard treatment. This

413: results in a stronger viscous heating, which somehow cancel the

414: effect of $\xi_{T_e}=1.5>1$. This is why the electron temperature is

415: roughly the same with the standard treatment while higher than that

416: of Case A.

417: 

418: From Fig. \ref{xiz} to \ref{mratedyn2}, we find that within the

419: range of the values of parameters we adopt to describe the

420: outflow, the strength of the outflow has the most significant influence

421: on the dynamics of inflow (i.e., Fig. \ref{xiz}).

422: The influences of all other properties of outflow, namely

423: the angular momentum, temperature,

424: and radial velocity, are much smaller (Figs. \ref{mratedyn1}

425: and \ref{mratedyn2}). This is the reason why the discrepancy

426: between our model (with different properties of outflow) and

427: the standard treatment, which only considers the strength of outflow but

428: assume the properties of outflow are the same with inflow, is small.

429: 

430: \section{Summary}

431: 

432: Outflow is now believed to be very significant in

433: advection-dominated accretion flow thus it is important to

434: investigate their influence on the dynamics of inflow. Previously

435: this was done by using a rather simple way. In the ``standard

436: treatment'' the only change compared to the case of no outflow is

437: that the mass accretion rate is not a constant, but a power-law

438: function of radius, $\dot{M}\propto r^s$ ($s>0$). All other

439: equations describing the accretion flow remain unchanged (e.g.,

440: Quataert \& Narayan 1999; Yuan, Quataert \& Narayan 2003).

441: 

442: In this paper, we investigate the influence of outflow in more

443: detail to check how good the above ``standard treatment'' is. We

444: have derived the height-integrated accretion equations including the

445: coupling between the inflow and outflow, to investigate the

446: influence of outflow on the dynamics of inflow. We assume

447: hydrostatic equilibrium for the inflow. For the outflow, we assume

448: they are launched just above the surface of the inflow. We

449: parameterize and estimate the quantities of outflow in terms of the

450: quantities of inflow mainly from the results of numerical

451: simulations. In this way, we reduce the numbers of unknown

452: quantities in the above inflow/outflow equations thus we are able to

453: get their global solution.

454: 

455: We have studied the influences on the dynamics of inflow of the

456: strength (via vertical velocity), the (ion and electron)

457: temperature, specific angular momentum, and the radial velocity of

458: the outflow. We find that among them the strength of the outflow is

459: the most important quantity. It can produce orders of magnitude

460: difference for the density of inflow. If the strength of outflow is

461: fixed, all other quantities of outflow can only produce a difference

462: for the density and temperature within a factor of $\sim$ two, if

463: our estimations to the properties of outflow are correct. Therefore,

464: the ``standard treatment'' is usually a good approximation.

465: 

466: The largest uncertainty in our model comes from the

467: estimations to the properties of outflow such as their temperature,

468: specific angular momentum, azimuthal and radial velocity. We

469: estimate these values from numerical simulations to accretion flows

470: which is still not exact. With the rapid development of numerical

471: simulation, our results will be significantly improved. Especially,

472: if the properties of outflow is found to be far more deviated

473: from the inflow than those adopted in the present paper

474: (e.g., $\xi_{T_e} >> 1$), the standard treatment will be not enough

475: although we believe that our conclusion that the strength of

476: outflow is the most influential quantity should remain correct. In

477: that case, the influence of the other properties of outflow

478: must be taken into account as well.

479: 

480: \acknowledgments We thank our referee, Chris Done,

481: for constructive suggestions. This work was supported in part by the

482: Natural Science Foundation of China (grant 10773024),

483: One-Hundred-Talent Program of China, and Shanghai Pujiang Program.

484: 

485: 

486: \begin{appendix}

487: 

488: \section{The height-integrated equations describing inflow/outflow}

489: 

490: %As in the standard theory of accretion disks, we adopt a cylinder

491: %coordinate system ($r, \phi, z$) to describe steady axisymmetric

492: %($\partial /\partial t= \partial / \partial\phi = 0 $) accretion

493: %flow. Basic dynamical equations are those of continuity,

494: %three-dimensional momentum, and energy. We neglect the vertical

495: %differences of all quatities (e.g., $v_r, v_\phi, T_e, T_i$) in

496: %inflow except those related with density $\rho$(e.g. Manmoto, 1997),

497: %and we assume that inflow's vertical velocity $v_z$, which is

498: %considered in the outflow, is neglectable in inflow's dynamics.

499: %Similar to Manmoto et al (1997), the vertical density distribution

500: %in inflow follows $\rho(r,z) = \rho(r,0) \exp(-\frac{z^2}{2 H^2})$,

501: %where $H = C_s/\Omega_k$, is the characteristic height of accretion

502: %flow and is treated as the dividing height of inflow and outflow.

503: 

504: For a steady axisymmetric accretion flow, the equation of mass conservation is:

505: \begin{equation}

506: \label{cont} \frac{1}{r}\frac{\partial}{\partial r} (r \rho

507: \upsilon_r) + \frac{\partial}{\partial z} ( \rho \upsilon_z) = 0.

508: \end{equation}

509: Defining the mass accretion rate $\mdot \equiv -4 \pi r \rho

510: \upsilon_r H$, we integrate eq. (\ref{cont}) in $z$ from $0$ to

511: $H^+$, here $H^+$ denotes just above the surface $z=H$ where the

512: outflow is just launched. Noting $\upsilon_z=0$ for inflow and

513: $\upsilon_z=\upsilon_{z,w}$ for outflow at $H^+$, we get:

514: \begin{equation}

515: \frac{d \mdot}{d r} = \eta_1 4 \pi r \rho \upsilon_{z,w},

516: \end{equation}

517: where \be \eta_1 = \frac{\rho_w}{\overline{\rho}} =

518: \frac{e^{-1/2}\rho(r,0)}{\frac{1}{H} \int_0^H \rho(r,0)

519: exp(-\frac{z^2}{2 H^2}) d z} = 0.7089, \ee which gives the ratio of

520: the density of outflow and height-averaged inflow.

521: 

522: The radial and azimuthal momentum equations read as follows:

523: \begin{equation}

524: \label{rmom} \rho (\upsilon_r \frac{\partial \upsilon_r}{\partial r}

525: - \frac{\upsilon_\phi^2}{r} + \upsilon_z \frac{\partial

526: \upsilon_r}{\partial z}) = -\frac{\partial P}{\partial r} + \rho

527: g_r,

528: \end{equation}

529: \begin{equation}

530: \label{phimom} \rho (\upsilon_r \frac{\partial

531: \upsilon_\phi}{\partial r} + \frac{\upsilon_r \upsilon_\phi}{r} +

532: \upsilon_z \frac{\partial \upsilon_\phi}{\partial z}) =

533: \frac{1}{r^2}\frac{\partial}{\partial r}(r^2 \tau_{\phi r}).

534: \end{equation}

535: Here $g_r$ is the radial component of the gravitational force.

536: The energy equation is

537: \begin{equation}

538: \rho Tds/dt\equiv \rho\left(dU/dt - P/\rho^2 d \rho/ d

539: t\right) = q^+ - q^-,

540: \end{equation}

541: which for steady flow reduces to:

542: \begin{equation}

543: \label{energy} \rho \left[ \upsilon_r \left(\frac{\partial

544: U}{\partial r} - \frac{P}{\rho^2} \frac{\partial \rho}{\partial

545: r}\right) + \upsilon_z \left(\frac{\partial U}{\partial z} -

546: \frac{P}{\rho^2} \frac{\partial \rho}{\partial z}\right)\right]=q^+

547: - q^-.

548: \end{equation}

549: 

550: We integrate the above equations \ref{rmom}-\ref{energy} for $z$

551: from $z=0$ to $z=H^+$ using the following general result:

552: \begin{equation}

553: \label{hint}\int_0^{H^+} f \upsilon_z \frac{\partial g}{\partial z}

554: d z = f \upsilon_{z,w} (g_w - g_{z=H}),

555: \end{equation}

556: where $f$ and $g$ are functions of ($r, z$). Note that if $g(r,z)$

557: is continuous at $z = H$ (e.g., density $\rho$), the right side of

558: equation (\ref{hint}) is equal to 0. Then we will get eqs.

559: (\ref{radial}) - (\ref{ion}).

560: 

561: \end{appendix}

562: 

563: 

564: %\bibliography{}

565: \begin{thebibliography}{}

566: \def\refpar{\hangindent=3em\hangafter=1}

567: %\def\reference{\refpar\noindent}

568: %\def\ref{\refpar\noindent}

569: %\def\ref{\bibitem[]{587}}

570: %\def\reference{\bibitem[]{588}}

571: \def\apj{ApJ}

572: \def\apjs{ApJS}

573: \def\apss{Ap\&SS}

574: \def\mnras{MNRAS}

575: \def\aa{A\&A}

576: \def\aas{A\&A Suppl. Ser.}

577: \def\aj{AJ}

578: \def\araa{ARA\&A}

579: \def\nat{Nature}

580: \def\pasj{PASJ}

581: \def\pasp{PASP}

582: 

583: %\bibitem[]{} Appl, S., \& Camenzind, M. 1992, \aa, 256, 354

584: %\bibitem[]{} Appl, S., \& Camenzind, M. 1993, \aa, 274, 699

585: %\bibitem[]{} Bisnovatyi-Kogan, G. S., Lovelace, R. V. E. 1997, \apj,

586: %486, L43

587: %\bibitem[]{} Beckert, T. 2000, \apj, 539, 223

588: \bibitem[]{} Beckwith, K., Hawley, J. F., Krolik, J. H. 2008,

589: astro-ph/0709.3833

590: \bibitem[]{} Blandford, R. D., Begelman, M. C. 1999, \mnras, 303, L1

591: (BB99)

592: \bibitem[]{} Blandford, R. D., Begelman, M. C. 2004, \mnras, 349, 68

593: \bibitem[]{} Blandford, R. D., Payne, D. G. 1982, \mnras, 199, 883

594: \bibitem[]{} Chartas, G., Brandt, W. N., Gallagher, S. C. 2003, \apj, 595, 85

595: \bibitem[]{} De Villiers, J.-P., Hawley, J. F., Krolik, J. H., Hirose, S. 2005, \apj, 620, 878

596: \bibitem[]{} Granato, G. L., De Zotti, G., Silva, L., Bressan, A., Danese, L. 2004, \apj, 600, 580

597: \bibitem[]{} Igumenshchev, I. G., Narayan, R., Abramowicz, M. A.

598: 2003, \apj, 592, 1042

599: \bibitem[]{} Kaspi, S., Brandt, W. N., George, I. M., et al. 2001,

600: \apj, 554, 216

601: \bibitem[]{} Kuncic, Z. \& Bicknell, G. V. 2007, \apss, 311, 127

602: \bibitem[]{} Gu, W., Lu, J. 2007, \apj, 660, 541

603: %\bibitem[]{} Manmoto, T., Mineshige, S., Kusunose, M. 1997, \apj, 489,

604: %791 (MMK)

605: \bibitem[]{} Marrone, D. P., Moran, J. M., Zhao, J. H., Rao, R. 2006,

606: \apj, 640, 308

607: \bibitem[]{} McKinney, J. C. 2006, \mnras, 368, 1561

608: \bibitem[]{} Misra, R., \& Taam, R. E. 2001, \apj, 553, 978

609: %\bibitem[]{} Narayan, R. 2005, \apss, 300, 177

610: %\bibitem[]{} Narayan, R., Mahadevan, R., Quataert, E. 1998, in The

611: %Theory of Black Hole Accretion Discs, eds. M. A. Abramowicz, G.

612: %Bjornsson, \& J. E. Pringle (Cambridge University Press), 148

613: \bibitem[]{} Narayan, R., Yi, I. 1994, \apj, 428, L13

614: \bibitem[]{} Narayan, R., Yi, I. 1995, \apj, 444, 231

615: \bibitem[]{} Paczy\'nski, B., Wiita, P. J. 1980, \aa, 88, 23

616: \bibitem[]{} Proga, D. 2003, \apj, 585, 406

617: %\bibitem[]{} Proga, D. 2007, in The Central Engine of Active Galactic

618: %Nuclei, eds. L. C. Ho and J.-M. Wang (San Francisco: ASP),

619: %astro-ph/0701100

620: \bibitem[]{} Quataert, E., Narayan, R. 1999, \apj, 520, 298

621: %\bibitem[]{} Quataert, E. 2004 ASPC, 311, 131Q

622: \bibitem[]{} Silk, J., Rees, M. J. 1998, \aa, 331, L1

623: \bibitem[]{} Shakura, N. I., Syunyaev, R. A. 1973, \aa, 24, 337

624: \bibitem[]{} Springel, V., Di Matteo, T., Hernquist, L. 2005,

625: \mnras, 361, 776

626: \bibitem[]{} Spruit, H. C. 1996, in Physical processes in Binary Stars,

627: eds. R.A.M.J. Wijers, M.B. Davies and C.A. Tout, (NATO ASI Series,

628: Kluwer Dordrecht), astro-ph/9602022

629: \bibitem[]{} Stone, J. M., Pringle, J. E. 2001, \mnras, 322, 461

630: \bibitem[]{} Stone, J. M., Pringle, J. E., Begelman, M. C. 1999,

631: \mnras, 310, 1002

632: \bibitem[]{} Xu, G., Chen, X. 1997, \apj, 489, L29

633: \bibitem[]{} Xue, L., Wang, J. 2005, \apj, 623, 372

634: \bibitem[]{} Yuan, F., Quataert, E., Narayan, R. 2003, \apj, 598,

635: 301

636: \bibitem[]{} Vlahakis, N., K$\ddot{o}$nigl, A. 2003, \apj, 596, 1080

637: \end{thebibliography}{}

638: 

639: %\begin{figure} \epsscale{1.} \plotone{mratesp.eps} \vspace{0.5in}

640: %\caption{The spectrum of different models at same accretion rate

641: %($\mdot_{r_{out}},\mdot_{r_{in}}$). Red curve presents the standard

642: %treatment, color black and blue stand for $\xi_{T_i}$ equals $1.0$

643: %and $1.5$, respectively. curves in dot, dashed, long dashed present

644: %$\xi_\phi$ to be $0.8, 1.0, 1.2$, respectively. $\xi_z$ is adjusted

645: %to meet the accretion rate required. Other parameters are set to

646: %ordinary values as $\xi_r = 0.2, \xi_{T_e} = 1.5$.} \label{mratesp}

647: %\end{figure}

648: 

649: %==========================scheme.eps====================

650: \begin{figure}

651: \centering

652: %emulateapj%

653: \includegraphics[width=0.9\columnwidth,clip=true]{f1.eps}

654: %aastex

655: %\epsscale{0.67}

656: %\plotone{scheme.eps}

657: \caption{Schematic diagram of inflow/outflow model. Outflow launches

658: from the surface $z = H$.} \vspace{0.1in}\label{scheme}

659: \end{figure}

660: %========================================================

661: 

662: %========================figure mdot=============================

663: \begin{figure}

664: \epsscale{1.} \plotone{f2.eps}

665: %\vspace{0.5in}

666: \caption{The change of accretion rate as a function of radius for

667: our model (solid line) and the standard treatment (dashed line) in

668: which $\dot{M}(r)\propto r^s$ with $s$ being a constant.}

669: \vspace{0.1in}\label{mdot}

670: \end{figure}

671: %================================================================

672: 

673: %========================== figure xiz-dyn========================

674: \begin{figure*} \epsscale{1.} \plotone{f3.eps}

675: %\vspace{0.5in}

676: \caption{Influence of the vertical velocity of the outflow described

677: by $\xi_z$ on the Mach number, density, temperature, and specific

678: angular momentum of inflow. The solid, dashed, and long-dashed lines

679: are for $\xi_z=0.01, 0.05, 0.15$, respectively. The Keplerian

680: angular momentum $l_k$ is show as the dot-dashed line in the bottom

681: right panel. Other parameters are $\upsilon_{r,w} = \upsilon_r,

682: \xi_\phi= \xi_{T_e}= \xi_{T_i} = 1.0$.} \vspace{0.1in} \label{xiz}

683: \end{figure*}

684: %=================================================================

685: 

686: %======================== figure mdotconst.-1=====================

687: \begin{figure*} \epsscale{1.} \plotone{f4.eps}

688: %\vspace{0.5in}

689: \caption{Influence of the specific angular momentum of the outflow

690: described by $\xi_{\phi}$ on the dynamics of inflow and their

691: comparison with the standard treatment (solid line). The dotted,

692: dashed, and long-dashed lines are for $\xi_{\phi}=0.8, 1.0, 1.2$,

693: respectively. Other parameters are $\xi_{T_i}=1$ (Case A),

694: $\xi_r=0.2, \xi_{T_e}=1.5$. We adjust $\xi_z$ so that all the four

695: models have the same accretion rates at the inner and outer

696: boundary.} \vspace{0.1in}\label{mratedyn1}

697: \end{figure*}

698: %==================================================================

699: 

700: %========================= figure mdotconst.-2=====================

701: \begin{figure*} \epsscale{1.} \plotone{f5.eps}

702: %\vspace{0.5in}

703: \caption{Influence of the specific angular momentum of the outflow

704: on the dynamics of inflow and their comparison with the standard

705: treatment. All parameters are the same as Fig. \ref{mratedyn1} (Case

706: A) except $\xi_{T_i}=1.5$.} \vspace{0.1in}\label{mratedyn2}

707: \end{figure*}

708: %==================================================================

709: 

710: \end{document}

711: