‘laes as a probe of the high-z universe’, ringberg, march, 2009 lyman alpha emitters (laes) as a...

‘LAEs as a Probe of the High-z Universe’, Ringberg, March, 2009

Lyman Alpha Emitters (LAEs) as a Probe of the High-Redshift Universe

Mark Dijkstra (CfA)Collaborators: Stuart Wyithe (Melbourne), Zoltan Haiman, (Columbia) Avi Loeb, Adam Lidz (CfA),

Andrei Z Mesinger (Prinzeton), Marco Spaans (Groningen)

Image credit: Kim Nillson


• Lya is produced following recombination in HII regions surrounding O + B stars.

Llya scales in proportion with SFR*(1-fesc) (Llya=7-25%Lbol!)

• Ionization state of the IGM may leave imprint on observed Lya flux.

Introduction

A neutral IGM may suppress the number of observed Lya emitters beyond the redshift corresponding to the end of the EoR (e.g. Haiman & Spaans 99).

HI


• Observed number density of LAEs drops suddenly beyond z=6?

Are existing known LAEs probing the EoR?

Log (Lya Luminosity)

Cum

. Num

ber

dens

ity

89 LAEs observed at z=5.7 (Shimasaku+06, blue squares), 57 LAEs observed at z=6.5 (Kashikawa+06, red circles)

(also see Ota+08, and talks by Ota & M. Ouchi)




Cum

. Num

ber

dens

ity

89 LAEs observed at z=5.7 (Shimasaku+06, blue squares), 57 LAEs observed at z=6.5 (Kashikawa+06, red circles)

Restframe UV LF remains constant!




• For galaxies of a given restframe UV flux density, their corresponding measured Lya flux from galaxies at z=6.5 is lower than at z=5.7 (Kashikawa+06). How much lower?


Cum

. Num

ber

dens

ity



• Observations imply we receive ~ 10-80% (~95% CL) Lya photons per restframe UV continuum photon from z=6.5 compared to z=5.7 (D, Wyithe & Haiman+07). ~ 30% see M. Ouchi’s talk.

• Why?

– Evolution in f_esc? (Because Llya ~ [1-fesc]). See Yajima’s talk.– Dust?– These effects involve a detailed understanding of galaxies at z>5.5 – Less Lya is transmitted through IGM?

• Gas densities evolve as (1+z)3; nHI~(1+z)6

• (Re)Ionized gas can be significantly more opaque at z=6.5 than at z=5.7.

1.0 1.2 1.4 1.6 1.8

Opacity Ratio



• Observations imply we receive ~ 10-80% (~95% CL) Lya photons per restframe UV continuum photon from z=6.5 compared to z=5.7 (D, Wyithe & Haiman+07)

• Why?

– Evolution in f_esc? (Because Llya ~ [1-fesc]). – Dust?– These effects involve a detailed understanding of galaxies at z>5.5 – Less Lya is transmitted through IGM?

• Gas densities evolve as (1+z)3; nHI~(1+z)6

• (Re)Ionized gas can be significantly more opaque at z=6.5 than at z=5.7.

HI

HIHI

Residual HI gas inside a reionized IGM may be comprise a (large) evolving source of opacity for LAEs



• What is the opacity of residual HI gas in a reionized patch of the IGM to LAEs? • ‘First-order’ treatment the IGM: Lya line before IGM processing assumed to be a Gaussian with FWHM set by bulk motions of HII

regions within galaxy (~vcirc of host DM halo)

• Photons emitted blueward of Lya resonance eventually redshift into Lya resonance where IGM is opaque: transmission is ~TIGM (1) blueward (redward) of Lya resonance, I.e T>0.5.

-ln TIGM

Faucher-Giguere+ 08

Blue Red

The Opacity of the Ionized IGM to LAEs


• This prescription for the IGM is only valid when galaxies are randomly distributed throughout the Universe.

• However, galaxies preferentially form in overdense regions of the Universe and are highly clustered.

• When quantifying the opacity of the IGM around Lya emitting galaxies one must account for (e.g. D. Lidz & Wyithe 07, Iliev+08):

– local overdensity of IGM gas around galaxies

– Infall of IGM gas near galaxies (gas is *not* comoving with Hubble flow )

– Enhancement of local ionizing background (due to source clustering)



• Chicken Chicken Chicken Chicken Chicken Chicken Chicken Chicken Chicken Chicken Chicken

Chicken.


Chicken.


Chicken (Chicken+07,Chicken+08):

– Chicken Chicken Chicken Chicken Chicken

– Chicken Chicken Chicken Chicken (Chicken Chicken Chicken Chicken )

– Chicken Chicken Chicken (Chicken Chicken Chicken )

Chicken Chicken Chicken Chicken


• Impact IGM in more realistic model.

• Account for ‘average’ overdensity + peculiar velocities of gas around virialized halo (Barkana 04) (in red, schematically)

• IGM transmits 10-30% of emitted flux (D, Lidz & Wyithe +07)• IGM at z=6.5 can be up to 30% times more opaque (10-80% was observed)-> cannot conclude that the observed evolution ‘proof’ of probing EoR.

Absorption redward of systemic, by infalling gas, which is denser



• However, the impact of the IGM is depends on prominence ‘back-scattering’ mechanism.

• Depending on velocity + HI column density, majority of Lya photons can escape from galaxy with large enough redshift for the IGM to become irrelevant.

Lya source

RED BLUE

Verhamme+06

~1-10 kpc

The Impact of Winds on the Lya Line Shape


• Backscattering mechanism can nicely reproduce some observed Lya line shapes (Verhamme+08,Schaerer & Verhamme+08).

Blue Red

‘Backscattering’ transforms originally Gaussian emission line into a redshifted (few hundred km/s) Lya emission line.

The Impact of Winds on the Lya Line Shape


• Can we constrain ‘wind properties’ (NHI and vexp) from the Lya line shape?

Blue Red

Spectrum associated with specific IGM model (completely different model..)

D, Haiman & Spaans 06

Degeneracies likely exist when modeling the Lya line shape.

Furthermore, outflows may be ‘clumpy’, which allows a larger fraction of Lya to escape at the systemic velocity.

Do We Understand the Winds + Their Impact on the Lya Line?


Furthermore, outflows may be ‘clumpy’, which allows a larger fraction of Lya to escape at the systemic velocity (Hansen & Oh 06).

VS

Do We Understand the Winds + Their Impact on the Lya Line?


• Opacity + redshift its evolution of (re)ionized IGM are not well constrained.

– Main source of opacity is gas at 1-5 rvir. Variation from sightline-to-sightline is expected.How much? To be investigated with high resolution hydro simulations (Mesinger+ in prep) .

• How important are HI outflows in redshifting the Lya line emerging from LAEs?

– If very important: I.e. all Lya emerges with a systemic redshift of few hundred km/s, then the ionized IGM may not provide an important source of opacity to LAEs.

– However, if only a small fraction (~10 %) of Lya still emerges at systemic velocity, then studying the resonant opacity of the IGM is important (which is likely the case).

– Interesting that existing single-shell outflow models (e.g. Verhamme+08) for LAEs predict high level of linear polarization (D & Loeb 08) > testable.

• Note that a significantly neutral ‘interbubble’ IGM suppresses the flux regardless of winds etc.

Which Model Aspects Need to be Improved?


• Lyman Alpha Emitting Galaxies/ Lyman Alpha Emitters (LAEs) can probe the ionization state of H in the intergalactic medium (IGM).

• The H ionization state may be affected by the process of Helium reionization.

He Reionization with LAEs

Faucher-Giguere+08, Bernardi+03

‘Feature’ in average IGM opacity

--> He+ reionization?

~ 1 million LAEs at z=2-4 by ~2011.


Appendix


III: Polarization of Scattered Lya

• Scattered photons can appear polarized to an observer (electric vectors of photons have some preferred directions).

• Consider photon whose path is indicated with



• Scattered photons can appear polarized to an observer (electric vectors of photons have some preferred directions).

• Lya scattering can in practise be described accurately by Rayleigh scattering, for which scattering by deg, results in 100[sin2 /(1+cos2 )] % polarization.

Electric vector of photon

Propagation direction of photon



• Compute polarization of backscattered Lya radiation using a Monte-Carlo radiative transfer code (D & Loeb ‘08, also see Lee & Ahn ‘98). In this code:

– the trajectories of individual photons are simulated as they scatter off H atoms (microphysics of scattering is accurate)

– can attach a polarization vector to each photon, and– compute observed quantities such as the Lya spectrum, surface brightness profile,

and the polarization

• Polarization quantified as P=|Il-Ir|/(Il+Ir). Single photon contributes cos2 to Il and sin2 to Ir (Rybicki & Loeb 99).

• Apply Monte-Carlo code to a central Lya emitting source, completely surrounded

by a thin, single, expanding shell of HI gas (as in Verhamme+06,08). Free parameters are NHI and vexp.



• Lya can reach high levels of polarization (~40%, D & Loeb ‘08)

• Polarization depends on NHI and vsh, and therefore provides additional constraints on scattering medium (frequency dependence of polarization also constrains sign of vsh , see D & Loeb ‘08).

Pol

ariz

atio

n

Impact parameter

45%

18%Impact parameter

‘laes as a probe of the high-z universe’, ringberg, march, 2009 lyman alpha emitters (laes) as a...

Documents