1: %% v150108_ms.tex from HJV on 15 01 2008
2: %% From VZ on Jan. 11, 08
3: %% Modified by HJV January 8, 2008
4: %%
5: %% Beginning of file 'sample.tex'
6: %%
7: %% Modified 2004 January 9
8: %%
9: %% This is a sample manuscript marked up using the
10: %% AASTeX v5.x LaTeX 2e macros.
11:
12: %% The first piece of markup in an AASTeX v5.x document
13: %% is the \documentclass command. LaTeX will ignore
14: %% any data that comes before this command.
15:
16: %% The command below calls the preprint style
17: %% which will produce a one-column, single-spaced document.
18: %% Examples of commands for other substyles follow. Use
19: %% whichever is most appropriate for your purposes.
20: %%
21: %%\documentclass[12pt,preprint]{aastex}
22:
23: %% manuscript produces a one-column, double-spaced document:
24:
25: %\documentclass[manuscript]{aastex}
26:
27: %% preprint2 produces a double-column, single-spaced document:
28:
29: \documentclass[preprint2]{aastex}
30:
31: %% Sometimes a paper's abstract is too long to fit on the
32: %% title page in preprint2 mode. When that is the case,
33: %% use the longabstract style option.
34:
35: %% \documentclass[preprint2,longabstract]{aastex}
36:
37: %% If you want to create your own macros, you can do so
38: %% using \newcommand. Your macros should appear before
39: %% the \begin{document} command.
40: %%
41: %% If you are submitting to a journal that translates manuscripts
42: %% into SGML, you need to follow certain guidelines when preparing
43: %% your macros. See the AASTeX v5.x Author Guide
44: %% for information.
45:
46: %% You can insert a short comment on the title page using the command below.
47:
48: \slugcomment{To appear in ApJ}
49:
50: %% If you wish, you may supply running head information, although
51: %% this information may be modified by the editorial offices.
52: %% The left head contains a list of authors,
53: %% usually a maximum of three (otherwise use et al.). The right
54: %% head is a modified title of up to roughly 44 characters.
55: %% Running heads will not print in the manuscript style.
56:
57: \shorttitle{Modeling Bell's Non-resonant Cosmic Ray Instability}
58: \shortauthors{V.N. Zirakashvili, V.S.Ptuskin and H.J.V\"olk}
59:
60: %% This is the end of the preamble. Indicate the beginning of the
61: %% paper itself with \begin{document}.
62:
63: \begin{document}
64:
65: %% LaTeX will automatically break titles if they run longer than
66: %% one line. However, you may use \\ to force a line break if
67: %% you desire.
68:
69: \title{\bf Modeling Bell's
70: Non-resonant Cosmic Ray Instability}
71:
72: %% Use \author, \affil, and the \and command to format
73: %% author and affiliation information.
74: %% Note that \email has replaced the old \authoremail command
75: %% from AASTeX v4.0. You can use \email to mark an email address
76: %% anywhere in the paper, not just in the front matter.
77: %% As in the title, use \\ to force line breaks.
78:
79: \author{V.N.Zirakashvili}
80: \affil{Pushkov Institute for Terrestrial Magnetism, Ionosphere and Radiowave
81: Propagation, 142190, Troitsk, Moscow Region, Russia}
82: \affil{Max-Planck-Institut f\"{u}r\ Kernphysik, Postfach 103980, 69029
83: Heidelberg, Germany}
84: %\email{zirak@mpimail.mpi-hd.mpg.de}
85: \author{V.S.Ptuskin}
86: \affil{Pushkov Institute for Terrestrial Magnetism, Ionosphere and Radiowave
87: Propagation, 142190, Troitsk, Moscow Region, Russia}
88: \and
89: \author{H.J.V\"olk}
90: \affil{Max-Planck-Institut f\"{u}r\ Kernphysik, Postfach 103980, 69029
91: Heidelberg, Germany}
92:
93: %% Notice that each of these authors has alternate affiliations, which
94: %% are identified by the \altaffilmark after each name. Specify alternate
95: %% affiliation information with \altaffiltext, with one command per each
96: %% affiliation.
97:
98: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
99: %% style, abstract will output a Received/Accepted line after the
100: %% title and affiliation information. No date will appear since the author
101: %% does not have this information. The dates will be filled in by the
102: %% editorial office after submission.
103:
104: \begin{abstract}
105: We have studied
106: the non-resonant streaming instability of charged energetic particles moving
107: through a background plasma, discovered by Bell
108: \cite{bell04}. We confirm his numerical results regarding a
109: significant magnetic field amplification in the system. A detailed physical
110: picture of the instability
111: development and of the magnetic field evolution is given.
112: \end{abstract}
113:
114: %% Keywords should appear after the \end{abstract} command. The uncommented
115: %% example has been keyed in ApJ style. See the instructions to authors
116: %% for the journal to which you are submitting your paper to determine
117: %% what keyword punctuation is appropriate.
118:
119: %% Authors who wish to have the most important objects in their paper
120: %% linked in the electronic edition to a data center may do so in the
121: %% subject header. Objects should be in the appropriate "individual"
122: %% headers (e.g. quasars: individual, stars: individual, etc.) with the
123: %% additional provision that the total number of headers, including each
124: %% individual object, not exceed six. The \objectname{} macro, and its
125: %% alias \object{}, is used to mark each object. The macro takes the object
126: %% name as its primary argument. This name will appear in the paper
127: %% and serve as the link's anchor in the electronic edition if the name
128: %% is recognized by the data centers. The macro also takes an optional
129: %% argument in parentheses in cases where the data center identification
130: %% differs from what is to be printed in the paper.
131:
132: \keywords{cosmic rays--
133: acceleration--
134: instabilities}
135:
136: %% From the front matter, we move on to the body of the paper.
137: %% In the first two sections, notice the use of the natbib \citep
138: %% and \citet commands to identify citations. The citations are
139: %% tied to the reference list via symbolic KEYs. The KEY corresponds
140: %% to the KEY in the \bibitem in the reference list below. We have
141: %% chosen the first three characters of the first author's name plus
142: %% the last two numeral of the year of publication as our KEY for
143: %% each reference.
144:
145: \section{Introduction}
146:
147: The diffusive shock acceleration process (Krymsky
148: \cite{krymsky77}; Axford et al. \cite{axford77}; Bell
149: \cite{bell78}; Blandford and Ostriker \cite{blandford78}) is
150: considered as the principal mechanism for the production of the
151: galactic cosmic rays in supernova remnants (SNRs). A great
152: strength of the random magnetic fields that provide the scattering
153: of energetic particles both upstream and downstream of a supernova
154: shock is necessary for efficient acceleration. It was originally
155: suggested that this may be the result of a gyroresonant streaming
156: instability that develops due to the presence of a diffusive
157: streaming of accelerated particles (Bell \cite{bell78}, Blandford
158: \& Ostriker \cite{blandford78}). The corresponding
159: magnetohydrodynamic (MHD) waves have wavelengths of the order of
160: the gyroradii of the energetic particles. This kind of resonant
161: cyclotron instability had been suggested earlier to regulate the
162: propagation of the galactic cosmic rays (Lerche \cite{lerche67},
163: Wentzel \cite{wentzel74}). If the generated random fields are
164: amplified up to wave amplitudes that correspond to the strength
165: of the mean magnetic field, the scattering mean free path of
166: resonant energetic particles will decrease to a value comparable
167: with their gyroradius. This regime of diffusion is called Bohm
168: diffusion and is often used to estimate the maximum energy of the
169: accelerated particles.
170:
171: However, even this rather optimistic regime of diffusion may ensure the
172: acceleration of cosmic ray protons only up to energies of about $10^{14}$ eV
173: (Lagage \& Cesarsky \cite{lagage83}, Berezhko \cite{ber96}) if the magnetic
174: field in the remnant is comparable with the interstellar magnetic field, which
175: is typically less than $10$ $\mu $G. Significant magnetic field amplification
176: is necessary for acceleration up to higher energies. Amplification might be
177: possible since quasilinear theory of the resonant streaming instability allows
178: it (see e.g. McKenzie \& V\"olk \cite{mckenzie82}). Since this
179: perturbation theory breaks down in the case of resonant interaction of
180: particles with high amplitude MHD waves, Lucek and Bell \cite{lucek00}
181: performed MHD simulations combined with calculations of energetic particle
182: trajectories and found that the magnetic field can be amplified to a level
183: where the random field exceeds the initial mean field. The necessity to
184: perform detailed trajectory calculations leads to strong numerical limitations
185: of such simulations.
186:
187: Different qualitative treatments of magnetic field amplification
188: at supernova shocks were suggested on an analytical level (see
189: e.g. Bell \& Lucek \cite{bell01}, Ptuskin \& Zirakashvili
190: \cite{ptuskin03}, Vladimirov et al. \cite{vladimirov06}, Amato \&
191: Blasi \cite{amato06}). Using the observed synchrotron emission,
192: field amplification was included on a phenomenological level in
193: numerical solutions of the coupled gas dynamic and particle
194: acceleration equations (e.g. Berezhko et al. \cite{berezhko02},
195: Berezhko \& V\"olk \cite{berezhko04a, berezhko04b}, V\"olk et al.
196: \cite{voelk07}). It became clear that such amplification can lead
197: to particle accelaration to knee energies at $\approx 3\cdot
198: 10^{15}$ eV and -- according to speculative extrapolations --
199: possibly even beyond.
200:
201: On the basis of the dispersion relation for collisionless MHD waves derived by
202: Achterberg \cite{achterberg83}, Bell \cite{bell04} found a non-resonant
203: streaming instability that had been overlooked before. He
204: argued that the diffusive streaming of particles accelerated at supernova
205: shocks may be so strong that it modifies the dispersion relation of MHD waves
206: in an essential way. Then a non-oscillatory purely
207: growing MHD mode appears at scales smaller than the gyroradius of the
208: particles which excite the instability. The growth rate of this mode can be
209: larger than the growth rate of the resonant mode. Since the particle
210: trajectories are only weakly deflected by small-scale magnetic
211: inhomogeneities, one can avoid complicated trajectory
212: calculations. Only the mean cosmic-ray flux needs to be specified. Bell \cite
213: {bell04} performed corresponding MHD simulations and showed that the magnetic
214: field can indeed be significantly amplified.
215:
216: In the following we investigate this non-resonant instability in considerably
217: more detail. We have performed MHD simulations of the instability with very
218: good numerical resolution and have found a simplified
219: analytical description for the magnetic field amplification. Our results may be
220: applied to different astrophysical objects, where energetic particles exist.
221:
222: The present work deals with the instability in the presence of a
223: non-resonant energetic particle population streaming through the thermal
224: plasma. And it simulates the nonlinear instability development. In
225: a companion paper (Zirakashvili \& Ptuskin \cite{zirakashvili08},
226: Paper II) this modeling will be combined with
227: an analytical treatment of the diffusive acceleration of
228: particles in a plane, steady shock with its
229: intrinsically non-uniform energetic particle distribution in the precursor.
230:
231: The paper is organized as follows: The basic equations are given
232: in the next section. The scattering of cosmic ray particles by
233: small-scale random magnetic fields is considered in Sect.3.
234: The cosmic ray electric current is calculated in Sect. 4. The
235: non-resonant instability and the MHD simulations are described in
236: Sect.5 and 6. The summary is given in the last Section.
237:
238: \section{Basic equations}
239:
240: We consider a system that consists of a thermal plasma and a cosmic ray gas
241: with a negligible mass density. We shall treat the gas
242: as a magnetized fluid with frozen-in magnetic field ${\bf B}$ and induced
243: electric field ${\bf E=-[u\times B]}/c$. Here ${\bf u}$ is the mass velocity
244: (essentially equal to the thermal gas velocity) and $c$ is the velocity of
245: light. The electric charge density $\rho _\mathrm{th}$ and electric current
246: density ${\bf j}_\mathrm{th}$ of the thermal plasma that determine the Lorentz
247: force ${\bf F}=[({\bf j}_\mathrm{th}-\rho _\mathrm{th}{\bf u})\times {\bf
248: B}]/c$ may be found from the quasi-neutrality condition $\rho
249: _\mathrm{th}=-\rho _\mathrm{cr}$ and Amp\`ere's law $[\nabla \times {\bf
250: B}]=4\pi ({\bf j}_\mathrm{th}+{\bf j}_\mathrm{cr})/c$, respectively. Here $\rho
251: _\mathrm{cr}$ and ${\bf j}_\mathrm{cr}$ are the electric charge density and
252: electric current density of the cosmic ray gas, respectively. The Euler
253: equation of the gas motion may then be written as
254:
255: \[
256: \rho \left( \frac {\partial {\bf u}}{\partial t}+({\bf u\nabla )u}\right) =
257: -\frac {[\bf B\times [\nabla \times B]]}{4\pi }-\nabla P
258: \]
259: \begin{equation}
260: -\frac 1c[({\bf j}_\mathrm{cr}-\rho _\mathrm{cr}{\bf u})\times {\bf B}],
261: \end{equation}
262: where $P$ is the thermal gas pressure.
263: %The meaning of the last term in this equation can
264: %be understood from the following arguments. The trajectory of a
265: %cosmic ray particle is deflected in the electromagnetic fields of
266: %the thermal plasma. The corresponding oppositely directed force acts
267: %on the plasma. The force produced by all cosmic ray particles is
268: %reduced to the last term of this equation.
269:
270: The evolution of the mass density $\rho $ and the magnetic field ${\bf B}$ are
271: governed by the continuity equation
272: \begin{equation}
273: \frac {\partial \rho }{\partial t}+\nabla (\rho {\bf u})=0
274: \end{equation}
275: and Faraday's law
276:
277: \begin{equation}
278: \frac {\partial {\bf B} }{\partial t}=[\nabla \times [{\bf u}\times {\bf B}].
279: \end{equation}
280: The equation for the plasma energy density
281: $e=(\rho {u^2}/2)+(B^2/(8\pi ))+P/(\gamma -1)$
282: may be written as:
283: \[
284: \frac {\partial e }{\partial t}+\nabla \left( \rho {\bf u} \frac {u^2}{2}
285: +\frac {\gamma P {\bf u}}{\gamma -1} +\frac {[{\bf B\times [u\times B]}}{4\pi
286: }\right)
287: \]
288: \begin{equation}
289: =-\frac {\bf u}{c}[{\bf j}_\mathrm{cr}\times {\bf B}].
290: \end{equation}
291: Here $\gamma $ is the adiabatic index of the gas. The last term in this
292: equation describes the mechanical work produced by cosmic rays. The electric
293: charge density $\rho _\mathrm{cr}$ and the electric current density of cosmic
294: rays ${\bf j}_\mathrm{cr}$ may be found from the momentum distribution of
295: cosmic rays $f({\bf p,r},t)$. It obeys the equation
296: \begin{equation}
297: \frac {\partial f }{\partial t}+{\bf v}\nabla f+\frac qc
298: [({\bf v}-{\bf u})\times {\bf B}]\frac {\partial f }{\partial {\bf p}}=0.
299: \end{equation}
300: Here $q$ is the charge of cosmic ray particles.
301:
302: The system of Eqs. (1)-(5) describes the interaction of cosmic rays and
303: magnetized thermal plasma. We shall use this system in the next sections.
304:
305: \section{Scattering by the small-scale field}
306:
307: The cosmic ray momentum distribution may be written as $f=f_0+\delta f$. Here
308: $f_0=\left< f\right> $ is the momentum distribution averaged over the
309: fluctuations of the magnetic field $\delta {\bf B}$ and of the plasma velocity
310: $\delta {\bf u}$; $\delta f$ is the fluctuation of the momentum
311: distribution. One can use perturbation theory for the calculation of $\delta f$
312: when the fluctuations of the magnetic field are small in comparison with the
313: mean field ${\bf B}_0$. Since we are interested in the investigation of
314: considerable magnetic field amplification, the theory of perturbations with
315: small magnetic field amplitude is, generally speaking, not
316: applicable. Fortunately in the case of the non-resonant instability that is
317: most interesting for the present consideration, only small-scale fields are
318: generated (see also below) and the scale of the random field is smaller than
319: the particle gyroradius in the total magnetic field. Perturbation theory is
320: applicable in this case because the particles are only weakly deflected on the
321: characteristic scale of the random field. The theory of cosmic ray diffusion
322: in such magnetic fields was developed by Dolginov and Toptygin
323: \cite{dolginov67}.
324:
325: The equation for the fluctuation of the momentum distribution can be found from
326: Eq. (5):
327:
328: \[
329: \frac {\partial \delta f }{\partial t}+{\bf v}\nabla \delta f+\frac qc [({\bf
330: v}-{\bf u}_0)\times {\bf B}_0]\frac {\partial \delta f }{\partial {\bf p}}
331: \]
332: \begin{equation}
333: =-\frac qc [({\bf v}-{\bf u}_0)\times {\bf \delta B}]\frac {\partial f_0
334: }{\partial {\bf p}}.
335: \end{equation}
336: Here ${\bf u}_0$ is the mean mass velocity and we neglect
337: the velocity perturbations $\delta {\bf u}$. It is assumed that they
338: are small in comparison with
339: the mean velocity ${\bf u}_0$.
340: Such perturbations
341: would only result in
342: second order Fermi acceleration which is
343: a factor of $\left< \delta u^2\right> /c^2$ slower
344: in comparison with particle scattering.
345:
346: Since the random field ${\bf \delta B}$ has small spatial scales, the last
347: term on the left-hand side of this equation may be
348: also neglected. This means that to lowest order the particles
349: can be considered to move in straight-line orbits. The
350: calculations are
351: in this case significantly simplified in comparison with the
352: general case, when the integration along the helical orbits of
353: particles results in the appearance of series containing
354: Bessel functions (see e.g. Berezinskii et al.
355: \cite{berezinsky90}). Although the zero-order orbits are
356: different for these two methods, they give similar results in the
357: case of almost any small-scale random magnetic field. The formal
358: limit $B_0\to 0$ is not trivial and was considered by Tsytovich
359: \cite{tsytovich77}.
360:
361: As long as the gradient lengths of $f_0$ and ${\bf u}_0$ are large compared to
362: the spatial fluctuation scale $k^{-1}$, the fluctuation amplitudes
363: carry a corresponding parametrical spatial dependence. We shall assume this
364: in the following. Fourier transforming then in time and space we find the
365: expression for the Fourier transform of the distribution $\delta f_{\omega
366: ,{\bf k}}=\int d^3rdt \delta f\exp (i(\omega t-{\bf kr}))(2\pi )^{-4}$:
367:
368: \begin{equation}
369: \delta f_{\omega ,{\bf k}}=\frac {q/c}{i(\omega -{\bf kv})} [({\bf v}-{\bf
370: u}_0)\times {\bf \delta B_{\omega ,k}}]\frac {\partial f_0 }{\partial {\bf p}}.
371: \end{equation}
372:
373: Here ${\bf \delta B_{\omega ,k}}$ is the Fourier transform of the random
374: magnetic field. We shall further assume that the
375: magnetic field changes slowly in time in the frame moving with the mean
376: mass velocity ${\bf u}_0$ and therefore write ${\bf \delta B_{\omega
377: ,k}}=\delta {\bf B_k}\delta (\omega -{\bf ku}_0)$.
378:
379: Averaging now Eq. (5) and using Eq.(7) we
380: obtain the following kinetic equation for the average cosmic ray
381: distribution function $f_0$:
382:
383: \[
384: \frac {\partial f_0 }{\partial t}+{\bf v}\nabla f_0+q\left( {\bf E}_0+\frac 1c
385: [({\bf v}-{\bf u}_0)\times {\bf B}_0]\right) \frac {\partial f_0 }{\partial
386: {\bf p}}
387: \]
388:
389: \begin{equation}
390: =\frac \partial {\partial p_i}\nu _{ij}\frac {\partial f_0}{\partial p_j}
391: \end{equation}
392: Here ${\bf E}_0=-c^{-1}\left< \bf \delta u\times \delta B\right> $ is the mean
393: electric field in the frame of reference moving with the mean plasma
394: velocity ${\bf u}_0$.
395:
396: The scattering tensor $\nu _{ij}$ in the last equation is determined by the
397: spectrum of the random magnetic field $B_{ij}({\bf k})=\int \frac {d^3r}{(2\pi
398: )^3}\left< B_i{\bf (r+r}_0)B_j{\bf (r}_0)\right> \exp (-i{\bf kr)}$:
399:
400: \[
401: \nu _{ij}=\frac {q^2}{c^2}\pi \int d^3k\delta ({\bf
402: k(v-u}_0))e_{ilm}e_{jrs}
403: \]
404: \begin{equation}
405: \times B_{lr}({\bf k})
406: (v_m-u_{0m})(v_s-u_{0s})
407: \end{equation}
408: Here $e_{ijk}$ is the antisymmetric tensor. The scattering tensor makes the
409: cosmic ray distribution isotropic in the frame moving with the velocity ${\bf
410: u}_0$. Expression (9) may be simplified in the case
411: of an isotropic random magnetic field $B_{ij}({\bf k})=\frac
412: 12B_{\mathrm{isotr}}(k)(\delta _{ij}-k_ik_j/k^2)$ (Dolginov \& Toptygin
413: \cite{dolginov67}):
414: \begin{equation}
415: \nu _{ij}=p^2\nu (p) \frac
416: {({\bf v-u_0})^2\delta _{ij}-(v_i-u_{0i})(v_j-u_{0j})}
417: {2v|{\bf v-u}_0|}
418: \end{equation}
419: where the scattering frequency $\nu (p)$ is given by the formula
420: \begin{equation}
421: \nu (p)=\frac \pi 4\frac {q^2v}{p^2c^2}\int d^3k B_{\mathrm{isotr}}(k)/k
422: \end{equation}
423: Here the spectrum of the isotropic magnetic field is normalized as $\left< \delta
424: B^2\right> =\int d^3kB_{\mathrm{isotr}}(k)$.
425:
426: In the diffusion approximation the average cosmic ray
427: distribution function $f_0$ may be written as
428:
429: \begin{equation}
430: f_0({\bf p})=N(p)+\frac 3{pv}{\bf pJ}(p),
431: \end{equation}
432: where $N(p)$ is the isotropic part and ${\bf J}(p)$ is the cosmic ray flux density,
433: which is the sum of the diffusive and advective flux densities and is given by
434: Eq. (A.3) in Appendix A.
435:
436: \section{Calculation of the electric current}
437:
438: We shall use Eq. (7) for the calculation of the electric current density of
439: the cosmic ray gas. Substitution of the expression
440: (12) into Eq. (7), multiplication by $q{\bf v}$, and integration over momentum
441: space give the fluctuating part of the Fourier transform of the cosmic ray
442: electric current $\delta {\bf j}_\mathrm{cr, \omega ,{\bf k}}$. Disregarding
443: terms of the order $u_0/v$ we obtain
444:
445: \begin{equation}
446: \delta {\bf j}_\mathrm{cr, \omega ,{\bf k}}=
447: 3\pi \frac {q^2}c\delta (\omega -{\bf ku}_0)\int \frac {d^3p}{pv}\delta ({\bf
448: kv)(J}_\mathrm{d}[{\bf v\times \delta B_k])v}.
449: \end{equation}
450: Here ${\bf J}_\mathrm{d}={\bf J+u}_0\frac p3\frac {\partial
451: N}{\partial p}$ is the average cosmic ray diffusion flux. The
452: appearance of the $\delta $-function in this equation is due
453: to the Landau resonance $\omega ={\bf kv}$ (cf. Lifshitz and
454: Pitaevskii \cite{lifshitz81}) in Eq.(7).
455:
456: The total electric current of the cosmic ray gas is
457: the flux {$\bf J$} multiplied by the particle charge $q$.
458:
459: Performing the integration on the two angles in momentum space,
460: and
461: calculating the inverse Fourier transformation of the last
462: equation, we obtain the expression for the diffusive electric current ${\bf
463: j}_\mathrm{d}={\bf j}_\mathrm{cr}-\rho _\mathrm{cr}{\bf u}$ that appears in
464: the right-hand side of Eq. (1):
465:
466: \[
467: {\bf j}_\mathrm{d}=q\int d^3p{\bf J}_\mathrm{d}+
468: \frac {3\pi}4 \frac {q^2}c\int d^3k\exp (i{\bf k(r-u}_0t))
469: \]
470: \begin{equation}
471: \times \int \frac {d^3p}{pk^3} \left( k^2{\bf [\delta B_k\times J}_\mathrm{d}]-{\bf
472: k(k[\delta B_k\times J}_\mathrm{d}])\right) .
473: \end{equation}
474:
475: The first term on the right-hand-side of this equation is simply
476: the zero-order term of the expansion in the magnetic fluctuation
477: $\delta {\bf B}$, while the second term is the linear term of the
478: expansion. The latter is always smaller than the first term if the
479: small-scale field approximation is valid, that is if $qB/pck<<1$,
480: which means that the particle gyroradius $pc/qB$ is large compared
481: to the scale $k^{-1}$ of the magnetic field.
482:
483: %The first integral is always larger than the second one when this
484: %condition is fulfilled. This justifies Bell's MHD modeling of magnetic field
485: %amplification \cite{bell04}, where only the first term from the right-hand side
486: %of Eq. (14) was taken into account.
487:
488: However, in some cases the second integral may play a r\^ole. When the cosmic
489: ray streaming is not strong, it results only in a small change of the
490: dispersion relation of MHD waves. The first integral in Eq. (14) in this case
491: produces only a small shift of the frequency of MHD waves. The second term then
492: gives a small imaginary part of the frequency and describes a resonant wave
493: instability based on the Landau resonance in Eq. (7). Within the limits of the
494: small-scale field approximation, and for strong cosmic ray streaming,
495: this resonant instability is ineffective in comparison to the well-known
496: gyroresonant streaming instability.
497:
498: Another important point is that the second integral on the right-hand side of
499: Eq. (14) appears in the calculation of the mean force acting on the thermal
500: plasma (Ptuskin \cite{ptuskin84}):
501: ${\bf F}=c^{-1}\left< [{\bf B\times (\bf j}_\mathrm{cr}-\rho _\mathrm {cr}{\bf
502: u}_0)]\right> $.
503:
504: Using Eq. (14) and averaging we obtain
505:
506: \[
507: {\bf F}=\frac \pi 4\frac {q^2}{c^2}\int \frac {d^3p}{p}{\bf J}_d\int d^3k B_{\mathrm{isotr}}(k)/k
508: \]
509: \begin{equation}
510: =\int {d^3p}\frac {p}v\nu (p){\bf J}_d
511: \end{equation}
512: It was assumed here that the random field is isotropic and
513: expression (11) was used. Since the diffusive flux is equal to
514: ${\bf J}_d=-\frac {v^2}{3\nu }\nabla N$ in the diffusion
515: approximation, ${\bf F}=-\nabla P_\mathrm{cr}$, where
516: $P_\mathrm{cr}=\int d^3p pvN(p)/3$ is the cosmic ray pressure. In
517: the sense of MHD theory, Eq. (1) then describes the overall
518: momentum balance of the system, where the forces on the r.h.s. are
519: the Lorentz force and the gradient of the overall pressure,
520: thermal plus nonthermal. Indeed $c^{-1}\left< [({\bf
521: j}_\mathrm{cr}-\rho _\mathrm{cr}{\bf u}_0) \times {\bf B}]\right>
522: = -\nabla P_\mathrm{cr}$ is the general form of the average
523: cosmic ray momentum balance.
524:
525: In the sequel we will neglect the 2nd term of Eq. (14) (see also Bell
526: \cite{bell04}). This is justified in the linear analysis and in the
527: nonlinear simulation of the non-resonant instability if the small-scale field
528: approximation is valid.
529:
530: \begin{figure}[t]
531: \includegraphics[width=7.5cm]{f1.eps}
532: \caption{Explanation of the non-resonant streaming instability. The magnetic
533: spiral (thin solid line) is stretched by the Lorentz force ${\bf F}=-[{\bf
534: j}_\mathrm{d}\times {\bf B}]c^{-1}$ that appears due to
535: the diffusive cosmic ray electric current ${\bf j}_\mathrm{d}$.}
536: \end{figure}
537:
538: \section{Non-resonant streaming instability}
539:
540: The dispersion relation for small-scale MHD perturbations may be
541: found from Eqs. (1)-(5). In the case when the
542: cosmic ray diffusion flux ${\bf J}_\mathrm{d}$ and the wavenumber
543: ${\bf k}$ are parallel to the mean magnetic field ${\bf B}_0$ this
544: dispersion relation may be written as (Bell \cite{bell04}):
545: \begin{equation}
546: (\omega -{\bf ku}_0)^2=V^2_\mathrm{a}k^2\mp \frac {j_\mathrm{d}B_0}{c\rho _0}k.
547: \end{equation}
548: Here $\rho _0$ is the mean plasma density, $V_\mathrm{a}=B_0/\sqrt{4\pi \rho
549: _0}$ is the Alfv\'en velocity,
550: ${\bf j}_\mathrm{d}=q\int d^3p{\bf J}_\mathrm{d}$ is the average
551: diffusive electric current of cosmic rays,
552: and the two signs $\mp $ correspond to the two
553: circular polarizations. A non-resonantly unstable MHD mode appears if the
554: condition $k< k_\mathrm{c}=j_\mathrm{d}B_0/c\rho _0V^2_\mathrm{a}$ is
555: fulfilled. The unstable magnetic field line spiral expands in the direction
556: perpendicular to the mean magnetic field (see Fig.1). The mode with
557: $k=k_\mathrm{c}/2$ has the maximum growth rate $\gamma _{\max }$ :
558:
559: \begin{equation}
560: \gamma _{\max }=\frac {j_\mathrm{d}B_0}{2c\rho _0V_\mathrm{a}},
561: \end{equation}
562: which does not depend on the magnetic field strength.
563:
564: Since we assumed that the scale $k^{-1}$ of the perturbations is smaller than
565: the gyroradius {\bf $pc/qB_0$} of the energetic particles,
566: the wavenumber $k$ should obey the condition $qB_0/pc<< k_\mathrm{c}$. This
567: means that the necessary condition for instability is
568: $k_\mathrm{c}> qB_0/pc$. This condition may be rewritten as
569: \begin{equation}
570: \frac {u_\mathrm{cr}}{v}\frac {\epsilon _\mathrm{cr}}{B^2_0/4\pi }>> 1
571: \end{equation}
572: Here $u_\mathrm{cr}$ and $\epsilon _\mathrm{cr}$ are the bulk velocity and the
573: energy density of the cosmic ray gas, respectively, and $v$ is the velocity of
574: energetic particles. This condition is easily fulfilled at the shocks of SNRs,
575: where $u_\mathrm{cr}$ is of the order of the shock velocity $u_1$ and the
576: energy density of the relativistic particles may be comparable with $\rho
577: u^2_1$.
578:
579:
580: \begin{figure}
581: \includegraphics[width=7.5cm]{f2.eps}
582: \caption{ Numerical results of the modeling of the non-resonant instability
583: with the dimensionless cosmic ray electric current $J=16$ as a function of
584: normalized time $\tilde{t}$. The r.m.s. values of the magnetic field
585: fluctuation $\left< \delta B^2\right> ^{1/2}$ divided by the mean magnetic
586: field strength $B_0$ and of the mass velocity fluctuations $\left< \delta
587: u^2\right> ^{1/2}$ divided by the Alfv\'en velocity $V_a$ are shown by the {\it
588: solid} and {\it dashed lines}, respectively. The ratio of the sonic velocity
589: $c_\mathrm{s}$ and the Alfv\'en velocity $V_a$ is shown by the {\it dotted
590: line}. The square root of the ratio of the mean electric field $E_0$ and the
591: mean magnetic field $B_0$ multiplied on $(c/V_a)^{1/2}$ is shown by the {\it
592: thin solid line}. The characteristic scale $L_m$ of the magnetic field is also
593: shown by the {\it thin dashed line}.
594: %For the normalizations, see text.
595: }
596: %The ratio of rms of magnetic fluctuations in
597: %the parallel and perpendicular to the mean magnetic field
598: %direction is
599: %also shown by the dotted line
600:
601: \end{figure}
602:
603:
604: \begin{figure}
605: \includegraphics[width=7.6cm]{f3.eps}
606: \caption{Slices of the magnetic field strength through the center of the box,
607: obtained at $\tilde{t}=0.82$ (top), $\tilde{t}=1.03$ (middle) and
608: $\tilde{t}=1.48$ (bottom), in the run with $J=16$. The scaling of the magnetic
609: field strength is logarithmic between $0.32\left< B^2\right> ^{1/2}$ (white) and
610: $3.2\left< B^2\right> ^{1/2}$ (black). The mean
611: magnetic field and the diffusive electric current are in the $z$ direction.}
612: \end{figure}
613:
614: \section{Numerical modeling of the non-resonant instability}
615:
616: We have numerically modeled
617: the non-resonant instability similar in spirit to the modeling of
618: Bell \cite{bell04}. The MHD Eqs. (1)-(4), written
619: in dimensionless form, were solved numerically. We used the numerical method of
620: Pen et al. \cite{pen03}. It is a second order in
621: space and time, flux-conservative total variation diminishing
622: MHD scheme which enforces the $\nabla {\bf B}=0$ constraint to
623: machine precision. The nonlinear flux limiter "minmod" was used.
624:
625: \begin{figure}[t]
626: \includegraphics[width=8.0cm]{f4.eps}
627: \caption{The one-dimensional spectra of the perpendicular component of
628: the random magnetic field $B_\perp (\tilde{k}_z)$ obtained at $\tilde{t}=0$ (thin solid
629: line), $\tilde{t}=0.54$ (solid
630: line), $\tilde{t}=0.74$ (dashed line), $\tilde{t}=0.97$ (dotted line) and
631: $\tilde{t}=1.26$ (thick solid line).
632: All spectra are normalized to the square of the mean magnetic field. The
633: critical wavenumber $\tilde{k}_c=16$ is marked on the $x$-axis.}
634: \end{figure}
635:
636: %Only the first term on the right-hand side of Eq. (14) was taken into account.
637: The dimensionless time $\tilde{t}$, the space coordinate $\tilde{z}$ and the
638: velocity $\tilde{u}$ are defined as
639: $\tilde{t}=tV_\mathrm{a}k_0$, $\tilde{z}=k_0z$,
640: $\tilde{u}=u/V_\mathrm{a}$, respectively. Here $k_0$ is the wavenumber that corresponds to
641: the real size of the numerical box $2\pi /k_0$. The dimensionless density
642: $\tilde{\rho }$ and the electric current $J$ can be expressed via the magnetic
643: field $B_0$ and the Alfv\'en velocity $V_\mathrm{a}$ as $\tilde{\rho }=4\pi \rho
644: V_\mathrm{a}^2/B_0^2$ and $J=4\pi j/ck_0B_0$. The dimensionless wavenumber
645: $\tilde{k}$ is simply $\tilde{k}=k/k_0$.
646:
647: The simulations were performed in a cubic box with size
648: $2\pi $. Periodic boundary conditions were imposed at the sides of
649: the box. We used $256^3$ grid cells in our simulations (to be
650: compared with $128^3$ cells in Bell's case).
651:
652: At $\tilde{t}=0$ the plasma pressure and density are uniformly distributed in
653: space. Small random magnetic perturbations corresponding to isotropically
654: distributed Alfv\'en waves with a one dimensional spectrum $\propto k^{-1}$ and
655: $\left< B^2\right> ^{1/2}/B_0=0.09$
656: were added to the mean unit
657: strength magnetic field that is in $z$ direction. Here $\left< \right> $
658: denote the spatial average over the simulation volume.
659: We use the values
660: $\gamma =5/3$ and $\beta =1$. Here $\beta = 4\pi P/B_0^2$.
661:
662: The evolution of the magnetic field and the mass velocity fluctuations,
663: together with the evolution of the sound speed, the mean electric field $E_0$
664: and the characteristic scale $L_m$ of the magnetic field, for a
665: dimensionless cosmic ray current $J=16$, is shown in Fig.2. The characteristic
666: scale $L_m=\int dk_zB_{\perp }(k_z)|k_z|^{-1}/\int dk_zB_{\perp }(k_z)$ is
667: determined via the spectrum $B_{\perp }(k_z)$ of the perpendicular component of
668: the random magnetic field $B_{\perp }(k_z)=\int dk_xdk_y(B_{xx}({\bf
669: k})+B_{yy}({\bf k}))$.
670:
671: After a brief initial stage the fluctuations grow exponentially with a growth
672: rate that is slightly smaller than $\tilde{\gamma }_{\max }=8$ in dimensionless
673: units. The initial growth of the magnetic fluctuations is not exponential,
674: since only a part of the initial perturbation corresponds to unstable
675: modes.
676:
677: At $\tilde{t}=0.7$ the magnetic perturbations are already comparable with the
678: mean magnetic field. Parts of the magnetic spiral expanding into the
679: $XY$-plane begin to collide with their surroundings.
680: When this happens multiple shocks are formed. The shape of these shocks may be
681: seen in the top panel of Fig.3, where the magnetic field strength in
682: perpendicular slices through the center of the box are shown. In the
683: $YZ$-plane, which contains the initial magnetic field vector, they look like
684: bow shocks. The shocks corresponding to the adjacent turns of the same
685: magnetic spiral are clearly seen. These shocks are almost
686: circular in the $XY$-plane. Low density cavities appear inside these
687: shocks. The size of these cavities in the $XY$ plane is larger than the size in
688: the $z$ direction. At later times the shocks collide with each other in the
689: $XY$ plane and the gas motion becomes strongly turbulent (middle and bottom
690: panels of Fig.3).
691:
692: The MHD turbulence has rather small spatial scales in the early stage of the
693: magnetic field growth. The scale of the magnetic field increases with time (see
694: Fig.2). This increase is slower in comparison with what was suggested by Bell
695: $k\sim 4\pi j_\mathrm{d}/Bc$ (or $\tilde{k}\sim JB_0/B$ in dimensionless units)
696: when the magnetic tension forces are comparable with the
697: Lorentz force produced by the cosmic ray electric current. This effect is
698: illustrated in Fig.4, where the magnetic spectra obtained for several instants
699: of time, are shown. It is clear from this figure that a nonlinear transfer of
700: magnetic energy takes place in the system. The increase of the magnetic energy
701: at wavenumbers $\tilde{k}>\tilde{k}_c$ that are not excited in the linear
702: approximation demonstrates the non-linear transfer of energy to smaller scales.
703: The magnetic energy in the small wavenumbers also grows faster than predicted
704: by the analytical growth rate formula. For example the amplitude of the
705: harmonics with wavenumber $\tilde{k}=1$ increases by a factor of 4.4 during the
706: period from $\tilde{t}=0.74$ up to $\tilde{t}=0.97$ (see Fig.4). The linear
707: growth rate for this harmonics $\tilde{\gamma
708: }=\sqrt{J\tilde{k}-\tilde{k}^2}=\sqrt{15}$ corresponds to an amplification
709: factor of about 2.2 during this period. This demonstrates the nonlinear
710: transfer of energy to larger scales. Thus the magnetic energy is non-linearly
711: transferred to both smaller and larger scales compared to the linearly excited
712: spatial scales.
713:
714: Towards the end of the simulation, at $\tilde{t}\sim 1.5$, the scale of the
715: magnetic field is comparable with the size of the box. At this point in time
716: the internal energy of the thermal gas is roughly equal to the kinetic and
717: magnetic energy in our simulation. Bell continued his calculation beyond this
718: point and found that at later times the magnetic field reaches a saturation
719: value. Also continuing our simulations we found that the magnetic field
720: continues to grow. We believe that this difference is due to the fact that
721: different MHD codes were used. The difference is not important however, since
722: the simulation does not model the real situation any more when the scale of the
723: field has become comparable with the size of the spatial simulation domain. In
724: this sense our results are similar to Bell's results.
725:
726: The growth of the MHD perturbations decreases with increasing magnetic field
727: amplification. As estimated in Appendix B, the magnetic field is amplified only
728: linearly at large times, $B/B_0\sim J\tilde{t}$ in dimensionless units, whereas
729: the gas thermal energy density increases $\propto t^3$. These dependencies are
730: derived using the equation for the evolution of the magnetic helicity.
731:
732:
733: \section{Conclusion}
734:
735: We have modeled the non-resonant
736: instability produced by a flux of charged
737: energetic particles that is driven through a scattering thermal
738: plasma. Using a significantly better numerical resolution we
739: basically confirm the results obtained earlier by Bell
740: \cite{bell04}. The magnetic field may be amplified significantly. The unstable
741: magnetic spirals collide with each other and shocks of moderate strength are
742: formed as the instability
743: develops (see previous Section). These shocks lead to significant gas heating.
744: Since free expansion of the magnetic
745: spirals after collision is impossible, the field grows only linearly in time at
746: later epochs (see Eq. (B3)). If the system has enough time to
747: evolve, the magnetic field growth will be stopped when the gyroradius $pc/qB$
748: of the energetic particles in the amplified field $B$ will drop down to the
749: scale of the amplified field $k^{-1}\sim cB/4\pi j_\mathrm{d}$. This determines
750: the value of the saturated magnetic field (Bell
751: \cite{bell04}, Pelletier et al. \cite{pelletier06}):
752:
753: \begin{equation}
754: \frac {B^2}{4\pi }\sim \frac {u_\mathrm{cr}}{v}\epsilon _\mathrm{cr}.
755: \end{equation}
756:
757: We should note that the small-scale approximation considered in Sect. 3
758: becomes invalid when the magnetic field reaches
759: this saturation value.
760:
761: Since the instability is driven by the Lorentz force, the corresponding MHD
762: turbulence has specific properties. It has non-zero magnetic helicity and
763: a non-zero mean electric field ${\bf E}_0$ parallel to the mean magnetic
764: field.
765:
766: We expect that the formation of multiple shocks in three-dimensional
767: MHD turbulence will also occur for other instabilities, in particular
768: for the resonant streaming instability, driven by cosmic rays.
769:
770: The scattering of energetic particles by the small-scale magnetic
771: inhomogeneities can be described using the Dolginov-Toptygin
772: approximation (Dolginov \& Toptygin \cite{dolginov67}). The
773: appearance of a mean second order electric field
774: ${\bf E}_0=-c^{-1}\left< \bf \delta
775: u\times \delta B\right> $, which is
776: oppositely directed to the electric current of the energetic
777: particles, modifies the cosmic ray transport equation (see
778: Appendix A).
779:
780: This non-resonant streaming instability may be important in any
781: astrophysical site where a strong electric current of energetic particles
782: exists and where the initial magnetic strength is small enough (see condition
783: (18)). Supernova remnants, starburst galaxies, galaxy cluster accretion shocks
784: and AGN jets are possible candidates for an application
785: of this instability.
786:
787: The results obtained will be used
788: in a companion paper by Zirakashvili \& Ptuskin \cite{zirakashvili08} (Paper
789: II) for a model of diffusive shock acceleration
790: in young SNRs in the presence of the non-resonant streaming
791: instability.
792:
793: \begin{acknowledgements}
794: We thank the anonymous referee for a number of valuable comments.
795: VSP and VNZ acknowledge the hospitality of the Max-Planck-Institut f\"ur
796: Kernphysik, where this work was mainly carried out. The work was also supported
797: by the RFBR grant in Troitsk.
798:
799: \end{acknowledgements}
800: %
801:
802:
803:
804: \appendix
805: \section{Diffusion approximation in the presence of the
806: additional electric field}
807:
808: The cosmic ray transport equation is modified in the presence of
809: the additional mean electric field ${\bf E}_0=-c^{-1}\left< \bf \delta
810: u\times \delta B\right> $ (Fedorov et al. \cite{fedorov92}).
811: Let us substitute the cosmic ray momentum
812: distribution (12) into Eq. (8) with the scattering tensor (10).
813: Performing the expansion up to the second order in $u/v$ and
814: collecting the terms independent of the direction of the particle
815: velocity ${\bf v}$ and separately those proportional to the
816: velocity we obtain after some algebra
817:
818: \begin{equation}
819: \frac {\partial N}{\partial t}+\nabla {\bf J}+ \frac {1}{p^2}\frac
820: {\partial }{\partial p}\frac {p^2}{v}q({\bf E}_0{\bf J}) =\frac
821: {{\bf u}_0}{p^2}\frac {\partial }{\partial p}\frac {p^3}{v^2}
822: \left( \nu \left( {\bf J+u}_0\frac {p}{3}\frac {\partial
823: N}{\partial p}\right) +\left[ {\bf \Omega \times J}\right] \right)
824: ,
825: \end{equation}
826:
827: \begin{equation}
828: \frac {\partial {\bf J}}{\partial t}+\frac {v^2}{3}\nabla N +q{\bf
829: E}_0\frac {v}{3}\frac {\partial N}{\partial p} = -\nu \left( {\bf
830: J+u}_0\frac {p}{3}\frac {\partial N}{\partial p}\right) + \left[
831: {\bf \Omega }\times \left( {\bf J+u}_0\frac {p}{3}\frac {\partial
832: N}{\partial p}\right) \right]
833: \end{equation}
834:
835: Here ${\bf \Omega }=qv{\bf B}_0/pc$. Assuming a slow time evolution we
836: neglect the time derivative in the last equation. Then it can be
837: used to find the cosmic ray flux ${\bf J}$:
838: \begin{equation}
839: J_i=-D_{ij}\left( \nabla _jN+E_{0j}\frac {q}{v}\frac {\partial
840: N}{\partial p}\right) -{u_{0i}}\frac {p}{3}\frac {\partial
841: N}{\partial p},
842: \end{equation}
843: where the diffusion tensor $D_{ij}$ has the following form
844: \begin{equation}
845: D_{ij}=(D_{\parallel }-D_{\perp })b_ib_j+D_{\perp }\delta
846: _{ij}+D_Ae_{ijk}b_k.
847: \end{equation}
848: Here ${\bf b}={\bf B}_0/B_0$ is the unit vector in the direction of the mean
849: field ${\bf B}_0$, $D_{\parallel }$ and $D_{\perp }$ denote the parallel and
850: perpendicular diffusion coefficients, respectively, and $D_A$ is the
851: antisymmetric diffusion coefficient. They are given
852: by Dolginov \& Toptygin {\cite{dolginov67}):
853: \begin{equation}
854: D_{\parallel }=\frac {v^2}{3\nu }, \ D_{\perp }=\frac {v^2\nu
855: /3}{\Omega ^2+\nu ^2}, \ D_{A}=\frac {v^2\Omega /3}{\Omega ^2+\nu
856: ^2}.
857: \end{equation}
858: Then Eq. (A1) reduces to
859: \begin{equation}
860: \frac {\partial N}{\partial t}+{\bf u}_0\nabla N-\frac {p}{3}\frac
861: {\partial N}{\partial p}\nabla {\bf u}_0 =\left( \nabla _i+\frac
862: {1}{p^2}\frac {\partial }{\partial p}\frac {p^2q}{v}E_{0i} \right)
863: D_{ij}\left( \nabla _jN+\frac qvE_{0j}\frac {\partial N}{\partial
864: p} \right) .
865: \end{equation}
866: Eq. (A6) shows how the presence of the mean electric field ${\bf E}_0$ modifies
867: the diffusion term of the cosmic ray transport equation.
868:
869: \section{Magnetic helicity}
870:
871: As noted first by Pelletier et al. \cite{pelletier06}, the MHD
872: turbulence generated by the non-resonant instability has nonzero
873: magnetic helicity $H=\left< \delta {\bf A} \delta {\bf B}\right> $,
874: where $\delta {\bf A}$ is the perturbation of the magnetic
875: potential. The magnetic helicity $H$ is a useful quantity and it
876: is often used in the theory of MHD turbulence and dynamo theories
877: (see e.g. Biskamp \cite{biskamp03}). Faraday's equation (3) may be
878: used for the determination of the time evolution of this quantity.
879: For the periodic system we have (cf. Subramanian \&
880: Brandenburg \cite{subramanian04})
881: \begin{equation}
882: \frac {\partial H}{\partial t}=-2c{\bf B}_0{\bf E}_0
883: \end{equation}
884:
885: We neglect magnetic dissipation here. This seems well justified because the
886: magnetic helicity is an integral quantity and is not transferred by nonlinear
887: interactions to smaller scales where dissipation is essential. In this sense
888: the magnetic helicity is different from the nonthermal (kinetic + magnetic)
889: energy that may be transferred to smaller and smaller scales where it is
890: transformed into gas thermal energy even in the case of infinitely small
891: viscosity.
892:
893: The mean electric field appears in the system as a response of the medium to
894: the cosmic ray electric current. The evolution of the energy density of the
895: plasma takes place according to Eq. (4). The right-hand side of this equation
896: is simply $-{\bf E}_0{\bf j}_\mathrm{d}$. Comparing with Eq. (B1) for a time
897: independent $j_\mathrm{d}$, we obtain the relation
898: \begin{equation}
899: H-2\frac {B_0\left< e\right> }{j_\mathrm{d}}=\mathrm{const}(t)
900: \end{equation}
901: It is worth emphasizing that Eq. (B2) is valid for the total plasma energy
902: density $e$ which appears in Eq. (4). During the stage of exponential growth
903: the instability produces magnetic field and velocity perturbations, whereas gas
904: heating is important at later times.
905:
906: We may use Eq. (B1) for a derivation of the equation for the magnetic field
907: amplification. Since the turbulence is helical, the magnetic helicity (that is
908: the product of the magnetic field and the vector magnetic potential) is $H\sim
909: B^2/k$, the electric field is $E_0\sim V_\mathrm{a}B^2/cB_0$, and $k\sim 4\pi
910: j_\mathrm{d}/Bc$. Then Eq. (B.1) yields the equation for the amplification of
911: the magnetic field:
912:
913: \begin{equation}
914: \frac {\partial B}{\partial t}\sim \frac {4\pi V_\mathrm{a}j_\mathrm{d}}{c}
915: \end{equation}
916: The numerical factor in this equation is of order unity, according to our
917: numerical results.
918:
919: There is another way to obtain this last equation. The field is amplified by
920: turbulent motions of the medium, that is $\partial B/\partial t\sim B_0ku_k$,
921: where $u_k$ is the turbulent velocity with wavenumber $k$. Assuming
922: equipartition, $u_k\sim V_\mathrm{a}B/B_0$, and the estimate $kB\sim 4\pi
923: j_\mathrm{d}/c$, we arrive at Eq. (B3).
924:
925: Since $H\propto B^3$ at late times, also $\left< e\right> \propto
926: t^3$, cf. Eq. (B2). On the other hand $B^2\propto t^2$, and
927: therefore at very late times the gas internal energy dominates.
928: %
929: \begin{thebibliography}{99}
930:
931: \bibitem[1983]{achterberg83} Achterberg, A., 1983, A\&A, 119, 274
932:
933: \bibitem[2006]{amato06} Amato, E., \& Blasi, P., 2006, MNRAS, 371, 1251
934:
935: \bibitem[1977]{axford77} Axford, W.I., Leer, E., Skadron, G., 1977, Proc. 15th
936: Int. Cosmic Ray Conf., Plovdiv, 90, 937
937:
938: % \bibitem[2003]{bamba03} Bamba, A., Yamazaki, R., Ueno, M., \& Koyama, K.,
939: % 2003, ApJ, 589, 827
940:
941: % \bibitem[2005]{bamba05} Bamba, A., Yamazaki, R., \& Hiraga, J.S., 2005, ApJ,
942: % 632, 294
943:
944: \bibitem[1978]{bell78} Bell, A.R., 1978, MNRAS, 182, 147
945:
946: \bibitem[2001]{bell01} Bell, A.R., \& Lucek, S.G., 2001, MNRAS, 321, 433
947:
948: \bibitem[2004]{bell04} Bell, A.R., 2004, MNRAS, 353, 550
949:
950: \bibitem[1996]{ber96} Berezhko, E.G., 1996, Astropart. Phys. 5, 367
951:
952: \bibitem[2002]{berezhko02} Berezhko, E.G., Ksenofontov L.G., \& V\"olk, H.J.,
953: 2002, A\&A 395, 943
954:
955: \bibitem[2004a]{berezhko04a} Berezhko, E.G., \& V\"olk, H.J., 2004a, A\&A 419,
956: L27
957:
958: \bibitem[2004b]{berezhko04b} Berezhko, E.G., \& V\"olk, H.J., 2004b, A\&A 427, 525
959:
960: \bibitem[1990]{berezinsky90} Berezinskii V.S., Bulanov, S.V., Dogiel, V.A.,
961: Ginzburg, V.L., \& Ptuskin, V.S., 1990, Astrophysics of
962: Cosmic Rays, North Holland, NY, Chapter IX
963:
964: \bibitem[2003]{biskamp03} Biskamp, D., 2003, Magnetohydrodynamic
965: turbulence, Cambridge, Cambridge Univ. Press
966:
967: \bibitem[1978]{blandford78} Blandford, R.D., \& Ostriker, J.P., 1978, ApJ, 221,
968: L29
969:
970: \bibitem[2005]{chevalier05} Chevalier, R., 2005, ApJ, 619, 839
971:
972: \bibitem[1967]{dolginov67} Dolginov, A.Z., \& Toptygin, I.N., 1967, JETP, 24,
973: 1195
974:
975: \bibitem[1992]{fedorov92} Fedorov, Yu.I., Katz, M.E., Kichatinov, L.L., \&
976: Stehlik, M., 1992, A\&A 260, 499
977: % \bibitem[2001]{gotthelf01} Gotthelf, E.V., Halpem, J.P., Camilo, F. et
978: % al. 2001, ApJ, 552, L125
979:
980: %\bibitem[1975]{gull75} Gull, S.F., 1975, MNRAS, 171, 263
981:
982: %\bibitem[2002]{hwang02} Hwang, U., Decourchelle, A., Holt, S.S., \& Petre, R.,
983: % 2002, ApJ, 581, L101
984:
985: \bibitem[1977]{krymsky77} Krymsky, G.F., 1977, Soviet Physics-Doklady, 22, 327
986:
987: \bibitem[1983]{lagage83} Lagage, P.O., \& Cesarsky, C.J., 1983, A\&A, 118, 223
988:
989: \bibitem[1981]{lifshitz81} Lifshitz, E.M., \& Pitaevskii, L.P., 1981, Physical kinetics,
990: Oxford: Pergamon Press
991:
992: \bibitem[1967]{lerche67} Lerche I., 1967, ApJ, 147, 689
993:
994: \bibitem[2000]{lucek00} Lucek,S.G.,\& Bell, A.R., 2000, MNRAS, 314, L65
995:
996: \bibitem[2001]{malkov01} Malkov, M.A., \& Drury, L.O'C, 2001, Reports on
997: Progress in Physics, 64, 429
998:
999: \bibitem[1982]{mckenzie82} McKenzie, J.F., \& V\"olk, H.J., 1982, A\&A, 116,
1000: 191
1001:
1002: \bibitem[2006]{pelletier06} Pelletier, G., Lemoine, M., \& Marcowith, A., 2006,
1003: A\&A, 453, 181
1004:
1005: \bibitem[2003]{pen03} Pen, U.L., Arras, P., \& Wong, S.K., 2003, ApJSS, 149,
1006: 447
1007:
1008: % \bibitem[2005]{pohl05} Pohl, M., Yan, H., \& Lazarian, A., 2005,
1009: % ApJ, 626, L101
1010:
1011: \bibitem[1984]{ptuskin84} Ptuskin, V.S., 1984, JETP, 59, 281
1012:
1013: \bibitem[2003]{ptuskin03} Ptuskin, V.S., \& Zirakashvili, V.N., 2003, A\&A, 403,
1014: 1
1015:
1016: \bibitem[2005]{ptuskin05} Ptuskin, V.S., \& Zirakashvili, V.N., 2005, A\&A, 429,
1017: 755
1018:
1019: \bibitem[2004]{subramanian04} Subramanian, K., \& Brandenburg, A., 2004
1020: Phys. Rev. Lett., 93, 205001
1021:
1022: \bibitem[1977]{tsytovich77} Tsytovich, V.N., 1977, Theory of Turbulent Plasma,
1023: Consultant Bureau, New York
1024: % \bibitem[2003]{vink03} Vink, J., \& Laming, J.M., 2003, ApJ, 584, 758
1025:
1026: \bibitem[2006]{vladimirov06} Vladimirov, A., Ellison, D.C., \& Bykov, A., 2006,
1027: ApJ, 652, 1246
1028:
1029: % \bibitem[2003]{voelk03} V\"olk, H.J., Berezhko, E.G., \& Ksenofontov, L.T.,
1030: % 2003, A\&A, 409, 563
1031:
1032: % \bibitem[2005]{voelk05} V\"olk, H.J., Berezhko, E.G., \& Ksenofontov, L.T.,
1033: % 2005, A\&A, 433, 229
1034:
1035: \bibitem[2007]{voelk07} V\"olk, H.J., Berezhko, E.G., \& Ksenofontov, L.T.,
1036: 2007, submitted to A\&A
1037:
1038: \bibitem[1974]{wentzel74} Wentzel, D.G., 1974, ARA\&A 12, 71--96
1039:
1040: \bibitem[2008]{zirakashvili08} Zirakashvili, V.N., \& Ptuskin, V.S., 2008, ApJ,
1041: submitted (Paper II)
1042:
1043:
1044: \end{thebibliography}
1045: \end{document}
1046: