Stellar processes near AGN

1
Stellar processes nearAGN

• Sergei Nayakshin
• Bus 6, and University of Leicester, UK
• Jorge Cuadra (MPA ? JILA, Colorado)
• Rashid Sunyaev, Volker Springel (MPA)
• Reinhard Genzel (MPE)

2
Theoretical Astroph. in Leicester need 3
postdocs(next year)

• Accretion discs in AGN/X-ray binaries (A. King,
  G. Wynn, SN)
• Self-gravitating discs, star formation (SN, G.
  Lodato)
• Galaxy formation (W. Dehnen, M. Wilkinson, A.
  King)

3
Outline

• Motivation
• Observational constraints on the young stars near
  Sgr A
• Theory and numerical simulations
• AGN implications
• Warped discs for AGN torii
• Conclusions

4
Binaries
Accretion in a binary system
5
AGN fuelling

• Mergers are believed to fuel most luminous
  quasars
• Antennae a merging galaxy pair

Credit Brad Whitmore (STScI) and NASA
6
Centre of the Galaxy
Chandra X-ray map/radio emission overlay
7
Sgr A SMBH in the GC

• Only stellar orbits give out Sgr A
• Luminosity of Sgr A is only 100 L
• Very little gas in the inner parsec
  now
• Young stars, M gt 20 M_sun, L gt 10 5 L_sun
• Where did young stars come from?

Eisenhauer et al 05
8
Stellar rings around Sgr A

• Levin Beloborodov 03, Genzel et al. 2003

9
View of the rings on the sky
(Genzel et al. 2003) view of the rings on the
sky
10
Orbital precession in axisymmetric potential
Circular orbits with different R precess at
different rates. A planar disk will be warped.
11
Constraints on stellar rings from orbits
Nayakshin 2005
warping of a disk by an inclined stellar ring
12
Disk(s)

• Too much warping destroys disks

Nayakshin, Dehnen, Cuadra Genzel 2006
13
Top-heavy IMF
1) K-band luminosity function the data require
a mass function flatter than a Salpeter function
by ?1 to 1.5 dex. Paumard et al. (2006) 2)
X-ray YSO constrains (Nayakshin Sunyaev 2005)
(Baganoff et al. 2003)
Chandra upper limit

• Stars with M lt 3 Msun are
• under-abundant by a factor of 5-10

14
Cluster disruption model
Paumard et al 2006

• A young 10(5-6) M_sun cluster formed at 10 pc
  could sink in to Sgr A rapidly (Gerhard 2001, )

• Cluster disruption leaves a trail of stars
• But the disks have outer edges
• X-ray emission from young low-mass stars isnt
  observed (Nayakshin Sunyaev 2005)

15
Observational facts

• Stars are years old in two
  co-eval discs
• Total stellar mass Solar masses
• Circular (e 0.2) and thin clockwise disc
• Eccentric and fluffy counter-clockwise disc
• Top-heavy IMF
• Inner disc edge 1 or 0.03 pc
• Outer edge 0.5 pc
• Stars have formed in situ
• Some facts may not be facts

16
Star formation in an accretion disc
Paczynski 1978, Kolykhalov Sunyaev 1980,
Shlosman Begelman 1989, Collin Zahn 1999,
Gammie 2001, Goodman et al. 2003

BH gravity/self-gravity
Toomre (1964)

• For star formation to occur, need
• 1) Q lt 1
• 2) Short cooling time (Gammie 2001, Rice et al.
  2005)

17
Sgr A case
Nayakshin 2006
Nayakshin Cuadra 2005

• Theory
• mass
• inner radius 0.03 pc
• circular initial stellar orbits

18
Observational facts

• Stars are years old in two
  co-eval discs
• Total stellar mass Solar masses
• Circular (e 0.2) and thin clockwise disc
• Eccentric and fluffy counter-clockwise disc
• Top-heavy IMF
• Inner disc edge 1 or 0.03 pc
• Outer edge 0.5 pc
• Stars have likely formed in situ

??
?
?
?
??
?
19
Top - heavy IMF Slow fragmentation or SF
Feedback?
Nayakshin 2006
Levin 2006 gas clumps do not collapse until they
grow much larger by agglomeration.

• M_dot gt M_Edd,
• Disc is heated up from inside by the stars
• Q increases to gt 1,
• Further disc fragmentation stops,
• But accretion onto proto-stars continues

20
Results of SPH simulations
(Nayakshin, Cuadra Springel, in preparation)

• Use P-Gadget2. Parallel code with SPH to model
  gas physics, and collisionless particles to model
  stars.
• Modifications
• turn off cosmology/galaxy formation
  prescriptions
• sink particles to model individual star
  formation and accretion
• Finite collapse time for gas clumps
• proto-stellar clumps can merge
• locally constant cooling time
• Eddington-limited accretion rate
• (2-10) x 106 particles running on 16-64 CPUs
• Confirmed Rice, Lodato Armitage 05 no
  fragmentation for beta gt 6

21
(No Transcript)
22

23
MF of stars in simulations
24
Fragmentation rate
Disc heats up due to N-body heating from
stars Fragmentation stops Stars can grow to M gtgt
Jeans mass
25
Theory vs observations

• Stars are years old in two
  co-eval discs
• Total stellar mass Solar masses
• Circular (e 0.2) and thin clockwise disc (H/R
  0.2)
• Eccentric and fluffy counter-clockwise disc
• Top-heavy IMF
• Inner disc edge 1 or 0.03 pc
• Outer edge 0.5 pc
• Stars have likely formed in situ

??
?
?
?
?
??
?
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Theory vs observations

• Stars are years old in two
  co-eval discs
• Total stellar mass Solar masses
• Circular (e 0.2) and thin clockwise disc (H/R
  0.2)
• Eccentric and fluffy counter-clockwise disc
• Top-heavy IMF
• Inner disc edge 1 or 0.03 pc
• Outer edge 0.5 pc
• Stars have likely formed in situ

??
?
?
?
?
?
??
?

• Scenario A massive cloud interacted with
  circum-nuclear disk, fell in in two separate
  pieces
• Two disks ? Moral Accretion disks can be eaten
  up by SF very quickly

32
AGN perspective accretion please, no SF
All disc models survive for times ltlt viscous
times except for F2 (strong feedback)
33
How do AGN disks feed SMBHs????

Sirko Goodman03 SED for accretion disc models
with M8 1, lE 0.5, a 0.01. The five SEDs, in
order of increasing total luminosity, correspond
to the five values of rmax/RS 103, 104, 105,
106 (diagonally hatched) and 107 (vertically
hatched). Multiple solutions for Teff(r) are
responsible for the ambiguity of the SEDs plotted
as bounded regions.
Personal opinion disk survives in a very clumpy
state
34
Conclusions (Star formation)

• AGN discs can form stars
• (Sgr A, and N. Levenson and R. Davies talks)
• SF enriches AGN disks with metals (e.g. Collin
  Zahn 1999)
• Sgr A was robbed of fuel by SF (would make a
  great AGN..)
• Problem How do SMBH get their food if SF steals
  it ???

?
35
Is torus inflated by star formation?

• Wada Norman 2002 Supernova explosions
• R 30 pc
• SN rate 1 per year (!) ? O(opt. thick) 0.2
• Accretion rate 0.4 Solar mass/year
• Works for a quasar but not AGN.

36
Is torus a stellar-driven wind?
movie
Model star-forming AGN disk. Stellar winds break
through the disk and escape to infinity.
SPH/N-body code Gadget-2 is used (Cuadra et al
2005,2006) Optically thin radiative cooling is
included GC model total mass loss rate is 0.1
M_sun/year, disk from R 0.04 to 0.3
parsec Stellar wind velocities of 300 km/s and
700 km/sec
37
Is torus a disk-driven wind?
Nayakshin Cuadra 2006
38
Is torus a disk-driven wind?
Solar mass per year
39
Is torus a disk-driven wind?

• Wind is great because
• Provides vertical support (Elitzur 05, Elitzur
  Shlosman 2006)
• Cloud collisions are avoided

• To make winds optically thick, super-Eddington
  mass loss rates are needed. Is it realistic for L
  1043 erg/s?
• Why would wind be equally powerful for low L and
  high L sources?
• Can the winds provide the needed obscuration ?

40
Is torus a disk-driven wind?

• If winds dont work, what does?
• Constraint
• Suggestion a warped and clumpy accretion disk
  (Nayakshin 2005)
• No problems with mass budget (mass is not
  wasted, its recycled)
• Clumps are naturally created collisions are
  mild
• Can the warps be maintained long enough?

type 1
type
2

Elitzur et al 03
41
Warped discs
Nayakshin 2005
warping of a disc by an inclined stellar ring
Other warped disk models F. Meyer et al, J.
Pringle et al.
42
4.3 Accretion in non-spherical potential
43
4.3 Warped star-forming eccentric discs
44
Conclusions

• parsec-scale SF is a major process regulating
• AGN feeding
• metal enrichment
• obscuration (?)

?
A lot of work to do!
45
Top-Heavy IMF why?

• Is the Jeans mass larger in an accretion disc? -
  No

Masses of first stars in a disc.
Levin 2006
46
Low mass YSO near Sgr A Chandra limits
Inner 40 parsec of the Galaxy (Baganoff et al.
2003)

• YSOs X-ray emission properties from Feigelson et
  al 2005,
• No more than 10000 YSOs (cf. Muno et al. 2004)
• Cluster model is ruled out (Nayakshin Sunyaev
  2005)

47
AGN accretion
A massive AGN disc would have short t_visc,
but Solar mass/year For M_BH 108, M_d
0.1 M_BH If BH expells most of mass in an outflow
(King, Pringle, Begelman), the BH doesnt grow