You are on page 1of 23

Cosmology Notes

Murtaza Jafry1
1
Department of Physics, University of Washington, 98195-1560, Seattle, USA

Winter 2020

Abstract

These are notes for Cosmology, ASTRO 425, at UW. These will be a composition of the notes
throughout the course as a whole. The notes will also have some extra sections further explaining
specific details if necessary. These notes will cover the Friedman-Walker (FRW) Metric and its
implication toward describing universes of different cosmologies. The notes will also go into detail
describing the General Relativity behind the cosmologies of different universes.
Contents
1 Critical Density 2
1.1 Cosmological Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Using Friedmann’s Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Use of Friedmann’s Equations for Epoch Times . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Solving a(t) for Cosmological Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Different Universes 5
2.1 Flat Universe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Comoving Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Proper Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Radiation-Dominated, Flat Universe (ω = 31 ) . . . . . . . . . . . . . . . . . . . . . . . 6
2.5 Matter Dominated Universe (ω = 0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.6 Cosmological Constant, Λ, Dominated Universe . . . . . . . . . . . . . . . . . . . . . . 6
2.7 Current Planck Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3 Making Measurements of Our Current Universe: Distance Ladder 7


3.1 Stellar Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Cosmological Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.3 Distance as a Cosmological Constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.4 Outer Milky Way Distances: Indirect . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.5 Extragalactic Distance Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

4 Lensing and Dark Matter 10


4.1 Milky Way . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.2 Measuring Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.3 Newtonian Gravitational Lensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.4 Gravitational Deflection From General Relativity . . . . . . . . . . . . . . . . . . . . . 11
4.5 Point-Mass Lens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
4.6 General Case for Lensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.7 Lensing Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.8 Time Delays and the Hubble Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.9 Mass Models for Lenses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5 Thermal History of Universe 15


5.1 Hot Dense Universe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.2 For Relativistic Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.3 Neutrino Decoupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5.4 Electron-Positron Annihilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5.5 Big Bang Nucleosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.6 Neutron Abundance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.7 Neutron Decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.8 Helium Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.9 Helium Abundance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.10 Sensitivity to Baryon Fraction η . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.11 Observational Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.12 Recombination and Decoupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.13 Discovery of CMB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.14 Shape of Cosmic Microwave Background . . . . . . . . . . . . . . . . . . . . . . . . . . 21

1
1 Critical Density
In a flat universe we know that from Firedman’s equations that
 2
3c2 ȧ 2
 
ȧ 8πG
= , c = (1.1)
a 3c2 8πG a

Here we see that c is the critical energy density when the universe is flat. If we now consider a time
dependent energy density, we can find the function as

3c2 2
c = H (t) (1.2)
8πG
Under this functional representation, we can classify our universe by the characteristics that if  > c ,
then k = +1 which means that we live in a closed universe. If  = c , k = 0 which means that our
universe is flat. Finally, if  < c , then k = −1, then we live in an open universe. We also note that
at t = t0 , our energy density takes the form

3c2 2
c,0 = H ≈ 8.3 × 10−10 JM −3 (1.3)
8πG 0

which means that we can write the function Ω(t) = (t)c (t)
. Here Ω(t) is the energy density parameter.
Here this unitless energy density lies in the regime 0.9 < Ω < 1.1 are the current measured bounds.

1.1 Cosmological Constant


An energy density of vacuum was originally posed by Einstein to create a static universe. This addition
would impact the Friemdan equations by
 2
ȧ 8πG kc2 Λ
= − + (1.4)
a 3c2 a2 R03 3

Here we have that  = −P and ω = −1. This would lead the dyanmics of the energy density will
become

˙ + 3 ( + P ) = 0 (1.5)
a
Where ˙ = 0 and  = constant. By convention we often treat Λ as a component of the energy density.
More precisely,
c2
Λ = Λ (1.6)
8πG
This leads Friedman’s equations to become
 2
2 ȧ 8πG 8πG kc2
H (t) = =  + Λ − (1.7)
a 3c2 3c2 a2 R02

1.2 Using Friedmann’s Equations


We know that the net density and pressure is the sum of all the constituent canonical density and
pressures, respectively. This means that
X X
P = Pi = ωi i (1.8)
i i

2
In each component of the pressure or density, the (matter, radiation, neutrino, ...) is conserved. Also,
each component of the respective quantities obey the Horid equation. This equation is defined as

i = i,0 a−3(1+ωi )
(1.9)
Ωi = Ωi,0 a−3(1+ωi )

8πG kc2 3c2 2


If we know that today, H(t) = 3c2
 − a2 R02
, at t = t0 and i,0 = 8πG H0 . This means that

H(t) 0 kc2
= − 2 2 2 (1.10)
H0 c,0 a0 H0 R0

kc2 0
Where H02 R02
= i,0 − 1 = 1 − Ω0 . With this simplification, we can then redefine our Hubble constant
as
8πG H2
H 2 (t) =
2
 + 20 (1 − Ω0 ) (1.11)
3c a
We can also divide by the Hubble parameter, today, to find that

H 2 (t) 1 − Ω0
= Ω(t) + (1.12)
H0 a2
If we want to write the Hubble constant as a sum, we will find that the equation becomes

H 2 (t) X 1 − Ω0
= Ωi,0 a−3(1+ωi ) +
H02 i
a2
H 2 (t)
= Ωr,0 a−4 + ΩM,0 a−3 + ΩΛ,0 + (1 − Ω0 )a−2 (1.13)
H02
H 2 (z)
= Ωr,0 (1 + z)4 + ΩM,0 (1 + z)3 + ΩΛ,0 + (1 − Ω0 )(1 + z)2
H02

From these equations, we can see that at t = t0 , a = 1 such that the Friedman’s equation becomes

1 = Ωr,0 + ΩM,0 + ΩΛ,0 + (1 − Ω0 ) =⇒ Ω0 = Ωr,0 + ΩM,0 + ΩΛ,0 (1.14)

This means that 1 − Ω0 = Ωk , where Ωk would be defined as the curvature density. For a flat universe,
this means that Ωk = 0.We also know that Ω0 = 1, where ȧa = H(t). This means that H(t) when
solved becomes ˆ a
da
H0 t = 1 (1.15)
0 (Ωr,0 a −2 −1
+ ΩM,0 a + ΩΛ,0 a2 + (1 − Ω0 )) 2
The question now becomes what are the different time periods for the early universe epochs.

1.3 Use of Friedmann’s Equations for Epoch Times


Dynamics are dominated by the component with largest energy density. We transition from a universe
that is radiation dominated to a matter dominated universe. One can actually think about this in
terms of the density parameter.
Ωr Ωr,0 (1 + z)4
= (1.16)
ΩM Ωm,0 (1 + z)3
The current measurement of Ωr,0 = 8.12 × 10−5 done from the CMB by COBE. From this, we can
derive the energy density as being  = aT 4 where a = 7.565 × 10−16 JM −3 K −4 with T = 2.73 K. We

3
also have that Ωm,0 = 0.315. Here we can see that the current universe is dominated by matter rather
than radiation due to the relative magnitudes of Ωm,0 and Ωr,0 . If we now set that

Ωr
= 1, z = 3879 (1.17)
Ωm
We can also look at the situation when, Λ dominates the dynamics of the universe. More specifically,
this will mean that
ΩΛ
ΩΛ,0 = 0.685, = 1 at z ∼ 0.3 (1.18)
ΩM
At late times, z < 0.3, Λ dominates dynamics of our universe. We can now solve for a(t) for our
cosmological model.

1.4 Solving a(t) for Cosmological Model


In an empty universe (Milne Universe), we have that

Ωr,0 = Ωm,0 = ΩΛ = 0 (1.19)

This leads us to assume that


 2
H 2 (t) 1 ȧ
= 2 = (1 − Ωk,0 )a−2 = a−2 (1.20)
H02 H0 a

If we now integrate this expression, this leads to the quantitiy,

ȧ = H0
ˆ a ˆ t
ȧ 1 (1.21)
da = H0 dt =⇒ a = H0 t =⇒ H(t) = =
0 0 a t

At time t = t0 , the Hubble constant equals H0 = t10 . Here the Hubble Time ≡ Age of Universe. Thus,
we calculate the time at which a photon is emitted by a galaxy at redshift z. This means that

a(t0 ) t0
1+z = =
a(te ) te
(1.22)
t0 H −1
te = = 0
1+z 1+z
For distances, in the comoving distance frame, we can define the radial distance as
ˆ t0 ˆ t0  
dt c dt c t0 c
r=c = =⇒ r = ln = ln(1 + z) (1.23)
te a(t) H0 te t H0 te H0

This means that the proper distance then becomes


ˆ r
c
dp (t) = a(t) dr = a(t)r = a(t) ln(1 + z) (1.24)
0 H0
c ln(1+z)
1 c
When t = t0 , dp (t0 ) = Hc0 ln(1 + z) and that dp (te ) = 1+z H0 ln(1 + z) = H0 1+z . We also know that
at small z, dp (t0 ) = Hcz0 , Hubble’s Law. This means that at high redshifts, the universe is smaller by
a factor of 1 + z.

4
2 Different Universes
2.1 Flat Universe
A flat universe is characterized as having no matter. For a radiation only universe, we can define the
function

= H02 a−2
a
ˆ a ˆ t
a da = H02 dt (2.1)
0 0
√ 1
a(t) = 2H0 t 2

We can generalize a single component universe with the equation

ȧ2 = H02 a−(1+3ωi ) (2.2)

Here, ωi : equation of state parameter. If ω = 0, it is a matter only universe. If ω = 13 , it is a radiation


only universe. We can now integrate this expression to get that
ˆ a ˆ t
(r+3ω)
a 2 da = H0 dt
0 0
2 3
a 2 H(t)ω = H0 t (2.3)
3(1 + ω)
  2   2
3 3(1+ω) t 3(1+ω) 2
a(t) = H0 (1 + ω)t = , t0 = H −1
2 t0 3(1 + ω) 0

This means that if ω > − 31 then t0 < H0−1 . If ω < − 13 , then t0 > H0−1 . This will then allow us to
calculate the comoving distance.

2.2 Comoving Distance


We will first define the function for the radial distance within the comoving frame.
ˆ t0 ˆ t0 2
 t0    1+3ω 
dt 2 3(1 + ω) 3H(t)ω
2
− 3H(t)ω 1+3ω 3(1 + ω) te 3(1+ω)
r=c t
= ct 3H(t)ω dt = ct0 t 3(1+ω) = ct0 1−
te a(t) te 1 + 3ω te 1 + 3ω t0
(2.4)
We can now reexpress this function in terms of redshift observed. This is critical as experiments
measure red shift measurements rather than time of emission. With this resubstituion, we will find
that  
c 2 − 1+3ω
r= 1 − (1 + z) 2 (2.5)
H0 1 + 3ω
Here 1 + z = a(t1e ) . We can now also transform to the proper distance frame. The expression for the
proper distance becomes

2.3 Proper Distance


The Proper distance is defined as being equal to the comoving distance times the scale factor a(t).
This means that the distance measured equals
 
c 2 − 1+3ω
dp (t0 ) = 1 − (1 + z) 2 (2.6)
H0 1 + 3ω

5
From this, we can now define the horizon distance as the separation distance between any causal
contact between matter within the universe. We can now define the horizon distance as when z → ∞
or when te → 0. This means that
c 2
dHorizon (t0 ) = (2.7)
H0 1 + 3ω

2.4 Radiation-Dominated, Flat Universe (ω = 13 )


The equation for the scale factor in this universe becomes
 1
t 2 1
a(t) = , t0 =
te 2H0
  (2.8)
c c z
dp (t0 ) = 1 − (1 + z)−1 =
H0 H0 1 + z
c c z
This means that dHorizon (t0 ) = H0 and dp (te ) = H0 (1+z)2

2.5 Matter Dominated Universe (ω = 0)


For this matter dominated universe, we can now define the scale factor as
 2
t 3 2
a(t) = , t0 = H0−1
t0 3
 
2c − 21
dp (t0 ) = 1 − (1 + z)
H0 (2.9)
 
2c − 12
dp (te ) = 1 − (1 + z)
H0 (1 + z)
2c
dHorizon (t0 ) =
H0
A matter dominated universe (flat) is called an Einstein-de Sitter universe. This was the kind of
universe that was primarily theorized as being our universe during the 1990s. This universe expands
more rapidly than radiation dominated universe, but less than an empty universe.

2.6 Cosmological Constant, Λ, Dominated Universe


Here we know that the Hubble constant will be defined as being equal to
 2

= H0
a
ˆ a ˆ t
da
= H0 dt
0 a 0 (2.10)
H0 (t−t0 )
a(t) = e , Exponential Growth
ˆ te  
H0 (t−t0 ) c H0 (t−t0 )
dp (t0 ) = c e dt = e −1
t0 H0

In this kind of universe, the phase shift can be expressed as 1+z = eH0 (t−t0 ) . This means that Hubble’s
law will state that dp (t0 ) = Hc0 z, d Horizon (t0 ) → ∞, d Horizon (te ) = Hc0 which is a finite quantity. For
each of these universes, we can use the equations from Section 1.2 to calcualte the age of the universe.

6
2.7 Current Planck Measurements
The current measurenments for the various densities in the universe are
ΩN,0 = 0.685 ± 0.017
ΩM,0 = 0.315 ± 0.017
(2.11)
Ωk = 0
Ωr,0 = 5 × 10−5

Cosmologists often use these parameters to assess the different types of cosmologies of our universe.
Types of universes which are of interest are Big Bounce, Loitering, and Big Crunch universes. To
summarize the scale factors as a function of time for different types of universes, we find that for a
1
radiation dominated universe a(t) ∝ t 2 , this means that t ≤ 5 × 104 Yrs. For a matter dominated
2
universe, a(t) ∝ t 3 with an age of t ≤ 9.8 billion years. Finally, for a Λ dominated universe, a(t) ∝ eH0 t .

3 Making Measurements of Our Current Universe: Distance Ladder


First we will look at age determination of our universe. We will first focus on local to distant mea-
surements. We will do this by making age measurements of solar systems. These techniques stem
from radioactive decay dating from the early 1900s. To conduct this analysis, we first assume the
relative abundance of a certain element in nature. One example of this would ratios of isotopes for
235
Uranium. If we assume the initial abundance appears to be U U 238
≈ 1.3. The current mreasurenment
U 235
of the relative ratio would be U 238 ≈ 0.0072. This relative ratio come from the R process. The half
lives of the isotopes of Uranium are U 235 ∼ 704 MyR and U 238 ∼ 4.46 GyR. From this analysis, we
can assess the approximate age of the universe. With the analysis of Uranium, one finds that the age
of the universe would be ≈ 6.3 GyR. This procedure can also be done with other elements within our
universe. More specifically, if this analysis is run on Th238 which is quite prevalently found within the
earth, one would fine that the age of the universe would be ∼ 4.5 GyR.
The assumptions in this procedure is that we assume that the system is closed. In other words,
the assumption is that the production of new Uranium or other radioactive elements is negligible.
Another assumption is that the material is homogenous when measuring the relative abundances. In
reality, these assumptions have to be taken into account when making measurements of our universe.

3.1 Stellar Evolution


If we assume that the stars (or groups of stars) form at the same time, then we can estimate the age
from the “main sequence turnoff.” The main sequence will plot, Temperature vs. Luminosity. The
temperature scale will span from about T ∈ [40, 000 K, 2, 500 K] and the luminosity will span from
L ∈ [10−4 , 105 ]. In these scales, stars with Luminosity > 1.1M0 would be massive stars which are hot
and very luminous. On the other hand, if the luminosity < 1.1M0 , then the stars of low mass and are
cool. This means that as stars start burning hydrogen, they enter the main sequence of hydrostatic
+ thermal equilibrium. In this situation, the pressure of the star balances the gravitation force, and
the energy density balances the luminosity. This means that L ∝ M α , where L: Luminosity and M :
Mass. From this analysis, one could then derive a nuclear timescale of the age of the star. More
precisely,
f M c2
τN uc = (3.1)
L
Here f : fraction of fuel available, : efficiency of converting matter → energy. This means that the
time for the main sequence would take τM S ∝ M −2.5 . As an example of some stars, a star with mass

7
10M0 will have a main sequence time of τM S ∼ 106 Yrs. A star with a mass of 0.1M0 would have a
mass of τM S ∼ 1011 Yrs. This means that if we measure this turn off, then we can measure the age
of the universe. The major caveat of this procedure is that we must know the distance as we need
to calculate the luminosity. The current age of the universe through this procedure predicts the age
begin around τAge ∼ 5 − 12 GyR.

3.2 Cosmological Measurements


In these kinds of measurements, we are measuring the age of high redshift galaxies. The most distant
galaxy currently known is z = 11.09. This was measured by the Hubble telescope. For our benchmark
cosmology,
τLook Time = 13.06 Gyr
ˆ z
dz (3.2)
τLook Back =
0 (1 + z)H(z)

In order to use assessment, we must assume that there has to be sufficient time for galaxy to form
given age of universe.

3.3 Distance as a Cosmological Constraint


Earlier, we met two different types of distances that could be used to describe the expanding Friedman
metric. For the luminosity distance
• Luminosity Distance is defined as being
dL = SK (r)(1 + z)
where Sk (r) = R sin( Rr ) for k = +1, Sk (r) = r for k = 0, and Sk (r) = R sinh( Rr )
• Angular Diameter Distance is defined as being
Sk (r)
dA =
1+z
and
dL
dA =
(1 + z)2
Even though one can use both of these distance metrics, no single distance method can measure
distances to all sources. We rely on distance ladders of different techniques. One example of this is
using Local Measurements for our solar system. One means of doing this is to use laser and radar
ranging.
c∆t
• Measuring time delay d = 2

• Accurate wetting solar system


One can also use Direct Stellar Measurements: parallax. In this case we can define a parallax angle
which equals p = d1 , where d: distance (parsecs) and p: parallax angle (arsec). With this definition of
distance, stars “move” relative to background stars. A parsec here is defined by a change in position
of larsecs, Lpc ≡ 3.26 Lyrs. In the case of Alpha Centauri, p = 0.76 arcsec and d = 1.3 pc. Accuracy
(distance) depends on our ability to measure positions (astrometry). Ground based telescope (optical)
image quality limited by atmosphere (∼ 0.7 arcsec). Space based telescopes (e.g Gaia) has made the
measurement of v ∼ 15 mag with ∆p ∼ 20 parsecs, and v ∼ 20 mag with ∆p ∼ 250 parsecs. Here,
20 µas ≡ 50 kpc, and we expect to measure distances accurately to ∼ 20 kpc for a billion stars.

8
3.4 Outer Milky Way Distances: Indirect
Cepheid variables vary in brightness with regular period (2-60 days). Pulsation due to variation in
size - outer envelope is low density and varies by 10-20% in size and temperature

L = 4πR2 σT 4 (3.3)

Here R: Radius, L: Luminosity, T: Temperature, if R+T vary then so does L. If we know that Cepheid
is composed and opaque, then we know that

• Temperature and pressure increase

• Outer part of envelope expand and become transparent.

• Photons stream out; then, the gas cools and contracts due to gravity.

From Henrietta Leavitt, a supercomputer centered in Cambridge Massachusetts,1 it was found that

Mv = (−2.43 ± 0.12)(log10 (P ) − 1) (3.4)

Here the Mv : the V=band absolute magnitude, and P : period measured in days. If we measure the
period, we infer the distance from the luminosity and observed brightness. This is known as secondary
distance indicator.

3.5 Extragalactic Distance Indicators


This is a tertiary distance indicator, where we set the distance by using Cepheids in nearby galaxies.

• For spiral galaxies, we use the Tully-Fisher relation in which

L ∝ Vrα , α:2−4

Here Vr is the rotation velocity, and we assume that brighter galaxies rotate more quickly. If we
measure Vr , we infer L and derive distance: accurate up to 10%.

• For Elliptical galaxies, we use the Faber-Jackson relation in which

L ∝ σv4

where σv is the velocity dispersion. Here brighter ellipticals have larger velocity dispersion with
accuracy of ∼ 20%.

• For Type IA supernova, there are no hydrogen lines, no He4 lines, strong silicon lines in spectrum.
It is visible for weeks after explosion and luminosity comparable to brightness of galaxy. It is
also thought to arise from white dwarfs through merging or material accelerating onto the white
dwarfs (causes a thermonuclear reaction). The brightness is a function of time (light curve)
correlates with intrinsic brightness. This will then allow us to infer the intrinsic Luminosity +
estimate distance.
1
Henrietta Leavitt was also a famous astronomer in Massachusetts whose work allowed for this indirect measurement
technique.

9
In each of the Type IA, Tully-Fisher, and Faber Jackson, we assume the sources are standard candles,
and we use the luminosity distance to measure distances. For Baryon Acoustic Oscillations which
are currently measured at the CMB. In the early universe, photons and baryons are tightly coupled.
Here this is the scattering of the photons and the electrons; this is known as Thompson scattering.
Fluctuations in density propagate at close to the speed of light. Once the universe cools, the protons
and the neutrons will combine for the first time. This means that there is no longer any scattering
with the photons; thus the photons then stream away from over-densities. We are then just left with
fluctuations in the baryons. This feature or maximum scale for the fluctuations (∼ 100 M pC) is
imprinted on clustering of galaxies and CMB. The scale is fixed in size (proper) and so if we measure
angular scale of fluctuations and how this changes with distance (angular diameter distance), we can
compare to cosmological models.

4 Lensing and Dark Matter


The dynamics of dark matter depend primarily on mass density. We measure light distribution. We
relate mass and light through mass-to-light ratio.

• M/L ratio: ∼ 2 × 10−5 M


L0 , measured in solar unites M0 : Mass of sun, L0 : luminosity of sun.
0

• Solar neighborhood ∼ 4 M
L0
0

Here we can measure luminosity in local solar neighborhood gives a mass density of ρ∗ = 5 ×
108 M0 M pC −1 and Ωk,0 = 0.004  1.

4.1 Milky Way


If we consider gas, stars, then the total mass density is ∼ 0.5% of critical density. If we now look at
clusters of galaxies, we know that they contain hot gas (106 K) with 10× the mass of stars in cluster.
This leads to Ωclusters = 3% of critical mass. One can also use cosmological scales to measure Big
Bang nucleosynthesis (BBN) to show that ΩBaryon ∼ 0.04 ± 0.01. Most of the mass is dark, and we
measure it through its dynamics. We would like now what techniques that are used to measure mass
within the universe.

4.2 Measuring Mass


In the 1970s, Vera Rubin (name sake for Vera C. Rubin) showed a relation between mass and rotation
of galaxies. For a point mass on a galaxy, we can write that

v2 GM (R)
= (4.1)
R R2
Here v: rotational velocity, R: Radius, M : Enclosed Mass, G: Gravitational Constant. If we assume
1
that M (R) = 34 πR3 ρ, then we expect that V (R) ∝ (R2 ρ) 2 , where ρ: constant and V (R) ∝ R. Also,
1
if Boyos and radius (size) of galaxy M : Constant then we know that V (R) ∝ R− 2 . At some critical
1
point, we have transition of orbits from V (R) ∝ R to Keplerian orbits in which V (R) ∝ R− 2 . A
keplerian orbit is similar to the dynamics of planets in our solar system. In reality, we must account
for inclination of galaxies (face-on systems have no observable rotational component along the line
of sight). Thus, we say that V (R) = VObserved
sin(i) where i: Inclination angle. The outer part of a
rotation curve is approximately “flat” meaning that it is not a Keplerian orbit. This means that
R2 ρ ∼ Constant, ρ ∝ R−2 , or M ∝ R.

10
Thus if we consider the example in which a spiral galaxy is rotating at a constant velocity, then
we know that
v2R
  
10 v R
M (R) = = 9.6 × 10 M0 (4.2)
G 220 km s−1 8.5 KpC
If we assume that R = 75 KpC and M ∼ 8×1010 L0 , then we have a Luminosity of about ∼ 2.3×1010 L0 .
Thus for this galaxy, the ratio of Mass to Luminosity becomes M L ∼ 35. For much larger L on
M

large scales than measured lcoally implies Ωgal ∼ 0.35. More mass than means that Luminosity ≡
dark matter. The name dark matter comes from Fritz Zwicky who in the 1930s showed the velocity
dispersion of galaxies in clusters (∼ 1000 km s−1 )  than could be explained by Luminosity of galaxies.
He gave the name “dunkle materie.”

4.3 Newtonian Gravitational Lensing


From the Newtonian perspective, if we assume that there is some particle with small mass, the particle
passes within b of a mass M traveling at near the speed of light. In this case the particle experiences
the force
GM m
F ∼ (4.3)
b2
For a time scale ∆t ∼ 2b GM m 2b
c , we can estimate the change in momentum as begin equal to ∆p = b2 c ≈
2GM m 2GM m 2GM
bc , and the deflection (defined as the change in angle) takes the form α = bc = bc2 . We now
also consider the gravitational deflection within General Relativity.

4.4 Gravitational Deflection From General Relativity


If we first assume a metric of the form
Metric: ds2 = −(1 + 2Φ)dτ 2 + (1 − 2Φ)(dx2 + dy 2 + dz 2 )
Possion’s Equation: ∇2 Φ = 4πGρ
d2 xµ ρ
µ dx dx
σ
(4.4)
Geodesic Equation: + Γ ρσ
dλ2 dλ dλ
dxµ dxν
Null Path: 0 = gµν
dλ dλ
If we assume the deflection is small; it can be treated as a small perturbation and integrated along
the path of the photon, then we find that
ˆ
2
α = 2 ∇⊥ Φ ds (4.5)
c
Here ∇⊥ is the gradient of the gravitational potential orthogonal to the line of sight. Thus, if we
assume that the potential is Φ(r) = − GM √GM ∂Φ
r = − b2 +z 2 , then we will get that ∇⊥ Φ = ∂b =
GM b
3 .
(b2 +z 2 ) 2
This means that α then equals
ˆ ˆ ˆ
2 2 ∞ GM b 2GM ∞ dx 4GM
α = 2 ∇⊥ Φ ds = 2 dz = = (4.6)
c c −∞ (b2 + z 2 ) 32 bc2 −∞ (1 + x2 ) 23 bc2

This then becomes the desired value of α for the Newtonian gravitational potential. From this, one can
construct the Thin Lens Equations to assess the impact of gravitational lensing on measured radiation.
Given this definition of the deflection angle, we find that

θDs = βDs + DLS (Small angle formula) (4.7)

11
q
4GM DLS
For β = 0 the deflection angle becomes θ = θe = Dd Ds c2 where θe : Einstein radius. This is a
solution ring with angular
q radius θe . In the case for extended sources, we have that the deflection
4GM (θe ) DLS
angle becomes θe = c2 Dd Ds .

4.5 Point-Mass Lens


θ2
For this situation, we can write down that β = θ − θt = θ − α̂(g). Here α̂(g) is define as being equal
to
DLS 4GM DLS
α̂(g) = 2
= α(g) (4.8)
Dg Dd c θ Ds
If we now solve for this quadratic equation, we will find that
 
1 q
2 2
θ1,2 = β ± β + 4gf (4.9)
2

Here the magnification is a change in solid angle of a source. This leads to 3 different types of lensed
sources. More precisely,

• Unlensed source: annulus with area equal to dA = 2πβdβ



2πθdθ
• Lensed source: area of the form dA = 2πθdθ with magnification of µ which equals µ = 2πβdβ =


θ dθ
β dβ . If we now integrate over this expression, we will find two different µ1,2 values. These

   q 
θe4 β β 2 +4θf2
1
then equal µ1,2 = 1 − θ4 = 4 (β 2 +4θ2 ) + β ±2
e

The solution of these equations leads to two images. One inside the Einstein radius and one outside
the radius. For the inside radius, inside the θe will have a mirror inverted images such that θf < θe .
We often write angles normalizable θe . This leads to the expressions,

µ = βθe−1
u2 + 2
µ1,2 = √
2v u2 + 4 (4.10)
u2
+2
µ = |µ1 | + |µ2 | = √
u u2 + 4
If a source is on θe then θe = β, which means that µ = 1.17 + 0.17 = 1.34 with u = 1. We can now take
certain examples of galaxies to assess the different characteristic scales. We know that for a Galactic
stars, the angle becomes s s
M D
θe = 0.9M as (4.11)
M 10 KpC
For an extragalactic galaxy, the deflection angle becomes
s s
M D
θe = 0.900 00
(4.12)
10 M 1 GpC

12
4.6 General Case for Lensing
We assume a “thin lens” which means that there is a thin sheet of mass. We care about mass surface
density that takes the form ˆ
Σ(ξ) = ρ(ξ, z)dz (4.13)

where z : direction along line of sight, and ξ : distance vector along plane of lens. For weak deflection,
they simply add linearly. More precisley,
ˆ
4G (ξ − ξ 0 )Σ(ξ 0 ) 2
α(ξ) = 2 d ξ (4.14)
c |ξ − ξ 0 |2

If this is circularly symmetric, then we know that

4GM (ξ)
α(ξ) =
ξc2
ˆ ξ (4.15)
M (ξ) = 2π Σ(ξ 0 )ξ 0 dξ 0
0

Here deflection is defined by mass enclosed within ξ. If Σ(ξ) is a constant, then we can define a new
variable α̂ which equals
4G DLS DLS Dd 4πG
α̂ = 2 Σ(ξ)πξ 2 = σ(ξ)θ (4.16)
c ξ Ds Ds c2
We then write our lens equation as β = θ − α̂(θ) where σ(ξ) = Σ is a constant, and that β ∝ θ. We
can define a critical surface mass density so that α̂ = θ. For this case, β = 0 for all θ ∈ R, i.e. lens
will focus all of the light at one point where β = 0. From this, the critical density will become
−1
c2

Ds D Dd DLS
Σcr = ≈ 0.35 g cm−2 , D≡ (4.17)
4πG DLS Dd 1GpC Ds

If the ratio for a specific mass density ΣΣcr = k, then if Σ > Σcr this means there are multiple images
(strong lensing). If Σ < Σcr , then we have small distortions to shapes, but not multiple images (weak
lensing).

4.7 Lensing Potential


We can introduce the idea of a lensing potential where we project and scale the Newtonian potential
(by distance). This leads to the equation
ˆ
DLS 2
ψ(g) = Φ(Dd θ, z)dz (4.18)
Ds Dd c2
The gradient of ψ gives the deflection angle. In other words Ddθ = ξ.
ˆ
2 DLS
∇g ψ = Dd ψ ξ ψ = 2 ∇⊥ Φ dz = α̂ (4.19)
c Ds
This allows us to construct the Lagrangian of ψ. More precisely,
ˆ
2 2 Dd DLS 2 2 Dd DLS Σ(θ)
∇g ψ = 2 ∇ Φ dz = 2 4πGΣ(θ) = 2 = 2k(θ) (4.20)
c Ds |{z} c Ds Σcr
Poisson’s Equation

13
We can now think about lensing as mapping a point in the source pane to a point in the lens plane.
If we write this as a matrix, one finds that
∂ 2 ψ(θ)
   
∂B ∂αi (θ)
A= = δij − = δij − (4.21)
∂θ ∂θj ∂θi ∂θj
Here Ĥ is the inverse of magnification and is known as a shear matrix. If ∂α
∂g = 0, then there is no
change in position. Here A is a 2 × 2 matrix. This means that we can write our convergence as a
1 Σ(θ)
K = (ψ11 + ψ22 ) = (4.22)
2 ΣCr (θ)
and then we can define a pseudo-vector that takes the form γ = (γ1 , γ2 ). This means that
1
γ1 (θ) = (ψ11 − ψ22 ) = γ cos(2φ(θ))
2
γ2 (θ) = ψ12 = ψ21 = γ sin(2φ(θ)) (4.23)
q
γ = γ12 + γ22
This means that we can write A in terms of k, γ . This becomes
   
1 − k − γ1 −γ2 A11 A12
A= = (4.24)
−γ2 1 − k + γ1 A21 A22
We could also write this matrix in terms of a matrix equation.
   
1 0 cos(2φ) sin(2φ)
A = (1 − k) − γ (4.25)
0 1 sin(2φ) − cos(2φ)
| {z } | {z }
Increase size of image Increase size and distance of “shear” image

Lensing in other words stretches or “shears” objects along a preferred direction φ. This imparts an
1
ellipticity to image shapes. This can be seen if we define a variable µ = det(A) = (1−k)12 −γ 2 . This shear
induced ellipticity leads the major and minor axes to
• Semi-Major axis: (1 − k − γ)−1 ≡ A.
• Semi-Minor axis: (1 − k + γ)−1 ≡ B.
For gravitational weak lensing, k ∼ 0 (no multiple lenses), which leads to a = (1 − γ)−1 and b =
(1 + γ)−1 . If we now measure shape (ellipticity) of a galaxy and we assume on average all galaxies are
circular. From this, we get γ → ψij → α → ∇⊥ Φ.

4.8 Time Delays and the Hubble Constant


The passage of light through a gravitational field leads to a time delay of the form
ˆ
∆t = − Φ ds (4.26)
C
This was derived by Shapiro in 1964. Here Φ : Potential and s : Path of the photon. The total time
delay is the sum of the deflection delay and gravitational delay to the photon. More precisely, the net
time delay equation becomes
 
1 + z Dd Ds 1 2
t(θ) = (θ − β) − ψ (4.27)
c DLS |2 {z } |{z}
∆t Grav
∆tGeom

We can now measure time delay from light curves of multiple imaged quasars e.g Q50 095 + 5Gl A,
B. Due to this, Refsdal noted that

14
• Geometric time delay is proportional to path length of photons (angular diameter distance) and
is proportional to H0 .

• Gravitational time delay also scales as H0−1 as the size and mass of lens scale as H0−1 .

We also know that H0 ∆t depends on geometry of lens. If we have a mass model then one can make a
prediction regarding the value of H0 ∆t. Thus, if we measure ∆t, then we can approximate the value
of H0 . In the case for M ∼ 1012 M , there exists an approximate time delay of ∆t ∼ 107 s ∼ 0.3 Yrs.

4.9 Mass Models for Lenses


Often, we assume a mass profile of a singular, isothermal sphere. The density of this profile would
then be defined as
σ2 1
ρ(r) = v 2 (4.28)
2πG r
here σv : Velocity Dispersion. The rotation curves will take the form
GM
vr2 = ≡ 2σv2 (4.29)
r
From this, the surface mass density can then be defined as

σv2 1
Σ(ξ) =
2G ξ
4πσ 2
α = 2 v , Constant (4.30)
c
 2  
00 σv DLS
θc = 1.6
200 km s−1 Ds

This then becomes the deflection angle relative to surface mass densities of this form.

5 Thermal History of Universe


5.1 Hot Dense Universe
Interaction rates are high. If we consider 1 second after the big bang, the mean free path is approx-
imately size of an atom. If the interaction rate is greater than the expansion of the universe, then
the system has reached equilibrium. In other words, Γ > H. If we are now in equilibrium, we can
use statistical mechanics to define density of particles. For non-relativistic particles, abundances are
mc 2
−K
governed by Boltzmann equation. This will lead to the fact that n ∝ e bT
as T and n decrease. We
can now consider the case for relativistic particles.

5.2 For Relativistic Particles


In the case for photons, the number scales as n ∝ a−3 . We can also define a baryon to photon ratio
which today approximately equals η0 = 6 × 10−10 as na ∝ a−3 and nb ∝ a−3 . This quantity from
current measurements is apparently constant. The apparent inconsistency that appears in this model
is that if this were true, then our universe would be composed of mostly photons. This is clear as when
the universe expands after the big bang, the universe cools and the there is greater decay of baryonic
matter. However, this is not the case from current measurements. Thus, this apparent contradiction
is due to our initial condition that the interaction rate is greater than the expansion rate. If we drop

15
out of equilibrium, then we will reach a critical point. We call this the “freeze out” point and this
occurs when Γ < H. This process is important for dark matter, BBN, and recombination. After the
electroweak phase transition and the QCD phase transition, we have a certain set of particle species
in the universe. In the relativistic case, we have pions, electrons, muons, neutrinos, and photons. For
the non-relativistic case, we have protons and neutrons. As the temperature decreases in the universe,
the particles begin to annihilate.
Due to this, we describe the universe in terms of “effective” number of relativistic particles. This
means that the approximate relation will be defined by the equation
7
g∗ = gb + gf (5.1)
8
Here gb = 28 for photons, W ± , Z 0 , gluons, and Higgs. For gf = 90 for quarks, charged leptons, and
neutrinos. At the beginning of QCD, we have that g∗ = 106.75. At the end of QCD, we find that
g∗ = 10.75.

5.3 Neutrino Decoupling


e− + e+ → νe + ν̄e (5.2)
Here the decoupling/freeze out occurs when Γ < H. This leads to the equation that
Γ = Ahvφi (5.3)
Here A is the number density of particles, v is the velocity, and σ is the cross section area. Finally, h i
is the average value of the quantity over phase space. For the weak interaction, the velocity is defined
as
v ∼ G2f T 2 (5.4)
where Gf = 1.17 × 10−5 GeV−2 is the Fermi constant. For relativistic particles, where v ∼ c, n ∝ T 3
and Γ ∝ G2f T 5 . In this case, the Fiedman’s equations for a radiation dominated universe take the
from Sec. 2.4 r
t 1
a(t) = , H(t) = ∝ T2 (5.5)
t0 2t
We can now substitute this quantity into Γ in order to find its temperature dependence.
 3
Γ 2 3 T
∼ Gf T ∼ (5.6)
H MeV
Here the neutrinos decouple at around T ∼ 1 Mev.

5.4 Electron-Positron Annihilation


In this situation we have that
e+ + e− −→ γ + γ (5.7)
This drops out of equilibrium. at around T ∼ 0.5 MeV. This means that
a(tdecouple )
Tν = Tν (5.8)
a(t)
Here Te tracks Tγ . here entropy density is conserved.2 From this, one can derive the equations that
+P π2 
S= , = g∗ T 4 , P (5.9)
T 30 3
In the case of before annihilation, we have that Sγ and Se+ ,e− . After annihilation, we only have Sγ .
2
This was found by KdB and Turner in the 1990s.

16
5.5 Big Bang Nucleosynthesis
The same as freeze out. It possess ν, e+ , e− and has a time scale of t ∼ 1s, T ∼ 1 MeV, and γ, e− , e+
equilibrium. Proton mass is approximately equal to the neutron mass which is about 740 MeV for
non-realtiavistic particles. In thermal equilibrium we have that

n + νe → p + e−
(5.10)
n + e+ → p + ν̄e

The density of non-relativistic particles obeys the equation,


r 2
mc2 − (mk −µ)
n= e bT (5.11)
2π~
Here g : 2 for protons and neutrons, and µ is the chemical potential for the system. In addition to
this, for abundances we will make use of Saha’s equation. Which states that
nΛ ne+ np nµ̄
0 0 = 0 0 (5.12)
nΛ ne+ np nν̄

In this situation, ne+ = n0e+ and nν̄ = n0ν̄ as both are relativistic in equilibrium. In the last case we
have that 3 2
n0n

nn Mn 2 − (mnk−mT p )c
= 0 = e b (5.13)
np np Mp
This means that Mn ∼ Mp which gives the rough value that Q = (mn − mp )c2 = 1.293 MeV. This
− kQT
then leads to nnnp = e b . In the case for T  1 MeV, nn ∼ np , and for T  1 MeV, nn decreases
exponentially.3

5.6 Neutron Abundance


If we write that Xn = nnn+n
n
p
nn
= nBaryons and define that nb ≡ nBaryons , we will find that the equation
for equilibrium abundance becomes
− kQT
e b
Xn0 (T ) = (5.14)
− Q
1 + e kb T
When neutrinos decouple because of the weak interactions at around T ∼ 0.8 MeV, we have that
Xn (0.6 MeV) = 0.16. This means that 16% of baryon abundance is in the form of neutrinos.

5.7 Neutron Decay


t
After the process of freeze out, we will find that Xn∞ = 0.15, Xn = Xn∞ e− τn , where τn is the lifetime
for the neutron. From current measurenments, this is approximately 887s.

Binding Energy: Reaction:


2.2 MeV p(nγ)D
(5.15)
8.48 MeV D(n, γ) H3 , D(D, p) H3
7.72 MeV D(p, γ)H3e , D(D, n)H3e ,
3
If weak interactions are efficient, then there would also be no neutrinos. The lack of neutrino produciton would also
lead to the lack of He4 production.

17
We require D to initiate the process of n + p → D + γ. From Saha’s equation, we know that (nγ = n2γ )
becomes 3 3
3 2πm0 ~2 2 − (m0 −mkA −m
 p )c
n0
= e BT (5.16)
nn np 4 mn mp kb T
Here g0 = 3, gp = gn = 2, and m0 = 2Mn . We can now relate n0 to baryon density. Currently, the
current measurement of np ≈ 0.75nb . The Baryon-to-photon ratio equals nb = ηnγ , which means that
 2
−10 Ωh
η = 5.5 × 10 (5.17)
0.02

Here η is small so this inhibits D production i.e reaction is slow. This leads to neutron decay.

5.8 Helium Production


In this process D(p, γ)He3 or H3e (D, p)H4e energies. The binding energy for helium 4 is approximately
equal to 28.3 MeV. Neutron abundance determines He4 abundance based on the equation
t t
Xn (t) = Xn∞ e− τn = 0.15e− τn (5.18)

Here we need that X0 > 10−3 in order to produce Helium in “volume.” This occurs at T ∼ 0.075 MeV
or t ∼ 210s.

5.9 Helium Abundance


The approximate ratio of Helium to neutron abundance is given by the equation
1
XHe = Xn (5.19)
2

In terms of mass fraction of YHe . This means that yHe = 4ΛnHe b


∼ 2Xn = 0.24. One can actually
calculate these abundances. If one is given all of the cross sections for all reactions, then we can define
(compute) relative abundance of light elements (from initial conditions).4 In the BBN process, we
have predominantly t ∼ 300s, T ∼ 0.8 × 109 K ∼ 0.07 MeV. If we predict that 24%(by mass) of Helium
4 baryons, then in the BBN process, we will find that the reaction
−3 −6
He4 →×10 He3 →×10 Li (5.20)

Here the abundance depends on η, τn , g∗ at T ∼ 1010 K.

5.10 Sensitivity to Baryon Fraction η


During nucleo-synthesis of D, η increases. We also know that He3 and He4 start earlier, and there are
more neutrons with a larger He4 abundance. In this case, we have a weak relation which means that
D, He3 burn more efficiently, and D, He3 abundance decreases. In addition, Li7 has two reactions.
One reaction is through direct formation which occurs at small η. Then there is indirect formation
which occurs via Be7 . This indirect process occurs at high η.

• τn increases: In this case nn increases and leads to an increase of He4 abundance. On top of this
∆yHe ∼ 2.4 × 10−4 (∆τn /ls).
4
This work was founded by Wagener et. al in 1967.

18
−1
• gs increases: In this case,we have that Σr ∝ g∗ T 4 as H ∝ −1 and t ∝ g∗ 2 . From this, we can
see that as g∗ increases, H also increases i.e faster expansion. When freeze out occurs earlier,
the n-p ratio is higher. This then leads to an increase of He4 abundance. Ultimately, this will
mean that
∆yHe4 ∼ 0.01∆g∗ (5.21)
with light constraints on number of neutrino families.

We would now like to understand how this can be assess through observational data.

5.11 Observational Constraints


He4 will typically form from ionized states of HII clouds (recombination lines). These measurements
are made in metal poor clouds. This is done in order to avoid any mixed measurements that come
from metal reacting to produce more helium. This is the case as helium is formed within stars. From
this, we can extrapolate to zero metallicity the abundance of helium 4. Current measurements predict

yHe = 0.24 ± 0.01 (5.22)


D
We can now define a variable D which is a form of OV absorption in ISM. This means that H ∼
−5
1.6 × 10 . This is a lower bound because it is easy to destroy D within the universe. This is
estimated from Ly-α clouds which have high redshift and low metallicity. This would then give a
value of
D
∼ 2.82 × 10−5 (5.23)
H
which is the strongest constraint on ΩB . Measurements for Helium 3 are also done. In this case, one
measures HII regions (He3+ hyper-fine line equivalent to 21 cm emission in hydrogen.) within bodies
such as meteorite. From current measurements, one finds an approximate abundance of

He3
= 1.4 × 10−5 (5.24)
H
This assessment is also done on Li7 which comes from circulation through the centers of stars. Mea-
surements is typically done on high metallicity stars. This gives a relative abundance of
 7
7 Li
[Li ] = 12 ± ln 2.1 ± 0.1 (5.25)
H

This also gives a measurement of η = 6.23 ± 0.17 × 10−10 = 6.1 ± 0.6 × 10−10 .

5.12 Recombination and Decoupling


From the Big Bang Nucleosythnesis to recombination, we know that the temperature of the universe
goes from T ∼ 1 MeV to about T ∼ 1 eV. Due to this transition, we will find that the scale factor a(t)
1 2
within Friedman’s equation will transform into a(t) ∝ t 2 → a(t) ∝ t 3 . In this transitional period, we
will find that there is extensive Compton scattering. This occurs due to the free baryons (electrons
+ nuclei) in the plasma. These kinds of particle interactions look like e− + γ → γ + e− . There will
also exist Coulomb scattering which takes the form e− + P → e− + p. Due to this, we know that as
temperature decreases, e− and P combine (recombine) which leads to a decrease in the density of e− .
This also means that the mean free path for γ also increases which leads to a universe which is more
transparent. We also find that the γ decouple and propagate outward. These light rays then become
part of the Cosmic Microwave Background.

19
We will first consider what happens in the simplified version of P +e− scattering. If the temperature
is T > 1 MeV, one kind of scattering process that will be present is P +e− → H +γ. Then from Saha’s
n n n0 n0
equation, we know that neHp = ne 0 P . We note that at nγ = n0γ , we will find that Saha’s equation
H
becomes  3
ne np ge gP me kb T 2 − k BT
= e B (5.26)
nH gH 2π~2
In this situation, we assume that MP ∼ MH , ge = gp = 2, and gH = 4. In addition, B = 13.6 MeV,
ne = np , and nb = nP + nH = ne + nH . The question now arises is at what temperature this process
of recombination occurs? At around T ∼ 32, 000 K, recombination occurs with a redshift of about
z ∼ 12, 000. We can also define an ionization fraction of xn . First we know that
ne np 1−x
xe = = , where nH = np (5.27)
nb np + nH x

We can then use Saha’s equation and the fact that if ne = np , then
 3
me kB T − 2 k BT

1−x
= np e B (5.28)
x 2π~2
np
where η = xnγ and assuming that the universe only has hydrogen. In the Big Bang spectrum, we will
find that
2.404 kB T 3 kB T 3
   
nγ = = 0.243 (5.29)
π2 ~c ~c
If we now substitute back into our original expression, we will find that
 3
1−x kB T 2 k BT
= 3.84η e B (5.30)
x me c2

1+4s−1
The quadratic solution to this equation would lead to x = 2s and that
 3
kB T 2 B
S = 3.84η e kB T (5.31)
me c2

This then allows us to solve for x(t), T (z), and x(z). During recombination, we can define a few of
the necessary parameters. If we assume that xe = 0.5, kB TRec = 0.32 eV, or T ∼ 3760 K. This means
that as TRec = T0 (HzRec ) and T0 = 2.725 K. If we assume that zRec ∼ 1380, then the age of the
universe will become t ∼ 250, 000 Yrs. This means that recombination is fast, but not instantaneous.
The x(t) para-mater transitions from x = 0.9 to x = 0.1 in about 70, 000 Yrs. After this, the universe
transitions into a decoupling phase. In this case, there is scattering of γ through Thompson scattering
(low energy limit of Compton scattering). The associated particle equation is e− + γ → γ + e− . This
α2 ~62 −29 M2 .
will lead the decay of γ to be defined as Γe = ne σT c = nb xe σT c. Here σT ∼ 8π3 m2e c2 ∼ 6.6510
For massive particles, σT → 0. As the density of e decreases, Γγ also decreases. This means that the
decay equation can be defined as

Γ(z) = nb,0 (1 + z)3 x(z)σT c (5.32)

In a matter dominated universe, the Hubble constant will then be defined as H(z) = H0 (1 + z)3 . With
3
the initial conditions for this decoupling phase, H(z) becomes H(z) = 1.23 × 10−18 (1 + z) 2 . From

20
this, one finds that Γ(z) = 5.0 × 10−21 x(z)(1 + z)3 where 1 + z = 39.3
3 . In this case, z ∼ 1100. In
x(z)z 2
this recombination phase, we can also find the optical depth. This is defined as
ˆ
τ (t) = Γ(t) dt (5.33)

This equation states the probability that a γ will scatter off of a e− . We can also repress t as a function
of z due to the realtion dt = da
ȧ . This means that
ˆ z ˆ z ˆ z
Γ(z) dz 3 1 dz √
τ (z) = = σT cnb,0 (1 + z) 3 = 0.0041 1 + z dz (5.34)
0 H(z) 1 + z 0 H0 (1 + z) 2 1 + z 0

This expression will integrate to τ (z) = 1 at z = 1072 which becomes the last scattering surface.

5.13 Discovery of CMB


It was discovered in 1964 by Penzias and Wislon at λ = 7.35 cm. Horn reflection antenna used for
measuring reflection signals from communication satellite. An excess signal was found when directed
at all directions in the sky. It was 3.5 K was higher than expected. This was assumed to be indicative
of the cosmic microwave background and not noise due to the isotropic nature of the measurement.
Robert Dickie (Princeton) was building a receiver to measure the CMB which he started after the Bell
Labs measurements. Both the experimental and theory measurement were published back to back.
Wilson and Penzias received the nobel prize for their work.

5.14 Shape of Cosmic Microwave Background


The cosmic microwave background has a black-body spectrum. It was first measured by the FIRSAS
instrument on the COBE Satellite in 1989. The peak of the black-body occurs at a wavelength of
λPeak = 1.51+z
mm
. The “earlier” discovery in 1941 when the CN transition at λ = 2.6 mm. The
cosmic microwave background has a uniform flux at a temperature of ∆T = 3.365 K. This shift in
the temperature occurs due to the Doppler motion of satellite with the CMB. This means that the
CMB defines a “rest frame.” The motion of the local group due to gravity then becomes 620 km/s2 .
The higher order terms remain after subtracting the dipole and the galaxy emission. This leads the
 2  1
2
∆T
hT i = 2.725 K, and the T . There is a 20 µK variation found at any point in the sky. We
will now assess at what scales one expects to see fluctuations within the CMB. If we assume that we
live in a matter dominated universe, we will find that the horizon red-shift occurs at an approximate
value of z = 1100. From the Fiedman’s equations, we will find that
2c
dHorizon (zrec ) = (5.35)
H(z = 1100)

The horizon (the region in causal contact) is the proper distance as z → dHorizon (zrec ) ∼ 0.24 MpC. The
horizon scale scale in degrees in which θhorizon = 1.5 degrees. If we calculate this for our benchmark
universe, we get 2 degrees. Another question of interest is what causes the thermal fluctuations in the
CMB. At the time of last scattering, there are three possible processes. The gravitational potential
which states that photons climb out of the gravitational potential, lose energy (gravitational red-shift)
and the temperature of the photons goes down. The doppler shift, photons with their last scatter are
red-shifted (scattered away) or blue-shifted (scattered towards) relative to the observer. Or there is
compression, where the photons and baryons are tightly coupled before decoupling. The temperature
of the photons is higher in compressed regions of the universe. In an adiabatic universe η = nnγb , which

21
δn δn
means that nγγ = δn 3 δT 1 γ δT 1 δnb
nb as nγ ∝ T . This means that T = 3 nγ and T = 3 nb . Thus, an increase in
b

δρ means an increase in δT (density and temperature are tightly coupled). This leads to the fact that
δT 1 δρb
T = 3 ρb . The formation of structure on large scales (scales larger than the horizon)

22

You might also like