1: %\documentclass[12pt,preprint]{aastex}
2: %\documentclass[10pt,preprint]{aastex}
3: %\documentclass[manuscript]{aastex}
4: %\documentclass[preprint2]{aastex}
5: \documentclass[12pt,twocolumn]{emulateapj}
6: \usepackage{apjfonts}
7: % Version 08/07/2003
8:
9: %% less or similar to
10: %\def\lsim{\raisebox{-.4ex}{$\stackrel{<}{\scriptstyle \sim}$\,}}
11: \def\lsim{\lower 2pt \hbox{$\, \buildrel {\scriptstyle <}\over
12: {\scriptstyle \sim}\,$}}
13: %%
14:
15:
16: %% You can insert a short comment on the title page using the command
17: %% below.
18:
19: %\slugcomment{Not to appear in Nonlearned J., 45.}
20: %\slugcomment{Submitted to ApJ Letter}
21: %\slugcomment{Revised Auguest 7, 2003}
22: %\slugcomment{Accepted for publication in the Astrophysical Journal}
23: \slugcomment{To appear in ApJ 2004}
24:
25: \shorttitle{Formation of Semi-relativistic Jets from Magnetospheres of
26: Accreting Neutron Stars}
27: \shortauthors{Kato, Hayashi, \& Matsumoto}
28:
29: \begin{document}
30:
31: \title{Formation of Semi-relativistic Jets from Magnetospheres of
32: Accreting Neutron Stars: Injection of Hot Bubbles into a Magnetic
33: Tower}
34:
35: \author{Y. Kato\thanks{Department of Physics, Graduate School of
36: Science and Technology, Chiba University, Inage-ku, Chiba 263-8522,
37: Japan}, M. R. Hayashi\thanks{National Astronomical Observatory,
38: Mitaka, Tokyo 181-8588, Japan}, and R. Matsumoto\thanks{Department of
39: Physics, Faculty of Science, Chiba University, Inage-ku, Chiba
40: 263-8522, Japan}}
41:
42: \begin{abstract}
43: We present the results of 2.5-dimensional resistive
44: magnetohydrodynamic (MHD) simulations of the magnetic interaction
45: between a weakly magnetized neutron star and its accretion disk.
46: General relativistic effects are simulated by using the
47: pseudo-Newtonian potential. We find that well-collimated jets
48: traveling along the rotation axis of the disk are formed by the
49: following mechanism: (1) The magnetic loops connecting the neutron
50: star and the disk are twisted due to the differential rotation between
51: the neutron star and the disk. (2) Twist injection from the
52: disk initiates expansion of the loop. (3) The expanding magnetic
53: loops create a magnetic tower in which accelerated disk material
54: travel as collimated bipolar jets. The propagation speed of the
55: working surface of the jet is the order of 10\% of the speed of light
56: ($\sim 0.1c$). (4) Magnetic reconnection taking place inside the
57: expanding magnetic loops injects hot bubbles intermittently into the
58: magnetic tower. The ejection speed of the bubble is the order of the
59: local Alfv\'{e}n speed of the launching point and $\sim 0.2c$ in our
60: simulations. (5) The hot bubbles moving inside the tower catch up
61: with the working surface of the jet. High energy electrons created by
62: the magnetic reconnection are a plausible source of radio emission.
63: Our model can explain the formation process of a narrow jet from a
64: weakly magnetized ($|\mbox{\boldmath$B_{*}$}|\le 10^{9}$ gauss)
65: neutron star and the correlation between radio flares of the core and
66: of the lobe observed in Sco X-1.
67: \end{abstract}
68:
69: \keywords{accretion, accretion disks --- relativity --- MHD --- stars:
70: neutron --- ISM: jets and outflows}
71:
72: \section{INTRODUCTION}
73: The mechanism of jet formation is one of the most important subjects
74: of research in astrophysics. Blandford \& Payne (1982) studied
75: magneto-centrifugal acceleration along a magnetic field line threading
76: an accretion disk. Uchida \& Shibata (1985) and Shibata \& Uchida
77: (1986) showed by time-dependent magnetohydrodynamical (MHD)
78: simulations that bipolar jets are produced from an accretion disk
79: threaded by open magnetic field lines.
80:
81: Observational evidence indicate the presence of semi-relativistic jets
82: in some galactic X-ray binaries (XRBs) such as SS433 (Hjellming \&
83: Johnston 1981), Cygnus X-3 (Mart{\'{\i}} {\it et al.} 2000,
84: Mioduszewski {\it et al.} 2001), Sco X-1 (Fomalont {\it et al.}
85: 2001a,b), and Circinus X-1 (Fender {\it et al.} 1998). The speed of
86: jets in such XRBs indicates that the jets are launched close to the
87: last stable orbit of the accretion disk. Therefore, we should include
88: general relativistic effects around a compact object. In addition,
89: the magnetic field of the compact object may play an essential role in
90: the dynamics of the jets and the accretion flow. Recent observations
91: of XRBs by the Very Large Baseline Interferometer (VLBI) revealed
92: their energetic activities such as radio flares and radio jets. VLBI
93: discovered the motions and variabilities of radio lobes in Sco X-1
94: (Fomalont {\it et al.} 2001a,b). According to their observations,
95: radio lobes are ejected from the core of Sco X-1. After the radio
96: flare of the core, the intensity of the advancing radio lobe
97: increases. The correlation between the core flare and the lobe flare
98: in Sco X-1 indicates that explosive events at the core transports
99: energy to the lobe.
100:
101: A plausible model of advancing lobes is the working surface of jets
102: ejected from the core. However, the existence of large scale open
103: magnetic fields assumed in the magneto-centrifugal model of jet
104: formation is not evident in XRBs. Instead, we should take into
105: account the magnetic interaction between the neutron star and its
106: accretion disk.
107:
108: Lynden-Bell \& Boily (1994) studied the evolution of force-free
109: magnetic loops anchored to the star and the disk. They obtained
110: self-similar solutions for the evolution of the magnetic loops. They
111: found that the loops are unstable against the twist injection from the
112: rotating disk and that the loops expand along a direction of 60
113: degrees from the rotation axis of the disk. Lovelace {\it et al.}
114: (1995) pointed out that the dipole magnetic field of the star deforms
115: itself into an open magnetic field due to the differential rotation
116: between the star and the disk. Subsequently, Lynden-Bell (1996;
117: hereafter referred to as LB96) showed that a cylindrical magnetic
118: tower is formed when the disk is surrounded by external plasma with
119: finite pressure.
120:
121: Hayashi, Shibata, and Matsumoto (1996; hereafter referred to as HSM96)
122: carried out the first numerical calculations of magnetic interaction
123: between a protostar and its accretion disk. By assuming Newtonian
124: gravity and neglecting the rotation of the star, they showed that the
125: magnetic interaction can explain the X-ray flares and outflows
126: observed in protostars. Later, Goodson, B\"{o}hm, and Winglee;
127: Goodson \& Winglee (1999); hereafter referred to as GBW99 and GW99,
128: respectively, showed that the recurring magnetic reconnection creates
129: periodic outflow along the rotation axis of the disk. However, these
130: works have not clearly shown the formation of a magnetic tower.
131:
132: In this paper, we numerically studied the formation process of
133: magnetic towers around a weakly magnetized neutron star. In \S 2 we
134: present basic equations and models. Numerical results are shown in \S
135: 3. \S 4 is devoted for discussion and conclusion.
136:
137: \section{SIMULATION MODEL}
138: We solved resistive magnetohydrodynamic equations in cylindrical
139: coordinates. For normalization, we use the Schwarzschild radius
140: $r_{s}=2GM/c^{2}=1$ and the speed of light $c=1$. General
141: relativistic effects in the innermost region of the accretion disk
142: around the neutron star are taken into account by using the
143: pseudo-Newtonian potential (Paczy\'{n}sky \& Wiita 1980)
144: $\psi=-1/[2(R-1)]$ where $R=\sqrt{r^{2}+z^{2}}$. The basic equations
145: in conservative form are as follows:
146: %
147: \begin{equation}
148: {\partial\rho\over \partial t} +
149: \mbox{\boldmath$\nabla$}\cdot(\rho\mbox{\boldmath$v$}) = 0
150: \label{continuity}
151: \end{equation}
152: %
153: \begin{eqnarray}
154: {\partial\over\partial t}\left(\rho\mbox{\boldmath$v$}\right)
155: +\mbox{\boldmath$\nabla$}\cdot
156: \left(\rho\mbox{\boldmath$v$}\mbox{\boldmath$v$} -
157: {\mbox{\boldmath$B$}\mbox{\boldmath$B$}\over 4\pi}\right)
158: =-\mbox{\boldmath$\nabla$}\left(p+{B^{2}\over 8\pi}\right) -
159: \rho\mbox{\boldmath$\nabla$}\psi
160: \label{momentum}
161: \end{eqnarray}
162: %
163: \begin{eqnarray}
164: {\partial\over\partial t}\left(\epsilon +{B^{2}\over 8\pi}\right)
165: +\mbox{\boldmath$\nabla$}\cdot\left[\left(\epsilon +
166: p\right)\mbox{\boldmath$v$} + {\mbox{\boldmath$E$}\times
167: \mbox{\boldmath$B$}\over 4\pi}\right]
168: =-\rho\mbox{\boldmath$v$}\cdot\mbox{\boldmath$\nabla$}\psi
169: \label{energy}
170: \end{eqnarray}
171: %
172: \begin{equation}
173: {\partial \mbox{\boldmath$B$}\over \partial t} =
174: -\mbox{\boldmath$\nabla$} \times \mbox{\boldmath$E$}
175: \label{induction}
176: \end{equation}
177: where $\epsilon=\rho v^{2}/2+p/(\gamma-1)$ is the total energy of
178: the gas ($\gamma=4/3$), $\mbox{\boldmath$E$} = -\mbox{\boldmath$v$}
179: \times \mbox{\boldmath$B$} + \eta\mbox{\boldmath$J$}$ and
180: $\mbox{\boldmath$J$} = (\mbox{\boldmath$\nabla$} \times
181: \mbox{\boldmath$B$})/4\pi$. The other symbols have their usual
182: meanings.
183:
184: The resistivity $\eta$ is assumed to be uniform in the entire
185: computational domain and parametrized by the magnetic Reynolds
186: number $R_{m}=c r_{s}/\eta = 1000$, where we take the characteristic
187: speed and length as $c$ and $r_{s}$, respectively. The resistivity
188: is taken to be small enough to ensure the coupling between plasma
189: and magnetic fields. The diffusion time scale of magnetic fields
190: $\tau_{d}\equiv {\cal A}/\eta = ({\cal A}/c r_{s})R_{m}$ (${\cal A}$
191: is the area of the diffusion region) is much longer than the
192: dynamical time scale except in the localized current sheet. In this
193: paper, we assumed a uniform resistivity rather than an anomalous
194: resistivity, which incorporates plasma instabilities due to kinetic
195: effects, adopted in Hayashi {\it et al.} (1996). As we show later,
196: essential results of Hayashi {\it et al.}'s simulation such as a
197: topological change of magnetic field lines and ejection of plasmoids
198: are reproduced by numerical simulations with small uniform
199: resistivity.
200:
201: The basic equations are solved by the 2-D axisymmetric MHD code
202: (Matsumoto {\it et al.} 1996; Hayashi {\it et al.} 1996) based on a
203: modified Lax-Wendroff scheme with artificial viscosity.
204:
205: The initial rotating torus is obtained by assuming a polytropic
206: equation of state $p=K\rho^{1+1/n}$ where $n=3$ and the angular
207: momentum distribution is $l=l_{0}(r/r_{0})^{a}$ where
208: $l_{0}=r_{0}^{3/2}/[(r_{0}-1)\sqrt{2}]$ and $a=0.35$. The subscript
209: $0$ signifies the values at the density maximum of the torus
210: $r_{0}=13$ and the density is normalized at this point. The initial
211: density distribution of the torus is given by:
212: %
213: \begin{equation}
214: \rho_{t}=\rho_{0}\left(1-{\gamma\over {v^{2}_{\rm s,0}}}
215: {\tilde{\psi}-\tilde{\psi}_{0}\over n+1}\right)^{n}
216: \end{equation}
217: %
218: %\begin{equation}
219: %p_{t}=\rho_{0}{v^{2}_{\rm s,0}\over\gamma}
220: % \left({\rho_{t}\over\rho_{0}}\right)^{1+1/n}
221: %\end{equation}
222: where $\rho_{0}$, $\tilde{\psi}=\psi+(l/r)^{2}/[2(1-a)]$, and
223: $v_{\rm s,0}$ are the maximum density of the torus, the effective
224: potential, and the sound speed of the torus, respectively. The
225: thermal energy of the torus is parametrized by $E_{\rm
226: th,0}=v^{2}_{\rm s,0}/\gamma|\psi_{0}|$.
227:
228: Outside the torus, we assume a non-rotating, spherical, and isothermal
229: hot corona in hydrostatic equilibrium. The density
230: distribution of the corona is:
231: %
232: \begin{equation}
233: \rho_{c}=\rho_{*}\exp{\left[-{\gamma\over
234: v^{2}_{\rm s,*}}\left(\psi-\psi_{*}\right)\right]}
235: \end{equation}
236: %
237: %\begin{equation}
238: %p_{c}=\rho_{c}\left({v^{2}_{\rm s,*}\over\gamma}\right)
239: %\end{equation}
240: The subscript $*$ signifies the values at the surface of the neutron
241: star $R_{*}=2.8$. The thermal energy of the corona is parametrized by
242: $E_{\rm th,*}=v^{2}_{\rm s,*}/\gamma|\psi_{*}|$.
243:
244: The initial magnetic field of the neutron star is the dipole magnetic
245: field described by the toroidal component of the vector potential
246: $A_{\phi}=-m_{*} r/R^{3}$. The strength of the magnetic field is
247: defined by the magnetic dipole moment:
248: \begin{equation}
249: m_{*}=r_{*}^{3}\sqrt{{2\pi\rho_{*}\over
250: \beta_{*}}{v^{2}_{s,*}\over\gamma}}
251: \end{equation}
252: parametrized by the value of the plasma $\beta$ defined as the
253: ratio of gas pressure of the corona to magnetic pressure.
254:
255: Initial conditions are determined by choosing the non-dimensional
256: parameters $\rho_{*}/\rho_{0}=0.006$, $E_{\rm th,0}=0.006$, $E_{\rm
257: th,*}=0.21$, and $\beta_{*}=2$. We use $600\times 600$ non-uniform
258: meshes. The grid size is uniform ($\Delta r=\Delta z=0.025$) in the
259: region $0 < r < 10$ and $0 < z < 10$. Otherwise, it increases with
260: $r$ and $z$ where $r>10$ or $z>10$. The size of the entire
261: computational box is $0\leq r\leq 100$ and $0\leq z\leq 100$. At the
262: inner boundary $R=R_{*}$, we imposed an absorbing boundary condition.
263: The deviation from the initial values are artificially decreased in
264: the damping layer between $R=R_{*}$ and $R=R_{*}+\Delta R$, where
265: $\Delta R=0.2$. We imposed a symmetric boundary condition at the
266: equatorial plane and a reflecting boundary condition at the rotation
267: axis. The outer boundaries are free boundaries where waves can
268: transmit.
269:
270: \section{SIMULATION RESULTS}
271: Figure \ref{fig1:eps} shows the time evolution of poloidal magnetic
272: field lines (solid lines) and the distribution of $T=\gamma P/\rho$
273: (color contours). The arrows show the velocity vectors. In the upper
274: panels, the time interval between each panel is about one rotation
275: period $\tau_{rot} \sim 350r_{s}/c$ at $r=r_{0}$. The bottom panels
276: enlarge the innermost region.
277:
278: \begin{figure*}[t]
279: %\plotone{f1.eps}
280: \centerline{\epsfxsize=\hsize \epsfbox{f1.eps}}
281: \caption{Formation of a magnetic tower, working surface 'A', and
282: intermittent ejection of hot bubbles 'B', 'C', and 'D' and their
283: time evolution inside the magnetic tower (upper panels). The color
284: contour shows the temperature distribution and arrows show the
285: velocity vectors. The disk material is launched from the inner part
286: of the accretion disk (bottom-left panel) to a direction of 60
287: degrees from the rotation axis and collimated into the rotation
288: axis. The initial outflow velocity is about 20\% of the speed of
289: light. Hot bubbles are created inside the magnetic loops anchored
290: to the disk and the star (bottom-right panel.)}
291: \label{fig1:eps}
292: \end{figure*}
293:
294: As the magnetic field lines connecting the neutron star and the disk
295: are twisted due to the rotation of the disk, they begin to inflate in
296: the direction 30-60 degrees from the rotation axis. The magnetic
297: fields cease to splay out when the magnetic pressure balances with the
298: ambient gas pressure. Afterwards, the expanding magnetic loops form a
299: cylindrical tower of helical magnetic fields whose height increases
300: with time. In previous MHD simulations of disk-star magnetic
301: interaction (e.g., HSM96; GBW99 and GW99; Ustyugova et al. 2000), the
302: tower structure was not so prominent because the ambient gas pressure
303: was too low to confine the magnetic tower inside the computational
304: box.
305:
306: Numerical results indicate that the top interface between the tower
307: and the external medium propagates with a speed $\sim 0.1c$. This
308: propagation speed is consistent with that of the theory of magnetic
309: towers (LB96). Disk materials are accelerated and launched along the
310: magnetic tower. Inside the magnetic tower, current sheets are
311: formed because magnetic field lines extend upward from the disk and
312: then go downward to the star. Magnetic reconnection taking place
313: in the current sheet injects hot plasmoids intermittently into the
314: magnetic tower. In figure \ref{fig1:eps}a, symbols 'B', 'C', and 'D'
315: denote such plasmoids created by magnetic reconnection. Region 'A' is
316: the hot region between the top of the magnetic tower and the jet
317: terminal shock. Beside this working surface, the jet flow changes its
318: direction away from the rotation axis and creates backflows. The
319: radial interface between the magnetic tower and the ambient matter is
320: stable for the growth of the Kelvin-Helmholtz instability because
321: magnetic fields stabilize the instability.
322:
323: Figure \ref{fig2:eps}a shows the time evolution of the distribution of
324: toroidal magnetic fields. Figure 2b visualizes the three-dimensional
325: structure of magnetic field lines at {\it t}=739.5. The magnetic field
326: lines are strongly twisted due to the rotation of the disk. Note the
327: similarity between this figure and the analytical model of magnetic
328: towers (figure 2 of LB96.)
329:
330: \begin{figure*}[t]
331: %\plotone{f2.eps}
332: \centerline{\epsfxsize=\hsize \epsfbox{f2.eps}}
333: \caption{(a) Magnetic field configuration of the magnetic tower in
334: {\it r-z} plane (top panels). Solid curves show poloidal magnetic
335: field lines projected onto this plane. The color contour shows the
336: strength of toroidal magnetic fields and arrows show velocity
337: vectors. The alphabetical symbols indicate the individual magnetic
338: islands. (b) The bottom panel shows the 3-D image of the magnetic
339: field lines (solid lines) and density distribution in the equatorial
340: plane (color contour).}
341: \label{fig2:eps}
342: \end{figure*}
343:
344: Figure \ref{fig3:eps}a shows the trajectories of the working surface of
345: the magnetic tower 'A' and the bubbles 'B', 'C', and 'D'. The plasmoids
346: are ejected intermittently with interval of the rotation period around
347: $5 - 10 r_{s}$. The maximum speed of the plasmoid moving inside the
348: magnetic tower is about $0.2c$. The plasmoid 'B' catches up with the
349: working surface 'A' at $t\sim 800$ and merges after $t\sim 900$.
350:
351: Figure \ref{fig3:eps}b shows the distribution of velocity,
352: density, temperature, gas pressure, magnetic pressure, and the
353: vertical energy flux [kinetic energy flux $(\rho v^{2}/2)v_{z}$,
354: enthalpy flux $[\gamma p/(\gamma - 1)]v_{z}$, the Poynting flux
355: $(\mbox{\boldmath$E$}\times \mbox{\boldmath$B$})_{z}/4\pi$] at
356: $r=10$ (inside the magnetic tower) and $r=30$ (outside the magnetic
357: tower). The radius of the tower is about 20. The total pressure of
358: the tower slightly exceeds the ambient gas pressure. It is
359: clear that the Poynting flux dominates other fluxes inside the
360: jets.
361:
362: The density distribution has a minimum in the working
363: surface 'A'. It corresponds to the region between the contact
364: discontinuity (CD) and the jet terminal shock (JTS).
365: We should remark that the bow shock ahead of the working surface
366: is not visible because the sound speed of the ambient plasma
367: ($v_{\rm s,*} \sim 0.4$) is larger than the propagation speed of the
368: jet.
369:
370: \begin{figure*}[t]
371: %\plotone{f3.eps}
372: \centerline{\epsfxsize=\hsize \epsfbox{f3.eps}}
373: \caption{(a) The trajectories of the working surface 'A' and the
374: bubbles 'B', 'C', and 'D' in $(t,z)$ plane. The shaded
375: region indicates the working surface 'A'. At $t=900$, the plasmoid
376: 'B' collides with 'A' and merges. The speed of the bubbles is about
377: 10\% of the speed of light. The dotted line indicates 10\% of
378: the speed of light. (b) The snapshot of velocity, density,
379: temperature, gas pressure (thin), magnetic pressure (thick), and
380: distribution of energy flux inside the magnetic tower are shown from
381: top to bottom. Solid and dashed curves indicate the value at $r=10$
382: and $r=30$, respectively. A contact discontinuity (CD) locates
383: at $z=67$ and a jet terminal shock (JTS) is placed at $z=52$.}
384: \label{fig3:eps}
385: \end{figure*}
386:
387: \section{DISCUSSION}
388: In this letter, we have demonstrated that highly collimated magnetic
389: towers are formed along the rotation axis of the accretion disk
390: surrounding a weakly magnetized neutron star. We also found that hot
391: plasmoids created by the intermittent magnetic reconnections are
392: injected into the magnetic tower (Figure \ref{fig1:eps}). The
393: magnetic towers are confined to the radius where the magnetic pressure
394: of the expanding magnetic loops is comparable to the gas pressure of
395: the ambient plasma.
396:
397: Recently, {\it Poynting jet}, in which the energy and angular momentum
398: are carried predominantly by the electromagnetic field, has been
399: studied numerically by Romanova {\it et al.} (1998) and Ustyugova {\it
400: et al.} (2000). The magnetic tower jet is a Poynting jet,
401: because the Poynting flux dominates the energy flux of the bulk
402: flow inside the jet in our simulation (Figure \ref{fig3:eps}b). The
403: electromagnetic extraction of angular momentum from the disk drives
404: the accretion of the disk material. It is interesting to compare the
405: amount of angular momentum extracted electromagnetically by the jet
406: and the angular momentum transported inside the disk due to the
407: magnetic turbulence driven by the magneto-rotational instability
408: (MRI). We have to carry out three-dimensional MHD simulations to
409: determine the MRI driven angular momentum transport rate
410: self-consistently. We intend to report the results of such
411: simulations in subsequent papers. Our simulation demonstrated the
412: formation mechanism of a semi-relativistic Poynting jet around a
413: neutron star.
414:
415: We can identify the jet terminal shock (JTS) behind the working
416: surface of the jet. Our numerical results show that a hot, low
417: density region is created between the JTS and the discontinuity (CD)
418: separating the jet material and the ambient plasma. The structure
419: of the working surface of the jet is qualitatively consistent with
420: those in previous MHD simulations of jet propagation assuming
421: force-free magnetic fields initially uniform in the $z$-direction
422: (e.g., Todo et al. 1992), although the detailed structure is
423: different because our numerical simulations started with a dipole
424: magnetic field.
425:
426: We found that hot plasmoids are injected intermittently into the
427: magnetic tower. The injection speed of the blobs is the order of the
428: local Alfv\'{e}n speed of the reconnection region ($v_{A}\sim 0.2c$),
429: which is comparable to the rotation speed of the innermost region of
430: the disk. Since this speed is faster than the propagation speed of
431: the working surface of the jet, the blobs catch up with the working
432: surface and release magnetic energy. Magnetic reconnection taking
433: place in the core and in the working surface could generate high
434: energy electrons which emit synchrotron radiation.
435:
436: Let us discuss the correlation between the radio flare of the
437: core and the lobe observed in Sco X-1. Magnetic reconnection
438: taking place in the core will eject hot plasmoids, which
439: subsequently release magnetic energy again at the working surface of
440: the magnetic tower. This event can be observed as the lobe flares
441: correlated with the core flares. In our simulations, however, the
442: maximum speed of the hot plasmoids ($\sim 0.2c$) is smaller than the
443: speed estimated from the time lag between the core flare and the
444: lobe flare. This discrepancy may be resolved if the Alfv\'{e}n
445: speed inside the magnetic tower is faster than that attained in our
446: simulation. This happens when the plasma density inside the
447: magnetic tower is lower than the ambient plasma. Plasma heating due
448: to magnetic reconnection helps increasing the Alfv\'{e}n speed and
449: ejection speed of hot bubbles.
450:
451: We should point out the limitation of using non-relativistic
452: MHD. Although we take into account general relativistic effects by
453: using a pseudo-Newtonian potential, we solved Newtonian MHD
454: equations by neglecting special relativistic corrections. By
455: extending our work to fully general relativistic MHD, we will be
456: able to simulate models having Alfv\'{e}n speed closer to the light
457: speed.
458:
459: The work presented here is the first numerical simulation which
460: clearly shows the formation process of the axisymmetric magnetic
461: tower and semi-relativistic jets. We expect that
462: semi-relativistic jets observed in some XRBs are driven by the
463: magnetic interaction between the dipole magnetic field of a weakly
464: magnetized neutron star and its accretion disk. 3-D simulations
465: are necessary to study the stability of magnetic towers against
466: non-axisymmetric perturbations. We would like to report the results
467: of such simulations in the near future.
468:
469: \acknowledgments
470:
471: The authors would like to thank K. Shibata, S. Mineshige, and D. Meier
472: for valuable discussions. One of us (YK) appreciate stimulating
473: discussion with R. E. Pudritz in Les Houches Euro Summer School 2002.
474: Numerical computations were carried out on VPP5000 at the Astronomical
475: Data Analysis Center of the National Astronomical Observatory, Japan
476: (myk18b). The work is supported in part by ACT-JST of Japan Science
477: and Technology corporation and Grants-in-Aid of the Ministry of
478: Education, Culture, Sports, Science, and Technology of Japan
479: (15037202, RM).
480:
481: \begin{thebibliography}{}
482: \bibitem[Blandford & Payne (1982)]{1982MNRAS.199..883B} Blandford,
483: R.~D.~\& Payne, D.~G.\ 1982, \mnras, 199, 883
484: \bibitem[Fender et al. (1998)]{1998ApJ...506L.121F} Fender, R.,
485: Spencer, R., Tzioumis, T., Wu, K., van der Klis, M., van Paradijs,
486: J., \& Johnston, H.\ 1998, \apjl, 506, L121
487: \bibitem[Fomalont, Geldzahler, & Bradshaw (2001)]{2001ApJ...553L..27F}
488: Fomalont, E.~B., Geldzahler, B.~J., \& Bradshaw, C.~F.\ 2001a, \apjl,
489: 553, L27
490: \bibitem[Fomalont, Geldzahler, & Bradshaw (2001)]{2001ApJ...558..283F}
491: Fomalont, E.~B., Geldzahler, B.~J., \& Bradshaw, C.~F.\ 2001b, \apj,
492: 558, 283
493: \bibitem[A. P. Goodson, K. Borhm, and R. M. Winglee (1999)]{goodson99a}
494: Goodson, A. P., B\"{o}hm, K., \& Winglee, R. M. 1999, \apj, 524, 142
495: \bibitem[A. P. Goodson and R. M. Winglee (1999)]{goodson99b}
496: Goodson, A. \& Winglee, R. M. 1999, \apj, 524, 149
497: \bibitem[M. R. Hayashi, K. Shibata, and R. Matsumoto (1996)]{hayashi96}
498: Hayashi, M. R., Shibata, K., \& Matsumoto, R. 1996, \apjl, 468, L37
499: \bibitem[Hjellming & Johnston (1981)]{1981Natur.290..100H} Hjellming,
500: R.~M.~\& Johnston, K.~J.\ 1981, \nat, 290, 100
501: \bibitem[Lovelace, Romanova, &
502: Bisnovatyi-Kogan (1995)]{1995MNRAS.275..244L} Lovelace, R.~V.~E.,
503: Romanova, M.~M., \& Bisnovatyi-Kogan, G.~S.\ 1995, \mnras, 275, 244
504: \bibitem[Lynden-Bell & Boily (1994)]{1994MNRAS.267..146L} Lynden-Bell,
505: D.~\& Boily, C.\ 1994, \mnras, 267, 146
506: \bibitem[Lynden-Bell (1996)]{1996MNRAS.279..389L} Lynden-Bell, D.\
507: 1996, \mnras, 279, 389
508: \bibitem[Mart{\'{\i}}, Paredes, &
509: Peracaula (2000)]{2000ApJ...545..939M} Mart{\'{\i}}, J., Paredes,
510: J.~M., \& Peracaula, M.\ 2000, \apj, 545, 939
511: \bibitem[R. Matsumoto, Y. Uchida, S. Hirose, K. Shibata,
512: M. R. Hayashi, A. Ferrari, G. Bodo, and C. Norman (1996)]{matsumoto96}
513: Matsumoto, R., Uchida, Y., Hirose, S., Shibata, K., Hayashi, M. R.,
514: Ferrari, A., Bodo, G., \& Norman, C. 1996, \apj, 461, 115
515: \bibitem[Mioduszewski et al. (2001)]{2001ApJ...553..766M}
516: Mioduszewski, A.~J., Rupen, M.~P., Hjellming, R.~M., Pooley, G.~G.,
517: \& Waltman, E.~B.\ 2001, \apj, 553, 766
518: \bibitem[Romanova et al. (1998)]{1998ApJ...500..703R} Romanova, M.~M.,
519: Ustyugova, G.~V., Koldoba, A.~V., Chechetkin, V.~M., \& Lovelace,
520: R.~V.~E.\ 1998, \apj, 500, 703
521: \bibitem[B. Paczynsky and P. J. Wiita (1980)]{pw80} Paczy\'{n}sky,
522: B. \& Wiita, P. J. 1980, Astronomy and Astrophysics, 88, 23
523: \bibitem[Shibata & Uchida (1986)]{1986PASJ...38..631S} Shibata, K.~\&
524: Uchida, Y.\ 1986, \pasj, 38, 631
525: \bibitem[Todo, Uchida, Sato, & Rosner (1992)]{1992PASJ...44..245T}
526: Todo, Y., Uchida, Y., Sato, T., \& Rosner, R.\ 1992, \pasj, 44, 245
527: \bibitem[Uchida & Shibata (1985)]{1985PASJ...37..515U} Uchida, Y.~\&
528: Shibata, K.\ 1985, \pasj, 37, 515
529: \bibitem[Ustyugova et al. (2000)]{2000ApJ...541L..21U} Ustyugova,
530: G.~V., Lovelace, R.~V.~E., Romanova, M.~M., Li, H., \& Colgate,
531: S.~A.\ 2000, \apjl, 541, L21
532: \end{thebibliography}
533:
534: \clearpage
535:
536: %% Use the figure environment and \plotone or \plottwo to include
537: %% figures and captions in your electronic submission.
538:
539: \end{document}
540: