Star Formation at Very Low Metallicities

1
Star Formation at Very Low Metallicities

• Jonathan Tan
• Dept. of Astronomy
• University of Florida

Primordial Mini-Halo (Abel ea.) Orion
Nebula Cluster (McCaughrean)
2
Importance of First Stars and their Mass
Observations
CMB polarization (WMAP Page et al. 06) H 21cm
(LOFAR Morales Hewitt 03)
Reionization (H, He) Metal Enrichment proto-gala
xies and IGM Illumination SN
GRBs? Progenitors of SMBHs?
Z of low Z halo stars (e.g. Christlieb et al.
03) Z of Lya forest?(Schaye et al. 03 Norman et
al. 04)
NIR bkg. intensity (Santos et al. 03
Fernandez Komatsu 06) NIR bkg. fluctuations
(Kashlinsky et al. 04)
JWST (Weinmann Lilly 2005) SWIFT (Bromm Loeb
2002)
Sloan (Fan et al. 03 Willott et al. 03)
Influence on structure formation SMBHs, Globular
Clusters, Galaxies? Critical Z 5x10-3Z? (Bromm
et al. 2001)
Simple problem initial conditions,
chemistry, no feedback from other stars, weak
B-fields(?)
3
Star Formation A Complicated, Non-linear Process
Physics gravity vs pressure (thermal, magnetic,
turbulence, radiation, cosmic rays). Equation of
state heating and cooling, decay and sources of
turbulence, diffusion of B-fields, generation of
B-fields (dynamo), etc. Chemical evolution of
dust and gas.
Numerical models
Analytic theory
Observations
PopIII Initial Conditions
Wide range of scales (10 dex in space, time) and
multidimensional. Uncertain initial
conditions/boundary conditions.
Computers cannot yet solve the complete problem
time step Dt ? 1/(Gr)1/2 Need to break it up
into separable parts, tractable with analytic and
simple numerical methods (e.g. Stahler, Shu, Taam
1980). Need to understand physics, make
reasonable approximations, compare to
observations and numerical simulations.

4
Outline

• The formation of zero metallicity stars
• Initial conditions
• Accretion and protostellar evolution
• Feedback
• Implications for the next generation
• Ionization feedback and HD
• Critical metallicity
• Transition to present-day star formation

5
Numerical Simulations Results
Abel, Bryan, Norman (2002)

• Form pre-galactic halo 106M?
• Form quasi-hydrostatic gas core inside halo
• M4000M?, r 10 pc,
• nH 10 cm-3, fH2 10-3, Tgt200K
• Rapid 3-body H2 formation at nHgt1010cm-3.
• Strong cooling -gt supersonic inflow.

4. 1D simulations (Omukai Nishi 1998) Form
quasi-hydrostatic protostar nH 1016-17cm-3, T
2000 K optically thick, adiabatic contraction
-gt hydrostatic core with m0.005M?, r14R?
(also Ripamonti et al. 2002)
6
The initial conditions for primordial star
formation
H2 cooling -gt Tmin 200 K, ncrit 104
cm-3 MBE 380M? cs1.2 km/s
Centrally concentrated cloud, inefficient cooling
-gt quasi-hydrostatic contraction Density
structure self-similar, ??r-2.2
Yoshida et al. 2006
7
Slow contraction damps turbulent motions and
inhibits fragmentation
Dt 200 yr
Dt 1.5 kyr
Dt 3 kyr
Dt 30 kyr
Dt 0.3 Myr
Dt 9 Myr
z 19
8
Heating by Dark Matter Annihilation?
Chen Kamionkowski 2004 Ripamonti ea. 2007
Spolyar ea. 2007 A.Natarajan Tan 2007
baryons
dark matter
Main uncertainties ?DM(rlt10-3pc), m?, lt?vgt,
ftrap Could potentially delay/halt collapse
Abel, Bryan, Norman (2002)
9
The Accretion RateCollapse of an Isentropic
Sphere
Density structure self-similar,
k2.2 singular polytropic sphere in virial and
hydrostatic equilibrium Can model with a
polytropic equation of state

?1.1
10
Comparison withNumerical Models
Omukai Nishi (1998)
Within factor 2 of Bromm Loeb 03 for 1st 104 yr
(Mlt50 M?)
Tan McKee (2004)
Ripamonti et al. (2002)
Yoshida et al. (2006)
Note ABN curve based on estimate from infall
rate prior to protostar formation
ABN (2002)
11
Collapse to a disk around a central protostar
Rotation core forms from mergers and collapse
along filaments expect Jgt0 fKep ? vcirc / vKep
? 0.5
12
Disk Models
(Tan McKee 2004 Tan Blackman 2004)
?0.3
Surface density
Thickness
Ionization
Temperature Tc , Teff
Toomre
.
m
17x10-3M?/yr
6.4x10-3M?/yr
2.4x10-3M?/yr
13
Magnetic Fields Inverse Helical Dynamo
Blackman Field (2002)
Turbulence in disk amplifies field strength,
eventually saturating due to back reaction of
B-field (Tan Blackman 2004). Large-scale
helicity is generated in each hemisphere, leading
to strong, coherent, large-scale B-fields, which
rise up to form a corona and outflow.
14
Magnetic Fields Saturation Points
Tan Blackman (2004)
Exponential growth of seed field (B10-16G)
eventually becomes resistively limited (Blackman
Field 2002). t m/m
Gravito-Turbulence
MRI-Turbulence
Unsaturated
Saturated
.
Growth limited by disk lifetime
Growth limited by local accretion time in disk
(?ss0.01, 0.3)
Either Gravito- or MRI-driven turbulence can
provide viscosity to remove angular momentum.
15
Evolution of the Protostar depends on Accretion
Rate
(Stahler et al. 1980 Palla Stahler 1991
Nakano 1995 Omukai Palla 2003 Tan McKee
2003)
Assume polytropic stellar structure and
continuous sequence of equilibria
One zone model follow energy of protostar as
it accretes
Deuterium burning for Tcgt106K Structural
rearrangement after tKelvin Eddington model for
? Solve for r(m), until reach main sequence
16
Evolution of the Protostar
Radius
Comparison with Stahler et al. (1986), Omukai
Palla (2001)
Initial condition m 0.04 M? r 14
R? (Ripamonti et al. 02)
Protostar is large (100 R?) until it is older
than tKelvin Contraction to Main
Sequence Accretion along Main Sequence
17
Evolution of the Protostar
Luminosity
Boundary Layer
Accretion Disk
18
Evolution of the Protostar
Ionizing Luminosity
fKep0.5
Spectrum depends on initial rotation c.f. Omukai
Palla (2003)
19
Overview of Radiative Feedback
20
Growth of the HII Region with Rotation
McKee Tan 2007
Balance ionizing flux vs recombinations and
infall
HII Region
Find stellar mass at breakout rHII rg, where
vescci polar equatorial
21
Disk Photoevaporation
Hollenbach et al. 94
.
thin neutral disk
Mass loss rate for zero age main sequence
22
Accretion rate reduced by expansion of HII region
Accretion continues from regions shadowed by the
accretion disk. Need to calculate vertical
structure (i.e. thickness) of disk.
23
Radial profile of disk
m100M?
24
Feedback-limited accretion
McKee Tan 2007

• Self-consistent model for
• growth and evolution of
• protostar, including
• declining accretion rate
• accretion disk (r,z)
• protostellar structure
• ionizing feedback

Expect masses 60-400M? depending on K, fKep
25
Implications of IMF for supernovae and metal
enrichment
Fate of NON-ROTATING metal-free stars (Heger
Woosley 2002)
However, rapid rotation can lead to significant
mass loss (Meynet et al. 2004, 2006 Hirschi
2006), so initial stellar mass is not necessarily
the final pre-SN mass.
26
Impact of Massive PopIII Stars

• Radiative and Mechanical Feedback
• - Unbinding of gas from host and neighboring
  mini halos (ionization, winds, SNe)
• (Shapiro ea. 2004 Kitayama ea. 2004 Whalen ea.
  2004 Iliev ea. 2005 Alvarez ea. 2006 Yoshida
  ea. 2007)
• - Dissociation of H2 while star shines -gt
  contribution to FUV background
• (Haiman ea 1997, 2000 Omukai Nishi 1999
  Glover Brand 2001 Machacek ea. 2001 Oh
  Haiman 2002 Yoshida ea. 2003 Ahn Shapiro
  2006 Susa 2007 OShea Norman 2007b)
• - (Relic) Ionization can promote H2 and HD
  formation promote star formation
• (Shapiro Kang 1987 Kang Shapiro 1992 Susa
  ea 1998 Uehara Inutsuka 2000 Nakamura
  Umemura 2002 Ricotti ea. 2002 Nagakura Omukai
  2005 Johnson Bromm 2006, 2007 Yoshida 2006
  Ripamonti 2006 Susa Umemura 2006 Yoshida ea.
  2007a,b)
• -Cosmic ray ionization can promote H2 and HD
  formation
• (Stacy Bromm 2007 Jasche ea. 2007)

27
Impact of Massive PopIII Stars
2. Production of metals and dust to change gas
cooling properties mf? SNe yields (mf)? Dust
formation in SNe? Mixing? - Supernova
remnants (Madau ea. 2001 Mori ea. 2002 Bromm
ea. 2003 Wada Venkatesan 2003 Norman ea.
2004 Greif ea. 2007) - Effect on cooling
function (Omukai 2000 Bromm ea. 2001
Schneider ea. 2002 Bromm Loeb 2003 Frebel ea.
2007 Smith Sigurdsson 2007 Smith ea. 2007)
28
Comparison to Local Massive Star Formation
e.g. Beuther et al. (2007, PPV)
Orion Nebula Cluster (McCaughrean)
29
Overview of Physical Scales
AV7.5 NH1.6x1022cm-2 ?180 M? pc-2
AV1.4 NH3.0x1021cm-2 ?34 M? pc-2
30
Overview of Physical Scales
AV200 NH4.2x1023cm-2 ?4800 M? pc-2
AV7.5 NH1.6x1022cm-2 ?180 M? pc-2
AV1.4 NH3.0x1021cm-2 ?34 M? pc-2
31
Overview of Physical Scales
PopIII core mostly supported by thermal pressure.
Weak cooling. No fragmentation.
AV200 NH4.2x1023cm-2 ?4800 M? pc-2
AV7.5 NH1.6x1022cm-2 ?180 M? pc-2
AV1.4 NH3.0x1021cm-2 ?34 M? pc-2
32
Transition from PopIII

• PopIII cores supported by thermal pressure
• Present-day cores and protoclusters supported by
  magnetic fields and turbulence
• Magnetic fields and turbulence unlikely to be
  effective in post-PopIII regions
• (B-fields probably too weak turbulence damps
  out)

• Post-PopIII regions may be able to cool
• effectively via HD and/or metals-dust
• If cs ltlt virial velocity of halo, then gas
• may collapse to form a rotationally
• supported thin disk possible fragmentation.
• (Lodato P. Natarajan 2006 Clark, Glover
  Klessen 2007)
• However, continued infall and merging of
  halos
• may maintain turbulence (Wise Abel 2007)

33
Conclusions
Approximately convergent initial conditions for
star formation set by H2 cooling. Mostly thermal
pressure support slow cooling -gt no (disk)
fragmentation -gt single massive star in each
mini-halo (no clusters of primordial stars). DM
heating?
Need analytic model to follow the growth of the
protostar accretion rate accretion disk
protostellar evolution feedback. Large (AU)
protostar, contracts to main sequence for
mgt30M? This is when feedback processes start to
become important.
Feedback processes depend on core rotation and
accretion rate Gradual reduction in SF
efficiency because of ionizing and
radiation pressure feedback for mgt30M?. Final
mass in fiducial case (K1, fKep0.5), is
160M?, set by ionizing feedback on core and disk.
Changing accretion rate by factor of 5
(K0.5-2) leads to final masses in the range 60
- 400M? (seeds for SMBHs?)
Impact of PopIII stars on subsequent SF
suppressed cooling due to FUV background
enhanced cooling from metals/dust and ionization
catalyzed H2/HD -gt rotational supported,
fragmenting cores? Eventual build up of B-field
support to resemble present-day SF clusters.