21cm Constraints on Reionization Benedetta Ciardi MPA T. Di Matteo (CMU), A. Ferrara (SISSA), I. Iliev (CITA), P. Madau (UCSC), A. Maselli (MPA), F. Miniati (ETH), R. Salvaterra (UInsubria), E. Scannapieco (UCSB), P. Shapiro (UAustin), F. Stoehr (IAU), M. Valdes (SISSA), S. White (MPA)

The gas needs to cool to be available for star formation The first sites of ionizing radiation Atomic hydrogen cooling Molecular hydrogen cooling DM halos virial temperature Cooling function (no metals!) Temperature [K] Λ/n² [erg cm³/s] H2 cooling Atomic H cooling HD cooling Temperature [K] 1-σ 2-σ 3-σ (Barkana & Loeb 2001) Redshift H2 is a key species in the early universe

Feedback effects on primordial objects The first generation of stars affects the subsequent star formation process Radiative feedback Mechanical feedback Chemical feedback H2 photodissociation (L-W photons, eV) Radiative feedback Photoheating (Gnedin 2000) Photoevaporation (Shapiro et al. 2003)

Feedback effects on primordial objects Blowout/blowaway Blow-away Blow-out (Mac Low & Ferrara 99) Radiative feedback Mechanical feedback Chemical feedback Mechanical feedback

Feedback effects on primordial objects Change in fragmentation mode due to Z Radiative feedback Mechanical feedback Chemical feedback ~ 100 Msun ~ 1 Msun ~ 0.1 Msun Fraction of metals depleted onto dust grains Metallicity / Solar Metallicity (Schneider et al) (Bromm et al, Schneider et al)

Feedback effects on primordial objects Radiative feedback Mechanical feedback Chemical feedback IGM reionization See Ciardi & Ferrara 2005 for a review

Constraints on the epoch of reionization High-z QSOs  latest stages of reionization at z~6 CMB  global amount of electrons Which are the first sources of ionizing radiation? How did the reionization process evolve? Which is its interplay with the galaxy formation process?

21 cm line diagnostic 21 cm Ideal probe of neutral H at high-z different observed frqs.  different z (ν~150,120,80 MHz  z~8,11,18) DECOUPLING HI ground state CMB photons are absorbed by HI  thermal equilibrium u l Scattering with Lyalpha photons emitted by the same stars that reionize

Absorption or emission? Differential brightness temperature: absorption emission 30<z<200  absorption z_ion<z<30  emission Heating by Lyalpha or X-ray photons, shock heating (Madau, Meiksin & Rees 1997; Giroux & Shull 2001; Glover & Brand 2002; Chen & Miralda-Escude’ 2003; Gnedin & Shaver 2004) (1+z) (Loeb & Zaldarriaga 2004) IGM CMB

Simulations of galaxy formation (feedback effects) Galaxies Quasars Properties of the sources of ionizing radiation Radiative transfer of ionizing radiation Modeling of cosmic reionization: ingredients computationally demanding

Tsu3: cosmological radiative transfer codes comparison (Iliev, BC et al. 2006) Stromgren sphere Dense clumpCosmological field

Simulations of galaxy formation (feedback effects) Galaxies Quasars Properties of the sources of ionizing radiation Radiative transfer of ionizing radiation Modeling of cosmic reionization: ingredients computationally demanding

Spectral Energy Distribution ? Source emission properties: ? Escape Fraction: Modeling of cosmic reionization: parameters Zero or higher metallicity? Salpeter or Larson IMF? Fesc <20% but there is a big variation in the number both theoretically & observationally ? IMF and spectrum:

Simulations of reionization Simulation properties Source properties - - metal-free stars - L=20/h Mpc com. - Salpeter/Larson IMF - Fesc=5-20% l Simulations of galaxy formation  gas & galaxy properties (Springel et al. ‘00; Stoehr ‘04) l Stellar type sources  emission properties l  propagation of ionizing photons (BC et al. ’01; Maselli, Ferrara & BC ’03)

(BC, Stoehr & White 2003) Redshift Evolution of HI density ‘Proto-Cluster’: 10/h Mpc ‘Field’: 20/h Mpc z=18 z=16 z=14 z=12 z=10 z= The environment affects the reionization process

S5: Salpeter IMF+fesc=5% (late reion. case) S20: Salpeter IMF+fesc=20% L20: Larson IMF+fesc=20% (early reion. case) Early/Late Reionization 68% CL (Kogut et al. 2003) (BC, Ferrara & White 2003) The simulations are consistent with the WMAP results Source properties:

- MHs are sinks of UV photons - MHs photoevaporation has been studied in hydro-simulation and implemented in semi-analytic models (e.g. Shapiro, Iliev & Raga 2004; Iliev, Scannapieco & Shapiro 2005) Effect of mini-halos cell sub-grid physics Fraction of gas collapsed in MHs UV photons needed to photoevaporate the MHs 1. No MHs re-formation allowed once they are evaporated 2. MHs are allowed to form again

Effect of mini-halos Ionization fraction no MHs MHs, re-formation MHs, no re-formation With no MHs re-formation there is no substantial difference form the no-MHs case If MHs are allowed to re-form, reionization can be delayed by Δz~2 (BC et al 2006)

cm line diagnostic (BC & Madau 2003) The 21cm line is observed in emission if:

Fluctuations of brightness temp. The fluctuations are due to variations in HI distribution (density distrib. + ionized regions) l Late/Early reionization show similar behaviour l The peak of the emission is ~10 mK l Early reion. 90MHz, late reion. 115MHz S5L20 Future radio telescopes should be able to detect such signal

LOw Frequency ARray

(Valdes, BC et al., 2006) Instrument sampling Instrument sensitivity Convolution with a Gaussian beam (  =3 arcmin) LOFAR will be able to map the reionization history, especially its latest stages LOFAR expected response (late reion.) SimulatedSynthetic z=10.6, ν=122 MHz z=9.89, ν=130 MHz z=9.26, ν=138 MHz

S5 L20  CMB anisotropies are produced by free electrons  21cm line is emitted by neutral hydrogen The power spectrum is dominated by the secondary anis. at l > 4000 (θ < 3’). They reach a peak of ~ which could be detected by e.g. ALMA, ACT CMB secondary anisotropies CMB/21cm line correlation (Salvaterra, BC, Ferrara & Baccigalupi 2005) S5 L20

Characteristic angular scale of the cross-correlation function S5 L20 Mpc/h The characteristic angular scale of the cross-correlation function gives an estimate of the typical dimension of the HII regions at redshift of the 21cm emission line. CMB/21cm line correlation We find an anti-correlation below a characteristic angular scale, θ 0, when the correlation function becomes < 0.

Extra-galactic foreground contamination (Di Matteo, BC & Miniati 2004)  Point Sources: - Free-free emission from IS HII regions - Low-z radio galaxies  Extended Sources: - Free-free emission from IG HII regions - Syncrotron emission from cluster radio halos & relics Late reionization MHz F>0.1 mJy

Extra-galactic foreground contamination Radio galaxies Radio halos After removal of bright sources (F>0.1 mJy), at scales >1 arcmin 21cm emission line is free from extra-galactic foreground contamination

Conclusions Simulations of cosmic reionization 21cm line emission from neutral IGM (from reionization simulations) CMB/21cm line cross-correlation Extra-galactic foreground contamination calculated self-consistently with emission signal LOFAR should be able to map the reionization process, at least its latest stages More detailed information will be acquired implementing the above observations with CMB anisotropies measurements Observations of 21cm line at angles > 1 arcmin could be free from extra-galactic foreground contamination