Introduction to Cosmology

C.P. Dullemond
Based in substantial part on script by M. Bartelmann 1

1 http://www.ita.uni-heidelberg.de/research/bartelmann/Lectures/cosmology/index.shtml

Chapter 1

Introduction
1.1

A brief history

If we look at the night sky, the most conspicuous thing to observe is also the one that
we always take for granted: that the night sky is black. There are, of course, stars in
the sky, but their combined radiation is not enough to read a book by. At first sight this
may seem natural, but it isnt. If you stand in a forest, and the forest is large enough,
then no matter where you look, your sight is blocked by a tree. If the Universe is
infinitely large, then no matter where we look, our sight should be blocked by a star.
If the average temperature of these stars is, say, 6000 K, then our entire sky should
look like the surface of a star of about 6000 K. In other words, we would fry. Since
the sky is dark, something must be wrong with this argument. This is called Olbers
Paradox. It shows that our Universe cannot be a static, infinitely large and infinitely
old Universe. A better model of the Universe must be constructed.
The history of modern cosmology starts with the discovery that the Universe is expanding. The spectra of galaxies were found to be redshifted, and that this redshift
increases with the galaxys distance. The most obvious interpretation was that there
exists a linear relation between velocity and distance. Hubble (1929, Astrophysical
Journal 74, 43) is customarily credited with discovery, because his observations established this relation beyond reasonable doubt. In fact, the expansion coefficient
is called the Hubble constant. However, the book by Nussbaumer & Bieri (2009,
Discovery of the expanding Universe, Cambridge University Press) points out that a
tentative discovery of this velocity-distance relation was already made by Wirtz (1922,
Astronomische Nachrichten 215, 349) and Lundmark (1925, Monthly Notices of the
Royal Astronomical Society 85, 865). Even earlier, Slipher (1917 Proc. Amer. Phil.
Soc., 56, 403) already detected the redshift of spectral lines in spiral nebulae before
they were proven by Hubble (1925) to be extragalactic objects.
In the same decade, Alexander Friedmann found a solution to the equations of Einsteins theory of General Relativity that describes the universe (1922, Zeitschrift fur
Physik 10, 377; 1924, Zeitschrift fur Physik 21, 326). The spatial geometry of that
universe is either flat, or positively or negatively curved. And most importantly: It
expands. At first, Einstein did not believe in the expansion predicted by these solutions and dismissed them in an article. He later realized his mistake and rehabilitated
Friedman in the same journal. George Lematre, a Belgian monk and physicist, was
apparently unaware of Friedmanns solution, but he was well aware of the observational evidence of receding galaxies. In an article written in the somewhat obscure
journal Annales de la Societe Scientifique de Bruxelles in 1927 (an English version
can be found on the web) he presented a similar model as Friedmanns to explain
these receding motions. In the 1930s it was shown by Robertson in 1933 and Walker
in 1934 that the spatial geometry assumed by Friedmann and Lematre (flat, or pos2

itively/negatively curved) are in fact the only homogeneous and isotropic geometries
possible, and thereby showing that the Friedmann-Lematre equations (with a cosmological constant) give a complete description of the dynamics of the universe. Indeed,
these models are still valid and in active use today!
If the universe is expanding, it is logical (though not necesssarily imperative) to assume that at some finite time in the past all the matter in the universe must have
converged into a single point, or in other words that the universe was born out of a
huge point explosion. Fred Hoyle was not particularly fond of this idea and called this
scathingly the Big Bang in a radio interview. It has been called the Big Bang ever
since.
So, how do we verify if this model is correct, and how do we know how the contents
of the Universe developed over time? Fortunately for us, the finite light travel time
acts as a kind of time machine: By looking at the sky with ever bigger and better
telescopes we can look at objects ever farther away, and as a consequence, further
back in time. In fact, we can look back in time to the point where the Universe was
just 3 105 times the present age, as we shall see later.
Going back in time there must therefore be a point in time when the universe was
extremely dense. Using the laws of adiabatic compression, we must conclude that the
universe must also have been extremely hot in these early days of its existence. It turns
out that in about the first 20 minutes of the Universes existence the temperatures and
densities were so high that neutrons and protons were formed and paired up to form
(in addition to Hydrogen of course) Deuterium, Helium and a bit of Lithium. The
predictions from these primordial cosmic abundances by this theory of Big Bang
Nucleosynthesis are very well confirmed by observations. All heavier elements that
are around today are all made by stars.
Apart from these elemental abundances, is there some other evidence of the Big Bang?
We know that hot dense matter tends to produce a lot of thermal radiation, like the
surface of the sun. When the universe became tenuous and cool enough for electrons
to recombine with the H and He atoms, the universe become transparent (roughly half
a million years after the Big Bang). This radiation would continue to flow through
space. Gamov, Alpher and Herman realized in the 1940s that this radiation should still
be all around us today, albeit redshifted to very low temperatures: just a few Kelvin.
In the 1960s the group around Dicke at Princeton University essentially rediscovered
that such low temperature background radiation must exist, and started creating the
necessary equipment to detect it. Unfortunately for them, another team accidently
stumbled on this radiation before Dickes team found it. Arno Penzias and Robert
Wilson were technicians of Bell Labs who were testing new radio equipment. They
were trying to get rid of a perpetual noise, but whatever they tried, they could not get
rid of it. They happened to work at Crawford Hill in New Jersey, which is just a few
miles from Princeton, so they called Dicke to ask for advice. When Dicke put up the
phone he is famously quoted as saying Boys, we have been scooped. Penzias and
Wilson won the Nobel Prize for their discovery of the cosmic microwave background
radiation (often abbreviated as CMB). This radiation is still today the most distant
signal that we can observe with telescopes, dating from redshift z # 1100, when the
universe was only half a million years old. Compare that to the current age of the
universe, 13.7 0.13 billion years, this means that the CMB is really a relic from the
infant universe.
Interestingly, the CMB is very homogeneous. The temperature is 2.725 Kelvin, no
matter into which direction you look. The current universe is, however, highly structured, at least on scales of galaxy clusters and smaller. So how did the universe get
so structured? And what seeded these structures? If the universe was born perfectly
homogeneously, then no structures could ever form. So it was a huge discovery when
the COBE Satellite detected tiny perturbations on the CMB in the early 1990s, making George Smoot and John Mather Nobel laureats in 2006 (the latter for the proof
3

that the CMB is a near-perfect blackbody curve). These anisotropies were tiny: on
scales of a few K! But they are the signatures of the initial seeds that created the rich
structure of the cosmos we have today. Later measurements with the BOOMERanG
balloon observatory and the Wilkinson Microwave Anisotropy Probe (WMAP) space
telescope have since dramatically improved the measurements of these anisotropies,
allowing detailed analysis of their spatial power spectra and thereby narrowing down
many of the previously unknown characteristics of the universe. Currently the Planck
space telescope is redoing this at even higher precision.
Now, what caused these anisotropies in the first place? For that, we have to go to
much earlier times: to times around 1036 seconds after the Big Bang. Before that
time it is believed that several of the fundamental forces of Nature were unified into
a single Grand Unified Theory. Then, according to current theories, a period of extremely rapid exponential growth followed, known as inflation. This period lasted
from roughly 1036 seconds until about 1032 seconds after the Big Bang during which
the volume of the Universe expanded by about a factor of 1080 . This inflation period
had several effects: (a) regions in the universe became mutually disconnected, (b)
spatial curvature got flattened away, (c) inhomogeneities were mostly washed out.
But during this period quantum fluctuations in the material density also got frozen
in as the universe expanded so rapidly. Tiny fluctuations at the quantum level could
thus cause large scale effects. Also, because of the many orders of magnitude expansion, while quantum fluctuations kept occurring, the macroscopic perturbations they
caused should be on all scales: The earlier fluctuations having expanded further and
being responsible for the larger scale perturbations. The power spectrum of these perturbations therefore are expected to be scale-free (Harrison-Zeldovich-Peebles power
spectrum). If this theory is correct, then we also expect the anisotropies of the CMB to
be Gaussian: the T should have a probability distribution that is a Gauss curve. This
is a prediction from quantum field theory. So far the CMB appears to be consistent
with Gaussianity. The search for non-Gaussianity is on-going.
Going back to the era after the production of the CMB, around half a million years
after the Big Bang. The first small perturbations had already reached the few percent
level and continued to grow by gravitational contraction. Most of the matter is in the
form of Dark Matter, i.e. presumably weakly interacting massive particles that form
a kind of collisionless self-gravitating fluid. In fact, only roughly 1/6 of the matter is
thought to be baryonic matter (stars, planets, gas, dust etc), while 5/6 of the matter in
the universe is thought to be in the form of this collisionless (cold) dark matter (often
written as CDM). The formation of structure is therefore mostly a matter of the nonlinear evolution of a self-gravitating collisionless fluid in which the initial seed perturbations are imprinted. Large scale simulations such as the Millennium Simulation
(Springel, 2005, Nature 435, 629) then follow the non-linear evolution in detail. What
comes out is the formation of a cosmic web with huge voids, walls, filaments and
massive galaxy clusters. However, initially these structure do not produce any light
because no stars have yet formed. The period between the CMB and the first stars is
often called the Dark Ages. The first stars are thought to have formed roughly a few
hundred million years after the Big Bang. These stars produced so many UV photons
that they started to reionize the gas in the universe. One of the hot topics today is to
search for indications of this reionization process happening, as it would give evidence
for early star formation as well as the structure of the early universe and the end of
the Dark Ages. In particular the LOFAR radio array currently under construction, and
the future SKA array, will search for neutral hydrogen from the Dark Ages through
measurements of the redshifted 21 cm hydrogen fine structure line.
It will be very difficult, if not impossible, to directly observe the first stars themselves
during their formation. But there are many indirect pieces of evidence, such as very
low metallicity stars that are still present today. But the search for the highest redshift
galaxies is of course still very much ongoing.

Time after BB
< 1043 sec
1043 -1035 sec
1035 -1032 sec
1032 -1012 sec
1012 -106 sec
106 -101 sec
0.2 sec
2 sec
40 sec
3 - 5 min
60 103 year
300 103 year
0.3 400 Myr
0.4 Gyr
0.65 Gyr
9.17 Gyr
9.6 Gyr
13.74 Gyr

1.15 1010
3.3 109
8 108
3 108
3233
110080

3.2 1010 K
9.3 109 K
2 109 K
8 108 K
8590 K
3100 K

11
8
0.43
0.39
0

30 K
25 K
3.9 K
3.8 K
2.725 K

Event
Planck era
GUT era
Inflation
Electroweak epoch
Quark epoch
Hadron epoch
Freeze-out of + e+ + e
Freeze-out of n+e p+e
Freeze-out of e+ + e 2
Nucleosynthesis
Matter starts dominating
Recombination (Origin of CMB)
Dark Ages
Re-ionization by first stars
Farthest known galaxies (Sep 2011)
Formation of Earth
Dark energy starts dominating
Today

Table 1.1: Time line of the Universe, computed according to the Friedmann model using the parameters from Jarosik et al. (2011), see table
4.1.
Information about the very early Universe can be found at e.g.
http://en.wikipedia.org/wiki/Timeline_of_the_Big_Bang.

In 1998 a new chapter in cosmology was written with the discovery that the Universe is currently expanding at an accelerated rate. By observing and studying type Ia
supernovae at high redshifts it was found that the expansion is faster than expected.
Apparently the Universe must be permeated with a mysterious quantity called dark
energy. On Tuesday, October 4, 2011 it was announced that Saul Perlmutter, Brian
Schmidt and Adam Riess won the Nobel Prize of Physics 2011 for this discovery. The
implication of this discovery is that our Universe will not recollapse, but will expand
forever. The current standard model of cosmology, which includes this dark energy, is
called the -CDM model.

1.2

Some fundamental concepts

Before we are going to start with developing a model of the Universe we must first
introduce a few concepts.
1.2.1 Redshift
If we make a spectrum of a galaxy and find that the location of the spectral lines has
been shifted by
observed real
z=
(1.1)
real
then we call z the redshift of this galaxy. The ratio of the energy of the photon as it is
observed observed to when it was emitted hreal is
1
hobserved
=
hreal
1+z

(1.2)

You can see why the radiation is redshifted in two ways. The easiest is to realize that
if you send a photon from a moving object, it will be Doppler-shifted with respect to
a non-moving object. This explains redshift for the near Universe, where z & 1, but
for galaxies at z ! 1 this classical concept is not self-consistent. A special-relativisitic
treatment would be necessary. However, a much easier way to understand it is if one
5

regards the expansion of the Universe as a real expansion of space itself. If a galaxy
emits an electromagnetic wave with wavelength real propagating through space, but
by the time it arrives in our telescope space has been stretched by a factor A then the
wavelength is also stretched by this factor: obs = Areal .
1.2.2 The Hubble constant
If we can find a reliable way of determining the distance to the galaxies we observe
(and this is not an easy task!), then we can make a direct connection between the
redshift and the distance (at least for the near Universe for which z & 1). This turns
out to be linear, and can thus be written as:
D=

c
z
H0

(1.3)

where H0 is called the Hubble constant. The Hubble constant is a way to express the
expansion rate of the Universe. If we write the redshift z as a velocity z = v/c where v
is the velocity by which the galaxy we observe is moving away from us, then
H0 =

v
D

(1.4)

i.e. the Hubble constant is the velocity of recession per unit distance. Currently this
value is thought to be H0 = 70 km/s/Megaparsec with reasonable confidence. The
unit of H0 is 1/second, so it gives already a rough estimate of the age of the universe:
1/H0 = 13.9 Gyr. This simple estimate is called the Hubble time. If you multiply
this with the light speed c you get a distance called the Hubble distance (also called
Hubble radius which is rH = c/H0 = 4.2 Gpc.
The word Hubble constant is a bit misleading. The constant refers to the fact that
the observed ratio v/D appears not to vary with D. But if we follow the evolution of
the Universe over its entire life time, then H(t) is not constant at all. Thats why for
the present-day Hubble constant we write H0 with index 0. We will do this for many
other variables as well.
It is customary to introduce a constant h defined as
H0 = 100 h km/s/Mpc

(1.5)

With the current estimate being h = 0.7.