1: %\documentstyle[epsfig,referee]{mn}
2: \documentstyle[onecolumn,epsfig]{mn}
3: %\oddsidemargin 0 mm \evensidemargin 0 mm
4: %\topmargin -10 mm
5: \textheight 235 mm
6: % \textwidth 163 mm
7:
8:
9:
10: %\documentstyle[epsfig]{mn}
11: \begin{document}
12:
13:
14: \title[Self-similar structure of disks]{Self-similar structure of the magnetized radiation-dominated accretion disks}
15:
16:
17: \author[Shadmehri and Khajenabi]
18: {M. Shadmehri\thanks{E-mail:
19: mshadmehri@science1.um.ac.ir} and F. Khajenabi\thanks{E-mail:fkhajenabi@science1.um.ac.ir}\\
20: Department of Physics, School of Science, Ferdowsi University,
21: Mashhad, Iran}
22:
23: \maketitle
24:
25: \date{Received 10 April 2005 / Accepted _________________ }
26: %
27: %\maketitle
28: %
29: \markboth{Shadmehri & Khajenabi: Self-similar structure of
30: disks}{}
31: %
32: \begin{abstract}
33: %
34: We investigate the effects of a large-scale magnetic field with
35: open field lines on the steady-state structure of a
36: radiation-dominated accretion disk, using self-similarity
37: technique. The disk is supposed to be turbulent and possesses an
38: effective viscosity and an effective magnetic diffusivity. We
39: consider the extreme case in which the generated energy due to
40: viscous and magnetic dissipation is balanced by the advection
41: cooling. While the magnetic field outside of the disk is treated
42: in a phenomenological way, the internal field is determined
43: self-consistently. Magnetized and nonmagnetized solutions have
44: the same radial dependence, irrespective of the values of the
45: input parameters. Generally, our self-similar solutions are very
46: sensitive to the viscosity or diffusivity coefficients. For
47: example, the density and the rotation velocity increase when the
48: viscosity coefficient decreases. The gas rotates with
49: sub-Keplerian angular velocity with a factor less than unity
50: which depends on the magnetic field configuration. Magnetic field
51: significantly reduce disk thickness, however, tends to increase
52: the radial velocity comparing to the nonmagnetic self-similar
53: solutions.
54: %
55: \end{abstract}
56:
57: \begin{keywords}
58: accretion, accretion disks - black hole physics - MHD
59: \end{keywords}
60: %
61: %________________________________________________________________
62:
63: \section{Introduction}
64: %
65: %
66:
67:
68: Advection-dominated accretion flows (ADAFs) have been studied by
69: many authors during recent years (e.g., Abramowicz et al. 1988;
70: Narayan \& Yi 1994; Chen 1995; Narayan, Kato \& Honma 1997). A
71: key feature of radiatively inefficient accretion flows is that
72: radiative energy losses are small so that most of the energy is
73: advected with the gas. However, advection-dominated accretion
74: flows can occur in two regimes, depending on the accretion rate
75: and the optical depth. When the accretion rate is high, the
76: optical depth becomes very high and the radiation is trapped in
77: the gas. These kinds of solutions which are known under the name
78: 'slim accretion disk' have been studied in detail by Abramowicz
79: et al. (1988). On other hand, we may have optically thin
80: accretion flows with very low mass accretion rate (e.g., Rees et
81: al. 1982; Narayan \& Yi 1994; Abramowitz et al. 1995; Chen 1995).
82:
83: Because of the complexity of the equations, similarity technique
84: can help us to explore the relevant physics of radiatively
85: inefficient accretion flows. As long as we are not interested in
86: the boundaries of the problem, such solutions that describe the
87: behavior of the flow in an intermediate region far from the radial
88: boundaries. Originally, Narayan \& Yi (1994) studied optically
89: thin advection-dominated accretion disk using their self-similar
90: solutions. They speculated, on the basic of some numerical
91: calculations, that the self-similar solution is the natural state
92: for an advection-dominated flow. Subsequent analysis (e.g.,
93: Narayan, Kato \& Honma 1997) showed that the global solutions
94: achieve approximate self-similar behavior within a short distance
95: from the outer boundary and the approach to self-similarity is
96: quite impressive. Considering these achievements, Wang \& Zhou
97: (1999) constructed self-similar solutions for optically {\it
98: thick} advection dominated accretion flow, in which photon
99: trapping and advection dominate over surface diffusion cooling and
100: used these to explore its different properties from optically
101: thin self-similar solutions. However, both solutions have the same
102: power index of radius.
103:
104:
105: A remarkable problem arises when the accretion disk is threaded by
106: magnetic field. There are good reasons for believing that magnetic
107: fields are important to the physics of accretion disk. Schwartzman
108: (1971) was the first to point out the importance of the magnetic
109: field in an accretion process. He proposed a hypothesis of
110: equipartition between the magnetic and kinetic energy densities
111: and this picture is usually accepted in the modern picture of
112: viscous ADAF models. Bisnovatyi-Kogan \& Lovelace (2000) suggested
113: that recent papers discussing ADAF as a possible solution for
114: astrophysical accretion should be treated with caution,
115: particularly because of ignorance of the magnetic field. While
116: they obtained a solution for a time-averaged magnetic field in a
117: quasispherical accretion flow, an analysis of energy dissipation
118: and equipartition between magnetic and flow energies has been
119: presented (Bisnovatyi-Kogan \& Lovelace 2000). Numerical
120: simulations of {\it magnetized}, radiatively inefficient flows
121: have been done recently by several authors (e.g., Machida,
122: Matsumoto \& Mineshige 2001; Igumenshchev, Narayan \& Abramowicz
123: 2003). However, in most of these the resistive terms in the MHD
124: equations have been neglected, or the resistivity has been
125: considered only in the induction equation without accounting the
126: corresponding dissipation in the energy equation.
127:
128: Some authors tried to study magnetized accretion flow
129: analytically. For example, Kaburaki (2000) presented a set of
130: analytical solutions for a fully advective accretion flow in a
131: global magnetic field and the conductivity is assumed to be
132: constant for simplicity. Shadmehri (2004) extended this analysis
133: by considering non-constant resistivity. He obtained a set of
134: self-similar solutions in spherical coordinates that describes
135: quasi-spherical magnetized accretion flow. Lai (1998) and Lee
136: (1999) calculated transonic disk accretion flows around a weekly
137: magnetized neutron star, where it was assumed the disk is fully
138: advective. While Lee (1999) considered both the thermal and the
139: radiation pressures, Lai (1998) assumed the radiation pressure
140: dominates over the gas pressure.
141:
142: In this study, we present self-similar solutions of an idealized
143: height-integrated set of equations that describe magnetized
144: radiation-dominated accretion flow. In fact, this analysis extends
145: self-similar solutions of Wang \& Zhou (1999) for optically thick
146: advection-dominated accretion flow to the magnetized case. For a
147: steady state disk, there can be a final, steady configuration of
148: magnetic field, in which the inward dragging of field lines by the
149: disk is balanced everywhere by the outward movement of filed lines
150: due to magnetic diffusivity. We show that the radial structure of
151: a magnetized radiation-dominated accretion flow does not differ
152: from the non-magnetic solutions. But we can see significantly
153: different behaviors, because of the effects of the magnetic
154: fields. The equations of the model are presented in the second
155: section. We obtain and solve the set of self-similar equations
156: analytically in the third section. For a set of illustrative
157: parameters the solutions will be discussed in this section.
158:
159:
160:
161: \section{General Formulation}
162: %
163: %
164: We employ a cylindrical coordinate system $(R,\varphi,z)$ centered
165: on a central object (e.g., a black hole) with mass $M_{\ast}$
166: which accretes matter at a steady state $\dot{M}$ from a
167: geometrically thin axisymmetric accretion disk in steady state
168: threaded by an ordered magnetic field. Our model generalize the
169: usual slim disks around black holes (e.g., Muchotrzeb \&
170: Paczy\'{n}ski 1982; Matsumoto et al. 1984; Abramowicz et al. 1988)
171: by including the effect of magnetic fields. General relativistic
172: effects are neglected and outside of the disk, dissipative effects
173: are assumed to be negligible. For the magnetic field geometry, we
174: are following a general approach presented by Lovelace, Romanova
175: \& Newman 1994 (hereafter LRN), in which even filed symmetry is
176: assumed so that $B_{\rm R}(R,z)=-B_{\rm R}(R,-z)$,
177: $B_{\varphi}(R,z)=-B_{\varphi}(R,-z)$ and $B_{\rm z}(R,z)=+B_{\rm
178: z}(R,-z)$. However, we note that the solution for the magnetic
179: field outside of disk should match to the field solution inside
180: the disk at its surface. But our model does not present a
181: self-consistent model for the magnetic field outside of the disk.
182: In analogy to LRN, we parameterize the magnetic field outside of
183: the disk. Also, it is assumed that the accreting matter is
184: confined to a thin disk, and we do not formally introduce a
185: magnetosphere into our model.
186:
187:
188: The basic equations are integrated over the vertical thickness of
189: the disk. The mass continuity equation takes the form
190: %
191: \begin{equation}
192: -2\pi R \Sigma v_{\rm R}=\dot{M},\label{eq:masscon}
193: \end{equation}
194: %
195: where $v_{\rm R}$ is the radial velocity and $\Sigma=\int dz \rho
196: \simeq 2 h \rho $ is surface density of the disk. The
197: half-thickness is denoted by $h$, where we consider the magnetic
198: field effect on the disk thickness. We will see that the disk can
199: be compressed or flattened depending on the field configuration.
200: Also, we note that since the radial velocity is negative for
201: accretion (i.e., $v_{\rm R}<0$), the accretion rate $\dot{M}$ as
202: an input parameter of our model is positive.
203:
204: The radial momentum equation reads
205: %
206: \begin{equation}
207: \Sigma v_{\rm R}\frac{d v_{\rm R}}{d
208: R}-\Sigma\frac{v_{\varphi}^2}{R}=-\frac{d P}{d
209: R}-\Sigma\frac{GM_{\ast}}{R^2}+\int F_{\rm R}^{\rm mag}
210: dz,\label{eq:rcom}
211: \end{equation}
212: %
213: where $P=\int dz p$ is the integrated disk pressure. We consider
214: an extreme case in which the radiation pressure dominates over
215: the gas pressure and the generated energy is balanced by the
216: advection cooling. The first assumption implies $P\simeq 2h a
217: T^4/3$, where $a$ is black body constant and $T$ denotes the
218: midplane temperature of the disk. The energy equation is written
219: based on the second assumption. The last term of equation
220: (\ref{eq:rcom}) represents the height integrated radial magnetic
221: force which can be written as (LRN)
222: %
223: \begin{displaymath}
224: \int F_{\rm R}^{\rm mag} dz=\frac{1}{2\pi}(B_{\rm R} B_{\rm
225: z})_{\rm h}-\frac{1}{4\pi R^2}\frac{d}{d R}[h
226: R^{2}<B_{\varphi}^{2}-B_{\rm R}^2>]
227: \end{displaymath}
228: %
229: \begin{equation}
230: -\frac{1}{4\pi}\frac{d}{d R}[h <B_{\rm
231: z}^{2}>]+\frac{1}{4\pi}\frac{dh}{dR}(B_{\varphi}^{2}+B_{\rm z
232: }^{2}-B_{\rm R}^{2})_{\rm h},
233: \end{equation}
234: %
235: where $<\cdots>\equiv\int_{-h}^{h}dz(\cdots)/(2h)$, and the $h$
236: subscript denotes that the quantity is evaluated at the upper
237: disk plane, i.e. $z=h$. Similarly, integration over $z$ of the
238: azimuthal equation of motion gives
239:
240: %
241: \begin{equation}
242: R\Sigma v_{\rm R}\frac{d}{d R}(Rv_{\varphi})=\frac{d}{d
243: R}[R^{3}\nu\Sigma\frac{d}{d R}(\frac{v_{\varphi}}{R})]+\int
244: F_{\varphi}^{\rm mag} dz,\label{eq:phicom}
245: \end{equation}
246: %
247: where
248: %
249: \begin{displaymath}
250: \int F_{\varphi}^{\rm mag}
251: dz=\frac{1}{2\pi}(R^{2}B_{\varphi}B_{\rm z})_{\rm
252: h}-\frac{1}{2\pi}\frac{dh}{dR} (R^{2}B_{\rm R}B_{\varphi})_{\rm h}
253: \end{displaymath}
254: %
255: \begin{equation}
256: +\frac{1}{2\pi}\frac{d}{dR}[hR^{2} <B_{\rm R}B_{\varphi}>].
257: \end{equation}
258: %
259: Here, the last term of equation (\ref{eq:phicom}) represents the
260: height integrated toroidal component of magnetic force. Note that
261: this form of azimuthal equation of motion is not exactly similar
262: to the original slim disk model (e.g., Abramowicz et al. 1988),
263: in which the well-know $\alpha p$ prescription of Shakura \&
264: Sunyaev (1973) has been used as a general approximate form for
265: the $\varphi r$ component of the viscous stress tensor
266: ($\tau_{\varphi\rm r}=-\alpha p$). Here, we replaced $\alpha p$
267: prescription by a diffusive viscosity prescription, i.e.
268: %
269: \begin{equation}
270: \tau_{\varphi\rm r}=\rho\nu r\frac{\partial\Omega}{\partial r},
271: \end{equation}
272: %
273: where $\rho$ is the density, $\nu$ is a kinematic viscosity
274: coefficient, and $\Omega$ is the angular velocity of matter in
275: the disk. The above prescription leads to equation
276: (\ref{eq:phicom}). In our model, we also assume
277: %
278: \begin{equation}
279: \nu=\alpha c_{\rm s} h,\label{eq:vis}
280: \end{equation}
281: %
282: where $c_{\rm s}$ is the local sound speed and $\alpha$ is a
283: constant less than unity.
284:
285:
286: The $z$ component of equation of motion gives the condition for
287: vertical hydrostatic balance, which can be written as (Lovelace
288: et al. 1987)
289: %
290: \begin{equation}
291: (\frac{h}{R})^{2}+q(\frac{h}{R})-\frac{2P}{\Sigma v_{\rm
292: K}^2}=0,\label{eq:zcom}
293: \end{equation}
294: %
295: where $v_{\rm K}=\sqrt{GM_{\ast}/R}$ is the Keplerian velocity and
296: %
297: \begin{equation}
298: q=\frac{R[(B_{\varphi})_{\rm h}^{2}+(B_{\rm R})_{\rm
299: h}^{2}]}{4\pi\Sigma v_{\rm K}^2}.
300: \end{equation}
301: %
302:
303: Now, we can treat the internal magnetic field using the induction
304: equation. LRN showed that the variation of $B_{\rm z}$ with $z$
305: within the disk is negligible for even field symmetry. Moreover,
306: $B_{\rm R}$ and $B_{\varphi}$ are odd functions of $z$ and
307: consequently $\partial B_{\rm R}/\partial z\approx (B_{\rm
308: R})_{\rm h}/h$ and $\partial B_{\varphi}/\partial z\approx
309: (B_{\varphi})_{\rm h}/h$. Thus,
310: %
311: \begin{displaymath}
312: B_{\rm R}(R,z)=\frac{z}{h}(B_{\rm R})_{\rm h},
313: B_{\varphi}(R,z)=\frac{z}{h}(B_{\varphi})_{\rm h},
314: \end{displaymath}
315: %
316: \begin{equation}
317: B_{\rm z}(R,z)=B_{\rm z}(R),
318: \end{equation}
319: %
320: and the induction equation reads
321: %
322: %
323: \begin{equation}
324: -RB_{\rm z} v_{\rm R}-\frac{\eta R}{h}(B_{\rm R})_{\rm h}+\eta
325: R\frac{d B_{\rm z}}{d R}=0,\label{eq:induction}
326: \end{equation}
327: %
328: where the magnetic diffusivity $\eta$ has the same units as
329: kinematic viscosity. We assume that the magnitude of $\eta$ is
330: comparable to that of the turbulent viscosity $\nu$ (e.g.,
331: Bisnovatyi-Kogan \& Ruzmaikin 1976; Shadmehri 2004). Exactly in
332: analogy to alpha prescription for $\nu$, we are using a similar
333: form for the magnetic diffusivity $\eta$,
334: %
335: \begin{equation}
336: \eta=\eta_{0}c_{\rm s}h,
337: \end{equation}
338: %
339: where $\eta_{0}$ is a constant. Note that $\eta$ is {\it not}
340: constant and depends on the physical variables of the flow, and in
341: our self-similar solutions, as we will show, $\eta$ scales with
342: radius as a power law. This form of scaling for diffusivity has
343: been widely used by many authors (e.g., Lovelace, Wang \& Sulkanen
344: 1987; LRN ; Ogilvie \& Livio 2001; R\"{u}diger \& Shalybkov 2002).
345:
346:
347: While equation (\ref{eq:induction}) describes transport of a
348: large-scale magnetic field (here, $B_{\rm z}(R)$), the values of
349: $(B_{\rm R})_{\rm h}$ and $(B_{\varphi})_{\rm h}$ are determined
350: by the filed solutions external to the disk. Instead, we are
351: following approach of LNR, in which the external field solutions
352: obey the relations
353: %
354: \begin{equation}
355: (B_{\rm R})_{\rm h}=\beta_{\rm r} B_{\rm z}, (B_{\varphi})_{\rm
356: h}=\beta_{\varphi} B_{\rm z},
357: \end{equation}
358: where $\beta_{\rm r}$ and $\beta_{\varphi}$ are constants of
359: order unity ($\beta_{\varphi}<0$). Thus, one can simply show that
360: $<B_{\rm R}^{2}>=\beta_{\rm r}^{2} B_{\rm z}^{2}/3$,
361: $<B_{\varphi}^{2}>=\beta_{\varphi}^{2} B_{\rm z}^{2}/3$ and
362: $<B_{\rm R}B_{\varphi}>=\beta_{\rm r}\beta_{\varphi}B_{\rm
363: z}^{2}/3$.
364: %
365: %
366: %
367:
368:
369: To close the equations of our model, we can write the energy
370: equation describing the thermal state of the flow as
371: %
372: \begin{equation}
373: \rho T v_{\rm R}\frac{dS}{dR}=Q_{\rm vis}+Q_{\rm
374: Joule},\label{eq:energy}
375: \end{equation}
376: %
377: where $S$ is the specific entropy (per unit mass) and $T$ is
378: midplane temperature of the disk. For the heating term, we may
379: have two sources of dissipation: the viscous and resistive
380: dissipations due to a turbulence cascade. So, $Q_{\rm vis}$ and
381: $Q_{\rm Joule}$ represent viscous dissipation due to the radial
382: motion and the Joule heating rate, respectively,
383: %
384: \begin{equation}
385: Q_{\rm vis}=\rho\nu R^{2}(\frac{d\Omega}{dR})^{2},
386: \end{equation}
387: %
388: and
389: %
390: \begin{equation}
391: Q_{\rm Joule}=\frac{\eta}{4\pi h^{2}}[2(B_{\rm R})_{\rm
392: h}^{2}+\frac{3}{5}(B_{\varphi})_{\rm h}^{2}]
393: \end{equation}
394: %
395:
396:
397: Now we have constructed our model and the main equations of the
398: model are equations (\ref{eq:masscon}), (\ref{eq:rcom}),
399: (\ref{eq:phicom}), (\ref{eq:zcom}), (\ref{eq:induction}) and
400: (\ref{eq:energy}). In the next section, we will present
401: self-similar solutions of these equations.
402: %
403: \section{self-similar solutions}
404: %
405: The equations of our model are reduced to standard equations of
406: the slim disk, if we set all the magnetic terms equal to zero. As
407: we discussed, we are considering a radiation-dominated disk, in
408: which the gas pressure has been neglected comparing to the
409: radiation pressure. After some algebraic manipulations we get to
410: the following set of self-similar solutions:
411: %
412: \begin{equation}
413: \Sigma(R)=a \Sigma_{0}(\frac{R}{R_0})^{-1/2},
414: \end{equation}
415: %
416: \begin{equation}
417: v_{\varphi}(R)=b
418: \sqrt{\frac{GM_{\ast}}{R_0}}(\frac{R}{R_0})^{-1/2},
419: \end{equation}
420: %
421: \begin{equation}
422: v_{\rm R}(R)=- c
423: \sqrt{\frac{GM_{\ast}}{R_0}}(\frac{R}{R_0})^{-1/2},
424: \end{equation}
425: %
426: \begin{equation}
427: P(R)= d \frac{\Sigma_{0}GM_{\ast}}{R_0}(\frac{R}{R_0})^{-3/2},
428: \end{equation}
429: %
430: \begin{equation}
431: B_{\rm z}(R)=e
432: \sqrt{4\pi\Sigma_{0}\frac{GM_{\ast}}{R_{0}^{2}}}(\frac{R}{R_0})^{-5/4},
433: \end{equation}
434: %
435: \begin{equation}
436: h(R)=f R_{0} (\frac{R}{R_0}),
437: \end{equation}
438: %
439: %
440: %
441: %
442: where $\Sigma_{\rm 0}$ and $R_{\rm 0}$ provide convenient units
443: with which the equations can be written in non-dimensional form.
444: Thus, we obtain the following system of dimensionless equations,
445: to be solved for $a$, $b$, $c$, $d$, $e$ and $f$:
446:
447: \begin{equation}
448: ac=\dot{m},
449: \end{equation}
450: %
451: \begin{equation}
452: -\frac{1}{2}ac^{2}-ab^{2}=\frac{3}{2}d-a+[2\beta_{\rm
453: r}+(\frac{5+\beta_{\varphi}^{2}-\beta_{\rm r}^{2}}{2})f]e^{2},
454: \end{equation}
455: %
456: \begin{equation}
457: -\frac{1}{2}abc=-\frac{3\alpha}{4}\sqrt{\frac{d}{a}}fab+\beta_{\varphi}(2-\beta_{\rm
458: r }f)e^{2},
459: \end{equation}
460: %
461: \begin{equation}
462: af^{2}+(\beta_{\varphi}^{2}+\beta_{\rm r}^{2})fe^{2}-2d=0,
463: \end{equation}
464: %
465: \begin{equation}
466: c-\eta_{0}\beta_{\rm
467: r}\sqrt{\frac{d}{a}}-\frac{5}{4}\eta_{0}f\sqrt{\frac{d}{a}}=0,
468: \end{equation}
469: %
470: \begin{equation}
471: \frac{1}{4}c\sqrt{ad}=\frac{3\alpha}{8}fab^{2}+\frac{1}{3}\eta_{0}(2\beta_{\rm
472: r }^{2}+\frac{3}{5}\beta_{\varphi}^{2})e^{2},
473: \end{equation}
474: %
475: where $\dot{m}=\dot{M}/(2\pi \Sigma_{0}\sqrt{GM_{\ast}R_{0}})$ is
476: nondimensional mass accretion rate. Thus, our input parameters
477: are $\alpha$, $\eta_{0}$, $\beta_{\rm r}$, $\beta_{\varphi}$ and
478: $\dot{m}$. We will present values of the physical quantities in
479: non-dimensional form.
480:
481:
482: In the limit of nonmagnetic case, the above equations can be
483: solved analytically:
484: %
485: \begin{equation}
486: a=\frac{3\sqrt{2}\dot{m}\alpha}{\sqrt{25+36\alpha^{2}}-5}\simeq
487: \frac{5\sqrt{2}\dot{m}}{6\alpha},
488: \end{equation}
489: %
490: \begin{equation}
491: b=\frac{1}{3\alpha}\sqrt{\sqrt{25+36\alpha^{2}}-5}\simeq\sqrt{\frac{2}{5}},
492: \end{equation}
493: %
494: \begin{equation}
495: c=\frac{\sqrt{25+36\alpha^{2}}-5}{3\sqrt{2}\alpha}\simeq
496: \frac{6\alpha}{5\sqrt{2}},
497: \end{equation}
498: %
499: \begin{equation}
500: d=\frac{\sqrt{2}}{3\alpha}\dot{m},
501: \end{equation}
502: %
503: \begin{equation}
504: f=\frac{\sqrt{2}}{3\alpha}\sqrt{\sqrt{25+36\alpha^{2}}-5}\simeq\frac{2}{\sqrt{5}}.
505: \end{equation}
506: %
507: The second relation in each equation refers to the limit
508: $\alpha\ll 1$. The scaling of our solutions are different from
509: Wang \& Zhou (1999) who analyzed self-similar solution of
510: optically thick advection-dominated flows. Because they applied
511: $\alpha p$ prescription for viscous stress tensor, but a diffusive
512: prescription has been used in our model, i.e. equation
513: (\ref{eq:vis}).
514:
515: The above nonmagnetic solutions show that the surface density
516: increases with accretion rate, and decreases inversely with
517: $\alpha$. However, the radial velocity is directly proportional
518: to the viscosity coefficient $\alpha$. The gas rotates with
519: sub-Keplerian angular velocity, i.e.
520: $\Omega\approx\sqrt{\frac{2}{5}}\Omega_{\rm K}$. Note that except
521: for the surface density and the pressure, the other physical
522: quantities are independent of the accretion rate but depend only
523: on the viscosity coefficient $\alpha$. An interesting feature is
524: that the opening angle of the disk is fixed $h/R\approx
525: 2/\sqrt{5}$, independent of $\alpha$ and of the mass accretion
526: rate $\dot{m}$.
527:
528:
529: %
530: %
531: %
532: %
533: %
534: %
535: \begin{figure}
536: \epsfig{figure=f1.eps,angle=0,width=\hsize} \caption{Profiles of
537: some height-averaged physical quantities for the disk, as a
538: function of $\beta_{\rm r}$, for $\dot{m}=1$, $\eta_0=0.1$,
539: $\alpha=0.01$ and $\beta_{\varphi}=-0.8$ (solid line),
540: $\beta_{\varphi}=-0.9$ (dashed line) and $\beta_{\varphi}=-1.1$
541: (dotted line). The surface density $\Sigma$, the radial velocity
542: $v_{\rm R}$, the rotational velocity $v_{\varphi}$, the pressure
543: $P$, half-thickness $h$ are drawn in dimensionless form. The
544: ratio of the magnetic pressure at the surface of the disk to the
545: pressure is presented by $\beta_{m}$.}\label{fig:figure1}
546: \end{figure}
547: %
548: %
549: %
550: %
551: %
552: %
553: \begin{figure}
554: \epsfig{figure=f2.eps,angle=0,width=\hsize} \caption{Same as
555: Figure 1, but in the case of $\alpha=0.001$, $\dot{m}=1$,
556: $\eta_0=0.1$. Solid, dashed, and dotted lines refer to
557: $\beta_{\varphi}=-0.8$, $\beta_{\varphi}=-0.9$ and
558: $\beta_{\varphi}=-1.1$.}\label{fig:figure2}
559: \end{figure}
560:
561:
562: Although the radial scaling of the solutions are similar to the
563: nonmagnetic case, we can see significant differences because of
564: the magnetic field effect. The first important effect of the
565: magnetic field on the disk structure is a squeezing effect, where
566: the scale height $h$ is reduced comparing to the nonmagnetic
567: case. In fact, the squeezing effect of the large-scale magnetic
568: field counterbalances the thickening of the disk generated by
569: advection. Figure 1 shows the behaviour of physical disk
570: quantities as a function of $\beta_{\rm r}$, for three different
571: values of $\beta_{\varphi}=-0.8, -0.9, -1.1$. The other input
572: parameters are assumed as $\dot{m}=1$, $\eta_0=0.01$,
573: $\alpha=0.01$. Generally, $\beta_{\rm r}$ and $\beta_{\varphi}$
574: are parameters of order unity, however, we consider
575: $\beta_{\varphi}$ around unity but changing $\beta_{\rm r}$ from
576: 0.5 to 1.2. For these input parameters, the surface density
577: significantly reduces from $a\simeq 117$ to a value between 25
578: and 45 depending on the magnetic field configuration. However, as
579: $\beta_{\varphi}$ increases, the surface density slightly
580: increases for a fixed $\beta_{\rm r}$. The rotation velocity is
581: below Keplerian and as $\beta_{\rm r}$ increases, the rotation
582: rate reaches to a maximum and then decreases. Also, the radial
583: velocity increases with $\beta_{\rm r}$ or $\beta_{\varphi}$,
584: however, is significantly below free-fall velocity. But comparing
585: to the nonmagnetic solution, the magnetic field tends to increase
586: the radial velocity. For example, the above input parameters gives
587: $c=8.48\times 10^{-3}$ for nonmagnetic flow, but the field causes
588: $c$ increases to a value between $0.02$ to $0.04$. Magnetic field
589: causes the opening angle of the disk decreases.
590:
591: Figure 2 is the same as Figure 1, but with lower viscosity
592: coefficient, i.e. $\alpha=0.001$. We see this decrease of
593: $\alpha$ causes the surface density increases. While the radial
594: velocity decreases with $\alpha$, the rotation velocity becomes
595: closer to the Keplerian rate. We see that the opening angle of
596: disk for $\alpha=0.001$ is smaller than $\alpha=0.01$. Generally,
597: the physical quantities are sensitive to the viscosity coefficient
598: $\alpha$.
599:
600: We repeated the above calculations for $\alpha=\eta_0=0.01$ and
601: found that the solutions weakly change because of the variations
602: of $\beta_{\rm r}$ and $\beta_{\varphi}$. In this case, the
603: solutions are qualitatively similar to Figures 1 and 2, but the
604: scaling is somewhat different. For example, for these input
605: parameters, we find the ratio of $v_{\varphi}/v_{\rm K}$ between
606: $0.66$ and $0.69$ depending on $\beta_{\rm r}$ and
607: $\beta_{\varphi}$. Regarding to Figure 1 which is for
608: $\eta_0=0.1$, we can say as the resisitivity coefficient $\eta_0$
609: decreases, the dependence of solutions on the outside field
610: solutions becomes weaker. The nonmagnetic solutions show that the
611: rotational and the radial velocity and the opening angle are
612: independent of the mass accretion rate. This result is valid even
613: in magnetic case, as we found by changing the accretion rate and
614: keeping other input parameters constant.
615: %
616: %
617: %
618:
619:
620: \section{conclusion}
621: %
622: Although investigating the behavior of the magnetic field and
623: associated currents within the disk was not the main purpose of
624: our study, we studied the effect of a large-scale magnetic field
625: with open field lines on the structure of an optically thick
626: accretion disk. While the field outside of the disk treated in a
627: phenomenological way, we solved the height-averaged MHD equations
628: self-consistently using similarity technique in analogy to the
629: original study of optically thin ADAF by Narayan \& Yi (1994).
630: Our self-similar solutions reduce to the nonmagnetic solutions of
631: Wang \& Zhou (1999) for optically thick advection-dominated
632: accretion flow. The disk structure and the field geometry are
633: closely linked. The magnetic field is dragged by the accreting
634: flow, however, the field tends to squeeze the disk and to
635: increase the radial velocity. The angular velocity of the flow is
636: less than the local Keplerian angular velocity by a factor which
637: depends on the magnetic field configuration.
638:
639: Our simple self-similar solutions show that the effect of the
640: magnetic field can not be ignored in a realistic accretion model.
641: The present solutions may be applied to the X-ray galactic and
642: extragalactic sources, when the accretion rate is high and
643: radiation dominated regime takes place. The emergent spectrum of
644: such a disk can be calculated using our magnetized self-similar
645: solutions and considering the energy transfer in the vertical
646: direction. We will discuss this problem in future.
647:
648:
649: %
650: %
651: \begin{thebibliography}{}
652:
653: \bibitem[]{} Abramowicz M. A., Chen X., Kato S., Lasota J. P.,
654: Regev O., 1995, ApJ, 438, L37
655:
656: \bibitem[]{} Abramowicz M. A., Czerny B., Lasota J. -P.,
657: Szuszkiewicz E., 1988, ApJ, 332, 646
658:
659: \bibitem[]{} Bisnovatyi-Kogan G. S., Lovelace R. V. E., 2000, ApJ,
660: 529, 978
661:
662: \bibitem[]{} Bisnovatyi-Kogan G. S., Ruzmaikin A. A., 1976,
663: Ap\&SS, 42, 401
664:
665:
666: \bibitem[]{} Chen X., 1995, MNRAS, 275, 641
667:
668: \bibitem[]{} Igumenshchev I. V., Narayan R., Abramowicz M. A.,
669: 2003, ApJ, 592, 1042
670:
671: \bibitem[]{} Kaburaki O., 2000, ApJ, 531, 210
672:
673: \bibitem[]{} Lai D., 1998, ApJ, 502, 721
674:
675: \bibitem[]{} Lee U., 1999, ApJ, 511, 359
676:
677: \bibitem[]{} Lovelace R. V. E., Romanova M. M., Newman W. I.,
678: 1994, ApJ, 437, 136 (LRN)
679:
680: \bibitem[]{} Lovelace R. V. E., Wang J. C. L., Sulkanen M.
681: E., 1987, ApJ, 315, 504
682:
683: \bibitem[]{} Machida M., Matsumoto R., Mineshige S., 2001, PASJ,
684: 53, L1
685:
686: \bibitem[]{} Matsumoto R., Kato S., Fukue J., Okazaki A. T.,
687: 1984, PASJ, 36, 71
688:
689: \bibitem[]{} Muchotrzeb B., Paczy\'{n}ski B., 1982, Acta
690: Astron., 32, 1
691:
692: \bibitem[]{} Narayan R., Yi I., 1994, ApJ, 428, L13
693:
694: \bibitem[]{} Narayan R., Kato S., Honma F., 1997, ApJ, 476, 49
695:
696: \bibitem[]{} Ogilvie G. I., Livio M., 2001, ApJ, 553, 158
697:
698: \bibitem[]{} Rees M. J., Begelman M. C., Blandford R. D., Phinney
699: E. S., 1982, Nature, 295, 17
700:
701: \bibitem[]{} R\"{u}diger G., Shalybkov D. A., 2002, A\&A, 393,
702: L81
703:
704: \bibitem[]{} Shvartsman V. F., 1971, Soviet Astron., 15, 377
705:
706: \bibitem[]{} Shadmehri M., 2004, A\&A, 424, 379
707:
708: \bibitem[]{} Shakura N. I., Sunyaev R. A., 1973, A\&A, 24, 337
709:
710: \bibitem[]{} Wang J.-M., Zhou Y.-Y., 1999, ApJ, 516, 420
711:
712:
713: \end{thebibliography}
714:
715:
716:
717: \end{document}
718: