The 1 st galaxies and the cosmic web: the clustering of galaxy hosts from dark matter simulations Darren Reed Los Alamos National Laboratory arxiv: Katrin Heitmann (LANL) Salman Habib (LANL) Zarija Lukic (Illinois) Richard Bower (ICC-Durham) Carlos Frenk (ICC-Durham) Adrian Jenkins (ICC-Durham) Tom Theuns (ICC-Durham) (see astro-ph/ ) soccorro, 2008

Overview High redshift dark matter simulations – capture entire luminous mass range mini-halos “1 st stars” (T vir > ~2000K) “galaxy halos” (T vir > 10 4 K) We found how many halos (universal mass function)……… now need to know where are the halos? Clustering background and its importance Halo clustering results Conclusions

The dark ages……. L. Hernquist ionized, z > 1100 (CMB) reionized, z 6 neutral --> 21cm Spitzer Subaru LOFAR,SKA...

star formation in metal free gas T vir >10 4 K: Atomic cooling Efficient T vir >~ 2000K: H 2 line cooling Inefficient Mhalo ≥ ~10 8 Msun Mhalo ≥ ~ Msun H+e -  H - +hυ H - +H  H 2 +e- mini-halo ~ M sun ? cools lots of stars? 1 st galaxies?

high z (reionizing era) halo clustering -JWST (or sooner) to measure z >~ 6 clustering --galaxies form within halos -galaxy clustering at high redshift --sensitive to cosmology (e.g. σ 8, D.M. type) --sensitive to galaxy formation physics (e.g. SNe)  probe cosmology and galaxy form. physics

Simulation Techniques Cosmological parameters: (  =0.25,  =0.75,  8 =0.9, H 0 =73, n s =1)  8 : amplitude of power spectrum, (rms mass fluctuation of 8 Mpc/h spheres) + Hi-z SNe, 2df, etc. NASA/WMAP High-z (linear) gaussian random realization Evolve particles (gravity) L-GADGET2 (PM-tree code V. Springel)

z=10 12 Mpc/h

z=10

What is a halo? friends-of-friends ~iso-density link length ~ 0.2 lmean -->~“universal” halo mass function, f( σ mass (m)) Lukic et al z=0

Is f(σ) redshift invariant? “almost” f sim /f(σ) f sim /f(σ,n eff ) (z dependent) n eff ≡slope of P(k) at scale of halo P(k) α k neff Reed et al. 2007

highly clustered, “biased” weakly clustered bias(r)=(ξ halos (r)/ξ mass (r)) 1/2 ξ (r) = N pair /N pair_random -1

halo bias auto-correlation function of “galaxy” halos mass b =(ξ halos /ξ mass ) 1/2 halos mass

Scale-dependence of halo bias z=10 Sheth, Mo & Tormen large scale prediction (b SMT ) Small scales strongly biased “non-universal” vs lo-z sims M vir > our fit lo-z sim fit by Diaferio et al. (2003)

Scale-dependence of halo bias z=10 Sheth, Mo & Tormen large scale prediction (b SMT ) Small scales strongly biased “non-universal” vs lo-z sims b(r) =f( σ mass (r)) M vir > our fit lo-z sim fit by Diaferio et al. (2003)

z=10 12 Mpc/h

z=3

z=1

z=0

z=10

z=3

z=1

z=0

z=0 galaxiesz=10 galaxies Why is the scale dependence of halo bias so steep during the dark ages?

Universality of large-scale bias “universal” mass variable,, MW=predictions of Mo & White (1996) SMT=Sheth, Mo & Tormen (2001) Millennium run data from Gao et al. (2005).

mass halos Finite-box effects on velocities. Pairwise velocity dispersion along the line of separation

Conclusions Clustering measured in dark matter simulations  halos of 1 st “galaxies” and 1 st stars Spatial clustering  steep scale dependence, “non-universal”  Large-scale “universal” & consistent with analytic predictions Impending Observational uses:  cosmological probe on small scales, e.g. σ 8, DM type  physics of galaxy formation, local feedback effects, how galaxies populate halos