Embed Size (px)
Citation preview
Available in: http://www.redalyc.org/articulo.oa?id=57115743031
Red de Revistas Científicas de América Latina, el Caribe, España y Portugal
Sistema de Información Científica
M. Brüggen, M. Ruszkowski, A. Simionescu, M. Hoeft
SIMULATIONS OF MAGNETIC FIELDS IN CLUSTERS AND FILAMENTS
Revista Mexicana de Astronomía y Astrofísica, vol. 36, 2009, pp. 216-221,
Instituto de Astronomía
México
How to cite Complete issue More information about this article Journal's homepage
Revista Mexicana de Astronomía y Astrofísica,
ISSN (Printed Version): 0185-1101
Instituto de Astronomía
México
www.redalyc.orgNon-Profit Academic Project, developed under the Open Acces Initiative
© 2
009:
Inst
ituto
de
Ast
rono
mía
, UN
AM
- M
ag
netic
Fie
lds
in th
e U
nive
rse
II: F
rom
La
bo
rato
ry a
nd S
tars
to th
e P
rimo
rdia
l Uni
vers
eEd
. A. E
squi
vel,
J. F
ranc
o, G
. Ga
rcía
-Se
gur
a, E
. M. d
e G
ouv
eia
Da
l Pin
o, A
. La
zaria
n, S
. Liz
ano
, & A
. Ra
ga
RevMexAA (Serie de Conferencias), 36, 216–221 (2009)
SIMULATIONS OF MAGNETIC FIELDS IN CLUSTERS AND FILAMENTS
M. Bruggen,1 M. Ruszkowski,2 A. Simionescu,3 and M. Hoeft1
RESUMEN
El origen de los campos magneticos afuera de las galaxias es aun controversial. Para el medio intra-cumulo sehan reportado magnitudes de los campos del orden de µG. Es posible que los flujos de nucleos de galaxias activashayan magnetizado el medio intra-cumulo. El campo magnetico intergalactico dentro de filamentos deberıaser menos contaminado por flujos magnetizados de galaxias activas. Mas aun, se piensa que los filamentoscontienen el medio intergalactico tibio-caliente, el cual representa una gran fraccion de la materia barionica deluniverso. La emision difusa de sincrotron de los rayos cosmicos en estos filamentos revelara la presencia de talmedio. Presentamos simulaciones de alta resolucion con malla adaptativa de campos magneticos cosmicos yhacemos prediciones de la evolucion y estructura de dichos campos en filamentos. Los campos magneticos encumulos de galaxias pueden tambien tener su origen en nucleos de galaxias activas, cuya evolucion es a su vezfuertemente afectada por los campos magneticos. Mostramos como la dinamica de burbujas producidas pornucleos de galaxias activas, mecanismo vital de retroalimentacoon en cumulos de galaxias, es gobernado porlos campos magneticos del ambiente, los cuales suprimen de manera efectiva inestabilidades hidrodinamicas.
ABSTRACT
The origin of cosmic magnetic fields outside of galaxies remains controversial. In the intracluster medium fieldstrengths of the order of µG have been reported. Possibly the outflows of AGN have magnetised the ICM.The intergalactic magnetic field within filaments should be less polluted by magnetised outflows from activegalaxies than magnetic fields in clusters. Therefore, filaments may be a better laboratory to study magneticfield amplification by structure formation than galaxy clusters, since they host less active galaxies. Moreover,filaments are thought to contain the Warm-Hot Intergalactic Medium which represents a large fraction of thebaryonic matter in the universe. Diffuse synchrotron emission by cosmic rays in these filaments will revealthe presence of this medium. We present highly resolved cosmological adaptive mesh refinement simulationsof magnetic fields in the cosmos and make predictions about the evolution and structure of magnetic fields infilaments. Magnetic fields in clusters may also have their origin in AGN, whose evolution, in turn, is stronglyaffected by magnetic fields. We show how the dynamics of AGN-blown bubbles, that are vital for feedbackmechanisms in clusters and galaxies, is governed by ambient magnetic fields that are effective at suppressinghydrodynamic instabilities.
Key Words: H II regions — ISM: jets and outflows — stars: mass loss — stars: pre-main sequence
1. INTRODUCTION
Clusters of galaxies are known to host magneticfields with strengths that are of the order of 1 µGand with coherence scales that are of the order of10 kpc (Carilli & Taylor 2002; Govoni et al. 2004a).In cool cores of clusters remarkably high fields havebeen found: For example, Blanton et al. (2003) havefound magnetic fields as high as 11 µG in the coolcore of A2052. Knowledge about cluster magneticfields comes from synchrotron and inverse Comptonradiation, radio halos, as well as Faraday rotation
1Jacobs University Bremen, Campus Ring 1, 28759 Bre-
men, Germany.2Department of Astronomy, The University of Michigan,
500 Church Street, Ann Arbor, MI 48109, USA.3Max-Planck-Institute for Extraterrestial Physics,
Giessenbachstr, 85748 Garching, Germany.
measurements. The field estimates do not all agree.In particular, the inverse Compton estimates tend tobe lower by almost an order of magnitude than theestimates derived from Faraday rotation (for a re-view see Govoni & Feretti 2004b). The latter methodhas also been used to make inferences about thestructure of magnetic fields in clusters. Based onanalyses of rotation measure maps in three clusters,Vogt & Enßlin (2003) have derived spectral indices(slopes of the 1D power spectrum of the magneticfield) of 1.6 to 2.0 (Vogt & Enßlin 2005; Enßlin &Vogt 2003).
Meanwhile, the origin of cluster magnetic fieldsin the intergalactic medium (IGM) remains unclear.It has been suggested that they are of primordialorigin (see e.g. Widrow 2002 for a review of the
216
© 2
009:
Inst
ituto
de
Ast
rono
mía
, UN
AM
- M
ag
netic
Fie
lds
in th
e U
nive
rse
II: F
rom
La
bo
rato
ry a
nd S
tars
to th
e P
rimo
rdia
l Uni
vers
eEd
. A. E
squi
vel,
J. F
ranc
o, G
. Ga
rcía
-Se
gur
a, E
. M. d
e G
ouv
eia
Da
l Pin
o, A
. La
zaria
n, S
. Liz
ano
, & A
. Ra
ga
SIMULATIONS OF MAGNETIC FIELDS IN CLUSTERS AND FILAMENTS 217
origin of cosmic magnetic fields), i.e. a seed field thathas formed prior to recombination is subsequentlyamplified by compression and turbulence. In Dolaget al. (2002), the average fields in the core of clustersat redshift 0 were found to range between 0.4 and2.5 µG if the initial seed fields at redshift 15 were ofthe order of a nanogauss. The question, of course,remains, where the seed fields come from. It is quitepossible that the first stars and the first AGN haveproduced magnetic fields. The magnitude of theseseed fields, though, is as uncertain as most processesat this epoch.
Other mechanisms have been proposed (see refer-ences in Bruggen et al. 2005). Also, it has been sug-gested that the IGM has been magnetised by bubblesof radio plasma that are ejected by active galacticnuclei (AGN) (Furlanetto & Loeb 2001). A growingnumber of large, magnetised bubbles are being dis-covered in clusters of galaxies (Bırzan et al. 2004),and simulations show that they may not be verylong-lived (Bruggen et al. 2005; Bruggen & Kaiser2002). Presumably, the magnetised bubble plasmagets mixed with the IGM and is thus likely to con-tribute to present cluster fields. Here one may notethat the life time of the bubbles is itself dependent onthe magnetic field in the IGM as magnetic fields tendto stabilise the bubble against instabilities (Jones &De Young 2005). AGN activity favours the centresof massive clusters, so that the IGM in filamentsought to be much less affected by magnetised plasmaejected from AGN.
An interesting question that follows from studiesof cluster magnetic fields is whether significant mag-netic fields also exist on other scales in the cosmiclarge-scale structure, as for example in filaments ofgalaxies. This is particularly interesting against thebackground of the elusive Warm-Hot IntergalacticMedium (WHIM). As pointed out by Cen & Ostriker(1999), hydrodynamic simulations suggest that up tohalf of the baryons at present should have tempera-tures between 105 and 107 K. Accretion shocks asso-ciated with filamentary structures in the cosmic webare believed to have heated gas up to these temper-atures. These shocks may also have amplified mag-netic fields and helped to accelerate particles to rela-tivistic energies. Such relativistic particles will emitfaint and diffuse radio emission that will be detectedwith the upcoming generation of radio telescopes.The detection of this diffuse radio emission by theWHIM will be an important milestone because thestudy of the structure and distribution of the WHIMis notoriously difficult as its main tracers are highlyexcited Oxygen lines that are difficult to observe.
Fig. 1. The contours show the X-ray emission from agalaxy cluster. The blue denotes the diffuse radio emis-sion. For details see Hoeft et al. (2008).
Moreover, it has been noted (Kronberg et al. 2008)that diffuse radio emission offers an alternative wayof probing large-scale structure that is quite differentfrom that defined by the distribution of galaxies. Aprediction of the radio emission around filaments andin the outskirts of clusters has also been performedby Hoeft et al. (2008) (see Figure 1).
Bagchi et al. (2002) have discovered diffuse radioemission from a large network of filaments of galaxiesthat span several Mpc containing at least 80 galax-ies. Minimising the total non-thermal energy, Bagchiet al. (2002) estimated a minimum field strength ofB ∼ 0.3 µG. This estimate hinges on several assump-tions, e.g. that the emission is synchrotron emission,that the energy ratio of protons to electrons is 1,that the spectral index of the radio-frequency spec-trum is −0.5 and that the volume filling factor isof order unity. However, meanwhile it is being de-bated whether the filament described in Bagchi etal. (2002) is really a filament or rather a group ofgalaxies.
Kronberg et al. (2008) have discovered faint dif-fuse radio emission in the vicinity of the Coma clus-ter. It indicates intergalactic magnetic fields in therange 0.2−0.4 µG on scales of up to 4 Mpc. While apossible relationship to the WHIM awaits confirma-tion, the observations suggest that some cosmic rayacceleration mechanism must exist in this medium.
Bruggen et al. (2005) have simulated the mag-netic field in filaments with sufficient resolution inorder to make predictions about their magnitude andstructure. There have been few cosmological simula-tions that include magnetic fields, and most of themfocus solely on galaxy clusters.
© 2
009:
Inst
ituto
de
Ast
rono
mía
, UN
AM
- M
ag
netic
Fie
lds
in th
e U
nive
rse
II: F
rom
La
bo
rato
ry a
nd S
tars
to th
e P
rimo
rdia
l Uni
vers
eEd
. A. E
squi
vel,
J. F
ranc
o, G
. Ga
rcía
-Se
gur
a, E
. M. d
e G
ouv
eia
Da
l Pin
o, A
. La
zaria
n, S
. Liz
ano
, & A
. Ra
ga
218 BRUGGEN
2. SIMULATIONS OF MAGNETIC FIELDS INFILAMENTS
In order to bridge the large range of scales in acosmological simulation to get down to the scales ina galaxy filament, we used the adaptive-mesh piece-wise parabolic method (PPM) code FLASH thatprovides full support for particles and cosmology.
The initial conditions for our simulations are thepublicly available conditions of the Santa Barbaracluster (Frenk et al. 1999) at redshift z = 50. Conse-quently, our cosmological parameters are those of theSanta Barbara cluster, i.e. H0 = 50 km s−1 Mpc−1,Ωm = 1 and ΩΛ = 0. The cluster perturbationcorresponds to a 3σ peak in density smoothed over10 Mpc. We assumed standard cosmic composition,i.e. the hydrogen and helium fractions were 0.7589and 0.2360, respectively. Star formation has beenneglected.
Our computational domain was a cubic box ofside L = 64 h−1 Mpc. FLASH is a modularblock-structured adaptive mesh refinement (AMR)code, parallelised using the Message Passing Inter-face (MPI) library. FLASH solves the Riemannproblem on a Cartesian grid using the Piecewise-Parabolic Method (PPM) and, in addition, includesparticles that represent the dark matter. Our simu-lation included 2,097,152 dark matter particles witha mass of 7.8 · 109 M each. Thus, in clusters andfilaments, the mean separation between dark matterparticles is less than the grid spacing. We chose ablock size of 163 zones and used periodic boundaryconditions. The minimal level of refinement was setto 4, which means that the minimal grid contains16 · 2(4−1) = 128 zones in each direction. The max-imum level of refinement was 10, which correspondsto an effective grid size of 16 · 2(10−1) = 8192 zonesor an effective resolution of 7.8 h−1 kpc. This wasthe maximum that we could afford computationally.
Depending on the dark matter density, the meshis refined and derefined automatically. It was set upsuch that no more than 8 dark matter particles occu-pied one computational cell. We first ran a simula-tion with only 8 levels of refinement, then identifiedthree regions that contained filaments, and reran thesimulation with 10 levels of refinement. In additionto refining on the dark matter we enforced full re-finements in the regions that contain the filaments.However, the dark matter distribution, representedby the particles, is not refined.
In order to achieve satisfactory accuracy even onsmall scales we decided to implement a passive mag-netic field solver in FLASH that keeps the divergenceof the magnetic field very small. The price for this
is that, currently, we are unable to compute the fullMHD equations. However, we find that the magneticfields in the intracluster medium (ICM) will rarelybe dynamically important, and, at this level, theyare described with sufficient accuracy by a passivefield solver. We do not include thermal conductionwhich may have important effects in the intraclustermedium. For a treatment of heat transfer by turbu-lent motions see Lazarian (2006).
We chose to implement the passive field solverdescribed by Pen (2003), which solves the magneticfield on a staggered mesh. It is straightforwardto show that if the magnetic field is evolved on astaggered instead of a centred grid, the magneticfield will remain divergence-free provided that it wasdivergence-free to start with (Evans & Hawley 1988).At every time step, the magnetic field is evolved us-ing a total-variation diminishing (TVD) scheme thatsolves
∂B = ∇× (v × B) −3
2
a
aB , (1)
where B = Bphysical/a and a is the scale-factor. Thethus evolved field is then interpolated with a third-order scheme to the cell centres.
Initially we set up a divergence-free field usingthe vector potential. The magnetic vector potentialA(k) was composed of 512 Fourier modes with ran-dom phases and amplitudes that were drawn from aGaussian distribution. A was then scaled as k−α,where k is the wavenumber of the mode. Thisleads to a magnetic power spectrum of PB(k) dk ∼
k2(2−α) dk. Here, α was chosen to be 2. The vectorpotential was then Fourier-transformed into physi-cal space with a Fast Fourier Transform. The fieldwas normalised such that the physical mean field atredshift 50 was 3 · 10−11 G.
In agreement with what Dolag et al. (2002) havefound using smoothed particle hydrodynamic (SPH)simulations, we verified that the resulting magneticfields are very insensitive to the topology of the ini-tial field at z = 50. Even if one starts with an ini-tially uniform magnetic field, the results are not sig-nificantly different. The field at any time obviouslydepends on the strength of the initial field. However,since the field is only evolved passively, the normal-isation is arbitrary.
3. RESULTS IN FILAMENTS
Roughly speaking, we found that the magneticfield correlates with density as one would expectfrom the conservation of magnetic flux, which im-plies B ∝ ρ2/3. However, we find that in some re-gions the magnetic field is amplified significantly be-yond B ∝ ρ2/3, which is what one would expect from
© 2
009:
Inst
ituto
de
Ast
rono
mía
, UN
AM
- M
ag
netic
Fie
lds
in th
e U
nive
rse
II: F
rom
La
bo
rato
ry a
nd S
tars
to th
e P
rimo
rdia
l Uni
vers
eEd
. A. E
squi
vel,
J. F
ranc
o, G
. Ga
rcía
-Se
gur
a, E
. M. d
e G
ouv
eia
Da
l Pin
o, A
. La
zaria
n, S
. Liz
ano
, & A
. Ra
ga
SIMULATIONS OF MAGNETIC FIELDS IN CLUSTERS AND FILAMENTS 219
compression alone. This is caused by shear flows andoccurs mainly in the outskirts of the cluster and nearaccretion shocks. An analytic calculation of mag-netic field amplification by shear has been presentedby King & Coles (2005).
If indeed filaments host magnetic fields of the or-der of B ≥ 0.3 µG, as suggested by Bagchi et al.(2002), this would require seed fields of the orderof 10−9 G at z ∼ 50. Seed fields of this order ofmagnitude may be hard to produce (see also Dolaget al. 2005). Consequently, this would call for al-ternative origins for IGM magnetic fields (Kronberg2004). Future observing campaigns with instrumentssuch as the Low Frequency Array (LOFAR) and theSquare Kilometer Array (SKA) are expected to shedsome light on magnetic fields in the early universe.
As is evident from Figure 2, we find that mag-netic fields are predominantly oriented in a directionparallel to the filaments. This is true for all filamentsin our simulation volume. As material accretes ontothe filaments, the magnetic fields are compressedparallel to the filaments, and, as clusters and groupsform, the filaments are stretched. This field topologycould have several interesting implications: For one,the diffusion length of cosmic rays along filamentsmay be much greater than the diffusion length outof the filaments. This makes filaments efficient trapsfor cosmic rays and could help to explain the dif-fuse radio emission seen along the entire length ofthe filament in Bagchi et al. (2002). Secondly, heatconduction is likely to be more efficient along fila-ments than elsewhere. This could lead to heat beingconducted from the outskirts of the clusters into thefilaments. Further details on these simulations canbe found in Bruggen et al. (2005).
4. SIMULATIONS OF MAGNETISEDAGN-BLOWN BUBBLES
In the previous section we reported on simu-lations of magnetogenesis in filaments of galaxies.Meanwhile, magnetic fields in clusters of galaxieshave received wider attention in recent years withsome numerical work, both using SPH and AMRsimulations. Here, we do not have sufficient space toreview this work. Instead, we would like to presentsome work on the effects of magnetic fields on the in-teraction between active galactic nuclei and the ICM.This is a very important issue, as the absence of cool-ing flows in galaxy clusters as well as the cut-off ofthe galaxy luminosity function is ascribed to feed-back by AGN. Radio-loud active galactic nuclei drivestrong outflows in the form of jets that inflate bub-bles or lobes. The lobes are filled with hot plasma,
Fig. 2. Magnitude of the magnetic field in a slice throughtwo filaments in our computational box. The colors rep-resent the logarithm of the magnitude of the magneticfield, and the tick marks denote Mpc. The lines showthe orientation of the magnetic field in the plane of theplot.
and can heat the cluster gas. However, the physicsof heating by AGN is still poorly understood. Oneof the main questions is how the energy in the AGN-inflated bubbles that appear as cavities in the X-raysurface brightness is transferred to the ICM.
One problem is that these cavities all appear tobe intact, even after inferred ages of several 108 yrs,as for example the outer cavities in Perseus. How-ever, hydrodynamic simulations fail to reproduce theobserved morphology as hydrodynamic instabilitiesshred the bubbles in relatively short time.
Although magnetic fields in clusters are knownto have plasma β > 1; (e.g. Blanton, Sarazin,& McNamara 2003), they may in principle have astrong effect on suppressing Kelvin-Helmholtz andRayleigh-Taylor instabilities. Jones & De Young(2005) considered the evolution of bubbles in a mag-netized ICM by performing two-dimensional numeri-cal magnetohydrodynamical (MHD) simulations andfound that bubbles could be prevented from shred-ding even when β is as high as ∼ 120. However,two-dimensional simulations cannot easily be gen-eralised to 3D problems. Here we focus on a laterstage in the evolution of the AGN-blown cavities andconsider more realistic (stochastically tangled) fieldconfigurations in fossil bubbles (i.e. after the transi-
© 2
009:
Inst
ituto
de
Ast
rono
mía
, UN
AM
- M
ag
netic
Fie
lds
in th
e U
nive
rse
II: F
rom
La
bo
rato
ry a
nd S
tars
to th
e P
rimo
rdia
l Uni
vers
eEd
. A. E
squi
vel,
J. F
ranc
o, G
. Ga
rcía
-Se
gur
a, E
. M. d
e G
ouv
eia
Da
l Pin
o, A
. La
zaria
n, S
. Liz
ano
, & A
. Ra
ga
220 BRUGGEN
Fig. 3. Natural logarithm of density distribution. Theleft-hand column shows density for the random (i) case(lower panel t = 15, upper t = 25 code time units) buttwice as high mean magnetic pressure. The right-handcolumn corresponds to the draping case for twice themean magnetic pressure compared to the original drapingcase.
tion from the momentum-driven to buoyancy-drivenstage) than in previous buoyant stage MHD simula-tions. We show that only when the power spectrumcut-off of magnetic field fluctuations is larger thanthe bubble size can the bubble shredding be sup-pressed.
The simulations were performed with the PEN-CIL code (e.g. Haugen, Brandenburg, & Dobler2003). Although PENCIL is non-conservative, it isa highly accurate grid code that is sixth order inspace and third order in time. It is particularlysuited for weakly compressible turbulent MHD flows.Magnetic fields are implemented in terms of a vectorpotential so the field remains solenoidal throughoutsimulation.
We consider magnetic fields inside the bubblesand in the ICM that are dynamically unimportantin the sense that their plasma β parameter is greaterthan unity, that is, magnetic pressure may becomeimportant when compared to the bubble ram pres-sure associated with the gas motions in the ICMbut it is generally small compared to the pressure ofthe ICM in our simulations. Regarding the geome-try of magnetic fields, we consider random isotropicfields with coherence length smaller than the bub-ble sizes (hereafter termed random), and the “drap-
ing” case of isotropic fields characterized by coher-ence length exceeding bubble size as well as a non-magnetic case. Magnetic draping has been consid-ered previously in the context of merging clustercores and radio bubbles using analytical approachby Lyutikov (2006). He found that even when mag-netic fields are dynamically unimportant through-out the ICM, a thin layer of dynamically importantfields can form around merging dense substructureclumps (bullets) and prevent their disruption. Wesuggest that, depending on the unknown value ofmagnetic diffusivity, there may be some relic mag-netic power spectrum with a smaller amplitude thanthe freshly injected one (either by AGN-driven bub-bles or dynamo-driven) that extends to scales largerthan the bubble size and that provides draping fieldsto stabilize the bubbles.
We considered three-dimensional MHD simula-tions of buoyant bubbles in cluster atmospheres forvarying magnetic field strengths characterized byplasma β > 1 and for varying field topologies. Wefind that field topology plays a key role in controllingthe mixing of bubbles with the surrounding ICM. Weshow that large-scale external fields are more likelyto stabilize bubbles than internal ones but a mod-erate stabilizing effect due to magnetic helicity canmake internal fields play a role too. The simulationsshow that a bubble morphology that resembles thefossil bubbles in the Perseus cluster can be repro-duced if the coherence of magnetic field is greaterthan the typical bubble size. While it is not clearif such a “draping” case is representative of typi-cal cluster fields, Vogt & Ensslin (2005) find thatlength-scale of magnetic fields in Hydra A is smallerthan typical bubble size. If this also holds true inother clusters then other mechanisms, such as vis-cosity, would be required to keep the bubbles stable.Unfortunately, the Faraday rotation method used byVogt & Ensslin (2005) is not very sensitive to large-scale magnetic fields if aligned with the bubble sur-face. Moreover, their maximum-likelihood methodassumes a power-law relation between magnetic fieldand density, statistical isotropy for the purpose of de-projection and a particular jet angle with respect tothe line of sight. Smaller angles and different mag-netic field configurations might yield a weaker de-cline of the power spectrum at larger scales. Takinginto account the above limitations, it is entirely pos-sible that magnetic draping provides a solution to theproblem of bubble stability. We also suggest thata hybrid model that combines helical fields insidethe bubble with external draping fields could be suc-cessful in explaining morphologies of X-ray bubbles
© 2
009:
Inst
ituto
de
Ast
rono
mía
, UN
AM
- M
ag
netic
Fie
lds
in th
e U
nive
rse
II: F
rom
La
bo
rato
ry a
nd S
tars
to th
e P
rimo
rdia
l Uni
vers
eEd
. A. E
squi
vel,
J. F
ranc
o, G
. Ga
rcía
-Se
gur
a, E
. M. d
e G
ouv
eia
Da
l Pin
o, A
. La
zaria
n, S
. Liz
ano
, & A
. Ra
ga
SIMULATIONS OF MAGNETIC FIELDS IN CLUSTERS AND FILAMENTS 221
in clusters. Another possibility is that dynamicallysignificant fields are present inside the bubbles andthe consequences of the high-β case for bubble dy-namics and stability should be investigated further.We note that the bubbles will most likely eventu-ally get disrupted (partially helped by cosmologicalsloshing gas motions in clusters). A generic featurefound in our simulations is the formation of a mag-netic wake where fields are ordered and amplified.We suggest that this effect could prevent evapora-tion by thermal conduction of cold Hα filaments ob-served in the Perseus cluster. The physical process ofbubble mixing in the presence of magnetic fields hasimportant consequences also for modelling of massdeposition and star formation rates in cool core clus-ters as well as the particle content of bubbles andcosmic-ray diffusion from them. These issues willbe further complicated by the effects of anisotropyof transport processes due to magnetic fields. Thismay give rise to the onset of magnetothermal insta-bility on the bubble-ICM interface (Parrish & Stone2005). Further details on these simulations can befound in Ruszkowski et al. (2007).
5. CONCLUSIONS
With the upcoming instruments LOFAR andSKA, the study of cosmic magnetism and cosmic rayswill take centre stage for years to come. Magnetismholds the clue for understanding the physics of theintergalactic medium and the dynamics of cosmicrays. Diffuse synchrotron emission by cosmic raysin magnetic fields will reveal the Warm-Hot Inter-galactic Medium which represents a large fractionof the baryonic matter in the universe. Here wehave presented cosmological adaptive-mesh simula-tions that predict the evolution of magnetic fieldsin filaments. In the second part of this review, wehave shown how even weak magnetic fields can havea strong impact on the dynamics of the intraclus-ter medium. In particular, we have shown how thedynamics of AGN-blown bubbles, that are vital forfeedback mechanisms in clusters and galaxies, is gov-erned by ambient magnetic fields that are effectiveat suppressing hydrodynamic instabilities.
MB gratefully acknowledges support from DFGgrant BR 2026/2 and the supercomputing grant NIC1658. This material is based upon work supported bythe National Science Foundation under the following
NSF programs: Partnerships for Advanced Compu-tational Infrastructure, Distributed Terascale Facil-ity (DTF) and Terascale Extensions: Enhancementsto the Terascale Facility. The software used in thiswork was in part developed by the DOE-supportedASCI/Alliance Center for Astrophysical Thermonu-clear Flashes at the University of Chicago.
REFERENCES
Bagchi, J., Enßlin, T. A., Miniati, F., Stalin, C. S., Singh,M., Raychaudhury, S., & Humeshkar, N. B. 2002,NewA, 7, 249
Bırzan, L., Rafferty, D. A., McNamara, B. R., Wise,M. W., & Nulsen, P. E. J. 2004, ApJ, 607, 800
Blanton, E. L., Sarazin, C. L., & McNamara, B. R. 2003,ApJ, 585, 227
Bruggen, M., & Kaiser, C. R. 2002, Nature, 418, 301Bruggen, M., Ruszkowski, M., & Hallman, E. 2005, ApJ,
630, 740Bruggen, M., Ruszkowski, M., Simionescu, A., Hoeft, M.,
& Dalla Vecchia, C. 2005, ApJ, 631, L21Carilli, C. L., & Taylor, G. B. 2002, ARA&A, 40, 319Cen, R., & Ostriker, J. P. 1999, ApJ, 514, 1Dolag, K., Bartelmann, M., & Lesch, H. 2002, A&A, 387,
383Dolag, K., Grasso, D., Springel, V., & Tkachev, I. 2005,
J. Cosmol. Astropart. Phys., 1, 9Enßlin, T. A., & Vogt, C. 2003, A&A, 401, 835Evans, C. R., & Hawley, J. F. 1988, ApJ, 332, 659Frenk, C. S., et al. 1999, ApJ, 525, 554Furlanetto, S. R., & Loeb, A. 2001, ApJ, 556, 619Govoni, F., & Feretti, L. 2004b, Int. J. Mod. Phys. D,
13, 1549Govoni, F., Markevitch, M., Vikhlinin, A., VanSpey-
broeck, L., Feretti, L., & Giovannini, G. 2004a, ApJ,605, 695
Haugen, N. E. L., Brandenburg, A., & Dobler, W. 2003,ApJ, 597, L141
Jones, T. W., & De Young, D. S. 2005, ApJ, 624, 586King, E. J., & Coles, P. 2005, arXiv:astro-ph/0508370Kronberg, P. P. 2004, J. Korean Astron. Soc., 37, 343Kronberg, P. P., Bernet, M. L., Miniati, F., Lilly, S. J.,
Short, M. B., & Higdon, D. M. 2008, ApJ, 676, 70Lazarian, A. 2006, ApJ, 645, L25Lyutikov, M. 2006, MNRAS, 373, 73Parrish, I. J., & Stone, J. M. 2007, ApJ, 664, 135Pen, U., Arras, P., & Wong, S. 2003, ApJS, 149, 447Ruszkowski, M., Enßlin, T. A., Bruggen, M., Heinz, S.,
& Pfrommer, C. 2007, MNRAS, 378, 662Vogt, C., & Enßlin, T. A. 2003, A&A, 412, 373
. 2005, A&A, 434, 67Widrow, L. M. 2002, Rev. Mod. Phys., 74, 775
