1. Cooling, AGN Feedback and Star Formation in Simulated Cool-Core Galaxy Clusters Yuan Li University of Michigan Collaborators: Greg L. Bryan (Columbia) Mateusz Ruszkowski (University of Michigan) G. Mark Voit, Brian W. O’Shea, Megan Donahue (MSU) Grant Tremblay (Yale) and more Shanghai, China June 2015

2. Outline The cooling flow problem Cooling—How does gas cool? Can AGN feedback balance cooling? The role of Star Formation Yuan Li, University of Michigan2

3. The Cooling Flow Problem Sanderson et al. (2006) Yuan Li, University of Michigan3

4. The Cooling Flow Problem Sanderson et al. (2006) Temperature: 1/2-1/3 T virial Cooling time: t cool ~ (5nkT / 2) / n 2 Λ << Hubble Time Steady state: cooling flow of 100s- 1000 M  /yr Artist's conception of a classical cooling flow Yuan Li, University of Michigan4

5. The Cooling Flow Problem Observed SFR ~ 10 M  /yr << 100s-1000 M  /yr Lack of cool gas below 1-2 keV in the X-ray Fabian et al. (2006) No Classical Cooling Flow! Artist's conception of a classical cooling flow Sanderson et al. (2006) Yuan Li, University of Michigan

6. Composite X-ray image of Perseus Core Why AGN Feedback? Yuan Li, University of Michigan 100 kpc 6 Fabian et al. (2006)

7. Why AGN Feedback? Fabian 2012 Yuan Li, University of Michigan Cavagnolo et al. 2008 7

8. The Origin of the Cold Filaments Conselice et al. (2001) Hα image of the Perseus cluster core. Yuan Li, University of Michigan Artist's conception of a classic cooling flow Composite X-ray image of Perseus Core Fabian et al. (2006) 8

9. Yuan Li, University of Michigan9 McDonald et al. 2012 Hα FILAMENTS

10. Hα Luminosity vs. ICM Entropy Yuan Li, University of Michigan10 Voit & Donahue 2014

11. Hα Luminosity vs. min(t cool /t ff ) Yuan Li, University of Michigan11 Voit & Donahue 2014

12. Key Questions Cooling--What is the origin of the cold filaments? Can AGN feedback balance cooling? The Role of Star Formation Yuan Li, University of Michigan12

13. Pure Cooling Flow Simulation Enzo (an AMR code) Box size L = 16 Mpc/h N root = 256, l max = 15 10 6  smallest cell ~ 2pc. 3D, spherically symmetric Gravity: NFW Dark Matter + BCG + SMBH + gas Initial gas density and temperature: observations of the Perseus Cluster Initial pressure: Ideal Gas Law Yuan Li, University of Michigan13

14. 16.6 kpc Pure Cooling Flow t=300 Myr Yuan Li & Greg Bryan (2012) ApJ 747 26 Density PressureL X-ray Temperature Yuan Li, University of Michigan14

15. Jet Modeling Add mass, momentum and energy to cells within the xy plane of the jet launching region (a cylinder with r = 100 pc and h = 300 pc) following Omma et al. (2004) Outflow rate (ṁ) = cold accretion rate (M cold /t accretion ) Jet power Ė= εṁc 2 Kinetic fraction: f = 0.5 V jet ≈ 10000 km/s (ε=0. 1%) Precessing jets: P = 10 Myr l max = 12  smallest cell size ~ 15pc Yuan Li, University of Michigan15 Li & Bryan (2014) ApJ 798 54 & ApJ 798 153 SMBH

16. 80 kpc Yuan Li, University of Michigan16

17. Hα Luminosity vs. min(t cool /t ff ) Yuan Li, University of Michigan17 Voit & Donahue 2014

18. Idealized Simulations with Volumetric Heating Sharma et al. 2011 (Spherical geometry): t cool /t ff =10 McCourt et al. 2012 (Cartesian geometry): t cool /t ff =1 Meece et al. 2015 (Cartesian geometry): t cool /t ff =10 Yuan Li, University of Michigan18 Simulations with AGN Jets Gaspari et al. 2012: t cool /t ff ~ a few to 20 Deovrat Prasad et al. 2015 (poster) : t cool /t ff ~10 Thermal Instability Criterion

19. Gas is Linearly Stable! What Makes it Cool? Yuan Li, University of Michigan19 min( t cool /t ff )~3

20. Yuan Li, University of Michigan20

21. What is Special about the Gas that Cools into Clumps? Yuan Li, University of Michigan21

22. Gas is Linearly Stable! Jets Make it Cool! Yuan Li, University of Michigan22

23. Yuan Li, University of Michigan t dyn t cool of the cooling particles 23

24. The Origin of the Filaments AGN jets (non-linear perturbation) Low entropy gas with short cooling time (but still linearly stable!) t cool /t ff ~ 10 (a few to 20) Yuan Li, University of Michigan24

25. Key Questions Cooling--What is the origin of the cold filaments? Can AGN feedback balance cooling? The Role of Star Formation Yuan Li, University of Michigan25

26. 80 kpc Yuan Li, University of Michigan26

27. McDonald et al. 2010 Filamentary Gas Turbulence: low Jet Power: low Yuan Li, University of Michigan27

28. McDonald et al. 2010 Extended Cold Gas Turbulence: high Jet Power: high Yuan Li, University of Michigan28

29. Rotating Disk Turbulence: high Jet Power: Steady Yuan Li, University of Michigan Disk Keplerian Velocity Rotational Velocity 29

30. McDonald et al. 2010 Yuan Li, University of Michigan Rotating Disk Turbulence: high Jet Power: Steady 30

31. Can AGN Feedback Balance Cooling? Yes! Yuan Li, University of Michigan31

32. Success of the Model Balancing Cooling Consistent results when varying model parameters: kinetic fraction f (0.1-1) feedback efficiency ε (0.1%-1%) Converging results when smallest cell size < 500 pc Yuan Li, University of Michigan32

33. But… Disk is too large (~ 10 11 M  ) and stays forever Consistent with simulations (Gaspari et al. 2012), but not the observations. Yuan Li, University of Michigan33

34. Missing Physics Self-gravity? – help accrete the disk Viscosity? – help accrete the disk Conduction? – dissolve the cold disk Magnetic Field? – complicated Star Formation? – consumes and locally heats up the cold gas Yuan Li, University of Michigan34

35. Missing Physics Self-gravity? – help accrete the disk Viscosity? – help accrete the disk Conduction? – dissolve the cold disk Magnetic Field? – complicated Star Formation? – consumes and locally heats up the cold gas Yuan Li, University of Michigan35

36. Missing Physics Self-gravity? – help accrete the disk Viscosity? – help accrete the disk Conduction? – dissolve the cold disk Magnetic Field? – complicated Star Formation? – consumes and locally heats up the cold gas Yuan Li, University of Michigan36

37. Filamentary Star Formation Yuan Li, University of Michigan37 Donahue et al., 2015 50 kpc by 50 kpc UV images of BCGs in CLASH

38. Star Formation Criteria the gas density exceeds a certain value the gas mass exceeds the local Jeans mass the flow is convergent the cooling time of the gas is shorter than the dynamical time of the gas in the cell Yuan Li, University of Michigan38 => Form a star particle m * Cen & Ostriker (1992) http://arxiv.org/abs/1503.02660

39. Stellar Feedback Locally inject mass and thermal energy: Yuan Li, University of Michigan39 ε SNe m * c 2, ε SNe =10 -5 Resolution~ 250 pc SMBH feedback efficiency ε SMBH = 1% The Simulation

40. 80 kpc Yuan Li, University of Michigan40

41. Yuan Li, University of Michigan41

42. Dynamical Balance Between Cooling and AGN Heating 42 Temperature Projection 300 kpc

43. What Happened to the Cold Gas? Yuan Li, University of Michigan43

44. Comparison with Observations Yuan Li, University of Michigan44

45. 45 Tremblay et al. 2015

46. Hα Luminosity vs. min(t cool /t ff ) Yuan Li, University of Michigan46 Voit & Donahue 2014

47. SFR vs. Cold Dense Gas Mass Yuan Li, University of Michigan47

48. Radio Power vs. Molecular Gas Mass Yuan Li, University of Michigan48 Salomé&Combes 2003 No Correlation

49. Radio Power vs. Cooling Time Yuan Li, University of Michigan49 Mittal et al. 2009 No Correlation

50. SFR vs. Ṁ SMBH Yuan Li, University of Michigan50 No Correlation

51. The T-Rex Effect Yuan Li, University of Michigan51 ICM SMBH

52. Yuan Li, University of Michigan52

53. Yuan Li, University of Michigan53 SFR vs.

54. SMBH feedback efficiency 0.1% Yuan Li, University of Michigan54

55. Conclusion AGN jets drive non-liner perturbations, causing the marginally stable ICM to precipitate. AGN feedback regulates star formation by heating the ICM; star formation and stellar feedback regulate the long-term AGN cycles. AGN feedback balances ICM cooling in a dynamical way. The instantaneous SMBH accretion rate shows large variations on short timescales, but the average correlates with the SFR. Constraints on SMBH feedback efficiency. Too low => too many stars. Yuan Li, University of Michigan55

56. Young Stars & All Stars Yuan Li, University of Michigan56

57. Comparison with Observations Yuan Li, University of Michigan Artist's conception of a classic cooling flow Composite X-ray image of Perseus Core Fabian et al. (2006) Synthetic X-ray image from the simulation 57

58. Comparison with Observations Synthetic Hα map from the simulation Yuan Li, University of Michigan Artist's conception of a classic cooling flow Hα image of the Perseus cluster core. 58