Ancient Universe with GRB, Kyoto, 2010 T HE F IRST S TARS Naoki Yoshida
Contents ✦ Star formation in the early universe - from big bang ripples to protostars ✦ The fate of very massive stars ✦ Low-metallicity star formation References: NY, Omukai, Hernquist, 2008, Science Bromm, NY, McKee, Hernquist, 2009, Nature (review) Ohkubo, Umeda, Nomoto, NY, Tsuruta, 2009, ApJ Umeda, NY, Nomoto, Ohkubo, Sasaki, Tsuruta, 2009, JCAP Omukai, Hosokawa, NY in prep.
From the ancient universe GRB 090423 @z=8.3 A massive star’s death just 600 million years after the Big Bang!
In the backyard... Frebel et al. 2005
Theory of Primordial Star Formation
density fluctuations The Standard Cosmology length scale An ab initio approach is possible CMB + LSS + SNe tell us about the initial state of the universe, and the energy content now and then precisely .
Primordial Star Formation A ‘Simple’ Problem The Physics - understood Gravity, hydrodynamics, atomic/molecular processes 14 species (H, He, D, ions) ~50 important reactions + many radiative processes Density evolution to ~10 21 /cc The Initial Condition - “observed” dark matter + hydrogen-helium gas + CMB In the Dark Age... Cosmologically determined; CDM model
In the beginning, there was a sea of light elements and dark matter ....and some ripples.
From diffuse gas to protostar The Omukai diagram A proto-star (hydrostatic core) The Physics 10 4 collision H 2 formation induced line cooling emission T [K] 3-body (NLTE) reaction opaque to 10 3 continuum loitering and (~LTE) opaque to dissociation Heat molecular adiabatic release line contraction M J ~1000 M sun 10 2 number density
An early universe “experiment” A complete picture of the formation process of a primordial protostar. Dynamic range 10 13 Resolving planetary scale structures in a cosmological volume! NY, Omukai, Hernquist Science 2008
The structure Composition Velocity The central 0.01 Msun core is a seed for the subsequent formation of a massive star. atomic core fully molecular
A hyper-accreting protostar “hot” infalling gas T ~ 500-10000K The gas mass accretion rate is estimated to be dM/dt = 0.01-0.1 M sun /yr. hydrostatic core This is enough to make a 10-100 M sun star within 1000 years! outer envelope
NY, Omukai, Hernquist, Abel, 2006, ApJ Protostellar evolution H burning starts dM/dt = 0.01-0.1 M sun/yr (time dependent) M ZAMS ~ 100 M sun
Primordial Star Formation 1. The large mass (~1000Msun) at the onset of collapse. 2. High temperature (~1000K) gas surrounding the protostar = Very large accretion rate 3. Lack of opacity source (no dust) = accretion continues Bromm+99 Nakamura-Umemura 01 Omukai-Palla 03 NY+ 06, 08 Omukai-Palla 03 Tan-McKee 04 Hosokawa’s poster
Massive PopIII Stars
A stick in our throat Very massive Pop III... Models and numerical simulations suggest that the first stars were rather massive, > 100 Msun However, there is no indication, no single evidence that pair-instability supernovae contributed Galactic chemical evolution. WHY ?
Pair-instability? No way. From N. Tominaga Observed metal-poor stars
Some resolutions 1. Blackholes are formed, without expelling metals (Madau&Rees01, Ohkubo+09) 2. Selection effect in metal-poor star surveys? (Karlsson+08) 3. There is another mode of Pop III star formation (Johnson&Bromm06, NY, Omukai 07)
Some resolutions 1. Blackholes are formed, without expelling metals (Madau&Rees01, Ohkubo+09) 2. Selection effect in metal-poor star surveys? (Karlsson+08) star formation (Johnson&Bromm06, NY+07) 3. There is another mode of Pop III
Model accretion rate 0.1 0.01 0.001 1 10 100 1000 M sun M sun /year Accretion rates from cosmo. simulations Y06, 1st Y07, 2nd gen. Ohkubo, Umeda, Nomoto, NY, Tsuruta, ApJ, 2009
Hyper Massive Star exhausted Ohkubo, Umeda, Nomoto, NY, Tsuruta 2009 Direct collapse in the core Hydrogen Stellar radius N S y t i l i b a t s n I r i a P 1 10 Mass 100 1000
Core evolution
NY, Oh, Kitayama, Hernquist (2007) ApJ Star formation in a reionized gas Temperature profile HD cools the gas! CMB plays a role! PopIII.2 after reionization
PopIII.2 after reionization mass Protostellar radius NY, Omukai, Hernquist 2007, ApJL Nakamura & Umemura 2003 NY, Omukai, Hernquist (2007, ApJL, 667, 117) Nagakura & Omukai 2005 McGreer & Bryan 2008 “conventional” 1st star Primordial stars formed by HD cooling are not very massive 2nd. gen. star because of low-T Stars formed by HD cooling with HD cooling ~ low dM/dt Hydrogen-burning starts at M ~ 30 M sun M ZAMS ~ 40 M sun M cloud ~ 40 M sun
What if the gas is enriched with heavy elements...
PopIII to PopII Is there a “critical metallicity” for cloud fragmentation ? If so, what determines it ? Bromm et al. cooling by C, O @low-density Omukai, Schneider cooling by dust @high density vs.
Toward a direct simulation Chemistry and radiative transfer in a gas with heavy elements and dust : １ Cooling by CI, CII, OI ２ Dust thermal emission ３ Molecular cooling by H 2 O, OH, CO ４ New cooling rates for H 2 , HD
Dust is a nightmare
Dust cooling temperature evolution T dust determined by the thermal balance: 4 T 4 = L gr (gas -> dust) Next talk by Kaz! T dust
Results: Z=-5 密度 温 度 H 2 HD H 2 O T dust
Oxygen chemistry : Z=-5 密度 [/cc] H 2 O OI OH H 2 + OH -> H 2 O + H O + H -> OH +
Fragmentation For Z=-5, Rapid cooling by dust at high density (n~10 14 ) leads to core fragmentation. Fragment mass ~ 0.1 Msun 5AU
Summary massive supermassive blackholes at high-z cloud fragments by dust cooling, • Primordial stars are predominantly • Remnant blackholes might seed the • At metallicity as low as Z=-5, gas
Binary formation (?) Turk+ (2009) simulation: the core breaks up at n ~ 10 12 /cc. Formation of massive (not very massive) star pair with ~ tens Msun. 1case over 5 samples. A large fraction produces PopIII binaries ?
Radiative transfer A direct comparison of density evolution a factor of ~100 at a given time Case with a fitting function ( a la Turk et al.) With 3D radiative transfer
Radiative transfer effects Molecular line escape probability Velocity gradient Similarly important for continuum The escape probability is not just a function of density; it varies with temperature (via Doppler width) and with velocity gradients, hence is highly time- and direction-dependent. NY, Omukai in prep
Recommend
More recommend
