AGN Feedback in Massive Ellipticals

1. Jerusalem 16 Dec 2009; JPO AGN Feedback in Massive Ellipticals Co-conspirators*: L. Ciotti M.-G. Park M.-S. Shin F. Yuan D. Proga S.Y.Sazonov R.A.Sunyaev G. Novak

2. Cartoon of Co-Evolution of Elliptical and MBH/AGN • 1) Black holes postulated in the 1960s to explain rare events in distant galaxies - huge EM output from small sources. • 2) Now we know that ALL massive galaxies, near and far, contain MBH’s and that most are in the “off” state, ie small duty cycle. • 3) But modern treatments almost all ignore the radiative output! • 4) Biggest source of mass to feed MBH is recyled gas from stellar evolution - typically ignored.

3. Prologue - Historical • 1) Black holes postulated in the 1960s to explain rare events in distant galaxies - huge EM output from small sources. But modern treatments almost all ignore this radiative output! • 2) Now we know that ALL massive galaxies, near and far, contain MBH’s and that most are in the “off” state, ie small duty cycle. • 3) Biggest source of mass to feed MBH is recyled gas from stellar evolution - typically ignored.

4. A physics problem – what happens when a BH in a galaxy accretes matter? Plan of presentation • What are the physically dominant processes in AGNs and Starbursts ? • Why radiative preheating may be important. • Spherical accretion onto BHs: theoretical aspects and accurate calculations. • Simulations with recycled gas and radiative feedback. • Starburst phenomena added. • Preliminary 2-D simulations • Mechanical energy input (AGN wind and jet) added. • Summary, future calculations and tests.

5. AGNs and Starbursts: physical processesconsequent to normal galactic evolution • Co-incident and co-terminous: gas added to the center of galaxies feeds central black hole AND also fuels starburst. Processes comparable in importance. • Gas source from recycled gas is 25% of stellar mass; galactic merger induced gaseous in-fall may be comparable but less, especially at late times (z < 1.5). • Energy input due to • Radiative input: UV from stars, and UV-Xray from AGN • Mechanical: winds&SN from stars, and winds&jets from AGN • Efficiency of input varies based on • Amount of energy input • Timescale of energy input • Location of energy input

6. Mergers of relatively low importance: Title: The clustering of narrow-line AGN in the local Universe Authors: Cheng Li, Guinevere Kauffmann, Lan Wang, Simon D.M. White, Timothy M. Heckman, Y.P. Jing Comments: 14 pages, 11 figures, submitted to MNRAS \\ We have analyzed the clustering of ~ 90,000 narrow-line AGN drawn from the Data Release 4 (DR4) of the SDSS. We compute the cross-correlation between AGN and a reference sample of galaxies, and compare this to results for control samples of inactive galaxies matched simultaneously in redshift,stellar mass,concentration, velocity dispersion and the 4000A break strength. We also compare near-neighbour counts around AGN and around the control galaxies. On scales larger than a few Mpc, AGN have almost the same clustering amplitude as the control sample. This demonstrates that AGN host galaxies and inactive galaxies populate dark matter halos of similar mass.On scales between 100kpc and 1Mpc,AGN are clustered more weakly than the control galaxies. We use mock catalogues constructed from high-resolution N-body simulations to interpret this anti-bias, showing that the observed effect is easily understood if AGN are preferentially located at the centres of their dark matter halos. On scales less than 70 kpc, AGN cluster marginally more strongly than the control sample, but the effect is weak. When compared to the control sample, we find that only one in a hundred AGN has an extra neighbour within a radius of 70 kpc. This excess increases as a function of the accretion rate onto the black hole, but it does not rise above the few percent level. Although interactions between galaxies may be responsible for triggering nuclear activity in a minority of nearby AGN, some other mechanism is required to explain the activity seen in the majority of the objects in our sample.

7. Stellar Sources Energy Input • Winds and SN II each give roughly 1051 erg per star having M > 8 Msolar (assume Chabrier IMF): • mech, = 5 x 10-6 (E = mechM* c2 ) • UV, = 6 x 10-5 (E = UVM* c2 ) • mech,UV= 1 x 107 yrs • Rstarburst = 100 pc • SN I roughly give: • mech,SNI= 2 x 10-6 (E = mechM* c2 ) • mech,SNI= 2 x 109 yrs • RSNI=3000 pc

8. AGN Input • Multiply efficiency by rBH = MBH/M* = 0.0013 • EM,AGN = 0.12 x 0.0013 = 1.6 x 10-4 • UV,AGN = EM x 0.2 = 3.2 x 10-5 • mech,AGN= 0.004 x 0.0013 = 5.2 x 10-6 * • mech,UV= Edd / fDC = 5x107 yrs /.025 = 2.0 x 10+9 yrs • RAGN= = 0.1 - 1.0 pc • --------------------------------------------------- • *Maximal estimate, better estimate presented subsequently

9. Comparison of Stellar and AGN Feedback Processes • Ultraviolet Energy Input:UV,* >~ UV,AGN • Mechanical Energy Input:mech,AGN ~ mech,* • AGN Radiative Energy input dominates (?). • Time-scale:mech,SNI > mech,AGN > mech,* • Spatial Scale: RSNI > Rstarburst > RAGN AGN and Starburst comparably important with respect to mechanical energy input, but AGN radiative input dominant if it can couple. SNI and SNII best wrt driving super-winds.

10. Sequence of problems to be solved: • 1) Idealized spherical accretion onto an isolated MBH allowing for radiative feedback effects. • 2) Embed MBH in a typical elliptical galaxy to compute winds, duty cycle and limiting growth of MBH. • 3) Allow for star-formation in regions of dense cold gas to estimate properties and recurrence of starbursts. • 4) Add dust, radiation pressure on the dust, a central disk, with star-formation and wind from AGN + disk. • 5) Add mechanical energy input from AGN wind & jet. • 6) Add interaction with ambient (eg cluster) medium.

11. 1) Physics included and excludedin typical BH accretion calc’s • Included: • Dynamics: gravity and hydrodynamics. • Atomic physics: cooling, electron-ion coupling etc. • Radiative emission processes. • Sometimes included: • Accretion to the region of the central engine. • Visous effects, MHD. • Not included (usually): • Shocks, shock heating & cosmic ray generation. • Radiative heating due to luminosity of interior material. This omission is strange, given the enormous observed hard X-ray luminosities of accreting black holes !

12. 2) The Physics (cartoon version) of Spherical Infall to an Isolated MBH • For high flow rates (Mdot ->Mdot,Edd), gas cools at large radius: Tgas-> 104.3 K << Tgrav = (GMbhmp/kr). • Then, independent of the radiative processes, rad << Egrav < mc2. • But, shocks (or radiative preheating) can raise T -> Tgrav allowing higher rad . • Or, radiative preheating can raise the temperature to T > Tgrav , reversing the flow and causing relaxation oscillations. Allowance for radiative preheating can qualitatively change the solution.

13. Quasi-spherical Accretion: theory • Independent of Mbh in Eddington units. • Thompson -> (mdot = Mdot /Mdot,Edd) as r -> Rsch. . • Self-consistent solution determines output luminosity and spectrum given mdot . • Non-unique: hot and cold solutions exist at given mdot . • Preheating and unstable flows at higher luminosity. • Feedback from luminous output => relaxation oscillations if l = L/LEdd ≥ 0.01. • Overall duty cycle expected to be small. • “Efficiency” is determined self-consistently and is typically quite small: ADAF (since 1973)

14. High efficiency solutions are all unstable! (time-scale for fluctuation ~ GM/Cx3 ~ 1 yr ) Log e

15. Conclusion: most energy emitted in UV, but weighted by photon energy it is in the X-Ray region Observational determination of the mean AGN emitted spectrum from individual objects and the X-Ray-background.

16. Compton Temp and Radiative Heating = 2 107 K Independent of Optical/UV absorption and of direction (for isotropic initial emission). Elementary Thermodynamics: kTgas -> <h> ~ kTC

17. 3) Accretion of MBH within elliptical galaxy (minimal complications) • Detailed spherically, symmetric, time-dependent hydro of accreting MBH in E galaxy with • Assumed accretion efficiency (0.001 -> 0.100). • Assumed Spectrum with Tcompt = 107.5->109.0. • No star formation. • ------------------------------------------------------------------------------------- • Relaxation oscillations: cooling, infall, photo-heating, expanding hot bubble, cooling…. • Most time in non-accreting state (fduty ~ 0.006), galaxy looks like normal elliptical. • In burst mode looks like a quasar. • Final MBH masses reasonable. • Elliptical gas luminosity reasonable.

18. QSO luminosity in grouped bursts.

19. 10 40 ≤LX,gas ≤ 1042 “Cooling flows”

20. 4) Repeat with allowance for starburst; feedback from both AGN and stars. • Accurate stellar dynamical model (separate Hernquist profiles for DM and stars). Orbits evolved as mass profile changes.M87 type gal. • Accurate stellar evolution (Renzini) for mass loss in PN, red-giant winds, SN etc. Gives mass, energy and metals input vs time. • Star-formation via conventional formalism from cold collapsing regions. Gives additional (M,E,Z). • AGN radiative feedback with detailed atomic phys.

21. Energy Input Rates from MBH, old stars and gas Red: MBH Green: Old stars Black: Gas emission Note: “quiet” level now computed by high resolution code. Not “sub-Bondi”.

22. There is detailed fine-structure on a 107 yr timescale t = 2 x 10^7 yrs; see Forman (2006)

23. Mass flow rates from recycled gas, to MBH and wind Recycled gas from planetary nebulae. Galactic wind into ambient medium driven by AGN bursts with peak value at 100 Msolar/yr. See Fabian(2006)

24. Luminosity Distribution as Observed “Cooling flow” model

25. Overall Results • Lx, gas in right range. • MBH appropriate (not too large). • Modest mass outflow. • Duty cycle ~ 0.006. • ------------------------------------------------- • Appears most times as a normal elliptical, some time as an incipient cooling flow and during very brief intervals as a quasar.

26. 5) Add nuclear disk, dust and radiation pressure on the dust • Innermost shell dumps gas onto gaseous disk, within which stars form a la Schmidt law with bulk going to wind and central BH in observed ratios. • Radiation pressure from newly formed stars in central regions (disk + nuclear starburst) allowed to act on inflowing dusty cold gas dominates over gas pressure and Eddington effects during star-bursts.

27. Nuclear region A stars outlast the bursts -> “E + A” Spectrum Optical starburst red: BH (solid= Eddington; dotted=accretion lum., dashed = absorbed) green: STARS (solid = LsnIa; dotted = LsnII; dashed = thermalization of stell. mass losses) green: STARS (dot-dashed=optical starburst; long-dashed = UV starburst) black: ISM (solid = X-ray, dotted = bolometric)

28. Very low final gas fraction At late times the gas mass in the galaxy is less than the BH mass TOP: solid = total BH accreted mass; dotted = ISM mass in the galaxy BOTTOM: same as above, but for rates; dashed: mass loss rate from the galaxy as a wind

29. Nuclear stellar disks outlast the starbursts somewhat. Domination by high mass stars leads to fading in ~ 107.5 years (“E+A”) RED: BH mass BLACK: solid (spikes) disk gas mass; mass lost as a wind: dotted GREEN: solid & dotted: low and high mass stars ; horizontal dashed: remnants

30. Detailed Hydro Sequence(nb very high spatial and temporal resolution needed, very large dynamic range required in the code for radius, temperature and density) • Stellar evolution leads to gas density increase and inefficient SNI driven wind. • Cooling, collapsing shell forms leading to nuclear starburst and oscillatory AGN phase. • Gas depletion in starburst coupled with energy input first from SNII and then from AGN produces a hot expanding bubble driving out a cool shell, new wind phase. • Bubble cools, recycled gas accumulates and the process repeats.

31. Pre cooling collapse: Normal X-ray emission from gas and low level hot accretion onto the central BH

32. Cooling shell forms: Essentially the Field instability. Would be filamentary in 3-D calculation.

33. Post starburst: Out-flowing wind starts at thousands of km/sec at radius of 500 pc. Fabian: X-ray “shells” made Strong shock weakens to strong sound wave as it propagates into ambient gas.

34. Stellar density distributions: starburst component in cusp New stars: Rate Cumulative Old stars:

35. Post starburst, expanding cold shell Detailed X-ray profiles as a function of energy will provide the strongest test of the model. Cold collapsing shell Before starburst

36. 6) Add AGN Wind, Jet, CR and Look at Disc Geometry • Jet is observed to be relativistic with low mass loading - origin near BH. • Wind from BLR has modest velocity but mass flux comparable to Mdot,acc - origin from circum-blackhole disk. • Both components specified by output of mass energy and momentum.

37. Small changes in burst structure but reduction in BH growth and strong shocks produced.

38. Brief Non-Thermal Phase after an Outburst and CR Generation: A giant SNR! Y.-F. Jiang, L. Ciotti & JPO (2009)

39. Effects of Mechanical Energy Input • Winds from disk blow central bubbles, and added feedback reduces somewhat the final BH mass. Shocks and CR generation. • Jet drills through ISM of isolated galaxy but is significant for BCG in cluster gas. • Little qualitative change but significant quantitative change (lower AGN luminosity, lower BH mass, lower X-Ray luminosity from ISM gas etc). Both radiative and mechanical feedback are needed

40. Many present treatments of mechanical feedback are not consistent with mass and energy conservation ! • (dM/dt)Acc = (dM/dt)Inf - (dM/dt)Out • dE/dt = mech * (dM/dt)Acc *c2 • = ( ½)* (dM/dt)Out VOut2 • => • (dM/dt)Acc = (dM/dt)Inf /(1+ b), and • dE/dt = mech * (dM/dt)Inf *c2/(1+ b) • where b = (2 mech c2 /VOut)) >> 1 • Many extant calculations, based on computing (dM/dt)Acc and (dM/dt)Acc to agree w observations do not include the 1/(1+ b) factor

41. 7) First Efforts of 2-D Solution (Proga &O) • Wind produced with Mdot,w >~ Mdot,acc • Small solid angle until Edd Luminosity approached. • Energy efficiency ~ 1 x 10-4 • Cold cloud ejection at high Eddington rates

42. Computational domain bounded by inner and outer BC

43. No Rotation Rotation Vel & Density Temperature

44. No X-ray Heating X-ray Heating Higher density at outer boundary (x 10) And Possibility of background X-ray heating

46. Zoom Out – Effects on Ambient Medium Two-D hydrodynamic simulation (Novak & JPO; 2009) cf Fabian (2009)

47. Primary Predictions • Low duty cycle outbursts expected with coupled AGN and starburst effects in normal ellipticals from recycled metal rich gas. • Longer timescale (~ cooling time) low density high temperature central bubbles should be observable. • Radioactivity likely from particle acceleration in the outgoing shocks -> synchrotron emission, eg like SNR. • Cooling shell phases are unstable to R-T and will appear as filamentary cool gas. • Stellar population of central regions of elliptical galaxies should be red (metal rich) and relatively young. “ E +A star” spectra will survive the nuclear starburst as evidence.

48. Summary • Feedback must be important. Both mechanical and radiative feedback regulate the growth of massive black holes. Magorian relation. Duty cycle <~ 1%. • Re-cycled gas from late stellar mass loss (25%) must produce outbursts. Star-formation, SNII and AGN feedback use up and drive gas out of the nucleus. • Central wind pulses are driven into ambient medium subsequent to nuclear starburst and AGN episode heat the ambient gas. CRs made. Cycle recurs after cooling flow. • Remnant nuclear stellar discs survive the outbursts. • Much of the most dramatic episodic activity is shrouded in dust and will only be seen as radio or far-IR or hard X-ray activity. • Winds and weak shocks propagate into ambient medium producing X-Ray shell structures.

49. Caveats • Work so far is exploratory. Problems are far from solved at present. Every time an important new physical effect is included the results change in a significant way: • Check for additional physical effects needed, including cosmological infall, merger etc • More accurate treatment of current physics needed including self-consistent, 2-D, time dependent calculations, radiative transfer etc • Better comparison to observations needed to check validity, enable predictions etc