Franoise Combes March 25 2015 Gas Flows in

Françoise Combes March 25, 2015 Gas Flows in Galactic Nu Perseus cooling flow NGC 4258 1 NGC 1433 Mrk 231

Outflows more visible than inflow Ceverino et al Outflows are violent events (~1 Myr) Inflow is a slow process (~1 Gyr) Cosmic filament size~50 kpc Larger than a galaxy, diffuse gas Settles down in the disk in rotation Secular evolution, slow infall in the disk Except fountain effect, HVC. . Companions, gas bridges Agertz et al 2011 2

Necessity of outflows for quenching Star formation SN, stellar winds Baugh 2006, Eke et al 2006, Jenkins et al 2001 AGN feedback Radio jets 3

IR SFR BHAR x 3300 X-ray Madau & Dickinson 2014 Star formation history 4

Outline AGN energy is largely enough MBH=1 -2 10 -3 Mgal Egal ~Mgal s 2 EBH ~0. 1 MBHc 2 EBH/Egal > 80 AGN Pounds & King 2013 1 - AGN feedback, quenched red nuggets 2 - SN+AGN molecular outflows in galaxies 3 - Mechanisms -- Energetics 5

1 - Two main modes for AGN feedback Quasar mode: radiative or winds When luminosity close to Eddington, young QSO, high z LEdd = 4 p. GMBHmpc/s. T MBH ~f s. Ts 4, f gas fraction Same consideration with radiation pressure on dust, with sd sd /s. T ~1000, limitation of Mbulge to 1000 MBH ? Radio mode, or kinetic mode, jets When L < 0. 01 Ledd, low z, Massive galaxies, Radio E-gal Not destructive: keeps a balance cooling-heating Radiatively inefficient flow ADAF High frequency of cooling flows in clusters, Low-luminosity AGN Seyferts. . 6

Radiative mode in simulations SFR ~rn with n=1, 1. 5, 2 SN feedback+ BH growth and associated feedback Sub-grid physics How much feedback? Springel et al. (2003 -2005), Hopkins et al. 2006 Gabor & Bournaud 2014: No quenching effect 7

Compact red and blue galaxies M > 1010 Mo Black lines: density of c. SFG required to explain the red nugget tburst=0. 3 -1 Gyr Barro et al 2013 8 CANDELS

Possible scenarios Two evolutionary tracks of QG formation: (1) early (z>2), formation path of rapidly quenched c. SFGs fading into c. QGs that later enlarge, (2) late-arrival (z<2) path in which larger SFGs form extended QGs without passing through a compact state Σ 1. 5 ≡ M/r 1. 5 Barro et al 2013 9

X-ray Perseus A , Fabian et al 2003 Gas flow in cool core clusters Star formation (green) Canning et al 2014 Molecular Gas Salomé et al 2006 10

Cold gas in filaments Inflow and outflow coexist The molecular gas coming from previous cooling is dragged out by the AGN feedback The bubbles create inhomogeneities and further cooling The cooled gas fuels the AGN Velocity much lower than free-fall 11 Salome et al 2008

Large variety of simulations Brueggen et al 2007, Cattaneo & Teyssier 2007, Dubois et al 2010, Gaspari et al 2011, 2012 for clusters, or massive elliptical galaxies Cooling rate ~ boosted Bondi rate, but Cold gas accretion better Radiation pressure insufficient Mechanical feedback with jets or winds Success in moderating the cooling, keeping the CC structure Efficiency scaled to the structure scale Gaspari et al 2012 3 e-4 (E-gal) 5 e-3 (clus) 12

Comparison between models During a merger, accreted gas fuels SFR and BHAR delay, since SN-feedback too strong at the beginning (Wild et al 2010) Bondi accretion SDH Springel+05 BS Booth+09 WT Wurster+13 HPMK Hobbs+12 Thacker et al 2014 Chen, Hickox et al 2013 Other accretion models DQM Debuhr +11 ONB Okamoto+08 PNK Power+1 13

2 - Molecular outflows J 1148 Z=6. 4 CII Mrk 231 AGN and also nuclear Starburst, 107 -108 Mo Outflow 700 Mo/yr IRAM Ferruglio et al 2010 Maiolino et al 2012 On kpc scales, affects the galaxy, quenches SF? Blue wing Red wing CO Cicone et al 2012 d. M/dt = 3 v MOF/ROF ~1000 Mo/yr, (5 x. SFR) Kinetic power ~2 1044 erg/s AGN High density, HCN, HCO+, Aalto et al 2012 14

Relations outflows with AGN For AGN-hosts, the outflow rate Correlates with the AGN power Cicone et al 2014 d. M/dt v ~20 LAGN/c Can be explained by energy-driven outflows (Zubovas & King 2012) 15

Molecules in outflows? Zubovas & King 2014 Gas shock-heated to 106 -107 K Molecules dissociated Cooling efficient (free-free, metals) Multiphase, with RT instabilities Time-scale for cooling << 1 Myr At kpc scales, SF induced The SF results in a Luminosity Comparable to LAGN 100 Mo/yr! This means that SB or AGN outflows are difficult to disentangle 16 All could be due to AGN

3 - Energy-conserving outflows? If the cooling is very efficient, momentum-conserving outflow But for very fast winds > 10 000 km/s, radiative losses are slow energy-conserving flow (Faucher-Giguère & Quataert 2012) Push by the hot post-shock gas, boost the momentum Vs of the swept-up material (conservation of msvs 2 = minvin 2) Boost of vin /2 Vs ~50! Explains why momentum flux >> LAGN/c // Adiabatic phase, or Sedov-Taylor phase in SN remnant 17

Slow cooling --High momentum fluxes Faucher-Giguère & Quataert 2012 c UFO b Tombesi 10, 14 a v/c Costa, Sijacki, Haehnelt, 2014 18

Impact of a fast wind (UFO) on the ISM Two phases 1 - Starting with MBH below the M-s relation, the v=0. 1 c wind is stopped by an ISM shock << 1 kpc Momentum-driven flow, energy radiated, not enough to stop the gas, the BH grows 2 - When M-s relation reached, the flow can extend to larger radii > 1 kpc, the flow becomes energy-driven The pressure is >> rv 2, the ISM is easily ejected, with Vesc (seen in molecular outflows) Regulation of Mbulge A disk is a too massive obstacle, the jet is diverted bipolar King & Pounds 2015 19

Several modes simulated Quasar mode, when d. MBH/dt > 0. 01 Edd – Energy released spherically Radio-jets otherwise: V= 104 km/s, in a cylinder perp. to the disk Energy-driven, much more efficient than momentum-driven AGN outflows with > 10 Ledd/c Entrained cold gas > 109 Mo If after-shock cooling with metals Costa et al 2014 Comparison between realistic cosmological simulations and idealised ones in spherically gravitational potentiels A factor 10 larger momentum flow is required for AGN feedback 20 to be efficient

Quasar mode: Two-phase simulations Most of the outflow kinetic energy escapes through the voids Positive and negative feedback Cold gas is pushed by ram-pressure More feedback on low-density gas Could lead to M-s relation Nayakshin 2014 21

AGN wind feedback fails May be AGN positive feedback SF feedback? Starburst in a ring AGN could trigger a starburst by compressing the gas Gaibler et al 2012 22

AGN-triggered star formation The AGN provides an extra-pressure forming more clumps in the molecular gas Bieri et al 2015 23

Radio mode: Fractal structure 2 pc 1 kpc Efficient relativistic jets; Influence of the porosity Wagner & Bicknell 2011 24

Jet in the disk plane NGC 4258 Cecil et al 2000 25

Positive AGN feedback Radio jets triggered SF Silk 2005, Dubois et al 2013 Young, restarted radio loud AGN 4 C 12. 50 The outflow is located 100 pc from the nucleus where the radio jet interacts with the ISM Morganti et al 2013, Dasyra & Combes 2012 26

Feedback in low-luminosity AGN NGC 1433: barred spiral, CO(3 -2) with ALMA Molecular gas fueling the AGN, + outflow // the minor axis MH 2= 5. 2 107 Mo in FOV=18’’ 100 km/s flow 7% of the mass= 3. 6 106 Mo Smallest flow detected Lkin=0. 5 d. M/dt v 2 ~2. 3 1040 erg/s Lbol (AGN)= 1. 3 1043 erg/s Flow momentum > 10 LAGN/c Combes et al 2013 27

SUMMARY: AGN outflows Mechanisms: Quasar modes (winds), powerful AGN Or Radio modes (jets), for low-luminosity AGN, lower z Molecular outflows are now observed frequently, around AGN, v=200 -1200 km/s 107 -109 Mo, load factors 1 -5 Energy-conserving flow: Momentum boost 20 LAGN/c However, not efficient to quench SF (or even trigger SF? ) AGN radio mode is very efficient in Cool Core clusters to moderate the cooling: mechanical with radio jets, cold gas accretion 28