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

2: %\documentstyle[12pt,aaspp4]{article}

3: 

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

5: 

6: \usepackage{emulateapj5}

7: %\usepackage{amsmath}

8: %\usepackage{graphicx}

9: %\usepackage{subfigure}

10: \usepackage{apjfonts}

11: 

12: \begin{document}

13: \def\gapprox{\mathrel{\vcenter{\offinterlineskip \hbox{$>$}

14:     \kern 0.3ex \hbox{$\sim$}}}}

15: \def\lapprox{\mathrel{\vcenter{\offinterlineskip \hbox{$<$}

16:     \kern 0.3ex \hbox{$\sim$}}}}

17: 

18: \title{The Magnetic Rayleigh-Taylor Instability in Three Dimensions}

19: 

20: \author{James M. Stone and Thomas Gardiner\altaffilmark{1}}

21: \affil{Department of Astrophysical Sciences, Princeton University, Princeton,

22: NJ 08544}

23: \altaffiltext{1}{Present address: 3915 Rayado Pl NW, Albuquerque, NM 87114}

24: 

25: \begin{abstract}

26: We study the magnetic Rayleigh-Taylor instability in three dimensions,

27: with focus on the nonlinear structure and evolution that results from

28: different initial field configurations.  We study strong fields in the

29: sense that the critical wavelength $\lambda_c$ at which perturbations

30: along the field are stable is a large fraction of the size of the

31: computational domain.  We consider magnetic fields which are initially

32: parallel to the interface, but have a variety of configurations, including

33: uniform everywhere, uniform in the light fluid only, and fields which

34: change direction at the interface.  Strong magnetic fields do not suppress

35: instability, in fact by inhibiting secondary shear instabilities, they

36: reduce mixing between the heavy and light fluid, and cause the rate of

37: growth of bubbles and fingers to increase in comparison to hydrodynamics.

38: Fields parallel to, but whose direction changes at, the interface produce

39: long, isolated fingers separated by the critical wavelength $\lambda_c$,

40: which may be relevant to the morphology of the optical filaments in the

41: Crab nebula.

42: 

43: \end{abstract}

44: 

45: \keywords{MHD, instabilities, ISM:magnetic fields}

46: 

47: \section{Introduction}

48: 

49: There are a number of astrophysical systems in which the magnetic

50: Rayleigh-Taylor instability (RTI) is expected to be important, for

51: example accretion onto magnetized compact objects (Arons \& Lea 1976;

52: Wang \& Nepveu 1983; Wang, Nepveu, \& Robertson 1984), buoyant bubbles

53: generated by radio jets in clusters of galaxies (Robinson et al. 2005;

54: Jones \& De Young 2005; Ruszkowski et al. 2007)), the emergence of

55: magnetic flux from the solar interior and the formation of flux tubes

56: (Isobe et al. 2005; 2006; and references therein), and at both the contact

57: discontinuity between the shocked circumstellar medium and ejecta in

58: supernovae remnants (Jun \& Norman 1996a; b), and in the thin shell of

59: ejecta swept up by a pulsar wind (Hester et al. 1996, hereafter H96;

60: Bucciantini et al.  2004).  For the idealized case of two inviscid,

61: perfectly conducting fluids separated by a contact discontinuity with

62: a uniform magnetic field ${\bf B}$ parallel to the interface undergoing

63: constant acceleration $g$, then a linear analysis (Chandrasekhar 1961)

64: demonstrates that for modes parallel to the magnetic field there is a

65: critical wavelength

66: \begin{equation}

67: \lambda_c = \frac{B^2}{g(\rho_h - \rho_l)}

68: \end{equation}

69: below which instability is completely suppressed, where $\rho_h$ and

70: $\rho_l$ are the densities in the heavy and light fluids respectively,

71: and we have chosen a system of units in which the magnetic permeability

72: $\mu=1$.  At larger wavelengths, the growth rate is reduced compared

73: to the hydrodynamic case, and there is a peak growth rate occurring at

74: a wavelength $\lambda_{\rm max} = 2\lambda_c$.  Equation (1) can also

75: be thought of as a condition on the magnetic field: instability on a

76: scale $L$ parallel to the field requires $B < B_c \equiv [Lg(\rho_h -

77: \rho_l)]^{1/2}$.  Modes perpendicular to the field are unaffected,

78: and have the same growth rate and stability condition as in pure

79: hydrodynamics.  The highly anisotropic nature of the growth rate of

80: modes parallel versus perpendicular to the field suggests that it is

81: important to study the nonlinear regime of the magnetic RTI in full

82: three dimensions.

83: 

84: One of the most compelling applications of the magnetic RTI is to the

85: structure of the optical filaments in the Crab nebula (H96).  As the

86: low density, highly magnetized synchrotron nebula powered by the Crab

87: pulsar sweeps up the stellar ejecta, the interface between the two

88: is RT unstable, resulting in radially orientated filaments that point

89: to the center of the synchrotron nebula.  H96 have suggested the long,

90: widely spaced filaments observed by HST are a consequence of suppression

91: of short wavelength modes due to the magnetic field in the synchrotron

92: plasma, since the filaments bear no resemblance to the turbulent mixing

93: layer that results from the RTI in hydrodynamics (Dimonte et al 2004,

94: hereafter D04), but are a better fit to the morphology that results from

95: the magnetic RTI in two-dimensions (Jun, Norman, \& Stone 1995).  However,

96: in purely hydrodynamic simulations of the RTI in the spherically expanding

97: shell swept up by the pulsar wind, Jun (1998) was able to reproduce the

98: morphology and separation of the fingers remarkably well, suggesting

99: than geometrical effects might be important.  Since the simulations were

100: performed in two-dimensions assuming axial symmetry, it is unclear if

101: the isolated fingers will persist in three-dimensional hydrodynamics, or

102: whether MHD effects are indeed essential.  More recently, Bucciantini

103: et al. (2004) have presented the most realistic numerical models

104: of the filaments in the Crab nebula to date, using two-dimensional

105: MHD simulations of the expanding shell and nebula.  In their

106: more realistic treatment of the conditions at the unstable interface,

107: they find fields near equipartition can completely suppress the RTI.

108: However, as they point out, because of the anisotropic nature of the

109: magnetic RTI, three-dimensional effects are critical and need to be

110: included in future studies.

111: 

112: Since fully three-dimensional MHD simulations in a spherical geometry

113: which can follow the expanding shell of ejecta are computationally challenging,

114: it is worthwhile to begin investigation of three-dimensional effects

115: in the idealized plane parallel case.  Recently, we have described an

116: extensive study of the nonlinear evolution of the magnetic RTI in a

117: three-dimensional planar geometry, focusing on the effect of varying the

118: field strength on the growth rate of fingers and bubbles at the interface,

119: and on the amount of mixing between the heavy and light fluids (Stone

120: \& Gardiner 2007, hereafter Paper I).  To facilitate comparison with

121: previous experimental and computational studies of the hydrodynamic RTI

122: (D04), a relatively modest difference in density between the fluids

123: was chosen, that is $\rho_h/\rho_l = 3$.  In this paper, we extend the

124: study by considering a more astrophysically relevant density ratio,

125: $\rho_h/\rho_l = 10$, and by focusing on the effect of strong magnetic

126: fields (in the sense that $\lambda_c \sim L$, where L is the size of the

127: computational domain) on the suppression of the RTI in three dimensions.

128: 

129: A number of studies of magnetic buoyancy instabilities in three

130: dimensions have been reported, both in the context of the emergence

131: of new magnetic flux from the solar photosphere (Wissink et al. 2000;

132: Fan 2001; Isobe et al. 2005; 2006), and the nonlinear evolution of the

133: Parker instability in the galactic disk (Kim, Ostriker, \& Stone 2002;

134: Kosi\'{n}ski \& Hanasz 2007).  In these studies, the magnetic field is strong

135: enough for the ratio of thermal to magnetic pressure $\beta \sim 1$, so

136: that the magnetic field not only plays a significant role in the support

137: of the initial equilibrium state, but also is responsible for driving

138: buoyant motions.  In contrast, we study weak fields in the sense that

139: $\beta \gg 1$, so that the magnetic field plays almost no role in the

140: vertical equilibrium, and the RTI is driven by the buoyancy of the fluid.

141: Our goal is to study how magnetic fields affect the evolution

142: of the classical RTI.

143: 

144: Our primary conclusions are that in three dimensions, uniform magnetic

145: fields do not suppress the RTI due to the growth of interchange modes

146: perpendicular to the field.  In fact, since magnetic fields suppress

147: secondary Kelvin-Helmholtz instabilities and therefore mixing between

148: the heavy and light fluids, the growth rate of bubbles and fingers is

149: in fact enhanced in the magnetic RTI compared to the hydrodynamic case.

150: We explore a variety of initial field

151: configurations, including uniform fields, uniform fields in the light

152: fluid only, and fields with a rotation at the interface, and we show that

153: well separated, long fingers reminiscent of the optical filaments in

154: the Crab nebula can be generated if the magnetic field direction changes

155: through large angles over a distance short compared to $\lambda_c$.

156: 

157: \section{Method}

158: We solve the equations of ideal MHD with a constant vertical

159: acceleration ${\bf g} = (0,0,g)$ 

160: %

161: \begin{eqnarray}

162: \frac{\partial \rho}{\partial t} +

163: {\bf\nabla\cdot} \left(\rho{\bf v}\right) & = & 0

164: \label{eq:cons_mass} \\

165: %

166: \frac{\partial \rho {\bf v}}{\partial t} +

167: {\bf\nabla\cdot} \left(\rho{\bf vv} - {\bf BB}\right) +

168: {\bf \nabla} P^{*} & = & \rho {\bf g}  \\

169: %

170: \frac{\partial {\bf B}}{\partial t} +

171: {\bf\nabla}\times \left({\bf v} \times {\bf B}\right) & = & 0 \\

172: %

173: \frac{\partial E}{\partial t} +

174: \nabla\cdot((E + P^*) {\bf v} - {\bf B} ({\bf B \cdot v})) & = & \rho {\bf v} \cdot {\bf g}

175: \label{eq:cons_energy}

176: \end{eqnarray}

177: %

178: %

179: The total pressure $P^* \equiv P + ({\bf B \cdot B})/2$,

180: where $P$ is the gas pressure.  The total energy density $E$ is 

181: %

182: \begin{equation}

183: E \equiv \epsilon + \rho({\bf v \cdot v})/2 + ({\bf B \cdot B})/2 ~.

184: \end{equation}

185: %

186: where $\epsilon$ is the internal energy density.  We use an ideal gas

187: equation of state for which $P = (\gamma - 1) \epsilon$, where $\gamma$

188: is the ratio of specific heats.  We use $\gamma=5/3$ in this paper.

189: In relativistic plasmas such as synchrotron nebulae $\gamma=4/3$ would

190: be more appropriate.  However, given our choice for the numerical value

191: of $g$ and the size of the computational domain (see below), the flows

192: induced by the magnetic RTI are subsonic and nearly incompressible.

193: Thus, varying the adiabatic index should have little effect on the

194: results reported here.

195: 

196: The three-dimensional computational domain is of size $L

197: \times L \times 2L$, where $L=0.1$.

198: Periodic boundary conditions are used in the transverse

199: ($x-$ and $y-$) directions, and reflecting boundary conditions are used at

200: the top and bottom.  The origin of the $z-$coordinate

201: is centered in the domain, so that the computations span $-0.1 \leq z

202: \leq 0.1$  The upper half of the domain ($z>0$) is filled with heavy

203: fluid of density $\rho_h=10$, while in the lower half ($z<0$) the density

204: of the light fluid is $\rho_l=1$.  Thus, the Atwood number

205: %

206: \begin{equation}

207:  A \equiv \frac{\rho_h - \rho_l}{\rho_h + \rho_l} = \frac{9}{11}.

208: \end{equation}

209: %

210: Most of the experimental studies of the hydrodynamic RTI used

211: to validate computational methods (D04) use $A=1/2$.  In Paper I we

212: studied the magnetic RTI with $A=1/2$;

213: in this paper we study the high Atwood number regime which is more

214: relevant to most astrophysical systems.

215: 

216: Initially the gas is in magnetohydrostatic equilibrium, with the amplitude

217: of the gas pressure chosen so that the sound speed in the light fluid

218: $c_s=1$ at the interface, thus

219: %

220: \begin{equation}

221: P^*(z) = \frac{3}{5} - g \rho z + B^{2}/2

222: \end{equation}

223: %

224: The sound crossing time in the light fluid at the interface $t_s =0.1$.

225: We choose $g=0.1$, thus the ratio of the free-fall velocity to the sound

226: speed $\sqrt{gL}/c_s = 0.1$, implying the induced flows should be

227: nearly incompressible.

228: 

229: The magnetic field is initialized with an amplitude $B_{0}$ that is chosen

230: to be a fixed fraction of the critical field strength $B_c$ at which

231: there are no unstable modes within $L$, we choose $B_{0} \approx 0.6B_c$.

232: From equation (1), the critical

233: wavelength at which all modes are suppressed $\lambda_c/L \approx 0.35$.

234: The field is always initially parallel to the interface, but has a variety

235: of different initial configurations which will be described along with the

236: results of each individual simulation in \S 3.  The ratio of the gas to

237: magnetic pressure at the interface $\beta = 480$ in all the runs.  Thus,

238: although we study strong fields in the sense that $\lambda_c \sim L$,

239: the energy density in the field is far below equipartition, and the field

240: plays little role in the initial vertical equilibrium.  Increasing the

241: size of the computational domain $L$ to accommodate a larger $\lambda_c$

242: associated with stronger, near-equipartition fields, or simply lowering

243: the sound speed to decrease $\beta$ in the present simulations, will

244: both produce flows in which $\sqrt{gL}/c_s$ is increased, and therefore

245: are more compressible.

246: 

247: To seed the RTI, zone-to-zone perturbations are added to the vertical

248: velocity $v_z$ throughout the volume with an amplitude $A$ that is kept

249: small compared to the sound speed, and is decreased toward the vertical

250: boundaries; thus $A = A_0 R (1 + \cos{2\pi z/L})$ where $A_0=0.005$, and

251: $R$ is a random number between -1 and 1.  The maximum perturbed velocity

252: is only 1\% of the sound speed in the light fluid at the interface.

253: 

254: The computations presented in this paper use Athena, a new MHD code

255: that implements a recently developed

256: Godunov method for compressible MHD (Gardiner \& Stone 2005; 2007).

257: A complete description of the algorithm, including the results of an

258: extensive series of test problems, is given in these references.  All of

259: the simulations use a grid of $256 \times 256 \times 512$, which means the

260: critical wavelength $\lambda_c$ is resolved with nearly 100 grid points.

261: Our numerical resolution is much higher than used in previous work

262: (for example, the few 3D simulations reported

263: in Jun, Norman \& Stone 1995), and uses stronger initial fields.

264: 

265: In Paper I, we presented a comprehensive convergence study of our

266: numerical algorithms for the magnetic RTI in two dimensions, focused

267: on the amount of mass mixing due to numerical effects.  For single

268: mode perturbations, features such as the shape and height of the

269: interface at a fixed time were converged with 32 or more grid points

270: per wavelength.  The amount of mixing between heavy and light fluids

271: was also found to converge to zero at first order, independent of the

272: magnetic field strength.  First order convergence is consistent with

273: mixing being proportional to the width of the interface between the

274: heavy and light fluids (which cannot be smaller than one grid cell).

275: With multimode perturbations, the degree of mixing does not converge to

276: zero, because at higher resolution there are more small scale distortions

277: in the interface which increase its surface area.  Convergence of

278: the mixing to zero with multimode perturbations therefore requires

279: the introduction of surface tension or viscosity to create a fixed

280: small scale below which the interface is smooth.  Instead, in this

281: paper we compute all solutions at the highest resolution we can afford

282: (so that they are all at the same, high Reynolds and magnetic Reynolds

283: numbers), and focus on the {\em comparison} of features with different

284: field strengths and geometries that occur at these Reynolds numbers.

285: In this way, we can isolate the effects of changing field strength or

286: geometry from the effect of changing the numerical diffusion.

287: 

288: \section{Results}

289: 

290: We describe the results from simulations that use a variety of different

291: initial magnetic field configurations. 

292: 

293: \subsection{Uniform Field versus Hydrodynamics}

294: 

295: We begin with the evolution in a uniform magnetic

296: field parallel to the interface and along the $x-$axis, ${\bf B} =

297: (B_{0},0,0)$.  For comparison purposes, we also describe the results of

298: a hydrodynamical calculation, computed with the same parameters, grid,

299: and numerical algorithm.  Hereafter, we refer to the uniform field case

300: as run U, and the hydrodynamical simulation as run H.

301: 

302: Figure 1 shows isosurfaces of the density, along with slices of the

303: density at the edges of the computational domain, at two times during

304: the evolution for both runs H and U.  The hydrodynamic case shows

305: the typical evolution of the RTI into a turbulent mixing layer (D04).

306: In hydrodynamics, short wavelength modes grow fastest, thus at early times the

307: instability is dominated by bubbles and fingers at small scales.

308: Secondary Kelvin-Helmholtz instabilities, associated with the shear

309: between the rising and descending plumes, give the tips a ``mushroom-cap"

310: appearance, and cause some of the fingers to break up.  At late times,

311: mergers between bubbles favors growth of structure at larger scales, while

312: secondary instabilities continue to distort the plumes and cause mixing.

313: Note the large fraction of fluid at intermediate densities (green colors)

314: at late times in the hydrodynamic case.

315: 

316: In the uniformly magnetized simulation run U, the early nonlinear phase

317: of the RTI shows the strongly anisotropic structure of modes introduced by

318: the magnetic field.  Perpendicular to the field (along the $y-$axis), 

319: interchange modes grow fastest at short wavelength, whereas along the field

320: short wavelengths are suppressed.  As a result, the interface develops a

321: filamentary structure that is strongly reminiscent of the structure

322: reported by Isobe et al (2005; 2006) in simulations of flux tubes emerging from

323: the solar photosphere.  At late times, fluid flowing along flux tubes

324: collects at bubbles and fingers at the tips (similar to the nonlinear

325: evolution of the Parker instability, Kim et al. 1998), which are then

326: wrinkled by interchange instability at their surface.  This produces

327: large-scale smooth bubbles.  Slices along the edges of the domain reveal

328: far less mixing than in the hydrodynamic case.  Note in three-dimensions

329: the magnetic RTI in a uniform field does not result in isolated, long

330: fingers comparable to the observations of the Crab (H96).

331: 

332: One measure of the rate of growth of the RTI is the time evolution of the

333: height $h$ of bubbles from the interface.  Self-similar arguments (D04)

334: predict that

335: \begin{equation}

336: h = \alpha Agt^2

337: \end{equation}

338: where $\alpha$ is a dimensionless constant.  The experimentally measured

339: value is $\alpha = 0.057 \pm 0.008$.  Without specialized

340: front-tracking algorithms that can prevent mixing between the fluids

341: at the grid scale, most numerical methods give a value for $\alpha$ which

342: is about a factor of two smaller (D04, Paper I).

343: 

344: Figure 2 plots the location of the tips of the rising

345: bubbles as a function of time in both runs H and U.  At any instant in

346: time, we define the vertical location of the tips of the fingers as the

347: point where the horizontally averaged fraction of the heavy fluid

348: \begin{equation}

349:  \langle f_h \rangle = \int_x \int_y f_h dx dy/L^2

350: \end{equation}

351: is 0.95, where for incompressible fluid with $\rho_h=10$ and $\rho_l=1$

352: the fraction of heavy fluid in any cell is $f_h = (\rho -1)/9$.

353: (To account for the effects of compressibility, we choose $f_h=0.95$

354: rather than one to mark the boundary of the mixing region.)  From figure

355: 2, we see that after an initial rise, the increase in $h$ in both

356: hydrodynamics and MHD follows the expected self-similar scaling equation

357: (9).  In hydrodynamics, the slope $\alpha =0.03$, whereas in MHD the

358: slope $\alpha =0.05$ where we have ignored the final point in both cases,

359: since it is undoubtedly affected by the reflecting boundary conditions

360: at the top of the domain $h/L=1$.  It is clear that the bubbles rise

361: {\em faster} in MHD than in the hydrodynamic RTI, in agreement with the

362: results at $A=1/2$ (Paper I).  As discussed in \S 3.5, this is primarily

363: due to the reduction of mixing in the MHD case.

364: 

365: \subsection{Field in Light Fluid Only}

366: 

367: In the magnetic RTI associated with some astrophysical systems, such

368: as the interface between the pulsar wind nebula and the supernova

369: ejecta in the Crab nebula, only the light fluid is expected to be

370: strongly magnetized.  Given that the results in section \S 3.1 show

371: that strong, uniform fields do not suppress the RTI, it is

372: unlikely that a strong field in the light fluid only will inhibit

373: instability.  Nonetheless, it is of interest to investigate the

374: structure of the nonlinear regime in this case.

375: 

376: Figure 3 plots isosurfaces of the density, along with slices of the

377: density at the edges of the computational domain, at two times during

378: the evolution of a simulation in which the magnetic field is uniform,

379: parallel to the interface, and along the $x-$axis, ${\bf B} = (B_{0},0,0)$

380: in the light ($\rho=1$) fluid only, with ${\bf B} = 0$ everywhere else.

381: As before, we choose $B_0 = 0.6 B_{c}$.

382: The gas pressure is increased in the heavy fluid above the interface

383: so that the total pressure is continuous, that is exact vertical

384: equilibrium is maintained initially.  We refer to this calculation as

385: run LF hereafter.

386: 

387: It is instructive to compare the structures observed in figure 3 with

388: the uniform field case (bottom row of figure 1).  At the early time in

389: run LF, the fingers and bubbles are not elongated along the field as

390: in run U.  Instead, the structure is nearly isotropic, similar to the

391: hydrodynamic case but with less small scale structure.  At late time,

392: large smooth bubbles emerge in run LF that appear isotropic.  Overall,

393: the three-dimensional structure of the fingers and bubbles in run LF is

394: intermediate between the hydrodynamic and uniformly magnetized runs.

395: The density slices at the edge of the domain show much less mixing than

396: run H.  The height of the bubbles and degree of mixing (revealed by the

397: density slices at the edge of the domain) show much more similarity to

398: run U; these will be analyzed further in \S 3.5.  Once again, we see that

399: in three-dimensions, strong uniform fields in the light fluid

400: are unable to inhibit the RTI.

401: 

402: \subsection{Fields with a Discontinuous Rotation}

403: 

404: In most astrophysical systems, there is no reason to expect the

405: magnetic field has the same geometry in both the light and heavy fluids.

406: Since only unstable modes parallel to the magnetic field are suppressed,

407: rotating the field near the interface will inhibit

408: modes in multidimensions.

409: To investigate this regime we have performed simulations in

410: which the magnetic field is rotated discontinuously through large angles

411: at the interface.

412: In the first simulation, hereafter referred to as run R45, the field is

413: rotated through $45^{\circ}$, that is ${\bf B} = (B_{0},0,0)$ in the

414: light fluid ($z<0$), and ${\bf B} = (B_{0}/\sqrt{2},B_0/\sqrt{2},0)$

415: in the heavy fluid ($z>0$).  In the second simulation, hereafter referred

416: to as run R90, the field is rotated through $90^{\circ}$, that is ${\bf

417: B} = (B_{0},0,0)$ in the light fluid, and ${\bf B} = (0,B_0,0)$ in the

418: heavy fluid.  In both cases, there is a current sheet at the

419: interface.

420: 

421: Figure 4 plots isosurfaces of the density, along with slices of the

422: density at the edges of the computational domain, at two times during

423: the evolution of both runs R45 and R90.  At early times in both cases,

424: filamentary structures appear at an angle roughly half-way between the

425: direction of the field in the heavy and light fluids (about $22^{\circ}$

426: with respect to the $x-$axis in R45, $45^{\circ}$ with respect to

427: the $x-$axis in R90), most likely because the magnetic tension forces

428: which are proportional to ${\bf k}\cdot{\bf B}$ are minimized in this

429: direction.  Analysis of the magnetic field and velocity perturbations

430: at this time shows flow occurs along the field lines into the ridges.

431: Pure interchange modes are no longer possible with rotated fields,

432: and the growth of perturbations requires ${\bf k}\cdot{\bf B} \ne 0$ in

433: either the light or heavy fluids, or both.  Note the amplitude of perturbations

434: is much smaller in R90 at early times in comparison to R45, and only

435: long wavelength modes are present.

436: 

437: By $t/t_s=40$, the interface in both R45 and R90 is strongly distorted

438: by RTI.  Interestingly, the structure of modes at late times is quite

439: different from previous cases.  Isolated, large scale bubbles dominate,

440: with very smooth surfaces, and bulbous tips.  The spacing between bubbles

441: is roughly the critical wavelength $\lambda_c$.  The structure of R90 is

442: particularly interesting.  The fingers in this case are nearly isotropic,

443: and have a length which significantly exceeds $\lambda_c$.  The surface

444: of the bubbles is extremely smooth, whereas in R45 there is some evidence

445: for wrinkling due to interchange modes.  The interface between the light

446: and heavy fluids is remarkably thin in R90.  At the faces of the volume,

447: the density slices reveal very little material at densities intermediate

448: to the values of the isosurfaces (at $\rho=9.9$ and 1.1 respectively).

449: Thus, the faces of the volume are transparent, and the interior of the

450: bubbles is clearly visible.  Contrast this to run H in figure 1, where

451: the slice at the edge of the domain revealed a turbulent mixing layer.

452: A more quantitative analysis of mixing in all the runs

453: will be presented in \S 3.5.

454: 

455: \subsection{Fields with Continuous Rotation}

456: 

457: In the previous section, the direction of the magnetic field was changed

458: discontinuously at the interface, resulting in a current sheet.  It is

459: possible that in many astrophysical systems, the direction of the field

460: varies smoothly on many different scales.  To investigate the effect

461: this might have on the magnetic RTI, we consider the case where the field

462: amplitude is constant everywhere, while the direction is rotated through

463: a large angle (we choose $90^{\circ}$) over a finite vertical distance

464: $L_{rot}$. More specifically, for $z<-L_{rot}/2$ the field is ${\bf B} =

465: (B_{0},0,0)$, for $-L_{rot}/2<z<L_{rot}/2$ the direction of the field

466: varies linearly with $z$ from along the $x-$axis to along the $y-$axis

467: while the amplitude is fixed at $B_{0}$, and for $z>L_{rot}/2$ the field

468: is ${\bf B} = (0,B_{0},0)$.  Note this geomtry results in a current

469: layer with constant amplitude in the region $-L_{rot}/2<z<L_{rot}/2$.

470: If $L_{rot} \ll \lambda_{c}$ we expect this initial configuration to

471: evolve similar to the discontinous rotation run R90 studied in \S3.3,

472: while if $L_{rot} \gg \lambda_{c}$ it will evolve like the uniform field

473: case run U studied in \S3.1.  Here we choose $L_{rot}/\lambda_{c} =

474: 0.5$, and hereafter refer to this calculation as run C90.

475: 

476: In fact, we find at late times the structure that emerges from the

477: magnetic RTI in run C90 is remarkably similar to that produced in run R90.

478: For example, at $t/t_s=40$s, isolated smooth bubbles are produced with

479: similar sizes and spacing as observed in figure 4.  Conversely, we

480: find at early times there is little suppression of interchange modes.

481: This is not suprising: the fastest growing modes occur at the largest

482: wavenumbers, and therefore have wavelengths much smaller than $L_{rot}$.

483: On these scales, the early evolution of the interface is as if the field were

484: uniform (run U).

485: Figure 5 plots the height of bubbles in run C90 versus the uniform field

486: case run U.  The evolution of both is very similar.  Our results confirm

487: the intuition that changes in the direction of the field at the interface

488: must be on very small scales to inhibit the interchange modes.

489: 

490: \subsection{Mixing}

491: 

492: The amount of mixing between the heavy and light fluids strongly affects

493: the rate at which bubbles and fingers are displaced from the interface

494: (D04, Paper I).  The presence of even a weak field can, through the action

495: of tension forces at small scales, significantly reduce mixing in 

496: comparison to hydrodynamics (Paper I).  Here we investigate mixing in the

497: simulations presented above.

498: 

499: Figure 6 plots the height of bubbles above the interface, defined using

500: the point at which $\langle f_h \rangle=0.95$, for runs U, LF, R45

501: and R90.  In every case, at late times the height $h$ grows as $t^2$,

502: as expected (equation 9).  However, in R45, and especially R90, growth

503: is delayed.  The slope of the lines, as measured by the dimensionless

504: constant $\alpha$ are remarkably similar, $\alpha = 0.050 \pm 0.005$.

505: The decrease in the slope at late time in each model is most likely an

506: influence of the upper (reflecting) boundary condition rather than a

507: divergence from the self-similar evolution.

508: 

509: It is useful to define a mixing parameter $\Theta$ as

510: \begin{equation}

511: \Theta =4\langle f_h f_l \rangle

512: \end{equation}

513: The peak value of $\Theta$ is one, and occurs when $f_h=f_l=1/2$, that

514: is in regions that are fully mixed.  In regions that are not mixed,

515: $\Theta=0$.  Figure 7 plots the profile of $\Theta$ versus height in

516: runs H, U, and R90.  Note in the hydrodynamic case run H, the mixing

517: parameter is close to the theoretical maximum near the original location

518: of the interface $z=0$.  This quantifies the result which is evident from

519: a visual inspection of figure 1, namely the hydrodynamic RTI results in

520: a turbulent mixing zone which is dominated by material at intermediate

521: densities.  On the other hand, the uniformly magnetized case run U

522: shows far less mixing than run H, again a fact which is evident from

523: the lower panels of figure 1.  Finally, run R90 shows the least mixing,

524: with a peak value of $\Theta$ which is five times smaller than the peak

525: value in run H.  At the peak of $\Theta$ in run R90, the horizontally

526: averaged fraction of heavy fluid $\langle f_h \rangle = 0.2$, indicating

527: the fingers of heavy fluid occupy a much smaller volume than the bubbles

528: of light fluid.  Again, all of these results are evident from figure 4,

529: where density slices at the edge of the domain show the mixing layer

530: between the two fluids to be very thin, and that the bubbles of light

531: fluid fill most of the volume.

532: 

533: \subsection{Magnetic Field Evolution}

534: 

535: Self-similar arguments predict that the rate of growth of the height $h$

536: of bubbles and fingers should be proportional to $t^2$ (equation 9).

537: Since the amount

538: of gravitational binding energy released by the descending plumes of heavy

539: fluid is proportional to $h^2$ (the energy released is the product of

540: the mass involved in the flow and the distance it falls, both of which

541: are proportional to $h$), we expect the rate of growth of

542: the kinetic and magnetic energies in the RTI should be proportional

543: to $t^{4}$.

544: 

545: Figure 8 plots each component of the volume averaged

546: kinetic and magnetic energies, normalized by the initial volume averaged

547: magnetic energy $B_{0}^{2}/2$, in runs U and R90 versus $t^4$.  The magnetic energy

548: associated with the horizontal components of the field have their initial

549: values subtracted as appropriate, thus the plot shows the fractional change

550: in the energies.  Note that at early times, the curves are straight lines,

551: indicating the expected scaling with $t^4$ is recovered.  In each

552: case the vertical components of the energies dominate, and in the

553: horizontal directions there is rough equipartition between kinetic and

554: magnetic energies.  The kinetic energy associated with the $y-$component

555: of the kinetic energy is larger in comparison to the $x-$component

556: in run U since motions perpendicular to the field are favored by

557: interchange modes, which we have shown are important in strong uniform

558: fields.   The amplification of the vertical field is larger in run R90

559: in comparison to run U, although the total magnetic energy in all 

560: components of the field is roughly the same at late times in both cases.

561: This is another indication that run R90 leads to ordered, vertical flows

562: and columns, whereas larger amplitude horizontal flows (and therefore

563: more mixing and less ordered fingers) are produced in uniform fields.

564: In both cases the magnetic RTI leads to significant amplification of

565: magnetic energy.

566: 

567: It is worth emphasizing that the time evolution of volume averaged

568: quantities as shown in figure 8 is controlled by a number of dimensionless

569: parameters, including the ratio of the critical wavelength to the size

570: of the computational domain $\lambda_c/L$ and the ratio of the free fall

571: to the sound crossing time $\sqrt{\lambda_c/g}/t_{s}$.  We have studied

572: strong fields in the sense that $\lambda_c/L \sim 1$.  If the calculations

573: were repeated with identical parameters but in a much larger domain,

574: then the evolution would resemble the weak field simulations presented

575: in Paper I.  That is, if the calculations presented here were continued in

576: a much larger domain, so that the height of the fingers and bubbles $h

577: \gg \lambda_c$, then the flow would become more hydrodynamic, a turbulent

578: mixing zone would emerge, and once the vertical field is a large fraction

579: of the initial horizontal value, the $t^4$ scaling of energies is broken

580: (Paper I).

581: 

582: \section{Summary and Discussion}

583: 

584: We have shown that strong, uniform magnetic fields cannot suppress the

585: RTI in three dimensions.  In the linear regime only long wavelength modes

586: parallel to the magnetic

587: field are unstable; interchange modes perpendicular to the

588: field are unaffected, and grow at the same rate as in

589: hydrodynamics.   We have shown that in the nonlinear regime this leads

590: to a highly anisotropic structure.  At late times, flow of plasma along

591: field lines produces large bubbles much as in the Parker instability (Kim

592: et al 1998), which in turn become wrinkled by secondary interchange modes.

593: 

594: In fact, in one respect strong magnetic fields actually increase

595: the growth rate of the RTI in the nonlinear regime, in comparison to

596: hydrodynamics.  Magnetic fields inhibit secondary instabilities and

597: mixing between the light and heavy fluid.  In turn, the reduction of

598: mixing causes bubbles (fingers) to rise (fall) more rapidly.  In fact,

599: the tension force associated with even weak fields can suppress mixing

600: on small scales, and increase the growth rate of bubbles and fingers

601: (Paper I).  The suppression of a turbulent mixing layer with even a weak

602: magnetic field could be relevant to a number of astrophysical systems,

603: for example the evolution of supernovae remnants (Jun \& Norman 1996a;

604: b), or the mixing of metals from early generations of stars into the

605: intergalactic medium.

606: 

607: Although uniform magnetic fields do not suppress the RTI, we have shown

608: that if the direction of the field changes through a large angle at

609: the interface, this can delay instability, and significantly alter

610: the structures that emerge in the nonlinear regime.  We have studied

611: field geometries that have both discontinuous rotations of the field

612: at the interface, and continuous rotations over a finite vertical

613: length $L_{rot}$ at the interface.  When $L_{rot}/\lambda_c \leq 1$,

614: the nonlinear regime in both these cases is similar, and consists of

615: isolated, smooth, long fingers and bubbles.

616: 

617: There are several obvious applications of the magnetic RTI to

618: astrophysical systems.  The first is to the penetration of infalling

619: plasma into the magnetosphere of an accreting neutron star (Arons \&

620: Lea 1976; Wang \& Nepveu 1983; Wang, Nepveu, \& Robertson 1984), or to

621: the confinement of the plasma along field lines at the polar caps (Litwin,

622: Brown, \& Rosner 2001).  A related problem, confinement of strong vertical

623: flux tubes at the galactic center, has been investigated by Chandran

624: (2001).  In each of these cases (except the last), the field is rigidly

625: anchored at a boundary, whereas we have studied the magnetic RTI with

626: periodic boundary conditions in both horizontal directions.  The nonlinear

627: evolution of interchange modes will probably be strongly affected no-slip

628: boundary conditions on the magnetic field at the edges of the domain,

629: so our results may only have limited applicability to these systems.

630: 

631: The second is to the stability of buoyant bubbles generated by radio

632: jets in clusters of galaxies.  Robertson et al. (2004) and Jones \& De Young

633: (2005) have presented

634: two-dimensional simulations of the morphology of magnetized, buoyant

635: bubbles.  However, it is clear that three-dimensional effects will

636: be very important in this problem, due to the very different behavior

637: of modes perpendicular versus parallel to the field.  Recent work in 3D

638: by Ruszkowski et al. (2007) confirms that magnetic fields are unable to

639: suppress shredding of bubbles in three dimensions unless the coherence length

640: of the field is larger than the size of the bubble.  In fact,

641: magnetic fields in cluster gas can alter the dynamics in ways

642: than go beyond the obvious effects of magnetic stresses.  Due to the long

643: mean-free-paths of particles, anisotropic heat conduction and viscosity

644: are important in hot cluster gas.  Balbus (2000) has

645: shown that the convective stability criterion is fundamentally altered

646: in a plasma with anisotropic heat conduction (see also Chandran \&

647: Dennis 2006).  Numerical simulations of the nonlinear regime of this

648: instability (Parrish \& Stone 2005; 2007) reveal vigorous convective motions

649: that are quenched only when the plasma becomes isothermal.  Thus, inclusion

650: of magnetic fields in the dynamics of buoyant bubbles

651: alters the basic plasma dynamics in ways that warrant

652: further investigation.

653: 

654: Finally, our results have application to the optical filaments being swept

655: up by the pulsar wind in the Crab nebula (H96).  It is tempting to compare

656: the long, well-separated fingers generated with rotated fields (run R90,

657: figure 4) with the filaments.  However, due to its orientation, figure 4

658: shows the morphology of the rising bubbles of light fluid, whereas the

659: observations reveal the morphology of the descending fingers of heavy

660: fluid.  In figure 9 we plot isosurfaces of the density at $\rho=1.1$,

661: and slices along the face of the computational domain showing only regions

662: where $\rho > 1.1$, with the orientation flipped relative to figure 4,

663: that is with the descending fingers of heavy fluid oriented upward.

664: Note the long, thin fingers of dense gas along the $y-z$ plane are in

665: good agreement with the morphology of the fingers in the Crab.

666: This calculation includes field in both the heavy and light fluids,

667: although in the Crab only the light fluid (synchrotron nebula) is expected

668: to be strongly magnetized.  It is an open issue as to whether a strong

669: field in the light fluid only, whose direction changes on scales small

670: compared to $\lambda_c$, can reproduce the structures seen in figure 9.

671: Note that a uniform field in the light fluid only (run LF, see figure 3)

672: results in structure markedly different than in figure 9. 

673: 

674: Of course, to accurately model the fingers in the Crab nebula, it is

675: important to include density compression due to cooling, to study

676: fields near equipartition (which will also increase the importance

677: of compressibility), and perhaps most importantly, to include the

678: geometrical effects produced by the spherically expanding shell.

679: Previous two-dimensional studies have shown that purely hydrodynamical

680: instability in the appropriate geometry can produce the structure of the

681: Crab filaments (Jun 1998), and more recently it has shown that strong

682: fields in this same geometry significantly alter the picture (Bucciantini

683: et al 2005).  In this paper, we have emphasized the importance of three

684: dimensional effects on the magnetic RTI.  It will be important to extend

685: these fully three-dimensional results to the expanding wind geometry

686: appropriate to the Crab.

687: 

688: \acknowledgements

689: We thank Jeff Hester for discussions.  Simulations were performed on

690: the Teragrid cluster at NCSA, the IBM Blue Gene at Princeton University, and

691: on computational facilities supported by NSF grant AST-0216105.  Financial

692: support from DoE grant DE-FG52-06NA26217 is acknowledged.

693: 

694: \begin{references}

695: 

696: \reference{} Arons, J., \& Lea, 1976, ApJ, 207, 914.

697: 

698: \reference{} Balbus, S.A., 2000, ApJ, 534, 420.

699: 

700: \reference{} Bucciantini, N., Amato, E., Bandiera, R., Blondin, J.M.,

701: \& Del Zanna, L., 2004, A\& A, 423, 253.

702: 

703: \reference{} Chandrasekhar, S., 1961. {\em Hydrodynamic and Hydromagnetic

704: stability}, (Oxford, Oxford University press).

705: 

706: \reference{} Chandran, B.D., 2001, ApJ, 562, 737.

707: 

708: \reference{} Chandran, B.D., \& Dennis, T.J., 2006, ApJ, 642, 140.

709: 

710: \reference{} Dimonte, G., Youngs, D.L., Dimits, A., et al., 2004,

711: Phys. Fluids 16, 1668.

712: 

713: \reference{} Fan, Y., 2001, ApJ, 546, 509.

714: 

715: \reference{} Gardiner, T., \& Stone, J.M., 2005, JCP, 205, 509.

716: 

717: \reference{} Gardiner, T., \& Stone, J.M., 2007, JCP, submitted.

718: 

719: \reference{} Isobe, H., Miyagoshi, T., Shibata, K., \& Yokoyama, T., 2005,

720: Nature, 434, 478.

721: 

722: \reference{} Isobe, H., Miyagoshi, T., Shibata, K., \& Yokoyama, T., 2006,

723: PASJ, 58, 423.

724: 

725: \reference{} Jones, T. W., \& De Young, D. S, 2005, ApJ, 624, 586.

726: 

727: \reference{} Jun, B.-I., 1998, ApJ, 499, 282.

728: 

729: \reference{} Jun, B.-I., Norman, M.L., \& Stone, J.M., 1995, ApJ, 453, 332.

730: 

731: \reference{} Jun, B.-I., \& Norman, M.L., 1996a, ApJ, 465, 800.

732: 

733: \reference{} Jun, B.-I., \& Norman, M.L., 1996b, ApJ, 472, 245.

734: 

735: \reference{} Kim, J., Hong, S.S., Ryu, D., \& Jones, T.W., 1998, ApJ 506, L139.

736: 

737: \reference{} Kim, W.-T., Ostriker, E., \& Stone, J.M., 2002, ApJ, 581, 1080.

738: 

739: \reference{} Kosi\'{n}ski, R., \& Hanasz, M., 2007, MNRAS, 376, 861.

740: 

741: \reference{} Litwin, C., Brown, E.F., \& Rosner, R., 2001, ApJ, 553, 788.

742: 

743: \reference{} Parrish, I.J., \& Stone, J.M., 2005, ApJ, 633, 334.

744: 

745: \reference{} Parrish, I.J., \& Stone, J.M., 2007, ApJ, in press.

746: 

747: \reference{} Ruszkowski, M., En{\ss}lin, T. A., Br\"{u}ggen, M., Heinz, S., \&

748: Pfrommer, C., 2007, MNRAS, 378, 662.

749: 

750: \reference{} Robinson, K., Dursi, L.J., Ricker, P.M., et al.,

751: 2004, ApJ, 601, 621.

752: 

753: \reference{} Stone, J.M., \& Gardiner, T.A., 2007, Phys. Fluids, in press (Paper I), arXiv:0707.1022v1 [astro-ph].

754: 

755: \reference{} Wang, Y.-M., \& Nepveu, M., 1983, A\& A, 118, 267

756: 

757: \reference{} Wang, Y.-M., Nepveu, M., \& Robertson, J.A., 1984, A\& A, 135, 66

758: 

759: \reference{} Wissink, J.G., Hughes, D.W., Matthews, P.C., \& Proctor, M.R.E., 2000, MNRAS, 318, 501.

760: 

761: \end{references}

762: 

763: \clearpage

764: 

765: % Figure 1 -- color images of density

766: \begin{figure}

767: \epsscale{1.0}

768: \plotone{newfig01.ps}

769: \figcaption

770: {Isosurfaces of the density at $\rho=9.9$ and $\rho=1.1$ at times

771: $t/t_s = 20$ (left panels) and $t/t_s=40$

772: (right panels) in runs H (top, pure hydro) and

773: U (bottom, uniform field).  Also shown are slices of the density at the

774: edges of the computational domain.}

775: \end{figure}

776: 

777: % Figure 2 -- height vs time

778: \begin{figure}

779: \epsscale{0.6}

780: \plotone{fig2.ps}

781: \figcaption

782: {Height of bubbles as a function of time in hydrodynamic (run H,

783: solid line and triangles) and magnetic RTI in a uniform field

784: (run U, dotted line and squares).  Note bubbles rise {\em faster} in MHD.}

785: \end{figure}

786: 

787: % Figure 3 -- color images of density, B in light fluid

788: \begin{figure}

789: \epsscale{1.0}

790: \plotone{newfig03.ps}

791: \figcaption

792: {Isosurfaces of the density at $\rho=9.9$ and $\rho=1.1$ at times $t/t_s =

793: 20$ (left panel) and $t/t_s=40$ (right panel) in run LF (uniform field

794: in light fluid only).  Also shown are slices of the density at the edges

795: of the computational domain.}

796: \end{figure}

797: 

798: % Figure 4 -- color images of density, rotated B

799: \begin{figure}

800: \epsscale{1.0}

801: \plotone{newfig04.ps}

802: \figcaption

803: {Isosurfaces of the density at $\rho=9.9$ and $\rho=1.1$ at times $t/t_s

804: = 20$ (left panels) and $t/t_s=40$ (top right panel) or $t/t_s=50$

805: (bottom right panel) in runs R45 (top, field rotated by $45^{\circ}$)

806: and U (bottom, field rotated by $90^{\circ}$).  Also shown are slices

807: of the density at the edges of the computational domain.}

808: \end{figure}

809: 

810: % Figure 5 -- Height of bubbles C90 versus U

811: \begin{figure}

812: \epsscale{0.6}

813: \plotone{fig5.eps}

814: \figcaption

815: {Height of bubbles as a function of time in runs U (dashed line, uniform

816: field) and C90 (solid line, continuous rotation over a vertical distance 

817: $L_{rot}/\lambda_{c} = 0.5$ at the interface).  }

818: \end{figure}

819: 

820: 

821: % Figure 6 -- Height of bubbles

822: \begin{figure}

823: \epsscale{0.6}

824: \plotone{fig6.eps}

825: \figcaption

826: {Height of bubbles as a function of time in runs U (solid line, uniform

827: field), LF (long dashed line, field in light fluid only), R45 (dotted

828: line, field rotated by $45^{\circ}$) and R90 (short dashed line, field

829: rotated by $90^{\circ}$).  }

830: \end{figure}

831: 

832: % Figure 7 -- Mixing parameter

833: \begin{figure}

834: \epsscale{0.6}

835: \plotone{fig7.ps}

836: \figcaption

837: {Vertical profile of the mixing parameter, defined by equation (11),

838: for runs H (solid line), U (dotted line), and R90 (dashed line) at 

839: $t/t_s=40$.  A

840: value of zero indicates no mixing, while one indicates fully mixed.

841: The magnetic field clearly suppresses mixing, especially in the rotated

842: field case (R90).

843: }

844: \end{figure}

845: 

846: % Figure 8 -- Evolution of energies

847: \begin{figure}

848: \epsscale{0.6}

849: \plotone{fig8a.eps}

850: \plotone{fig8b.eps}

851: \figcaption

852: {Evolution of the volume averaged kinetic and magnetic energies in

853: {\em (a)} run U, uniform field case, and {\em (b)} run R90, field

854: rotated by $90^{\circ}$ at the interface.  All components of the energy

855: are normalized by the volume averaged magnetic energy in the initial

856: conditions $B_0^{2}/2$.   In addition, the energies associated with the

857: $x-$ and $y-$components of the magnetic field have their initial values

858: subtracted as appropriate.

859: }

860: \end{figure}

861: 

862: % Figure 9 -- R90 upside down

863: \begin{figure}

864: \epsscale{0.6}

865: \plotone{newfig09.ps}

866: \figcaption

867: {Isosurface of the density at $\rho=1.1$ at time $t/t_s=45$

868: in run R90 (field rotated by $90^{\circ}$).

869: Also shown are slices

870: of the density at the edges of the computational domain for $\rho>1.1$.

871: Compare to the lower RH panel in figure 4; noting the

872: orientation of the image is reversed, that is

873: descending fingers of heavy fluid point upward here.}

874: \end{figure}

875: 

876: 

877: 

878: \end{document}

879: