Lyman-α Galaxies at High Redshift
James E. Rhoads
(Space Telescope Science Institute)
with
Sangeeta Malhotra,
Steve Dawson, Arjun Dey, Buell Jannuzi, Emily
Landes, Hy Spinrad, Dan Stern, JunXian Wang,
and Chun Xu
Lyman α as a Signpost of Primordial
Galaxies
We are looking for the first galaxies to form from
primordial gas. Such objects should have…
• Young stars, with high ultraviolet luminosities.
• Considerable amounts of gas.
• Very low abundances of heavy elements, implying
– Stars that are hotter than “normal” (for fixed mass);
– Low dust abundances.
Lyman α Line Emission
• Ionizing flux + gas → 2 Lyman α photons for
every 3 ionizing photons absorbed by hydrogen.
• In principle, up to 6-7% of a young galaxy’s
luminosity may emerge in the Lyman α line (for a
Salpeter IMF).
The Narrowband Search Method
• We take pictures of a
large area of sky in
both broad and
narrow filters.
• Emission line
sources appear
bright in narrow
filters (demand 5σ).
• The blue “veto
filter” eliminates
foreground emission
line objects (demand
< 2σ).
Windows for Narrowband Surveys
Red lines
mark our
survey filter
wavelengths.
Line searches
are highly
efficient in
far red!
The Large Area Lyman Alpha Survey
(LALA): an Overview
Ω; Volume
Sensitivity
z
Candidates,
Spectroscopic
Success rate
4.5
1.4x106 Mpc3
(1/2 in Bootes)
1.7x10-17 ergs/s/cm2
350; > 70%
5.7
6 x105 Mpc3
(1/3 in Bootes)
1x10-17 ergs/s/cm2
~50; ~ 70%
6.6
1.5x105 Mpc3
(all in Bootes)
2x10-17 ergs/s/cm2
3; 1 of 3
confirmed.
z=4.5: Rhoads et al 2000, Malhotra & Rhoads 2002, Dawson et al 2004 in prep;
z=5.7: Rhoads & Malhotra 2001, Rhoads et al 2003, Wang et al 2004 in prep.
z=6.5: Rhoads et al 2004
Physical Nature of Ly-α Galaxies
These are young star forming galaxies:
• High equivalent widths require youth, plus either
low metallicity or top-heavy IMF (Malhotra &
Rhoads 2002)
• AGN are ruled out by a combination of X-ray
observations (see JunXian Wang’s talk [48.14] in
this session) and optical spectroscopy (see Steve
Dawson’s talk [48.04] in this session).
Candidate z=6.5 LALA galaxies
Ic (NDWFS) z’SDSS
NB918
All data from NOAO 4m telescopes;
NB918 stack is 24 hours’ integration.
LALA J142442.24+353400.2 at z=6.535
• Gemini spectrum
shows an
Asymmetric line and
no continuum.
• Nod and shuffle
helps eliminate the
possibility of other
lines if [OIII] (5007)
(Rhoads et al. 2004;
Gemini spectrum
reduced by Chun
Xu.)
Physical properties of this galaxy
• Line luminosity 1043 erg s-1.
• Star formation rate: 11 Msun / year
– Based on Lyman-α line + Kennicutt IMF + Case B
recombination… Some modelling uncertainty.
• Equivalent width > 530Å observer frame (2σ
lower bound from photometry)
– Rest frame > 70 Å (naively)
– Boosting z’ flux to account for IGM absorption under
Madau’s (1995) prescription gives > 40Å
• Spatially unresolved in 1” seeing  < 6 kpc.
“May you live in interesting times…”
• The z = 6.5 universe was just 6% of its present age.
• Moreover, this precedes the redshift of the SDSS
quasars with Gunn-Peterson troughs.
• Thus the 9200Å window affords a view of late
reionization.
The Lyman-α ReionizationTest
Ionized IGM
Continuum
Photons
To
Observer
Lyman α
photons
Young starburst
The Lyman-α Reionization Test
Neutral IGM
Continuum
Photons
To
Observer
Young starburst
Lyman α
photons
(Miralda-Escude 1998; Miralda-Escude & Rees 1998;
Haiman & Spaans 1999; Loeb & Rybicki 1999)
Constraining Reionization with LALA
• Lyman α emitters at
z=5.7 and 6.5 suggest
zre > 6.5.
(Rhoads & Malhotra 2001; ApJ
Lett 563, L5; Hu et al 2002;
Rhoads et al 2003; Rhoads et
al 2004)
• Large equiv. width of
the LALA object 
line attenuation factor
is a few at most.
Figure: z=5.7 Ly-α sources
from Rhoads et al 2003
Charting Reionization
Current evidence: Combine the Lyman α and Gunn-Peterson
tests so far to study the evolution of the mass avgeraged
neutral fraction, x:
There is no
contradiction
between the GP
effect at z=6.2
and the Ly α at
z=6.56.
(Figure from
Rhoads et al
2003)
Lyman α Test Details
• HII regions: R>1.2 pMpc  line center τ0< 1.
– Residual neutrals in HII region matter.
• τ < τ0 in the red wing of lines (Haiman 2002)
• Evolution affects galaxy counts… LBG
comparison samples could resolve this.
• Gas motions alter Ly-α radiative transfer (Santos
2003).
• Bottom line: Factors of ~ 3 reduction in Ly-α
luminosity are unavoidable in neutral universe.
Reduction larger in many models but not all.
Star Formation and Metal Production
• Using LF constraint from HCM 6a  ~3e-2
Msun/yr/Mpc3 star formation, and ~ 200 Msun/Myr/Mpc3
of metal production… comparable to z ~ 4 numbers from
UV continuum!
• Using instead lens+blank slit data of Santos et al could
reduce both numbers by a factor ~ 30.
• Compare SFRD estimates of 5.2e-4 (Kodaira et al, z=6.5
Ly-α sample); 4.6e-4 (Stanway et al, z ~ 6 LBG sample);
neither integrates a LF much below the observed limit.
• The usual conclusion: More good data wanted!
Summary
• Lyman-α provides an efficient way of finding
young galaxies at high redshift.
• These galaxies afford a new probe of reionization,
complementing the Gunn-Peterson test.
• Ly-α from LALA z=6.535 galaxy requires either a
largely ionized IGM, or an intrinsic redshift of the
line relative to systemic velocity.
• Remaining uncertainties in Lyman-α reionization
test can be overcome using statistical samples.