1: \documentclass[12pt,preprint]{aastex}
2: %\usepackage[dvips]{graphicx}
3:
4: \newcommand{\vdag}{(v)^\dagger}
5: \newcommand{\myemail}{marikok@geo.titech.ac.jp}
6:
7: %\slugcomment{to appear in Astrophys. J.}
8:
9: \shorttitle{Disk Angular Velocity Modified by MRI}
10: \shortauthors{Kato et al.}
11:
12: \begin{document}
13:
14: \title{Modification of Angular Velocity by Inhomogeneous MRI Growth
15: in Protoplanetary Disks}
16:
17: \author{M. T. Kato, K. Nakamura, R. Tandokoro}
18: \affil{Department of Earth and Planetary Science, Tokyo Institute of Technology, Ookayama 2-1-12-I2-10, Meguro-ku, Tokyo}
19: \email{marikok@geo.titech.ac.jp}
20:
21: \author{M. Fujimoto}
22: \affil{Institute of Space and Astronomical Science, Japan Aerospace Exploration Agency, Yoshinodai 3-1-1, Sagamihara, Kanagawa}
23:
24: \and
25:
26: \author{S. Ida}
27: \affil{Department of Earth and Planetary Science, Tokyo Institute of Technology, Ookayama 2-1-12-I2-10, Meguro-ku, Tokyo}
28:
29: \begin{abstract}
30: We have investigated evolution of magneto-rotational instability (MRI) in protoplanetary disks that have radially non-uniform magnetic field such that stable and unstable regions coexist initially,
31: and found that a zone in which the disk gas rotates
32: with a super-Keplerian velocity emerges as a result of
33: the non-uniformly growing MRI turbulence.
34: We have carried out two-dimensional resistive MHD simulations
35: with a shearing box model.
36: We found that if the spatially averaged magnetic Reynolds number, which is determined by widths of the stable and unstable regions in the initial conditions and values of the resistivity, is smaller than unity, the original Keplerian shear flow is transformed to the quasi-steady flow
37: such that more flattened (rigid-rotation in extreme cases) velocity profile emerges locally and the outer part of the profile tends to be super-Keplerian.
38: Angular momentum and mass transfer due to temporally generated MRI turbulence
39: in the initially unstable region is responsible for the transformation.
40: In the local super-Keplerian region,
41: migrations due to aerodynamic gas drag and
42: tidal interaction with disk gas are reversed.
43: The simulation setting corresponds to the regions near
44: the outer and inner edges of a global MRI dead zone in a disk.
45: Therefore, the outer edge of dead zone, as well as the inner edge,
46: would be a favorable site to accumulate dust particles to
47: form planetesimals and retain planetary embryos against type I migration.
48: \end{abstract}
49:
50: \keywords{accretion disks --- instabilities ---
51: MHD --- planetary systems: formation --- turbulence}
52:
53:
54: \section{Introduction}
55:
56: The ubiquity of extrasolar planets strongly suggests that planet formation is a common process associated with
57: star formation.
58: However, two serious barriers are recognized in
59: planet formation theory: meter-size and
60: type I migration barriers.
61: Terrestrial planets or icy cores for
62: gas giants that are embedded in protoplanetary
63: disks tend to lose orbital angular momentum and
64: migrate inward through tidal interaction with the disk gas
65: (``type I migration'').
66: Linear calculations \citep[e.g.,][]{tan02}
67: predict that the planets spiral into the host stars
68: on timescales $\la 10^5$ years for the minimum-mass solar nebula model,
69: which is much shorter than observationally inferred disk lifetime.
70: This is the type I migration barrier for survival of
71: planets with an Earth mass or more.
72:
73: The meter-size barrier is the barrier for formation of
74: planetesimals.
75: Since rotation speed of disk gas is slightly slower
76: than dust grains or planetesimals due to negative
77: radial pressure gradient due to global structure of the gas disk
78: and the motions of meter-sized bodies (boulders)
79: are marginally coupled
80: to gas motion through aerodynamical gas drag,
81: the meter-sized boulders suffer
82: the fastest inward orbital migration
83: and the timescale to spiral into the protostar is
84: $\sim$ a hundred orbits \citep{adachi76,wei77}.
85: This timescale would be too short compared to the sticking timescale
86: for the boulders to form planetesimals of 1--10 km radius or more,
87: the motions of which are decoupled from the gas motion, since
88: the meter-sized boulders are expected to
89: stick together only poorly \citep{benz00}.
90:
91: One way to bypass the meter-size barrier is formation of
92: clumps through self-gravitational instability of a dust layer,
93: which can occur on orbital periods \citep{saf69,gold73}.
94: However, local Kelvin-Helmholtz instability due to
95: difference in rotation velocities between the dust-rich layer
96: (Keplerian) and
97: an overlaying dust-poor layer (sub-Keplerian)
98: would prevent dust grains
99: from settling to the midplane to form the thin enough layer for
100: the self-gravitational instability
101: \citep{weiden93}.
102: Furthermore, if global turbulence exists in the disk,
103: it also stirs up the dust grains from the midplane.
104: Commonly observed strong H$\alpha$ emission from young stars
105: implies protoplanetary disks are viscously evolving
106: accretion disks with turbulent viscosity.
107: One of the most likely sources for the turbulence is
108: magnetorotational instability (MRI) in weakly ionized disk gas
109: \citep{bal91}.
110:
111: Although turbulence inhibits formation of the thin dust
112: layer near the midplane, dust grains could be trapped
113: into turbulent eddies.
114: It is expected that meter-sized boulders are concentrated
115: in anticyclonic eddies \citep[e.g.,][]{barge95,chav00,joha04,ina06}.
116: In particular, \citet{joha07} carried out a simulation of
117: evolution of self-gravitating boulders in an MRI turbulent disk
118: and found that 1000km-sized clumps are formed in eddies.
119: One of the biggest issues in this model is
120: lifetime of eddies.
121: To form clumps, the eddies must persist until boulders
122: are accumulated into the eddies and produce highly dust-rich regions.
123: Since rapid formation of highly dust-rich regions is required,
124: \citet{joha07} assumed meter-sized boulders that show
125: the most favorable behavior in the disks.
126:
127: Trapping of dust grains is also possible at a
128: locally high-pressure disk radius, because a positive radial pressure gradient
129: induces super-Keplerian gas flow in which dust grains suffer tail winds,
130: while a normal negative gradient induces head winds
131: \citep[e.g.,][]{naka86,kla05}.
132: \citet{rice04} demonstrated that meter-sized boulders
133: can be accumulated in high-density spiral arms of
134: self-gravitating disks.
135:
136: The pressure maximum also exists near the inner boundary of
137: MRI "dead" zone.
138: Innermost disk regions are thermally ionized, so that
139: MRI is active there.
140: In outer regions, X rays from host stars and cosmic rays
141: penetrate all the way to the midplane of the disk,
142: and the ionization degree may be high enough to activate MRI.
143: On the other hand, in the intermediate regions, only surface layers are
144: MRI active and the regions near the midplane may be inactive ("dead")
145: \citep{gam96,sano00}.
146: Assuming steady gas accretion through a disk, disk gas surface density
147: jumps up at the inner boundary of the dead zone
148: according to the change in effective viscosity due to turbulence,
149: so pressure maximum emerges there.
150:
151: The pressure maximum, in other words, the outer boundary of
152: a local super-Keplerian region
153: also halts inward type I migration \citep{tan02,mass06}.
154: If residual and/or secondary-generated dust grains maintain
155: the dead zone when accretion of planetary embryos proceeds,
156: the inner boundary of the dead zone is a favorable
157: cite to retain planetary embryos.
158: However, the inner boundary is generally located
159: well inside 1AU.
160: It cannot provide building blocks for terrestrial
161: planets in habitable zones and icy planets nor retain the cores
162: to form gas giants around solar-type stars,
163: although it would play an important role in
164: architecture of short-period extrasolar planets.
165:
166: \citet{kret07}, however, pointed out that
167: a local dead zone can appear near the ice line and
168: the ice line can be a spatial barrier for dust migration due to gas drag.
169: Because dust grains are the most efficient argents for charge recombination,
170: ionization rates and thickness of the active layers
171: rapidly decreases across the ice line due to condensation of
172: icy dust grains.
173: Correspondingly, in a range of disk accretion rates,
174: a local dead zone appears near the ice line and
175: at the outer boundary of the dead zone large amount of icy grains
176: can be accumulated and cores are retained to form gas giants.
177:
178: \citet{brau08} performed simulations of dust settling and
179: coagulation with radial migration due to gas drag
180: and the ice line effect.
181: They found that the dust to gas ratio increases and formation of
182: planetesimals may be efficient near the ice line.
183: \citet{idalin08a} showed that if the ionization rate is
184: order of magnitude larger than that predicted by
185: \citet{kret07} due to dust growth [\citet{kret07} assumed
186: that all the dust grains have $\mu$m sizes],
187: cores stop migration at the ice line and
188: they efficiently form gas giants without significant reduction
189: of type I migration speed.
190: Note that in order for the ice line effect to
191: actually work, disk accretion rate and dust population
192: must be within some ranges of parameters such that
193: the thickness of the active layer is comparable to
194: that of dead zone near the ice line, although the ranges are not
195: too restricted.
196:
197: We here show through MHD simulations
198: that the pressure maximum associated with
199: quasi-steady local super-Keplerian rotation may be
200: created in the MRI marginal regions, such as the
201: outer boundary of the global dead zone as well as its inner boundary,
202: without requirements of persistent turbulent eddies nor
203: the ice line effect.
204: The accumulation of dust grains and retention of planetary
205: embryos at the outer boundary of the dead zone have
206: a great importance for formation of terrestrial and jovian planets.
207: MRI is controlled by the vertical component of magnetic
208: field ($B_z$) penetrating the disks as well as by ionization degree
209: \citep[e.g.,][]{sano00}.
210: As shown in later sections,
211: non-uniform temporal MRI turbulence that occurs
212: in the marginally stable regions
213: transforms disk gas flow into quasi-steady flow that
214: has local rigid-rotation regions.
215: This flow pattern is sustained by non-uniform pressure gradient
216: produced by mass transfer associated with the temporal turbulence.
217: In the outer regions of the local rigid-rotation regions,
218: gas rotation is super-Keplerian.
219:
220: Since dead zones may shrink as dust grains grow
221: and surface areas for charge recombination decrease \citep{sano00},
222: such marginal state sweeps from outer disk regions to inner regions.
223: Furthermore, the density fluctuations due to MRI turbulence
224: induced by the enhanced ionization degree lead to disruptive
225: collisions of small planetesimals and the collisions
226: reproduce dust grains \citep{ida08}.
227: The grains lower the ionization degree so that
228: disk gas becomes marginally stable against MRI.
229: Then, grain growth starts again.
230: Thus, marginally MRI stable state could be maintained
231: in significant regions at $\la 10$AU until disk gas
232: is depleted to some degree \citep{ida08}.
233: \citet{oishi07} showed the random torques due to
234: the density fluctuations in the surface active layers
235: affect planetesimals in the dead zone near the disk midplane.
236: This effect may result in more continuous dust production
237: and help maintenance of the marginally stable state.
238: Thus, it is expected that broad regions at $\la 10$AU
239: may once experience such local super-Keplerian motions
240: and accumulate dust grains to form planetesimals
241: during disk evolution.
242:
243: In section \ref {numerical}, we describe the numerical method
244: and the initial conditions. We use a local shearing box with
245: non-uniform $B_z$.
246: In section 3.1, the results with a fiducial model are shown.
247: In section 3.2, the dependence of the results on initial settings
248: is presented.
249: Summarizing the numerical results, we show in section 3.3
250: that creation of the rigid rotation and super-Keplerian regions
251: is regulated by spatially averaged magnetic Reynolds number.
252: Section 4 is conclusion and discussion.
253:
254:
255: \section{Numerical Model}
256: \label{numerical}
257:
258: \subsection{Shearing box model}
259:
260: We carry out local two-dimensional MHD simulations
261: of protoplanetary disks in the shearing box model
262: \citep{wisdom88,haw95}.
263: The coordinates are centered at $r_{0}$ from a host star
264: and rotating with Keplerian angular velocity at $r_0$ ($\Omega_{0}$).
265: The radial, azimuthal and vertical directions are $x$, $y$ and $z$,
266: respectively.
267: Assuming uniformity in the $y$-direction,
268: we simulate 2D flow in the $x$-$z$ plane.
269: From the flow in the $x$-$z$ plane, the evolution of $v_y$
270: is calculated.
271: We will discuss coordinate sizes in the $x$-$z$ directions
272: ($L_x$ and $L_z$) later.
273: Periodic boundary conditions are applied for the $x$ and $z$ directions.
274: In the $x$ direction, Keplerian shear motion in the $y$-direction
275: is taken into account in the boundary condition.
276:
277:
278: \subsection{Basic equations}
279:
280: We adopt compressible resistive magnetohydrodynamic (MHD) equations:
281: \begin{eqnarray}
282: \frac{\partial \mathbf{v}}{\partial t}
283: +\left(\mathbf{v}\cdot\nabla\right)\mathbf{v}&=&
284: -\frac{1}{\rho}\nabla\left(P+\frac{\mathbf{B}^2}{8\pi}\right)
285: +\frac{1}{4\pi\rho}\left(\mathbf{B}\cdot\nabla\right)\mathbf{B}
286: -2\mathbf{\Omega}_{0}\times\mathbf{v}+3\Omega^2_{0}x\hat{\mathbf{x}}, \label{motion}\\
287: \frac{\partial \rho}{\partial t}
288: +\nabla\cdot\left(\rho\mathbf{v}\right)&=&0, \label{continuity}\\
289: \frac{\partial \mathbf{B}}{\partial t} &=&
290: \nabla\times\left[\left(\mathbf{v}\times\mathbf{B}\right)-
291: \eta\nabla\times\mathbf{B} \right],
292: \label{induction}\\
293: P&=&c^2_{s}\rho, \label{isothermal}
294: \end{eqnarray}
295: where $\mathbf{v}$, $P$, $\rho$ and $c_{s}$ are
296: velocity in the rotating frame, pressure, density and sound speed
297: of the gas, respectively.
298: The third and forth terms in r.h.s. of equation~(\ref{motion})
299: are the Coriolis force and tidal force (the difference between
300: the centrifugal force and gravitational force from the central star),
301: in which $\hat{\mathbf{x}}$ is a unit vector in the $x$-direction.
302: $\mathbf{B}$ and $\eta$ are magnetic field and resistivity.
303: For simplicity, we omit the vertical component of the gravity.
304: We include ohmic dissipation in equation~(\ref{induction}),
305: while we neglect ambipolar diffusion.
306: Its effect is weaker than the ohmic dissipation
307: in midplane of inner ($\la 100$AU) disk regions \citep{jin96},
308: although in the surface layer or disk inner edge,
309: dissipation due to ambipolar diffusion may be important \citep{chi07}.
310:
311: We scale length and time by disk scale height
312: ($H=c_{s}/\Omega_{0}$) and $\Omega_{0}^{-1}$, respectively.
313: Then, the independent parameters in the equations
314: are plasma beta ($\beta$)
315: and magnetic Reynolds number ($R_{\rm m}$)
316: defined respectively by
317: \begin{eqnarray}
318: \beta = \frac{2 c_s^2}{v_{\rm A}^2}, \label{eq:plasma_beta}\\
319: R_{\rm m} = \frac{v_{{\rm A}z}^2}{\eta \Omega_0} \label{eq:mag_Re},
320: \end{eqnarray}
321: where $v_{\rm A} = B/\sqrt{4 \pi \rho}$ is Alfven velocity
322: and $v_{{\rm A}z}$ is its $z$-component.
323: We assume spatially uniform and time-independent $c_s$ and $\eta$.
324:
325: We solve the resistive MHD equations
326: using the CIP-MOCCT method \citep{yab91,sto92} with
327: grid sizes of $0.01H$ (for dependence of our results on resolution,
328: see Appendix 1).
329: We usually integrate the evolution until $t \simeq 100 \Omega_0^{-1}$.
330: Although most of numerical studies on nonlinear stages of
331: MRI have assumed ideal MHD, we consider low ionization state
332: with finite resistivity.
333: Several numerical simulations \citep{fle00, haw96} showed
334: that the finite magnetic resistivity reduces growth rates of the MRI.
335:
336:
337: \subsection{Initial conditions}
338:
339: We assume that gas density is initially uniform ($\rho_0$),
340: so pressure is also uniform ($P_0$) due to the assumption of
341: isothermal gas (constant $c_s$).
342: We also assign the initial value of
343: plasma beta as $\beta = 400$ uniformly.
344: According to the large $\beta$, we set initial steady flow
345: as a uniform Keplerian shear flow, $v_{y}=-(3/2) x \Omega_0$.
346:
347: The remaining parameter is $R_{\rm m}$.
348: This value determines the MRI growth rate \citep{jin96, sam99}.
349: For $R_{\rm m} \la 1$, short wavelength modes are stabilized
350: by ohmic dissipation and the growth rate declines.
351: For $\Omega_0 \propto r^{-3/2}$,
352: the the most unstable wavelength is \citep{sam99}
353: \begin{equation}
354: \lambda_{\rm m.u.} \simeq 2 \pi \frac{\eta}{v_{{\rm A}z}}
355: \simeq \frac{2\sqrt{2}\pi}{\sqrt{\beta} R_{\rm m}}H
356: \simeq \frac{0.44}{R_{\rm m}}H.
357: \label{eq:lambda.m.u}
358: \end {equation}
359: Since MRI would not occur in the disk if
360: $H \la \lambda_{\rm m.u.}$, the stabilization
361: condition is $R_{\rm m} \la R_{\rm m,crit} \simeq 0.44$.
362:
363: We consider marginally stable state for MRI,
364: in which stable and unstable regions co-exist non-uniformly.
365: In this study, we assume constant $\eta$, and
366: the non-uniformity of $R_{\rm m}$ is set by that of
367: $v_{{\rm A}z}$ (equivalently, $B_z$).
368: Since non-uniform MRI is an essential point for
369: the emergence of the super-Keplerian flow,
370: similar results are obtained in the case of non-uniform $\eta$.
371: We will show the results with non-uniform $\eta$ in a next paper.
372:
373: In order to represent the non-uniformity of $B_z$ with uniform $\beta$,
374: we set initial $\mathbf{B}_0$ such that
375: \begin{equation}
376: \mathbf{B}_0=(0, B_{0}\sin\theta, B_{0}\cos\theta)
377: \label{eq:initial_B}
378: \end{equation}
379: with uniform $B_{0}$ and non-uniform $\theta$.
380: As shown in Figure \ref{fig:initial}, we set
381: $\theta = 0$ ($\cos \theta = 1$) in the middle region (unstable)
382: with radial size $L_{\rm u}$
383: and $\theta = 85$ degrees ($\cos \theta \simeq 0.087$)
384: in the side regions (stable) with individual radial size $L_{\rm s}$.
385: The box size in the horizontal direction is given by
386: $L_x = L_{\rm u} + 2 L_{\rm s}$.
387: Transition zones between $\theta = 0$ and
388: 85 degrees are set to avoid numerical instability.
389: We include non-zero azimuthal magnetic component $B_y$
390: because the plasma beta and therefore the magnetic pressure are
391: set to be spatially uniform to establish the initial equilibration.
392: The azimuthal component $B_y$ is calculated even in these two-dimensional
393: simulations on the $x$-$z$ plane, but the assumption of axisymmetry excludes
394: the occurrence of magnetic dynamo on the $x-y$ plane due to
395: reconnection of $B_y$ and the results here do not change
396: even if $B_y$ is set to be zero.
397: In our preliminary three-dimensional simulations,
398: we find that the MRI growth due to reconnection of $B_y$
399: in the side areas hardly affects the features
400: of azimuthally averaged fluid motion.
401: This is because the essential process
402: of the transformation from Keplerian flow to
403: quasi-steady non-uniform rotation flow
404: is temporal MRI growth and the stabilization of MRI
405: due to established rigid rotation
406: but not due to dynamo of magnetic field (see section 3).
407: The effects of three-dimensional flow
408: will be addressed in detail in the next paper.
409:
410: The above choice indeed situates both stable and unstable regions
411: in the simulation box.
412: Substituting equations~(\ref{eq:plasma_beta}) and (\ref{eq:initial_B})
413: into equation~(\ref{eq:mag_Re}), we obtain
414: \begin{equation}
415: R_{\rm m} = \frac{v_{\rm A}^2 \cos^2 \theta}{\eta \Omega_0}
416: = \frac{2 \beta^{-1} H^2 \Omega_0 \cos^2 \theta}{\eta}
417: = 2.5 \left( \frac{\eta}{0.002 H^2 \Omega_0}\right)^{-1} \cos^2 \theta.
418: \label{eq:mag_Re_fudcial}
419: \end{equation}
420: With a fiducial value $\eta=0.002H^{2}\Omega_0$, MRI is
421: initially activated ($R_{\rm m} > R_{\rm m,crit}$)
422: in the middle region with
423: $\theta < \theta_{\rm crit} \simeq 65$ degrees.
424: The exact dispersion relation
425: (Appendix 2) show that
426: wavelengths of all the growing modes exceed the vertical box size $L_z$
427: for $\theta < \theta_{\rm crit} \simeq 79$ degrees.
428: The vertical size $L_z$ is set to be $\simeq 0.5 H$
429: that is comparable to
430: the most unstable wavelength $\lambda_{\rm m.u.}$
431: (equation~[\ref{eq:lambda.m.u}]).
432:
433: We have carried out 79 runs in total with various
434: $\eta$, $L_{\rm u}$, and $L_{\rm s}$.
435: The detailed simulation parameters are given in Table \ref{tab:1}.
436: In some cases, we adopted larger values of $\eta$,
437: in which $\lambda_{\rm m.u.}$ is larger
438: (equation~[\ref{eq:lambda.m.u}]).
439: Then, we adopt $L_z$ as large as $\lambda_{\rm m.u.}$ with
440: $R_{\rm m} \sim 1$.
441: In all runs,
442: random fluctuations are initially given in the velocity field
443: with a maximum amplitude of $\left|\delta v_{x}\right|=0.001c_{s}$.
444:
445:
446: \section{Results}
447: \subsection{The Fiducial Model}\label{modelA}
448:
449: We first present the detailed results from a fiducial model with
450: $\eta=0.002H^{2}\Omega_0$, $L_{\rm u} = 1.43H$,
451: $L_{\rm s} = 4.05H$, and $L_z = 0.5H$.
452: Compared with the other models, this model initially has the
453: largest stable regions.
454: If the whole region has uniform and strong enough
455: magnetic field, MRI turbulence is self-sustaining.
456: However, evolution of turbulence is quite different
457: in the case of non-uniform magnetic field that we consider here.
458: MRI turbulence appears only tentatively, but velocity field
459: is transformed from Keplerian shear flow to another quasi-steady flow.
460:
461: Time evolution of the magnetic field on the $x$-$z$ plane
462: is shown in Figure~\ref{fiducial}a.
463: In the panel of $t\Omega_{0}=21$, the field lines
464: start to be stretched only in the initially unstable region.
465: At $t\Omega_{0}=27$, MRI turbulence develops and
466: the turbulence is transported into the initially stable regions.
467: Accordingly, the magnetic perturbations become weaker and
468: eventually vanish before they reach the boundary
469: of the computing box
470: (before they fully fill the stable region).
471:
472: Figure~\ref{fiducial}b shows evolution of $B_z$ averaged over $z$.
473: Turbulent diffusion smooths out $B_z$, however,
474: its level does not go below the critical value for stabilization of MRI
475: ($\sim 0.2 B_0$) in the initially unstable region
476: (the middle region) even at $t\Omega_{0}=70$.
477: Nevertheless, MRI turbulence ceases after $t\Omega_{0} \ga 40$.
478: This is because in the middle region,
479: rigid rotation is realized
480: as a result of the angular momentum transfer
481: during the MRI turbulence there (Fig.~\ref{fiducial}c) and
482: the free-energy source for MRI (differential rotation) is lost.
483: In the other areas, $B_z$ is smaller than the critical value
484: and MRI does not grow.
485: The stabilization due to rigid rotation is discussed
486: with linear analysis in Appendix 2.
487:
488: As shown in Figure~\ref{fiducial}a, the region of the rigid rotation
489: expands as propagation of turbulence.
490: This local rigid rotation is self-sustaining, at least,
491: until the remaining $B_z$ is diffused out by ohmic dissipation
492: on timescale $\sim L_u^2/\eta \sim 10^3 \Omega_0$
493: (see discussion in section 4). As long as certain level of $B_z$ is maintained,
494: deviation from the rigid rotation excites MRI turbulence
495: and this transfers angular momentum to bring the velocity profile back to the rigid rotation state.
496:
497: Figure \ref{fiducial}d is the vertically averaged pressure distribution.
498: The MRI turbulence radially transports mass, associated with
499: the angular momentum transfer.
500: Since in the unstable region, mass transfer is efficient
501: and gas density is proportional to pressure in our isothermal model,
502: pressure is decreased in the unstable region
503: while it is increased in the stable regions
504: adjacent to the unstable region.
505: The resultant pressure gradients near the boundaries between
506: the stable and unstable regions equilibrates with
507: the Coriolis and tidal forces raised by
508: a deviation from Keplerian rotation.
509: In Figure~\ref{force}, we plot radial components
510: of individual terms in the r.h.s. of
511: the equation of motion (equation~[\ref{motion}])
512: near the right boundary between the stable and unstable regions
513: ($x/H=0.7$, $z/H=0.25$).
514: At $t \Omega_0 = 20$--40, the amplitude of magnetic pressure
515: and tension are increased by the development of MRI turbulence
516: and gas pressure gradient is also raised
517: according to mass transfer associated with the turbulence.
518: After the MRI turbulence ceases, the pressure gradient
519: almost equilibrates with the Coriolis and tidal forces.
520: Thus, the rigid rotation with pressure variation caused by the
521: MRI turbulence is quasi-steady.
522:
523: As stated in section 1, dust grains can be trapped in
524: outer edge of local super-Keplerian regions, because
525: dust grains suffer tail wind and migration outward
526: in the super-Keplerian regions (see sections 1 and 4)
527: while they suffer head wind in the other sub-Keplerian regions.
528: In the result of Figure~\ref{fiducial}, the super-Keplerian region
529: exists at $0.0<x/H \la 2.0$ in the quasi-steady state.
530:
531:
532: \subsection{Dependence on widths of stable and unstable regions}
533:
534: In the limit of $L_{\rm s} \rightarrow 0$ or $L_{\rm u} \rightarrow \infty$,
535: the system would have uniform turbulence and
536: would not acquire the local rigid rotation.
537: To derive the condition for establishment of the local rigid rotation,
538: we perform runs with different values of $L_{\rm s}$ and $L_{\rm u}$.
539:
540: In the fiducial case, $L_{\rm u}=1.4H$ and $L_{\rm s}=4.0H$.
541: We first present the results of a series of runs with
542: various values of $L_{\rm s}$ and the fixed $L_{\rm u}$.
543: In model-s11, model-s055, and model-s005,
544: $L_{\rm s} = 1.1H$.
545: $0.55H,$ and $0.05H$, respectively.
546: The other parameters are the same as those in the fiducial model.
547: Figure~\ref{s1} describes the evolution of the magnetic field in model-s11.
548: Since the stable region is narrower than the fiducial case,
549: the magnetic perturbations that arise in the unstable region
550: propagate to the boundary of the computational box
551: before they are dissipated.
552: The whole region including the initially stable region becomes temporarily turbulent.
553: However, the perturbations that reach the boundary are
554: weakened and do not have enough momentum to go into the unstable region again.
555: The turbulence ceases after $t\Omega_{0} \ga 40$ and
556: the quasi-steady rigid rotation is obtained as the result, although
557: the pressure contrast is smaller than that in the fiducial model
558: (the minimum pressure is $\simeq 0.65P_0$, while it is
559: $\simeq 0.33 P_0$ in the fiducial model).
560:
561: The evolution of the MRI in model-s055 is shown in Figure~\ref{s05}.
562: The narrower stable region allows the magnetic perturbations to
563: come back to the unstable region because of the periodic boundary condition.
564: The turbulence lives longer, but it eventually vanishes
565: by $t\Omega_{0} \sim 70$.
566: The quasi-steady velocity profile which is more flattened than the Keplerian is established, although it is not as distinctive as the rigid rotation
567: in the fiducial model.
568:
569: Figure \ref{s005} shows the result of model-s005.
570: Because of the small stable regions, turbulence expanded to the entire
571: region does not cease and uniform turbulence is maintained.
572: Efficient angular momentum transfer in the entire region prevents
573: velocity field from having a quasi-steady super-Keplerian region.
574:
575: So far, we have changed values of $L_{\rm s}$ while $L_{\rm u}$ is constant.
576: We also performed a run (model-u34) with $L_{\rm u}=3.4H$ that is
577: enlarged from the fiducial model.
578: The other parameters are the same as those in the fiducial model.
579: The results are shown in Figure~\ref{u34}.
580: In the enlarged unstable region, the magnetic field is stretched enough for reconnection.
581: After the reconnection, long-lived magnetic perturbations are sustained
582: around the boundary between unstable and stable regions
583: (the panel at $t\Omega_{0}=58.5$).
584: The magnetic perturbations reach the computational boundary, because
585: the conversion from magnetic energy to kinetic energy
586: during the magnetic reconnection adds horizontal motion to the fluid element.
587: However, the magnetic perturbations are not strong enough
588: to pass through the stable region
589: and come back to the unstable region.
590: As a result, the evolution is similar to model-s11, except that the
591: rigid rotation is established in the regions
592: near the boundary between unstable and stable regions.
593: The panels at $t\Omega_{0}=61.5$ and $t\Omega_{0}=99.0$ show that
594: this flow pattern is also quasi-steady.
595:
596: These results suggest that dissipation of magnetic perturbations
597: during the passage through stable regions plays an
598: essential role in creation of rigid rotation regions.
599: We performed additional series of runs with
600: a larger value of the resistivity, $\eta=0.0028\Omega_{0}H^2$.
601: The result of model-$\eta0$ with $L_{\rm u}=1.4H$ and $L_{\rm s}=4.0H$ (the same as
602: the fiducial model) is shown in Figure~\ref {eta}.
603: Because of the faster dissipation, the quasi-steady rigid rotation is
604: established more early, although the weaker angular momentum transfer leads to
605: smaller pressure contrast in the quasi-steady state.
606: Even if we reduce $L_{\rm s}$ to $\sim 0.5H$,
607: the magnetic perturbations do not arrive at the boundary.
608: We also performed runs with further larger values of
609: the resistivity, $\eta/(\Omega_{0}H^2) = 0.0012, 0.0036, 0.0044, 0.0052$
610: and 0.0060 (Table 1) to confirm that the quasi-steady rigid rotation
611: is always established for these values of $\eta$.
612: We discuss the dependence on the resistivity in the next subsection.
613:
614:
615:
616: \subsection{Condition for the rigid rotation}\label{type}
617:
618: As we have shown, establishment of the quasi-steady rigid rotation depends
619: on $L_{\rm u}$, $L_{\rm s}$, and $\eta$.
620: Through the results with various $L_{\rm u}$, $L_{\rm s}$, and $\eta$
621: as listed in Table 1, we found that the results are
622: summarized by a single parameter,
623: a spatially averaged magnetic Reynolds number, defined by \citep{sam99, sai04}
624: \begin{equation}
625: R_{\rm m,ave}=\frac{v^2_{{\rm A}z,{\rm ave}}}{\eta\Omega_{0}},
626: \end{equation}
627: where $v_{{\rm A}z,{\rm ave}}$ is evaluated by
628: a spatially averaged vertical magnetic field ($B_{z,{\rm ave}}$)
629: at the initial stage.
630: In our simulation setting, $B_{z,{\rm ave}} \simeq
631: B_0(L_{\rm u} + 2 \cos 85^{\circ} L_{\rm s})/(L_{\rm u} + 2 L_{\rm s})$.
632:
633: In many cases, the initial distributions of vertical magnetic field
634: are spatially smoothed out by turbulent diffusion after $\sim 10$ orbits.
635: If the remaining magnetic field ($\sim B_{z,{\rm ave}}$)
636: is still large enough to globally cause MRI turbulence,
637: the transition to the quasi-steady state
638: does not occur, as seen in the results of model-s005.
639: Linear theory shows that the critical magnetic Reynolds number
640: for occurrence of MRI is $R_{\rm m} \sim 1$ \citep{sam99, sai04}.
641: Thus, $R_{\rm m,ave}$ regulates the establishment of the quasi-steady
642: rigid rotation in our simulations.
643:
644: We summarize the results of 91 runs with different parameters
645: in Figure~\ref{class} and found that evolution of the magnetic
646: field is indeed classified into four types by the values of
647: $R_{\rm m,ave}$ as follows:
648: \begin{description}
649: \item[Type A ($R_{\rm m,ave} \la 0.1$):]
650: Local MRI turbulence generated in the initially unstable region
651: propagates both inward and outward, but
652: the magnetic perturbations are dissipated before they reach
653: the boundary of the simulation box.
654: After a few tens of orbits, the turbulence
655: vanishes in the entire region and the quasi-steady flow
656: is established.
657: In the initially unstable region, rigid rotation flow is
658: resulted in by angular momentum transfer due to
659: the MRI turbulence.
660: This class is represented by filled circles in Figure~\ref{class}
661: and it includes the fiducial model and model-$\eta0$.
662:
663: \item[Type B ($0.1\la R_{\rm m,ave} \la 0.5$):]
664: The magnetic perturbations reach the boundary, but they do not
665: intrude back into the original unstable region.
666: The quasi-steady rigid-rotation region appears as in Type A.
667: This class is represented by triangles in Figure~\ref{class}
668: and it includes model-s11 and model-u34.
669:
670: \item[Type C ($0.5 \la R_{\rm m,ave} \la 1.0$):]
671: The magnetic perturbations intrude the unstable region
672: after the passage through the stable regions.
673: However, the diffused $B_{z}$ is not large enough to globally maintain
674: turbulence and the quasi-steady rigid rotation region is still formed,
675: although their locations are not necessarily the same as in Type A and B.
676: This class is represented by daggers in Figure~\ref{class}
677: and it includes model-s055.
678:
679: \item[Type D ($1.0 \la R_{\rm m,ave}$):]
680: Even after the turbulent diffusion, $B_{z}$ is able to maintain
681: the turbulence in the entire region.
682: Because of the uniform turbulent state,
683: the quasi-steady rigid rotation state is not generated.
684: This class is represented by crosses in Figure~\ref{class}
685: and it includes model-s005.
686:
687: \end{description}
688:
689:
690: \section{Conclusion and Discussion}\label{discuss}
691:
692: We have investigated evolution of
693: patchy magneto-rotational instability (MRI) due to radially
694: non-uniform magnetic field and found that, under some conditions,
695: the original Keplerian shear flow is transformed into
696: quasi-steady profile involving a local
697: rigid-rotation regions.
698: The outer parts of the rigid-rotation regions
699: are generally super-Keplerian.
700: Such a situation would arise in the outer boundary of
701: MRI dead zone as well as the inner boundary in a protoplanetary disk,
702: as discussed in section 1.
703:
704: Assuming uniformity in the azimuthal direction of disks,
705: we have carried out two-dimensional resistive MHD simulations
706: in a shearing box model with periodic boundary conditions.
707: We set up both stable and unstable regions in the box,
708: changing direction of the vertical seed magnetic field ($B_z$) non-uniformly.
709: In the initially unstable region,
710: MRI turbulence is generated locally and magnetic perturbations
711: propagate both radially inward and outward by the turbulent diffusion.
712: If the unstable region is sufficiently large compared with
713: the stable region, the turbulence eventually covers the entire
714: region and the initial non-uniformity vanishes.
715: However, if the stable region is relatively large, diffused
716: magnetic perturbations no more maintain MRI turbulence.
717: After the turbulence ceases,
718: the initial flow of uniform Keplerian shear is transformed
719: into a different quasi-steady state.
720: In the quasi-steady state, rigid-rotation is established locally.
721: The deviation from Keplerian shear motion is supported
722: by pressure gradient that has been produced also by mass transport
723: associated with the tentative turbulence.
724: Through simulations with various initial conditions, we found
725: that the quasi-steady rigid rotation is established
726: if the spatially averaged
727: magnetic Reynolds number satisfies $R_{\rm m,ave} \la 1$
728: in the initial state.
729:
730: Because the center of the local rigid rotation is often
731: Keplerian, super-Keplerian flow
732: appears in the outer parts of the rigid rotation region.
733: As explained in section 1, dust grains and planetary embryos
734: can be trapped in the boundary between regions of
735: sub- and super-Keplerian motion
736: through radial migration induced by aerodynamic gas drag
737: and type I migration.
738: The boundary is coincident with pressure maxima in the quasi-steady state.
739:
740: The effect of global pressure gradient
741: is included by shifting the initial gas velocity
742: from the pure Keplerian speed to slight sub-Kepler.
743: The sift, which is the velocity difference between the disk gas
744: and the dust grains, is
745: \citep[e.g.,][]{adachi76,wei77}
746: \begin{equation}
747: \Delta v_y = \frac{c_s^2}{2 v_{\rm K}^2} \frac{d \ln P}{d \ln r} v_{\rm K}
748: \simeq - 5 \times 10^{-2} \left(\frac{r}{\rm 1AU} \right)^{3/4} c_s
749: \end {equation}
750: where $d \ln P/d \ln r$ is the global pressure gradient and
751: the temperature distribution in the limit of optically thin disks
752: around solar-luminosity stars, $T = 280 (r/1{\rm AU})^{-1/2}$, is assumed.
753: Since maximum values of $\Delta v_{y}$ in the super-Keplerian regions
754: are $\ga 0.4 c_s$ in our results,
755: the MRI effects can easily surpass the global pressure gradient effects
756: and super-Keplerian regions can emerge even when the initial down-shift
757: in the gas rotation velocity is present.
758:
759: If the super-Keplerian region is sustained long enough for
760: dust grains to accumulate, planetesimals can be formed through
761: self-gravitational instability \citep{you02,joha06}.
762: In the case of $R_{\rm m,ave} \la 1$ in which MRI turbulence
763: ceases in the entire region after a few tens of orbits,
764: we found that $B_z$ is still large enough in the rigid-rotation region.
765: In that region, MRI is suppressed by disappearance of shear motion
766: but not by dissipation of $B_z$ (in the regions other than
767: the rigid-rotation region, diffused-out $B_z$ is smaller than
768: the value for MRI to occur).
769: If the rigid rotation tries to go back to the original
770: Keplerian shear motion, MRI turbulence again occurs
771: and it transfers angular momentum to recover the rigid rotation.
772: Thus, the rigid rotation and hence the associated super-Keplerian
773: rotation are self-sustaining.
774: When the remaining $B_z$ is diffused out by ohmic dissipation
775: on timescale $\sim L_u^2/\eta \sim 10^3 \Omega_0^{-1}$, such
776: stabilization mechanism is no more effective.
777: Then, the rigid rotation can go back to the original
778: Keplerian shear motion by the residual uniform viscosity.
779: However, since MRI no more occurs, the residual viscosity
780: would be very small.
781: Thus, it is expected that the super-Keplerian regions
782: would survive long enough for accumulation of dust grains and
783: formation of planetesimals.
784: We also did a calculation starting from the end result of the
785: super-Keplerian rotation state, artificially modifying $B_z$ to
786: uniform distribution.
787: However, we do not see any relaxation of the velocity field
788: back to Keplerian rotation within the timescales of 40 orbits.
789:
790: In the next paper, we will demonstrate the accumulation of dust grains and discuss the effect of the azimuthal magnetic field especially in stable region by three-dimensional simulation.
791: We will also show the results of non-uniform resistivity case
792: with constant $B_{z}$, which may be more likely
793: to occur at the outer edge of a dead zone.
794: We find similar appearance of super-Keplerian regions,
795: because intrinsic physics to
796: transform initial Keplerian flow to
797: quasi-steady non-uniform rotation flow
798: is temporal generation of MRI and the stabilization of the MRI
799: due to the established
800: rigid rotation but not due to dynamo of magnetic field (see section 3).
801:
802: The appearance of the super-Keplerian region
803: also halts inward type I migration of planetary
804: embryos \citep{tan02,mass06}.
805: Since the process we found also works at the outer
806: boundary of a dead zone and the outer boundary
807: migrate from $\sim 10$AU to the proximity of the central star,
808: this process may also help the formation of cores massive enough
809: to onset runaway gas accretion and retain terrestrial planets
810: against type I migration.
811: This may play an important role in frequency of
812: extrasolar gas giants and habitable planets.
813: We will address this issue with sequential planet formation
814: model \citep{idalin04,idalin08a,idalin08b} in a future paper.
815:
816:
817:
818: %Acknowlegemets
819: We thank detailed helpful comments by an anonymous referee.
820:
821: \appendix
822: \section*{Appendix 1. Resolution test}
823:
824: We have investigated the effects of numerical resolution on our results
825: using our fiducial model. Four cases are studied. The distribution of
826: angular velocity and the average Maxwell stress are shown in Figure
827: ~\ref{resolution}. The most important issue in our results is the emergence
828: of the quasi-steady state in which sub- and super-Keplerian areas exist.
829: Figure~\ref{resolution}a shows this to be seen even in the worst resolution
830: case. In addition, the Maxwell stress in Figure~\ref{resolution}b appears to
831: be converging at resolutions above $dx=0.01H$ while this quantity tends to
832: increase with resolution.
833:
834: \citet{fro07} showed the decrease of Maxwell stress with increasing resolution
835: and stated that the MRI turbulence activity is notoriously ill-behaved in high
836: resolution calculation. Our resolution test, however, shows the convergence
837: with increasing resolution. This could be because the ohmic dissipation
838: (resistivity) is kept constant in our resistive MHD simulations while the
839: simulation of \citet{fro07} included only the numerical resistivity and it
840: deacreases with increasing resolution. However, more detailed study is needed
841: to clarify the difference in convergence between our simulation and the
842: simulation of \citet{fro07}, which is left to our future study.
843:
844: This fact will be
845: important when we investigate the motion of dust particles. The degree of
846: particle concentration may be depend on the turbulence activity. While the
847: details are left for a future study, one may reasonably expect the
848: possibility of dust concentration at the outer-edge of super-Keplerian area.
849:
850: Figure~\ref{resolution}b also shows the eligibility of the integration time.
851: The saturation level varies only slightly in $t\Omega_{0}>60.0$
852: in the two high resolution cases. The quasi-steady state has already been
853: created by this time. These indicate that our choice the integration time
854: $t\Omega<100.0$
855: is validated.
856:
857:
858: \section*{Appendix 2. Dispersion relations}
859:
860: Our simulations show that the MRI stabilization in strong magnetic field
861: with nearly rigid rotation, which is consistent with the linear analyses.
862: The linear analysis using ideal MHD equations
863: in \citet{bal91} gave the critical wavelength (equation (2.14b)
864: in the paper, with neglect of the Brunt-V\"{a}is\"{a}l\"{a} frequency):
865: \begin{equation}
866: \left|\lambda_{z,{\rm crit}}\right| = v_{{\rm A}z}
867: \left|\frac{d\Omega^{2}}{d\ln r}\right|^{-1/2}
868: =\frac{v_{{\rm A}z}}{\Omega} (2\left|q\right|)^{-1/2},
869: \label{linear-ideal}
870: \end{equation}
871: where $q$ is defined as $\Omega(r) \propto r^{-q}$.
872: The perturbations with wavelength shorter than $\lambda_{z,{\rm crit}}$ are stable.
873: When the rotation becomes rigid rotation ($q \rightarrow 0$),
874: $\lambda_{z,{\rm crit}}$ becomes large.
875: If $\lambda_{z,{\rm crit}}$ is larger than the scale height of a disk,
876: the system is stable, irrespective of magnetic field strength.
877:
878: With the effect of ohmic dissipation,
879: the dispersion relation was obtained by
880: \citet{jin96} and \citet{sam99} as
881: \begin{equation}
882: \sigma^{4}+2\xi\sigma^{3}
883: +\left(2q_{z}^2+\xi^{2}+\frac{\kappa^{2}}{\Omega^2}\right)\sigma^{2}
884: +2\xi\left(q_{z}^{2}+\frac{\kappa^{2}}{\Omega^{2}}\right)\sigma
885: +\left(-4+\frac{\kappa^{2}}{\Omega^{2}}\right)q_{z}^{2}
886: +q_{z}^{4}+\frac{\kappa^{2}}{\Omega^{2}}\xi^{2}=0,
887: \label{linear-eta}
888: \end{equation}
889: where $\sigma$ is a growth rate in units of orbital frequency,
890: $q_{z}=k_{z}v{{\rm A}z}/\Omega$, $\xi=k_{z}^{2}\eta/\Omega$, and
891: $\kappa$ is an epicyclic frequency defined by
892: \begin{equation}
893: \kappa^{2}=\frac{2\Omega}{r}\frac{d\left(r^{2}\Omega\right)}{dr}
894: =\left(2-q\right)2\Omega^{2}.
895: \end{equation}
896: This dispersion relation is derived with
897: the assumption of uniform density and negligible
898: Brunt-V\"{a}is\"{a}l\"{a} frequency.
899: Since the density is almost uniform in the rigid rotation in our simulation,
900: we apply this dispersion relation to calculate the predicted $\sigma$
901: with the quantities obtained by our simulation.
902: We plot the temporally and vertically averaged $q$ ($=-dv_{x}/dx$),
903: Maxwell stress ($-\langle B_{x}B_{y} \rangle/4\pi P_{0}$)
904: and the evaluated growth rate $\sigma$ in Figure~\ref{growth}.
905: The growth rate is very small in the middle region of nearly rigid rotation
906: ($q \ll 1$), although Maxwell stress is not small there.
907: Thus, we conclude that MRI is suppressed by
908: established nearly rigid rotation, but not by dissipation of magnetic field.
909:
910: %\section{Appendix material}
911:
912: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
913: \begin{thebibliography}{}
914: %
915: \bibitem[Adachi(1976)]{adachi76}
916: Adachi, I., Hayashi, C., \& Nakazawa, K. 1976, Prog. Theor. Phys, 56, 1756
917: \bibitem[Balbus \& Hawley(1991)]{bal91}
918: Balbus, S. A., \& Hawley, J. F. 1991, \apjs, 376, 214
919: \bibitem[Barge \& Sommeria(1995)]{barge95}
920: Barge, P., \& Sommeria, J. 1995, \aap, 295, L1
921: \bibitem[Benz(2000)]{benz00}
922: Benz, W. 2000, \ssr, 92, 279
923: \bibitem[Brauer et al.(2008)]{brau08}
924: Brauer, F., Henning, Th., \& Dullemond, C. P. 2008, \aap accepted
925: \bibitem[Chavanis(2000)]{chav00}
926: Chavanis, P. H. 2000, \aap, 356, 1089
927: \bibitem[Chiang \& Murray-Clay(2007)]{chi07}
928: Chiang, E. I., \& Murray-Cley, R. A. 2007, Nature Phys., 3, 604
929: \bibitem[Fleming \& Stone(2000)]{fle00}
930: Fleming, T. P., Stone, J. M., \& Hawley, J. F. 2000, \apjs, 530, 464
931: \bibitem[Fromang \& Papaloizou(2007)]{fro07}
932: Fromang, S., \& Papaloizou, J. 2007, \aap, 476, 1113
933: \bibitem[Fromang \& Papaloizou(2007)]
934: Fromang, S., \& Papaloizou, J. 2007, \aap, 476, 1123
935: \bibitem[Gammie(1996)]{gam96}
936: Gammie, C. F. 1996, \apjs, 457, 355
937: \bibitem[Goldreich \& Ward(1973)]{gold73}
938: Goldreich, P., \& Ward, W. R. 1973, \apjs, 183, 1051
939: \bibitem[Hawley et al.(1995)]{haw95}
940: Hawley, J. F., Gammie, C. F., \& Balbus, S. A. 1995, \apj, 440, 742
941: \bibitem[Hawley et al.(1996)]{haw96}
942: Hawley, J. F., Gammie, C. F., \& Balbus, S. A. 1996, \apjs, 464, 690
943: \bibitem[Ida et al.(2008)]{ida08}
944: Ida, S., Guillot, T., \& Morbidelli, A. 2008, \apj, in press
945: \bibitem[Ida \& Lin(2004)]{idalin04}
946: Ida, S., \& Lin, D. N. C. 2004, \apj, 604, 388
947: \bibitem[Ida \& Lin(2008a)]{idalin08a}
948: ------. 2008, \apj, 673, 487
949: \bibitem[Ida \& Lin(2008b)]{idalin08b}
950: ------. 2008, \apj, in press
951: \bibitem[Inaba \& Barge(2006)]{ina06}
952: Inaba, S., \& Barge, P. 2006, \apjs, 649,415
953: \bibitem[Jin(1996)]{jin96}
954: Jin, L. 1996, \apjs, 457, 798
955: \bibitem[Johansen et al.(2004)]{joha04}
956: Johansen, A., Anderson, A. C., \& Brandenburg, A. 2004, \aap, 417, 361
957: \bibitem[Johansen et al.(2006)]{joha06}
958: Johansen, A., Klahr, H., \& Henning, Th. 2006, \apjs, 636, 1121
959: \bibitem[Johansen et al.(2007)]{joha07}
960: Johansen, A., Oishi, J. S., Mac Low, M,-M., Klahr, H., Henning, Th., \& Youdin, A. 2007, \nat, 448, 1022
961: \bibitem[Klahr \& Lin(2005)]{kla05}
962: Klahr, H., \& Lin, D. N. C. 2005, \apjs, 632, 1113
963: \bibitem[Kretke \& Lin(2007)]{kret07}
964: Kretke, K. A., \& Lin, D. N. C. 2007, \apjl, 664, L55
965: \bibitem[Masset et al.(2006)]{mass06}
966: Masset, F. S., D'Angelo, G., \& Kley, W. 2006, \apjs, 652, 730
967: \bibitem[Nakagawa et al.(1986)]{naka86}
968: Nakagawa, Y., Sekiya, M., \& Hayashi, C. 1986, Icarus, 67, 375
969: \bibitem[Oishi et al.(2007)]{oishi07}
970: Oishi, J. S., Mac Low, M-M., \& Menou, K. 2007, \apjs, 670, 805
971: \bibitem[Rice et al.(2004)]{rice04}
972: Rice, W. K. M., Lodato, G., Pringle, J. E., Armitage, P. J., \& Bonnell, I. A.
973: 2004, \mnras, 355, 543
974: \bibitem[Safronov(1969)]{saf69}
975: Safronov, V. S. 1969, Evolution of the Protoplanetary Cloud
976: and the Planets, NASA Tech. Transl. F-677
977: \bibitem[Sano \& Miyama(1999)]{sam99}
978: Sano, T., \& Miyama,S. M. 1999, \apjs, 515, 776
979: \bibitem[Sano et al.(2000)]{sano00}
980: Sano, T., Miyama, S. M., Umebayashi, T., \& Nakano, T. 2000, \apjs, 543, 486
981: \bibitem[Sano \& Inutsuka(2004)]{sai04}
982: Sano, T., Inutsuka, S. M., Turner, N. J. \& Stone, J. M. 2004, \apjs, 605, 321
983: \bibitem[Sekiya(1998)]{sekiya98}
984: Sekiya, M. 1998, Icarus, 133, 298
985: \bibitem[Stone \& Norman(1992)]{sto92}
986: Stone, J. M., \& Norman, M. L. 1992a, \apjs, 80, 753, 1992b, \apjs, 80, 791
987: \bibitem[Tanaka et al.(2002)]{tan02}
988: Tanaka, H., Takeuchi, T., \& Ward, W. 2002, \apjs, 565, 1257
989: \bibitem[Yabe \& Aoki(1991)]{yab91}
990: Yabe, T., \& Aoki, T. 1991, Comput. Phys. Comm., 66, 219
991: \bibitem[Youdin \& Shu(2002)]{you02}
992: Youdin, A. N., \& Shu, F. H. 2002, \apjs, 580, 494
993: \bibitem[Weidenschilling(1977)]{wei77}
994: Weidenschilling, S. J. 1977, \mnras, 180, 57
995: \bibitem[Weidenschilling \& Cuzzi(1993)]{weiden93}
996: Weidenschilling, S. J., \& Cuzzi, J. N. 1993, in Protostars and
997: Planets III, 1031-1060
998: \bibitem[Wisdom \& Tremaine(1988)]{wisdom88}
999: Wisdom, J., \& Tremaine, S. 1988, \aj, 95, 925
1000: %
1001: \end{thebibliography}
1002:
1003: \clearpage
1004:
1005: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1006: \begin{figure}[tbp]
1007: \begin{center}
1008: \includegraphics[width=100mm]{f1.eps}
1009: \end{center}
1010: \caption{The configuration of the initial magnetic field.
1011: $\theta$ represents the angle from the vertical direction ($z$)
1012: to the azimuthal one ($y$).
1013: It radially varies as shown in the right panel.
1014: With constant $B_0$, the initial magnetic field
1015: is given by $\mathbf{B}=(0, B_{0}\sin\theta, B_{0}\cos\theta)$.
1016: For $\eta = 0.002 H^2 \Omega_0$,
1017: linear calculations suggest that MRI occurs for
1018: $\theta < \theta_{\rm crit} \simeq 79$ degrees.}
1019: \label{fig:initial}
1020: \end{figure}
1021: %%
1022: %%
1023: \clearpage
1024: %%
1025: %%
1026: \begin{figure}[tbp]
1027: \begin{center}
1028: \includegraphics[width=100mm]{f2.eps}
1029: \end{center}
1030: \caption{Results of the fiducial model.
1031: (a) Time evolution of the magnetic field (black lines) and
1032: angular velocity $v_y$ (contours) on the $x$-$z$ plane.
1033: The maximum (brightest) or minimum (darkest) of the tone
1034: on the contours are set at the unstable and stable region boundaries.
1035: Time evolution of vertically averaged values of
1036: (b) the vertical magnetic component $B_z$,
1037: (c) angular velocity $v_y$,
1038: and (d) pressure $P$, as functions of $x$.
1039: In panels (b), (c) and (d),
1040: thin solid, dotted, dashed and bold lines express
1041: the snapshots at $t\Omega_{0} = 0.0, 27.0, 40.0$ and $70.0$,
1042: respectively.
1043: The unstable region is initially set between the two
1044: vertical dotted-lines.
1045: }
1046: \label{fiducial}
1047: \end{figure}
1048: %
1049: \clearpage
1050: %
1051: \begin{figure}[tbp]
1052: \begin{center}
1053: \includegraphics[width=80mm]{f3.eps}
1054: \end{center}
1055: \caption{Time variation of forces exerted on fluid at $x/H=0.7$ and $z/H=0.25$,
1056: at which super-Keplerian flow is established.
1057: The dotted, dash-dotted, thin dashed and dashed lines express gas radial components of
1058: pressure gradient,
1059: magnetic pressure gradient,
1060: magnetic tension, sum of gravity and Coriolis forces
1061: in the r.h.s. of the equation of motion (equation~[\ref{motion}]), respectively.
1062: The thick solid line is total force.
1063: The unit of the forces is $H \Omega_0^2$.
1064: The pressure gradient and the gravity/Coriolis forces
1065: are dominated and approximately equilibrate with each other.
1066: }
1067: \label{force}
1068: \end{figure}
1069: %
1070: %\clearpage
1071: %
1072: \begin{figure}[tbp]
1073: \begin{center}
1074: \includegraphics[width=100mm]{f4.eps}
1075: \end{center}
1076: \caption{Results of model-s11.
1077: (a) Time evolution of the magnetic field (solid lines) and
1078: angular velocity $v_y$ (contours) on the $x$-$z$ plane.
1079: (b) Time evolution of vertically averaged
1080: angular velocity ($v_y$).
1081: The bold, dashed and thin solid lines express
1082: the snapshots at $t\Omega_{0} = 0.0, 40.0$ and $70.0$,
1083: respectively.
1084: The meanings of lines are
1085: the same as Fig.~\ref{fiducial}.
1086: }
1087: \label{s1}
1088: \end{figure}
1089: %
1090: %\clearpage
1091: %
1092: \begin{figure}[tbp]
1093: \begin{center}
1094: \includegraphics[width=100mm]{f5.eps}
1095: \end{center}
1096: \caption{
1097: The same plots as Fig.~\ref{s1} except for model-s055.
1098: }
1099: \label{s05}
1100: \end{figure}
1101: %
1102: %\clearpage
1103: %
1104: \begin{figure}[tbp]
1105: \begin{center}
1106: \includegraphics[width=60mm]{f6.eps}
1107: \end{center}
1108: \caption{
1109: The same plots as Fig.~\ref{s1}a except for model-s005.
1110: }
1111: \label{s005}
1112: \end{figure}
1113: %
1114: %\clearpage
1115: %
1116: \begin{figure}[tbp]
1117: \begin{center}
1118: \includegraphics[width=90mm]{f7.eps}
1119: \end{center}
1120: \caption{
1121: The same plots as Fig.~\ref{s1} except for model-u34.
1122: }
1123: \label{u34}
1124: \end{figure}
1125: %
1126: %\clearpage
1127: %
1128: \begin{figure}[tbp]
1129: \begin{center}
1130: \includegraphics[width=90mm]{f8.eps}
1131: \end{center}
1132: \caption{
1133: The same plots as Fig.~\ref{s1} except for model-$\eta 0$.
1134: }
1135: \label{eta}
1136: \end{figure}
1137: %
1138: %\clearpage
1139: %
1140: \begin{figure}[tbp]
1141: \begin{center}
1142: \includegraphics[width=100mm]{f9.eps}
1143: \end{center}
1144: \caption{
1145: The classification of results by the magnetic Reynolds number.
1146: Runs with (a)$\eta=0.0020\Omega_{0}H^2$ and (b)$\eta=0.0028\Omega_{0}H^2$
1147: on the $L_{\rm u}-L_{\rm s}$ plane
1148: and (c)all the runs on the $\eta-L_{\rm s}/L_{\rm u}$ plane.
1149: The four types of results, A, B, C and D, are represented by
1150: filled circles, triangles, daggers and crosses.
1151: The solid, dashed and dotted lines express $R_{\rm m,ave}=0.1,
1152: 0.5$ and $1.0$ respectively.
1153: The runs with numbers express
1154: (1)fiducial model; (2)model-s11; (3)model-s054; (4)model-s005
1155: (5)model-u34; (6)model-$\eta 0$.
1156: }
1157: \label{class}
1158: \end{figure}
1159: %
1160: \clearpage
1161: %
1162: \begin{figure}[tbp]
1163: \begin{center}
1164: \includegraphics[width=200mm]{f10.eps}
1165: \end{center}
1166: \caption{
1167: Results for various resolutions.
1168: (a)Close-up snapshots of vertically averaged angular velocity($v_y$)
1169: at $t\Omega_{0}=150.0$.
1170: (b)Time evolution of the volume-averaged Maxwell stress which is normalized by
1171: the initial pressure $P_{0}$.
1172: Resolutions corresponding to $dx=dz=0.005H$, $0.01H$, $0.0156H$ and $0.025H$ are represented by dashed, bold, dash-dotted and dotted lines respectively.
1173: }
1174: \label{resolution}
1175: \end{figure}
1176: %
1177: \clearpage
1178: %
1179: \begin{figure}[tbp]
1180: \begin{center}
1181: \includegraphics[width=200mm]{f11.eps}
1182: \end{center}
1183: \caption{
1184: Temporally and vertically averaged shear rate $q=-dv_{x}/dt$ ({\it top}),
1185: Maxwell stress normalized by initial pressure
1186: $- \langle B_{x}B_{y} \rangle /4\pi P_{0}$ ({\it middle}) and the
1187: growth rate estimated by the dispersion relation
1188: (equation (2) in Appendix 2) with the simulated values ({\it bottom})
1189: of the fiducial model. Bold lines are values temporally averaged
1190: over $t\Omega_{0}=45.0-70.0$ and dashed lines mean the initial values.
1191: }
1192: \label{growth}
1193: \end{figure}
1194: %
1195: %
1196: \begin{table}
1197: \begin{tabular}{llllll}
1198: \hline \hline
1199: $\eta/\Omega_{0}H^2$ & $L_{\rm u}/H$ & $L_{\rm s}/H$ &
1200: $R_{\rm m,ave}$ & {\rm result} & $\Delta v_{y}/c_s$ \\ \hline
1201:
1202: 0.0012 & 1.5 & 6.5, 6.0, 5.0, 4.0 & 0.10, 0.11, 0.13, 0.16
1203: & A, B, B, B & 0.82, 1.2, 0.91, 0.86 \\
1204: & & 1.0, 0.51 & 0.61, 1.1 & C, C & 0.66, 0.51\\
1205: & 2.0 & 7.5, 6.0, 4.0 & 0.13, 0.16, 0.23 & B, B, B
1206: & 1.7, 1.1, 0.98 \\
1207: & & 1.0, 0.51 & 0.89, 1.4 & C, D & 0.68, ... \\
1208: & 2.5 & 0.51 & 1.7 & D & ... \\ \hline
1209: 0.002 & 1.0 & 1.1, 0.54 & 0.23, 0.42 & B, B & 0.40, 0.38 \\
1210: & 1.2 & 3.1 & 0.10 & A & 0.57 \\
1211: & 1.4 & $4.0^{1}$, 3.1, $1.1^{2}$, $0.55^{3}$
1212: & 0.096, 0.13, 0.37, 0.64
1213: & A, B, B, C & 0.73, 0.55, 0.48, 0.41 \\
1214: & & 0.15, 0.10, $0.05^{4}$ & 1.2, 1.4, 1.5
1215: & D, D, D & ..., ..., ... \\
1216: & 1.9 & 6.0, 4.0, 3.0 & 0.094, 0.14, 0.19
1217: & A, B, B & 0.97,0.88, 0.54 \\
1218: & & 2.0, 1.0, 0.35 & 0.28, 0.53, 1.1
1219: & B, C, C & 0.87, 0.72, 0.43 \\
1220: & 2.4 & 2.0, 0.35 & 0.38, 1.3 & B, D & 1.2, ... \\
1221: & 2.9 & 10.0, 8.0, 1.0 & 0.092, 0.11, 0.81 & A, B, C
1222: & 1.2, 1.4, 0.99 \\
1223: & 3.4 & 12.0, 8.0, $4.0^{5}$, 0.24 & 0.090, 0.14, 0.28, 1.8
1224: & A, B, B, D& 1.0, 1.2, 1.1, ... \\
1225: & 3.9 & 3.1 & 0.44 & B & 0.77 \\ \hline
1226: 0.0028 & 1.4 & 5.1, $4.0^{6}$, 2.1, 0.57 & 0.15, 0.069, 0.14, 0.45
1227: & A, A, B, C & 0.35, 0.45, 0.28, 0.23 \\
1228: & & 0.37, 0.17, 0.11 & 0.61, 0.87, 1.00 & C, C, D
1229: & 0.34, 0.14, ... \\
1230: & 1.9 & 4.0, 3.1, 2.1, 1.1 & 0.099, 0.14, 0.20, 0.38
1231: & A, B, B, B & 0.48, 0.31, 0.49, 0.42 \\
1232: & & 0.57, 0.17, 0.11 & 0.61, 1.05, 1.2 & C, C, D
1233: & 0.40, 0.25, ... \\
1234: & 2.4 & 6.1, 5.1, 4.0, 1.1 & 0.088, 0.11, 0.13, 0.49
1235: & A, B, B, B & 0.50, 0.62, 0.66, 0.53 \\
1236: & & 0.57, 0.37, 0.11 & 0.75, 0.92, 1.25 & C, C, D
1237: & 0.46, 0.28, ... \\
1238: & 2.9 & 7.1, 6.1, 5.1, 2.1 & 0.093, 0.11, 0.12, 0.47
1239: & A, B, B, B & 0.60, 0.42, 0.50, 0.48 \\
1240: & & 1.1, 0.57, 0.27, 0.17 & 0.58, 0.85, 1.13, 1.26 & C, C, D, D
1241: & 0.56, 0.37, ..., ... \\
1242: & 3.4 & 8.1, 7.1, 6.1, 2.1 & 0.0097, 0.11, 0.13, 0.39
1243: & A, B, B, B & 1.00, 0.79, 0.58, 0.78 \\
1244: & & 1.1, 0.37, 0.27 & 0.66, 1.1, 1.3 & C, C, D
1245: & 0.48, 0.61, ... \\
1246: & 3.9 & 8.1, 2.1, 1.1 & 0.11, 0.45, 0.73
1247: & B, B, B & 0.46, 0.43, 0.49 \\ \hline
1248: 0.0036 & 1.8 & 4.1 & 0.077 & A & 0.28 \\
1249: 0.0044 & 1.7 & 4.1, 3.1 & 0.063, 0.085
1250: & A, A & 0.17, 0.18\\
1251: & & 2.1, 1.1, 0.43 & 0.13, 0.24, 0.50
1252: & B, B, B & 0.20, 0.18, 0.18 \\
1253: 0.0052 & 2.3 & 1.3 & 0.24 & B & 0.15 \\
1254: 0.0060 & 2.4 & 3.3, 2.3, 1.3, 0.32
1255: & 0.084, 0.36, 0.21, 0.50 & A, B, B, B & 0.22, 0.17, 0.21, 0.20 \\
1256: \hline
1257: \end{tabular}
1258: \end{table}
1259: %
1260: %
1261: %\clearpage
1262: %
1263: \begin{table}
1264: \caption{Simulation parameters and results for 91 runs.
1265: $\eta$, $L_{\rm u}$, $L_{\rm s}$, $R_{\rm m,ave}$,
1266: and $\Delta v$ are
1267: magnetic resistivity, radial width of unstable region,
1268: that of stable regions, initial averaged magnetic Reynolds number
1269: and the maximum deviation from Kepler velocity, respectively.
1270: For highly turbulent cases (marked ``D''), $\Delta v_{y}$ is omitted.
1271: The fifth column ``result'' indicates classification of
1272: the results (\S 3.3).
1273: Multiple values in the columns correspond to different runs.
1274: (1) fiducial model;
1275: (2) model-s11; (3) model-s055; (4) model-s005;
1276: (5) model-u34; (6) model-$\eta 0$.
1277: }
1278: \label{tab:1}
1279: \end{table}
1280: %
1281: %
1282: \end{document}
1283:
1284: %%
1285: %% End of file `sample.tex'.
1286:
1287:
1288: