8 Dynamics of Expansion

In this chapter we discuss the dynamics of expansion, i.e., the evolution of the scale factor a(t). The ground
works for this discussion has been done partly in the previous chapters. The important ingredients to move
forward is to calculate the components of affine connection µ⌫ , Ricci tensor Rµ⌫ , energy momentum tensor
Tµ⌫ , and the Ricci scalar R. For the metric of the universe

dr2
2 2
ds = (dt) + a(t) 2
+ r2 (d✓2 + sin2 ✓d 2 ) .
1 Kr2
all the nonzero affine connections are
aȧ
0
11 = , 0
22 = aȧr2 , 0
33 = aȧr2 sin2 ✓ ,
1 Kr2
1 1 2 2 3 3 ȧ
01 = 10 = 02 = 20 = 03 = 30 = ,
a
1
22 = r(1 Kr2 ) , = 1
33 r(1 Kr ) sin2 ✓ ,
2
(152)
2 2 3 3 1
12 = 21 = 13 = 31 = ,
r
2 3 3
33 = sin ✓ cos ✓ , 23 = 32 = cot ✓ .
All the non-zero components of the Ricci tensor are
ä
R00 = 3 ,
a
aä + 2ȧ2 + 2K
R11 = , (153)
1 Kr2
R22 = r2 (aä + 2ȧ2 + 2K) ,
R33 = r2 (aä + 2ȧ2 + 2K) sin2 ✓ ,
and the Ricci scalar is ✓ ◆
6 2
R= aä + ȧ + K . (154)
a2
The calculations are given as an assignment.

8.1 Friedmann Equation

The components are now substituted in the Einstein’s field equation (124)
1
Rµ⌫ Rgµ⌫ = 8⇡GTµ⌫ ⇤gµ⌫ .
2
and derive the equations for the time evolution of a(t) for di↵erent µ⌫ combinations. If µ 6= ⌫ then the
RHS of the Einstein’s equation vanishes and it is an empty universe devoid of mass and energy. In general,
we are not interested in an empty universe, so we discuss only the µ = ⌫ cases. There are only two distinct
equations, µ⌫ = 00, and µ⌫ = ii where i = 1, 2, 3. Due to isotropy, all the nine equations for µ⌫ = ii are
the same.
For µ⌫ = 00 we get from the Einstein’s field equation
1
R00 Rg00 = 8⇡GTµ⌫ ⇤gµ⌫ ,
2
which gives, after substituting the 00 components of the metric and energy-momentum tensor, with g00 =
1, T00 = ⇢
✓ ◆
ä 1 6 2
3 + aä + ȧ + K g00 = 8⇡G⇢ + ⇤ ,
a 2 a2
ȧ2 K 8⇡G ⇤
) + 2 = ⇢+ . (155)
a2 a 3 3

53
As all the nine µ⌫ = ii gives the same equation, we take µ⌫ = 11. With g11 = 1/(1 Kr2 ), T11 =
P/(1 Kr2 ) we get
1
R11 Rg11 = 8⇡GT11 ⇤g11 ,
2 ✓ ◆
aä + 2ȧ2 + 2K 1 6 2 a2 a2 P ⇤a2
) 2
. 2
aä + ȧ + K 2
= 8⇡G ,
1 Kr 2 a 1 Kr 1 Kr2 1 Kr2
) 2aä ȧ2 K = 8⇡Ga2 ⇤a2 ,

Finally, we replace ȧ2 using eq.(155) to get

ä 4⇡G ⇤
= (⇢ + 3P ) + . (156)
a 3 3

The equations (155) and (156) are Freidmann equations in cosmology describing the expansion of the
universe. The second equation, often called the acceleration equation tells us that the cosmological constant
⇤ accelerates the universe.
Previously, we have associated a energy density ⇢⇤ = ⇤/8⇡G. Absorbing the ⇢⇤ into ⇢ we re-write the
Friedemann equations as
✓ ◆2
ȧ 8⇡G K
H2 = = ⇢ ,
a 3 a2 (157)
ä 4⇡G
= (⇢ + 3P ) .
a 3
Here, ⇢ in general indicate the energy density of matter, radiation, and cosmological constant. The Fried-
mann equations, paired with the fluid/continuity equation (144), are the fundamental equations of cosmol-
ogy.

8.2 Critical density and K

The first Friedmann equation (157) can be rearranged to write
✓ ◆2
ȧ 8⇡G K
H2 = = ⇢ .
a 3 a2

Rearranging we get
⇢ K ⇢ K
1= = .
3H 2 /8⇡G a2 H 2 ⇢crit a2 H 2
This equation tells that for any value of H, i.e., at any epoch of the cosmic evolution, there is a critical
density
3H 2
⇢crit = , (158)
8⇡G
of the universe such that if the energy density of the universe ⇢ is

• ⇢ = ⇢crit then K = 0, the universe is flat

• ⇢ > ⇢crit then K > 0, the universe is closed
• ⇢ < ⇢crit then K < 0, the universe is open
At the present epoch t0 , the Hubble constant is H0 = 72Km/s/M pc and the corresponding critical density
is
3H02
⇢crit,0 = = 1.9h2 ⇥ 10 29 g cm 3 , (159)
8⇡G

54
where the number to the left is obtained by substituting the present Hubble parameter H0 = 100hkm/s/Mpc
where h = 0.72.
There is something strange associated with the critical density of the universe. The energy density ⇢
of mass and radiation scales as 1/a3 and 1/a4 , respectively, while the energy density associated with ⇤
remains constant. So, if we extrapolate the Friedmann equation (157) to early times, then ⇢ would be large
and dominated by radiation. One can then neglect the K/a2 term which is negligible compared to 1/a4
or 1/a3 terms. Hence
✓ ◆2
ȧ 8⇡G ⇢
H2 = = ⇢ ) 1= . (160)
a 3 ⇢crit,0
It means that the density of the universe was equal to its critical density at very early times. Experimental
observations show that even today the density of the universe ⇢0 is fairly close to the critical density ⇢0,crit
implying a flat K = 0 universe. This is a puzzle because one normally expects that billion years of evolution
would make ⇢0 di↵erent from ⇢crit,0 . It referred to as the flatness problem.

8.3 Deceleration parameter

The second Hubble equation (157) contains ä. It describes how the expansion of the universe is changing
with time. If the right hand side is positive, i.e., ⇢ + 3P < 0 then ä >. There is a parameter called the
deceleration parameter
äa2
q= , (161)
ȧ2
which measures the state of acceleration. Analysis of type Ia supernovea observations show that at the
present universe q0 < 0, indicating and accelerated expansion.

8.4 Model Solutions

In this section we discuss a few model solutions of Friedmann equations where the solutions are exact.

8.4.1 Einstein Static Universe

A static universe implies that ȧ = 0 and ä = 0. From the Friedmann equation (157) we can write
8⇡G K 8⇡G K
0= ⇢ , =) ⇢= 2.
3 a2 3 a
As ⇢ is positive K must be equal to +1. The fluid equation gives

⇢ + 3P = 0 ,

which implies the pressure must be negative. So we can consider a new kind of entity such that it has
energy density. We see that the cosmological constant fit the bill. It has energy density ⇢⇤ and pressure
P = ⇢⇤ . The fluid equation can be re-written as

⇢m + ⇢⇤ 3⇢⇤ = 0 , ⇢m = 2⇢⇤ .

Hence the scale factor is

1 1
a= p =p .
4⇡G⇢m ⇤

8.4.2 de Sitter Universe

It is a cosmological constant dominated universe, i.e., ⇢ = 0. In this case it is convenient to solve the
acceleration equation (156)
✓r ◆ ✓ r ◆
⇤ ⇤ ⇤
ä = a , a(t) = C1 exp t + C2 exp t .
3 3 3

55
where C1,2 are constants of integrations. The ȧ2 and a2 are
 ✓ r ◆ ✓ r ◆
2 ⇤ 2 ⇤ 2 ⇤
ȧ = C1 exp 2 t + C2 exp 2 t 2C1 C2 ,
3 3 3
✓ r ◆ ✓ r ◆
⇤ ⇤
a2 = C12 exp 2 t + C22 exp 2 t + 2C1 C2 ,
3 3

Combining these two equations give

⇤ 2
ȧ2 = (a 4C1 C2 ) .
3
The Friedmann equation (155)
ȧ2 ⇤ K
2
= ,
a 3 3
and the expressions of ȧ2 and ȧ2 can be used to obtain the constraints
4⇤ 3K
C1 C2 = K , =) C2 = .
3 4⇤C1
So we have the following three solutions for K = ±1, 0
r ✓r ◆
3 3
a(t) = sinh t , for K = 1
⇤ ⇤
✓r ◆
⇤
a(t) = C1 exp t , for K = 0 (162)
3
r ✓r ◆
3 3
a(t) = cosh t , for K = +1
⇤ ⇤

To write the solutions the C1 has been appropriately chosen to obtain these forms.
In the de Sitter universe, there is no big-bang for K = 1. The universe has a bounce. For K = 0 the
universe has big bang at t = 1. For K = 1 the scale factor can be negative – so the universe has a big
bang.

8.4.3 Matter Dominated Universe

If the universe is matter dominated, then ⇤ = 0, and the energy density ⇢ / 1/a3 . Denoting the density at
the present time t0 by ⇢0 ⌘ ⇢(t0 ), we can write
✓ ◆3
a0
⇢ = ⇢0 ,
a

where a0 is the scale factor at the present time t0 a0 = a(t0 ). The first Freidmann equation (157) as
✓ ◆3
2 8⇡G a0
ȧ = ⇢0 a2 K ,
3 a
2A 4⇡G
= K, A= ⇢0 a30 (163)
a 3
To solve it, it is convenient to introduce a new variable ⌘, called the conformal time
d⌘ 1
= . (164)
dt a
This definition of the conformal time does not fix the constant of the integration. Unless we are studying
physics of inflation, ⌘ = 0 at the time of the big bang t = 0. The conformal time is not the age of the

56
universe. If the conformal time since the big bang is multiplied by the speed of light, it gives the total
comoving distance that is traveled by a light ray. In terms of conformal time, the equation (163) becomes
✓ ◆2
da
= 2Aa Ka2 .
d⌘
This equation can be integrated from the time of the big bang a = 0 to any time
Z a
da
⌘= p . (165)
0 2Aa Ka2
Lets consider solutions for three di↵erent scenarios
For K = 0: In this case the solution is
4⇡G 2
a= ⌘ ,
3
and Z Z ⌘
4⇡G 2 4⇡G 3
t = ad⌘ = ⌘ d⌘ = ⌘ .
0 3 9
These two parametric solutions can be combined to give
✓ ◆1/3
4⇡G
a= (3t)2/3 . (166)
3
The matter dominated (⇤ = 0) flat universe is called the Einstein-de Sitter universe.
For K = 1 the integration eq.(165) is
Z a
da 8⇡G 8⇡G
⌘= p , a= cosh ⌘ 1 , t = sinh ⌘ ⌘ . (167)
0 2Aa + a 2 3 3
For K = 1 the integration eq.(165) is
Z a
da 8⇡G 8⇡G
⌘= p , a= 1 cos ⌘ , t = ⌘ sin ⌘ . (168)
0 2Aa a 2 3 3
The variation of scale factor with cosmological time is shown in figure 17. As can be seen from the figure,
for K = 1 the universe expands to a mximum size and then collapses to a = 0. In the K = 1 open
universe, the universe expands for ever, where as K = 0 is an intermediate scenario between the two.

8.4.4 Radiation Dominated Universe

For a radiation dominated universe, the energy density varies as
✓ ◆4
a0
⇢ = ⇢0 .
a
Substituting in the first Friedmann equation we get
✓ ◆4
2 8⇡G a0
ȧ = ⇢0 a2 K ,
3 a
8⇡G⇢0 a40 1
= K (169)
3 a4
Imposing the initial condition a(t = 0) = 0 we get the following solutions
✓ r ◆1/2
8⇡G⇢0 a40
a= 2 t t2 , K = 1,
3
✓ r ◆1/2
8⇡G⇢0 a40
a= 2 t , K = 0,
3
✓ r ◆1/2
8⇡G⇢0 a40 2
a= 2 t+t , K = 1,
3

57
 2π A 2 2π A 3
k=0 a= 2
η , t= 2
η , (4.27)
3MPl 9MPl
4π A 4π A
k = −1 a= 2
(cosh η − 1) , t = 2
(sinh η − η) . (4.28)
3MPl 3MPl

Figure 4.1 shows the scale factor a(t) as a function of the cosmological time t for the
three scenarios. A closed universe expands up to a critical point and then recollapses.
An open universe expands forever. A flat universe is the boundary case between

k = -1
0.8 k=0

0.6

a(t)
0.4

k=1
0.2

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
t

Fig. 4.1 Scale factor a as a function of the cosmological time t for the three types of matter
Figure 16: Scale factor as a function
dominated of universe
universes: closed cosmological time for
(k = 1), flat universe (k = matter
0), and opendominated universe: closed universe
universe (k = −1).
2
K = 1, flat universe KHere and open universe K = 1.= 1
and a are presented in units in which 8π A/MPl
=t 0,

Just like in the case of radiation dominated universe, for K = 1 the scale factors reaches and maximum
and then collapses to a = 0. This is closed universe. In the open universe K = 1, the universe expands
for ever. The flat universe is an intermediate scenario. These are shown in figure
4.3 Cosmological Models 61

k = -1
0.8 k=0

0.6
a(t)

0.4

k=1
0.2

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
t

Fig. 4.2 Scale factor a as a function of the cosmological time t for the three types of radiation
Figure 17: Scale factor as a universes:
dominated function ofuniverse
closed cosmological time in
(k = 1), flat universe (k =radiation dominated
0), and open universe (k = −1).universe: closed universe
2
= 0, and open universe K = 1. = 1
The scale factor is expressed in units in which 8π B/MPl
K = 1, flat universe K
4.3.4 Vacuum Dominated Universe

In a vacuum dominated universe, there is no matter, so ρ = P = 0, but we allow for

a non-vanishing cosmological constant. As we will show in Chap. 11, in quantum
field theory the vacuum is not empty and in General Relativity it should act as an
effective cosmological constant or, more precisely, the vacuum energy is equivalent,
up to a constant factor, to a cosmological constant.
In the case of Λ > 0, the Friedmann equations provide the following solutions
! "! #
3 Λ
k=1 a= cosh t , (4.34)
Λ 3
"! #
Λ
k = 0 a = a(0) exp t , (4.35)
3
! "! #
3 Λ
k = −1 a = sinh t . (4.36)
Λ 3

If Λ < 0, the solution is

! "! #
3 Λ
k = −1 a= − cos − t , (4.37)
Λ 3

58 when Λ = 0, we recover the flat

and there is no solution for k = 0 and 1. Lastly,
spacetime with k = 0 and a constant.