1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3: \usepackage{apjfonts}
4:
5: \usepackage{graphicx,natbib}
6:
7: \begin{document}
8:
9: \title{Simulations of AGN feedback in galaxy clusters and groups: impact on
10: gas fractions and the $L_{\rm X}-T$ scaling relation}
11:
12: \shorttitle{Simulations of AGN feedback in galaxy clusters and groups}
13: \shortauthors{E. Puchwein, D. Sijacki and V. Springel}
14:
15: \author{E. Puchwein\altaffilmark{1},
16: D. Sijacki\altaffilmark{2},
17: V. Springel\altaffilmark{1}}
18:
19: \affil{$^{1}$Max-Planck-Institut f\"{u}r Astrophysik,
20: Karl-Schwarzschild-Stra\ss{}e 1, 85740 Garching bei
21: M\"{u}nchen, Germany}
22: \affil{$^{2}$Institute of Astronomy, Madingley
23: Road, Cambridge, CB3 0HA, United Kingdom}
24:
25: %\altaffiltext{1}{Max-Planck-Institut f\"{u}r Astrophysik,
26: % Karl-Schwarzschild-Str. 1, 85740 Garching bei
27: % M\"{u}nchen, Germany}
28: %\altaffiltext{2}{Institute of Astronomy, Madingley
29: % Road, Cambridge, CB3 0HA, UK}
30:
31: \email{puchwein@mpa-garching.mpg.de}
32:
33: \begin{abstract}
34: Recently, rapid observational and theoretical progress has established that
35: black holes (BHs) play a decisive role in the formation and evolution of
36: individual galaxies as well as galaxy groups and clusters. In particular,
37: there is compelling evidence that BHs vigorously interact with their
38: surroundings in the central regions of galaxy clusters, indicating that any
39: realistic model of cluster formation needs to account for these
40: processes. This is also suggested by the failure of previous
41: generations of hydrodynamical simulations without BH physics to
42: simultaneously account for the paucity of strong cooling flows in clusters,
43: the slope and amplitude of the observed cluster scaling relations, and the
44: high-luminosity cut-off of central cluster galaxies. Here we use
45: high-resolution cosmological simulations of a large
46: cluster and group sample to study how BHs affect their host systems. We focus on
47: two specific properties, the halo gas fraction and the X--ray luminosity-temperature scaling relation, both of which are notoriously difficult to
48: reproduce in self-consistent hydrodynamical simulations. We show that BH
49: feedback can solve both of these issues, bringing them in
50: excellent agreement with observations, without alluding to the `cooling
51: only' solution that produces unphysically bright central galaxies. By
52: comparing a large sample of simulated AGN-heated clusters with observations,
53: our new simulation technique should make it possible to reliably calibrate
54: observational biases in cluster surveys, thereby enabling various
55: high-precision cosmological studies of the dark matter and dark energy
56: content of the universe.
57: \end{abstract}
58:
59: \keywords{methods: numerical, galaxies: clusters: general, cosmology: theory,
60: black hole physics}
61:
62:
63: \section{Introduction}
64:
65: Numerous observational and theoretical studies of galaxy clusters and groups
66: show that the astrophysical processes relevant for determining their
67: properties are still poorly understood. Hydrodynamical simulations of
68: cluster formation which only include radiative cooling processes
69: \citep[e.g.][]{Lewis2000, Muanwong2001, Yoshida2002} suffer from excessive
70: overcooling within the densest cluster regions, where gas cooling times are
71: short. This translates into a very large fraction of cold gas and consequently
72: a large amount of stars that would form out of it, reaching at least $40-50\%$
73: of the baryonic cluster content. While the associated removal of low-entropy
74: gas from cluster centers breaks the self-similarity of the cluster scaling
75: relations in a way that resembles observations
76: \citep{Bryan2000, Dave2002, Nagai2007}, this `cooling-only' scenario is
77: generally discounted because observed clusters do not show such strong cooling
78: flows and enormously bright central galaxies. Instead, it is widely believed
79: that some non-gravitational energy input must strongly affect the
80: physics of the intracluster medium (ICM).
81:
82:
83: There has been considerable effort \citep[see][and references
84: therein]{Borgani2004} to include feedback mechanisms associated with star
85: formation in hydrodynamical simulations in order to reduce excessive
86: overcooling in cluster cores. However, so far simulation
87: models have failed to simultaneously reproduce the masses and colors of
88: central galaxies, the observed temperature and metallicity profiles, as well
89: as the observed X--ray luminosity-temperature ($L_{\rm X}-T$) relation. In
90: particular, the X--ray luminosities have been found to be substantially larger
91: than the observed values \citep{Borgani2004} on the group
92: scale. Furthermore the observed baryon fractions in clusters and
93: groups are typically smaller than in simulations, which for `standard' physics invariably predict a value close to the universal cosmic baryon fraction \citep[e.g.][]{Frenk1999,OShea2005,Kravtsov2005}.
94:
95:
96: Several attempts have been made to tackle these
97: problems with so-called `preheating' schemes
98: \citep[e.g.][]{Bialek2001,Borgani2005}, that are meant to mimic more energetic
99: astrophysical processes than the relatively inefficient supernovae feedback.
100: However, the preheating scenarios are physically not well motivated and rely
101: on an ad-hoc choice of the amount and epoch of energy injection into the ICM.
102:
103: On the other hand, it is becoming increasingly clear that active galactic
104: nuclei (AGN) play an important role for understanding the properties of
105: clusters and groups \citep[see ][for a recent
106: review]{McNamara2007}. Also analytic work
107: \citep[e.g.][]{Valageas1999, Bower2001, Cavaliere2002} suggests that including
108: self-consistent feedback from AGN in simulations might resolve the
109: discrepancies by providing a heating mechanism that offsets cooling, lowers
110: star formation rates and removes gas from the centers of poor clusters and
111: groups, thereby reducing their X--ray luminosities to values compatible with
112: the observed $L_{\rm X}-T$ relation.
113:
114: In this work we investigate whether AGN feedback can indeed solve these
115: problems by performing hydrodynamical simulations of galaxy clusters and
116: groups that employ a state-of-the-art model \citep{Sijacki2007} for BH growth and associated feedback processes.
117:
118: \section{Methodology} \label{SecMethods}
119:
120: We have selected a large sample of cluster- and group-sized halos from the
121: Millennium simulation \citep{Springel2005a} and resimulated them at higher
122: resolution, including a gaseous component and accounting for hydrodynamics,
123: radiative cooling, heating by a UV background, star formation and supernovae
124: feedback. For each halo, two kinds of resimulations were performed. One
125: containing the physics just described and an additional one including a model
126: for BH growth and associated feedback processes as in \cite{Springel2005b} and
127: \cite{Sijacki2007}. This allows us to compare the very same clusters simulated
128: with and without AGN heating in order to clearly pin down the imprints of AGN
129: activity on galaxy cluster and group properties, and to which extent this
130: resolves discrepancies with X--ray observations.
131:
132: \subsection{The simulations}
133:
134: In total we have selected $21$ Millennium run dark matter (DM) halos at
135: $z=0$, and resimulated them at much higher mass and force resolution. The selection was only based on mass and otherwise random, aiming to
136: approximately uniformly cover a large mass range from $8\!\times\! 10^{12} M_\odot$
137: to $1.5 \! \times \! 10^{15} M_\odot$. New initial conditions were created by
138: populating the Lagrangian region of each halo in the original initial
139: conditions with more particles and adding additional small-scale power, as
140: appropriate. At the same time, the resolution has been progressively reduced
141: in regions that are sufficiently distant from the forming halo. Gas
142: has been introduced into the high-resolution region by splitting
143: each parent particle into a gas and a DM particle.
144:
145:
146: We have adopted the same flat $\Lambda$CDM
147: cosmology as in the parent Millennium simulation, namely $\Omega_{\rm m}=0.25$, $\Omega_{\rm \Lambda}=0.75$, $h=0.73$,
148: $n_s=1$ and $\sigma_8=0.9$. $\Omega_{\rm b}=0.04136$ has been chosen so as to
149: reproduce the cosmic baryon fraction inferred from current cosmological
150: constraints \citep{Komatsu2008}.
151:
152:
153: In our high-resolution resimulations of halos with virial masses below $2
154: \! \times \! 10^{14} h^{-1} M_\odot$, the DM particle mass is $m_{\rm DM} =
155: 3.1 \! \times \! 10^7 h^{-1} M_\odot$, the gas particle mass is $m_{\rm gas} = 6.2
156: \! \times \! 10^6 h^{-1} M_\odot$ and the gravitational softening is $\epsilon = 15
157: \, h^{-1} \rm kpc$ comoving (Plummer-equivalent) for redshift $z>5$, which is
158: then replaced with a physical softening of $2.5 \, h^{-1} \rm kpc$ for
159: $z<5$. For four of these halos we have additionally performed very high resolution simulations
160: with mass resolution improved by a factor $(3/4)^3$. The change in halo properties resulting from this resolution increase is negligibly small, especially compared to the impact of the AGN heating, indicating that results are approximately converged and our comparison is not affected by resolution issues.
161: For the four most massive clusters we used a somewhat lower
162: resolution with a DM particle mass of $m_{\rm DM} = 1.1 \! \times \! 10^8 h^{-1} M_\odot$ and a gas particle mass of $m_{\rm gas} = 2.1
163: \! \times \! 10^7 h^{-1} M_\odot$. This makes these simulations computationally
164: affordable, while assuring reasonable convergence.
165:
166: The simulations were run with the {\small GADGET-3} code \citep[based
167: on][]{Springel2005c}, which employs an entropy-conserving formulation of
168: smoothed particle hydrodynamics (SPH). Radiative cooling
169: and heating was calculated for an optically thin plasma of hydrogen and helium,
170: and for a time-varying but spatially uniform UV background. We did not include metal-cooling as we wanted to study the impact of AGN feedback independently from the problems involved in following the metal-enrichment of the ICM. One should keep in mind, however, that metal-cooling might have some effect on halo properties on the group scale.
171: Star formation and supernovae feedback were modeled with a subresolution multi-phase model
172: for the interstellar medium as in \cite{Springel2003}.
173:
174: \subsection{The black hole growth and feedback model}
175:
176: Here we summarize the main features of our model for incorporating BH
177: growth and feedback in simulations; full details are given in
178: \cite{Springel2005b}, \cite{Sijacki2006}, and \cite{Sijacki2007}. In short, we assume that
179: low-mass seed BHs are produced sufficiently frequently that any halo above a
180: certain threshold mass contains one such BH at its center. In the simulations,
181: an on-the-fly friends-of-friends group finder puts seed BHs with a mass of
182: $10^5 h^{-1} M_\odot$ into halos when they exceed a mass of $5
183: \! \times \! 10^{10} h^{-1} M_\odot$ and do not contain any BH yet. The BHs are
184: represented by collisionless sink particles and are allowed to grow by mergers
185: with other BHs and by accretion of gas at the Bondi-Hoyle-Lyttleton rate, but with the Eddington limit
186: additionally imposed. Two BHs are merged when they fall within their
187: local SPH smoothing lengths and have small relative velocities.
188:
189: Motivated by growing theoretical and observational evidence that AGN feedback
190: is composed of two modes \citep[e.g.][and references therein]{Chrurazov2005}, we use two
191: distinct feedback models depending on the BH accretion rate (BHAR) itself
192: \citep[see][]{Sijacki2007}. For large accretion rates above $0.01$ of the Eddington rate, the bulk of AGN heating is
193: assumed to be in the form of radiatively efficient quasar activity with only a
194: small fraction of the luminosity being thermally coupled to the ICM. We adopt this thermal heating efficiency to be
195: $0.5\%$ of the rest mass-energy of the accreted gas, which reproduces the
196: observed relation between BH mass and bulge stellar velocity dispersion
197: \citep{DiMatteo2005}. For BHARs below $0.01$ of the Eddington rate, we assume
198: that feedback is in a so-called ``radio-mode'', where AGN jets inflate hot,
199: buoyantly rising bubbles in the surrounding ICM. The duty cycle
200: of bubble injection, energy content of the bubbles as well as their initial
201: size are determined from the BHAR. We assume the mechanical
202: feedback efficiency provided by the bubbles to be $2\%$ of the
203: accreted rest mass-energy, which is in good agreement with observations of
204: X--ray luminous elliptical galaxies \citep{Allen2006}.
205:
206: It was shown in \cite{Sijacki2007} that this model leads to a self-regulated BH growth and
207: brings BH and stellar mass densities into broad agreement with observational
208: constraints.
209:
210: \subsection{X--ray properties}
211: \label{sec:X-ray properties}
212:
213: We obtain realistic X--ray luminosities and spectroscopic
214: temperatures for each simulated halo by sorting the gas particles inside
215: $r_{500}$, the radius that encloses a mean density 500 times the critical
216: density today, into temperature bins and summing up the emission measure for each
217: bin. Using the XSPEC package \citep{Arnaud1996}, we then simulate a spectrum of the X-ray emission in that region as a sum of MEKAL emission models \citep{Liedahl1995}, one for each temperature bin, assuming a constant metallicity of 0.3 times the solar value
218: and using {\em Chandra}'s response function. In order not to be limited by photon
219: noise we adopted a large exposure time of $10^6 \textrm{s}$. The combined
220: spectrum of the emission inside $r_{500}$ is then fit by a single temperature
221: MEKAL model with temperature, metallicity and normalization of the spectrum as
222: free parameters. The resulting emission model then yields an estimate of the
223: bolometric luminosity $L_{500}$, while the spectroscopic temperature $T_{500}$
224: is taken directly from the fitted model. Note, however,
225: that we use the gas particles inside the three-dimensional radius $r_{500}$ to
226: calculate $L_{500}$, while we use the gas particles inside the
227: projected radius $r_{500}$ to obtain $T_{500}$. Also note
228: that we exclude very cold high-density gas particles with
229: $T \! < 3 \!\times\! 10^4 \rm K$ and densities above 500 times the mean baryon density as well as multiphase particles to avoid spurious contributions from the multiphase model for the star-forming gas.
230:
231: \pagebreak
232:
233: \section{Results} \label{SecResults}
234:
235: We focus in this work on how AGN feedback affects the $L_{\rm X}-T$ relation and the gas mass fractions in
236: clusters. Additional cluster properties and simulations where the feedback energy in the ``radio mode'' is not injected
237: thermally but in the form of cosmic rays \citep{Sijacki2008} will be
238: discussed in a forthcoming companion paper.
239:
240: \subsection{Halo gas fractions}
241:
242: In Figure~\ref{fig:gas_fractions}, we show the ratio of
243: gas mass to total mass inside the radius $r_{500}$ as a function of halo
244: X--ray temperature. For each simulated halo, arrows connect the
245: results obtained without and with AGN heating. For
246: comparison, we show constraints on halo gas fractions obtained from X--ray
247: observations \citep{Vikhlinin2006,Sun2008}. Also shown are gas fractions we
248: computed from the gas density and temperature profile parameters given in
249: \cite{Sanderson2003}.
250:
251: The most obvious effect of the AGN feedback is the significantly reduced gas
252: fraction at the low temperature end of our sample, i.e. in poor clusters and
253: groups. There the AGN heating drives a substantial fraction of the gas to
254: radii outside of $r_{500}$. This lowers halo gas fractions in spite of the
255: reduced fraction of gas that is converted into stars in the runs with
256: AGN. The potential wells of massive clusters are, on the other hand,
257: too deep for AGN heating to efficiently remove gas from them. Thus the effect
258: of the suppressed star formation becomes more important or even dominant
259: towards more massive systems. While the gas fraction in the very inner regions
260: of massive clusters is somewhat reduced by the AGN, we find it
261: unchanged or slightly increased within $r_{500}$.
262:
263: \begin{figure}
264: %\begin{center}
265: \scalebox{0.72}{\includegraphics{f1.eps}}
266: %\end{center}
267: %\plotone{f1.eps}%
268: \caption{Halo gas mass fractions within $r_{500}$ of clusters
269: and groups simulated without AGN feedback (circles) and
270: with the feedback model included (squares). The arrows illustrate the
271: effect of the AGN heating for each halo.}
272: \label{fig:gas_fractions}
273: \end{figure}
274:
275: \subsection{The $L_{\rm X}-T$ relation}
276:
277: Given that AGN heating removes gas from the centers of poor clusters and
278: groups, it is no surprise that it also suppresses their X--ray luminosities
279: and affects the $L_{\rm X}-T$ relation. In Figure~\ref{fig:L_X-T_relation}, we
280: plot the X--ray luminosities $L_{500}$ against the spectroscopic temperatures
281: $T_{500}$, for all halos of our
282: sample. The arrows indicate the change due to the AGN feedback for each
283: halo. Data from a number of observational X--ray studies is shown for
284: comparison
285: \citep{Horner2001, Helsdon2000, Osmond2004, Arnaud1999, Markevitch1998}.
286:
287: Without the AGN feedback, we obtain substantially larger X--ray luminosities
288: for poor clusters than observed, while for massive clusters there
289: is reasonable agreement. This finding is in line with previous numerical
290: studies \citep[see e.g.][]{Borgani2004}, and is a manifestation of the
291: long-standing problem to explain the scaling relations of galaxy clusters
292: in hydrodynamical simulations. However, when we employ the model for BH growth and feedback the discrepancies between simulated and observed
293: $L_X-T$ relation are resolved. In particular, the $L_X-T$ relation on the
294: group scale is steepened significantly, as AGN heating removes a larger
295: fraction of gas from smaller halos and thereby reduces their X--ray luminosity. For massive clusters, the effect of the
296: feedback is less important. Overall, the $L_X-T$ relation obtained from the
297: simulations with AGN feedback is consistent with observations at all mass
298: scales, from massive clusters to small groups.
299:
300:
301: Note that while the $L_X-T$ relation of massive clusters is in a
302: reasonable agreement with observations even without the AGN feedback, this is
303: only because an unrealistically large amount of cold gas, which is eventually converted into stars, is produced so that the stellar
304: fractions within the virial radius reach of order of $35-45\%$, in clear
305: conflict with observations. On the other hand, AGN-heated massive clusters
306: not only lie closer to the observed $L_X-T$ relation but also
307: have much lower stellar fractions which are reduced by at least one
308: third. Especially their central galaxy is prevented from becoming too bright,
309: due to the suppression of strong cooling flows. Also, the mass fraction of stars bound to cluster galaxies agrees very well
310: with observations in our simulations with AGN. There is also a significant component of intra-cluster stars of up to 50\% which
311: lies at the upper end of the observational estimates \citep[see e.g.][for an overview]{Lin2004}.
312:
313: \begin{figure*}
314: \plotone{f2.eps}%
315: \caption{X--ray luminosity as a function of spectroscopic temperature for
316: clusters and groups simulated without AGN
317: feedback (circles) and with the feedback model included (squares).
318: For each halo, the arrow illustrates the effect of the AGN
319: heating. Data from several observational X-ray studies is shown for comparison. For a subset also the best-fit power-law
320: $L_{\rm X}-T$ relation is plotted. Including AGN feedback into the
321: simulations drastically improves the agreement with observations and
322: resolves the discrepancy that otherwise exists on the group scale.}
323: \label{fig:L_X-T_relation}
324: \end{figure*}
325:
326: \section{Summary and conclusions} \label{SecConclusions}
327:
328: We have performed very high-resolution numerical simulations of a
329: mass-selected sample of galaxy clusters and groups. In order to investigate
330: whether AGN feedback makes simulations and X--ray observations of cluster
331: and groups compatible, we have carried out two kinds of simulations for each
332: halo, namely (1) hydrodynamical simulations with a treatment of radiative
333: cooling, star formation, supernovae feedback, and heating by a UV background,
334: and (2) simulations that additionally employ a model for black hole growth and
335: associated feedback processes. Our main findings are:
336: \begin{itemize}
337: \item AGN feedback significantly lowers the gas mass fractions in poor
338: clusters and groups, even though fewer baryons are turned into stars at the
339: same time. This is because the AGN heating drives gas from
340: halo centers to their outskirts. In massive clusters, on the other hand, it mainly lowers the central gas density and
341: substantially reduces the amount of stars formed. Overall, both the gas and
342: stellar fractions of the whole sample of our simulated groups and clusters
343: are in a much better agreement with observations when AGN are included
344: than in simulations without AGN.
345: \item AGN feedback significantly reduces the X--ray luminosities of poor
346: clusters and groups, while the X--ray temperature within $r_{\rm 500}$ stays
347: roughly the same or is even slightly reduced. This results in a steepening
348: of the $L_{\rm X}-T$ relation on the group scale.
349: \item The $L_{\rm X}-T$ relation obtained from simulations with AGN feedback
350: is in excellent agreement with observations at all mass scales.
351: \end{itemize}
352:
353: We find it extremely encouraging that this simple model for BH growth and
354: feedback is capable of bringing the analyzed
355: properties of galaxy clusters and groups into much better agreement with
356: observations. This not only resolves a
357: long-standing problem in their hydrodynamical modeling, but
358: also opens up new exciting possibilities for using numerical simulations to
359: investigate the properties of clusters and groups, and their
360: co-evolution with central supermassive BHs. Also, it considerably brightens
361: the prospects to use simulations of cluster formation to accurately
362: calibrate and correct systematic effects in future X-ray and
363: Sunyaev-Zel'dovich cluster surveys. This is essential for the exploitation of
364: the full potential of clusters as cosmological probes to accurately constrain
365: the expansion history of the Universe.
366:
367: \acknowledgements We thank Simon White, Martin Haehnelt, Klaus Dolag and Gabriel Pratt for constructive discussions. D.S. acknowledges Postdoctoral Fellowship from the STFC. Part of the simulations were run on the Cambridge HPC cluster Darwin.
368:
369: \bibliographystyle{apj}
370:
371: \begin{thebibliography}{37}
372: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
373:
374: \bibitem[{{Allen} {et~al.}(2006){Allen}, {Dunn}, {Fabian}, {Taylor}, \&
375: {Reynolds}}]{Allen2006}
376: {Allen}, S.~W., {Dunn}, R.~J.~H., {Fabian}, A.~C., {Taylor}, G.~B., \&
377: {Reynolds}, C.~S. 2006, \mnras, 372, 21
378:
379: \bibitem[{{Arnaud}(1996)}]{Arnaud1996}
380: {Arnaud}, K.~A. 1996, in ASP Conference Series,
381: Vol. 101, Astronomical Data Analysis Software and Systems V, ed. G.~H.
382: {Jacoby} \& J.~{Barnes}, 17
383:
384: \bibitem[{{Arnaud} \& {Evrard}(1999)}]{Arnaud1999}
385: {Arnaud}, M. \& {Evrard}, A.~E. 1999, \mnras, 305, 631
386:
387: \bibitem[{{Bialek} {et~al.}(2001){Bialek}, {Evrard}, \& {Mohr}}]{Bialek2001}
388: {Bialek}, J.~J., {Evrard}, A.~E., \& {Mohr}, J.~J. 2001, \apj, 555, 597
389:
390: \bibitem[{{Borgani} {et~al.}(2005){Borgani}, {Finoguenov}, {Kay}, {Ponman},
391: {Springel}, {Tozzi}, \& {Voit}}]{Borgani2005}
392: {Borgani}, S., {Finoguenov}, A., {Kay}, S.~T., {Ponman}, T.~J., {Springel}, V.,
393: {Tozzi}, P., \& {Voit}, G.~M. 2005, \mnras, 361, 233
394:
395: \bibitem[{{Borgani} {et~al.}(2004){Borgani}, {Murante}, {Springel}, {Diaferio},
396: {Dolag}, {Moscardini}, {Tormen}, {Tornatore}, \& {Tozzi}}]{Borgani2004}
397: {Borgani}, S., {Murante}, G., {Springel}, V., {Diaferio}, A., {Dolag}, K.,
398: {Moscardini}, L., {Tormen}, G., {Tornatore}, L., \& {Tozzi}, P. 2004, \mnras,
399: 348, 1078
400:
401: \bibitem[{{Bower} {et~al.}(2001){Bower}, {Benson}, {Lacey}, {Baugh}, {Cole}, \&
402: {Frenk}}]{Bower2001}
403: {Bower}, R.~G., {Benson}, A.~J., {Lacey}, C.~G., {Baugh}, C.~M., {Cole}, S., \&
404: {Frenk}, C.~S. 2001, \mnras, 325, 497
405:
406: \bibitem[{{Bryan}(2000)}]{Bryan2000}
407: {Bryan}, G.~L. 2000, \apjl, 544, L1
408:
409: \bibitem[{{Cavaliere} {et~al.}(2002){Cavaliere}, {Lapi}, \&
410: {Menci}}]{Cavaliere2002}
411: {Cavaliere}, A., {Lapi}, A., \& {Menci}, N. 2002, \apjl, 581, L1
412:
413: \bibitem[{{Churazov} {et~al.}(2005){Churazov}, {Sazonov}, {Sunyaev}, {Forman},
414: {Jones}, \& {B{\"o}hringer}}]{Chrurazov2005}
415: {Churazov}, E., {Sazonov}, S., {Sunyaev}, R., {Forman}, W., {Jones}, C., \&
416: {B{\"o}hringer}, H. 2005, \mnras, 363, L91
417:
418: \bibitem[{{Dav{\'e}} {et~al.}(2002){Dav{\'e}}, {Katz}, \&
419: {Weinberg}}]{Dave2002}
420: {Dav{\'e}}, R., {Katz}, N., \& {Weinberg}, D.~H. 2002, \apj, 579, 23
421:
422: \bibitem[{{Di Matteo} {et~al.}(2005){Di Matteo}, {Springel}, \&
423: {Hernquist}}]{DiMatteo2005}
424: {Di Matteo}, T., {Springel}, V., \& {Hernquist}, L. 2005, \nat, 433, 604
425:
426: \bibitem[Frenk et al.(1999)]{Frenk1999} Frenk, C.~S., et al.\
427: 1999, \apj, 525, 554
428:
429: \bibitem[{{Helsdon} \& {Ponman}(2000)}]{Helsdon2000}
430: {Helsdon}, S.~F. \& {Ponman}, T.~J. 2000, \mnras, 315, 356
431:
432: \bibitem[{{Horner}(2001)}]{Horner2001}
433: {Horner}, D.~J. 2001, PhD thesis, University of Maryland College Park
434:
435: \bibitem[{{Komatsu} {et~al.}(2008){Komatsu}, {Dunkley}, {Nolta}, {Bennett},
436: {Gold}, {Hinshaw}, {Jarosik}, {Larson}, {Limon}, {Page}, {Spergel},
437: {Halpern}, {Hill}, {Kogut}, {Meyer}, {Tucker}, {Weiland}, {Wollack}, \&
438: {Wright}}]{Komatsu2008}
439: {Komatsu}, E., et~al. 2008,
440: ArXiv e-prints, 0803.0547
441:
442: \bibitem[{{Kravtsov} {et~al.}(2005){Kravtsov}, {Nagai}, \&
443: {Vikhlinin}}]{Kravtsov2005}
444: {Kravtsov}, A.~V., {Nagai}, D., \& {Vikhlinin}, A.~A. 2005, \apj, 625, 588
445:
446: \bibitem[{{Lewis} {et~al.}(2000){Lewis}, {Babul}, {Katz}, {Quinn}, {Hernquist},
447: \& {Weinberg}}]{Lewis2000}
448: {Lewis}, G.~F., {Babul}, A., {Katz}, N., {Quinn}, T., {Hernquist}, L., \&
449: {Weinberg}, D.~H. 2000, \apj, 536, 623
450:
451: \bibitem[{{Liedahl} {et~al.}(1995){Liedahl}, {Osterheld}, \&
452: {Goldstein}}]{Liedahl1995}
453: {Liedahl}, D.~A., {Osterheld}, A.~L., \& {Goldstein}, W.~H. 1995, \apjl, 438,
454: L115
455:
456: \bibitem[Lin
457: \& Mohr(2004)]{Lin2004} Lin, Y.-T., \& Mohr, J.~J.\ 2004, \apj, 617, 879
458:
459: \bibitem[{{Markevitch}(1998)}]{Markevitch1998}
460: {Markevitch}, M. 1998, \apj, 504, 27
461:
462: \bibitem[{{McNamara} \& {Nulsen}(2007)}]{McNamara2007}
463: {McNamara}, B.~R. \& {Nulsen}, P.~E.~J. 2007, \araa, 45, 117
464:
465: \bibitem[{{Muanwong} {et~al.}(2001){Muanwong}, {Thomas}, {Kay}, {Pearce}, \&
466: {Couchman}}]{Muanwong2001}
467: {Muanwong}, O., {Thomas}, P.~A., {Kay}, S.~T., {Pearce}, F.~R., \& {Couchman},
468: H.~M.~P. 2001, \apjl, 552, L27
469:
470: \bibitem[{{Nagai} {et~al.}(2007){Nagai}, {Vikhlinin}, \&
471: {Kravtsov}}]{Nagai2007}
472: {Nagai}, D., {Vikhlinin}, A., \& {Kravtsov}, A.~V. 2007, \apj, 655, 98
473:
474: \bibitem[O'Shea et al.(2005)]{OShea2005} O'Shea, B.~W., Nagamine,
475: K., Springel, V., Hernquist, L., \& Norman, M.~L.\ 2005, \apjs, 160, 1
476:
477: \bibitem[{{Osmond} \& {Ponman}(2004)}]{Osmond2004}
478: {Osmond}, J.~P.~F. \& {Ponman}, T.~J. 2004, \mnras, 350, 1511
479:
480: \bibitem[{{Sanderson} {et~al.}(2003){Sanderson}, {Ponman}, {Finoguenov},
481: {Lloyd-Davies}, \& {Markevitch}}]{Sanderson2003}
482: {Sanderson}, A.~J.~R., {Ponman}, T.~J., {Finoguenov}, A., {Lloyd-Davies},
483: E.~J., \& {Markevitch}, M. 2003, \mnras, 340, 989
484:
485: \bibitem[{{Sijacki} {et~al.}(2008){Sijacki}, {Pfrommer}, {Springel}, \&
486: {En{\ss}lin}}]{Sijacki2008}
487: {Sijacki}, D., {Pfrommer}, C., {Springel}, V., \& {En{\ss}lin}, T.~A. 2008,
488: \mnras, 387, 1403
489:
490: \bibitem[{{Sijacki} \& {Springel}(2006)}]{Sijacki2006}
491: {Sijacki}, D. \& {Springel}, V. 2006, \mnras, 366, 397
492:
493: \bibitem[{{Sijacki} {et~al.}(2007){Sijacki}, {Springel}, {di Matteo}, \&
494: {Hernquist}}]{Sijacki2007}
495: {Sijacki}, D., {Springel}, V., {di Matteo}, T., \& {Hernquist}, L. 2007,
496: \mnras, 380, 877
497:
498: \bibitem[{{Springel}(2005)}]{Springel2005c}
499: {Springel}, V. 2005, \mnras, 364, 1105
500:
501: \bibitem[{{Springel} {et~al.}(2005{\natexlab{a}}){Springel}, {Di Matteo}, \&
502: {Hernquist}}]{Springel2005b}
503: {Springel}, V., {Di Matteo}, T., \& {Hernquist}, L. 2005{\natexlab{a}}, \mnras,
504: 361, 776
505:
506: \bibitem[{{Springel} \& {Hernquist}(2003)}]{Springel2003}
507: {Springel}, V. \& {Hernquist}, L. 2003, \mnras, 339, 289
508:
509: \bibitem[{{Springel} {et~al.}(2005{\natexlab{b}}){Springel}, {White},
510: {Jenkins}, {Frenk}, {Yoshida}, {Gao}, {Navarro}, {Thacker}, {Croton},
511: {Helly}, {Peacock}, {Cole}, {Thomas}, {Couchman}, {Evrard}, {Colberg}, \&
512: {Pearce}}]{Springel2005a}
513: {Springel}, V., et~al. 2005{\natexlab{b}}, \nat, 435, 629
514:
515: \bibitem[{{Sun} {et~al.}(2008){Sun}, {Voit}, {Donahue}, {Jones}, \&
516: {Forman}}]{Sun2008}
517: {Sun}, M., {Voit}, G.~M., {Donahue}, M., {Jones}, C., \& {Forman}, W. 2008,
518: ArXiv e-prints, 805.2320
519:
520: \bibitem[{{Valageas} \& {Silk}(1999)}]{Valageas1999}
521: {Valageas}, P. \& {Silk}, J. 1999, \aap, 350, 725
522:
523: \bibitem[{{Vikhlinin} {et~al.}(2006){Vikhlinin}, {Kravtsov}, {Forman}, {Jones},
524: {Markevitch}, {Murray}, \& {Van Speybroeck}}]{Vikhlinin2006}
525: {Vikhlinin}, A., {Kravtsov}, A., {Forman}, W., {Jones}, C., {Markevitch}, M.,
526: {Murray}, S.~S., \& {Van Speybroeck}, L. 2006, \apj, 640, 691
527:
528: \bibitem[{{Yoshida} {et~al.}(2002){Yoshida}, {Stoehr}, {Springel}, \&
529: {White}}]{Yoshida2002}
530: {Yoshida}, N., {Stoehr}, F., {Springel}, V., \& {White}, S.~D.~M. 2002, \mnras,
531: 335, 762
532:
533: \end{thebibliography}
534:
535: \end{document}
536: