1: \documentclass[longnamesfirst,apjl]{emulateapj}
2: %\documentclass[preprint,12pt]{aastex}
3: \usepackage{apjfonts,natbib,epsfig}
4: \bibliographystyle{apj3}
5: \tighten
6:
7: \newcommand{\beq}{\begin{equation}}
8: \newcommand{\eeq}{\end{equation}}
9: \newcommand{\bea}{\begin{eqnarray}}
10: \newcommand{\eea}{\end{eqnarray}}
11:
12: \begin{document}
13: \title{Cluster Merger Shock Constraints on Particle Acceleration and Nonthermal Pressure in the Intracluster Medium}
14: \author{
15: Ehud Nakar\altaffilmark{1},
16: Milo\v s Milosavljevi\'c\altaffilmark{2}, and
17: Daisuke Nagai\altaffilmark{1}} \altaffiltext{1}{Theoretical
18: Astrophysics, Mail Code 130-33, California Institute of Technology,
19: 1200 East California Boulevard, Pasadena, CA 91125.}
20: \altaffiltext{2}{Department of Astronomy, University of Texas, 1
21: University Station C1400, Austin, TX 78712.}
22:
23: \righthead{SHOCK CONSTRAINTS ON NONTHERMAL PRESSURE IN CLUSTERS}
24:
25: \lefthead{NAKAR ET AL.}
26:
27: \begin{abstract}
28:
29: X-ray observations of galaxy cluster merger shocks can be used to
30: constrain nonthermal processes in the intracluster medium (ICM). The
31: presence of nonthermal pressure components in the ICM, as well as the
32: shock acceleration of particles and their escape, all affect shock
33: jump conditions in distinct ways. Therefore, these processes can be
34: constrained using X-ray surface brightness and temperature maps of
35: merger shock fronts. Here we use these observations to place
36: constraints on particle acceleration efficiency in intermediate Mach
37: number (${\cal M} \approx 2-3$) shocks and explore the potential to
38: constrain the contribution of nonthermal components (e.g., cosmic
39: rays, magnetic field, and turbulence) to ICM pressure in cluster
40: outskirts. We model the hydrodynamic jump conditions in merger shocks
41: discovered in the galaxy clusters A520 (${\cal M} \approx 2$) and 1E
42: 0657$-$56 (${\cal M} \approx 3$) using a multifluid model comprised of
43: a thermal plasma, a nonthermal plasma, and a magnetic field. Based on
44: the published X-ray spectroscopic data alone, we find that the
45: fractional contribution of cosmic rays accelerated in these shocks is
46: $\lesssim 10\%$ of the shock downstream pressure. Current observations
47: do not constrain the fractional contribution of nonthermal components
48: to the pressure of the undisturbed shock upstream. Future X-ray
49: observations, however, have the potential to either detect particle
50: acceleration in these shocks through its effect on the shock dynamics,
51: or to place a lower limit on the nonthermal pressure contributions in
52: the undisturbed ICM. We briefly discuss implications for models of
53: particle acceleration in collisionless shocks and the estimates of
54: galaxy cluster masses derived from X-ray and Sunyaev-Zel'dovich effect
55: observations.
56:
57: \keywords{
58: cosmic rays --- galaxies: clusters: general ---
59: galaxies: clusters: individual (A520, 1E 0657--06) ---
60: intergalactic medium ---
61: shock waves --- turbulence --- X-rays: galaxies: clusters }
62:
63: \end{abstract}
64:
65: \section{Introduction}
66:
67: \setcounter{footnote}{0}
68:
69: Astrophysical collisionless shocks are the likely sources of the
70: observed extra-solar high energy cosmic rays
71: \citep[e.g.,][]{Bell:78,Blandford:78}. Continuous observational
72: effort has not yet yielded direct evidence for acceleration of
73: cosmic ray nuclei in collisionless shocks, although recently,
74: tentative indirect evidence for such acceleration was identified in
75: the morphology of the Tycho supernova remnant \citep{Warren:05}, and
76: in the high-energy gamma-ray emission near the RX J1713.7$-$3946
77: remnant \citep{Aharonian:07}. The theory of diffusive shock
78: acceleration \citep[e.g.,][and references therein]{Blandford:87}
79: predicts that the acceleration efficiency and the spectrum of
80: accelerated particles depend on the Mach number and other parameters
81: of the shock \citep[e.g.,][and references therein]{Giacalone:97}.
82: The presence of fossil cosmic rays in the pre-shock medium, e.g.,
83: from previous shocks, can also affect the acceleration efficiency
84: \citep[e.g.,][and references therein]{Kang:07a,Kang:07b}. We here
85: argue that merging galaxy clusters are laboratories in which
86: theories of cosmic ray acceleration in intermediate Mach number
87: shocks can be tested. Observational constraints on particle
88: acceleration in such shocks are especially interesting as numerical
89: simulations suggest that these are the source of a large fraction of
90: the cosmic rays accelerated in galaxy clusters
91: \citep[e.g.,][]{Ryu:03}.
92:
93: Fossil cosmic rays,\footnote{Potential acceleration sites of fossil
94: cosmic rays are past accretion and merger shocks, giant radio
95: sources, supernovae, and turbulence in the ICM
96: \citep[e.g.,][]{Berezinsky:97}} magnetic field, and turbulence may
97: all contribute to the pressure of the intracluster medium (ICM),
98: thereby modifying its hydrodynamic behavior. Such contribution may
99: alter the interpretation of observations that ignore them. For
100: example X-ray observations have recently been used to estimate shock
101: velocities in two merging clusters (see below) neglecting nonthermal
102: components; improved estimates of shock velocities in these and
103: other clusters may have to take into account the cosmic ray and
104: other nonthermal contributions to the ICM pressure. Nonthermal
105: pressure is also a source of systematic bias when cluster masses are
106: estimated from X-ray and Sunyaev-Zel'dovich effect (SZE)
107: measurements that are made assuming an hydrostatic equilibrium
108: between gravitational forces and thermal pressure gradients in the
109: ICM \citep[e.g.,][and references
110: therein]{Ensslin:97,Rasia:06,Nagai:07a}. These nonthermal biases
111: limit the effectiveness of upcoming cluster surveys in the quest to
112: place constraints on the expansion history of the universe.
113:
114: Evidence for the nonthermal activity in clusters is growing.
115: Observed radio and hard X-ray emissions in clusters suggest a
116: presence of relativistic electrons with Lorentz factor of $\sim
117: 10^4$. This also suggests a presence of relativistic protons that
118: could have been accelerated by the same mechanism that has
119: accelerated the electrons. Direct evidence for cosmic ray ions in
120: the ICM is, however, still lacking. The nondetection in {\it EGRET}
121: data of gamma-ray emission expected from neutral pion decay in
122: cosmic ray collisions in the ICM \citep[e.g.,][]{Reimer:03} has so
123: far placed upper limits on the fraction of cosmic ray pressure to
124: $\lesssim 20\%-30\%$ in several nearby rich clusters
125: \citep{Ensslin:97,Pfrommer:04}. Most cluster atmospheres are also
126: substantially magnetized, with typical field strengths of order a
127: few $\mu$G out to Mpc radii \citep[][and references
128: therein]{Carilli:02,Govoni:04,Govoni:06}. There is likely to be
129: considerable variation in field strengths ($\sim 0.1-40
130: \mu\textrm{G}$) and topologies within clusters. Thus while magnetic
131: fields are likely to provide a significant contribution to the
132: pressure in some regions \citep[e.g., along some cold fronts;
133: see][]{Vikhlinin:01}, it is yet unclear what is the average
134: energetics in the cluster outskirts. Numerical simulations of
135: cluster formation also suggest that subsonic gas motions
136: (turbulence) contribute substantial nonthermal pressure in clusters
137: \citep[e.g.,][]{Norman:99,Ricker:01,Nagai:03,Faltenbacher:05,
138: Dolag:05,Rasia:06,Nagai:07a}. Further investigations of nonthermal
139: phenomena in clusters are hence critical for the success of upcoming
140: X-ray and SZE cluster surveys, as our ability to estimate cluster
141: masses hinges on a precise characterization of the nonthermal
142: components.
143:
144: In this work, we show that shock waves that form during merging of
145: galaxy clusters can provide unique constraints on nonthermal processes
146: in clusters. Recent {\it Chandra} X-ray observations have revealed
147: that shock waves with Mach numbers ${\cal M} \approx 2-3$ that form
148: during the merging of galaxy clusters are accompanied by distinct
149: X-ray surface brightness and temperature discontinuities. To date, a
150: density and temperature jump along axis of symmetry of a cluster
151: merger bow shock has been recovered in the cluster A520
152: \citep{Markevitch:05} and in 1E 0657$-$56
153: \citep{Markevitch:02,Markevitch:06}. The relative magnitudes of the
154: density and temperature jump depend on the contribution of nonthermal
155: components (e.g., cosmic rays, magnetic field, turbulence) to the
156: pressure of the undisturbed, upstream ICM. They also depend on the
157: efficiency of particle acceleration in the shock, on the escape of the
158: accelerated particles from the shock, and on the amplification of
159: turbulence in the shock.
160:
161: Here we make the first attempt to use the X-ray observations of
162: merging clusters to constrain their particle acceleration and the
163: fractional contribution of nonthermal components to the pressure
164: budget of the ICM. In \S~\ref{sec:theory}, we model the hydrodynamic
165: jump conditions and Mach number using the multi-fluid approximation.
166: We model the effect of cosmic rays, a tangled magnetic field, and
167: turbulence, on observed gas jump conditions in the shock. In
168: \S~\ref{sec:results}, we utilize this model to derive constraints on
169: nonthermal components in merger shocks in galaxy clusters A520 and 1E
170: 0657$-$56. We show that current observations can limit the efficiency
171: of particle acceleration, and that the fractional contribution of
172: nonthermal fluids to ICM pressure may be constrained with future,
173: improved, X-ray spectroscopy and SZE observations. In
174: \S~\ref{sec:discussion}, we discuss implications of these results in
175: context of particle acceleration models and for the estimates of
176: galaxy clusters masses with X-ray and SZE observations. In
177: \S~\ref{sec:conclusions}, we summarize our main conclusions.
178:
179:
180: \section{Limits on Particle Acceleration in Merger Shocks and Nonthermal Pressure in ICM}
181: \label{sec:theory}
182:
183: \subsection{Shock Jump Conditions}
184:
185: The ICM fluid consists of a thermal component (electrons and ions in
186: mutual thermal equilibrium) and number of nonthermal components
187: (nonthermal cosmic rays and magnetic fields comoving with the
188: thermal component). Turbulent gas motions and electromagnetic waves
189: can also contribute pressure to the ICM. We ignore the
190: electromagnetic waves in the current treatment. We also temporarily
191: ignore turbulence, but return to discuss its role on qualitative
192: level in \S~\ref{sec:turbulence}. The ICM fluid can be modeled with
193: an adiabatic equation of state with an effective adiabatic index
194: $\gamma$.\footnote{We define the adiabatic indices via
195: $\gamma=1+P/\varepsilon$, where $\varepsilon$ is the internal energy
196: density. The adiabatic index thus calculated may differ from
197: $\partial \ln P/\partial \ln\rho$.} In what follows, we use indices
198: ``u'' and ``d'' to denote the shock upstream and downstream,
199: respectively. Assuming cylindrical symmetry, conservation of
200: density, momentum, and energy fluxes across the shock dictates jump
201: conditions on the symmetry axis of the bow shock that read
202: \begin{eqnarray} \label{EQ:shock_jump}
203: \rho_{\rm u} v_{\rm u}&=&\rho_{\rm d} v_{\rm d} , \nonumber\\
204: P_{\rm u}+\rho_{\rm u}v_{\rm u}^2&=&P_{\rm d}+\rho_{\rm d}v_{\rm d}^2 , \nonumber\\
205: v_{\rm u}\left(\frac{1}{2}\rho_{\rm u}v_{\rm u}^2+\frac{\gamma_{\rm u}}{\gamma_{\rm u}-1}P_{\rm u}\right)&=&v_{\rm d}\left(\frac{1}{2}\rho_{\rm d}v_{\rm d}^2+\frac{\gamma_{\rm d}}{\gamma_{\rm d}-1}P_{\rm d}\right) ,
206: \end{eqnarray}
207: where $\rho_i$ is the mass density of the thermal gas (the mass
208: density of the nonthermal particles is negligible), $v_i$ is the
209: fluid velocity, and $P_i$ is the total pressure ($i=\{u,d\}$)
210:
211: The pressure is a sum of electronic, ionic, cosmic ray, and magnetic
212: field contributions, $P_i=P_{{\rm e},i}+P_{{\rm i},i}+P_{{\rm
213: CR},i}+P_{B,i}$. The effective adiabatic index $\gamma_i$ equals
214: \begin{eqnarray}
215: \label{EQ:gamma}
216: \frac{\gamma_i}{\gamma_i-1}&=&\frac{1-\epsilon_{{\rm nt},i}}{2}
217: \left(\frac{\gamma_{{\rm i},i}}{\gamma_{{\rm i},i}-1}
218: +\frac{\gamma_{{\rm e},i}}{\gamma_{{\rm e},i}-1}\right)\nonumber\\& &
219: +\epsilon_{{\rm CR},i}\frac{\gamma_{\rm CR}}{\gamma_{\rm CR}-1}
220: +\epsilon_{B,i}\frac{\gamma_{B,i}}{\gamma_{B,i}-1}.
221: \end{eqnarray}
222: where $\epsilon_{{\rm CR},i}\equiv P_{{\rm CR},i}/P_i$ and
223: $\epsilon_{B,i}\equiv P_{B,i}/P_i$ are, respectively, the fractional
224: contribution of cosmic rays and magnetic fields to the total pressure,
225: while $\epsilon_{{\rm nt},i}=\epsilon_{{\rm CR},i}+\epsilon_{B,i}$ is
226: the total fractional contribution of nonthermal pressure
227: components. The temperature jump across the shock of the thermal
228: electrons is then
229: \begin{equation}\label{EQ:Tjump}
230: \tau\equiv \frac{T_{\rm e,d}}{T_{\rm e,u}}=\frac{(1-\epsilon_{\rm
231: nt,d})[2\gamma_{\rm u}/(\gamma_{\rm
232: u}-1)-r^{-1}-1]}{(1-\epsilon_{\rm nt,u})[2\gamma_{\rm
233: d}/(\gamma_{\rm d}-1)-r-1]},
234: \end{equation}
235: where $r\equiv \rho_{\rm d}/\rho_{\rm u}$ is the compression ratio.
236:
237: \subsection{Thermal Electron Relativistic Corrections}
238:
239: Thermal ions are nonrelativistic in cluster shocks and thus
240: $\gamma_{{\rm i},i}=5/3$. The adiabatic index of thermal electrons
241: may differ from this value because of relativistic corrections. The
242: adiabatic index of thermal electrons equals the ratio of the
243: rest-frame pressure to internal energy density, $\gamma_{\rm
244: e}=1+P_{\rm e}/ \varepsilon_{\rm e}$, where $\varepsilon_{\rm e}$ is
245: the internal electron energy density (e.g., \citealt{Achterberg:84})
246: \begin{equation}
247: \label{eq:gamma_e} \gamma_{\rm e} = 1 + \frac{1}{3}
248: \frac{\int_0^\infty 4\pi p^4 (p^2/c^2+m_{\rm e}^2)^{-1/2}f_{\rm
249: e}(p) dp}{\int_0^\infty 4\pi p^2 [(p^2c^2+m_{\rm
250: e}^2c^4)^{1/2}-m_{\rm e}c^2]f_{\rm e}(p) dp} ,
251: \end{equation}
252: where
253: \begin{equation} f_{\rm e}(p) \propto \exp\left[-\frac{(p^2c^2+m_{\rm
254: e}^2c^4)^{1/2}}{kT_{\rm e}}\right]
255: \end{equation}
256: is the thermal electron momentum distribution. Defining the electron
257: temperature in units of the electron rest energy $\Theta_{\rm
258: e}\equiv kT_{\rm e}/ m_{\rm e}c^2$, the adiabatic index in equation
259: (\ref{eq:gamma_e}) can conveniently be expressed in terms of
260: modified Bessel functions of the second kind (e.g.,
261: \citealt{Kunik:03})
262: \begin{equation}
263: \label{eq:gamma_e_bessel} \gamma_{{\rm e},i}=1+\Theta_{{\rm e},i}
264: \left[3\Theta_{{\rm e},i}+ \frac{K_1(\Theta_{{\rm
265: e},i}^{-1})}{K_2(\Theta_{{\rm e},i}^{-1})}-1\right]^{-1}
266: \end{equation}
267: In the limit $\Theta_{\rm e}\ll 1$, relevant to cluster merger shocks,
268: the adiabatic index is $\gamma_{\rm e}= 5/3-(5/6) \Theta_{\rm e}+{\cal
269: O}(\Theta_{\rm e}^2)$. Note that when $kT_{\rm e,d}$ or $kT_{\rm e,u}$
270: becomes comparable to $m_{\rm e}c^2$, the right side of equation
271: (\ref{EQ:Tjump}) depends on the electron temperatures. For example,
272: if the nonthermal pressures are ignored (i.e., $\epsilon_{{\rm
273: nt},i}=0$), a relativistic correction of about $10\%$ applies to the
274: expected temperature jump and inferred Mach number for conditions
275: similar to those in the bow shock in 1E 0657$-$56 ($T_{\rm
276: u}=10\textrm{ keV}$ and $r=3$), while a correction of only about $1\%$
277: applies for the lower-temperature and weaker shock in A520 ($T_{\rm
278: u}=5\textrm{ keV}$ and $r=2.3$).
279:
280: \subsection{Fossil Cosmic Rays}
281:
282: The undisturbed ICM may contain an intracluster population of fossil
283: cosmic rays that could have been produced in the high Mach number
284: accretion shocks (${\cal M}\sim10-100$), in previous merger shocks, in
285: active galactic nuclei, and in starburst-associated phenomena
286: \citep[see,
287: e.g.,][]{Berezinsky:97,Fujita:01,Miniati:01,Gabici:03,Sarazin:04}.
288: After a merger bow shock passes a fluid element in the ICM, the fluid
289: element will contain the original fossil cosmic rays, some of which
290: may have been further accelerated in the shock. The shock may also
291: accelerate new cosmic rays drawn from the thermal pool and accelerated
292: in the shock for the first time.
293:
294: The cosmic ray adiabatic index depends on whether their pressure is
295: dominated by Newtonian or relativistic particles, which in turn
296: depends on the details of the cosmic ray spectrum
297: \citep[e.g.,][]{Achterberg:84}. The effective cosmic ray adiabatic
298: index lies in the range $4/3 \leq \gamma_{\rm CR} \leq 5/3$; the
299: electron cosmic rays are mildly or fully relativistic while the
300: protons that dominate the cosmic rays pressure can be Newtonian or
301: relativistic. Thus, assuming that electron and proton cosmic rays
302: are close to mutual equipartition, we expect $\gamma_{\rm CR}$ to
303: range between $\gamma_{\rm CR}\approx 4/3$ (for relativistic
304: electrons and protons) and $\gamma_{\rm CR} \approx 13/9$
305: \citep[relativistic electrons and Newtonian protons;
306: e.g.,][]{Konigl:80}. While in principle $\gamma_{\rm CR}$ may differ
307: between the upstream and the downstream, when there is a significant
308: population of upstream cosmic rays, it is reasonable to expect that
309: these will also dominate the downstream, and thus that $\gamma_{\rm
310: CR}$ is approximately constant across the shock.
311:
312: In addition to being reaccelerated by shock acceleration mechanism,
313: fossil cosmic rays may be accelerated adiabatically during the shock
314: compression of the magnetic field. However, shock compression may
315: very well be nonadiabatic as the Larmor radius of a mildly
316: relativistic electron in a microgauss field, $r_{\rm L}\sim
317: 10^9\textrm{ cm}$, may be much larger than the transition layer over
318: which the fluid density jumps.\footnote{If pressure across the
319: transition layer is not mediated by cosmic rays but by plasma
320: instabilities, the width of the transition layer will be of the
321: order of the plasma skin depth of the thermal protons, which is
322: $\delta \sim 10^7\textrm{ cm}$. On the other hand, adiabatic
323: acceleration will take place when $r_{\rm L}\ll \delta$.} Therefore,
324: it is possible that cosmic rays do not gain energy during
325: compression. Nevertheless, the compression increases the cosmic ray
326: number density and therefore its pressure by at least a factor of
327: $r$. Thus we express the fractional cosmic ray pressure in the
328: downstream as
329: \begin{equation} \label{eq:nofossil} \epsilon_{\rm
330: CR,d}=r\left(\frac{P_{\rm u}}{P_{\rm d}}\right)\epsilon_{\rm
331: CR,u}+\epsilon_{\rm acc}
332: \end{equation}
333: where the first term is the pressure of the fossil cosmic rays if
334: the energy of individual cosmic rays remains unchanged as they pass
335: through the shock, and the second term is any additional pressure of
336: cosmic rays that are accelerated in the shock. The acceleration term
337: includes pressure due to accelerated particles from the upstream
338: thermal pool that are added to the nonthermal pool, and also due to
339: fossil nonthermal particles that are reaccelerated by a shock
340: acceleration process and/or by adiabatic compression.
341:
342: \subsection{The Cluster Magnetic Field}
343: \label{sec:magnetic_field}
344:
345: The effect of the magnetic field on the shock depends on its
346: strength and topology, both of which are constrained very poorly,
347: especially in the cluster outskirts. Faraday rotation measurements
348: suggest that the magnetic pressure can reach $\sim 10\%$ of the
349: thermal pressure in some clusters or regions within clusters, and
350: that in others it is less than $1\%$
351: \citep[e.g.,][]{Carilli:02,Govoni:04,Govoni:06}.\footnote{A magnetic
352: field that is tangled on very small scales is not well constrained
353: by the Faraday rotation measurements.} Given the large uncertainty
354: in the field strength, we treat $\epsilon_B$ as a free parameter in
355: what follows.
356:
357: We assume that the magnetic field of the undisturbed ICM is tangled
358: and isotropic on scales relevant to the hydrodynamics of the cluster
359: merger shock. This approximation allows us to use the unmagnetized
360: form of the jump conditions, equation (\ref{EQ:shock_jump}), to
361: describe the effect of magnetic field on shock hydrodynamics. The
362: approximate coherence length of the field in some galaxy clusters,
363: $\sim 10\textrm{ kpc}$, is shorter than the curvature scale of the
364: merger shock, $\sim 100\textrm{ kpc}$, so that one can average over
365: the fluctuating field orientation near the shock.\footnote{The same
366: approximation was employed by \citet{Markevitch:05}.} Furthermore,
367: the measurements of density and temperature jumps across the shock
368: are based on deprojection of the X-ray map assuming a cylindrical
369: symmetry of a bow shock---thus an averaging of observables on the
370: shock curvature scale is implicit in the reported shock jump data.
371:
372:
373: Even if the upstream field is isotropic, the downstream field may be
374: anisotropic as a result of a preferential amplification of the
375: magnetic field component that is perpendicular to the shock,
376: $B_\perp$. In this case the perpendicular component will still be
377: isotropic in the plane of the shock, and can thus be parameterized
378: by the ratio of the average parallel and perpendicular field energy
379: densities $b\equiv \langle B_\perp^2\rangle/2\langle
380: B_\parallel^2\rangle$. The effective adiabatic index of the magnetic
381: fluid, $\gamma_B$, which is defined as the parameter that correctly
382: quantifies the behavior of the magnetic pressure in equation
383: (\ref{EQ:gamma}) and may depend on the orientation of the magnetic
384: field, can be expressed in terms of the averages of the components
385: ${\cal T}^{\rm em}_{\mu\nu}$ of the electromagnetic part of the
386: energy-momentum tensor in the shock frame via
387: \begin{equation}
388: \frac{\gamma_B}{\gamma_B-1}=\frac{\langle {\cal T}^{\rm em}_{10}\rangle}{\beta \langle {\cal T}^{\rm em}_{11}\rangle} .
389: \end{equation}
390: Here, $\beta$ is the shock velocity in units of the speed of light,
391: and $\Gamma\equiv (1-\beta^2)^{-1/2}$ is the Lorentz factor of the
392: shock. Components of the energy-momentum tensor read
393: \begin{eqnarray}
394: {\cal T}^{\rm em}_{11}&=&\frac{1}{8\pi}[(2\Gamma^2-1)B_\perp^2-B_\parallel^2] ,\nonumber\\
395: {\cal T}^{\rm em}_{10}&=&\frac{1}{4\pi} \Gamma^2\beta B_\perp^2 .
396: \end{eqnarray}
397: With these, for a nonrelativistic shock we obtain
398: \begin{equation}\label{eq:gammaB}
399: \gamma_B=\frac{4b}{2b+1} .
400: \end{equation}
401: The value of $\gamma_B$ in equation (\ref{eq:gammaB}) equals
402: $\gamma_B=2$ in a perpendicular shock ($b=\infty$) and vanishes in a
403: parallel shock ($b=0$). If the field is isotropic on average in the
404: shock upstream ($b=1$), we recover the value $\gamma_{B,{\rm
405: u}}=4/3$. For an isotropic upstream field, the downstream field will
406: have $b \geq 1$ and thus $4/3 \leq \gamma_{B,{\rm d}}\leq 2$.
407:
408: Relating $\epsilon_{B,{\rm d}}$ to $\epsilon_{B,{\rm u}}$ requires
409: an understanding of the shock structure. But even if the shock
410: structure is unknown, we can limit $\epsilon_{B,{\rm d}}$ in two
411: extreme cases, when field generation in the shock itself (as may
412: take place in unmagnetized shocks) is neglected. Assuming that
413: magnetohydrodynamic jump conditions apply within the shock
414: transition, and that the parallel and the perpendicular fields do
415: not transform into each other, then $B_{\|,{\rm d}}=B_{\|,{\rm u}}$
416: and $B_{\bot,{\rm d}}=rB_{\bot,{\rm u}}$, implying that
417: $\epsilon_{B,{\rm d}}=(2r^2-1)(P_{\rm u}/P_{\rm d})\epsilon_{B,{\rm
418: u}}$. The other extreme assumption, which was previously made by
419: \citet{Markevitch:05}, is that the field is isotropic and remains
420: isotropic throughout the shock compression, in which case
421: $\epsilon_{B,{\rm d}}=r^{4/3}(P_{\rm u}/P_{\rm d})\epsilon_{B,{\rm
422: u}}$, and thus the magnetic field behaves as a relativistic or
423: photon gas that is adiabatically compressed in the shock. We expect
424: $\epsilon_{B,{\rm d}}$ to lie between these two limits, assuming
425: that no new field is generated within the shock.
426:
427:
428: \subsection{Escape of the Highest Energy Cosmic Rays}
429: \label{sec:escape}
430:
431: Another process that may affect the shock and that we have ignored so
432: far is the escape of the highest-energy cosmic rays that are
433: accelerated in the shock. This process may take place if the shock
434: efficiently produces a hard cosmic ray spectrum, and the
435: highest-energy cosmic rays can escape and remove energy from the shock
436: transition \citep[see, e.g.,][]{Achterberg:84}. This energy leakage
437: can be parameterized by the fraction $Q$ of the incoming energy flux that
438: escapes the shock. The leakage affects the shock jump conservation
439: relations such that the left-hand side of the energy flux conservation
440: relation---the last of equations (\ref{EQ:shock_jump})---must be
441: multiplied by the factor $1-Q$. With this, equation (\ref{EQ:Tjump})
442: generalizes into
443: \begin{equation}\label{EQ:Tjump_gen}
444: \tau=\frac{(1-\epsilon_{\rm
445: nt,d})[2(1-Q)\gamma_{\rm u}/(\gamma_{\rm
446: u}-1)-r^{-1}-1+Qr/(r-1)]}{(1-\epsilon_{\rm nt,u})[2\gamma_{\rm
447: d}/(\gamma_{\rm d}-1)-r-1+Qr^2/(r-1)]} .
448: \end{equation}
449: The effect of escape is similar to the effect of particle
450: acceleration, as it removes a fraction of the incoming energy from
451: the downstream thermal fluid, thereby reducing the temperature jump.
452: Note, however, that if the measured value of $\tau$ is below the one
453: expected in a thermal shock, the deviation from the thermal
454: temperature jump can only partially be attributed to the escape,
455: since when $\epsilon_{\rm acc}=0$, the value of $Q$ must by
456: definition be zero.
457:
458:
459: \subsection{Sensitivity of the Temperature Jump}
460:
461: We have one relation (equation \ref{EQ:Tjump} or its generalized
462: form, equation \ref{EQ:Tjump_gen}) and several unknown parameters
463: describing the upstream and downstream contribution of the cosmic
464: rays and magnetic field to the fluid pressure. Without making
465: assumptions about the nature of the nonthermal fluids, given a
466: measurement of the density jump $r$ and the temperature jump $\tau$,
467: we can place constraints in the joint parameter space spanned by
468: these parameters, but cannot recover the fractional pressures
469: themselves. Additionally, when only soft X-ray data are available
470: (as from {\it Chandra} and {\it XMM-Newton}), the uncertainties in
471: $\tau$ typically exceed the uncertainties in $r$. Therefore, before
472: we proceed to explore the joint parameter space, we discuss the
473: effect of the variation of various parameters on $\tau$, assuming
474: that $r$ is accurately measured and is held fixed. The temperature
475: jump is most sensitive to the fractional pressure in cosmic rays
476: accelerated or reaccelerated in the shock, $\epsilon_{\rm acc}$. A
477: nonzero value of $\epsilon_{\rm acc}$ reduces the value of $\tau$,
478: e.g., in the ideal case in which the upstream nonthermal pressure
479: vanishes, $\epsilon_{\rm nt,u}=0$, we find that $\tau$ drops by
480: about a factor of two between $\epsilon_{\rm acc}=0$ and
481: $\epsilon_{\rm acc}=0.3$ ($r=2-3$). The reason for this sensitive
482: dependence is that the production of accelerated particles in the
483: shock saps a fraction of the incoming energy flux out of the thermal
484: component of the downstream, thereby reducing $\tau$. Increasing $Q$
485: results in a similar effect, for similar reason, on $\tau$.
486:
487: The effect of changing $\epsilon_{\rm CR,u}$ and $\epsilon_{B,{\rm
488: u}}$ on the temperature jump is more subtle since a high nonthermal
489: pressure in the shock upstream implies that there will also be a
490: high nonthermal pressure in the downstream; it is the balance
491: between the two that determines $\tau$. Therefore, $\tau$ depends
492: only weakly on the upstream nonthermal components. Increasing
493: $\epsilon_{\rm CR,u}$ while keeping the rest of the parameters
494: constant always increases $\tau$ slowly. For example assuming that
495: the magnetic pressure vanishes ($\epsilon_{B,{\rm u}}=0$) and that
496: $\gamma_{\rm CR}=4/3$, $\tau$ increases by a factor of $\approx 1.2$
497: between $\epsilon_{\rm CR,u}=0$ and $\epsilon_{\rm CR,u}=0.3$ for
498: $r=2.3$ and $\epsilon_{\rm acc}=0$. Having a constant $\epsilon_{\rm
499: acc}=0.3$ reduces the change in $\tau$ to a factor of $\approx 1.1$.
500: Increasing $\epsilon_{B,{\rm u}}$ also mostly results in a slightly
501: larger temperature jump. If the perpendicular field is strongly
502: amplified in the shock, $\gamma_B$ jumps in the shock, thereby
503: increasing the effective downstream adiabatic index $\gamma_{\rm
504: d}$, and thus increasing $\tau$. If on the other hand we treat the
505: field as a relativistic gas, then $\epsilon_{B,{\rm u}}$ and $\tau$
506: are positively correlated for $\epsilon_{\rm acc}=0$ and are weakly
507: anticorrelated for $\epsilon_{\rm acc} \approx 0.15$.
508:
509: In conclusion, $\tau$ is strongly anticorrelated with $\epsilon_{\rm
510: acc}$ and is typically weakly positively correlated with
511: $\epsilon_{\rm CR,u}$ and $\epsilon_{B,{\rm u}}$. Therefore an
512: accurate measurement of $\tau$ and $r$ can tightly constrain
513: $\epsilon_{\rm acc}$. If the measured value of $\tau$ falls below
514: that expected in a purely thermal hydrodynamic shock given a precise
515: measurement of $r$, then a nontrivial lower and upper bound can be
516: placed on $\epsilon_{\rm acc}$, but such a constraint cannot be
517: obtained for $\epsilon_{\rm CR,u}$ and $\epsilon_{B,{\rm u}}$. If,
518: on the other hand, the measured value of $\tau$ is higher than that
519: expected in a purely thermal hydrodynamic shock, only an upper limit
520: on $\epsilon_{\rm acc}$ can be placed, i.e., $\epsilon_{\rm acc}=0$
521: remains viable. Then, however, the measurement places a lower limit
522: on the upstream nonthermal pressure. However, based on the
523: measurement of $r$ and $\tau$ alone, one cannot separate the partial
524: contributions of the cosmic ray and magnetic components.
525:
526: \subsection{Implications for Shock Velocity Estimation}
527: \label{sec:mach_number}
528:
529: The velocity of cluster merger shock is of great interest since it
530: is a stepping stone toward relating the dynamics of the ICM to the
531: dynamics of the dark matter in galaxy cluster mergers
532: \citep[e.g.,][]{Hayashi:06,Farrar:06,Milosavljevic:07,Springel:07}.
533: Typically the shock Mach number is inferred from the compression
534: ratio $r$, which can be measured relatively accurately in X-ray
535: maps, under the strict assumption that all pressures are thermal
536: (the observed constraints on $\tau$ are typically used for
537: consistency check with this assumption). Taking the nonthermal
538: pressure into account (but ignoring cosmic ray escape; see
539: \S~\ref{sec:escape}), the Mach number is given by
540: \begin{equation}\label{eq:mach}
541: {\cal M}^2=\frac{2r\gamma_{\rm u}/(\gamma_{\rm u}-1)-2\gamma_{\rm
542: d}/(\gamma_{\rm d}-1)}{\gamma_{{\rm g,u}}(1-\epsilon_{{\rm nt,
543: u}})(1-r^{-1})[2\gamma_{\rm d}/(\gamma_{\rm d}-1)-r-1]},
544: \end{equation}
545: where $\gamma_{\rm g,u} = 5/3$ is the adiabatic index of the
546: upstream thermal gas, for which relativistic corrections are
547: negligible. This relation implies that high nonthermal pressure in
548: the upstream (downstream) increases (decreases) ${\cal M}$. For
549: example, assuming that $\epsilon_{\rm nt, u}=0$ and cosmic rays are
550: rather efficiently accelerated in the shock, $\epsilon_{\rm
551: acc}=0.15$, the inferred value of the Mach number must be revised by
552: $\approx 10\%-20\%$ downward of the value inferred for a thermal
553: shock, when $r$ is in the range $2-3$.
554:
555:
556: If, in addition to $r$, the value of $\tau$ is accurately measured,
557: the constraints that can be placed on the Mach number are tighter,
558: since the upstream and downstream nonthermal components are no longer
559: entirely free. In this case, when $r$ and $\tau$ are held constant, a
560: larger nonthermal pressure results in a higher Mach number. For
561: example if the nonthermal component behaves as relativistic gas
562: ($\gamma_{\rm nt}=4/3$) and $r$ and $\tau$ are related as they would
563: be in a purely thermal shock, $\tau=(4r-1)/r(4-r)$, the shock Mach
564: number equals
565: \begin{equation}\label{eq:mach_thermal}
566: {\cal M}^2=\frac{3(4r-1)(1-\epsilon_{\rm nt,u})/(4-r)+5r+3r\epsilon_{\rm nt,
567: u}-8}{\gamma_{\rm u}(1-\epsilon_{\rm nt,u})(1-r^{-1})(7-r)} ,
568: \end{equation}
569: in which case ${\cal M}$ increases by $\approx 10\%$ for
570: $\epsilon_{\rm nt,u}=0.3$, compared to the purely thermal shock, for
571: $r$ in the range $2-3$.
572:
573:
574: \subsection{Turbulence}
575: \label{sec:turbulence}
576:
577: In addition to the cosmic rays and the magnetic field, turbulence
578: also contributes pressure to the ICM. Turbulence in the ICM is
579: expected to be driven by gravitational clustering (accretion and
580: merging), and by outflows associated with active galactic nuclei.
581: Hydrodynamical simulations of galaxy cluster formation in which the
582: ICM is treated as an ideal fluid universally demonstrate that
583: turbulent pressure in the ICM is non-negligible, and that its
584: contribution to the total pressure is an increasing function of
585: radius from the center of the cluster. The fractional turbulent
586: pressure measured in the simulations is $\epsilon_{\rm tur}\sim
587: 0.06-0.36$ \citep{Norman:99}, $\epsilon_{\rm tur}\sim 0.05-0.1$
588: \citep{Ricker:01}, $\epsilon_{\rm tur}\sim 0.04-0.09$
589: \citep{Nagai:03,Faltenbacher:05}, and $\epsilon_{\rm tur} \sim
590: 0.05-0.3$ \citep{Dolag:05}.\footnote{Hydrodynamic simulations of
591: intracluster turbulence were also recently carried out by \citet
592: {Fujita:04}, \citet{Subramanian:06}, and \citet{Vazza:06}.}
593: Spatially-resolved gas pressure maps obtained from {\it XMM-Newton}
594: observations of the Coma galaxy cluster show a scale-invariant
595: pressure fluctuation spectrum on the scales of $40$ to $90\textrm{
596: kpc}$, which was analyzed to place a lower limit on the fractional
597: turbulent pressure of $\epsilon_ {\rm tur}\gtrsim 0.1$
598: \citep{Schuecker:04}.\footnote{Turbulence in the ICM can be detected
599: in other clusters given an X-ray detector with high spectral
600: resolution \citep{Inogamov:03,Sunyaev:03}.}
601:
602:
603: The physics of the interaction of a shock wave with a turbulent
604: upstream is complex and poorly understood even in unmagnetized,
605: ideal fluids. The primary theoretical uncertainties are the amount of
606: amplification of turbulence in the shock and the affect of
607: turbulence on the shock structure. The shock transition becomes
608: nonplanar in the presence of turbulence \citep[e.g.,][]{Rotman:91},
609: and this nonplanarity affects the local shock jump conditions that
610: short-wavelength fluctuations experience while crossing the shock
611: \citep[e.g.,][]{Zank:02}. Different analytical approximations (e.g.,
612: the rapid distortion theory and the local interaction analysis) and
613: direct numerical simulations do not always agree with each other
614: \citep [e.g.,][and references therein]{Andreopoulos:00}. Therefore
615: we do not attempt to include turbulent pressure in our quantitative
616: calculations. However, since turbulent pressure can affect the
617: observed jump conditions, we discuss it qualitatively using a simple
618: model \citep{Lele:92} based on the rapid distortion theory applied
619: to homogeneous turbulence \citep[see, e.g.,][]{Batchelor:53,Jacquin:93}.
620:
621:
622: \citet{Lele:92} derives the averaged density, momentum and energy
623: conservation equations of ideal fluid in the shock frame (his
624: equations $9-11$). Assuming homogeneous turbulence and a
625: cylindrically symmetric distribution of turbulent fluctuations, the
626: conservation equations are reduced to the form of equations
627: (\ref{EQ:shock_jump}) with a turbulent pressure and effective
628: adiabatic index of
629: \begin{eqnarray}
630: % \nonumber to remove numbering (before each equation)
631: P_{{\rm tur},i} = \overline{\rho_i}\widetilde{v''_{\|,i}
632: v''_{\|,i}} , \ \ \ \
633: %\\\nonumber
634: \gamma_{\rm tur} = \frac{3+2b_i}{1+2b_i} .
635: \end{eqnarray}
636: Here, the notation is such that any fluctuating quantity $f$ is
637: decomposed in two ways $f=\overline{f}+f'=\widetilde{f}+f''$, with
638: $\overline{f}$ denoting the average value of the quantity,
639: $\widetilde{f}\equiv \overline{\rho f}/\overline{\rho}$ denoting the
640: mass-weighted average, and $f'$ and $f''$ denoting the corresponding
641: fluctuating parts, while as before, $i=({\rm u},{\rm d})$. Just as we did for
642: the magnetic field (see \S~\ref{sec:magnetic_field}), we use
643: cylindrical symmetry to parameterize the turbulent field with a
644: single parameter $b_i=\widetilde{v''_{\perp,i}
645: v''_{\perp,i}}/(2\widetilde{v''_{\|,i} v''_{\|,i}})$, which in the
646: isotropic case equals unity.
647:
648:
649: The temperature jump $\tau$ and the shock Mach number ${\cal M}$ can
650: be calculated for a given upstream turbulent pressure fraction
651: $\epsilon_{\rm tur,u} \equiv P_{\rm tur,u}/P_{\rm u}$ and anisotropy
652: parameter $b_{\rm u}$ if the amplification of the turbulence in the
653: shock is known.\footnote{In case of isotropic turbulence
654: $\epsilon_{\rm tur}=\gamma_{\rm g}{\cal M}_{\rm
655: tur}^2/(3+\gamma_{\rm g}{\cal M}_{\rm tur}^2)$, where ${\cal M}_{\rm
656: tur}$ is the turbulence Mach number and $\gamma_{\rm g}$ is the
657: thermal gas adiabatic index.} \cite{Lele:92} uses the rapid
658: distortion theory to derive the shock amplification. This
659: approximation assumes that the mean turbulent flux amplitudes are
660: much smaller than their mean flow counterparts, that turbulent
661: fluctuations cross the shock much faster than the corresponding eddy
662: turnaround times, and that the mean flow does not vary much on the
663: length scale of an eddy. In particular, the otherwise planar shock
664: transition is assumed to have not been distorted, and rendered
665: nonplanar, by the fluctuations. In this theory, the parallel and
666: perpendicular turbulence is amplified according to \citep{Lele:92}
667: \begin{eqnarray}
668: % \nonumber to remove numbering (before each equation)
669: \frac{\widetilde{v''_\| v''_\|}_{\rm u}}{\widetilde{v''_\|
670: v''_\|}_{\rm d}} &=&
671: \frac{3}{4}r^2\left[\frac{1}{r^2-1}+\frac{r^2-2}{(r^2-1)^{3/2}}\tan^{-1}\sqrt{r^2-1}\right]\nonumber\\&\approx& \frac{6r-1}{5},\nonumber \\
672: \frac{\widetilde{v''_\perp v''_\perp}_{\rm u}}{\widetilde{v''_\perp
673: v''_\perp}_{\rm d}} &=&
674: \frac{3}{8}\left[1-\frac{1}{r^2-1}+\frac{r^4}{(r^2-1)^{3/2}}\tan^{-1}\sqrt{r^2-1}\right]
675: \nonumber\\&\approx& \frac{r+1}{2},
676: \end{eqnarray}
677: which in the case of isotropic upstream turbulence, $b_{\rm u}=1$
678: and $\gamma_{\rm tur,u}=5/3$, yields effective adiabatic index
679: and downstream pressure fraction for turbulence
680: \begin{eqnarray}
681: % \nonumber to remove numbering (before each equation)
682: \gamma_{\rm tur,d} &\approx& \frac{23r+2}{11r+4} ,\\\nonumber %~~; ~~\left(b_{\rm d} \approx \frac{5(r+1)}{2(6r-1)}\right)\\\nonumber
683: \epsilon_{\rm tur,d} &\approx& \frac{\epsilon_{\rm
684: tur,u}(6r-1)}{5\tau(1-\epsilon_{\rm tur,u})+\epsilon_{\rm
685: tur,u}(6r-1)} .
686: \end{eqnarray}
687:
688:
689: In this simple model of amplification of turbulence in the shock,
690: the effect of a pre-existing turbulent component is similar to that
691: of the pre-existing cosmic rays and magnetic pressure, namely, the
692: upstream turbulence is only weakly, positively correlated with
693: $\tau$ and ${\cal M}$. For example, assuming that turbulence is the
694: only nonthermal component, the temperature jump increases by a
695: factor $\approx 1.15$ and ${\cal M}$ increases by a factor $\approx
696: 1.3$ between $\epsilon_{\rm tur,u}=0$ and $\epsilon_{\rm tur,u}=0.3$
697: for compression ratio $r=2.3$. Very similar results are obtained if
698: instead of the rapid distortion theory, we employ the linear
699: interaction analysis to calculate the jump conditions
700: \citep[e.g.,][]{Lee:93,Lee:97}. Therefore, we conclude that for a
701: weak, subsonic, unmagnetized turbulence of the ICM, the effect of
702: turbulence on the thermal gas temperature jump of the shock is
703: similar to that expected in the presence of other nonthermal
704: components.
705:
706: \section{Results}
707: \label{sec:results}
708: \subsection{A520}
709:
710: \cite{Markevitch:05} analyze a $67$ kilosecond observation with {\it
711: Chandra} ACIS-I of the bow shock in the galaxy cluster merger A520
712: at z=0.203 and estimate the density and temperature jump along the
713: axis of symmetry of the shock. The upstream and downstream
714: temperatures are, respectively, $T_{\rm u}=4.8_{-0.8}^{+1.2}\textrm{
715: keV}$ and $T_{\rm d}=11.5_{-3.1}^{+6.7}\textrm{ keV}$, while the
716: density jump is $r=2.3 \pm 0.3$ (all errors are at the $90\%$
717: confidence levels). These values are consistent with a shock Mach
718: number of ${\cal M}\approx 2$. We use equation (\ref{EQ:Tjump}) to
719: explore the constraints that can be placed on the acceleration of
720: particles in the shock and on the fractional pressure in the
721: nonthermal components in the shock upstream, and equation
722: (\ref{EQ:Tjump_gen}) to constrain energy leakage from the shock in
723: A520.
724:
725: The predicted downstream temperature in the case of a pure thermal
726: gas, $T_{\rm d}=10\textrm{ keV}$ for $T_{\rm u}=4.8\textrm{ keV}$
727: and $r=2.3$, is consistent with the measured value. However, given
728: that the expected temperature is below the median measured
729: temperature, little room is left for significant particle
730: acceleration in the shock. To place constraints on $\epsilon_{\rm
731: acc}$, we first take $\epsilon_{B,i}=Q=0$, and carry a Monte Carlo
732: search in the remaining parameter space $(\epsilon_{\rm
733: CR,u},\epsilon_{\rm acc})$. We draw a large set ($10^5$) of the
734: observed parameters (density and temperature jump) from the observed
735: distributions.\footnote{We approximate each observed distribution by
736: two half-gaussians that peak at the median observed value and
737: satisfy the $90\%$ confidence range reported by
738: \citet{Markevitch:05}.} For each set of observed values we find all
739: the combinations of $(\epsilon_{\rm CR,u},\epsilon_{\rm acc})$ in
740: the domain $(0<\epsilon_{\rm CR,u}<0.3,0<\epsilon_{\rm acc}<0.25)$
741: that are compatible with the observations. For each point in the
742: $(\epsilon_{\rm CR,u},\epsilon_{\rm acc})$ plane, we calculate the
743: number of instances that the corresponding shock is compatible with
744: the generated ``observed'' values of $r$ and $\tau$. The resulting
745: number, properly normalized, provides the Bayesian likelihood of the
746: shock being characterized by a given pair $(\epsilon_{\rm
747: CR,u},\epsilon_{\rm acc})$, assuming a uniform prior in the domain
748: considered here.
749:
750: Figure \ref{FIG A520} shows the probability distribution for
751: $(\epsilon_{\rm CR,u},\epsilon_{\rm acc})$ for the relativistic
752: cosmic rays ($\gamma_{\rm CR}=4/3$). The two contours divide the
753: plane so that the cumulative probability constrained above each is
754: $0.33$ and $0.05$ of the total. The figure shows that $\epsilon_{\rm
755: acc} \lesssim 0.1$ at $95\%$ confidence levels for any value of
756: $\epsilon_{\rm CR,u}$, and that current observations do not provide
757: a constraint on $\epsilon_{\rm CR,u}$. Carrying out a similar
758: analysis for relativistic electrons and Newtonian proton cosmic rays
759: ($\gamma_{\rm CR}=13/9$), which would be expected if the cosmic ray
760: pressure were dominated by particles with typical energies of
761: $10^1-10^4$ times the thermal energy, yields an upper limit of
762: $\epsilon_{\rm acc} \lesssim 0.15$. Finally, the limit on the
763: efficiency of the acceleration of Newtonian cosmic rays,
764: $\gamma_{\rm CR}=5/3$, as expected if the cosmic rays are
765: accelerated only to several times the thermal temperature, is
766: $\epsilon_{\rm acc} \lesssim 0.25$. Naturally, allowing for $Q>0$
767: when deriving the constraints on $\epsilon_{\rm acc}$ yields a
768: tighter upper limit on $\epsilon_{\rm acc}$. The minor effect of
769: varying upstream magnetic field and upstream cosmic ray pressure on
770: the limits that can be placed on $\epsilon_{\rm acc}$ is explored in
771: Figure \ref{Fig eps_acc_A520}.
772:
773: \begin{figure}
774: \includegraphics[width=9cm]{f1.eps}
775: \caption{ \label{FIG A520} The probability distribution of
776: $(\epsilon_{\rm CR,u}$,$\epsilon_{\rm acc})$ for relativistic cosmic
777: rays in the bow shock in the galaxy cluster A520, assuming
778: $\epsilon_{B,{\rm u}}=0$ and $\epsilon_{\rm CR,u}<0.3$. The contours
779: divide the plane so that the cumulative distribution above the
780: contour includes only 0.33 ({\it lower contour}) and 0.05 ({\it
781: upper contour}) of the total probability. The probability is
782: calculated via a Monte Carlo simulation (see text).}
783: \end{figure}
784:
785: \begin{figure}
786: \includegraphics[width=9cm]{f2.eps}
787: \caption{ \label{Fig eps_acc_A520} The marginalized probability
788: distribution of $\epsilon_{\rm acc}$ for several different
789: nonthermal contributions to the gas pressure. The probability is
790: calculated via a Monte Carlo simulation (see text) for various
791: ranges of flat priors on $\epsilon_{\rm CR,u}$ and $\epsilon_{B,{\rm
792: u}}$, and various levels of magnetic field amplification in the
793: shocks, as indicated in the legend. The cosmic rays are relativistic
794: ($\gamma_{\rm CR}=4/3$), unless noted otherwise. The solid and
795: dashed lines contain $90\%$ of the distributions (the rest is in
796: dotted line tail). This figure shows that the limits on
797: $\epsilon_{\rm acc}$ are insensitive to the assumptions and the
798: priors that we choose for the upstream nonthermal components.}
799: \end{figure}
800:
801: Repeating the same analysis for various values of $0<\epsilon_{B,{\rm
802: u}}<0.3$ while assuming $\epsilon_{\rm CR,u}=0$, or taking a constant
803: $\epsilon_{\rm CR,u}$ and a flat prior on $\epsilon_{B,{\rm u}}$ in
804: the same range, for the two limiting forms magnetic field behavior at the shock
805: ($B_{\bot,{\rm d}}=rB_{\bot,{\rm u}}$ and $B_{\rm d}=r^{2/3}B_{\rm
806: u}$; see \S~\ref{sec:magnetic_field}), yields a similar upper
807: limit on $\epsilon_{\rm acc}$. The reason for the weak impact of
808: magnetic pressure on the constraints that can be placed
809: on particle acceleration is the
810: strong dependence of $\tau$ on $\epsilon_{\rm acc}$ and its weak
811: dependence on $\epsilon_{B,{\rm u}}$. Therefore, just as we found for
812: $\epsilon_{\rm CR,u}$, current observations do not provide a significant
813: constraint on $\epsilon_{B,{\rm u}}$.
814:
815: We also carried out the same analysis assuming no particle
816: acceleration, $\epsilon_{\rm acc}=0$, but allowing for an escape of
817: cosmic rays from the shock, in order to constrain energy leakage
818: from the shock, $Q$. Although this scenario is artificial (since for
819: $\epsilon_{\rm acc}=0$ we also expect $Q=0$), this analysis provides
820: an upper limit to the value of $Q$. We find that the observation of
821: A520 limit $Q\lesssim 0.1$ in its merger shock.
822:
823: The dependence of the temperature jump in equation (\ref{EQ:Tjump})
824: on $\epsilon_{\rm CR,u}$ and $\epsilon_{B,{\rm u}}$ is weak.
825: Therefore current observations do not put significant constraint the
826: presence of a relativistic component in the pre-merger ICM. Figure
827: \ref{FIG A520} shows that a future, improved X-ray spectroscopy
828: across the bow shock, such as with a longer exposure with {\it
829: Chandra} or {\it XMM-Newton}, may exclude the purely thermal
830: scenario, $\epsilon_{\rm acc}=\epsilon_{\rm CR,u}=0$. If measured
831: value of $\tau$ falls below the thermal prediction, our analysis
832: will yield a lower limit on $\epsilon_{\rm acc}$ to accompany the
833: current upper limit. If the measured value of $\tau$ falls above the
834: thermal prediction, the measurement will imply a lower limit on
835: $\epsilon_{\rm CR,u}$. For example, if we artificially reduce the
836: present measurement uncertainties in $T_{\rm u}$, $T_{\rm d}$, and
837: $r$ by a factor of $3$ while assuming that the mean values of these
838: observables remained unchanged, the data would require
839: $\epsilon_{\rm acc}<0.05$ and $\epsilon_{\rm nt,u}>0.05$ at $95\%$
840: confidence levels. However decoupling $\epsilon_{\rm nt,u}$ into its
841: cosmic ray and magnetic field components cannot be accomplished
842: given a measurement of $r$ and $\tau$ alone.
843:
844: From the density jump data alone and assuming a purely thermal
845: shock, \cite{Markevitch:05} estimate the Mach number of the merger
846: shock in A520 to be ${\cal M} =2.1_{-0.3}^{+0.4}$. As we discuss in
847: \S~\ref{sec:mach_number}, the likely presence of a non-negligible
848: nonthermal pressure requires a modification of the Mach number
849: estimate as in equation (\ref{eq:mach}); adding a nonthermal
850: pressure in the upstream increase the Mach number. Figure \ref{Fig
851: Mach} shows the Mach number probability for several scenarios in
852: which a nonthermal fluid is present with a fractional contribution
853: to the pressure that is allowed by the present observations. The
854: true value of the Mach number depends on the fractional nonthermal
855: pressure and the amplification of the magnetic field and turbulence
856: in the shock. For example, for $\epsilon_{\rm CR,u}=\epsilon_{B,{\rm
857: u}}=0.15$ and assuming significant magnetic field amplification in
858: the shock, the true Mach number can be as high as ${\cal M} \approx
859: 2.7$
860:
861: \subsection{1E 0657$-$56}
862:
863: Gas properties across the merger shock in 1E 0657$-$56 (the
864: ``bullet'' cluster at z=0.296) were measured by \cite{Markevitch:06}
865: using a $500$ kilosecond observation with {\it Chandra} ACIS-I. They
866: find a density jump of $r \approx 3$, which corresponds to a Mach
867: number of ${\cal M} \approx 3$. The measured temperatures are
868: $T_{\rm u} \approx 9\textrm{ keV}$ and a lower limit $T_{\rm
869: d}>32\textrm{ keV}$ at $1\sigma$ confidence levels. The shock in 1E
870: 0657$-$56 is stronger than that in A520 and is thus more propitious
871: for detecting particle acceleration. Unfortunately, the high
872: downstream temperature $T_{\rm d}$ complicates an accurate
873: measurement of the temperature jump with the high resolution X-ray
874: telescopes ${\it Chandra}$ and ${\it XMM-Newton}$.
875:
876: \cite{Markevitch:06} does not report the errors on some of the
877: measurements and therefore we cannot quantitatively constrain the
878: presence of nonthermal components in the shock upstream and
879: downstream. Figure \ref{FIG bullet} shows the model prediction for
880: $T_{\rm d}$ as a function of $\epsilon_{\rm CR,u}$ and
881: $\epsilon_{\rm acc}$ for the case of relativistic cosmic rays,
882: $\gamma_{\rm CR}=4/3$, assuming that $\epsilon_{B,{\rm u}}=Q=0$ and
883: taking $r=3$ and $T_{\rm u}=9\textrm{ keV}$. The figure shows that
884: constraints that can be made in 1E 0657$-$56 are qualitatively
885: similar to, although less stringent than, those in A520. The minimum
886: value of the temperature jump $\tau$ allowed by the measurement is
887: high and barely consistent with a purely thermal shock. It does not
888: leave much room for particle acceleration; we tentatively infer
889: $\epsilon_{\rm acc}<0.15$.
890:
891: The nonthermal components may affect also the Mach number of the
892: bullet cluster merger shock, which recently stirred a discussion about
893: its compatibility with standard cosmological models
894: \citep{Hayashi:06,Farrar:06,Milosavljevic:07,Springel:07}.
895: \cite{Markevitch:06} finds ${\cal M}=3.0 \pm 0.4$ , which corresponds to
896: $r=3.0^{+0.17}_{-0.23}$, assuming no nonthermal contribution and
897: neglecting relativistic corrections due to the high electron
898: temperature. Allowing for cosmic rays with pressure up to
899: equipartition, $0<\epsilon_{\rm CR,u}<0.3$ and $\epsilon_B=0$, while
900: requiring temperature and density jump
901: consistent with observations, $T_{\rm d}> 20\textrm{ keV}$
902: and $2.77<r<3.17$, yields the limits
903: $2.3<{\cal M}<3.7$. The lower bound
904: is obtained for $r=2.77$, $\epsilon_{\rm acc}=0.07$,
905: $\epsilon_{\rm CR,u}=0$, and $T_{\rm d}=20\textrm{ keV}$. The upper bound
906: corresponds to $r=3.17$, $\epsilon_{\rm acc}=0$, $\epsilon_{\rm
907: CR,u}=0.3$, and $T_{\rm d}=45\textrm{ keV}$.
908:
909:
910:
911:
912:
913: \begin{figure}
914: \includegraphics[width=9cm]{f3.eps}
915: \caption{ \label{Fig Mach} The Mach number probability distribution
916: based on the observed density and temperature jumps across the shock
917: in A520 for several different contributions of nonthermal components
918: to the gas pressure. The solid lines contain $90\%$ of the
919: distributions (the rest is in dotted line tails). In all cases the
920: cosmic rays are relativistic and the upstream magnetic field is
921: isotropic ($\gamma_{\rm CR}=\gamma_{B,{\rm u}}=4/3)$. The legend
922: indicates the fractional upstream nonthermal pressure and the shock
923: amplification of the field in each case.}
924: \end{figure}
925:
926: \begin{figure}
927: \includegraphics[width=9cm]{f4.eps}
928: \caption{ \label{FIG bullet} $T_{\rm d}$ as a function of
929: $\epsilon_{\rm CR,u}$ and $\epsilon_{\rm acc}$ ($\gamma_{\rm
930: CR}=4/3$) for $r=3$ and $T_{\rm d}=9\textrm{ keV}$, which correspond
931: to the values reported for the bullet cluster 1E 0657$-$56. The
932: $32\textrm{ keV}$ and $20 \textrm{ keV}$ contours are the $1\sigma$
933: and $2\sigma$ lower limits on $T_{\rm d}$ as measured by
934: \cite{Markevitch:06} (the $2\sigma$ limit is a rough estimate
935: derived from the plots in \citealt{Markevitch:06}).}
936: \end{figure}
937:
938: \section{Discussion}
939: \label{sec:discussion}
940:
941: \subsection{Particle Acceleration in Collisionless Shocks}
942: \label{sec:accel}
943:
944: Current measurements place constraints on the efficiency of particle
945: acceleration in cluster merger shocks. This offers a very unique
946: opportunity to test models of particle acceleration in astrophysical
947: collisionless shocks. Estimating the acceleration efficiency in
948: collisionless shocks is a difficult problem, which is severely
949: complicated by the fact that it remains unknown which among a number
950: of possibilities is the primary acceleration mechanism. Even in the
951: leading candidate mechanism, the diffusive shock acceleration,
952: estimates of acceleration efficiency range widely because of a
953: number fundamental theoretical uncertainties, concerning the
954: fraction of thermal particles that are injected into the
955: acceleration process \citep[e.g.,][and references
956: therein]{Malkov:01}, the nonlinear influence of the accelerated
957: particles on the hydrodynamic profile of the shock wave
958: \citep[e.g.,][]{Drury:81,Achterberg:84,Giacalone:97,Kang:02,Kang:07a},
959: and the amplification of the magnetic field by plasma instabilities
960: \citep
961: [e.g.,][]{Lucek:00,Bell:01,Schlickeiser:03,Bell:04,Bell:05,Schekochihin:05,Medvedev:06}.
962: Therefore, any estimate of particle acceleration efficiency is
963: strongly affected by the specific assumptions and approximations
964: employed in a self-consistent shock model. Here we do not attempt to
965: carry out a detailed comparison of our results with various
966: scenarios for particle acceleration. But, to demonstrate the power
967: of the constraints that can be obtained from merger shock dynamics,
968: we discuss our results in the context of the predictions of the
969: particle acceleration model of \citet[][see also
970: \citealt{Kang:07b}]{Kang:07a}.
971:
972: A variety of investigations point to a strong dependence of
973: acceleration efficiency on the shock Mach number, whereby stronger
974: shocks produce higher efficiencies. \citet{Kang:07a} report an
975: investigation of acceleration efficiency behavior in diffusive shock
976: acceleration simulations in quasi-parallel shocks with a Bohm
977: diffusion coefficient, a self-consistent treatments of particle
978: injection from the thermal pool into the acceleration process, and
979: Alfv\'en wave propagation. In their Figure 5, \citet{Kang:07a} plot
980: the dependence of the acceleration efficiency $\eta ({\cal M})$ on
981: the shock Mach number ${\cal M}$, where $\eta$ is defined as the
982: energy flux in downstream cosmic rays, divided by the bulk kinetic
983: energy flux in the upstream medium entering the shock. In the limit
984: $\epsilon_{\rm acc}\ll1$, the parameter $\eta$ in Kang et al. is
985: related to our $\epsilon_{\rm acc}$ via
986: \begin{equation}
987: \eta \approx \frac{3}{10} \frac{({\cal M}^2+3)(5{\cal M}^2-1)}{{\cal
988: M}^4} \epsilon_{\rm acc}\ \ \ \ (\epsilon_{\rm acc} \ll 1) .
989: \end{equation}
990: Thus, we have $\eta\approx 2.5\epsilon_{\rm acc}$ for ${\cal M}=2$ and
991: $\eta\approx 2\epsilon_{\rm acc}$ for ${\cal M}=3$.\footnote{We have
992: here ignored the modification of shock Mach number by the nonthermal
993: pressure, see \S~\ref{sec:mach_number}.} \citet{Kang:07a} detect a
994: strong dependence of $\eta$ on the presence of preexisting cosmic rays
995: in the shock upstream, $\epsilon_{\rm CR,u}$. The strong dependence
996: can be attributed to inefficient injection at low Mach numbers in
997: their model. For ${\cal M}=2$, the parameter $\eta$ jumps from $0$ to
998: $0.15$ as $\epsilon_{\rm CR,u}$ increases from $0$ to $0.23$. In
999: stronger shocks, for ${\cal M}=3$, the parameter $\eta $ jumps from
1000: $0.1$ to $0.25$ for the same increase in $ \epsilon_{\rm CR,u}$.
1001:
1002: Converting our current limits on $\epsilon_{\rm acc}$ into limits on
1003: $\eta$, we find $\eta \lesssim 0.2$ for ${\cal M} \approx 2$ and $\eta
1004: \lesssim 0.3$ for ${\cal M} \approx 3$. These limits are marginally
1005: consistent with the predictions of \citet{Kang:07a} for any assumed
1006: value of $\epsilon_{\rm CR,u}$. However, if as \citet{Kang:07a} argue,
1007: $\epsilon_{\rm acc}$ is itself a sensitive function of $\epsilon_{\rm
1008: CR,u}$, then improved measurements of $\epsilon_{\rm acc}$ that can be
1009: obtained with additional observations with existing X-ray telescopes
1010: can provide tighter limits on $\epsilon_{\rm CR,u}$. Moreover, a
1011: positive measurement of $\epsilon_{\rm acc}$ can provide a
1012: model-dependent constraint on $\epsilon_{\rm CR,u}$.
1013:
1014:
1015: \subsection{X-ray and SZE Cluster Mass Estimates}
1016: \label{sec:mass}
1017:
1018: Galaxy cluster surveys in which the cluster masses are measured
1019: accurately can be used as powerful cosmological probes of dark matter
1020: and dark energy. The mass estimates are plagued by systematic
1021: uncertainties that must be understood and quantified before the
1022: requisite mass measurement accuracy is achieved. Nonthermal pressure
1023: due to cosmic rays, magnetic fields, and turbulence, is a source of a
1024: systematic bias when cluster masses are estimated on the basis of the
1025: assumption of hydrostatic equilibrium between gravitational forces and
1026: thermal pressure gradients in the ICM \citep*[e.g.,][and references
1027: therein]{Ostriker:05,Rasia:06,Nagai:07a}. The hydrostatic mass
1028: profile of a spherically-symmetric cluster is given by
1029: \begin{equation}
1030: M(<r) = \frac{-r^2}{G\rho_{\rm g}} \left( \frac{dP_{\rm
1031: g}}{dr}+\frac{dP_{\rm nt}}{dr} \right),
1032: \label{eq:HSE}
1033: \end{equation}
1034: where $M(<r)$ is the mass enclosed within radius $r$, while $P_{\rm
1035: g}$ and $P_{\rm nt}$ are the thermal and the nonthermal
1036: contributions to the pressure. The thermal gas provides a
1037: significant fraction of the total pressure support, and this
1038: pressure is measured directly with current X-ray and SZE
1039: observations. The contribution of the nonthermal pressure, on the
1040: other hand, is customarily assumed to be relatively small ($\lesssim
1041: 10\%$) outside of a cluster core \citep[see e.g.,][]{Nagai:07b}, and
1042: it is often ignored in the hydrostatic mass estimates based on X-ray
1043: and SZE data. However, present observations do not yet constrain the
1044: nonthermal pressure in the regime in which it dramatically affects
1045: the calibration of the hydrostatic mass estimates. If not accounted
1046: for, these nonthermal biases limit the effectiveness of upcoming
1047: X-ray and SZE cluster surveys to accurately measure the expansion
1048: history of the universe. Detailed investigations of the sources of
1049: nonthermal pressure in clusters are thus critical for understanding
1050: their effect on the properties of the ICM and the utility of
1051: clusters as precision cosmological probes. We proceed to discuss how
1052: future observations of cluster merger shocks will have the potential
1053: to place unique constraints on the nonthermal pressure in the
1054: unshocked ICM, thereby improving cluster mass estimates.
1055:
1056: \subsection{Prospects for Future Constraints of Nonthermal Pressure}
1057: \label{sec:future}
1058:
1059: As discussed in \S\ref{sec:results}, current measurements alone do
1060: not place strong constraints on the presence of a nonthermal
1061: component in the unshocked ICM in both systems. However, the
1062: improved constraints that can be obtained with existing X-ray
1063: telescopes can provide useful lower limits on nonthermal pressure
1064: and their effects on the X-ray and SZE cluster mass estimates. In
1065: the case of the shock in the cluster merger A520, the current
1066: constraints, which are based on a 67 kilosecond observation with
1067: \emph{Chandra}, can be improved significantly with follow-up
1068: observations with \emph{Chandra} or \emph{XMM-Newton}. Therefore, we
1069: identify this system as the most promising one in which our method
1070: may yield a positive shock-hydrodynamic detection of a nonthermal
1071: component. While the stronger shock in the cluster merger 1E
1072: 0657$-$56 may be more efficient at accelerating particles, it will
1073: be more difficult to improve the measurement of the shock
1074: temperature jump in 1E 0657$-$56. This is because the very high
1075: temperature of the shock downstream medium ($T_{\rm d}\approx
1076: 30-50\textrm{ keV}$) lies far outside of spectral sensitivity window
1077: of X-ray telescopes with arcsecond resolution, which in turn renders
1078: it difficult to measure the temperature jump in the narrow
1079: post-shock layer. Hard X-ray observations
1080: \citep[e.g.,][]{Petrosian:06} with {\it RXTE}, \emph{Integral}, or
1081: {\it Suzaku}, combined with a model of ICM fluid flow
1082: \citep[e.g.,][]{Milosavljevic:07,Springel:07}, may help to pin down
1083: the downstream temperature. Alternative intriguing possibility is
1084: that future high resolution SZE observations may be able to measure
1085: the downstream temperature on the basis of the relativistic SZE,
1086: which should be prominent in the high temperature downstream.
1087:
1088: In the next few years, gamma-ray observations of galaxy clusters may
1089: provide tight constraints on the fractional contribution of nonthermal
1090: particles to the pressure of the ICM. Assuming that gamma-ray emission
1091: from the decay of neutral pions is the primary emission channel,
1092: measurements with the new gamma-ray telescope \emph{GLAST} can be used
1093: to place population-averaged limits on the hadronic cosmic-ray
1094: pressure support in clusters \citep[][see also
1095: \citealt{Berrington:03,Blasi:07}]{Ando:07}. These forthcoming
1096: constraints, combined with improved X-ray spectroscopic measurements
1097: of the jump conditions in merger shocks, may enable a separation of
1098: the upstream nonthermal pressure into its constituent components.
1099:
1100: Finally, comparisons of the hydrostatic mass estimates with those
1101: derived from gravitational lensing observations can, in principle,
1102: provide an important handle on nonthermal biases. Note that this
1103: approach is not practical for individual clusters, because lensing
1104: measures the mass in a projected aperture that cannot be directly
1105: compared to the mass within a sphere of the same radius, to which
1106: the hydrostatic mass is sensitive. But, it might be possible to
1107: compare different estimators in an average sense, while accounting
1108: for the effects of asphericity of clusters and projection effects
1109: \citep[see \S~5.2 in][]{Nagai:07a}.
1110:
1111: \section{Conclusions}
1112: \label{sec:conclusions}
1113:
1114: We model the effect of particle acceleration and nonthermal pressure
1115: components on shock jump conditions in nonrelativistic shocks. We
1116: focus on intermediate Mach number shocks, with Mach numbers in the
1117: range ${\cal M}=2-3$. We apply this to the merger shocks in galaxy
1118: clusters A520 and 1E 0657$-$56 and place the first constraints on the
1119: efficiency of particle acceleration in these shocks. Our main results
1120: are as follows.
1121:
1122: 1. The temperature jump of the thermal gas in the shock depends
1123: strongly on the efficiency of shock particle acceleration.
1124: Efficient acceleration can reduce the temperature jump by more than
1125: a factor of two for a constant compression ratio in the range
1126: $r=2-3$.
1127:
1128: 2. The correct effect of nonthermal pressure in the upstream, such as
1129: fossil cosmic rays, magnetic field, and turbulence, on the shock jump
1130: observed in the thermal gas cannot be derived at this point, because
1131: we lack an understanding of the interaction of these components with
1132: the shock. However, for a wide range of reasonable assumptions and
1133: analytic approximations, we find that nonthermal pressure in the
1134: unshocked ICM has only a minor effect on the downstream temperature
1135: (at a fixed compression ratio), and that in general, a high upstream
1136: nonthermal pressure increases the temperature jump in the thermal gas.
1137:
1138: 3. The combination of strong dependence of the temperature jump on
1139: particle acceleration and weak dependence on upstream nonthermal
1140: pressure enables derivation of meaningful constraints on the
1141: efficiency of particle acceleration in cluster merger shocks, even
1142: with current observations. Future, more accurate X-ray and SZE
1143: observations of these shocks may yield meaningful constraints on the
1144: upstream nonthermal pressure as well.
1145:
1146: 4. Nonthermal pressure and shock particle acceleration can also affect
1147: the Mach number that is inferred from the observed compression ratio
1148: $r$ by tens of percent. When the temperature jump is poorly
1149: constrained, the Mach number is anticorrelated with efficient
1150: particle acceleration and is positively correlated with upstream
1151: nonthermal pressure.
1152:
1153: 5. In the two observed high contrast galaxy cluster merger shocks,
1154: A520 and 1E 0657$-$56, we constrain the efficiency of acceleration
1155: of relativistic particles to be $\epsilon_{\rm acc} \lesssim 0.1$
1156: and $\epsilon_{\rm acc} \lesssim 0.15$, respectively. We find that
1157: considerable upstream pressure can increase the Mach number of the
1158: shock in A520 to reach ${\cal M} \approx 2.7$, much higher than the
1159: inferred value $2.1$ obtained assuming an absence of nonthermal
1160: components. The true Mach number of the shock in 1E 0657$-$56 can
1161: be in the range $2.3<{\cal M}<3.7$ with the compression ratio of
1162: $r=3$ allowing for nonthermal pressure components.
1163:
1164: \acknowledgements
1165:
1166: The research was supported in part by the Sherman Fairchild
1167: Foundation. We would like to thank A. K\"onigl, E. Komatsu,
1168: A. Kravtsov, P. Kumar, Y. Rephaeli, and J. Scalo for helpful
1169: discussions.
1170:
1171: \begin{thebibliography}{}
1172:
1173: \bibitem[Achterberg et al.(1984)]{Achterberg:84} Achterberg, A., Blandford, R., \& Periwal, V.\ 1984, \aap, 132, 97
1174:
1175: \bibitem[Aharonian et al.(2007)]{Aharonian:07} Aharonian, F., et al.\ 2007, \aap, 464, 235
1176:
1177: \bibitem[Ando \& Nagai (2007)]{Ando:07} Ando, S. \& Nagai, D. \ 2007, preprint (arXiv/0705.2588)
1178:
1179: \bibitem[Andreopoulos et al.(2000)]{Andreopoulos:00} Andreopoulos, Y., Agui, J.~H., \& Briassulis, G.\ 2000, Annual Review of Fluid
1180: Mechanics, 32, 309
1181:
1182: \bibitem[Batchelor(1953)]{Batchelor:53} Batchelor, G.~K.\ 1953, The Theory of Homogeneous Turbulence, Cambridge: Cambridge University Press, 1953,
1183:
1184: \bibitem[Berezinsky et al.(1997)]{Berezinsky:97} Berezinsky, V.~S., Blasi, P., \& Ptuskin, V.~S.\ 1997, \apj, 487, 529
1185:
1186: \bibitem[Berrington \& Dermer(2003)]{Berrington:03} Berrington, R.~C., \& Dermer, C.~D.\ 2003, \apj, 594, 709
1187:
1188: \bibitem[Bell(1978)]{Bell:78} Bell, A.~R.\ 1978, \mnras, 182, 147
1189:
1190: \bibitem[Bell(2004)]{Bell:04} Bell, A.~R.\ 2004, \mnras, 353, 550
1191:
1192: \bibitem[Bell(2005)]{Bell:05} Bell, A.~R.\ 2005, \mnras, 358, 181
1193:
1194: \bibitem[Bell \& Lucek(2001)]{Bell:01} Bell, A.~R., \& Lucek, S.~G.\ 2001, \mnras, 321, 433
1195:
1196: \bibitem[Blandford \& Eichler(1987)]{Blandford:87} Blandford, R., \& Eichler, D.\ 1987, \physrep, 154, 1
1197:
1198: \bibitem[Blandford \& Ostriker(1978)]{Blandford:78} Blandford, R.~D., \& Ostriker, J.~P.\ 1978, \apjl, 221, L29
1199:
1200: \bibitem[Blasi et al.(2007)]{Blasi:07} Blasi, P., Gabici, S., \& Brunetti, G.\ 2007, preprint (astro-ph/0701545)
1201:
1202: \bibitem[Carilli \& Taylor(2002)]{Carilli:02} Carilli, C.~L., \& Taylor, G.~B.\ 2002, \araa, 40, 319
1203:
1204: \bibitem[Dolag et al.(2005)]{Dolag:05} Dolag, K., Vazza, F., Brunetti, G., \& Tormen, G.\ 2005, \mnras, 364, 753
1205:
1206: \bibitem[Drury \& V\"olk(1981)]{Drury:81} Drury, L.~O., \& V\"olk, J.~H.\ 1981, \apj, 248, 344
1207:
1208: \bibitem[En\ss lin et al.(1997)]{Ensslin:97} En\ss lin, T.~A., Biermann, P.~L, Kronberg, P.~P., \& Wu, X.-P. \ 1997, \apj, 477, 560
1209:
1210: \bibitem[Faltenbacher et al.(2005)]{Faltenbacher:05} Faltenbacher, A., Kravtsov, A.~V., Nagai, D., \& Gottl{\"o}ber, S.\ 2005, \mnras, 358, 139
1211:
1212: \bibitem[Farrar \& Rosen(2006)]{Farrar:06} Farrar, G.~R., \& Rosen,
1213: R.~A.\ 2006, preprint (astro-ph/0610298)
1214:
1215: \bibitem[Fujita \& Sarazin(2001)]{Fujita:01} Fujita, Y., \& Sarazin,
1216: C.~L.\ 2001, \apj, 563, 660
1217:
1218: \bibitem[Fujita et al.(2004)]{Fujita:04} Fujita, Y., Matsumoto, T., \& Wada, K.\ 2004, \apjl, 612, L9
1219:
1220: \bibitem[Gabici \& Blasi(2003)]{Gabici:03} Gabici, S., \& Blasi, P.\
1221: 2003, \apj, 583, 695
1222:
1223: \bibitem[Giacalone et al.(1997)]{Giacalone:97} Giacalone, J., Burgess, D., Schwartz, S.~J., Ellison, D.~C., \& Bennett, L.\ 1997,
1224: \jgr, 102, 19789
1225:
1226: \bibitem[Govoni \& Feretti(2004)]{Govoni:04} Govoni, F., \& Feretti, L.\ 2004, International Journal of Modern Physics D, 13, 1549
1227:
1228: \bibitem[Govoni et al.(2006)]{Govoni:06} Govoni, F., Murgia, M., Feretti, L., Giovannini, G., Dolag, K., \& Taylor, G.~B.\ 2006, \aap,
1229: 460, 425
1230:
1231: \bibitem[Hayashi \& White(2006)]{Hayashi:06} Hayashi, E., \& White,
1232: S.~D.~M.\ 2006, \mnras, 370, L38
1233:
1234: \bibitem[Inogamov \& Sunyaev(2003)]{Inogamov:03} Inogamov, N.~A., \& Sunyaev, R.~A.\ 2003, Astronomy Letters, 29, 791
1235:
1236: \bibitem[Jacquin et al.(1993)]{Jacquin:93} Jacquin, L., Cambon, C., \& Blin, E.\ 1993, Physics of Fluids, 5, 2539
1237:
1238: \bibitem[Kang et al.(2002)]{Kang:02} Kang, H., Jones, T.~W., \& Gieseler, U.~D.~J.\ 2002, \apj, 579, 337
1239:
1240: \bibitem[Kang et al.(2007)]{Kang:07a} Kang, H., Ryu, D., Cen, R., \& Ostriker, J.~P.\ 2007, preprint (arXiv:0704.1521)
1241:
1242: \bibitem[Kang \& Jones(2007)]{Kang:07b} Kang, H., \& Jones,
1243: T.~W.\ 2007, preprint (arXiv:0705.3274)
1244:
1245: \bibitem[Konigl(1980)]{Konigl:80} Konigl, A.\ 1980, Physics of
1246: Fluids, 23, 1083
1247:
1248: \bibitem[Kunik et al.(2003)]{Kunik:03} Kunik, M., Qamar, S., \& Warnecke, G.\ 2003, Journal of Computational Physics, 187, 572
1249:
1250: \bibitem[Lee et al.(1993)]{Lee:93} Lee, S., Lele, S.~K., \& Moin, P.\ 1993, Journal of Fluid Mechanics, 251, 533
1251:
1252: \bibitem[Lee et al.(1997)]{Lee:97} Lee, S., Lele, S.~K., \& Moin, P.\ 1997, Journal of Fluid Mechanics, 340, 225
1253:
1254: \bibitem[Lele(1992)]{Lele:92} Lele, S.~K.\ 1992, Physics of Fluids, 4, 2900
1255:
1256: \bibitem[Lucek \& Bell(2000)]{Lucek:00} Lucek, S.~G., \& Bell, A.~R.\ 2000, \mnras, 314, 65
1257:
1258: \bibitem[Malkov \& Drury(2001)]{Malkov:01} Malkov, M.~A., \&
1259: O'C Drury, L.\ 2001, Reports of Progress in Physics, 64, 429
1260:
1261: \bibitem[Markevitch et al.(2002)]{Markevitch:02} Markevitch, M., Gonzalez, A.~H., David, L., Vikhlinin, A., Murray, S., Forman, W., Jones, C., \& Tucker, W.\ 2002, \apjl, 567, L27
1262:
1263: \bibitem[Markevitch et al.(2005)]{Markevitch:05} Markevitch, M., Govoni, F., Brunetti, G., \& Jerius, D.\ 2005, \apj, 627, 733
1264:
1265: \bibitem[Markevitch(2006)]{Markevitch:06} Markevitch, M.\ 2006, ESA
1266: SP-604: The X-ray Universe 2005, 723
1267:
1268: \bibitem[Medvedev et al.(2006)]{Medvedev:06} Medvedev, M.~V., Silva, L.~O., \& Kamionkowski, M.\ 2006, \apjl, 642, L1
1269:
1270: \bibitem[Milosavljevi\'c et al.(2007)]{Milosavljevic:07} Milosavljevi{\'c}, M., Koda, J., Nagai, D., Nakar, E., \& Shapiro, P.~R.\ 2007, \apjl, 661, L131
1271:
1272: \bibitem[Miniati et al.(2001)]{Miniati:01} Miniati, F., Ryu, D., Kang, H., \& Jones, T.~W.\ 2001, \apj, 559, 59
1273:
1274: \bibitem[Nagai et al.(2003)]{Nagai:03} Nagai, D., Kravtsov, A.~V., \& Kosowsky, A.\ 2003, \apj, 587, 524
1275:
1276: \bibitem[Nagai et al.(2007)]{Nagai:07a} Nagai, D., Vikhlinin, A., \& Kravtsov, A.~V., \ 2007, ApJ, 655, 98
1277:
1278: \bibitem[Nagai et al.(2007)]{Nagai:07b} Nagai, D., Kravtsov, A.~V., \& Vikhlinin A., \ 2007{\natexlab{b}}, preprint (astro-ph/0703661)
1279:
1280: \bibitem[Norman \& Bryan(1999)]{Norman:99} Norman, M.~L., \& Bryan, G.~L.\ 1999, The Radio Galaxy Messier 87, 530, 106
1281:
1282: \bibitem[Ostriker et al.(2005)]{Ostriker:05} Ostriker, J.~P., Bode, P., \& Babul, A.\ 2005, \apj, 634, 964
1283:
1284: \bibitem[Petrosian et al.(2006)]{Petrosian:06} Petrosian, V., Madejski, G., \& Luli, K.\ 2006, \apj, 652, 948
1285:
1286: \bibitem[Pfrommer \& En{\ss}lin(2004)]{Pfrommer:04} Pfrommer, C., \& En{\ss}lin, T.~A.\ 2004, \aap, 413, 17
1287:
1288: \bibitem[{{Rasia} et~al.(2006){Rasia}, et~al.}]{Rasia:06} {Rasia}, E., et~al. 2006, \mnras, 369, 2013
1289:
1290: \bibitem[Reimer et al.(2003)]{Reimer:03} Reimer, O., Pohl, M., Sreekumar, P., \& Mattox, J.~R.\ 2003, \apj, 588, 155
1291:
1292: \bibitem[Ricker \& Sarazin(2001)]{Ricker:01} Ricker, P.~M., \& Sarazin, C.~L.\ 2001, \apj, 561, 621
1293:
1294: \bibitem[Rotman(1991)]{Rotman:91} Rotman, D.\ 1991, Physics of Fluids, 3, 1792
1295:
1296: \bibitem[Ryu et al.(2003)]{Ryu:03} Ryu, D., Kang, H., Hallman,
1297: E., \& Jones, T.~W.\ 2003, \apj, 593, 599
1298:
1299: \bibitem[Sarazin(2004)]{Sarazin:04} Sarazin, C.~L.\ 2004, Journal of
1300: Korean Astronomical Society, 37, 433
1301:
1302: \bibitem[Schekochihin et al.(2005)]{Schekochihin:05} Schekochihin, A.~A., Cowley, S.~C., Kulsrud, R.~M., Hammett, G.~W., \& Sharma, P.\ 2005, \apj, 629, 139
1303:
1304: \bibitem[Schlickeiser \& Shukla(2003)]{Schlickeiser:03} Schlickeiser, R., \& Shukla, P.~K.\ 2003, \apjl, 599, L57
1305:
1306: \bibitem[Schuecker et al.(2004)]{Schuecker:04} Schuecker, P., Finoguenov, A., Miniati, F., B{\"o}hringer, H., \& Briel, U.~G.\ 2004, \aap, 426, 387
1307:
1308: \bibitem[Springel \& Farrar(2007)]{Springel:07} Springel, V., \&
1309: Farrar, G.\ 2007, preprint (astro-ph/0703232)
1310:
1311: \bibitem[Subramanian et al.(2006)]{Subramanian:06} Subramanian, K., Shukurov, A., \& Haugen, N.~E.~L.\ 2006, \mnras, 366, 1437
1312:
1313: \bibitem[Sunyaev et al.(2003)]{Sunyaev:03} Sunyaev, R.~A., Norman, M.~L., \& Bryan, G.~L.\ 2003, Astronomy Letters, 29, 783
1314:
1315: \bibitem[Vazza et al.(2006)]{Vazza:06} Vazza, F., Tormen, G., Cassano, R., Brunetti, G., \& Dolag, K.\ 2006, \mnras, 369, L14
1316:
1317: \bibitem[Vikhlinin et al.(2001)]{Vikhlinin:01} Vikhlinin, A., Markevitch, M., \& Murray, S.~S.\ 2001, \apjl, 549, L47
1318:
1319: \bibitem[Warren et al.(2005)]{Warren:05} Warren, J.~S., et al.\ 2005, \apj, 634, 376
1320:
1321: \bibitem[Zank et al.(2002)]{Zank:02} Zank, G.~P., Zhou, Y., Matthaeus, W.~H., \& Rice, W.~K.~M.\ 2002, Physics of Fluids, 14, 3766
1322:
1323: \end{thebibliography}
1324:
1325: \end{document}
1326: