Black Hole Astrophysics. Cole Miller, University of Maryland

Transcription

1 Black Hole Astrophysics Cole Miller, University of Maryland 1

2 Outline Why do we think BHs exist? The feeding of BH The dynamics of BH Ask questions any time! Would you be comfortable with group discussion? Approaches in theoretical astrophysics 2

3 Range of BH Masses We are confident that 3-20 M sun BH exist Similarly for M sun BH Not as clear: M sun BH ( intermediate-mass black holes ) More generally, what observational evidence suggests that black holes exist? 3

4 Accretion Disks Matter spirals onto BH from companion or surrounding gas Produces X-rays, UV, opt,... Need to ensure that object is too massive to be NS (spectra, timing very similar) 4

5 Stellar Orbits Easiest in our Galaxy (see right) Otherwise, see blended lines (look for Keplerian) Must argue that nothing else could explain data (such as dense cluster) 5

6 Masers Natural masers provide high intensity, angular localization is excellent Can see Keplerian motion NGC 4258: Swinburne astronomy group, Australia 6

7 Gravitational Lensing Achromatic brightening of background star Duration depends on lens mass but also on angular speed BH in galaxy minimal compared to galaxy mass 7

8 Brightness and Variability Very bright things could be accreting BHs, but to establish size need to have fast variability D<~cΔt, with possible beaming modifications 8

9 Feeding Black Holes As in, how do you get gas to a BH to light it up? 9

10 Wind Accretion From massive or giant stars Little net angular momentum Short-lived phase (little mass transferred) 10

11 Roche Lobe Overflow Donor overflows L1 Easier to get high accretion rates Can have from low-mass donor Can thus last longer Still doesn t add much mass or spin to BH (does to NS) 11

12 Small opening angle outflows Seen in protostars, white dwarfs, NS, BH Seems to require spin, probably mag field But much is not known Jets Image credit: NASA 12

13 Bondi-Hoyle Accretion Accretion from ISM/IGM Easiest in center of galaxy; cold flows 13

14 More About Bondi-Hoyle Mass accretion rate: 0 M = B GM c 2 s+v1 2 1 C A2 (c 2 s + v1) 2 1/2 ρ=density at infinity, c s =sound speed, v=relative speed at infinity, M=mass of BH, λ~1=eigenvalue. But note: if luminous, feedback increases c s, decreases ρ. Extremely difficult to grow stellar mass BH (straightforward to grow SMBH because T~M/(dM/dt)~1/M). 14

15 Eddington Luminosity In sph symm, outward rad force balances inward gravitational force L Edd =1.3x10 38 erg/s (M/M sun ) typically Surprisingly well-obeyed 15

16 Efficiency of Accretion At luminosities ~ L Edd, L~( )c 2 dm/dt Much below L Edd, get radiatively inefficient flow Much above, trapped radiation? 16

17 Characteristic Growth Time If L=ηc 2 (dm/dt), then M/(dM/dt)=45 Myr (η/0.1) -1 at L=L Edd Interesting puzzle: 10 9 M sun BH seen at z=7, but not enough time to grow from 10 M sun seed Supermassive star seed? Runaway cluster collapse? Slightly super-edd accretion? Currently under debate 17

18 Transport of Angular Momentum To accrete: mass in, ang mom out But how? Molecular viscosity too small Shakura+Sunyaev (1973): anom. viscosity T rφ =αp (α-model), α~0.1; physical origin? 18

19 Magnetorotational Instability Velikhov 1959; astro appl. Balbus+Hawley 1991 Spring analogy In disk: weak magnetic field lines tangle, amplify, transport angular momentum Turbulence! Don t find in α-model Credit: John Hawley 19

20 MRI Movie (John Hawley) 20

21 Open Questions Is a net vertical magnetic field needed for observed angular momentum transport? What about for jets? Quasiperiodic brightness oscillations are seen from many stellar-mass BH. What causes them? MRI destroys some modes Quasar microlensing suggests that the standard model isn t quite right; annular fluctuations (Dexter+Agol)? Cause? 21

22 Spin Alignment of SMBH Binaries When spinning BHs of comparable mass merge, kick could be thousands of km/s If the spins are aligned, kick is <200 km/s; implications for retention of BH postmerger Are there processes that might tend to align SMBH spin with each other and orbit? 22

23 Bardeen-Petterson Effect CXC Back-reaction of frame-dragging of disk by black hole causes hole to realign large lever-arm leads to effective realignment Bardeen-Petterson (1975) Natarajan & Armitage (1999) Sorathia et al. 2013a,b Tremaine & Davis /18/14 23

24 Miller+Krolik 2013 Circumbinary Disk Initial spins generally misaligned with each other, orbit, and the normal to the circumbinary plane at large distance. Spins dictate gas plane close to holes; binary dictates gas plane far from holes but close enough to binary. Net result: fast alignment of spins with each other and orbit. 9/18/14 24

25 What Could Prevent Alignment? If stellar dynamics is important, will tilt orbit but (usually) not individual spins Thus mergers of gas-poor galaxies might lead to large kicks If gas is prevented from getting to individual black holes, spins aren t aligned (Gerosa & Lodato) If gas all flows to one hole, the other is not aligned TBD whether these are realistic 25

26 Coevolution of SMBH, Galaxies Properties of SMBH are pretty well correlated with galaxies (e.g., M-σ) Many explanations! Deviations at high and low mass Important, but causes, scatter are Gultekin et al not well understood 26

27 Black Hole Feedback: Galaxies Galaxy formation is inefficient at low, high masses Low mass: star formation feedback? High mass: AGN feedback? Seems necessary to explain colors, too Dark matter halo Galaxy mass function 27

28 Black Hole Feedback: Clusters Centers of many galaxy clusters should have cool gas flow, SFR~10 3 M sun /yr But they don t; best guess is feedback from central AGN Perseus Cluster 28

29 Questions About Accretion? 29

30 The Dynamics of Black Holes How do black holes interact gravitationally with stars and gas? 30

31 Dynamical Friction Related to Bondi- Hoyle accretion Heavy objects sink in grav. potential T~1/M SMBH in galaxies have bulge around them, ~500x more massive; speedup! Until bulge stripped 31

32 Final Parsec Problem In galaxy merger, SMBHs with bulges drift rapidly to center But at ~0.1-1 parsecs, SMBHs have kicked out most stars Requires relaxation time to restock; can be >10 10 years for many galaxies Will stall unless BHs can get to ~10-3 pc Solutions? Triaxiality, massive perturbers, gas interactions 32

33 Extreme Mass Ratio Inspirals EMRIs Stellar-mass object (star, WD, NS, BH) spiraling into supermassive black hole One of main gravitational wave targets of elisa; very precise mapping of spacetime Also important for tidal disruption events But what are the important dynamical processes? 33

34 Two-Body Relaxation (largely) distant grav. 2-bod interactions Time when de/e~1 is energy relaxation time For M BH ~few x 10 6 M sun, t r ~few Gyr (10M sun /m). Less for smaller SMBH For equal-mass, core contracts, halo expands (increases entropy) Binaries, MBH have effects on distribution Evol. of low mass star dist. Decressin et al /18/14 34

35 Mass Segregation For distributed masses, heavies sink Seen in many globulars Boosts BH EMRI rate dramatically Combined with greater visibility, BH-SMBH dominate rate Pre-segregation or topheavy IMF? M22 (GC) radial annuli Albrow, De Marchi, Sahu /18/14 35

36 Net Rotation? 2-bod relaxation and mass segregation are enhanced when relative speeds of stars are decreased If there is net rotation in the inner ~1 pc, relaxation times are therefore decreased Some simulations suggest this could make huge difference to rates, properties E.g., work by Spurzem and colleagues 9/18/14 36

37 Resonant Relaxation Near SMBH, orbits are nearly Kepler ellipses If orbit orientation is ~fixed, torques can change angular momentum faster than 2-bod relaxation Issue: GR precession Schwarzschild barrier Rauch & Tremaine /18/14 37

38 Triaxiality Galaxy collisions can cause cores to be triaxial Then no symmetry preserves angular momentum of individual orbits Not as true close to SMBH Increased feeding rates to center, boosting EMRIs? M87 9/18/14 38

39 Expected Eccentricities Gravitational radiation circularizes orbits except very close to ISCO For highly eccentric, circularize at roughly constant pericenter distance Interactions with stars can increase or decrease eccentricities How do these play out in different circumstances? 9/18/14 39

40 Circularization Under pure GW evolution, time is Here µ is the reduced mass, M is total, f is GW frequency. Time is ~same for inspiral, circularization. f=10-4 Hz, µ=10 M sun, M=10 6 M sun, e=0.5: τ GW ~8x10 5 yr. Shorter than relaxation time. 9/18/14 40

41 EMRI 1: High Eccentricity Inspiral High apocenter orbit 2-body rel -> plunge Small pericenter means loss of energy Inspiral over orb Eccentric in LISA band Arbitrary inclination Triaxiality unlikely to boost Eccentric in LISA band Energy dissipation by gravitational radiation Related to tidal disruptions Courtesy V. Lauburg 9/18/14 41

42 EMRI 2: Binary Tidal Separation What if BH in binary? Miller et al 2005 Binary separates No energy loss needed High pericenter, low apocenter Low e, arb i in LISA band Triaxiality might boost Circular in LISA band. Arbitrary inclination. No energy dissipation required for capture Courtesy of V. Lauburg 9/18/14 42

43 EMRI Scenario 3: Settling in Accretion Disk Miralda-Escude & Kollmeier 2005 (also Yuri Levin) Star plunging through disk settles in disk Zero eccentricity Zero inclination Yunes et al. 2009: e=0, i=0, a/m=0 EOB EMRI 9/18/14 43

44 Rates? Very uncertain! Estimates of LISA detection rates of EMRIs range from per year Lots of theoretical and observational work needed to reduce the uncertainties 9/18/14 44

45 Questions about BH Dynamics? 45

46 Summary Far from being loners, black holes affect the evolution of objects as large as galaxy clusters Observations are coming at an increasing rate, as are computational models Rich range of astrophysics! 46