Physics Unlimited Premier Competition

2022 Examination

November 6, 2022

Fill in these two lines only if competing in-person:

Competitor ID (Exam Code):
Initials:

This exam contains 7 pages (including this cover page) and 4 questions worth 105 points total.
Only start working on the exam when you are told to do so.
You are not expected to complete all of the exam. So you are encouraged to read through the exam
first and start with the problems you find easiest. The problems and their subparts are NOT
ordered according to their difficulty, and it is possible that you can work out some
later parts of a question when you are stuck on a former part. Don’t spend too long
on any one problem. Partial credit will be awarded.

Please prepare at least 10 sheets of blank paper (for online competitors) or an empty exam booklet
(provided to in-person competitors), a ruler for drawing graphs (in one of the problems), and a
simple numeric calculator. Please be sure to not have anything else on your desk.

Note: if you are competing virtually, all work to be graded must be on blank sheets
of paper that you will take photos of and submit as instructed immediately after the
test. If you are competing in-person, all work should be in the exam workbook given
to you. Box all answers, and try to work as clearly and neatly as possible.
Distribution of Marks
Question Points Score
1 23
2 32
3 28
4 22
Total: 105

1
 November 6, 2022

1. Computing electric fields (23 points)

Electrostatics relies on multiple methods for computing electric fields and potentials. In this
problem, we will explore two of them, Gauss’s Law and Legendre polynomials.

Uniform charge distributions

Let us consider a hollow conducting sphere of radius R charged with the electric charge Q,
uniformly distributed on its surface. In order to calculate its potential, we can use Gauss’s
Law, which states that the flux of the electric field dF = E · dA across a closed surface is
proportional to the charge enclosed by that surface: F = Q/ε0 . We have denoted dA = dAn
the elementary oriented (towards the exterior) surface element.

(a) (7 points) Compute the electric potential inside and outside the sphere. Which equivalent
configuration in terms of point charges provides us with the same potential at distances
greater than R?
Legendre polynomials and non-uniform charge distributions

Legendre polynomials are a type of orthogonal polynomials essential in mathematical physics.

One of their applications is in computing electric potentials for more complicated charge con-
figurations. We will denote the n-th Legendre polynomial (having degree n) as Pn . Legendre
polynomials are defined on [−1, 1] and we can express their scalar product as
Z1
⟨Pm (x), Pn (x)⟩ = Pm (x)Pn (x)dx. (1)
−1

The first two Legendre polynomials are P0 (x) = 1 and P1 (x) = x.

(a) (4 points) Knowing that Legendre polynomials are orthogonal (⟨Pm (x), Pn (x)⟩ = 0 if
m ̸= n) and deg Pn (x) = n, obtain P2 (x) and P3 (x). For reaching the usual and most
2
convenient form of these polynomials, divide your results by the norm: ||Pn (x)|| = 2n+1 .

Let us now consider a sphere of radius R centered at the origin. Suppose a point charge q is
put at the origin and that this is the only charge inside or outside the sphere. Furthermore,
the potential is Φ = V0 cos θ on the surface of the sphere.
We know that we can write the potential induced by the charge on the sphere (without
taking into account q) in the following way:
X∞
Φ− = An rn Pn (cos θ), r < R
n=0
∞
X Bn
Φ+ = Pn (cos θ), r>R
n=0
rn+1
(b) (12 points) Compute the electric potential both inside and outside the sphere.

Page 2 of 7
 November 6, 2022

2. Johnson-Nyquist noise (32 points)

In this problem we study thermal noise in electrical circuits. The goal is to derive the Johnson-
Nyquist spectral (per-frequency, f ) density of noise produced by a resistor, R:
d⟨V 2 ⟩
= 4kT R. (2)
df
Here, ⟨⟩ denotes an average over time, so ⟨V 2 ⟩ is the mean-square value of the voltage fluctu-
ations due to thermal noise. f is the angular frequency, k is Boltzmann’s constant, and T is
temperature. It says that every frequency range [f, f + df ] contributes a roughly equal amount
of noise to the total noise in the resistor; this is called white noise.

Electromagnetic modes in a resistor

We first establish the properties of thermally excited electromagnetic modes

Vn (x) = V0 cos(kn x − ωn t) (3)
in a resistor of length L. The speed of light c′ = ωn /kn in the resistor is independent of n.
(a) (3 points) The electromagnetic modes travel through the ends, x = 0 and x = L, of the
resistor. Show that the wavevectors corresponding to periodic waves on the interval [0, L]
are kn = 2πn
L
.
dn L
Then, show that the number of states per angular frequency is dω n
= 2πc′.

(b) (2 points) Each mode n in the resistor can be thought of as a species of particle, called a
bosonic collective mode. This particle obeys Bose-Einstein statistics: the average number
of particles ⟨Nn ⟩ in the mode n is
1
⟨Nn ⟩ = . (4)
exp ℏωn
kT
−1
In the low-energy limit ℏωn ≪ kT , show that
kT
⟨Nn ⟩ ≈ . (5)
ℏωn
You can use the Taylor expansion ex ≈ 1 + x for small x.
(c) (3 points) By analogy to the photon, explain why the energy of each particle in the mode
n is ℏωn .
(d) (6 points) Using parts (a), (b), and (c), show that the average power delivered to the
resistor (or produced by the resistor) per frequency interval is
P [f, f + df ] ≈ kT df. (6)
Here, f = ω/2π is the frequency. P [f, f + df ] is known as the available noise power of the
resistor. (Hint: Power is delivered to the resistor when particles enter at x = 0, with speed
c′ , and produced by the resistor when they exit at x = L.)

Page 3 of 7
 November 6, 2022

Nyquist equivalent noisy voltage source

2
The formula d⟨Vdf ⟩ = 4kT R is the per-frequency, mean-squared value of an equivalent noisy
voltage source, V , which would dissipate the available noise power, dP
df
= kT , from the resistor
R into a second resistor r.
(a) (6 points) Assume that resistors R and r are in series with a voltage V . R and V are
fixed, but r can vary. Show the maximum power dissipation across r is

V2
Pmax = . (7)
4R
Give the optimal value of r in terms of R and V .
d⟨Pmax ⟩
(b) (3 points) If the average power per frequency interval delivered to the resistor r is df
=
d⟨V 2⟩
dE
df
= kT , show that df
= 4kT R.

Other circuit elements

d⟨V 2 ⟩
We derived the Johnson-Nyquist noise due to a resistor, R. It turns out the equation df
=
4kT R is not generalizable to inductors or capacitors.
(a) (4 points) Explain why no Johnson-Nyquist noise is produced by ideal inductors or ca-
pacitors. There are multiple explanations; any explanation will be accepted. (Hint: the
impedance of an ideal inductor or capacitor is purely imaginary.)
(b) (5 points) Any real inductor has undesired, or parasitic, resistance. We can model the real
inductor as an ideal inductor L in series with a parasitic resistance R.
d⟨V 2 ⟩
Due to the thermal noise df
= 4kT R of its parasitic resistance, this (real) inductor will
d⟨I 2 ⟩
support a nonzero per-frequency mean-squared current, df
, even when both sides of the
d⟨I 2 ⟩
inductor are grounded. Compute df
as a function of f, L, T and R.

Page 4 of 7
 November 6, 2022

3. The circular restricted three-body problem (28 points)

In general, there is no exact solution of the three-body problem, in which three masses move
under their mutual gravitational attraction. However, it is possible to make some progress by
adding some constraints to the motion.

Two-body problem

Let’s start with the motion of two masses, M1 and M2 . Assume both masses move in cir-
cular orbits about their center of mass. Consider the inertial frame whose origin coincides with
the center of mass of the system.

(a) (5 points) Express the equations of motion of M1 and M2 in terms of the gravitational
constant G and the position vectors r⃗1 and r⃗2 which point from the origin to M1 and M2 ,
respectively.

(b) (2 points) Find the period T and angular frequency ω of the orbital motion.

Circular restricted three-body problem

Let us transform to a non-inertial frame rotating with angular velocity ω ⃗ = (0, 0, ω) about
an axis normal to the orbital plane of masses M1 and M2 , with the origin at their center of
mass. In this frame, M1 and M2 are stationary at the Cartesian coordinates (−αR, 0, 0) and
((1−α)R, 0, 0) respectively. The third mass m is not stationary in this frame; in this non-inertial
frame its position is ⃗r(t) = (x(t), y(t), 0).
The masses satisfy M1 , M2 ≫ m. Consider m to be so small that it does not affect the motion
of M1 or M2 .
(a) (3 points) Express α in terms of M1 and M2 .

(b) (3 points) Let ρ1 (t) and ρ2 (t) be the distances from m to M1 and M2 respectively. Express
ρ1 (t) and ρ2 (t) in terms of the coordinates and constants given.

(c) (6 points) By considering the centrifugal acceleration ω 2⃗r and Coriolis acceleration −2ω ×
2
⃗v , find the acceleration dtd 2 ⃗r of the third mass in terms of the coordinates and constants
given, including ρ1 and ρ2 .

d2 x d2 y ω2
(d) (4 points) Express dt2
and dt2
in terms of U , where U = − GM
ρ1
1
− GM2
ρ2
− 2
(x2 + y 2 ).

(e) (5 points) Hence, write down an expression of the motion of m which is a constant.

Page 5 of 7
 November 6, 2022

4. The fundamental rocket equation (22 points)

In this problem, we will investigate the acceleration of rockets.

Rocket repulsion

In empty space, an accelerating rocket must “throw” something backward to gain speed from
repulsion. Assume there is zero gravity.
The rocket ejects fuel from its tail to propel itself forward. From the rocket’s frame of reference,
the fuel is ejected at a constant (relative) velocity u. The rate of fuel ejection is µ = dm/dt < 0,
and this rate is constant until the fuel runs out.

Figure 1: Rocket (and fuel inside) with mass m, ejecting fuel at rate µ = dm/dt with relative
velocity u.

(Note: the m in µ = dm/dt denotes the mass of rocket plus fuel, not the mass of an empty
rocket.)

(a) (3 points) Consider a small time interval ∆t in the rocket’s frame. ∆t is small enough that
the rocket’s frame can be considered an inertial frame (i.e., the frame has no acceleration).
The amount of fuel ejected in this time is ∆m = |µ|∆t. (See Figure 2.)

Figure 2: Rocket gaining speed ∆v during interval ∆t by ejecting mass ∆m with relative velocity
u.

Write down the momentum-conservation relation between u, ∆v, ∆m and m (the total
mass of the rocket and the fuel inside at the moment). What is the acceleration a = dv/dt
(independent of reference frame) in terms of u, µ and m?

(b) (3 points) Suppose that the rocket is stationary at time t = 0. The empty rocket has mass
m0 and the rocket full of fuel has mass 9m0 . The engine is turned on at time t = 0, and
fuel is ejected at the rate µ and relative velocity u described previously. What will be the
speed of the rocket when it runs out of fuel, in terms of u = |u|?

Page 6 of 7
 November 6, 2022
Rb1
(Hint: useful integral formula: a x
dx = ln( ab ), and ln(10) ≈ 2.302, ln(9) ≈ 2.197.)

(c) (2 points) Discuss the factors that limit the final speed of the rocket.
Chemical rockets

Nearly all rockets get energy from the chemical reaction of the fuel (burning with the oxi-
dant) they carry. This is called a combustion reaction. For example, the combustion reaction
for hydrogen is
2H2 + O2 −→ 2H2 O. (8)
(a) (3 points) Suppose that the product of combustion is just one species of molecule (H2 O,
for example) with mass mp and average kinetic energy E. What is the upper limit of
exhaust speed u when these molecules are ejected from the rocket?

(b) (4 points) To manufacture fast rockets, would you recommend using hydrogen as the fuel?
Assume that we only use oxygen as oxidant. If so, please give your reasoning. If not, what
fuel would you suggest and why? (There are no numerical values here, so please give the
best reasoning based on your own knowledge. Any reasonable answer is acceptable.)

(c) (3 points) The enthalpy of combustion of the reaction 2H2 + O2 −→ 2H2 O is approxi-
mately 15.76 × 106 J/kg. That is, for every kilogram of water (H2 O) produced, the energy
released by the reaction is 15.76 × 106 J. Assuming that all the energy released can be
converted
√ to the kinetic energy of the water molecules, what is their exhaust speed u?
(Hint: 15.76 × 106 ≈ 3970)

(d) (4 points) Now, consider the effects of gravity. Given a rocket of mass m0 and mass
ejection rate µ, what ejection speed u would be required to launch satellites in the Earth’s
gravitational field, of strength g? (An estimate, ignoring numerical factors, is acceptable.)
What about launching space probes to other planets or out of the solar system? Can you
guess why most rockets are multistage?

Page 7 of 7