arXiv:astro-ph/0602117v16Feb2006
ADVANCED TOPICS IN COSMOLOGY: A PEDAGOGICAL INTRODUCTION
T. Padmanabhan
IUCAA, Post Bag 4, Ganeshkhind, Pune - 411 007
email: nabhan@iucaa.ernet.in
These lecture notes provide a concise, rapid and pedagogical introduction to several advanced topics
in contemporary cosmology. The discussion of thermal history of the universe, linear perturbation
theory, theory of CMBR temperature anisotropies and the inflationary generation of perturbation
are presented in a manner accessible to someone who has done a first course in cosmology. The
discussion of dark energy is more research oriented and reflects the personal bias of the author. Contents:
(I) The cosmological paradigm and Friedmann model; (II) Thermal history of the universe;
(III) Structure formation and linear perturbation theories; (IV) Perturbations in dark matter and
radiation; (V) Transfer function for matter perturbations; (VI) Temperature anisotropies of CMBR;
(VII) Generation of initial perturbations from inflation; (VIII) The dark energy.
Contents
I. The Cosmological Paradigm and Friedmann model 1
II. Thermal History of the Universe 4
A. Neutrino background 4
B. Primordial Nucleosynthesis 5
C. Decoupling of matter and radiation 9
III. Structure Formation and Linear Perturbation Theory 10
IV. Perturbations in Dark Matter and Radiation 14
A. Evolution for λ ≫ dH 15
B. Evolution for λ ≪ dH in the radiation dominated phase 15
C. Evolution in the matter dominated phase 16
D. An alternative description of matter-radiation system 17
V. Transfer Function for matter perturbations 19
VI. Temperature anisotropies of CMBR 20
A. CMBR Temperature Anisotropy: More detailed theory 22
B. Description of photon-baryon fluid 24
VII. Generation of initial perturbations from inflation 26
VIII. The Dark Energy 31
A. Gravitational Holography 34
B. Cosmic Lenz law 34
C. Geometrical Duality in our Universe 35
D. Gravity as detector of the vacuum energy 36
Acknowledgement 37
References 37
I. THE COSMOLOGICAL PARADIGM AND FRIEDMANN MODEL
Observations show that the universe is fairly homogeneous and isotropic at scales larger than about 150h−1
Mpc
where 1 Mpc ≃ 3×1024
cm and h ≈ 0.7 is a parameter related to the expansion rate of the universe. The conventional
— and highly successful — approach to cosmology separates the study of large scale (l 150h−1
Mpc) dynamics of
2
the universe from the issue of structure formation at smaller scales. The former is modeled by a homogeneous and
isotropic distribution of energy density; the latter issue is addressed in terms of gravitational instability which will
amplify the small perturbations in the energy density, leading to the formation of structures like galaxies. In such an
approach, the expansion of the background universe is described by the metric (We shall use units with with c = 1
throughout, unless otherwise specified):
ds2
≡ dt2
− a2
dx2
≡ dt2
− a2
(t) dχ2
+ S2
k(χ) dθ2
+ sin2
θdφ2
(1)
with Sk(χ) = (sin χ, χ, sinh χ) for the three values of the label k = (1, 0, −1). The function a(t) is governed by the
equations:
˙a2
+ k
a2
=
8πGρ
3
; d(ρa3
) = −pda3
(2)
The first one relates expansion rate of the universe to the energy density ρ and k = 0, ±1 is a parameter which
characterizes the spatial curvature of the universe. The second equation, when coupled with the equation of state
p = p(ρ) which relates the pressure p to the energy density, determines the evolution of energy density ρ = ρ(a) in
terms of the expansion factor of the universe. In particular if p = wρ with (at least, approximately) constant w then,
ρ ∝ a−3(1+w)
and (if we further assume k = 0, which is strongly favoured by observations) the first equation in Eq.(2)
gives a ∝ t2/[3(1+w)]
. We will also often use the redshift z(t), defined as (1 + z) = a0/a(t) where the subscript zero
denotes quantities evaluated at the present moment. in a k = 0 universe, we can set a0 = 1 by rescaling the spatial
coordinates.
It is convenient to measure the energy densities of different components in terms of a critical energy density (ρc)
required to make k = 0 at the present epoch. (Of course, since k is a constant, it will remain zero at all epochs if it
is zero at any given moment of time.) From Eq.(2), it is clear that ρc = 3H2
0 /8πG where H0 ≡ (˙a/a)0 — called the
Hubble constant — is the rate of expansion of the universe at present. Numerically
ρc =
3H2
0
8πG
= 1.88h2
× 10−29
gm cm−3
= 2.8 × 1011
h2
M⊙ Mpc−3
= 1.1 × 104
h2
eV cm−3
= 1.1 × 10−5
h2
protons cm−3
(3)
The variables Ωi ≡ ρi/ρc will give the fractional contribution of different components of the universe (i denoting
baryons, dark matter, radiation, etc.) to the critical density. Observations then lead to the following results:
(1) Our universe has 0.98 Ωtot 1.08. The value of Ωtot can be determined from the angular anisotropy spectrum
of the cosmic microwave background radiation (CMBR; see Section VI) and these observations (combined with the
reasonable assumption that h > 0.5) show[1] that we live in a universe with critical density, so that k = 0.
(2) Observations of primordial deuterium produced in big bang nucleosynthesis (which took place when the universe
was about few minutes in age) as well as the CMBR observations show[2] that the total amount of baryons in the
universe contributes about ΩB = (0.024±0.0012)h−2
. Given the independent observations[3] which fix h = 0.72±0.07,
we conclude that ΩB
∼= 0.04 − 0.06. These observations take into account all baryons which exist in the universe
today irrespective of whether they are luminous or not. Combined with previous item we conclude that most of the
universe is non-baryonic.
(3) Host of observations related to large scale structure and dynamics (rotation curves of galaxies, estimate of
cluster masses, gravitational lensing, galaxy surveys ..) all suggest[4] that the universe is populated by a non-luminous
component of matter (dark matter; DM hereafter) made of weakly interacting massive particles which does cluster at
galactic scales. This component contributes about ΩDM
∼= 0.20 − 0.35 and has the simple equation of state pDM ≈ 0.
The second equation in Eq.(2), then gives ρDM ∝ a−3
as the universe expands which arises from the evolution of
number density of particles: ρ = nmc2
∝ n ∝ a−3
.
(4) Combining the last observation with the first we conclude that there must be (at least) one more component to
the energy density of the universe contributing about 70% of critical density. Early analysis of several observations[5]
indicated that this component is unclustered and has negative pressure. This is confirmed dramatically by the
supernova observations (see Ref. [6]; for a critical look at the current data, see Ref. [7]). The observations suggest
that the missing component has w = p/ρ −0.78 and contributes ΩDE
∼= 0.60 − 0.75. The simplest choice for
such dark energy with negative pressure is the cosmological constant which is a term that can be added to Einstein’s
equations. This term acts like a fluid with an equation of state pDE = −ρDE; the second equation in Eq.(2), then
gives ρDE = constant as universe expands.
(5) The universe also contains radiation contributing an energy density ΩRh2
= 2.56 × 10−5
today most of which
is due to photons in the CMBR. The equation of state is pR = (1/3)ρR; the second equation in Eq.(2), then gives
ρR ∝ a−4
. Combining it with the result ρR ∝ T 4
for thermal radiation, it follows that T ∝ a−1
. Radiation is
3
dynamically irrelevant today but since (ρR/ρDM ) ∝ a−1
it would have been the dominant component when the
universe was smaller by a factor larger than ΩDM /ΩR ≃ 4 × 104
ΩDM h2
.
(6) Taking all the above observations together, we conclude that our universe has (approximately) ΩDE ≃
0.7, ΩDM ≃ 0.26, ΩB ≃ 0.04, ΩR ≃ 5 × 10−5
. All known observations are consistent with such an — admittedly
weird — composition for the universe.
Using ρNR ∝ a−3
, ρR ∝ a−4
and ρDE=constant we can write Eq.(2) in a convenient dimensionless form as
1
2
dq
dτ
2
+ V (q) = E (4)
where τ = H0t, a = a0q(τ), ΩNR = ΩB + ΩDM and
V (q) = −
1
2
ΩR
q2
+
ΩNR
q
+ ΩDEq2
; E =
1
2
(1 − Ωtot) . (5)
This equation has the structure of the first integral for motion of a particle with energy E in a potential V (q). For
models with Ω = ΩNR + ΩDE = 1, we can take E = 0 so that (dq/dτ) = V (q). Based on the observed composition
of the universe, we can identify three distinct phases in the evolution of the universe when the temperature is less
than about 100 GeV. At high redshifts (small q) the universe is radiation dominated and ˙q is independent of the
other cosmological parameters. Then Eq.(4) can be easily integrated to give a(t) ∝ t1/2
and the temperature of the
universe decreases as T ∝ t−1/2
. As the universe expands, a time will come when (t = teq, a = aeq and z = zeq,
say) the matter energy density will be comparable to radiation energy density. For the parameters described above,
(1 + zeq) = ΩNR/ΩR ≃ 4 × 104
ΩDM h2
. At lower redshifts, matter will dominate over radiation and we will have
a ∝ t2/3
until fairly late when the dark energy density will dominate over non relativistic matter. This occurs at a
redshift of zDE where (1 + zDE) = (ΩDE/ΩNR)1/3
. For ΩDE ≈ 0.7, ΩNR ≈ 0.3, this occurs at zDE ≈ 0.33. In this
phase, the velocity ˙q changes from being a decreasing function to an increasing function leading to an accelerating
universe. In addition to these, we believe that the universe probably went through a rapidly expanding, inflationary,
phase very early when T ≈ 1014
GeV; we will say more about this in Section VII. (For a textbook description of these
and related issues, see e.g. Ref. [8].)
Before we conclude this section, we will briefly mention some key aspects of the background cosmology described
by a Friedmann model.
(a) The metric in Eq.(1) can be rewritten using the expansion parameter a or the redshift z = (a0/a)−1
− 1 as the
time coordinate in the form
ds2
= H−2
(a)
da
a
2
− a2
dx2
=
1
(1 + z)2
H−2
(z)dz2
− dx2
(6)
This form clearly shows that the only dynamical content of the metric is encoded in the function H(a) = (˙a/a). An
immediate consequence is that any observation which is capable of determining the geometry of the universe can only
provide — at best — information about this function.
(b) Since cosmological observations usually use radiation received from distant sources, it is worth reviewing briefly
the propagation of radiation in the universe. The radial light rays follow a trajectory given by
rem(z) = Sk(α); α ≡
1
a0
z
0
H−1
(z)dz (7)
if the photon is emitted at rem at the redshift z and received here today. Two other quantities closely related to
rem(z) are the luminosity distance, dL, and the angular diameter distance dA. If we receive a flux F from a source
of luminosity L, then the luminosity distance is defined via the relation F ≡ L/4πd2
L(z). If an object of transverse
length l subtends a small angle θ, the angular diameter distance is defined via (l = θdA ). Simple calculation shows
that:
dL(z) = a0rem(z)(1 + z) = a0(1 + z)Sk(α); dA(z) = a0rem(z)(1 + z)−1
(8)
(c) As an example of determining the spacetime geometry of the universe from observations, let us consider how
one can determine a(t) from the observations of the luminosity distance. It is clear from the first equation in Eq. (8)
that
H−1
(z) = 1 −
kd2
L(z)
a2
0(1 + z)2
−1/2
d
dz
dL(z)
1 + z
→
d
dz
dL(z)
1 + z
(9)
where the last form is valid for a k = 0 universe. If we determine the form of dL(z) from observations — which can
be done if we can measure the flux F from a class of sources with known value for luminosity L — then we can use
this relation to determine the evolutionary history of the universe and thus the dynamics.
4
II. THERMAL HISTORY OF THE UNIVERSE
Let us next consider some key events in the evolutionary history of our universe [8]. The most well understood phase
of the universe occurs when the temperature is less than about 1012
K. Above this temperature, thermal production of
baryons and their strong interaction is significant and somewhat difficult to model. We can ignore such complications
at lower temperatures and — as we shall see — several interesting physical phenomena did take place during the later
epochs with T 1012
.
The first thing we need to do is to determine the composition of the universe when T ≈ 1012
K. We will certainly
have, at this time, copious amount of photons and all species of neutrinos and antineutrinos. In addition, neutrons and
protons must exist at this time since there is no way they could be produced later on. (This implies that phenomena
which took place at higher temperatures should have left a small excess of baryons over anti baryons; we do not
quite understand how this happened and will just take it as an initial condition.) Since the rest mass of electrons
correspond to a much lower temperature (about 0.5×1010
K), there will be large number of electrons and positrons at
this temperature but in order to maintain charge neutrality, we need to have a slight excess of electrons over positrons
(by about 1 part in 109
) with the net negative charge compensating the positive charge contributed by protons.
An elementary calculation using the known interaction rates show that all these particles are in thermal equilibrium
at this epoch. Hence standard rules of statistical mechanics allows us to determine the number density (n), energy
density (ρ) and the pressure (p) in terms of the distribution function f:
n = f(k)d3
k =
g
2π2
∞
m
(E2
− m2
)1/2
EdE
exp[(E − µ)/T ] ± 1
(10)
ρ = Ef(k)d3
k =
g
2π2
∞
m
(E2
− m2
)1/2
E2
dE
exp[(E − µ)/T ] ± 1
(11)
p =
1
3
d3
kf(k)kv(k) =
1
3
|k|2
E
f(k)d3
k =
g
6π2
∞
m
(E2
− m2
)3/2
dE
e[(E−µ)/T ] ± 1
(12)
Next, we can argue that the chemical potentials for electrons, positrons and neutrinos can be taken to be zero. For
example, conservation of chemical potential in the reaction e+
e−
→ 2γ implies that the chemical potentials of electrons
and positrons must differ in a sign. But since the number densities of electrons and positrons, which are determined
by the chemical potential, are very close to each other, the chemical potentials of electrons and positrons must be
(very closely) equal to each other. Hence both must be (very close to) zero. Similar reasoning based on lepton number
shows that neutrinos should also have zero chemical potential. Given this, one can evaluate the integrals for all the
relativistic species and we obtain for the total energy density
ρtotal =
i=boson
gi
π2
30
T 4
i +
i=fermion
7
8
gi
π2
30
T 4
i = gtotal
π2
30
T 4
(13)
where
gtotal ≡
boson
gB +
fermion
7
8
gF . (14)
The corresponding entropy density is given by
s ∼=
1
T
(ρ + p) =
2π2
45
qT 3
; q ≡ qtotal =
boson
gB +
7
8
fermion
gF . (15)
A. Neutrino background
As a simple application of the above result, let us consider the fate of neutrinos in the expanding universe. From the
standard weak interaction theory, one can compute the reaction rate Γ of the neutrinos with the rest of the species.
When this reaction rate fall below the expansion rate H of the universe, the reactions cannot keep the neutrinos
coupled to the rest of the matter. A simple calculation [8] shows that the relevant ratio is given by
Γ
H
≃
T
1.4MeV
3
=
T
1.6 × 1010K
3
(16)
5
Thus, for T 1.6 × 1010
K, the neutrinos decouple from matter. At slightly lower temperature, the electrons and
positrons annihilate increasing the number density of photons. Neutrinos do not get any share of this energy since
they have already decoupled from the rest of the matter. As a result, the photon temperature goes up with respect
to the neutrino temperature once the e+
e−
annihilation is complete. This increase in the temperature is easy to
calculate. As far as the photons are concerned, the increase in the temperature is essentially due to the change in the
degrees of freedom g and is given by:
(aTγ)3
after
(aTγ)3
before
=
gbefore
gafter
=
7
8 (2 + 2) + 2
2
=
11
4
. (17)
(In the numerator, one 2 is for electron; one 2 is for positron; the 7/8 factor arises because these are fermions. The
final 2 is for photons. In the denominator, there are only photons to take care of.) Therefore
(aTγ)after =
11
4
1/3
(aTγ)before =
11
4
1/3
(aTν)before
=
11
4
1/3
(aTν)after≃1.4(aTν)after. (18)
The first equality is from Eq. (17); the second arises because the photons and neutrinos had the same temperature
originally; the third equality is from the fact that for decoupled neutrinos aTν is a constant. This result leads to the
prediction that, at present, the universe will contain a bath of neutrinos which has temperature that is (predictably)
lower than that of CMBR. The future detection of such a cosmic neutrino background will allow us to probe the
universe at its earliest epochs.
B. Primordial Nucleosynthesis
When the temperature of the universe is higher than the binding energy of the nuclei (∼ MeV), none of the heavy
elements (helium and the metals) could have existed in the universe. The binding energies of the first four light nuclei,
2
H, 3
H, 3
He and 4
He are 2.22 MeV, 6.92 MeV, 7.72 MeV and 28.3 MeV respectively. This would suggest that these
nuclei could be formed when the temperature of the universe is in the range of (1 − 30)MeV. The actual synthesis
takes place only at a much lower temperature, Tnuc = Tn ≃ 0.1MeV. The main reason for this delay is the ‘high
entropy’ of our universe, i.e., the high value for the photon-to-baryon ratio, η−1
. Numerically,
η =
nB
nγ
= 5.5 × 10−10 ΩBh2
0.02
; ΩBh2
= 3.65 × 10−3 T0
2.73 K
3
η10 (19)
To see this, let us assume, for a moment, that the nuclear (and other) reactions are fast enough to maintain thermal
equilibrium between various species of particles and nuclei. In thermal equilibrium, the number density of a nuclear
species A
NZ with atomic mass A and charge Z will be
nA = gA
mAT
2π
3/2
exp −
mA − µA
T
. (20)
From this one can obtain the equation for the temperature TA at which the mass fraction of a particular species-A
will be of order unity (XA ≃ 1). We find that
TA ≃
BA/(A − 1)
ln(η−1) + 1.5 ln(mB/T )
(21)
where BA is the binding energy of the species. This temperature will be fairly lower than BA because of the large
value of η−1
. For 2
H, 3
He and 4
He the value of TA is 0.07MeV, 0.11MeV and 0.28MeV respectively. Comparison
with the binding energy of these nuclei shows that these values are lower than the corresponding binding energies BA
by a factor of about 10, at least.
Thus, even when the thermal equilibrium is maintained, significant synthesis of nuclei can occur only at T 0.3MeV
and not at higher temperatures. If such is the case, then we would expect significant production (XA 1) of nuclear
species-A at temperatures T TA. It turns out, however, that the rate of nuclear reactions is not high enough to
maintain thermal equilibrium between various species. We have to determine the temperatures up to which thermal
6
equilibrium can be maintained and redo the calculations to find non-equilibrium mass fractions. The general procedure
for studying non equilibrium abundances in an expanding universe is based on rate equations. Since we will require
this formalism again in Section II C (for the study of recombination), we will develop it in a somewhat general context.
Consider a reaction in which two particles 1 and 2 interact to form two other particles 3 and 4. For example,
n+νe ⇋ p+e constitutes one such reaction which converts neutrons into protons in the forward direction and protons
into neutrons in the reverse direction; another example we will come across in the next section is p + e ⇋ H + γ
where the forward reaction describes recombination of electron and proton forming a neutral hydrogen atom (with
the emission of a photon), while the reverse reaction is the photoionisation of a hydrogen atom. In general, we are
interested in how the number density n1 of particle species 1, say, changes due to a reaction of the form 1 +2 ⇋ 3 +4.
We first note that even if there is no reaction, the number density will change as n1 ∝ a−3
due to the expansion
of the universe; so what we are really after is the change in n1a3
. Further, the forward reaction will be proportional
to the product of the number densities n1n2 while the reverse reaction will be proportional to n3n4. Hence we can
write an equation for the rate of change of particle species n1 as
1
a3
d(n1a3
)
dt
= µ(An3n4 − n1n2). (22)
The left hand side is the relevant rate of change over and above that due to the expansion of the universe; on the
right hand side, the two proportionality constants have been written as µ and (Aµ), both of which, of course, will
be functions of time. (The quantity µ has the dimensions of cm3
s−1
, so that nµ has the dimensions of s−1
; usually
µ ≃ σv where σ is the cross-section for the relevant process and v is the relative velocity.) The left hand side has to
vanish when the system is in thermal equilibrium with ni = neq
i , where the superscript ‘eq’ denotes the equilibrium
densities for the different species labeled by i = 1 − 4. This condition allows us to rewrite A as A = neq
1 neq
2 /(neq
3 neq
4 ).
Hence the rate equation becomes
1
a3
d(n1a3
)
dt
= µneq
1 neq
2
n3n4
neq
3 neq
4
−
n1n2
neq
1 neq
2
. (23)
In the left hand side, one can write (d/dt) = Ha(d/da) which shows that the relevant time scale governing the
process is H−1
. Clearly, when H/nµ ≫ 1 the right hand side becomes ineffective because of the (µ/H) factor and
the number of particles of species 1 does not change. We see that when the expansion rate of the universe is large
compared to the reaction rate, the given reaction is ineffective in changing the number of particles. This certainly does
not mean that the reactions have reached thermal equilibrium and ni = neq
i ; in fact, it means exactly the opposite:
The reactions are not fast enough to drive the number densities towards equilibrium densities and the number densities
“freeze out” at non equilibrium values. Of course, the right hand side will also vanish when ni = neq
i which is the
other extreme limit of thermal equilibrium.
Having taken care of the general formalism, let us now apply it to the process of nucleosynthesis which requires
protons and neutrons combining together to form bound nuclei of heavier elements like deuterium, helium etc.. The
abundance of these elements are going to be determined by the relative abundance of neutrons and protons in the
universe. Therefore, we need to first worry about the maintenance of thermal equilibrium between protons and the
neutrons in the early universe. As long as the inter-conversion between n and p through the weak interaction processes
(ν + n ↔ p + e), (e + n ↔ p + ν) and the ‘decay’ (n ↔ p + e + ν), is rapid (compared to the expansion rate of the
universe), thermal equilibrium will be maintained. Then the equilibrium (n/p) ratio will be
nn
np
=
Xn
Xp
= exp(−Q/T ), (24)
where Q = mn − mp = 1.293 MeV. At high (T ≫ Q) temperatures, there will be equal number of neutrons and
protons but as the temperature drops below about 1.3 MeV, the neutron fraction will start dropping exponentially
provided thermal equilibrium is still maintained. To check whether thermal equilibrium is indeed maintained, we
need to compare the expansion rate with the reaction rate. The expansion rate is given by H = (8πGρ/3)1/2
where
ρ = (π2
/30)gT 4
with g ≈ 10.75 representing the effective relativistic degrees of freedom present at these temperatures.
At T = Q, this gives H ≈ 1.1 s−1
. The reaction rate needs to be computed from weak interaction theory. The neutron
to proton conversion rate, for example, is well approximated by
λnp ≈ 0.29 s−1 T
Q
5
Q
T
2
+ 6
Q
T
+ 12 . (25)
At T = Q, this gives λ ≈ 5 s−1
, slightly more rapid than the expansion rate. But as T drops below Q, this decreases
rapidly and the reaction ceases to be fast enough to maintain thermal equilibrium. Hence we need to work out the
neutron abundance by using Eq. (23).
7
FIG. 1: The evolution of mass fraction of different species during nucleosynthesis
Using n1 = nn, n3 = np and n2, n4 = nl where the subscript l stands for the leptons, Eq. (23) becomes
1
a3
d(nna3
)
dt
= µneq
l
npneq
n
neq
p
− nn . (26)
We now use Eq. (24), write (neq
l µ) = λnp which is the rate for neutron to proton conversion and introduce the
fractional abundance Xn = nn/(nn + np). Simple manipulation then leads to the equation
dXn
dt
= λnp (1 − Xn)e−Q/T
− Xn . (27)
Converting from the variable t to the variable s = (Q/T ) and using (d/dt) = −HT (d/dT ), the equations we need to
solve reduce to
−Hs
dXn
ds
= λnp (1 − Xn)e−s
− Xn ; H = (1.1 sec−1
) s−4
; λnp =
0.29 s−1
s5
s2
+ 6s + 12 . (28)
It is now straightforward to integrate these equations numerically and determine how the neutron abundance changes
with time. The neutron fraction falls out of equilibrium when temperatures drop below 1 MeV and it freezes to about
0.15 at temperatures below 0.5 MeV.
As the temperature decreases further, the neutron decay with a half life of τn ≈ 886.7 sec (which is not included in
the above analysis) becomes important and starts depleting the neutron number density. The only way neutrons can
survive is through the synthesis of light elements. As the temperature falls further to T = THe ≃ 0.28MeV, significant
amount of He could have been produced if the nuclear reaction rates were high enough. The possible reactions which
produces 4
He are [D(D, n) 3
He(D, p) 4
He, D(D, p) 3
H(D, n) 4
He, D(D, γ) 4
He]. These are all based on D, 3
He and
3
H and do not occur rapidly enough because the mass fraction of D, 3
He and 3
H are still quite small [10−12
, 10−19
and 5 × 10−19
respectively] at T ≃ 0.3MeV. The reactions n + p ⇋ d + γ will lead to an equilibrium abundance ratio
of deuterium given by
npnn
ndn
=
4
3
mpmn
md
3/2
(2πkBT )3/2
(2π )
3
n
e−B/kBT
= exp 25.82 − ln ΩBh2
T
3/2
10 −
2.58
T10
. (29)
The equilibrium deuterium abundance passes through unity (for ΩBh2
= 0.02) at the temperature of about 0.07 MeV
which is when the nucleosynthesis can really begin.
8
So we need to determine the neutron fraction at T = 0.07 MeV given that it was about 0.15 at 0.5 MeV. During
this epoch, the time-temperature relationship is given by t = 130 sec (T/0.1 MeV)−2
. The neutron decay factor
is exp(−t/τn) ≈ 0.74 for T = 0.07 MeV. This decreases the neutron fraction to 0.15 × 0.74 = 0.11 at the time of
nucleosynthesis. When the temperature becomes T 0.07MeV, the abundance of D and 3
H builds up and these
elements further react to form 4
He. A good fraction of D and 3
H is converted to 4
He (See Fig.1 which shows the
growth of deuterium and its subsequent fall when helium is built up). The resultant abundance of 4
He can be easily
calculated by assuming that almost all neutrons end up in 4
He. Since each 4
He nucleus has two neutrons, (nn/2)
helium nuclei can be formed (per unit volume) if the number density of neutrons is nn. Thus the mass fraction of
4
He will be
Y =
4(nn/2)
nn + np
=
2(n/p)
1 + (n/p)
= 2xc (30)
where xc = n/(n + p) is the neutron abundance at the time of production of deuterium. For ΩBh2
= 0.02, xc ≈ 0.11
giving Y ≈ 0.22. Increasing baryon density to ΩBh2
= 1 will make Y ≈ 0.25. An accurate fitting formula for the
dependence of helium abundance on various parameters is given by
Y = 0.226 + 0.025 log η10 + 0.0075(g∗ − 10.75) + 0.014(τ1/2(n) − 10.3 min) (31)
where η10 measures the baryon-photon ratio today via Eq. (19) and g∗ is the effective number of relativistic degrees of
freedom contributing to the energy density and τ1/2(n) is the neutron half life. The results (of a more exact treatment)
are shown in Fig. 1.
As the reactions converting D and 3
H to 4
He proceed, the number density of D and 3
H is depleted and the reaction
rates - which are proportional to Γ ∝ XA(ηnγ) < σv > - become small. These reactions soon freeze-out leaving a
residual fraction of D and 3
H (a fraction of about 10−5
to 10−4
). Since Γ ∝ η it is clear that the fraction of (D,3
H)
left unreacted will decrease with η. In contrast, the 4
He synthesis - which is not limited by any reaction rate - is
fairly independent of η and depends only on the (n/p) ratio at T ≃ 0.1MeV. The best fits, with typical errors, to
deuterium abundance calculated from the theory, for the range η = (10−10
− 10−9
) is given by
Y2 ≡
D
H p
= 3.6 × 10−5±0.06 η
5 × 10−10
−1.6
. (32)
The production of still heavier elements - even those like 16
C, 16
O which have higher binding energies than 4
He
- is suppressed in the early universe. Two factors are responsible for this suppression: (1) For nuclear reactions to
proceed, the participating nuclei must overcome their Coulomb repulsion. The probability to tunnel through the
Coulomb barrier is governed by the factor F = exp[−2A1/3
(Z1Z2)2/3
(T/1MeV)−1/3
] where A−1
= A−1
1 + A−1
2 . For
heavier nuclei (with larger Z), this factor suppresses the reaction rate. (2) Reaction between helium and proton would
have led to an element with atomic mass 5 while the reaction of two helium nuclei would have led to an element with
atomic mass 8. However, there are no stable elements in the periodic table with the atomic mass of 5 or 8! The 8
Be,
for example, has a half life of only 10−16
seconds. One can combine 4
He with 8
Be to produce 12
C but this can occur at
significant rate only if it is a resonance reaction. That is, there should exist an excited state 12
C nuclei which has an
energy close to the interaction energy of 4
He + 8
Be. Stars, incidentally, use this route to synthesize heavier elements.
It is this triple-alpha reaction which allows the synthesis of heavier elements in stars but it is not fast enough in the
early universe. (You must thank your stars that there is no such resonance in 16
O or in 20
Ne — which is equally
important for the survival of carbon and oxygen.)
The current observations indicate, with reasonable certainty that: (i) (D/H) 1 × 10−5
. (ii) [(D +3
He)/H]
≃ (1 − 8) × 10−5
and (iii) 0.236 < (4
He/H) < 0.254. These observations are consistent with the predictions if
10.3 min τ 10.7 min, and η = (3 − 10) × 10−10
. Using η = 2.68 × 10−8
ΩBh2
, this leads to the important
conclusion: 0.011 ≤ ΩBh2
≤ 0.037. When combined with the broad bounds on h, 0.6 h 0.8, say, we can constrain
the baryonic density of the universe to be: 0.01 ΩB 0.06. These are the typical bounds on ΩB available today. It
shows that, if Ωtotal ≃ 1 then most of the matter in the universe must be non baryonic.
Since the 4
He production depends on g, the observed value of 4
He restricts the total energy density present at
the time of nucleosynthesis. In particular, it constrains the number (Nν) of light neutrinos (that is, neutrinos with
mν 1MeV which would have been relativistic at T ≃ 1MeV). The observed abundance is best explained by Nν = 3,
is barely consistent with Nν = 4 and rules out Nν > 4. The laboratory bound on the total number of particles
including neutrinos, which couples to the Z0
boson is determined by measuring the decay width of the particle Z0
;
each particle with mass less than (mz/2) ≃ 46 GeV contributes about 180 MeV to this decay width. This bound is
Nν = 2.79 ± 0.63 which is consistent with the cosmological observations.
9
C. Decoupling of matter and radiation
In the early hot phase, the radiation will be in thermal equilibrium with matter; as the universe cools below
kBT ≃ (ǫa/10) where ǫa is the binding energy of atoms, the electrons and ions will combine to form neutral atoms
and radiation will decouple from matter. This occurs at Tdec ≃ 3 × 103
K. As the universe expands further, these
photons will continue to exist without any further interaction. It will retain thermal spectrum since the redshift of the
frequency ν ∝ a−1
is equivalent to changing the temperature in the spectrum by the scaling T ∝ (1/a). It turns out
that the major component of the extra-galactic background light (EBL) which exists today is in the microwave band
and can be fitted very accurately by a thermal spectrum at a temperature of about 2.73 K. It seems reasonable to
interpret this radiation as a relic arising from the early, hot, phase of the evolving universe. This relic radiation, called
cosmic microwave background radiation, turns out to be a gold mine of cosmological information and is extensively
investigated in recent times. We shall now discuss some details related to the formation of neutral atoms and the
decoupling of photons.
The relevant reaction is, of course, e + p ⇋ H + γ and if the rate of this reaction is faster than the expansion rate,
then one can calculate the neutral fraction using Saha’s equation. Introducing the fractional ionisation, Xi, for each
of the particle species and using the facts np = ne and np + nH = nB, it follows that Xp = Xe and XH = (nH/nB)
= 1 − Xe. Saha’s equation now gives
1 − Xe
X2
e
∼= 3.84η(T/me)3/2
exp(B/T ) (33)
where η = 2.68 × 10−8
(ΩBh2
) is the baryon-to-photon ratio. We may define Tatom as the temperature at which 90
percent of the electrons, say, have combined with protons: i.e. when Xe = 0.1. This leads to the condition:
(ΩBh2
)−1
τ− 3
2 exp −13.6τ−1
= 3.13 × 10−18
(34)
where τ = (T/1eV). For a given value of (ΩBh2
), this equation can be easily solved by iteration. Taking logarithms
and iterating once we find τ−1 ∼= 3.084 − 0.0735 ln(ΩBh2
) with the corresponding redshift (1 + z) = (T/T0) given by
(1 + z) = 1367[1 − 0.024 ln(ΩBh2
)]−1
. (35)
For ΩBh2
= 1, 0.1, 0.01 we get Tatom
∼= 0.324eV, 0.307eV, 0.292eV respectively. These values correspond to the
redshifts of 1367, 1296 and 1232.
Because the preceding analysis was based on equilibrium densities, it is important to check that the rate of the
reactions p + e ↔ H + γ is fast enough to maintain equilibrium. For ΩBh2
≈ 0.02, the equilibrium condition is only
marginally satisfied, making this analysis suspect. More importantly, the direct recombination to the ground state of
the hydrogen atom — which was used in deriving the Saha’s equation — is not very effective in producing neutral
hydrogen in the early universe. The problem is that each such recombination releases a photon of energy 13.6 eV
which will end up ionizing another neutral hydrogen atom which has been formed earlier. As a result, the direct
recombination to the ground state does not change the neutral hydrogen fraction at the lowest order. Recombination
through the excited states of hydrogen is more effective since such a recombination ends up emitting more than one
photon each of which has an energy less than 13.6 eV. Given these facts, it is necessary to once again use the rate
equation developed in the previous section to track the evolution of ionisation fraction.
A simple procedure for doing this, which captures the essential physics, is as follows: We again begin with Eq. (23)
and repeating the analysis done in the last section, now with n1 = ne, n2 = np, n3 = nH and n4 = nγ, and defining
Xe = ne/(ne + nH) = np/nH one can easily derive the rate equation for this case:
dXe
dt
= β(1 − Xe) − αnbX2
e = α
β
α
(1 − Xe) − nbX2
e . (36)
This equation is analogous to Eq. (27); the first term gives the photoionisation rate which produces the free electrons
and the second term is the recombination rate which converts free electrons into hydrogen atom and we have used the
fact ne = nbXe etc.. Since we know that direct recombination to the ground state is not effective, the recombination
rate α is the rate for capture of electron by a proton forming an excited state of hydrogen. To a good approximation,
this rate is given by
α = 9.78r2
0c
B
T
1/2
ln
B
T
(37)
10
where r0 = e2
/mec2
is the classical electron radius. To integrate Eq. (36) we also need to know β/α. This is easy
because in thermal equilibrium the right hand side of Eq. (36) should vanish and Saha’s equation tells us the value of
Xe in thermal equilibrium. On using Eq. (33), this gives
β
α
=
meT
2π
3/2
exp[−(B/T )]. (38)
We can now integrate Eq. (36) using the variable B/T just as we used the variable Q/T in solving Eq. (27). The
result shows that the actual recombination proceeds more slowly compared to that predicted by the Saha’s equation.
The actual fractional ionisation is higher than the value predicted by Saha’s equation at temperatures below about
1300. For example, at z = 1300, these values differ by a factor 3; at z ≃ 900, they differ by a factor of 200. The
value of Tatom, however, does not change significantly. A more rigorous analysis shows that, in the redshift range of
800 < z < 1200, the fractional ionisation varies rapidly and is given (approximately) by the formula,
Xe = 2.4 × 10−3 (ΩNRh2
)1/2
(ΩBh2)
z
1000
12.75
. (39)
This is obtained by fitting a curve to the numerical solution.
The formation of neutral atoms makes the photons decouple from the matter. The redshift for decoupling can be
determined as the epoch at which the optical depth for photons is unity. Using Eq. (39), we can compute the optical
depth for photons to be
τ =
t
0
n(t)Xe(t)σT dt =
z
o
n(z)Xe(z)σT
dt
dz
dz ≃ 0.37
z
1000
14.25
(40)
where we have used the relation H0dt ∼= −Ω
−1/2
NR z−5/2
dz which is valid for z ≫ 1. This optical depth is unity at
zdec = 1072. From the optical depth, we can also compute the probability that the photon was last scattered in the
interval (z, z + dz). This is given by (exp −τ) (dτ/dz) which can be expressed as
P(z) = e−τ dτ
dz
= 5.26 × 10−3 z
1000
13.25
exp −0.37
z
1000
14.25
. (41)
This P(z) has a sharp maximum at z ≃ 1067 and a width of about ∆z ∼= 80. It is therefore reasonable to assume
that decoupling occurred at z ≃ 1070 in an interval of about ∆z ≃ 80. We shall see later that the finite thickness of
the surface of last scattering has important observational consequences.
III. STRUCTURE FORMATION AND LINEAR PERTURBATION THEORY
Having discussed the evolution of the background universe, we now turn to the study of structure formation. Before
discussing the details, let us briefly summarise the broad picture and give references to some of the topics that we will
not discuss. The key idea is that if there existed small fluctuations in the energy density in the early universe, then
gravitational instability can amplify them in a well-understood manner leading to structures like galaxies etc. today.
The most popular model for generating these fluctuations is based on the idea that if the very early universe went
through an inflationary phase [9], then the quantum fluctuations of the field driving the inflation can lead to energy
density fluctuations[10, 11]. It is possible to construct models of inflation such that these fluctuations are described
by a Gaussian random field and are characterized by a power spectrum of the form P(k) = Akn
with n ≃ 1 (see Sec.
VII). The models cannot predict the value of the amplitude A in an unambiguous manner but it can be determined
from CMBR observations. The CMBR observations are consistent with the inflationary model for the generation of
perturbations and gives A ≃ (28.3h−1
Mpc)4
and n = 0.97±0.023 (The first results were from COBE [12] and WMAP
has reconfirmed them with far greater accuracy). When the perturbation is small, one can use well defined linear
perturbation theory to study its growth. But when δ ≈ (δρ/ρ) is comparable to unity the perturbation theory breaks
down. Since there is more power at small scales, smaller scales go non-linear first and structure forms hierarchically.
The non linear evolution of the dark matter halos (which is an example of statistical mechanics of self gravitating
systems; see e.g.[13]) can be understood by simulations as well as theoretical models based on approximate ansatz
[14] and nonlinear scaling relations [15]. The baryons in the halo will cool and undergo collapse in a fairly complex
manner because of gas dynamical processes. It seems unlikely that the baryonic collapse and galaxy formation can be
understood by analytic approximations; one needs to do high resolution computer simulations to make any progress
[16]. All these results are broadly consistent with observations.
11
As long as these fluctuations are small, one can study their evolution by linear perturbation theory, which is
what we will start with [17]. The basic idea of linear perturbation theory is well defined and simple. We perturb
the background FRW metric by gF RW
ik → gF RW
ik + hik and also perturb the source energy momentum tensor by
T F RW
ik → T F RW
ik + δTik. Linearising the Einstein’s equations, one can relate the perturbed quantities by a relation
of the form L(gF RW
ik )hik = δTik where L is second order linear differential operator depending on the back ground
metric gF RW
ik . Since the background is maximally symmetric, one can separate out time and space; for e.g, if k = 0,
simple Fourier modes can be used for this purpose and we can write down the equation for any given mode, labelled
by a wave vector k as:
L(a(t), k)hab(t, k) = δTab(t, k) (42)
To every mode we can associate a wavelength normalized to today’s value: λ(t) = (2π/k)(1+z)−1
and a corresponding
mass scale which is invariant under expansion:
M =
4πρ(t)
3
λ(t)
2
3
=
4πρ0
3
λ0
2
3
=1.5 × 1011
M⊙(Ωmh2
)
λ0
1 Mpc
3
. (43)
The behaviour of the mode depends on the relative value of λ(t) as compared to the Hubble radius dH(t) ≡ (˙a/a)−1
.
Since the Hubble radius: dH(t) ∝ t while the wavelength of the mode: λ(t) ∝ a(t) ∝ (t1/2
, t2/3
) in the radiation
dominated and matter dominated phases it follows that λ(t) > dH(t) at sufficiently early times. When λ(t) = dH (t),
we say that the mode is entering the Hubble radius. Since the Hubble radius at z = zeq is
λeq
∼=
H−1
0
√
2
Ω
1/2
R
ΩNR
∼= 14Mpc(ΩNRh2
)−1
(44)
it follows that modes with λ0 > λeq enter Hubble radius in MD phase while the more relevant modes with λ < λeq
enter in the RD phase. Thus, for a given mode we can identify three distinct phases: First, very early on, when
λ > dH, z > zeq the dynamics is described by general relativity. In this stage, the universe is radiation dominated,
gravity is the only relevant force and the perturbations are linear. Next, when λ < dH and z > zeq one can describe
the dynamics by Newtonian considerations. The perturbations are still linear and the universe is radiation dominated.
Finally, when λ < dH, z < zeq we have a matter dominated universe in which we can use the Newtonian formalism;
but at this stage — when most astrophysical structures form — we need to grapple with nonlinear astrophysical
processes.
Let us now consider the metric perturbation in greater detail. When the metric is perturbed to the form: gab →
gab + hab the perturbation can be split as hab = (h00, h0α ≡ wα, hαβ). We also know that any 3-vector w(x) can be
split as w = w⊥
+ w in which w = ∇Φ is curl-free (and carries one degree of freedom) while w⊥
is divergence-free
(and has 2 degrees of freedom). This result is obvious in k−space since we can write any vector w(k) as a sum of two
terms, one along k and one transverse to k:
w(k) = w (k) + w⊥
(k) = k
w(k) · k
k2
along k
+ w(k) − k
k · w(k)
k2
transverse to k
; k × w = 0; k · w⊥
= 0 (45)
Fourier transforming back, we can split w into a curl-free and divergence-free parts. Similar decomposition works for
hαβ by essentially repeating the above analysis on each index. We can write:
hαβ = ψδαβ
trace
+ ∇αu⊥
β + ∇βu⊥
α
traceless from vector
+ ∇α∇β −
1
3
δαβ∇2
Φ1
traceless from scalar
+h⊥⊥
αβ ⇒ 1 + 2 + 1 + 2 = 6 (46)
The u⊥
α is divergence free and h⊥⊥
αβ is traceless and divergence free. Thus the most general perturbation hab (ten
degrees of freedom) can be built out of
hab = (h00, h0α ≡ wα, hαβ) = [h00, (Φ , w⊥
), (ψ, Φ1, u⊥
α , h⊥⊥
αβ )] ⇒ [1, (1, 2), (1, 1, 2, 2)] (47)
We now use the freedom available in the choice of four coordinate transformations to set four conditions: Φ = Φ1 = 0
and u⊥
α = 0 thereby leaving six degrees of freedom in (h00 ≡ 2Φ, ψ, w⊥
, h⊥⊥
αβ ) as nonzero. Then the perturbed line
element takes the form:
ds2
= a2
(η) [{1 + 2Φ(x, η)} dη2
− 2w⊥
α (x, η) dη dxα
− {(1 − 2ψ(x, η)) δαβ + 2h⊥⊥
αβ (x, η)} dxα
dxβ
] (48)
12
To make further simplification we need to use two facts from Einstein’s equations. It turns out that the Einstein’s
equations for w⊥
and h⊥⊥
αβ decouple from those for (Φ, ψ). Further, in the absence of anisotropic stress, one of the
equations give ψ = Φ. If we use these two facts, we can simplify the structure of perturbed metric drastically. As far
as the growth of matter perturbations are concerned, we can ignore w⊥
α and h⊥⊥
αβ and work with a simple metric:
ds2
= a2
(η)[(1 + 2Φ)dη2
− (1 − 2Φ)δαβdxα
dxβ
] (49)
with just one perturbed scalar degree of freedom in Φ. This is what we will study.
Having decided on the gauge, let us consider the evolution equations for the perturbations. While one can directly
work with the Einstein’s equations, it turns out to be convenient to use the equations of motion for matter variables,
since we are eventually interested in the matter perturbations. In what follows, we will use the over-dot to denote
(d/dη) so that the standard Hubble parameter is H = (1/a)(da/dt) = ˙a/a2
. With this notation, the continuity
equation becomes:
˙ρ + 3 aH − ˙Φ (ρ + p) = −∇α [(ρ + p)vα
] (50)
Since the momentum flux in the relativistic case is (ρ + p)vα
, all the terms in the above equation are intuitively
obvious, except probably the ˙Φ term. To see the physical origin of this term, note that the perturbation in Eq. (49)
changes the factor in front of the spatial metric from a2
to a2
(1 − 2Φ) so that ln a → ln a − Φ; hence the effective
Hubble parameter from (˙a/a) to (˙a/a) − ˙Φ which explains the extra ˙Φ term. This is, of course, the exact equation for
matter variables in the perturbed metric given by Eq. (49); but we only need terms which are of linear order. Writing
the curl-free velocity part as vα
= ∇α
v, the linearised equations, for dark matter (with p = 0) and radiation (with
p = (1/3)ρ) perturbations are given by:
˙δm =
d
dη
δnm
nm
= ∇2
vm + 3 ˙Φ;
3
4
˙δR =
d
dη
δnR
nR
= ∇2
vR + 3 ˙Φ (51)
where nm and nR are the number densities of dark matter particles and radiation. The same equations in Fourier
space [using the same symbols for, say, δ(t, x) or δ(t, k)] are simpler to handle:
˙δm =
d
dη
δnm
nm
= −k2
vm + 3 ˙Φ;
3
4
˙δR =
d
dη
δnR
nR
= −k2
vR + 3 ˙Φ (52)
Note that these equations imply
d
dη
δnR
nR
−
δnm
nm
=
d
dη
δ ln
nR
nm
=
d
dη
δ ln
s
nm
= −k2
(vR − vm) (53)
For long wavelength perturbations (in the limit of k → 0), this will lead to the conservation of perturbation δ(s/nm)
in the entropy per particle.
Let us next consider the Euler equation which has the general form:
∂η[(ρ + p)vα
] = −(ρ + p)∇α
Φ − ∇α
p − 4aH(ρ + p)vα
(54)
Once again each of the terms is simple to interpret. The (ρ + p) arises because the pressure also contributes to inertia
in a relativistic theory and the factor 4 in the last term on the right hand side arises because the term vα
∂η(ρ + p)
on the left hand side needs to be compensated. Taking the linearised limit of this equation, for dark matter and
radiation, we get:
˙vm = Φ − aHvm; ˙vR = Φ +
1
4
δR (55)
Thus we now have four equations in Eqs. (52), (55) for the five variables (δm, δR, vm, vR, Φ). All we need to do is to
pick one more from Einstein’s equations to complete the set. The Einstein’s equations for our perturbed metric are:
0
0 component : k2
Φ + 3
˙a
a
˙Φ +
˙a
a
Φ = −4πGa2
A
ρAδA = −4πGa2
ρbgδtotal (56)
0
α component : ˙Φ +
˙a
a
Φ = −4πGa2
A
(ρ + p)AvA; v = ∇v (57)
α
α component : 3
˙a
a
˙Φ + 2
¨a
a
Φ −
˙a2
a2
Φ + ¨Φ = 4πGa2
δp (58)
13
where A denotes different components like dark matter, radiation etc. Using Eq. (57) in Eq. (56) we can get a modified
Poisson equation which is purely algebraic:
−k2
Φ = 4πGa2
A
ρAδA − 3
˙a
a
(ρA + pA)vA (59)
which once again emphasizes the fact that in the relativistic theory, both pressure and density act as source of gravity.
To get a feel for the solutions let us consider a flat universe dominated by a single component of matter with the
equation of state p = wρ. (A purely radiation dominated universe, for example, will have w = 1/3.) In this case the
Friedmann background equation gives ρ ∝ a−3(1+w)
and
˙a
a
=
2
(1 + 3w)η
;
¨a
a
=
2(1 − 3w)
(1 + 3w)2η2
(60)
The equation for the potential Φ can be reduced to the form:
¨Φ +
6(1 + w)
1 + 3w
˙Φ
η
+ k2
wΦ = 0 (61)
The second term is the damping due to the expansion while last term is the pressure support that will lead to
oscillations. Clearly, the factor kη determines which of these two terms dominates. When the pressure term dominates
(kη ≫ 1), we expect oscillatory behaviour while when the background expansion dominates (kη ≪ 1), we expect the
growth to be suppressed. This is precisely what happens. The exact solution is given in terms of the Bessel functions
Φ(η) =
C1(k)Jν/2(
√
wkη) + C2(k)Yν/2(
√
wkη)
ην/2
; ν =
5 + 3w
1 + 3w
(62)
From the theory of Bessel functions, we know that:
lim
x→0
Jν/2(x) ≃
xν/2
2ν/2Γ(ν/2 + 1)
; lim
x→0
Yν/2(x) ∝ −
1
xν/2
(63)
This shows that if we want a finite value for Φ as η → 0, we can set C2 = 0. This gives the gravitational potential to
be
Φ(η) =
C1(k)Jν/2(
√
wkη)
ην/2
; ν =
5 + 3w
1 + 3w
(64)
The corresponding density perturbation will be:
δ = −2Φ −
(1 + 3w)2
k2
η2
6
C1(k)Jν/2(
√
wkη)
ην/2
+ (1 + 3w)
√
wkη
C1(k)J(ν/2)+1(
√
wkη)
ην/2
(65)
To understand the nature of the solution, note that dH = (˙a/a)−1
∝ η and kdH ≃ dH/λ ∝ kη. So the argument of
the Bessel function is just the ratio (dH/λ). From the theory of Bessel functions, we know that for small values of the
argument Jν(x) ∝ xν
is a power law while for large values of the argument it oscillates with a decaying amplitude:
lim
x→∞
Jν/2(x) ∼
cos[x − (ν − 1)π/4]
√
x
; (66)
Hence, for modes which are still outside the Hubble radius (k ≪ η−1
), we have a constant amplitude for the potential
and density contrast:
Φ ≈ Φi(k); δ ≈ −2Φi(k) (67)
That is, the perturbation is frozen (except for a decaying mode) at a constant value. On the other hand, for modes
which are inside the Hubble radius (k ≫ η−1
), the perturbation is rapidly oscillatory (if w = 0). That is the pressure
is effective at small scales and leads to acoustic oscillations in the medium.
14
A special case of the above is the flat, matter-dominated universe with w = 0. In this case, we need to take the
w → 0 limit and the general solution is indeed a constant Φ = Φi(k) (plus a decaying mode Φdecay ∝ η−5
which
diverges as η → 0). The corresponding density perturbations is:
δ = −(2 +
k2
η2
6
)Φi(k) (68)
which shows that density perturbation is “frozen” at large scales but grows at small scales:
δ =
−2Φi(k) = constant (kη ≪ 1)
−1
6 k2
η2
Φi(k) ∝ η2
∝ a (kη ≫ 1)
(69)
We will use these results later on.
IV. PERTURBATIONS IN DARK MATTER AND RADIATION
We shall now move on to the more realistic case of a multi-component universe consisting of radiation and collisionless
dark matter. (For the moment we are ignoring the baryons, which we will study in Sec. VI). It is convenient
to use y = a/aeq as independent variable rather than the time coordinate. The background expansion of the universe
described by the function a(t) can be equivalently expressed (in terms of the conformal time η) as
y ≡
ρM
ρR
=
a
aeq
= x2
+ 2x, x ≡
ΩM
4aeq
1/2
H0η (70)
It is also useful to define a critical wave number kc by:
k2
c =
H2
0 Ωm
aeq
= 4(
√
2 − 1)2
η−2
eq = 4(
√
2 − 1)2
k2
eq; k−1
c = 19(Ωmh2
)−1
Mpc (71)
which essentially sets the comoving scale corresponding to matter-radiation equality. Note that 2x = kcη and y ≈ kcη
in the radiation dominated phase while y = (1/4)(kcη)2
in the matter dominated phase.
We now manipulate Eqs. (52), (55), (56), (57) governing the growth of perturbations by essentially eliminating the
velocity. This leads to the three equations
yΦ′
+ Φ +
1
3
k2
k2
c
y2
1 + y
Φ = −
1
2
y
1 + y
δm +
1
y
δR (72)
(1 + y)δ′′
m +
2 + 3y
2y
δ′
m = 3(1 + y)Φ′′
+
3(2 + 3y)
2y
Φ′
−
k2
k2
c
Φ (73)
(1 + y)δ′′
R +
1
2
δ′
R +
1
3
k2
k2
c
δR = 4(1 + y)Φ′′
+ 2Φ′
−
4
3
k2
k2
c
Φ (74)
for the three unknowns Φ, δm, δR. Given suitable initial conditions we can solve these equations to determine the
growth of perturbations. The initial conditions need to imposed very early on when the modes are much bigger than
the Hubble radius which corresponds to the y ≪ 1, k → 0 limit. In this limit, the equations become:
yΦ′
+ Φ ≈ −
1
2
δR; δ′′
m +
1
y
δ′
m ≈ 3Φ′′
+
3
y
Φ′
; δ′′
R +
1
2
δ′
R ≈ 4Φ′′
+ 2Φ′
(75)
We will take Φ(yi, k) = Φi(k) as given value, to be determined by the processes that generate the initial perturbations.
First equation in Eq. (75) shows that we can take δR = −2Φi for yi → 0. Further Eq. (53) shows that adiabaticity
is respected at these scales and we can take δm = (3/4)δR = −(3/2)Φi;. The exact equation Eq. (72) determines Φ′
if (Φ, δm, δR) are given. Finally we use the last two equations to set δ ′
m = 3Φ ′
, δ ′
R = 4Φ ′
, Thus we take the initial
conditions at some y = yi ≪ 1 to be:
Φ(yi, k) = Φi(k); δR(yi, k) = −2Φi(k); δm(yi, k) = −(3/2)Φi(k) (76)
with δ ′
m(yi, k) = 3Φ ′
(yi, k); δ ′
R(yi, k) = 4Φ ′
(yi, k).
Given these initial conditions, it is fairly easy to integrate the equations forward in time and the numerical results
are shown in Figs 2, 3, 4, 5. (In the figures keq is taken to be aeqHeq.) To understand the nature of the evolution, it
is, however, useful to try out a few analytic approximations to Eqs. (72) – (74) which is what we will do now.
15
-6 -4 -2 0 2 4
Log10 a aeq
0
0.2
0.4
0.6
0.8
1
kkeq
32 k 0.01keq
k keq
k 100keq
-6 -4 -2 0 2 4
Log10 a aeq
0
1
2
3
4
5
Log101Σ
k 100keq
k keq
k 0.01keq
FIG. 2: Left Panel:The evolution of gravitational potential Φ for 3 different modes. The wavenumber is indicated by the label
and the epoch at which the mode enters the Hubble radius is indicated by a small arrow. The top most curve is for a mode
which stays outside the Hubble radius for most of its evolution and is well described by Eq. (78). The other two modes show
the decay of Φ after the mode has entered the Hubble radius in the radiation dominated epoch as described by Eq. (79). Right
Panel: Evolution of entropy perturbation (see Eq. (87) for the definition). The entropy perturbation is essentially zero till the
mode enters Hubble radius and grows afterwards tracking the dominant energy density perturbation.
A. Evolution for λ ≫ dH
Let us begin by considering very large wavelength modes corresponding to the kη → 0 limit. In this case adiabaticity
is respected and we can set δR ≈ (4/3)δm. Then Eqs. (72), (73) become
yΦ′
+ Φ ≈ −
3y + 4
8(1 + y)
δR; δ′
R ≈ 4Φ′
(77)
Differentiating the first equation and using the second to eliminate δm, we get a second order equation for Φ. Fortunately,
this equation has an exact solution
Φ = Φi
1
10y3
16 (1 + y) + 9y3
+ 2y2
− 8y − 16 ; δR ≈ 4Φ − 6Φi (78)
[There is simple way of determining such an exact solution, which we will describe in Sec. IV D.]. The initial condition
on δR is chosen such that it goes to −2Φi initially. The solution shows that, as long as the mode is bigger than the
Hubble radius, the potential changes very little; it is constant initially as well as in the final matter dominated phase.
At late times (y ≫ 1) we see that Φ ≈ (9/10)Φi so that Φ decreases only by a factor (9/10) during the entire evolution
if k → 0 is a valid approximation.
B. Evolution for λ ≪ dH in the radiation dominated phase
When the mode enters Hubble radius in the radiation dominated phase, we can no longer ignore the pressure terms.
The pressure makes radiation density contrast oscillate and the gravitational potential, driven by this, also oscillates
with a decay in the overall amplitude. An approximate procedure to describe this phase is to solve the coupled δR −Φ
system, ignoring δm which is sub-dominant and then determine δm using the form of Φ.
When δm is ignored, the problem reduces to the one solved earlier in Eqs (64), (65) with w = 1/3 giving ν = 3.
Since J3/2 can be expressed in terms of trigonometric functions, the solution given by Eq. (64) with ν = 3, simplifies
to
Φ = Φi
3
l3y3
[sin(ly) − ly cos(ly)] ; l2
=
k2
3k2
c
(79)
Note that as y → 0, we have Φ = Φi, Φ′
= 0. This solution shows that once the mode enters the Hubble radius, the
potential decays in an oscillatory manner. For ly ≫ 1, the potential becomes Φ ≈ −3Φi(ly)−2
cos(ly). In the same
16
limit, we get from Eq. (65) that
δR ≈ −
2
3
k2
η2
Φ ≈ −2l2
y2
Φ ≈ 6Φi cos(ly) (80)
(This is analogous to Eq. (68) for the radiation dominated case.) This oscillation is seen clearly in Fig 3 and Fig.4
(left panel). The amplitude of oscillations is accurately captured by Eq. (80) for k = 100keq mode but not for k = keq;
this is to be expected since the mode is not entering in the radiation dominated phase.
Let us next consider matter perturbations during this phase. They grow, driven by the gravitational potential
determined above. When y ≪ 1, Eq.(73) becomes:
δ′′
m +
1
y
δ′
m = 3Φ′′
+
3
y
Φ′
−
k2
k2
c
Φ (81)
The Φ is essentially determined by radiation and satisfies Eq. (61); using this, we can rewrite Eq. (81) as
d
dy
(yδ′
m) = −9(Φ′
+
2
3
l2
yΦ) (82)
The general solution to the homogeneous part of Eq. (82) (obtained by ignoring the right hand side) is (c1 + c2 ln y);
hence the general solution to this equation is
δm = (c1 + c2 ln y) − 9
dy
y
y
dy1[Φ′
(y1) +
2
3
l2
y1Φ(y1)] (83)
For y ≪ 1 the growing mode varies as ln y and dominates over the rest; hence we conclude that, matter, driven by Φ,
grows logarithmically during the radiation dominated phase for modes which are inside the Hubble radius.
C. Evolution in the matter dominated phase
Finally let us consider the matter dominated phase, in which we can ignore the radiation and concentrate on
Eq. (72) and Eq. (73). When y ≫ 1 these equations become:
yΦ′
+ Φ ≈ −
1
2
δm −
k2
y
3k2
c
Φ; yδ′′
m +
3
2
δ′
m = −
k2
k2
c
Φ (84)
These have a simple solution which we found earlier (see Eq. (69)):
Φ = Φ∞ = const.; δm = −2Φ∞ −
2k2
3k2
c
Φ∞y ∼ y (85)
In this limit, the matter perturbations grow linearly with expansion: δm ∝ y ∝ a. In fact this is the most dominant
growth mode in the linear perturbation theory.
-6 -4 -2 0 2 4
Log10 a aeq
-6
-4
-2
0
2
4
6
kkeq
32
∆R
k 100keq
FIG. 3: Evolution of δR for a mode with k = 100keq. The mode remains frozen outside the Hubble radius at (k/keq )3/2
(−δR) ≈
(k/keq)3/2
2Φ = 2 (in the normalisation used in Fig. 2 ) and oscillates when it enters the Hubble radius. The oscillations are
well described by Eq. (80) with an amplitude of 6.
17
-6 -4 -2 0 2 4
Log10 a aeq
1
2
3
4
5
kkeq
32
∆R
k keq
-6 -4 -2 0 2 4
Log10 a aeq
2
2.25
2.5
2.75
3
3.25
3.5
kkeq
32
∆R
k 0.01keq
FIG. 4: Evolution of δR for two modes k = keq and k = 0.01 keq. The modes remain frozen outside the Hubble radius at
(−δR) ≈ 2 and oscillates when it enters the Hubble radius. The mode in the right panel stays outside the Hubble radius for
most part of its evolution and hence changes very little.
D. An alternative description of matter-radiation system
Before proceeding further, we will describe an alternative procedure for discussing the perturbations in dark matter
and radiation, which has some advantages. In the formalism we used above, we used perturbations in the energy
density of radiation (δR) and matter (δm) as the dependent variables. Instead, we now use perturbations in the total
energy density, δ and the perturbations in the entropy per particle, σ as the new dependent variables. In terms of
δR, δm, these variables are defined as:
δ ≡
δρtotal
ρtotal
=
ρRδR + ρmδm
ρR + ρm
=
δR + yδm
1 + y
; y =
ρm
ρR
=
a
aeq
(86)
σ ≡
δs
s
=
3δTR
TR
−
δρm
ρm
=
3
4
δR − δm =
δnR
nR
−
δnm
nm
(87)
-6 -4 -2 0 2 4
Log10 a aeq
1
2
3
4
5
Log10kkeq
32
∆M
k 0.01keq
k keq
k 100keq
FIG. 5: Evolution of |δm| for 3 different modes. The modes are labelled by their wave numbers and the epochs at which they
enter the Hubble radius are shown by small arrows. All the modes remain frozen when they are outside the Hubble radius and
grow linearly in the matter dominated phase once they are inside the Hubble radius. The mode that enters the Hubble radius
in the radiation dominated phase grows logarithmically until y = yeq. These features are well approximated by Eqs. (83), (85).
18
Given the equations for δR, δm, one can obtain the corresponding equations for the new variables (δ, σ) by straight
forward algebra. It is convenient to express them as two coupled equations for Φ and σ. After some direct but a bit
tedious algebra, we get:
yΦ′′
+
yΦ′
2(1 + y)
+ 3(1 + c2
s)Φ′
+
3c2
sΦ
4(1 + y)
+ c2
s
k2
k2
c
y
1 + y
Φ =
3c2
sσ
2(1 + y)
(88)
yσ′′
+
yσ′
2(1 + y)
+ 3c2
sσ′
+
3c2
sy2
4(1 + y)
k2
k2
c
σ =
c2
sy3
2(1 + y)
k
kc
4
Φ (89)
where we have defined
c2
s =
(4/3)ρR
4ρR + 3ρm
=
1
3
1 +
3
4
ρm
ρR
−1
=
1
3
1 +
3
4
y
−1
(90)
These equations show that the entropy perturbations and gravitational potential (which is directly related to total
energy density perturbations) act as sources for each other. The coupling between the two arises through the right
hand sides of Eq. (88) and Eq. (89). We also see that if we set σ = 0 as an initial condition, this is preserved to
O(k4
) and — for long wave length modes — the Φ evolves independent of σ. The solutions to the coupled equations
obtained by numerical integration is shown in Fig.(2) right panel. The entropy perturbation σ ≈ 0 till the mode enters
Hubble radius and grows afterwards tracking either δR or δm whichever is the dominant energy density perturbation.
To illustrate the behaviour of Φ, let us consider the adiabatic perturbations at large scales with σ ≈ 0, k → 0; then
the gravitational potential satisfies the equation:
yΦ′′
+
yΦ′
2(1 + y)
+ 3(1 + c2
s)Φ′
+
3c2
sΦ
4(1 + y)
=
3c2
sσ
2(1 + y)
≈ 0 (91)
which has the two independent solutions:
f1(y) = 1 +
2
9y
−
8
9y2
−
16
9y3
, f2(y) =
√
1 + y
y2
(92)
both of which diverge as y → 0. We need to combine these two solutions to find the general solution, keeping in mind
that the general solution should be nonsingular and become a constant (say, unity) as y → 0. This fixes the linear
combination uniquely:
f(y) =
9
10
f1 +
8
5
f2 =
1
10y3
16 (1 + y) + 9y3
+ 2y2
− 8y − 16 (93)
Multiplying by Φi we get the solution that was found earlier (see Eq. (78)). Given the form of Φ and σ ≃ 0 we can
determine all other quantities. In particular, we get:
δR =
−2(1 + y)d(yΦ)/dy + yσ
1 + (3/4)y
≃ −
2(1 + y)
1 + (3/4)y
d
dy
(yΦ) (94)
The corresponding velocity field, which we quote for future reference, is given by:
vα = −
3c2
s
2(˙a/a)
(1 + y)∇α
d(yΦ)
dy
(95)
We conclude this section by mentioning another useful result related to Eq. (88). When σ ≈ 0, the equation for Φ
can be re-expressed as
a
dζ
da
= −
2c2
s
3
k2
/a2
H2
ρ
ρ + p
Φ ≈ 0 (for
k
aH
≪ 1) (96)
where we have defined:
ζ =
2
3
ρ
ρ + p
a
˙a
˙Φ +
˙a
a
Φ + Φ =
H
ρ + p
ikα
k2
δT 0
α + Φ (97)
19
(The i factor arises because of converting a gradient to the k space; of course, when everything is done correctly, all
physical quantities will be real.) Other equivalent alternative forms for ζ, which are useful are:
ζ =
2
3[1 + w(a)]
d
da
(aΦ) + Φ =
H2
a(ρ + p)
d
dt
aΦ
H
(98)
For modes which are bigger than the Hubble radius, Eq. (96) shows that ζ is conserved. When ζ=constant, we can
integrate Eq. (98) easily to obtain:
Φ = c1
H
a
+ c2 1 −
H
a
a
0
da′
H(a′)
(99)
This is the easiest way to obtain the solution in Eq. (78).
The conservation law for ζ also allows us to understand in a simple manner our previous result that Φ only
deceases by a factor (9/10) when the mode remains bigger than Hubble radius as we evolve the equations from
y ≪ 1 to y ≫ 1. Let us compare the values of ζ early in the radiation dominated phase and late in the matter
dominated phase. From the first equation in Eq. (98), [using Φ′
≈ 0] we find that, in the radiation dominated phase,
ζ ≈ (1/2)Φi+Φi = (3/2)Φi; late in the matter dominated phase, ζ ≈ (2/3)Φf +Φf = (5/3)Φf . Hence the conservation
of ζ gives Φf = (3/5)(3/2)Φi = (9/10)Φi which was the result obtained earlier. The expression in Eq. (99) also works
at late times in the Λ dominated or curvature dominated universe.
One key feature which should be noted in the study of linear perturbation theory is the different amount of growths
for Φ, δR and δm. The Φ either changes very little or decays; the δR grows in amplitude only by a factor of few. The
physical reason, of course, is that the amplitude is frozen at super-Hubble scales and the pressure prevents the growth
at sub-Hubble scales. In contrast, δm, which is pressureless, grows logarithmically in the radiation dominated era and
linearly during the matter dominated era. Since the later phase lasts for a factor of 104
in expansion, we get a fair
amount of growth in δm.
V. TRANSFER FUNCTION FOR MATTER PERTURBATIONS
We now have all the ingredients to evolve the matter perturbation from an initial value δ = δi at y = yi ≪ 1 to the
current epoch y = y0 = a−1
eq in the matter dominated phase at y ≫ 1. Initially, the wavelength of the perturbation
will be bigger than the Hubble radius and the perturbation will essentially remain frozen. When it enters the Hubble
radius in the radiation dominated phase, it begins to grow but only logarithmically (see section IV B ) until the
universe becomes matter dominated. In the final matter dominated phase, the perturbation grows linearly with
expansion factor. The relation between final and initial perturbation can be obtained by combining these results.
Usually, one is more interested in the power spectrum Pk(t) and the power per logarithmic band in k−space ∆k(t).
These quantities are defined in terms of δk(t) through the equations:
Pk(t) ≡ |δk(t)|2
; ∆2
k(t) ≡
k3
Pk(t)
2π2
(100)
It is therefore convenient to study the evolution of k3/2
δk since its square will immediately give the power per
logarithmic band ∆2
k in k−space.
Let us first consider a mode which enters the Hubble radius in the radiation dominated phase at the epoch aenter.
From the scaling relation, aent/k ∝ tent ∝ a2
ent we find that yent = (keq/k). Hence
k3/2
δm(k, a = 1) =
1
aeq
MD
ln
aeq
aent
RD
k3/2
δent(k)
at entry
∝ ln
k
keq
[k3/2
δent(k)] (101)
where two factors — as indicated — gives the growth in radiation (RD) and matter dominated (MD) phases. Let
us next consider the modes that enter in the matter dominated phase. In this case, aent/k ∝ tent ∝ a
3/2
ent so that
yent = (keq/k)2
. Hence
k3/2
δm(k, a = 1) =
1
aent
MD
k3/2
δent(k)
at entry
∝ k2
[k3/2
δent(k)] (102)
20
To proceed further, we need to know the k−dependence of the perturbation when it enters the Hubble radius which,
of course, is related to the mechanism that generates the initial power spectrum. The most natural choice will be
that all the modes enter the Hubble radius with a constant amplitude at the time of entry. This would imply that the
physical perturbations are scale invariant at the time of entering the Hubble radius, a possibility that was suggested
by Zeldovich and Harrison [18] (years before inflation was invented!). We will see later that this is also true for
perturbations generated by inflation and thus is a reasonable assumption at least in such models. Hence we shall
assume
k3
|δent(k)|2
= k3
Pent(k) = C = constant, (103)
Using this we find that the current value of perturbation is given by
P(k, a = 1) ∝ δm(k, a = 1)
2
∝
k (for k ≪ keq)
k−3
(ln k)2
(for k ≫ keq)
(104)
The corresponding power per logarithmic band is
∆2
(k, a = 1) ∝ k3
δm(k, a = 1)
2
∝
k4
(for k ≪ keq)
(ln k)2
(for k ≫ keq)
(105)
The form for P(k) shows that the evolution imprints the scale keq on the power spectrum even though the initial
power spectrum is scale invariant. For k < keq (for large spatial scales), the primordial form of the spectrum is
preserved and the evolution only increases the amplitude preserving the shape. For k > keq (for small spatial scales),
the shape is distorted and in general the power is suppressed in comparison with larger spatial scales. This arises
because modes with small wavelengths enter the Hubble radius early on and have to wait till the universe becomes
matter dominated in order to grow in amplitude. This is in contrast to modes with large wavelengths which continue
to grow. It is this effect which suppresses the power at small wavelengths (for k > keq) relative to power at larger
wavelengths.
VI. TEMPERATURE ANISOTROPIES OF CMBR
We shall now apply the formalism we have developed to understand the temperature anisotropies in the cosmic
microwave background radiation which is probably the most useful application of linear perturbation theory. We
shall begin by developing the general formulation and the terminology which is used to describe the temperature
anisotropies.
Towards every direction in the sky, n = (θ, ψ) we can define a fractional temperature fluctuation ∆(n) ≡
(∆T/T )(θ, ψ). Expanding this quantity in spherical harmonics on the sky plane as well as in terms of the spatial
Fourier modes, we get the two relations:
∆(n) ≡
∆T
T
(θ, ψ)=
∞
l,m
almYlm(θ, ψ)=
d3
k
(2π)
∆(k)eik·nL
(106)
where L = η0 −ηLSS is the distance to the last scattering surface (LSS) from which we are receiving the radiation. The
last equality allows us to define the expansion coefficients alm in terms of the temperature fluctuation in the Fourier
space ∆(k) . Standard identities of mathematical physics now give
alm =
d3
k
(2π)3
(4π) il
∆(k) jl(kL) Ylm(ˆk) (107)
Next, let us consider the angular correlation function of temperature anisotropy, which is given by:
C(α) = ∆(n)∆(m) = alma∗
l′m′ Ylm(n)Y ∗
l′m′ (m). (108)
where the wedges denote an ensemble average. For a Gaussian random field of fluctuations we can express the ensemble
average as alma∗
l′m′ = Clδll′ δmm′ . Using Eq. (107), we get a relation between Cl and ∆(k). Given ∆(k), the Cl’s
are given by:
Cl =
2
π
∞
0
k2
dk |∆(k)|2
j2
l (kL) (109)
21
Further, Eq. (108) now becomes:
C(α) =
l
(2l + 1)
4π
ClPl(cos α) (110)
Equation (110) shows that the mean-square value of temperature fluctuations and the quadrupole anisotropy corresponding
to l = 2 are given by
∆T
T
2
rms
= C(0) =
1
4π
∞
l=2
(2l + 1) Cl,
∆T
T
2
Q
=
5
4π
C2. (111)
These can be explicitly computed if we know ∆(k) from the perturbation theory. (The motion of our local group
through the CMBR leads to a large l = 1 dipole contribution in the temperature anisotropy. In the analysis of CMBR
anisotropies, this is usually subtracted out. Hence the leading term is the quadrupole with l = 2.)
It should be noted that, for a given l, the Cl is the average over all m = −l, ... − 1, 0, 1, ...l. For a Gaussian
random field, one can also compute the variance around this mean value. It can be shown that this variance in Cl is
2C2
l /(2l + 1). In other words, there is an intrinsic root-mean-square fluctuation in the observed, mean value of Cl’s
which is of the order of ∆Cl/Cl ≈ (2l +1)−1/2
. It is not possible for any CMBR observations which measures the Cl’s
to reduce its uncertainty below this intrinsic variance — usually called the “cosmic variance”. For large values of l,
the cosmic variance is usually sub-dominant to other observational errors but for low l this is the dominant source of
uncertainty in the measurement of Cl’s. Current WMAP observations are indeed only limited by cosmic variance at
low-l.
As an illustration of the formalism developed above, let us compute the Cl’s for low l which will be contributed
essentially by fluctuations at large spatial scales. Since these fluctuations will be dominated by gravitational effects, we
can ignore the complications arising from baryonic physics and compute these using the formalism we have developed
earlier.
We begin by noting that the redshift law of photons in the unperturbed Friedmann universe, ν0 = ν(a)/a, gets
modified to the form ν0 = ν(a)/[a(1 + Φ)] in a perturbed FRW universe. The argument of the Planck spectrum will
thus scale as
ν0
T0
=
ν(a)
aT0(1 + Φ)
=
ν(a)
a T0 [1 + (δR/4)](1 + Φ)
∼=
ν(a)
a T0 [1 + Φ + (δR/4)]
(112)
This is equivalent to a temperature fluctuation of the amount
∆T
T obs
=
1
4
δR + Φ (113)
at large scales. (Note that the observed ∆T/T is not just (δR/4) as one might have naively imagined.) To proceed
further, we recall our large scale solution (see Eq. (78)) for the gravitational potential:
Φ = Φi
1
10y3
16 (1 + y) + 9y3
+ 2y2
− 8y − 16 ; δR = 4Φ − 6Φi (114)
At y = ydec we can take the asymptotic solution Φdec ≈ (9/10)Φi. Hence we get
∆T
T obs
=
1
4
δR + Φ
dec
= 2Φdec −
3
2
Φi ≈ 2Φdec −
3
2
10
9
Φdec =
1
3
Φdec (115)
We thus obtain the nice result that the observed temperature fluctuations at very large scales is simply related to
the fluctuations of the gravitational potential at these scales. (For a discussion of the 1/3 factor, see [19]). Fourier
transforming this result we get ∆(k) = (1/3)Φ(k, ηLSS
) where ηLSS
is the conformal time at the last scattering
surface. (This contribution is called Sachs-Wolfe effect.) It follows from Eq. (109) that the contribution to Cl from
the gravitational potential is
Cl =
2
π
k2
dk|∆(k)|2
j2
l (kL) =
2
π
∞
0
dk
k
k3
|Φk|2
9
j2
l (kL) (116)
with
L = η0 − ηLSS
≈ η0 ≈ 2(ΩmH2
0 )−1/2
≈ 6000 Ω−1/2
m h−1
Mpc (117)
22
For a scale invariant spectrum, k3
|Φk|2
is a constant independent of k. (Earlier on, in Eq. (103) we said that scale
invariant spectrum has k3
|δk|2
= constant. These statements are equivalent since δ ≈ −2Φ at the large scales because
of Eq. (85) with the extra correction term in Eq. (85) being about 3 × 10−4
for k ≈ L−1
, y = ydec.) As we shall see
later, inflation generates such a perturbation. In this case, it is conventional to introduce a constant amplitude A and
write:
∆2
Φ ≡
k3
|Φk|2
2π2
= A2
= constant (118)
Substituting this form into Eq. (116) and evaluating the integral, we find that
l(l + 1)Cl
2π
=
A
3
2
(119)
As an application of this result, let us consider the observations of COBE which measured the temperature fluctuations
for the first time in 1992. This satellite obtained the RMS fluctuations and the quadrupole after smoothing over an
angular scale of about θc ≈ 10◦
. Hence the observed values are slightly different from those in Eq. (111). We have,
instead,
∆T
T
2
rms
=
1
4π
∞
l=2
(2l + 1)Cl exp −
l2
θ2
c
2
;
∆T
T
2
Q
=
5
4π
C2e−2θ2
c . (120)
Using Eqs. (118), (119) we find that
∆T
T Q
∼= 0.22A;
∆T
T rms
∼= 0.51A. (121)
Given these two measurements, one can verify that the fluctuations are consistent with the scale invariant spectrum
by checking their ratio. Further, the numerical value of the observed (∆T/T ) can be used to determine the amplitude
A. One finds that A ≈ 3 × 10−5
which sets the scale of fluctuations in the gravitational potential at the time when
the perturbation enters the Hubble radius.
Incidentally, note that the solution δR = 4Φ − 6Φi corresponds to δm = (3/4)δR = 3Φ − (9/2)Φi. At y = ydec,
taking Φdec = (9/10)Φi, we get δm = 3Φdec − (9/2)(10/9)Φdec = −2Φdec. Since (∆T/T )obs = (1/3)Φdec we get
δm = −6(∆T/T )obs. This shows that the amplitude of matter perturbations is a factor six larger that the amplitude
of temperature anisotropy for our adiabatic initial conditions. In several other models, one gets δm = O(1)(∆T/T )obs.
So, to reach a given level of nonlinearity in the matter distribution at later times, these models will require higher
values of (∆T/T )obs at decoupling. This is one reason for such models to be observationally ruled out.
There is another useful result which we can obtain from Eq. (109) along the same lines as we derived the SachsWolfe
effect. Whenever k3
|∆(k)|2
is a slowly varying function of k, we can pull out this factor out of the integral and
evaluate the integral over j2
l . This will give the result for any slowly varying k3
|∆(k)|2
l(l + 1)Cl
2π
≈
k3
|∆(k)|2
2π2
kL≈l
(122)
This is applicable even when different processes contribute to temperature anisotropies as long as they add in quadrature.
While far from accurate, it allows one to estimate the effects rapidly.
A. CMBR Temperature Anisotropy: More detailed theory
We shall now work out a more detailed theory of temperature anisotropies of CMBR so that one can understand
the effects at small scales as well. A convenient starting point is the distribution function for photons with perturbed
Planckian distribution, which we can write as:
f(xα
, η, E, nα
) =
Iν
2πν3
= fP
aE
1 + ∆
; fP (ǫ) ≡ 2 [exp (ǫ/T0) − 1]
−1
(123)
The fp(ǫ) is the standard Planck spectrum for energy ǫ and we take ǫ = aE(1+∆)−1
to take care of the perturbations.
In the absence of collisions, the distribution function is conserved along the trajectories of photons so that df/dη = 0.
So, in the presence of collisions, we can write the time evolution of the distribution function as
df
dη
=
aE
1 + ∆
f′
P
aE
1 + ∆
d ln(aE)
dη
−
d∆
dη
=
df
dη coll
(124)
23
where the right hand side gives the contribution due to collisional terms. Equivalently, in terms of ∆, the same
equation takes the form:
d∆
dη
−
d ln(aE)
dη
= −
1 + ∆
aE
[f′
P ]−1 df
dη coll
≡
d∆
dη coll
(125)
To proceed further, we need the expressions for the two terms on the left hand side. First term, on using the standard
expansion for total derivative, gives:
d∆
dη
=
∂∆
∂η
+
∂∆
∂xα
dxα
dη
nα
+
∂∆
∂E
zero
dE
dη
+
∂∆
∂nα
dnα
dη
O(∆2)=0
∼= ∂η∆ + nα
∂α∆ (126)
(Note that we are assuming ∂∆/∂E = 0 so that the perturbations do not depend on the frequency of the photon.)
To determine the second term, we note that it vanishes in the unperturbed Friedmann universe and arises essentially
due to the variation of Φ. Both the intrinsic time variation of Φ as well as its variation along the photon path will
contribute, giving:
d ln(aE)
dη
= −nα
∂αΦ + ∂ηΦ (127)
(The minus sign arises from the fact that the we have (1 + 2Φ) in g00 but (1 − 2Φ) in the spatial perturbations.)
Putting all these together, we can bring the evolution equation Eq. (125) to the form:
d∆
dη
= −nα
∂αΦ + ∂ηΦ +
d∆
dη coll
(128)
Let us next consider the collision term, which can be expressed in the form:
d∆
adη coll
= −NeσT
∆ + NeσT
1
4
δR + NeσT
(v·n)
= NeσT
−∆ +
1
4
δR + v·n (129)
Each of the terms in the right hand side of the first line has a simple interpretation. The first term describes the
removal of photons from the beam due to Thomson scattering with the electrons while the second term gives the
scattering contribution into the beam. In a static universe, we expect these two terms to cancel if ∆ = (1/4)δR which
fixes the relative coefficients of these two terms. The third term is a correction due to the fact that the electrons
which are scattering the photons are not at rest relative to the cosmic frame. This leads to a Doppler shift which
is accounted for by the third term. (We denote electron number density by Ne rather than ne to avoid notational
conflict with nα
.)
Formally, Eq. (128) is a first order linear differential equation for ∆. To eliminate the −NeσT ∆ term which is linear
in ∆ in the right hand side, we use the standard integrating factor exp(−τ) where
τ(χ) ≡
χ
0
dη (aNeσT
) (130)
We can then formally integrate Eq. (128) to get:
∆(n) =
η0
0
dχe−τ(χ)
−nα
∂αΦ + ∂ηΦ + aNeσT
1
4
δR + v·n (131)
We can write
e−τ
(−nα
∂αΦ) = −
dΦ
dη
e−τ
+ (∂ηΦ)e−τ
= −
d
dη
(Φe−τ
) + (aNeσT
Φ)e−τ
+ (∂ηΦ)e−τ
(132)
On integration, the first term gives zero at the lower limit and an unimportant constant (which does not depend on
n). Using the rest of the terms, we can write Eq. (131) in the form:
∆(n) =
η0
0
dχ e−τ
2∂ηΦ + aNeσT
Φ +
1
4
δR + v·n
=
η0
0
dχ e−τ
[2∂ηΦ] +
η0
0
dχ (e−τ
aNeσT
) Φ +
1
4
δR + v·n (133)
24
The first term gives the contribution due to the intrinsic time variation of the gravitational potential along the path
of the photon and is called the integrated Sachs-Wolfe effect. In the second term one can make further simplifications.
Note that e−τ
is essentially unity (optically thin) for z < zrec and zero (optically thick) for z > zrec; on the other
hand, NeσT is zero for z < zrec (all the free electrons have disappeared) and is large for z > zrec. Hence the product
(aNee−τ
) is sharply peaked at χ = χrec (i.e. at z ≃ 103
with ∆z ≃ 80). Treating this sharply peaked quantity as
essentially a Dirac delta function (usually called the instantaneous recombination approximation) we can approximate
the second term in Eq. (133) as a contribution occurring just on the LSS:
∆(n) =
1
4
δR + v · n + Φ
LSS
+ 2
η0
ηLSS
dχ∂ηΦ (134)
In the second term we have put τ = ∞ for η < ηLSS and τ = 0 for η > ηLSS.
Once we know δR, Φ and v on the LSS from perturbation theory, we can take a Fourier transform of this result to
obtain ∆(k) and use Eq. (109) to compute Cl. At very large scales the velocity term is sub-dominant and we get back
the Sachs-Wolfe effect derived earlier in Eq. (118). For understanding the small scale effects, we need to introduce
baryons into the picture which is our next task.
B. Description of photon-baryon fluid
To study the interaction of photons and baryons in the fluid limit, we need to again start from the continuity equation
and Euler equation. In Fourier space, the continuity equation is same as the one we had before (see Eq. (52)):
3
4
˙δR = −k2
vR + 3 ˙Φ; ˙δB = −k2
vB + 3 ˙Φ (135)
The Euler equations, however, gets modified; for photons, it becomes:
˙vR = (
1
4
δR + Φ)− ˙τ(vR − vB); ˙τ = NeσT
a (136)
The first two terms in the right hand side are exactly the same as the ones in Eq. (55). The last term is analogous to
a viscous drag force between the photons and baryons which arises because of the non zero relative velocity between
the two fluids. The coupling is essentially due to Thomson scattering which leads to the factor ˙τ. (The notation, and
the physics, is the same as in Eq. (130)). The corresponding Euler equation for the baryons is:
˙vB = −
˙a
a
vB + Φ+
˙τ(vR − vB)
R
(137)
where
R ≡
pB + ρB
pR + ρR
≃
3ρB
4ρR
≈ 30 ΩBh2 a
10−3
(138)
Again, the first two terms in the right hand side of Eq. (137) are the same as what we had before in Eq. (55). The last
term has the same interpretation as in the case of Euler equation Eq. (136) for photons, except for the factor R. This
quantity essentially takes care of the inertia of baryons relative to photons. Note that the the conserved momentum
density of photon-baryon fluid has the form
(ρR + pR)vR + (ρB + pB)vB ≈ (1 + R)(ρR + pR)vR (139)
which accounts for the extra factor R in Eq. (137).
We can now combine the Eqs. (135), (136), (137) to obtain, to lowest order in (k/ ˙τ) the equation:
¨δR +
˙R
(1 + R)
˙δR + k2
c2
sδR = F (140)
with
F = 4 ¨Φ +
˙R
(1 + R)
˙Φ −
1
3
k2
Φ ; c2
s =
1
3(1 + R)
(141)
25
An exact solution to this equation is difficult to obtain. However, we can try to understand several features by an
approximate method in which we treat the time variation of R to be small. In that case, we can drop the ˙R terms on
both sides of the equation. Since we know that the physically relevant temperature fluctuation is ∆ = (1/4)δR + Φ,
we can recast the above equation for ∆ as:
¨∆ + k2
c2
s∆ ≈ −k2
c2
sRΦ + 2¨Φ (142)
Let us further ignore the time variation of all terms (especially ¨Φ on the right hand side). Then, the solution is just
∆ = −RΦ + A cos(kcsηLSS
) + B sin(kcsηLSS
). To fix the initial conditions which determine A and B, we recall that
early on (η → 0), we have ∆ → Φ/3 (see Eq. (115)) and corresponding velocity should vanish. This gives the solution:
1
4
δR + Φ =
Φi
3
(1 + 3R) cos(kcsηLSS
) − ΦiR; v = −Φi (1 + 3R) cs sin(kcsηLSS
) (143)
(One can do a little better by using WKB approximation in which (kcsηLSS ) can be replaced by the integral of kcs
over η but it is not very important.) Given this solution, one can proceed as before and compute the Cl’s. Adding the
effects of [Φ + (1/4)δR] and that of [v · n] in quadrature and noticing that the angular average of (v · n)2
= (1/3)v2
we can estimate the Cl for scale invariant ( k3
|Φk|2
= 2π2
A2
) spectrum to be:
l(l + 1)Cl = 2π2
A2 (1 + 3R)
3
cos(k∗
csηLSS
) − R
2
+
(1 + 3R)2
3
c2
s sin2
(k∗
csηLSS
) (144)
with k∗
L ≈ l with L = η0 − ηLSS ≃ η0. The key feature is, of course, the maxima and minima which arises from the
trigonometric functions. The peaks of Cl are determined by the condition k∗
csηLSS
= lcsηLSS
/η0 = nπ; that is
lpeak =
nπ
cs
η0
ηLSS
= nπ
√
3(1 + zdec)1/2
≈ 172n (145)
More precise work gives the first peak at lpeak ≃ 200. It is also clear that because of non zero R the peaks are larger
when the cosine term is negative; that is, the odd peaks corresponding to n = 1, 3, ... have larger amplitudes than the
even peaks with n = 2, 4, ....
Incidentally, the above approximation is not very good for modes which enter the Hubble radius during the radiation
dominated phase since Φ does evolve with time (and decays) in the radiation dominated phase. We saw that Φ ≈
−3Φi(ly)−2
cos(ly) asymptotically in this phase (see Eq. (80)). From Eq. (80) we find that during this phase, for
modes which are inside the Hubble radius, we can take δR ≈ 6Φi cos(ly), so that ∆ ≈ δR/4 ≈ (3/2)Φi cos(ly). On the
other hand, at very large scale, the amplitude was ∆ = Φ/3 = (1/3)(9/10)Φi = (3/10)Φi. Hence the amplitude of the
modes that enter the horizon during the radiation dominated phase is enhanced by a factor (3/2)(10/9) = 5, relative
to the large scale amplitude contributed by modes which enter during matter dominated phase. This is essentially
due to the driving term ¨Φ which is nonzero in the radiation dominated phase but zero in the matter dominated phase.
(In reality, the enhancement is smaller because the relevant modes have k keq rather than k ≫ keq; see Figs. 3 and
4.)
If this were the whole story, we will see a series of peaks and troughs in the temperature anisotropies as a function of
angular scale. In reality, however, there are processes which damp out the anisotropies at small angular scales (large
-l) so that only the first few peaks and troughs are really relevant. We will now discuss two key damping mechanisms
which are responsible for this.
The first one is the finite width of the last scattering surface which makes it uncertain from which event we are
receiving the photons. In general, if P(z) is the probability that the photon was last scattered at redshift z, then we
can write:
∆T
T obs
= dz
(∆T/T ) if the last
scattering was at z
× P(z). (146)
From Eq. (41) we know that P(z) is a Gaussian with width ∆z = 80. This corresponds to a length scale
∆l = c
dt
dz
∆z · (1 + zdec) ≈ H−1
0
∆z
Ω1/2z
3/2
dec
≈8 Ωh2 −1/2
Mpc. (147)
over which the temperature fluctuations will be smoothed out.
26
It turns out that there is another effect, which is slightly more important. This arises from the fact that the photonbaryon
fluid is not tightly coupled and the photons can diffuse through the fluid. This diffusion can be modeled as a
random walk and the root mean square distance traveled by the photon during this diffusion process will smear the
temperature anisotropies over that length scale. This photon diffusion length scale can be estimated as follows:
(∆x)2
= N
numberofcollisions
q
a
2
comovingmeanfreepath
=
∆t
q(t)
q2
a2
=
∆t
a2
q(t) (148)
Integrating, we find the mean square distance traveled by the photon to be
x2
≡
tdec
0
dt
a2(t)
q(t) =
3
5
tdecq(tdec)
a2(tdec)
(149)
The corresponding proper length scale below which photon diffusion will wipe out temperature anisotropies is:
qdiff = a(tdec)x =
3
5
tdecq(tdec)
1
2
≃ 35 Mpc
ΩBh2
0.02
− 1
2
Ωh2
50
− 1
4
. (150)
It turns out that this is the dominant sources of damping of temperature anisotropies at large l ≈ 103
.
VII. GENERATION OF INITIAL PERTURBATIONS FROM INFLATION
In the description of linear perturbation theory given above, we assumed that some small perturbations existed
in the early universe which are amplified through gravitational instability. To provide a complete picture we need a
mechanism for generation of these initial perturbations. One such mechanism is provided by inflationary scenario which
allows for the quantum fluctuations in the field driving the inflation to provide classical energy density perturbations
at a late epoch. (Originally inflationary scenarios were suggested as pseudo-solutions to certain pseudo-problems;
that is only of historical interest today and the only reason to take the possibility of an inflationary phase in the early
universe seriously is because it provides a mechanism for generating these perturbations.) We shall now discuss how
this can come about.
The basic assumption in inflationary scenario is that the universe underwent a rapid — nearly exponential —
expansion for a brief period of time in very early universe. The simplest way of realizing such a phase is to postulate
the existence of a scalar field with a nearly flat potential. The dynamics of the universe, driven by a scalar field
source, is described by:
1
a2
da
dt
2
= H2
(t) =
1
3M2
Pl
V (φ) +
1
2
dφ
dt
2
;
d2
φ
dt2
+ 3H
dφ
dt
= −
dV
dφ
(151)
where Mpl = (8πG)−1/2
. If the potential is nearly flat for certain range of φ, we can introduce the “ slow roll-over”
approximation, under which these equations become:
H2
≃
V (φ)
3M2
Pl
; 3H
dφ
dt
≃ −V ′
(φ) (152)
For this slow roll-over to last for reasonable length of time, we need to assume that the terms ignored in the Eq. (151)
are indeed small. This can be quantified in terms of the parameters:
ǫ(φ) =
M2
Pl
2
V ′
V
2
; η(φ) = M2
Pl
V ′′
V
(153)
which are taken to be small. Typically the inflation ends when this assumption breaks down. If such an inflationary
phase lasts up to some time tend then the universe would have undergone an expansion by a factor exp N(t) during
the interval (t, tend) where
N ≡ ln
a(tend)
a(t)
=
tend
t
H dt ≃
1
M2
Pl
φ
φend
V
V ′
dφ (154)
27
One usually takes N ≃ 65 or so.
Before proceeding further, we would like to make couple of comments regarding such an inflationary phase. To
begin with, it is not difficult to obtain exact solutions for a(t) with rapid expansion by tailoring the potential for the
scalar field. In fact, given any a(t) and thus a H(t) = (˙a/a), one can determine a potential V (φ) for a scalar field
such that Eq. (151) are satisfied (see the first reference in [27]). One can verify that, this is done by the choice:
V (t) =
1
16πG
H 6H +
2
H
dH
dt
; φ(t) = dt
−2
8πG
dH
dt
1/2
(155)
Given any H(t), these equations give (φ(t), V (t)) and thus implicitly determine the necessary V (φ). As an example,
note that a power law inflation, a(t) = a0tp
(with p ≫ 1) is generated by:
V (φ) = V0 exp −
2
p
φ
MPl
(156)
while an exponential of power law
a(t) ∝ exp(Atf
), f =
β
4 + β
, 0 < f < 1, A > 0 (157)
can arise from
V (φ) ∝
φ
MPl
−β
1 −
β2
6
M2
Pl
φ2
(158)
Thus generating a rapid expansion in the early universe is trivial if we are willing to postulate scalar fields with tailor
made potentials. This is often done in the literature.
The second point to note regarding any inflationary scenarios is that the modes with reasonable size today originated
from sub-Planck length scales early on. A scale λ0 today will be
λend = λ0
aend
a0
= λ0
T0
Tend
≈ λ0 × 10−28
(159)
at the end of inflation (if inflation took place at GUT scales) and
λbegin = λendA−1
≈ λ0 × 10−58
(A/1030
)−1
(160)
at the beginning of inflation if the inflation changed the scale factor by A ≃ 1030
. Note that λbegin < LP for λ0 < 3
Mpc!! Most structures in the universe today correspond to transplanckian scales at the start of the inflation. It is not
clear whether we can trust standard physics at early stages of inflation or whether transplanckian effects will lead to
observable effects [20, 21].
Let us get back to conventional wisdom and consider the evolution of perturbations in a universe which underwent
exponential inflation. During the inflationary phase the a(t) grows exponentially and hence the wavelength of any
perturbation will also grow with it. The Hubble radius, on the other hand, will remain constant. It follows that,
one can have situation in which a given mode has wavelength smaller than the Hubble radius at the beginning of the
inflation but grows and becomes bigger than the Hubble radius as inflation proceeds. It is conventional to say that
a perturbation of comoving wavelength λ0 “leaves the Hubble radius” when λ0a = dH at some time t = texit(λ0).
For t > texit the wavelength of the perturbation is bigger than the Hubble radius. Eventually the inflation ends and
the universe becomes radiation dominated. Then the wavelength will grow (∝ t1/2
) slower than the Hubble radius
(∝ t) and will enter the Hubble radius again during t = tenter(λ0). Our first task is to relate the amplitude of the
perturbation at t = texit(λ0) with the perturbation at t = tenter(λ0).
We know that for modes bigger than Hubble radius, we have the conserved quantity (see Eq. (97)
ζ =
2
3
ρ
ρ + p
a
˙a
˙Φ +
˙a
a
Φ + Φ =
H
ρ + p
ikα
k2
δT 0
α + Φ (161)
At the time of re-entry, the universe is radiation dominated and ζentry ≈ (2/3)Φ. On the other hand, during inflation,
we can write the scalar field as a dominant homogeneous part plus a small, spatially varying fluctuation: φ(t, x) =
φ0(t) + f(t, x). Perturbing the equation in Eq. (151) for the scalar field, we find that the homogeneous mode φ0
satisfies Eq. (151) while the perturbation, in Fourier space satisfies:
d2
fk
dt2
+ 3H
dfk
dt
+
k2
a2
fk = 0 (162)
28
Further, the energy momentum tensor for the scalar field gives [with the “dot” denoting (d/dη) = a(d/dt)]:
ρ =
˙φ2
0
2a2
+ V ; p =
˙φ2
0
2a2
− V ; δT α
0 =
ikα
a
˙φ0 f (163)
It is easy to see that Φ is negligible at t = texit since
Φ ∼
4πGa2
k2
∼
4πG
H2
δρ ∼
δρ
ρ
∼
ρ + p
ρ
δρ
ρ + p
≪
H
(ρ + p)
˙φ0
a
fk (164)
Therefore,
ζexit ≈
H
( ˙φ2
0/a2)
−
˙φ0fk
a
= −aH
fk
˙φ0
≈
3H2
V ′
fk (165)
Using the conservation law ζexit = ζentry, we get
Φk
entry
=
9H2
2V ′
fk
exit
(166)
Thus, given a perturbation of the scalar field fk during inflation, we can compute its value at the time of re-entry,
which — in turn — can be used to compare with observations.
Equation (166) connects a classical energy density perturbation fk at the time of exit with the corresponding
quantity Φk at the time of re-entry. The next important — and conceptually difficult — question is how we can
obtain a c-number field fk from a quantum scalar field. There is no simple answer to this question and one possible
way of doing it is as follows: Let us start with the quantum operator for a scalar field decomposed into the Fourier
modes with ˆqk(t) denoting an infinite set of operators:
ˆφ(t, x) =
d3
k
(2π)3
ˆqk(t)eik.x
. (167)
We choose a quantum state |ψ > such the expectation value of ˆqk(t) vanishes for all non-zero k so that the expectation
value of ˆφ(t, x) gives the homogeneous mode that drives the inflation. The quantum fluctuation around this
homogeneous part in a quantum state |ψ > is given by
σ2
k(t) =< ψ|ˆq2
k(t)|ψ > − < ψ|ˆqk(t)|ψ >2
=< ψ|ˆq2
k(t)|ψ > (168)
It is easy to verify that this fluctuation is just the Fourier transform of the two-point function in this state:
σ2
k(t) = d3
x < ψ|ˆφ(t, x + y)ˆφ(t, y)|ψ > eik.x
. (169)
Since σk characterises the quantum fluctuations, it seems reasonable to introduce a c-number field f(t, x) by the
definition:
f(t, x) ≡
d3
k
(2π)3
σk(t)eik.x
(170)
This c-number field will have same c-number power spectrum as the quantum fluctuations. Hence we may take this
as our definition of an equivalent classical perturbation. (There are more sophisticated ways of getting this result but
none of them are fundamentally more sound that the elementary definition given above. There is a large literature
on the subject of quantum to classical transition, especially in the context of gravity; see e.g.[22]) We now have all
the ingredients in place. Given the quantum state |ψ >, one can explicitly compute σk and then — using Eq. (166)
with fk = σk — obtain the density perturbations at the time of re-entry.
The next question we need to address is what is |ψ >. The free quantum field theory in the Friedmann background
is identical to the quantum mechanics of a bunch of time dependent harmonic oscillators, each labelled by a wave
vector k. The action for a free scalar field in the Friedmann background
A =
1
2
d4
x
√
−g ∂aφ∂a
φ =
1
2
dt d3
x a3 ∂φ
∂t
2
−
1
a2
(∇φ)2
→
1
2
dt d3
k a3 dqk
dt
2
−
k2
a2
q2
k (171)
29
can be thought of as the sum over the actions for an infinite set of harmonic oscillators with mass m = a3
and
frequency ω2
k = k2
/a2
. (To be precise, one needs to treat the real and imaginary parts of the Fourier transform as
independent oscillators and restrict the range of k; just pretending that qk is real amounts the same thing.) The
quantum state of the field is just an infinite product of the quantum state ψk[qk, t] for each of the harmonic oscillators
and satisfies the Schrodinger equation
i
∂ψk
∂t
= −
1
2a3
∂2
ψk
∂q2
k
+
1
2
ak2
ψk (172)
If the quantum state ψk[qk, t′
] of any given oscillator, labelled by k, is given at some initial time, t′
, we can evolve it
to final time:
ψ [qk, t] = dq′
k K[qk, t; q′
k, t′
]ψ [q′
k, t′
] (173)
where K is known in terms of the solutions to the classical equations of motion and ψ[q′
k, t′
] is the initial state.
There is nothing non-trivial in the mathematics, but the physics is completely unknown. The real problem is that
unfortunately — in spite of confident assertions in the literature occasionally — we have no clue what ψ[q′
k, t′
] is. So
we need to make more assumptions to proceed further.
One natural choice is the following: It turns out that, Gaussian states of the form
ψk = Ak(t) exp[−Bk(t)q2
k] (174)
preserve their form under evolution governed by the Schrodinger equation in Eq. (172). Substituting Eq. (174) in
Eq. (172) we can determine the ordinary differential equation which governs Bk(t). (The Ak(t) is trivially fixed by
normalization.) Simple algebra shows that Bk(t) can be expressed in the form
Bk = −
i
2
a3 1
fk
dfk
dt
(175)
where fk is the solution to the classical equation of motion:
d2
fk
dt2
+ 3H
dfk
dt
+
k2
a2
fk = 0 (176)
For the quantum state in Eq. (174), the fluctuations are characterized by
σ2
k =
1
2
(Re Bk)−1
= |fk|2
(177)
Since one can take different choices for the solutions of Eq. (176) one get different values for σk and different spectra
for perturbations. Any prediction one makes depends on the choice of mode functions. One possibility is to choose
the modes so that ψk represents the instantaneous vacuum state of the oscillators at some time t = ti. (That is Re
Bk(ti) = (1/2)ω2
k(ti), say). The final result will then depend on the choice for ti. One can further make an assumption
that we are interested in the limit of ti → −∞; that is the quantum state is an instantaneous ground state in the
infinite past. It is easy to show that this corresponds to choosing the following solution to Eq. (176):
fk =
1
a
√
2k
(1 + ix)eix
; x =
k
Ha
(178)
which is usually called the Bunch-Davies vacuum. For this choice,
|fk|2
=
1
2ka2
1 +
k2
a2H2
; |fk|2
k=aH
≈
H2
k3
(179)
where the second result is at t = texit which is what we need to use in Eq. (166), (Numerical factors of order unity
cannot be trusted in this computation). We can now determine the amplitude of the perturbation when it re-enters
the Hubble radius. Eq. (166) gives:
|Φk|2
entry =
9H2
2V ′
2
|fk|2
=
1
k3
9H6
4V ′2
; k3
|Φk|2
entry ≃
H3
V ′
2
exit
(180)
30
One sees that the result is scale invariant in the sense that k3
|Φk|2
entry is independent of k.
It is sometimes claimed in the literature that scale invariant spectrum is a prediction of inflation. This is simply
wrong. One has to make several other assumptions including an all important choice for the quantum state (about
which we know nothing) to obtain scale invariant spectrum. In fact, one can prove that, given any power spectrum
Φ(k), one can find a quantum state such that this power spectrum is generated (for an explicit construction, see the
last reference in [20]). So whatever results are obtained by observations can be reconciled with inflationary generation
of perturbations.
To conclude the discussion, let us work out the perturbations for one specific case. Let us consider the case of the
λφ4
model for which
d2
φ
dt2
+ 3H
dφ
dt
+ V ′
(φ) = 0; V (φ) ≈ V0 −
λ
4
φ4
(181)
Using
N ∼= 8πG
φf
φ
V0
[−V ′]
dφ =
3H2
2λ
1
φ2
−
1
φ2
f
≈
3H2
2λφ2
(182)
we can write
−V ′
(φ) = λφ3
≃
H3
λ1/2N3/2
. (183)
so that the result in Eq. (180) becomes:
k3/2
Φk ≃ H3 λ1/2
N3/2
H3
≈ λ1/2
N3/2
. (184)
We do get scale invariant spectrum but the amplitude has a serious problem. If we take N 50 and note that
observations require k3/2
Φk ∼ 10−4
we need to take λ 10−15
for getting consistent values. Such a fine tuning of a
dimensionless coupling constant is fairly ridiculous; but over years inflationists have learnt to successfully forget this
embarrassment.
Our formalism can also be used to estimate the deviation of the power spectrum from the scale invariant form. To
the lowest order we have
∆2
Φ ∼ k3
|Φk|2
∼
H6
(V ′)2
∼
V 3
m6
P V ′2
(185)
Let us define the deviation from the scale invariant index by (n − 1) = (d ln ∆2
φ/d ln k). Using
d
d ln k
= a
d
da
=
˙φ
H
d
dφ
= −
m2
p
8π
V ′
V
d
dφ
(186)
one finds that
1 − n = 6ǫ − 2η (187)
Thus, as long as ǫ and η are small we do have n ≈ 1; what is more, given a potential one can estimate ǫ and η and
thus the deviation (n − 1).
Finally, note that the same process can also generate spin-2 perturbations. If we take the normalised gravity wave
amplitude as hab =
√
16πG eabφ, the mode function φ behaves like a scalar field. (The normalisation is dictated by the
fact that the action for the perturbation should reduce to that of a spin-2 field.) The corresponding power spectrum
of gravity waves is
Pgrav(k) ∼=
k3
|hk|2
2π2
=
4
π
H
mP
2
, Ωgrav(k)h2
≃ 10−5 M
mP
3
(188)
Comparing the two results
∆2
scalar ∼
H6
(V ′)2
∼
V 3
m6
P V ′2
; ∆2
tensor ∼
H2
m2
P
∼
V
m4
P
(189)
we get (∆tensor/∆scalar)2
≈ 16πǫ ≪ 1. Further, if (1 − n) ≈ 4ǫ (see Eq. (187) with ǫ ∼ η) we have the relation
(∆tensor/∆scalar)2
≈ O(3)(1−n) which connects three quantities, all of which are independently observable in principle.
If these are actually measured in future it could act as a consistency check of the inflationary paradigm.
31
VIII. THE DARK ENERGY
It is rather frustrating that the only component of the universe which we understand theoretically is the radiation!
While understanding the baryonic and dark matter components [in particular the values of ΩB and ΩDM ] is by no
means trivial, the issue of dark energy is lot more perplexing, thereby justifying the attention it has received recently.
In this section we will discuss several aspects of the dark energy problem.
The key observational feature of dark energy is that — treated as a fluid with a stress tensor T a
b = dia (ρ, −p, −p, −p)
— it has an equation state p = wρ with w −0.8 at the present epoch. The spatial part g of the geodesic acceleration
(which measures the relative acceleration of two geodesics in the spacetime) satisfies an exact equation in general
relativity given by:
∇ · g = −4πG(ρ + 3p) (190)
This shows that the source of geodesic acceleration is (ρ + 3p) and not ρ. As long as (ρ + 3p) > 0, gravity remains
attractive while (ρ + 3p) < 0 can lead to repulsive gravitational effects. In other words, dark energy with sufficiently
negative pressure will accelerate the expansion of the universe, once it starts dominating over the normal matter.
This is precisely what is established from the study of high redshift supernova, which can be used to determine the
expansion rate of the universe in the past [6].
The simplest model for a fluid with negative pressure is the cosmological constant (for some recent reviews, see
[23]) with w = −1, ρ = −p = constant. If the dark energy is indeed a cosmological constant, then it introduces
a fundamental length scale in the theory LΛ ≡ H−1
Λ , related to the constant dark energy density ρDE
by H2
Λ ≡
(8πGρDE
/3). In classical general relativity, based on the constants G, c and LΛ, it is not possible to construct any
dimensionless combination from these constants. But when one introduces the Planck constant, , it is possible
to form the dimensionless combination H2
Λ(G /c3
) ≡ (L2
P /L2
Λ). Observations then require (L2
P /L2
Λ) 10−123
. As
has been mentioned several times in literature, this will require enormous fine tuning. What is more, in the past,
the energy density of normal matter and radiation would have been higher while the energy density contributed by
the cosmological constant does not change. Hence we need to adjust the energy densities of normal matter and
cosmological constant in the early epoch very carefully so that ρΛ ρNR around the current epoch. This raises
the second of the two cosmological constant problems: Why is it that (ρΛ/ρNR) = O(1) at the current phase of the
universe ?
Because of these conceptual problems associated with the cosmological constant, people have explored a large variety
of alternative possibilities. The most popular among them uses a scalar field φ with a suitably chosen potential V (φ)
so as to make the vacuum energy vary with time. The hope then is that, one can find a model in which the current
value can be explained naturally without any fine tuning. A simple form of the source with variable w are scalar fields
with Lagrangians of different forms, of which we will discuss two possibilities:
Lquin =
1
2
∂aφ∂a
φ − V (φ); Ltach = −V (φ)[1 − ∂aφ∂a
φ]1/2
(191)
Both these Lagrangians involve one arbitrary function V (φ). The first one, Lquin, which is a natural generalization of
the Lagrangian for a non-relativistic particle, L = (1/2) ˙q2
−V (q), is usually called quintessence (for a small sample of
models, see [24]; there is an extensive and growing literature on scalar field models and more references can be found
in the reviews in ref.[23]). When it acts as a source in Friedmann universe, it is characterized by a time dependent
w(t) with
ρq(t) =
1
2
˙φ2
+ V ; pq(t) =
1
2
˙φ2
− V ; wq =
1 − (2V/ ˙φ2
)
1 + (2V/ ˙φ2)
(192)
The structure of the second Lagrangian (which arise in string theory [25]) in Eq. (191) can be understood by a
simple analogy from special relativity. A relativistic particle with (one dimensional) position q(t) and mass m is
described by the Lagrangian L = −m 1 − ˙q2. It has the energy E = m/ 1 − ˙q2 and momentum k = m ˙q/ 1 − ˙q2
which are related by E2
= k2
+ m2
. As is well known, this allows the possibility of having massless particles with
finite energy for which E2
= k2
. This is achieved by taking the limit of m → 0 and ˙q → 1, while keeping the ratio in
E = m/ 1 − ˙q2 finite. The momentum acquires a life of its own, unconnected with the velocity ˙q, and the energy
is expressed in terms of the momentum (rather than in terms of ˙q) in the Hamiltonian formulation. We can now
construct a field theory by upgrading q(t) to a field φ. Relativistic invariance now requires φ to depend on both space
and time [φ = φ(t, x)] and ˙q2
to be replaced by ∂iφ∂i
φ. It is also possible now to treat the mass parameter m as a
function of φ, say, V (φ) thereby obtaining a field theoretic Lagrangian L = −V (φ) 1 − ∂iφ∂iφ. The Hamiltonian
structure of this theory is algebraically very similar to the special relativistic example we started with. In particular,
32
the theory allows solutions in which V → 0, ∂iφ∂i
φ → 1 simultaneously, keeping the energy (density) finite. Such
solutions will have finite momentum density (analogous to a massless particle with finite momentum k) and energy
density. Since the solutions can now depend on both space and time (unlike the special relativistic example in which q
depended only on time), the momentum density can be an arbitrary function of the spatial coordinate. The structure
of this Lagrangian is similar to those analyzed in a wide class of models called K-essence [26] and provides a rich
gamut of possibilities in the context of cosmology [27, 28].
Since the quintessence field (or the tachyonic field) has an undetermined free function V (φ), it is possible to choose
this function in order to produce a given H(a). To see this explicitly, let us assume that the universe has two forms
of energy density with ρ(a) = ρknown(a) + ρφ(a) where ρknown(a) arises from any known forms of source (matter,
radiation, ...) and ρφ(a) is due to a scalar field. Let us first consider quintessence. Here, the potential is given
implicitly by the form [27, 29].
V (a) =
1
16πG
H(1 − Q) 6H + 2aH′
−
aHQ′
1 − Q
; φ(a) =
1
8πG
1/2
da
a
aQ′
− (1 − Q)
d ln H2
d ln a
1/2
(193)
where Q(a) ≡ [8πGρknown(a)/3H2
(a)] and prime denotes differentiation with respect to a. (The result used in
Eq. (155) is just a special case of this when Q = 0) Given any H(a), Q(a), these equations determine V (a) and φ(a)
and thus the potential V (φ). Every quintessence model studied in the literature can be obtained from these equations.
Similar results exists for the tachyonic scalar field as well [27]. For example, given any H(a), one can construct a
tachyonic potential V (φ) so that the scalar field is the source for the cosmology. The equations determining V (φ) are
now given by:
φ(a) =
da
aH
aQ′
3(1 − Q)
−
2
3
aH′
H
1/2
; V =
3H2
8πG
(1 − Q) 1 +
2
3
aH′
H
−
aQ′
3(1 − Q)
1/2
(194)
Equations (194) completely solve the problem. Given any H(a), these equations determine V (a) and φ(a) and thus
the potential V (φ). A wide variety of phenomenological models with time dependent cosmological constant have been
considered in the literature all of which can be mapped to a scalar field model with a suitable V (φ).
While the scalar field models enjoy considerable popularity (one reason being they are easy to construct!) it is very
doubtful whether they have helped us to understand the nature of the dark energy at any deeper level. These models,
viewed objectively, suffer from several shortcomings:
• They completely lack predictive power. As explicitly demonstrated above, virtually every form of a(t) can be
modeled by a suitable “designer” V (φ).
• These models are degenerate in another sense. The previous discussion illustrates that even when w(a) is
known/specified, it is not possible to proceed further and determine the nature of the scalar field Lagrangian.
The explicit examples given above show that there are at least two different forms of scalar field Lagrangians
(corresponding to the quintessence or the tachyonic field) which could lead to the same w(a). (See the second
paper in ref.[7] for an explicit example of such a construction.)
• All the scalar field potentials require fine tuning of the parameters in order to be viable. This is obvious in
the quintessence models in which adding a constant to the potential is the same as invoking a cosmological
constant. So to make the quintessence models work, we first need to assume the cosmological constant is zero.
These models, therefore, merely push the cosmological constant problem to another level, making it somebody
else’s problem!.
• By and large, the potentials used in the literature have no natural field theoretical justification. All of them are
non-renormalisable in the conventional sense and have to be interpreted as a low energy effective potential in
an ad hoc manner.
• One key difference between cosmological constant and scalar field models is that the latter lead to a w(a) which
varies with time. If observations have demanded this, or even if observations have ruled out w = −1 at the
present epoch, then one would have been forced to take alternative models seriously. However, all available
observations are consistent with cosmological constant (w = −1) and — in fact — the possible variation of w is
strongly constrained [30] as shown in Figure 6.
• While on the topic of observational constraints on w(t), it must be stressed that: (a) There is fair amount of
tension between WMAP and SN data and one should be very careful about the priors used in these analysis.
(b) There is no observational evidence for w < −1. (c) It is likely that more homogeneous, future, data sets
of SN might show better agreement with WMAP results. (For more details related to these issues, see the last
reference in [30].)
33
FIG. 6: The observational constraints on the variation of dark energy density as a function of redshift from WMAP and SNLS
data (see [30]). The green/hatched region is excluded at 68% confidence limit, red/cross-hatched region at 95% confidence level
and the blue/solid region at 99% confidence limit. The white region shows the allowed range of variation of dark energy at
68% confidence limit.
The observational and theoretical features described above suggests that one should consider cosmological constant
as the most natural candidate for dark energy. Though it leads to well know fine tuning problems, it also has certain
attractive features that need to kept in mind.
• Cosmological constant is the most economical [just one number] and simplest explanation for all the observations.
We stress that there is absolutely no evidence for variation of dark energy density with redshift, which is
consistent with the assumption of cosmological constant .
• Once we invoke the cosmological constant classical gravity will be described by the three constants G, c and
Λ ≡ L−2
Λ . It is not possible to obtain a dimensionless quantity from these; so, within classical theory, there
is no fine tuning issue. Since Λ(G /c3
) ≡ (LP /LΛ)2
≈ 10−123
, it is obvious that the cosmological constant is
telling us something regarding quantum gravity, indicated by the combination G . An acid test for any quantum
gravity model will be its ability to explain this value; needless to say, all the currently available models — strings,
loops etc. — flunk this test.
• So, if dark energy is indeed cosmological constant this will be the greatest contribution from cosmology to
fundamental physics. It will be unfortunate if we miss this chance by invoking some scalar field epicycles!
In this context, it is worth stressing another peculiar feature of cosmological constant when it is treated as a clue
to quantum gravity. It is well known that, based on energy scales, the cosmological constant problem is an infra
red problem par excellence. At the same time, it is a relic of a quantum gravitational effect or principle of unknown
nature. An analogy [31] will be helpful to illustrate this point. Suppose one solves the Schrodinger equation for the
Helium atom for the quantum states of the two electrons ψ(x1, x2). When the result is compared with observations,
one will find that only half the states — those in which ψ(x1, x2) is antisymmetric under x1 ←→ x2 interchange
— are realized in nature. But the low energy Hamiltonian for electrons in the Helium atom has no information
about this effect! Here is low energy (IR) effect which is a relic of relativistic quantum field theory (spin-statistics
theorem) that is totally non perturbative, in the sense that writing corrections to the Helium atom Hamiltonian in
some (1/c) expansion will not reproduce this result. The current value of cosmological constant could very well be
related to quantum gravity in a similar way. There must exist a deep principle in quantum gravity which leaves its
non perturbative trace even in the low energy limit that appears as the cosmological constant .
34
Let us now turn our attention to few of the many attempts to understand the cosmological constant. The choice
is, of course, dictated by personal bias and is definitely a non-representative sample. A host of other approaches exist
in literature, some of which can be found in [32].
A. Gravitational Holography
One possible way of addressing this issue is to simply eliminate from the gravitational theory those modes which
couple to cosmological constant. If, for example, we have a theory in which the source of gravity is (ρ + p) rather
than (ρ + 3p) in Eq. (190), then cosmological constant will not couple to gravity at all. (The non linear coupling of
matter with gravity has several subtleties; see eg. [33].) Unfortunately it is not possible to develop a covariant theory
of gravity using (ρ + p) as the source. But we can probably gain some insight from the following considerations. Any
metric gab can be expressed in the form gab = f2
(x)qab such that det q = 1 so that det g = f4
. From the action
functional for gravity
A =
1
16πG
d4
x(R − 2Λ)
√
−g =
1
16πG
d4
xR
√
−g −
Λ
8πG
d4
xf4
(x) (195)
it is obvious that the cosmological constant couples only to the conformal factor f. So if we consider a theory of
gravity in which f4
=
√
−g is kept constant and only qab is varied, then such a model will be oblivious of direct
coupling to cosmological constant. If the action (without the Λ term) is varied, keeping det g = −1, say, then one
is lead to a unimodular theory of gravity that has the equations of motion Rab − (1/4)gabR = κ(Tab − (1/4)gabT )
with zero trace on both sides. Using the Bianchi identity, it is now easy to show that this is equivalent to the usual
theory with an arbitrary cosmological constant. That is, cosmological constant arises as an undetermined integration
constant in this model [34].
The same result arises in another, completely different approach to gravity. In the standard approach to gravity
one uses the Einstein-Hilbert Lagrangian LEH ∝ R which has a formal structure LEH ∼ R ∼ (∂g)2
+ ∂2
g. If the
surface term obtained by integrating Lsur ∝ ∂2
g is ignored (or, more formally, canceled by an extrinsic curvature
term) then the Einstein’s equations arise from the variation of the bulk term Lbulk ∝ (∂g)2
which is the non-covariant
Γ2
Lagrangian. There is, however, a remarkable relation between Lbulk and Lsur:
√
−gLsur = −∂a gij
∂
√
−gLbulk
∂(∂agij)
(196)
which allows a dual description of gravity using either Lbulk or Lsur! It is possible to obtain the dynamics of gravity
[35] from an approach which uses only the surface term of the Hilbert action; we do not need the bulk term at all!.
This suggests that the true degrees of freedom of gravity for a volume V reside in its boundary ∂V — a point of
view that is strongly supported by the study of horizon entropy, which shows that the degrees of freedom hidden
by a horizon scales as the area and not as the volume. The resulting equations can be cast in a thermodynamic
form T dS = dE + PdV and the continuum spacetime is like an elastic solid (see e.g. [36]) with Einstein’s equations
providing the macroscopic description. Interestingly, the cosmological constant arises again in this approach as a
undetermined integration constant but closely related to the ‘bulk expansion’ of the solid.
While this is all very interesting, we still need an extra physical principle to fix the value (even the sign) of
cosmological constant. One possible way of doing this is to interpret the Λ term in the action as a Lagrange multiplier
for the proper volume of the spacetime. Then it is reasonable to choose the cosmological constant such that the total
proper volume of the universe is equal to a specified number. While this will lead to a cosmological constant which
has the correct order of magnitude, it has several obvious problems. First, the proper four volume of the universe is
infinite unless we make the spatial sections compact and restrict the range of time integration. Second, this will lead
to a dark energy density which varies as t−2
(corresponding to w = −1/3 ) which is ruled out by observations.
B. Cosmic Lenz law
Another possibility which has been attempted in the literature tries to “cancel out” the cosmological constant by
some process, usually quantum mechanical in origin. One of the simplest ideas will be to ask whether switching on
a cosmological constant will lead to a vacuum polarization with an effective energy momentum tensor that will tend
to cancel out the cosmological constant. A less subtle way of doing this is to invoke another scalar field (here we
go again!) such that it can couple to cosmological constant and reduce its effective value [37]. Unfortunately, none
of this could be made to work properly. By and large, these approaches lead to an energy density which is either
35
ρUV
∝ L−4
P (where LP is the Planck length) or to ρIR
∝ L−4
Λ (where LΛ = H−1
Λ is the Hubble radius associated with
the cosmological constant ). The first one is too large while the second one is too small!
C. Geometrical Duality in our Universe
While the above ideas do not work, it gives us a clue. A universe with two length scales LΛ and LP will be
asymptotically De Sitter with a(t) ∝ exp(t/LΛ) at late times. There are some curious features in such a universe
which we will now describe. Given the two length scales LP and LΛ, one can construct two energy scales ρUV
= 1/L4
P
and ρIR = 1/L4
Λ in natural units (c = = 1). There is sufficient amount of justification from different theoretical
perspectives to treat LP as the zero point length of spacetime [38], giving a natural interpretation to ρUV
. The
second one, ρIR also has a natural interpretation. The universe which is asymptotically De Sitter has a horizon
and associated thermodynamics [39] with a temperature T = HΛ/2π and the corresponding thermal energy density
ρthermal ∝ T 4
∝ 1/L4
Λ = ρIR . Thus LP determines the highest possible energy density in the universe while LΛ
determines the lowest possible energy density in this universe. As the energy density of normal matter drops below
this value, the thermal ambience of the De Sitter phase will remain constant and provide the irreducible ‘vacuum
noise’. Note that the dark energy density is the the geometric mean ρDE
=
√
ρIR
ρUV
between the two energy densities.
If we define a dark energy length scale LDE such that ρDE
= 1/L4
DE then LDE =
√
LP LΛ is the geometric mean
of the two length scales in the universe. (Incidentally, LDE ≈ 0.04 mm is macroscopic; it is also pretty close to the
length scale associated with a neutrino mass of 10−2
eV; another intriguing coincidence ?!)
Using the characteristic length scale of expansion, the Hubble radius dH ≡ (˙a/a)−1
, we can distinguish between
three different phases of such a universe. The first phase is when the universe went through a inflationary expansion
with dH = constant; the second phase is the radiation/matter dominated phase in which most of the standard
cosmology operates and dH increases monotonically; the third phase is that of re-inflation (or accelerated expansion)
governed by the cosmological constant in which dH is again a constant. The first and last phases are time translation
invariant; that is, t → t+ constant is an (approximate) invariance for the universe in these two phases. The universe
satisfies the perfect cosmological principle and is in steady state during these phases!
In fact, one can easily imagine a scenario in which the two De Sitter phases (first and last) are of arbitrarily long
duration [40]. If ΩΛ ≈ 0.7, ΩDM ≈ 0.3 the final De Sitter phase does last forever; as regards the inflationary phase,
nothing prevents it from lasting for arbitrarily long duration. Viewed from this perspective, the in between phase —
in which most of the ‘interesting’ cosmological phenomena occur — is of negligible measure in the span of time. It
merely connects two steady state phases of the universe.
While the two De Sitter phases can last forever in principle, there is a natural cut off length scale in both of them
which makes the region of physical relevance to be finite [40]. Let us first discuss the case of re-inflation in the late
universe. As the universe grows exponentially in the phase 3, the wavelength of CMBR photons are being redshifted
rapidly. When the temperature of the CMBR radiation drops below the De Sitter temperature (which happens when
the wavelength of the typical CMBR photon is stretched to the LΛ.) the universe will be essentially dominated by
the vacuum thermal noise [39] due to the horizon in the De Sitter phase. This happens when the expansion factor is
a = aF determined by the equation T0(a0/aF ) = (1/2πLΛ). Let a = aΛ be the epoch at which cosmological constant
started dominating over matter, so that (aΛ/a0)3
= (ΩDM /ΩΛ). Then we find that the dynamic range of the phase
3 is
aF
aΛ
= 2πT0LΛ
ΩΛ
ΩDM
1/3
≈ 3 × 1030
(197)
Interestingly enough, one can also impose a similar bound on the physically relevant duration of inflation. We know
that the quantum fluctuations, generated during this inflationary phase, could act as seeds of structure formation
in the universe. Consider a perturbation at some given wavelength scale which is stretched with the expansion of
the universe as λ ∝ a(t). During the inflationary phase, the Hubble radius remains constant while the wavelength
increases, so that the perturbation will ‘exit’ the Hubble radius at some time. In the radiation dominated phase,
the Hubble radius dH ∝ t ∝ a2
grows faster than the wavelength λ ∝ a(t). Hence, normally, the perturbation will
‘re-enter’ the Hubble radius at some time. If there was no re-inflation, all wavelengths will re-enter the Hubble radius
sooner or later. But if the universe undergoes re-inflation, then the Hubble radius ‘flattens out’ at late times and
some of the perturbations will never reenter the Hubble radius ! If we use the criterion that we need the perturbation
to reenter the Hubble radius, we get a natural bound on the duration of inflation which is of direct astrophysical
relevance. Consider a perturbation which leaves the Hubble radius (H−1
in ) during the inflationary epoch at a = ai. It
will grow to the size H−1
in (a/ai) at a later epoch. We want to determine ai such that this length scale grows to LΛ
just when the dark energy starts dominating over matter; that is at the epoch a = aΛ = a0(ΩDM /ΩΛ)1/3
. This gives
36
H−1
in (aΛ/ai) = LΛ so that ai = (H−1
in /LΛ)(ΩDM /ΩΛ)1/3
a0. On the other hand, the inflation ends at a = aend where
aend/a0 = T0/Treheat where Treheat is the temperature to which the universe has been reheated at the end of inflation.
Using these two results we can determine the dynamic range of this phase 1 to be
aend
ai
=
T0LΛ
TreheatH−1
in
ΩΛ
ΩDM
1/3
=
(aF /aΛ)
2πTreheatH−1
in
∼= 1025
(198)
where we have used the fact that, for a GUTs scale inflation with EGUT = 1014
GeV, Treheat = EGUT , ρin = E4
GUT
we have 2πH−1
in Treheat = (3π/2)1/2
(EP /EGUT ) ≈ 105
. If we consider a quantum gravitational, Planck scale, inflation
with 2πH−1
in Treheat = O(1), the ranges in Eq. (197) and Eq. (198) are approximately equal.
This fact is definitely telling us something regarding the duality between Planck scale and Hubble scale or between
the infrared and ultraviolet limits of the theory. The mystery is compounded by the fact the asymptotic De Sitter
phase has an observer dependent horizon and related thermal properties [39]. Recently, it has been shown — in a series
of papers, see ref.[35] — that it is possible to obtain classical relativity from purely thermodynamic considerations.
It is difficult to imagine that these features are unconnected and accidental; at the same time, it is difficult to prove
a definite connection between these ideas and the cosmological constant.
D. Gravity as detector of the vacuum energy
Finally, we will describe an idea which does lead to the correct value of cosmological constant. The conventional
discussion of the relation between cosmological constant and vacuum energy density is based on evaluating the zero
point energy of quantum fields with an ultraviolet cutoff and using the result as a source of gravity. Any reasonable
cutoff will lead to a vacuum energy density ρvac which is unacceptably high. This argument, however, is too simplistic
since the zero point energy — obtained by summing over the (1/2) ωk — has no observable consequence in any
other phenomena and can be subtracted out by redefining the Hamiltonian. The observed non trivial features of the
vacuum state of QED, for example, arise from the fluctuations (or modifications) of this vacuum energy rather than
the vacuum energy itself. This was, in fact, known fairly early in the history of cosmological constant problem and
is stressed by Zeldovich [41] who explicitly calculated one possible contribution to fluctuations after subtracting away
the mean value. This suggests that we should consider the fluctuations in the vacuum energy density in addressing
the cosmological constant problem.
If the vacuum probed by the gravity can readjust to take away the bulk energy density ρUV
≃ L−4
P , quantum
fluctuations can generate the observed value ρDE. One of the simplest models [42] which achieves this uses the
fact that, in the semi-classical limit, the wave function describing the universe of proper four-volume V will vary
as Ψ ∝ exp(−iA0) ∝ exp[−i(ΛeffV/L2
P )]. If we treat (Λ/L2
P , V) as conjugate variables then uncertainty principle
suggests ∆Λ ≈ L2
P /∆V. If the four volume is built out of Planck scale substructures, giving V = NL4
P , then the
Poisson fluctuations will lead to ∆V ≈
√
VL2
P giving ∆Λ = L2
P /∆V ≈ 1/
√
V ≈ H2
0 . (This idea can be a more
quantitative; see [42]).
Similar viewpoint arises, more rigorously, when we study the question of detecting the energy density using gravitational
field as a probe. Recall that an Unruh-DeWitt detector with a local coupling LI = M(τ)φ[x(τ)] to the field
φ actually responds to 0|φ(x)φ(y)|0 rather than to the field itself [43]. Similarly, one can use the gravitational field
as a natural “detector” of energy momentum tensor Tab with the standard coupling L = κhabT ab
. Such a model
was analysed in detail in ref. [44] and it was shown that the gravitational field responds to the two point function
0|Tab(x)Tcd(y)|0 . In fact, it is essentially this fluctuations in the energy density which is computed in the inflationary
models (see Eq. (170)) as the seed source for gravitational field, as stressed in ref. [11]. All these suggest treating the
energy fluctuations as the physical quantity “detected” by gravity, when one needs to incorporate quantum effects. If
the cosmological constant arises due to the energy density of the vacuum, then one needs to understand the structure
of the quantum vacuum at cosmological scales. Quantum theory, especially the paradigm of renormalization group
has taught us that the energy density — and even the concept of the vacuum state — depends on the scale at which
it is probed. The vacuum state which we use to study the lattice vibrations in a solid, say, is not the same as vacuum
state of the QED.
In fact, it seems inevitable that in a universe with two length scale LΛ, LP , the vacuum fluctuations will contribute
an energy density of the correct order of magnitude ρDE
=
√
ρIR
ρUV
. The hierarchy of energy scales in such a universe,
as detected by the gravitational field has [40, 45] the pattern
ρvac =
1
L4
P
+
1
L4
P
LP
LΛ
2
+
1
L4
P
LP
LΛ
4
+ · · · (199)
37
The first term is the bulk energy density which needs to be renormalized away (by a process which we do not
understand at present); the third term is just the thermal energy density of the De Sitter vacuum state; what is
interesting is that quantum fluctuations in the matter fields inevitably generate the second term.
The key new ingredient arises from the fact that the properties of the vacuum state depends on the scale at which it
is probed and it is not appropriate to ask questions without specifying this scale. If the spacetime has a cosmological
horizon which blocks information, the natural scale is provided by the size of the horizon, LΛ, and we should use
observables defined within the accessible region. The operator H(< LΛ), corresponding to the total energy inside a
region bounded by a cosmological horizon, will exhibit fluctuations ∆E since vacuum state is not an eigenstate of
this operator. The corresponding fluctuations in the energy density, ∆ρ ∝ (∆E)/L3
Λ = f(LP , LΛ) will now depend
on both the ultraviolet cutoff LP as well as LΛ. To obtain ∆ρvac ∝ ∆E/L3
Λ which scales as (LP LΛ)−2
we need to
have (∆E)2
∝ L−4
P L2
Λ; that is, the square of the energy fluctuations should scale as the surface area of the bounding
surface which is provided by the cosmic horizon. Remarkably enough, a rigorous calculation [45] of the dispersion in
the energy shows that for LΛ ≫ LP , the final result indeed has the scaling
(∆E)2
= c1
L2
Λ
L4
P
(200)
where the constant c1 depends on the manner in which ultra violet cutoff is imposed. Similar calculations have been
done (with a completely different motivation, in the context of entanglement entropy) by several people and it is
known that the area scaling found in Eq. (200), proportional to L2
Λ, is a generic feature [46]. For a simple exponential
UV-cutoff, c1 = (1/30π2
) but cannot be computed reliably without knowing the full theory. We thus find that the
fluctuations in the energy density of the vacuum in a sphere of radius LΛ is given by
∆ρvac =
∆E
L3
Λ
∝ L−2
P L−2
Λ ∝
H2
Λ
G
(201)
The numerical coefficient will depend on c1 as well as the precise nature of infrared cutoff radius (like whether it is LΛ
or LΛ/2π etc.). It would be pretentious to cook up the factors to obtain the observed value for dark energy density.
But it is a fact of life that a fluctuation of magnitude ∆ρvac ≃ H2
Λ/G will exist in the energy density inside a sphere
of radius H−1
Λ if Planck length is the UV cut off. One cannot get away from it. On the other hand, observations
suggest that there is a ρvac of similar magnitude in the universe. It seems natural to identify the two, after subtracting
out the mean value by hand. Our approach explains why there is a surviving cosmological constant which satisfies
ρDE
=
√
ρIR
ρUV
which — in our opinion — is the problem.
Acknowledgement
I thank J. Alcaniz for his friendship and warm hospitality during the X Special Courses at Observatorio Nacional,
Rio de Janeiro, Brazil and for persuading me to write up my lecture notes. I am grateful to Gaurang Mahajan for
help in generating the figures.
[1] P. de Bernardis et al., Nature 404, 955 (2000); A. Balbi et al., Ap.J. 545, L1 (2000); S. Hanany et al., Ap.J. 545, L5
(2000); T.J. Pearson et al., Ap.J. 591, 556 (2003); C.L. Bennett et al, Ap. J. Suppl. 148, 1 (2003); D. N. Spergel et
al., Ap.J.Suppl. 148, 175 (2003); B. S. Mason et al., Ap.J. 591, 540 (2003). For a summary, see e.g., L. A. Page,
astro-ph/0402547.
[2] For a review of Big Bang Nucleosynthesis (BBN), see S.Sarkar, Rept.Prog.Phys., 59, 1493 (1996); G.Steigman,
astro-ph/0501591. The consistency between CMBR observations and BBN is gratifying since the initial MAXIMABOOMERANG
data gave too high a value as stressed by T. Padmanabhan and Shiv Sethi, Ap. J. 555, 125 (2001),
[astro-ph/0010309].
[3] W. Freedman et al., Ap.J. 553, 47 (2001); J.R. Mould et al., Ap.J. 529, 786 (2000).
[4] For a discussion of the evidence, see P.J.E. Peebles, astro-ph/0410284.
[5] G. Efstathiou et al., Nature 348, 705 (1990); J. P. Ostriker, P. J. Steinhardt, Nature 377, 600 (1995); J. S. Bagla et al.,
Comments on Astrophysics 18, 275 (1996), [astro-ph/9511102].
[6] S.J. Perlmutter et al., Astrophys. J. 517, 565 (1999); A.G. Reiss et al., Astron. J. 116, 1009 (1998); J. L. Tonry et al.,
ApJ 594, 1 (2003); B. J. Barris,Astrophys.J. 602, 571 (2004); A. G.Reiss et al., Astrophys.J. 607, 665(2004).
[7] T. Roy Choudhury, T. Padmanabhan, Astron.Astrophys. 429, 807 (2005), [astro-ph/0311622]; T. Padmanabhan,T.
Roy Choudhury, MNRAS 344, 823 (2003) [astro-ph/0212573]. H.K.Jassal et al., Phys.Rev. D 72, 103503 (2005)
[astro-ph/0506748].
38
[8] T. Padmanabhan, Structure Formation in the Universe, (Cambridge University Press, Cambridge, 1993); T. Padmanabhan,
Theoretical Astrophysics, Volume III: Galaxies and Cosmology, (Cambridge University Press, Cambridge, 2002);
J.A. Peacock, Cosmological Physics, (Cambridge University Press, Cambridge, 1999); S. Dodelson, Modern Cosmology,
(Academic Press, 2003); J. Bernstein, Kinetic Theory in the Expanding Universe (Cambridge University Press, Cambridge,
(1988).
[9] D. Kazanas, Ap. J. Letts. 241, 59 (1980); A. A. Starobinsky, JETP Lett. 30, 682 (1979); Phys. Lett. B 91, 99 (1980);
A. H. Guth, Phys. Rev. D 23, 347 (1981); A. D. Linde, Phys. Lett. B 108, 389 (1982); A. Albrecht, P. J. Steinhardt,
Phys. Rev. Lett. 48, 1220 (1982); for a review, see e.g.,J. V. Narlikar and T. Padmanabhan, Ann. Rev. Astron. Astrophys.
29, 325 (1991).
[10] S. W. Hawking, Phys. Lett. B 115, 295 (1982); A. A. Starobinsky, Phys. Lett. B 117, 175 (1982); A. H. Guth, S.-Y. Pi,
Phys. Rev. Lett. 49, 1110 (1982); J. M. Bardeen et al., Phys. Rev. D 28, 679 (1983); L. F. Abbott, M. B. Wise, Nucl.
Phys. B 244, 541 (1984).
[11] T. Padmanabhan, Phys. Rev. Letts. , (1988), 60 , 2229; T. Padmanabhan, T.R. Seshadri and T.P. Singh, Phys.Rev. D 39
, 2100 (1989).
[12] G.F. Smoot. et al., Ap.J. 396, L1 (1992); T. Padmanabhan, D. Narasimha, MNRAS 259, 41P (1992); G. Efstathiou et
al., MNRAS, 258, 1 (1992).
[13] For a review, see T.Padmanabhan, Phys. Rept. 188, 285 (1990); Astrophys. Jour. Supp., 71 , 651 (1989);
astro-ph/0206131.
[14] Ya.B. Zeldovich, Astron.Astrophys. 5, 84 (1970); Gurbatov, S. N. et al, MNRAS 236, 385 (1989); T.G. Brainerd et al.,
Astrophys.J. 418, 570 (1993); J.S. Bagla, T.Padmanabhan, MNRAS 266, 227 (1994), [gr-qc/9304021]; MNRAS, 286,
1023 (1997), [astro-ph/9605202]; T.Padmanabhan, S.Engineer, Ap. J. 493, 509 (1998), [astro-ph/9704224]; S. Engineer
et.al., MNRAS 314, 279 (2000), [astro-ph/9812452]; for a recent review, see T.Tatekawa, [astro-ph/0412025].
[15] A. J. S. Hamilton et al., Ap. J. 374, L1 (1991); R. Nityananda, T. Padmanabhan, MNRAS 271, 976 (1994),
[gr-qc/9304022]; T. Padmanabhan, MNRAS 278, L29 (1996), [astro-ph/9508124]. T.Padmanabhan et al., Ap. J. 466,
604 (1996), [astro-ph/9506051]; D. Munshi et al., MNRAS, 290, 193 (1997), [astro-ph/9606170]; J. S. Bagla, et.al., Ap.J.
495, 25 (1998), [astro-ph/9707330]; N.Kanekar et al., MNRAS , 324, 988 (2001), [astro-ph/0101562].
[16] For a pedagogical description, see J.S. Bagla, astro-ph/0411043; J.S. Bagla, T. Padmanabhan, Pramana 49, 161-192
(1997), [astro-ph/0411730].
[17] There is extensive literature on linear perturbation theory as well as generation of perturbations from inflation. I am not
giving original references while discussing these topics. In addition to the textbooks mentioned in [8], the following review
articles will be useful: V.F. Mukhanov et al., Physics Reports 215, 203 (1992); D. H. Lyth and A. Riotto, Physics Reports
314, 1 (1999); A. R. Liddle and D. H. Lyth, Physics Reports 231, 1 (1993); H. Kodama and M. Sasaki, Prog. Theor. Phys.
Suppl. 78, 1 (1984).
[18] E.R. Harrison, Phys. Rev. D1, 2726 (1970); Zeldovich Ya B. MNRAS 160, 1p (1972).
[19] J. Hwang et al., Phys. Rev. D65, 043005 (2002) [astro-ph/0107307].
[20] One of the earliest attempts to include transplanckian effects in inflation is in [11]. A sample of more recent papers are
J.Martin, R.H. Brandenberger, Phys.Rev. D 63, 123501 (2001); A. Kempf, Phys.Rev. D 63 , 083514 (2001); U.H. Danielsson,
Phys.Rev. D 66, 023511 (2002); A. Ashoorioon et al., Phys.Rev. D 71 023503 (2005); R. Easther, astro-ph/0412613;
for a more extensive set of references, see L. Sriramkumar et al., Phys. Rev., D 71, 103512 (2005) [gr-qc/0408034].
[21] S. Corley, T. Jacobson, Phys.Rev. D 54, 1568 (1996); T. Padmanabhan, Phys.Rev.Lett. 81, 4297 (1998), [hep-th/9801015];
Phys.Rev. D 59 124012 (1999), [hep-th/9801138]; A.A. Starobinsky, JETP Lett. 73 371-374 (2001); J.C. Niemeyer,
R. Parentani, Phys.Rev. D 64 101301 (2001); J. Kowalski-Glikman, Phys.Lett. B 499 1 (2001); G.Amelino-Camelia,
Int.J.Mod.Phys. D 11 35 (2002).
[22] V.G. Lapchinsky, V.A. Rubakov, Acta Phys.Polon. B 10, 1041 (1979); J. B. Hartle, Phys. Rev. D 37, 2818 (1988); D
38, 2985 (1988); T.Padmanabhan, Class. Quan. Grav. 6, 533 (1989); T.P. Singh, T. Padmanabhan, Annals Phys. 196,
296(1989). T. Padmanabhan, Phys.Rev. D 39, 2924 (1989); T.Padmanabhan and T.P. Singh, Class. Quan. Grav.. 7, 441
(1990); J.J. Hallwell, Phys.Rev. D 39, 2912 (1989).
[23] L. Perivolaropoulos, astro-ph/0601014; S. Hannestad, astro-ph/0509320; P. J. E. Peebles and B. Ratra, Rev. Mod. Phys.
75, 559 (2003); T. Padmanabhan, Phys. Rept. 380, 235 (2003) [hep-th/0212290]; Current Science, 88,1057, (2005)
[astro-ph/0411044]; gr-qc/0503107; J. R. Ellis, Phil. Trans. Roy. Soc. Lond. A 361, 2607 (2003).
[24] A small sample of papers covering varied aspects are: S. Nojiri, S.D. Odintsov. hep-th/0601213; Mian Wang.
hep-th/0601189; Hao Wei et al., Phys.Rev. D 72, 123507 (2005); S. Capozziello, astro-ph/0508350; D. Polarski,
A. Ranquet, Phys.Lett. B 627, 1 (2005); A. A. Andrianov et al., Phys.Rev. D 72, 043531 (2005); H.Stefancic,
astro-ph/0504518; Martin Sahlen et al., astro-ph/0506696; Zhuo-Yi Huang et al., astro-ph/0511745; Z. K. Guo, N. Ohta
and Y. Z. Zhang, astro-ph/0505253; Shin’ichi Nojiri, S. D. Odintsov, hep-th/0506212; hep-th/0408170; W. Godlowski et
al.,astro-ph/0309569; V. F. Cardone et al., Phys.Rev. D69, (2004), 083517; I.P. Neupane, Class.Quant.Grav. 21, 4383
(2004); M.C. Bento et al.,, astro-ph/0407239; A. DeBenedictis et al., gr-qc/0402047; M. Axenides and K.Dimopoulos,
hep-ph/0401238; M. D. Maia et al., astro-ph/0403072; J. A. S. Lima, J. S. Alcaniz, Phys.Lett. B 600 191 (2004); J. S.
Alcaniz, astro-ph/0312424; Xin-Zhou Li et al., Int.J.Mod.Phys. A18, 5921 (2003); P. J. Steinhardt, Phil. Trans. Roy. Soc.
Lond. A 361, 2497 (2003); S. Sen and T. R. Seshadri, Int. J. Mod. Phys. D 12, 445 (2003); C. Rubano and P. Scudellaro,
Gen. Rel. Grav. 34, 307 (2002); S. A. Bludman and M. Roos, Phys. Rev. D 65, 043503 (2002); R. de Ritis and
A. A. Marino, Phys. Rev. D 64, 083509 (2001); A. de la Macorra and G. Piccinelli, Phys. Rev. D 61, 123503 (2000);
L. A. Urena-Lopez and T. Matos, Phys. Rev. D 62, 081302 (2000); P. F. Gonzalez-Diaz, Phys. Rev. D 62, 023513 (2000);
Phys. Lett. B 562, 1 (2003); T. Barreiro, E. J. Copeland and N. J. Nunes, Phys. Rev. D 61, 127301 (2000). B. Ratra,
39
P.J.E. Peebles, Phys.Rev. D 37, 3406 (1988).
[25] A. Sen, JHEP 0204 048 (2002), [hep-th/0203211].
[26] A small sample of papers are Hui Li et al., astro-ph/0601007; L. P. Chimento and A. Feinstein, Mod. Phys. Lett. A 19,
761 (2004); R. J. Scherrer, Phys. Rev. Lett. 93 ) 011301 (2004; P. F. Gonzalez-Diaz,hep-th/0408225; L. P. Chimento,
Phys.Rev. D69, 123517, (2004); O.Bertolami, astro-ph/0403310; R.Lazkoz, gr-qc/0410019; J.S. Alcaniz, J.A.S. Lima,
astro-ph/0308465; M. Malquarti et al., Phys. Rev. D 67, 123503 (2003); T. Chiba, Phys. Rev. D 66, 063514 (2002);
C. Armendariz-Picon et al., Phys. Rev. D 63, 103510 (2001).
[27] T. Padmanabhan, Phys. Rev. D 66, 021301 (2002), [hep-th/0204150]; T. Padmanabhan and T. R. Choudhury, Phys. Rev.
D 66, 081301 (2002) [hep-th/0205055]; J. S. Bagla, et al., Phys. Rev. D 67, 063504 (2003) [astro-ph/0212198].
[28] For a varied sample, see H. Singh, hep-th/0505012; hep-th/0508101; Anupam Das et al., Phys.Rev. D 72, 043528 (2005);
I.Ya. Aref’eva et al., astro-ph/0505605; astro-ph/0410443; J. M. Aguirregabiria and R. Lazkoz, hep-th/0402190; C. Kim et
al., hep-th/0404242; R. Herrera et al.,astro-ph/0404086; A. Ghodsi and A.E.Mosaffa,hep-th/0408015; D.J. Liu and X.Z.Li,
astro-ph/0402063; V. Gorini et al., Phys.Rev. D 69 123512 (2004); M. Sami et al., Pramana 62, 765 (2004); D.A. Steer,
Phys.Rev. D70, 043527 (2004); L.Raul W. Abramo, Fabio Finelli, Phys.Lett. B 575, 165 (2003); L. Frederic, A. W. Peet,
JHEP 0304, 048 (2003); M. Sami, Mod.Phys.Lett. A 18, 691 (2003); G. W. Gibbons, Class. Quant Grav. 20, S321 (2003);
C. J. Kim et al., Phys. Lett. B 552, 111 (2003); G. Shiu and I. Wasserman, Phys. Lett. B 541, 6 (2002); D. Choudhury et
al., Phys. Lett. B 544, 231 (2002); A. V. Frolov, et al., Phys. Lett. B 545, 8 (2002); G. W. Gibbons, Phys. Lett. B 537,
1 (2002).
[29] G.F.R. Ellis and M.S.Madsen, Class.Quan.Grav. 8, 667 (1991).
[30] H.K. Jassal et al., MNRAS 356, L11-L16 (2005), [astro-ph/0404378]; [astro-ph/0601389].
[31] T. Padmanabhan and T. Roy Choudhury, Mod. Phys. Lett. A 15, 1813 (2000), [gr-qc/0006018].
[32] There is extensive literature on different paradigms for solving the cosmological constant problem, like e.g., those based
on QFT in CST: E. Mottola, Phys.Rev. D 31, 754 (1985); N.C. Tsamis,R.P. Woodard, Phys.Lett. B301, 351 (1993); E.
Elizalde and S.D. Odintsov, Phys.Lett. B 321 199 (1994); B 333 331 (1994); J.V. Lindesay, H.P. Noyes, astro-ph/0508450;
Shi Qi, hep-th/0505109. Non-ideal fluids mimicking cosmological constant, like e.g., E. Baum, Phys. Letts. B 133, 185
(1983); T. Padmanabhan, Phys. Letts., A104, 196 (1984); S.W. Hawking, Phys. Letts. B 134, 403 (1984); T.Padmanabhan
and S.M. Chitre, Phys. Letts. A 120, 433 (1987); S. Coleman, Nucl. Phys. B 310 643 (1988); Quantum cosmological considerations:
T. Mongan, Gen. Rel. Grav., 33 1415 (2001) [gr-qc/0103021]; Gen.Rel.Grav. 35 685 (2003). Holographic dark
energy: Yungui Gong, Yuan-Zhong Zhang, hep-th/0505175; astro-ph/0502262; hep-th/0412218; Xin Zhang, Feng-Quan
Wu, astro-ph/0506310; M.G.Hu, Xin-He Meng, astro-ph/0511615. Those based on renormalization group, running coupling
constants and more general time dependent decay schemes: I.L. Shapiro, Phys.Lett. B 329, 181 (1994); I.L. Shapiro and
J. Sola, Phys.Lett. B475, 236 (2000); hep-ph/0305279; astro-ph/0401015; I.L. Shapiro et al., hep-ph/0410095; Cristina
Espana-Bonet, et.al., Phys.Lett. B 574 149 (2003); JCAP 0402, 006 (2004); F.Bauer, gr-qc/0501078; gr-qc/0512007; I. L.
Shapiro, J. Sola, JHEP, 0202 006 (2002), hep-ph/0012227; J.Sola and H.Stefancic, astro-ph/0505133; astro-ph/0507110;
J. S. Alcaniz, J. A. S. Lima, Phys.Rev. D 72 063516 (2005), [astro-ph/0507372]; J. S. Alcaniz , J. M. F. Maia, Phys.Rev.
D 67 043502 (2003); F. Bauer, Class.Quant.Grav., 22, 3533 (2005) and many more.
[33] T. Padmanabhan (2004) [gr-qc/0409089]; R. R. Caldwell, astro-ph/0209312; T.Padmanabhan, Int.Jour. Mod.Phys. A 4,
4735 (1989), Sec.6.1.
[34] A. Einstein, Siz. Preuss. Acad. Scis. (1919), translated as ”Do Gravitational Fields Play an essential Role in the Structure
of Elementary Particles of Matter,” in The Principle of Relativity, by edited by A. Einstein et al. (Dover, New York, 1952);
J. J. van der Bij et al., Physica A 116, 307 (1982); F. Wilczek, Phys. Rep. 104, 111 (1984); A. Zee, in High Energy
Physics, proceedings of the 20th Annual Orbis Scientiae, Coral Gables, (1983), edited by B. Kursunoglu, S. C. Mintz, and
A. Perlmutter (Plenum, New York, 1985); W. Buchmuller and N. Dragon, Phys.Lett. B 207, 292, (1988); W.G. Unruh,
Phys.Rev. D 40 1048 (1989).
[35] T. Padmanabhan, Int.Jour.Mod.Phys D (in press)[gr-qc/0510015]; Brazilian Jour.Phys. (Special Issue) 35, 362 (2005)
[gr-qc/0412068]; Class. Quan. Grav., 21, L1 (2004) [gr-qc/0310027]; Gen.Rel.Grav., 35, 2097 (2003); ibid.,34 2029 (2002)
[gr-qc/0205090].
[36] A. D. Sakharov, Sov. Phys. Dokl., 12, 1040 (1968); T. Jacobson, Phys. Rev. Lett. 75, 1260 (1995); T. Padmanabhan, Mod.
Phys. Lett. A 17, 1147 (2002) [hep-th/0205278]; 18, 2903 (2003) [hep-th/0302068]; Class.Quan.Grav., 21, 4485 (2004)
[gr-qc/0308070]; G.E. Volovik, Phys.Rept., 351, 195 (2001); T. Padmanabhan, Int.Jour.Mod.Phys. D 13, 2293-2298 (2004),
[gr-qc/0408051]; G. E. Volovik, The universe in a helium droplet, (Oxford University Press, 2003); B.L. Hu, gr-qc/0503067
and references therein.
[37] A.D. Dolgov, in The very early universe: Proceeding of the 1982 Nuffield Workshop at Cambridge, ed. G.W. Gibbons,
S.W. Hawking and S.T.C. Sikkos (Cambridge University Press), (1982), p. 449; S.M. Barr, Phys. Rev. D 36, 1691 (1987);
Ford, L.H., Phys. Rev. D 35, 2339 (1987); A. Hebecker and C. Wetterich, Phy. Rev. Lett., 85 3339 (2000); A. Hebecker
hep-ph/0105315; T.P. Singh, T. Padmanabhan, Int. Jour. Mod. Phys. A 3, (1988), 1593; M. Sami, T. Padmanabhan,
(2003) Phys. Rev. D 67, 083509 [hep-th/0212317].
[38] H. S. Snyder Phys. Rev., 71, 38 (1947); B. S. DeWitt, Phys. Rev. Lett., 13, 114 (1964); T. Yoneya Prog. Theor. Phys.,
56, 1310 (1976); T. Padmanabhan Ann. Phys. (N.Y.), 165, 38 (1985); Class. Quantum Grav. 4, L107 (1987); A. Ashtekar
et al., Phys. Rev. Lett., 69, 237 (1992 ); T. Padmanabhan Phys. Rev. Lett. 78, 1854 (1997) [hep-th/9608182]; Phys. Rev.
D 57, 6206 (1998); K.Srinivasan et al., Phys. Rev. D 58 044009 (1998) [gr-qc/9710104]; M. Fontanini et al. Phys.Lett. B
633, 627 (2006) hep-th/0509090. For a review, see L.J. Garay, Int. J. Mod. Phys. A10, 145 (1995).
[39] G. W. Gibbons and S.W. Hawking, Phys. Rev. D 15, 2738 (1977); T.Padmanabhan, Mod.Phys.Letts., A 17, 923
(2002) [gr-qc/0202078]; Class. Quant. Grav., 19, 5387 (2002) [gr-qc/0204019]; Mod.Phys.Letts. A 19, 2637 (2004)
40
[gr-qc/0405072]; For a recent review see e.g., T. Padmanabhan, Phys. Reports, 406, 49 (2005) [gr-qc/0311036].
[40] T.Padmanabhan Lecture given at the Plumian 300 - The Quest for a Concordance Cosmology and Beyond meeting at
Institute of Astronomy, Cambridge, July 2004; [astro-ph/0510492].
[41] Y.B. Zel’dovich, JETP letters 6, 316 (1967); Soviet Physics Uspekhi 11, 381 (1968).
[42] T. Padmanabhan, Class.Quan.Grav. 19, L167 (2002), [gr-qc/0204020]; for an earlier attempt, see D. Sorkin,
Int.J.Theor.Phys. 36, 2759 (1997); for related ideas, see Volovik, G. E., gr-qc/0405012; J. V. Lindesay et al.,
astro-ph/0412477; Yun Soo Myung, hep-th/0412224; E.Elizalde et al., hep-th/0502082.
[43] S. A. Fulling, Phys. Rev. D7 2850 (1973); W.G. Unruh, Phys. Rev. D14, 870 (1976); K. Srinivasan et al., Phys. Rev. D
60, 24007 (1999) [gr-qc-9812028]; L. Sriramkumar et al., Int. Jour. Mod. Phys. D 11,1 (2002) [gr-qc/9903054].
[44] T. Padmanabhan and T.P. Singh, Class. Quan. Grav., (1987), 4 , 1397.
[45] T. Padmanabhan, Class.Quan.Grav., 22, L107, (2005) [hep-th/0406060].
[46] L. Bombelli et al., Phys. Rev. D34, 373 (1986); M. Srednicki, Phys. Rev. Lett. 71, 666 (1993); R. Brustein et al., Phys.
Rev. D65, 105013 (2002); A. Yarom, R. Brustein, hep-th/0401081. This result can also be obtained from those in ref. [44].