February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp.

Chapter 2

Dark ages and cosmic reionization

Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Tirthankar Roy Choudhury

Harish-Chandra Research Institute,
Chhatnag Road, Jhusi, Allahabad 211 019, India.
E-mail: [email protected]

Abstract: About 300,000 years after the Big Bang, the protons and the electrons
combined for the first time in the Universe to form hydrogen (and helium) atoms,
which is known as the recombination epoch. Following that, the Universe entered
a phase called the “dark ages” where no significant radiation sources existed.
The dark ages ended once the first structures collapsed and luminous sources like
stars and accreting black holes started forming. The radiation from these sources
then ionized hydrogen atoms in the surrounding medium, a process known as
“reionization”. Reionization is thus the second major change in the ionization
state of hydrogen (and helium) in the Universe (the first being the recombination).
The study of dark ages and cosmic reionization has acquired increasing sig-
nificance over the last few years because of various reasons. On the observational
front, we now have good quality data of different types at high redshifts (quasar
absorption spectra, radiation backgrounds at different frequencies, number counts
of galaxies, cosmic microwave background polarization, Lyα emitters and so on).
Theoretically, the importance of the reionization lies in its close coupling with the
formation of first cosmic structures, and there have been numerous progresses in
modeling the process. In this article, we introduce the basic concepts involving
the formation of first structures and evolution of the ionization history of the
Universe. We also discuss the possibility of constraining the reionization history
by matching theoretical models with observations.

15
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

16 Vignettes in gravitation and cosmology

2.1 Introduction

Study of reionization mostly concerns with the ionization and thermal his-
tory of the baryons (hydrogen and helium) in our Universe [36, 2, 14, 19].
Within the framework of hot Big Bang model, hydrogen formed for the first
time when the age of the Universe was about 3 × 105 years, its size being
one-thousandth of the present (corresponding to a scale factor a ≈ 0.001
and a redshift z = 1/a − 1 ≈ 1000). The epoch at which the protons and
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

the electrons combined for the first time to form hydrogen atoms is known
as the recombination epoch and is well-probed by the Cosmic Microwave
Background Radiation (CMBR).
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Right after the recombination epoch, the Universe entered a phase called
the “dark ages” where no significant radiation sources existed. The hydro-
gen remained largely neutral at this phase. The small inhomogeneities in
the dark matter density field which were present during the recombination
epoch started growing via gravitational instability giving rise to highly non-
linear structures like the collapsed haloes. It should, however, be kept in
mind that most of the baryons at high redshifts do not reside within these
haloes, they rather reside as diffuse gas within the intergalactic space which
is known as the intergalactic medium (IGM) [47, 45].
The collapsed haloes form potential wells whose depth depend on their
mass and the baryons (i.e, hydrogen) then “fall” in these wells. If the mass
of the halo is high enough (i.e., the potential well is deep enough), the gas
will be able to dissipate its energy, cool via atomic or molecular transitions
and fragment within the halo. This produces conditions appropriate for
condensation of gas and forming stars and galaxies. Once these luminous
objects form, the era of dark ages can be thought of being over.
The first population of luminous stars and galaxies can generate ultravi-
olet (UV) radiation through nuclear reactions. In addition to the galaxies,
perhaps an early population of accreting black holes (quasars) also gener-
ated some amount UV radiation. The UV radiation contains photons with
energies > 13.6 eV which are then able to ionize hydrogen atoms in the sur-
rounding medium, a process known as “reionization”. Reionization is thus
the second major change in the ionization state of hydrogen (and helium)
in the Universe (the first being the recombination).
As per our current understanding, reionization started around the time
when first structures formed, which is currently believed to be around z ≈
20 − 30. In the simplest picture, each source first produced an ionized
region around it; these regions then overlapped and percolated into the
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 17

IGM. This era is usually called the “pre-overlap” phase. The process of
overlapping seemed to be completed around z ≈ 6 − 8 at which point the
neutral hydrogen fraction fell to values lower than 10−4 . Following that
a never-ending “post-reionization” (or “post-overlap”) phase started which
implies that the Universe is largely ionized at present epoch. Reionization
by UV radiation is also accompanied by heating: electron which are released
by photoionization will deposit an extra energy equivalent to hp ν − 13.6 eV
to the IGM, where ν is the frequency of the ionizing photon and hp is
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

the Planck constant. This reheating of the IGM can expel the gas and/or
suppress cooling in the low mass haloes – thus, there is a considerable
reduction in the cosmic star formation right after reionization.
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

The process of reionization is of immense importance in the study of

structure formation since, on one hand, it is a direct consequence of the
formation of first structures and luminous sources while, on the other, it
affects subsequent structure formation. Observationally, the reionization
era represents a phase of the Universe which is yet to be probed; the earlier
phases (z ≈ 1000) are probed by the CMBR while the post-reionization
phase (z < 6) is probed by various observations based on galaxies, clusters,
quasars and other sources. In addition to the importance outlined above,
the study of dark ages and cosmic reionization has acquired increasing
significance over the last few years because of the availability of good quality
data in different areas.
In this article, we will introduce various concepts which go into mod-
elling reionization. The main aim would be to systematically discuss the
set of equations which are crucial in understanding the process highlight-
ing the major physical processes and assumptions. We shall also highlight
the relevant observational probes at appropriate places. In Section 2.2, we
shall give a pedagogic introduction to the basic theoretical formalism for
studying reionization and IGM in different phases of evolution. Section 2.3
would be devoted to discussing detailed modelling of reionization using the
formalism developed. We shall illustrate on how to constrain the models
by comparing with a wide variety of available data sets. In Section 2.4, we
shall briefly discuss the current numerical simulations and observations re-
lated to reionization. We shall also briefly highlight what to expect in this
field in near future. More detailed discussions on this topic and detailed
reference lists can be found elsewhere [12].
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

18 Vignettes in gravitation and cosmology

2.2 Theoretical formalism

In this section, we discuss the basic theoretical formalism required for mod-
elling reionization of the IGM. The main aim here would be to highlight
the physical processes which are crucial in understanding reionization and
comparing with observations. In what follows, we shall assume that the
IGM consists only of hydrogen and neglect the presence of helium. It is
straightforward to include helium into the formalism.
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

Essentially, in presence of a ionizing radiation, the evolution of the mean

neutral hydrogen density nHI 1 is given by
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

ṅHI = −3H(t)nHI − ΓHI nHI + Cα(T )nHII ne (2.1)

where overdots denote the total time derivative d/dt, H ≡ ȧ/a is the Hub-
ble parameter, ΓHI is the photoionization rate per hydrogen atom, α(T )
is the recombination rate coefficient and ne represents the mean electron
density. The first term in the right hand side of equation (2.1) corresponds
to the dilution in the density because of cosmic expansion, the second term
corresponds to photoionization by the ionizing flux and the third term
corresponds to recombination of protons and free electrons into neutral
hydrogen. The quantity C is called the clumping factor and is defined as

hnHII ne i hn2H i
C≡ = (2.2)
hnHII ihne i hnH i2
where the last equality holds for the case when the IGM contains only
hydrogen (i.e., no helium) and is highly ionized, i.e, ne = nHII ≈ nH . The
clumping factor takes into account the fact that the recombination rate
in an inhomogeneous (clumpy) IGM is higher than a medium of uniform
density.
The ionization equation is usually supplemented by the evolution of the
IGM temperature T , which is given by

Ėkin = −2H(t)Ekin + Λ (2.3)

where Ekin = 3kB T nH is the kinetic energy of the gas and Λ is the net
heating rate including all possible heating and cooling processes. The first
term on the right hand side takes into account the adiabatic cooling of the
gas because of cosmic expansion.
1 In astrophysical notation, HI stands for neutral hydrogen while HII denotes ionized

hydrogen (proton).
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 19

2.2.1 Cosmological radiation transfer

The equation of radiation transfer, which describes propagation of radiation
flux through a medium, is written as an evolution equation for the specific
intensity of radiation Iν ≡ I(t, x, ν, n̂) which has dimensions of the energy
per unit time per unit area per unit solid angle per frequency range. It is
a function of time and space coordinates (t, x), the frequency of radiation
ν and the direction of propagation n̂. The radiation transfer equation in a
cosmological scenario has the form
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

∂Iν c ∂Iν
+ n̂ · ∇x Iν − H(t)ν + 3H(t)Iν
∂t a(t) ∂ν
c
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

= −cκν Iν + ν , (2.4)
4π
where κν is the absorption coefficient and ν is the emissivity. The above
equation is essentially the Boltzmann equation for photons with Iν being
directly proportional to the phase space distribution function [45]. The
terms on the left hand side of equation (2.4) add up to the total time
derivative of Iν ; in particular, the third term corresponds to dilution of
the intensity and the fourth term accounts for shift of frequency ν ∝ a−1
because of cosmic expansion. The effect of scattering (which is much rarer
than absorption in the IGM) can, in principle, be included in the κν term if
required. If the medium contains absorbers with number density nabs each
having a cross-section σν , the absorption coefficient is given by κν = nabs σν .
The mean free path of photons in the medium is given by λν (t) ≡ κ−1 ν (t).
We define the mean specific intensity by averaging Iν over a large volume
and over all directions
Z Z
d3 x dΩ
Jν (t) ≡ Iν (t, x, n̂) (2.5)
V V 4π
Then the spatially and angular-averaged radiation transfer equation be-
comes [48]
∂Jν ∂Jν c
J˙ν ≡ − H(t)ν = −3H(t)Jν − cκν Jν + ν (2.6)
∂t ∂ν 4π
where the coefficients κν and ν are now assumed to be averaged over the
large volume. The quantity Jν is essentially the energy per unit time per
unit area per frequency interval per solid angle.
The integral solution of the above equation along a line of sight can be
written as [29]
Z t  3 0 
c 0 0 a (t ) −τ (t,t0 ;ν)
Jν (t) = dt ν 0 (t ) 3 e . (2.7)
4π 0 a (t)
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

20 Vignettes in gravitation and cosmology

where ν 0 = νa(t)/a(t0 ), ν 00 = νa(t)/a(t00 ) and

Z t Z t
0 00 00 dt00
τ (t, t ; ν) ≡ c dt κν 00 (t ) = c 00
(2.8)
t0 t0 λν 00 (t )
0 0
is the optical depth along the line of sight from t to t > t Clearly, the
intensity at a given epoch is proportional to the integrated emissivity with
an exponential attenuation due to absorption in the medium. The intensity
attenuates by 1/e when the radiation travels a distance equal to the mean
free path.
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

The absorption is “local” when the mean free path of photons is much
smaller than the horizon size of the Universe, i.e., λν (t)  c/H(t). In ad-
dition, if we also assume that the emissivity ν does not evolve significantly
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

over the small time interval λ/c, then the specific intensity is related to the
emissivity through a simple form [38, 43]
ν (t)λν (t) ν (t)
Jν (t) ≈ = (2.9)
4π 4πκν (t)
Note that in the case of local absorption, J˙, HJ  c. In this approxima-
tion, the background intensity depends only on the instantaneous value of
the emissivity (and not its history) because all the photons are absorbed
shortly after being emitted (unless the sources evolve synchronously over
a timescale much shorter than the Hubble time). We shall discuss later in
Section 2.2.2 that this is a useful approximation for the IGM for redshifts
z & 3.

2.2.2 Post-reionization epoch

Let is first study the radiation transfer in the post-reionization epoch. Com-
pared to the pre-overlap era, this epoch is much easier to study because the
IGM can be treated as a highly ionized single-phase medium (whereas dur-
ing the pre-overlap era, one is looking into two distinct phases – ionized
and neutral). The optical depth can be written as
Z z0
dz 00
τ (z, z 0 ; ν) = c nHI (z 00 ) σabs (ν 00 ) (2.10)
z (1 + z 00 )H(z 00 )
where σabs (ν) is the total absorption cross section of neutral hydrogen and
we have changed the time coordinate to the redshift z. Various processes
can, in principle, contribute to σabs (ν), most dominant being the resonant
Lyman series absorption corresponding to excitation of hydrogen atoms
from the ground state to higher ones (1s → np) and the continuum absorp-
tion of photons above the ionization threshold via photoionization process.
Let us discuss them in the following:
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 21

2.2.2.1 Resonant Lyman series absorption

The Lyman series absorption arises from the electronic excitation of neu-
tral hydrogen atoms from the 1s ground state to higher ones. The most
dominant of these are the Lyα (1s → 2p, rest wavelength λα ≈ 1216 Å )
and Lyβ (1s → 3p, λβ ≈ 1206 Å ) transitions, and hence they are the most
relevant ones as far as observations are concerned. For simplicity, we shall
present results for the Lyα absorption only, the others can be calculated in
identical manner. The Lyα absorption cross section is given by
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

 
ν
σabs (ν) = σα V −1 (2.11)
να
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

where να = c/λα is the resonant frequency of transition, σα = 4.45 ×

10−18 cm2 is the cross section at νalpha and V is a function which determines
the profile of the absorption line. It is called the Voigt profile function and
is a convolution of the Lorentzian shape for the natural broadening and
the Gaussian shape for the thermal broadening. For the purpose of this
article, it is sufficient to note that V is a sharply peaked function about
ν/να = 1; for most our discussion, we shall take it to be a Dirac-delta
function V (ν/να − 1) = δD (ν/να − 1).
The optical depth between the redshifts z and z 0 is then given by
c
τ (z, z 0 ; ν) = τ (zα ) = σα nHI (zα );
H(zα )
να
1 + zα = (1 + z) (2.12)
ν
If we put this into equation (2.7), we see that the Lyα absorption at a
redshift zα reduces the specific intensity observed at z at a frequency να (1+
z)/(1 + zα ) by a factor e−τ (zα ) . The value of τ (zα ) along a given line of
sight would depend upon the distribution of nHI (zα ). However, we would
mostly be interested in the mean value of specific intensity averaged over a
number of lines of sight. The corresponding reduction can be described by
a line-of-sight-averaged optical depth
e−τeff (zα ) ≡ he−τ (zα ) iLOS (2.13)
where h iLOS denotes averaging over lines of sight. The quantity τeff is
usually known as the “effective optical depth”.
Theoretically, the value of τeff can be calculated if we know the dis-
tribution of optical depth P (τ ) [which can be calculated from the neutral
hydrogen distribution P (nHI )]:
Z ∞
−τeff (z)
e = dτ P (τ ; z) e−τ (2.14)
0
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

22 Vignettes in gravitation and cosmology

Of course, one requires detailed understanding of the evolution of the bary-

onic density field to model the distribution P (nHI ), which we shall avoid
discussing here. However, we can still make some inference assuming the
distribution is uniform, i.e., τeff (z) = cσα nHI (z)/H(z). If we define the neu-
tral hydrogen fraction to be xHI ≡ nHI /nH , then we can calculate τeff ∝ xHI
given a set of cosmological parameters (which would uniquely determine
H(z) and nH ).
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Fig. 2.1 The effective optical depth of Lyα absorption as function of redshift z. The
points with errorbars represent the observational data. The dashed curves, from top
to bottom, represent the predictions for a uniform IGM with neutral hydrogen fraction
xHI = (3, 1, 0.3) × 10−5 , respectively.

Observationally, τeff can be determined by looking at the spectra of

bright sources like quasars at high redshifts. These spectra show a series
of absorption features at frequencies larger than the Lyα frequency in the
quasar rest frame. Since one has a good knowledge of the unabsorbed
quasar spectra (from looking at nearby quasars and also having some un-
derstanding about the physical processes), one can calculate the amount of
absorption happening because of the intervening IGM between the quasar
and the observer; this absorption, averaged over numerous lines of sight, is
essentially the quantity e−τeff . This has been done to quite high redshifts
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 23

z ∼ 6.5 and the values of τeff observed are shown as points with errorbars
[61, 20] in Figure 2.1.
To understand what these values imply, we have plotted with dashed
lines the calculated value of τeff for a uniform IGM assuming three values of
xHI = (3, 1, 0.3) × 10−5 from top to bottom. This immediately tells us that
the fraction of neutral hydrogen has to be ∼ 10−5 in order to reproduce the
observed values of τeff . In fact, had xHI been slightly (say ∼ 10−4 ) higher,
one would have obtained τeff much higher than unity (∼ 10 − 100) and
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

hence the flux from the quasar would be completely absorbed. If that were
the case, it would show up as a absorption “trough” at frequencies larger
than the rest frame Lyα frequency. In reality, a considerable amount of
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

transmitted flux is found at these frequencies alongwith a series of absorp-

tion features arising from the Lyα transition of residual neutral hydrogen.
These absorption signatures are known as the “Lyα” forest and are powerful
probes of the neutral hydrogen distribution in the IGM at z < 6 [52].
The absence of a absorption trough is a direct proof of the fact that
hydrogen is completely reionized in the diffuse IGM at redshifts z . 6. This
is known as the Gunn-Peterson effect [28]. Note that the actual inferred
value of xHI might be slightly different if one models with an appropriate
density distribution, however, the basic conclusion remains unchanged.
We can also see from Figure 2.1 that for quasars at redshifts z & 6, the
observed value of τeff & 5; this would imply an attenuation & 0.01 and hence
one actually observes absorption troughs as predicted by Gunn-Peterson ef-
fect [20, 68]. Unfortunately, finding such troughs does not necessarily imply
that the IGM is highly neutral as even a xHI ∼ 10−4 could be sufficient
to absorb all the flux. However, one can use much detailed modelling to
improve the constraint, which we shall discuss later in Section 2.3.2.
The values of τeff . 1 at z < 4 means that the diffuse IGM is highly
transparent (also called optically thin) to Lyα photons. Only about ∼ 10%
of the Lyα photons are absorbed, mostly within the high density regions.
These high density systems are often modelled as a set of discrete absorbers
of some size. If we consider an absorber having neutral hydrogen density
nHI and a size L  c/H(z) along the line of sight at a redshift zabs , the
optical depth is given by

τ (z, z 0 ; ν) = NHI σabs (νabs ) (2.15)

where νabs = ν(1 + zabs )/(1 + z) and NHI ≡ nHI L is the column density
of neutral hydrogen within the absorber. Hence, each absorber reduces the
specific intensity by a factor e−NHI σabs (νabs ) . If we assume that the absorbers
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

24 Vignettes in gravitation and cosmology

are Poisson-distributed, then it is straightforward to show that the effective

optical depth is given by [46]
τeff (z, z 0 ; ν) =
Z z0 Z ∞
∂2N 00
dz 00
dNHI 00 [1 − e−NHI σabs (ν ) ] (2.16)
z 0 ∂z ∂N HI
where [∂ 2 N/∂z∂NHI] dz dNHI is the number of absorbers within (z, z + dz)
having column densities in the range (NHI , NHI + dNHI ).
In case of Lyα resonant absorption, we can use the cross section in
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

(2.11) to calculate τeff . Since V is a function which is sharply peaked

around ν/να = 1, we can approximate the above integral as
Z
1 + zα ∞ ∂2N
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

0
τeff (z, z ; ν) = dNHI Wα (NHI ) (2.17)
λα 0 ∂zα ∂NHI
where Z h 00
i
Wα (NHI ) ≡ dλ00 1 − e−NHI σα V (λα /λ −1) (2.18)
is called the “equivalent width” of the absorber.

2.2.2.2 Continuum absorption

In case of continuum absorption of radiation by photoionization, the cross
section is given by
σabs (ν) = σHI (ν)Θ(ν − νHI ) (2.19)
where σHI (ν) is the photoionization cross section and Θ is the Heaviside step
function taking into account that only photons with frequencies ν > νHI =
13.6 eV/hp would be absorbed by the photoionization process. The exact
form of σHI (ν) is rather complicated, however one can approximate it by a
power-law of the form σHI (ν) = σ0 (ν/νHI )−3 where σ0 = 6.3 × 10−18 cm2 .
Since σ0 ∼ σα , one can show that the absorption due to a diffuse IGM in
this case too is negligibly small. The only significant absorption can be seen
in very high density regions which have a large fraction of their hydrogen
in neutral form. In that case, we can use the relations obtained for a set
of Poisson-distributed absorbers in the vase of resonant transition. We
essentially have an optical depth of the form (2.17), and the corresponding
mean free path of ionizing photons due to these discrete absorbers is found
to be [38]
c
λν (z) =
H(z)(1 + z)
Z ∞ −1
∂2N −NHI σHI (ν)
× dNHI [1 − e ] (2.20)
0 ∂z∂NHI
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 25

At this point, let us introduce the concept of Lyman-limit systems which

have column densities NHI > σ0−1 = 1.6 × 1017 cm−2 ; these absorbers con-
tribute an optical depth of unity to the ionizing photons. The average
distance between these systems is given by
 −1
c dNLLS
λLLS = (2.21)
H(z)(1 + z) dz
where dNLLS /dz is the redshift distribution of the Lyman-limit systems
Z ∞
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

dNLLS ∂2N
= dNHI (2.22)
dz 1/σHI (νHI ) ∂z∂NHI
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

−1.5
For the observed distribution ∂ 2 N/∂z∂NHI ∝ NHI [49], one can show
from equations (2.20) and (2.21) that the mean free path is related to the
distance and redshift distribution Lyman-limit systems as
λLLS c
λνHI = √ = √ . (2.23)
π πH(z)(1 + z) dNLLS /dz
The redshift distribution of Lyman-limit systems dNLLS /dz is a quantity
which has been measured for 2 < z < 4.5 by observations of quasar ab-
sorption spectra. Though the observational constraints are poor, one can
still obtain a value dNLLS /dz ≈ 0.3(1 + z)−1.55 [65], which in turn gives the
mean free path as λνHI /[c/H(z)] ≈ 0.1[(1 + z)/4]−2.55 Hence the mean free
path of ionizing photons is much smaller than the horizon size for z > 3,
which implies that we can use the local absorption approximation at these
redshifts.
We can summarise the main results of this section as: the post-
reionization epoch is characterized by a highly ionized IGM as observed
by the quasar absorption spectra. The IGM is largely transparent to ion-
izing photons at these redshifts. However, there exist regions with high
column densities (NHI > 1017 cm−2 ) which are optically thick to the ioniz-
ing radiation; these regions determine the photon mean free path. We shall
see later how to use this information to obtain an improved model of the
IGM.

2.2.3 Pre-overlap epoch

We now turn our attention towards the IGM in the pre-overlap era. In
this era, the overlap of individual ionized regions is not complete and hence
the IGM is partially ionized. So the radiation transfer equation has to be
modified to account for the multi-phase nature of the IGM.
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

26 Vignettes in gravitation and cosmology

Let us define the volume filling factor of ionized regions to be QHII ;

this is the fraction of volume that is ionized and reionization is said to be
complete when QHII = 1. Next, note that the number density of photons
present in the background flux is
Z
4π ∞ Jν
nJ (t) = dν (2.24)
c νHI hp ν
Since there is no ionizing flux within the neutral regions (otherwise they
would not remain neutral), the photoionization rate per hydrogen atom
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

within the ionized (HII) regions is

Z ∞
II 1 Jν
ΓHI = 4π dν σHI (ν) (2.25)
QHII h pν
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

νHI

where the factor Q−1HII accounts for the fact that the radiation is limited to
a fraction of the total volume. The emission rate of ionizing photons per
unit volume from sources of emissivity ν is
Z ∞
ν
ṅph = dν (2.26)
νHI hp ν
Then the equation of radiation transfer becomes [43, 11]
4π JνHI
ṅJ = −3H(t)nJ − H(t) + ṅph
c hp
II dQHII
− nII II
HI QHII ΓHI − nHII (2.27)
dt
where nII II
HI and nHII are the number densities of neutral and ionized hydro-
gen within the HII regions, respectively. The first term in the right hand
side of equation (2.27) corresponds to the dilution in density due to cosmic
expansion while the second term accounts for the loss of ionizing radiation
because of a photon being redshifted below the ionization edge of hydro-
gen νHI . The third term is essentially the source of ionizing photons. The
fourth term accounts for the loss of photons in ionizing the residual neutral
hydrogen within the ionized regions. The fifth term, which is only relevant
for the pre-overlap stages, accounts for the photons which ionize hydrogen
for the first time and hence increase the filling factor QHII . For QHII = 1,
equation (2.27) reduces to that for the post-reionization phase.
If we now assume that the photons are absorbed locally, then J, ˙ HJ 
c and Jν is essentially given by equation (2.9). We can then ignore terms
containing J and nJ in equation (2.27). This gives an equation describing
the evolution of the filling factor QHII
dQHII ṅph
= II − QHII ΓII HI (2.28)
dt nHII
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 27

If we further assume photoionization equilibrium within the ionized region

d(nII 3 II II II II
HI a )/dt → 0, then we have from equation (2.1) nHI ΓHI = Cα(T )nHII ne
and the evolution of QHII can be written in the form [38]
dQHII ṅph
= II − QHII Cα(T )nII
e (2.29)
dt nHII
In this description, reionization is complete when QHII = 1 and equation
(2.29) cannot be evolved further on. Clearly the assumptions of local ab-
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

sorption and photoionization equilibrium (both of which are reasonably

accurate) has given us an equation which can be solved once we have a
model for estimating ν and C. Of course, there is a dependence of the
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

recombination rate coefficient α(T ) on temperature, however that depen-

dence is often ignored while studying the volume filling factor. In case
one is interested in temperature evolution, one has to solve equation (2.3)
taking into account all the heating and cooling processes in the IGM, in
particular, the photoheating by ionizing photons whose rate is given by
Z ∞
Jν
Γph,HI = 4π dν hp (ν − νHI )σHI (ν). (2.30)
νHI h pν

2.2.4 Reionization of the inhomogeneous IGM

The description of reionization in the previous section is not adequate as it
does not take into account the inhomogeneities in the IGM appropriately
(except for a clumping factor C in the effective recombination rate). To see
this, consider the post-reionization phase where we know from observations
that there exist regions of high density which are neutral; these regions
are being gradually ionized and hence one would ideally like to write an
equation similar to (2.29) for studying the post-reionization phase. Since
the ionization state depends on the density, one should have to account for
the density distribution of the IGM.
In order to proceed, first note that the volume filling factor may not
be the appropriate quantity to study for evolution of reionization because
most of the photons are consumed in regions with high densities (which
might be occupying a small fraction of volume). In other words, if we
neglect recombination for the moment,
R we have from equation (2.29) that
the volume filling factor QHII = dt ṅph /nH = nph /nH ; however, in reality
the photon to hydrogen ratio should be equal to the ionized mass fraction
M M
FHII , i.e., nph /nH = FHII . Hence, we must replace the volume filling factor
by the mass filling factor in the description of the previous section, in
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

28 Vignettes in gravitation and cosmology

particular equation (2.29) should have the form

M
dFHII ṅph M
= II − FHII Cα(T )nII
e (2.31)
dt nHII
M
One can relate FHII to the IGM density distribution by using the fact
that regions of lower densities will be ionized first, and high-density regions
will remain neutral for a longer time. The main reason for this is that the
recombination rate (which is ∝ n2H ) is higher in high-density regions where
dense gas becomes neutral very quickly. If we assume that hydrogen in all
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

regions with overdensities ∆ < ∆HII is ionized while the rest is neutral,
then the mass ionized fraction is clearly [43]
Z ∆HII
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

M
FHII ≡ F M (∆HII ) = d∆ ∆ P (∆) (2.32)
0
where P (∆) is the (volume-weighted) density distribution of the IGM. The
term describing the effective recombination rate gets contribution only from
the low density regions (high density neutral regions do not contribute) and
is then given by
Z ∆HII
α(T )nII
e d∆ ∆2 P (∆) ≡ α(T )nIIe R(∆HII ) (2.33)
0
The evolution for the mass ionized fraction is then
dF M (∆HII ) ṅph
= II − R(∆HII )α(T )nII
e (2.34)
dt nHII
The evolution equation essentially tracks the evolution of ∆HII which rises
as F M (∆HII ) increases with time (i.e., more and more high density regions
are getting ionized). Since the mean free path is determined by the high
density regions, one should be able to relate it to the value of ∆HII [43]. It
is clear that a photon will be able to travel through the low density ionized
volume
Z ∆HII
FV (∆HII ) = d∆ P (∆) (2.35)
0
before being absorbed. When a very high fraction of volume is ionized,
one can assume that the fraction of volume filled up by the high density
regions is 1 − FV , hence their size is proportional to (1 − FV )1/3 , and the
separation between them along a random line of sight will be proportional
to (1 − FV )−2/3 , which, in turn, will determine the mean free path. Then
one has
λ0
λν (a) = (2.36)
[1 − FV (∆HII )]2/3
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 29

where we can fix λ0 by comparing with low redshift observations like the
distribution of Lyman-limit systems [equation (2.23)].
The situation is slightly more complicated when the ionized regions
are in the pre-overlap stage. At this stage, a volume fraction 1 − QHII
of the universe is completely neutral (irrespective of the density), while
the remaining QHII fraction of the volume is occupied by ionized regions.
However, within this ionized volume, the high density regions (with ∆ >
∆HII ) will still be neutral. Once QHII becomes unity, all regions with ∆ <
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

∆HII are ionized and the rest are neutral; this can be thought of as the end
of reionization. The generalization of equation (2.34), appropriate for this
description is given by [43]
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

d[QHII FM (∆HII )] ṅph (z)

= II − QHII αR (T )ne R(∆HII ) (2.37)
dt nHII
Note that there are two unknowns QHII and FM (∆HII ) in equation (2.37)
which is impossible to solve without more assumptions. One assumption
which is usually made is that ∆HII does not evolve significantly with time
in the pre-overlap stage, i.e., it is equal to a critical value ∆c . This critical
density is determined from the the mean separation of the ionizing sources.
To have some idea about the value of ∆c , two arguments have been put
forward in the literature: In the first, it is argued that ∆c is determined by
the distribution of sources [43]. When the sources are very numerous, every
low-density region (void) can be ionized by sources located at the edges, and
hence the overlap of ionized regions can occur (i.e., QHII approaches unity)
when ∆c ∼ 1 is the characteristic overdensity of the thin walls separating
the voids. For rare and luminous sources, the mean separation is much
larger and hence the value of ∆c has to be higher before QHII can be close
to unity. In the second approach, it is assumed that the mean free path
is determined by the distance between collapsed objects (which manifest
themselves as Lyman-limit systems) and hence ∆c should be similar to the
typical overdensities near the boundaries of the collapsed haloes [10, 13]. It
usually turns out to be ∼ 50 − 60 depending on the density profile of the
halo. Interestingly, results do not vary considerably as ∆c is varied from
∼ 10 to ∼ 100. Once ∆c is fixed, one can follow the evolution of QHII until
it becomes unity. Following that, we enter the post-overlap stage, where
the situation is well-described by equation (2.34).
Of course, the above description is also not fully adequate as there will
be a dependence on how far the high density region is from an ionizing
source. A dense region which is very close to an ionizing source will be
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

30 Vignettes in gravitation and cosmology

ionized quite early compared to, say, a low-density region which is far away
from luminous sources. However, it has been found that the above descrip-
tion gives a reasonable analytical description of the reionization process,
particularly for the post-reionization phase. The main advantages in this
approach are (i) it takes into account both the pre-overlap and post-overlap
phases under a single formalism, (ii) once we have some form for the IGM
density distribution P (∆), we can calculate the clumping factor and the
effective recombination rate self-consistently without introducing any extra
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

parameter; in addition we can also compute the mean free path using one
single parameter (λ0 , which can be fixed by comparing with low-redshift
observations).
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

2.3 Modelling of reionization

Given the formalism we have outlined in the previous section, we can now go
forward and discuss some other details involved in modelling reionization.

2.3.1 Reionization sources

The main uncertainty in any reionization model is to identify the sources.
The most natural sources which have been observed to produce ionizing
photons are the star-forming galaxies and quasars. Among these, the quasar
population is seen to decrease rapidly at z > 3 and there is still no evidence
of a significant population at higher redshifts. Hence, the most common
sources studied in this area are the galaxies.
The subject area of formation of galaxies is quite involved in itself deal-
ing with formation of non-linear structures (haloes and filaments), gas cool-
ing and generation of radiation from stars.

2.3.1.1 Mass function of collapsed haloes

The crucial ingredient for galaxy formation is the collapse and virialization
of dark matter haloes. This can be adequately described by the Press-
Schechter formalism for most purposes. It can be shown that the number
density of collapsed objects per unit comoving volume (which is physical
volume divided by a3 ) within a mass range (M, M + dM ) at an epoch t is
given by [50]
r
∂n(M, t) 2 −ν 2 /2 ρ̄m d ln σ(M ) ν
dM = − e dM, (2.38)
∂M π M d ln M M
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 31

where ρ̄m is the comoving density of dark matter, σ(M ) is defined as the
rms mass fluctuation at a mass scale M at z = 0, ν ≡ δc /[D(t)σ(M )],
D(t) is the growth factor for linear dark matter perturbations and δc is
the critical overdensity for collapse, usually taken to be equal to 1.69 for a
matter-dominated flat universe (Ωm = 1). This formalism can be extended
to calculate the comoving number density of collapsed objects having mass
in the range (M, M + dM ), which are formed within the time interval
(tform , tform + dtform ) and observed at a later time t [56, 11] is given by
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

∂ 2 n(M, t; tform ) ∂ 2 n(M, tform )

dM dtform =
∂M ∂tform ∂M ∂tform form
× psurv [t|tform ] dM dtform
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

(2.39)
2
where ∂ n(M, tform )/∂M ∂tform |form is the formation rate of haloes at tform
and psurv [t|tform ] is their survival probability till time t. Assuming that
haloes are destroyed only when they merge to a halo of higher mass, both
these quantities can be calculated from the merger rates of haloes. The
merger rates can be calculated using detailed properties of Gaussian random
field. The quantities can also be calculated in a more simplistic manner by
assuming that the merger probability is scale invariant; in that case [56]
∂ 2 n(M, t) ∂ 2 n(M, t) ∂n(M, t) Ḋ
= +
∂M ∂t form ∂M ∂t ∂M D
∂n(M, t) 2 Ḋ
= ν (2.40)
∂M D
and
D(t)
psurv [t|tform ] = . (2.41)
D(tform )

2.3.1.2 Star formation rate

If these dark matter haloes are massive enough to form huge potential wells,
the baryonic gas will simply fall into those wells. As the gas begins to settle
into the dark matter haloes, mergers will heat it up to the virial temperature
via shocks. However, to form galaxies, the gas has to dissipate its thermal
energy and cool. If the gas contains only atomic hydrogen, it is unable
to cool at temperatures lower than 104 K because the atomic hydrogen
recombines and cannot be ionized by collisions. The gas can cool effectively
for much lower temperatures in the presence of molecules – however, it is
not straightforward to estimate the amount of molecules present in the gas
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

32 Vignettes in gravitation and cosmology

at high redshifts. Hence the lower mass cutoff for the haloes which can host
star formation will be decided by the cooling efficiency of the baryons.
Let Ṁ∗ (M, t, tform ) denote the rate of star formation at time t within a
halo of mass M which has formed at tform . Then we can write the cosmic
SFR per unit volume at a time t,
Z t Z ∞
1
ρ̇∗ (t) = 3 dtform dM 0 Ṁ∗ (M 0 , t, tform )
a (t) 0 Mmin (t)
∂ 2 n(M 0 , t; tform )
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

× (2.42)
∂M 0 ∂tform
where the a−3 is included to covert from comoving to physical volume. The
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

lower mass cutoff Mmin (t) at a given epoch is decided by the cooling criteria
as explained above. However, once reionization starts and regions are re-
heated by photoheating, the value of Mmin (t) is set by the photoionization
temperature ' 104 K. This can further suppress star formation in low mass
haloes and is known as radiative feedback. We shall discuss this later in
this section.
The form of Ṁ∗ (M, t, tform ) contains information about various cooling
and star-forming processes and hence is quite complex to deal with. It
can be obtained from semi-analytical modelling of galaxy formation [60] or
constrained from observations of galaxy luminosity function [54]. A very
simple assumption that is usually made for modelling reionization is that
the duration of star formation is much less than the Hubble time H −1 (t)
which is motivated by the fact that most of the ionizing radiation is pro-
duced by hot stars which have shorter lifetime. In that case, Ṁ∗ (M, t, tform )
can be approximated as
ρ̄b
Ṁ∗ (M, t, tform ) ≈ M∗ δD (t − tform ) = f∗ M δD (t − tform ) (2.43)
ρ̄m
where ρ̄b /ρ̄m M is the mass of baryons within the halo and f∗ is the fraction
of baryonic mass which has been converted into stars. The cosmic SFR per
comoving volume is then
Z ∞
1 ρ̄b ∂ 2 n(M 0 , t)
ρ̇∗ (t) = 3 dM 0 f∗ M 0 (2.44)
a (t) ρ̄m Mmin (t) ∂M 0 ∂t form
One should keep in mind that many details of the star-formation process
have been encoded within a single parameter f∗ . This should, in principle,
be a function of both halo mass M and time t. However, it is not clear at
all what the exact dependencies should be. Given such uncertainties, it is
usual to take it as a constant.
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 33

In addition, one finds that the merger rate of haloes at high redshifts
is much less than the formation rate (which follows if ν  1), hence
∂ 2 n(M, t)/∂M ∂t|form ≈ ∂ 2 n(M, t)/∂M ∂t. Then, one can write the SFR
in terms of the fraction of collapsed mass in haloes more massive than
Mmin (t)
Z ∞
1 ∂n(M 0 , t)
fcoll (t) = dM 0 M 0
ρ̄m Mmin (t) ∂M 0
 
δc
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

= erfc √ (2.45)
2 D(t)σ(Mmin )
as
ρ̄b dfcoll (t)
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

ρ̇∗ (t) = f∗ 3 (2.46)

a (t) dt

2.3.1.3 Production of ionizing photons

Given the SFR, we can calculate the emissivity of galaxies, or equivalently
the rate of ionizing photons in the IGM per unit volume per unit frequency
range:
 
dNν
ṅν,ph (t) = fesc ρ̇∗ (t) (2.47)
dM∗
where dNν /dM∗ gives the number of photons emitted per frequency range
per unit mass of stars and fesc is the fraction of ionizing photons which
escape from the star forming haloes into the IGM. The emissivity is simply
ν = hp ν ṅν,ph .
Given the spectra of stars of different masses in a galaxy, and their
Initial Mass Function (IMF), dNν /dM∗ can be computed in a straightfor-
ward way using “population synthesis” codes [35, 7]. The IMF and spectra
will depend on the details of star formation (burst formation or continu-
ous) and metallicity. In fact, it is possible that there are more than one
population of stellar sources which have different dNν /dM∗ . For example,
there are strong indications, both from numerical simulations and analyti-
cal arguments [6], that the first generation stars were metal-free, and hence
massive, with a very different kind of IMF and spectra than the stars we
observe today [57]; they are known as the Population III (PopIII) stars.
A fraction of photons produced in a galaxy would be consumed in ion-
izing the neutral matter within the galaxy itself. Hence only a fraction
of photons escape into the IGM and is available for reionization, which is
encoded in the parameter fesc . This parameter is again not very well mod-
elled and its observed value is also quite uncertain; typical values assumed
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

34 Vignettes in gravitation and cosmology

are ∼ 0.1. It is most likely fesc , like f∗ , is also a function of halo mass
and the time of halo formation, however since the dependences are not well
understood, it is taken to be a constant.
The total number of ionizing photons is then obtained by integrating
the above quantity over all energies above the ionization threshold:
Z ∞
dfcoll (t)
ṅph (t) = dν ṅν,ph (t) = Nion nb (2.48)
νHI dt
where nb is the total baryonic number density in the IGM (equal to nH if
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

we ignore the presence of helium) and

Z ∞  
dNν
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Nion ≡ f∗ fesc mp dν (2.49)

νHI dM∗
is the number of photons entering the IGM per baryon in collapsed objects.
In case there are more than one population of stars, one has to use different
values of Nion for the different populations.

2.3.1.4 Feedback processes

The moment there is formation of stars and other luminous bodies, they
start to affect the subsequent formation of structures – this is known as
feedback [17]. The process is intrinsically non-linear and hence quite com-
plex to model. The feedback processes can be categorized roughly into
three categories.
The first of them is the radiative feedback which is associated with the
radiation from first stars which can heat up the medium and can photoionize
atoms and/or photodissociate molecules. Once the first galaxies form stars,
their radiation will ionize and heat the surrounding medium, increasing the
mass scale (often referred to as the filtering mass [27]) above which baryons
can collapse into haloes within those regions. The minimum mass of haloes
which are able to cool is thus much higher in ionized regions than in neutral
ones. Since the IGM is multi-phase in the pre-overlap phase, one needs to
take into account the heating of the ionized regions right from the beginning
of reionization. In principle, this can be done self-consistently from the
evolution of the temperature of the ionized region in equation (2.3).
The low mass haloes can be subjected to mechanical feedback too, which
is mostly due to energy injection via supernova explosion and winds. This
can expel the gas from the halo and suppress star formation. As in the
case for radiative feedback, one can parametrize this through the minimum
mass parameter Mmin (t).
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 35

Finally, we also have chemical feedback where the stars expel metals into
the medium and hence change the chemical composition. This would mean
that the subsequent formation of stars could be in a completely different
environment and hence the nature of stars would be highly different.

2.3.1.5 Quasars
Besides the stellar sources, a population of accreting black holes or quasars
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

are also known to produce significant amount of ionizing radiation. Hence

it is possible that they too have contributed to reionization. The fraction
of their contribution would depend on the number of quasars produced at
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

a particular redshift. Observationally, the luminosity function of quasars

is quite well-probed till a redshift z ∼ 6 [53]. It turns out that the popu-
lation peaks around z ≈ 3 and decreases for higher redshifts. Hence their
contribution at higher redshifts is highly debated.
The difference between reionization by stellar sources and quasars lie
in the fact that quasars produce significant number of high energy pho-
tons compared to stars. This would imply that quasars can contribute
significantly to double-reionization of helium (which requires photons with
energies > 54.4 eV, not seen in galaxies). In addition, quasars produce
significant amount of X-ray radiation. Since the absorption cross section
of neutral hydrogen varies with frequency approximately as ν −3 , the mean
free path for photons with high energies would be very large. A simple
calculation will show that for photons with energies above 100–200 eV,
the mean free path would be larger than the typical separation between
collapsed structures [39] (the details would depend upon the redshift and
exact description of collapsed haloes). These photons would not be asso-
ciated with any particular source at the moment when they are absorbed,
and thus would ionize the IGM in a more homogeneous manner (as opposed
to the overlapping bubble picture for UV sources).

2.3.2 Illustration of a semi-analytical model

The physics described above in the preceding sections can all be combined
to construct a semi-analytical model for studying the thermal and ionization
history of the IGM. We shall give an explicit example of one such model [13,
15] whose main features are: The model accounts for IGM inhomogeneities
by adopting a lognormal distribution for P (∆); reionization is said to be
complete once all the low-density regions (say, with overdensities ∆ < ∆c ∼
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

36 Vignettes in gravitation and cosmology

Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Fig. 2.2 Comparison of analytical model predictions with observations for the best-fit
model. The different panels indicate: (a) The volume-averaged neutral hydrogen fraction
xHI , with observational lower limit from QSO absorption lines at z = 6 and upper limit
from Lyα emitters at z = 6.5 (shown with arrows). In addition, the ionized fraction
xe is shown by the dashed line. (b) SFR ρ̇∗ for different stellar populations. (c) The
number of source counts above a given redshift, with the observational upper limit from
NICMOS HUDF is shown by the arrow. The contribution to the source count is zero at
low redshifts because of the J-dropout selection criterion. (d) Electron scattering optical
depth, with observational constraint from WMAP 3-year data release. (e) Lyα effective
optical depth. (f) Lyβ effective optical depth. (g) Evolution of Lyman-limit systems. (h)
Photoionization rates for hydrogen, with estimates from numerical simulations (shown
by points with error-bars; [4]). (i) Temperature of the mean density IGM.

60) are ionized. The ionization and thermal histories of neutral, HII and
HeIII regions are followed simultaneously and self-consistently, treating the
IGM as a multi-phase medium. Three types of reionization sources have
been assumed: (i) metal-free (i.e. PopIII) stars having a Salpeter IMF in
the mass range 1 − 100M : they dominate the photoionization rate at high
redshifts; (ii) PopII stars with sub-solar metallicities also having a Salpeter
IMF in the mass range 1−100M ; (iii) quasars, which are significant sources
of hard photons at z . 6; they have negligible effects on the IGM at higher
redshifts.
As discussed earlier, reionization is accompanied by various feedback
processes, which can affect subsequent star formation. In this model, ra-
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 37

diative feedback is computed self-consistently from the evolution of the

thermal properties of the IGM. Furthermore, the chemical feedback induc-
ing the PopIII → PopII transition is implemented using a merger-tree “ge-
netic” approach which determines the termination of PopIII star formation
in a metal-enriched halo [59].
The predictions of the model are compared with a wide range of obser-
vational data sets, namely, (i) redshift evolution of Lyman-limit absorption
systems dNLL /dz [66], (ii) the effective optical depths τeff ≡ − ln F for
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

Lyα and Lyβ absorption

R in the IGM [61], (iii) electron scattering optical
depth τel = σT c dt ne (where σT is the Thomson scattering cross section)
as measured from CMBR experiments [63], (iv) temperature of the mean
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

intergalactic gas [58], (v) cosmic star formation history ρ̇∗ [44], and (vi)
source number counts at z ≈ 10 from NICMOS HUDF [5].
The data constrain the reionization scenario quite tightly. We find that
hydrogen reionization starts at z ≈ 15 driven by metal-free (PopIII) stars,
and it is 90 per cent complete by z ≈ 8. After a rapid initial phase,
the growth of the volume filled by ionized regions slows down at z . 10
due to the combined action of chemical and radiative feedback, making
reionization a considerably extended process completing only at z ≈ 6.
The number of photons per hydrogen at the end of reionization at z ≈ 6 is
only a few, which implies that reionization occurred in a “photon-starved”
manner [3].

2.4 Current status and future

In this section, let us review the current status of various approaches to

studying reionization and their future prospects.

2.4.1 Simulations
Though the analytical studies mentioned above allow us to develop a good
understanding of the different processes involved in reionization, they can
take into account the physical processes only in some approximate sense.
In fact, a detailed and complete description of reionization would require
locating the ionizing sources, resolving the inhomogeneities in the IGM,
following the scattering processes through detailed radiative transfer, and
so on. Numerical simulations, in spite of their limitations, have been of
immense importance in these areas.
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

38 Vignettes in gravitation and cosmology

The ionizing photons during early stages of reionization mostly originate

from smaller haloes which are far more numerous than the larger galaxies
at high redshifts. The need to resolve such small structures requires the
simulation boxes to have high enough resolution. On the other hand, these
ionizing sources were strongly clustered at high redshifts and, as a conse-
quence, the ionized regions they created are expected to overlap and grow
to very large sizes, reaching upto tens of Mpc. As already discussed, the
many orders of magnitude difference between these length scales demand
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

extremely high computing power from any simulations designed to study

early structure formation from the point of view of reionization.
To simulate reionization, one usually runs a N-body simulation (either
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

dark matter only or including baryons) to generate the large-scale density

field, identifies haloes within the density field and assign ionizing photons
to the haloes using a assumption like equation (2.48). It turns out that
the most difficult step is to solve the radiative transfer equation and study
the growth of ionized regions. In principle, one could solve equation (2.4)
directly for the intensity at every point in the seven-dimensional (t, x, n, ν)
space, given the absorption coefficient and the emissivity. However, the
high dimensionality of the problem makes the solution of the complete
radiative transfer equation well beyond our capabilities, particularly since
we do not have any obvious symmetries in the problem and often need high
spatial and angular resolution in cosmological simulations. At present,
most of the simulations do not have enough resolution to reliably identify
the low mass ∼ 108 M sources which were probably responsible for early
stages of reionization. Also, there are difficulties in resolving the small scale
structures which contribute significantly to the clumpiness in the IGM and
hence extend the reionization process.

2.4.2 Various observational probes

Finally, we review certain observations which shape our understanding of
reionization.

2.4.2.1 Absorption spectra of high redshift sources

We have already discussed that the primary evidence for reionization comes
from absorption spectra of quasars (Lyα forest) at z < 6. We have also dis-
cussed that the effective optical depth of Lyα photons becomes significantly
large at z & 6 implying regions with high transmission in the Lyα forest
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 39

becoming rare at high redshifts [20, 68]. Therefore the standard methods
of analyzing the Lyα forest (like the probability distribution function and
power spectrum) are not very effective. Amongst alternate methods, one
can use the the distribution of dark gaps [18, 62] which are defined as con-
tiguous regions of the spectrum having an optical depth above a threshold
value [62, 20]. It has been found that the current observations constrain
the neutral hydrogen fraction xHI < 0.36 at z = 6.3 [25]. It is expected that
the SDSS and Palomar-Quest survey [37] would detect ∼ 30 QSOs at these
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

redshifts within the next few years and hence we expect robust conclusions
from such studies in very near future.
Like quasars, one can also use absorption spectra of other high redshift
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

energetic sources like gamma ray bursts (GRBs) and supernovae. In fact,
analyses using the damping wing effects of the Voigt profile have been
already performed on the GRB detected at a redshift z = 6.3, and the wing
shape is well-fit by a neutral fraction xHI < 0.17 [67]. The dark gap width
distribution gives a similar constraint xHI = (6.4±0.3)×10−5 [26]. In order
to obtain more stringent limits on reionization, it is important to increase
the sample size of z > 6 GRBs.
In addition to hydrogen reionization, the Lyα forest in the quasar ab-
sorption lines at z ≈ 3 can also be used for studying reionization of singly
ionized helium to doubly ionized state (the reionization of neutral helium to
singly ionized state follows hydrogen for almost all types of sources). The
helium reionization coincides with the rise in quasar population at z ∼ 3
and it effects the thermal history of the IGM at these redshifts. However,
there are various aspects of the observation that are not well understood
and requires much detailed modelling of helium reionization.

2.4.2.2 CMBR observations

As we have discussed already, the first evidence for an early reionization
epoch came from the CMBR polarization data. This data is going to be
much more precise in future with experiments like PLANCK, and is ex-
pected to improve the constraints on τel . With improved statistical errors,
it might be possible to distinguish between different evolutions of the ion-
ized fraction, particularly with E-mode polarization auto-correlation, as is
found from theoretical calculations [31, 8]. An alternative option to probe
reionization through CMBR is through the small scale observations of tem-
perature anisotropies. It has been well known that the scattering of the
CMBR photons by the bulk motion of the electrons in clusters gives rise
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

40 Vignettes in gravitation and cosmology

to a signal at large multipoles ` ∼ 1000, known as the kinetic Sunyaev Zel-

dovich (SZ) effect. Such a signal can also originate from the fluctuations in
the distribution of free electrons arising from cosmic reionization. It turns
out that for reionization, the signal is dominated by the patchiness in the ne -
distribution. Now, in most scenarios of reionization, it is expected that the
distribution of neutral hydrogen would be quite patchy in the pre-overlap
era, with the ionized hydrogen mostly contained within isolated bubbles.
The amplitude of this signal is significant around ` ∼ 1000 and is usually
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

comparable to or greater than the signal arising from standard kinetic SZ

effect. Theoretical estimates of the signal have been performed for various
reionization scenarios, and it has been predicted that the experiment can
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

be used for constraining reionization history [55, 42]. Also, it is possible to

have an idea about the nature of reionization sources, as the signal from
UV sources, X-ray sources and decaying particles are quite different. With
multi-frequency experiments like Atacama Cosmology Telescope (ACT)2
and South Pole Telescope (SPT)3 coming up in near future, this promises
to put strong constraints on the reionization scenarios.

2.4.2.3 Lyα emitters

In recent years, a number of groups have studied star-forming galaxies at
z ∼ 6 − 7, and measurements of the Lyα emission line luminosity function
evolution provide another useful observational constraint [40, 64]. While the
quasar absorption spectra probe the neutral hydrogen fraction regime xHI ≤
0.01, this method is sensitive to the range xHI ∼ 0.1 − 1.0. Lyα emission
from galaxies is expected to be suppressed at redshifts beyond reionization
because of the absorption due to neutral hydrogen, which clearly affects the
evolution of the luminosity function of such Lyα emitters at high redshifts
[30, 40, 22]. Thus a comparison of the luminosity functions at different
redshifts could be used for constraining the reionization. Through a simple
analysis, it was found that the luminosity functions at z = 5.7 and z = 6.5
are statistically consistent with one another thus implying that reionization
was largely complete at z ≈ 6.5. More sophisticated calculations on the
evolution of the luminosity function of Lyα emitters [40, 24, 30] suggest
that the neutral fraction of hydrogen at z = 6.5 should be less than 50
per cent [41]. Unfortunately, the analysis of the Lyα emitters at high
redshifts is complicated by various factors like the velocity of the sources
2 http://www.hep.upenn.edu/act/
3 http://spt.uchicago.edu/
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 41

with respect to the surrounding IGM, the density distribution and the size
of ionized regions around the sources and the clustering of sources. It is
thus extremely important to have detailed models of Lyα emitting galaxies
in order to use them for constraining reionization.

2.4.2.4 Sources of reionization

As we discussed earlier, a major challenge in our understanding of reioniza-
tion depends on our knowledge of the sources, particularly at high redshifts.
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

As we understand at present, neither the bright z > 6 quasars discovered by

the SDSS group [21] nor the faint AGN detected in X-ray observations [1]
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

produce enough photons to reionize the IGM. The discovery of star-forming

galaxies at z > 6.5 [32, 34, 33] has resulted in speculation that early galax-
ies produce bulk of the ionizing photons for reionization. Unfortunately,
there are significant uncertainties in constraining the amount of ionizing
radiation at these redshifts because the bulk of ionizing photons could be
produced by faint sources which are beyond the present sensitivities. In
fact, some models have predicted that the z > 7 sources identified in these
surveys are relatively massive (M ≈ 109 M ) and rare objects which are
only marginally (≈ 1%) contributing to the reionization photon budget [16].
A much better prospect of detecting these sources would be through the
Ultra-Deep Imaging Survey using the future telescope JWST.

2.4.2.5 21cm observations

Perhaps the most promising prospect of detecting the fluctuations in the
neutral hydrogen density during the (pre-)reionization era is through the
21 cm emission experiments [23], some of which are already taking data
(GMRT 4 , 21CMA 5 ), and some are expected in future (MWA 6 , LOFAR
7
, SKA 8 ). The basic principle which is central to these experiments is the
neutral hydrogen hyperfine transition line at a rest wavelength of 21 cm.
This line, when redshifted, is observable in radio frequencies (∼ 150 MHz
for z ∼ 10) as a brightness temperature:

TS − TCMB 3c3 Ā10 nHI (z, n̂)(1 + z)3

δTb (z, n̂) = (2.50)
1+z 16kboltz ν02 TS H(z)
4 http://www.gmrt.ncra.tifr.res.in
5 http://web.phys.cmu.edu/∼past/
6 http://www.haystack.mit.edu/arrays/MWA
7 http://www.lofar.org
8 http://www.skatelescope.org/
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

42 Vignettes in gravitation and cosmology

where TS is the spin temperature of the gas, TCMB = 2.76(1 + z) K is the

CMBR temperature, A10 is the Einstein coefficient and ν0 = 1420 MHz is
the rest frequency of the hyperfine line.
The observability of this brightness temperature against the CMB back-
ground will depend on the relative values of TS and TCMB . Depending on
which processes dominate at different epochs, TS will couple either to ra-
diation (TCMB ) or to matter (determined by the kinetic temperature Tk )
[51]. Almost in all models of reionization, the most interesting phase for
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

observing the 21 cm radiation is 6 . z . 20. This is the phase where the

IGM is suitably heated to temperatures much higher than CMB (mostly
due to X-ray heating [9]) thus making it observable in emission. In that
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

case, we have δTb ∝ nHI /nH , which means that the observations would
directly probe the neutral hydrogen density in the Universe. Furthermore,
this is the era when the bubble-overlapping phase is most active, and there
is substantial neutral hydrogen to generate a strong enough signal. At low
redshifts, after the IGM is reionized, nHI falls by orders of magnitude and
the 21 cm signal vanishes except in the high density neutral regions. Since
the observations directly probe the neutral hydrogen density, one can use
it to probe the detailed topology of the ionized regions in the pre-overlap
phase. It is therefore essential to model the clustering of the sources accu-
rately so as to predict the reionization topology.
There are essentially two complementary approaches to studying reion-
ization using 21 cm signal. The first one is through global statistical prop-
erties of the neutral hydrogen signal, like the power spectrum. The second
one is to directly detect the ionized bubbles around sources, either through
blind surveys or via targetted observations.
The major difficulty in obtaining the cosmological signal from these
experiments is that it is expected to be only a small contribution buried
deep in the emission from other astrophysical sources (foregrounds) and in
the system noise. It is thus a big challenge to detect the signal which is
of cosmological importance from the other contributions that are orders of
magnitude larger. Once such challenges are dealt with, this probe will be
the strongest probe for not only reionization, but of the matter distribution
at very small scales during the dark ages.
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 43

2.5 Concluding remarks

We have discussed the analytical approaches to model different aspects of

reionization which will help in understanding the most relevant physical
processes. In an explicit example, we have shown how to apply this formal-
ism for constraining the reionization history using a variety of observational
data. These constraints imply that reionization is an extended process over
a redshift range 15 > z > 6. It is most likely driven by the first sources
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

which form in small mass haloes. However, there are still uncertainties
about the exact nature of these sources and the detailed topology of ion-
ized regions. Such details are going to be addressed in near future as new
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

observations, both space-borne and ground-based, are likely to settle these

long-standing questions. From the theoretical point of view, it is thereby
important to develop detailed analytical and numerical models to extract
the maximum information about the physical processes relevant for reion-
ization out of the expected large and complex data sets.

References

[1] Barger, A. J., Cowie, L. L., Capak, P., Alexander, D. M., Bauer, F. E.,
Brandt, W. N., Garmire, G. P. and Hornschemeier, A. E. (2003). Very
High Redshift X-Ray-selected Active Galactic Nuclei in the Chandra
Deep Field-North, ApJ 584, pp. L61–L64.
[2] Barkana, R. and Loeb, A. (2001). In the beginning: the first sources of
light and the reionization of the universe, Phys. Rep. 349, pp. 125–238,
doi:10.1016/S0370-1573(01)00019-9.
[3] Bolton, J. S. and Haehnelt, M. G. (2007). The observed ionization
rate of the intergalactic medium and the ionizing emissivity at z >=
5: evidence for a photon-starved and extended epoch of reionization,
MNRAS 382, pp. 325–341, doi:10.1111/j.1365-2966.2007.12372.x.
[4] Bolton, J. S., Haehnelt, M. G., Viel, M. and Springel, V. (2005). The
Lyman α forest opacity and the metagalactic hydrogen ionization rate
at z˜ 2-4, MNRAS 357, pp. 1178–1188, doi:10.1111/j.1365-2966.2005.
08704.x.
[5] Bouwens, R. J., Illingworth, G. D., Thompson, R. I. and Franx, M.
(2005). Constraints on z˜10 Galaxies from the Deepest Hubble Space
Telescope NICMOS Fields, ApJ 624, pp. L5–L8.
[6] Bromm, V., Kudritzki, R. P. and Loeb, A. (2001). Generic Spectrum
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

44 Vignettes in gravitation and cosmology

and Ionization Efficiency of a Heavy Initial Mass Function for the First
Stars, ApJ 552, pp. 464–472, doi:10.1086/320549.
[7] Bruzual, G. and Charlot, S. (2003). Stellar population synthesis at the
resolution of 2003, MNRAS 344, pp. 1000–1028.
[8] Burigana, C., Popa, L. A., Salvaterra, R., Schneider, R., Choudhury,
T. R. and Ferrara, A. (2008). Cosmic microwave background polar-
ization constraints on radiative feedback, MNRAS 385, pp. 404–410,
doi:10.1111/j.1365-2966.2008.12845.x.
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

[9] Chen, X. and Miralda-Escudé, J. (2004). The Spin-Kinetic Temper-

ature Coupling and the Heating Rate due to Lyα Scattering before
Reionization: Predictions for 21 Centimeter Emission and Absorption,
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

ApJ 602, pp. 1–11.

[10] Chiu, W. A., Fan, X. and Ostriker, J. P. (2003). Combining Wilkinson
Microwave Anisotropy Probe and Sloan Digital Sky Survey Quasar
Data on Reionization Constrains Cosmological Parameters and Star
Formation Efficiency, ApJ 599, pp. 759–772, doi:10.1086/379318.
[11] Chiu, W. A. and Ostriker, J. P. (2000). A Semianalytic Model for
Cosmological Reheating and Reionization Due to the Gravitational
Collapse of Structure, ApJ 534, pp. 507–532.
[12] Choudhury, T. R. (2009). Analytical Models of the Intergalactic
Medium and Reionization, Curr. Sci. 97, pp. 841–857.
[13] Choudhury, T. R. and Ferrara, A. (2005). Experimental constraints
on self-consistent reionization models, MNRAS 361, pp. 577–594.
[14] Choudhury, T. R. and Ferrara, A. (2006). Physics of cosmic reion-
ization, in R. Fabbri (ed.), Cosmic Polarization (Research Signpost,
India), p. 205.
[15] Choudhury, T. R. and Ferrara, A. (2006). Updating reionization sce-
narios after recent data, MNRAS 371, pp. L55–L59.
[16] Choudhury, T. R. and Ferrara, A. (2007). Searching for the reioniza-
tion sources, MNRAS 380, pp. L6–L10.
[17] Ciardi, B. and Ferrara, A. (2005). The First Cosmic Structures and
Their Effects, Space Sci. Rev. 116, pp. 625–705.
[18] Croft, R. A. C. (1998). Characterization of Lyman Alpha Spectra and
Predictions of Structure Formation Models: A Flux Statistics Ap-
proach, in Olinto A. V., Frieman J. A., Schramm D. N. ed., Eighteenth
Texas Symposium on Relativistic Astrophysics. (World Scientific, River
Edge, N. J.), pp. 664–+.
[19] Fan, X., Carilli, C. L. and Keating, B. (2006). Observational Con-
straints on Cosmic Reionization, ARA&A 44, pp. 415–462.
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 45

[20] Fan, X., Strauss, M. A., Becker, R. H., White, R. L., Gunn, J. E.,
Knapp, G. R., Richards, G. T., Schneider, D. P., Brinkmann, J. and
Fukugita, M. (2006). Constraining the Evolution of the Ionizing Back-
ground and the Epoch of Reionization with z˜6 Quasars. II. A Sample
of 19 Quasars, AJ 132, pp. 117–136.
[21] Fan, X., Strauss, M. A., Richards, G. T., Hennawi, J. F., Becker,
R. H., White, R. L., Diamond-Stanic, A. M., onley, J. L. D., Jiang, L.,
Kim, J. S., Vestergaard, M., Young, J. E., Gunn, J. E., Lupton, R. H.,
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

Knapp, G. R., Schneider, D. P., Brandt, W. N., Bahcall, N. A., Bar-

entine, J. C., Brinkmann, J., Brewington, H. J., ukugita, M. F., Har-
vanek, M., Kleinman, S. J., Krzesinski, J., Long, D., eilsen, E. H. N.,
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Jr., Nitta, A., Snedden, S. A. and Voges, W. (2005). Preprint: astro-

ph/0512080.
[22] Furlanetto, S. R., Hernquist, L. and Zaldarriaga, M. (2004). Constrain-
ing the topology of reionization through Lyα absorption, MNRAS 354,
pp. 695–707.
[23] Furlanetto, S. R., Oh, S. P. and Briggs, F. H. (2006). Cosmology at
low frequencies: The 21 cm transition and the high-redshift Universe,
Phys. Rep. 433, pp. 181–301.
[24] Furlanetto, S. R., Zaldarriaga, M. and Hernquist, L. (2006). The effects
of reionization on Lyα galaxy surveys, MNRAS 365, pp. 1012–1020.
[25] Gallerani, S., Ferrara, A., Fan, X. and Choudhury, T. R. (2008).
Glimpsing through the high-redshift neutral hydrogen fog, MNRAS
386, pp. 359–369, doi:10.1111/j.1365-2966.2008.13029.x.
[26] Gallerani, S., Salvaterra, R., Ferrara, A. and Choudhury, T. R. (2008).
Testing reionization with gamma-ray burst absorption spectra, MN-
RAS 388, pp. L84–L88, doi:10.1111/j.1745-3933.2008.00504.x.
[27] Gnedin, N. Y. and Hui, L. (1998). Probing the Universe with the Lyal-
pha forest - I. Hydrodynamics of the low-density intergalactic medium,
MNRAS 296, pp. 44–55.
[28] Gunn, J. E. and Peterson, B. A. (1965). On the Density of Neutral
Hydrogen in Intergalactic Space. ApJ 142, pp. 1633–1641.
[29] Haardt, F. and Madau, P. (1996). Radiative Transfer in a Clumpy Uni-
verse. II. The Ultraviolet Extragalactic Background, ApJ 461, p. 20.
[30] Haiman, Z. and Cen, R. (2005). Constraining Reionization with the
Evolution of the Luminosity Function of Lyα Emitting Galaxies, ApJ
623, pp. 627–631.
[31] Holder, G. P., Haiman, Z., Kaplinghat, M. and Knox, L. (2003). The
Reionization History at High Redshifts. II. Estimating the Optical
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

46 Vignettes in gravitation and cosmology

Depth to Thomson Scattering from Cosmic Microwave Background

Polarization, ApJ 595, pp. 13–18.
[32] Hu, E. M., Cowie, L. L., McMahon, R. G., Capak, P., Iwamuro, F.,
Kneib, J.-P., Maihara, T. and Motohara, K. (2002). A Redshift z=6.56
Galaxy behind the Cluster Abell 370, ApJ 568, pp. L75–L79.
[33] Kneib, J.-P., Ellis, R. S., Santos, M. R. and Richard, J. (2004). A
Probable z˜7 Galaxy Strongly Lensed by the Rich Cluster A2218: Ex-
ploring the Dark Ages, ApJ 607, pp. 697–703.
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

[34] Kodaira, K., Taniguchi, Y., Kashikawa, N., Kaifu, N., Ando, H.,
Karoji, H., Ajiki, M., Akiyama, M., Aoki, K., Doi, M., Fujita, S. S., Fu-
rusawa, H., Hayashino, T., Imanishi, M., Iwamuro, F., Iye, M., Kawa-
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

bata, K. S., Kobayashi, N., Kodama, T., Komiyama, Y., Kosugi, G.,
Matsuda, Y., Miyazaki, S., Mizumoto, Y., Motohara, K., Murayama,
T., Nagao, T., Nariai, K., Ohta, K., Ohyama, Y., Okamura, S., Ouchi,
M., Sasaki, T., Sekiguchi, K., Shimasaku, K., Shioya, Y., Takata, T.,
Tamura, H., Terada, H., Umemura, M., Usuda, T., Yagi, M., Yamada,
T., Yasuda, N. and Yoshida, M. (2003). The Discovery of Two Lyman
α Emitters beyond Redshift 6 in the Subaru Deep Field, PASJ 55, pp.
L17–L21.
[35] Leitherer, C., Schaerer, D., Goldader, J. D., Delgado, R. M. G.,
Robert, C., Kune, D. F., de Mello, D. F., Devost, D. and Heckman,
T. M. (1999). Starburst99: Synthesis Models for Galaxies with Active
Star Formation, ApJS 123, pp. 3–40.
[36] Loeb, A. and Barkana, R. (2001). The Reionization of the Universe by
the First Stars and Quasars, ARA&A 39, p. 19.
[37] López-Cruz, O., Djorgovski, S. G., Carrasco, L., Baltay, C., Recil-
las, E., Mayya, Y. D., Escobedo, G. A., Castillo, E., Mahabal, A.,
Thompson, D. J., Rabinowitz, D., Bauer, A., Graham, M., Williams,
R., Rengstorf, A., Brunner, R., Andrews, P., Ellman, N., Duffau, S.,
Lauer, R., Bogosavljevic, M., Musser, J., Mufson, S. and Gebhard, M.
(2005). A Search for z ˜ 6 QSOs in the Palomar-Quest Sky Survey,
in A. M. Hidalgo-Gámez, J. J. González, J. M. Rodrı́guez Espinosa
and S. Torres-Peimbert (eds.), Revista Mexicana de Astronomia y As-
trofisica Conference Series (Instituto de Astronoma, Universidad Na-
cional Autnoma de Mxico, Mexico), pp. 164–169.
[38] Madau, P., Haardt, F. and Rees, M. J. (1999). Radiative Transfer in
a Clumpy Universe. III. The Nature of Cosmological Ionizing Sources,
ApJ 514, p. 648.
[39] Madau, P., Rees, M. J., Volonteri, M., Haardt, F. and Oh, S. P. (2004).
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 47

Early Reionization by Miniquasars, ApJ 604, pp. 484–494.

[40] Malhotra, S. and Rhoads, J. E. (2004). Luminosity Functions of Lyα
Emitters at Redshifts z=6.5 and z=5.7: Evidence against Reionization
at z<=6.5, ApJ 617, pp. L5–L8.
[41] Malhotra, S. and Rhoads, J. E. (2006). The Volume Fraction of Ionized
Intergalactic Gas at Redshift z=6.5, ApJ 647, pp. L95–L98, doi:10.
1086/506983.
[42] McQuinn, M., Furlanetto, S. R., Hernquist, L., Zahn, O. and Zaldar-
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

riaga, M. (2005). The Kinetic Sunyaev-Zel’dovich Effect from Reion-

ization, ApJ 630, pp. 643–656.
[43] Miralda-Escudé, J., Haehnelt, M. and Rees, M. J. (2000). Reionization
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

of the Inhomogeneous Universe, ApJ 530, pp. 1–16.

[44] Nagamine, K., Cen, R., Hernquist, L., Ostriker, J. P. and Springel, V.
(2004). Is There a Missing Galaxy Problem at High Redshift? ApJ
610, pp. 45–50.
[45] Padmanabhan, T. (2002). Theoretical Astrophysics, Volume III:
Galaxies and Cosmology (Cambridge, England: Cambridge University
Press, UK).
[46] Paresce, F., McKee, C. F. and Bowyer, S. (1980). Galactic and extra-
galactic contributions to the far-ultraviolet background, ApJ 240, pp.
387–400, doi:10.1086/158244.
[47] Peacock, J. A. (1999). Cosmological Physics (Cambridge, UK: Cam-
bridge University Press, UK).
[48] Peebles, P. J. E. (1993). Principles of physical cosmology (Princeton,
NJ: Princeton University Press, USA).
[49] Petitjean, P., Bergeron, J., Carswell, R. F. and Puget, J. L. (1993). De-
tailed structure of expanding photoionized Ly-alpha clouds, MNRAS
260, pp. 67–76.
[50] Press, W. H. and Schechter, P. (1974). Formation of galaxies and clus-
ters of galaxies by self-similar gravitational condensation, ApJ 187,
pp. 425–438.
[51] Pritchard, J. R. and Loeb, A. (2008). Evolution of the 21cm signal
throughout cosmic history, Phys. Rev. D 78, 10, pp. 103511–+, doi:
10.1103/PhysRevD.78.103511.
[52] Rauch, M. (1998). The Lyman Alpha Forest in the Spectra of QSOs,
ARA&A 36, p. 267.
[53] Richards, G. T., Strauss, M. A., Fan, X., Hall, P. B., Jester, S., Schnei-
der, D. P., Vanden Berk, D. E., Stoughton, C., Anderson, S. F., Brun-
ner, R. J., Gray, J., Gunn, J. E., Ivezić, Ž., Kirkland, M. K., Knapp,
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

48 Vignettes in gravitation and cosmology

G. R., Loveday, J., Meiksin, A., Pope, A., Szalay, A. S., Thakar,
A. R., Yanny, B., York, D. G., Barentine, J. C., Brewington, H. J.,
Brinkmann, J., Fukugita, M., Harvanek, M., Kent, S. M., Kleinman,
S. J., Krzesiński, J., Long, D. C., Lupton, R. H., Nash, T., Neilsen,
E. H., Jr., Nitta, A., Schlegel, D. J. and Snedden, S. A. (2006). The
Sloan Digital Sky Survey Quasar Survey: Quasar Luminosity Function
from Data Release 3, AJ 131, pp. 2766–2787, doi:10.1086/503559.
[54] Samui, S., Srianand, R. and Subramanian, K. (2007). Probing the star
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

formation history using the redshift evolution of luminosity functions,

MNRAS 377, pp. 285–299, doi:10.1111/j.1365-2966.2007.11603.x.
[55] Santos, M. G., Cooray, A., Haiman, Z., Knox, L. and Ma, C.-P. (2003).
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

Small-Scale Cosmic Microwave Background Temperature and Polariza-

tion Anisotropies Due to Patchy Reionization, ApJ 598, pp. 756–766.
[56] Sasaki, S. (1994). Formation rate of bound objects in the hierarchical
clustering model, PASJ 46, pp. 427–430.
[57] Schaerer, D. (2002). On the properties of massive Population III stars
and metal-free stellar populations, A&A 382, pp. 28–42.
[58] Schaye, J., Theuns, T., Leonard, A. and Efstathiou, G. (1999). Mea-
suring the equation of state of the intergalactic medium, MNRAS 310,
p. 57.
[59] Schneider, R., Salvaterra, R., Ferrara, A. and Ciardi, B. (2006). Con-
straints on the initial mass function of the first stars, MNRAS 369,
pp. 825–834.
[60] Somerville, R. S. and Primack, J. R. (1999). Semi-analytic modelling
of galaxy formation: the local Universe, MNRAS 310, p. 1087.
[61] Songaila, A. (2004). The Evolution of the Intergalactic Medium Trans-
mission to Redshift 6, AJ 127, pp. 2598–2603.
[62] Songaila, A. and Cowie, L. L. (2002). Approaching Reionization: The
Evolution of the Ly α Forest from z=4 to z=6, AJ 123, pp. 2183–2196.
[63] Spergel, D. N., Bean, R., Doré, O., Nolta, M. R., Bennett, C. L.,
Dunkley, J., Hinshaw, G., Jarosik, N., Komatsu, E., Page, L., Peiris,
H. V., Verde, L., Halpern, M., Hill, R. S., Kogut, A., Limon, M.,
Meyer, S. S., Odegard, N., Tucker, G. S., Weiland, J. L., Wollack, E.
and Wright, E. L. (2007). Three-Year Wilkinson Microwave Anisotropy
Probe (WMAP) Observations: Implications for Cosmology, ApJS 170,
pp. 377–408.
[64] Stern, D., Yost, S. A., Eckart, M. E., Harrison, F. A., Helfand, D. J.,
Djorgovski, S. G., Malhotra, S. and Rhoads, J. E. (2005). A Galaxy
at z = 6.545 and Constraints on the Epoch of Reionization, ApJ 619,
February 8, 2012 14:59 World Scientific Book - 9in x 6in 00˙Vignettes/Chp. 2

Dark ages and cosmic reionization 49

pp. 12–18.
[65] Storrie-Lombardi, L. J., McMahon, R. G. and Irwin, M. J. (1996).
Evolution of neutral gas at high redshift: implications for the epoch of
galaxy formation, MNRAS 283, p. L79.
[66] Storrie-Lombardi, L. J., McMahon, R. G., Irwin, M. J. and Hazard, C.
(1994). Evolution of Lyman-limit absorption systems over the redshift
range 0.40 less than Z less than 4.69, ApJ 427, pp. L13–L16.
[67] Totani, T., Kawai, N., Kosugi, G., Aoki, K., Yamada, T., Iye, M.,
Vignettes in Gravitation and Cosmology Downloaded from www.worldscientific.com

Ohta, K. and Hattori, T. (2006). Implications for Cosmic Reionization

from the Optical Afterglow Spectrum of the Gamma-Ray Burst 050904
at z = 6.3, PASJ 58, pp. 485–498.
by KAINAN UNIVERSITY on 01/09/15. For personal use only.

[68] Willott, C. J., Delorme, P., Reylé, C., Albert, L., Bergeron, J., Cramp-
ton, D., Delfosse, X., Forveille, T., Hutchings, J. B., McLure, R. J.,
Omont, A. and Schade, D. (2009). Six More Quasars at Redshift 6
Discovered by the Canada-France High-z Quasar Survey, AJ 137, pp.
3541–3547, doi:10.1088/0004-6256/137/3/3541.