MASSACHUSETTS INSTITUTE OF TECHNOLOGY

Physics Department
Physics 8.286: The Early Universe August 30, 2020
Prof. Alan Guth
PROBLEM SET 1
DUE DATE: Friday, September 11, 2020, 5:00 pm.
READING ASSIGNMENT: The First Three Minutes, Chapters 1 and 2.
NOTE ABOUT EXTRA CREDIT: This problem set contains 40 points of regular
problems and 15 points extra credit, so it is probably worthwhile for me to clarify the
operational definition of “extra credit”. We will keep track of the extra credit grades
separately, and at the end of the course I will first assign provisional grades based
solely on the regular coursework. I will consult with our teaching assistant, Bruno
Scheihing, and we will try to make sure that these grades are reasonable. Then I will
add in the extra credit, allowing the grades to change upwards accordingly. Finally,
Bruno and I will look at each student’s grades individually, and we might decide to
give a higher grade to some students who are slightly below a borderline. Students
whose grades have improved significantly during the term, students whose average
has been pushed down by single low grade, and students who have been affected by
adverse personal or medical problems will be the ones most likely to be boosted.
The bottom line is that the extra credit problems are OPTIONAL. You should
feel free to skip them, and you will still get an excellent grade in the course if you
do well on the regular problems. However, if you have some time and enjoy an extra
challenge, then I hope that you will find the extra credit problems interesting and
worthwhile.

PROBLEM 1: NONRELATIVISTIC DOPPLER SHIFT, SOURCE AND

OBSERVER IN MOTION (15 points)

Consider the Doppler shift of sound waves, for a case in which both the source and
the observer are moving. Suppose the source is moving with a speed vs relative to the
air, while the observer is receding from the source, moving in the opposite direction with
speed vo relative to the air. Calculate the Doppler shift z. (Recall that z is defined by
1 + z ≡ λo /λs , where λo and λs are the wavelengths as measured by the observer and
by the source, respectively.) Hint: while this problem can be solved directly, you can
save time by finding a way to determine the answer by using the cases that are already
calculated in Lecture Notes 1.
8.286 PROBLEM SET 1, FALL 2020 p. 2

PROBLEM 2: THE TRANSVERSE DOPPLER SHIFT (25 points)

Consider the Doppler shift observed by a stationary observer, from a source that
travels in a circular orbit of radius R about the observer. Let the speed of the source be
v.

(a) (5 points) If the wave in question is sound, and both the source speed v and the
wave speed u are very small compared to the speed of light c, what is the Doppler
shift z? Assume that the observer is at rest relative to the air.

(b) (5 points) If the wave is light, traveling with speed c, and v is not small compared
to c, what is the Doppler shift z? This is called the transverse Doppler shift, since
the velocity of the light ray is perpendicular to the velocity of the source at the time
of emission, as seen in the reference frame of the observer.

(c) (5 points) Still considering light waves and the same pattern of
motion as shown in the figure, suppose that the source and the
observer were reversed. That is, suppose a light ray is sent from
the person at the center of the circle to the person traveling
around the circle at speed v. In this case, what would be the
Doppler shift z?

(d) (5 points) Now suppose that the motion is linear instead of

circular. Again we consider light rays, and as in part (b)
we assume that the source is moving with a speed v that is
not small compared to c. If the light ray is emitted by the
source at the moment of its closest approach to the observer,
as shown in the diagram, what is the Doppler shift z?

(e) (5 points) Again consider linear motion, with light rays. As

in part (c), assume that the observer is moving with a speed
v that is not small compared to c. If the light ray is received
by the observer at the moment of its closest approach to the
source, as shown in the diagram, what is the Doppler shift
z?
8.286 PROBLEM SET 1, FALL 2020 p. 3

PROBLEM 3: A HIGH-SPEED MERRY-GO-ROUND

(This problem is not required, but can be done for 15 points extra credit.)
Now consider the Doppler shift as it would be observed in a high-speed “merry-go-
round.” Four evenly-spaced cars travel around a central hub at speed v, all at a distance
R from a central hub. Each car is sending waves to all three of the other cars.

(a) If the wave in question is sound, and both the source speed v and the wave speed u
are very small compared to the speed of light c, with what Doppler shift z does a
given car receive the sound from (i) the car in front of it; (ii) the car behind it; and
(iii) the car opposite it?
(b) In the relativistic situation, where the wave is light and the speed v may be compa-
rable to c, what is the answer to the same three parts (i)-(iii) above?

Total points for Problem Set 1: 40, plus 15 points of extra credit.
 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe September 12, 2020
Prof. Alan Guth
PROBLEM SET 2
DUE DATE: Friday, September 18, 2020, 5:00 pm.

SEPTEMBER/OCTOBER
MON TUES WED THURS FRI
September 14 15 16 17 18
Class 3 Class 4

21 22 23 24 25
Class 5 Class 6 PS 3 due
PS 2 due
28 29 30 October 1 2
Class 7 Quiz 1
— “in class”

READING ASSIGNMENT: Barbara Ryden, Introduction to Cosmology, Chapters

1-3.
PLANNING AHEAD: If you want to read ahead, the reading assignment with Prob-
lem Set 3 will be Weinberg, The First Three Minutes, Chapter 3. Problem Sets
1 through 3, including the reading assignments, will be included in the material
covered on Quiz 1, on Wednesday, September 30.

INTRODUCTION TO THE PROBLEM SET

In this problem set we will consider a universe in which the scale factor is given by
a(t) = bt2/3 ,
where b is an arbitrary constant of proportionality which should not appear in the answers
to any of the questions below. (We will see in Lecture Notes 3 that this is the behavior of
a flat universe with a mass density that is dominated by nonrelativistic matter.) We will
suppose that a distant galaxy is observed with a redshift z. As a concrete example we
will consider the most distant known object with a well-determined redshift, the galaxy
GN-z11, which has a redshift z = 11.1. The discovery of this galaxy was announced in
March 2016 by an international group of astronomers, using the Hubble Space Telescope*

* P. A. Oesh et al., “A Remarkably Luminous Galaxy at z = 11.1 Measured with Hub-

ble Space Telescope Grism Spectroscopy,” The Astrophysical Journal 819, 129 (2016),
https://arxiv.org/abs/1603.00461.
8.286 PROBLEM SET 2, FALL 2020 p. 2

The rate at which the highest measured redshift

has been growing has been dramatic. In 1960 the
highest measured redshift was only z=0.461. The di-
agram at the right shows a graph of the highest con-
firmed redshift by year of discovery, using the listing
in the Wikipedia‡ The red circles represent quasars,
the green circles represent galaxies, and the one or-
ange circle at 2009 is a gamma ray burst. The search
for high redshift objects continues to be an exciting
area of research, as astronomers try to sort out the
conditions in the universe when the first galaxies be-
gan to form.

PROBLEM 1: DISTANCE TO THE GALAXY (10 points)

Let t0 denote the present time, and let te denote the time at which the light that we
are currently receiving was emitted by the galaxy. In terms of these quantities, find the
present value of the physical distance `p between this distant galaxy and us.

PROBLEM 2: TIME OF EMISSION (10 points)

Express the redshift z in terms of t0 and te . Find the ratio te /t0 for the z = 11.1
galaxy.

PROBLEM 3: DISTANCE IN TERMS OF REDSHIFT z (10 points)

Express the present value of the physical distance in terms of the present value of
the Hubble expansion rate H0 and the redshift z. Taking H0 ≈ 67 km-sec−1 -Mpc−1 , how
far away is the galaxy? Express your answer both in light-years and in Mpc.

PROBLEM 4: SPEED OF RECESSION (10 points)

Find the present rate at which the physical distance `p between the distant galaxy
and us is changing. Express your answer in terms of the redshift z and the speed of light
c, and evaluate it numerically for the case z = 11.1. Express your answer as a fraction of
the speed of light. [If you get it right, this “fraction” is greater than one! Our expanding
universe violates special relativity, but is consistent with general relativity.]

‡ “List of the most distant astronomical objects.” In Wikipedia, The Free Encyclo-
pedia. Retrieved 16:35, September 12, 2020, from https://en.wikipedia.org/wiki/
List_of_the_most_distant_astronomical_objects.
8.286 PROBLEM SET 2, FALL 2020 p. 3

PROBLEM 5: APPARENT ANGULAR SIZES (20 points)

Now suppose for simplicity that the galaxy is spherical, and that its physical diameter
was w at the time it emitted the light. (The actual galaxy is seen as an unresolved point
source, so we don’t know it’s actual size and shape.) Find the apparent angular size θ
(measured from one edge to the other) of the galaxy as it would be observed from Earth
today. Express your answer in terms of w, z, H0 , and c. You may assume that θ  1.
Compare your answer to the apparent angular size of a circle of diameter w in a static
Euclidean space, at a distance equal to the present value of the physical distance to the
galaxy, as found in Problem 1. [Hint: draw diagrams which trace the light rays in the
comoving coordinate system. If you have it right, you will find that θ has a minimum
value for z = 1.25, and that θ increases for larger z. This phenomenon makes sense if
you think about the distance to the galaxy at the time of emission. If the galaxy is very
far away today, then the light that we now see must have left the object very early, when
it was rather close to us!]

PROBLEM 6: RECEIVED RADIATION FLUX

(This problem is not required, but can be done for 15 points extra credit.)
At the time of emission, the galaxy had a power output P (measured, say, in ergs/sec)
which was radiated uniformly in all directions. This power was emitted in the form of
photons. What is the radiation energy flux J from this galaxy at the earth today? Energy
flux (which might be measured in ergs-cm−2 -sec−1 ) is defined as the energy per unit area
per unit time striking a surface that is orthogonal to the direction of energy flow. The
easiest way to solve this problem is to consider the trajectories of the photons, as viewed
in comoving coordinates. You must calculate the rate at which photons arrive at the
detector, and you must also use the fact that the energy of each photon is proportional
to its frequency, and is therefore decreased by the redshift. You may find it useful to
8.286 PROBLEM SET 2, FALL 2020 p. 4

think of the detector as a small part of a sphere that is centered on the source, as shown
in the following diagram:

Total points for Problem Set 2: 60, plus 15 points of extra credit.
 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe September 19, 2020
Prof. Alan Guth
PROBLEM SET 3
DUE DATE: Friday, September 25, 2020, 5:00 pm.

READING ASSIGNMENT: Steven Weinberg, The First Three Minutes, Chapter 3.

SHORT-TERM CALENDAR:

SEPTEMBER/OCTOBER
MON TUES WED THURS FRI
September 14 15 16 17 18
Class 3 Class 4

21 22 23 24 25
Class 5 Class 6 PS 3 due
PS 2 due
28 29 30 October 1 2
Class 7 Quiz 1
— “in class”

QUIZ DATES FOR THE TERM:

Quiz 1: Wednesday, September 30, 2020
Quiz 2: Wednesday, October 28, 2020
Quiz 3: Wednesday, December 2, 2020

FIRST QUIZ: The first of three quizzes for the term will be given on Wednesday,
September 30, 2020.
Coverage: Lecture Notes 1, 2, and 3; Problem Sets 1, 2, and 3; Weinberg, Chapters
1, 2, and 3; Ryden, Chapters 1, 2, and 3. While all of Ryden’s Chapter 3 has been
assigned, questions on the quiz will be limited to Sections 3.1 (The Way of Newton) and
3.3 (The General Way of Einstein). Section 3.2 (The Special Way of Einstein) describes
special relativity. Ryden’s approach is somewhat different from our Lecture Notes 1 —
for the quiz, you will be responsible only for the issues discussed in Lecture Notes 1. The
material in Sections 3.4–3.6 will be discussed in lecture later in the course, and you will
not be responsible for it until then.
Quiz Logistics: The quiz will be closed book, no calculators, no internet, and 85 minutes
long. I assume that most of you will take it during our regular class time on September
30, but you will have the option of starting it any time during a 24-hour window from
8.286 PROBLEM SET 3, FALL 2020 p. 2

11:05 am EDT on September 30 to 11:05 am EDT on Thursday, October 1. If you want to

start later than 11:05 am 9/30/2020, you should email me your choice of starting time by
11:59 pm on 9/29/2020. The quiz will be contained in a PDF file, which I am planning to
distribute by email. You will each be expected to spend up to 85 minutes working on it,
and then you will upload your answers to Canvas as a PDF file. I won’t place any precise
time limit on scanning or photographing and uploading, because the time needed for that
can vary. If you have questions about the meaning of the questions, I will be available
on Zoom during the September 30 class time, and we will arrange for either Bruno or
me to be available by email as much as possible during the other quiz times. If you
have any special circumstances that might make this procedure difficult, or if you need a
postponement beyond the 24-hour window, please let me ([email protected]) know.

PROBLEM 1: A CYLINDRICAL UNIVERSE (25 points)

The following problem originated on Quiz 2 of 1994, where it counted 30 points.

The lecture notes showed a construction of a Newtonian model of the universe that
was based on a uniform, expanding, sphere of matter. In this problem we will construct
a model of a cylindrical universe, one which is expanding in the x and y directions but
which has no motion in the z direction. Instead of a sphere, we will describe an infinitely
long cylinder of radius Rmax,i , with an axis coinciding with the z-axis of the coordinate
system:

We will use cylindrical coordinates, so

p
r= x2 + y 2

and
~r
~r = xı̂ + y̂ ; r̂ = ,
r
8.286 PROBLEM SET 3, FALL 2020 p. 3

where ı̂, ̂, and k̂ are the usual unit vectors along the x, y, and z axes. We will assume
that at the initial time ti , the initial density of the cylinder is ρi , and the initial velocity
of a particle at position ~r is given by the Hubble relation

~vi = Hi~r .

(a) (5 points) By using Gauss’ law of gravity, it is possible to show that the gravitational
acceleration at any point is given by

Aµ
~g = − r̂ ,
r
where A is a constant and µ is the total mass per length contained within the radius
r. Evaluate the constant A.
(b) (5 points) As in the lecture notes, we let r(ri , t) denote the trajectory of a particle
that starts at radius ri at the initial time ti . Find an expression for r̈(ri , t), expressing
the result in terms of r, ri , ρi , and any relevant constants. (Here an overdot denotes
a time derivative.)
(c) (5 points) Defining
r(ri , t)
u(ri , t) ≡ ,
ri
show that u(ri , t) is in fact independent of ri . This implies that the cylinder will
undergo uniform expansion, just as the sphere did in the case discussed in the lecture
notes. As before, we define the scale factor a(t) ≡ u(ri , t).
(d) (5 points) Express the mass density ρ(t) in terms of the initial mass density ρi and
the scale factor a(t). Use this expression to obtain an expression for ä in terms of a,
ρ, and any relevant constants.
(e) (5 points) Find an expression for a conserved quantity of the form

1 2
E= ȧ + V (a) .
2

What is V (a)? Will this universe expand forever, or will it collapse?

8.286 PROBLEM SET 3, FALL 2020 p. 4

PROBLEM 2: A FLAT UNIVERSE WITH UNUSUAL TIME EVOLUTION

(10 points)

Consider a flat universe which is filled with some peculiar form of matter, so that
the Robertson–Walker scale factor behaves as

a(t) = bt3/4 ,

where b is a constant.
(a) (5 points) For this universe, find the value of the Hubble expansion rate H(t).
(b) (5 points) What is the mass density of the universe, ρ(t)? (In answering this question,
you will need to know that the equation for ȧ/a in Lecture Notes 3,
 2
ȧ 8π kc2

   = Gρ − 2 ,
a 3 a
holds for all forms of matter, while the equation for ä,

4π
ä = − Gρ(t)a ,
3
requires modification if the matter has a significant pressure. The ä equation is
therefore not applicable to this problem.)

PROBLEM 3: ENERGY AND THE FRIEDMANN EQUATION (30 points)

The Friedmann equation,

 2
ȧ 8π kc2

   = Gρ − 2 , (1)
a 3 a
was derived in Lecture Notes 3 as a first integral of the equations of motion. The equation
was first derived in a different form,

1 2 4π Gρi
E= ȧ − = constant, (2)
2 3 a
where k = −2E/c2 . In this form the equation looks more like a conservation of energy
relation, although the constant E does not have the dimensions of energy. There are two
ways, however, in which the quantity E can be connected to the conservation of energy.
It is related the energy of a test particle that moves with the Hubble expansion, and it
is also related to the total energy of the entire expanding sphere of radius Rmax , which
was discussed in Lecture Notes 3 as a method of deriving the Friedmann equations. In
this problem you will derive these relations.
8.286 PROBLEM SET 3, FALL 2020 p. 5

First, to see the relation with the energy of a test particle moving with the Hubble
expansion, define a physical energy Ephys by

Ephys ≡ mri2 E , (3)

where m is the mass of the test particle and ri is its initial radius. Note that the
gravitational force on this particle is given by

GmM (ri )
F~ = − ~ eff (r) ,
r̂ = −∇V (4)
r2
where M (ri ) is the total mass initially contained within a radius ri of the origin, r is the
present distance of the test particle from the origin, and the “effective” potential energy
Veff (r) is given by
GmM (ri )
Veff (r) = − . (5)
r
The motivation for calling this quantity the “effective” potential energy will be explained
below.
(a) (10 points) Show that Ephys is equal to the “effective” energy of the test particle,
defined by
1
Eeff = mv 2 + Veff (r) . (6)
2

We understand that Eeff is conserved because it is the energy in an analogue problem in

which the test particle moves in the gravitational field of a point particle of mass M (ri ),
located at the origin, with potential energy function Veff (r). In this analogue problem the
force on the test particle is exactly the same as in the real problem, but in the analogue
problem the energy of the test particle is conserved.
We call (6) the “effective” energy because it is really the energy of the analogue
problem, and not the real problem. The true potential energy V (r, t) of the test particle
is defined to be the amount of work we must supply to move the particle to its present
location from some fixed reference point, which we might take to be r = ∞. We will
not bother to write V (r, t) explicitly, since we will not need it, but we point out that it
depends on the time t and on Rmax , and when differentiated gives the correct gravitational
force at any radius. By contrast, Veff (r) gives the correct force only at the radius of the
test particle, r = a(t)ri . The true potential energy function V (r, t) gives no conservation
law, since it is explicitly time-dependent, which is why the quantity Veff (r) is useful.
To relate E to the total energy of the expanding sphere, we need to integrate over the
sphere to determine its total energy. These integrals are most easily carried out by
dividing the sphere into shells of radius r, and thickness dr, so that each shell has a
volume
dV = 4πr2 dr . (7)
8.286 PROBLEM SET 3, FALL 2020 p. 6

(b) (10 points) Show that the total kinetic energy K of the sphere is given by
 
2 1 2
K= cK M Rmax,i ȧ (t) , (8)
2

where cK is a numerical constant, M is the total mass of the sphere, and Rmax,i is
the initial radius of the sphere. Evaluate the numerical constant cK .
(c) (10 points) Show that the total potential energy of the sphere can similarly be written
as  
2 4π ρi
U = cU M Rmax,i − G . (9)
3 a
(Suggestion: calculate the total energy needed to assemble the sphere by bringing
in one shell of mass at a time from infinity.) Show that cU = cK , so that the total
energy of the sphere is given by

2
Etotal = cK M Rmax,i E. (10)

PROBLEM 4: A POSSIBLE MODIFICATION OF NEWTON’S LAW OF

GRAVITY (20 points)

READ THIS: This problem was Problem 2 of Quiz 1 of 2011, and the solution is
posted as http://web.mit.edu/8.286/www/quiz11/ecqs1-1.pdf. Unlike the situation with
other problems, in this case you are encouraged to look at these solutions and benefit from
them. When you write your solution, you can even copy it verbatim from these solutions
if you wish, although obviously you will learn more if you think about the solution and
write your own version.
In Lecture Notes 3 we developed a Newtonian model of cosmology, by considering a
uniform sphere of mass, centered at the origin, with initial mass density ρi and an initial
pattern of velocities corresponding to Hubble expansion: ~vi = Hi~r:
8.286 PROBLEM SET 3, FALL 2020 p. 7

We denoted the radius at time t of a particle which started at radius ri by the

function r(ri , t). Assuming Newton’s law of gravity, we concluded that each particle
would experience an acceleration given by
GM (ri )
~g = − r̂ ,
r2 (ri , t)
where M (ri ) denotes the total mass contained initially in the region r < ri , given by
4π 3
M (ri ) = r ρi .
3 i

Suppose that the law of gravity is modified to contain a new, repulsive term, pro-
ducing an acceleration which grows as the nth power of the distance, with a strength
that is independent of the mass. That is, suppose ~g is given by
GM (ri )
~g = − r̂ + γrn (ri , t) r̂ ,
r2 (ri , t)
where γ is a constant. The function r(ri , t) then obeys the differential equation
GM (ri )
r̈ = − + γrn (ri , t) .
r2 (ri , t)

(a) (4 points) As done in the lecture notes, we define

u(ri , t) ≡ r(ri , t)/ri .
Write the differential equation obeyed by u. (Hint: be sure that u is the only time-
dependent quantity in your equation; r, ρ, etc. must be rewritten in terms of u, ρi ,
etc.)
(b) (4 points) For what value of the power n is the differential equation found in part
(a) independent of ri ?
(c) (4 points) Write the initial conditions for u which, when combined with the differ-
ential equation found in (a), uniquely determine the function u.
(d) (8 points) If all is going well, then you have learned that for a certain value of n, the
function u(ri , t) will in fact not depend on ri , so we can define
a(t) ≡ u(ri , t) .
Show, for this value of n, that the differential equation for a can be integrated once
to obtain an equation related to the conservation of energy. The desired equation
should include terms depending on a and ȧ, but not ä or any higher derivatives.

Total points for Problem Set 3: 85.

 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe October 4, 2020
Prof. Alan Guth
PROBLEM SET 4

DUE DATE: Friday, October 9, 2020, at 5:00 pm.

READING ASSIGNMENT: Steven Weinberg, The First Three Minutes, Chapter
4; Barbara Ryden, Introduction to Cosmology, Chapters 4 and Chapter 5 through
Section 5.4.1 (Matter + Curvature). (In the first edition, Chapters 4, 5, and Section
6.1.) In Weinberg’s Chapter 4 (and, later, Chapter 5) there are a lot of numbers
mentioned. You certainly do not need to learn all these numbers, but you should be
familiar with the orders of magnitude. In Ryden’s Chapters 4 and 5, the material
parallels what we either have done or will be doing in lecture. For these chapters
you should consider Ryden’s book as an aid to understanding the lecture material,
and not as a source of new material. On the upcoming quizzes, there will be no
questions based specifically on the material in these chapters.

SHORT-TERM CALENDAR:

OCTOBER/NOVEMBER
MON TUES WED THURS FRI
October 5 6 7 8 9
Class 9 Class 10 PS 4 due

October 12 13 14 15 16
Columbus Day Class 11 Class 12 PS 5 due

October 19 20 21 22 23
Class 13 Class 14 PS 6 due

October 26 27 28 29 30
Class 15 Class 16
Quiz 2
8.286 PROBLEM SET 4, FALL 2020 p. 2

QUIZ DATES FOR THE TERM:

Quiz 1: Wednesday, September 30, 2020
Quiz 2: Wednesday, October 28, 2020
Quiz 3: Wednesday, December 2, 2020

PROBLEM 1: PHOTON TRAJECTORIES AND HORIZONS IN A FLAT

UNIVERSE WITH a(t) = bt1/2 (20 points)

The following questions all pertain to a flat universe, with a scale factor given by

a(t) = bt1/2 ,

where b is a constant and t is the time. We will learn later that this is the behavior of a
radiation-dominated flat universe.
(a) (2 points) If physical lengths are measured in meters, and coordinate lengths are
measured in notches, what are the units of a(t) and the constant b?
(b) (2 points) Find the Hubble expansion rate H(t).
(c) (2 points) Find the physical horizon distance `p,hor (t). Your answer should give
the horizon distance in physical units (e.g., meters) and not coordinate units (e.g.,
notches).
Consider two pieces of comoving matter, A and B, at a coordinate distance `c from each
other. We will consider a photon that is emitted by A at some early time tA , traveling
toward B. The physical distance between A and B at the time of emission is of course
1/2
`p,AB (tA ) = btA `c , which approaches zero as tA → 0.
(d) (2 points) What is the rate of change of the physical distance between A and B,
d`p,AB (t)/dt, at t = tA ? Is the physical distance increasing or decreasing? Does the
rate of change approach zero, infinity, negative infinity, or a nonzero finite number
as tA → 0?
(e) (3 points) At what time tB is the photon received by B? As tA → 0, does tB
approach zero, infinity, or a nonzero finite number?
(f) (3 points) Calculate `p,γB (t), the physical distance between the photon and B at
time t, for tA ≤ t ≤ tB .
(g) (3 points) What is the rate of change of the physical distance between the photon
and B, d`p,γB (t)/dt, at the instant tA when the photon is emitted?
(h) (3 points) At what value of tA is this rate of change d`p,γB (t)/dt equal to zero? For
earlier values of tA , is the physical distance between the photon and B increasing
or decreasing at the time of emission? As tA → 0, does d`p,γB (t)/dt at the time of
emission approach zero, infinity, minus infinity, or a nonzero finite number?
8.286 PROBLEM SET 4, FALL 2020 p. 3

PROBLEM 2: EVOLUTION OF AN OPEN, MATTER-DOMINATED UNI-

VERSE (35 points)

The following problem originated on Quiz 2 of 1992 (ancient history!), where it counted
30 points.
The equations describing the evolution of an open, matter-dominated universe were
given in Lecture Notes 4 as
ct = α (sinh θ − θ)

and
a
√ = α (cosh θ − 1) ,
κ
where α is a constant with units of length. The following mathematical identities, which
you should know, may also prove useful on parts (e) and (f):

eθ − e−θ eθ + e−θ
sinh θ = , cosh θ =
2 2

θ θ2 θ3
eθ = 1 + + + + ... .
1! 2! 3!

a) (5 points) Find the Hubble expansion rate H as a function of α and θ.

b) (5 points) Find the mass density ρ as a function of α and θ.
c) (5 points) Find the mass density parameter Ω as a function of α and θ. As with
part (c) of the previous problem, the answer to this part appears in Lecture Notes 4.
However, you should show that you get the same answer by combining your answers
to parts (a) and (b) of this question.
d) (6 points) Find the physical value of the horizon distance, `p,horizon , as a function of
α and θ.
e) (7 points) For very small values of t, it is possible to use the first nonzero term of a
power-series expansion to express θ as a function of t, and then a as a function of t.
Give the expression for a(t) in this approximation. The approximation will be valid
for t  t∗ . Estimate the value of t∗ .
f) (7 points) Even though these equations describe an open universe, one still finds that
Ω approaches one for very early times. For t  t∗ (where t∗ is defined in part (e)),
the quantity 1 − Ω behaves as a power of t. Find the expression for 1 − Ω in this
approximation.
8.286 PROBLEM SET 4, FALL 2020 p. 4

PROBLEM 3: THE CRUNCH OF A CLOSED, MATTER-DOMINATED

UNIVERSE (25 points)

This is Problem 5.7 (Problem 6.5 in the first edition) from Barbara Ryden’s Introduc-
tion to Cosmology, with some paraphrasing to make it consistent with the language
used in lecture.
Consider a closed universe containing only nonrelativistic matter. This is the closed
universe discussed in Lecture Notes 4, and it is also the “Big Crunch” model discussed
in Ryden’s section Section 5.4.1 (Section 6.1 in the first edition). At some time during
the contracting phase (i.e., when θ > π), an astronomer named Elbbuh Niwde discovers
that nearby galaxies have blueshifts (−1 ≤ z < 0) proportional to their distance. He
then measures the present values of the Hubble expansion rate, H0 , and the mass density
parameter, Ω0 . He finds, of course, that H0 < 0 (because he is in the contracting phase)
and Ω0 > 1 (because the universe is closed). In terms of H0 and Ω0 , how long a time
will elapse between Dr. Niwde’s observation at t = t0 and the final Big Crunch at
t = tCrunch = 2πα/c? Assuming that Dr. Niwde is able to observe all objects within his
horizon, what is the most blueshifted (i.e., most negative) value of z that Dr. Niwde is
able to see? What is the lookback time to an object with this blueshift? (By lookback
time, one means the difference between the time of observation t0 and the time at which
the light was emitted.)

PROBLEM 4: THE AGE OF A MATTER-DOMINATED UNIVERSE AS

Ω → 1 (15 points)

The age t of a matter-dominated universe, for any value of Ω, was given in Lecture
Notes 4 as
  √  √ 
Ω 2 1 − Ω 2 1 − Ω
− arcsinh if Ω < 1


2(1 − Ω)3/2 Ω Ω





|H| t = 2/3 if Ω = 1 (4.47)

   √  √ 
Ω 2 Ω−1 2 Ω−1


arcsin ± ∓


 if Ω > 1
2(Ω − 1)3/2 Ω Ω

It was claimed that this formula is continuous at Ω = 1. In this problem you are asked to
show half of this statement. Specifically, you should show that as Ω approaches 1 from
below, the expression for |H|t approaches 2/3. In doing this, you may find it useful to
use the Taylor expansion for arcsinh(x) about x = 0:

(1)2 3 (3 · 1)2 5 (5 · 3 · 1)2 7

arcsinh(x) = x − x + x − x + ... .
3! 5! 7!
The proof of continuity as Ω → 0 from above is of course very similar, and you are not
asked to show it.
8.286 PROBLEM SET 4, FALL 2020 p. 5

PROBLEM 5: ISOTROPY ABOUT TWO POINTS IN EUCLIDEAN

SPACES

(This problem is not required, but can be done for 15 points extra credit. I’d like to give
you two weeks to think about it, so you should turn it in with Problem Set 5 on October
16.)
In Steven Weinberg’s The First Three Minutes, in Chapter 2 on page 24, he gives
an argument to show that if a space is isotropic about two distinct points, then it is
necessarily homogeneous. He is assuming Euclidean geometry, although he is not explicit
about this point. (The statement is simply not true if one allows non-Euclidean spaces
— we’ll discuss this.) Furthermore, the argument is given in the context of a universe
with only two space dimensions, but it could easily be generalized to three, and we will
not concern ourselves with remedying this simplification. The statement is true for two-
dimensional Euclidean spaces, but Weinberg’s argument is not complete. To show that
isotropy about two galaxies, 1 and 2, implies that the conditions at any two points A
and B must be identical, he constructs two circles. One circle is centered on Galaxy 1
and goes through A, and the other is centered on Galaxy 2 and goes through B. He then
argues that the conditions at A and B must both be identical to the conditions at the
point C, where the circles intersect. The problem, however, is that the two circles need
not intersect. One circle can be completely inside the other, or the two circles can be
separated and disjoint. Thus Weinberg’s proof is valid for some pairs of points A and
B, but cannot be applied to all cases. For 15 points of extra credit, devise a proof that
holds in all cases. We have not established axioms for Euclidean geometry, but you may
use in your proof any well-known fact about Euclidean geometry.

Total points for Problem Set 4: 95, plus 15 points of extra credit.
 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe October 10, 2020
Prof. Alan Guth
PROBLEM SET 5

DUE DATE: Friday, October 16, 2020, at 5:00 pm.

READING ASSIGNMENT: Steven Weinberg, The First Three Minutes, Chapters 5
and 6, and also Barbara Ryden, Introduction to Cosmology, Chapter 7 (Dark Matter)
(First edition: Chapter 8).

SHORT-TERM CALENDAR:

OCTOBER
MON TUES WED THURS FRI
October 12 13 14 15 16
Indigenous Class 11 Class 12 PS 5 due
Peoples Day

19 20 21 22 23
Class 13 Class 14 PS 6 due

26 27 28 29 30
Class 15 Class 16
Quiz 2
8.286 PROBLEM SET 5, FALL 2020 p. 2

PROBLEM 1: A CIRCLE IN A NON-EUCLIDEAN GEOMETRY

(15 points)

Consider a three-dimensional space described by the following metric:

dr2
 
2 2
+ r2 dθ2 + sin2 θdφ2


ds = R 2
.
1 − kr

Here R and k are constants, where k will always have one of the values 1, −1, or 0. θ
and φ are angular coordinates with the usual properties: 0 ≤ θ ≤ π and 0 ≤ φ ≤ 2π,
where φ = 2π and φ = 0 are identified. r is a radial coordinate, which runs from 0 to 1 if
k = 1, and otherwise from 0 to ∞. (This is the Robertson-Walker metric of Eq. (5.27) of
Lecture Notes 5, evaluated at some particular time t, with R ≡ a(t). You should be able
to work this problem, however, whether or not you have gotten that far. The problem
requires only that you understand what a metric means.) Consider a circle described by
the equations

z=0
x2 + y 2 = r02 ,

or equivalently by the angular coordinates

r = r0
θ = π/2 .

(a) (5 points) Find the circumference S of this circle. Hint: break the circle into in-
finitesimal segments of angular size dφ, calculate the arc length of such a segment,
and integrate.
(b) (5 points) Find the radius ρ of this circle. Note that ρ is the length of a line which
runs from the origin to the circle (r = r0 ), along a trajectory of θ = π/2 and φ =
constant. Hint: Break the line into infinitesimal segments of coordinate length dr,
8.286 PROBLEM SET 5, FALL 2020 p. 3

calculate the length of such a segment, and integrate. Consider the case of open and
closed universes separately, and take k = ±1. (If you don’t remember why we can
take k = ±1, see the section called “Units” in Lecture Notes 3,). You will want the
following integrals:
Z
dr
√ = sin−1 r
1−r 2

and
Z
dr
√ = sinh−1 r .
1 + r2

(c) (5 points) Express the circumference S in terms of the radius ρ. This result is
independent of the coordinate system which was used for the calculation, since S
and ρ are both measurable quantities. Since the space described by this metric
is homogeneous and isotropic, the answer does not depend on where the circle is
located or on how it is oriented. For the two cases of open and closed universes,
state whether S is larger or smaller than the value it would have for a Euclidean
circle of radius ρ.

PROBLEM 2: VOLUME OF A CLOSED UNIVERSE (15 points)

Calculate the total volume of a closed universe, as described by the metric of

Eq. (5.14) of Lecture Notes 5:

ds2 = R2 dψ 2 + sin2 ψ dθ2 + sin2 θdφ2



.

Break the volume up into spherical shells of infinitesimal thickness, extending from ψ to
ψ + dψ:
8.286 PROBLEM SET 5, FALL 2020 p. 4

By comparing Eq. (5.14) with Eq. (5.8),

ds2 = R2 dθ2 + sin2 θ dφ2

,

the metric for the surface of a sphere, one can see that as long as ψ is held fixed, the
metric for varying θ and φ is the same as that for a spherical surface of radius R sin ψ.
Thus the area of the spherical surface is 4πR2 sin2 ψ. To find the volume, multiply this
area by the thickness of the shell (which you can read off from the metric), and then
integrate over the full range of ψ, from 0 to π.

PROBLEM 3: SURFACE BRIGHTNESS IN A CLOSED UNIVERSE (25

points)

The spacetime metric for a homogeneous, isotropic, closed universe is given by the
Robertson-Walker formula:

dr2
 
2 2 2 2 2 2 2 2 2 2


ds = −c dτ = −c dt + a (t) + r dθ + sin θ dφ ,
1 − r2

where I have taken k = 1. To discuss motion in the radial direction, it is more convenient
to work with an alternative radial coordinate ψ, related to r by

r = sin ψ .

Then
dr
√ = dψ ,
1 − r2
so the metric simplifies to

ds2 = −c2 dτ 2 = −c2 dt2 + a2 (t) dψ 2 + sin2 ψ dθ2 + sin2 θ dφ2



.

The form of a(t) depends on the nature of the matter in the universe, but for this problem
you should consider a(t) to be an arbitrary function. You should simplify your answers
as far as it is possible without knowing the function a(t).
(a) (10 points) Suppose that the Earth is at the center of these coordinates, and that
we observe a spherical galaxy that is located at ψ = ψG . The light that we see was
emitted from the galaxy at time tG , and is being received today, at a time that we
call t0 . At the time of emission, the galaxy had a power output P (which could
be measured, for example, in watts, where 1 watt = 1 joule/sec). The power was
radiated uniformly in all directions, in the form of photons. What is the radiation
energy flux J from this galaxy at the Earth today? Energy flux (which might be
measured in joule-m−2 -sec−1 ) is defined as the energy per unit area per unit time
8.286 PROBLEM SET 5, FALL 2020 p. 5

striking a surface that is orthogonal to the direction of the energy flow. [Hint: it is
easiest to use a comoving coordinate system with the radiating galaxy at the origin.]
(b) (10 points) Suppose that the physical diameter of the galaxy at time tG was w. Find
the apparent angular size ∆θ (measured from one edge to the other) of the galaxy
as it would be observed from Earth today.

(c) (5 points) The surface brightness σ of the distant galaxy is defined to be the energy
flux J per solid angle subtended by the galaxy.* Calculate the surface brightness σ
of the galaxy described in parts (a) and (b). [Hint: if you have the right answer, it
can be written in terms of P , w, and the redshift z, without any reference to ψG .
The rapid decrease in σ with z means that high-z galaxies are difficult to distinguish
from the night sky.]

PROBLEM 4: TRAJECTORIES AND DISTANCES IN AN OPEN UNI-

VERSE (30 points)

The spacetime metric for a homogeneous, isotropic, open universe is given by the
Robertson-Walker formula:

dr2
 
2 2 2 2 2 2 2 2 2 2


ds = −c dτ = −c dt + a (t) + r dθ + sin θ dφ ,
1 + r2

where I have taken k = −1. As in Problem 3, for the discussion of radial motion it is
convenient to introduce an alternative radial coordinate ψ, which in this case is related
to r by
r = sinh ψ .

* Definition of solid angle: To define the solid angle subtended by the galaxy, imagine
surrounding the observer by a sphere of arbitrary radius r. The sphere should be small
compared to cosmological distances, so that Euclidean geometry is valid within the sphere.
If a picture of the galaxy is painted on the surface of the sphere so that it just covers the
real image, then the solid angle, in steradians, is the area of the picture on the sphere,
divided by r2 .
8.286 PROBLEM SET 5, FALL 2020 p. 6

Then
dr
√ = dψ ,
1 + r2
so the metric simplifies to

ds2 = −c2 dτ 2 = −c2 dt2 + a2 (t) dψ 2 + sinh2 ψ dθ2 + sin2 θ dφ2



.

You should treat the function a(t) as a given function. You should simplify your answers
as far as it is possible without knowing explicitly the function a(t).

(a) (5 points) Suppose that the Earth is at the origin of the coordinate system (ψ = 0),
and that at the present time, t0 , we receive a light pulse from a distant galaxy G,
located at ψ = ψG . Write down an equation which determines the time tG at which
the light pulse left the galaxy. (You may assume that the light pulse travels on a
“null” trajectory, which means that dτ = 0 for any segment of it. Since you don’t
know a(t) you cannot solve this equation, so please do not try.)

(b) (5 points) What is the redshift zG of the light from galaxy G? (Your answer may
depend on tG , as well as ψG , t0 , or any property of the function a(t).)

(c) (5 points) To estimate the number of galaxies that one expects to see in a given
range of redshifts, it is necessary to know the volume of the region of space that
corresponds to this range. Write an expression for the present value of the volume
that corresponds to redshifts smaller than that of galaxy G. (You may leave your
answer in the form of a definite integral, which may be expressed in terms of ψG ,
tG , t0 , zG , or the function a(t).)

(d) (5 points) There are a number of different ways of defining distances in cosmology,
and generally they are not equal to each other. One choice is called proper dis-
tance, which corresponds to the distance that one could in principle measure with
rulers. The proper distance is defined as the total length of a network of rulers that
are laid end to end from here to the distant galaxy. The rulers have different veloci-
ties, because each is at rest with respect to the matter in its own vicinity. They are
arranged so that, at the present instant of time, each ruler just touches its neighbors
on either side. Write down an expression for the proper distance `prop of galaxy G.

(e) (5 points) Another common definition of distance is angular size distance, de-
termined by measuring the apparent size of an object of known physical size. In a
static, Euclidean space, a small sphere of diameter w at a distance ` will subtend an
8.286 PROBLEM SET 5, FALL 2020 p. 7

angle ∆θ = w/`:

Motivated by this relation, cosmologists define the angular size distance `ang of an
object by
w
`ang ≡ .
∆θ
What is the angular size distance `ang of galaxy G?
(f) (5 points) A third common definition of distance is called luminosity distance,
which is determined by measuring the apparent brightness of an object for which
the actual total power output is known. In a static, Euclidean space, the energy flux
J received from a source of power P at a distance ` is given by J = P/(4π`2 ):

Cosmologists therefore define the luminosity distance by

r
P
`lum ≡ .
4πJ
Find the luminosity distance `lum of galaxy G. (Hint: the Robertson-Walker coor-
dinates can be shifted so that the galaxy G is at the origin.)
8.286 PROBLEM SET 5, FALL 2020 p. 8

PROBLEM 5: THE KLEIN DESCRIPTION OF THE G-B-L GEOMETRY

(This problem is not required, but can be done for 15 points extra credit.)
I stated in Lecture Notes 5 that the space invented by Klein, described by the distance
relation  
d(1, 2) 1 − x1 x2 − y1 y2
cosh =p p ,
a 1 − x21 − y12 1 − x22 − y22
where
x2 + y 2 < 1 ,
is a two-dimensional space of constant negative curvature. In other words, this is just a
two-dimensional Robertson–Walker metric, as would be described by a two-dimensional
version of Eq. (5.27), with k = −1:
dr2
 
2 2 2 2
ds = a + r dθ .
1 + r2
The problem is to prove the equivalence.
(a) (5 points) As a first step, show that if x and y are replaced by the polar coordinates
defined by
x = u cos θ
y = u sin θ ,
then the distance equation can be rewritten as
 
d(1, 2) 1 − u1 u2 cos(θ1 − θ2 )
cosh = p p .
a 1 − u21 1 − u22

(b) (5 points) The next step is to derive the metric from the distance function above.
Let
u1 = u θ1 = θ ,

u2 = u + du θ2 = θ + dθ ,
and
d(1, 2) = ds .
Insert these expressions into the distance function, expand everything to second order
in the infinitesimal quantities, and show that
( )
2 2 2
du u dθ
ds2 = a2 2 + 1 − u2 .
2
(1 − u )
(This part is rather messy, but you should be able to do it.)
(c) (5 points) Now find the relationship between r and u and show that the two metric
functions are identical. Hint: The coefficients of dθ2 must be the same in the two
cases. Can you now see why Klein had to impose the condition x2 + y 2 < 1?
8.286 PROBLEM SET 5, FALL 2020 p. 9

REMINDER: The following extra credit problem from Problem Set 4 is to be turned
in with this problem set, if you choose to do it:

PROBLEM 5 (PROBLEM SET 4): ISOTROPY ABOUT TWO POINTS IN

EUCLIDEAN SPACES

(This problem is not required, but can be done for 15 points extra credit. It was first
posted with Problem Set 4, but is to be turned in with Problem Set 5.)
In Steven Weinberg’s The First Three Minutes, in Chapter 2 on page 24, he gives
an argument to show that if a space is isotropic about two distinct points, then it is
necessarily homogeneous. He is assuming Euclidean geometry, although he is not explicit
about this point. (The statement is simply not true if one allows non-Euclidean spaces
— we’ll discuss this.) Furthermore, the argument is given in the context of a universe
with only two space dimensions, but it could easily be generalized to three, and we will
not concern ourselves with remedying this simplification. The statement is true for two-
dimensional Euclidean spaces, but Weinberg’s argument is not complete. To show that
isotropy about two galaxies, 1 and 2, implies that the conditions at any two points A
and B must be identical, he constructs two circles. One circle is centered on Galaxy 1
and goes through A, and the other is centered on Galaxy 2 and goes through B. He then
argues that the conditions at A and B must both be identical to the conditions at the
point C, where the circles intersect. The problem, however, is that the two circles need
not intersect. One circle can be completely inside the other, or the two circles can be
separated and disjoint. Thus Weinberg’s proof is valid for some pairs of points A and
B, but cannot be applied to all cases. For 15 points of extra credit, devise a proof that
holds in all cases. We have not established axioms for Euclidean geometry, but you may
use in your proof any well-known fact about Euclidean geometry.

Total points for Problem Set 5: 85, plus up to 15 points extra credit.
Also up to 15 points extra credit for Problem Set 4.
 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe October 17, 2020
Prof. Alan Guth
PROBLEM SET 6

DUE DATE: Friday, October 23, 2020, at 5:00 pm.

READING ASSIGNMENT: Steven Weinberg, The First Three Minutes, Chapter 7

(The First One-Hundredth Second), and also Barbara Ryden, Introduction to Cos-
mology, Chapter 9 (Nucleosynthesis and the Early Universe) [First edition: Chapter
10]. For now we are skipping Chapter 8 (The Cosmic Microwave Background), but
we will come back to it. Chapter 9 contains some references back to Eqs. (8.26)
and (8.28) of Section 8.3 (The Physics of Recombination), but I will post some notes
that will give you the necessary information. Ryden’s treatment in Sections 9.2 and
9.3 of particle number densities in thermal equilibrium is not quite correct, as she
leaves out the role of chemical potentials. Her statement in footnote 3 on p. 170
that “chemical potentials are small enough to be safely neglected” is in fact wildly
incorrect. Equations for number densities such as Eqs. (9.11) and (9.12) [(10.11) and
(10.12) in the first edition] are incorrect, but all of the equations for ratios of num-
ber densities (such as (9.13), (9.25), or (9.26) [(10.13), (10.25), or (10.26) in the first
edition])) are correct, with the chemical potential factors canceling. The notes that
will be posted will correct these errors, and will hopefully give a clear explanation
of how chemical potentials are treated.

CALENDAR FOR THE REST OF THE TERM:

OCTOBER/NOVEMBER
MON TUES WED THURS FRI
19 20 21 22 23
Class 13 Class 14 PS 6 due

26 27 28 29 30
Class 15 Class 16
Quiz 2

November 2 3 4 5 6
Class 17 Class 18 PS 7 due
8.286 PROBLEM SET 6, FALL 2020 p. 2

NOVEMBER/DECEMBER
MON TUES WED THURS FRI
9 10 11 12 13
Class 19 Veteran’s
Day

16 17 18 19 20
Class 20 Class 21 PS 8 due

23 24 25 26 27
Thanksgiving — — — —
Week

30 December 1 2 3 4
Class 22 Class 23
Quiz 3

7 8 9 10 11
Class 24 Class 25
PS 9 due
Last Class

PROBLEM 1: GEODESICS IN A FLAT UNIVERSE (25 points)

According to general relativity, in the absence of any non-gravitational forces a

particle will travel along a spacetime geodesic. In this sense, gravity is reduced to a
distortion in spacetime.
Consider the case of a flat (i.e., k = 0) Robertson–Walker metric, which has the
simple form
ds2 = −c2 dt2 + a2 (t) dx2 + dy 2 + dz 2 .
 

Since the spatial metric is flat, we have the option of writing it in terms of Cartesian
rather than polar coordinates. Now consider a particle which moves along the x-axis.
(Note that the galaxies are on the average at rest in this system, but one can still discuss
the trajectory of a particle which moves through the model universe.)
(a) (8 points) Use the geodesic equation to show that the coordinate velocity computed
with respect to proper time (i.e., dx/dτ ) falls off as 1/a2 (t).
(b) (8 points) Use the expression for the spacetime metric to relate dx/dt to dx/dτ .
8.286 PROBLEM SET 6, FALL 2020 p. 3

(c) (9 points) The physical velocity of the particle relative to the galaxies that it is
passing is given by
dx
v = a(t) .
dt
(Note that this is just a generalization of what we have previously said for photons,
dx/dt = c/a(t).) Show that the momentum of the particle, defined relativistically
by
mv
p= p
1 − v 2 /c2
falls off as 1/a(t). (This implies, by the way, that if the particle were described as
a quantum mechanical wave with wavelength λ = h/|~ p |, then its wavelength would
stretch with the expansion of the universe, in the same way that the wavelength of
light is redshifted.)

PROBLEM 2: METRIC OF A STATIC GRAVITATIONAL FIELD (25

points)

In this problem we will consider the metric

3
2
 2  2 X
2
ds = − c + 2φ(~x) dt + dxi ,
i=1

which describes a static gravitational field. Here i runs from 1 to 3, with the identifications
x1 ≡ x, x2 ≡ y, and x3 ≡ z. The function φ(~x) depends only on the spatial variables
~x ≡ (x1 , x2 , x3 ), and not on the time coordinate t.
(a) (5 points) Suppose that a radio transmitter, located at ~xe , emits a series of evenly
spaced pulses. The pulses are separated by a proper time interval ∆Te , as measured
by a clock at the same location. What is the coordinate time interval ∆te between
the emission of the pulses? (I.e., ∆te is the difference between the time coordinate
t at the emission of one pulse and the time coordinate t at the emission of the next
pulse.)
(b) (5 points) The pulses are received by an observer at ~xr , who measures the time
of arrival of each pulse. What is the coordinate time interval ∆tr between the
reception of successive pulses?
(c) (5 points) The observer uses his own clocks to measure the proper time interval ∆Tr
between the reception of successive pulses. Find this time interval, and also the
redshift z, defined by
∆Tr
1+z = .
∆Te
8.286 PROBLEM SET 6, FALL 2020 p. 4

First compute an exact expression for z, and then expand the answer to lowest order
in φ(~x) to obtain a weak-field approximation. (This weak-field approximation is in
fact highly accurate in all terrestrial and solar system applications.)
(d) (5 points) A freely falling particle travels on a spacetime geodesic xµ (τ ), where τ is
the proper time. (I.e., τ is the time that would be measured by a clock moving with
the particle.) The trajectory is described by the geodesic equation

dxν dxλ dxσ

 
d 1
gµν = (∂µ gλσ ) ,
dτ dτ 2 dτ dτ

where the Greek indices (µ, ν, λ, σ, etc.) run from 0 to 3, and are summed over when
repeated. Calculate an explicit expression for

d2 xi
,
dτ 2

valid for i = 1, 2, or 3. (It is acceptable to leave quantities such as dt/dτ or dxi /dτ
in the answer.)
(e) (5 points) In the weak-field nonrelativistic-velocity approximation, the answer to the
previous part reduces to
d2 xi
= −∂i φ ,
dt2
so φ(~x) can be identified as the Newtonian gravitational potential. Use this fact to
estimate the gravitational redshift z of a photon that rises from the floor of this room
to the ceiling (say 4 meters). (One significant figure will be sufficient.)

PROBLEM 3: CIRCULAR ORBITS IN A SCHWARZSCHILD METRIC (30

points)

READ THIS: This problem was Problem 16 of Review Problems for Quiz 2 of 2011,
and the solution is posted as http://web.mit.edu/8.286/www/quiz11/ecqr2-1.pdf. Like
Problem 4 of Problem Set 3, but unlike all other homework problems so far, in this case
you are encouraged to look at the solutions and benefit from them. When you write
your solution, you can even copy it verbatim from these solutions if you wish, although
obviously you will learn more if you think about the solution and write your own version.
The Schwarzschild metric, which describes the external gravitational field of any
spherically symmetric distribution of mass (including black holes), is given by
   −1
2GM 2GM
2 2 2
ds = −c dτ = − 1 − 2 2
c dt + 1 − dr2 + r2 dθ2 + r2 sin2 θ dφ2 ,
rc2 rc2
8.286 PROBLEM SET 6, FALL 2020 p. 5

where M is the total mass of the object, 0 ≤ θ ≤ π, 0 ≤ φ < 2π, and φ = 2π is

identified with φ = 0. We will be concerned only with motion outside the Schwarzschild
horizon RS = 2GM/c2 , so we can take r > RS . (This restriction allows us to avoid the
complications of understanding the effects of the singularity at r = RS .) In this problem
we will use the geodesic equation to calculate the behavior of circular orbits in this metric.
We will assume a perfectly circular orbit in the x-y plane: the radial coordinate r is fixed,
θ = 90◦ , and φ = ωt, for some angular velocity ω.
(a) (7 points) Use the metric to find the proper time interval dτ for a segment of the
path corresponding to a coordinate time interval dt. Note that dτ represents the
time that would actually be measured by a clock moving with the orbiting body.
Your result should show that
r
dτ 2GM r2 ω2
= 1− − .
dt rc2 c2
Note
p that for M = 0 this reduces to the special relativistic relation dτ /dt =
1 − v 2 /c2 , but the extra term proportional to M describes an effect that is new
with general relativity— the gravitational field causes clocks to slow down, just as
motion does.
(b) (7 points) Show that the geodesic equation of motion (Eq. (5.65)) for one of the
coordinates takes the form
 2  2
1 ∂gφφ dφ 1 ∂gtt dt
0= + .
2 ∂r dτ 2 ∂r dτ

(c) (8 points) Show that the above equation implies

 2  2
dφ GM dt
r = 2 ,
dτ r dτ
which in turn implies that
GM
rω 2 = .
r2
Thus, the relation between r and ω is exactly the same as in Newtonian mechanics.
[Note, however, that this does not really mean that general relativity has no effect.
First, ω has been defined by dφ/dt, where t is a time coordinate which is not the
same as the proper time τ that would be measured by a clock on the orbiting body.
Second, r does not really have the same meaning as in the Newtonian calculation,
since it is not the measured distance from the center of motion. Measured distances,
you will recall, are calculated by integrating the metric, as for example in Problem
1 of Problem Set 5, A Circle in a Non-Euclidean Geometry. Since the angular (dθ2
and dφ2 ) terms in the Schwarzschild metric are unaffected by the mass, however, it
can be seen that the circumference of the circle is equal to 2πr, as in the Newtonian
calculation.]
(d) (8 points) Show that circular orbits around a black hole have a minimum value of
the radial coordinate r, which is larger than RS . What is it?
8.286 PROBLEM SET 6, FALL 2020 p. 6

PROBLEM 4: GAS PRESSURE AND ENERGY CONSERVATION (25

points)

In this problem we will pursue the implications of the conservation of energy. Con-
sider first a gas contained in a chamber with a movable piston, as shown below:

Let U denote the total energy of the gas, and let p denote the pressure. Suppose that the
piston is moved a distance dx to the right. (We suppose that the motion is slow, so that
the gas particles have time to respond and to maintain a uniform pressure throughout
the volume.) The gas exerts a force pA on the piston, so the gas does work dW = pAdx
as the piston is moved. Note that the volume increases by an amount dV = Adx, so
dW = pdV . The energy of the gas decreases by this amount, so

dU = −pdV . (P4.1)

It turns out that this relation is valid whenever the volume of a gas is changed, regardless
of the shape of the volume.
Now consider a homogeneous, isotropic, expanding universe, described by a scale
factor a(t). Let u denote the energy density of the gas that fills it. (Remember that
u = ρc2 , where ρ is the mass density of the gas.) We will consider a fixed coordinate
volume Vcoord , so the physical volume will vary as

Vphys (t) = a3 (t)Vcoord . (P4.2)

The energy of the gas in this region is then given by

U = Vphys u . (P4.3)

(a) (9 points) Using these relations, show that

d 3 2
d
a ρc = −p (a3 ) , (P4.4)
dt dt
8.286 PROBLEM SET 6, FALL 2020 p. 7

and then that

ȧ  p
ρ̇ = −3 ρ+ 2 , (P4.5)
a c
where the dot denotes differentiation with respect to t.
(b) (8 points) The scale factor evolves according to the relation
 2
ȧ 8π kc2
= Gρ − 2 . (P4.6)
a 3 a
Using Eqs. (P4.5) and (P4.6), show that
 
4π 3p
ä = − G ρ + 2 a . (P4.7)
3 c
This equation describes directly the deceleration of the cosmic expansion. Note that
there are contributions from the mass density ρ, but also from the pressure p.
(c) (8 points) So far our equations have been valid for any sort of a gas, but let us now
specialize to the case of black-body radiation. For this case we know that ρ = bT 4 ,
where b is a constant and T is the temperature. We also know that as the universe
expands, aT remains constant. Using these facts and Eq. (P4.5), find an expression
for p in terms of ρ.
PROBLEM 5: THE EFFECT OF PRESSURE ON COSMOLOGICAL EVO-
LUTION (25 points)
A radiation-dominated universe behaves differently from a matter-dominated uni-
verse because the pressure of the radiation is significant. In this problem we explore the
role of pressure for several fictitious forms of matter.
(a) (8 points) For the first fictitious form of matter, the mass density ρ decreases as the
scale factor a(t) grows, with the relation
1
ρ(t) ∝ 6 .
a (t)
What is the pressure of this form of matter? [Hint: the answer is proportional to
the mass density.]
(b) (9 points) Find the behavior of the scale factor a(t) for a flat universe dominated
by the form of matter described in part (a). You should be able to determine the
function a(t) up to a constant factor.
(c) (8 points) Now consider a universe dominated by a different form of fictitious matter,
with a pressure given by
1
p = ρc2 .
2
As the universe expands, the mass density of this form of matter behaves as
1
ρ(t) ∝ n .
a (t)
Find the power n.
8.286 PROBLEM SET 6, FALL 2020 p. 8

PROBLEM 6: TIME EVOLUTION OF A UNIVERSE WITH MYSTERI-

OUS STUFF (15 points)

Suppose that a model universe is filled with a peculiar form of matter for which

1
ρ∝ .
a5 (t)

Assuming that the model universe is flat, calculate

(a) (4 points) The behavior of the scale factor, a(t). You should be able to find a(t) up
to an arbitrary constant of proportionality.
(b) (3 points) The value of the Hubble parameter H(t), as a function of t.
(c) (4 points) The physical horizon distance, `p,horizon (t).
(d) (4 points) The mass density ρ(t).

Total points for Problem Set 6: 145.

 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe October 31, 2020
Prof. Alan Guth
PROBLEM SET 7

DUE DATE: Friday, November 6, 2020, at 5:00 pm.

READING ASSIGNMENT: Steven Weinberg, The First Three Minutes, Chapter 8
(Epilogue: The Prospect Ahead), and Afterword: Cosmology Since 1977. Barbara
Ryden, Introduction to Cosmology, Chapter 8 [First edition: Chapter 7] (The Cosmic
Microwave Background).

CALENDAR FOR THE REST OF THE TERM:

NOVEMBER
MON TUES WED THURS FRI
November 2 3 4 5 6
Class 17 Class 18 PS 7 due

9 10 11 12 13
Class 19 Veteran’s
Day

16 17 18 19 20
Class 20 Class 21 PS 8 due

23 24 25 26 27
Thanksgiving — — — —
Week

30 December 1 2 3 4
Class 22 Class 23
Quiz 3

7 8 9 10 11
Class 24 Class 25
PS 9 due
Last Class
8.286 PROBLEM SET 7, FALL 2020 p. 2

PROBLEM 1: EFFECT OF AN EXTRA NEUTRINO SPECIES (15 points)

According to the standard assumptions (which were used in the lecture notes), there
are three species of effectively massless neutrinos. In the temperature range of 1 MeV <
kT < 100 MeV, the mass density of the universe is believed to have been dominated by
the black-body radiation of photons, electron-positron pairs, and these neutrinos, all of
which were in thermal equilibrium.
(a) (5 points) Under these assumptions, how long did it take (starting from the instant
of the big bang) for the temperature to fall to the value such that kT = 1 MeV? (In
this part and the next, you may assume that the period when kT > 100 MeV was so
short that one can calculate as if the value of g that you find for 1 MeV < kT < 100
MeV can be used for earlier times as well.)
(b) (5 points) How much time would it have taken if there were one other species of
massless neutrino, in addition to the three which we are currently assuming?
(c) (5 points) What would be the mass density of the universe when kT = 1 MeV under
the standard assumptions, and what would it be if there were one other species of
massless neutrino?

PROBLEM 2: ENTROPY AND THE BACKGROUND NEUTRINO TEM-

PERATURE (15 points)

The formula for the entropy density of black-body radiation is given in Lecture Notes
6. The derivation of this formula has been left to the statistical mechanics course that
you either have taken or hopefully will take. For our purposes, the important point is
that the early universe remains very close to thermal equilibrium, and therefore entropy
is conserved. The conservation of entropy applies even during periods when particles,
such as electron-positron pairs, are “freezing out” of the thermal equilibrium mix. Since
total entropy is conserved, the entropy density falls off as 1/a3 (t).
When the electron-positron pairs disappear from the thermal equilibrium mixture as
kT falls below me c2 = 0.511 MeV, the weak interactions have such low cross sections that
the neutrinos have essentially decoupled. To a good approximation, all of the energy and
entropy released by the annihilation of electrons and positrons is added to the photon
gas, and the neutrinos are unaffected. Use the conservation of entropy to show that as
electron-positron pair annihilation takes place, aTγ increases by a factor of (11/4)1/3 ,
while aTν remains constant. It follows that after the disappearance of the electron-
positron pairs, Tν /Tγ = (4/11)1/3 . As far as we know, nothing happens that significantly
affects this ratio right up to the present day. So we expect today a background of thermal
neutrinos which are slightly colder than the 2.7◦ K background of photons.
8.286 PROBLEM SET 7, FALL 2020 p. 3

Added note: In principle the heating of the photon gas due to electron-positron
annihilation can also be calculated by using energy conservation, but it is much more
difficult. Since  p
ρ̇ = −3H ρ + 2
c
(this was Eq. (6.36) of Lecture Notes 6), one needs to know p(t) to understand the changes
in energy density. But as the electron-positron pairs are disappearing, kT is comparable
to the electon rest mass me c2 , and the formula for the thermal equilibrium pressure under
these circumstances is complicated.

PROBLEM 3: FREEZE-OUT OF MUONS (25 points)

A particle called the muon seems to be essentially identical to the electron, except
that it is heavier— the mass/energy of a muon is 106 MeV, compared to 0.511 MeV for
the electron. The muon (µ− ) has the same charge as an electron, denoted by −e. There
is also an antimuon (µ+ ), analogous to the positron, with charge +e. The muon and
antimuon have the same spin as the electron. There is no known particle with a mass
between that of an electron and that of a muon.
(a) The formula for the energy density of black-body radiation, as given by Eq. (6.48)
of the lecture notes,
π 2 (kT )4
u=g ,
30 (h̄c)3
is written in terms of a normalization constant g. What is the value of g for the
muons, taking µ+ and µ− together?
(b) When kT is just above 106 MeV as the universe cools, what particles besides the
muons are contained in the thermal radiation that fills the universe? What is the
contribution to g from each of these particles?
(c) As kT falls below 106 MeV, the muons disappear from the thermal equilibrium
radiation. At these temperatures all of the other particles in the black-body radiation
are interacting fast enough to maintain equilibrium, so the heat given off from the
muons is shared among all the other particles. Letting a denote the Robertson-
Walker scale factor, by what factor does the quantity aT increase when the muons
disappear?

PROBLEM 4: THE REDSHIFT OF THE COSMIC MICROWAVE BACK-

GROUND (25 points)

It was mentioned in Lecture Notes 6 that the black-body spectrum has the peculiar
feature that it maintains its form under uniform redshift. That is, as the universe expands,
even if the photons do not interact with anything, they will continue to be described by a
black-body spectrum, although at a temperature that decreases as the universe expands.
Thus, even though the cosmic microwave background (CMB) has not been interacting
8.286 PROBLEM SET 7, FALL 2020 p. 4

significantly with matter since 350,000 years after the big bang, the radiation today still
has a black-body spectrum. In this problem we will demonstrate this important property
of the black-body spectrum.
The spectral energy density ρν (ν, T ) for the thermal (black-body) radiation of pho-
tons at temperature T was stated in Lecture Notes 6 as Eq. (6.75), which we can rewrite
as
16π 2 h̄ν 3 1
ρν (ν, T ) = 3 hν/kT
, (4.1)
c e −1
where h = 2πh̄ is Planck’s original constant. ρν (ν, T ) dν is the energy per unit volume
carried by photons whose frequency is in the interval [ν, ν + dν]. In this problem we will
assume that this formula holds at some initial time t1 , when the temperature had some
value T1 . We will let ρ̃(ν, t) denote the spectral distribution for photons in the universe,
which is a function of frequency ν and time t. Thus, our assumption about the initial
condition can be expressed as

ρ̃(ν, t1 ) = ρν (ν, T1 ) . (4.2)

The photons redshift as the universe expands, and to a good approximation the
redshift and the dilution of photons due to the expansion are the only physical effects
that cause the distribution of photons to evolve. Thus, using our knowledge of the
redshift, we can calculate the spectral distribution ρ̃(ν, t2 ) at some later time t2 > t1 . It
is not obvious that ρ̃(ν, t2 ) will be a thermal distribution, but in fact we will be able to
show that

ρ̃(ν, t2 ) = ρ ν, T (t2 ) , (4.3)
where in fact T (t2 ) will agree with what we already know about the evolution of T in a
radiation-dominated universe:
a(t1 )
T (t2 ) = T1 . (4.4)
a(t2 )

To follow the evolution of the photons from time t1 to time t2 , we can imagine select-
ing a region of comoving coordinates with coordinate volume Vc . Within this comoving
volume, we can imagine tagging all the photons in a specified infinitesimal range of fre-
quencies, those between ν1 and ν1 + dν1 . Recalling that the energy of each such photon
is hν, the number dN1 of tagged photons is then

ρ̃(ν1 , t1 ) a3 (t1 ) Vc dν1

dN1 = . (4.5)
hν1

(a) We now wish to follow the photons in this frequency range from time t1 to time t2 ,
during which time each photon redshifts. At the latter time we can denote the range
8.286 PROBLEM SET 7, FALL 2020 p. 5

of frequencies by ν2 to ν2 + dν2 . Express ν2 and dν2 in terms of ν1 and dν1 , assuming

that the scale factor a(t) is given.
(b) At time t2 we can imagine tagging all the photons in the frequency range ν2 to
ν2 + dν2 that are found in the original comoving region with coordinate volume
Vc . Explain why the number dN2 of such photons, on average, will equal dN1 as
calculated in Eq. (4.5).
(c) Since ρ̃(ν, t2 ) denotes the spectral energy density at time t2 , we can write

ρ̃(ν2 , t2 ) a3 (t2 ) Vc dν2

dN2 = , (4.6)
hν2

using the same logic as in Eq. (4.5). Use dN2 = dN1 to show that

a3 (t1 )
ρ̃(ν2 , t2 ) = ρ̃(ν1 , t1 ) . (4.7)
a3 (t2 )

Use the above equation to show that Eq. (4.3) is satisfied, for T (t) given by Eq. (4.4).

Total points for Problem Set 7: 80.

 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe November 7, 2020
Prof. Alan Guth
PROBLEM SET 8
DUE DATE: Friday, November 20, 2020, at 5:00 pm. This is the last problem set before
Quiz 3, which will be Wednesday, December 2. There will also be a Problem Set 9, to be
due Wednesday, December 9, the last day of classes.
READING ASSIGNMENT: Barbara Ryden, Introduction to Cosmology, Chapter
10 (Inflation and the Very Early Universe) [First edition: Chapter 11]. Also read
Inflation and the New Era of High-Precision Cosmology, by Alan Guth, written for
the MIT Physics Department annual newsletter, 2002. It is available at
http://web.mit.edu/physics/news/physicsatmit/physicsatmit_02_cosmology.pdf

The data quoted in the article about the nonuniformities of the cosmic microwave
background radiation has since been superceded by much better data, but the con-
clusions have not changed. They have only gotten stronger.

CALENDAR FOR THE REST OF THE TERM:

NOVEMBER/DECEMBER
MON TUES WED THURS FRI
9 10 11 12 13
Class 19 Veteran’s
Day

16 17 18 19 20
Class 20 Class 21 PS 8 due

23 24 25 26 27
Thanksgiving — — — —
Week

30 December 1 2 3 4
Class 22 Class 23
Quiz 3

7 8 9 10 11
Class 24 Class 25
PS 9 due
Last Class
8.286 PROBLEM SET 8, FALL 2020 p. 2

PROBLEM 1: BIG BANG NUCLEOSYNTHESIS (20 points)

The calculations of big bang nucleosynthesis depend on a large number of measured

parameters. Below you are asked to qualitatively describe the effects of changing some
of these parameters. Include a sentence or two to explain each of your answers. (These
topics have not been discussed in class, but you are expected to be able to answer the
questions on the basis of your readings in Weinberg’s and Ryden’s books.)
(a) (5 points) Suppose an extra neutrino species is added to the calculation. Would the
predicted helium abundance go up or down?
(b) (5 points) Suppose the weak interactions were stronger than they actually are, so that
the thermal equilibrium distribution between neutrons and protons were maintained
until kT ≈ 0.25 MeV. Would the predicted helium abundance be larger or smaller
than in the standard model?
(c) (5 points) Suppose the proton-neutron mass difference were larger than the actual
value of 1.29 MeV/c2 . Would the predicted helium abundance be larger or smaller
than in the standard calculation?
(d) (5 points) The standard theory of big bang nucleosynthesis assumes that the matter
in the universe was distributed homogeneously during the era of nucleosynthesis,
but the alternative possibility of inhomogeneous big-bang nucleosynthesis has been
discussed since the 1980s. Inhomogeneous nucleosynthesis hinges on the hypothesis
that baryons became clumped during a phase transition at t ≈ 10−6 second, when
the hot quark soup converted to a gas of mainly protons, neutrons, and in the early
stages, pions. The baryons would then be concentrated in small nuggets, with a
comparatively low density outside of these nuggets. After the phase transition but
before nucleosynthesis, the neutrons would have the opportunity to diffuse away from
these nuggets, becoming more or less uniformly distributed in space. The protons,
however, since they are charged, interact electromagnetically with the plasma that
fills the universe, and therefore have a much shorter mean free path than the neu-
trons. Most of the protons, therefore, remain concentrated in the nuggets. Does this
scenario result in an increase or a decrease in the expected helium abundance?
8.286 PROBLEM SET 8, FALL 2020 p. 3

PROBLEM 2: BRIGHTNESS VS. REDSHIFT WITH A POSSIBLE COS-

MOLOGICAL CONSTANT (25 points)

In Lecture Notes 7, we derived the relation between the power output P of a source
and the energy flux J, for the case of a closed universe:

P H02 |Ωk,0 |
J= ,
4π(1 + zS )2 c2 sin2 ψ(zS )

where
Z zS
dz
q
ψ(zS ) = |Ωk,0 | p .
0 Ωm,0 (1 + z)3 + Ωrad,0 (1 + z)4 + Ωvac,0 + Ωk,0 (1 + z)2

Here zS denotes the observed redshift, H0 denotes the present value of the Hub-
ble expansion rate, Ωm,0 , Ωrad,0 , and Ωvac,0 denote the present contributions to Ω
from nonrelativistic matter, radiation, and vacuum energy, respectively, and Ωk,0 ≡
1 − Ωm,0 − Ωrad,0 − Ωvac,0 .
(a) Derive the corresponding formula for the case of an open universe. You can of course
follow the same logic as the derivation in the lecture notes, but the solution you write
should be complete and self-contained. (I.e., you should NOT say “the derivation
is the same as the lecture notes except for . . . .”)
(b) Derive the corresponding formula for the case of a flat universe. Here there is of
course no need to repeat anything that you have already done in part (a). If you
wish you can start with the answer for an open or closed universe, taking the limit as
k → 0. The limit is delicate, however, because both the numerator and denominator
of the equation for J vanish as Ωk,0 → 0.

PROBLEM 3: AGE OF A UNIVERSE WITH MYSTERIOUS STUFF (20

points)

READ THIS: This problem was Problem 8 of Review Problems for Quiz 3 of 2011,
and the solution is posted as http://web.mit.edu/8.286/www/quiz11/ecqr3-1.pdf. Like
Problem 4 of Problem Set 3 and Problem 3 of Problem Set 6, but unlike all other homework
problems so far, in this case you are encouraged to look at the solutions and benefit from
them. When you write your solution, you can even copy it verbatim from these solutions
if you wish, although obviously you will learn more if you think about the solution and
write your own version.
Consider a universe that contains nonrelativistic matter, radiation, vacuum energy,
and the same mysterious stuff that was introduced in Problem 7 of Review√Problems
for Quiz 3, from 2011. Since the mass density of mysterious stuff falls off as 1/ V , where
V is the volume, it follows that in an expanding universe the mass density of mysterious
stuff falls off as 1/a3/2 (t).
8.286 PROBLEM SET 8, FALL 2020 p. 4

Suppose that you are given the present value of the Hubble expansion rate H0 , and
also the present values of the contributions to Ω ≡ ρ/ρc from each of the constituents:
Ωm,0 (nonrelativistic matter), Ωr,0 (radiation), Ωv,0 (vacuum energy density), and Ωms,0
(mysterious stuff). Our goal is to express the age of the universe t0 in terms of these
quantities.

(a) (10 points) Let x(t) denote the ratio

a(t)
x(t) ≡
a(t0 )

for an arbitrary time t. Write an expression for the total mass density of the universe
ρ(t) in terms of x(t) and the given quantities described above.

(b) (10 points) Write an integral expression for the age of the universe t0 . The expression
should depend only on H0 and the various contributions to Ω0 listed above (Ωm,0 ,
Ωr,0 , etc.), but it might include x as a variable of integration.

PROBLEM 4: SHARED CAUSAL PAST (20 points)

Recently several of my colleagues published a paper (Andrew S. Friedman, David

I. Kaiser, and Jason Gallicchio, “The Shared Causal Pasts and Futures of Cosmolog-
ical Events,” http://arxiv.org/abs/arXiv:1305.3943, Physical Review D, Vol. 88, arti-
cle 044038 (2013)) in which they investigated the causal connections in the standard
cosmological model. In particular, they calculated the present redshift z of a distant
quasar which has the property that a light signal, if sent from our own location at the
instant of the big bang, would have just enough time to reach the quasar and return
to us, so that we could see the reflection of the signal at the present time. They found
z = 3.65, using Ωmatter,0 = 0.315, Ωrad,0 = 9.29 × 10−5 , Ωvac,0 = 0.685 − Ωrad,0 , and
H0 = 67.3 km-s−1 -Mpc−1 . Feel free to read their paper if you like. Your job, however, is
to carry out an independent calculation to find out if they got it right.

(a) (15 points) Write an equation that determines this redshift z. The equation may
involve one or more integrals which are not evaluated, and the equation itself does
not have to be solved.

(b) (5 points) The integrals that should appear in your answer to part (a) can be eval-
uated numerically, and the whole equation you found in part (a) can be solved
numerically. Do this, and see how your z compares with 3.65.
8.286 PROBLEM SET 8, FALL 2020 p. 5

PROBLEM 5: MASS DENSITY OF VACUUM FLUCTUATIONS (25 points)

The energy density of vacuum fluctuations has been discussed qualitatively in lecture.
In this problem we will calculate in detail the energy density associated with quantum
fluctuations of the electromagnetic field. To keep the problem finite, we will not consider
all of space at once, but instead we will consider the electromagnetic field inside a cube
of side L, defined by coordinates
0≤x≤L,
0≤y≤L,
0≤z≤L.
Our goal, however, will be to compute the energy density in the limit as the size of the
box is taken to infinity.
(a) (10 points) The electromagnetic waves inside the box can be decomposed into a
Fourier sum of sinusoidal normal modes. Suppose we consider only modes that
extend up to a maximum wave number kmax , or equivalently modes that extend
down to a minimum wavelength λmin , where

2π
kmax = .
λmin

How many such modes are there? I do not expect an exact answer, but your approx-
imations should become arbitrarily accurate when λmin  L. (These mode counting
techniques are probably familiar to many of you, but in case they are not I have
attached an extended hint after part (c).)
(b) (10 points) When the electromagnetic field is described quantum mechanically, each
normal mode behaves exactly as a harmonic oscillator: if the angular frequency of
the mode is ω, then the quantized energy levels have energies given by

1


En = n + 2 h̄ω ,

where h̄ is Planck’s original constant divided by 2π, and n is an integer. The integer
n is called the “occupation number,” and is interpreted as the number of photons in
the specified mode. The minimum energy is not zero, but instead is 12 h̄ω, which is
the energy of the quantum fluctuations of the electromagnetic field. Assuming that
the mode sum is cut off at λmin equal to the Planck length (as defined in the Lecture
Notes), what is the total mass density of these quantum fluctuations?
(c) (5 points) How does the mass density of the quantum fluctuations of the electromag-
netic field compare with the critical density of our universe?
Extended Hint:
8.286 PROBLEM SET 8, FALL 2020 p. 6

The electromagnetic fields inside a closed box can be expanded as the sum of modes,
each of which has a sinusoidally varying time dependence, but the precise form of these
modes depends on the nature of the boundary conditions on the walls of the box. Phys-
ically reasonable boundary conditions, such as total reflection, are in fact difficult to use
in calculations. However, it is known that in the limit of an infinite-sized box, the na-
ture of the boundary conditions will not make any difference. We are therefore free to
choose the simplest boundary conditions that we can imagine, and for this purpose we
will choose periodic boundary conditions. That is, we will assume that the fields and
their normal derivatives on each wall are fixed to precisely match the fields and their
normal derivatives on the opposite wall.
To begin, we consider a wave in one dimension, moving at the speed of light. Such
waves are most easily described in terms of complex exponentials. If A(x, t) represents
the amplitude of the wave, then a sinusoidal wave moving in the positive x-direction can
be written as h i
ik(x−ct)
A(x, t) = Re Be ,

where B is a complex constant and k is a real constant. Defining ω = c|k|, waves in

either direction can be written as
h i
A(x, t) = Re Bei(kx−ωt) ,

where the sign of k determines the direction. To be periodic with period L, the parameter
k must satisfy
kL = 2πn ,
where n is an integer. So the spacing between modes is ∆k = 2π/L. The density of
modes dN/dk (i.e., the number of modes per interval of k) is then one divided by the
spacing, or 1/∆k, so
dN L
= (one dimension) .
dk 2π

In three dimensions, a sinusoidal wave can be written as

h i
~
A(~x, t) = Re Bei(k·~x−ωt) ,

where ω = c|~k|, and

kx L = 2πnx , ky L = 2πny , kz L = 2πnz ,

where nx , ny , and nz are integers. Thus, in three-dimensional ~k-space the allowed values
of ~k lie on a cubical lattice, with spacing 2π/L. In counting the modes, one should also
remember that for photons there is an extra factor of 2 associated with the fact that
electromagnetic waves have two possible polarizations for each allowed value of ~k.
8.286 PROBLEM SET 8, FALL 2020 p. 7

PROBLEM 6: PLOTTING THE SUPERNOVA DATA (EXTRA CREDIT, 20

pts)

The original data on the Hubble diagram based on Type Ia supernovae are found in
two papers. One paper is authored by the High Z Supernova Search Team,* and the other
is by the Supernova Cosmology Project.† More recent data from the High Z team, which
includes many more data points, can be found in Riess et al., http://arXiv.org/abs/astro-
ph/0402512.¶ (By the way, the lead author Adam Riess was an MIT undergraduate
physics major, graduating in 1992.)
You are asked to plot the data from either the 2nd or 3rd of these papers, and to
include on the graph the theoretical predictions for several cosmological models.
The plot will be similar to the plots contained in these papers, and to the plot on
p. 121 of Ryden’s book, showing a graph of (corrected) magnitude m vs. redshift z. Your
graph should include the error bars. If you plot the Perlmutter et al. data, you will be
plotting “effective magnitude” m vs. redshift z. The magnitude is related to the flux
J of the observed radiation by m = − 25 log10 (J) + const. The value of the constant
in this expression will not be needed. The word “corrected” refers both to corrections
related to the spectral sensitivity of the detectors and to the brightness of the supernova
explosions themselves. That is, the supernova at various distances are observed with
different redshifts, and hence one must apply corrections if the detectors used to measure
the radiation do not have the same sensitivity at all wavelengths. In addition, to improve
the uniformity of the supernova as standard candles, the astronomers apply a correction
based on the duration of the light output. Note that our ignorance of the absolute
brightness of the supernova, of the precise value of the Hubble constant, and of the
constant that appears in the definition of magnitude all combine to give an unknown but
constant contribution to the predicted magnitudes. The consequence is that you will be
able to move your predicted curves up or down (i.e., translate them by a fixed distance
along the m axis). You should choose the vertical positioning of your curve to optimize
your fit, either by eyeball or by some more systematic method.
If you choose to plot the data from the 3rd paper, Riess et al. 2004, then you should
see the note at the end of this problem.
For your convenience, the magnitudes and redshifts for the Supernova Cosmology
Project paper, from Tables 1 and 2, are summarized in a text file on the 8.286 web page.
The data from Table 5 of the Riess et al. 2004 paper, mentioned above, is also posted on
the 8.286 web page.

* http://arXiv.org/abs/astro-ph/9805201, later published as Riess et al., Astronomical

Journal 116, 1009 (1998).
† http://arXiv.org/abs/astro-ph/9812133, later published as Perlmutter et al., Astro-
physical Journal 517:565–586 (1999).
¶ Published as Astrophysical Journal 607:665-687 (2004).
8.286 PROBLEM SET 8, FALL 2020 p. 8

For the cosmological models to plot, you should include:

(i) A matter-dominated universe with Ωm = 1.
(ii) An open universe, with Ωm,0 = 0.3.
(iii) A universe with Ωm,0 = 0.3 and a cosmological constant, with Ωvac,0 = 0.7. (If you
prefer to avoid the flat case, you can use Ωvac,0 = 0.6. Or, if you want to compare
directly with Figure 4 of the Riess et al. (2004) paper, you should use Ωm,0 = 0.29,
Ωvac,0 = 0.71.)
You may include any other models if they interest you. You can draw the plot with either
a linear or a logarithmic scale in z. I would recommend extending your theoretical plot
to z = 3, if you do it logarithmically, or z = 2 if you do it linearly, even though the data
does not go out that far. That way you can see what possible knowledge can be gained
by data at higher redshift.
NOTE FOR THOSE PLOTTING DATA FROM RIESS ET AL. 2004:
Unlike the Perlmutter et al. data, the Riess et al. data is expressed in terms of the
distance modulus, which is a direct measure of the luminosity distance. The distance
modulus is defined both in the Riess et al. paper and in Ryden’s book (p. 118) [First
edition: p. 120] as  
dL
µ = 5 log10 + 25 ,
1 Mpc
where Ryden uses the notation m − M for the distance modulus, and dL is the luminosity
distance. The luminosity distance, in turn, is really a measure of the observed brightness
of the object. It is defined as the distance that the object would have to be located to
result in the observed brightness, if we were living in a static Euclidean universe. More
explicitly, if we lived in a static Euclidean universe and an object radiated power P in a
spherically symmetric pattern, then the energy flux J at a distance d would be

P
J= .
4πd2
That is, the power would be distributed uniformly over the surface of a sphere at radius
d. The luminosity distance is therefore defined as
r
P
dL = .
4πJ
Thus, a specified value of the distance modulus µ implies a definite value of the ratio
J/P .
In plotting a theoretical curve, you will need to choose a value for H0 . Riess et al. do
not specify what value they used, but I found that their curve is most closely reproduced
if I choose H0 = 66 km-sec−1 -Mpc−1 . This seems a little on the low side, since the value
8.286 PROBLEM SET 8, FALL 2020 p. 9

is usually estimated as 70–72 km-sec−1 -Mpc−1 , but Riess et al. emphasize that they
were not concerned with this value. They were concerned with the relative values of the
distance moduli, and hence the shape of the graph of the distance modulus vs. z. In their
own words, from Appendix A, “The zeropoint, distance scale, absolute magnitude of the
fiducial SN Ia or Hubble constant derived from Table 5 are all closely related (or even
equivalent) quantities which were arbitrarily set for the sample presented here. Their
correct value is not relevant for the analyses presented which only make use of differences
between SN Ia magnitudes. Thus the analysis are independent of the aforementioned
normalization parameters.”

Total points for Problem Set 8: 110, plus an optional 20 points of extra credit.
 MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Physics Department
Physics 8.286: The Early Universe November 28, 2020
Prof. Alan Guth
PROBLEM SET 9 (The Last!)
DUE DATE: Wednesday, December 9, 2020, at 12:30 pm.
READING ASSIGNMENT: None.

CALENDAR FOR THE REST OF THE TERM:

NOVEMBER/DECEMBER
MON TUES WED THURS FRI
23 24 25 26 27
Thanksgiving — — — —
Week

30 December 1 2 3 4
Class 22 Class 23
Quiz 3

7 8 9 10 11
Class 24 Class 25
PS 9 due
Last Class

PROBLEM 1: THE HORIZON PROBLEM (20 points)

The success of the big bang predictions for the abundances of the light elements
suggests that the universe was already in thermal equilibrium at one second after the big
bang. At this time, the region that later evolves to become the observed universe was,
in the context of the conventional (non-inflationary) cosmological model, many horizon
distances across. Try to estimate how many. You may assume that the universe is flat,
that it was radiation-dominated for t ∼< 50,000 yr, and for this crude estimate you can
also assume that it has been matter-dominated for all t ∼ > 50,000 yr, and that a(t)T (t)
was about equal to its present value for the whole period from 1 second to the present.

PROBLEM 2: THE FLATNESS PROBLEM (20 points)

Although we now know that Ω0 = 1 to an accuracy of about half a percent, let us
pretend that the value of Ω today is 0.1. It nonetheless follows that at 10−37 second
after the big bang (about the time of the grand unified theory phase transition), Ω must
have been extraordinarily close to one. Writing Ω = 1 − δ , estimate the value of δ at
t = 10−37 sec (using the standard cosmological model). You may again use any of the
approximations mentioned in Problem 1.
8.286 PROBLEM SET 9, FALL 2020 p. 2

PROBLEM 3: THE MAGNETIC MONOPOLE PROBLEM (20 points)

In Lecture Notes 9, we learned that Grand Unified Theories (GUTs) imply the
existence of magnetic monopoles, which form as “topological defects” (topologically sta-
ble knots) in the configuration of the Higgs fields that are responsible for breaking the
grand unified symmetry to the SU(3)×SU(2)×U(1) symmetry of the standard model
of particle physics. It was stated that if grand unified theories and the conventional
(non-inflationary) cosmological model were both correct, then far too many magnetic
monopoles would have been produced in the big bang. In this problem we will fill in the
mathematical steps of that argument.
At very high temperatures the Higgs fields oscillate wildly, so the fields average to
zero. As the temperature T falls, however, the system undergoes a phase transition.
The phase transition occurs at a temperature Tc , called the critical temperature, where
kTc ≈ 1016 GeV. At this phase transition the Higgs fields acquire nonzero expectation
values, and the grand unified symmetry is thereby spontaneously broken. The monopoles
are twists in the Higgs field expectation values, so the monopoles form at the phase
transition. Each monopole is expected to have a mass MM c2 ≈ 1018 GeV, where the
subscript “M ” stands for “monopole.” According to an estimate first proposed by T.W.B.
Kibble, the number density nM of monopoles formed at the phase transition is of order

nM ∼ 1/ξ 3 , (3.1)

where ξ is the correlation length of the field, defined roughly as the maximum distance
over which the field at one point in space is correlated with the field at another point in
space. The correlation length is certainly no larger than the physical horizon distance
at the time of the phase transition, and it is believed to typically be comparable to this
upper limit. Note that an upper limit on ξ is a lower limit on nM — there must be at
least of order one monopole per horizon-sized volume.
Assume that the particles of the grand unified theory form a thermal gas of blackbody
radiation, as described by Eq. (6.48) of Lecture Notes 6,

π 2 (kT )4
u=g , (3.2)
30 (h̄c)3

with gGUT ∼ 200. Further assume that the universe is flat and radiation-dominated from
its beginning to the time of the GUT phase transition, tGUT .
For each of the following questions, first write the answer in terms of physical con-
stants and the parameters Tc , MM , and gGUT , and then evaluate the answers numerically.
(a) (5 points) Under the assumptions described above, at what time tGUT does the phase
transition occur? Express your answer first in terms of symbols, and then evaluate
it in seconds.
8.286 PROBLEM SET 9, FALL 2020 p. 3

(b) (5 points) Using Eq. (3.1) and setting ξ equal to the horizon distance, estimate the
number density nM of magnetic monopoles just after the phase transition.
(c) (5 points) Calculate the ratio nM /nγ of the number of monopoles to the number of
photons immediately after the phase transition. Refer to Lecture Notes 6 to remind
yourself about the number density of photons. You may assume that the temperature
after the phase transition is still approximately Tc .
(d) (5 points) For topological reasons monopoles cannot disappear, but they form with an
equal number of monopoles and antimonopoles, where the antimonopoles correspond
to twists in the Higgs field in the opposite sense. Monopoles and antimonopoles can
annihilate each other, but estimates of the rate for this process show that it is
negligible. Thus, in the context of the conventional (non-inflationary) hot big bang
model, the ratio of monopoles to photons would be about the same today as it was
just after the phase transition. Use this assumption to estimate the contribution
that these monopoles would make to the value of Ω today.

PROBLEM 4: EXPONENTIAL EXPANSION OF THE INFLATIONARY

UNIVERSE (15 points)
Recall that the evolution of a Robertson-Walker universe is described by the equation
 2
ȧ 8π kc2
= Gρ − 2 . (4.1)
a 3 a
Suppose that the mass density ρ is given by the constant mass density ρf of the false
vacuum. For the case k = 0, the growing solution is given simply by
a(t) = const eχt , (4.2)
where r
8π
χ= Gρf (4.3)
3
and const is an arbitrary constant. Find the growing solution to this equation for an
arbitrary value of k. Be sure to consider both possibilities for the sign of k. You may
find the following integrals useful:
Z
dx
√ = sinh−1 x . (4.4a)
1+x 2
Z
dx
√ = sin−1 x . (4.4b)
1−x 2
Z
dx
√ = cosh−1 x . (4.4c)
x2 − 1
Show that for large times one has
a(t) ∝ eχt (4.5)
for all choices of k.
8.286 PROBLEM SET 9, FALL 2020 p. 4

PROBLEM 5: THE HORIZON DISTANCE FOR THE PRESENT UNI-

VERSE (25 points)
We have not discussed horizon distances since the beginning of Lecture Notes 4,
when we found that Z t
c
`p,horizon (t) = a(t) 0)
dt0 . (5.1)
0 a(t
This formula was derived before we discussed curved spacetimes, but the formula is valid
for any Robertson-Walker universe, whether it is open, closed, or flat.
(a) (5 points) Show that the formula above is valid for closed universes. Hint: write the
closed universe metric as it was written in Eq. (7.27):
ds2 = −c2 dt2 + ã2 (t) dψ 2 + sin2 ψ dθ2 + sin2 θ dφ2


, (5.2)
where
a(t)
ã(t) ≡ √ (5.3)
k
and ψ is related to the usual Robertson-Walker coordinate r by
√
sin ψ ≡ k r . (5.4)
Use the fact that the physical speed of light is c, or equivalently the fact that ds2 = 0
for any segment of the light ray’s trajectory.
(b) (20 points) The evaluation of the formula depends of course on the form of the
function a(t), which is governed by the Friedmann equations. For the Planck 2018
best fit to the parameters (see Table 7.1 of Lecture Notes 7, and Eq. (6.23) of Lecture
Notes 6),
H0 = 67.7 km · s−1 · Mpc−1 (5.5a)
Ωm,0 = 0.311 (5.5b)
Ωr,0 = 4.15 × 10−5 h−2
0 (Tγ,0 = 2.725 K)
= 9.05 × 10−5 (5.5c)
Ωvac,0 = 1 − Ωm,0 − Ωr,0 , (5.5d)
find the current horizon distance, expressed both in light-years and in Mpc. Hint:
find an integral expression for the horizon distance, similar to Eq. (7.23a) for the age
of the universe. Then do the integral numerically.
Note that the model for which you are calculating does not explicitly include inflation.
If it did, the horizon distance would turn out to be vastly larger. By ignoring the
inflationary era in calculating the integral of Eq. (5.1), we are finding an effective
horizon distance, defined as the present distance of the most distant objects that
we can in principle observe by using only photons that have left their sources after
the end of inflation. Photons that left their sources earlier than the end of inflation
have undergone incredibly large redshifts, so it is reasonable to consider them to be
completely unobservable in practice.
8.286 PROBLEM SET 9, FALL 2020 p. 5

PROBLEM 6: THE DEUTERIUM BOTTLENECK (30 points)

The “deuterium bottleneck” plays a major role in the description of big bang nucle-
osynthesis: all of the nuclear reactions involved in nucleosynthesis depend on deuterium
forming at the start, but, due to its small binding energy, deuterium does not become
stable until the temperature reaches a rather low value. In this problem we will explore
the statistical mechanics of the deuterium bottleneck. Ryden discusses this in Section
9.3 [First Edition: Section 10.3], Deuterium Synthesis, but the calculation here will be a
little more complete.
As discussed in Notes on Thermal Equilibrium (which I will refer to as NTE), a
dilute ideal gas of classical nonrelativistic particles of type X, in thermal equilibrium,
has a number density given by
3/2
mX c2
  
mX kT µ 
X
nX = ḡX 2 exp − exp , (6.1)
2πh̄ kT kT
where h̄ = h/2π, c, and k have their usual meanings: Planck’s constant, the speed of
light, and the Boltzmann constant. Here ḡX is the number of spin degrees of freedom
associated with the particle, mX is the mass of the particle, T is the temperature, and
µX is the chemical potential of the particle. Because the gas is dilute there are no
corrections associated with fermions and the Pauli exclusion principle, since it is highly
unlikely in any case for two particles to occupy the same quantum state. The dilute gas
approximation is valid provided that
 3/2
mX kT
nX  . (6.2)
2πh̄2
Eq. (6.1) is equivalent to Eq. (23) of NTE, or Eq. (8.26) of Ryden’s 2nd edition.
(a) (5 points) I mentioned in NTE that our textbook sometimes writes Eq. (6.1) in-
correctly, omitting the chemical potential factor. See for example Eqs. (9.11) and
(9.12). The author has a footnote about this (p. 170), which concludes that “in most
cosmological contexts, as it turns out, the chemical potential is small enough to be
safely neglected.” [First edition: Eqs. (10.11) and (10.12, footnote on p. 156.] We can
check this statement by using the author’s formula to calculate the proton density
at 3 minutes into the big bang, at the time of Steven Weinberg’s Fifth Frame, from
chapter 5 of The First Three Minutes. At that time the temperature was T = 109 K.
To find the right answer for comparison, we make use of the fact that the ratio of
the number density nb of baryons to the number density nγ of photons is given by*
nb
η≡ = (6.15 ± 0.35) × 10−10 (95% confidence). (6.3)
nγ

* Particle Data Group, Section 2: Astrophysical Constants and Parameters (2020),

https://pdg.lbl.gov/2020/reviews/rpp2020-rev-astrophysical-constants.pdf.
8.286 PROBLEM SET 9, FALL 2020 p. 6

According to Weinberg, at that time 14% of the baryons were neutrons, with 86%
protons. At the risk of appearing impertinent toward the author (but physicists are
known for their impertinence), I will phrase the question this way: By how many
kilo-orders of magnitude is the author’s formula for np in error?† (Be prepared to
have your calculators overflow — if they do, calculate the logarithm of the answer.)
(b) (15 points) For deuterium production, the relevant reaction is

n + p ←→ D + γ . (6.4)

Recall that chemical potentials are defined initially in terms of conserved quantities,
so the chemical potentials on both sides of any allowed process much match. Since
the photon carries no conserved quantities, its chemical potential must vanish. It
follows that µn + µp = µD . This equality implies that if we form the ratio
nD
, (6.5)
np nn

expressing each number density as in Eq. (6.1), then the chemical potential factors
will cancel out. (This is how the formula is normally used, and this is how Ryden uses
it on pp. 175–176 [First edition: pp. 180–181]. From here on Ryden’s treatment is
correct, but we will proceed with slightly more detail.) To describe the bookkeeping
for the reaction of Eq. (6.4), we need to define our variables. I am using nn , np ,
and nD to mean the number densities of free neutrons, free protons, and deuterium
nuclei. nb denotes the total baryon number density, so

nb = nn + np + 2nD . (6.6)

Finally, I will use nTOT

n and nTOT
p to denote the total number densities of neutrons
and protons respectively, whether free or bound inside deuterium. We assume that
deuterium production happens fast enough so that there is no further change in the
neutron-proton balance while deuterium is forming, so the ratio

nTOT
n
f≡ (6.7)
nb
is fixed. We will describe the extent to which the reaction has proceeded by specifying
the fraction x of neutrons that remain free,
nn
x≡ TOT
. (6.8)
nn

† I exchanged email with Barbara Ryden about this after the first edition came out,
and she said she would fix it in the next edition. She corrected Section 8.3, The Physics
of Recombination, but did not follow through consistently.
8.286 PROBLEM SET 9, FALL 2020 p. 7

Using these definitions, write the equation that equates the ratio nD /(np nn ) to
a function of temperature, using Eq. (6.1) for each of the number densities. The
deuteron is spin-1, with g = 3, and the proton and neutron are each spin- 21 , with
g = 2. Except in the exponential factor, you may approximate mn = mp = mD /2.
Manipulate this formula so that it has the form

F (η, f, x) = G(T ) , (6.9)

where F and G are functions that you must determine. You will need the binding
energy of deuterium,

B = (mp + mn − mD )c2 ≈ 2.22 MeV. (6.10)

Eq. (6.9) determines x as a function of T , or vice versa, but we will not try to write
x(T ) or T (x) explicitly.

(c) (5 points) Using your result in part (b), and taking f = 0.14 from Weinberg’s book,
find the value of x, the fraction of neutrons that have been bound in deuterium, at
the time of the Fifth Frame, when T = 109 K. You will probably want to solve the
equation numerically. Two significant figures will be sufficient.

(d) (5 points) Again using your result from part (b), and assuming that f = 0.14 is still
accurate, find the temperature at which x = 21 , i.e., the temperature for which half of
the neutrons have become combined into deuterium. Again you will presumably find
the answer numerically, and 2 significant figures will be sufficient. What is the value
of kT at this temperature? Qualitatively, what feature of the calculation causes this
number to be small compared to B?

PROBLEM 7: A ZERO MASS DENSITY UNIVERSE— GENERAL REL-

ATIVITY DESCRIPTION

(This problem is not required, but can be done for 20 points extra credit.)
In this problem and the next we will explore the connections between special rela-
tivity and the standard cosmological model which we have been discussing. Although we
have not studied general relativity in detail, the description of the cosmological model
that we have been using is precisely that of general relativity. In the limit of zero mass
density the effects of gravity will become negligible, and the formulas must then be com-
patible with the special relativity which we discussed at the beginning of the course. The
goal of these two problems is to see exactly how this happens.
These two problems will emphasize the notion that a coordinate system is nothing
more than an arbitrary system of designating points in spacetime. A physical object
might therefore look very different in two different coordinate systems, but the answer to
8.286 PROBLEM SET 9, FALL 2020 p. 8

any well-defined physical question must turn out the same regardless of which coordinate
system is used in the calculation.
From the general relativity point of view, the model universe is described by the
Robertson-Walker spacetime metric:

dr2
 
2 2 2 2
+ r2 dθ2 + sin2 θdφ2


ds = −c dt + a (t) 2
. (7.1)
1 − kr

This formula describes the analogue of the “invariant interval” of special relativity, mea-
sured between the spacetime points (t, r, θ, φ) and (t + dt, r + dr, θ + dθ, φ + dφ).
The evolution of the model universe is governed by the general relation
 2
ȧ 8π kc2
= Gρ − 2 , (7.2)
a 3 a

except in this case the mass density term is to be set equal to zero.
(a) (5 points) Since the mass density is zero, it is certainly less than the critical mass
density, so the universe is open. We can then choose k = −1. Derive an explicit
expression for the scale factor a(t).
(b) (5 points) Suppose that a light pulse is emitted by a comoving source at time te , and
is received by a comoving observer at time to . Find the Doppler shift ratio z.
(c) (5 points) Consider a light pulse that leaves the origin at time te . In an infinitesimal
time interval dt the pulse will travel a physical distance ds = cdt. Since the pulse is
traveling in the radial direction (i.e., with dθ = dφ = 0), one has

dr
cdt = a(t) √ . (7.3)
1 − kr2

Note that this is a slight generalization of Eq. (2.9), which applies for the case of
a Euclidean geometry (k = 0). Derive a formula for the trajectory r(t) of the light
pulse. You may find the following integral useful:
Z
dr
√ = sinh−1 r . (7.4)
1+r 2

(d) (5 points) Use these results to express the redshift z in terms of the coordinate r of
the observer. If you have done it right, your answer will be independent of te . (In
the special relativity description that will follow, it will be obvious why the redshift
must be independent of te . Can you see the reason now?)
8.286 PROBLEM SET 9, FALL 2020 p. 9

PROBLEM 8: A ZERO MASS DENSITY UNIVERSE— SPECIAL RELA-

TIVITY DESCRIPTION

(This problem is also not required, but can be done for 20 points extra credit.)
In this problem we will describe the same model universe as in the previous problem,
but we will use the standard formulation of special relativity. We will therefore use an
inertial coordinate system, rather than the comoving system of the previous problem.
Please note, however, that in the usual case in which gravity is significant, there is no
inertial coordinate system. Only when gravity is absent does such a coordinate system
exist.
To distinguish the two systems, we will use primes to denote the inertial coordinates:
(t , x0 , y 0 , z 0 ). Since the problem is spherically symmetric, we will also introduce “polar
0

inertial coordinates” (r0 , θ0 , φ0 ) which are related to the Cartesian inertial coordinates by
the usual relations:
x0 = r0 sin θ0 cos φ0
y 0 = r0 sin θ0 sin φ0 (8.1)
z 0 = r0 cos θ0 .
In terms of these polar inertial coordinates, the invariant spacetime interval of special
relativity can be written as

ds2 = −c2 dt02 + dr02 + r02 dθ02 + sin2 θ0 dφ02 .

(8.2)

For purposes of discussion we will introduce a set of comoving observers which travel
along with the matter in the universe, following the Hubble expansion pattern. (Although
the matter has a negligible mass density, I will assume that enough of it exists to define
a velocity at any point in space.) These trajectories must all meet at some spacetime
point corresponding to the instant of the big bang, and we will take that spacetime point
to be the origin of the coordinate system. Since there are no forces acting in this model
universe, the comoving observers travel on lines of constant velocity (all emanating from
the origin). The model universe is then confined to the future light-cone of the origin.
(a) (5 points) The cosmic time variable t used in the previous problem can be defined as
the time measured on the clocks of the comoving observers, starting at the instant
of the big bang. Using this definition and your knowledge of special relativity, find
the value of the cosmic time t for given values of the inertial coordinates— i.e., find
t(t0 , r0 ). [Hint: first find the velocity of a comoving observer who starts at the origin
and reaches the spacetime point (t0 , r0 , θ0 , φ0 ). Note that the rotational symmetry
makes θ0 and φ0 irrelevant, so one can examine motion along a single axis.]
(b) (5 points) Let us assume that angular coordinates have the same meaning in the two
coordinate systems, so that θ = θ0 and φ = φ0 . We will verify in part (d) below that
this assumption is correct. Using this assumption, find the value of the comoving
8.286 PROBLEM SET 9, FALL 2020 p. 10

radial coordinate r in terms of the inertial coordinates— i.e., find r(t0 , r0 ). [Hint:
consider an infinitesimal line segment which extends in the θ-direction, with constant
values of t, r, and φ. Use the fact that this line segment must have the same physical
length, regardless of which coordinate system is used to describe it.] Draw a graph
of the t0 -r0 plane, and sketch in lines of constant t and lines of constant r.
(c) (5 points) Show that the radial coordinate r of the comoving system is related to the
magnitude of the velocity in the inertial system by

v/c
r= p . (8.3)
1 − v 2 /c2

Suppose that a light pulse is emitted at the spatial origin (r0 = 0, t0 = anything)
and is received by another comoving observer who is traveling at speed v. With
what redshift z is the pulse received? Express z as a function of r, and compare your
answer to part (d) of the previous problem.
(d) (5 points) In this part we will show that the metric of the comoving coordinate
system can be derived from the metric of special relativity, a fact which completely
establishes the consistency of the two descriptions. To do this, first write out the
equations of transformation in the form:

t0 =?
r0 =?
(8.4)
θ0 =?
φ0 =? ,

where the question marks denote expressions in t, r, θ, and φ. Now consider an

infinitesimal spacetime line segment described in the comoving system by its two
endpoints: (t, r, θ, φ) and (t + dt, r + dr, θ + dθ, φ + dφ). Calculating to first order in
the infinitesimal quantities, find the separation between the coordinates of the two
endpoints in the inertial coordinate system— i.e., find dt0 , dr0 , dθ0 , and dφ0 . Now
insert these expressions into the special relativity expression for the invariant interval
ds2 , and if you have made no mistakes you will recover the Robertson-Walker metric
used in the previous problem.

DISCUSSION OF THE ZERO MASS DENSITY UNIVERSE:

The two problems above demonstrate how the general relativistic description of
cosmology can reduce to special relativity when gravity is unimportant, but it provides
a misleading picture of the big-bang singularity which I would like to clear up.
First, let me point out that the mass density of the universe increases as one looks
backward in time. So, if we imagine a model universe with Ω = 0.01 at a given time, it
8.286 PROBLEM SET 9, FALL 2020 p. 11

could be well-approximated by the zero mass density universe at this time. However, no
matter how small Ω is at a given time, the mass density will increase as one follows the
model to earlier times, and the behavior of the model near t = 0 will be very different
from the zero mass density model.
In the zero mass density model, the big-bang “singularity” is a single spacetime
point which is in fact not singular at all. In the comoving description the scale factor a(t)
equals zero at this time, but in the inertial system one sees that the spacetime metric
is really just the usual smooth metric of special relativity, expressed in a peculiar set
of coordinates. In this model it is unnatural to think of t = 0 as really defining the
beginning of anything, since the the future light-cone of the origin connects smoothly to
the rest of the spacetime.
In the standard model of the universe with a nonzero mass density, the behavior of
the singularity is very different. First of all, it really is singular— one can mathematically
prove that there is no coordinate system in which the singularity disappears. Thus, the
spacetime cannot be joined smoothly onto anything that may have happened earlier.
The differences between the singularities in the two models can also be seen by
looking at the horizon distance. We learned in Lecture Notes 4 that light can travel only
a finite distance from the time of the big bang to some arbitrary time t, and that this
“horizon distance” is given by
Z t
c
`p (t) = a(t) 0)
dt0 . (8.5)
0 a(t
For the scale factor of the zero mass density universe as found in the problem, one can
see that this distance is infinite for any t— for the zero mass density model there is no
horizon. For a radiation-dominated model, however, there is a finite horizon distance
given by 2ct.
Finally, in the zero mass density model the big bang occurs at a single point in
spacetime, but for a nonzero mass density model it seems better to think of the big
bang as occurring everywhere at once. In terms of the Robertson-Walker coordinates,
the singularity occurs at t = 0, for all values of r, θ, and φ. There is a subtle issue,
however, because with a(t = 0) = 0, all of these points have zero distance from each
other. Mathematically the locus t = 0 in a nonzero mass density model is too singular
to even be considered part of the space, which consists of all values of t > 0. Thus, the
question of whether the singularity is a single point is not well defined. For any t > 0 the
issue is of course clear— the space is homogeneous and infinite (for the case of the open
universe). If one wishes to ignore the mathematical subtleties and call the singularity
at t = 0 a single point, then one certainly must remember that the singularity makes
it a very unusual point. Objects emanating from this “point” can achieve an infinite
separation in an arbitrarily short length of time.

Total points for Problem Set 9: 130, plus an optional 40 points of extra credit.