2021-USAPhO Plus Solutions v2
2021-USAPhO Plus Solutions v2
2021-USAPhO Plus Solutions v2
• The exam is hosted by AAPT on the platform provided by Art of Problem Solving. It will
be proctored by the US Physics Team coaches via Zoom.
• Before you start the exam, make sure you are provided with blank paper, both for your
answers and scratch work, writing utensils, graph paper and a ruler, a hand-held scien-
tific calculator with memory and programs erased, and a computer for you to log into the
USAPhO+ testing page.
• At the end of the exam, you have 20 minutes to upload solutions to all of the problems for
that part. For each problem, scan or photograph each page of your solution, combine them
into a single PDF file, and upload them on the testing platform.
• USAPhO+ graders are not responsible for missing pages or illegible handwriting. No late
submissions will be accepted.
Congratulations again on your qualification for the USAPhO+. We wish you the best of luck on
the challenging problems to follow.
We acknowledge the following people for their contributions to this year’s exam (in alphabetical order):
JiaJia Dong, Mark Eichenlaub, Abijith Krishnan, Kye W. Shi, Brian Skinner, Mike Winer, and Kevin Zhou.
Question 1
The Jet Stream
The jet stream is an eastward wind current that moves over the continental United States at an
altitude of 23, 000 to 35, 000 feet (the range of typical cruising altitudes of commercial airlines).
This strong current affects flight times significantly: flights traveling eastward fly significantly faster
than flights traveling westward.
This problem consists of two independent parts. In the first part, you will consider a simple
model for airplane flight. In the second part, you will determine the jet stream speed on a fictitious
planet called Orb.
1. The power that a plane expends is used both to combat drag and to generate lift. Throughout
this part of the problem, you may assume that the plane travels with horizontal velocity vrel
relative to the air, the density of air is ρair , the mass of the plane is m, and the cross-sectional
area of the plane is Acs .
Solution
We have that P = F · v, so
1 3
Pdrag = cd ρair Acs vrel .
2
Solution
The mass flux is given by ρvrel , so the rate is
Solution
Suppose the deflected air has velocity u downward. Then the lift force is given by
mg = ρvrel Au.
This is the power that goes into accelerating the air downward, so by energy conser-
vation it must have come from the plane’s engine.
(c) Estimate the speed at which an airplane flies relative to the air by minimizing the power
expended by the plane. To get a numeric answer, you may use the following parameters:
mplane ∼ 8 × 104 kg, ρair ∼ 1 kg/m3 , cd ∼ 10−2 , Acs ∼ 100 m2 .
Solution
The total power is given by
m2 g 2 1 3
P ∼ + cd ρAcs vrel .
ρair vrel Acs 2
The minimum power occurs when the derivative vanishes, which roughly gives
!1/4
m2 g 2
vrel ∼ ∼ 300 m/s.
cd ρ2 A2
Next, we estimate the jet stream speed using flight times. Because the jet stream speed on
Earth varies greatly with location, time of year, and climate effects (such as El Niño and La Niña),
you will instead consider the fictitious planet Orb, where the jet stream is eastward and uniform
in the region of interest. At the end of the problem is a map of the region, whose area is much
smaller than the surface area of Orb (i.e., you can neglect the curvature of Orb).
2. At cruising altitude, we asume all airplanes travel at a fixed speed vrel relative to the air. (This is
not necessarily the same as your answer to 1(c), which was just a rough estimate.) Additionally,
we assume that flights occur in three stages – (1) taxi and takeoff, (2) flight at cruising altitude,
(3) landing and taxi – and that stages (1) and (3) take a fixed total time t0 for every flight.
(a) Suppose a plane, at cruising altitude, is traveling at an angle θ away from due east relative
to the ground. What is the speed of the plane relative to the ground? Give your answer in
terms of vrel , θ, and vw , the speed of the jet stream relative to the Earth’s surface.
Solution
The velocity of the plane relative to the ground, v, the velocity of the plane relative to
the air, vrel , and the jet stream velocity, vw , all form a triangle under tip-tail addition.
Then, from law of cosines,
2
vrel = v 2 + vw
2
− 2vvw cos θ.
Solving gives us q
v = vw cos θ + 2 − v 2 sin2 θ.
vrel w
Of course, we assumed vw < vrel , as otherwise flying westward would not even be possible.
(b) If the plane travels a distance D, what is the total travel time t, including taxi, takeoff, and
landing?
Solution
The answer is given by
D
t= + t0 ,
v
where v is our answer to the previous part.
(c) Below, we present some data on airplane flights on Planet Orb. Each of the flight times shown
below has an independent uncertainty of ∆t = 5 min. From the data and the map, determine
vw and vrel , giving your answers in km per hour with uncertainties. Indicate clearly what
two quantities you are plotting against each other on each graph that you plot.
Solution
Note that the pairs of cities given in the second table are directly north/south of each
other. For such cities, we expect the times in either direction to be equal, up to the
timing uncertainty, and the expression for the time simplifies to
D
tN S = q + t0 .
2 − v2
vrel w
We can read tN S off the second table, and calculate D from the figure. The unknown
quantity t0 goes into the intercept, while the slope is
1
q = 0.0768 ± 0.0048 min/km.
2 − v2
vrel w
Next, consider the pairs of cities given in the second table. We can cancel out the effect
of t0 by considering the difference in flight times in the two directions,
D D
t12 − t21 = − q + q .
2 − v 2 sin2 θ
vw cos θ + vrel −v cos θ + v 2 − v 2 sin2 θ
w w rel w
Therefore, plotting the time difference versus D cos θ, where θ can also be measured from
the graph, gives
2vw
2 2
= 0.0722 ± 0.0003 min/km.
vrel − vw
Combining these results gives the answers,
Uncertainty on the times should be propagated through the calculation using the usual
rules, while the uncertainties on the slopes are found by drawing the steepest and shal-
lowest fit lines.
Rubinstead Boltzmannburg
W E
Newton City
S
Einsteinopolis Maxwellbury
Question 2
The Dark Forest
Dark matter could be made of hypothetical, extremely light particles called axions. Because in-
dividual axions are so light, experiments do not search for individual axions, but rather for the
classical axion field formed by a large collection of axions, which oscillates as
a(t) = a0 sin(ωt).
This is analogous to how a large collection of photons can form a classical electromagnetic field. In
the presence of a magnetic field B and an axion field a, the axion field produces an effective current
J = g ȧB
where we define ȧ = da/dt. The effective current produces electromagnetic fields in exactly the same
way as ordinary current, though it does not come from the motion of actual charges. Experiments
can search for axion dark matter using systems which are resonantly driven by this current.
You may use fundamental constants in your answers, such as
You do not have to provide numeric answers unless asked. When asked to “estimate”, you may drop
constants of order one. The numeric values provided below are from standard references where ~,
c, µ0 , and 0 are set to one; to get correct numeric results, you must restore these factors yourself.
(a) Consider a single axion at rest, with mass m. Find its associated angular frequency ω. This
will be the angular frequency of the corresponding classical field, when there are many axions.
Solution
The axion is a quantum particle and satisfies the de Broglie relation E = ~ω, and its
energy comes from its rest mass, E = mc2 . Thus,
mc2
ω= .
~
(b) Suppose dark matter is distributed spherically symmetrically in the galaxy with uniform
density ρ. The solar system is a distance r from the center of the galaxy and orbits around
it with period T . Neglecting everything besides dark matter, find the dark matter density ρ.
Solution
Setting the centripetal acceleration equal to the gravitational acceleration,
GM 4
rω 2 = , M = πr3 ρ.
r2 3
(c) The energy density of the axion field is m2 a20 /(2~3 c). Find the axion field amplitude a0 .
Solution
The dark matter density comes from the mass-energy of the axion field, so
r
m 2 a2 2ρ~3 c3
ρc2 = 0
, a0 = .
2~3 c m2
(d) The radius and period of the Sun’s orbit, as well as a typical axion mass, are
r = 2.5 × 1020 m, T = 7.1 × 1015 s, m = 1.0 × 10−9 eV.
Numerically compute the axion field amplitude a0 .
Solution
Plugging in numbers and noting that the provided value of the axion mass is really its
energy (and hence dividing it by c2 ) gives the value
(e) In this problem, we treat the axion field as spatially uniform within a terrestrial laboratory.
To verify that this assumption is reasonable, numerically estimate the axion field’s wavelength
λ, assuming the axions have the same galactic speed as the Sun.
Solution
The axion speed is
2πr
v= = 2 × 105 m/s.
T
Using the de Broglie relation again,
h h
λ= = ∼ 106 m.
p mv
As expected, this is much larger than the apparatus considered below. (Note: a common
mistake was to write v = f λ, but this is only true for waves which travel at a constant
speed, such as light. To see the problem, note that if you apply the de Broglie relations,
this equation becomes v = E/p. That’s true for photons, where E = pc, but certainly not
true for massive particles. Some students also tried the Compton wavelength λ = h/mc,
but this is a different quantity, namely the wavelength of a photon if it had the same
energy as an axion at rest.)
(f) In part (a), you found ω by neglecting the axion’s speed. In reality, the axion’s finite speed
changes the frequency to ω + ∆ω, in a frame at rest with respect to the galactic center.
Numerically estimate ∆ω/ω to show that it is reasonable to neglect this effect.
Solution
Because E = ~ω, we need to see how the energy is changed. We have
∆E mv 2 /2 v2
≈ =
E mc2 2c2
where we used the fact that v, computed above, is nonrelativistic. We thus find
∆ω
∼ 3 × 10−7
ω
which is small as expected. (Note that the usual Doppler shift formula would not work
here, and trying to use it would get the wrong answer. The Doppler shift formula applies
to particles with constant speed, which satisfy ω = vk. The axion is massive; when at
rest, it has ω 6= 0 and k = 0. However, you could also get the answer by Lorentz
transforming the four-vector (ω, k).)
The ABRACADABRA1 experiment, currently taking data at MIT, is a toroidal solenoid with
inner and outer radius Rin and Rout and height h. You may assume h Rout for simplicity. A
superconducting wire carrying current I wraps N times around the toroid, where N is high enough
to neglect the discreteness of the wires. A circular pickup loop with radius slightly less than Rin is
placed at the center of the toroid.
2. Now, we will find the axion signal generated in the ABRACADABRA apparatus.
(a) Find the magnetic field B(r) inside the toroid due to the superconducting current.
1
aka, A Broadband/Resonant Approach to Cosmic Axion Detection with an Amplifying B-field Ring Apparatus.
Solution
Using Ampere’s law,
µ0 N I
B= θ̂.
2πr
(b) The superconducting wires lose their superconductivity when exposed to a magnetic field
greater than Bmax . Find the maximum possible current Imax that can be used, and assume
this current is used in later parts.
Solution
The maximum field is at r = Rin , so
µ0 N Imax 2πRin Bmax
Bmax = , Imax = .
2πRin µ0 N
(c) Assuming that ω is small, find the magnetic flux ΦB (t) through the pickup loop due to the
axion field in terms of a0 , g, ω, Bmax , and the dimensions of the apparatus. (You may ignore
any currents induced on the surfaces of the superconducting wires. Accounting for them
makes the problem much harder, but does not substantially affect the final result.)
Solution
The effective current is
Rin
J = g ȧB = ga0 Bmax ω cos(ωt)θ̂.
r
The resulting field is like that of an array of concentric solenoids, where we may neglect
fringe fields because h Rout . Therefore, the axion-produced field Bax inside the pickup
loop is uniform. For simplicity, we define
J0 = ga0 Bmax ω.
(d) If ω is too large, the result above breaks down due to radiation effects. Estimate the frequency
ωc where this happens.
Solution
In the calculations above, we have assumed that the fields are quasistatic, neglecting the
Solution
Plugging numbers in, we find
Then we have
ω/ωc ∼ 0.0025
which is small as expected. Any answer within an order of magnitude is acceptable.
(f) Let Φ0 be the amplitude of the time-varying axion flux. Using the typical values
Solution
First, we have to restore the dimensions of g. Notice that since current g ȧB has the
same units as displacement current 0 Ė, the quantities ga and (E/B)0 must have the
same dimensions, and furthermore E/B has dimensions of speed.
Now, in 1(c) we found that a has dimensions of energy times velocity, while the value
of g given here has dimensions of inverse energy. Thus, the combination ga only has
dimensions of speed, which means a factor of 0 was left out. In other words, if we
denote the value given above as g 0 , the true value of g is 0 g 0 .
To avoid mistakes, it’s best to group terms in the long expression for Φ0 , so that each
piece has simple units and a reasonable magnitude. We have
Rout
2
Φ0 = (πRin B0 ) log Rin µ0 0 g 0 a0 ω
Rin
ωRin g 0 a0
2 Rout
Φ0 = (πRin B0 ) log
Rin c c
= (3.9 Wb)(0.693)(0.0025)(4.91 × 10−19 )
= 3.3 × 10−21 Wb.
The pickup loop has inductance L and is attached to a capacitor, forming a circuit with resonant
frequency equal to the axion frequency ω. The circuit also has a small internal resistance R in series,
and is at temperature T . The axion signal can be detected by monitoring the current in the circuit.
The main source of noise is thermal noise, which causes fluctuations in the current.
3. We will now estimate the sensitivity of ABRACADABRA to axions.
(a) The axion produces a current which oscillates sinusoidally. Find the signal current amplitude
Is in terms of ω, Φ0 , and the circuit parameters.
Solution
Since the axion is at the resonant frequency, the impedance of the circuit is approximately
R. Therefore the current is
E 1 dΦB
I= =− .
R R dt
The amplitude is
ωΦ0
Is = .
R
(b) Find the average value of the current squared hI 2 i in the circuit due to thermal noise.
Solution
The energy stored in the inductor is LI 2 /2. By the equipartition theorem, the average
energy stored in any such quadratic degree of freedom is kB T /2, so
kB T
hI 2 i = .
L
This effect is also called Johnson noise.
(c) At anypmoment in time, the noise current is oscillating sinusoidally with typical amplitude
In = hI 2 i, which is much larger than Is . However, the phase of the noise current also
fluctuates randomly, so that after a typical time tc , its phase will be roughly independent of
the phase it had before. Find an estimate for tc in terms of ω and the circuit parameters.
(Hint: at any given moment, the thermal noise current is simultaneously being produced by
the random motion of electrons in the circuit, and damped by the resistor.)
Solution
The noise current that exists in the circuit at any given moment, by definition, came
from the thermal motion of electrons in the circuit up to a time tc ago. Therefore, tc is
roughly the characteristic time for decay of current in the circuit. There are many ways
(d) Suppose the experiment runs for a total time te tc . Roughly estimate the average amplitude
of the noise current over this period of time.
Solution
√
Uncertainty goes down by n when averaging n independent trials. Here we have
n = te /tc , so r
tc
I¯n = In .
te
(e) The axion is detectable if the signal current amplitude is larger than the averaged noise
current amplitude, and the circuit parameters are
Roughly numerically estimate the time needed to potentially detect the axion. (Hint: if your
answer seems strange, note that in reality, the axion’s phase also fluctuates over time, because
of the effect of part 1(f). In addition, we don’t know ω ahead of time, so the experiment
needs to be run many times. We ignored these effects here to keep things simple.)
Solution
By setting Is = I¯n and solving for te , we find
kB T R
te ∼ = 550 s.
ω 2 Φ20
In reality, you would need to repeat this experiment millions of times to scan a sufficient
range of ω, and the time in each step would be longer, because after some point the
√
average axion signal also starts going down as 1/ n.
Question 3
Great Hall
The classical Hall effect was first measured by Edwin Hall in 1879, shortly after the publication of
Maxwell’s equations. In all parts of this problem, materials contain nV electrons per unit volume,
and each electron has charge qe < 0 and mass me . You may use these quantities in all of your
answers. We will begin by investigating the implications of the classical Hall effect.
1. An infinite plate in the xy plane, with thickness d in the z direction, is placed in a uniform
magnetic field B = B ẑ as shown. An electric field E = E x̂ is applied in the plane of the plate
and the system is allowed to reach a steady state.
x
B
E
(a) If the electrons have velocity v at steady state, what is the current density J? Recall that J
is defined as the total flow of charge through a unit cross-section area per unit time.
Solution
The current density is given by nV qe v.
(b) In the Drude model, electrons are subject to both the Lorentz force and a damping force
−γv, where γ is a constant that depends on the material. In the above system, what is the
current density in the steady state? Give both the magnitude and direction of J, e.g. in polar
coordinates.
Solution
The equation of motion for an electron is
γqe E −qe2 EB
vx = , vy = .
γ + qe2 B 2
2 γ 2 + qe2 B 2
Then,
γnV qe2 E −nV qe3 EB
Jx = , Jy = .
γ 2 + qe2 B 2 γ 2 + qe2 B 2
qe2 nV E
Bqe
|J| = p , θ = − arctan .
γ 2 + B 2 qe2 γ
Solution
The two resistivities are
γ B
ρ0 = , ρH = .
nV qe2 n V qe
(d) A Hall effect sensor detects the strength of magnetic fields. Consider the following circuit
consisting of a square plate of side length L and thickness d in a perpendicular uniform
magnetic field B.
B
− − − − − − − −
VH I
V A
+ + + + + + + +
d L
L
+ −
E
A longitudinal emf E is applied to the plate. At steady state, a Hall voltage VH is measured
across the plate due to the buildup of charge on either side of the plate. If the electrical
resistivity of the plate at zero magnetic field is ρ0 , what is the Hall voltage VH and the
current I through the plate? Express your answer in terms of ρ0 , E, B, and the dimensions
of the plate.
Solution
The Hall voltage is such that the electric field from the external EMF and from the
charge buildup causes the current to flow in the x direction. Our equation of motion is
thus
qe E/Lx̂ + qe VH /Lŷ − qe vx Bŷ − γvx x̂ = 0.
Experiments in the 20th century revealed that in many materials, the Hall resistivity could only
take certain discrete values. We will now show how this follows from Bohr quantization. (These
next parts are independent of the first part of the problem.)
(a) If the electrons all have the same speed, what is the angular momentum of each electron?
Solution
The current is I = qe nV aAw , where v is the speed of each electron. Then the angular
momentum is
Ime R
me vRẑ = ẑ.
qe nV Aw
Note that the angular momentum points into the page, because the electrons have neg-
ative charge.
(b) If we allow the flux in the solenoid to change, the usual, “mechanical” angular momentum L
of each electron is not conserved. Instead, a quantity called the canonical angular momentum,
Lcan = L + Cqe Φ, for some constant C, is conserved. Find C.
Solution
We compute the change in angular momentum associated with a change in flux ∆Φ. By
Faraday’s Law, there is an induced electric field resulting from the change in magnetic
flux.
dΦ
2πRE = − .
dt
The electric field results in a force qe E on each electron, and thus we find that
me dv dΦ
2πR =− .
qe dt dt
Then, the rate of change in angular momentum is
dv qe dΦ
Rme = .
dt 2π dt
Therefore,
qe
∆L = − ∆Φ.
2π
Then, for L + Cqe Φ to be conserved,
1
C= .
2π
(c) The Bohr quantization condition says that for a closed circular orbit, an integer number of
de Broglie wavelengths must fit in its circumference. The de Broglie wavelength is
h
λ= ,
pcan
where h is Planck’s constant, and pcan = Lcan /R is the canonical momentum. For a given
solenoid flux Φ, what is the set of allowed mechanical angular momenta L?
Solution
The circumference of such an orbit is 2πR. A de Broglie wavelength is given by
hR
λ= .
Lcan
Then,
2πLcan Lcan
n= = ,
h ~
or Lcan = n~ where ~ = h/2π. Then the allowed mechanical momenta satisfy
qe Φ
L = n~ − ,
2π
where n is an integer.
(d) What is the minimum possible change in the magnetic flux for which the same set of me-
chanical angular momenta is allowed? This is known as the flux quantum.
Solution
|qe |∆Φ
We require = ~ such that we get the same set of allowed mechanical angular
2π
momenta. Then,
h
∆Φ = .
|qe |
Note that for a superconducting ring, the flux quantum is h/2|qe | because the electrons
come in Cooper pairs, but this wasn’t part of this problem.
3. Now, consider an annulus held perpendicular to a fixed, uniform external magnetic field B, and
suppose an additional, tunable magnetic flux Φ threads the center of the annulus, with both
pointing out of the page. The annulus has a transverse Hall resistance RH (i.e., an EMF of E
around the annulus generates a perpendicular current E/RH via the Hall effect) and you may
neglect its self-inductance.
Φ
B
(a) Suppose Φ begins to increase slowly and steadily in time. After a short time, the electrons
will begin flowing steadily from one side of the annulus to the other. Do the electrons move
inward or outward? Justify your answer.
Solution
The increasing flux induces a clockwise EMF. The Lorentz force on an electron is
qe (E + v × B).
Then, v × B must oppose E. Therefore, v must point radially inward by the right hand
rule. Therefore, electrons must be flowing radially inward.
(b) If the threaded flux increases by ∆Φ, how many electrons pass from one edge of the annulus
to the other? You may use RH , among other variables, in your answer.
Solution
We have that
1 dΦ
=⇒ ∆Q = 1 ∆Φ.
|I| =
RH dt RH
Therefore, the number of electrons that pass from one edge to another is
∆Φ
.
RH |qe |
(c) As we showed in 2(d), if the magnetic flux changes by the flux quantum Φq , the allowed orbits
from Bohr quantization are unchanged. Quantum mechanics thus tells us that in conventional
materials, if the magnetic flux changes by Φq , an integer number k of electrons must pass
from one edge to another. What constraint does this place on the Hall resistance?
Solution
We set
h h
|qe | ν = =⇒ RH = 2 .
RH |qe | kqe
This is known as the integer quantum Hall effect, and it applies to any material shape.