The simulated 21 cm signal during the EoR : Ly-α and X-ray fluctuations

Transcription

1 The simulated 21 cm signal during the EoR : Ly-α and X-ray fluctuations Sunghye BAEK Collaborators : B. Semelin, P. Di Matteo, F. Combes, Y. Revaz LERMA - Observatoire de Paris 9 Dec 2008

2 Physics of the 21 cm Line LICORICE for the 21 cm Transition Simulations X-ray effect on T k Future Work

3 Physics of the 21 cm Line The Differential Brightness Temperature ( 1 + z δt b 28.1 mk x HI (1 + δ) 10 ) 1 2 TS T CMB H(z)/(1 + z) T S dv r /dr The usual assumption T s T k T CMB No need for computing T s or T k No signal in absorption

4 Is T s T k T CMB always true? T 1 S 1. T s T k is true either = T 1 CMB + x αtc 1 + x c T 1 K 1 + x α + x c x α 1 : sufficient Ly-α scattering (not in the early EoR) x c 1 : sufficient collision (effective where δρ/ρ) 2. T k T CMB is true when neutral IGM in the voids is sufficiently pre-heated. = In some cases(early EoR), we need to compute T k ( x, z) and T s ( x, z) as well as x HI ( x, z) for an exact estimation of 21 cm transition.

5 The code : LICORICE General RT methods Monte Carlo ray-tracing Adaptive grid

6 The code : LICORICE 1. Compute T k ( x, z) and x HI ( x, z) in UV Continuum and now X-rays continuum Hydrogen and Helium Heating and Cooling process Adiabatic expansion Adaptive time step for ionization and cooling General RT methods Monte Carlo ray-tracing Adaptive grid +... shock heating(future work): need coupled hydro radiative simulation!

7 The code : LICORICE 2. Compute T s ( x, z) with Ly-α line transfer (Semelin et al. 2007) General RT methods Monte Carlo ray-tracing Adaptive grid local x α value from Ly-α line transfer Fully cosmological (redshifting photons, retarded time) Several acceleration schemes(a few tens of scatterings instead of 10 6 )

8 The code : LICORICE A typical run particles, 130 snapshots from z 40 to z 6 Dynamics with GADGET (Y.Revaz) Continuum RT 1000 CPU hours 10 Go shared memory(openmp) 10 8 photon packets Ly-α RT 1000 CPU hours 10 Go shared memory(openmp) 10 8 to 10 9 photon The next step particles in a 100 Mpc/h box ( 10 9 M halos) Dynamics : done RT : Possible on Vargas(IDRIS). 256 Go on a single node.

9 Simulations(Baek et al. 2008) DM Baryons(no He) 20 Mpc/h(S20) and 100 Mpc/h(S100) box size M and M resolved halos for S20 and S100 simulations The simulation pipeline 1. Dynamic (GADGET) Baryon overdensity, Star formation 2. UV continuum RT (LICORICE) T k ( x, z), x HI ( x, z) 3. Ly-α RT (LICORICE) T s ( x, z) movie movie

10 Result 1. 3D Line transfer is necessary Local Ly-α flux x α ( x, z) vs. homogeneous flux x α (z) show up to 50% difference in δt b locally Visible effect in the 3D powerspectrum when < x α >= 1 (Directly observable by interferometers)

11 Result 2. x α is not always + When x α > 10 (< x HII > 0.04), the error made on δt b by assuming x α = + is smaller than 10%

12 Result 3. Signal in absorption Figure: Early (moderate Ly-α)

13 Result 3. Signal in absorption Figure: Later(x HI = 0.5)

14 X-ray heating The Source Model QSOs, X binaries, SNe Soft X-ray photon 100eV to 2keV (Prichard & Furlanetto 2007) Spectral power index α = 1.6 (Telfer et al 2002) 0.1% of L tot to L QSO and 99.9% to L stellar (Glover & Brand 2003)

15 X-ray heating Method Ray-tracing of X-ray photons (homogeneous background X) Redshifting photon, retarded time λ X 4.9 ( ) E 3 Mpc (comoving) 300eV Secondary ionization and heating by high energy electrons (Shull & van Steenberg 1985)

16 The evolution of T k with X-ray heating

17 The evolution of T k with X-ray heating

18 The evolution of T k with X-ray heating First conclusion Preheating takes time!

19 Reference S. Baek, P. Di Matteo, B. Semelin, F. Combes, Y. Revaz (A&A accepted) arxiv: Future Work in 100 to 250 Mpc/h Helium + X-ray heating dv r /dr effect Shock heating in coupled simulation