1: \documentstyle[12pt,aasms4]{article}
2:
3: \lefthead{M. Reyes-Ruiz}
4:
5: \righthead{Magnetorotational instability in the dead zone}
6:
7: \begin{document}
8:
9: \title{The magnetorotational instability across the dead zone of
10: protoplanetary disks}
11:
12: \author{M. Reyes-Ruiz}
13: \affil{Instituto de Astronom\'{\i}a, UNAM, Apdo. Postal 877,
14: Ensenada, B.C. 22800, M\'{e}xico.}
15:
16: \begin{abstract}
17: We examine the linear stability of a flow threaded by a weak,
18: vertical magnetic field in a disk with a keplerian rotation profile and
19: a vertical stratification of the ionization degree as that predicted
20: for vast portions of protoplanetary disks. A quasi-global analysis is
21: carried out, where the form of the perturbations in the vertical
22: direction is determined. Considering the ohmic magnetic diffusivity
23: of the gas, the conditions leading to the magnetorotational
24: instability are analyzed as a function of the diffusivity at
25: the disk surfaces, its vertical profile and the strength of the
26: unperturbed magnetic field. For typical conditions believed to prevail in
27: protoplanetary disks at radial distances between 0.1 and 10 AU, where the
28: so-called dead zone is proposed to exist, we find that generally the
29: instability is damped. This implies that, if the
30: MRI is considered the only possible source of turbulence in
31: protoplanetary disks, no viscous angular momentum transport occurs
32: at those radii.
33: \end{abstract}
34:
35: \keywords{accretion, accretion disks -- magnetic fields --
36: MHD -- planetary systems -- solar system: formation}
37:
38: Disks around classical T-Tauri stars, also called protoplanetary
39: disks (or PP disks hereafter), are commonly
40: modeled as viscous accretion disks.
41: Observed evolutionary trends for
42: the global properties of PP disks have been shown to be
43: roughly consistent with those predicted by accretion disk models
44: (Hartmann et al. 1998, Stepinski 1998). However, the acceptance
45: of such models is justifiably not
46: universal and their precise properties are far from established.
47: Nevertheless, partially on account of the existing theoretical
48: framework, accretion disk models constitute the basis of most
49: current models for protoplanetary disks.
50:
51: In such models,
52: molecular viscosity being drastically
53: insufficient to explain the observed evolutionary timescales for
54: these objects, an anomalous viscosity is invoked to drive
55: the evolution of PP disks. Albeit their precise origin was not readily
56: identified, turbulence and magnetic fields where suggested, and generally
57: accepted, as responsible of the required anomalous ``viscous'' torques
58: (Lynden-Bell 1969, Shakura \& Sunyaev 1973, Eardley \& Lightman 1976).
59: In the seminal work of Balbus \& Hawley
60: (1991) the instability of perfectly conducting, keplerian flows,
61: threaded by a weak magnetic field, was identified as an unavoidable
62: factor in the dynamics and magnetic field generation process in
63: accretion disks. Since that time, this so-called
64: magnetorotational instability (also called MRI hereafter) has been
65: shown to lead to
66: self-sustained MHD turbulence capable of transporting angular momentum
67: outwards through the disk while mass is driven toward the central
68: object (Hawley, Gammie \& Balbus 1995, Brandenburg et al. 1995,
69: Stone et al. 1996).
70:
71: As long as magnetic diffusive processes are small enough and a weak magnetic
72: field threads a keplerian disk, the MRI arises and leads to
73: self-sustained turbulence capable of driving disk evolution
74: (Jin 1996, Balbus \& Hawley 1998). These conditions are generally met in
75: accretion disks around compact objects but apparently not in extended
76: regions of protoplanetary disks (Gammie 1996, D'alessio et al. 1998).
77: Temperatures less than 1000 K outwards of $\sim 1$ AU predicted by models
78: of PP disks result in a sharp drop of the ionization degree (Stepinski 1992).
79: Since the magnetic ohmic diffusivity is inversely
80: proportional to the ionization degree this implies the importance of
81: diffusive processes beyond such distances (see also Stepinski, Reyes-Ruiz
82: \& Vanhala 1993). There, galactic cosmic rays and other external
83: sources become the principal ionizing agents.
84:
85: An important feature of
86: the ionization due to such extraneous factors is that, since their
87: effect is shielded as they travel into the
88: disk, the ionization fraction is strongly stratified decreasing
89: from the disk surfaces to the midplane (Dolginov \& Stepinski 1994).
90: Under the assumption that the MRI is the only source of self-sustained
91: turbulence and ``viscous'' torques in accretion disks,
92: Gammie (1996) proposed that mass transfer in PP disks, beyond
93: the inner hot region, takes place in
94: a layered manner. In the layered accretion scenario angular momentum
95: and mass transfer occurs only through active layers near the
96: disk surfaces, where the magnetic diffusivity is low enough for
97: the MRI to develop, while the gas stagnates in the strongly
98: diffusive ``dead zone'' around the midplane. The evolution of
99: protoplanetary disks via layered accretion has been calculated by
100: Stepinski (1999) finding that it can be dramatically
101: different from that predicted by ``standard'' models
102: of PP disks, with vertically uniform viscosity, as those
103: presented by Ruden \& Lin (1986),
104: Ruden \& Pollack (1991) and Reyes-Ruiz \& Stepinski (1995)
105: among others.
106:
107: Given the importance of having reliable, detailed models
108: for the structure and evolution of PP disks as foundation for
109: theories of planet formation, we believe a critical revision
110: of the layered accretion scenario is justified. We begin
111: this work in the present paper by studying the effect of
112: diffusivity stratification on the linear development of the MRI.
113:
114: We consider the most important implicit assumptions
115: leading to the scenario of layered accretion
116: (Gammie 1996) are the following.
117: First, that the MRI develops in the active regions leading to
118: MHD turbulence, and angular momentum transport, as it does in
119: homogeneous disks, i.e. disks without diffusivity stratification.
120: Secondly, the stagnant condition giving
121: the dead zone its name in the model of
122: Gammie (1996) follows from the assumption that no significant
123: stresses, of Reynolds or Maxwell type, are present in such
124: region. The first assumption is motivated by the
125: results of the local, linear analysis carried out by Jin
126: (1996) who found that although ohmic magnetic diffusivity
127: can quench the MRI, at least for the minimum mass solar
128: nebula model, the ionization degree in the active
129: layers may be sufficient for the instability to develop.
130: However, as discussed by Sano \& Miyama (1999) a quasi-global
131: analysis, incorporating the boundary
132: conditions of the problem in the $z$ direction and
133: the stratification of the magnetic diffusivity, could yield
134: new constraints on the instability criterion. Moreover, at
135: least this level of nonlocality in the analysis of the MRI
136: is required if one is to say anything about the velocity and
137: magnetic field perturbations, and resulting stresses,
138: induced in the dead zone of PP disks.
139:
140: The aim of this paper is to analyze in detail
141: the emergence of the MRI across the dead zone of
142: protoplanetary disks. We address the main assumptions of
143: the layered accretion scenario outlined above. In
144: contrast to the recent, similar work by Sano \&
145: Miyama (1999) we adopt the simplest geometry and
146: perturbation type for which the MRI is known to arise
147: and focus
148: on the specific effects of a magnetic diffusivity
149: profile as that expected in standard, accretion disk
150: models of PP disks. We seek to analyze the
151: stability criterion and
152: $z$ dependence of the perturbations through the active
153: layers and dead zone. The implications
154: of our results on the dynamics of PP disks
155: are discussed for a wide range of disk
156: models.
157:
158: \section[]{Formulation of the problem}
159:
160: We consider the quasi-global, linear stability of a
161: keplerian flow threaded by a uniform vertical magnetic
162: field in a medium with $z$-dependent ohmic diffusivity.
163: Our quasi-global study follows in spirit the analysis
164: by Gammie \& Balbus (1994) but we include a variable
165: ohmic diffusivity in the magnetic induction equation.
166: This configuration is considered as a model of the middle
167: portions of protoplanetary disks where the dead zone is
168: proposed to exist (Gammie 1996). As a first approximation
169: to this problem, for reasons purely of mathematical
170: simplicity, we consider only the effect of ohmic magnetic
171: diffusion. The influence of other diffusive processes
172: on the MRI, such as ambipolar diffusion,
173: has been studied by Wardle (1997, 1999) in a local approach.
174: Although such processes are probably important in
175: regions of PP disks, their incorporation into a
176: global analysis is left for future contributions.
177:
178: In our quasi-global analysis the unperturbed system
179: is considered homogeneous in the radial and azimuthal
180: directions in all properties except the angular velocity,
181: which is a function of the radial distance from the
182: central star. We concentrate on the effect of
183: diffusion on the most unstable modes found in the
184: ideal MHD analysis, those corresponding to destabilized
185: Alfven waves traveling in the $z$-direction
186: (Balbus \& Hawley 1998). To this end we assume
187: an unperturbed weak magnetic field configuration
188: like ${\mathbf B} = B \hat{z}$ where $B$ is a constant.
189: The equilibrium condition for the flow, in the presence
190: of the gravitational force due to the central star,
191: corresponds to circular rotation with velocity
192: ${\mathbf U} = r \Omega \hat{\phi}$, where $\Omega$ is
193: equal to $\sqrt{G M_\star}/r^{3/2}$. This is the
194: simplest disk configuration where the instability
195: arises in the ideal MHD analysis (Balbus \& Hawley 1998).
196:
197: Our starting point are the governing continuity equation,
198:
199: \begin{equation}
200: \frac{\partial \rho}{\partial t} \ + \
201: {\mathbf \nabla} \cdot (\rho {\mathbf U})
202: \ = \ 0 ,
203: \label{conti}
204: \end{equation}
205:
206: \noindent momentum conservation equation,
207:
208: \begin{equation}
209: \frac{\partial {\mathbf U}}{\partial t} \ +
210: ({\mathbf U} \cdot {\mathbf \nabla} {\mathbf U}) \ =
211: \ - \frac{1}{\rho} {\mathbf \nabla} \left( P +
212: \frac{|{\mathbf B}|^2}{8 \pi} \right)
213: \ + \ \frac{1}{4 \pi \rho} ({\mathbf B} \cdot
214: {\mathbf \nabla} {\mathbf B})
215: \ - \ {\mathbf \nabla} \Phi ,
216: \label{momen}
217: \end{equation}
218:
219: \noindent where $\Phi$ is the gravitational potential from the
220: central star, and magnetic induction equation,
221:
222: \begin{equation}
223: \frac{\partial {\mathbf B}}{\partial t} \ = \
224: {\mathbf \nabla} \times ({\mathbf U} \times {\mathbf B}
225: \ - \ \eta {\mathbf \nabla} \times {\mathbf B}),
226: \label{induc}
227: \end{equation}
228:
229: \noindent where $\eta$ is the ohmic magnetic diffusivity, a
230: function of $z$ in general. The system physical
231: properties; velocity, magnetic
232: field, density, $\rho$, and pressure, $P$, are perturbed
233: with axisymmetric linear disturbances of
234: the form $ g(z) {\rm e}^{-\gamma t}$, where $g(z)$
235: represents the $z$-dependent amplitude of the
236: perturbation on any of the physical properties
237: of the system.
238:
239: In the approximation for geometrically thin disks
240: and weak magnetic fields, the restrictions mentioned above
241: allow us to reduce the governing equations,
242: (\ref{conti})-(\ref{induc}),
243: to a system of coupled equations similar to (8)-(14) of
244: Sano \& Miyama (1999). An important simplification
245: in our approach is that, in absence of a
246: toroidal component of the background field, $B_{\phi}$,
247: the $r$ and $\phi$ components of the momentum and
248: magnetic induction equations, containing the Alfven
249: modes, decouple from the density, pressure and
250: vertical velocity perturbations.
251:
252: We represent the perturbations of velocity and
253: magnetic field components as $u_r, \ u_\phi$
254: and $b_r, \ b_\phi$ respectively. There is no radial
255: dependence of the perturbation since we restrict our
256: analysis to vertically traveling perturbations.
257: Future contributions will deal with a more general
258: derivation.
259:
260: To first
order, we can write the equations for
261: the amplitudes of perturbed quantities as:
262:
263: \begin{equation}
264: - \gamma u_r \ - \ 2 \Omega u_{\phi} \ - \ \frac{B}{4\pi \rho} D b_r
265: \ = \ 0 ,
266: \label{ur8}
267: \end{equation}
268:
269: \begin{equation}
270: - \gamma u_{\phi} \ + \frac{\Omega}{2} u_r \ - \ \frac{B}{4\pi
271: \rho} D b_{\phi} \ = \ 0 ,
272: \label{uphi8}
273: \end{equation}
274:
275: \begin{equation}
276: - \gamma b_r \ - B D u_r \ - [ \eta D^2 + D\eta D ] b_r \ = \ 0 ,
277: \label{br8}
278: \end{equation}
279:
280: \begin{equation}
281: - \gamma b_{\phi} \ - \ B D u_{\phi} \ + \ \frac{3}{2} \Omega b_r
282: - [ \eta D^2 + D\eta D ] b_{\phi} \ = \ 0 ,
283: \label{bphi8}
284: \end{equation}
285:
286: \noindent where $D$ represent the operator $d/dz$.
287: The density, $\rho$ appearing in equations
288: (\ref{ur8})-(\ref{bphi8}) is the
289: unperturbed value, taken from the condition of
290: hydrostatic equilibrium in the $z$ direction for an
291: isothermal disk,
292:
293: \begin{equation}
294: \rho (z) \ = \ \rho_o {\rm e}^{-z^2/H^2} ,
295: \label{rho1}
296: \end{equation}
297:
298: \noindent where $\rho_o$ is the value of the density at the
299: disk midplane and $H = \sqrt{2} C_s/\Omega$ is the isothermal
300: scale height, with $C_s$ the sound speed in the gas.
301:
302: These system of equations is solved subject to the following
303: boundary conditions:
304:
305: \begin{equation}
306: \left. D u_r \right|_{z=\pm H_t} = \left. D u_\phi \right|_{z=\pm H_t} = 0
,
307: \label{bcsu}
308: \end{equation}
309:
310: \begin{equation}
311: \left. b_r \right|_{z=\pm H_t} = \left. b_\phi \right|_{z=\pm H_t} = 0
,
312: \label{bcsb}
313: \end{equation}
314:
315: \noindent where $H_t$ is the height above the midplane where
316: the boundary conditions are applied. These boundary conditions
317: correspond to a hot, tenuous halo model for the exterior of the disk,
318: in a force-free state
319: with the assumption of vanishing stress as $|z| \rightarrow \infty$.
320: Such model has been previously used
321: in similar quasi-global analysis of the MRI (Gammie \& Balbus 1994,
322: Sano \& Miyama 1999). The derivation,
323: presented by Gammie \& Balbus (1994), is simplified
324: to equations (\ref{bcsu}) and (\ref{bcsb}) in our case
325: of perturbations dependent only on $z$. We will return to discuss
326: the assumed boundary conditions in section 4.
327:
328:
329: \section{Numerical solution}
330:
331: To ease manipulation in numerically computing the growth rates of the
332: instability, $\gamma$,
333: we first normalize all variables by dividing by $\Omega$ and making
334: the
335: following substitutions:
336:
337: \begin{equation}
338: \hat{\gamma} = \gamma / \Omega,
339: \end{equation}
340: \begin{equation}
341: \hat{b}_i = b_i / B \ \ {\rm for} \ i = r, \phi,
342: \end{equation}
343: \begin{equation}
344: \eta = \eta_o f(z),
345: \end{equation}
346: \begin{equation}
347: \hat{z} = z / H ,
348: \end{equation}
349: \begin{equation}
350: \hat{u}_i = u_i / C_s \ \ {\rm for} \ i = r, \phi,
351: \end{equation}
352: \begin{equation}
353: \beta = C_s^2 / V_A^2 \ \ {\rm where} \ V_A^2 = B^2/4\pi\rho,
354: \end{equation}
355: \noindent and
356: \begin{equation}
357: R_m = H^2 \Omega / \eta_o
358: .
359: \end{equation}
360:
361: Our parameter $\beta$ is that typically used in
362: plasma physics and our $R_m$ is a magnetic Reynolds
363: number based on the velocity $C_s$ over a scale $H$.
364: Note that this differs slightly from the traditional definition
365: of $R_m$ in terms of the flow velocity appearing in the magnetic
366: induction equation. Since $\rho$
367: depends on $z$ in our model, so does $\beta$. We write
368: $\beta = \beta_c / h(z)$, where $\beta_c$ is the value of the
369: parameter at the disk midplane and $h(z) = {\rm e}^{\hat{z}^2}$
370: in our assumed isothermal disk structure.
371: The
resulting system of equations can be written as
372:
373: \begin{equation}
374: - \hat{\gamma} \hat{u}_r \ - \ 2 \hat{u}_{\phi} \ - \
375: \frac{1}{\sqrt{2}} \beta_c^{-1} h D \hat{b}_r \ = \ 0 , \label{ur9}
376: \end{equation}
377:
378: \begin{equation}
379: - \hat{\gamma} \hat{u}_{\phi} \ + \frac{1}{2} \hat{u}_r \ - \
380: \frac{1}{\sqrt{2}} \beta_c^{-1} h D \hat{b}_{\phi} \ = \ 0 ,
381: \label{uphi9}
382: \end{equation}
383:
384: \begin{equation}
385: - \hat{\gamma} \hat{b}_r \ - \frac{1}{\sqrt{2}} D \hat{u}_r \ -
386: R_m^{-1} [ f D^2 + Df D ] \hat{b}_r \ = \ 0 ,
387: \label{br9}
388: \end{equation}
389:
390: \begin{equation}
391: - \hat{\gamma} \hat{b}_{\phi} \ - \ \frac{1}{\sqrt{2}} D \hat{u}_{\phi} \
392: + \ \frac{3}{2} \hat{b}_r - R_m^{-1} [ f D^2 + Df D ]
393: \hat{b}_{\phi} \ = \ 0.
394: \label{bphi9}
395: \end{equation}
396:
397: The convenience of this normalization is making explicit the dependence
398: of the solution to our problem on the parameters $\beta_c$ and $R_m$, it
399: will also depend on the form of the diffusivity profile and indirectly
400: on the boundary conditions.
401:
402: Now we discretize the variables and coefficients on a set of points
403: $\hat{z}[i] \in [-\hat{H_t},\hat{H_t}]$
404: using the notation $g[i]$ to denote the value of
405: the variable $g$ at point $i$. The derivatives $D$ and $D^2$ of the
406: variables are
407: approximated with 2nd order centered finite differences on an
408: equispaced grid ($\hat{z}[i] = \hat{z}_{\rm min} + i * \Delta \hat{z}$).
409: With this, the system of 4 differential equations (\ref{ur9})-(\ref{bphi9})
410: becomes a system of 4$N$ coupled algebraic equations, where $N$ is the number
411: points in the grid. Corresponding to each point there are 4 equations so that
412: the discretized system of difference equations can be recast in matrix
413: form as
414:
415: \[
416: {\mathbf A}
417: \left( \begin{array}{c}
418: \hat{u}_r[1] \\
419: \hat{u}_\phi[1] \\
420: \hat{b}_r[1] \\
421: \hat{b}_\phi[1] \\
422: \cdot \\ \cdot \\ \cdot \\
423: \hat{u}_r[N] \\
424: \hat{u}_\phi[N] \\
425: \hat{b}_r[N] \\
426: \hat{b}_\phi[N]
427: \end{array} \right)
428: \ - \ \hat{\gamma}
429: \left( \begin{array}{c}
430: \hat{u}_r[1] \\
431: \hat{u}_\phi[1] \\
432: \hat{b}_r[1] \\
433: \hat{b}_\phi[1] \\
434: \cdot \\ \cdot \\ \cdot \\
435: \hat{u}_r[N] \\
436: \hat{u}_\phi[N] \\
437: \hat{b}_r[N] \\
438: \hat{b}_\phi[N]
439: \end{array} \right)
440: \ = \ 0
441: \]
442:
443: \noindent where the growth rate of our perturbations, $\hat{\gamma}$, are
444: seen to be the eigenvalues of the matrix $\mathbf A$ made up of the
445: corresponding terms from the discretization of the
446: system of equations (\ref{ur9})-(\ref{bphi9}).
447:
448: We calculate eigenvalues numerically transforming the matrix to upper
449: Hessenberg form and using a QR algorithm. The number of points in the
450: grid was varied from 10 to 50 with no significant change (less than 1\%)
451: in the value
452: of the maximum growth rate of the instability. Results shown correspond
453: to a discretization over 20 points for $\hat{z}$.
454:
455:
456: The boundary conditions are
457: incorporated rewriting the corresponding
458: equations (\ref{ur9})-(\ref{bphi9}) for the first and last points
459: of our grid,
460: upon discretization of the conditions (\ref{bcsu}) and (\ref{bcsb})
461: with 2nd order forward and backward differences at the bottom and top
462: disk surfaces respectively.
463:
464: \section{Results}
465:
466: With the application to the middle, poorly ionized
467: portions of PP disks in mind,
468: we compute growth rates and eigenmodes (the
469: form of the perturbations) for weak unperturbed magnetic fields ($\beta_c
470: \geq 0.1$) and moderately diffusive disks.
471: We consider the later condition as $ 1 < R_m \leq 10^{3}$
472: since the parameter $R_m$
473: can be seen as the ratio of the characteristic diffusive timescale
474: over a length $H$ to the dynamical timescale.
475:
476: \subsection{Case of uniform diffusivity}
477:
478: A global analysis for a homogeneous disk with ohmic diffusivity
479: has recently been presented by
480: Sano \& Miyama (1999). Here we repeat their calculation following a
481: slightly different approach, and a different computational method,
482: to ease the comparison to the results for a stratified disk.
483:
484: To consider a homogeneous diffusivity we simply put $Df = 0$
485: and $f = 1$ in the system (\ref{ur9})-(\ref{bphi9}).
486: Figure \ref{fig1} shows the value of the maximum growth rate of the
487: instability, given by the minimum value of $\hat{\gamma}$ with no
488: imaginary part, as a
489: function of the resistivity reflected in
$R_m$, and the
490: unperturbed magnetic field strength, entering through
491: the parameter $\beta_c$.
492:
493: Fitting a line through the nodes of the function depicted in Figure
494: \ref{fig1} indicates that the MRI arises as long as the following
495: relation holds:
496:
497: \begin{equation}
498: \beta_c < R_m^2
499: \label{coneta1}
500: \end{equation}
501:
502: \noindent with the stated weak field, weak diffusion restrictions,
503:
504: \begin{equation}
505: \beta_c > 1
\ {\rm and} \ R_m
> 1
506: \end{equation}
507:
508: We can recast condition (\ref{coneta1}) in terms of the
509: characteristic timescales
510: of the problem: the diffusion timescale,
511: $t_{\rm dif} = H^2/\eta$ and the
512: Alfven wave $H$-crossing time, $t_{\rm A} = H/V_A$. For a density
513: stratified disk with constant magnetic diffusivity, the MRI will
514: arise at a given radius if:
515:
516: \begin{equation}
517: t_{\rm A} \lesssim t_{\rm dif}
518: \label{coneta2}
519: \end{equation}
520:
521: In order for the instability to develop an Alfven wave, with the
522: velocity at the disk midplane, must have
523: time to cross the disk before the field diffuses over the same
524: length scale. This result, summarized in
525: condition (\ref{coneta1}), is equivalent to the requirement
526: that the Lundquist number be greater than unity. The
527: Lundquist number is the ratio of the Alfven velocity to the
528: ``velocity'' of drift of magnetic field lines through the
529: conductor due to diffusion.
530: These results are consistent with Sano \& Miyama (1999).
531:
532: \subsection{Effect of diffusivity stratification}
533:
534: The calculation of growth rates for a stratified disk is now a simple
535: procedure of prescribing a particular analytic profile for $f(z)$ and
536: following the steps described above.
537:
538: As we have mentioned we expect
539: the diffusivity to increase rapidly from its minimum value at the disk
540: surfaces. The results we present consider a $z$-dependence of
541: the ohmic diffusivity like:
542:
543: \begin{equation}
544: f(\hat{z}) = {\rm e}^{(\hat{H}_t^2-\hat{z}^2)} \
545: {\rm e}^{({\rm Erf}(\hat{H}_t) - {\rm Erf}(\hat{z}))/l_o}
546: \label{perfil}
547: \end{equation}
548:
549: \noindent where Erf refers to the well known Error function.
550: The precise form
551: for this profile and particular
552: choices for $l_o$ will be justified in the discussion section.
553: In addition to $\beta_c$ and $R_m$,
554: in a disk with diffusivity stratification two new parameters
555: $l_o$ and $H_t$, determine
556: the diffusivity profile at a given radius in the disk and
557: the growth rates of the MRI.
558:
559:
560: Figures \ref{fig2} and \ref{fig3} show the maximum growth rate of the
561: MRI in disks with different $\beta_c$, as a function of the
562: stratification parameters $R_m$ (a measure of the ohmic magnetic
563: diffusivity at the surface of the disk) and $l_o$. The range in
564: both parameters is chosen with the application to PP disks in mind,
565: as we discuss in the next section. Figure
566: \ref{fig2} corresponds to an unperturbed magnetic field strength with
567: $\beta_c = 10^{2}$, which we call the strong magnetic field case.
568: Figure \ref{fig3} corresponds to $\beta_c = 10^{4}$, referred to as the
569: weak magnetic field case.
In both cases the surface of the disk is
570: considered to be at $z=H$.
571:
572: Depending on the strength
of the unperturbed magnetic field
573: the effect of the stratification can be determinant for the existence
574: of the instability. Regions of the disk with characteristic
575: diffusive timescales at the surface
576: less than 10 dynamical timescales, will not be unstable for all
577: unperturbed magnetic fields considered, $\beta_c > 10^{2}$.
578: The instability will
579: also be generally damped if the stratification is strong
580: $l_o^{-1} \ga 10$. In the next section we discuss the application
581: of our results to PP disks.
582:
583: Stratification can also affect the form of the perturbations
584: for the fastest growing modes. Figure \ref{fig4} shows the
585: $z$-dependent amplitude of the three fastest growing unstable
586: modes for a region with moderate stratification ($l_o^{-1} = 1$) and
587: diffusivity near the stabilizing limit ($R_m = 20$). The
588: surface of the disk is placed at $H_t = 2 H$ and the unperturbed
589: magnetic field corresponds to $\beta_c = 10^{2}$. An interesting
590: feature of this result is the ``stratification'' of the instability,
591: i.e. the excitation of the instability preferentially in the
592: regions away from the disk midplane. The largest
593: amplitudes in the linear regime, both for velocity and magnetic
594: field perturbations, appear at $z \approx H$.
595:
596:
597: \section{Discussion}
598:
599: We have seen that the behavior of the MRI in stratified media depends
600: crucially on the value and form of the diffusivity profile. It also
601: depends on the magnitude of the unperturbed magnetic field originally
602: present in the disk. In this section we argue what are the likely
603: values of this parameters in the context of PP disk models and discuss
604: plausible consequences of our results.
605: However, before doing this
606: we briefly discuss the implications of modifying some of the
607: main assumptions of our analysis.
608:
609: \subsection{Boundary conditions and field geometry}
610:
611: As discussed in section 1 the adopted boundary conditions
612: follow from modeling the exterior medium
613: as a hot tenuous corona,
614: also called a hot halo by Gammie \& Balbus (1994).
615: The continuity of the stress tensor resulting from
616: the perturbations across the disk-halo boundary, and
617: the condition that it vanish at infinity lead to
618: restrictions (\ref{bcsu}) and (\ref{bcsb}).
619: While this is apparently a natural choice for isolated
620: disks, different sets of boundary conditions
621: could certainly be constructed and, in principle,
622: these could affect the results presented so far.
623:
624: Gammie \& Balbus (1994)
625: propose two other models for the disk exterior:
626: a tenuous atmosphere, actually the continuation of the disk
627: with an isothermal density profile, and the so-called
628: rigid conductor model, introduced to mimic the effect of
629: having a sink of angular momentum outside the disk.
630: In the former, infinite disk case, no new unstable
631: modes are found by Gammie \& Balbus (1994) in comparison
632: to the hot halo model.
633: In the second case, when a load is connected to the
634: field lines, a new mode is found which can be
635: unstable even when all other modes are stable.
636: As discussed by Gammie \& Balbus (1994) the new
637: unstable global mode arises for strong seed magnetic
638: fields ($\beta_c \sim 1$). Although the results of
639: Gammie \& Balbus (1994) assume ideal
640: MHD for the disk material, we believe similar conclusions will be
641: found in our case since the introduction of resistivity
642: has not resulted in the appearance of new unstable modes
643: (Jin 1996, Sano \& Miyama 1999).
644:
645: In addition to the nature of the interface,
646: the precise location of the disk-halo boundary is also
647: assumed. This is done in our model
648: through the parameter $H_t$ in equation (\ref{perfil}),
649: assumed equal to
650: the isothermal scale height in the results of the previous section.
651: Our choice is motivated by the calculations by
652: D'alessio et al. (1999) of the detailed vertical
653: structure of PP disks in which the density drops
654: rapidly by orders of magnitude after
655: $\sim$ 1-2 isothermal scale heights. The termination
656: of the disk at the scale height
657: $H$ is also an appropriate choice if one considers
658: a polytropic rather
659: than isothermal model for the vertical structure.
660: Alternatively one could argue that the transition
661: from the disk to the halo should be located at the
662: point where $\beta$ equals unity, i.e. where the magnetic
663: field passes from being a weak dynamical agent,
664: with respect to gas pressure, to being the dominant
665: dynamical factor, as is expected in the corona.
666: Such assumption, with our isothermal disk
667: structure, places $H_t$ at approximately 2 and 3
668: scaleheights for $\beta_c = 10^{2}$ and $10^{4}$
669: respectively. Figure \ref{fig5} shows the regions
670: of stability and instability as a function of
671: $R_m$ and $l_o^{-1}$ for different values of
672: the parameter $\beta_c$.
673: In comparison to figures
674: \ref{fig2} and \ref{fig3}, and to figure \ref{fig7}
675: to be discussed below, we see that our results
676: are not changed drastically, although the trend to
677: make the disk more unstable as one increases the
678: vertical extent is significant. As one incorporates into the
679: model regions of lower $\beta$, which are
680: more unstable, the damping effect of the poorly
681: ionized middle regions is less important.
682:
683: \subsection{The value of $\beta_c$}
684:
685: The parameter $\beta_c$ depends on the strength of the unperturbed
686: magnetic field seeding the instability. In PP disks a
687: weak seed magnetic field can be expected as a remnant of the
688: disk formation process out of a magnetized molecular cloud.
689:
690: Measurements of magnetic fields in molecular clouds indicate that the
691: densest regions, where PP disks are being formed, are commonly threaded
692: by ordered fields with magnitudes around $\sim 10^{-2}$ Gauss
693: (Heiles et al. 1993). If we consider this as the seed field
694: for the MRI, our $B$, for the typical midplane
695: densities $10^{-9}-10^{-11}$ gm/cm$^3$
696: and temperatures $10 - 500$ K predicted for PP disks around a few AU,
697: this implies that $10^{2} \la \beta_c \la 10^{5}$.
698:
699: Somewhat stronger fields would serve to seed the instability
700: near the central star, if it has a significant magnetic field.
701: However, at radial distances $\ga 1$ AU where the dead zone
702: is proposed to exist, such field would most likely be weak,
703: with $\beta_c$ in the range considered.
704:
705: \subsection{Ionization state of PP disks}
706:
707: We now attempt to justify the range of parameters $l_o^{-1}$ and
708: $R_m$ explored in section 3 on the basis of the expected
709: properties of protoplanetary disks.
710:
711: Traditionally the evolution
712: of the solar nebula, the archetype of PP disks, has been thought
713: of as consisting of 3 stages. In the first stage, infall from
714: the molecular cloud leads to the formation and growth of an
715: hydrostatic disk. In this so-called formation stage, efficient
716: angular momentum transport resulting from self gravity instabilities
717: leads to a fast evolution which, after some $\sim 10^5$ years,
718: renders most of the disk stable to gravitational instabilities.
719: What follows has usually been called the viscous stage as it is
720: believed that the long term evolution of the disk in this phase
721: is mainly governed by anomalous ``viscous" torques resulting on the
722: accretion of mass through the disk and onto the central object.
723: Finally, in the dispersal stage the remaining gas and
724: dust from the process of planet formation is somehow
725: removed. Partly because diffusive, viscous processes act on an ever
726: increasing timescale in the disk, it has been suggested that
727: external agents dominate this final stage of PP disk evolution
728: (see Hollenbach et al. 2000 and references therein).
729:
730: As mentioned in the introduction, it is in connection to the
731: origin of the anomalous torques, dominant through the $\sim 10^7$
732: year lifetime of the viscous stage, that the MRI has become such
733: an important part of accretion disk theory. In this paper,
734: our aim is to determine whether the MRI will arise in poorly ionized
735: portions of PP disks (a prerequisite for generating viscous torques)
736: at the beginning and possibly during the viscous stage. Hence, we
737: consider models for viscous stage protoplanetary disks.
738: Over the last 20 years $\alpha$-disk
739: models have been constructed addressing the influence of different
740: agents on the disk's viscous evolution (see Ruden \& Lin 1986, Ruden
741: \& Pollack 1991, Sterzik \& Morfill 1994, and Reyes-Ruiz \&
742: Stepinski 1995, among others). Apart from temporal and detailed
743: differences in the structure of the disk, the overall properties
744: of the disk generally agree well with the models
745: of the Ruden \& Lin (1986). In fact with the exception of the
746: outermost regions, near the outer boundary of the disk,
747: steady state models as those described in detail in
748: Stepinski et al. (1993) reproduce satisfactorily the
749: properties of PP disks at a given instant.
750:
751: In this paper we adopt such steady state models of PP disks
752: (Stepinski et al. 1993) to describe the radial variation
753: of physical properties as the disks emerge from the formation
754: stage of their evolution. The models incorporate the Shakura \&
755: Sunyaev (1973) viscosity prescription and temperature dependent
756: opacity laws as proposed by Ruden \& Pollack (1991).
757: The physical properties of the disk are given by a
758: sequence of power laws, described in Stepinski et al. (1993),
759: distributed radially
760: depending on the two basic parameters of the model:
761: the turbulence viscosity parameter of Shakura \& Sunyaev,
762: $\alpha_{ss}$, and the uniform mass accretion rate through the
763: disk, $\dot{M}$. Values of $\alpha_{ss}$ between
764: $10^{-2}$ and $10^{-3}$ are
765: usually quoted from simulations of the MRI in
766: shearing boxes with keplerian differential rotation. As
767: such values seem to be also consistent with observations
768: of T-Tauri disks (D'alessio et al. 1998) we will present
769: results for the limiting values. Again both in models and
770: observations (see review by Calvet et al. 2000) the mass
771: accretion rate ranges from $\sim 10^{-6} M_\odot$/yr, at the
772: beginning of the viscous stage, to $\sim 10^{-8} M_\odot$/yr
773: through most of their lifetime.
774:
775: A particular choice of model parameters
776: determines the profiles of midplane density
777: and temperature, surface density, scale height,
778: etc. These, and our isothermal vertical structure model,
779: can be used to calculate the ionization
780: degree throughout the disks following
781: Stepinski et al. (1993). In the innermost disk
782: regions, inwards of $\sim 1$ AU, we consider
783: thermal ionization of Potassium assumed present
784: in PP disks with solar abundance.
785: As found by Stepinski (1992) in regions with temperature
786: above $\sim 10^{3}$ K, the ionization degree is
787: high enough to render the disk unstable to the
788: MRI ($R_m \gg 1$). The development of the MRI
789: in this region, called the inner active region (IAR)
790: by Stepinski (1999) proceeds as analyzed by Gammie
791: \& Balbus (1994) and is not addressed in this paper.
792:
793: We concentrate on the regions outwards of the IAR,
794: defined by $r > R_{IAR}$, where $R_{IAR}$ is defined
795: as the radius where the temperature drops below
796: 1000 K. There, ionization is mainly due to the action
797: of galactic cosmic rays penetrating through the
798: disk surfaces. We obtain an upper limit to the ionization
799: fraction neglecting the
800: recombination of electrons onto dust grains and
801: considering only ion reactions as a sink of electrons.
802: Using the ionization and recombinations rates of
803: Stepinski (1992), the ionization fraction, $x$, can be
804: obtained from the condition of ionization equilibrium,
805:
806: \begin{equation}
807: x \ = \ 4 \times 10^{-18} \rho(z)^{- 1/2}
808: \ {\rm e}^{\frac{-S(z)}{2 S_o}} ,
809: \end{equation}
810:
811: \noindent where $S_o$ is the characteristic cosmic ray
812: attenuation density, $\sim 100$ gm/cm$^2$, and
813: $S(z)$ is the integrated surface density from the disk
814: surface to the height $z$:
815:
816: \begin{equation}
817: S(z) \ = \ \int_z^{H_t} \rho_c \ {\rm e}^{- \frac{z^2}{H^2}} \ dz.
818: \end{equation}
819:
820: Considering the rapid decrease of the argument, for mathematical
821: simplicity we approximate this function as:
822:
823: \begin{equation}
824: S(\hat{z}) \ \approx \ \frac{\sqrt{\pi}}{2} \ \rho_c \ H \
825: (1 - {\rm Erf}(\hat{z})),
826: \end{equation}
827:
828: \noindent so that the ionization degree can be written as:
829:
830: \begin{equation}
831: x(\hat{z}) \ = \ 4 \times 10^{-18} \ \rho_c^{-1/2} \
832: h^{-1/2} \
833: {\rm e}^{-(1 - {\rm Erf}(\hat{z}))/l_o} ,
834: \end{equation}
835:
836: \noindent where $l_o^{-1} = \sqrt{\pi} \rho_c H/4 S_o$. With
837: ohmic diffusivity given by $\eta = 7 \times 10^3 / x$
838: (Stepinski 1992) we arrive at equation (\ref{perfil}) for
839: the diffusivity profile.
840:
841: In figure \ref{fig6} we show the radial profiles of the parameters
842: $R_m$ and $l_o^{-1}$ characterizing the diffusivity profile,
843: for a series of specific models of PP disks. Two values of the
844: turbulent viscosity parameter are considered, $\alpha = 10^{-2}$
845: and $10^{-3}$. The mass accretion rate through the disk
846: takes the values $10^{-6}$, $10^{-7}$ and $10^{-8} \ M_\odot$/yr.
847: Models with lower $\alpha$ and higher accretion rate are
848: characterized by a higher temperature and surface density at
849: a given radius. For such models the IAR extends farther
850: out in the disk in comparison to corresponding models with
851: higher $\alpha$ and lower $\dot{M}$. Also, in comparison
852: beyond $R_{\rm IAR}$ their diffusivity is
853: higher, i.e. smaller $R_m$, as is their stratification reflected
854: in the parameter $l_o$.
855:
856: We summarize the results of this section in Figure \ref{fig7}
857: where a direct comparison of the properties of PP disk models and
858: the regions of stability for different values of $\beta_c$ is shown.
859: Depending on the value of $\beta_c$, marked radii allow us to
860: determine in what regions of the disk the MRI will arise.
861: For example, for the disk model with $\alpha = 10^{-2}$ and
862: $\dot{M} = 10^{-8} M_\odot $/yr (short dashed line), even if the
863: seed field is as strong as $\beta_c = 10^{2}$, in a broad region
864: between several AU and the $R_{\rm IAR}$ (square), located at less than
865: 0.3 AU from Figure \ref{fig6}, the MRI will not arise. For the same model,
866: if $\beta_c = 10^{4}$, outwards of $R_{\rm IAR}$ the whole disk will
867: be stable. A similar analysis
868: can be performed for the reader's preferred specific model.
869:
870: Finally, it is worth stressing that the calculated diffusivities
871: represent a lower bound as we neglect recombination onto dust grains,
872: which can be dominant if their size distribution is similar to
873: that of the ISM (see Reyes-Ruiz \& Stepinski 1995).
874:
875: \section[]{Summary and Conclusions}
876:
877: We have conducted a quasi-global analysis of the
878: behavior of the MRI in the linear regime considering
879: the effects of diffusivity stratification as that
880: expected in protoplanetary disks. As our main result
881: we find that stratification can add
882: to the stabilizing effect of ohmic diffusion found in
883: previous calculations.
884:
885: Ohmic diffusivity alone, without stratification, will
886: damp the MRI if $R_m \la 10$, regardless of the strength
887: of the unperturbed magnetic field. Weak seed magnetic fields,
888: $\beta_c \ga 10^{4}$, will become stable if $R_m \la 100$,
889: i.e. if the diffusion timescale is up to 100 times longer
890: than the dynamical timescale. In addition, the stabilizing
891: effect of stratification is significant when it
892: is characterized by values of the
893: parameter $l_o^{-1} \ga 1$, i.e. when the surface density if greater
894: than $S_o \approx 100$ gm/cm$^2$.
895:
896: \subsection{Revised model of PP disks}
897:
898: If the MRI is the only source of
899: turbulence and anomalous viscosity in protoplanetary
900: disks, our results suggest the necessity of a
901: revision of standards models. The calculation
902: of properties of these revised models, time dependent
903: in essence, is out of the scope of this paper.
904: However, one can speculate
905: on a plausible structure of the revised PP disk models
906: based on the dynamical agents operating at different radii.
907:
908: In most of the models, outwards of $R_{\rm IAR}$ there
909: is a broad region where the MRI does not operate.
910: The flow in such region would not be turbulent and,
911: in absence of anomalous viscous torques, mass would
912: accumulate in a stable ``dead'' region (DR).
913:
914: In some models,
915: for example the one with $\alpha = 10^{-2}$ and
916: $\dot{M} = 10^{-7} M_\odot$/yr, the instability
917: can be excited also in the outermost parts of the
918: disk. This could occur first in a ``stratified''
919: manner (as shown in figure \ref{fig4}) due to
920: the strong stratification of the diffusivity.
921: Whether such excitation leads, in the nonlinear regime,
922: to a layered accretion scenario as proposed by
923: Gammie (1996), is the subject of numerical
924: simulations. We propose the existence of a region of
925: layered accretion (LAR) with
926: the cautionary comment that, according to our linear
927: analysis, significant velocity and magnetic field
928: perturbations will also be present in the middle
929: ``dead" zone. These could possibly lead, in the
930: nonlinear regime, to somewhat smaller, but
931: non-zero viscous torques in the so-called dead zone.
932:
933: Regions at a still greater
934: distance from the central object are less
935: diffusive and less stratified, so the operation of
936: the MRI in such regions would be again similar to
937: that in homogeneous disk and
938: turbulence can be generated across the whole vertical
939: extent of the disk in what we call the outer active region (OAR).
940: In conclusion, the results derived in this paper with the
941: assumptions discussed above lead to a revised model
942: for the structure of protoplanetary disks as
943: schematically illustrated in figure \ref{fig8}.
944:
945: \subsection{Caveats of our analysis}
946:
947: We end this contribution pointing out potentially
948: important inconsistencies in our analysis which
949: could affect our conclusions.
950:
951: First, it is clear that our disk models are dynamically
952: inconsistent. They are derived assuming a constant
953: turbulent viscosity parameter $\alpha$ which, as we
954: show in this paper, is not justified if turbulence is
955: due to the MRI. At most, we could argue that such
956: models reflect the properties of PP disks as they
957: emerge from the self-gravitating formation stage
958: to the viscous evolutionary phase where the MRI is
959: believed to be the dominant player in the transport
960: of angular momentum. Our assumption is made
961: as a means to show the necessity of revising the
962: standard, constant $\alpha$ models of protoplanetary disks
963: as well as the layered accretion scenario.
964:
965: A similar inconsistency is also present in the
966: calculation of the vertical diffusivity profile since,
967: as shown by Dolginov \& Stepinski (1994), if the
968: turbulence in the disk involves a tangled magnetic
969: field, the shielding of cosmic rays can be significantly
970: larger than estimated here. This effect will certainly be
971: important in the nonlinear regime of the instability.
972:
973: %
974: %*******FIGURE 2***********
975:
976: \vspace{10mm}
977:
978: \noindent{\bf Acknowledgements}
979: The author is thankful to Adriana Gazol for helpful comments
980: at the beginning of this work.
981: This work has been supported by CONACYT project J22990E.
982:
983:
984: \begin{thebibliography}{}
985: \bibitem{b1} Balbus, S.A. \& Hawley, J.F. 1991, ApJ, 376, 214
986: \bibitem{b2} Balbus, S.A. \& Hawley, J.F. 1998, Rev. Modern Physics, 70, 1
987: \bibitem{b3} Brandenburg, A. et al. 1995, ApJ, 446, 741
988: \bibitem{b34} Calvet, N., Hartmann, L. \& Strom, S. 2000, in
989: Mannings, V., Boss, A.P., Russell, S.S., eds, Protostars and Planets IV,
990: University of Arizona Press: Tucson.
991: \bibitem{b4} D'alessio, P., Canto, J., Calvet, N. \& Lizano, S.,1998, ApJ, 500,
992: 411
993: \bibitem{b4} D'alessio, P., Calvet, N., Hartmann, L., Lizano, S.
994: \& Canto, J.,1999, ApJ, 527, 893
995: \bibitem{b5} Dolginov, A. \& Stepinski, T.F. 1994, ApJ, 427, 377
996: \bibitem{b5b} Eardley, D.M. \& Lightman, A.P. 1976, ApJ, 200, 181
997: \bibitem{b6} Gammie, C.F. 1996, ApJ, 457, 355
998: \bibitem{b7} Hartmann, L., Calvet, N., Gullbring, E. \& D'alessio, P.
999: 1998, ApJ, 495,
1000: 385
1001: \bibitem{b8} Hawley, J.F., Gammie, C.F. \& Balbus S.A. 1995, ApJ, 440, 742
1002: \bibitem{b9} Hawley, J.F., Gammie, C.F. \& Balbus S.A. 1996, ApJ, 464, 690
1003: \bibitem{b10} Heiles, C. et al 1993, in Levy, E.H. \& Lunine, J.I., eds,
1004: Protostars and Planets III. University of Arizona Press, Tucson, p279
1005: \bibitem{b34} Hollenbach, D.J., Yorke, H.W. \& Johnstone, D. 2000, in
1006: Mannings, V., Boss, A.P., Russell, S.S., eds, Protostars and Planets IV,
1007: University of Arizona Press: Tucson.
1008: \bibitem{b11} Jin, L. 1996, ApJ, 457, 798
1009: \bibitem{b11b} Lynden-Bell, D. 1969, Nature, 223, 690
1010: \bibitem{b12} Reyes-Ruiz, M. \& Stepinski, T.F. 1995, ApJ, 438, 750
1011: \bibitem{b13} Ruden, S.P. \& Pollack, J.B. 1991, ApJ, 375, 740
1012: \bibitem{b14} Ruden, S.P. \& Lin, D.N.C. 1986, ApJ, 308, 883
1013: \bibitem{b24} Sano, T. \& Miyama, S.M. 1999, ApJ, 515, 776
1014: \bibitem{b24b} Shakura, N.I., \& Sunyaev, R.A.1973, \aap, 24, 337
1015: \bibitem{b25} Stepinski, T.F. 1992, Icarus, 97, 130
1016: \bibitem{b26} Stepinski, T.F. 1998, ApJ, 507, 361
1017: \bibitem{b26} Stepinski, T.F. 1999, in Proceedings of the 30th Annual Lunar
1018: and Planetary Science Conference, abstract no.1205
1019: \bibitem{b27} Stepinski, T.F., Reyes-Ruiz, M. \& Vanhala, H.A.T. 1993, Icarus,
1020: 106, 71
1021: \bibitem{b31} Stone, J.M., Hawley, J.F. \& Gammie C.F., Balbus S.A.
1022: 1996, ApJ, 463, 656
1023: \bibitem{b33} Umebayashi, T. \& Nakano, T. 1988, Prog. Theor. Phys. Suppl., 96, 151
1024: \bibitem{b34} Wardle, M. 1997, in Wickamasinghe D., Ferrario L.,
1025: Bicknell G., eds, Proc. IAU Colloq. 163, Accretion Phenomena and Related
1026: Outflows. ASP Press: San Francisco, p.561
1027: \bibitem{b49} Wardle, M. 1999, MNRAS, 307, 849
1028: %\bibitem[Iben 1991]{ib91} Iben, I. 1991, \apjs, 76, 55.
1029: \end{thebibliography}
1030:
1031: \clearpage
1032: %
1033: %*******FIGURE 1*************************************************
1034: %
1035: \begin{figure}
1036: %\epsfysize=6.5cm
1037: %\epsfxsize=8.5cm
1038: %\epsffile{fig1.ps}
1039: \plotone{fig1.ps}
1040: \caption{Maximum growth rate of the magnetorotational
1041: instability as a function of
1042: $\beta_c$ and $R_m$, in a medium with uniform ohmic
1043: diffusivity.}
1044: \label{fig1}
1045: \end{figure}
1046: %
1047: %**********END FIGURE 1******************************************
1048: %
1049:
1050: \clearpage
1051:
1052: %
1053: %*******FIGURE 2************************************************
1054: %
1055: \begin{figure}
1056: \plotone{fig2.ps}
1057:
1058: \caption{Maximum growth rate of the MRI in a stratified disk with
1059: $\beta_c = 10^{2}$ as a function of $l_o^{-1}$ and $R_m$. The
1060: diffusivity profile is given by equation (\ref{perfil}) with the
1061: surface of the disk placed at $z = H$. }
1062: \label{fig2}
1063: \end{figure}
1064: %
1065: %**********END FIGURE 2*****************************************
1066: %
1067:
1068: \clearpage
1069:
1070: %
1071: %*******FIGURE 3************************************************
1072: %
1073: \begin{figure}
1074: \plotone{fig3.ps}
1075:
1076: \caption{Same as figure \ref{fig2} but for an unperturbed magnetic field
1077: corresponding to $\beta_c = 10^{4}$. Notice the change of scale in
1078: the $z$ axis.}
1079: \label{fig3}
1080: \end{figure}
1081: %
1082: %**********END FIGURE 3*****************************************
1083: %
1084:
1085: \clearpage
1086:
1087:
1088: %
1089: %*******FIGURE 4************************************************
1090: %
1091: \begin{figure}
1092: \plotone{fig4.ps}
1093:
1094: \caption{Fastest growing modes (from top to bottom) of the MRI in
1095: a stratified disk with $R_m = 20$, $l_o^{-1} = 1$ and
1096: $H_t = 2 H$. The unperturbed magnetic field corresponds to
1097: $\beta_c = 10^{2}$. Left and right panels correspond to velocity and
1098: magnetic field perturbations respectively. In all cases the solid
1099: (dashed) line indicates the azimuthal (radial) component.
1100: }
1101: \label{fig4}
1102: \end{figure}
1103: %
1104: %**********END FIGURE 4 *****************************************
1105: %
1106:
1107: \clearpage
1108:
1109: %
1110: %*******FIGURE 5************************************************
1111: %
1112: \begin{figure}
1113: \plotone{fig5.ps}
1114:
1115: \caption{Regions of $l_o^{-1} - R_m^{-1}$ parameter
1116: space where the MRI arises for $H_t = 3 H$. The shaded region
1117: illustrates unstable combinations of the diffusivity
1118: parameters for $\beta_c = 10^{4}$. The 2 thin solid lines indicate the
1119: boundary of stability in disks with $\beta_c = 10^{3}$ (lower line)
1120: and $\beta_c = 10^{2}$ (upper line).}
1121: \label{fig5}
1122: \end{figure}
1123: %
1124: %**********END FIGURE 5*****************************************
1125: %
1126: \clearpage
1127:
1128: %
1129: %*******FIGURE 6************************************************
1130: %
1131: \begin{figure}
1132: \plotone{fig6.ps}
1133:
1134: \caption{Radial profiles of ionization properties, reflected in the parameters
1135: $R_m$ and $l_o^{-1}$ for different models
1136: of protoplanetary disks. In the legend of linestyles the
1137: first number indicates the value the $\alpha$ parameter and
1138: the second the mass accretion rate in solar masses per year.
1139: For each model the square indicates the position of the $R_{\rm IAR}$.}
1140: \label{fig6}
1141: \end{figure}
1142: %
1143: %**********END FIGURE 6*****************************************
1144: %
1145:
1146:
1147: \clearpage
1148:
1149: %
1150: %*******FIGURE 7************************************************
1151: %
1152: \begin{figure}
1153: \plotone{fig7.ps}
1154:
1155: \caption{Comparison of $R_m$ and $l_o^{-1}$ values for various PP disk
1156: models with the stability conditions resulting from our analysis. Disk
1157: models and corresponding linestyles are the same as those presented
1158: in figure \ref{fig5}. The leftmost edge of such lines
1159: correspond to the value of $R_m^{-1}$ and $l_o^{-1}$ at 60 AU.
1160: Over each line the square marks the $R_{\rm IAR}$ and the triangle
1161: and diamond mark where $R = 1$ and $R = 10$ AU respectively.
1162: }
1163: \label{fig7}
1164: \end{figure}
1165: %
1166: %**********END FIGURE 7 *****************************************
1167: %
1168:
1169: \clearpage
1170:
1171: %
1172: %*******FIGURE 8************************************************
1173: %
1174: \begin{figure}
1175: \plotone{fig8.ps}
1176:
1177: \caption{Illustration of the general structure of PP disks,
1178: indicating active and dead regions as suggested by our results.}
1179: \label{fig8}
1180: \end{figure}
1181: %
1182: %**********END FIGURE 8*****************************************
1183: %
1184:
1185: \end{document}
1186:
1187:
1188:
1189:
1190:
1191:
1192:
1193:
1194: