1: %\documentstyle[emulateapj]{article}
2: %\input{psfig.sty}
3: %
4: %\catcode`\@=11 % This allows us to modify PLAIN macros.
5: %\def\@versim#1#2{\vcenter{\offinterlineskip
6: % \ialign{$\m@th#1\hfil##\hfil$\crcr#2\crcr\sim\crcr } }}
7: %\newcommand{\beq}{\begin{equation}}
8: %\newcommand{\eeq}{\end{equation}}
9: %\def\lsim{\mathrel{\mathpalette\@versim<}}
10: %\def\gsim{\mathrel{\mathpalette\@versim>}}
11: %\def\vp{v_\phi}
12: %\def\la{\langle}
13: %\def\ra{\rangle}
14: %\def\ch{{\em Chandra} }
15: %\def\mpy{\rm \ M_\odot \ {\rm yr^{-1}}}
16: %\newcommand{\kmsec}{\,{\rm km\,s^{-1}}}
17: %\newcommand{\msun}{\,{\rm M_\odot}}
18: %\newcommand{\etal}{{et al.\ }}
19: %\def\spose#1{\hbox to 0pt{#1\hss}}
20: %\newcommand{\lta}{\mathrel{\spose{\lower 3pt\hbox{$\mathchar"218$}}
21: % \raise 2.0pt\hbox{$\mathchar"13C$}}}
22: %\newcommand{\gta}{\mathrel{\spose{\lower 3pt\hbox{$\mathchar"218$}}
23: % \raise 2.0pt\hbox{$\mathchar"13E$}}}
24: %\lefthead{Ruszkowski, Br\"{u}ggen \& Begelman}
25: %\righthead{3D simulations of viscous dissipation}
26: %
27: %\makeatletter
28: %\newenvironment{tablehere}
29: % {\def\@captype{table}}
30: % {}
31: %\newenvironment{figurehere}
32: % {\def\@captype{figure}}
33: % {}
34: %\makeatother
35:
36: \documentclass[12pt,preprint]{aastex}
37: \usepackage[numberedappendix]{emulateapj5} % manuscript mode, emulate ApJ
38:
39: % \documentclass[manuscript]{aastex}
40: %% preprint2 produces a double-column, single-spaced document:
41: % \documentclass[preprint2]{aastex}
42:
43: \newcommand{\vdag}{(v)^\dagger}
44: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
45:
46: \slugcomment{accepted for ApJ}
47:
48:
49: \shorttitle{Viscous dissipation in the ICM}
50: \shortauthors{Ruszkowski et al.}
51:
52:
53: \received{2003 October 27}
54: \begin{document}
55:
56: \title{3D simulations of viscous dissipation in the intracluster medium}
57:
58: \author{Mateusz Ruszkowski\altaffilmark{1}}
59: \affil{JILA, Campus Box 440,
60: University of Colorado at Boulder, CO 80309-0440; \email{mr@quixote.colorado.edu}}
61:
62: \author{Marcus Br\"{u}ggen}
63: \affil{International University Bremen, Campus Ring 1, 28759 Bremen, Germany;
64: \email{m.brueggen@iu-bremen.de}}
65: \and
66: \author{Mitchell C. Begelman\altaffilmark{2}}
67: \affil{JILA, Campus Box 440, University of Colorado at Boulder, CO 80309-0440;
68: \email{mitch@jila.colorado.edu}}
69:
70:
71: \altaffiltext{1}{{\it Chandra} Fellow}
72: \altaffiltext{2}{also at Department of Astrophysical and Planetary Sciences,
73: University of Colorado at Boulder}
74:
75: %\begin{document}
76: %\submitted{Submitted to ApJ}
77: %\title{3D simulations of viscous dissipation in the intra-cluster medium}
78: %\author{Mateusz Ruszkowski}
79: %\vspace{0.5\baselineskip}
80: %\affil{{\it Chandra} Fellow; JILA, UCB 440,
81: %University of Colorado at Boulder, CO 80309-0440; mr@quixote.colorado.edu}
82: %\vspace{0.5\baselineskip}
83: %\author{Marcus Br\"{u}ggen}
84: %\affil{International University Bremen, Campus Ring 1, 28759 Bremen,
85: %Germany; m.brueggen@iu-bremen.de}
86: %\and
87: %%\vspace{-0.5\baselineskip}
88: %\author{Mitchell C. Begelman}
89: %\affil{JILA, UCB 440, University of Colorado at Boulder, CO 80309-0440;\\
90: %Department of Astrophysical and Planetary Sciences, University of Colorado at Boulder;
91: %mitch@jila.colorado.edu}
92:
93: \begin{abstract}
94: We present three-dimensional simulations of viscous dissipation of
95: AGN induced gas motions and waves in clusters of galaxies. These
96: simulations are motivated by recent detections of ripples in the
97: Perseus and Virgo clusters. Although
98: the sound waves generated by buoyant bubbles decay with
99: distance from the cluster center, we show that these waves can contribute substantially
100: to offsetting the radiative cooling at distances
101: significantly exceeding the bubble size.
102: The energy flux of the waves declines more steeply with radius
103: than the
104: inverse-square law predicted by energy conservation, implying
105: that dissipation plays an important role in tapping the wave energy.
106: We show that such dispersing sound waves/weak shocks are
107: detectable as ripples on unsharp-masked X-ray cluster maps, and point out that
108: the interfaces between the intracluster medium and old bubbles
109: are also clearly detectable in unsharp-masked X-ray maps. This opens up
110: the possibility of detecting
111: fossil bubbles that are difficult to detect in radio emission.
112: This mode of heating is
113: consistent with other observational constraints, such as the
114: presence of cool rims around the bubbles and the absence of strong
115: shocks. Thus,
116: the mechanism offers a way of heating clusters in a spatially
117: distributed and gentle fashion. We also discuss the energy transfer
118: between the central AGN and the surrounding medium. In our
119: numerical experiments, we find that roughly 65 per cent of the energy
120: injected by the AGN is transferred to the intracluster medium and
121: approximately 25 percent of the injected energy is dissipated
122: by viscous effects and contributes to heating of the gas. The
123: overall transfer of heat from the AGN to the gas is comparable to
124: the radiative cooling losses. The
125: simulations were performed with the FLASH adaptive mesh refinement
126: code.
127: \end{abstract}
128:
129:
130: \keywords{cooling flows --- galaxies: active --- waves --- galaxies: clusters: general --- methods: numerical --- intergalactic medium}
131:
132:
133: \section{Introduction}
134: The long-standing problem of cooling flow clusters of
135: galaxies, in which the central cooling time is much shorter than
136: the Hubble time, is how to prevent the intracluster medium (ICM) from collapsing
137: catastrophically on a short timescale. The original idea for maintaining the
138: overall cluster stability (Fabian 1994) was to postulate that a
139: certain amount of gas decouples from the flow and does not contribute
140: to the cooling of the remaining gas. This model would require up
141: to 1000 $M_{\odot}$ yr$^{-1}$ in mass deposition rates to guarantee
142: cluster stability. This has been found to be inconsistent with recent
143: {\it Chandra} (e.g., McNamara et al. 2000, Blanton et
144: al. 2001) and XMM-{\it Newton} observations (e.g., Peterson et
145: al. 2001, 2003; Tamura et al. 2001). {\it Chandra} observations
146: reveal a number of clusters with X-ray cavities/bubbles created by
147: the central active galactic nuclei (AGN). It has been suggested
148: by many authors that AGN feedback may play a crucial role in
149: self-regulating cooling flows (e.g. Churazov et al. 2001,
150: Ruszkowski \& Begelman 2002, Brighenti \& Mathews 2003). One
151: of the main outstanding issues is how the AGN heating comes about in
152: detail. In principle, strong shocks generated by AGN
153: outbursts can dissipate in the ICM and heat
154: the gas. However, imaging observations of cooling flow cores do
155: not give evidence for this mode of heating. Recent {\it Chandra}
156: observations of two well-known clusters, the Perseus cluster
157: (Fabian et al. 2003a,b) and the Virgo cluster (Forman et
158: al. 2004), suggest that dissipation of sound waves and weak shocks
159: could be an important source of gas heating --- an idea first proposed
160: by Fabian et al. (2003a). Further support for the idea that
161: viscosity may play an important role in the ICM
162: comes from a recent study of density profiles in clusters (Hansen
163: \& Stadel 2003). Recently, a number of papers have described
164: simulations of bubble-heated clusters (e.g., Churazov et al. 2001,
165: Br\"{u}ggen et al. 2002, Br\"{u}ggen \& Kaiser 2002, Br\"{u}ggen
166: 2003, Quilis et al. 2001). Numerical simulations of viscous
167: dissipation of AGN energy in ICM were previously considered by
168: Ruszkowski et al. (2004) (Paper I) and Reynolds et al. (2004).\\
169: \indent The main purpose of this paper is to extend our
170: previous work on viscous heating of the ICM by waves to three
171: dimensions. The results of a simulation of viscous dissipation in
172: three dimensions could differ from our previous two-dimensional
173: results given that the amplitudes of waves decrease faster with radius in
174: three dimensions. This would directly affect the spatial
175: distribution of the viscous dissipation rate.
176: Apart from performing our simulations in three dimensions, we
177: extend our previous analysis and that of Reynolds et al. (2004) to
178: include aspects of heating that were previously neglected. First,
179: we show that the conclusions drawn from our 2D simulations carry over to
180: three dimensions. Second, our new results include (i) X-ray maps and
181: unsharp-masked X-ray images, (ii) extend the discussion of spatial
182: distribution of energy dissipation and overall heating rate, (iii) present
183: details of the flow of heat between the bubbles and the intracluster
184: medium and (iv) discuss the wave decay rates.\\
185: \indent
186: The outline of
187: this paper is as follows. In the next section we describe the
188: assumptions of the model. Section 3 presents and discusses our
189: results, focusing on the spatial distribution of heating. In
190: particular, we discuss the detectability of the ripples/sound
191: waves, their decay rate with distance from the center/bubble
192: surface, the energy transfer from the AGN to the kinetic and thermal
193: energy of the ICM, and the fraction of the dissipated energy that
194: heats the gas. The fourth section discusses the limitations of our model.
195: The final section summarizes our findings.
196: \section{Assumptions of the Model}
197:
198: The initial conditions, details of energy injection, dissipation and radiative
199: cooling assumed in our simulations are similar to
200: those in Ruszkowski et al. (2003). Here we only summarize the
201: differences from Paper I.\\
202: \indent
203: Calculations were done in three dimensions in Cartesian geometry using
204: the PPM adaptive mesh refinement code FLASH (version 2.3).
205: Starting from a single
206: top-level block and using block sizes of 16$^3$ zones we allowed for 5
207: levels of refinement, giving an effective number of 256$^3$ zones. The
208: size of the computational domain was (200 kpc)$^3$, which corresponds
209: to an effective resolution of $\sim 0.8$ kpc. As previously, we
210: employed outflow boundary conditions on all boundaries.\\
211: \indent
212: When the source is active, each of the active regions
213: has a constant luminosity of
214: $L=1.6\times 10^{45}$ erg s$^{-1}$ and the rate at
215: which mass is injected into the active region is $\dot{\rho}V=2.8$
216: M$_{\odot}$yr$^{-1}$. The energy injection
217: is intermittent with a period of $3\times 10^7$ years, within which the
218: source is active for $0.5\times 10^{7}$ years.
219: Thus, since there are two active regions in the source, the time-averaged
220: luminosity is $\sim 5.3\times 10^{44}$ erg s$^{-1}$.
221: The motivation behind a different choice
222: of source parameters in the 3D case is as follows.
223: In 2D the code takes the input density (g cm$^{-3}$) and luminosity
224: density (erg cm$^{-3}$ s$^{-1}$) and interprets them as surface
225: quantities. That is, the hydrodynamic equations are effectively integrated
226: along one axis by multiplying these quantities by a unit length. As
227: the size of the active regions are smaller then unity, the code
228: effectively assigns a bigger volume to the active region when the
229: simulation is performed in 2D. The source luminosity density that
230: enters the energy equation is a constant equal to $L/V_{\rm b}$, where $L$
231: is the source luminosity and $V_{\rm b}=(4/3)\pi r^{3}_{\rm b}$ is the
232: bubble volume (the same formula is used in the code in 2D and 3D
233: simulations). This means that in order to supply the same energy
234: to the cluster in 3D as in a 2D simulation, one has to convert
235: the 2D luminosity to the 3D (real) luminosity using an approximate prescription
236: $L_{{\rm 3d}} \sim (3/4)$(unit length/$r_{\rm b}$)$L_{\rm{2d}}$.
237: The size of the active region in the 3D case was $r_{\rm b}=3$ kpc.
238: As far as the mass injection is concerned, we used higher values for
239: purely pragmatic reasons. Higher injection rates assure that the
240: densities in the active regions are reduced to a lesser degree then
241: they would have been if the injection rate had been lower. This means that
242: the constraints on the Courant timestep are less severe and the simulations
243: can be performed in a shorter time.\\
244: \indent
245: We use the standard Spitzer viscosity for an
246: unmagnetized plasma (Braginskii 1958), for which $\mu =
247: 7.1\times 10^{-17}(\ln\Lambda/31)^{-1}T^{5/2}$ g cm$^{-1}$ s$^{-1}$.
248: As conditions inside the buoyantly rising bubbles
249: are very uncertain and because we want to focus on energy
250: dissipation in the ambient ICM, we assume that dissipation occurs only
251: in the regions surrounding the buoyant gas. To this end we impose
252: a condition that switches on viscous effects provided that the
253: fraction of the injected gas in a given cell is smaller than
254: $10^{-3}$. The gas obeys a
255: polytropic equation of state with an adiabatic index of $\gamma=5/3$.
256: Unlike in Paper I we do not assume that the active regions are initially
257: underdense and overheated.
258:
259: \section{Results}
260:
261: \subsection{Morphology of X-ray Emission and Dissipation}
262: The top row in Figure 1 shows X-ray maps of the heated cluster. The panels show five
263: different epochs from earliest on the left to the latest on the
264: right. We assumed that the main contributor to X-ray emissivity
265: is free-free emission. The maps correspond to emission integrated in
266: the energy band $E\in (2-10)$ keV. The axis of injection is
267: located in the plane of the sky. The emissivity contrast between
268: the injected material and the surrounding medium is strongest at
269: the center. As the bubbles rise buoyantly in the cluster
270: atmosphere, the contrast diminishes and so does the probability of
271: detecting the bubbles. \\
272: \indent
273: The middle row in Figure 1 presents unsharp-masked X-ray images
274: corresponding to the images on the top row. These images were
275: generated by smoothing the original X-ray map and subtracting the
276: original X-ray map. Smoothing was done by convolving the original
277: image with a Gaussian filter centered on a given point. The full
278: width at half maximum of the adopted Gaussian distribution was
279: $\sim 6$ kpc. It is
280: evident from this figure that ripples outside bubble locations are
281: visible. As discussed below, these perturbations are
282: the sound waves/weak shocks. Note also that the interfaces between the bubble location
283: and the ICM are clearly present in these maps. This means that the ripples
284: in the Perseus cluster can be due to a combination of sound waves, weak shocks and interfaces between
285: the ICM and fossil bubbles.\\
286: \indent
287: The bottom row in Figure 1 shows the viscous dissipation rate
288: (erg s$^{-1}$g$^{-1}$) associated with the dispersing waves seen in
289: the middle row. These maps show cross-sections through the cluster center
290: that are perpendicular to the line of sight.
291:
292: \subsection{Spatial Distribution of Heating and Total Dissipation}
293: In Figure 2 we show the ratio of the viscous heating rate
294: to the radiative cooling rate as a function of time for a range
295: of radii. More distant regions are heated at later times and hence the
296: curves corresponding to these regions rise at later times.
297: It is interesting to note that the heating-to-cooling
298: ratio is of order unity even for more distant regions close to the
299: cooling radius. Thus, heating is well distributed spatially, even
300: though the bubbles occupy a smaller volume than the waves. Note that
301: the curves exhibit a clear periodic behavior which is due to the
302: intermitency of the central source. We also note that this figure is
303: similar to Fig. 2 in Paper I, thus confirming that
304: one of the main conclusions of their paper holds in three-dimensional
305: geometry.\\
306: \indent
307: Figure 3 shows the ratio of viscous heating to radiative cooling rate as a
308: function of distance from the cluster
309: center for equally-spaced time intervals of
310: $\Delta t=10^{7}$ yr until $2\times 10^{8}$ yr.
311: As time increases the curves start to decline at progressively larger radii.
312: From this figure one can also deduce the characteristic speed of
313: the wave pattern (see below) and compare it with analytical estimates.\\
314: \indent
315: Figure 4 shows the ratio of the volume-integrated heating
316: rate to the volume-integrated cooling rate as a function of
317: time. Note that this ratio is of order unity. Therefore,
318: this heating mechanism has the potential for significantly
319: affecting the rate at which the gas loses its internal energy or
320: perhaps even offsetting radiative cooling altogether.\\
321: \indent
322: We stress that the actual value of the ratio of heating to cooling (Figs. 2--4)
323: depends on the parameters
324: adopted in the model. However, fine tuning may not be necessary
325: as already argued by Paper I.\\
326: \indent
327: Additional heating may result from damping of large scale motions caused
328: by entrainment and lifting of the gas surrounding the bubbles. This is
329: the general mechanism proposed by Begelman (2001) and Ruszkowski
330: \& Begelman (2002) in their discussion of ``effervescent
331: heating''. The widespread spatial distribution of heat in
332: this mechanism can be achieved when the buoyant bubbles rise in
333: the ICM perturbed by preceding bubbles. That is, subsequent
334: bubbles find lower resistance to move in all directions once they
335: encounter underdense regions ``drilled'' by earlier bubbles.
336: Moreover, when the bubbles move a substantial distance from the
337: cluster center, the lateral spreading is further enhanced as
338: the bubble entropy becomes comparable to the entropy of the
339: ambient medium. Spatial spreading
340: of heat could also be facilitated if the jet (or other form of outflows) precesses
341: or if the black hole
342: in the center of the AGN adjusts in response to a change in the orientation of the accretion
343: disk. This could be triggered by a merger with a substructure clump
344: in the cluster or by the oscillatory motion of the AGN around
345: the cluster center of gravity (see e.g. Johnstone et al. 2002).
346:
347: \subsection{Energy Transfer Between the Bubbles and the ICM}
348: In Figure 5 we present the cumulative
349: injected energy $E_{\rm inj}$ (solid curve), energy contained in the
350: rising bubbles $E_{\rm bubb}$ (dashed) and the energy transferred
351: to the ICM ($E_{\rm tran}\equiv E_{\rm inj}-E_{\rm bubb}$; dotted). The bubble energy
352: $E_{\rm bubb}$ is defined as
353:
354: \begin{eqnarray}
355: E_{\rm bubb} & \equiv & \int_{\rm bubb}
356: [e_{\rm pot}(t)+e_{\rm int}(t)+e_{\rm kin}(t)]\rho dV\nonumber\\
357: & - & [e_{\rm pot}(0)+e_{\rm int}(0)]\rho_{0} (2V_{0}) ,
358: \end{eqnarray}
359:
360: \noindent
361: where $e_{\rm pot}$, $e_{\rm kin}$ and $e_{\rm int}$ are the gravitational potential,
362: specific kinetic energy density and specific internal energy density, respectively.
363: The $2V_{0}$ factor is the total volume of the initial injection regions.
364: The integration is performed over the volume of the bubbles.
365: The bubble energy increases during
366: outbursts and then decreases during dormant phases.
367: The decrease of the energy in the bubbles with time can be attributed to
368: two factors: (i) energy is transferred to the ambient medium
369: via $PdV$ work and because the rising bubble experiences drag from the
370: surrounding gas and (ii) mixing of the
371: bubble and the ICM gas. Recall that the bubble is defined as the region
372: where the fraction of the injected gas is greater than $10^{-3}$
373: (this region overlaps exactly with the low density bubble region).
374: This assures that essentially no energy injected into the bubble is omitted in
375: the calculation of the bubble energy. The contribution to the bubble energy
376: from the gravitational potential is relatively small. Since mixing occurs without
377: a significant change in pressure and the bubble internal energy is
378: proportional to pressure, the change in bubble energy comes mostly
379: from the $PdV$ work done by the bubble against its surroundings
380: and the drag on the bubble from the ambient ICM.\\
381: \indent
382: The fraction of the AGN energy that is transferred to the
383: surrounding ICM is plotted in Figure 6 (solid curve). The ratio
384: of viscously dissipated energy to the energy
385: injected by the AGN is plotted as the bottom (dashed) curve in Figure
386: 6. This figure indicates that approximately 35 per cent of the
387: energy injected by the AGN remains in the buoyant bubbles while the
388: rest is transferred to the surrounding ICM and that about 25 per cent of the
389: injected energy ends up heating the ambient ICM viscously.
390: We point out that not all of the energy that is transferred to the ICM
391: is converted to heat. It is viscous dissipation that is responsible for
392: heating the gas.
393: The fraction of the injected energy that is transferred to the ambient
394: ICM agrees qualitatively with simple analytical estimates presented in
395: Paper I.
396: For the adopted adiabatic index and assuming pressure equilibrium between the bubble
397: and the ICM about $40\%$ of the input energy can be transferred to the ambient
398: gas.
399: %and we repeat this argument here for the
400: %sake of clarity. Suppose $dQ$ is the heat injected into the
401: %cavity, $dU$ the increase in internal energy, $P$ the pressure,
402: %$dV$ the change in volume, $dH$ the change in enthalpy and $dW$
403: %the work done on the surroundings. Applying the first law of
404: %thermodynamics to the cavities and assuming local pressure
405: %equilibrium, $dQ = dU + P \, dV = dH - V \, dP \simeq dH,$ so that $dW
406: %= P \, dV \simeq d (PV) = {\gamma - 1 \over \gamma} dH \simeq
407: %{\gamma - 1 \over \gamma} dQ$. This means that about $40\%$ of the
408: %energy input can be transferred to the ambient medium.
409: Because the cavities are mildly overpressured, the fraction of the
410: input power transferred to the ICM in the actual simulation is a
411: little larger.
412:
413: \subsection{Wave decay}
414: From Figure 3 one can deduce the characteristic speed of
415: the wave pattern which we found to have a Mach number of $\sim 1.3$.
416: Thus, the waves can be interpreted as strong sound waves or weak shocks.
417: In the weakly nonlinear regime, the speed of the wave is
418: $c_{s}(1+[(\gamma +1)/2]\alpha)$, where $c_{s}$ is the sound speed and
419: $\alpha\sim\delta\rho/\rho$ is the normalized wave density amplitude
420: (Stein \& Schwartz 1972, Mihalas \& Mihalas 1984).
421: For the typical amplitudes seen in the simulation, the wave speed of
422: Mach 1.3 is consistent with these estimates.\\
423: \indent
424: In Figure 7 we show the energy flux of the decaying wave
425: corresponding to the initial outburst. The simulation results are
426: denoted by filled squares connected by a solid line. Also shown
427: for reference is the decay profile corresponding to $\sim r^{-2}$. All
428: curves have arbitrary units.
429: The period-averaged wave energy per unit time that is
430: streaming through a surface $S$ in a direction perpendicular to
431: this surface is $L_{w}\sim (\delta P)^{2}S/(\rho v_{w})$, where
432: $v_{w}$ is the wave speed. In the absence of any dissipation, the
433: energy flux should scale as $\sim r^{-2}$.
434: However, the slope of the energy flux in our simulations
435: is steeper. This means that viscous dissipation plays an important role in
436: tapping the wave energy. \\
437: \indent
438: The characteristic dissipation length $l$ can be estimated from
439: $l\sim 70\lambda_{10}^{2}n_{0.02} T_{4}^{-2}$ kpc, where
440: $\lambda=10\lambda_{10}$kpc, $n=0.02n_{0.02}$ cm$^{-3}$
441: and $T=4T_{4}$ keV (Fabian et al. 2003a, cf. Landau and Lifshitz 1975).
442: The dissipation length can also be estimated from
443: $l\sim\ [\partial\ln(\rho\alpha^{2})/\partial r + 2/r]^{-1}$,
444: where $\alpha\equiv \delta\rho/\rho$ is the normalized density amplitude of the wave
445: and can be directly derived from the simulation results.
446: At $r\sim 55$ kpc, the dissipation length is of order 40 kpc, which is
447: qualitatively consistent with the above simple analytical estimates.
448:
449: \subsection{Caveats}
450: We stress that our
451: use of the Spitzer viscosity is meant to be illustrative and may
452: not accurately represent the momentum transport in the magnetized
453: intracluster medium. For one thing, magnetic shear stresses are
454: likely to dominate over molecular viscosity in the transport of
455: bulk momentum. This could either enhance or suppress the
456: dissipation of sound waves, and will almost certainly make the
457: dependence of stress on the velocity field more complicated. For
458: another, in this macroscopic form of momentum transport the rate
459: of dissipation (due to reconnection) would be nonlocally related to
460: the stress tensor. Treatment of these effects will require
461: high-resolution magnetohydrodynamical simulations. Moreover,
462: magnetic fields could introduce effects similar to bulk viscosity,
463: as a result of plasma microinstabilities. In our simulations we
464: neglected bulk viscosity since it vanishes for an ideal gas. We
465: note that bulk viscosity, if present, could dissipate waves even
466: more efficiently. Finally, we have neglected the effects of
467: thermal conduction, which (assuming Spitzer conductivity) could
468: damp the sound waves more quickly than Spitzer viscosity (since
469: the conductive dissipation rate exceeds the viscous one by a factor
470: $\sim 10$ under the simplified assumption that waves are
471: linear and that the gas has constant density and pressure and
472: gravity can be neglected; see Landau and Lifshitz 1975). We
473: point out that, as long as the waves are linear, the nature of the
474: wave decay due to Spitzer viscosity or Spitzer conductivity is the
475: same, i.e., the only change is the constant damping coefficient.
476: Note that the waves considered here are either linear (sound
477: waves) or weakly non-linear (weak shocks) and characterized
478: by a relatively small Mach number. Since conductivity is expected
479: to be suppressed by magnetic fields, a realistic assessment of
480: whether conduction enhances the damping rate of sound waves is
481: beyond the scope of this investigation.
482: If the dissipation rate significantly exceeds the Spitzer
483: value, then the sound waves will be damped more efficiently and only the
484: gas close to the bubbles will be heated efficiently.\\
485: \indent
486: We now argue that the choice of boundary conditions has negligible
487: effects on our results.
488: First of all, the issue of reflection has no effect at all on our
489: results presented in Figures 4, 5, 6 and 7. This is because the waves
490: had no time to reach the boundary for the times considered in these
491: figures.\\
492: \indent
493: The waves reach the boundary approximately at time $\sim 6\times 10^{7}$ yr. However,
494: there is
495: no clear rapid rise in the dissipation rate for the times before and
496: after this time. No such jump is seen in Figure 2 (e.g., for the
497: curves that correspond to outer shells in the cluster, where the
498: reflection effects should be seen immediately) or in Figure 4, which
499: shows the total dissipation rate. Had the reflection been important,
500: such a jump would have been clearly visible.\\
501: \indent
502: In the last row of Figure 1, 2nd (earlier time) and 3rd (later time) columns, shows the
503: dissipation wave in the process of crossing the boundary. No
504: reflection is seen in either 2nd or 3rd column figure. Also, the wave
505: fronts in the corners propagate across the boundary without any
506: obvious signs of strong reflection.\\
507: \indent
508: It is known that
509: for supersonic fluctuations the outflow boundary conditions are
510: exact and no reflections
511: are expected. The fact that our waves are strong sound waves or weak
512: shocks traveling at mildly supersonic velocities helps to minimize
513: reflections.\\
514: \indent
515: In principle, the outflow boundary conditions can be made arbitrarily
516: close to the exact ones if the simulation resolution is sufficiently
517: high. That is, the error made by replacing the ghost zone solution
518: based on some exact method by the value copied from the last active
519: zone (outflow boundary condition) tends to zero as the resolution
520: increases. Similarly, when the perturbations in question are smoother,
521: the outflow boundary conditions give a more accurate answer because
522: the waves are better resolved. We note that the effect of dissipation
523: is to disperse the waves and make them more easily
524: resolvable. Moreover, the amplitude of the waves decays faster in 3D
525: than in 2D. Thus, in 3D the effects of reflection are reduced.\\
526: \indent
527: The outflow boundary conditions are exact for a wave traveling along
528: the boundary. Therefore, the degree to which a wave is reflected
529: decreases for waves moving at an angle to the boundary.
530:
531: \section{Summary}
532: To summarize, we have analyzed the energy deposition in the cluster due to
533: rising bubbles, sound waves and weak shocks.
534: This was motivated by the recent discovery of such waves in the Perseus
535: cluster by Fabian et al. (2003a) and in the Virgo cluster by Forman et al. (2004).
536: We found that the dissipated energy may be comparable to the cooling rate,
537: thereby significantly affecting the cooling flow or even quenching it
538: altogether. We showed that about 65 per cent of the energy injected by the central
539: source can be transferred to the ICM.
540: Approximately 25 per cent of the energy injected by the AGN
541: can be converted to heat, assuming Spitzer viscosity.
542: We discussed the wave decay rates and showed that a significant fraction of
543: wave energy is deposited within the cooling radius. The computed decay
544: rates are consistent with linear theory estimates of the damping
545: length. The damped sound waves or weak shocks are still detectable
546: in unsharp-masked X-ray images.
547: Old bubbles become increasingly difficult to detect in the
548: X-ray maps as the contrast between the rising bubbles and the surrounding
549: gas diminishes.
550: However, apart from sound waves and weak shocks in
551: unsharp-masked X-ray maps,
552: the interfaces between the intracluster medium and old bubbles
553: are also clearly visible. This opens up the possibility of detecting
554: fossil bubbles that are difficult to detect in radio emission.
555:
556:
557: %Spaans cooling to low temp.,
558: %jets, cosmpological context, in preparation
559:
560: \acknowledgments{
561: MR thanks Daniel Proga for discussions.
562: Support for this work was provided by National
563: Science Foundation grant AST-0307502 and the National Aeronautics and Space
564: Administration through {\it Chandra} Fellowship Award Number PF3-40029
565: issued by the Chandra X-ray Observatory Center, which is operated by
566: the Smithsonian Astrophysical Observatory for and on behalf of the
567: National Aeronautics and Space Administration under contract
568: NAS8-39073.
569: MB acknowledges support by DFG contract BR2026/2 and thanks
570: JILA, University of Colorado at Boulder for their hospitality.
571: The software used in this work was in part developed by the DOE-supported
572: ASCI/Alliance Center for Astrophysical Thermonuclear Flashes at the
573: University of Chicago. Preliminary computations were performed
574: on JILA Keck cluster sponsored by the Keck Foundation.}
575:
576: \begin{thebibliography}{zz}
577:
578: \item Batchelor, G.K. 1967, {\it An Introduction to Fluid Dynamics}, Cambridge University Press, Cambridge
579:
580: %\item Begelman, M.C. 2001, in ASP Conf. Proc., 240, ``Gas and Galaxy Evolution'', ed. J.E. Hibbard, M.P. Rupen, \& J.H. van Gorkom (San Francisco: ASP), 363
581:
582: %\item Begelman, M.C. 2003, in {\it Carnegie Observatories Astrophysics Series},
583: %Vol. 1: ``Coevolution of Black Holes and Galaxies'' ed. L. C. Ho (Cambridge:
584: %Cambridge Univ. Press), in press, astro-ph/0303040
585:
586: \item Blanton, E.L., Sarazin, C.L., McNamara, B.R., \& Wise, M.W. 2001, {\it ApJ}, 558, 15
587:
588: \item Braginskii, S.L. 1958, {\it Sov. Phys.}, {\it JETP}, 6, 358
589:
590: \item Brighenti, F., \& Mathews, W.G. 2003, ApJ, 587, 580
591:
592: \item Br\"uggen, M. 2003, {\it ApJ}, 592, 839
593:
594: \item Br\"uggen, M., \& Kaiser, C.R. 2002, {\it Nature}, 418, 301
595:
596: \item Br\"uggen, M., Kaiser, C.R., Churazov, E., \& En{\ss}lin 2002, {\it MNRAS}, 331, 545
597:
598: %\item Burns, J.O. 1990, {\it AJ}, 99, 14
599:
600: \item Churazov, E., Br\"uggen, M., Kaiser, C.R., B\"ohringer, H., \& Forman, B. 2001, {\it ApJ}, 554, 261
601:
602: \item Churazov, Sunyaev, Forman, \& B{\"o}hringer 2002, {\it MNRAS}, 332, 729
603:
604: %\item Croom, S. at al. 2003, in ASP Conf. Proc., ``AGN Physics with the Sloan
605: %Digital Sky Survey'', ed. G. T. Richards and P. B. Hall (San Francisco: ASP), in press, astro-ph/0310533
606:
607: \item Fabian, A.C. 1994, {\it ARA\&A}, 32, 277
608:
609: %\item Fabian, A.C. et al. 2000, {\it MNRAS}, 318, 65
610:
611: %\item Fabian, A.C., Celotti, A., Blundell, K.M., Kassim, N.E., \& Perley, R.A. 2002, {\it MNRAS}, 331, 369
612:
613: \item Fabian, A.C., Sanders, J.S., Allen, S.W., Crawford, C.S., Iwasawa, K., Johnstode, R.M., Schmidt, R.W., \& Taylor, G.B. 2003a, {\it MNRAS}, 344, 43
614:
615: \item Fabian, A.C., Sanders, J.S., Crawford, C.S., Conselice, C.J., Gallagher, J.S., \& Wyse, R.F.G. 2003b, {\it MNRAS}, 344, 48
616:
617: \item Forman, W. et al. 2004, {\it ApJ}, submitted, astro-ph/0312576
618:
619: \item Hansen, H.H. \& Stadel, J. 2003, {\it ApJ}, 595, 37
620:
621: \item Johnstone, R.M., Allen, S.W., Fabian, A.C., \& Sanders, J.S. 2002, MNRAS, 336, 299
622:
623: %\item Kaiser, C.R., Dennet-Thorpe, J., \& Alexander, P. 1997, {\it MNRAS}, 292, 723
624:
625: %\item Mazzotta, P., Kaastra, J. S., Paerels, F. B., Ferrigno, C.,
626: %Colafrancesco, S., Mewe, R., \& Forman, W. R. 2002, {\it ApJ}, 567, 37
627:
628: \item Landau, L.D \& Lifshitz, E.M. 1975, {\it Fluid Mechanics}, Pergamon Press
629:
630: \item McNamara, B.R., et al. 2000, {\it ApJ}, 534, 135
631:
632: \item Mihalas, D., \& Mihalas, B.W. 1984, Foundations of Radiation Hydrodynamics
633: (New York: Oxford Univ. Press)
634:
635: \item Navarro, J., Frenk, C., \& White, S. 1995, {\it MNRAS}, 275, 720
636:
637: \item Navarro, J., Frenk, C., \& White, S. 1997, {\it ApJ}, 490, 493
638:
639: \item Peterson, J.R., Paerels, F.B.S., Kaastra, J.S., et al. 2001, {\it A}\&{\it A}, 365, 104
640:
641: \item Peterson, J.R., Kahn, S.M., Paerels, F.B.S., Kaastra, J.S., et al. 2003, {\it ApJ}, 590, 207
642:
643: %\item Porter, D.H., \& Woodward, P.R. 1994, {\it ApJS}, 93, 309
644:
645: %\item Pringle, J.E. 1989, {\it MNRAS}, 239, 479
646:
647: \item Quilis, V., Bower, R., \& Balogh, M.L. 2001, {\it MNRAS}, 328, 1091
648:
649: %\item Reynolds, C.S., Heinz, S., \& Begelman, M.C. 2001, {\it ApJ}, 594, 197
650:
651: \item Reynolds, C.S. et al. 2004, submitted to {\it MNRAS}, astro-ph/0402632
652:
653: \item Ruszkowski, M., \& Begelman, M.C. 2002, {\it ApJ}, 581, 223
654:
655: \item Ruszkowski, M., Br\"{u}ggen, M., \& Begelman, M.C., 2004, {\it ApJ},
656: in press, astro-ph/0310760, Paper I
657:
658: \item Shu, F.H., 1992, {\it The Physics of Astrophysics}, Vol.2, {\it Gas Dynamics},
659: University Science Books, Sausalito, CA
660:
661: \item Stein, R.F., \& Schwartz, R.A. 1972, ApJ, 177, 807
662:
663: \item Sutherland, R.S., \& Dopita, M.A. 1993, ApJS, 88, 253
664:
665: %\item Sytine, I.V., Porter, D.H., Woodward, P.R., Hodson, S.W., \& Winkler, K.-H. 2000, J. Comput. Phys., 158, 225
666:
667: %\item Tabor, G., \& Binney, J. 1993, {\it MNRAS}, 263, 323
668:
669: \item Tamura, T., Kaastra, J.S., Peterson, J.R., Paerels, F., et al. 2001, {\it A}\&{\it A}, 365, 87
670:
671: \item Tozzi, P., \& Norman, C. 2001, {\it ApJ}, 546, 63
672:
673: \end{thebibliography}
674:
675: \newpage
676:
677: %\begin{figurehere}
678: \begin{figure}
679: \centerline{
680: %\psfig{file=f1.eps,width=5.0in,angle=0}
681: \includegraphics[width=7.0in,angle=0]{f1.eps}}
682: \caption{\footnotesize
683: Top row shows the X-ray emissivity maps of the AGN-heated cluster.
684: Snapshots correspond to $3.0\times 10^{7}$, $1.15\times 10^{8}$,
685: $1.25\times 10^{8}$, $1.55\times 10^{8}$ and $1.85\times 10^{8}$ years,
686: from left to right, respectively.
687: Middle row shows X-ray unsharp masked maps corresponding to X-ray maps.
688: Bottom row shows the map of the viscous dissipation pattern.
689: Whereas the maps in the bottom row show cross-sections through the cluster center
690: that are perpendicular to the line of sight, the first two rows correspond to
691: projections onto the plane of the sky.
692: \label{fig1}}
693: %\end{figurehere}
694: \end{figure}
695:
696: %\begin{figurehere}
697: \begin{figure}
698: \centerline{
699: \includegraphics[width=3.5in,angle=90]{f2.eps}}
700: %\psfig{file=f4.eps,width=3.5in,angle=90}}
701: %\psfig{file=HC_t.ps,width=5.0in,angle=90}}
702: \caption{\footnotesize The ratio of viscous heating to radiative cooling rate as a
703: function of time for a number of concentric shells around the cluster
704: center. The curves that start rising at later times correspond to
705: shells located further away from the center. The heating-to-cooling
706: ratio was calculated in ten shells starting from the first shell at 5
707: kpc and the remaining shells located in increments of 10 kpc away from
708: the cluster center. Note that the heating rate is comparable to the
709: cooling rate.
710: \label{fig2}}
711: %\end{figurehere}
712: \end{figure}
713:
714: %\begin{figurehere}
715: \begin{figure}
716: \centerline{
717: \includegraphics[width=3.5in,angle=90]{f3.eps}}
718: %\psfig{file=f5.eps,width=3.5in,angle=90}}
719: %\psfig{file=HC_r.ps,width=5.0in,angle=90}}
720: \caption{\footnotesize
721: The ratio of viscous heating to radiative cooling rate as a
722: function of distance from the cluster
723: center for equally-spaced time intervals of $\Delta t=10^{7}$ yr until $2\times 10^{8}$ yr.
724: \label{fig3}}
725: %\end{figurehere}
726: \end{figure}
727:
728: %\begin{figurehere}
729: \begin{figure}
730: \centerline{
731: \includegraphics[width=3.5in,angle=90]{f4.eps}}
732: %\psfig{file=f6.eps,width=3.5in,angle=90}}
733: %\psfig{file=HC_tot.ps,width=5.0in,angle=90}}
734: \caption{\footnotesize The ratio of volume-integrated heating rate (within
735: 100 kpc from the center) to volume-integrated cooling rate
736: as a function of time.
737: \label{fig4}}
738: %\vspace{+0.5cm}
739: %\end{figurehere}
740: \end{figure}
741:
742: %\begin{figurehere}
743: \begin{figure}
744: \centerline{
745: \includegraphics[width=3.5in,angle=90]{f5.eps}}
746: %\psfig{file=f7.eps,width=3.5in,angle=90}}
747: %\psfig{file=energies1.ps,width=5.0in,angle=90}}
748: \caption{\footnotesize The cumulative injected energy $E_{\rm inj}$ (solid curve),
749: bubble energy $E_{\rm bubb}$ (dashed curve) and
750: energy transferred to the ambient ICM ($\equiv E_{\rm inj}-E_{\rm bubb}$; dotted curve)
751: as a function of time. All plots are in arbitrary units.
752: We only plot data until 60 Myr as for later times
753: the waves start to escape the computational box.
754: \label{fig5}}
755: %\end{figurehere}
756: \end{figure}
757:
758: %\begin{figurehere}
759: \begin{figure}
760: \centerline{
761: \includegraphics[width=3.5in,angle=90]{f6.eps}}
762: %\psfig{file=f8.eps,width=3.5in,angle=90}}
763: %\psfig{file=energies3.ps,width=5.0in,angle=90}}
764: \caption{\footnotesize The ratio of cumulative transferred energy to the cumulative injected
765: energy (top curve)
766: and the ratio of the cumulative viscously dissipated energy to the cumulative injected energy
767: as a function of time.
768: We only plot data until 60 Myr as for later times
769: the waves start to escape the computational box.
770: \label{fig6}}
771: %\end{figurehere}
772: \end{figure}
773:
774: %\begin{figurehere}
775: \begin{figure}
776: \centerline{
777: \includegraphics[width=3.0in,angle=270]{f7.eps}}
778: %\psfig{file=fig1.eps,width=5.0in,angle=90}}
779: %\psfig{file=decay.ps,width=3.0in,angle=270}}
780: \caption{\footnotesize The wave energy flux as a function of the distance from the cluster center.
781: The $\sim r^{-2}$
782: decay profile is shown for comparison. All quantities are in arbitrary units.
783: \label{fig7}}
784: %\end{figurehere}
785: \end{figure}
786:
787: \end{document}
788: