Professional Documents
Culture Documents
Ka Chun Yu: Lecture Notes For Introduction To Cosmology
Ka Chun Yu: Lecture Notes For Introduction To Cosmology
Ka Chun Yu
Denver – 2005
DMNS Technical Report 2005-12, 167 pages total
by
Dr. Ka Chun Yu
Available at https://scientists.dmns.org/sites/kachunyu
Contents
3 Theoretical Universes 53
3.1 The Curvature of Space-Time . . . . . . . . . . . . . . . . . . . . . . 56
3.2 The Distribution of Matter and Energy in the Universe . . . . . . . . 58
3.3 Modeling the Universe . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3.1 Einstein’s Universe . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3.2 The de Sitter Universe . . . . . . . . . . . . . . . . . . . . . . 62
3.3.3 The Friedmann-Robertson-Walker Universes . . . . . . . . . . 62
3.4 Cosmological Redshifts and the Hubble Constant . . . . . . . . . . . 67
3.5 The Critical Density . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.6 The Age of the Universe . . . . . . . . . . . . . . . . . . . . . . . . . 70
iii
iv CONTENTS
v
vi LIST OF FIGURES
2.1 Population Model of M 31 Bulge: The derived mass and V-band luminosity
fraction of stars within each spectral type range. . . . . . . . . . . . . . . 24
ix
Chapter 1
Cosmology is the branch of science concerned with the study of the entire Universe as
a whole. The fundamental questions that are asked by cosmologists include: “What
is the Universe made of?” “What is its structure?” “How did the Universe begin?”
“What is its eventual fate?” Long before we had the technology to enable us to firmly
answer them, these questions have interested philosophical and religious thinkers since
ancient times.
We are therefore very fortunate to be living in a time when we have the technology
and know-how to begin answering these questions. Cosmology is not a static, stodgy
science. Progresss in discoveries are coming quicker than ever before. Although we
will be covering much of what has been discovered from the early years of the 20th
century, we will also concentrate on major discoveries in the last ten years. We are
entering a period of “precision cosmology,” first coined by the American cosmologist
Michael Turner. Cosmological theories are being put to the test by observations.
Instead of theorists conjecturing wildly about the past, current and future states of
the Universe, they are being constrained by what observers are finding from the latest
telescopes and instrumentation. What is also exciting is that from multiple lines of
independent evidence, we are starting to piece together a very coherent, consistent
picture of the Universe.
For this course, I am assuming a basic understanding of astronomy as a prerequi-
site. If you have an amateur’s interest in the field, read the occasional issue of Sky &
Telescope or Astronomy at the dentist’s office, or took an Astro 101 course a decade or
two ago, then you are probably prepared. If this is your first exposure to astronomy
in the formal school or informal museum education sense, then there might be some
rough patches. Although there will be opportunities for defining terms, more often
than not, we will be plunging ahead assuming you know basic definitions like star,
galaxy, spectrum, and Doppler shift. If you feel rusty about what a light year or an
absolute magnitude is, you may want to check out the Appendix to get brief primers.
1
2 CHAPTER 1. COSMOLOGY FROM ANCIENT TO MODERN TIMES
of cosmology. Although the ancient Greek philosophers had a variety of ideas about
the nature of the universe—many of which we know to be incorrect—we can begin to
sense a thread of discovery and inquiry that will eventually lead to modern science.
Thales of Miletus (634–546 BCE) believed the Earth was a flat disk surrounded
by water.
Anaxagoras (ca. 500–ca. 428 BCE) believed the world was cylindrically shaped,
and we lived on the flat-topped surface. This world cylinder floats freely in space
on nothingness, with the fixed stars in a spherical shell that rotated about the
cylinder. The Moon shone as a result of reflected light from the Sun, and lunar
eclipses were the result of the Earth’s shadow falling on the Moon.
Eudoxus of Cnidus (ca. 400–ca. 347 BCE) also had a geocentric model for the
Earth, but added in separate concentric spheres for each of the planets, the Sun,
and the Moon, to move in, with again the fixed stars located on an outermost
shell. Each of the shells for the seven heavenly bodies moved at different rates
to account for their apparent motions in the sky. To keep the model consistent
with observations of the planets’ motions, Eudoxus’ followers added more circles
to the mix—for instance, seven were needed for Mars. The complexity of this
system soon made his model unpopular.
Aristotle (384–322 BCE) refined the Eudoxus model, by adding more spheres
to make the model match the motions of the planets, especially that of the ret-
rograde motions seen in the outermost planets. Aristotle believed that “nature
abhors a vacuum,” so he postulated a universe that was filled with crystalline
spheres moving about the Earth. Aristotle also believed that the universe was
eternal and unchanging. Outside of the fixed sphere of stars was “nothingness.”
Aristarchus (ca. 310–ca. 230 BCE) made a first crude determination of the
relative distance between the Moon and the Sun. His conclusion was that the
4 CHAPTER 1. COSMOLOGY FROM ANCIENT TO MODERN TIMES
Sun was 20× further, and the only reason they appeared to be of the same size
was that the Sun was also 20× larger in diameter. Aristarchus then wondered,
if the Sun was so much larger, does it really make sense for it to move around
the Earth? Would it make more sense for the Earth to circle it?
Claudius Ptolemy (ca. 100–ca. 170 CE) writing in Syntaxis (aka Almagest;
∼ 140 CE) took the basic ideas of Eudoxus’ and Aristotle’s cosmology, but had
the planets move in circular epicycles, the centers of which then moved around
the Earth on the deferent, an even bigger orbit. Ptolemy’s ideas gave the most
accurate explanations for the motion of the planets (as best as their positions
were known at the time). (Ptolemy’s and Aristotle’s ideas about the universe
and its laws of motion remained the dominant idea in Western thought until
the 15th century CE!)
Although Copernicus made an immense leap by moving the displacing the Earth from
the center of the Universe, in other ways he was still stuck with the past. The heavens
were still moving via perfect circular motions. In order for his model to accurately
reflect the actual motions of the planets, Copernicus still had to use the motion of
smaller circles, known as an epicyclet, that orbited an offset circle.
However it cannot be understated how revolutionary was Copernicus’ De Revo-
lutionibus. Displacing the Earth from the center of the Universe was an attack on
a worldview held by all serious medieval European thinkers, and one which had a
lineage that could be traced back more than two millenia. Copernicus was attacked
immediately by both Catholic Church officials and by leaders of the Protestant move-
ment. However he died soon after the publication of his work and so was spared the
wrath of the authorities. Others who also had heretical views about the Universe
were not so lucky. Galileo Galilei (see below) was forced to recant by the Church,
and Giordano Bruno (also more below) was burned at the stake.
Thomas Digges (1546–1595), a leading English admirer of Copernicus, published
A Perfect Description of the Celestial Orbes, which re-stated Copernicus’ heliocentric
theory. However Digges went further by claiming that the universe is infinitely large,
and filled uniformly with stars. This is one of the first pre-modern statements of the
Cosmological Principle.
1.3. THE BEGINNINGS OF MODERN SCIENCE 5
Giordano Bruno (1548–1600) goes even further: not only are there an infinite
number of stars in the sky, but they are also suns with their own solar systems, and
orbited by planets filled with life. These and other heretical ideas (e.g., that all these
other life-forms, planets, and stars also had their own souls) resulted in him being
imprisoned, tortured, and finally burned at the stake by the Church.
Tycho Brahe (1546–1601) made and recorded very careful naked eye observations
of the planets, which revealed flaws in their positions as tabulated in the Ptolemaic
system. He played with a variety of both geocentric and heliocentric models.
Johannes Kepler (1571–1630) finally was able to topple the Ptolemaic system
by proposing that planets orbited the Sun in ellipses, and not circles. He proposed
his three laws of planetary motion. In 1610, Kepler also first pointed out that an
infinite universe with an infinite number of stars would be extremely bright and hot.
This issue was taken up again by Edmund Halley (1656–1742) in 1720 and Heinrich
Olbers (1758–1840) in 1823. Olbers suggested that the universe was filled with dust
that obscured light from the most distant stars. Only 20 years later, John Herschel
showed that this explanation would not work. The problem of Olber’s paradox would
not be resolved until the 20th century.
Figure 1.3: Johannes Kepler, Galileo Galilei, & Sir Isaac Newton
Left to right: Johannes Kepler, Galileo Galilei, and Sir Isaac Newton.
gravitation that today bear his name. His Philosophiae Naturalis Principia Math-
ematica—or simply, the Principia—was the first book on theoretical physics, and
provided a framework for interpreting planetary motion. He was thus the first to
show that the laws of motion which applied in laboratory situations, could also apply
to the heavenly bodies.
Newton also wrote about his own view of a cosmology with a static universe in
1691: he claimed that the universe was infinite but contained a finite number of stars.
Self gravity would cause such a system to be unstable, so Newton believed (incor-
rectly) that the finite stars would be distributed infinitely far so that the gravitational
attraction of stars exterior to a certain radius would keep the stars interior to that
radius from collapsing.
The English astronomer Thomas Wright (1711–1786) published An Original
Theory or New Hypothesis of the Universe (1750), in which he proposed that the
Milky Way was a grouping of stars arranged in a thick disk, with the Sun near the
center. The stars moved in orbits similar to the planets around our Sun.
Immanuel Kant (1724–1804), the German philosopher, inspired by Wright, pro-
posed that the Milky Way was just one of many “island universes” in an infinite
space. In his General Natural History and Theory of Heaven (1755), he writes of the
nebulous objects that had been observed by others (including Galileo!), and reflects
on what the true scale of the universe must be:
Because this kind of nebulous stars must undoubtedly be as far away from
us as the other fixed stars, not only would their size be astonishing (for
in this respect they would have to exceed by a factor of many thousands
the largest star), but the strangest point of all would be that with this
extraordinary size, made up of self-illuminating bodies and suns, these
stars should display the dimmest and weakest light.
1.4. THE COPERNICAN REVOLUTION 7
Sir William Herschel (1738–1822) and his son John (1792–1871) used a tele-
scope, based on a design by Newton, to map the nearby stars well enough to conclude
that the Milky Way was a disk-shaped distribution of stars, and that the Sun was
near the center of this disk. He mapped some 250 diffuse nebulae, but thought they
were really gas clouds inside our own Milky Way. Others however took Kant’s view
that the nebulae were really distant galaxies. The German mathematician Johann
Heinrich Lambert (1728–1777) adopted this idea, plus he discarded heliocentrism,
believing the Sun to orbit the Milky Way like all of the other stars.
equations:
T,
G = 8πT
predicted a Universe to be either expanding or contracting. This contradicted what
was known about the Universe at the time, and it was also against Einstein’s sensibil-
ities. Einstein as a result added a term into his equations, the cosmological constant
to keep his model Universe from being dynamic:
T,
G + Λ = 8πT
Depending on its sign, a cosmological constant can keep a growing universe from
getting bigger, and it keep keep a shrinking universe from getting smaller. However to
keep the universe perfectly static, the cosmological constant has to balance out exactly
the other terms. Why it should have such an arbitrary value cannot be explained
from first principles, but is more of an ad hoc solution to fit the requirement of a
static universe. This is therefore not a very satisfactory solution for physicists.
Dutch astronomer Willem de Sitter (1872–1934) used Einstein’s General Rel-
ativity equations with a low (or zero) matter density but without the cosmological
constant to arrive at an expanding universe (1916–1917). His view was that the
cosmological constant:
In 1920, Harlow Shapley and Herbert Curtis held a debate on the “Scale of
the Universe,” or really about the nature of the “spiral” nebulae. Shapley argued that
these were gas clouds inside our own Milky Way and that the universe consisted just
of our Milky Way. Curtis on the other hand argued that they were other galaxies
just like the Milky Way, but much further away. Although the debate laid open
the positions of the two sides, nothing was immediately resolved. (That same year,
Johannes Kapteyn was arguing that the Sun was in the center of a small Milky Way,
based on star counts.) It was only in the following decade that as Edwin Hubble
and other astronomers found novae and Cepheid variable stars in nearby galaxies,
that Curtis’ view was slowly adopted. (When a letter from Hubble describing the
period-luminosity relation for Cepheids in M 31 arrived at Shapley’s office, Shapley
held out the letter and said, “Here is the letter that destroyed my universe!”)
Hubble found was particular, since it also depended on distance. Galaxies twice as
far away moved twice as fast from us; those three times as far traveled three times as
fast, and so on. These observations can be summarized by “Hubble’s Law:”
v = H◦ d,
which relates the velocity v of a galaxy to its distance d by a scale factor, H◦ , or the
Hubble constant. Hubble worked out a value for the eponymous constant to be
H◦ = 500 km s−1 Mpc−1 , nearly 10 times the current accepted value.
rejected it, we shall see the cosmological constant, or Λ crop up again later in the
story.)
In 1927, the Belgian astronomer Georges Lemaı̂tre (1894–1966) independently
arrived at Friedmann’s solutions to Einstein’s equations, and realized they must cor-
rectly describe the universe, given Hubble’s recent discoveries. Lemaı̂tre was the first
person to realize that if the universe has been expanding, it must have had a begin-
ning, which he called the “Primitive Atom.” This is the precursor to what is today
known as the “Big Bang.”
Chapter 2
13
14 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
The Hubble class of elliptical galaxies are objects that look elliptical in shape (ob-
viously!), with a bright core and a luminosity that drops off smoothly away from the
center, with little or no clumpiness. The Hubble types range from E0, for ellipticals
that look circular, to E7, for ellipticals that are highly elongated. Note that this
classification is made by the elliptical’s apparent shape. It has nothing to do with its
actual three-dimensional shape, since a long, cigar-shaped object might look circular
with the right perspective. Ellipticals are the most common type of galaxies, making
up more than 60% of all those that are observed. Ellipticals also come in a range
of sizes. The most prominent are the giant elliptical galaxies, which can be 3–4
times larger than the Milky Way. Most ellipticals belong to the faint dwarf ellipti-
cal class, containing only a few million stars. Dwarf galaxies may be undercounted
in galaxy surveys because they are more difficult to find than their brighter cousins.
The stars in elliptical galaxies tend to be old, with no significant star formation in
the last 10 billion years. This makes sense since there is little gas and dust left in
them.
Spiral galaxies are categorized by how tightly the spiral arms wrap around the
bulge, and whether the galaxy has a bar at the center of the spiral. The Hubble types
Sa, Sb, and Sc are for unbarred spirals, with Sa galaxies having the tightest arms and
largest bulges, and Sc galaxies having the loosest wound arms and smallest bulges.
Because of the relationship to bulge size, the exact Hubble type for a spiral galaxy can
be guessed even if the galaxy is nearly edge-on. Barred spirals have the inner parts
of the spiral arms meeting at the ends of a stellar bar. The barred spiral types are
SBa, SBb, and SBc, with SBa tightly wound, and SBc loosely wound. Barred spirals
are only evident if the galaxy is seen face-on. Spiral galaxies are often seen close
to edge-on, such as M 31, the Andromeda Galaxy. However many spiral galaxies,
including our own Milky Way, show evidence of a bar, and it may turn out that
all spirals have bars, with many too small to be seen. And because of the physical
and chemical similarities between barred and unbarred spirals, some extra-galactic
astronomers1 do not even bother differentiating between the two categories.
The disks in spirals have copious gas and dust allowing for ongoing star formation.
As a result, both young and old stars can be found mixed together in the disk. The
lack of gas in the bulge and halo means only older stars are found in those regions.
Lenticular galaxies are similar to spiral galaxies by having a disk and a bulge.
However their disks possess no spiral arms. The bulge may also be relatively large,
compared to the disk, and may contain a bar-like structure as well.
Spirals and lenticulars make up about 30% of all galaxies, with about 60% of each
class having noticeable bars.
Irregular galaxies do not have the clear symmetries or regular shapes found in the
previously described Hubble classes. Some have traces of spiral arms or hints of a
disk. These are the Type I irregulars, which have regions of vigorous star formation,
1
“Extra-galactic” here refers to outside our own Galaxy, i.e., other galaxies. “Galactic” as-
tronomers study our own Milky Way.
2.1. THE EXTRA-GALACTIC ZOO 15
(c) M 99 (Sc)
(a) M 85 (S0)
with vast H ii regions and brilliant OB stars. The Magellanic Clouds are the two
closest examples of this sub-class. Type II irregulars are much more chaotic looking,
often containing odd arms, loops, or explosive filaments. These may be the result of
mergers or collisions between neighboring galaxies. Irregulars are the least common of
the traditional galaxy classes, making up about 15% of all observed galaxies. Finally
there is a small fraction of peculiar galaxies as well, that can not be made to easily
fit any of the Hubble categories.
2
Protons and neutrons fall under the subatomic particle class of baryons, with masses mp =
1.673 × 10−27 kg and mn = 1.675 × 10−27 kg, respectively. Electrons are classified as leptons, but
have a mass of me = 9.109 × 10−31 kg, or roughly 1/1800th the mass of a proton or a neutron.
Therefore ordinary matter is made up almost entirely by mass of baryons, and hence is called
baryonic matter.
22 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
Table 2.1: Population Model of M 31 Bulge: The derived mass and V-band luminosity
fraction of stars within each spectral type range.
Determining the stellar mass of galaxies is not as easy as simply measuring a single
spectral line from the gas. The luminous visible light emission of galaxies is from the
cumulative glow of all of its member stars. Each star will have a different spectrum
depending on its mass. The total light from a galaxy is therefore due to the total
numbers of stars of each spectral type in that galaxy. The method of population
synthesis looks at the spectrum of an entire galaxy. Using assumptions of what
fraction of stars are to be expected in each spectral type, a theoretical spectrum can
be constructed and compared with the actual observed spectrum. The numbers of
stars within each spectral class can be adjusted until the population model gives
a reasonable match to the observed spectrum. Then from the total luminosity of the
galaxy, total numbers of stars in each spectral class can be inferred, and by adding
up the mass associated with each spectral type, the galaxy’s total stellar mass can be
determined.
Table 2.1 shows the results of such a population model calculated for the An-
dromeda Galaxy, M 31. As to be expected, the orange-ish bulge of a spiral galaxy is
dominated by cool, low-mass red stars (the M-dwarfs) which contribute to more than
80% of the total bulge mass, but because they are so faint, only add to a few percent
of the total luminosity. What is visible in the bulges of spiral galaxies (including our
own Milky Way) are the red giant and subgiant stars. K giants are responsible for
2.3. DARK MATTER COMPOSITION OF GALAXIES 25
Finally rotation curves for spiral galaxies can also be constructed by measuring
the Doppler velocities of stars within the disk. Such a rotation curve for the Milky
Way is shown in Fig. 2.11. The shape of the curve looks fundamentally different from
26 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
the previous two examples because although the stars in the Galaxy are not rotating
like a rigid body, they also do not all orbit around the same common center of mass
as do the planets around the Sun. Instead a star orbits around the collective mass of
everything inside its orbit. The further a star is from the center of the Milky Way,
the more mass its orbit encloses.
The velocities of stars within much of the disk are consistent with the observed
luminous mass of the Galaxy. If the visible edge of the Galactic disk was indeed where
there was a sharp drop-off in mass, then one would expect the rotation curve for stars
beyond the edge to drop like that for planetary systems. However what is actually
observed are orbital velocities that stay constant with distance rather than falling off
(Fig. 2.12). This suggests that in addition to the visible matter in the outer reaches
of a spiral galaxy, there is also a dark matter halo, consisting of matter that has
not been observed in any (including non-visible) wavelengths. Although the visible
edge of the Milky Way’s disk is located about 50,000 light years from its center, dark
matter extends out at least to 65,000 light years. Depending on the assumptions used
in setting the boundaries of the Galaxy and the distributions of the dark matter, the
total mass of our Milky Way, including the dark matter, might be anywhere from
4–60 times the luminous mass from stars. That is, given the total luminous mass of
the Milky Way is 1011 M , then the luminous plus dark matter mass can be anywhere
from 4 × 1011 M to 6 × 1012 M .
2.3. DARK MATTER COMPOSITION OF GALAXIES 27
As clusters go, the Local Group is rather sparse and small. Other clusters are
30 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
far richer, meaning not only do they have larger memberships (up to thousands of
observable galaxies), but they are also more densely packed. In fact the typical rich
cluster is comparable in diameter to the Local Group, but there are far more galaxies
at the cluster cores. Two well-known clusters are the Coma cluster and the Virgo
cluster. Each has over a thousand known members.3 The Coma cluster is about
250–300 million light years from the Milky Way, and is a spherically shaped cluster
consisting mainly of elliptical and lenticular galaxies. The Virgo cluster is closer, at
slightly more than 50 million light years, and is irregular in shape, with a mix of
spirals as well as ellipticals.
At larger scales, even larger superclusters of galaxies can be seen. The Local
Supercluster is centered on the Virgo cluster, includes the Local Group, and is
roughly 100 million light years across. The galaxy clusters that make up the super-
cluster are not gravitationally bound to each other, so the boundaries of a supercluster
depend on the cutoff number density that separates regions of high galaxy count ver-
sus low galaxy count. At larger scales of 800 million light years, other superclusters
are evident.
At the largest scales, superclusters appear to be organized in a vast network
of sheets and filaments that surround nearly-empty, great voids. The large-scale
structure of the Universe therefore appears to consist of chains of superclusters of
galaxies, arranged around voids, of order 600 million light years across, like the pores
inside sponges. The largest structures appear to have a scale of about 600 million
light years. At even larger scales, the Universe finally starts to appear uniform or
homogeneous—because any one 600 million light year sized region will look like any
other 600 million light year patch.
L = d × θ,
where θ is the angular size that we measure for the galaxy in the sky in units of
radians.4 Once we know its true distance, we can calibrate the flux received from the
3
Remember that these are observed galaxies. Many galaxies, such as dwarf ellipticals, are below
the current observable threshold, and therefore would not be counted in surveys.
4
This relationship holds as long as the angular size θ is small, which is true for all observable
external galaxies.
2.5. THE COSMIC DISTANCE LADDER 39
galaxy by our instruments and determine its true light output. A distance measure-
ment allows us to fix not only the size of the galaxy, but many other derived physical
quantities that depend on the size.
It turns out that there are many different methods for determining distances,
and new methods are being devised and old methods are being revised constantly.
Each distinct methodology has its own advantages and disadvantages. When quoting
a distance using a particular technique, the careful astronomer is always aware of
the errors and uncertainties associated with that technique. It is important to keep
these errors in mind, since any one particular distance measurement method is good
over a limited range. A nearby distance measurement technique is used to calibrate
a second distance method that overlaps it slightly. This second technique is then
used to calibrate a third method that works on objects yet more distant, and so
on. Uncertainties associated with one method can propagate to other methods if the
former is used as the calibration.
d = R × p, (2.1)
where p, measured in radians, is the parallax angle or half of the total angular shift that
we see the foreground star make, and R is the Sun-Earth distance (or 1 astronomical
unit). This equation also shows us where the term parsec (“parallax-second”) arises.
An object is at a distance of 1 parsec if it has a parallax p = 100 :
1 AU
1 parsec = = 3.0856 × 1018 cm
(2π/1.296 × 106 arcsec)
40 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
where λobs is the observed wavelength of the line, and λem is the original emitted
wavelength. Assuming a linear relationship between the redshift and distance, we
2.5. THE COSMIC DISTANCE LADDER 47
get:
H◦
z= d, (2.3)
c
where c is the speed of light (2.9979 × 108 m sec−1 ) and H◦ is the Hubble constant.
Rewriting Eq. 2.3 as cz = H◦ d, and assuming that cz is the same as the veloc-
ity (which is correct when the redshift is small), we can use Eq. 2.3 to give us a
relationship between velocity and distance:
v = H◦ d. (2.4)
48 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
This equation states that the recessional speed of a galaxy is exactly proportional to
its distance. This is what would be expected from a uniformly expanding Universe.
Thus if the Hubble constant, H◦ , was known accurately, one can use this equation to
determine the distance to a galaxy. This is the primary reason why an accurate deter-
mination of H◦ has long been a goal of astronomers. However as we see in Fig. 2.29,
even for galaxies with distances determined via standard candle techniques, there is
enough “noise” in the data to not quite match the smooth Hubble flow predicted by
Eq. 2.4. Some is due to uncertainties and unaccounted-for errors in measuring the
distances. However other uncertainty is due to intrinsic velocity differences that a
galaxy might have. If a galaxy was part of a large, rich cluster and was gravitation-
ally interacting with the other cluster members, its orbital motions could result in
its recessional velocity to be quite different than the overall recessional speed of the
cluster.
and black holes. The white dwarfs and neutron stars would have stopped glowing
after a few billion years. If no matter was available to be accreted, a black hole
would be invisible. Attempts have been made to survey the halo for small, compact
objects via gravitational lensing, and such a population has been found. However the
numbers of these objects are only about 20% of what is necessary to account for all
of the dark matter halo mass.
In addition to this relatively cold, baryonic dark matter, physicists and astronomers
have also suggested that non-baryonic matter could make up the bulk of the dark
matter. This is the so-called hot dark matter component, because such particles
would move at velocities very close to the speed of light. The neutrino has been men-
tioned as a candidate for such a particle. Although neutrinos were thought to have
zero rest mass, and would therefore travel at the speed of light, physicists realized
that there was no fundamental reason why they should be massless. Therefore if
neutrinos did have a very small mass, they could provide some fraction of the dark
matter mass by sheer numbers. Recent experiments involving neutrinos from the Sun
showed that they have a mass about 5 million times less massive than the electron.
This however is not enough; the neutrinos would constitute on order only 1% of all
dark matter.
Physicists have also suggested WIMPs, the weakly interacting massive par-
ticles as a dark matter candidate. The “weak” in the name refers to one of the four
fundamental forces in the Universe that affects such particles.5 According to proposed
supersymmetry theories of particle physics, relationships between fundamental par-
ticles and the forces of nature implies the existence of new classes of undiscovered
particles which interact only by gravity and the weak force. One such postulated
particle, the neutralino, has a mass 20–1000 times that of a proton. It is expected
that if supersymmetry is correct, many neutralinos would have been created in the
early universe. Currently there are efforts (unsuccessful so far) to generate artificially
neutralinos in particle accelerators, as well as to detect any cosmic neutralinos that
might be wandering through the Solar System from the Galactic halo.
5
The other three forces are: the strong force which mediates interactions between protons
and neutrons in atomic nuclei, and keeps them bound together; the electromagnetic force which
controls interactions between charged particles; and gravity. See more on this in § 8.3.
52 CHAPTER 2. WHAT IS IN THE OBSERVABLE UNIVERSE?
Chapter 3
Theoretical Universes
To understand how the Universe got to its current state, we need a basic under-
standing of some of the basic physical laws that are most important to its evolution.
The first of these are Einstein’s Special Theory of Relativity and General Theory of
Relativity. In deriving these theories, Einstein showed how space and time could be
radically re-thought.
Before Einstein’s theories, space and time merely contained matter and energy.
Every particle of matter and each bit of energy could be located within a single
position in space and at a single moment of time. Furthermore, space and time were
passive: they made up the stage in which matter and energy interacted, but played
no part otherwise in the Universe. Einstein showed that the individual dimensions of
space and time were really components of a single space-time, and it was impossible
to disentangle space completely from time. Furthermore matter and energy affected
the geometry of this four-dimensional space-time, and the shape of space-time could
affect matter and energy in return. This radical restructuring of the fabric of reality
also showed how gravity became perhaps the most important force in shaping the
evolution of the Universe.
According to Newton, gravity was a force between masses, which acted instantly
across intervening space. The Sun exerted a gravitational force on the Earth, and
the Earth also exerted a far weaker gravitational force (because of its much smaller
mass) on the Sun. Newton’s law of universal gravitation can be described by:
m1 m2
F =G , (3.1)
r2
where the force F between two bodies is given by the product of the Gravitational
constant G and the masses of the two bodies, m1 and m2 , divided by the square of
the distance r between the two. Newton’s laws of gravity and motion were highly
successful for more than two centuries in explaining the motions of Solar System
bodies. However Newton could not explain the origin of gravity; his laws could only
describe accurately the motions.
53
54 CHAPTER 3. THEORETICAL UNIVERSES
However Einstein also showed from his Special Theory of Relativity that matter
and energy were equivalent and convertible so that one could turn one into the other.
The maximum amount of energy that can result from a 100% perfect conversion from
some lump of matter is given by perhaps the most famous equation in all of science:
E = mc2 . (3.2)
This equivalence suggests that matter is not solely responsible for distorting space-
time. In fact the distribution of all energy and momentum throughout a four-
55
dimensional volume of space-time will have an effect on the local curvature of that
volume. Since the motion due to gravitational forces can be described as being caused
by the curvature of space-time, then gravity is due not only to matter, but also energy
and momentum as well. The energy and momentum of matter and radiation cause
space-time to warp, and the warped space-time tells how the matter and radiation
move through space-time.
Einstein published his General Theory of Relativity in 1916. The prediction that
energy in the form of electromagnetic radiation would be affected by the curvature
of space-time was tested only a few years later. Observations of stars during a total
eclipse of the Sun in 1919 showed them to be exactly where General Relativity pre-
dicted they would be. This validation of Einstein’s theory made him internationally
famous.
56 CHAPTER 3. THEORETICAL UNIVERSES
c 2 = a2 + b 2 , (3.3)
where c is the length of the hypotenuse of the right triangle, and a and b are the
lengths of the other two sides. One can generalize the Pythagorean theorem to three-
dimensions as well:
c 2 = a2 + b 2 + c 2 , (3.4)
For a mathematical geometrician, these two equations are all that is required to com-
pletely describe the surface. Such flat geometry is suitable for architects, surveyors,
and most everything that we deal with in our everyday life.
However this is not the only geometry that is possible. Although the flat Euclidean
geometry is familiar to us, it is not a completely accurate description of our world.
The Earth is not flat, but spherical. But because the Earth is so large relative to
the scales and distances that we are used to normally, a small patch of its surface
will look extremely flat to us. But on a larger scale, expanding to a size where the
3.1. THE CURVATURE OF SPACE-TIME 57
curvature of the Earth becomes important, our geometry is now curved—in fact it is
positively curved. Here lines that start off parallel eventually meet. A circle with
radius r will have a circumference C < 2πr—for a flat geometry, recall that circles
have circumferences C = 2πr. The interior angles of a triangle add up to be a value
greater than 180◦ . (In fact it is possible to draw a triangle on a sphere with each
angle being a right angle of 90◦ !)
In the 19th century, mathematicians Karl Friedrich Gauss (1777–1855) and Bern-
hard Riemann (1826–1866) generalized geometry so that it was not restricted to just
flat surfaces where Euclidean rules applied. To derive rules for these other geometries,
Gauss assumed as a starting axiom that the Pythagorean theorem, Eqs. 3.3 and 3.4,
would be true for any geometry as long as the triangle you studied was small enough.
So even for a highly curved, non-flat surface, one could shrink down and look at a
tiny, local part of that surface that would look flat. A triangle drawn on that would
still follow the Pythagorean theorem. This triangle would have sides of length dx, dy,
and ds (where the d prefix comes from calculus and is used to denote variables that
are very small), and they would be related by:
(ds)2 = (dx)2 + (dy)2 . (3.5)
Again in a three-dimensional space, we have the equivalent equation
(ds)2 = (dx)2 + (dy)2 + (dz)2 . (3.6)
Eq. 3.5 would apply to a flat sheet that extends infinitely in all directions. If however
you were on the surface of the Earth which only looks locally flat, then one could also
write an equation similar to 3.5, but only more complicated to completely describe
the curved geometry.
Similarly one can use Eq. 3.6 to completely describe a three-dimensional space
that is flat. But if that three-dimensional space is curved, then you would need a
variant of Eq. 3.6 that was similarly much more complicated. Never mind how you
can imagine a three-dimensional space that is curved—it’s not possible so don’t even
try! Just realize that it can be mathematically described.
Actually one can understand this by looking at a two-dimensional analogy. Imag-
ine a two-dimensional universe filled with 2D creatures. If their universe was flat, then
they would appear to us—3D beings—as living on an infinite flat sheet. However if
their universe was actually positively curved, then we might see their universe as
curved like a sphere. However it would be impossible for the 2D creatures to visualize
how their 2D universe could be positively curved, although they would be able to use
mathematics to describe it.
In addition to a positively curved sphere, there can also be negatively curved
surface as well. A negatively curved two-dimensional surface is usually represented
by a saddle shape. The rules for geometry are now the opposite of what they were in
the positively curved case. Parallel lines always diverge. Circles have circumferences
C > 2πr. Triangles have interior angles which add up to less than 180◦ .
58 CHAPTER 3. THEORETICAL UNIVERSES
When constructing his Theory of Relativity, Einstein took this previous work
on generalizing geometry and instead of just dealing with three-dimensional space,
he expanded it to four-dimensional space-time. Instead of measuring the physical
distance of points in three-dimensional spaces, Einstein’s theory dealt with the space-
time separation of events. Einstein’s version of Eq. 3.6 for flat space-time is:
(ds)2 = (dx)2 + (dy)2 + (dz)2 − c2 (dt)2 , (3.7)
where c is the speed of light and dt is the time separation of the two events. If space-
time was curved in either the positive or negative direction, then there would be a
corresponding equation that would also be much more complicated.
appeared mostly empty, with an occasional galaxy as an island of stars in the sea
of space. This lumpiness certainly does not look smoothly distributed as would be
required by homogeneity. We now have surveys (Figs. 2.19 and 2.20) that show the
Universe to look uniform and isotropic at large enough scales. Early in the 20th
century, it was a leap of faith to assume that this was true given the state of the
observational data at the time.
However cosmologists working with Einstein’s General Theory of Relativity found
themselves facing an extremely complicated set of equations. The theoretical cos-
mologists had to make whatever simplifications they could, in order to get tractable
solutions from Einstein’s equations. They therefore modeled the Universe as com-
pletely filled by a uniform gas or fluid. (And at large enough scales, superclusters
of galaxies will look uniform enough to be described by this “uniform fluid.”) From
these simple model assumptions, a cosmologist can now describe properties of the
gas by its density ρ (Greek letter ‘rho’). Since the fluid is uniform everywhere, the
pressure and density is the same everywhere. However if the Universe is expanding
or collapsing, the density will change with time t, which we can express as ρ(t).
Given such a simplified model, early theoretical cosmologists could now use Ein-
stein’s equations to predict the behavior of the Universe over time. Einstein’s field
equations are represented by:
T,
G + Λ = 8πT (3.8)
where as we noted back in § 1.5 that Λ represents the cosmological constant. We can
now also state that G is a mathematical construct called a tensor that describes the
curvature of space, and T is another tensor that describes the distribution of energy
and momentum in space.
3.3. MODELING THE UNIVERSE 61
However Eq. 3.8 is far more complicated than Newton’s universal law of gravity
(Eq. 3.1). Using Newton’s equation is difficult enough, since you have to apply the
effect of gravity of all of particles in your system on all of the other particles. This
can be quite a burden—though not impossible using computers—to calculate for re-
alistic descriptions of physical systems. But this is simple compared with Einstein’s
equation, which is actually not one, but a family of ten equations that have to be all
solved simultaneously. Furthermore as noted before, matter and energy are equiva-
lent, so both matter and energy create gravitational fields. But a gravitational field
is also a form of energy, so that also provides an additional warpage to space. This
non-linearity of the equations means that there are not that many instances where
one can come up with exact solutions to the field equations!
where k is the curvature parameter, r is the distance of the two events from the origin
of the coordinate system, and R(t) is a scale-factor that is time-dependent (hence the
“(t)”). For descriptions of either finite or infinite universes, it is almost meaningless
to talk about the “radius” of a universe. However one can define a coordinate system
that is attached to space-time, and can expand or contract with it. This co-moving
coordinate system will change its scale, and the distances between matter attached
to the coordinate system, by the scale factor R(t).
To determine the evolution of a model universe, the cosmic scale-factor R(t) must
be determined by solving something known as Friedmann’s equation:
8πG 2 Λ 2
Ṙ2 = ρR + R − k c2 . (3.10)
3 3
The solution will depend on the value of k, the density ρ, and the cosmological
constant Λ. Where this equation comes from and its solution is beyond the scope
3.3. MODELING THE UNIVERSE 63
of this course but the results are shown in Fig. 3.8, which show how the scale-factor
changes over time.
Again it is important to re-emphasize that only the positive curvature, k = +1
universes are finite. All the other k = 0, −1 universes are infinite in extent. In
any case, R(t) is not the radius or curvature of a universe, which is meaningless.
It is instead a scale-factor that tells you how a coordinate system attached to the
expanding or contracting space-time changes over time.
Note that in most of the model universes, running time back to t = 0 results in
R = 0 as well. This means that all of these universes have a beginning in a big
bang. Some universes return back to a state where the scale-factor is 0, R = 0, but
at a future time. These universes suffer a big crunch. Other universes continue
expanding forever without ever contracting again.
Λ < 0 Universes
For universes with negative Λ, the cosmological constant causes all the universes,
regardless of curvature to stop expanding and collapse back on themselves in big
crunches.
Λ = 0 Universes
Ignoring the cosmological constant or letting it be very small means effectively Λ = 0.
Until recently, it was thought that such models best described the Universe. If so, then
there are some general statements that we can make about the Universe depending
on what the curvature:
1. k = 00: The simplest case is the flat universe with zero curvature. But if
there is matter and energy in the universe, then why do they not warp space into
a non-flat curvature? It turns out that the curvature from the matter-energy
in the universe can be negated if the universe is also expanding. The energy
from the expansion of the universe will create a curvature that can cancel the
curvature from the matter and energy that is already in the universe. Note
that the expansion has to balance out the matter-energy density in the universe
exactly for the universe to stay flat. Hence this solution is also known as the
critical case. Since gravity from the mass-energy is negated and cannot slow
or stop the expansion, a flat universe will expand forever.
A flat universe also has several other properties. A pair of light beams that
start out parallel will stay parallel forever.
Also a flat universe is infinite in extent. You can travel in any direction forever
without coming back to a part of the universe that you have visited before.
2. k = +1
+1: If the matter-energy density dominates, then the universe has a posi-
tive curvature and is closed. The gravity of the mass-energy warps space-time
66 CHAPTER 3. THEORETICAL UNIVERSES
into a analog of the surface of a sphere. For a flat universe, the expansion will
eventually slow and the universe will collapse back on itself.
If you send out two laser beams that start off parallel, the beams will eventually
converge and cross in a closed universe.
Finally a closed universe is also finite in size. If you travel indefinitely in one
direction, you will not hit a wall or run into the edge of the universe, but you
will eventually arrive back at your starting point.
3. k = −1
−1: If the universe has negative curvature—for instance, because the
expansion dominates over the positive curvature from the matter-energy—then
the universe is open, and the universe will expand forever.
Light beams that start off parallel will diverge.
An open universe, like the flat universe, is also infinite in extent.
Λ > 0 Universes
We first consider the k = +1 models together:
2. 0 < Λ < ΛE : If the cosmological constant is greater than zero but less than ΛE ,
then for k = +1, there are two possible behaviors. One is the familiar big bang
at time zero followed by a big crunch (the bottom line in the plot). The second
line in the plot however does not have a definite t = 0. The universe contracts
from some infinite time in the past, but rebounds before it reaches R = 0, and
starts expanding again.
3. Λ > ΛE : Finally for the case where the cosmological constant is greater than
for Einstein’s static universe, the k = +1 scenario is known as the Lemaı̂tre
model, advocated by the Belgian cosmologist and priest Georges Lemaı̂tre
(1894–1966). After the initial big bang, the Lemaı̂tre universe’s scale-factor
evolves to the “flat” portion of the graph where it stays virtually static and
doesn’t grow very much. This was popular for a time in the 1930s when it was
3.4. COSMOLOGICAL REDSHIFTS AND THE HUBBLE CONSTANT 67
thought that the early phase after the big bang could be the time when the
chemical elements were created, while the “coasting” phase was when stars and
galaxies had time to form.
The k = 0, Λ > 0 model is right now thought to be the best description we have for
our Universe. We will see in later chapters evidence for why this is so. This universe
model starts off with a big bang and continues to expand, accelerating forever. There
is a slight slowdown, but not as long as the Lemaı̂tre model’s “coasting” phase.
Finally the k = −1, Λ > 0 model results in an infinitely large universe, that again
expands forever, with the expansion accelerating over time.
Ṙ(t)
H(t) = (3.13)
R(t)
where H(t) is a time-dependent version of the Hubble constant and Ṙ(t) represents the
rate of change of the scale-factor over time—you can think of it relating to the velocity
68 CHAPTER 3. THEORETICAL UNIVERSES
of the expansion. This means that the Hubble’s constant is not really constant. As
the Universe grows and expands, its value changes. The present day value of the
Hubble constant, at time t◦ after the Big Bang, is given by:
Ṙ(t◦ )
H(t◦ ) = H◦ = . (3.14)
R(t◦ )
The Hubble constant that we measure today is therefore the fractional rate of change
of the scale-factor R(t) at the present time t◦ .
If the density in the Universe matches the critical density, then Ω = 1. If the Universe
over-dense which will eventually lead to a big crunch, then Ω > 1. And if there is
less actual density than the critical density, then Ω < 1.
We can also define a density parameter for just the matter (which includes the
baryonic and the dark matter):
ρm (t)
Ωm (t) = . (3.15)
ρcrit (t)
If the density of matter at some time in the Universe was one-third of the critical
density, then Ωm (t) = 1/3; if it was one-half of the critical density, then Ωm (t) = 1/2..
Notice that in Friedmann’s equation (Eq. 3.10), the cosmological constant plays a
similar role as the density ρ. This suggests that we can also define a density parameter
for Λ as well:
ρΛ
ΩΛ (t) = . (3.16)
ρcrit (t)
70 CHAPTER 3. THEORETICAL UNIVERSES
The density ρΛ can be thought of as the energy density of vacuum or of space. Unlike
the matter density which decreases as space expands, the energy density of vacuum
stays constant. But if space expanded and increased in volume, then the total amount
of vacuum energy will increase since there is more vacuum which leads to a greater
total vacuum energy!
As we will see later, there is growing observational evidence that the Universe does
contain a cosmological constant-like energy. Since there are many hypotheses that
have been proposed to explain it, with the cosmological constant being only one of
the possibilities, it has been generically called dark energy. You will therefore often
see in the literature references to ΩΛ as the density parameter for dark energy, not for
the cosmological constant (although Λ did originally refer only to the cosmological
constant).
The density parameters Ωm and ΩΛ can be related directly to the curvature pa-
rameter k:
< 1, k = −1;
Ωm + ΩΛ = 1, k = 0; (3.17)
> 1, k = +1.
Therefore whether the Universe is closed, flat, or open can be determined by mea-
suring Ωm and ΩΛ . All the possible different scenarios for different combinations of
Ωm and ΩΛ is given in Fig. 3.10. The dashed line represents all possibilities where
ΩΛ + Ωm = 1, and k = 0. Above the line is where k is positive; below the line has a
negative k.
So what are the values of ΩΛ and Ωm ? The best measurements imply that ΩΛ +
Ωm = 1, while,
ΩΛ ≈ 0.7, (3.18)
Ωm ≈ 0.3.
Thus the Universe is flat, and will continue to expand and accelerate in its expansion
forever.
If we lived in such a universe (which we probably do not, since all recent observations
point to Λ 6= 0), then we would be able to determine how old the universe is by
3.6. THE AGE OF THE UNIVERSE 71
Other relationships between the Hubble constant and the age of the universe can
be found for the other FRW models. However the forms of the equations are much
more complicated than 3.20. How the scale-factor evolves with time t for several
different FRW cases is plotted in Fig. 3.11. The present time is represented by t0 .
The Hubble constant is the same for each model universe, but the total elapsed time
from the big bang to the present ends up being different.
The closed universe with k = +1 and Λ = 0 has the shortest age, while the critical
universe with k = 0 and Λ = 0 is older, and the open universe with k = −1 and Λ = 0
3.6. THE AGE OF THE UNIVERSE 73
is even older still. The oldest is the model with Λ > 0 and k = 0, which is a universe
that accelerates as it expands.
Cosmologists have calculated possible ages for a range of different cosmological
parameters. Given our best measurements of Ωm and ΩΛ (Eqs 3.19), the age of our
Universe is just shy of one Hubble time, or t◦ = 13.7 billion years old.
74 CHAPTER 3. THEORETICAL UNIVERSES
Chapter 4
If the Universe is expanding, then if we ran the movie of the Universe backward, the
volume of space-time would decrease the further back we went in time. The density of
matter and radiation would increase, and so would the temperatures and pressures.
From basic physics cosmologists can determine what the physical conditions must
have been like in the early Universe, while from knowledge of nuclear and particle
physics, they can also determine what sort of processes must have occurred in the
past.
γ + γ
e− + e+ (4.1)
Within this sea of radiation, electrons, and positrons, the following reactions were
reversible (and hence the arrows pointing left and right):
e− + p
n + ν
ν̄ + p
n + e+ (4.2)
n
p + e− + ν̄,
Thus as long as the temperature was hot enough, protons were being converted into
neutrons, but there also existed a process which converted neutrons back into protons.
As the temperature dropped below 30 billion K, about 2 seconds after the Big
Bang, primordial nucleosynthesis became possible: the creation of atomic nuclei
75
76 CHAPTER 4. THE BIG BANG
in the aftermath of the Big Bang. The important difference between this and the
stellar nucleosynthesis that is still occurring today inside stars is that the conditions
after the Big Bang were constantly changing. As the Universe grew larger after the
Big Bang, the pressures and temperatures dropped. The intense initial temperatures
meant however the nuclear reactions occurred at rates much higher than that found
inside stars.
The first important nuclear reaction created deuterium or 21 H:
p + n
21 H + γ.
Note that this reaction is again reversible, so that sufficiently energetic photon can de-
stroy the deuterium nucleus. However once the temperature of the Universe dropped
below 1 billion K, the average photon energies dropped to the point where quantities
of deuterium could build up. At about t ≈ 0.7 seconds after the Big Bang, the ratio
of neutrons to protons “freezes out” at:
n
= 0.22. (4.3)
p
Free neutrons have a decay half-life of time of tN = 10.5 minutes, so the number of
neutrons will decrease a little bit more. However after the temperature drops well
below 1 billion K, or about 170 seconds after the Big Bang, the photons are no longer
energetic enough to destroy deuterium nuclei. The electron-positron pairs all end up
annihilating, and the neutron-proton ratio ends up at:
n
≈ 0.14. (4.4)
p
At this point the deuterium abundance has grown large enough for the deuterons
to produce helium, via the sequence of reactions:
d+d ↔ t+p
d+d ↔ 32 He + n
t+d ↔ 32 He + n
3
2 He + d ↔ 42 He + p, (4.5)
where t is tritium or 31 H.
Essentially most of the neutrons wind up in Helium-4 (42 He). Based on the final
number of neutrons after they freeze out (Eq. 4.4), knowing that the number of 42 He
nuclei is half the number of neutrons, and ignoring any other atomic nuclei (other
than Hydrogen), then the mass fraction of 42 He is:
4n4 He
x4 He ≈
nN
4(nn /2) 2(n/p)
= = (4.6)
nn + np 1 + (n/p)
' 0.25,
4.2. THE COSMIC MICROWAVE BACKGROUND 77
After about 1000 seconds (17 minutes) after the Big Bang, the temperature was below
500 million K and all element creation ended.
There are no known processes in the present Universe that can create substantial
amounts of deuterium; stars can only destroy deuterium as a fuel source in reactions
at their cores. All of the deuterium today is the result of these primordial reactions.
As a result, there are many studies underway today to determine the exact amount
of present day deuterium in order to understand the conditions of the early Universe.
Observations to measure the abundance of deuterium, usually expressed as the D/H
ratio, has been made for Solar System objects and in the local interstellar medium
(ISM). For instance, the best Solar System measurement comes from the atmosphere
of Jupiter, with D/H ' 1–4 × 10−5 , which is similar to values from the local ISM.
More sophisticated modeling of the nucleosynthesis period in the early Universe
showed that the total amount of 42 He created should be about 24% (still not too far off
from our simplistic result in Eq. 4.6. This agrees remarkably well with observations
of not just interstellar gas not polluted by heavy element enrichment from stars, but
with the rough proportion of observed helium abundance everywhere in the Universe.
That the Big Bang (and not any other model) can predict elemental abundances so
close to what is actually observed is one of its underlying strengths as a theory.
freely, instead of interacting with the electrons and nuclei. Because the photons could
no longer easily share their energy with other particles, they became decoupled from
the matter. The matter and photons now evolved separately from each other.
One consequence of the cosmological expansion of the Universe is the cosmologi-
cal redshift of photons, where their wavelengths are stretched out by the expansion
of space-time (see § 3.4). Because the photons were originally coupled to the dense
hot matter soup (containing protons, electrons, and heavier nuclei), they retained a
blackbody spectrum. A blackbody spectrum originates whenever a source perfectly
absorbs and perfectly re-emits radiation, which aptly describes the early Universe
since the photons are continually absorbed and re-emitted by the matter. Another
term for this radiation is thermal radiation, since the perfect absorption and emis-
sion brings the source and the radiation into thermal equilibrium. Blackbody spectra
have a unique shape as we shall see below.
After decoupling, the blackbody spectrum continued to red-shift to lower tem-
peratures and longer wavelengths. From a temperature of 3000–4500 K at a time
300,000–400,000 years after the Big Bang, the radiation has cooled from the infrared
to the microwave portion of the electromagnetic spectrum today. Today this radiation
makes up the cosmic microwave background (CMB), with a temperature just a
few degrees above absolute zero (equivalent to its emission mostly in the microwaves).
The CMB radiation consists of photons coming from the last-scattering sur-
face. These photons last encountered another matter particle when the Universe was
opaque, just before recombination. Immediately after recombination, they were able
to free-stream through space to reach our detectors. Although this last-scattering
surface appears as a spherical shell (like the celestial sphere) that surrounds us, this
does not mean we are at the center of the Universe. A useful analogy is if you go out
walking on a foggy day and there is a visibility of 50 feet: light from objects does
not reach you from beyond a distance of 50 feet. Instead the photons are scattered
by the suspended water droplets that make up the fog. As you walk around this
“foggy universe,” it will feel as if the universe has a radius of 50 feet, defined by
how far you can see. Another observer taking a separate walk outside will see a
different last-scattering surface. Each observer will have her own observable universe
that is slightly different than any other observer’s. Thus there is a limit to how far
we can observe in the Universe! The Universe might be much larger than our visible
horizon as defined by the CMB, but there is no way for us to detect it.
The CMB is the third observational triumph that supports the Big Bang theory.
It was predicted by theoretical calculations in the 1940s, and was first detected in
1965 by Arno Penzias and Robert Wilson of Bell Laboratories. They were using
a horn-shaped microwave antenna to study microwave emission from the Earth’s
atmosphere, in order to identify sources of interference for satellite communications.
However Penzias and Wilson discovered a uniform background source of noise that did
not vary by the time of day, which suggested that it did not originate from the Earth.
It was highly uniform across the sky, suggesting it was also not galactic in nature
80 CHAPTER 4. THE BIG BANG
has never observed in a laboratory. However since the matter creation rate required
was only about one atom of Hydrogen per cubic meter of space over the age of the
universe, this postulated effect could not be absolutely ruled out.
The discovery of the CMB in 1965 was perhaps the strongest piece of evidence
that killed the Steady State model, since it had no explanation for such a highly
uniform blackbody emission. From the Perfect Cosmological Principle, the Steady
State model also predicted that the Universe should look the same no matter how
deeply in space (and hence how far back in time since light has a finite speed) one
observed. The discovery of quasars in 1966 provided more evidence that directly
contradicted Steady State predictions, since these ultra-luminous compact objects
did not have any counterparts in the nearby Universe.
Chapter 5
After Hubble’s discovery of the expansion of the Universe, the cosmological constant
Λ was left out of Einstein’s Field Equation when describing our Universe. It felt
somewhat ad hoc to include it in, since its value had to be finely tuned to balance out
the rest of the matter-energy in the Universe to keep the Universe absolutely static.
For decades, cosmologists assumed Λ = 0 mostly as deference to Occam’s Razor.
There was no evidence that a cosmological constant was needed, so it was left out of
serious models of the Universe. Cosmologists took it for granted that the Universe
was gradually slowing due to the gravity of the matter-energy that it contained. The
only question was whether the expansion would continue forever or eventually slow,
stop, and reverse itself. However starting in the mid-1990s, work would be done that
would completely change this view.
Ṙ(t◦ )
H◦ = , (5.1)
R(t◦ )
where t◦ is the present time, R is the scale-factor, and Ṙ is the rate of change in
the scale factor. The Hubble constant can be thought of as a indicator of the speed
of expansion of the Universe. However as the above equation shows, it is also not
necessarily constant since it can depend on time; its value now at time t◦ might not be
what it is in the past, or what it will be in the future. The Hubble constant H◦ should
therefore really be thought of the present-day value of the Hubble parameter H(t):
H◦ = H(t◦ ). (5.2)
One way to think of how the current value of the Hubble parameter relates to
the expansion of the Universe is to look at Fig. 5.1 which shows the evolution of the
85
86 CHAPTER 5. THE ACCELERATING UNIVERSE
be indicated at R(t) = 0, or where the gray line intersects the horizontal time axis.
An accelerating universe would be older than the constantly expanding universe, since
in the past it was expanding at a slower rate than indicated by the current value of
H◦ . A decelerating universe was expanding faster in the past—it had a larger value
for the Hubble parameter—and so is younger than either of the other two universes.
In all three of these universes, the value of the Hubble constant H◦ tells you
about the rate of expansion at the current time t◦ . As long as galaxies observed
are relatively nearby, they will follow the expansion of space-time given by H◦ . The
redshift z, distance d, and Hubble constant H◦ are therefore related simply by
H◦
z= d, (5.3)
c
or re-writing to give the distance:
cz
d= . (5.4)
H◦
In our universe, this linear relationship between the redshift and distance works up to
about z = 0.1–0.2. For instance, Hubble’s 1929 figure showing the measured distances
and velocities of galaxies (Fig. 5.2) can be fit by a straight line using an equation of
the same form as Eq. 5.4. Similarly data from measuring galaxies at distances several
hundred times further can be plotted as shown in Fig. 2.29 where the farthest galaxies,
receding at velocities 2–3 × 104 km s−1 , have redshifts of z < ∼ 0.1. In this case, the
data is also fit well by a straight line of the form of Eq. 5.4.
However if we observe more distant galaxies (which are galaxies existing at a much
earlier time), then a different value of H might be at work, since the universe is likely
to be expanding at a faster or slower rate. A version of Eq. 5.4 that includes a current
value of the deceleration parameter q◦ can be given as:
cz h 1 i
d= 1 + (1 − q◦ )z . (5.5)
H◦ 2
The measured value of the deceleration parameter q◦ is merely the current-day value of
the time-variable q(t). Its relationship to the scale-factor for a Friedmann-Robertson-
Walker universe can be shown to be:
R(t)
q(t) = − R̈(t), (5.6)
[Ṙ(t)]2
where R̈(t) (“R double dot at time t”) is a measure of the acceleration or deceleration
of the scale-factor. Since there is a minus sign (−) in Eq. 5.6, a negative deceleration
parameter means the universe is accelerating, while a positive value for q means it is
decelerating.
For many decades, cosmologists (notably Allan Sandage) have focused their efforts
on determining precisely the value of q◦ . The deceleration parameter is linked to
88 CHAPTER 5. THE ACCELERATING UNIVERSE
the ultimate fate of the Universe, whether it will continue to expand, or halt and
recollapse. However to determine q◦ , much more distant, high redshift galaxies must
be observed and their distances need to be measured to see the deviation from the
simple Hubble relation. Nearby galaxies will show a simple, straight relationship
between distance and z (see Fig. 5.3) that is based on the current value of the Hubble
parameter, H◦ .
Attempts to measure q◦ began in the late 1940s. George Abell focused on very
rich galaxy clusters that he was discovering—the so-called “Abell clusters” that were
named after him. He could not observe individual stars in these distant clusters
to identify Cepheid variables, so he assumed that statistically, the 10th brightest in
each galaxy had roughly the same intrinsic luminosity. These were then assumed to
be standard candles, which Abell used to determine approximate distances to the
clusters.
Work following along these lines led to a range of values for q◦ appearing in the
astronomical literature, with no clear agreement. Later work showed that there was
90 CHAPTER 5. THE ACCELERATING UNIVERSE
substantial evolution among galaxies. Galaxies in clusters further away are observed
when they are younger, but also possibly with a very different luminosity than they
would have today. These evolutionary effects made it impossible to use whole galaxies
as standard candles in this fashion.
The supernovae had smaller redshifts than would be expected for their brightnesses,
given calibrations with nearby Type Ia supernovae and Hubble’s law. This implied
that the Universe was accelerating in its expansion, something not predicted by the
Λ = 0 Friedmann-Robertson-Walker models.
This and other corroborating work after 1998 suggests a cosmological constant
(or something like it) is an important component of the Universe. However instead
of referring to it merely as a cosmological constant, cosmologists commonly give this
factor the generic name of dark energy, since Einstein’s cosmological constant is only
one of a number of proposed models for explaining the origins of the acceleration. We
will therefore refer to the symbol Λ interchangeably as dark energy or the cosmological
constant.
From the SCP and HZSNS data, estimates can be made for the value of the density
5.3. MORE ON DARK ENERGY 93
parameters ΩΛ and Ωm . (This was how Fig. 3.10 was generated.) The SCP results
can be used to construct the plot in Fig. 5.6. Plotted are ellipses showing confidence
levels, statistical measures of the uncertainty of the locations of the values for Ωm and
ΩΛ . For instance there is a 99% chance that the true values for Ωm and ΩΛ lie within
the dotted line ellipse. The smaller ellipses more tightly constrain the possible values
of the density parameters, but there is less certainty in these results. This (and plots
like this using combined data from both teams) show that our Universe appears to
be flat, meaning
ΩΛ + Ωm = 1, (5.7)
while the total contributions from dark energy and matter to the density of the
Universe is given by
ΩΛ ≈ 0.7 (5.8)
Ωm ≈ 0.3.
Ωb ≈ 0.02−0.04. (5.9)
That is, the observed number of baryons is only 2–4% of the critical density necessary
for a flat universe.
From several independent methods of accounting for dark matter (some are re-
counted in § 2.6, we find roughly 10 times more dark matter than ordinary baryonic
matter:
Ωd ≈ 0.20−0.30. (5.10)
Eqs. 5.9 and 5.10 are consistent with Ωm ≈ 0.3 in Eq. 5.9.
The discovery of dark energy makes a fascinating denouement to the Copernican
Revolution originally discussed in § 1.4. From the time of the Renaissance and the
beginnings of modern science, we have seen the Copernican principle periodically
manifest itself in our knowledge. The Earth was displaced from the center of the
Solar System and became just another planet that orbited the Sun. Our sense of the
size of the Universe also increase over time. By the early 20th century, the Sun was
further removed from the center of our collective system of stars that made up the
Milky Way Galaxy. Our own Galaxy soon became just another member of a small,
insignificant cluster in an odd corner of a Universe filled with other galaxies. And as
a final insult, we have learned that the baryonic matter that we are composed of is
94 CHAPTER 5. THE ACCELERATING UNIVERSE
only a few percent at most of all matter-energy, and most of the Universe is made of
a bizarre dark energy that we are only beginning to understand!
So what is dark energy? One class of energies that might explain it is vacuum
energy, the energy of empty space devoid of traditional forms of matter-energy.
Although energy is absolutely conserved in classical physics, this is not the case in
quantum mechanics, where energy can appear and disappear out of nowhere, spon-
taneously and unpredictably. Thus even “empty” space is never empty, but is filled
with virtual particles, pairs of a particle and its anti-particle, appearing out of
nowhere and destroying themselves and disappearing before they can be detected.
The amount of energy ∆E available to create these virtual particles and the time ∆t
that they exist follow Heisenberg’s Uncertainty Principle, one form of which is:
h
∆E∆t ∼
> , (5.11)
2π
where h is the Planck constant, h = 6.626 × 10−27 erg sec. Even when there is
not enough energy to create a particle pair, it is “borrowed” for a very brief amount
of time, and after the particles annihilate, the energy “debt” is paid back. The
Uncertainty Principle implies that the more massive the virtual particles, the shorter
lived they will be. Such virtual particles filling space can have a calculable effect on
the energy levels of atoms, and this has been predicted as well as observed as the
Lamb shift. Another phenomenon, the Casimir effect whereby two plates hanging
close together in a vacuum are attracted toward each other, can also be explained by
virtual particles.
Physicists can calculate from first principles what the energy density of the quan-
tum vacuum energy must be. They add up all the contributions from virtual particles
that could arise based on energies from Eq. 5.11. There is still uncertainty what cut-
off one should make in the maximum virtual particle energies. Clearly one cannot
add up virtual particles with near-infinitely large energies. Whatever maximum par-
ticle energy they decide to pick, the resulting answer is much, much larger than what
is observed for the dark energy density. For example, using one typical maximum
cut-off gives a vacuum energy density of 10119 times bigger than observed—that is,
100, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000,
000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000, 000,
000, 000, 000, 000, 000, 000, 000, 000, 000, 000!
This discrepancy between the theoretical value for the vacuum energy density and
the observed dark energy density has been described by physicist Steven Weinberg as
“the worst failure of an order-of-magnitude-estimate in the history of science.” There
must be some additional restrictions on the maximum energy (or perhaps other terms
that cancel out the largest energies) that is not clear to the theoreticians working on
this problem presently. Dark energy may yet turn out to be quantum mechanical
vacuum energy, but a clear theoretical case for it does not yet exist.
96 CHAPTER 5. THE ACCELERATING UNIVERSE
If dark energy cannot be explained quantum mechanically right now, can it still
be Einstein’s cosmological constant? If it is, then dark energy would be true to its
name: it is a constant for any particular volume of space at any time. If your volume
of space doubles in size because of the expansion of space-time, then the amount
of cosmological constant present will also be doubled. However some theoretical
cosmologists think that if the cosmological constant had been present in its current
form from the start of the Big Bang, then the evolution of structure in the Universe
would not lead to what we see today. The cosmological constant would have to be
far weaker to allow for the formation of galaxies and stars that we see.
As a result, many other forms of dark energy have been proposed, including a class
called quintessence. Quintessence is named after the fifth element of the Ancient
Greeks (after earth, water, air, and fire), and which was thought to hold the stars
and planets in place in the sky in their old cosmogonies. The difference between
quintessence and Einstein’s cosmological constant is that quintessence is allowed to
be changing, over time and space.
There are many different models for quintessence, and they all arise to some extent
from particle physics. One idea is that it is a scalar field. A scalar field defines a
strength of the field at every point in space, but does not specify a preferred direction
for the field. Einstein’s cosmological constant can be thought of as a special case of a
scalar field, where the strength of the field is the same everywhere. In more general
scalar fields, the strength of the field can vary in time and position. Quintessence
is thought to vary in strength over time. One scenario has it increasing in strength
until it reaches the constant value that it still has today.
Chapter 6
As first discussed in § 4.2, a highly uniform radiation field pervades the Universe. This
cosmic microwave background (CMB) radiation originates when the Universe first be-
came transparent, about 300,000–400,000 years after the Big Bang. Before this time,
the radiation was strongly coupled to the matter via collisions with free electrons. In
fact, the radiation dominated over the matter, with roughly 109 (1 billion) photons
for every baryon. The constant collision, absorption, and re-emission of photons with
the matter particles meant the radiation retained a blackbody spectrum, which can be
characterized by a single temperature. Fig. 6.1 show several example blackbody spec-
tra that would be expected from objects with temperatures 3000–6000 K. Notice that
as the temperature decreases, two things happen: (1) the flux density decreases—less
total radiation is emitted, and (2) the peak of the emission moves to longer (redder)
wavelengths. The peak of the blackbody spectrum is in fact given by Wien’s Law,
which in SI units is:
5.1 × 10−3
λpeak = , (6.1)
T
where the wavelength λpeak is in meters and the temperature T is in Kelvin. From
Fig. 4.4 the peak of the COBE spectrum can be measured accurately to be at a
wavelength of 1.87 mm. From Wien’s Law, the implied temperature is T = 5.1 ×
10−3 /λpeak , where λpeak = 1.87 mm = 1.87 × 10−3 m or T ≈ 2.73 K.
As also noted before, this is the current temperature of the radiation. When it
was emitted at the time of recombination at z = 1100, the Universe was 1100 times
smaller. The expansion of space-time has resulted also expanded the CMB radiation
by this factor. Since Wien’s Law is linear, a 1100 factor decrease in wavelength of
the radiation implies a temperature at recombination with a temperature 1100 times
The specific way in which the CMB is treated as an acoustic phenomenon in this
chapter follows the method developed by Mark Whittle. See his excellent website
http://www.astro.virginia.edu/ dmw8f/ for more information.
97
98CHAPTER 6. ANISOTROPIES IN THE COSMIC MICROWAVE BACKGROUND
The temperature is highly uniform, after foreground emission and the dipole
anistropy has been subtracted out. However as seen in the bottom image in Fig. 4.3,
the CMB is not completely smooth: it has tiny temperatures on the order of 1:100,000
of the mean temperature. These fluctuations further appear at different sizes in the
sky in the COBE maps. They therefore have different angular scales. COBE’s
detectors had a resolution of 7◦ , meaning this was the smallest angular-size of an
object that it could detect. Fluctuations larger than this scale could be seen, but not
anything smaller.
There have been many follow-up attempts to measure finer scale fluctuations
than could be observed by COBE. Many telescopic experiments were planned and
executed from the ground, while a couple were planned for space. One ground-based
experiment was BOOMERanG (“Balloon Observations Of Millemetric Extragalactic
Radiation and Geophysics”), a balloon-borne telescope that floated above Antarctica,
and which observed a 40◦ × 25◦ patch of sky for 10 days.
However the highest resolution maps to date are being provided by the Wilkin-
son Microwave Anisotropy Probe (WMAP), which was launched in June 2001. The
angular resolution has been improved by almost a factor of ten over BOOMERanG,
resulting in structures that are resolved down to 0.25◦ . The first detailed maps from
WMAP were released in 2003, but the satellite continues to gather more data.
6.1. ANALYZING THE FLUCTUATIONS 99
For cosmologists, the locations of a specific warmer patch versus the location of
another cooler patch is not that important. What will reveal information about
the Universe as a whole is a statistical study of all of the fluctuations at once. This
can give us a global view of what is occurring in the Universe, rather than a local view
of the temperature variation in any one patch of sky. The method that cosmologists
have used to analyze the CMB is to study not just the temperature fluctuations,
but their wavelengths or angular size, and to see how strongly each sized fluctuation
contributes to the overall CMB.
The Universe before recombination should have been a hot plasma of protons,
neutrons, and electrons, mixed into a sea of photons, neutrinos, and dark matter
particles. Back then as now, the dark matter should dominate over the ordinary
baryonic matter. However the dark matter does not interact with the radiation field,
but interacts only gravitationally with the baryonic matter. But since the baryonic
matter is coupled tightly with the photons (to form a photon-baryon fluid), the
dark matter will be interacting indirectly with the radiation.
Non-uniformities in the dark matter means that the dark matter particles will
100CHAPTER 6. ANISOTROPIES IN THE COSMIC MICROWAVE BACKGROUND
pool together in slightly higher density halos.1 The photon-baryon fluid will have a
tendency to flow toward these halos by the action of gravity. However the collapsing
photon-baryon fluid will resist being compressed by its internal pressure: it will tend
to “bounce back.” The photon-baryon fluid rebounding from falling into the dark
matter halos will oscillate, which create variations in density and pressure that spread
out as acoustic or sound waves. Thus the chaotic primordial stew in the aftermath
of the Big Bang can be viewed as being criss-crossed by sound waves.
Sound waves are compression waves that increase and decrease the pressures of
the medium that it travels through (Fig. 6.5). The amplitude of a wave is the height
of the peaks and the depth of the troughs as seen in a traditional line plot representa-
tion (Fig. 6.5a). In a sound wave, this amplitude difference is a difference in pressure
between the regions compressed and rarefied by the wave (Fig. 6.5b). In the com-
pressed zones of a sound wave that is passing through the air, the air molecules are
squeezed closer together, while in the rarefaction zones, the air molecules are spread
further apart. The wavelength is the distance between successive peaks (or troughs)
within the wave.
The wavelengths of these sound waves depend on the time after the Big Bang that
the oscillations that generated them begin. At time t, any disturbances in pressure in
the photon-baryon fluid are not felt at distances greater than the distance ct, since the
maximum speed by which any signals can propagate is fixed by the speed of light, c.
Early enough in the Universe’s history, ct is smaller than the size of the dark matter
halos, so no fluctuations will originate from the halos. However given enough time
as the Universe expands, ct will be greater than the size of the halos and the sound
waves can begin to propagate from the halos. A useful point to remember is that the
largest wavelength of the waves will be roughly ct.
Early in the Universe, many waves with a large range of wavelengths and am-
plitudes can exist, just like the waves on the surface of a choppy sea (Fig. 6.6).
Looking at the CMB is therefore like looking at the multitude of ripples that are on
the surface of a wind-whipped ocean. The wavelengths measured so far by COBE
and its successor telescopes and experiments are in the range 20,000–200,000 light
years. The variations in pressure are about 1:10,000, which correspond to 110 dB
(decibels). Because this period in the Universe’s history is radiation-dominated, the
wave speeds are very high, about 0.5–0.6c. The number of waves passing per second
gives the pitch or sound frequency. Since these waves are so enormous and it would
take 40,000 years or more for a single wave to pass by, their frequencies are extremely
low, around f ≈ 10−12 –10−14 Hz.
To study the ripples in the CMB, cosmologists construct an angular power spec-
trum. This is a plot showing the contributions of all of the different wavelengths to
the CMB as seen in the entire sky. Using special mathematical techniques, they
extract out the amplitude of each wave with a wavelength corresponding to some
1
These slightly denser regions will be the seeds of future superclusters of galaxies, while the giant
voids between superclusters will come from the regions between the halos.
104CHAPTER 6. ANISOTROPIES IN THE COSMIC MICROWAVE BACKGROUND
angular size on the sky. The minimum size structure in the COBE maps for instance
are waves that correspond to structures 7◦ across. To make a more direct correspon-
dence with sound waves, instead of talking about the wavelength of the fluctuations,
we refer to their angular frequency or multipole number l, which is roughly
180◦
l≈ , (6.2)
λ
where λ is the wavelength. Therefore the minimum COBE-visible structure of 7◦
corresponds to l ∼ 25, while 1◦ -sized patches are equivalent to l ∼ 180. Thus small
waves have large l, while enormous waves show up at tiny l. Fig. 6.7 shows three
examples of very simple power spectra. The largest wave possible is one that fills
up the entire sky with the peak of the wave in hemisphere, and the trough in the
opposite hemisphere. The wavelength is λ = 180◦ , so l = 1. If this was the only wave
component in the CMB, then its power spectrum would show a single peak at l = 1.
If the CMB contained only fluctuations that were 7◦ across, then the sky might look
like the middle row in the figure, with a power spectrum spike at l ∼ 25. Finally
if there CMB was filled with structure 0.25◦ across, then the power spectrum would
have a single peak at l ∼ 700.
Fig. 6.8 shows a power spectrum from the CMB that summarizes the results
from seven different experiments, including BOOMERanG and WMAP. Plotted are
fluctuations with l in the range 2–1800. The data points show not only the amplitude
of each fluctuation, but also the corresponding error bars, which gives you a rough
idea of the uncertainties associated with each experiment.
Fig. 6.9 shows the result of averaging together all of the available data from
early 2003. The smooth curve shows the best fit from a cosmological model based
on Friedmann-Robertson-Walker universes that vary a large number of cosmological
parameters.
numbers l <∼ 50, variations in the CMB are caused by the Sachs-Wolfe effect. A
region with a higher density in the last-scattering surface implied a slightly over-
density, which had a deeper gravitational potential well. General Relativity predicts
that photons become gravitationally red-shifted if they are forced to climb out of a
potential well (and conversely become blue-shifted if they fall in). Additionally, Gen-
eral Relativity also predicts a time dilation effect when they climb out of a potential
well. The slightly lower temperature patches in the CMB are therefore locations that
are over-dense compared to the average, and these lead to the “plateau” in the CMB
angular power spectrum at low l-numbers.
At l-numbers in the range ∼ 50–1000, the power spectrum reflects the acoustic
waves at the time of decoupling. Some pockets of the photon-baryon fluid are reaching
their maximum state of compression as they oscillate in the dark matter halos. These
compressions heat the photon-baryon fluid, causing the photons leaving these pockets
to be at a higher temperature than photons elsewhere. The size of these anisotropies
will be similar to the wavelength of the waves, or hence, will be at most ctdec in size,
where tdec = 380, 000 years is the time of recombination. Thus the largest fluctuations
should be no more than 380,000 light years in size. In a flat universe, where k = 0,
these largest fluctuations are thought to be about 1◦ in size, and correspond to the
peak at l = 220.
Another way to think of this is to think of these fluctuations as due to the photon-
baryon fluid having a chance to compress once before it freezes out at recombination.
There are other fluctuations that are half as large; their wavelengths are 1/2 of the
6.2. ORIGINS OF THE POWER SPECTRUM 109
l = 220 mode, but as a result oscillates twice as fast. Then this anisotropy will have
a chance to compress and rarefy before combination. Finally for a mode that has
a wavelength 1/3 of the l = 220 mode, it can compress, rarefy, and compress itself
again. These three modes show up as the first, second, and third peaks in the CMB
power spectrum. The second, third, and higher number peaks can be thought of as
higher order harmonics of the first, fundamental peak.
For very small angular scales, with l >
∼ 1000, the acoustic waves are reduced by
an effect called Silk damping. The wavelengths of the oscillations are so short,
that they are comparable to the distance that photons travel during recombination.
Recombination does not happen instantly; during this period photons can bounce
collide with the charged electrons and baryons. During this random walk, photons can
scatter across the wavelength of the fluctuation, and the warmer and cooler photons
can mix and average out. The acoustic peaks therefore end up getting smeared out.
measurements are added. However the parameters derived from it are a remarkable
set of numbers, some of which have been pursued by cosmologists for much of the last
century. A measurement like the age of the Universe has been debated by thinkers for
millenia, and it is a testament to our scientific and technical prowess that we know it
to such a high degree of accuracy today.
Galaxy and structure formation in the Universe after the Big Bang is thought to
have proceeded from the non-uniformity “seeds” provided by the dark matter halos
discussed in the previous chapter. These localized regions of density enhancements
attracted additional matter because of its gravity, which continued to increase size of
these localized regions of density. The dark matter is important because left to itself,
seeds containing baryonic matter only would take too long to grow and amplify into
the structures we see today; some models show it should take over 40 billion years
for gravity to coalesce structure from only baryonic matter. Therefore not only is the
existence of dark matter inferred from observations of galaxies and galaxy clusters,
but they appear necessary for baryonic matter to settle into large-scale structure
within the lifetime of the Universe. Dark matter also behaves differently than the
matter. Its only interaction is via gravity so it does not interact with the photons, like
the baryons. Dark matter merely clumps up and forms the potential wells depicted
in Fig. 6.10. And because dark matter is anywhere from 5–10 times more abundant
than ordinary baryonic matter, it is primarily responsible for structure formation in
the early Universe.
Two general categories of dark matter have been proposed. The first, hot dark
matter (HDM), is now somewhat out of favor. The “hot” in its name refer to the
subatomic particles which are thought to travel near the speed of light. The neutrino
was the favored candidate for HDM, but recent experiments have shown that its
mass is too insignificant to make up but a small component of the dark matter.
Because of its speeds close to the speed of light, HDM tends to wash out the small
scale fluctuations. Cosmologists who create theoretical simulations with HDM see
extremely large-scale structures forming first, which then break apart into smaller
objects, in an hierarchical scenario which gives top-down structure formation.
However in order to produce galaxy-sized objects in the present-day Universe, HDM
models tend to also create much more large scale structure than is currently observed.
113
114 CHAPTER 7. STRUCTURE FORMATION IN THE UNIVERSE
The other class of models involve cold dark matter (CDM), which are typically
massive, slow moving particles such as the WIMPs. Computer simulations show
that a CDM-dominated universe tends to form small structures first, typically a
million solar masses. These then collect together to hierarchically build up to galaxy-
sized (and larger) structures. This is then a bottom-up scenario. The timeline of
structure formation would therefore have stars forming first, which then would gather
into globular cluster-sized structures, which would themselves gravitationally coalesce
into galaxy-sized objects.
The most difficult part of CDM simulations is modeling the stars. Once stars
turn on, the physics in the simulation will grow much more complex. Massive stars
radiate copious amounts of UV radiation, expel matter out in powerful winds, and
die in supernova explosions. All three of these processes can dramatically change the
environment that they are in. Winds and supernovae, for instance, disrupt or rip
gas from proto-galaxies. Furthermore the first stars had no heavy elements, so they
represent a class of objects which we have no examples of today.
Whether a clump of matter actually collapses or remains stable depends on several
parameters. For a given gas cloud of a certain size, mass, and temperature, the cloud
can have enough internal pressure to be stable against collapse, or its gravity can
overwhelm this pressure. The mass of a cloud right at the boundary between these
two scenarios is called the Jeans mass. A cloud of a given size that is as massive as
a Jeans mass will collapse. A cloud with less mass than its Jeans mass will have an
internal pressure that can counteract the collapse.
In the early Universe, we must also worry about the expansion. Although gravity
is pulling matter into regions with increasing density, the expansion is decreasing the
overall density. About 10 million years after the Big Bang, the lumpiness of the matter
is about 50% that of the lumpiness of the dark matter wells. After 100 million years
after the Big Bang, the density concentrations are now twice the average density of the
surrounding average density. At this point, gravity has overwhelmed the expansion
of space-time, and the clumps start to free-fall and collapse. The largest galactic
superclusters that we see in the Universe are thought to result from the clumps that
give the 1st and 2nd peaks in the CMB angular power spectrum.
black holes which have masses a million times the mass of the Sun (or larger) could
effectively be invisible. Increasing the number of particles means each particle in the
simulation can represent a smaller mass, and you will be able to follow the formation
of less massive objects, like small galaxies.
Although the growth of density fluctuations can be derived analytically by math-
ematics for awhile, numerical simulations are necessary after the fluctuations begin
their collapse into the large-scale structure found in galaxy superclusters today. Such
simulations represent the matter in the Universe as a large number of discrete par-
ticles, each of which react to the collective gravity of all of the other particles. The
matter particles further can interact with radiation. The Millenium simulation fol-
lowed 10 billion particles of ordinary matter and dark matter located in a volume
with a present width of 700 Mpc, from a redshift of z = 127 to the present, z = 0.
The simulation was able to follow the evolutionary histories for 20 million galaxies,
some as small as that of the Small Magellanic Cloud.
Additional cosmological parameters—such as the size of the Hubble constant, the
amount of dark matter relative to the baryonic matter, the size of the cosmological
constant, the size of the fluctuations expected from analyses of the CMB, etc.—are
input into the model as initial conditions. The computer model is then allowed to
evolve, with the interactions of all of the matter following realistic physics. The
simulation was then used to follow formation and growth of black holes (which lay at
the centers of most galaxies), quasars, and galaxies over the Hubble lifetime. Through
trial and error, the input parameters were adjusted and the model was repeatedly
run until it gave results for nearby, low-redshift structure that is similar to what is
observed in our local Universe.
Quasars and other active galactic nuclei are thought to be powered by central
black hole engines. Supermassive black holes have been discovered in the centers of
many galaxies, and they may be found in all galaxies. The formation of these enor-
mous black holes (which can be millions of times the size of “ordinary” black holes
that originate from the collapse of a massive star) is therefore thought by many extra-
galactic astronomers to occur in the early Universe, concurrent with the hierarchical
formation of structure.
The Sloan Digital Sky Survey (SDSS) has made the surprising discovery of some
extremely bright quasars up to a redshift z = 6.43. If the distances are correct, then
the luminosity of these quasars implies that the central black hole engines have a
mass of 109 M , which is a thousand times larger than the mass of the central black
hole in our Milky Way. Cosmologists were unsure whether such super-supermassive
black holes could form, but the Millenium Simulation does show a number of such
objects coalescing. These eventually evolve into the most massive elliptical galaxies
in the simulation.
The goal of follow-ups to the Millenium Simulation and its ilk is that as they
successfully model large-scale and smaller-scale structure in their simulations (the
Millenium Simulation can resolve objects down to the Small Magellanic Cloud in
116 CHAPTER 7. STRUCTURE FORMATION IN THE UNIVERSE
size), cosmologists and astronomers will be able to “fine-tune” the cosmological pa-
rameters. Some input parameters are well known, such as the Hubble constant (with
an uncertainty of 10%). Others are not as well known, such as the behavior of dark
energy at early times. By tweaking the parameter behavior—for instance by changing
the value of dark energy at different times—the cosmologists hope to find a simulation
that can be compared with the results from future, deep astronomical surveys, and
at the same time, deduce something about the nature of dark energy.
Chapter 8
The Big Bang theory is well established with predictions that are matched by ob-
servations. The three broad categories of evidence are the expanding universe, the
light element abundances, and the existence of a highly uniform cosmic microwave
background. However despite this set of consistent evidence, there are several issues
that remain unresolved with the basic model. The inflation model has therefore been
proposed to fix these problems. Inflation is added on so that it sets the initial condi-
tions for the Big Bang. There is yet any slamdunk evidence to prove inflation is right
beyond reasonable doubt, unlike the evidence for the Big Bang.
The Horizon Problem: Recall that in the cosmic microwave background is ex-
tremely uniform over the entire sky, about 1 part in 105 , meaning there is less
than a 0.0001◦ K difference in temperature from one part of the sky to another.
However at the time of last-scattering, when the photons were coupled to the
matter, the closest neighboring photons that could exchange energy with each
other at the speed of light is equivalent to only 2◦ on the sky today. This is
the horizon problem: how can parts of the cosmic microwave background 180◦
apart exchange energy to have the same temperature today? It could have been
sheer coincidence that all these disconnected regions had such close tempera-
tures, but that seems very unlikely. Somehow these very widely separated parts
of the Universe today must have been far closer together to be able to swap en-
ergy and reach thermal equilibrium, then is suggested by the normal Big Bang
121
122 CHAPTER 8. INFLATION AND THE EARLY UNIVERSE
model.
1.0000000000000000000000000000000000000000000000000000000000001,
or smaller than:
0.9999999999999999999999999999999999999999999999999999999999999.
symmetry breaking when the strong force splits off from the electroweak
force at the end of the Grand Unified era (see § 8.3). However there are ex-
pected to be slight timing differences when this symmetry breaking occurs,
which results in volumes of space-time that are separated by “topological de-
fects.” These interfaces in space-time can be “cosmic strings” (one-dimensional)
or “domain walls” (two-dimensional). However no topological defects have ever
been detected either.
One last problem with our model of the Universe is that there appears to be an
imbalance between matter and antimatter. Matter will annihilate perfectly with its
antimatter counterpart (electrons with positrons, protons with anti-protons, neutrons
with anti-neutrons, etc.) with high energy photons as a result. Similarly the reverse
reaction can occur: photons can spontaneously create matter-antimatter pairs, such
as an electron and a positron. One would there expect the early Big Bang to produce
equal amounts of matter and antimatter in the Universe. However by all accounts,
there is very little antimatter. The Universe is mostly matter.
quantum mechanics, one can make predictions about the behavior of the vacuum.
One of these is that at extremely high temperatures, such as that which you find
immediately after the Big Bang, the vacuum can change states. This is akin to a
property of water, which normally freezes and turns to ice during a phase transition
at its critical temperature of 0◦ C. However it is possible to cool undisturbed water
below 0◦ C. This supercool water remains liquid, but any slight disturbance will cause
the water to quickly turn to ice.
The vacuum in the early Universe reached its critical, unstable state when it
cooled below 1027 K. This version of the vacuum is known as the false vacuum.
The phase transition that would drive inflation is thought to come as the result of
new hypothetical particles that exist during the Grand Unified Theory (GUT) era,
when the strong force merges with the electroweak force. (Gravity is still separate,
and does not unify with the other forces until times t < 10−43 sec; see § 8.3.) These
particles are thought to be described by scalar fields, which is exactly the type of
force necessary to create a vacuum-driven expansion.
A postulated inflaton particle, which gives rise to inflation, has a quantum field
φ that describes it, which varies with temperature. The energy of this field can
be described by a “Mexican hat” potential, with a peak at φ = 0 and the energy
increasing with greater values of the field φ. At high enough temperatures, the
location of the inflaton is at a minimum at φ = 0 of a “Mexican hat” potential
where the false vacuum is located (Fig. 8.2). While the inflaton is in the φ = 0 state,
inflation occurred.
However the location of the false vacuum is unstable: just as a marble located at
the peak of a Mexican hat tends to roll down the side, given just a slight nudge, the
inflaton can drop down in energy with a nudge from a quantum fluctuation. Such
a fluctuation (similar to those that create virtual particles) causes a pocket of the
false vacuum to decay into a true vacuum. Very quickly (within 10−34 sec) the
true vacuum would fill up the Universe. During the transition from a false vacuum
to a true vacuum, an enormous amount of energy is released which forms particle-
anti-particle pairs. Thus according to the inflation model, the vast majority of the
particles in the Universe were created as a result of inflation.
This enormous exponential expansion solved many of the problems we mentioned.
The Universe started expanding because it was impelled to do so by inflation. The
horizon problem is no longer an issue: The regions now separated by great distances
in the present epoch were actually once much closer together —so close that they
allowed these regions to reach the same temperature within a light travel time. It
was inflation that spread out space-time by such a vast amount that such regions
could never be in contact with each other again.
Inflation solves the flatness problem by expanding the space-time geometry of
the Universe. Even if the Universe had not started out being flat, the vast rate of
expansion stretched the curvature of Universe until it appeared flat. Therefore it does
not matter if Ω started as 1, 10, 100, 106 , or 10−6 . As long as there is enough inflation,
126 CHAPTER 8. INFLATION AND THE EARLY UNIVERSE
Ω would have evolved to 1. A similar analogy is that a small balloon might have a
noticeably curved geometry, with lumpiness or even wrinkles. However if you were
to blow up the balloon until it was the size of the Earth, the surface of the balloon
would appear to be extremely flat, even in the parts that were wrinkled. (If inflation
were to occur for our balloon, it would blow up to a size about 1024 or 1,000,000,-
000,000,000,000,000,000 times larger than the current size of the observable Universe.
The surface of such a balloon would indeed look extremely flat!)
The problem of magnetic monopoles and other exotic objects is also solved by
such an enormous expansion. The original pool of magnetic monopoles would have
expanded along with the rest of the Universe. The expansion is so great that they
would be so spread out, that there would likely not be a single magnetic monopole
left within our observable Universe.
Inflationary theory does not quite solve the imbalance problem between matter
and antimatter. However Grand Unified Theories involving the unification of the
strong nuclear force with the electromagnetic and weak forces (see the next section)
suggests that at high enough energies, matter could have formed with a slight excess
over antimatter. As the Universe cooled after this very early period to the time
t = 10−5 sec, all of the protons annihilated with the anti-protons to create gamma
ray photons. For every billion (109 ) proton-anti-proton pairs, there was an extra
proton that did not annihilate. These residual protons would eventually become all
the matter that we see today. The photons that came out of the annihilations became
the radiation bath that expanded with the Universe and turned into the radiation
background discussed back in §4.2.
128 CHAPTER 8. INFLATION AND THE EARLY UNIVERSE
weakest of the four, it is more significant than either the strong or weak force, be-
cause it operates at a distance (instead of at subatomic length-scales). And although
electromagnetism also operates at long distances, gravity is the dominant force in the
Universe, since charge imbalances appear to be small, so that electromagnetic forces
tend to cancel each other out.
Particle physics experiments show that the strength of interactions vary with the
interaction energy. At higher temperatures and higher energies, the electromagnetic
and the weak nuclear forces are found to be closer and closer to each other in strength.
16
At T > ∼ 10 K, the strengths of the two interactions are expected to be the same,
and the electromagnetic and weak forces can therefore be thought of as variations
of a single electroweak force, via a theory that was first proposed by physicists
Abdus Salam and Steven Weinberg. Electroweak theory was verified in 1983 with the
discovery of the W + , W − , and Z 0 bosons in particle accelerators.
Because of this success with the electroweak theory, physicists are now pursuing
unification theories that will include the other two forces. The drive for unification
is in part philosophical: the belief that the Universe can be described by a minimal
set of particles and physical processes. The next goal for physicists is the Grand
Unified Theory (or GUT), which combines the strong and the electroweak forces
at T ∼ 1028 K. Inflation is thought to originate from the end of the GUT period, as
the strong and electroweak forces split. The current matter-antimatter imbalance is
also thought to be the result of processes from this period.
The final unification with gravity is believed not to occur until temperatures ex-
ceed T ∼ 1032 K. This is associated with a time t ∼ 10−43 sec (roughly the Planck
time—see below) after the Big Bang, during which quantum mechanical and gravi-
tational effects become equally important. Stephen Hawking and Roger Penrose have
shown that at time t = 0, the field equations of General Relativity break down, and
singularities, or infinities appear. Analogous to how electromagnetic theory breaks
down at high enough energies, it is thought that with sufficiently strong gravitational
fields or extremely small length scales, gravity must be described with a quantized
model to avoid these singularities.
133
134 CHAPTER 9. RAMPANT SPECULATION
of which goes through inflationary growth and evolves into a true vacuum, resulting
in a Big Bang and a universe similar to our observable universe. However at the same
time that the true vacuum universe is expanding, the space in the two remaining
regions of false vacuum has also expanded, so those two regions are now the same
size as the starting region of false vacuum. If we follow these regions for another
period of time, part of the false vacuum can also evolve into a local universe with a
true vacuum. This process continues forever with each false vacuum growing in size,
generating new universes, and growing in size again, ad infinitum.
What results are an infinite number of bubble universes or pocket universes.
None of the universes are accessible to any of the other universes, since inflation
will have expanded them into space-time volumes that are far beyond each of their
observable horizons. One consequence of eternal inflation is that the infinite web of
expanding universes are fractal in nature, with an infinite number of pocket universes
growing with different scale factors, while false vacuum pockets keep subdividing to
spawn new universes.
One possibility for eternal inflation is that there may not be a first universe or
an original false vacuum. One can imagine Fig. 9.2 extending forever upward as well
as forever downward so that there is no t = 0. In the plot, we have drawn only the
results of just one patch of false vacuum at one point in time. If eternal inflation
continues infinitely into the past, then this false vacuum may share a common false
vacuum ancestor with an infinite other false vacuums and pocket universes.
As we have seen from these two examples, inflation is still a model very much in a
state of flux. Even the exact mechanism for initiating inflation is not well understood.
Although inflation solves many fundamental problems of the Big Bang, this feature
does not guarantee that inflation actually occurred. The very ad hoc nature of the
many variants of inflation—people often come up with new variations to explain
different observations—may even argue against the idea as a whole. However many
physicists and cosmologists would argue that the concept of inflation has solved far
more problems than it introduces, and some version of it is here to stay.
ing in a world that is exactly the same as our world. This Parallel Earth orbits the
Parallel Sun with the same set of Solar System objects. Parallel universes may also
diverge from our own Universe, whether slightly or by a lot. Again jumping off from
the science fiction literature, there might be universes where JFK was not shot or
Hitler won World War II, or where Elvis is still alive (or perhaps this latter one is
our Universe).
Recall that the furthest that we can observe in the Universe is about 13.7 billion
light years, which is to the edge of the cosmic microwave background, at which point
the Universe becomes opaque. This Hubble volume is the furthest that we can see
now, but this has changed with time. When the Universe was half its current size,
our Hubble volume was then 7 billion light years in radius, because that was as far
light could travel to reach us from the beginning of the Universe. The expansion
of space-time is also constantly carrying galaxies out of our Hubble volume. If we
observe light from a galaxy whose distance we measure to be 10 billion light years
away, then that light was emitted 10 billion years ago. Today however, that galaxy is
not only 10 billion years older, but the expansion of space-time has carried it along
so that right now, it is 30 billion light years away. If we wanted to view it as it looks
today, it would not be possible since there would be no time for light being emitted
now to ever reach us. The expansion of the Universe is moving those galaxies outside
of our Hubble volume. An accelerating expansion makes it even worse: depending on
the exact amount of acceleration, we would lose sight of more galaxies faster.
There is also evidence that our Universe is much larger than our observable Hub-
ble volume. The curvature of space as measured by the CMB appears to be very
flat (§ 6.3). Even taking the upper limit of the curvature parameter based on the
uncertainties in the measurements, the Universe will contain at least 1000 other Hub-
ble volumes. The flatter the Universe, the more Hubble volumes will exist outside
our own Hubble volume. Completely flat Friedmann-Robertson-Walker universes are
infinitely large. If space is infinite in size now, then it was infinite to start with.
Inflation itself adds another complication, since it predicts the Universe to be
much, much larger than the observable Hubble volume. Some cosmologists, like
Jaume Garriga and Alexander Vilenkin, even argue that inflation can cause a Universe
that was originally finite in size to grow to infinite size.
Our Hubble volume however contains only a finite amount of space, with a finite
number of particles—estimated to be about 1090 . In an infinitely large universe, there
will be an infinite number of Hubble volumes. In classical physics, it turns out there
are an infinite number of ways to arrange a finite number of particles. But in quantum
mechanics, there will be only so many different ways you can arrange 1090 particles
within a single Hubble volume. Just from statistical arguments, one would expect
to run eventually into a repeat of that particle arrangement after a traveling a finite
distance away.
How far away would we find a repeat particle arrangement? Using basic quantum
mechanical arguments about how to arrange particles, Tegmark estimates that the
138 CHAPTER 9. RAMPANT SPECULATION
29
closest identical copy of any one of us is about 1010 meters away. For larger arrange-
ments of particles, the probability is lower for an exact duplicate, so you have to go
even further to find a duplicate. For a sphere about 100 light years across (meaning
a volume that exactly matches our Sun, the Solar System, and nearby stellar neigh-
91
borhood), a duplicate will be found on average about 1010 meters away. Finally to
find a Hubble volume that is identical to ours, statistically speaking, one has to travel
115
about 1010 meters.
These are ridiculously huge numbers (although still finite and less than infinity).
29
Even the smallest distance, 1010 meters, is so large that it is difficult to imagine,
and is impossible to write out.2 These numbers are therefore far, far beyond anything
that is currently or will be observable.3
Tegmark calls such a set of parallel universes “Level I” multiverse. His “Level II”
multiverse involves elements of eternal inflation theory.4 Recall that extensions of
inflation like eternal inflation predict that universes continue to sprout from the false
vacuum (Fig. 9.2). Taken to one logical extreme, there will be an infinite number
of these bubble universes, for eternal inflation forward in time. If eternal inflation
continues backward in time as well, then there will be infinitely still more pocket
universes. Each of these bubble universes will also be infinitely big if the view of the
Level I multiverse is correct.
A post-inflation universe that evolves into a big bang however may have different
sets of physical constants. The value of the Planck constant, the electron-proton mass
ratio, the ratio of the electron charge to its mass, the strength of the weak force, etc.,
do not necessarily have to be locked to the values found in our Universe. Their values
could be different depending on the types of symmetry breaking that occur as the
post-inflation bubble cools. The Level II multiverse will therefore have much more
variation in it than just a rearrangement of particles, when compared to our Universe.
The number of space dimensions could be different, as well as the number of quark
families, or any of the other numerical constants in the “Standard Model” of particle
physics.
Thus while the Level I multiverse involves all possible arrangements of particles
to create parallel universes, the Level II multiverse is a super-set of that, involving
all possible values of physical parameters.
29
2
That is, 1010 is 1 followed by 1029 zeroes. To give you an idea of how many zeroes this is,
there are slightly less than 1029 protons in the human body.
3
That is unless dark energy oscillates between a repulsive and an attractive state in the far future,
which has been suggested by some cosmologists. If it moves out of the repulsive realm, dark energy
will act to reinforce gravity, causing the expansion of space-time to slow, eventually stop, and then
reverse. As the Universe grows smaller in size once unobservable regions of the Universe will come
into the observable Hubble volume.
4
If you guessed that Tegmark has come up with higher level multiverses, you are correct. We
will not have time to go through them here but you can learn more about them in his articles on
the subject.
9.3. THE END OF THE UNIVERSE 139
1. The Stellar Era: This is the era that we are in now. Gas is locked up in stars,
and then expelled back into space when a star dies. However this process cannot
go on forever since eventually all the useful hydrogen will be used up. More and
more mass is locked up in white dwarfs, neutron stars, and black holes, until
there is not enough free gas to create new stars. The longest length of time that
a star can theoretically live is about 1014 years (or 100 million million years)
for objects with about one-tenth the mass of the Sun. We currently live in an
epoch 13.7 billion years after the Big Bang. After a length of time 10,000 times
longer than the current age of the Universe, all of the gas will be used up and
the Age of Stars will be at an end.
2. The Degenerate Era is reached once all matter is locked up in white dwarfs
and neutron stars. Here the word “degenerate” describes the quantum me-
chanical state of the electrons and protons in these compact objects. Over time
white dwarfs and neutron stars will cool off by radiation until they are the same
temperature as the cosmic microwave background (which itself will have cooled
off until it is just a fraction of a degree Kelvin above absolute zero). Galaxies
also gradually dissipate through a process called two-body relaxation, so that
the stars “evaporate” over time. What is left then is a Universe filled with free
floating black dwarfs and black holes. However even the black dwarfs them-
selves break down. According to Grand Unified Theories in particle physics,
the proton, one of the most stable of elementary particles, is expected to decay
and break down into a positron and a meson in 1032 years. Thus all of the
black dwarfs, planets, and whatever ordinary matter is left will have undergone
proton decay in 1037 years.
3. The Black Hole Era: Once all the protons have broken down, the only objects
left in the Universe with any appreciable mass are the black holes. However
even black holes themselves break down and evaporate through the process of
Hawking radiation emission. They will eventually disappear completely over
time as they are replaced by the pool of electrons, positrons, and photons that
are released. Black hole decay through Hawking radiation is an extremely slow
process, with the largest super massive black holes from the centers of galaxies
140 CHAPTER 9. RAMPANT SPECULATION
4. The Dark Era: After 10100 years, and after the disappearance of black holes,
the Universe is left with nothing but a sea of photons and neutrinos, whose
wavelengths get longer as the Universe continues to expand. We are at a point
of maximum entropy and chaos. This photon and neutrino sea is so uniform,
without sources of energy or any sinks, so that it is not possible to do work
to create ordered structures. Eventually the expansion of the Universe will
continue to the point where each individual remaining particle is expanded
outside of the light horizon of every other particle. That is, if we were to examine
one photon, there would be no other observable photon or other particle within
the observable Universe. They are too widely separated for one to reach another
even traveling at the speed of light.
The photon energy will continue to decrease until it reaches the lowest possible
value of that of the quantum vacuum state. Not only is the Universe now
completely dark and alone except for a solitary particle at the bare minimum
energy level, but we have reached the End of Time as well. There is nothing
to distinguish one moment from the next. This is truly the End of the Universe.
Appendix A
Because science often deals with extremely large or incredibly small numbers, a short-
hand notation for writing such figures has been developed. First we write an expo-
nent as a superscript after a number, which signifies the number of factors of that
number to be multiplied together. Thus, an exponent of 2 over a 10 means two tens
multiplied together. Here is a list of exponents up to 8:
10 = 10 = 101
100 = 10 × 10 = 102
1000 = 10 × 10 × 10 = 103
10, 000 = 10 × 10 × 10 × 10 = 104
100, 000 = 10 × 10 × 10 × 10 × 10 = 105
1, 000, 000 = 10 × 10 × 10 × 10 × 10 × 10 = 106
10, 000, 000 = 10 × 10 × 10 × 10 × 10 × 10 × 10 = 107
100, 000, 000 = 10 × 10 × 10 × 10 × 10 × 10 × 10 × 10 = 108 .
141
142 APPENDIX A. NOTATION AND ASTRONOMICAL UNITS
0.1 = 10−1
0.01 = 10−2
0.001 = 10−3
0.0001 = 10−4
0.00001 = 10−5
0.000001 = 10−6
0.0000001 = 10−7
0.00000001 = 10−8 .
1 = 100 .
We can now use this system to express very large numbers, such as the length of a
light year (the distance light travels in one year):
Very small numbers can also be expressed such as the density of gas out in interstellar
space between the stars:
1m = 39.37 inches
= = 100 cm = 102 cm (centimeters)
= = 1, 000 mm = 103 mm (millimeters)
= = 1, 000, 000, 000 nm = 109 nm (nanometers).
1 kg = 2.20 pounds
= 1000 g = 103 g.
A.3. DISTANCES 143
The SI unit for temperature is the degree Kelvin (K). It is similar to the Celsius or
centigrade degree, so that 1 K = 1◦ C But while 0◦ C is set to the freezing point of
water, 0 K is defined to be at absolute zero, when all thermal motion stops, and
therefore the coldest temperature possible. To convert from degrees Kelvin to degrees
Celsius:
Temp(◦ C) = Temp(K) − 273.
To do the opposite and go from Celsius to the Kelvin scale:
According to these formulae, the boiling point of water is then 100◦ C = 373 K.
A.3 Distances
Since the space sciences deal with the vast distances in the universe, a number
of length measurements have appeared in the astronomical sciences that are used
nowhere else in science. Although they have nothing to do with SI units, their use is
so widespread that it is unlikely they will go away in our lifetimes or perhaps ever.
The first is the astronomical unit or AU, and this is defined to be the distance
(technically the semi-major axis distance) between the Earth and the Sun. It is
defined to be:
You will find it used whenever distances in the Solar System are referred to. You may
even see this unit in measurements of other solar systems, or proto-solar systems:
a planet orbiting 6 AUs from its parent star, or an accretion disk 60,000 AUs in
diameter around a Sun-like star.
A far larger unit of measuring distance is the light year, the distance that light,
moving at 299,790 km s−1 , covers in a single Earth year. This is a unit that is
appropriate for describing the distances between stars, and is:
Another unit similar to the light year is the parsec. It derives from the measurement
of distances to stars using the technique of trigonometric parallax discussed in § 2.5.1.
A parsec (abbreviated as pc) is a little over 3 light years:
What about distances far larger than a parsec or a light year? As you have seen in
§ 2.4, the distances between galaxies and clusters of galaxies can be many millions of
light years. Instead of coming up with a new unit for this size scale, astronomers use
144 APPENDIX A. NOTATION AND ASTRONOMICAL UNITS
the SI method of attaching prefixes to existing units. Just as we can scale up a meter
by 1000 times and call it a “kilometer,” a distance 1000 times a parsec is a kiloparsec
or kpc. A kiloparsec or kilo-light year is appropriate for describing distances from one
end of a galaxy to the other.
For distances between clusters of galaxies, we must resort to the “Mega-” prefix,
where a Megaparsec (or Mpc) is 1 million = 1,000,000 parsecs. Figs. 2.15 and 2.18
could have been re-labeled using Mpc or Mly units. Finally for the scales of voids
and filamentary superclusters in the large-scale structure in the observable Universe
(such as Fig. 2.21, we can go to the even larger prefix of Gigaparsecs (Gpc), which is
1 billion parsecs.
A.4 Magnitudes
Traditionally, the brightness of astrophysical objects have been expressed not as
fluxes, but as magnitudes. This is a logarithmic scale, with the apparent magnitude
proportional to the natural logarithm of the incoming radiation flux, or m ∝ log f .
If two objects have observed fluxes f1 and f2 , then
The factor of 2.5 means that a difference in flux of a factor of 100 corresponds to 5
magnitudes.
The absolute magnitude M is defined as the magnitude a source would have if it
were at a standard distance of 10 parsecs. The absolute magnitude therefore gives a
measure of a star’s luminosity. The relation between apparent and absolute magnitude
is:
m − M = 5 log(D/10), (A.2)
where D is the distance in parsecs. The absolute magnitude of the Sun is 4.72, while
its apparent magnitude is m = −26.85. The difference between the two, m − M , is
called the distance modulus.
Many familiar Earth-bound units are based on degrees, arcminutes, and arcsec-
onds. For instance, the latitude and longitude of Denver is given as (39◦ 450 North,
105◦ 00 West), where the starting points for latitude is the equator (0◦ ), and the start-
ing point for longitude is the Greenwich meridian. Celestial coordinates are given
in right ascension (or R.A.) and declination (or Dec.). Declination is measured
from the celestial equator to the celestial poles, while R.A. is measured along the
celestial equator with the vernal equinox as the starting point. However to make
things more confusing to the novice,1 instead of degrees, arcminutes, and arcseconds
for right ascension, astronomers have introduced the units of hours (h ), minutes
(m ), and seconds (s ). These are each, respectively, 15 times larger than degrees,
arcminutes, and arcseconds. Thus,
1h = 60m = 60s , (A.4)
1h = 15◦ ,
1m = 150 ,
1s = 1500 .
In mathematics, it is often simpler to use an angular unit called the radian (rad),
and you will often see this refered to in the astronomical literature. Its relationship
with more familiar units is shown in the following,
π rad = 180◦ , (A.5)
1 rad = 57.2957795◦
= 3437.746771 arcmin
= 2.062648062 × 105 arcsec.
Finally just as one can proceed from measurements of lengths to measurements of
areas when we add an orthogonal dimension, we can go from angular measures of
length to angular measures of area. The standard unit is the steradian (sr), and it
is scaled in such a way that a single spherical surface (such as the celestial sphere of
the sky) is 4π sr in size. Here is how a steradian relates to square degrees, square
arcminutes, and square arcseconds:
1 sr = (180◦ /π)2 = 32400/π 2 deg2 (A.6)
= 1.1664 × 108 /π 2 arcmin2
= 4.19904 × 1011 /π 2 arcsec2 .
to that of the Sun. Therefore, you might read or hear that a particular star has 12 L ,
or 12 times the luminosity of the Sun. (The symbol refers to the Sun.) Similarly
one can also use the mass of the Sun, as a unit: you can say a neutron mass has a
mass of 2.5 M . Or you can claim a supermassive black hole has a radius of 10 R ,
or 10 times the Solar radius.
Just for reference, the solar mass, luminosity, and radius can be written in tradi-
tional SI units as:
The best part of the Internet is that—except for your connection charges—it’s free.
The worst part is that . . . it’s free. The fact that anyone can put up a webpage means
that there is so much content out there, that it can be difficult to sort out the useful
from the useless. Here is a minor attempt at finding and collecting together a list of
web resources that have some relevance to what was covered in class.
Einstein Online: Terrific website covering topics that range from the
elementary—like relativity—all the way to cosmology, black holes, extra-
dimensional space, holographic universes, and more. Recommended if you
are interested in the some of the latest ideas that physicists are pursuing
(http://www.einstein-online.info/en/index.html).
Einstein Year 2005: All about Einstein, put together to celebrate the 100th
anniversary of his “miracle year” when he wrote five ground-breaking pa-
pers on physics (two of which involved the Special Theory of Relativity)
(http://www.einsteinyear.org/).
147
148 FURTHER READING
The Official String Theory Web Site: I don’t know what makes this the “official”
site, but it’s a useful compendium of information on string theory and how it relates
to cosmology (http://superstringtheory.com/index.html).
Superstrings! Another string theory site with a tutorial and list of references
(http://www.sukidog.com/jpierre/strings/).
Ned Wright’s Cosmology Tutorial: One of the best online resources for
all things cosmological and relativistic. Includes both tutorials and FAQs
(http://www.astro.ucla.edu/∼wright/cosmolog.htm).
Martin White’s Cosmology Reading List: Links to dozens of other useful sites
(http://cfa-www.harvard.edu/∼mwhite/readinglist.html).
Cosmology: The Study of the Universe: An educational site from the WMAP
folks, covering the theoretical and observational aspects of the Big Bang and the
Universe (http://map.gsfc.nasa.gov/m uni.html)
150 FURTHER READING
The following is a list of books and magazines for learning more about topics of this
course. Most have popular books written for the interested layperson, at the same
level as articles appearing in Scientific American or Discover magazines. The few that
are textbooks with some mathematics are noted as such.
Fred C. Adams & Greg Laughlin, The Five Ages of the Universe: Inside the
Physics of Eternity, 1999, Free Press. [A history of the Universe starting from
the Big Bang and ending at a staggering 10100 years later. Good if you are really
intrigued by The End of the Universe section from class.]
Kristy Ferguson, Measuring the Universe: Our historic Quest to Chart the Horizons
of Space and Time, 1999, Walker. [A chronicle of the attempts from the last 2000
years to measure the size of the Universe. It was also written late enough to
cover in its last chapter the Type Ia supernovae results.]
George Gamow & Russell Stannard, The New World of Mr. Tompkins, 1999,
Cambridge University Press. [An update of a classic book for the layperson,
with explanations of the Special Theory of Relativity, quantum mechanics, and
the structure of the atom.]
Brian Greene, The Elegant Universe: Superstrings, Hidden Dimensions, and the
Quest for the Ultimate Theory, 2000, Vintage.
Brian Greene, The Fabric of the Cosmos: Space, Time, and the Texture of Reality,
2004, Knopf.
Alan H. Guth, The Inflationary Universe: The Quest for a New Theory of Cosmic
Origins, 1998, Perseus Books Group.
Edward R. Harrison Cosmology: The Science of the Universe, 2nd Edition, 2000,
Cambridge University Press. [Big and expensive, but a real cosmology textbook
but for the non-technical reader, with minimal mathematics.]
Lawrence Krauss, Quintessence: The Mystery of Missing Mass in the Universe, 2000,
Basic Books. [An update of his earlier book, The Fifth Essence.]
Mario Livio, The Accelerating Universe: Infinite Expansion, the Cosmological Con-
stant, and the Beauty of the Cosmos, 2000, Wiley.
Malcolm S. Longair, Our Evolving Universe, 2nd Edition, 1996, Cambridge University
Press.
Laurence Marschall, The Supernova Story, 1994, Princeton University Press.
Martin J. Rees, Before the Beginning: Our Universe and Others, 1997, Addison-
Wesley Longman.
Martin J. Rees, Just Six Numbers: The Deep Forces that Shape the Universe, 1999,
Basic Books, NY.
Martin J. Rees, Our Cosmic Habitat, 2001, Princeton Univ. Press, 2001.
Vera C. Rubin, Bright Galaxies, Dark Matters, 1997, American Institute of Physics.
The editors at Scientific American, The Once and Future Cosmos,
2002. [A reprint of the September 2002 issue of the magazine containing
more than a dozen articles by the field’s leading researchers on the
recent revolution in cosmology. You can order it through SciAm, e.g.,
http://www.sciam.com/special/toc.cfm?issueid=6&sc=rt nav list.]
Edwin F. Taylor & John A. Wheeler Spacetime Physics, 2nd Edition, 1997, W. H.
Freeman. [A very nice introductory---at the undergraduate science level---text
on the Special Theory of Relativity, including most of the famous ‘‘paradoxes’’
and other ‘‘weirdness.’’ The mathematics are all algebra-level.]
Kip Thorne, Black Holes & Time Warps: Einstein’s Outrageous Legacy, 1995,
W. W. Norton & Company. [An excellent mid-level introduction to General
Relativity, black holes, evidence for them in the Universe, and their many
strange properties. Excellent coverage of a topic that was just touched upon in
class!]
Neil DeGrasse Tyson & Donald Goldsmith, Origins: Fourteen Billion Years of Cosmic
Evolution, 2004, W. W. Norton & Company.
William H. Waller & Paul W. Hodge, Galaxies and the Cosmic Frontier, 2003,
Harvard University Press. [Covers all aspects of galaxies, from morphologies
to composition to formation and evolution. The subject of early chapters
concentrate on the Milky Way, and then gradually move focus to the Magellanic
Clouds, the Local Group, and so on until we get to the superclusters and large
scale structures in the Universe.]
152 FURTHER READING
Index
4
INDEX 5