Heating and Cooling in Clusters & Galaxies

Transcription

1 Physics 463, Spring 07 Lecture 12 Heating and Cooling in Clusters & Galaxies review of the white & rees model hydrodynamical simulations hot and cold flows multi-phase cooling feedback: photoionization, SN, AGN Key: dark matter hot gas t 1 t 2 dark matter: violent relaxation gas: shock heated to virial temperature gas within a cooling radius cools, and its collapse is halted by angular momentum to form a disk Shock Waves Basic idea of shocks: occurs in supersonic motion of a fluid: sound waves traveling against the flow cannot travel and pressure builds, creating a high pressure shock wave a sharp increase in density, pressure, temperature and speed at the shock cold gas White & Rees, White & Frenk 1991 t 3 cooling radius grows with time, feeding cold gas into the disk t 4 when Rcool < Rvir, accretion from a quasi-hydrostatic gas halo regulated by the cooling rate when Rcool > Rvir, gas accretion by infall rate

2 Virial Shock Heating A gas velocity map overlayed on the emission-weighted temperature map of the cluster undergoing major merger at z=0.43. The map shows the average projected velocity and entropy in a 125h-1kpc (comoving) slice centered on the central density peak. The size of the region shown is 8h-1Mpc. D Nagai as halos collapse, the dark matter undergoes violent relaxation within the virial radius, dm shells cross and the pressure of the gas increases, prevents shell crossing of the gas gas is infalling at greater than the sound speed. pressure makes this velocity vanish at the center. because it s supersonic, info about the boundary condition can t propagate outwards, and a shock is created. as the shock propagates outwards, gas that crosses the shock is heated. this increases the sound speed and makes the interior flow subsonic. net result is transfer of KE of the collapse into thermal energy of the gas in order to persist, the gas internal to the shock has to have pressure. Shocks in Cosmology!"#$%&'$()'&($#''*#+,&,()+,(-.""&,(/.%0()(1*23"&4(3)%%&'+(*-($0*15(/)6&$ ( 789!:7(9;(<=;;>?>@A(BC!8(@DBE>?(C?9D@<(C(!FD7A>??G#H(:)+IH(8)""2)+(J(K*+&$(LMNNOP Virial Shocks Shocks in merging clusters Shocks in moderately dense filaments -- heat the IGM Shocks in the ISM and ICM from eg. SN or other energy sources

3 !"#$%"$#&'()#*+",)-'./)%0.'%+-'1&'."$2,&2',-'.,"$',-'!34 Cooling Basics Cooling Rates In the White & Rees model (standard semianalytic) H HeText metal line-cooling K Bremstralung T>10 8 K cooling rate defines cooling time radius where cooling time = t Universe defines the cooling radius within which gas cools

4 basics of hydro simulations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

5 !"##$%$&'()*+%,+*&-."/-0(1".20-'",&(/,+$1(-%$(3%,-+0*("&(-4%$$.$&'5( -03$"'(6"')(1231'-&'"-0(1/-''$%(-&+(+"##$%$&/$1("&(+$'-"0 ( 789(:;<7;(=;>=;>;(?@A:79>(?BCD;>E:B<(D>BF9?7 G%$&H5(I)"'$(J(KL(/,M-2'),%1((NOPPPQ Agertz et al 2006 Figure 4. Gas density slices through the centre of the cloud at t = 0.25, 1.0, 1.75 and 2.5 τkh. From top to bottom we show Gasoline (Gas 10m), GADGET-2 (Gad 10m), Enzo (Enzo 256), FLASH (FLASH 256) and ART-Hydro (ART 256). The grid simulations clearly show dynamical instabilities and complete fragmentation after 2.5 τkh, unlike the SPH simulations in which most of the gas remains in a single cold dense blob. key observations The Galaxy Luminosity Function Bimodality in galaxy colors Early formation of galaxies in clusters Properties of gas in clusters galaxy formation is not equally efficient at all masses: need for feedback and/or non-standard cooling

6 10-1 Halo Baryons: f b M v Standard Cooling: M Rc Only ~8% of baryons are in stars (Fukagita, Hogan and Peeples 1998, Bell et. al. 2003, Fukagita et al. 2004) N(>M) Mpc Bell et al. 03 Halo Baryons: f b M v M(M O ) N(>M) Mpc Bell et al M(M O ) galaxy formation is not equally efficient at all masses: need for feedback and/or non-standard cooling A Maller A Maller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

8 SN Feedback Supernova Feedback time Mori et al. 1996ApJ T Photoionization squelching Supernova Feedback Scale Dekel & Silk 86 T SN feedback ineffective dwarfs/lsb H SN feedback effective +1 log gas density tcool<tff He 30 in presence of UV ionizing background, halos with virial temp < background radiation field are unable to accrete gas (! < km/s) gas can be boiled out of halos (!<20 km/s) cooling function modified (cooling suppressed at low T) naturally leads to supression of low mass galaxies. Brems. CDM H 100 virial velocity "Mreheat = # (4/3) ($SNESN/Vvir2) "Mstar Theul & Weinberg, Bullock, Kratsov & Weinberg 2002, Somerville 2001 Benson et al. 2002

9 Cold Flows filtering mass Gnedin 2000 Growth of a Massive Galaxy T K A Less Massive Galaxy T K M _ M _ M _ shock-heated gas disc cold infall shocked disc Spherical hydro simulation Birnboim & Dekel 03 Spherical hydro simulation Birnboim & Dekel 03

10 in our statistics has only a small impact, increasing the apparent importance of cold accretion at low redshift (see the dotted line in Fig. 2 below). Our measured rates of smooth accretion really represent the sum of genuinely smooth gas flow and mergers with galaxies below our resolution limit, including systems that are not even under-resolved bound objects in the simulation but would nevertheless exist in the real Universe. In Section 4.1 we extrapolate the measured distribution of merger mass ratios to show that this subresolution merging is unlikely to be an important correction, so that our measured smooth accretion rates indeed correspond mainly to smooth gas flows. We use the following redshift spacing for determining T max values and calculating accretion rates: for 0! z! 1 we use!z 0.125, for 1 < z! 3.5 we use!z 0.25, for 3.5 < z! 5 we use!z 0.5, and for higher redshifts we use!z 1. In the cosmology we consider here, these redshift intervals correspond to time intervals of roughly!t = 0.25 Gyr for z > 3.5,!t 0.5 Gyr at z = 1.5 and!t 1.5 Gyr at z = 0. These time intervals are close to the typical infall time-scales at the corresponding outputs, except at Hydro Simulation: ~Massive M=3x1011 Kravtsov et al. z=4 M=3x1011 Tvir=1.2x106 Rvir=34 kpc virial shock ρ/ρ and T K is comprised of shock-heated gas in virialized haloes and, at the lower-density end, around filaments. The narrow, downward-sloping locus at ρ/ρ > 103 and T 104 K represents radiatively cooled, dense gas in galaxies. Cooling times at these densities are short, so gas remains close to the equilibrium temperature where photoionization heating balances radiative cooling, which is a slowly decreasing function of density. According to the conventional sketch described in the introduction, gas that ends up in galaxies starts in the diffuse phase, enters the shock-heated phase, then cools and condenses to reach the galactic phase. Fig. 1(b) plots the ρ T trajectories of 15 randomly selected gas particles accreted on to galaxies near z = 3, starting near z = 15. Some of these particles follow just the path described above. However, approximately half of them start in the diffuse IGM phase and move directly to the dense galactic phase without ever heating above 105 K. The virial temperature of the smallest resolved halo in the simulation is K at z = 3, so these particles have never been close to the virial temperature of any resolved halo. The separation between outputs in Fig. 1(b) is typically Gyr, and Less Massive M=1.8x1010 Kravtsov et al. cold infall virial radius Figure 1. Left: distribution of gas particles in the ρ T plane at z = 3, in the L22/128 simulation. One can easily identify three major phases: low-density, low-temperature gas in the photoionized IGM, shock-heated overdense gas and high-density radiatively cooled gas within galaxies. Right: trajectories of 15 particles that accreted on to galaxies shortly before z = 3, illustrating the cold (solid lines, circles) and hot (dashed lines, triangles) accretion modes. Hot-mode particles are shock heated above K before cooling, while cold-mode particles move directly from the diffuse IGM phase to the dense, galactic phase without ever heating above 105 K. Trajectories start at z = 14.9 and end at z = 3. Points mark the individual redshift outputs, which have typical time separations of Gyr. $ C z=9 M=1.8x1010 Tvir=3.5x105 Rvir=7 kpc Keres et al 2005 Fraction of cold/hot accretion Mass Distribution of Halo Gas SPH simulation disk Keres, Katz, Weinberg, Dav e 2004 cold flows Z=0, underestimating Mshock density shockheated sharp transition adiabatic infall M cold! M #2 / 3 M tot " 2005 RAS, MNRAS 363, 2 28 M! M 2/3 L Temperature Analysis of Eulerian hydro simulations by Birnboim, Zinger, Dekel, Kravtsov fraction of cold vs hot accretion Keres et al

11 Shock-heating Scale Birnboim & Dekel 03 T M~6x1011M! cold infall into discs +1 log gas density -1 CDM typical halos: reside in relatively thick filaments, fed spherically no cold flows tcool<tff He H virial velocity 300 the millenium cosmological simulation 1000 Origin of dense filaments in hot halos (M$Mshock) at high z At low z, Mshock halos are typical residiing in thicker filaments of comparable density At high z, Mshock halos are high-_ peaks - fed by a few thinner filaments of higher density Large-scale filaments grow selfsimilarly with M*(t) and always have typical width ~R*!"#1/3 high-sigma halos: fed by relatively thin, dense filaments cold flows Brems. shock heating H Ms~M* Ms>>M* Multi-phase cooling

12 N(>M) Mpc -3 A Maller Bell et al. 03 Halo Baryons: f b M v Standard Cooling: M Rc M(M O ) 2/3 of the baryons cool forming a Milky Way mass of 12 x 10 10, twice what is observed. Supernova feedback is invoked to blowout half of the Milky Ways while not destroying the thin disk. A Maller Cooling and Fragmentation in Astrophysical Plasmas A cooling plasma is hydrodynamically unstable (Field 1965) Higher density regions will cool faster, becoming denser and therefore cooling faster. Low density regions won t cool as quickly, will expand into the space left by the high density regions, thus decreasing their density and the rate that they cool. One ends up with a two phase medium of low density hot gas and warm clouds. A Maller A Maller

13 Cloud-Cloud collisions Clouds will move around in the halo until they collide or lose energy from ram pressure. Only then will they merge with the galaxy. M h ρ c R 3 c Mass (10 10 M O ) Galaxy M g Hot Gas M h Clouds M cl Total M Rc Lookback time, t (Gyr) N(>M) Mpc Bell et al. 03 Halo Baryons: f b M v Standard Cooling: M Rc Multiphase Cooling: Mc Central Galaxy: M g M(M O ) Multi-phase cooling seems to predict this cutoff naturally. density in a pure disk galaxy Figure 9. Temperature maps of the gas in slices through the centre of the standard M33 simulation after 2.1 Gyr. Panels (a) and (b) show 200 and 32 kpc regions, respectively. Panel (c) shows a 40-kpc slice perpendicular to the disc plane. Kaufmann et al 2006 High Velocity Clouds: HI clouds detected around our galaxy not associated with the disk.

14 When gas cools we expect a two phase medium to arise. AGN Feedback To survive warm clouds must have masses of solar masses. To get the observed Milky Way mass clouds must have masses of To match the observed properties of high velocity clouds requires cloud masses of 3-7 x 10 6 solar masses. A similar range of cloud masses is needed to produce quasar absorption systems. Irrespective of the cloud mass, the hot low density core sets an upper limit on the amount of mass that can cool of 2 x solar masses explaining the exponential cutoff in the Luminosity function. A Maller M BH -! relation QSO luminosity function Haehnelt & Kauffmann 2000

15 The rise and fall of quasar activity The co-evolution of galaxies, quasars, and SMBH basic premise: major mergers feed gas to central BH and trigger AGN activity! built-in connection between BH and spheroid formation when galaxies merge, their central BH merge as well (conserving mass) energy injection from the AGN AGN radio mode The AGN radio mode: Gas Heating In the radio mode, quiescent hot gas accretes onto the central supermassive black hole. This ongoing accretion comes from the surrounding hot halo, where we capture the mean behaviour with an empirical equation. 200 z [ h -1 kpc ] x [ h -1 kpc ] T [K] z [ h -1 kpc ] x [ h -1 kpc ] Fig. 1. Mass-weighted temperature maps of a h 1 M isolated halo, subject to AGN bubble heating. The velocity field of the gas is over-plotted with white arrows T [K] Populating Galaxies with Black Holes AGN quasar mode The AGN This quasar accretion mode: is typically well below the Eddington limit In the quasar mode, super-massive black holes grow through merging events where black holes coalesce and cold disk gas (DC is driven et al. onto 2005) the central black hole. D Croton This is the primary mode of black hole growth (Kauffmann & Haeanelt 2000)

16 Luminosity Functions Galaxy Colours and Ages B-V colour bi-modality and mean stellar age The K and bj-band luminosity functions with and without AGN Why Does Such Heating Work? Shock Heating Triggers AGN Feedback Unlike other heating mechanisms (e.g. super-winds, starbursts,...), AGN heating suppresses star formation without itself requiring star formation to efficiently operate. Unlike event mechanisms (e.g. merger driven quasar winds), whatever quenches star formation needs to be an ongoing process (local massive ellipticals are not quasars!). M>M shock More than enough energy is available in AGNs Hot gas is vulnerable to AGN feedback, while cold streams are shielded An AGN-like low energy heating source, fed from the hot x-ray halo, is an energetically feasible candidate Ṡhock heating is the trigger for AGN feedback in massive halos Kravtsov et al.

17 joint formation of SMBH and spheroids leads naturally to scaling relations BUT strong differential feedback needed also needed to make big BH at z~6 without overproducing BH mass density rise and fall of quasar activity (maybe) tied to same physics as star formation rate in galaxies (merger rate/gas supply) feedback 1 0 photo-ionization UV on dust cold hot SN M vir [M _ ] AGN + hot medium dynamical friction in groups groups R Somerville While halos grow by mergers and accretion M<M crit : The Blue Sequence cold gas supply disk growth & star formation SN-fdbk regulates star formation long duration bursts very blue mergers & bar instability bulges M>M crit : The Red Sequence shock-heated gas +AGN fdbk no new gas supply +gas exhausted + AGNs especially in bulges no disk growth, star formation shuts off passive stellar evolution red & dead further growth of spheroids by gas-poor mergers!"#$%&$'%(')*+),-.//'012'$%34567%1($'10'86569#'012367%1( ' <6;%67%=:'&115%(8>'?@'A6&B8214(;'C"1318:(:14$D E4A2:$1547%1('3457%F"6$:'31;:5'012'7":'/EGH'E762'012367%1('6(;'0::;A6&B!":(13:(1518%&65'31;:5'012'8656&7%&'I%(;$ +:76%5:;'&":3%&65':(2%&"3:(7 -":2365'&1(;4&7%1( G68(:71."#;21;#(63%&$ J1(.7":2365'2:567%=%$7%&'&13F1(:(7'C&1$3%&'26#$D )21I7"'10'$4F:236$$%=:'A56&B'"15:$'6(;'*)J'0::;A6&B M4AA5:'":67%(8'6(;'0::;A6&B'A#'*)J E"1&B';:7:&7%1(!"#$%&65'=%$&1$%7#'=%6'J6=%:2.E71B:$':N467%1(