1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{natbib}
3: %
4: \newcommand{\var}{\epsilon}
5: \newcommand{\gr}{$\gamma$-ray }
6: \newcommand{\no}{\noindent}
7: \newcommand{\nn}{\nonumber}
8: \newcommand{\bea}{\begin{eqnarray}}
9: \newcommand{\ena}{\end{eqnarray}}
10: \newcommand{\be}{\begin{equation}}
11: \newcommand{\en}{\end{equation}}
12: \newcommand{\bed}{\begin{displaymath}}
13: \newcommand{\ed}{\end{displaymath}}
14: \newcommand{\ti}{\times}
15: \newcommand{\dd}{\mathrm{d}}
16: \newcommand{\pd}[2]{{{\partial #1}\over {\partial #2}}}
17: %
18: \shorttitle{Magnetic filaments in SNR}
19: \shortauthors{Pohl et al.}
20: %
21: %
22: \begin{document}
23: %
24: %
25: \title{Magnetically limited X-ray filaments in young SNR}
26:
27: \author{M. Pohl}
28: \affil{Department of Physics and Astronomy, Iowa State University, Ames, Iowa 50011-3160, USA}
29: \email{mkp@iastate.edu}
30: \author{H. Yan and A. Lazarian}
31: \affil{Department of Astronomy, University of Wisconsin, 5534 Sterling Hall,
32: 475 North Charter Street, Madison, WI 53706}
33:
34:
35: \begin{abstract}
36: We discuss the damping of strong magnetic turbulence downstream of the forward shock of young
37: supernova remnants (SNR). We find that strong magnetic fields, that have been produced by the
38: streaming instability in the upstream region of the shock, or by other kinetic instabilities at
39: the shock, will be efficiently reduced, so the region of
40: enhanced magnetic field strength would typically have a thickness of the order
41: $l_d=(10^{16}-10^{17})$~cm. The non-thermal X-ray
42: filaments observed in young SNR are thus likely limited by the magnetic field and not by the
43: energy losses of the radiating electrons. Consequently the thickness of the filaments would not
44: be a measure of the magnetic field strength and claims of efficient cosmic-ray acceleration
45: on account of a run-away streaming instability appear premature.
46: \end{abstract}
47: %
48: \keywords{acceleration of particles -- supernova remnants -- X-rays: ISM}
49: %
50: %________________________________________________________________
51: %
52: \section{Introduction}
53: Supernova remnant shocks are prime candidates for the acceleration sites of galactic cosmic rays.
54: A number of young shell-type SNR have been observed to emit nonthermal X-ray emission, which is
55: commonly interpreted as being synchrotron emission of freshly accelerated electrons in the 10--100~TeV
56: energy range. These electrons, and also newly accelerated cosmic-ray nucleons, should produce
57: GeV--TeV scale \gr emission, which has been detected from a few objects \citep[e.g.][]{aha04}.
58:
59: Likely though the inference may appear, evidence for electron acceleration in
60: SNR does not imply evidence for cosmic-ray nucleon acceleration. To solve the long-standing
61: problem of the origin of galactic cosmic rays one is therefore interested in finding unambiguous
62: signatures of a large flux of high-energy cosmic-ray nucleons in SNR, which obviously requires a
63: good understanding of the nonthermal X-ray emission from SNR, i.e. the spatial distribution
64: of magnetic field and the spatial and energy distribution of high-energy electrons, so the leptonic
65: \gr emission can be accurately modeled.
66:
67: High-resolution X-ray observations indicate that a large fraction of the nonthermal X-ray
68: emission on the rims of young SNR is concentrated in narrow filaments \citep{bamba05}.
69: In the literature these filaments are usually interpreted as reflecting the spatial distribution
70: of the high-energy electrons around their acceleration site, presumably the forward shock
71: \citep{yama04}. By comparing the electron energy-loss time with the downstream flow velocity one
72: then derives an estimate of the magnetic field strength that far exceeds the
73: typical interstellar magnetic field
74: strength modified according to the Rankine-Hugoniot conditions \citep{vl03,bamba04}.
75:
76: A number of authors have further advanced the argument by noting that strong near-equipartition
77: magnetic fields can be produced by the streaming instability \citep{lb00,bl01}, thus
78: apparently providing evidence for a very efficient acceleration of cosmic-ray
79: nucleons at SNR forward shocks \citep{ber03a,ber03b,bv04}.
80:
81: It has been pointed out before that the observed X-ray filaments may actually be magnetic filaments,
82: i.e. localized enhancements of the magnetic field, rather than limited by the distribution
83: of high-energy electrons. \citet{lp04} studied the pile-up of magnetic field at the contact
84: discontinuity and concluded that it could explain the X-ray filaments, if a significant
85: fraction of the electrons was accelerated close to the contact discontinuity.
86:
87: In this Letter we investigate the damping of turbulent magnetic field near
88: SNR shocks. In contrast to earlier studies \citep{vzz88,fed92,ddk96,pz03,pz05}
89: that concentrated on damping in the upstream region, where
90: it is in competition with the streaming instability, we only consider the downstream region
91: where the growth of turbulent magnetic field is negligible. We thus determine the spatial
92: extent of amplified turbulent magnetic field that is produced at or upstream of the forward shock.
93: Our analysis is independent of the amplification process, it applies to
94: magnetic field produced by the streaming instability in the upstream region
95: as well as to field generated by a kinetic plasma instability at the SNR shock
96: \citep{akh75}.
97:
98: Substantial progress has been made in the understanding of strong magnetic turbulence, e.g.
99: we have theoretical scalings that are supported by numerical simulations. Nevertheless,
100: many questions remain unanswered, and in order to derive a generally valid picture,
101: we use various models of
102: the cascading and damping behaviour resulting from nonlinear wave interactions. While the models
103: differ in their assumptions, they all predict a significant reduction of the turbulent
104: magnetic energy density on a timescale shorter than that of energy losses of electrons that could
105: synchrotron-radiate at a few keV X-ray energy.
106: Given typical plasma streaming velocities downstream of a SNR shock,
107: this would result in magnetic filaments of thickness $\lesssim 10^{17}$~cm.
108: So even if a strong magnetic field is produced at or upstream of the forward shock by any
109: mechanism, it is likely confined to a small volume around the shock and the observed
110: X-ray filaments would just reflect the structure of the magnetic fields.
111:
112: \section{Wave damping in the downstream medium of SNR shocks}
113: The energy density $W(k,x)$ of magnetic turbulence is related to the total
114: turbulent magnetic field as
115: \be
116: {{(\delta B)^2}\over {4\pi}} = \int dk\ W(k,x)
117: \label{1}
118: \en
119: In a plasma under
120: steady-state conditions $W(k,x)$ obeys the continuity equation
121: \be
122: U\,\pd{W}{x}=2\,(\Gamma_g -\Gamma_d)\,W
123: \label{2}
124: \en
125: where $\Gamma_g$ is the growth rate and $\Gamma_d$ is the damping rate of turbulence.
126: $U$ is the propagation velocity of the wave energy, i.e. the sum of the plasma
127: velocity and the component of the group velocity of waves along the direction of the plasma flow.
128: In the case under study we expect the excitation of turbulence by the
129: streaming instability in the upstream region at $x \le 0$, which is transported
130: to the downstream region ($x > 0$) and eventually damped. Alternatively one could consider
131: kinetic instabilities which are known to produce magnetic turbulence at the shock front
132: ($x\simeq 0$). The instabilities
133: should not operate in the downstream region and hence the growth rate would
134: vanish there.
135:
136: Equation \ref{2} must be separately solved on either side of the shock. As boundary
137: condition we may assume absence of turbulence far ahead of the shock,
138: $W(k,x=-\infty)=0$, whereas the MHD jump conditions applied to the upstream solution
139: at $x=0$ provide the boundary condition for the downstream solution at
140: $x\rightarrow 0$ \citep{vs99}.
141:
142: Earlier studies found that the streaming instability during the early stages
143: of SNR evolution can be so strong that the amplified turbulent
144: magnetic field $\delta B\gtrsim 100 \ {\rm \mu G}$ far exceed the undisturbed
145: magnetic field $B\approx (3-10)\,{\rm \mu G}$ upstream of the SNR blast wave
146: \citep{lb00,bl01}.
147: The cosmic-ray diffusion coefficient $\kappa$ should then be close to the Bohm limit, and
148: consequently both the cosmic rays with energy $E$ and the amplified magnetic field would
149: in the upstream region be confined to a shock precursor zone of thickness
150: \be
151: l_{\rm prec.}\approx {\kappa\over {U_s}}
152: \label{3}
153: \en
154: \bed
155: \simeq (10^{15}\ {\rm cm})\ \left({{U_s}\over {3000\,{\rm km/s}}}\right)^{-1}\,
156: \left({{E}\over {\rm TeV}}\right)\,
157: \left({{B}\over {100\,{\rm \mu G}}}\right)^{-1} \ed
158: where $U_s$ is the SNR shock velocity.
159:
160: Let us in the following assume that the streaming instability has efficiently
161: produced strong magnetic fields in the precursor zone. The magnitude of the
162: amplified magnetic field upstream of the shock depends on
163: the efficiency of the various possible damping mechanisms.
164: In any case a fraction
165: of the magnetic turbulence will propagate through the shock to the downstream region,
166: where it will convect away from the shock. We are interested in the
167: spatial scale, on which the amplified magnetic energy is damped downstream of
168: the shock. If this length scale is small compared with the dimensions of a supernova
169: remnant, then the streaming instability essentially produces a magnetic filament
170: at the location of the forward shock, that should be observable as a non-thermal
171: X-ray filament on account of enhanced synchrotron emission of very high-energy
172: electrons.
173: If, on the other hand, the damping length scale were comparable to or larger than
174: the typical SNR, then the entire interior of he remnant should be filled with
175: strong magnetic field and the observed non-thermal X-ray filaments must be limited
176: by electron energy losses.
177:
178: Ion-neutral collisions will not efficiently damp magnetic turbulence in the downstream
179: region on account of the high plasma temperature. Ordinary collisionless
180: damping does occur, but is very inefficient for Alfv\'en and fast-mode perturbations
181: that propagate along a large-scale magnetic field. Calculations of the damping rate
182: of obliquely propagating fast modes for low-$\beta$ plasma \citep{gin61} and high-$\beta$
183: systems \citep{fk79} suggest that collisionless damping is also slow for small k or large wavelengths.
184: In the following we will discuss
185: nonlinear wave-wave interactions and cascading as potential processes
186: that may limit the thickness of the high magnetic field layer in the downstream region of
187: the SNR shock.
188:
189: The cascading of wave energy in astrophysical environments is not well
190: understood to date. Much of the interstellar medium, and hence the upstream medium of SNR
191: shocks, is thought to be turbulent. \citet{gs95} have studied the cascading of wave energy
192: and found a concentration of wave energy in transverse modes, which are noted to be an inefficient
193: means of cosmic-ray scattering \citep{ch00,yl02}. In addition, wave damping may occur by interactions
194: with background MHD turbulence. Wave packages are distorted in
195: collisions with oppositely directed turbulent wave packets, resulting in a cascade of
196: wave energy to smaller scaler where it is ultimately dissipated \citep{yl02}.
197: Analoguous to MHD perturbations that can be decomposed into Alfv\'{e}nic, slow and fast
198: waves with well-defined dispersion relations, MHD perturbations that characterize turbulence
199: can apparently be separated into distinct modes.
200:
201: The separation into Alfv\'en and pseudo-Alfv\'en modes, which
202: are the incompressible limit of slow modes, is an essential element
203: of the \citet{gs95} model.
204: Even in a compressible medium, MHD turbulence is not an inseparable mess in spite of the fact that MHD
205: turbulence is a highly non-linear phenomenon \citep{LG01,CL02}.
206: The actual decomposition of MHD turbulence into Alfv\'en, slow and fast
207: modes was addressed in \citet{CL02,CL03}, who
208: used a statistical procedure of
209: decomposition in the Fourier space, where the basis of the Alfv\'en, slow
210: and fast perturbations was defined. For some particular cases, e.g. for a low $\beta$ medium,
211: the procedure was benchmarked successfully and therefore argued to be reliable.
212:
213: Here we list the results of three previous studies on wave damping by cascading to very small scales,
214: one using a very general picture of turbulence cascading, and the other two refering to the cascading of
215: fast-mode and Alfv\'en waves in background MHD turbulence, respectively.
216:
217: \subsection{Kolmogorov-type energy cascade}
218: For the case of a Kolmogorov-type nonlinearity \citet{pz03}
219: have considered the simplified expression
220: \be
221: \Gamma_{\rm nl}\approx {{V_A}\over {20}}\,k\,A(>k)\ ,\qquad A(>k)={{\sqrt{4\pi\,k\,W(k,x)}}\over {B_0}}
222: \label{4}
223: \en
224: where
225: $V_A$ is the Alfv\'en speed. This ansatz does not build on the disparity of the parallel and
226: perpendicular scales, and therefore its applicability may be limited, but it has been used in part of
227: the recent literature and thus merits our consideration.
228:
229: In conjunction with a vanishing growth rate $\Gamma_g=0$, inserting Eq.~\ref{4}
230: in Eq.~\ref{2} yields
231: the damping length scale
232: \be
233: l_k =20\,{{U\,\sqrt{\rho}}\over {\sqrt{W(k,0)}\,k^{3/2}}}
234: \label{7}
235: \en
236: with $W(k,0)$ being the energy density spectrum of turbulence that is transmitted through the
237: shock to the downstream region.
238:
239: Most of the magnetic energy density will reside at small $k$ or large wavelengths $\lambda$.
240: Inserting numbers that are typical of the downstream region of young SNR blast waves and
241: a highly amplified turbulent magnetic field the damping length scale is expected to be
242: \be
243: l_k\simeq (1.5\cdot 10^{16}\ {\rm cm})\,
244: \left({{U}\over {1000\,{\rm km/s}}}\right)\, \label{8}
245: \en
246: \bed\times\
247: \left({{n}\over {{\rm atoms/cm^3}}}\right)^{1/2}\,
248: \left({{\lambda}\over {10^{15}\,{\rm cm}}}\right)\,
249: \left({{\delta B (\lambda)}\over {100\,{\rm \mu G}}}\right)^{-1}
250: \ed
251: Here $\delta B$ is defined as
252: \be
253: {{\left[\delta B(\lambda)\right]^2}\over {4\pi}} =
254: {{\left[\delta B(k)\right]^2}\over {4\pi}} = k\, W(k,x)
255: \label{9}
256: \en
257: The wavelength of the turbulent magnetic field, $\lambda$, that is used in Eq.~\ref{8}
258: should be related to the Larmor radius $r_L$ of the high-energy particles, which are supposedly
259: accelerated at the shock front. Noting that
260: \be
261: r_L = (3\cdot 10^{13}\ {\rm cm})\,
262: \left({{E}\over {\rm TeV}}\right)\,
263: \left({{B}\over {100\,{\rm \mu G}}}\right)^{-1}
264: \label{10}
265: \en
266: we see that the value of $\lambda$ used in Eq.~\ref{8} corresponds to about 30 TeV
267: particle energy, in the case of electrons sufficient to account for X-ray synchrotron emission
268: at a few keV.
269:
270: For the smallest $k$ or largest wavelength $\lambda$ the magnetic fields components
271: used in Eqs.~\ref{8} and \ref{9} must be similar. Internal consistency should require that
272: the damping length scale is larger than the largest wavelength of the waves, lest the waves at
273: large wavelengths would
274: have very little energy density in contradiction to our assumption. In addition, an efficient backscattering
275: of cosmic rays to the upstream region requires that the high magnetic field layer in the downstream region
276: has a thickness much larger than the scattering mean free path, which should be similar to the Larmor radius
277: of the particles and hence to the wavelength of the waves. This leads to a limit for the cosmic-ray energy
278: that is independent of the actual strength of the amplified magnetic field.
279: \be
280: l_k > \lambda\approx r_L\quad\Rightarrow\ E\lesssim 500\ {\rm TeV}
281: \label{11}
282: \en
283:
284: \subsection{Cascading of fast-mode turbulence}
285: \citet{yl04} have studied the cascading of fast-mode turbulence, that is randomly driven by
286: turbulence on large scales $L$ in a compressible medium. They find a cascading timescale
287: \be
288: \Gamma_{\rm fast} \simeq \sqrt{k\over L}\,{{V_L^2}\over {V_\phi}}
289: \label{12}
290: \en
291: where $V_L$ is the turbulence velocity at the injection scale and $V_\phi$ is the phase velocity
292: of fast-mode wave and equal to the Alfv\'en and sound velocity for high- and low-$\beta$ plasma,
293: respectively. Typical values for the sound and Alfv\'en speed in the downstream region of the shock
294: of a young SNR are
295: \be
296: C_s \simeq 10^8\ {\rm cm\,s^{-1}}\label{13}
297: \en
298: \bed
299: V_A\simeq (2\cdot 10^7\ {\rm cm\,s^{-1}})\ \left({B\over {100\,{\rm \mu G}}}\right)
300: \,\left({{n_d}\over {\rm atoms\,cm^{-3}}}\right)^{-1/2}
301: \ed
302: so both characteristic velocities may be expected to be similar, corresponding to $\beta\approx 1$ in the
303: downstream plasma. Inserting the cascading rate (\ref{12}) as damping rate into the turbulence transport
304: equation (\ref{2}) we find that the turbulent magnetic field, if composed of fast-mode waves,
305: should exponentially decay on a length scale
306: \bed
307: l_f \simeq
308: (10^{16}\ {\rm cm})\,\left({{U}\over {1000\,{\rm km/s}}}\right)\,
309: \left({{V_\phi}\over {1000\,{\rm km/s}}}\right)\,
310: \ed
311: \be
312: \times\ \left({{V_L}\over {1000\,{\rm km/s}}}\right)^{-2}\,
313: \left({{L}\over {3\,{\rm pc}}}\right)^{1/2}\,
314: \left({{\lambda}\over {10^{15}\,{\rm cm}}}\right)^{1/2}\,
315: \label{14}
316: \en
317: where for the turbulence driving scale and velocity we have used values that are characteristic of the
318: SNR blast wave itself.
319:
320: \subsection{Cascading of Alfv\'en modes}
321: \citet{fg04} have investigated the cascading of Alfv\'en modes by MHD turbulence that
322: is driven on the outer scale $L$. The damping rate is
323: lowest, and the growth rate by cosmic-ray streaming is highest, for parallel-propagating Alfv\'en
324: waves. All other waves with the same (parallel) wavelength should damp faster than the parallel
325: propagating Alfv\'en waves, so a lower limit to the damping rate of Alfv\'en waves
326: would be
327: \be
328: \Gamma_A\gtrsim {{V_A}\over \sqrt{\lambda\,L}}
329: \label{17}
330: \en
331: Again, inserting the cascading rate (\ref{17}) as damping rate into the turbulence transport
332: equation (\ref{2}) we find that the turbulent magnetic field, if composed of Alfv\'en waves,
333: should exponentially decay on a length scale
334: \bed
335: l_A \lesssim
336: (10^{17}\ {\rm cm})\,\left({{U}\over {1000\,{\rm km/s}}}\right)\,
337: \left({{V_A}\over {500\,{\rm km/s}}}\right)^{-1}\ed
338: \be\times\
339: \left({{L}\over {3\,{\rm pc}}}\right)^{1/2}\,
340: \left({{\lambda}\over {10^{15}\,{\rm cm}}}\right)^{1/2}
341: \label{18}
342: \en
343: which is very similar to the length scale obtained for the cascading of fast-mode waves in
344: background MHD turbulence (Eq.~\ref{14}).
345:
346: \section{Discussion}
347: Limited and incomplete though they likely are, all models of strong turbulence cascading predict
348: a damping length of magnetic turbulence $l_d=(10^{16}-10^{17})$~cm that is generally small
349: enough to produce a magnetic filament. The
350: length scale of X-ray filaments, that are limited by electron energy losses, is larger
351: than $10^{17}$~cm even for a strongly amplified magnetic field with $B\simeq 100\ {\rm \mu G}$.
352:
353: To obtain a qualitative understanding of the X-ray properties of magnetic filaments we performed
354: a simplified calculation for a spherically symmetric SNR with a forward shock located at the radius
355: $r_s=10$~pc.
356: The magnetic is isotropic and its spatial distribution follows
357: \be
358: B=(5\ {\rm \mu G}) + (45\ {\rm \mu G})\,\exp\left({{r-r_s}\over {0.1\ \rm pc}}\right)
359: \label{20}
360: \en
361: The differential electron density is at first assumed to follow
362: \be
363: N(E)=N_0\,E^{-2}\,\exp\left(-{E\over {100\ \rm TeV}}\right)\,\Theta\left(r_s-r\right)
364: \label{19}
365: \en
366: We have also studied an exponential decrease of the electron density on the lengthscale $0.5$~pc,
367: i.e. $\propto \exp((r-r_s)/0.5\ {\rm pc})$, as a rough approximation to the effects of a finite
368: age of the high-energy electron population.
369:
370: The resulting X-ray intensity distribution as a function of the projected distance from the SNR center,
371: $r_p$,
372: is an integral over the synchrotron emission coefficient, $j_\nu$,
373: \be
374: I_\nu (r_p) = \int_{-\infty}^\infty dx\ j_\nu (r=\sqrt{x^2+r_p^2})
375: \label{21}
376: \en
377: and is shown in Fig.\ref{pyl-f1} for three cases.
378:
379: \begin{figure}
380: \plotone{f1.eps}
381: \caption{The nonthermal X-ray intensity as a function of the projected distance from the
382: SNR center. The acceleration site is assumed located at $r=10$~pc.
383: The solid line shows the intensity at 1 keV for a homogeneous electron density, and
384: the dotted line displays the (scaled) intensity at 10 keV X-ray energy. The dashed line is derived
385: for the same parameters as the solid line, except that the electron density is assumed to
386: exponentially decay on a length scale of $0.5$~pc.}
387: \label{pyl-f1}
388: \end{figure}
389: To be noted from the figure is a weak frequency dependence of the downstream width of the X-ray filaments,
390: which is caused by the variation of the cut-off frequency of the synchrotron spectrum on account of
391: the spatial variation of the magnetic field strength. Spectral modeling of the radio-to-X-ray spectra of
392: young SNR suggests that the cut-off frequencies are generally found in the X-ray band \citep{rk97},
393: so we must expect some frequency dependence of the width of X-ray filaments, even if
394: they are dominantly magnetic.
395:
396: One should also note that any ageing or finite age of the electron population would enhance the contrast
397: between the intensity of the filament and that of the plateau emission from the interior of the remnants.
398: For young SNR electron energy losses would be generally unimportant outside of the magnetic
399: filament, though.
400:
401: Our results have serious consequences for the modeling of leptonic TeV-scale \gr emission, as
402: that is produced anywhere, where high-energy electrons reside, and not only inside the filaments.
403: A careful study of the acceleration history and thus of the electron distribution
404: inside the remnant is required to obtain accurate estimates of the spectral energy distribution
405: of leptonic emission from young SNR.
406:
407: \acknowledgements
408:
409: Support for MP by NASA under award No. NAG5-13559 is gratefully acknowledged.
410: HY is supported by the NSF grant ATM 0312282. AL acknowledge support of the Center for Magnetic Self-Organization in Laboratory and Astrophysical Plasmas.
411:
412:
413: \begin{thebibliography}{}
414:
415: \bibitem[Aharonian et al.(2004)]{aha04} Aharonian F.A., Akhperjanian A.G., Aye K.-M. et al., 2004,
416: \nat, 432, 75
417:
418: \bibitem[Akhiezer et al.(1975)]{akh75} Akhiezer A.I., Akhiezer I.A., Polovin R.V., Sitenko A.G.,
419: Stepanov K.N., 1975, Plasma Electrodynamics, Pergamon Press
420:
421: \bibitem[Bamba et al.(2005)]{bamba05} Bamba A., Yamazaki R., Yoshida T., Terasawa T., Koyama K., 2005,
422: \apj, in press (astro-ph/0411326)
423:
424: \bibitem[Bamba et al.(2004)]{bamba04} Bamba A., Yamazaki R., Ueno M., Koyama K., 2004,
425: Adv. Sp. Res.,
426:
427: \bibitem[Bell \& Lucek(2001)]{bl01} Bell A.R., Lucek S.G., 2001, MNRAS, 321, 433
428:
429: \bibitem[Berezhko, Ksenofontov \& V\"olk(2003)]{ber03a} Berezhko E.G., Ksenofontov L.T.,
430: V\"olk H.J., 2003, \aap, 412, L11
431:
432: \bibitem[Berezhko, P\"uhlhofer \& V\"olk(2003)]{ber03b} Berezhko E.G., P\"uhlhofer G.,
433: V\"olk H.J., 2003, \aap, 400, 971
434:
435: \bibitem[Berezhko \& V\"olk(2004)]{bv04} Berezhko E.G., V\"olk H.J., 2004, \aap, 419, L27
436:
437: \bibitem[Chandran(2000)]{ch00} Chandran B.D.G., 2000, \prl, 85, 4656
438:
439: \bibitem[Cho \& Lazarian(2003)]{CL03} Cho J., Lazarian A., 2003, \mnras, 345, 325
440:
441: \bibitem[Cho \& Lazarian(2002)]{CL02} Cho J., Lazarian A., 2003, \prl, 88, 245001
442:
443: \bibitem[Drury, Duffy \& Kirk(1996)]{ddk96} Drury L.O'C, Duffy P., Kirk J.G., 1996, \aap, 309, 1002
444:
445: \bibitem[Farmer \& Goldreich(2004)]{fg04} Farmer A.J., Goldreich P., 2004, \apj, 604, 671
446:
447: \bibitem[Fedorenko(1992)]{fed92} Fedorenko V.N., 1992, Ap\&SS Rev., Vol.8:4, 1
448:
449: \bibitem[Foote \& Kulsrud(1979)]{fk79} Foote E.A., Kulsrud R.M., 1979, \apj, 233, 302
450:
451: \bibitem[Ginzburg(1961)]{gin61} Ginzburg V.L., 1961, Propagation of Electromagnetic Waves in Plasma,
452: New York, Gordon \& Breach
453:
454: \bibitem[Goldreich \& Sridhar(1995)]{gs95} Goldreich P., Sridhar S., 1995, \apj, 438, 763
455:
456: %\bibitem[Goldreich \& Sridhar(1997)]{gs97} Goldreich P., Sridhar S., 1997, \apj, 485, 680
457:
458: \bibitem[Lithwick \& Goldreich(2001)]{LG01} Lithwick Y., Goldreich P., 2001, \apj, 562, 279
459:
460: \bibitem[Lucek \& Bell(2000)]{lb00} Lucek S.G., Bell A.R., 2000, MNRAS, 314, 65
461:
462: \bibitem[Lyutikov \& Pohl(2004)]{lp04} Lyutikov M., Pohl M., 2004, \apj, 609, 785
463:
464: \bibitem[Ptuskin \& Zirakashvili(2005)]{pz05} Ptuskin V.S., Zirakashvili V.N., 2005, \aap, 429, 755
465:
466: \bibitem[Ptuskin \& Zirakashvili(2003)]{pz03} Ptuskin V.S., Zirakashvili V.N., 2003, \aap, 403, 1
467:
468: \bibitem[Reynolds \& Keohane(1999)]{rk97} Reynolds S.P., Keohane J.W., 1999, \apj, 525, 368
469:
470: \bibitem[Vainio \& Schlickeiser(1999)]{vs99} Vainio R., Schlickeiser R., 1999, \aap, 343, 303
471:
472: \bibitem[Vink \& Laming(2003)]{vl03} Vink J., Laming J.M., 2003, \apj, 584, 758
473:
474: \bibitem[V\"olk, Zank \& Zank(1988)]{vzz88} V\"olk H.J., Zank L.A., Zank G.P., 1988, \aap, 198, 274
475:
476: \bibitem[Yamazaki et al.(2004)]{yama04} Yamazaki R., Yoshida T., Terasawa T., Bamba A., Koyama K.,
477: 2004, \aap, 416, 595
478:
479: \bibitem[Yan \& Lazarian(2002)]{yl02} Yan H., Lazarian A., 2002, \prl, 89, 281101
480:
481: \bibitem[Yan \& Lazarian(2004)]{yl04} Yan H., Lazarian A., 2004, \apj, 614, 757
482:
483: \end{thebibliography}
484:
485: \end{document}
486: