How do Black Holes Get Their Gas?

Transcription

1 How do Black Holes Get Their Gas? Philip Hopkins Eliot Quataert, Lars Hernquist, T. J. Cox, Kevin Bundy, Jackson DeBuhr, Volker Springel, Dusan Keres, Gordon Richards, Josh Younger, Desika Narayanan, Paul Martini, Adam Lidz, Tiziana Di Matteo, Yuexing Li, Alison Coil, Adam Myers, Patrik Jonsson, Chris Hayward

2 ~5 kpc galaxy-galaxy mergers disk instabilities Focus: Most luminous QSOs (~1-10 M sun /yr) Bottleneck at <10-50pc: BH begins to dominate the potential (e.g. Goodman et al., Jogee et al., Martini et al.) 500 pc <10 pc bars within bars BH/nuclei merging gravitational instability? (NO...?)? clumps? (NO) viscosity? (NO) MHD wind? (NO) <0.1 pc Viscous disk/mri

3 Galaxy merger: good way to get lots of gas to small scales! If BHs trace spheroids, then *most* mass added in violent events that also build bulges

4 Problem: Scale of merger: ~100 kpc Galaxy merger: good way to Viscous disk: ~0.1 pc get lots of gas to small scales! Solution 1: simple prescription If Solution BHs trace 2: re-simulate spheroids, then *most* ( zoom mass in ) added and see in what violent events happens! that also build bulges

5 Simulations: FOLLOWING THE GAS IN... Here: Focus on robust conclusions

6 Simulations: FOLLOWING THE GAS IN... Need to include: Here: Focus on robust conclusions

7 Simulations: FOLLOWING THE GAS IN... Need to include: Gas+Stars Here: Focus on robust conclusions

8 Simulations: FOLLOWING THE GAS IN... Need to include: Gas+Stars Self-gravity! Here: Focus on robust conclusions

9 Simulations: FOLLOWING THE GAS IN... Need to include: Gas+Stars Self-gravity! Cooling Here: Focus on robust conclusions

10 Simulations: FOLLOWING THE GAS IN... Need to include: Gas+Stars Self-gravity! Cooling Star formation Here: Focus on robust conclusions

11 Simulations: FOLLOWING THE GAS IN... Need to include: Hicks et al. Gas+Stars Self-gravity! Krumholz & Tan Cooling Star formation Here: Focus on robust conclusions

12 Simulations: FOLLOWING THE GAS IN... Need to include: Hicks et al. Gas+Stars Self-gravity! Krumholz & Tan Cooling Star formation Feedback - Admit we don t understand it! Here: Focus on robust conclusions

13 Simulations: FOLLOWING THE GAS IN... Need to include: Hicks et al. Gas+Stars Self-gravity! Krumholz & Tan starbursts (Downes+Solomon, Scoville, et al.) masers (Greenhill, Kondratko) Cooling Star formation Feedback - Admit we don t understand it! Here: Focus on robust conclusions

14

15 Tidal torques large, rapid gas inflows (e.g. Barnes & Hernquist 1991)

16

17 Triggers Starbursts (e.g. Mihos & Hernquist 1996)

18

19 Fuels Rapid BH Growth? (e.g. Di Matteo et al., PFH et al. 2005)

20

21 Large-scale simulation: follow gas to sub-kpc scales

22

23 Now: Re-simulate central kpc at high-res Follow gas to ~10 pc

24

25 Continue, re-simulate central regions, down to 0.1pc resolution

26

27 How do massive BHs get their gas? CAN WE FUEL THE MONSTER? Cascade of instabilities: merger not efficient inside ~kpc Any mechanism that gets to similar densities at these scales will do the same Instabilities change form at BH radius of influence

28 Sub-kpc scales: Stuff within Stuff Diverse morphologies on sub-kpc scales: not just bars! Inflow is not smooth/continuous

29 Sub-kpc scales: Stuff within Stuff Diverse morphologies on sub-kpc scales: not just bars! Inflow is not smooth/continuous

30 Gas-rich merger (lots of inflow) Weakly bar-unstable disk (less inflow) Key parameter: Gas driven in, vs. pre-existing bulge/bh mass

31 Gas-rich merger (lots of inflow) Weakly bar-unstable disk (less inflow) Key parameter: Gas driven in, vs. pre-existing bulge/bh mass

32 Gravity dominates torques from ,000 pc: Stars torquing on gas stars (color) gas (contours)

33 How does this work? Build analytic models: Structure Growth rates Stability Inflow rates gas (contours) stars (color)

34 How does this work? Build analytic models: Structure Growth rates Stability Inflow rates gas (contours) stars (color) standard (dissipationless) formulation: spiral waves carry the angular momentum: (Lynden-Bell & Kalnajs 72) Ṁ inflow = [k, a ]/ R 2 a 2 kr 2 M disk M tot M gas t dyn ( kr 1)

35 How does this work? Build analytic models: Structure Growth rates Stability Inflow rates gas (contours) stars (color) standard (dissipationless) formulation: spiral waves carry the angular momentum: (Lynden-Bell & Kalnajs 72) Ṁ inflow = [k, a ]/ R 2 a 2 kr 2 M disk M tot M gas t dyn ( kr 1) with shocks & dissipation: >100x larger!!! a M gas t dyn

36 kpc 10 pc

37 kpc Actual inflow rate 10 pc

38 kpc Actual inflow rate Prediction (gravitational torques with shocks) 10 pc

39 kpc Actual inflow rate Prediction (gravitational torques with shocks) 10 pc No dissipation (Lynden-Bell+ 71)

40 Can we build a better accretion rate estimator?

41 Can we build a better accretion rate estimator? Derive Gravitational Torque Rate: Ṁ 10 M yr 1 Disk Total 5/2 1/6 M BH, 8 M gas, 9 R 3/2 0,100

42 Inflow from ~kpc to ~0.1 pc is NOT viscous or Bondi-Hoyle: Actual inflow rate onto BH [Msun/yr] from 1 kpc from 100 pc c 2 s gas 1 Viscous: Bondi: G 2 MBH 2 c 3 s

43 Actual inflow rate onto BH [Msun/yr] from 10 pc from 100 pc from 50 pc from 1 kpc Gravitational Prediction

44 True Inflow Rate [Msun/yr] Actual inflow rate onto BH [Msun/yr] from 10 pc from 100 pc from 50 pc from 1 kpc Gravitational Prediction Predicted (New Gravitational Scaling)

45 So, what about the small scales near the BH?

46 ~10 pc scales: Nuclear eccentric disks Inside BH radius of influence: develop thick, precessing disks Need both star formation and self-gravity

47 Eccentric/lopsided disks (m=1 modes) are special in a near-keplerian potential Keplerian potentials are special: = Hence, closed elliptical orbits! epicycle

48 Eccentric/lopsided disks (m=1 modes) are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: Keplerian potentials are special: = cos m Hence, closed elliptical orbits! epicycle

49 Eccentric/lopsided disks (m=1 modes) are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: Keplerian potentials are special: = Generically, force some deviations/torques/etc: Hence, closed elliptical orbits! v cos m M disk (<r) epicycle V c M BH

50 Eccentric/lopsided disks (m=1 modes) are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: Keplerian potentials are special: = Generically, force some deviations/torques/etc: Hence, closed elliptical orbits! v cos m M disk (<r) epicycle V c M BH But, if (and only if) m=1: v V c

51 Relic, ~pc-scale nuclear stellar disk... Gas-stellar exchange dramatically enhances torques Drives ~10 M sun /yr inflow Leave relic stellar disks?

52 These are observed! M31, NGC4486B, many candidates (NGC 404,507,1374,3706,4073,4291,4382,5055,5576,7619, VCC128, M32,83) M31: Lauer et al Kormendy & Bender 1999

53 These are observed! M31, NGC4486B, many candidates (NGC 404,507,1374,3706,4073,4291,4382,5055,5576,7619, VCC128, M32,83) M31: Lauer et al Kormendy & Bender 1999 M31 disk has ~0.1-1 MBH in old stellar mass Outer radius R~1-10 pc Moderate thickness, high eccentricity

54 These are observed! M31, NGC4486B, many candidates (NGC 404,507,1374,3706,4073,4291,4382,5055,5576,7619, VCC128, M32,83) M31 run backwards : the M31 disk implies accretion at ~0.5-3 M sun /yr (~L Edd ) for ~100 Myr (~ M BH )!

55 What about the obscuration from these disks? Lots of gas in this disk during the inflow stages...

56 What about the obscuration from these disks? Lots of gas in this disk during the inflow stages...

57 What about the obscuration from these disks? Lots of gas in this disk during the inflow stages... The eccentric disk IS the torus!

58 What about the obscuration from these disks? cs~5 km/s cs~10 km/s cs~15 km/s cs~20 km/s cs~30 km/s cs~50 km/s The eccentric disk IS the torus Occurs even if allow cooling and no stellar feedback! Heating by bending/warping modes, themselves excited by the eccentric pattern

59 Summary Ø Ø Ø Ø Fueling Most Luminous BHs: Global gravitational instabilities CAN power ~10 Msun/yr! Really! New Mdot estimator: neither viscous nor Bondi Stuff within Stuff : Cascade of instabilities with diverse morphology Doesn t matter how first get down from large scales Accretion rates & orientations are stochastic Vary on all timescales Angular momentum changes rapidly - no correlation with host disk The torus is the disk: a dynamical accretion driver Bending/warping instabilities: thick even without stellar feedback Ø Stellar nuclear disk relics : M31 & 4486b: Can we directly observe the fossil of the accretion driver & torus?

60 Obscuration and the torus Observed surface densities and kinematics arise naturally Masers (Kondratko, Greenhill, et al.) AO (Hicks, Davies, et al.) 5 pc

61 Compare column density distributions: edge-on gas has NO substructure quasi-virial clumps observed (Risaliti, Treister, Malizia)

62 Gravity dominates torques from ,000 pc Torques at R

63 m=1 slow modes are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: cos m number of arms

64 m=1 slow modes are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: cos m number of arms 1 Response: e = 2 m 2

65 m=1 slow modes are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: cos m number of arms 1 Response: e = 2 m 2 Near a BH: 1 1 (1 m) 2

66 m=1 slow modes are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: cos m number of arms 1 Response: e = 2 m 2 m =1: Near a BH: 1 1 (1 m) 2 e 2 r 3 : 1 0 M disk (<r) M BH

67 m=1 slow modes are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: cos m number of arms 1 Response: e = 2 m 2 Near a BH: 1 1 (1 m) 2

68 m=1 slow modes are special in a near-keplerian potential Disturb the stars with some perturbation in the disk: cos m number of arms 1 Response: e = 2 m 2 Near a BH: 1 1 (1 m) 2 e m =1: 0 (resonance) Strong torques can propagate to all r (even << 0.1pc) INDEPENDENT of Mdisk(<r)/MBH

69

70 Large-Scale Tides are Not Important for AGN:

71 The Effective Stellar Feedback on Small Scales: (REQUIRE SOME SUB-RESOLUTION MODEL)

72 A No Feedback ISM is Ruled Out on Small Scales:

73 But qualitative conclusions are insensitive to the gas microphysics

74 Vertical Torus Structure

75 Vertical Torus Structure

76 Mis-alignments with the parent disk are common Implications for: BH spin BH-BH mergers Recoils Variability

77 Torus-Host disk misalignments: