Cosmology todayA brief review

Jorge L. Cervantes-Cota

and George Smoot

Depto. de Fsica, Inst. Nac. de Investigaciones Nucleares, A.P. 18-1027, 11801 Mxico DF,
Berkeley Center for Cosmological Physics, University of California, Berkeley, USA.

Lawrence Berkeley National Laboratory and University of California at Berkeley, Berkeley, USA,
Institute for the Early Universe WCU at Ewha Womans University Seoul, Korea,
Paris Centre for Cosmological Physics, Universite Paris Diderot, France.
Abstract. This is a brief review of the standard model of cosmology. We rst introduce the FRW
models and their at solutions for energy uids playing an important role in the dynamics at different
epochs. We then introduce different cosmological lengths and some of their applications. The later
part is dedicated to the physical processes and concepts necessary to understand the early and very
early Universe and observations of it.
Keywords: Theoretical cosmology, Observational cosmology
PACS: 98.80.-k;98.80.Bp;98.80.Es
INTRODUCTION
The purpose of the present review is to provide the reader with an outline of modern cos-
mology. This science is passing through a revolutionary era, mainly because recent high
precision observations have severely constrained theoretical speculations, and opened
new windows to the cosmos. Tracking the recent history, in the late 1940s George
Gamow [1] predicted that the Universe should had begun from a very dense state, char-
acterized by a huge density at very high temperatures, a scenario dubbed the Big Bang,
that was conjectured by George Lematre in the early 30s. This scenario predicts that
matter and light were at very high energetic states, and both components behaved as a ra-
diation uid, in thermal equilibrium described as Planck blackbody (Bose-Einstein and
the related Fermi-Dirac distributions). This initial state remains today imprinted in the
Cosmic Microwave Background Radiation (CMBR). Gamows scenario predicted this
primeval radiation would be measured at a temperature of only few Kelvins degrees;
since the expansion of the Universe cools down any density component. Robert Dicke,
and others, begun the race to discover this radiation coming from the cosmos, and in
1965 A. A. Penzias and R.W. Wilson of the Bell telephone laboratories discovered, by
chance, this radiation form, conrming the general Big Bang scenario; see the original
references published in Ref. [2].
During this elapsed period the Big Bang scenario was generally accepted. However,
some key questions remained open, for instance, whether or not this radiation was of
Planckian nature to entirely conrm that the Universe was in thermal equilibrium at the
very beginning of time. In 1990 a modern version of the PenziasWilson experiment
was carried out. This experiment, lead by G. Smoot and J. Mather with the Cosmic
Background Explorer (COBE) satellite, started a new high precision experimental era
in cosmology [3]. The COBE team for the rst time revealed that the Universe was
a
r
X
i
v
:
1
1
0
7
.
1
7
8
9
v
1


[
a
s
t
r
o
-
p
h
.
C
O
]


9

J
u
l

2
0
1
1
in equilibrium (the radiation had the Planck form to high precision), and was almost
homogeneous and isotropic, but not completely[4]. The tiny (10
5
) anisotropies found
by COBE also imprinted in the matter distribution were lately the responsible for the
formation of stars, galaxies, clusters, and all large scale structures of our Universe.
The origin of these tiny anisotropies is presumably in quantum uctuations of funda-
mental elds of nature which were present in the very early Universe. Modern quantum
eld theories, together with cosmological models, help to understand how this small
uctuations evolved to become of cosmic scales, enabling COBE to detect them. This
satellite and other more recent cosmological probes, such as BOOMERANG, MAX-
IMA, WMAP and now PLANCK satellite that measured CMBR and the Two degree
Field (2dF) galaxy survey and the Sloan Digital Sky Survey (SDSS), among others, not
only conrmed with a great accuracy some of the theoretical predictions of the standard
big bang cosmological model, but also opened the possibility to test theories and sce-
narios applicable in very early Universe, such as ination, or in present times, such as
quintessence. In this way, cosmology that used to be a purely theoretical and often spec-
ulative science, is today subject to high precision tests in light of these new observations
[5, 6].
ON THE STANDARD MODEL OF COSMOLOGY
We begin our study by reviewing some aspects of the standard lore of physical and the-
oretical cosmology. In doing that we consider the Friedmann-Robertson-Walker (FRW)
model in Einsteins general relativity (GR) theory. We shall make use of natural units
h = c = k
B
= 1 and our geometrical sign conventions are as in Ref. [7].
FRW models
The cosmological principle states that the Universe is both spatially homogeneous and
isotropic on the large scale, which was originally assumed but now observed to be valid
for the very large large scale of the Universe. This homogeneous and isotropic space
time symmetry was originally studied by Friedmann, Robertson, and Walker (FRW), see
Refs. [8]. The symmetry is encoded in and denes the unique form of the line element:
ds
2
=dt
2
+a
2
(t)

dr
2
1kr
2
+r
2
(d
2
+sin
2
d
2
)

, (1)
where t is the cosmic time, r are polar coordinates, which can be adjusted so
that the constant curvature takes the values k = 0, +1, or 1 for a at, closed, or open
space, respectively. a(t) is the unknown potential of the metric that encodes the size at
large scales, more formally is the scale factor of the Universe.
To then solve this line element for the scale factor of the Universe a(t) and perturba-
tions one simply needs a valid theory of gravity. The standard model of cosmology is
based on GR, which can be derived from the Einstein-Hilbert Lagrangian
L =
1
16G
(R+L
m
)

g , (2)
where R is the Ricci scalar, G the Newton constant, and g =|g

| the determinant of the

metric tensor. By performing the metric variation of this equation, one obtains the well
known Einsteins eld equations
R

1
2
Rg

= 8GT

, (3)
where T

g
L
m

g
g

is the stress energymomentumtensor, which is a symmetric

tensor. Thus, Eq. (3) is a collection of ten coupled partial differential equations. How-
ever, the theory is diffeomorphism invariant, and one adds to them a gauge condition,
implying in general four extra equations to Eq. (3) that reduce the degrees of freedom.
The beautifully symmetric FRW solutions to the Einstein Eqs. (3) represent a cor-
nerstone in the development of modern cosmology, since with them it is possible to
understand the expansion of Universe as was realized in the 1920s through Hubbles
law of expansion [9]. With this FRW metric, the GR cosmological eld equations are,
H
2

a
a

2
=
8G
3

k
a
2
(4)
and
a
a
=
4G
3
( +3p) , (5)
where H is the Hubble parameter; and p are the density and pressure of the perfect
uid considered, that is, T

= u

+ p(u

+g

), where u

=
0

is the uids
four velocity in comoving coordinates. Dots stand for cosmic time derivatives.
Assuming energymomentum tensor conservation, T

;
=0, is valid, one obtains the
continuity equation,
+3H( + p) = 0 . (6)
The above system of equations (4), (5), and (6) implies three unknown variables (a,
, p) for three equations, but the equations are not all linearly independent, just two of
them. Thus, an extra assumption has to be made to close the system. The answer should
come from the micro-physics of the uids considered. For the moment let us assume a
barotropic equation of state that is characteristic for different cosmic uids, w = const.,
p

= w =

1
3
for radiation or relativistic matter
0 for dust
1 for stiff uid
1 for cosmological constant
(7)
to integrate Eq. (6), yielding
=
M
w
a
3(1+w)
or

x

x 0
=

a
0
a

3(1+w)
, (8)
where M
w
is the integration constant and is different dimensioned by considering differ-
ent wuids. With this equation the system is closed and can be solved once the initial
conditions are known.
In addition to the above uids, one can include an explicit cosmological constant ()
in Eq. (4) and arrange it in the following form:

R
+
M
+

= 1+
k
a
2
H
2
(9)
with
R
being the radiation component dominated by the CMB and small (510
5
at
present),
M

8G
m
3H
2
and

3H
2
. The parameter is called the density parameter,
which is composed of a matter part and a cosmological constant term. Thus, we see that
the different values of the density parameters (
m
,

) will impose different values for

the curvature term. If > 1, it turns out k is greater than zero, signifying a Universe
with a positive curvature (closed Universe). If < 1, then k < 1, this corresponds
to a negative curvature (open Universe). A critical value is obtained obviously when
= 1, then the spatial curvature is null, k = 0. The value of the energy density for
which = [ +/(8G)]/
c
=1 holds is known as the critical density,
c
3H
2
/8G.
The last term in Eq. (9) can be dened as
k
=
k
a
2
H
2
and thus the Friedman equation
becomes a constriction for the density parameters
i

i
= 1, and this expression holds
at any time. It is useful to dene an alternative measure of the expansion of the Universe
through the redshift (z), 1+z a
0
/a(t), where a
0
is the scale factor at present and is the
set to unity by convention. Today z
0
= 0 and in the early Universe the redshift grows. In
terms of the redshift the density parameters are, from Eq. (8),
i
=
(0)
i
(1 +z)
3(1+w)
;
quantities with a subindex or superindex 0 are evaluated at the present time.
Let us very briey recall which wvalues are needed to describe the different epochs
of the Universes evolution. The assumption that w = 1/3 is valid for a uid of
radiation and/or of ultra-relativistic matter (T m, m being its rest mass). This epoch is
of importance at the beginning of the hot big bang theory, where the material content of
the Universe consisted of photons, neutrinos, electrons, and other massive particles with
very high kinetic energy. After some Universe cooling, some massive particles decayed
and others survived (protons, neutrons, electrons) whose masses eventually dominated
over the radiation components (photon, neutrinos) at the equality epoch (
rel
=
m
)
at z
eq
3200 [10]. From this epoch and until recent efolds of expansion (z
DE
1/2)
the main matter component produced effectively no pressure on the expansion and,
therefore, one can accept a model lled with dust, w = 0, to be representative for the
energy content of the Universe in the interval 3200 <z <1/2. The dust equation of state
is then representative of inert, cold dark matter (CDM). Dark matter (DM) does not
(signicantly) emit light and therefore it is dark. Another possibility is that dark matter
interacts weakly and is generically called WIMP (Weakly Interacting Massive Particle),
being the neutralino the most popular WIMP candidate. Another popular dark matter
candidate is the axion, a hypothetical particle postulated to explain the conservation of
the CP symmetry in quantum chromodynamics (QCD). Back to the Universe evolution,
from z 1/2 and until now the Universe happens to be accelerating with an equation
of state w 1, due to some constant energy that induces a cosmological constant,
= 8G =const. The cosmological constant is the generic factor of an inationary
solution, see the k = 0 solution below, Eq. (12). The details of the expansion are
still unknown and it is possible that the expansion is due to some new fundamental
eld (e.g. quintessence) that induces an effective (t) const. One calls (M. Turner
dubbed it) dark energy (DE) to this new element. Dark energy does not emit light
nor any other particle, as so far known, it simply behaves as a (transparent) media
that gravitates with an effective negative pressure. The physics behind dark energy or
even the cosmological constant is unclear since theories of grand unication (or theories
of everything, including gravity) generically predict a vacuum energy associated with
fundamental elds, < 0|T

|0 >=< > g

, that turns out to be very large. This can

be seen by summing the zero-point energies of all normal modes of some eld of mass
m, to obtain < > M
4
/(16
2
), where M represents some cutoff in the integration,
M m. Then, assuming GR is valid up to the Planck (Pl) scale, one should take
M1/

8G, which gives < >= 10

71
GeV
4
. This term plays the role of an effective
cosmological constant of = 8G < > M
2
Pl
10
38
GeV
2
which must be added to
the Einstein equations (3) and yields an inationary solution Eq. (12). However, since
the cosmological constant seems to dominate the Universe dynamics nowadays, one has
that
8G
0
= 3H
2
0
10
83
GeV
2
. (10)
which is very small compared with the above value derived on dimensional grounds.
Thus, the cosmological constraint and theoretical expectations are rather dissimilar, by
about 121 orders of magnitude! Even if one considers symmetries at lower energy scales
the theoretical is indeed smaller, but never as small as the cosmological constraint:

GUT
10
21
GeV
2
,
SU(2)
10
29
GeV
2
. This problem has been reviewed since
decades ago [11, 12] and remains open.
The ordinary differential equations system described above needs a set of initial, or
alternatively boundary, conditions to be integrated. One has to assume a set of two
initial values, say, ((t

), a(t

)) (

, a

) at some (initial) time t

, in order to determine
its evolution. The full analysis of it can be found in many textbooks [13, 7]. Here, in
order to show some physical, early Universe consequences we assume k = 0, justied
as follows: From Eqs. (4) and (8) one notes that the expansion rate, given by the Hubble
parameter, is dominated by the density term as a(t) 0, since 1/a
3(1+w)
> k/a
2
for w >1/3, that is, the at solution is very well tted at the very beginning of times.
Therefore, assuming k = 0, Eq. (4) implies
a(t) = [6GM
w
(1+w)
2
]
1
3(1+w)
(t t

)
2
3(1+w)
=

(
32
3
GM1
3
)
1/4
(t t

)
1/2
for w =
1
3
radiation
(6GM
0
)
1/3
(t t

)
2/3
for w = 0 dust
(24GM
1
)
1/6
(t t

)
1/3
for w = 1 stiff uid
(11)
and
a(t) = a

e
Ht
for w =1 cosmological constant (12)
where the letters with a subindex are integration constants, representing quantities
evaluated at the beginning of times, t = t

. To obtain Eq. (12), the argument given

right above to neglect k is not anymore valid, since here = const.; that is, from the
very beginning it must be warranted that H
2

8G
3

> k/a
2

, otherwise k cannot be
ignored. Nevertheless if is present, it will eventually dominate over the other decaying
components, this is the so called cosmological no-hair theorem[14]. Ageneral feature of
all the above solutions is that they are expanding, at different Hubble rates, H =
2
3(1+w)
1
t
for Eqs. (11) and H =const. for Eq. (12).
From Eq. (11) one can immediately see that at t =t

, a

=0 and from Eq. (8),

=,
that is, the solution has a singularity at that time, at the Universes beginning; this initial
cosmological singularity is precisely the big bang singularity. As the Universe evolves
the Hubble parameter goes as H 1/t, i.e., the expansion rate decreases; whereas
the matter-energy content acts as an expanding agent, cf. Eq. (4), it decelerates the
expansion, however, asymptotically decreasing, cf. Eqs. (5) and (8). In that way, H
1
represents an upper limit to the age of the Universe; for instance, H
1
= 2t for w = 1/3
and H
1
= 3t/2 for w = 0, t being the Universes age.
The exponential expansion (12) possesses no singularity (at nite times), being the
Hubble parameter a constant. A fundamental ingredient of this ination is that the right
hand side of Eq. (5) is positive, a >0, and this is performed when +3p <0, that is, one
does not have necessarily to impose the stronger condition w = 1, but it sufces that
w < 1/3, in order to have a moderate inationary solution; for example, w = 2/3 it
implies a = a

t
2
, a mild power-law ination.
Since the scale factor evolves as a smooth function of time (most of the time!), one
is able to use it as a variable, instead of time, in such as a way that d/dt = aHd/da.
This change of variable helps to integrate the continuity equation for non-constant w(a)
to obtain:
(a) =
0
e
3

[1+w(a)]da/a
. (13)
If, for instance, one parametrizes dark energy through an analytic function of the scale
factor, w(a), one immediately obtains its solution in terms of
t =

8G(a)/3
da
a
. (14)
In cosmology, typical times and distances are determined mainly by the Hubble
parameter, and in practice measurements are often related to redshift, as measured
from stars, gas, etc. It is then useful to express the Friedmann Eq. (4) in terms of the
redshift. The standard model of cosmology considers a Universe lled baryons, photons,
neutrinos, CDM, and a cosmological constant (), and is termed CDM for short. For
this model one obtains:
H
2
= H
2
0
i

(0)
i
(1+z)
3(1+w
i
)
, (15)
where w
i
is the equation of state parameter for each of the uids considered. The present
contribution of the main energy components can be tted from different cosmological
probes, obtaining
(0)
b
= 0.046 0.002,
(0)
DM
= 0.23 0.01, and
(0)

= 0.73 0.02,
together with a Hubble constant of 70.4
+1.3
1.4
km/s/Mpc [10]. Around 96% of mater-
energy the Universe is composed of dark components!
In general, if dark energy is a function of the redshift, fromEq. (13) one can generalize
the above equation to:
H(z)
2
/H
2
0
=
(0)
m
(1+z)
3
+
(0)

(1+z)
4
+
(0)
k
(1+z)
2
+
(0)
DE
f(z), (16)
where m stands for dark matter and baryons, for photons, and
f(z) = exp

z
0
1+w(z

)
1+z

dz

. (17)
Eq. (14) gives the age of the Universe in terms of the redshift, H
0
, and the density
parameters:
t
0
= H
1
0

0
dz
(1+z)H(z)
. (18)
When combining different cosmological probes one obtains for the CDM model an
age of t
0
= 13.750.11 Gyr [10].
Cosmic distances and their measurements
It is useful to write the FRW metric, Eq. (1), in terms of a new distance coordinate ()
ds
2
=dt
2
+a
2
(t)

d
2
+ f
2
k
()(d
2
+sin
2
d
2
)

, (19)
where
f
k
() =

sin , k = +1,
, k = 0,
sinh , k =1.
(20)
Now we proceed to dene some cosmic distances necessary to understand the cosmic
physics.
Causal horizon. The region of space that can be connected to some other region by
causal physical processes, at most through the propagation of light, implies ds
2
= 0. For
the FRW Eq. (1) or (19), in spherical coordinates with , =const. implies that [15, 13]:

H
=

H
0
d =

t
0
t
dt

a(t

)
=
1
a
0
H
0

z
0
dz

E(z

)
(21)
this is the so-called comoving distance, being the distance between two points in the
Universe in which the expansion is factored out. Now, the causal or particle horizon, d
H
is given by:
d
H
(t) = a(t)

H
0
d = a(t)

t
t

dt

a(t

)
. (22)
In order to analyze the whole horizon evolution, fromnowadays (t
0
) to the Planck time
(t
Pl
), we have to consider all the Universe stages, but for brevity we shall not include the
current accelerated expansion. We rstly compute the horizon for the matter dominated
era t
eq.
t t
0
and secondly for the radiation era t t
eq.
, because they are differently
determined by Eq. (11), where we set t

= 0 for convenience. For the matter epoch one

has a(t) = a
0
(t/t
0
)
2/3
, then the rst equation above gives
H
=
3
a
0
(t
2
0
t)
1/3
; from the
second equation one obtains the horizon d
H
(t) = 3t = 2H
1
. For the radiation period,
one has that
H
=
2
a
eq.
(t
eq.
t)
1/2
and d
H
(t) =2t =H
1
. We see, for the matter dominated
era, the causal horizon is twice the Hubble distance, H
1
(sometimes called Hubble
horizon), and they are equal to each other during the radiation dominated era; therefore,
one uses them interchangeably. It is clearly seen for both eras that as t 0, the Universe
is causally disconnected, being a(t) > d
H
(t). But, on the other side, by that time the
CMBR was already highly isotropic and with a black body spectrum. Then, one has to
take for granted that the initial conditions for all small horizon volumes were very ne
tuned to account for the present observed large angle CMBR levels of isotropy, with
T/T few10
5
. This is the horizon problem.
Event Horizon. The event horizon, d
e
, determines the region of space which will
keep in causal contact after some time; that is, it delimits the region from which one can
ever receive (up to some time t
max
) information about events taking place now (at the
time t):
d
e
(t) = a(t)

t
max
t
dt

a(t

)
. (23)
For a at model during its matter dominated era (a t
2/3
), d
e
as t
max
.
Luminosity distance. The luminosity distance is the distance measured using the
energy ux (F) observed by a light source with absolute luminosity (L
s
):
d
2
L

L
s
4F
, (24)
where F = L
0
/S, being L
0
the observed luminosity and S = 4(a
0
f
k
())
2
is the sphere
area at z = 0. The luminosity distance becomes
d
2
L
= (a
0
f
k
())
2
L
s
L
0
. (25)
If we express the energy emitted by a light pulse in a time interval t
1
as E
1
, the
absolute luminosity is given by L
s
=E
1
/t
1
. Similarly we dene L
0
=E
0
/t
0
, where
E
0
is the detected energy in a time t
0
. On the other hand, since the photon energy can
be expressed in terms of its wavelenght (), one has that E
1
/E
0
=
0
/
1
= 1+z and
moreover c = 1, being constant implies that
1
/t
1
=
0
/t
0
, from which we nally
have that
L
s
L
0
=
E
1
E
0
t
0
t
1
= (1+z)
2
, (26)
and the luminosity distance becomes
d
L
= a
0
f
k
()(1+z). (27)
Since f
k
() depends on and this on the redshift, cf. Eq. (21), thus by measuring
the luminosity distance, we can determine the expansion rate of the Universe. For the
CDM model one nds that [16]
d
L
=
(1+z)
H
0

z
0
dz

0
i
(1+z

)
3(1+w)
. (28)
The luminosity distance becomes larger when the cosmological constant is present and
this is what was found to t better the supernovae Ia.
In practice, one uses the relationship of the apparent (m) and absolute (M) magnitude,
related to the luminosity measured at present and when emitted, respectively to have:
mM = 5Log
10

d
L
Mpc

+25. (29)
By adjusting best t curves to their data, two different supernova groups [17] found
a clear evidence for in the late 90s. The presence of a cosmological constant makes
the Universe not only expanding, but accelerating and, in addition, its age is older, not
conicting with globular cluster ages. With the course of the years, various supernova
groups have been getting more condence that the data is compatible with the presence
of dark energy, dark matter, and a high value of Hubble parameter. One of the latest data
release, the Union2 compilation [18], reports that the at concordance CDM model
remains an excellent t to the data with the best t constant equation of state parameter
w =0.997
+0.050
0.054
for a at Universe, and w =1.035
+0.055
0.059
with curvature. Also, they
found that
m
= 0.270 0.021 (including baryons and DM) for xed
k
= 0. That is,

= 0.7300.021.
Angular diameter distance. The angular diameter distance is given by
d
A

, (30)
where is the angular aperture of an object of size x orthogonal to the line of sight
in the sky. Usually this distance is used in the CMBR anisotropy observations, since the
source emitting the radiation is on a surface of a sphere of radius with the observer
located at the center; it is also used in the determination of the BAO feature, see below.
Thus, the size x at the time t
1
in the Friedmann metric, Eq. (19), is given by
x = a(t
1
) f
k
(). (31)
Thus, the angular diameter distance is
d
A
= a(t
1
) f
k
() =
a
0
f
k
()
1+z
(32)
and comparing it with Eq. (27) one has
d
A
=
d
L
(1+z)
2
, (33)
which is called duality relationship. Eq. (33) is valid beyond the FRW metric. In fact, it
is valid for any metric in which the ux is conserved.
THE PHYSICAL UNIVERSE
In the following we provide with theoretical tools to understand the physics of the early
Universe. We treat some micro-physics that rules the interactions of the particles and
elds.
Thermodynamics in the early Universe
In the early Universe one considers a plasma of particles and their antiparticles, as
originally was done by Gamow [1], who has rst considered a physical scenario for
the hot big bang model for the Universes beginning. Later on, with the development
of modern particle physics theories in the 70s it was unavoidable to think about a
physical scenario which should include the new physics for the early Universe. It was
also realized that the physics described by GR should not be applied beyond Planckian
initial conditions, because there the quantum corrections to the metric tensor become
very important, a theory which is still in progress. So the things, one assumes at some
early time, t
>

t
Pl
, that the Universe was lled with a plasma of relativistic particles which
include quarks, leptons, and gauge and Higgs bosons, all in thermal equilibrium at a very
high temperature, T, with some gauge symmetry dictated by a particle physics theory.
Theoretically, in order to work in that direction one introduces some thermodynamic
considerations necessary for the description of the physical content of the Universe,
which we would like to present here. Assuming an ideal-gas approximation, the number
density n
i
of the particles of type i, with a momentum q, is given by a Fermi or Bose
distribution [19]:
n
i
=
g
i
2
2

q
2
dq
e
(E
i

i
)/T
1
, (34)
where E
i
=

m
2
i
+q
2
is the particle energy,
i
is the chemical potential, the sign (+)
applies for fermions and () for bosons, and g
i
is the number of spin states. One has
that g
i
= 2 for photons, quarks, baryons, electron, muon, tau, and their antiparticles,
but g
i
= 1 for neutrinos because they are only left-handed. For the particles existing
in the early Universe one usually assumes that
i
= 0: one expects that in any particle
reaction the
i
are conserved, just as the charge, energy, spin, and lepton and baryon
number, as well. For photons, which can be created and/or annihilated after some
particles collisions, its number density, n

, must not be conserved and its distribution

with

= 0, E = q = h, reduces to the Planckian one. For other constituents, in order

to determine the
i
, one needs n
i
; one notes from Eq. (34) that for large
i
> 0, n
i
is
large too. One does not know n
i
, but the WMAP data constrains the baryon density at
nucleosynthesis such that [20]:

n
B
n

n
baryons
n
antibaryons
n

= 6.140.2510
10
. (35)
The smallness of the baryon number density, n
B
, relative to the photons, suggests that
n
leptons
may also be small compared with n

. Therefore, one takes for granted that

i
=0
for all particles. Why the ratio n
B
/n

is so small, but not zero, is one of the puzzles of

the standard model of cosmology called baryogenesis, as we will explain below.
The above approximation allows one to treat the density and pressure of all particles as
a function of the temperature only. According to the second law of thermodynamics, one
has [13]: dS(V, T) =
1
T
[d(V)+pdV], where S is the entropy in a volume V a
3
(t) with
= (T), p = p(T) in equilibrium. Furthermore, the following integrability condition

2
S
TV
=

2
S
VT
is also valid, which turns out to be
dp
dT
=
+ p
T
. (36)
On the other hand, the energy conservation law, Eq. (6), leads to
a
3
(t)
dp
dt
=
d
dt
[a
3
(t)( + p)] (37)
and using Eq. (36), the latter takes the form
d
dt
[
a
3
(t)
T
( + p)] = 0, and using Eq. (36)
again, the entropy equation can be written as dS(V, T) =
1
T
d[( + p)V]
V
T
2
( + p)dT.
Last two equations imply that the entropy is a constant of motion:
S =
a
3
T
[ + p] = const. . (38)
The density and pressure are given by

E
i
n
i
dq , p

q
2
3E
i
n
i
dq . (39)
For photons or ultra-relativistic uids, E = q, these equations become such that p =
1
3
,
and then conrming Eq. (7) with w = 1/3, and after integrating Eq. (36), it comes out
that
= bT
4
, (40)
with the constant of integration, b. In a real scenario there are many relativistic particles
present, each of which contributes like Eq. (40). By including all of them, =
i

i
and p =
i
p
i
over all relativistic species, one has that b(T) =

2
30
(N
B
+
7
8
N
F
), which
depends on the effective relativistic degrees of freedom of bosons (N
B
) and fermions
(N
F
); therefore, this quantity varies with the temperature; different ispecies remain
relativistic until some characteristic temperature T m
i
, after that the value N
F
i
(or
N
B
i
) contributes no more to b(T). The factor 7/8 accounts for the different statistics the
particles have, see Eq. (34). In the standard model of particles physics b 1 for T 1
MeV and b 35 for T > 300 GeV [19]. Also for relativistic particles, one obtains from
Eq. (34) that
n = cT
3
, with c =
(3)

2
(N
B
+
3
4
N
F
) . (41)
where (3) 1.2 is the Riemann zeta function of 3. Nowadays, n

422
cm
3
T
3
2.75
, where
T
2.75

0
2.75

K
.
From Eq. (38), and using the relativistic equation of state given above (w = 1/3), one
gets that T 1/a(t) and from its solution in Eq. (11) one has,
T =
4

M1
3
b
1
a(t)
=
4

3
32Gb
1
(t t

)
1
2
, (42)
a decreasing temperature behavior as the Universe expands. Then, initially at the big
bang t =t

implies T

=, the Universe was not only very dense but also very hot.
The entropy for an effective relativistic uid is given by Eq. (38) together with its
equation of state and Eq. (40), S =
4
3
b (a T)
3
= const. Combining this with Eq. (42),
one can compute the value of M1
3
to be M1
3
= (
3
4
S)
4/3
/b
1/3
10
116
, since b 35 and
the photon entropy S
0
=
4
3
b (a
0
T
0
)
3
10
88
for the nowadays evaluated quantities
a
0
d
H
(t
0
) = 10
28
cm and T

0
= 2.7

K. One denes the entropy per unit volume,
entropy density, to be s S/V =
4
3

2
30
(N
B
+
7
8
N
F
)T
3
, then, nowadays s 7n

. The
nucleosynthesis bound on , Eq. (35), implies that n
B
/s 10
11
.
Now we consider particles in their non-relativistic limit (m T). From Eq. (34) one
obtains for both bosons and fermions that
n = g

mT
2

3/2
e
m/T
. (43)
The abundance of equilibrium massive particles decreases exponentially once they be-
come non-relativistic; this situation is referred as in equilibrium annihilation. Their den-
sity and pressure are given through Eqs. (39) and (43) by = nm and p = nT .
Therefore, the entropy given by Eq. (38) for non-relativistic particles, using last two
equations, diminishes also exponentially during their in equilibrium annihilation. The
entropy of these particles is transferred to that of relativistic components by augmenting
their temperature. Hence, the constant total entropy is essentially the same as the one
given above, but the ispecies contributing to it are just those which are in equilibrium
and maintain their relativistic behaviour, that is, particles without mass such as photons.
Having introduced the abundances of the different particle types, we would like to
comment on the equilibriumconditions for the constituents of the Universe, as it evolves.
This is especially of importance in order to have an idea whether or not a given ispecies
disappears or decouples from the primordial brew. To see this, let us consider n
i
when
the Universe temperature, T, is such that (a) T m
i
, during the ultra-relativistic stage
of some particles of type i and (b) T m
i
, when the particles i are nonrelativistic, both
cases rst in thermal equilibrium. From Eq. (41) one has for the former case that n
i
T
3
;
the total number of particles, n
i
a
3
, remains constant. Whereas for the latter case, from
Eq. (43), n
i
T
3/2
e
m
i
/T
, i.e., when the Universe temperature goes down below m
i
, the
number density of the ispecies signicantly diminishes; it occurs an in equilibrium
annihilation. Let us take as example the neutronproton annihilation, one has
n
n
n
p
e
m
p
m
n
T
= e

1.510
10
K
T
, (44)
which drops with the temperature, from near to 1 at T 10
12
K to about 5/6 at
T 10
11
K, and 3/5 at T 3 10
10
K [21]. If this is forever valid, one ends
without massive particles, and our Universe should have consisted only of radiative
components; our own existence prevents that! Therefore, eventually the in equilibrium
annihilation had to be stopped. The quest is nowto freeze out this ratio to be n
n
/n
p
1/6
(due to neutron decays, until the time when nucleosynthesis begins, n
n
/n
p
reduces
to 1/7) in order to leave the correct number of hadrons for later achieving successful
nucleosynthesis. The answer comes from comparing the Universe expansion rate, H,
with particle physics reaction rates, . Hence, for H < , the particles interact with
each other faster than the Universe expansion rate, then equilibrium is established.
For H > the particles cease to interact effectively, then thermal equilibrium drops
out. This is only approximately true; a proper account of that involves a Boltzmann
equation analysis. For that analysis numerical integration should be carried out in which
annihilation rates are balanced with inverse processes, see for example [22, 19]. In
this way, the more interacting the particles are, the longer they remain in equilibrium
annihilation and, therefore, the lower their number densities are after some time, e.g.,
baryons vanish rst, then charged leptons, neutral leptons, etc.; nally, the massless
photons and neutrinos, whose particle numbers remain constant, as it was mentioned
above. Note that if interactions of an ispecies freeze out when it is still relativistic,
then its abundance can be signicant nowadays.
It is worth to mention that if the Universe would expand faster, then the temperature
of decoupling, when H , would be higher, then the xed ratio n
n
/n
p
must be greater,
and the
4
He abundance would be higher, thus leading to profound implications in the
nucleosynthesis of the light elements. Thus, the expansion rate cannot arbitrarily be
modied during the equilibrium era of some particles. Furthermore, if a particle species
is still highly relativistic (T m
i
) or highly non-relativistic (T m
i
), when decoupling
from primordial plasma occurs, it maintains an equilibrium distribution; the former
characterized by T
r
a =const. and the latter by T
m
a
2
=const., cf. Eq. (47).
There are also some other examples of decoupling, such as neutrino decoupling: dur-
ing nucleosynthesis there exist reactions, e.g. e
+
e

, which maintain neutrinos

efciently coupled to the original plasma ( >H) until about 1 MeV, since

H

T
MeV

3
.
Below1 MeVreactions are no more efcient and neutrinos decouple and continue evolv-
ing with a temperature T

1/a. Then, at T
>

m
e
=0.51MeVthe particles in equilibrium
are photons (with N
B
= 2) and electron and positron pairs (with N
F
= 4) to contribute
to the entropy with b(T) =

2
30
(11/2). Later, when the temperature drops to T m
e
, the
reactions are no more efcient ( < H) and after the e

pair annihilation there are only

photons in equilibrium with b(T) =

2
30
(2). Since the total entropy, S =
4
3
b(aT)
3
, must
be conserved, the decrease in b(T) must be balanced with an increase in the radiation
temperature, then one has that
T

11
4

1/3
, which should remain so until today, im-
plying the existence of a cosmic background of neutrinos with a temperature today of
T

0
= 1.96

K. This cosmic relic has not been measured yet.
Another example of that is the gravitation decoupling, which should be also present
if gravitons were in thermal equilibrium at the Planck time and then decouple. The
today background of temperature should be characterized at most by T
grav.
=

4
107

1/3

0.91

K.
For the matter dominated era we have stressed that effectively p = 0; next we will see
the reason of this. First consider an ideal gas (such as atomic Hydrogen) with mass m,
then = nm+
3
2
nT
m
and p = nT
m
. From Eq. (37) one obtains, equivalently, that
d
da
(a
3
(t)) =3pa
2
(t) (45)
and substituting the above and p, one has that
d
da
(nma
3
(t) +
3
2
nT
m
a
3
(t)) =3nT
m
a
2
(t) (46)
where nma
3
(t) is a const. This Eq. yields that
T
m
a
2
(t) = const. , (47)
the matter temperature drops faster than that of radiation as the Universe expands, cf. Eq.
(42). Now, if one considers both radiation and matter, one has that =nm+
3
2
nT
m
+bT
4
r
and p = nT
m
+
1
3
bT
4
r
; the source of Universes expansion is proportional to +3p =
nm+
9
2
nT
m
+2bT
4
r
, the rst term dominates the second, precisely because T
m
decreases
very rapidly. The third term diminishes as 1/a
4
, whereas the rst as 1/a
3
, and after
the time of densities equality,
m
=
r
, the matter density term is greater than the others,
that is why one assumes no pressure for that era.
From now on, when we refer to the temperature, T, it should be related to the ra-
diation temperature. The detailed description of the Universe thermal evolution for the
different particle types, depending on their masses, cross-sections, etc., is well described
in many textbooks, going from the physics known in the early 70s [13] to the late 80s
[19], and therefore it will not be presented here. However, we notice that as the Uni-
verse cools down a series of spontaneous symmetrybreaking (SSB) phase transitions
are expected to occur. The type and/or nature of these transitions depend on the specic
particle physics theory considered. Among the most popular ones are Grand Unica-
tion Theories (GUTs), which bring together all known interactions except for gravity.
One could also be more modest and just consider the standard model of particle physics
or some extensions of it. Ultimately, one should settle, in constructing a cosmologi-
cal theory, up to which energy scale one wants to describe physics. For instance, at a
temperature between 10
14
GeV to 10
16
GeV the transition of the SU(5) GUT should
took place, if this theory would be valid, in which a Higgs eld breaks this symmetry
to SU(3)
C
SU(2)
W
U(1)
HC
, a process through which some bosons acquired their
masses. Due to the gauge symmetry, there are color (C), weak (W) and hypercharge
(HC) conservation, as the subindices indicate. Later on, when the Universe evolved to
around 150 GeVthe electroweak phase transition took place in which the standard model
Higgs eld broke the symmetry SU(3)
C
SU(2)
W
U(1)
HC
to SU(3)
C
U(1)
EM
;
through this breaking fermions acquired their masses. At this stage, there were only
color and electromagnetic (EM) charge conservation, due to the gauge symmetry. Af-
terwards, around a temperature of 175 MeV the Universe should underwent a transition
associated to the chiral symmetrybreaking and color connement from which baryons
and mesons were formed out of quarks. Subsequently, at approximately 10 MeV begun
the synthesis of light elements (nucleosynthesis), when most of the today observed Hy-
drogen, Helium, and some other light elements abundances were produced. So far the
nucleosynthesis represents the earliest scenario tested in the standard model of cosmol-
ogy. After some thousand years (z 3200 [10]), the Universe is matter dominated, over
the radiation components. At about 380, 000 years (z 1090 [10]) recombination took
place, that is, the Hydrogen ions and electrons combined to compose neutral Hydrogen
atoms, then matter and EM radiation decoupled from each other; at this moment (bary-
onic) matter structure begun to form. Since that moment the surface of last scattering
of the CMBR evolved as an imprint of the early Universe. This is the light that Penzias
and Wilson rst measured, and was later measured in more detail by BOOMERANG,
MAXIMA, COBE, and WMAP, among other probes. PLANCK cosmological data will
be forthcoming in 2012.
Ination: the general idea
As we mentioned above, the FRW cosmological Eqs. (4)-(6) admit very rapid ex-
panding solutions for the scale factor. This is achieved when +3p, is negative, i.e.,
when the equation of state admits negative pressure such that w <1/3, to have a > 0.
For instance, if w = 2/3, one has that a t
2
and 1/a, that is, the source of rapid
expansion decreases inversely proportional with the expansion. Of special interest is the
case when w =1, = const., because this guarantees that the expansion rate will not
diminish. Thus, if =const. is valid for a period of time, , the Universe will experience
an expansion of N = H foldings, given by a = a

e
N
, Eq. (12). This is the well known
de Sitter cosmological solution [23], achieved here only for a -stage in a FRW model.
We shall now see how an inationary stage helps to solve the horizon and atness
problems of the old standard cosmology. Firstly consider the particle (causal) horizon,
given by Eq. (22), during ination, again with k = 0, one obtains
d
H
= H
1
(e
Ht
1) , (48)
the causal horizon grows exponentially, whereas H
1
remains constant. We compare the
horizon distance with that of any physical length scale, L(t) = L

a(t)
a

= L

e
Ht
, to get
d
H
L
=
H
1
(e
Ht
1)
L

e
Ht
>

1e
Ht
, (49)
for initial length scales L

<

H
1
. After a few e-fold times the causal horizon is as big
as any length scale that was initially subhorizon sized. Therefore, if the original patch
before ination is causally connected, and presumably in equilibrium, then after ination
this region of causality is exponentially bigger than it was, and all the present observed
(apparent) Universe can stemfromit, solving the horizon problem. In fact, if the ination
stage is sufciently large, there can exist nowadays regions which are so distant away
from each other that they are still not in contact, even though originally they come from
the same causal patch existing before ination. These regions will be for a time not in
contact since we currently are experiencing an accelerated expansion, and then, if this is
of exponential type, the event horizon is constant and light/information that shall come
to us will be from only a delimited region H
1
, as we explain below.
From Eq. (49) one can observe that if the initial physical length scale is greater than
the Hubble distance, L

> H
1
, then d
H
< L during ination. Events initially outside
the Hubble horizon remain acausal. This is better seen by considering the event horizon,
d
e
, dened in Eq. (23). This delimits the region of space which will keep in causal
contact after some time; that is, it delimits the region from which one can ever receive
(up to some time t
max
) information about events taking place now (at the time t). During
ination one has that
d
e
= H
1
(1e
(t
max
t)H
) H
1
, (50)
which implies that any observer sees only those events that take place within a distance
H
1
. In this respect, there is an analogy with black holes, from whose surface
no information can get away. Here, in an exponential expanding Universe, observers
encounter themselves in a region which apparently were surrounded by a black hole
[24, 25], since they receive no information located farther than H
1
.
The apparent horizon at present stems from a region delimited by the original patch
d
e
H
1
, which during ination remains almost constant and, afterwards, evolves as
H
1
t. At the end of ination a(t) H
1
(t)/H
1
0
. Subsequently, the scale factor
expands only with the power law solution t
1/2
(and later as t
2/3
), whereas the Hubble
horizon evolves faster, H
1
t. Then, at some later time the Hubble horizon is as large
as the scale factor, H
1
a(t)H
1
0
. Accordingly, there is a minimal number of efolds
of ination, N 60, necessarily to have this equality at present (this number depends
on the energy scale of ination, see for instance J. L. Cervantes-Cota in [5]); that is, the
original patch grown until now is as big as our apparent, Hubble horizon. Hence, some
time ago, say, at the last scattering surface (photon decoupling) the Universe consisted
of 10
5
Hubble horizon regions, yet all these regions stem from one original patch of size
H
1

at the start of ination.

A typical scale L

H
1
will increase exponentially its size as L(t) = L

a(t)
a

= L

e
N
.
That is, all physical inhomogeneities, anisotropies and/or perturbations of any kind
(including particles!) will be diluted away from a region d
e
H
1
, and its density
becomes insignicant, thus solving the monopole (and other relics) problem.
On the other hand, the atness problem in the old standard cosmology arises since
approaches closely to unity as one goes back in time in a way that one has to choose
very special initial density values, at the Planck time
Pl
1 10
59
, for explaining
our atness today, i.e.,
0
O(1). Now, imagine the Universe with initial conditions
such that

1 k. Now, if the exponential expansion occurs, (t = ) evolves to

() 1 =

c

c
=
k
a
2
H
2
= ke
2N
. (51)
If N is sufciently large, which will be case since typically N > 60, the Universe looks
after a de Sitter stage like an almost perfect at model. Therefore, it plays almost no
role what the initial density was, if the exponential expansion occurs -guaranteed by
the cosmological no hair- the Universe becomes effectively at. In this way, instead of
appealing to very special initial conditions, one starts with an Universe with more normal
conditions, that is, non ne tuned, which permit the Universe to evolve to an inationary
stage, after which it looks like it would had very special conditions, i.e., with 1 with
exponential accuracy.
After ination the Universe contains a very small particle density and is very cold,
even as cold as the CMBR is today! The transition to a radiation dominated era with
sufcient entropy and particle content comes from the decaying or transformation of
the energy source of ination, =V(0), into heat; a process called reheating (RH).
Reheating and baryogenesis
At the end of ination the -eld (the inaton) begins to oscillate around its stable,
global minimum, say v. Its oscillation frequency is given by the effective mass of the
Klein Gordon equation in a FRW Universe,

+3H

+V

() = 0. After ination the

term V

(V

= M
2
) is greater than 3H

, since the Hubble rate evolves hereafter
always decreasing, H 1/t. Thus, the oscillation frequency of the -eld is simply
given by the eld mass, M. The stored energy of the inaton eld,
H
=V +
1
2

2
, can
decay to give rise to quantum particle creation [26]. The reheating models depend on
the particle physics models but general features of the process have been understood.
We rstly explain the old scenario called reheating and secondly the preheating that
seems to be more realistic. In reheating, the state = v is considered as a coherent
state of scalar particles in rest. Then, this state decays through the ordinary decay of
the eld bosons and the decay rate coincides with the rate of decrease of the energy of
oscillations. Thus, the decay rate of the boson,
H
, introduces a friction term of the type

H

in the Klein Gordon equation that causes the scalar eld to vanish and the reheating
of the Universe [26]. The transformed energy goes into masses and kinetic energy of
the new particles produced. Typically, the produced particles (bosons, fermions) have
smaller masses than the eld boson, therefore much of their energy goes into kinetic
energy, and particles behave as a relativistic uid. If the decay rate is greater than the
Hubble rate after ination,
H
> H(t = t
f
) H
f
, then the reheating process occurs
within an expansion time, very rapidly. In this case, the reheating temperature is [27]
T
RH

H
b

1/4
f
=

V+
1
2

2
b

1/4
f
, where b =

2
30
(N
B
+
7
8
N
F
), N
B
(N
F
) stands for boson
(fermion) degrees of freedom. For high temperatures, T >300 GeV, one has that b 35,
see discussion after Eq. (40). The subindex f means to be evaluated at the end of
ination. There, the kinetic and potential energies are of the same order of magnitude.
Then, for a self-interaction potential with V V(0) =
1
24
v
4
, one has that
T
RH

v
4
24b

1/4
=

Mv
4

8b
. (52)
One the other hand, if the decay process is rather slow,
H
< H
f
, then the eld
continues to oscillate coherently until t =
1
H
. During this time the solution of the Klein
Gordon equation is
1
t
cosMt, H =
2
3t
. The coherent oscillations behave as non-
relativistic matter uid (w = 0), i.e., a t
2/3
[28]. If eld bosons do not completely
decay, the oscillations represent a sea of cold bosons with M T. They can account
for the cold dark matter, but some degree of ne tuning is necessary [29].
Without partial or total decaying of eld oscillations the Universe remains cold
and devoid of fermions and (other) bosons. Therefore, let us suppose that indeed re-
heating took place, but now with
H
<

H
f
, then the reheated temperature is T
RH

(

Mv/(
4

8b)

H
/H
f
, a factor

H
/H
f
smaller than the efcient reheating case,
Eq. (52). The reheating process occurs normally within one or few Hubble times. Then,
the scale factor does not increase signicantly during it.
The reheating scenario presented above is based on the original theory developed
in the context of the new inationary scenario, however, it is also applicable to other
models. In the course of the time important steps to consolidate the theory were made,
see for example Ref. [30]. But qualitative new ideas were introduced in Refs. [31].
Accordingly, the process of reheating should consist of three different stages. At the rst
phase, the -eld decays into massive bosons (fermions) due to a parametric resonance
given through a Mathieu equation that determines the regions of stability and instability
(particle production) of the quantumuctuations of the created particles. These can be -
particles or other bosons (fermions) coupled to the -eld. This process is very efcient,
even explosive, and much bosons can be created in this stage. Note that the original
theory is based upon the decay of the -particles, whereas in the present theory the -
eld decays into -particles, and perhaps others, and only after this process the decay
of these particles proceeds. Then, to distinguish this explosive process from the normal
stage of particle decay, the authors of Ref. [31] called it preheating. Bosons produced
at this stage are far away from thermal equilibrium and have very big occupational
numbers. The second stage of this scenario describes the decay of the already produced
particles. This phase is described as in the original reheating theory. Then, the methods
developed for the original theory are now applied to the product particles, but not itself
to the decay of the -eld. The third stage is the thermalization by which the system
reaches equilibrium [32].
The process of reheating is very complex and depends ne on the particle physics
theory one has in turn. As a matter of fact, one expects a reheat temperature T
RH
>
few MeV to be able to attain nucleosynthesis. A second, and more restrictive constraint
comes from baryogenesis. One can see this by noting that the number density of any
conserved quantity before reheating divided by the entropy density
n
s
becomes after
reheating insignicant because of the huge entropy produced. One gets
n
s
|
r f
= e
3N n
s
|
ri
,
where ri and r f denote the initial and nal state of reheating, respectively. In this
way, any baryon asymmetry initially present will be brought to unmeasurable values.
Therefore, after reheating the baryon asymmetry must be created. Note also that any
unwanted relic (x), accounted through
n
x
s
, will essentially disappear after reheating. The
correct baryon-antibaryon balance at nucleosynthesis is n
B
/s 10
11
, in consistency
with Eq. (35). There are some attempts to achieve baryogenesis at low energy scales, as
low as few GeV or TeV [33]. Recent attempts to solve this problem seek to yield a prior
a lepton asymmetry, leptogenesis, generated in the decays of a heavy sterile neutrino
[34], to later end with baryogenesis.
THE PERTURBED UNIVERSE
In the previous sections we have outlined how the evolution of a homogeneous Universe
can be described by means of fewequations and simple concepts such as the ideal perfect
uids. The next step is introducing in this scenario small inhomogeneities that can be
treated as rst order perturbations to those equations, the goal being the description of
the structures we see today in the Universe. This perturbative approach is sufcient to
accurately describe the small temperature anisotropies (T/T 10
5
) observed in the
CMBR today, but can describe the distribution of matter today only at those scales that
are still in the linear regime. At the present epoch, scales smaller than 30 Mpc h
1
[35] have already entered the non linear-regime (/ >> 1) due to the fact that matter
tends to cluster under the effect of gravity. These scales can therefore only be described
by means of numerical or semi-numerical approaches [36].
The approach is quite straightforward but involves a differential equation for the
density perturbation of each individual constituent: scalar elds in ination, or baryons,
radiation, neutrinos, DM, and DE (usually treated as cosmological constant) in later
times, and in general needs to be solved numerically. In the context of geometry metric
and/or GR the metric is treated as the general expansion term g
(0)

plus perturbation h

:
g

= g
(0)

+h

, (53)
with h

<< g
(0)

where
(0)
indicates the unperturbed homogeneous quantities.
Inhomogeneities in the distribution of the components of the Universe are a source of
scalar perturbations of the metric. Nevertheless vector or tensor perturbations can modify
the metric as well. The standard cosmological model do not predict vector perturbations,
that would introduce off-diagonal terms in the metric tensor. These perturbations would
produce vortex motions in the primordial plasma that are expected to rapidly decay.
Models with topological defects or inhomogeneous primordial magnetic elds instead
predict a consistent fraction of vector perturbations [37, 38, 39].
On the other hand, the standard cosmological model predicts the production of gravi-
tational waves during the epoch of ination, when the Universe expanded exponentially,
as we will see below. Gravitational waves induce tensor perturbations h
T

on the metric
of the type:
h
T

= a
2

0 0 0 0
0 h
+
h

0
0 h

h
+
0
0 0 0 0

where h
+
and h

are the polarization directions of the gravitational wave. This tensor

is traceless, symmetric and divergentless, i.e. it perturbs the time space orthogonally
to the propagation direction of the wave. The amplitude of these tensor perturbations
are expected to be small compared to scalar ones, and therefore negligible in rst
approximation as far as we are interested in studying the perturbations of the metric
tensor. Nevertheless these waves are expected to leave an imprint in the polarization
of the CMBR, and their eventual detection would unveil an extremely rich source of
information about an epoch of the Universe that is very hardly observable otherwise.
It is important to underline that choosing to model the metric perturbations corre-
sponds to choosing a gauge, i.e. a specic coordinate system in which the metric tensor
is represented. Changing the coordinate system of course do not change the physics, but
can remarkably vary the difculty of the calculations and the understanding of the phys-
ical meaning of the different quantities. To solve the perturbed equations one chooses
convenient gauges for the different expansion epochs and depending on whether the
formalism is theoretical or numerical, as we will see below.
The presence of weak inhomogeneous gravitational elds introduce small pertur-
bations in the metric tensor that can be modeled by introducing two scalar functions
(x, )) and (x, ) [40] in the Robertson-Walker metric as:
ds
2
= a
2
()

[1+2(x, )] d
2
+[1+2(x, )]dx
i
dx
i

, (54)
where therefore the perturbed part of the metric tensor is:
h
00
(x, ) =2(x, ), h
0i
(x, ) = 0, h
i j
(x, ) = a
2

i j
(2(x, )). (55)
This metric is just a generalization of the well known metric for a weak gravitational
eld usually presented in text books (e.g. Chapt. 18 in Misner [7]) for the case of
a static Universe (a() = 1). The function describes Newtons gravitational eld,
while is the perturbation of the space curvature. The above gauge is the Newtonian
conformal gauge, which has the advantage of having a diagonal metric tensor g

in
which the coordinates are totally xed with no residual gauge modes and therefore with
a straightforward interpretation of the functions introduced [40, 41, 42]. An example of
an alternative gauge particularly popular in literature is the synchronous gauge, generally
dened as:
ds
2
= a
2
()[d
2
+(
i, j
+h
i, j
)dx
i
dx
j
], (56)
which is especially used in codes computing the anisotropies and inhomogeneities in the
Universe, as better behaved numerically by choosing that observers fall freely without
changing their spatial coordinates. A further analysis is found in e.g. [43].
Perturbations during ination
The primeval uctuations are thought to be present at the very beginning of time, at
the inationary epoch. The generation of perturbations are produced by quantum uc-
tuations of the -eld during the accelerated stage, for a review see [40, 5, 43]. These
uctuations are usually studied in the comoving gauge in which the scalar eld is equal
to its perturbed value at any given time during ination and, therefore, the perturbation
information resides in the metric components. The perturbations cross outside the event
horizon during ination and re-enter into the horizon much later, at the radiation and
matter dominated epochs, to yield an almost scale invariant density perturbation spec-
trum (Harrison-Zeldovich, n
S
= 1), as the required for structure formation.
We introduce this topic by noting that the event horizon during a de Sitter stage is
d
e
H
1
, cf. Eq. (50). This means that microphysics can only operate coherently within
distances at most as big as the Hubble horizon, H
1
. Recall that the causal horizon, d
H
,
expands exponentially and it is very large compared to the almost constant H
1
during
ination, see Eq. (48). Hence, during the de Sitter stage the generation of perturbations,
which is a causal microphysical process, is localized in regions of the order of H
1
.
It was shown that the amplitude of inhomogeneities produced corresponds to the
Hawking temperature in the de Sitter space, T
H
= H/(2). In turn, this means that
perturbations with a xed physical wavelength of size H
1
are produced throughout
the inationary era. Accordingly, a physical scale associated to a quantum uctuation,

phys
= a(t), expands exponentially and once it leaves the event horizon, it behaves as
a metric perturbation; its description is then classical, general relativistic. If ination
lasts for enough time, the physical scale can grow as much as a galaxy or horizon
sized perturbation. The eld uctuation expands always with the scale factor and after
ination, it evolves according to t
n
(n = 1/2 radiation or n = 2/3 matter). On the other
hand, the Hubble horizon evolves after ination as H
1
t. This means, it will come a
time at which eld uctuations cross inside the Hubble horizon and re-enters as density
uctuations. Thus, ination produces a gross spectrum of perturbations, the largest scale
originated at the start of ination with a size H
1
i
, and the smallest with H
1
f
at the end
of ination. The power spectra for scalar (S) and tensor (T) perturbations are given by:
P
S
(k)

H
2
16
3

2
c

k=aH
, P
T
(k)

H
2
4
2
m
2
Pl

k=aH
, (57)
where

c
is the classical eld velocity. The equations are evaluated at the horizon
crossing (k = aH) during ination. Each of the kmodes generate also an anisotropy
pattern in the CMBR that was measured for scalar perturbations by the COBE [4]
and later probes. The PLANCK satellite may have the chance to detect the ratio of
tensor to scalar amplitudes r C
T
l
/C
S
l
< 0.36 [10], since the tensor modes modulate
CMBR photons coming from last scattering. Associated to these perturbations one has
the spectral indices, n
S
0.96 [10] and n
T
, and their runnings, dn
S
/dlnk 0.034 [10].
These density and other metric perturbations are small, but we discuss in the next
section how to include them so that the information contained can be recognized and
exploited.
Perturbations inside the horizon
In the early Universe, baryons were tightly coupled to photons in an expanding back-
ground. Baryonic matter and dark matter potential wells provoked the local collapse of
density uctuations up certain point, at which the radiation pressure was big enough to
pull out the matter apart, and smoothing the potential wells. These oscillations of the
plasma can be thought of as acoustic waves. As we know any wave can be decomposed
into a sum of modes with different wave numbers, k = 2/. Since these modes are
in the sky, their wavelengths are measured as angles rather than as distances. Accord-
ingly, instead of decomposing the wave in a Fourier series, what is normally done is to
decompose the wave in terms of spherical harmonics, Y
lm
( p). The angular power spec-
trum can be expanded in Legendre polynomials, since there is no preferred direction in
the Universe and that only angular separation is relevant. A mode l plays the same
role of the wavenumber k, thus l 1/. Ultimately, we are interested in the tempera-
ture uctuations that are analyzed experimentally in pairs of directions n and n

where
cos() = n n

. We then average these uctuations obtaining the multipole expansions:

T
T
=

l=1
l

m=l
a
lm
(x, )Y
lm
( p), P
S
() =

(2l +1)
4
C
l
P
l
(cos), (58)
where P
S
() is the angular power spectrum, P
l
are the Legendre polynomials and C
l
is
estimated as the average over m of a
lm
. All this information can be used to determine
the cosmological parameters
i
. We will not discuss detailed calculations nor the curve
that must be adjusted to obtain the best t values for such parameters. The peak of the
fundamental mode appears at approximately
l
200

. (59)
BOOMERANG [44] and MAXIMA [45] were two balloon-borne experiments de-
signed to measure the anisotropies at smaller scales than the horizon at decoupling
(
hordec
1

), hence measuring the acoustic features of the CMBR. The sensitivity

of the instruments allowed a measurement of the temperature uctuations of the CMBR
over a broad range of angular scales. BOOMERANG found a value of l = 1976 and
MAXIMA-1 found a value of l 220. This implies that the cosmological density pa-
rameter 1, see Eq. (9), implying that the Universe is practically at,
k
0. These
two experiments provided with the rst strong evidence for a at Universe from observa-
tions. Happily, this result was expected from Ination. These results were conrmed by
the Wilkinson Microwave Anisotropy Probe (WMAP) in a series of data releases in the
last decade, as well as by other cosmological probes: the Universe is at or pretty close
to be at. The problem in the exact determination of the curvature is because the CMBR
anisotropies show strong degeneracies among the cosmological parameters [46]. How-
ever, the satellite PLANCK will offer results of the density parameters with uncertainties
less than a percentage level.
Since baryons and photons were in thermal equilibrium until recombination (also
called last scattering), the acoustic oscillations were imprinted too in the matter per-
turbations, as they were in the CMBR anisotropies. These are known as baryon acoustic
oscillations (BAO). The sound horizon at the moment when the baryons decoupled from
photons plays a crucial role in determination of the position of the baryon acoustic peaks.
This time is known as drag epoch which happens at z
d
= a
0
/a
d
1. The sound horizon
at that time is dened in terms of the effective speed of sound of the baryon-photon
plasma, c
2
s
p

/(

+
b
),
r
s
(z
d
) =

d
0
d c
s
() =
1
3

a
d
0
da
a
2
H(a)

1+(3
b
/4

)a
. (60)
Note that drag epoch does not coincide with last scattering. In most scenarios z
d
< z
ls
[47]. The redshift at the drag epoch can be computed with a tting formula that is a
function of
(0)
m
h
2
and
(0)
b
h
2
[48]. The WMAP -5 year team computed these quantities
for the CDM model obtaining z
d
= 1020.51.6 and r
s
(z
d
) = 153, 32.0 Mpc [49].
What one measures is the angular position and the redshift [50, 6]:

s
(z) =
r
s
(z
d
)
(1+z)d
A
(z)
, (61)
z
s
(z) = r
s
(z
d
)H(z), (62)
where d
A
(z) is the proper (not comoving) angular diameter distance, Eq. (32), and H(z)
by Eq. (16). The angle
s
(z) corresponds to the direction orthogonal to the line-of-sight,
whereas z
s
(z) measures the uctuations along the line-of-sight. Observations of these
quantities are encouraging to determine both d
A
(z) and H(z). However, from the current
BAO data is not simple to independently measure these quantities. This will certainly
happen in forthcoming surveys [51]. Therefore, it is convenient to combine the two
orthogonal dimensions to the line-of-sight with the dimension along the line-of-sight to
dene [52]:
D
V
(z)

(1+z)
2
d
A
(z)
2
z
H(z)

1/3
, (63)
where the quantity D
M
d
A
/a = (1+z)d
A
(z) is the comoving angular diameter distance.
One also denes the BAO distance
r
BAO
(z) r
s
(z
d
)/D
V
(z). (64)
The BAOsignal has been measured in large samples of luminous red galaxies fromthe
SDSS [52]. There is a clear evidence (3.4) for the acoustic peak at 100h
1
Mpc scale.
Moreover, the scale and amplitude of this peak are in good agreement with the prediction
of the CDM given the WMAP data. One nds that D
V
(z = 0.35) = 1370 64 Mpc,
and more recently new determinations of the BAO signal has been published [53] in
which
s
(z = 0.55) = 3.90

0.38

, and w = 1.03 0.16 for the equation of state

parameter of the dark energy, or
M
= 0.26 0.04 for the matter density, when the
other parameters are xed.
Measuring the BAO feature in the matter distribution at different redshifts will help
to break the degeneracy that exists in the determination of the cosmological parameters.
And by combining line-of-sight with angular determinations of the BAO feature one will
constrain even more the parameter space. Further, a complete combination of BAO, the
full matter power spectrum, Supernovae Ia, and CMBR data shall certainly envisage the
true nature of the mysterious, dark Universe.
Finally, we show recent plots of the CMBR power spectrum from the WMAP 7-
year data and from the South Pole Telescope, from Ref. [54], and the mass variance
(
M
/M =

P(k)k
3
/(2
2
) ) of the reconstructed matter power spectrum from the Ata-
cama Cosmology Telescope and other observations, from Ref. [55]. The left panel shows
the extraordinary t of the CDM model and the importance of foregrounds for large
l-modes. The right panel shows the variance decrease as the mass increases, covering
10
12
10
13
10
14
10
15
10
16
10
17
10
18
10
19
10
20
10
21
10
22
10
23
Mass scale M [Msolar]
10
6
10
5
10
4
10
3
10
2
10
1
10
0
10
1
10
2
M
a
s
s
V
a
r
i
a
n
c
e

M
/
M
SDSS DR7 (Reid et al. 2010)
LyA (McDonald et al. 2006)
ACT CMB Lensing (Das et al. 2011)
ACT Clusters (Sehgal et al. 2011)
CCCP II (Vikhlinin et al. 2009)
BCG Weak lensing (Tinker et al. 2011)
ACT+WMAP spectrum (this work)
FIGURE 1. Left Panel: Recent CMBR angular power spectrum from the WMAP 7-year data and from
the South Pole Telescope observations showas band-averaged powers along with the best t CDMmodel
(CMB - dashed line) and (CMB + foreground solid line), taken from Ref. [54]. The detailed location,
amplitude, and shape of the peaks (bumps) provide information on the contents of the Universe and the
conditions at that early epoch plus secondary effects. Right panel: The mass variance of the reconstructed
matter power spectrum from the Atacama Cosmology Telescope; large masses correspond to large scales
and hence small values of k, taken from Ref. [55]. The BAO are barely visible in the detail of this arching
spectrum driven by the damping, mostly at low mass, of the original perturbations during the oscillations.
ten orders of magnitude in the range of masses. We also notice the effect of BAO at
intermediate scales and damping on the essentially scale invariant perturbations that one
anticipates from ination. These observations t remarkably well to the CDM model.
ACKNOWLEDGMENTS
We gratefully acknowledge support from CONACYT Grant No. 84133-F.
REFERENCES
1. G. Gamov, Phys. Rev. 70 (1946) 572; ibib 74 (1948) 505.
2. E. W. Kolb, and M. S. Turner, The Early Universe: Reprints, Frontiers in Physics # 70 (Addison-
Wesley, 1988).
3. J. C. Mather et al, Astrophys. J. Lett. 354 (1990) L37.
4. G.F. Smoot et al, Astrophys. J. Lett. 396 (1992) L1.
5. N. Breton, J. L. CervantesCota, and M. Salgado, Eds., An introduction to Standard Cosmology in
The Early Universe and Observational Cosmology LNP 646 (SpringerVerlag, 2004).
6. L. Amendola and S. Tsujikawa, Dark energy: theory and observations (Cambridge University Press,
2010).
7. C.W. Misner, K.S. Thorne, and J.A. Wheeler, Gravitation (Freeman and Company, 1973).
8. A. Friedmann, Zeit. f. Phys. 10 (1922) 377; ibid 21 (1924) 326; H.P. Robertson, Astrophys. J. 82
(1935) 284; ibid 83 (1936) 187, 257; A.G. Walker, Proc. Lond. Math. Soc. (2) 42 (1937) 90.
9. E. P. Hubble, Proc. Nat. Acad. Sci. 15 (1929) 168.
10. N. Jarosik et al Astrophys. J. Suppl. 192 (2011) 14.
11. S. Weinberg, Rev. Mod. Phys. 61 (1989) 1.
12. S. Carrol, W. Press, and E. Turner, Ann. Rev. Astron. Astrophys. 30 (1992) 499.
13. S. Weinberg, Gravitation and Cosmology: principles and applications of the general theory of
relativity (John Wiley & Sons, 1972); Cosmology (Oxford University Press, 2008).
14. C. M. Chambers and I. G. Moss, Phys. Rev. Lett. 73 617 (1994).
15. W. Rindler, Mon. Not. Roy. Astron. Soc. 116 (1956) 663.
16. E. J. Copeland, M. Sami, and S. Tsujikawa, Int. J. Mod. Phys. D15 (2006) 1753.
17. A. G. Riess et al., Astron. J. 116, (1998) 1009; Astron. J. 117 (1999) 707; S. Perlmutter et al.,
Astrophys. J. 517, (1999) 565.
18. R. Amanullah et al, Astrophys. J. 716 (2010) 712.
19. E.W. Kolb and M.S. Turner, The Early Universe Frontiers in Physics # 69 (Addison-Wesley, 1990).
20. R. H. Cyburt, B. D. Fields, K. A. Olive, E. Skillman, Astropart.Phys. 23 (2005) 313.
21. J.V. Narlikar, Introduction to Cosmology (Cambridge University press, Third Ed., 2002).
22. G. Steigman, Ann. Rev. Nucl Part. Sci. 29 (1979) 313.
23. W. de Sitter, Proc. Kon. Ned. Akad. Wet. 19 (1917) 1217; ibid 20 (1917) 229; Mon. Not. R. Astron.
Soc. 76 (1916) 699; ibid 77 (1916) 155; ibid 78 (1917) 3.
24. G.W. Gibbons and S.W. Hawking, Phys. Rev. D 15 (1977) 2738.
25. A.D. Linde, Particle Physics and Inationary Cosmology, (Harwood Ac., 1990).
26. A. Albrecht, P.J. Steinhardt, M.S. Turner, and F. Wilczek, Phys. Rev. Lett. 48 (1982) 1437; A.D.
Dolgov and A.D. Linde, Phys. Lett. B 116 (1982) 329; L.F. Abbott, E. Farhi, and M.B. Wise, Phys.
Lett. B 117 (1982) 29.
27. P.J. Steinhardt and M.S. Turner, Phys. Rev. D 29 (1984) 2162.
28. M.S. Turner, Phys. Rev. D 28 (1983) 1243.
29. A. R. Liddle and L. A. Urena-Lopez, Phys. Rev. Lett. 97 (2006) 161301.
30. J. H. Traschen and R.H. Brandenberger, Phys. Rev. D 42 (1990) 2491.
31. L. Kofman, A. Linde, and A.A. Starobinsky, Phys. Rev. Lett. 73 (1994) 3195; L. Kofman, A. Linde,
and A.A. Starobinsky, Phys. Rev. Lett. 76 (1996) 1011.
32. G. N. Felder and L. Kofman, Phys. Rev. D 63 (2001) 103503.
33. A.D. Dolgov, Phys. Reports 222 (1992) 309; A. G. Cohen, D.B. Kaplan, and A.E. Nelson, Ann. Rev.
Nucl. Part. Sci. 43 (1993) 27; J. L. Cervantes-Cota and H. Dehnen, Nucl. Phys. B 442 (1995) 391;
M. Trodden, Rev. Mod. Phys. 71 (1999) 1463; F.L. Bezrukov and M. Shaposhnikov, Phys.Lett. B
659 (2008) 703.
34. S. Davidson, E. Nardi, and Y. Nir, Phys. Rept. 466 105 (2008).
35. B. A. Reid et al, Mon. Not. R. Astron. Soc. 404 (2010) 60.
36. J. Carlson, M. White, and N. Padmanabhan, Phys. Rev. D 80 (2009) 043531.
37. U. Seljak and M. Zaldarriaga. Phys. Rev. Lett. 78 (1997) 2054.
38. N. Turok, U.-L. Pen, and U. Seljak, Phys. Rev. D 58 (1998) 023506.
39. J. Kim and P. Naselsky, JCAP 7 (2009) 41.
40. V.F. Mukhanov, H.A. Feldman, and R.H. Brandenberger, Phys. Reports 215 (1992) 203.
41. C.-P. Ma and E. Bertschinger, Astrophys. J. 429 (1994) 22.
42. C.-P. Ma and E. Bertschinger, Astrophys. J. 455 (1995) 7.
43. D. H. Lyth and A. R. Liddle, The primordial density perturbation: cosmology, ination and the origin
of structure, (Cambridge University Press, 2009).
44. P. de Bernardis et al., Nature (London) 404 (2000) 955.
45. S. Hanany et al, Astrophys. J. 545 (2000) L5.
46. J. R. Bond, G. Efstathiou, and M. Tegmark, MNRAS 291 (1997) L33; M. Zaldarriaga, D. Spergel,
and U. Seljak, ApJ 488 (1997) 1.
47. W. Hu and N. Sugiyama, Astrophys. J. 471 (1996) 542.
48. D. J. Eisenstein and W. Hu, Astrophys. J. 496 (1998) 605.
49. E. Komatsu et al Astrophys. J. Suppl. 180 (2009) 330.
50. H.-J. Seo and D. J. Eisenstein, Astrophys. J. 598 (2003) 720.
51. D. Schlegel et al, The BigBOSS Experiment, arXiv: 1106.1706.
52. D. J. Eisenstein et al, Astrophys. J. 633 (2005) 560.
53. A. Carnero et al, arXiv: 1104.5426.
54. R. Keisler et al, arXiv: 1105.3182.
55. R. Hlozek et al, arXiv: 1105.4887.