Dynamics of the Universe
1
Contents
1 A Brief Summary of Observational Facts 3
2 The Friedmann-Robertson-Walker Universe 3
3 Critical Density 7
4 Dark Matter and Dark Energy 8
5 Early Universe and Critical Universe 9
6 Redshift 13
7 Inflationary Universe 15
2
1 A Brief Summary of Observational Facts
1. 1 pc(parsec)=3.26 ly(light year)
2. The universe is homogeneous and isotropic in a large scale
3. Hubble’s law: v = H0 d, H0 := h0 × (100 km/s)/Mpc,
h0 = 0.72 ± 0.05.
4. H0−1 = 0.978 × 1010 year/h0 ' 13.6 Gy; 2 −1
H
3 0
' 9.07 Gy.
5. With gravitational pull, the universe should be decelerat-
ing. Instead, it is accelerating.
2 The Friedmann-Robertson-Walker Universe
The homogeneous, isotropic, and expanding universe produced
by the big bang is sometimes called the Friedmann-Robertson-
Walker (FRW) universe, in honour of the people who first inves-
tigated this situation theoretically. This classical big bang theory
describes how the expansion of the universe is slowed down by
the gravitational pull from, and the pressure exerted by, the rest
of the universe.
The shaded region in the Figure is a sphere of radius R(t) =
a(t)r. It represents a part of our homogeneous and isotropic uni-
verse. r is the coordinate of a point (say a galaxy) in the universe
if it were not expanding. It is called the co-moving coordinate.
The expansion of the universe, assumed to be homogeneous,
is independent of r. The expansion is described by the scale
factor a(t). The fractional expansion rate (dR(t)/dt)/R(t) =
(da(t)/dt)/a(t) ≡ H(t) is the instantaneous Hubble constant at
time t. If we take t = t0 to be the present time, then H0 ≡ H(t0 )
3
R(t) m
M
Figure 1:
is the presently observed Hubble constant. We shall take R to
be small so that dR/dt c, thus ensuring the non-relativistic
dynamics used below to be valid. It turns out that luckily, the
result derived below using Newtonian mechanics is identical to
that obtained in general relativity.
Imagine a galaxy of mass m sitting at the surface of the small
sphere. It forms part of the surface so it moves out with the
surface when the universe expands. Its total energy is conserved
because the total energy of the universe is, and the universe is
made up of the sum of these galaxies. Let us call this conserved
constant − 12 mr 2 k, then
2
1 dR(t) GMm 1
m − = − mr 2 k, (1)
2 dt R(t) 2
where G is the Newtonian gravitational constant, and M is the
4
mass of the shaded sphere. In arriving at this expression, we
have used the well-known fact that the mass of the universe
outside of the shaded sphere does not contribute to the gravita-
tional energy of that particular galaxy.
If ρ is the energy density of the universe, so ρ/c2 is the equiv-
alent mass density, then M = 4πρR3 /3c2 , so eq. (1) can be
written as
!2
da 8π
− 2
Gρa2 = −k. (2)
dt 3c
For k > 0, = 0, < 0, potential energy is respectively larger than,
equal to, and smaller than the kinetic energy. The expansion
of the universe will eventually stop and reverse itself in the first
case, but will expand forever in the last two cases. In the sec-
ond case, the asymptotic expansion velocity tends to zero, but
it remains finite in the third case. The universe in these cases is
respectively called closed, critical, and open. A critical universe
is spatially flat, but the other two types have a spatical curva-
ture. The present evidence is that the universe is critical, so it
is spatially flat.
The expansion of the shaded sphere is resisted by a pressure
P exerted on its surface from the rest of the universe. As a
result, the shaded region has to spend energy to overcome this
pressure. The work done in expanding the shaded sphere from
radius R to R + dR is P (4πR2 )dR. The energy needed to do
the work, or the resulting decrease of energy inside the shaded
region, is equal to the work done. Hence
!
4π 3
−d R ρ = P (4πR2 )dR,
3
5
and therefore
d 3
ρR = −3P R2 ,
dR
or equivalently,
d 3
ρa = −3P a2 . (3)
da
I shall refer to (2) as the energy equation (officially it is called the
Friedman equation), and (3) as the work equation (often called
the continuity equation).
Taking (d/da) of (2), use (3), and the fact that
d(da/dt) d2 a 1
= 2 ,
da dt da/dt
we obtain the force equation
d2 a 4π
2
= − 2 G(ρ + 3P )a. (4)
dt 3c
To see the meaning of this equation, multiply both sides by mr.
Then the left-hand-side is the acceleration of the galaxy on the
surface, so the right-hand-side should be the force exerted on
that galaxy. If we put P = 0 and write R = ar = R3 /R2 , we
recognize the right-hand-side to be nothing but the gravitational
force of the shaded universe acting on the galaxy. What is inter-
esting is that instead of ρ, in general we have ρ +3P , so pressure
appears to contribute to the gravitational force as well. What
really happens is that pressure, like gravitational pull from the
sphere, makes it harder for the galaxy to fly outward, so it acts
as if it were an additional attractive force.
If the pressure is negligible, then the gravitational pull from
the shaded universe gives rise to a deceleration of the amount
6
shown in the equation to slow down the velocity of the galaxy
and hence the expansion of the universe. On the other hand, if
for some reason P is negative and ρ + 3P < 0, the gaxlaxy will
accelerate because it is being sucked out! This bizzare situation
is actually present in our universe. More of that later.
3 Critical Density
To simplify writing let da/dt be denoted by ȧ. It follows from
(2) for k = 0 that
!2
ȧ 8πGρ
:= H 2 = . (5)
a 3c2
This equation gives the critical mass density to be
1 3H02
2
ρc = = 1.88h20 × 10−29 g/(cm)3 , (6)
c 8πG
where we have traded H0 with the parameter h0 defined by H0 =
(100 km/s/Mpc)h0 . At h0 = 0.72, this turns out to be 0.975 ×
10−26 kg/m3 , which is the density for about six protons per cubic
metre!
The actual density ρ0 of the universe at the present time,
expressed in units of ρc , is commonly denoted by Ω: Ω = ρ0 /ρc .
The universe is open, critical, or closed if Ω < 1, = 1, or > 1.
The value of ρc is extremely small. For example, the mass
density of the earth is 5.5 g/(cm)3 , of the moon is 3.35 g/(cm)2 ,
and even gaseous objects like the sun has a density of 1.4 g/(cm)3 ,
some 29 orders of magnitude larger than the critical density. So
surely Ω > 1 and the universe is closed?
This turns out not to be so because there are lots of empty
spaces in the universe. When averaged out, the density of the
7
universe turns out to be quite small. As a matter of fact, the Ω
value for luminous matter is only about 0.005 ± 0.002, though
there is dark matter and other forms of energy around to make
the total Ω larger.
4 Dark Matter and Dark Energy
Dead stars, planets, interstellar gas are objects that may not
be luminous. They are part of the dark matter in the universe.
The surprising thing is, that there are a lot more dark matter
around than those ordinary objects.
An indication of the presence of dark matter is shown on the
rotational curve. Stars rotate about the galactic centre in much
the same way as planets rotating about the sun. If a star of
mass m rotates about a galaxy of mass M at a distance r and
with a speed v, then the balance between the gravitational force
and the centrifugal force gives rise to the equation
GMm mv 2
= . (7)
r2 r
√
Thus the speed v should drop off like 1/ r for stars lying outside
of the galaxy. However, what is observed is that the speed v
remains roughly constant for quite a distance outside of the
visible edge of the galaxy. According to (7), this is possible only
when M ∝ r, meaning that dark matter is present beyond the
visible edge of the galaxy.
We will discuss later ways to measure the amount of dark
matter in the universe. The total amount of ordinary (‘hadronic’)
matter can be estimated from primordial nucleosynthesis (or
big bang nucleosynthesis) (to be discussed later) to have an Ω
value of .045 ± .002, some ten times that of luminous matter.
8
There are also lots of other dark matter around which we have
not been able to detect directly because of its very weak in-
teraction with ordinary matter. It amounts to an Ω value of
0.27 ± 0.01. In addition, lately it is discovered that there is also
dark energy around, giving rise to an additional Ω of 0.72 ±0.05.
Adding these together, one gets a total Ω for the universe to be
1.02 ± 0.02. So within observational error, our universe is con-
sistent with being a critical universe.
5 Early Universe and Critical Universe
In this section we look at the critical universe. The result is
applicable to the open or the closed early universe, because at
the beginning both terms on the left-hand-side of (2) are large,
so their difference −k is negligible.
When k = 0, or negligible, (2) becomes
v
u
da u 8πGρ
= at . (8)
dt 3c2
We can now solve for a(t) if the dependence of ρ on a is known.
This in turn can be obtained from (3) if the equation of state
determining the relation between P and ρ is given.
There are three kinds of equations of state we want to con-
sider. P = ρ/3, P = 0, and P = −ρ.
We are familiar with the ideal gas law relating P to ρ by P =
κ(T )ρ, where κ is a constant proportional to the temperature
T . We shall see later that at very high temperature, when the
particles in the universe is extremely relativistice, the constant
κ(T ) saturates to the value 1/3, so that the equation of state is
1
P = ρ. (9)
3
9
This situation happens right after the big bang when the uni-
verse is very hot, and the epoch is known as the radiation-
dominated era. According to (3),
d 3 ρa3 b
ρa = − ⇒ ρ= 4 (10)
da a a
for some constant b. Substituting this into (8), we get
v
u
da u 8πGb
t
a = . (11)
dt 3c2
Consequently a2 is linear in t, and assuming a(t) = 0 at t = 0,
we get
!
0
√ 32πGb 1/4
0
a(t) = b t, b = . (12)
3c2
Later on, when the universe expands and cools down, massive
particles will slow down to the point when the energy of the
universe is dominated by mc2 of the massive particles. This is
the matter-dominated era. We are presently at this epoch. In
the matter dominated era, there is very little pressure P because
the particles are hardly moving, so we may set P = 0 in (3) and
conclude that
b00
ρ = 3. (13)
a
This is the familiar situation where the total mass (4π/3c2 )ρR3
of the shaded sphere, and hence of the universe, is a constant in
time.
According to (8), the time-development of this era is given
by
v
u
√ da u 8πGb00 2 000
a =t ≡ b , (14)
dt 3c2 3
10
whose solution is
v
u 00
3ut 8πGb
a3/2(t) = a3/2(tm ) + b000 (t − tm ), b000 = , (15)
2 3c2
where tm is the time when the matter-dominated universe takes
over from the radiation-dominated universe. This time (56 kyr)
is very short compared to the age of the universe at the present,
which is of the order of 1010 years. Hence for most purposes it
can be set equal to zero. Using that approximation, the age of
the universe t0 at the present is related to the Hubble constant
H0 by
ȧ(t0 ) 2 2
H0 = = , ⇒ t0 = H0−1 . (16)
a(t0) 3t0 3
If the Hubble expansion rate H = ȧ/a were constant in time,
then H0−1 would be the age of the universe. The extra factor 2/3
in (16) comes about because the expansion rate is predicted to
decrease with time, as can be seen from (15). This deceleration
can also be seen directly from the force equation (4), since ρ+3P
are positive both in the radiation and the matter eras.
Taking a value of h0 = 0.72 in (16), we get the age of the
universe to be t0 = 0.9 × 1010 years. This is a bit worrisome
because there are stars in globular clusters with an age 11 ∼
12 × 109 years, seemingly older than the age of the universe!
This dilemma was solved in the last few years by the discov-
ery that the universe actually accelerates, not decelerates. In
that case the expansion rate was slower in the recent past than
present, hence the age of the universe is considerably longer than
(2/3)H0−1 .
According to (4), acceleration occurs if ρ + 3P < 0. This
happens if the present universe contains two components, ρ =
11
ρ1 +ρ2, P = P1 +P2, with ρ1 being the matter density with P1 =
0, and ρ2 being the dark energy density with P2 = −ρ2 . Thus
ρ + 3P = ρ1 − 2ρ2 is negative if ρ1 < 2ρ2 . From the discussion at
the end of Sec. 4, ρ1 = (0.27 ± 0.01)ρc , and ρ2 = (0.78 ± 0.02)ρc ,
so there is no doubt that ρ +3P < 0. In fact, acceleration would
occur even if P2 = wρ2, as long as ρ1 < −(3w + 1)ρ2 .
With w = −1, 8πGρ2 := Λ is known as the cosmological
constant. It was introduced by Einstein who regretted later as
being his biggest mistake. May be it is no mistake at all! It
was introduced by him to keep the universe static, at a time be-
fore Hubble’s discovery of the expanding universe. An amount
ρ + 3P = ρ1 − 2ρ2 = 0 would suffice to cancel the force acting
on every galaxy and make the universe static, hence his intro-
duction of ρ2 .
According to (3), the equation of state P = −ρ for the dark
energy is equivalent to the assertion that ρ is independent of a(t)
and hence independent of time. For whatever physical reason
when ρ is constant (or changes very slowly with time), then a
dark-energy like density can be considered to be present. A
(false) vacuum (which is translational invariant) would give rise
to such an equation of state.
With the equation of state P = −ρ, ρ is constant, so the
solution of (8) is
q
a(t) = a(0) exp(b̄t), b̄ = 8πGρ/3c2 . (17)
A universe that expands exponentially like this is called a de
Sitter universe. We will have more to say about it in the section
on Inflation.
Physically, we can think of the growth of total amount of
matter inside the sphere of radius R(t) = a(t)r as due to the
negative pressure sucking in matter from outside.
12
The present density ρ is not constant, as noted above, but
consists of two components ρ1 + ρ2 . Since P1 = 0, ρ1 ∝ 1/a3 .
Since P2 = −ρ2 , ρ2 is independent of a. Presently ρ2 ∼ 2ρ1 ,
so ρ + 3P < 0 and the universe is accelerating. Since ρ1 was
bigger in the past and smaller in the future, while ρ2 remains
constant, in the distant past ρ1 − 2ρ2 > 0 and the universe
was decelerating, like one expected for a FRW universe. In
the distant future, ρ → ρ2 and it becomes a de-Sitter universe
with an exponential growth. But note that this is so only when
w = −1 for the second component, and at the present time we
really do not know that this is true. We do not even know that
w does not depend on time.
6 Redshift
In a static universe, the Doppler redshift factor z is defined by
z + 1 = λ0 /λE = νE /ν0 , where λ and ν refer to wave length
and frequency, and the subscripts E and 0 refer to the emitting
galaxy and the receiver on earth. From z one can compute
the velocity v of a receding galaxy. In an expanding universe,
it becomes a bit complicated to see what that means because
we are moving and the galaxy is moving. Moreover, the light
reaching us was emitted quite some time ago. What relative
velocity we are talking about? Are these the velocities at the
emission time tE , or the receiving time t0 ? It turns out that the
correct answer is
a(t0)
z +1 = . (18)
a(tE )
To prove that, let us first examine how light travels in an
expanding universe. Refer to the Figure. Let the distance of
13
a galaxy from the centre of the shaded region be R(t) = a(t)r.
Imagine a light signal emitted at rE at time tE reaching us at
the centre of the sphere at time t0. At any time t in between,
the signal will travel a distance c(dt) = a(t)dr in a time interval
dt. In other words, in co-moving coordinates, it will travel an
amount dr = c(dt)/a(t). From time tE to time t0 , the total
amount it travels in the co-moving frame is given by
Z t
0
rE = c(dt)/a(t). (19)
tE
This is the distance between the observer and the emitting
galaxy in the co-moving coordinate. It does not change with
time!
Let νE and ν0 be the frequencies at emission and reception.
The period it takes for successive emissions is 1/νE at the galaxy,
and the time interval between their arrival on earth is 1/ν0 .
Light emitted at time tE + 1/νE will arrive on earth at time
t0 + 1/ν0 , having travelled exactly the same scaled distance rE .
Thus
Z t Z t +ν −1
0 0 0
c(dt)/a(t) = −1
c(dt)/a(t). (20)
tE tE +νE
Subtracting these two integrals, we can write the equation as
Z t +ν −1 Z t +ν −1
E E 0 0
c(dt)/a(t) − c(dt)/a(t) = 0. (21)
tE t0
The times ν0−1 and νE−1 are so short that the scale factor can be
assumed to be constant during those time intervals. Hence
νE−1 ν0−1
= ,
a(tE ) a(t0 )
or equivalently,
νE a(t0)
= = 1 + z, (22)
ν0 a(tE )
14
as claimed in (17).
In the special case when the galaxy is not too far away from
us, when t0 − tE is fairly small, we may expand a(tE ) as a(t0) +
ȧ(t0)(tE − t0 ). Hence (17) or (21) gives
a(t0 ) 1 ȧ(t0 ) v
z' −1' c(t0 − tE ) = , (23)
a(t0) + (tE − t0 )ȧ(t0 ) c a(t0 ) c
because d = c(t0 − tE ) is the distance to the galaxy, H0 =
ȧ(t0)/a(t0 ) is the Hubble constant, so v = H0 d is the velocity of
recession of the galaxy according to Hubble’s law.
Eq. (22) tells us that z is just the Doppler redshift factor if
v c. However, when the galaxy is far away and z is large,
then z should be interpreted to be the ratio of the two scale
factors a, as given in (17).
7 Inflationary Universe
The classical big bang theory discussed in the previous section
suffered from several drawbacks, the most severe of which is
probably the horizon problem. √This stems from √ the fact that
right after the big bang, a(t) ∝ t, so ȧ(t) ∝ 1/ t. This would
be larger than c for small t.
The fact that the universe expands faster than the speed of
light is by itself not a problem, since we can never make use
of this rate to propagate a signal which travels faster than the
speed of light. In other words, it is the space that expands fast
but a signal in an unexpanded space still cannot travel faster
than speed of light.
The problem occurs because we know by observation that
cosmic microwave background radiation is isotropic and homo-
geneous at least to 1 part in 103 (see details in later lectures),
15
and with this fast expansion rate in the beginning this isotropic
and homogeneous property simply does not make sense. To
understand why let us calculate how much distance D(t) light
travels after a time t in an expanding universe.
This calculation is very similar to the red shift calculation of
the last section. Refer to the Figure and eq. (19). If light is
emitted at time tE = 0 at the centre of the sphere, then at a
later time t0 = t it will have travelled a distance
Z t
dt0
D(t) = a(t)rE = a(t) c 0 . (24)
0 a(t )
√
In the radiation-dominated era, when a(t) = b0 t, we get D(t) =
2ct. In the matter-dominated era, when a(t) ' b00 t2/3, we get
D(t) = 3ct. For small t, light travels a distance in the expanding
universe smaller than the size of the universe, hence different
regions of the universe will not be causally connected. This
would be so even at the time of decoupling, when the microwave
background radiation emerged. This is inconsistent with the
observation that the radiation is isotropic up to one part in 103 .
This is the ‘horizon problem’.
There is also the ‘flatness problem’. Using (6), eq. (2) can be
written as
ρ k
1− = 1 − Ω = − 2. (25)
ρc ȧ
Before the dark energy component was known, Ω = Ωmatter ∼
0.35, giving a value .65 to the left-hand side. However, at the
early universe when ȧ is very very large, the left-hand-side is
practically zero. The question is why, at the early universe, the
density is so close to the critical density. Is it not more natural
to have it exactly the critical density then? If so, then k = 0
16
and Ω = 1 at all times, and the universe is flat. That is why the
‘unnatural’ closeness of the (previously known) denstiy to the
critcal density is called the ‘flatness problem’.
Both of these problems can be solved if k = 0 and P = −ρ.
In that case, a(t) = a(0) exp(b̄t) and D(t) = (c/b̄)(eb̄t − 1).
If c/d > a(0), then there is no problem for light propagation
to catch up with the expansion of the universe, and the hori-
zon problem will disappear. This inflationary universe scenario
which is supposed to occur even before the radiation-dominated
era of the FRW universe was proposed by Alan Guth to solve
this and other problems of the classical big bang theory. The
exponential growth of the size, the velocity, and the acceler-
ation of the inflationary (de-Sitter) universe also explains the
initial big-bang velocity of the universe. The false vacuum with
ρ =constant cannot last forever. When that decays towards
a smaller or zero ρ, the inflationary era stops, and the energy
released is channeled into heating the universe in the radiation-
dominated era.
17
