Draftversion January5,2012 PreprinttypesetusingLATEXstyleemulateapjv.5/2/11 CLEARING OUT A GALAXY Kastytis Zubovas1 and Andrew King1 Draft version January 5, 2012 ABSTRACT It is widely suspected that AGN activity ultimately sweeps galaxies clear of their gas. We work out the observable properties requiredto achieve this. Large–scaleAGN–driven outflows should have 2 kinetic luminosities ∼ ηL /2 ∼ 0.05L and momentum rates ∼ 20L /c, where L is the Edd Edd Edd Edd 1 Eddington luminosity of the central black hole and η ∼ 0.1 its radiative accretion efficiency. This 0 creates an expanding two–phase medium in which molecular species coexist with hot gas, which can 2 persist after the central AGN has switched off. This picture predicts outflow velocities ∼ 1000− n 1500 kms−1 and mass outflow rates up to 4000 M⊙yr−1 on kpc scales, fixed mainly by the host a galaxy velocity dispersion (or equivalently black hole mass). All these features agree with those of J outflows observed in galaxies such as Mrk231. This strongly suggests that AGN activity is what 4 sweepsgalaxiesclearoftheirgasonadynamicaltimescaleandmakesthemredanddead. Wesuggest future observational tests of this picture. ] Subject headings: galaxies: evolution — quasars: general — black hole physics — accretion A G 1. INTRODUCTION We return to this problem here, as it offers a clear . h observationaltest of the idea that AGN outflows are re- Recently, three groups (Feruglio et al. 2010; p sponsiblefor makinggalaxiesredanddead. To keepour Rupke & Veilleux 2011; Sturm et al. 2011) have - treatment as general as possible (specifically, indepen- o used molecular spectral line observations to re- r veal fast (v ∼ 1000 km s−1) kpc–scale, massive dent of the details of numerical simulations) we adopt st (M˙out ∼ 100o0utM⊙yr−1) outflows in the nearby quasar aprseidmicptlseoauntaflloywticvaelpopcritoiaecsh∼. W10e00fin−d1t5h0a0tktmhiss−p1r,ocaensds [a Mrk231. Other galaxies show indications of similar mass outflow rates up to ∼ 4000 M⊙yr−1, several hun- phenomena (e.g. Riffel & Storchi-Bergmann 2011b,a; dredtimes the Eddingtonvalue,ingoodagreementwith 1 Sturm et al. 2011). These appear to show how quasar observations. In addition, we find that the observable v feedback can transform young, star–forming galaxies momentum outflow rate is ∼ 20 times greater than L/c 6 into red and dead spheroids. All three groups reach ofthedrivingAGN,alsoinagreementwithobservations. 6 this conclusion for Mrk231 essentially by noting that We conclude thatAGN outflowsaregoodcandidatesfor 8 the mass outflow rate M˙out and the kinetic energy the agency sweeping galaxies clear of gas. 0 rate E˙ = M˙ v2 /2 of the outflow are too large . out out out 1 to be driven by star formation, but comparable with 2. WINDS 0 those predicted in numerical simulations of AGN 2 feedback. The kinetic energy rate is a few percent TodriveoutflowswithE˙out approachingLEdd atlarge 1 of the likely Eddington luminosity L = 4πGMc/κ radius, the active nucleus of a galaxy must somehow Edd : of the central black hole, of mass M (where κ is the communicate this luminosity from its immediate vicin- v ity. Direct transport by radiation is problematic, not electron–scattering opacity). The outflowing material i X must have a multi-phase structure, because v greatly least because galaxies are generally optically thin (How- out ever, dust opacity may be large enough to absorb this r exceeds the velocity corresponding to the molecular a dissociation temperature (v .10 km s−1; see Section radiationand provide feedback; see Murray et al. 2005). diss 5 below). Jets are sometimes invoked, but are relatively inefficient becausetheytendtodrillholesintheinterstellarmedium In a recent paper (King et al. 2011) we showed that ratherthandrivingitbodilyaway. Accordinglythemost large–scale flows of this type (technically, an energy– driven flow, see Section 3) can indeed drive much of the likely connection is via high–velocity wide–angle winds expelled from the vicinity of the nucleus by radiation interstellar gas out of a galaxy bulge on a dynamical timescale∼108yr,leavingitredanddead,providedthat pressure (e.g. Pounds et al. 2003a,b). Recent observa- the central supermassive black hole accretes for about tions suggest that such winds are very common in AGN (Tombesi et al. 2010a,b). In this paper, we use the term twice the Salpeter time after reaching the value set by the M −σ relation. In Power et al. (2011) we showed ‘wind’ to refer to the mildly relativistic (v ∼ 0.1c) ejec- that the remaining bulge mass is close to the value set tion of accretion disc gas from the immediate vicinity by the observed black–hole – bulge–mass relation (e.g. of the SMBH resulting from Eddington accretion, and Ha¨ring & Rix2004). Howeverwedidnotinvestigatethe ‘outflow’ (see Section 4) for the large–scale nonrelativis- observable features of this process, including in particu- ticflowscausedbytheinteractionbetweenthewindand lar the way that the interstellar gas is swept up. the galaxy’s ambient gas. The winds have simple properties. With mass rate 1TheoreticalAstrophysicsGroup,UniversityofLeicester,Le- M˙w ∼ M˙Edd, where M˙Edd = LEdd/ηc2 is the Ed- icesterLE17RH,U.K.;[email protected] dington accretion rate and η is the accretion efficiency, 2 Fig. 1.— Diagram of momentum–driven (top) and energy–driven (bottom) outflows. In both cases a fast wind (velocity ∼ ηc ∼ 0.1c) impactstheinterstellargasofthehostgalaxy,producinganinnerreverseshockslowingthewind,andanouterforwardshockaccelerating theswept–upgas. Inthemomentum–drivencase,theshocksareverynarrowandrapidlycool tobecomeeffectively isothermal. Onlythe ram pressureiscommunicated to the outflow, leading to very low kinetic energy ∼(σ/c)LEdd. In anenergy–driven outflow, the shocked regionsaremuchwideranddonotcool. Theyexpandadiabatically, communicatingmostofthekineticenergyofthewindtotheoutflow (insimplecases approximately 1/3rd isretained by the shocked wind). The outflow radial momentum flux is thereforegreater than that ofthewind. Momentum–drivenflowsoccurwhenshockshappenwithin∼1kpcoftheAGN,andestablishtheM−σrelation(King2003, 2005). Once the supermassive black hole mass attains the critical M −σ value, the shocks move further from the AGN and the outflow becomes energy–driven. Thisproduces theobservedlarge–scaleflows,whichprobablysweepthegalaxyclearofgas. a wide–angle wind has scattering optical depth ∼ 1 (cf King 2010). The likely ionization equilibrium of the (King & Pounds 2003), assuming that the covering fac- wind is such that it produces X–rays (King 2010). In tor of absorbing gas is close to unity (see the discussion line with these expectations, blueshifted X–ray iron ab- beloweqn2). Soeachdrivingphotononaveragescatters sorptionlines correspondingto velocities ∼0.1careseen aboutoncebefore escapingtoinfinity andgivesup allof in a significant fraction of local AGN (e.g. Pounds et al. its momentum to the wind, so that the wind mass flow 2003a,b; Tombesi et al. 2010a,b), justifying our assump- rate M˙ and velocity v obey tion of a covering factor close to unity. In all cases the w inferred wind mass flow rates agree with eqn. (1). So L M˙ v ∼ (1) theseblackholewindshavemomentumandenergyrates w c L L Defining m˙ =M˙w/M˙Edd ∼1 as the Eddington factor of P˙w ∼m˙ Ecdd ∼ Ecdd, (3) the wind, we immediately find and v η 1 η ∼ ∼0.1 (2) E˙ = M˙ v2 ∼ L ∼0.05L , (4) w w Edd Edd c m˙ 2 2m˙ 3 where we have used eqns (1, 2) in eqn (4). than the shocked wind and swept–up ISM: the cooling functiontheyuseextendsonlytotemperatures∼107 K, 3. SHOCKS rather than the wind shock temperature ∼1011 K – see The expression (4) for the energy rate of a black hole King et al. (2011) for details. wind is obviously promising for driving the observed large–scale outflows. Although the interstellar medium 4. OUTFLOWS is clumpy, the outflow bubble inflated by the wind eas- For an isothermal ISM density distribution with ve- ily sweeps pastthe clumps, affecting the diffuse gas(e.g. locity dispersion σ and gas fraction f (the ratio of gas c Mac Low & McCray 1988). Furthermore, the clouds are density to background potential density) one can solve shockedby the passing outflow and evaporate inside the analyticallytheequationofmotionfortheshockpattern hotwindbubble(Cowie & McKee1977),somostoftheir for both momentum–driven flow (King 2003, 2005) and material also joins the outflow. A detailed treatment of energy–drivenflow (King 2005; King et al. 2011). the interaction between the clumpy ISM and the wind– In the momentum–driven case there are two distinct drivenoutflowisbeyondthescopeofthispaper,although flow patterns, depending onthe black hole mass M. For we address some of the implications in Section 6. In the M <M , where σ present analysis we assume that most of the sightlines f κ fsrpoemctitvheeoSfMwhBeHthaerrethceoyvearreedawlsiothobdsiffcuurseedmbeydciulumm,pisr.re- Mσ = πGc 2σ4 ∼−4×108M⊙σ2400 (5) The question now is how efficiently the wind energy (with f = 0.16 (its cosmological value) and σ = c 200 is transmitted to the outflow. This depends crucially σ/(200 kms−1)) the wind momentum is too weak to on how the wind interacts with the diffuse interstellar driveawaytheswept–upISM,andtheflowstallsatsome medium of the host galaxy. Since the wind is hyper- point. For M > M the wind momentum drives the sonic,itmustdecelerateviolentlyinareverseshock,and σ swept–up matter far from the nucleus. It is intuitively simultaneously drive a forwardshock into the host ISM. reasonable to assume that the black hole cannot easily There are two possible outcomes, which are realised un- growits masssignificantlybeyondthe pointwhereitex- der different conditions in galaxies. pelsthelocalinterstellargasinthisway,i.e. beyondM . The first outcome (momentum–driven flow) occurs if Equation(5)isveryclosetotheobservedM−σrelationσ, the shockedwind gascan coolon a timescale shortcom- despite having no free parameter. Detailed calculations pared with the motion of the shock pattern. In this (Power et al.2011)showthattheSMBHislikelytogrow case the shocked wind gas is compressed to high density for 1-2 additional Salpeter times after it reaches M , in- and radiates away almost all of the wind kinetic energy σ creasing its final mass by a factor of a few. This process (i.e. E˙out << E˙w = (η/2m˙ )LEdd). This shocked wind is even more pronounced at higher redshift, as then it has gas pressure equal to the pre–shock ram pressure takes longer for the outflow to clear the galaxy, so the P˙w ∼−LEddm˙ /c∝M,andthis pushesinto the hostISM. SMBH must be active for longer. The second case (energy–driven flow) occurs if the We conclude that outflows drive gas far from the nu- shocked wind gas is not efficiently cooled, and instead cleus, and thus become energy–driven, once M & M . σ expands as a hot bubble. Then the flow is essentially This is evidently the case needed to explain the molecu- adiabatic, and has the wind energy rate, i.e. E˙ ∼− lar outflows seen in Mrk231 and other galaxies. out E˙ = (η/2m˙ )L ∼ 0.05L (from eqn 4). The hot w Edd Edd 5. LARGE–SCALEFLOWS bubble’s thermal expansion makes the driving into the host ISM more vigorous than in the momentum–driven Inanenergy–drivenflowtheadiabaticexpansionofthe case. Observedgalaxy–wide molecular outflows must be shocked wind pushes the swept–up interstellar medium energy–driven,as demonstrated directly by their kinetic in a ‘snowplow’. King et al. (2011) derive the analytic energy content (cf eqn 4). solutionfortheexpansionoftheshockedwindinagalaxy Which of these two very different cases occurs at a bulge with an isothermal mass distribution. With AGN given point depends on the cooling of the shocked gas. luminosity lLEdd, all such solutions tend to an attractor It is easy to show that the usual atomic cooling pro- cesses(free–freeandfree–boundradiation)arenegligible R˙ =v ∼− 2ηlfcσ2c 1/3 ∼−925σ2/3(lf /f )1/3 km s−1 in all cases. The dominant process tending to cool the e 3f 200 c g (cid:20) g (cid:21) shocked black hole wind is the inverse Compton effect (6) (Ciotti & Ostriker 1997). The quasar radiation field is until the central AGN luminosity decreases significantly muchcooler than the wind shock temperature (typically atsomeradiusR=R ,whentheexpansionspeeddecays 0 ∼ 107 K and ∼ 1011 K respectively), and so cools the as shocked wind provided that it is not too diluted by dis- 10 1 2 10 R˙2 =3 v2+ σ2 − − σ2 (7) tance. Equations 8 and 9 of King (2003) show that this e 3 x2 3x3 3 holds if and only if the shock is at distances R . 1 kpc (cid:18) (cid:19)(cid:18) (cid:19) fromthe active nucleus, since the Compton cooling time where x = R/R ≥ 1. In eq. (6), f is the gas fraction 0 g goes as R2 and the flow time typically as R. So we ex- relative to all matter. This may be lower than the value pect momentum–driven flow close to the nucleus, and f prevailingwhentheearliermomentum–drivenoutflow c energy–drivenflow if gascan be drivenfar awayfrom it. establishestheM−σrelation(5),asgasmaybedepleted We note that Silk & Nusser (2010) claim that an through star formation for example. energy–driven flow never occurs. However they seem to Thesolutions(6,7)describethemotionofthe contact have considered the cooling of the ambient gas, rather discontinuitywhere the shockedwindencounters swept– 4 TABLE 1 Observed outflowparameters Object MBH/M⊙ σ/kms−1 Lbol/ergs−1 (l) M˙out/M⊙ yr−1 vout/kms−1 Mrk231(a) 4.7·107(b) 120(b) 45.69(c) (0.80) 420 1100 Mrk231(d) 4.7·107 120 45.69(0.80) 700 750 Mrk231(e) 4.7·107 120 46.04(f) (1.8) 1200 1200 IRAS08572+3915(e) ∼4.5·107∗ 120∗∗∗ 45.66(1∗) 970 1260 IRAS13120–5453(e) 5.3·106∗ 70∗∗∗ 44.83(1∗) 130 860 IRAS17208–0014(e)∗∗ − − 45.11(≪1) 90 370 Mrk1157(g) 8.3·106 100 42.57(3.4·10−3) 6 350 2QZJ002830.4-281706(h) 5.1·109(i) 385∗∗∗ 46.58(5.8·10−2) 2000 2000 Outflows observed inmolecular gas (Mrk231, IRAS 08572+3915, IRAS 13120–5453 and IRAS 17208–0014), and warm ionisedgas (Mrk1157, 2QZJ002830.4-281706). ∗ - the AGN is assumed to be radiating at its Eddington limit; ∗∗ - the galaxy is starburst– dominated(Riffel&Storchi-Bergmann2011b),soweexpectalowEddingtonfactorandhencemakenoestimates;∗∗∗ -theSMBH isassumedtolieontheM−σ relation;notethatthismaybequestionable insomecases(cfMcConnelletal.2011). References: a-Rupke&Veilleux(2011);b-Tacconietal.(2002);c-Lonsdaleetal.(2003);d-Feruglioetal.(2010);e-Sturmetal. (2011);f -Veilleuxetal.(2009);g -Riffel&Storchi-Bergmann(2011b);h -Cano-Diazetal.(2011);i -Shemmeretal.(2004). TABLE 2 Observationallyderived versustheoretically predicted outflow parameters Object 0.0E˙5oLubtol M˙oLutbvoolutc fL≡ MM˙˙oaucct fL,pred. M˙pred./M⊙ yr−1 vpred./kms−1 Mrk231 0.66 18 490=222 840 880 810 Mrk231 0.51 20 820=292 840 880 810 Mrk231 1.0 25 1400=372 1110 1150 1060 IRAS08572+3915 2.1 50 1200=352 910 950 875 IRAS13120–5453 0.88 31 1080=332 1870 220 610 IRAS17208–0014 0.06 4.9 396=202 − − − Mrk1157 1.3 110 9270=962 170 85 115 2QZJ002830.4-281706 1.3 20 307=17.52 74 8200 740 The first three columns give quantities derived from observations of large–scale outflows summarized in Table 1. The last three columnsgivethemass–loadingparameter,masssweep–outrateandterminalvelocitypredictedbyourequations (14),(15)and(8) respectively. We assume the simplest case of an isotropic outflow. Collimation would reduce the predicted mass outflow rate and increase the predicted outflow velocity. With one outlier (see below), the outflow kineticenergy is always veryclose to 5% ofLbol (1stcolumn)aspredictedbyeq. (4),andthemomentumloading(2ndcolumn)isalwaysverysimilartothesquarerootofthemass loading(rhsof3rdcolumn),aspredictedbyeq. (18). Itisstrikingthattherelationholdsforlocalquasars(Mrk231),high-redshift quasars (2QZJ002830.4-281706) and low luminosity galaxies (Mrk1157). The last two columns can be directly compared with the lasttwocolumnsofTable1;thediscrepanciesariseduetostrongoutflowcollimation. Theonlysignificantoutlier,IRAS17208–0014, isknowntobeastarburst-dominatedgalaxy,sowewouldnotexpecttheoutflow tobedominatedbytheAGNcontribution. up interstellar gas (see Figure 1). The observed molecu- The outflow rate of shocked interstellar gas is lar lines are likely to come from the shocked interstellar gasaheadofthisdiscontinuity–itstemperatureismuch dM(R ) (γ+1)f σ2 lower (∼ 107 K) than that of the shocked wind, as we M˙out = out = g R˙. (10) dt G shall see. The outer shock must run ahead of the con- tact discontinuity into the ambient interstellar medium Assuming M =M , the wind outflow rate is σ in sucha way that the velocity jump acrossit is a factor (γ+1)/(γ−1) (where γ is the specific heat ratio). This 4f m˙ σ4 M˙ ≡m˙ M˙ = c . (11) fixes its velocity as w Edd ηcG γ+1 lf 1/3 Wecannowdefineamass–loadingfactorfortheoutflow, vout = 2 R˙ ∼−1230σ220/03 fc km s−1 (8) which is the ratio of the mass flow rate in the shocked (cid:18) g (cid:19) ISM to that in the wind: (where we have used γ = 5/3 in the last form). This M˙ η(γ+1)f R˙c corresponds to a shock temperature of order 107 K for f ≡ out = g . (12) the forward shock into the interstellar medium (as op- L M˙w 4m˙ fc σ2 posed to ∼ 1010−11 K for the wind shock). Since the Then the mass outflow rate is outer shock and the contact discontinuity are very close togetherwhenenergy–drivenflowstarts(seeFig. 1)this η(γ+1)f R˙c means that the outer shock is always at M˙ =f M˙ = g M˙ . (13) out L w 4 f σ2 Edd c γ+1 R = R. (9) If the AGN is still radiating at a luminosity close to Ed- out 2 5 dington, we have R˙ =v , and using (6) gives galaxy. The outflows should have mechanical luminosi- e ties E˙ ∼(η/2)L ∼ 0.05L , but (scalar) momen- out Edd Edd f = 2ηc 4/3 fg 2/3 l1/3 ∼−460σ−4/3l1/3, (14) tumratesP˙out ∼20LEdd/c. Thesepredictionsagreewell L 3σ f m˙ 200 m˙ withobservations(seeTables1and2). Weconcludethat (cid:18) (cid:19) (cid:18) c(cid:19) AGNoutflows maywellbe whatsweepsgalaxiesclearof and gas. M˙out ∼−3700σ280/03l1/3 M⊙yr−1 (15) Our picture predicts several other features that may aid in interpreting observations. It suggests that the for typical parameters, fg = fc and γ = 5/3. If the molecularoutflowscomefromclumpsofcoolgasembed- central quasar is no longer active, the mass outflow rate ded in the outflowing shocked ISM. They are entrained evidently declines as R˙/v times this expression, with R˙ bytheadvancingoutershockfrontandpersistforalong e given by (7). time. We note that this shock front is Rayleigh–Taylor It is easy to check from (8, 15) that the approximate stable since interstellar gas is compressed here. Fur- equality ther, the temperature of the shocked ISM is in the right 1 1 range for thermal instability (McKee & Ostriker 1977), 2M˙wvw2 ∼− 2M˙outvo2ut. (16) and Richtmyer-Meshkov instabilities (Kane et al. 1999) inducedbytheforwardshockmeanthatnewcoldclumps holds, i.e. most of the wind kinetic energy ultimately mayformintheoutflowbehindit. SimulationsbyZubo- goes into the mechanical energy of the outflow, as ex- vas & Nayakshin (in preparation) show that up to 10% pected for energy driving. One can show from the equa- ofthe totalmassinthe outflowmaybe lockedup inthis tions in King (2005) that if the quasar is still active cold phase. This agrees with the conversion factor of and accreting close to its Eddington rate, the shocked ∼10%usedinthepaperscitedinTable1toestimatethe windretains1/3ofthetotalincidentwindkineticenergy totalmass outflowrate fromobservedmolecularspecies. M˙ v2/2,giving2/3totheswept–upgas. Theenergyre- Our model therefore predicts both the total mass out- w w tainedinthewindandtheswept–upgashavepotentially flowrates(eqn15)andtheobservationalsignaturesused observable emission signatures (see Discussion). to estimate them, in good agreement with observation Equation(16)meansthattheswept–upgasmusthave (Table 2). a momentum rate greater than the Eddington value The inner wind shock presumably accelerates cosmic L /c, since we can rewrite it as ray particles, and gamma rays result when these hit the Edd colderISMandshockedwind. Theoutflowsaretherefore P˙2 P˙2 directly comparable with the gamma–ray emitting bub- w ∼− out , (17) blesinourGalaxyrecentlydiscoveredbyFermi(Su et al. 2M˙ 2M˙ w out 2010). One can interpret these as relics of the Milky where the P˙’s are the momentum fluxes. With P˙ = Way’s last quasar outburst about 6 Myr ago by noting w thatthegreaterdensityoftheGalacticplanemustpinch L /c, we have Edd a quasi–spherical quasar outflow into a bipolar shape (Su et al. 2010; Zubovas et al. 2011). The gamma–ray 1/2 P˙ =P˙ M˙out = LEddf1/2 ∼20σ−2/3l1/6LEdd emission from distant galaxies discussed here should be out w M˙w ! c L 200 c intrinsicallystrongerthanintheMilkyWay,butthelong integration time required to detect the Galactic bubbles (18) meansthattheseoutflowsmaybeundetectablewithcur- where f is the mass loading factor of the outflow. The L rent instruments. factor fL1/2 ∼ 20 is the reason why observations consis- Perhaps more promisingly, these cosmic ray electrons tently show M˙ v >L /c. cool and emit synchrotron radiation in the radio band. out out Edd This radiation may be observable and so it would be 6. DISCUSSION interesting to check whether there are kpc or sub–kpc scale radio bubbles associated with the outflows. We have shown that large–scale outflows driven by wide–angle AGN winds should have typical velocities vout ∼ 1000−1500 kms−1 and mass flow rates up to WethankSylvainVeilleuxandDavidRupkeforhelpful M˙out ∼ 4000 M⊙yr−1 (eqns 8, 15) if the central quasar discussions, and the referee for a very thoughtful and is still active, with lower values if it has become fainter. helpful report. Research in theoretical astrophysics at Our equations (8, 15) directly relate the outflow ve- Leicesterissupportedby anSTFC RollingGrant. KZis locities and mass rates to the properties of the host supported by an STFC research studentship. REFERENCES Cano-Diaz,M.,Maiolino,R.,Marconi,A.,Netzer,H.,Shemmer, —.2005, ApJ,635,L121 O.,&Cresci,G.2011,ArXive-prints King,A.R.2010,MNRAS,402,1516 Ciotti,L.,&Ostriker,J.P.1997,ApJ,487,L105+ King,A.R.,&Pounds,K.A.2003, MNRAS,345,657 Cowie,L.L.,&McKee,C.F.1977,ApJ,211,135 King,A.R.,Zubovas, K.,&Power,C.2011,MNRAS,L263+ Feruglio,C.,Maiolino,R.,Piconcelli,E.,&etal.2010, A&A, Lonsdale,C.J.,Lonsdale,C.J.,Smith,H.E.,&Diamond,P.J. 518,L155+ 2003,ApJ,592,804 H¨aring,N.,&Rix,H.-W.2004,ApJ,604,L89 MacLow,M.-M.,&McCray,R.1988,ApJ,324,776 Kane,J.,Drake,R.P.,&Remington, B.A.1999,ApJ,511,335 McConnell,N.J.,Ma,C.-P.,Gebhardt, K.&etal.2011,Nature, King,A.2003,ApJ,596,L27 480,215 6 McKee,C.F.,&Ostriker,J.P.1977,ApJ,218,148 Silk,J.,&Nusser,A.2010, ApJ,725,556 Murray,N.,Quataert,E.,&Thompson,T.A.2005,ApJ,618, Sturm,E.,Gonz´alez-Alfonso,E.,Veilleux,S.,&etal.2011,ApJ, 569 733,L16+ Pounds,K.A.,King,A.R.,Page,K.L.,&O’Brien,P.T.2003a, Su,M.,Slatyer,T.R.,&Finkbeiner,D.P.2010,ApJ,724,1044 MNRAS,346,1025 Tacconi,L.J.,Genzel,R.,Lutz,D.,Rigopoulou,D.,Baker,A.J., Pounds,K.A.,Reeves, J.N.,King,A.R.,&etal.2003b, Iserlohe,C.,&Tecza, M.2002, ApJ,580,73 MNRAS,345,705 Tombesi,F.,Cappi,M.,Reeves, J.N.,&etal.2010a, A&A,521, Power,C.,Zubovas,K.,Nayakshin,S.,&King,A.R.2011, A57+ MNRAS,413,L110 Tombesi,F.,Sambruna,R.M.,Reeves,J.N.,&etal.2010b, Riffel,R.A.,&Storchi-Bergmann,T.2011a,MNRAS,411,469 ApJ,719,700 —.2011b,MNRAS,1320 Veilleux,S.,Rupke,D.S.N.,Kim,D.-C.,&etal.2009, ApJS, Rupke,D.S.N.,&Veilleux,S.2011,ApJ,729,L27+ 182,628 Shemmer,O.,Netzer,H.,Maiolino,R.,Oliva,E.,Croom,S., Zubovas,K.,King,A.R.,&Nayakshin,S.2011,MNRAS,415, Corbett, E.,&diFabrizio,L.2004, ApJ,614,547 L21