Galactic Super Massive

1
Galactic Super Massive Binary Black Hole Mergers
Dr. Peter Berczik Astronomisches Rechen-Institut
(ARI), Zentrum für Astronomie Univ. Heidelberg,
Germany
berczik_at_ari.uni-heidelberg.de
Second RSDN meeting, 25-27 Nov. 2005, Hoher
List, Germany
2
Collaborators

• David Merritt, Rochester Institute of Technology,
  NY, USA
• Rainer Spurzem, ARI, Zentrum für Astronomie Univ.
  Heidelberg
• Gabor Kupi, ARI, Zentrum für Astronomie Univ.
  Heidelberg
• Stefan Harfst, Rochester Institute of Technology,
  NY, USA

Grants

• AST-0206031, AST-0420920 AST-0437519 from the
  NSF
• NNG04GJ48G from NASA
• HST-AR-09519.01-A from STScI
• SFB-439 from the Deutsche Forschungsgemeinschaft

Publications

• Berczik, Merritt Spurzem, 2005, ApJ, 633, 680,
  astro-ph/0507260
• Berczik, Merritt Spurzem, in prep

3
Galaxy Collisions
4
BHs in galaxies (MW - Sgr A)
5
Galaxy Collisions BHs collisions
Multiple Massive Black Holes NGC6240 strong
ongoing merger
6
Future Observations
Gravitational Wave Detection - LISA

• Two of the strongest potential sources in the
• low-frequency (LISA) regime are
• Coalescence of binary supermassive black holes
• Extreme-mass-ratio inspiral into supermassive
  black holes

7
Some of the previous works

• Milosavljevich M. Merritt D., 2001, ApJ, 563,
  34
• Hemsendorf M., Sigurdsson S. Spurzem R., 2002,
  ApJ, 581, 1256
• Chatterjee P., Hernquist L. Loeb A., 2003, ApJ,
  592, 32
• Makino J. Funato Y., 2004, ApJ, 602, 93
• Laun F. Merritt D., 2004, astro-ph/0408029
• Szell A., Merritt D. Seppo M., 2005,
  astro-ph/0502198

Dynamical Modeling Methods

• Direct N-body method
• As much as possible accurate
• Symmetry of the problem is irrelevant
• (-) Very compute intensive!!!

8
Basic idea of the N-body code
9
Our own GRAPEN-body1 parallel code
4th order Hermite scheme
Hierarchical Block Time Steps
10
GRAPE GRAvity PipE
N
N2
11
GRAPE GRAvity PipE more detail
12
GRAPE6a PCI board
GRAPE6a - PCI Board for PC-Clusters, recent
development of the University of Tokyo
128 Gflops for a price 5K USD Memory for N, up
to 128K particles
13
RIT ARI 32 node GRAPE6a clusters

• 32 dual-Xeon 3.0 GHz nodes
• 32 GRAPE6a
• 14 TB RAID
• Infiniband switch (10 Gb/s)
• Speed 4 Tflops
• N up to 4M
• Cost 500K USD
• Funding NSF/NASA/RIT

• 32 dual-Xeon 3.2 GHz nodes
• 32 GRAPE6a
• 32 FPGA
• 7 TB RAID
• Dual port Infiniband switch (20 Gb/s)
• Speed 4 Tflops
• N up to 4M
• Cost 350K EUR
• Funding Vwagen/BW/ARI

14
RIT ARI 32 node GRAPE6a clusters
15
Parallel PP on GRAPE6a cluster
N
Nact
N/Np
16
Parallel PP on GRAPE6a cluster
17
Parallel PP on GRAPE6a cluster
18
Parallel PP on GRAPE6a cluster
19
Parallel PP on GRAPE6a cluster
20
Initial Conditions - I
Z
Two equal-mass black holes near center of
Plummer-model galaxy
Y
X
21
Some Theory
N-body Integration of Binary Black Hole Dynamical
Evolution
star
Example loss-cone around a binary black hole.
Stars scattered into the binary are ejected via
the gravitational slingshot. The binary responds
by shrinking.
?
Full loss-cone
Diffuse regime
binary black hole
In a real galaxy, the shrinking rate (d/dt)(1/a)
would be limited by the rate of diffusion of
stars into the loss cone.
22
Results I (Plummer)
23
Results I (Plummer)
24
Initial Conditions - II
Z
Two equal-mass black holes near center of
King-model (W06) galaxy
Y
X
25
Results II (King)
26
Results I (Plummer) II (King)
27
Double check of the Results
28
Double check of the Results
29
BH collisions?
???
d 10R_BH
If we scaled up our numerical results, for the
typical galaxy bulge (109 Mo 3 kpc 10 Gyr
130) we see that the BHs separation never come
closer 1 0.1 pc For the typical BHs mass
(106 Mo) the gravitational merging regime
start with 10-6 pc!!!
30
Possible way of solution

• Initial data
• No equilibrium
• Higher initial eccentricity
• New code e0
• regularization (CHAIN - ?, KS - ?)
• Larger direct N simulations
• AC neighbor scheme
• N-body GAS (SPH)
• Hardware solution for SPH calculations (FPGA)

31
Conclusions

• First large direct N 1M parallel GRAPE6a cluster
  simulations
• The BBH decay rate is N dependent! 400K 1M
  particle is already enough to have a near
  diffuse regime
• The initial rotation of the host galaxy is very
  important for the BBH orbital evolution. For
  larger rotation we see the clear fixation of
  decay rate
• Some of the highly rotating models can produce
  the BBH with a very high eccentricity e1.
  Possible source of the low frequency GW (LISA)

Thank you for attention... ?