QualifyingExamProblemPhysicsUnivOfWashington 1992 2000

Basi
Physi s
Fall 1998 Qualifying Exam
Solutions maintained by Graeme Lufkin
gwlu.washington.edu
1 Capa itors
(30 pts.) A parallel plate apa itor is harged from a battery. The spa e between the plates of the apa itor
is lled with air. After the harging is omplete, the swit h is opened.
Referen es: Capa itors are dis ussed in any basi physi s text, for example Halliday and Resni k. This
problem is similar to the examples in Introdu tion to Ele trodynami s, GriÆths, Se tion 4.4.
1.1
(10 pts.) A pie e of a diele tri (whi h ts exa tly in the spa e between the plates) is pla ed at the edge
of the apa itor. Explain whether the diele tri is repelled, attra ted, or left unae ted if released. Negle t
the ee t of air resistan e and of the gravitational eld.
Solution: You need to know the energy stored in a apa itor. The system wants to lose energy, so the
diele tri will move to lower the energy stored in the apa itor. The rst formula you may remember is
U = 21 CV 2 . However, we have just dis onne ted the battery, so the potential V will not remain onstant.
So it is more prudent to use U = Q2 =2C , be ause the harge on the apa itor will remain onstant. When
adding a diele tri , the apa itan e goes as C = 0 C0 , so it in reases in this ase. So the energy is lower
with the diele tri , and it will get su ked into the apa itor.
1.2
(10 pts.) Repla e in the previous question the diele tri with a metalli pie e of a smaller width than the
distan e between the plates. Explain and des ribe what happens in this ase.
Solution: You need to realize that the work metalli means a good ondu tor, and we will ee tively
have a new ir uit ontaining two apa itors in series. We know that C1 = C11 + C12 and that for a parallel
plate apa itor, C = 0 A=d. Sin e the pie e of metal has some nite thi kness, the total separation between
the plates of our new, ee tive apa itors is less than the original separation distan e. So we see that the
new apa itan e is greater than the original. This is the same situation as the previous question, so we give
the same answer: the metalli pie e is su ked in.
1.3
(10 pts.) Explain and des ribe what would happen in the rst question if the swit h remains losed, and
al ulate how mu h work the battery does during this part of the pro ess.
Solution: You need to now the energy stored in a battery, U = QV . Sin e the battery remains
onne ted, we an use equations that assume V is held xed. Also remember the basi equation for apa itors,
Q = V C . So for the system before we introdu e the diele tri we have
1 2 2 1 2 1 2
U = Q0 V +
C V = C0 V + C0 V = C V
2 0 2 2 0
When we introdu e the diele tri , the harge and apa itan e hanges, and we get
1 2 1 2
U
0
= 0
QV C V =
+ 0 0
C V
2 2
Sin e C 0
= 0 C0 > C0 , the energy is again lower with the diele tri , so it gets su ked in.
1
2 Mean Free Path
(10 pts.)
The mean free path is dened as the average distan e traveled by a mole ule between two onse utive
ollisions with other mole ules. Assume that the absolute temperature of air in reases by a fa tor of two,
while the air pressure remains the same. Does the mean free path hange? If yes, by how mu h? If not,
why?
Solution: You need to know an expression for the mean free path. This expression was taken from
Halliday and Resni k: = d12 where the radius of the parti les is d and the number density (N=V ) is .
The form is valid for non-spheri al parti les: just repla e d2 by something with the units of area. So we
assume air is approximately an ideal gas, and P V = N kB T holds. The temperature doubles, but pressure
stays onstant, so the number density is halved.
! 21 ) ! 2
3 Classi al Me hani s
(60 pts.) A material point moves without fri tion on the interior surfa e of a ir ular one. The symmetry
axis (^z) is exa tly verti al. The angle between the verti al and the one surfa e is . The initial velo ity is
tangent to the one surfa e and in the horizontal plane ~v = v0 x^. See the gure on the original question.
3.1
(5 pts.) Show all the for es a ting on the material point.

Solution: You need to know how do draw a free-body diagram. Gravity is a ting with ^
mg z on the
point mass. There is also a normal for e perpendi ular to the one surfa e.
3.2
(10 pts.) Show that the z - omponent of the angular momentum is onserved and al ulate it.
Solution: We are given the hint to work in ylindri al oordinates, where = z tan . So,
Lz = m2 _ = mz 2 tan2 _

Now remember that torque an be des ribed as the hange in angular momentum, = ddtL . We know that
= r F. Use the right-hand rule to nd the dire tion of this ve tor, and you will see that has no omponent
in the z dire tion.
dLz
z =0 =
dt
So the z - omponent of angular momentum is indeed onserved, and we an write that Lz = m0 v0 =
mz0 tan v0 .
3.3
(10 pts.) Des ribe in words the motion of the material point.
Solution: You need to be good at visualization. The point mass will move in a oni se tion. For most
initial velo ities, it will move in an ellipse, going up and down along a sli e through the one. If the velo ity
is small, the starting point will be the highest point on the traje tory, and for large velo ities, it will be the
lowest point. For one value of the initial velo ity, the point mass will move in a perfe t ir le.
2
3.4
(25 pts.) Determine the highest and lowest value of the oordinate z rea hed by the material point during
its motion.
Solution: You need to be able to solve the right equations. We know that the initial point will be either
the lowest or highest, depending on the initial velo ity, so it is the other riti al point we are looking for.
First, get an expression for the total energy, in ylindri al oordinates. Remember that there an be motion
radially, azimuthally, and along the z -axis.
1
m(_ + _ + z_ ) + mgz
E =T +V = 2 2 2 2
2
Now you need to remember the tri k of writing the energy in terms of the angular momentum, to get an
ee tive potential. This is a ommon tri k in systems with some sort of radial symmetry.
1 L2z 1
E = m(_ 2 + 2 2 + z_ 2 ) + mgz = m(_ 2 + z_ 2 ) + Ve ()
2 m 2
Where, using the fa t that we are onstrained to the one, z = = tan ,
L2z mg
Ve () = +
2m2 tan
The angular momentum bit is the familiar entrifugal barrier. At the riti al points we are looking for, there
is no radial or z -axis velo ity. So we an write, using the initial point and velo ity as one of the riti al
points
L2z L2
2 + mgz0 = z 2 + mgz
2m0 2m
where we want to solve for or z (they are equivalent be ause we are onstrained to the one). Use the
value for Lz we obtained previously, and we get
mv02 mv02 20
+ mgz0 = + mgz
2 2 2
Move it around, and nally swit h to the z oordinate (you an do it using either or z , but the questions
asks for the height, so we use z here).
mv02 z02 z0
(1 ) = mgz (1 )
2 z2 z
Now fa tor the quadrati term and an el to get

2 mv02
mgz (z + z0 ) = 0
2
This is a quadrati equation whi h we an solve, and nally we get
v02 v0
pv2 + 8gz
0 0
z =
4g
Note that the solution provided by the question's author has v02 in the radi al, while it should be v04 in his
expression.
3
Basi Physi s
gwlu.washington.edu
1 Pendulums
(50 pts.) You are given a thin rod of length L with uniform mass density and uniform ross se tion A L2 .
Let M denote the mass of the rod. This rod is equipped with two movable massless pivots so that it an be
used as a physi al pendulum with an adjustable period about either of two positions. Let h1 and h2 (both
taken as positive numbers) denote the displa ement of the pivots with respe t to the enter of mass of the
rod. Let g denote the gravitational a eleration due to the Earth.
1.1 Moment of Inertia
(10 pts.) Cal ulate the moment of inertia about either pivot point as a fun tion of hi .
Solution: You need to know the equation for the moment of inertia, and the parallel axis theorem. You
might just remember the moment of inertia for a rod about its enter of mass, but the question asks for a
al ulation, not a good memory. So we remember the integral to get the moment of inertia about an obje t's
enter of mass Z
I m = d3 r(r)r2
where the integral must be performed about a prin iple axis of the obje t. In this ase, we have
Z L=2
1
I m = A x2 dx = AL3 =12 = M L2
L=2 12
Now remember the parallel axis theorem
1
I = I m + M h2 = M ( L2 + h2 )
12
where in our ase, h an be either h1 or h2 .
1.2 Os illation Frequen y
(15 pts.) Cal ulate the frequen y ! of small os illations about a pivot point.
Solution: You need to be able to solve simple Lagrangian problems. In this ase, the kineti energy
is just T = 12 I _2 , where we use the I we al ulated in the previous question. The potential energy is
V = M gh os . We onstru t our Lagrangian and take derivatives to get the Euler-Lagrange equation.
The question asks for small os illations, so we use the small angle approximation and get
I + M gh =0
We have solved this equation many times, and we see that
M gh gh
!2 = = 1 2 2
I 12 L + h
1.3
(10 pts.) Your result to the previous question should be multi-valued, i.e. for some frequen ies ! there are
multiple values for h < L=2. Given a pair of distin t pivot positions h1 < h2 with the same ! , show that
g = (h1 + h2 )! 2 .
1
Solution: You have to gure out the question, and then do some algebra. Our equation from the previous
question has an h2 in it, so from psy hology of the question, we an assume that \multi-valued" means that
there will be a quadrati equation somewhere. Look at the equation for ! 2 we just obtained. We an invert
it and write it in powers of h. This gives us a quadrati equation in h, whi h we solve to get
p
g g 2 ! 4 L2 =3
h=
2! 2
So, for a given ! , there will be two distin t h's. The next part of the question lets us assume that we indeed
have h1 and h2 , the two solutions to our quadrati equation, for some ! . Then we add the two together,
and the term an els, giving h1 + h2 = g=! 2 , whi h is the equation we were looking for.
1.4 Experimental Issues
(15 pts.) Brie y, qualitatively, dis uss whether determination of g using the formula with two pivot points
is experimentally superior or inferior to the immediate use of the result from one pivot point. Hint: how
sensitive is ea h formula to irregularities of the rod.
2
Solution: We have two formulas for the gravitational a eleration, g = ! (h1 + h2 ) and g = I! =M h.
2
The question asks for experimental issues that might have an ee t on these equations. From the psy hology
of the question, we guess that the two-pivot equation is better, or we would not have been asked to derive it.
Presumably, we an weigh the obje t reliably, so the mass is not an issue, and ! 2 is present in both equations.
So the key dieren e is that the se ond formula requires that we know the moment of inertia ahead of time,
while the rst only requires that we nd two pivot points that give the same os illation frequen y. Taking
the hint, if the rod were not quite so thin, or not really of uniform density, our expression for the moment
of inertia would be wrong, while we ould still nd the two pivot points.
2 S attering
(25 pts.) The angular-dependen e of mono hromati opti al-wavelength light s attering an be used to
determine the size of small spheri al parti les suspended in a water solution. However, this te hnique be omes
unreliable if a signi ant fra tion of the dete ted s attered radiation has been multiply-s attered (rather than
just singly-s attered) by the small spheri al parti les in the sample. Brie y propose an experimental test to
quantify the ontribution to the s attered intensity from multiple s attering. Ignore any absorption in the
sample. Your answer should in lude a brief des ription and/or a sket h of your apparatus. Explain your
answer. Available equipment in lude (but is not limited to) : polarizers, dete tors, lenses, pinholes, mirrors,
glassware, more lasers, et .
Solution: There are several solutions that make varying degrees of sense. I think the most lear is
suggested by the question's original author, and that is given here, slightly reworded. Measure the s attered
light at large angles from the in ident dire tion. Multiple s attering will yield approximately isotropi
s attering, while single s attering will largely be in the forward dire tion. So a measure of the intensity at
large angles will give you a rough idea of whether multiple s attering is o urring.
3 Short Answer
(25 pts.)
3.1 Dust
(10 pts.) Major vol ani explosions an push a large amount of dust parti les high into the stratosphere.
After major eruptions, the moon (as observed from the Earth) often has a red tint. Explain this observation.
In parti ular, also explain what this observation tells you about the size distribution of the dust parti les.
Referen es: S attering o spheres and the theory of blue sky is dis ussed in Se tions 10.2 and 10.4 of
Classi al Ele trodynami s, Ja kson.
2
Solution: You need to reason this out any way you an. If the Moon is tinted red, that must mean that
the blue omponents of the visible light are being s attered by the dust more than the red omponents.
So the s attering ross se tion must be dependent on the wavelength of the light. Ja kson gives the ross
se tion for s attering o a sphere for the long wavelength limit, where dd
/ 4 . We don't know the short
wavelength limit, but we guess that it is less dependent on wavelength than the long wavelength limit. So we
say that the parti les must be mu h smaller than the wavelength of visible light, so that the approximation
will hold.
3.2 Mirages
(15 pts.) Two physi s graduate students are taking a summer va ation in the New Mexi o desert. During
a daytime hike on level ground, the students reliably observe that the desert oor appears to be re e ting
(giving the false appearan e of a rippling mirror or pond of water) at a distan e of a few hundred meters
from where they are standing. Give a qualitative explanation for their observations.
Solution: You need to know that a mirage is a re e tion of the blue sky. It o urs only when there is
a severe temperature dierential between two substan es. In this ase, the desert ground is mu h hotter
than the air above it. This dieren e in temperature auses the index of refra tion of the air to vary with
height above the ground. The index of refra tion is inversely proportional to the temperature. A (possibly
spurious) explanation for this is to assume that the air is an ideal gas. Be ause the atmosphere is so big, the
pressure will remain onstant. So as the temperature goes up, the density goes down. Light beams moving
from one medium to a less dense medium will be bent away from the verti al. This relationship for the index
of refra tion will ause a light ray to slowly bend away from the ground, and re e t towards the hikers, who
see the mirage.
3
Basi Physi s
gwlu.washington.edu
1 Falling Rope
(25 pts.) An innitely exible rope is suspended above a s ale, with the bottom end of the rope just tou hing
the s ale (gravity points down). The rope has lenth L and total mass M , distributed evenly along its length.
The rope is released. After a length x has fallen on the s ale, what does the s ale read?
Solution: The weight measured by the s ale will have two parts: one fromn the weight of the rope already
lying on the s ale, and a se ond from the impulse of the rope landing on the s ale. The amount of rope
already on the s ale will be x=L, and its weight will be M gx=L be ause the mass is uniformly distributed.
For the se ond omponent, use F = dp=dt = dt d (mv ). We want our answer as a fun tion of x, not of t, so we
manipulate some equations:
v (t) = gt
1 2
x(t) = gt
2
1 2
x(v ) = v
2g
v 2 (x) = 2gx
So now we have
d
(mv ) = 2gx
p dm
dt dt
The amount of mass hitting the s ale during a time dt will depend on the length already fallen x.
dm dx=dt M
=M = v
dt L L
So the total weight measured by the s ale is
x
W (x) = M g + v 2 (x)M=L = 3M gx=L
L
2 Lens
(25 pts.) You are given a ylindri al baby food jar, a lear liquid of unknown index of refra tion, a measuring
sti k, and a distant light sour e (say a bright distant vista). You an negle t the wall thi kness of the baby
food jar.
2.1 Snell's Law
(20 pts.) Design a measurement to determine the index of refra tion n of the liquid. Sket h the physi al
setup, showing the paths of representative light rays. Indi ate how the index of refra tion is to be al ulated.
Solution:
2.2 Spe ial Case
(5 pts.) What is spe ial about n = 2?

Solution:
1
3 Ro ket
(50 pts.) A ro ket ship propels itself through spa e by emitting exhaust at a onstant rate. The exhaust's
velo ity, as it leaves the ro ket, is u. The initial mass of the ro ket is M0 .
3.1 Ro ket Equation
(15 pts.) In the inertial frame, in whi h the ro ket was originally at rest, use Newtonian me hani s to show
that the mass of the ro ket M (v ) at speed v is:
M (v )
= e v=u
M0
Solution:
3.2 Relativisti Ro ket Equation
(25 pts.) Assume that the speed of exhaust, u, is bounded only by the speed of light (0 u ). Derive
the relativisti equivalent of the equation you found in the previous part.
Solution:
3.3 Classi al Limit
(10 pts.) Show that, in the lassi al limit (u ), the relativisti expression redu es to the lassi al one.
Solution:
2
Classi al Me hani s
Spring 1995 Qualifying Exam
gwlu.washington.edu
1 Lagrangian and Hamiltonian Dynami s

(100 pts.) A bead of mass m slides without fri tion along a straight wire that is rotating with onstant
angular velo ity
about a verti al axis. See the gure on the original question. The wire makes a xed
angle 0 with the rotation axis. There is a uniform downward gravitational for e mg on the bead.
Referen es: The rst few questions are Problem 3.2 in Theoreti al Me hani s of Parti les and Continua,
Fetter and Wale ka.
1.1 Lagrangian
(12 pts.) Constru t the Lagrangian for the bead using as generalized oordinate the distan e s measured
along the wire from the point of interse tion with the rotation axis.
Solution: This is pretty straightforward. The potential is just mgh where h is the height the bead is
raised, so we have
V = mgs os 0 + onstants
The kineti energy has a rotational omponent of 21 I
2 and a omponent from the motion along the wire,
2 ms_ . The moment of inertia of a point mass is I = mR . In this ase, the bead is moving in a ir le of
1 2 2
radius R = s sin 0 , so we have
1
T = m(s_ 2 + s2
2 sin2 0 )
2
These expressions give us the Lagrangian, L = T V .
1.2 Cir ular Orbit
(14 pts.) Obtain the Lagrange equation of motion for s and use it to show that the ondition for an
equilibrium ir ular orbit is s0 = g os 0 =(
sin 0 )2 .
Solution: You need to know how to get the equation of motion from the Lagrangian. Take the appropriate
derivatives, and we see that
ms ms
2 sin2 0 + mg os 0 = 0
The equilibrium ondition of ir ular motion means that the bead is not moving along the wire, so s = s_ = 0.
In this ase, we only have two terms in our equation of motion
ms0
2 sin2 0 + mg os 0 =0
We an easily solve this for s0 to get the desired result.
1.3 Newton's Se ond Law
(10 pts.) Derive the above result for s0 by dire tly applying Newton's se ond law of motion (F~ = m~a) in an
inertial frame. Hint: analyze the verti al and horizontal omponents of the for e of the wire on the bead in
the ase of a ir ular orbit.
Solution: You need to know how to draw a free-body diagram. The wire exerts a normal for e N on
the bead, perpendi ular to the wire. In the equilibrium orbit, the bead is not moving along the wire, so the
for es must sum to zero. The verti al omponent of the normal for e is N sin 0 , whi h must be equal to
mg . The horizontal omponent of the normal for e is N os 0 , whi h must be equal to the entripetal for e,
mr
2 = ms sin 0
2 . We an solve for N in one equation, and use that in the other to get the exa t same
equation as in the prvious question, yielding the same result.
1
1.4 Small Displa ements
(12 pts.) Disuss the stability of this orbit against small displa ements along the wire by nding an expression
for (t) = s(t) s0 .
Solution: Noti e that s = + s0 , and hen e s _ = _ and s = . Then our equation of motion is
m m( + s0 )
2 sin2 0 + mg os 0 = 0
Use the given expression for s0 , and we nd that the onstant terms an el, to give

2 sin2 0 = 0
We an solve this dierential equation to get

(t) = Aekt + Be kt
where k =
sin 0 . So, in general, this system will exponentially grow away from the equilibrium solution
when slightly perturbed.
1.5 For es of Constraint
(14 pts.) Use Lagrangian methods to nd the for e of onstraint a ting on the bead in the ^ dire tion. (
is the azimuth angle, with _ =
.)
Solution: You need to know how to do for e of onstraint problems. The usual question on for es of
onstraint requires you to temporarily pretend that a xed variable is no longer xed. We are looking for
the for e of onstraint in the ^ dire tion, so we let be ome another generalized oordinate. In this ase, we
just repla e
with _ in our Lagrangian, and take derivatives to get the new equation of motion. The for e
of onstraint is an external for e, so it goes on the right-hand side of the equation of motion. Remember
to take time-derivatives of s as well as _ , and we get
ms2 sin2 0 + 2mss__ sin2 0 =
Suddenly, we remember that is not really a variable, so = 0 and _ =

, and we get
= 2mss_
sin2 0
Note that this for e of onstraint is a tually a torque on the system, to keep it rotating about the z -axis
with angle 0 . So to get the for e, we write = ~r F~ , or F = 2ms_
sin 0 .
1.6 Coriolis For e
(14 pts.) Suppose the bead is moving outward along the wire with an instantaneous speed s_. Find the
magnitude and dire tion of the oriolis a eleration of the bead in a frame rotating with the wire, and
dedu e the for e of the wire on the bead needed to ompensate the oriolis a eleration. What is the
relationship between this for e and the for e of onstraint found in the previous question.
Solution: You need to know what the oriolis a eleration is. In a non-inertial frame, a system will feel
extra for es. When rotating, one of these for es is alled the oriolis for e, with
~abead = ~ainertial ~a oriolis
An expression for the oriolis a eleration is ~a = 2~! ~v . In this ase, we do the ross produ t with the
right-hand rule, and see that
~a oriolis = 2
s_ sin 0 ^
The bead feels the negative of this a eleration, whi h is in the ^ dire tion. The for e needed to ompensate
this is just m~a oriolis, whi h is equal to the for e of onstraint we found in the previous question. So we say
that the oriolis for e and the for e required to keep the wire turning at onstant angular velo ity make a
Newton's Third Law for e pair.
2
1.7 Hamiltonian
(12 pts.) Find H , the Hamiltonian of the bead, in terms of a suitable oordinate and momentum.
Solution: You need to know how to get anoni al momenta and express the Hamiltonian in terms of
them. We know that p = qL , so we know that the anoni al mometum in this ase is p = ms_. The
Hamiltonian is just
H=
X p q_ L =
p2 1 ms2
2 sin2 + mgs os

2m 2 0 0
1.8 Conserved Quantities
(12 pts.) Show whether H is onserved and whether H is equal to the sum of the kineti and potential
energies.
Solution: You need to know how to take total derivatives. First, we know that H 6= T + V by inspe tion.
This is be ause the system is being fed energy from some external sour e to keep the system moving at
onstant angular velo ity, regardless of the motion of the bead. We take the time-derivative of H to see if it
is onserved, and get
dH
dt
= H p H s H
p t
+ s t + t
The Hamiltonian, the momentum, and the oordinate have no expli it time depeden e, so this expression is
equal to zero. Thus, the Hamiltonian is indeed onserved, even though it is not the total energy.
3
Classi al Me hani s
gwlu.washington.edu
1 Coupled Pendula
(100 pts.) Consider a system of two identi al fri tionless pendula of length l and mass m oupled with a
massless spring of for e onstant k and moving in a plane. See the gure on the original question.
Referen es: The rst few questions are done in Se tion 23 of Theoreti al Me hani s of Parti les and
Continua, Fetter and Wale ka.
1.1 Lagrangian
(25 pts.) Assuming small angle displa ement, write the Lagrangian for this system in terms of x1 and x2 .
Spe ify any assumptions and show all work.
Solution: You need to know how to do small os illation problems. We make the small angle approxima-
tion, and also ignore the kineti energy from the verti al motion of the masses, as it is higher order than
we are working to. The potential energy of the masses has a omponent due to the spring, 12 k (x1 x2 )2 ,
and omponents from the height the masses are raised to. This height is l(1 os ). Make the small angle
approximation, noting that x = l sin l, and we get
1 mg 2
V = k (x1 x2 )2 + (x + x22 ) + onstants
2 2l 1
The kineti energy is due just to the horizontal motion, so we have
1
T = m(x_ 21 + x_ 22 )
2
These expressions give us the Lagrangian L = T V.
1.2 Equation of Motion

(15 pts.) What are the equations of motion for this system?
Solution: You need to know whi h derivatives to take. The Euler-Lagrange equations are
d L L
=0
dt q_ q
So we have mg
1 + k (x1
mx x2 ) + x1 =0
l
and mg
2
mx k (x1 x2 ) + x2 =0
l
1.3 Normal Modes

(25 pts.) Using the equations of motion, nd the natural frequen ies !i for this system.
Solution: You need to know the normal modes. The rst mode has the two pendula moving together in
the same dire tion. In this ase, the spring is not ompressed at all, and we an imagine that the pendula are
not oupled. We know the solution for a single simple pendulum, !12 = g=l, whi h we an verify be adding
the two equations of motion and setting x1 = x2 . The se ond normal mode is also simple: the pendula move
against ea h other, in the opposite dire tion. In this ase, x1 = x2 . This time subtra t the equations of
motion, and we get !22 = g=l + 2k=m. These solutions ould also have been obtained by assuming solutions
of the form xi (t) = Ai ei!i t and solving the matrix equations for !i .
1
1.4 Eigenfun tions
(10 pts.) Determine the motions of the system whi h os illate at ea h of the natural frequen ies, and des ribe
the motion for ea h ase qualitatively.
Solution: We have already des ribed the motions, and said that the motion was harmoni . To be expli it:
In the rst ase, the amplitude of motion is the same for both pendula. In the se ond ase, the amplitudes
have the same magnitude and opposite sign.
1.5 Dierent Masses

(15 pts.) Suppose the mass of one pendulum is hanged to M , where M > m.
1.5.1 Qui k
(5 pts.) One an determine without doing any further detailed al ulations one of the natural frequen ies of
this system. Find the frequen y and des ribe the orresponding motion.
Solution: You need to remember that the period of a simple pendulum does not depend on the mass. The
frequen y for our rst normal mode did not ontain any mention of the mass m. Even with one pendulum
signi antly heavier than another, this mode will remain (if our approximations remain valid). So one
frequen y of the new system is !12 = g=l.
1.5.2 Big Mass Dieren e

(10 pts.) Consider the ase where M m. Des ribe qualitatively the motion orresponding to the other
natural frequen y of the system. What is the value of this frequen y in the limit that m=M ! 0?
Solution: You need to be able to visualize this system. If one pendulum is mu h more massive than the
other, it will not move mu h during the os illation. To the lighter pendulum, this will be like being oupled
to a xed wall. To al ulate the frequen y of this os illation, we rst rewrite our equations of motion. Let
x2 be the heavier pendulum, and note that the equation for x1 is un hanged.
Mg
2
Mx k (x1 x2 ) + x2 =0
l
The easiest way to al ulate the frequen y is to write the equations in matrix form, and solve the determinant
equation. The matrix we are interested in an be read o the equations of motion.

k + gl m m! 2 k
det =0
k k + gl M M !2
Let x g
l !2 , and our equation is
(mx + k )(M x + k ) k2 = 0 = mM x2 + kx(m + M )
We an solve this for x to get x = k + 1). Now apply the limit, and go ba k to ! 2 to see
m (m=M
g k
!2 = +
l m
1.6 Weak Coupling

(10 pts.) On e again assume two pendulums with equal masses, but now let the length of one of the
pendulums be hanged to L, where L l. Assuming weak oupling (k=m g=L), what are the natural
frequen ies of the system?
Solution: You need to know that, to rst order, weak oupling means no oupling at all. In this ase,
the two pendula behave independently, with natural frequen y related to their individual lengths. Thus, we
have !12 = g=l and !22 = g=L.
2
Classi al Me hani s
gwlu.washington.edu
1 Lagrangian Me hani s
(100 pts.) Consider a thin hoop of radius R and mass M , free to os illate in its plane around a xed point
P . On the hoop there is a point mass, also M , whi h an slide freely along the hoop. The system is in a
uniform gravitational eld ~g. See the gure on the original question.
Referen es: This is Problem 4.1 in Theoreti al Me hani s of Parti les and Continua, Fetter and Wale ka.
1.1 Coordinates
(20 pts.) Introdu e appropriate oordinates des ribing the ombined motion of the hoop and point mass,
showing them on the diagram. How many un onstrained oordinates are required.
Solution: You need to remember how to pi k good oordinates for me hani s problems. We pi k two
angles, 1 and 2 . The rst is the angle between the verti al and the line between the xed point P and
the enter of the hoop. The se ond is the angle between the verti al and the line between the enter of the
hoop and the point mass. These two angles are suÆ ient to des ribe the state of the system. The next
question asks us to onsider small os illations, where we will swit h to 1 and 2 , the small displa ements
from verti al of the hoop's enter and the point mass, respe tively. I think it is easier to see the angles at
rst, and then move to the displa ements.
1.2 Lagrangian
(50 pts.) Consider small os illations. Derive the Lagrangian, and the Lagrange equations. Find the normal
mode eigenfrequen ies.
Solution: You need to know what the Lagrangian is, and how to get the Lagrange equations from it.
Our anoni al oordinates will be 1 and 2 , the small displa ements des ribed in the answer to the previous
question. Some of the quantities are easiest to write down rst in terms of the angles, and then swit h to
the displa ements by using small angle approximations. We know that L = T V , so we need the kineti
and potential energies of the system. First the potential:
V = Vhoop + Vpoint = M gR os 1 M g (R os 1 + R os 2 ) = M gR(2 os 1 + os 2 )
Now use the relation = R sin R and the approximation os 1 2 =2 to get

Mg
V = (2 2 + 22 ) + onstants
2R 1
For the kineti energy, we will have 12 mv 2 for the point mass, and 21 I! 2 for the hoop. The moment of inertia
of a hoop about its enter is I m = M R2 , so we use the parallel axis theorem and see that we need to use
I = 2M R2 in this ase. Be ause of the small angle approximation, the motion of the point mass in the
verti al dire tion is of higher order than we are working to, so is ignored. Using the relation _ = R os _,
we have
1 1 M
T = I _12 + (_ 1 + _ 2 )2 = (3_ 12 + _ 22 + 2_ 1 _ 2 )
2 2 2
And this gives us our total Lagrangian. Now remember the Lagrange equation
d L L
=0
dt q_ q
This is straighforward, and some basi algebra is omitted here, with the result being
g
31 + 2 + 2 1 = 0
R
1
g
1 + 2 + 2 = 0
R
Now, we an get one of the normal modes qui kly, be ause of the symmetry of the problem. One of the
solutions has to be a simple pendulum motion, with the hoop and bead moving in the same dire tion at all
times. In this ase, the displa ements of the hoop's enter of mass and the bead will always be the same:
1 = 2 . So we an add (or subtra t) our two Lagrange equations to get
g
2 + =0
R
where is both 1 and 2 . The solution
p to this is the familiar (t) = A os !1 t + B sin !1 t where we have
found the rst eigenfrequen y, !1 = 2gR . Now, we will rst give the easy way to get the other solution.
Intuition tells us that the other normal mode will be when the bead and hoop swing in opposite dire tions.
To keep the enter of mass dire tly below the pivot point P , we must have 2 = 21 , be ause 2 is measured
from the verti al line below 1 . So we again add (or subtra t) our equations, and use the relation between
the two displa ements to get
g
2 + 4 = 0
R
q 2g
whi h gives the other eigenfrequen y, !2 = R . Now for the hard way. Remember the matrix formalism
for Lagrangian dynami s. The eigenfrequen ies an be obtained by solving det jV ! 2 Mj = 0 where V and
M are the potential and mass matri es, obtained from the relation with their elements
1X
T = M _ _
2 ; ;
2
V; = q q
V

In this ase, we an read these elements right o the Lagrange equations. The mass matrix elements are
the oeÆ ients of the double time-derivatives, and the potential matrix elements are the oeÆ ients of the
displa ements themselves. So we have

3 1 2 0
M= 1 1
V = Rg 0 1
Then our equation to solve the determinant is
2g g
( 3! 2 )( !2) !4 = 0
R R
This is a quadrati equation in ! 2 whi h an be solved, after some tedious algebra, to reveal the two solutions
we saw before.
1.3 Eigenmodes
(30 pts.) Find the normal mode eigenfun tions and sket h the motion.
Solution: You need to know that the eigenve tors of a system satisfy the equation (V ! 2 M) = 0
for both 1 orresponding to !1 and for 2 orresponding to !2 . We saw in the previous question that one
of the normal modes has the hoop and bead moving together with the same amplitude, and so the same
equation held for both 1 and 2 . The other normal mode has dierent amplitudes for the hoop and bead.
The matrix equation an a tually be solved, but it is easier to remember what we said before, how the bead
will move in the opposite dire tion with twi e the amplitude as the hoop. So for the se ond normal mode,
the solution is
1 (t) = A os !2 t + B sin !2 t
2 (t) = 21 (t)
These motions an be easily sket hed.
2
Classi al Me hani s
gwlu.washington.edu
1 Bead on a Hoop
(45 pts.) A ir ular wire hoop is rotating about a verti al axis (along a diameter) with onstant angular
velo ity ! . A bead of mass m is free to slide without fri tion on the hoop. A onvenient denition is
!02 = g=R.
Referen es: The same system is Problem 3.1 in Theoreti al Me hani s of Parti les and Continua, Fetter
and Wale ka.
1.1 Label For es
(5 pts.) Draw all the for es (in the Lab frame) on the bead when it is in an equilibrium position (for
0 < < ). Also show the net for e (label it learly).
Solution: The equilibrium point for this system is when the bead is not moving along the hoop. The
weight is pulling straight down, with mg . There is a normal for e N perpendi ular to the hoop, pointing
to the enter of the hoop. The net for e is then pointing perpendi ular to the axis of rotation. The bead is
moving in a ir le, and the entripetal for e is towards the enter of that ir le.
1.2 Lagrangian
(5 pts.) Write the Lagrangian.

Solution: The kineti energy has two parts: one from the bead moving along the hoop, and one from the
motion of the bead about the axis of rotation. The perpendi ular distan e between the bead and the axis is
R sin . So we write
1
T = mR2 (_2 + ! 2 sin2 )
2
The potential energy is from the height raised, so
1
L=T V = mR2 (_2 + ! 2 sin2 ) + mgR os
2
1.3 Equation of Motion
(10 pts.) Derive the equation of motion from the Lagrangian.

Solution: Take the derivatives to nd
mR2 mR2 ! 2 sin os + mgR sin = 0
Divide by ommon fa tors to get the ni er form

g
! 2 sin os + sin = 0
R
1.4 Equilibrium
(10 pts.) Find the stable equilibrium position q of the bead as a fun tion of ! . There is a riti al value of
! , all it ! , below whi h the nature of the equilibrium hanges. Find ! . Des ribe the nature of the hange.
Solution: At equilibrium, the bead is not moving along the hoop: = _ = 0. Plug into the equation of
motion
! 2 sin q os q + !02 sin q = 0
1
Solve for q
!0 2
os q =
!
We an use this expression to hara terize the riti al angular velo ity. A solution exists for q only if ! !0 .
As ! approa hes !0 from above, the equilibrium angle de reases towards zero. Previously, the equilibrium
point of = 0 was unstable. Now, the rotation velo ity is so slow that the two equilbria merge, as there is
not enough rotational energy to move the bead out along the hoop. So, the riti al velo ity is the natural
velo ity, ! = !0 .
1.5 Small Os illations
(15 pts.) Find the frequen y of small os illations of the mass about the equilibrium point q . Assume the
angular velo ity is above the riti al value ! and 0 < q < .
Solution: We are looking at a small deviation from the equilibrium point q . So we say that (t) = q + (t)
where (t) is small. Then _ = _ and = . So the equation of motion be omes
! 2 sin(q + ) os(q + ) + !02 sin(q + ) = 0
Obviously, we need to use trig identites and the small angle approximation. Also, drop terms of order 2 ,
and use the relation we found above between q , ! and !0 . After a lot of boring algebra, we nd
+ (! 2 sin2 q ) = 0
This is the harmoni os illation equation, so we an read o the frequen y of small os illations ! = ! sin q .
0
2 Moment of Inertia Tensor

(25 pts.) The inertia tensor of an obje t omposed of point masses mp is
!
X X3
k 2 i j
Iij = mp Æij (xp ) xp xp
p k=1
2.1 About an Axis
(10 pts.) Show that

^ I n^
In = n
is the moment of inertia about an axis spe ied by the unit ve tor n^ .
Solution:
2.2 Angular Momentum
(10 pts.) If Iij = Ii is diagonal and ~! is in an arbitrary dire tion, derive an expli it expression for the angular
momentum L ~ in terms of the omponents of ! ~ and Ii . Work this out using subs ript notation.
Solution:
2.3 Prin iple Axes
(5 pts.) Under what general onditions will the angular momentum ve tor point in the dire tion of the
angular velo ity ve tor?
Solution:
2
3 Spa eship Orbits
(30 pts.) You wish to send a spa e raft from Earth to Jupiter via a fuel-eÆ ient Hohmann transfer orbit.
This is an ellipti al orbit around the sun tangent to both the Earth's orbit and to Jupiter's orbit. Assume that
the orbits of Earth and Jupiter are ir ular and oplanar and ignore other planets as well as the gravitational
attra tion of the spa e raft to Earth and Jupiter. Call the radius of Jupiter's orbit RJ and the radius of
Earth's orbit Re = 1 AU. For general information (you don't need this number) RJ 5:2 AU. The velo ity
of the Earth in orbit is ve = 2 AU / year.
3.1 Ellipses and Energy
(10 pts.) A relationship between the semi-major axis a of an ellipse and the total energy E is
m
a=
2jE j
where = GMs and m is the mass of the orbiting obje t. Derive this for the spe ial ase of a ir ular orbit.
Solution: We an write an expression for the total energy, rst in general
1 m
E = T + V = mv 2
2 r
For a ir ular orbit the entripetal for e must an el out the attra tive gravitational for e: mv 2 =r = m=r2 .
Solve this relation for v 2 and plug into the total energy expression to get
1 m
E=
2 r
This an easily be solved for r to get the desired relation.
3.2 Orbit Speed
(15 pts.) What speed does the spa e raft have when it rea hes the orbit of Jupiter? Give your answer in
terms of the speed of Jupiter in its orbit vJ and the orbit radii RJ and Re . Is the spa e raft speed larger or
smaller than the speed of Jupiter in its orbit?
Solution: The spa e raft will travel in an ellipse with semi-major axis of a = (Re + RJ )=2. Use this in
the given relation to get
m
E=
Re + RJ
At Jupiter's orbit, we also know that E = 21 mv 2 m=RJ . Equate these relations and solve for v 2 . We are
assuming Jupiter's orbit is ir ular, so we an say the velo ity of Jupiter in its orbit is given by vJ2 = =RJ .
So we get
2Re
v 2 = vJ2
Re + RJ
Sin e Jupiter's orbit is greater than Earth's, the spa e raft is travelling slower than Jupiter at this point.
3.3 Speed Change
(5 pts.) How does the speed of the spa e raft when it rea hes Jupiter's orbit ompare with its speed when
it started its ellipti al path at Earth's orbit (give the answer in terms of orbit radii) ?
Solution: Use onservation of angular momentum. The spa e raft's velo ity at Earth's orbit is ve and at
Jupiter it is vJ . Then
L = mRe ve = mRJ vJ
Solve to see
Re
vJ = ve
RJ
3
Ele tri ity and Magnetism
gwlu.washington.edu
1 Radiation (Mostly)
For mono hromati ally varying harge distributions (~r; t) = (~r)e i!t the s alar potential is also mono hro-
mati and given in the radiation gauge (in gs units) by (~r; t) = (~r)e i!t where
Z i!j~r ~r0 j=
(~r) = d3 r 0 ~
r0
( ) e j~r r0
~ j
Similarly, the urrent distribution and ve tor potential are also mono hromati : ~j (~r; t) = ~j (~r)e i!t and
~ (~
A r ; t) = A
~ (~
r )e i!t . The equation for the ve tor potential A
~ (~
r ) is similar to that for (~
r ) but with (~
r)
repla ed by the urrent distribution j (~r) in the above integral.
~
1.1 Stati Limit
(25 pts.) Show that for \small" sour es at \short" distan es, the frequen y dependen e in (~r) is negligible,
and hen e, one re overs the \stati " limit given by Coulomb's Law. Roughly estimate the extent of this
\short distan e" regime for 1000 Mhz signals and omment on possible impli ations for high speed omputer
pu hips whi h have typi al dimensions of about 5 m.
Referen es: The stati limit (and the radiation zone limit) is dis ussed in Se tion 9.1 of Classi al Ele -
trodynami s, Ja kson.
Solution: You need to gure out what \small" and \short" mean. In this ase, we assume that the sour e
has dimensions d, and that \small" means r d. This allows us to ignore the ~r0 bit in the argument of
the exponential. However, we are also at a \short" distan e, so r or r 2! . We an write this as
!r 2 , so that the argument of the exponential is approximately zero. So the exponential gets repla ed
by a one, and we obtain Z 0
(~r) = d3 r0 j~r(~r ~r)0 j
whi h is just the familiar Coulomb's law, and has no frequen y dependen e. Now we need to plug in the
right numbers: we want an estimate of the distan e r that makes the exponential go to unity
r

=
= 3:0 108 m = 0:047m
! 2f 2 109
This is about the size of a pro essor, as given by the problem. So the radiation from a pro essor's lo k
will reate a stati eld throughout part of the pu, whi h ould alter the expe ted behavior of the other
elements of the pro essor.
1.2 Legendre Polynomials
(25 pts.) Consider a uniform harge distribution (~r) of net harge Q onned to a loop of radius a
in the x y plane with negligible wire thi kness. In speri al oordinates the harge density is (~r) =
(Q=2a2)Æ( os )Æ(r a). Determine the rst two non-vanishing terms of the expansion of the potential
(~r) in Legendre polynomials Pl ( os ) and powers of rl for short distan e r a. Hint: rst determine
the potential along the z -axis and then mat h; the leading Legendre polynomials are P0 (x) = 1, P1 (x) = x,
P2 (x) = (3x2 1)=2.
Referen es: This problem is the last example in Se tion 3.3 of Classi al Ele trodynami s, Ja kson.
Solution: First, you need to realize that this has nothing to do with the previous question. Then, you
need to get the potential or a ring of harge along the axis, the expansion of an azimuthally symmetri
1
system in Legendre polynomials, and the Taylor series expansion of (1 + x) 1=2 . We take the author's hint
and get the potential along the z -axis, whi h is a al ulation we did in high s hool, but have of ourse sin e
forgotten. Also, we drop the usual 1=40 be ause the author wants us to use gs units.
(z ) = p 2Q 2
a +z
The expansion is Legendre polynomials is a useful thing to remember,

X
(r; ) = (Al rl + rBl+1l )Pl ( os )
l
Sin e we are lose in, we set all the Bl to zero. Now equate the two expressions on the axis (r = z , = 0) :
Q X
p = Al rl Pl (1)
a2 + z 2 l
Now remember that Pl (1) = 1, and that

p11+ x = 1 x2 +
We only need two non-vanishing terms, so hopefully this is enough. Let x (z=a)2 , and
Q
p
1 Q
(1 1 (z=a)2 )
a 2
1 + (z=a) a 2
So we see that for all odd l, Al will be zero, whi h we might have guessed had we pondered the system's
symmetry a bit more. So the rst two non-vanishing terms are A0 = Q=a and A2 = Q=2a3. Then our
expression for the potential is
2
(r; ) = Qa Q2 ar 3 P2 ( os )
1.3
(25 pts.) From Maxwell's equations, show that the ele tri eld E~ (~r) / r B~ (~r) for all points outside the
harge and urrent distribution, and determine the proportionality onstant. For very large distan es (the
\radiation limit") and small sour es, you may approximate j~r ~r0 j r, eikj~r ~r j =j~r ~r0 j (eikr =r)e i~k~r ,
0 0
r = r^ r where k = r^!= . In the radiation limit B

~
R 3 0 0
~ = ~beikr =r. Determine an expli it expression for ~b in
terms of ~j~k = d r ~j (~r )e i~k
0
~r and the observation dire tion r^.
Solution: First realize that we are ba k to the radition problem. Now you need to apply the right
Maxwell equation. We are outside the urrent and harge region, so r B~ = 12 tE~ . The ele tri eld is
also mono hromati , so we have r B~ = i !2 E~ or E~ = i !2 r B~ . Now to get the radition magneti eld,
rst make approximations in the expression for the ve tor potential.
Z i!j~r ~r0 j= Z
eikr ikr
d3 r0~j (~
r0 ) d3 r0~j ~
r0 e ik~r
e
A( )=
~ ~
r
j~r r0 j
~
r
( ) ~ 0
= er ~j~
k
Now B~ = r A~ , so we get (dropping terms of order 1=r2 be ause we are in the radiation zone)
ikr ikr
~
B = (^r r ) A~ = ik e r (^r ~j~k ) ~b e r
where ~b = i~k ~j~k .
2
1.4
(25 pts.) Show that the elds E~ and B~ = ~beikr =r are orthogonal to ea h other and to the ve tor ~r in the
\radiation limit", and evaluate the time averaged power radiated into solid angle d
in dire tion k^ in this
limit in terms of ~b.
Solution: The magneti eld is easy, be ause B ^ = r^ ^j~k , so it is orthogonal to ~r. For the ele tri eld,
use the previous result, and remember that we an approximate r by r^ r , so we get
2 ~ 2
~
E = i ! (^r rB ) = i ! (ikr^ B~ ) = ^ B~
r
whi h is also orthogonal to ~r. Now you need to remember the formula for the radiated power
dP
= 1 <[r2 r^ E~ B~ ℄ = 1 <[ r2 r^ (^r B~ ) B~ ℄ = 1 <[ r2 eikr e ikr r^ (^r ~b) ~b ℄
d
20 20 20 r r
Rearrange the ross produ ts to an el the minus sign, and realize that ~b is orthogonal to ~r, and we get
dP ~2
d
= 2 jbj 0
3
Ele tri ity and Magnetism
gwlu.washington.edu
1 Fields, Energy, Momentum, and Torque
(50 pts.) An innitely long ylinder of insulating material with radius a, permeability = 0 , and permit-
tivity = 0 has uniform volume harge density > 0, surfa e harge density and is ele tri ally neutral.
It is pla ed in a onstant, uniform magneti eld B~ = B z^ that is ylindri ally symmetri of radius RB with
RB > a and of innite extent in the z -dire tion. The symmetry axis of the ylinder is ollinear with the
symmetry axis of the magneti eld. See the gure on the original question. The ylinder is free to rotate
about its symmetry axis.
1.1 Ele tri Field
(10 pts.) Find the ele tri eld E~ everywhere.

Solution: You need to know Gauss' Law. The symmetry of the system shows us that the ele tri eld
will be independent of z and be azimuthally symmetri . Consider a Gaussian ylinder of length l and radius
r. If r > a, there is no en losed harge (be ause the total ylinder is neutral), so E
~ = 0. Inside the ylinder,
2
the en losed harge is r l. The integral of the ele tri eld on the Gaussian surfa e is just 2rlE . Move
some fa tors around to get E~ = 2r0 r^.
1.2 Poynting Ve tor
(15 pts.) Compute the Poynting ve tor and momentum density everywhere. Find the angular momentum
per unit length, L~ , with the ylinder at rest.
Solution: You need to know an expression for the Poynting ve tor and the momentum density. We just
found E~ , and the system is stati , so the only magneti eld is the onstant external eld. The Poynting
ve tor is S~ = 10 E~ B~ . Be ause the ele tri eld is zero for r > a, so will the Poynting ve tor be. Then for
r < a, we get
~ = 1 E
S ~ B ~ = rB (^ r z^) =
rB 2 ^
0 002 2
1
The momentum density is just ~g = 2 S~ . To get the total angular momentum, we remember that
Z
~
L = ~
r ~g(~r)d3 r
We will perform this integral out to r = a, as the momentum density vanishes beyond that. We get a fa tor
of l for the integral along z , and a 2 from the azimuthal integral. Remember that the dierential element
in this ase will then be rdr. So the angular momentum per unit length is just
Za 4
~ = 2 B (^
L=l r ^) r3 dr = B a z^
2 0 4
1.3 Faraday's Law
(13 pts.) The magneti eld is now turned o with time dependen e B~ (t). Determine the ele tri eld indu ed
by the time-varying magneti eld. Assume that any velo ity at any point in the harge distribution is small
and relativisti ee ts an be negle ted.
Solution: You need to apply Faraday's Law. The ele tromotive for e is equal to the negative of the
hange in magneti ux. In this ase, the magneti ux is just
(r; t) = B (t)r2
1
This holds for r < RB . Outside this, just repla e r with RB . The ele tromotive for e will be 2rE . For
r < RB we will get
~ = 1 dB (t) r
E ^
2 dt
For r RB we get
2
E~ = 1 dB (t) RB ^
2 dt r
1.4 Torque
(12 pts.) Find the torque per unit length, ~ , on the ylinder as a fun tion of time and determine the angular
momentum per unit length of the ylinder as a fun tion of time. Assume that any velo ity at any point in
the harge distribution is small and relativisti ee ts an be negle ted.
Solution: You need to know that torque is the hange in angular momentum. Now that the magneti
eld is hanging, we have two omponents of the ele tri eld. Consequently, the Poynting ve tor now has
two omponents. By doing the ross produ ts, we see that S~ points in the ^ dire tion as before, and now also
in the r^ dire tion. To get the angular momentum, do the same al ulation as before. Lu kily for us, taking
the ross produ t r^ r^ gives us zero. So the angular momentum is the same, with onstant B repla ed by
B (t). Now take the time derivative to get
~ dB t a4
( ) z^
~
= ddtL =
dt 4
2 Green's Fun tions
(50 pts.) A loop of radius r0 and harge Q is above a grounded, ondu ting sphere, \S ", of radius a. See
the gure on the original question. The plane of the loop is displa ed verti ally from the plane of re e tion
symmetry of S by a distan e d. The entire system is embedded in a diele tri of innite extent with diele tri
onstant . Note: Some useful relations:
1 = 4 X 1 Y (
0 )Y (
) r<l
j~r ~r0 j l;m
2l + 1 lm lm l+1
r>
r
Yl0 = Æm0
2l + 1 P ( os )
4 l
where r< and r> represent the lesser and greater of j~rj and j~r0 j, respe tively.
The Green's fun tion G(~r; X~ )> for the equation determining the ele trostati potential in the diele tri
region outside the ondu ting sphere is given by
" #
G(~ ~ )> =
r; X
1 1 a 1
4 j~r
~ ~X j jX j j~r
~ Y
^ 2 =jX~ j, with the origin of the oordinate

where ~r is the eld position, X~ is the sour e position, and Y~ = Xa
system being the enter of the ondu ting sphere.
2.1 Boundary Conditions
(12 pts.) From Maxwell's equations, derive the boundary onditions satised by an ele trostati eld at the
interfa e between diele tri and ondu ting sphere.
Solution: You need to apply Gauss' and Faraday's Laws appropriately. From Maxwell's equations in
dierential form, we an integrate using Stoke's Theorem and the Divergen e Theorem to get the integral
forms of Faraday's and Gauss' Laws. We are in the stati ase, so a line integral of the ele tri eld must be
zero. Pi k a urve that runs parallel to the interfa e, with one side in the ondu tor, and the other outside.
2
Make the pie e of the urve perpendi ular to the surfa e go to zero, and we see that the parallel omponents
of the ele tri eld must vanish. Now draw a Gaussian pillbox, again with one side in and the other out of
the ondu tor. Let the sides perpendi ular to the surfa e go to zero again, and let the other surfa es have
area dA. The harge en losed is just dA. The ele tri eld inside a ondu tor is zero. So the ele tri eld
perpendi ular to the interfa e is just D? = .
2.2 Properties of the Green's Fun tion
(12 pts.) What properties does G(~r; X~ )> possess to make it the appropriate Green's fun tion? Draw a
s hemati to show the lo ation of the sphere, a point harge and any image harges that might arise.
Referen es: This problem is done in Se tion 2.2 of Classi al Ele trodynami s, Ja kson.
Solution: You need to know about Green's fun tions. The form of the given Green's fun tions is that of
a point harge outside, and the image harge we imagine at a parti ular lo ation in the sphere. Lu kily for
us, the Green's fun tion is given, and does not need to be derived. Noti e that the magnitude of the image
harge is a=jX~ j. The image harge is lo ated along the same line as from the enter of sphere to the outside
point harge. Noti e that as the harge outside goes to innity, the lo ation of the image harge goes to the
enter of the sphere, and the magnitude of the harge goes to zero. Now remember that a Green's fun tions
is the potential of a system due to a point harge. Maxwell's equations show that you an write r2 V = .
So the Green's fun tion must satisfy
r2 G(~r; X~ )> = Æ3 ~
(r r0
~ )
Another property of the Green's fun tion is the boundary onditions it satises. This is a Diri hlet Green's
fun tion, so it is zero on the boundary. You an see this by examining a point on the surfa e of the ondu ting
sphere, where the image harge appears in the same pla e with opposite harge, so there is zero potential.
2.3 Legendre Polynomials
(14 pts.) Express the potential in the diele tri resulting from the harge loop as a sum of Legendre
polynomials.
Solution: We are given the expansion in spheri al harmoni s for 1=j~ r0 j, so we should be able to work
r ~
with the Green's fun tion to get the desired form. Remember that if we are given the Green's fun tion, we
an get the potential for a whole distribution of harge by
Z
(~r) = (~r0 )G(~r; ~r0 )d3 r0
In this ase ~r0 X~ . We an write the harge distribution in spheri al oordinates as

Q
~
r( ) = 2r Æ (r R Æ ) ( os os 0 )
0
p
where R = d2 + r02 and os 0 = d=R. Using the enter of the sphere as the origin, we an write the Green's
fun tion in terms of sums of spheri al harmoni s. The system is azimuthally symmetri , so we must have
m = 0, so we an move to Legendre polynomials. Noti e that in the se ond term of the Green's fun tion,
the ve tor Y~ always has magnitude less than ~r, so we an write r< = jY~ j and r> = j~rj in that term. Now
perform the integral and we get
rl
Q 1 X a2l+1
(~r) = 2r 4 Pl ( os 0 )Pl ( os ) l+1 (rR)l+1
<
0 l
r>
where r< and r> are the smaller and greater, respe tively, of j~rj and R.
3
2.4 Indu ed Charge
(12 pts.) Find the harge indu ed on the ondu ting sphere.
Solution: There will be a surfa e harge indu ed on the sphere to simulate the behavior of our imagined
image harge. Read o the magnitude of the image harge from the Green's fun tion: qimage = a=jX~ j. In
this ase, jX~ j = R. So for ea h dQ of the ring, there will be an image harge dQa=R. Sin e the total harge
of the ring is Q, the total image harge will be Qa=R. This is the total harge indu ed on the surfa e of
the ondu tor.
4
Modern Physi s
gwlu.washington.edu
You may nd the following useful: h = 1:973 keV Angstroms, me 2 = 511 keV.
1 Leptons
(20 pts.) The ele tron (e), muon, (), and tau ( ) parti les are harged parti les whi h belong to the lass
of elementary parti les alled leptons. There are also neutral leptons alled neutrinos. The leptons are all
fermions.
1.1 For es
(10 pts.) Whi h of the four known intera tions (gravity, ele tromagneti , weak, and strong) do the harged
leptons exhibit?
Solution: Gravity a ts on all massive parti les, in luding the harged leptons. Charge is the unit of the
ele tromagneti for e. The de ay of the heavier harged leptons are all weak. The strong for e is between
quarks, and sin e the harged leptons are elementary parti les, they annot be omposed of quarks and do
not intera t via the strong for e.
1.2 De ays
(10 pts.) A positively harged pion is observed to de ay mainly into a muon and a neutrino
+ ! +
with the positive muon subsequently de aying into a positron and two neutrinos
+ ! e+
where the bar signies an anti-neutrino. Explain why you get one neutrino in the rst ase and two in the
se ond.
Solution: The de ays must onserve the lepton number of ea h avor. Ea h lepton has the opposite
lepton number as its sister neutrino. So in pion de ay, the lepton number before was zero, so we need one
- avored neutrino to make the lepton number of the nal state zero also. In the de ay of the muon, we
need a - avored anti-neutrino in the nal state to onserve lepton number in the muon family. Then we
need a positron-anti-ele tron neutrino pair to onserve ele tron avored lepton number.
2 Muoni Hydrogen
(50 pts.)
2.1 Energy Levels
(10 pts.) What is the approximate energy of muoni hydrogen in its n-th ex ited state; i.e., a proton whi h
has aptured a muon in an atomi orbit. The mass of the muon in units of the ele tron mass is m = 207me
and the mass of the proton is mp = 1836me .
Solution: You need to remember the formula for the energy levels of the hydrogen atom:
2 e4 1
En =
2h2 n2
1
where is the redu ed mass. In a plain hydrogen, this is approximated well by me . Sin e the muon is mu h
heavier than the ele tron, we must deal with its ontribution to the redu ed mass:
m mp m 207me
= = =
m + mp 1 + m =mp 1 + 207=1836
188me
So the energy levels of muoni hydrogen will be about 188 times those of plain hydrogen. We remember that
the binding energy of plain hydrogen is 13:6 eV, and we an write, for muoni hydrogen,
1 1
En = 188 13:6 2 = 2:5 keV
n n
2.2 Level Transition
(10 pts.) Cal ulate the energy of the photon whi h is emitted when the muoni hydrogen undergoes a
transition from the rst ex ited state (n = 2) to the ground state (n = 1).
Solution: This is just the dieren e between energy levels:
1 1

E = E2 E1 = 2:5keV 2 = 1:9keV
2 12
2.3 Bohr Radius
(10 pts.) What is the Bohr radius in the ground state of muoni hydrogen (radius in n = 1 state)?
Solution: You need to remember the denition of the Bohr radius
h 2 2
a0 =
e2
mh e4
e
So for muoni hydrogen, multiply the ele tron mass by 188 to get the new Bohr radius. If you remember
the Bohr radius of plain hydrogen is about half an Angstrom, we nd that the muoni hydrogen has a mu h
larger radius, almost 100 Angstroms.
2.4 Lifetime
(10 pts.) Compare the lifetime for the n = 2 ! n = 1 transition for muoni hydrogen and ele tri hydrogen.
Dis uss.
Solution:
2.5 Comparing Lifetimes
(10 pts.) Given muoni hydrogen in the state n = 2, ompare the lifetimes for the transition ! e versus
the transition n = 2 ! n = 1.
Solution:
3 Quark Model
(30 pts.)
3.1 Protons and Neutrons
(10 pts.) A ording to the quark model, what ombinations of quarks onstitute a proton and a neutron?
Solution: This problem is totally unfair, as you just have to know these things. The proton is made of
two up quarks and one down. The neutron is made of two down quarks and one up.
2
3.2 Charge and Spin
(10 pts.) What are the ele tri harges and spins of ea h of these onstituents?
Solution: Quarks are fermions, so they all have spin 1=2. The harge of an up quark is +2=3 and the
down is 1=3.
3.3 Masses
(10 pts.) What are the relative masses of the onstituents?

Solution: The down quark is about twi e as massive as the up.
3
Modern Physi s
gwlu.washington.edu
1 Big Bang Spe trum

(30 pts.) A ording to the Big Bang Theory, about 1 million years after the Bang, the Universe was about
1000 times smaller (in ea h dimension) than it is today, and lled with radiation of energy density des ribed
by the bla k-body spe trum (with T about 3000 K)
8h 1
(; T ) =
5 eh =kB T 1
1.1 Radiation Spe trum

(20 pts.) Derive the spe trum of that radiation today.
Referen es: The osmi ba kground radiation is dis ussed in Problems 4.17 and 4.18 of Thermal Physi s,
Kittel and Kroemer.
Solution: You have to not think too hard about this question. The tri k is to ask what has hanged sin e
the Big Bang. The only quantities in the energy spe trum that ould have hanged are and T . So, all the
expansion fa tor in ea h dimension S = 1000. We are assuming the Universe ontains modes of radiation,
just like a bla k box. So if we in rease the volume of our box, the wavelength of ea h mode in the box will
in rease, by a fa tor S in ea h dimension, for a total fa tor of S 3 . This means that the energy per mode
de reases. We know that E = h =, so we now want to use E = h = = h =S. We are looking for the
0 0
energy spe trum today, all it ( ; T ), whi h is related to the given spe trum by the fa tors just mentioned.
0 0
Consider the energy in a small pie e d of the spe trum

0
1 E 0
0 (0 ; T )d0 = (; T )d

S3 E
So we an write
1 d 1
0 (0 ; T ) =(; T ) = 5 (; T ) = ( ; T ) 0
S 4 d S 0
This shows that the spe trum of the ba kground radiation today is still that of a bla k-body. Further,
the ba kground radiation is now at a temperature of T 2:8K . The original author's solution says that
T = T=S , but does not mention why. Perhaps this an be explained with the Stefan-Boltzmann law, whi h
0
says / T 4 . However, this gives T = T=S 5=4 .

0
Beware this problem: The explanation behind the solution is sket hy! Further, the form of the bla k-body
spe trum given here is dierent from what I have seen before, so the entire question may be suspe t!
1.2 Ba kground Isotropy

(10 pts.) What information do we gain from measuring the deviation of the osmi ba kground radiation
from isotropy, i.e. from measuring the multipole moments: the dipole moment, and the higher terms (i.e.
small angular s ales) ?
Solution: You pretty mu h have to just know this. The dipole moment is primarily due to the motion
of the Sun with respe t to the rest frame of the universe. See for example the results of the COBE mission.
The higher order terms an provide a wealth of information, in luding the distribution of the primordial
ele tron-proton \soup." For example, su h information might validate the in ationary theory of the Universe.
2 Spe ial Relativity

(35 pts.) Consider the de ay of a parti le of mass m1 , energy E1 and three-momentum ~p1 into a parti le
of mass m2 , energy E2 and three-momentum p~2 and a photon with energy E and three-momentum p~. The
1
angle between the dire tion of the photon and the dire tion of parti le 1 is . See the gure on the original
question.
2.1 Parti le De ay
(20 pts.) Show that the photon energy is given by
m21 4 m22 4
E=
2m1 2 1 (1 1 os )
Hint: you will need 1 = p1 =E1 and 1 = E1 =m1 2 .
Solution: You need to nd the right ve tors to work with, and work some algebra. We know two things:
onservation of energy and momentum. From energy onservation we get E1 = E + E2 . From momentum
onservation we get ~p1 = ~p + ~p2 (the fa tor of makes the units the same as the energy equation). The
idea is to absolute square these equations, and then bring them together in the hopes that some stu will
an el. We are given the angle , but not the angle of the parti le with mass m2 . Sin e we will need to take
a dot produ t somewhere, it makes sense to move the equations around before squaring them, so that we
will end up with an expression involving , instead of the unknown angle. So now we have
E12 + E 2 2E1 E = E22
2 p21 + 2 p2 + 2 2 p1 p os = 2 p22
Subtra t the momentum equation from the energy equation. Note that if we had written the whole thing in
terms of momentum 4-ve tors, we would have obtained this equation a few steps sooner.
(E12 2 p21 ) + (E 2 2 p2 ) 2E1 E 2 2 p1 p os = (E22 2 p22 )
Now remember that E 2 = p2 2 + m2 4 , and that for the photon, E = p . So we have
m21 4 + 0 2E (E1 + p1 os ) = m22 4
Now use the relations E1 = m1 2 1 and 1 = p1 =E1 and solve for E to get the solution
m21 4 m22 4
E=
2m1 2 1 (1 1 os )
2.2 Doppler Ee t
(15 pts.) Use this result to obtain the expression for the Doppler ee t for ele tromagneti radiation.
Referen es: The relativisti Doppler shift is dis ussed in Se tion 11.3 of Classi al Ele trodynami s, Ja k-
son.
Solution: The wording of this question is un lear, but here goes. We are given the expression for the
photon's energy, all it E0 , in a parti ular frame of referen e. Now move to the rest frame of the de aying
parti le, where 1 = 0 and 1 = 1. Call the energy of the photon seen from this frame E , and we have 0
m21 4 m22 4
E0 =
2m21 2
Ba k in the lab, we see that E = E0 1 (1
0
1 os ). We know that E = h! , so we an write the standard
expression for the Doppler shift
! 0 = !0 1 (1 1 os )
3 Experimental Issues
(35 pts.)
2
3.1 Fixed vs. Collider Dete tors
(15 pts.) Explain brie y the main advantage, as well as the main disadvantage, of a xed-target experiment
as ompared with a olliding-beam experiment ( onsider only symmetri al olliders, i.e. same energy for
both beams).
Solution: Some of this you just have to know. The main advantage, obviously, is that the laboratory frame
is now also the enter of mass frame, so essentially all the energy of the beam is available for produ tion of
other parti les. The disadvantages are a bit more un lear, but an in lude te hni al diÆ ulty in building the
ma hines, and the result of lower intensity of parti les, be ause we no longer have one side of the experiment
extraordinarily more dense than the other.
3.2 Bubble Chambers

(20 pts.) Explain brie y the pro esses responsible for the two unrelated events in a Hydrogen Bubble
Chamber depi ted in the gure. The bubble density indi ates that all visible tra ks are ele trons and/or
positrons. See the gure on the original question.
Solution: You need to know something about bubble hambers. The two events seem to appear from
nowhere, so they are the de ays or ollisions of a neutral parti le (only harged parti les an ionize the
hydrogen and leave a trail in the bubble hamber). The three-pronged event ontains an ele tron and a
positron, as seen by the tightly urved tra ks. The magneti eld that permeates the hamber will ause the
ele tron to urve one way, and the positron the other. The third parti le has the same harge as the parti le
on the right, be ause it urves the same way. However, it is obviously mu h more energeti , as it does not
urve tightly. The solution suggests that this might be a neutron olliding with an ele tron, produ ing three
ele trons. The event at the top of the gure is two energeti , oppositely harged parti les. A good guess
might be pair produ tion from a gamma ray.
3
Quantum Me hani s
gwlu.washington.edu
1 Rotations
(30 pts.) In quantum me hani s, the ee t of a rotation on operators and states is given by the rotation
operator Rn() = e iJn, where hJ is the total angular momentum operator, and n is a unit ve tor spe ifying
the rotation axis. (The exponential operator an be dened by its Taylor expansion.)
1.1 S alars
(10 pts.) An operator is alled a s alar if it does not hange under spatial rotations. By onsidering an
innitesimal rotation d, show that a s alar operator must ommute with any omponent of J.
Solution: You need to know how to apply a rotation. Let our s alar operator be alled A, and we rotate
it by d about the n axis to get A0 = Rn(d)y ARn(d). By denition A0 must be equal to A, be ause it is a
s alar operator. Sin e this is an innitesimal rotation, we an expand the exponential to get
A0 = (1 + iJn d)A(1 iJn d) = A
where Jn is the omponent of J in the n dire tion. We ignore terms of order d2, and nd that
A = A + id(Jn A AJn ) = A + id[Jn ; A℄
The ommutator must therefore be zero. Sin e we did not spe ify the rotation axis n, this holds for any
omponent of J.
1.2 Ve tors
(10 pts.) A ve tor operator transforms under spatial rotations in the same way as an ordinary ve tor; e.g.
for a rotation about the z-axis by an angle , the new omponents of a ve tor operator K are given in terms
of the old omponents by:
Kx0 = Kx os Ky sin Ky0 = Kx sin + Ky os Kz0 = Kz
By onsidering innitesimal rotations as in the previous question, prove the following ommutation relations
between the omponents of J and a ve tor operator K:
[Jx; Ky ℄ = iKz and y li permutations
Solution: Consider a rotation by d about the z-axis. Applying small angle approximations to the given
expression for a rotated ve tor, we see that
Kx0 = Kx dKy Ky0 = dKx + Ky Kz0 = Kz
Now apply the rotation to the x- omponent of K to see

Kx0 = (1 + iJz d)Kx (1 iJz d) = Kx + id[Jz ; Kx ℄
Equate the two relations for Kx0 to see that [Jz ; Kx℄ = iKy . If we wanted, we ould do the same thing for
the other omponents, and for rotations about other axes, proving all the relations.
1
1.3 Sele tion Rules
(10 pts.) The omponents of the ele tri eld in ir ularly polarized light are E0(x iy) when the light
is propagating along the z-axis. Consider ele tri dipole absorption of su h ir ularly polarized light by an
atom of total angular momentum J. Use the ve tor ommutation relations to derive the sele tion rules for
the hange in MJ along z. The ele tri dipole intera tion is er E.
Solution: You have to gure out what the question is asking. We are asked to nd a sele tion rule for m
along the z-axis, given a spe i intera tion. Presumably, we will derive a relation between the new m0 and
the old m. The intera tion is given as H1 = er E. We are given the ele tri eld for ir ularly polarized
light. Commonly, when given expressions of the form Ax iAy it is bene ial to all these A. Using
this notation for r and E, we an write the Hamiltonian as e(r+ E + r E + + zEz ). In this ase, the z-
omponent will be zero, and we will have only one of the plus-or-minus terms, depending on the polarization
of the in oming light. The operator in question is r, and we an use the ommutation relations we proved
earlier to nd that
[Jz ; r ℄ = r
Finally, we an start to answer the question. The energy shift due to this intera tion is
E = hm0jH1jmi = hm0 j er E jmi
Now insert the ommutator in pla e of r , and we get
E = ehm0j([Jz ; r ℄)E jmi
This gives us a Jz a ting to the left, and another a ting to the right. This will yield the m0 and m that we
need. So we have
E = e(m0 m)hm0 jr E jmi = (m0 m)E
This relation gives us the sele tion rule: m0 = m 1 for E .
2 Time Evolution
(20 pts.) The magneti dipole intera tion of J with a magneti eld B is des ribed by the intera tion
Hamiltonian:
Hmag = h J B
where is proportional to the magneti dipole moment.
2.1 Pre ession
(10 pts.) Classi ally, the angular momentum pre esses about B. To see the quantum me hani al analog,
show that the time evolution aused by the magneti eld is given by the rotation operator. In other words,
using the intera tion pi ture in whi h the time dependen e of the state ve tor arises only from Hmag , and
writing
jti = Umag (t t0 )jt0 i
show that
Umag (t t0 ) = Rn ()
where n lies along B, and = B(t t0).
Solution: You need to know an expression for the evolution operator, U (t) = exp( iHt=h). We are given
the Hamiltonian, so we an write for the evolution operator
Umag (t t0 ) = e iHmag (t t0 )=h = ei JB(t t0 ) = ei B (t t0 )Jn
But this is the expression for Rn () with = B(t t0 ), so we have shown it.
2
2.2 Expe tation Value
(10 pts.) Show that the expe tation value, hKi, of a ve tor operator rotates with angular frequen y B.
Solution: We have been told that a state ve tor at time t is just the evolution operator times the state
ve tor at time t0. Let K be a ve tor operator, and we an write
hK it = htjK jti = ht0 jU y (t t0 )KU (t t0 )jt0 i = ht0 jRny ()KRn ()jt0 i
where = B(t t0) as in the previous question. So the expe tation value of K at time t is the expe tation
value at t0, rotated by . So the time dependen e is exp( i!(t t0)) where ! = B.
3 Pauli Spin Matri es
(20 pts.) The omponents of ~, the Pauli spin operator, are given here in the usual representation:

x =
0 1 y = 0 i z = 1 0
1 0 i 0 0 1
For J = 1=2, show that

Rn () = os
2 i~ n sin 2
Solution: You need to know that J = ~=2 for spin 1=2, and some properties of the Pauli matri es. Now
expand the exponential form of the rotation operator to get
X1 ( i =2)k
Rn () = e in =2 =
n
k=0
k!
Now remember that nk = 1 for k even, and nk = n for k odd. Then we an write
X ( i=2)k X ( i=2)n
Rn () = + n
keven
k! kodd
k!
Now re ognize these sums as the expansions for osine and i times sine, respe tively, and we have the desired
result.
4 Neutron Interferometer
(30 pts.) Consider a neutron interferometer (see gure on original question). A beam of neutrons (spin 1/2)
is split by a partial re e tor into two beams of equal intensity that travel along paths 1 and 2, and re ombine
at the dete tor D. A magneti eld B an be added along path 1. Under the semi- lassi al onditions that
are usual in su h an interferometer, the neutrons spend a well-dened time T traversing the magneti eld
region, and the ee t of applying B in that region is simply to make the following hange in the spinor wave
fun tion rea hing the dete tor along path 1:
1 (D ) ! Umag (T ) 1 (D )
(The dire tion of B relative to the propagation of the neutron beam makes no dieren e.)
4.1 Counting Rate
(10 pts.) Plot the neutron ounting rate you would expe t to see at D as a fun tion of BT over the range
0 to 8.
Solution: The ounting rate will be the intensity at the dete tor, or j 1 (D) + 2 (D)j2 . In the absen e
of the magneti eld, we would have 1(D) = 2 (D). So in the presen e of the magneti eld, we an say
3
that 1(D) = U (T ) 2(D). We showed that U was equal to Rn(), and we derived an expression for Rn()
in terms of sine and osines. So we an write

I = j 2 (D)j2 j1 + Rn ()j2 = j 2 (D)j2 j1 + os in sin j2
2 2
Perform the absolute square, and the terms linear in n an el. Remember that n2 = 1, and we get
= j 2(D)j2 (2 + 2 os 2 )
I
In this expression = BT , so the ounting rate is a osine wave with period 4.
4.2 Blo ked Path
(10 pts.) Repeat the previous question for the ase when path 2 is blo ked. Be quantitative in your
omparison.
Solution: Intuitively, the ount rate should now be a onstant, be ause there is no interferen e. The
alignment of the neutrons will still hange, but they will still rea h the dete tor. Quantitatively, we have
I = j 1(D)j2 = I0 jRn()j2 = I0j os 2 in sin 2 j2
where I0 would be the intensity of path 1 in the absen e of the magneti eld. Again perform the absolute
square, and we get I = I0 , a onstant. The j 2 (D)j2 in the previous ase is I0 in this ase. So the onstant
ounting rate with path 2 blo ked will be one-fourth of that when both paths are unobstru ted with no
magneti eld.
4.3 Explain
(10 pts.) Interpret your answers in terms of the behavior of spin 1=2 systems under spatial rotations of 2
and 4. Su h experiments illustrating the rotational properties of neutrons a tually have been arried out
in re ent years.
Solution: Take the expression for Rn () and plug in the two values. We see that Rn(2) = 1 and
Rn (4 ) = 1. So we on lude that spin 1/2 parti les have period 4 for spatial rotations. This might be
ounter-intuitive, as lassi al parti les always have 2 symmetry for rotations.
4
Quantum Me hani s
gwlu.washington.edu
1 Potential Well
(80 pts.) A deep and narrow attra tive potential well V (x) entered on the origin of a one-dimensional system
an be represented by boundary onditions on the wave fun tion (x) and its derivative 0 (x) = d (x)=dx,
of the form
(0+) = (0 ) 0 (0+) = 0 (0 ) (0)
where is a positive onstant. This equation says that the wave fun tion is ontinuous at the origin, but
its derivative has a dis ontinuity proportional to the value of the wave fun tion. Everywhere ex ept at the
origin the wave fun tion satises the time independent S hroedinger equation
h 2 d2
E =
2m dx2
1.1 Bound State

(20 pts.) Show that there is a bound stationary state for this potential, sket h its wave fun tion, and give
the analyti form (you are not required to normalize the wave fun tion).
Solution: To nd the bound state, we need to solve the dierential equation for (x) with negative
energy. This is an equation we have solved many times before, where the se ond derivative is a onstant
times the original wave fun tion. So we guess the solution

(x) = Aejxj = A (x)e x + ( x)ex
where A is some normalization. This is an exponential de ay on both sides of the origin. Now we have
E = h 2 2 =2m, whi h is learly negative, so this is indeed a bound state. Use the boundary ondition
equation for the derivative to nd
A = A A
whi h gives = =2. So we have an expression for (x) in terms of the given quantities.
1.2 Energy of Bound State

(10pts.) What is the energy of this bound state?
Solution: We have the energy
h 2 2 h 2 2
E= =
2m 8m
1.3 Symmetry
(5 pts.) What is the symmetry about the origin of this bound state?
Solution: Intuition tells us that the potential is so simple that the bound state should be an even fun tion.
Indeed, we an use our expression for to nd that (x) = ( x).
1.4 Stationary State

(10 pts.) For positive energy E = h2 k 2 =2m, show that there is a stationary state of the form
(x) = A os(k jxj + Æ+ )
and nd an equation for Æ+ in terms of k .
1
Solution: Given the form of the solution, it is lear that the se ond derivative of (x) is k 2 (x), so it
indeed solves the equation for the given energy. It may help to write the wave fun tion in two pie es
(x) = A ((x) os(kx + Æ+ ) + ( x) os( kx + Æ+ ))
and the derivative is
0 (x) = kA ((x) sin(kx + Æ+ ) ( x) sin( kx + Æ+ ))
Now the boundary ondition on the derivative is learly
kA sin Æ+ = kA sin Æ+ A os Æ+
Solve this to see that tan Æ+ = =2k .
1.5 Find a Stationary State

(10 pts.) Find a positive energy stationary state with E = h2 k 2 =2m whi h has odd symmetry about the
origin.
Solution: We need to solve the dierential equation again, with positive energy and odd symmetry. Sin e
the osine works, lets try the losest odd fun tion, the sine: (x) = A sin(kx). Clearly we have the required
positive energy. Sin e the sine is an odd fun tion, we don't need to bother with the absolute value. This wave
fun tion is ontinuous at the origin, and be ause its value is zero at the origin, its derivative is ontinuous
as well.
1.6 Re e tion and Transmission

(25 pts.) What is the probability that a parti le of energy E = h2 k 2 =2m in ident from the left will be
re e ted?
Solution: We don't have any restri tions on the symmetry now, so we an write the wave fun tion in
terms of omplex exponentials. On the left side, we will have the in oming and re e ted waves, and on the
right we will have only the transmitted wave. We an write this as

(x) = ( x) eikx + Re ikx + (x)T eikx
whi h satises the dierential equation with the required energy. Now he k the boundary onditions. The
ontinuity of the wave fun tion gives us
1 + R = T = (0)
The dis ontinuity of the derivative gives us
ik(1 R) = ikT (1 + R)
Eliminate T and solve for R to get
R = 1=(2ik= 1)
The probability of re e tion is thus
jRj2 = 1=(1 + 4k2 =2 )
2 Bosons in a 1-D Box
(20 pts.) The normalized wave fun tions for the lowest two eigenstates of a single spinless boson onned to
the one-dimensional regon 0 < x < L are
r r
2 x 2 2x
0 (x) = L
sin
L 1 (x) = L
sin
L
with energies
h2 2 2h2 2
E0 = E1 =
2mL2 mL2
2
2.1 Two Bosons: Energies
(10 pts.) What are the energies of the lowest two eigenstates for two non-intera ting bosons onned to su h
a region?
Solution: The key point is that we are dealing with bosons. So the lowest energy state of two parti les
is when both parti les are in the single-parti le ground state. The lowest energy is thus
h2 2
E2 bosons;0 = 2E1 boson;0 =
mL2
The next highest state has one parti le still in the ground state, and the other parti le in the next ex ited
state.
5h2 2
E2 bosons;1 = E1 boson;0 + E1 boson;1 =
2mL2
2.2 Two Bosons: Wave Fun tions

(10 pts.) What are the normalized eigenfun tions of the lowest two eigenstates for two non-intera ting bosons
onned to su h a region?
Solution: The wave fun tion of the two-boson system is the produ t of the wave fun tions for the single
parti le states. There is only one way to put both parti les in the ground state, so we have
2 x1 x2
0 (x1 ; x2 ) = 0 (x1) 0 (x2) = L sin L sin L
Sin e the single-parti le wave fun tions were normalized, this produ t is also. There are two ways we an
onstru t the next ex ited state, be ause the parti le are non identi al. The requires an extra normalization,
and we see that
p
x 2x2 2x1 x
1 (x1 ; x2 ) = p1 ( 0 (x1 ) 1 (x2) + 1 (x1 ) 0 (x2)) =
L
2
sin 1 sin
L L
+ sin
L
sin 2
L
2
3
Quantum Me hani s
gwlu.washington.edu
1 Density Matri es
(40 pts.)
Referen es: Density matri es, their properties, and appli ations thereof are dis ussed in Prin iple of
Quantum Me hani s , Shankar, end of se tion 4.2, and also in Modern Quantum Me hani s, Sakurai, Se tion
3.4.
1.1
(10 pts.) Explain brie y in what ir umstan es one should des ribe a physi al system using a density matrix
rather than a wavefun tion.
Solution: You just have to know this. A density operator is useful in situations that are slightly more
physi ally realisti than a system of pure states, where we pi k states from an ensemble, from whi h we know
the probabilities of pi king ea h type of state.
1.2
(10 pts.) Explain why the density matrix for an ensemble of spin-1=2 parti les an always be written in the
form
1
= (I + ~a ~
)
2
where ~a is a real ve tor and ~ are the Pauli matri es

x =
0 1 y =
0 i
z =
1 0
1 0 i 0 0 1
Solution: You need to know some things about the density matrix . First, the Pauli spin matri es and
the identity matrix form a omplete basis for 2 2 matri es. So any matrix A an be written as A = 0 I +~ ~ ,
where 0 ; ~ are, in general, omplex. Now we use properties of the density matrix: it is Hermitian ( = y )
and Tr = 1. These properties an be derived from the denitions given as \potentially useful results." So
we must have 0 = 1=2 and ~ a real ve tor. The fa tor of one-half (~ = 21 ~a) is by onvention, and you just
had to know it (or guess).
1.3
(5 pts.) Show that the ensemble average of the spin is given by hS~ i = (h=2)~a.
Solution: You need to know that the ensemble average of any quantity A ~ is given by hA
~ i = TrA
~ . Now
use S = (h=2)~ and properties of the Pauli matri es. These an be derived empiri ally, as they are given on
~
the exam, but are good to know.
Tri = 0 Tri j = 2Æij
So we break into omponents and get
X X
hS~i i = h4 Tr(Ii + ai i j ) = 0 + h4 ai 2Æij = h2 ai
j j
1
1.4
(15 pts.) The spins are pla ed in a magneti eld aligned in the z -dire tion, su h that the Hamiltonian is
H = !Sz , and are kept at temperature T . The density operator des ribing this situation is
=
1 exp H Z = Tr(exp H )
Z kB T kB T
Cal ulate hS~ i as a fun tion of T , and omment on the result in the limits T ! 0 and T ! 1.
Referen es: This question very similar to the last example in Se tion 3.4 of Sakurai.
Solution: You have to see that we an do our al ulations using the Sz basis. Then we have
1 1
exp h ! 0 exp h ! 0
=
Z
!Sz
exp k T = Z 2
0
k B T
exp h ! Z = Tr 2
0
k B T
exp h ! = 2 osh 2kh!T
B 2 kB T 2 kB T B
Now we an get the ensemble average of omponents

h
h 1
exp h ! 0

hSi i = 2 Tri = 2 Z Tr 2
0
k B T
exp 2khB!T i
Noti e that the matrix produ t will have zeros in the diagonal for x and y , so only hSz i will be non-zero.
We also ould have seen this by noting that we have made ~a lie along the z a-axis and using the result of the
previous question. So we have
hSz i = h2 Z1 ( 2 sinh 2kh !T ) = h2 tanh 2kh !T
B B
by using the expression for Z obtained earlier. As T ! 0 from above, the tanh ! 1, so we see all the spins
pointing along the z -axis, whi h is the lowest energy state. As T ! 1, however, the tanh ! 0, and we see
that the spins be ome randomly oriented, as there is no longer a predili tion for one dire tion over another.
2 S attering Theory
(60 pts.)
Referen es: The s attering ross se tion is dened in any quantum me hani s textbook.
2.1
(5 pts.) Dene the dierential s attering ross se tion d=d

in physi al terms.
Solution: This is something that should have been drilled into your head by the time you take the
qualifying exam. Then we imagine a beam of in oming parti les striking a target, and onsider a ross-
se tional are of the beam d. A \long" distan e away, a dete tor subtends a pie e of solid angle d
. Then
the denition of the dierential ross se tion is
d
d
=
Number of parti les s attered into d
perunittime
d
Number of in ident parti les rossing unit area per unit time
It is often helpful to write
d
d
= r2 jjjjs atteredjj
in ident
where j means parti le ux.
2
2.2
(20 pts.) The s attering wave fun tion at large distan es from the s attering enter has the form
ikr
(+)
(r) / eikr + e r f (; )
Explain why the dierential s attering ross se tion is given by
d

= jf (; )j
2
d
Referen es: This problem was on the PHYS 519 Midterm Exam in Spring 2000.
Solution: You need to know how to do ux al ulations on the wavefun tions. You are given the equation
for ux, J = mh =( r ). The outgoing wavefun tion is given. It is omposed of the in oming plane wave
(eikr ) and an outgoing spheri al wave. So we an immediately al ulate the in ident ux

J = =(e ikr ik r^eikr = k r^
h h
in
m m
2.3
(15 pts.) Parti les of mass m s atter o a ylindri al potential well of uniform strength V0 , radius R, and
length L, with R L. See gure on original question: this might be a rude representation of a deformed
nu leus.) In this part we make no assumption on erning the initial dire tion of these parti les. Assuming
the potential to be weak, given an approximate expression for the s attering amplitude, and explain brie y
the range of validity of this expression.
Solution: You need to re ognize that the assumption of a weak potential means we an use the rst Born
approximation. We are given expression for the s attering fun tion
f (k~0 ; ~
k) =
1 2m (2)3 hk~0 jV j (+) i
4 h 2
The rst Born approximation lets us substitute a plane wave (eik~ x~ =(2)3=2 ) for the full outgoing wavefun -
0 0
tion (hx~0 j (+) i). So some 's an el, and we get

f (k~0 ; ~
k) =
1 2m Z d3 x0 V (x~0 )ei(~k k~ )x~
0 0
4 h 2
2.4
(10 pts.) Evaluate the dierential ross-se tion, d=d

, in the lowe energy limit (kR 1), for parti les
approa hing along the axis of the ylinder (marked \1" on the original gure). Explain why the result does
not depend on the initial dire tion.
Solution:
2.5
(10 pts.) How would you distinguish experimentally between s attering in the dire tions marked \1" and \2"
in the original gure? In other words, how might you determine, from the distribution of s attered parti les,
whether the ylinder is oriented along the in oming dire tion or perpendi ular to it? At what order in kR
would su h an ee t rst appear?
Solution:
3
Quantum Me hani s
gwlu.washington.edu
1 Ele tron in a Magneti Field

(60 pts.) An ele tron in a spheri ally symmetri potential V (r) is in an energy eigenstate with quantum
numbers S = 1=2, L = 1, J = 1=2. (S is the spin quantum number, L is the orbital angular momentum
quantum number, J is the total angular momentum quantum number.) The ele tron magneti moment is
he ~
~ = L + g S~
2m
Here e and m are the harge and mass of the ele tron, is the speed of light, and g is the gyromagneti
ratio. At time t = 0 a magneti eld is swit hed on with magnitude B pointing in the z dire tion.
1.1 Hamiltonian
(4 pts.) Write down an expression for the Hamiltonian, in luding the spin-orbit intera tion.
Solution: The rst two terms should be automati : p =2m + V (r ). Next is the intera tion between the
2
magneti moment of the ele tron and the external magneti eld, whi h has the form ~ B ~ . Finally, the
~ S~ with ompli ated oeÆ ients, depending on the potential. So the total
spin-orbit intera tion has an L
Hamiltonian is
p2 1 1 V ~ ~
H= + V (r) ~ B
~ LS
2m 2m2 2 r r
You an now assume that the magneti eld is arbitarily weak.
1.2 Good Quantum Numbers
(8 pts.) Whi h ombination of the quantum numbers J; L; S; Jz ; Lz ; Sz are good quantum numbers for the
energy eigenstates of this system?
Solution:
1.3 Expe tation Value
(17 pts.) Cal ulate the expe tation value of the omponent of the magneti moment whi h is along the J~
dire tion. ~
J^. Note that we an write J^ = J~ jJJ2j . So we have
Solution: We need to nd ~
h e ~ jJ~j
~ J^ = L + gS~ J~ 2
2m J
Use J~ = L
~ + S~ and J 2 = L2 + S 2 + 2L ~ S~ to write this as
~
~ J^ =
he 1
L2 + gS 2 + (1 + g ) (J 2 L2 S2
jJ j
2m 2 J2
We are given the quantum numbers s, l, and j , so now use J 2 = j (j + 1); L2 = l(l + 1); S 2 = s(s + 1) to get
the result
he jJ~j
~ J^ = (4 g )
2m 3
1
1.4 Pre ession
(17 pts.) At time t = 0, assume Jx = +1=2. Cal ulate the time dependen e of the expe tation value of Jx .
Solution:
1.5 Strong Field
(12 pts.) Will your answer about the set of good quantum numbers hange if the magneti eld is made
strong? If not, explain why not. If so, explain how strong the magneti eld has to be in order to hange
your on lusion.
Solution:
2 Harmoni Os illator
(16 pts.) 4 neutrons and 3 protons are in a 3D harmoni os illator potential with
X ~p2i 1
H= + k~ri2
i
2m 2
where the sum on i is over the parti les. There are no intera tions between the parti les. Find the ground
state energy of this system.
Solution: Note that this is a three-dimensional os iallator, so the energy levels are
3
E (nx ; ny ; nz ) = ( + nx + ny + nz )h!
2
where ! 2 = k=m, and we are negle ting the small mass dieren e between neutrons and protons. The single
parti le ground state has energy E0 = 23 h! , and there are two of these, one for ea h spin state. The single
parti le rst ex ited state has energy E1 = 25 h! , and there are 6 of these, be ause we have three dimensions
to work with. We an onsider the neutrons and protons separately, as our ability to distinguish them means
that we an put them in the same state without violating the Pauli ex ulsion prin iple. So we an put 2
neutrons and 2 protons in the nx = ny = nz = 0 state. The other 2 neutrons and proton go in the single
parti le rst ex ited state. So the total energy is
3 5 27
E = 4( h ! ) + 3( h! ) = h!
2 2 2
3 States
(24 pts.) A quantum me hani al parti le an be found in one of three orthogonal states r , w , and b . These
three states form a omplete basis. R is an operator whi h y li ally permutes the states, e.g. R b = w ,
R w = r , R r = b . The parti le Hamiltonian is
H = !h(R + R ) y
3.1 Observables
(6 pts.) Is R an observable? Explain why or why not.

Solution: The Hamiltonian uses both R and R , implying that R is not Hermitian, as we will nd
y
expli itly in the next part. Observables are, by denition, Hermitian, so R annot be an observable.
2
3.2 Eigenvalues and Eigenstates
(18 pts.) Find the energy eigenvalues and eigenstates of the parti le.
Solution: Let the three orthogonal states be the olumn ve tors (1; 0; 0); (0; 1; 0); (0; 0; 1). Then using the
a tions of R given, we an onstru t 0 1
0 1 0
R= 0 0 1 A
1 0 0
from whi h we get the Hamiltonian 0 1
0 1 1
H = h! 1 0 1 A
1 1 0
Now we have a simple matrix, nd its eigenvalues and eigenve tors any way you hoose. We get
0 1 0 1
1 1
H 1 A = 2h! 1 A
1 1
0 1 0 1
1 1
H 0 A = h! 0 A
1 1
0 1 0 1
1 1
H 1 A = h! 1 A
0 0
3
Statisti al Me hani s
gwlu.washington.edu
1 Identities and Gases

(60 pts.)
1.1 Gibbs-Duhem Relation
(10 pts.) Given the rst law of thermodynami s dE = T dS + dN pdV and that the energy is extensive
E = T S + N pV , derive the Gibbs-Duhem relation dp = (S=V )dT + nd where n N=V .
Solution: You need to do some algebra, and take a total dierential. Take the total dierential of E to
get
dE = T dS + dT S + dN + dN pdV dpV
Combine this with the given expression for dE to nd that
SdT + Nd V dp = 0
Solve this for dp and use n = N=V to get
S
dp = dT + nd
V
1.2 Free Energy
(10 pts.) It is often more onvenient to work with the Helmholtz free energy per unit volume, f (E T S )=V .
Show that df = (S=V )dT + dn.
Solution: Again you have to take a dierential. We are given an expression for f , so take a dierential
of that. Be ause f is the free energy per unit volume, we do not need to take dierentials of V . So we have
1 1 S
df = (dE SdT T dS ) = (T dS + dN SdT T dS ) = dn dT
V V V
1.3 Imperfe t Gas
(10 pts.) Consider the free energy per unit volume whi h des ribes an imperfe t gas:
k T
f (T; n) = wnv0 (1 nv0 ) + B [(1 nv0 ) log(1 nv0 ) + nv0 log(nv0 )℄
v0
where v0 is a unit volume, and w is an energy per unit volume. Assume that the parti les in the gas attra t
one another, in whi h ase w > 0. Show that the pressure is given by
kB T
p(T; n) = w(nv0 )2 log(1 nv0 )
v0
Solution: You need to do some algebra and equation manipulation. Combining the expressions for f and
E , we an write p = n f . So now we need an expression for . Use the relation for df we obtained in the
previous question and divide by dn to write
f S T
=
n V n
Now realize that T=n is zero, and we have p = nf=n f . The derivative is ugly, but we an do it.
There is a lot of an ellation, and the desired result falls out.
1
1.4 Gas To Liquid
(15 pts.) At suÆ iently low temperatures, the gas is expe ted to ondense to a liquid. What fa tor(s) favor
this ondensation, and what oppose it? From your argument (and simple dimensional analysis), what is a
reasonable guess for an expression for the transition temperature?
Solution: You need to reason this out however you an. We have been told that w has units of energy
per unit volume, and is attra tive, or favorable to ondensation. The expression for f has an energy part
and an entropy part. The se ond term in the imperfe t gas expression for f must be the entropy part.
This is positive, and thus opposes the ondensation. Presumably, the transition will o ur around where the
favoring and opposing terms balan e out. For this to be the ase, the two terms must an el ea h other out.
The leading order omponent of the energy term is w, while kB T=v0 is the orresponding omponent of the
entropy term. So a guess as to the transition temperature might be kB T wv0 .
1.5 Criti al Point
(15 pts.) At the riti al point at whi h the transition rst o urs in the system,
p 2p
=0 =0
n n2
Using this, al ulate the lo ation, n and T , of the riti al point. At temperatures lower than this, there is
two-phase oexisten e. (You need not show this.)
Solution: You need to take derivatives arefully.
p k T
= 2wv02 n + B
n 1 nv0
and
2p vk T
= 2wv02 + 0 B 2
n 2 (1 nv0 )
We have two equations and two unknowns, so we an solve this. After some algebra, we see that n = 1=2v0
and kB T = v0 w=2. Armed with this result, we ould go ba k and answer the previous result, if it had eluded
us before.
2 Quantum Harmoni Os illator

(40 pts.)
2.1 Partition Fun tion
(20 pts.) The energy levels of a one-dimensional harmoni os illator are given by En = (n + 1=2)h! . Show
that the partition fun tion of the system is
1
Z=
2 sinh(h!=2kB T )
Solution: This is su h a ommon problem that we should be able to do it in our sleep. You need to know
Z=
Xe
the denition of the partition fun tion. In this ase we have
En =kB T = e h !=2kB T
Xe nh !=kB T
n n
Let x exp( h!=kB T ) and we have a geometri series in x, so we get
h !
e 2kB T
Z= h !
1 e kB T
Multiplying by exp( h!=2kB T ) on the top and bottom gives us the desired form of the hyberboli sine.
2
2.2 Entropy
(20 pts.) Ca lulate the entropy S (T ). Show that even in the lassi al limit, it still depends upon h. Brie y
give a reason why this should be so. This behavior of the entropy is in ontrast to the energy E (T ) whi h,
in this same limit, does not depend on h. (You need not show this.)
Solution: You need to know the relationships between Z , F , and S . We know that F = kB T log Z .
The exponential form is a bit easier to work with, giving
1
F = h ! + kB T log(1 e h !=kB T )
2
Now we use
F e h !=kB T h ! h! 1
S= = k log(1 e h !=kB T ) + kB T h !=k T = k log(1 e h !=kB T ) + h !=k T
T 1 e B kB T 2 T e B 1
Now you need to remember what the lassi al limit is for the quantum os illator: kB T h!. In this ase,
the arguments of the exponential will be small, so we an expand to get
S ! kB log(kB T=h!) + kB
So the entropy does indeed depend on h even in the lassi al limit. The entropy is related to the logarithm
of the number of possible ombinations for the state of the system. Re all that the lassi al energy of an
os illator is kB T , and this orroborates our expression for S . Sin e the os illator is truly a quantum system
at heart, the number of possible states should involve h, while the total energy may not.
3
gwlu.washington.edu
1 Quantum System / Partition Fun tion

(30 pts.) A parti ular quantum system has two dierent energy levels, the ground energy E0 = 0 and
the ex ited energy E1 . The ground state is unique (i.e. has degenera y of one), and the ex ited state has
degenera y n (i.e. there are n distinguishable states of energy E1 ). At a temperature of T :
Referen es: This problem is very similar to Thermal Physi s, Kittel and Kroemer, Chapter 3, Problem
1.
1.1
(6 pts.) Give the free energy of the system.

Solution: You need to know that F = kB T log Z where Z is the partition fun tion.
Z(T ) =
X e Ei=kB T
i
where the sum is over all the states of the system, and Ei is the energy of the ith state. In this ase we have
Z = e E0=kB T + ne E1 =kB T = 1 + ne E1 =kB T
So we have F .
1.2
(6 pts.) Give the probability that the ex ited state will be o upied.
Solution: You need to know that the probability of a given state being o upied is the Boltzmann fa tor
for that state diveded by the partition fun tion:
e Ei =kB T
P (Ei ) =
Z
In this system, the ex ited state is n-fold degenerate, so the answer is that
ne E1=kB T 1
P (E1 ) = =
Z 1 + n eE1 =kB T
1
The form of this expression tells us that as the energy of the ex ited state in reases, the system will more
likely be found in the ground state, as a transition be omes more ostly. However, as the ex ited state
be omes more degenerate, the extra possibilities de rease the ost of transition.
1.3
(6 pts.) Give the average energy of the system.

Solution: You need to know that
U = hE i =
X P (E )E
s s
s
where U is the average energy. This expression is exa tly an expe tation value from quantum me hani s We
know the probability from the previous question, and the ground state has energy equal to zero, so
E1
U = P (E0 )E0 + P (E1 )E1 =
1 + n eE1 =kB T
1
1
1.4
(12 pts.) Give the entropy of the system. Cal ulate or state the values as T ! 0 and T ! 1. Explain why
these results are reasonable.
Solution: You need to know either F = U T S or S = F T . We already have expressions for U and F ,
so we have
U F E1 =T
S= = + kB log(1 + ne E1=kB T )
T 1 + n1 eE1 =kB T
As T ! 0, the rst term goes like e 1=T =T to zero The se ond term also goes to log 1 = 0, be ause the
exponential is negative. As the temperature de reases, the entire system will move into the ground state,
and will be ompletely ordered, so the entropy should be zero. As T ! 1, the rst term goes to zero like
1=T , and the se ond term goes to kB log(1 + n). Note that this answer diers from the one provided by the
question's author! Another expression for entropy is S = kB log where is the number of possible states.
As the temperature in reases, the dieren e between the ground and ex ited states will be ome small in
omparison. So the number of possible states is n + 1, and our expression is orre t.
2 Quantum Harmoni Os illator

(25 pts.) Consider a harmoni os illator whi h has energy levels En = (n + 1=2)h!.
Referen es: This problem is very similar to Thermal Physi s, Kittel and Kroemer, Chapter 3, Problem
3. Note that the book problem ignores the ground state energy of 12 h!.
2.1
(8 pts.) Give an estimate of the temperature T , below whi h signi ant deviations from lassi al behavior
o ur.
Solution: Classi ally, we would hara terize the energy of the os illator as kB T . So we expe t deviations
from lassi al behavior when kB T h! is not true.
2.2
(8 pts.) Cal ulate the entropy of the harmoni os illator.

Solution: We do this al ulation mu h like the previous problem. First we obtain the partition fun tion:
Z=
X e En=kB T = e h !
2kB T
Xe h !
kB T n
n n
P
Remember the sum of a geometri series n xn = 1=(1 x) and we get
h !
e 2kB T
Z= h !
1 e kB T
Now we an get
1
F = kB T log Z = h ! + kB T log(1 e h !=kB T )
2
Finally we use the derivative to get the entropy
F h!=T
S= = kB log(1 e h !=kB T ) + h !=kB T
T e 1
2
2.3
(9 pts.) Take the lassi al limit for the entropy and give a physi al explanation for its value.
Solution: In the ase the lassi al limit means kB T h!. So we expand the exponential e h !=kB T !
1 khB!T and hen e
h! k T
S ! kB log(1 1 + ) + kB ! kB log B
kB T h!
We an ingnore the extra kB be ause the rst term will be mu h greater in the lassi al limit. We said
earlier that the lassi al energy is kB T , so the expression khB!T is the index of the average o upied energy
level. This ts with our previous knowldege that the entropy is proportional to the logarithm of the number
of possible states.
3
gwlu.washington.edu
1 Heat Capa ity
(30 pts.) Suppose that ions with spin s = 1=2, embedded in ertain weird media, undergo a transformation
from a paramagneti to a ferromagneti state. A phenomenologi al temperature dependen e for the molar
heat apa ity asso iated with these spins is given by
T n n
T
C (T ) = C0
1
T2 T2
for T2 T T1. The transition takes pla e between T1 and T2. A fully ferromagneti state exists below
T1 . Above T2 , a fully disordered paramagneti state exists. The nature of the weird medium determines the
parameters n, T1 , and T2 . However, C0 is fully determined by these parameters and fundamental onstants.
Referen es: Ferromagneti and paramagneti states are dened in many basi texts, for example Halliday
and Resni k.
1.1 Entropy
(7 pts.) What is the entropy per mole of the spins below T1?
Solution: You need to know what ferromagneti means. The system is ferromagneti below T1 . In a
ferromagneti material, the spins are all lined up, even in the absen e of an external magneti eld. Sin e
the ions are all lined up, there is perfe t order in the system. The number of ombinations that make up
this state is = 1, so the entropy, by S = kB log , is zero.
1.2 More Entropy
(7 pts.) What is the entropy per mole of the spins above T2?
Solution: You need to know what paramagneti means. The spins in a paramagneti material will line
up only in the presen e of an external magneti eld. If there is no applied eld, the spins will be randomly
oriented. In this ase, we know that spin is 1=2, so there are two possibilities for ea h ion. So for N ions,
the entropy is S = kB log 2N . We are looking for the entropy per mole, and we know that N kB = R, so
S = R log 2.
1.3 Even More Entropy
(8 pts.) Using the phenomenologi al molar heat apa ity given above, al ulate the entropy per mole between
T1 and T2 .
Solution: You need to know an expression for the heat apa ity in terms of the entropy: Cv = T
T . We
S
have an expression for Cv , so we should able to do

Z Cv
S = dT
T
and get the entropy. This integral is not too diÆ ult, and we get
n n
1 T T1
S = C0 log T
n T2 T2
However, the integral we wrote down was indenite, so we need a onstant of integration. In this ase, we
have already said that S (T1) = 0, so we mat h that to get, after some algebra
T n T n T n
C0 T
S = 1
n
1
log
n T2 T2 T2 T1
1
1.4
(8 pts.) Find an expression for C0 in terms of n, T1 , and T2 using the results of the previous questions.
Solution: You need to have gotten the previous questions. We used one end point to get the expression
for Cv in the last question. Now we use the other end point to get an expression for C0 . We know that
S (T2 ) = R log 2, so we do a bun h of algebra to get
nR log 2
C0 =
1 ( TT21 )n (1 n log TT12 )
2 Stephan-Boltzmann Law
(30 pts.)
2.1 Light Bulb
(6 pts.) The bulb of an ordinary 100 Watt in andes ant lamp, 3 m radius, is bla kened to absorb all
radiation emitted by its lament. If the bulb is surrounded by va uum, what is its temperature? The
Stephan-Boltzmann onstant is 5:67 10 8 in MKS units.
Solution: You need to know the Stephan-Boltzmann Law, J = B T . In this equation, J is the energy
4
ux density, so we need to multiply it by an area to get the units of power that we are given. The area of
interest is the surfa e area of the bulb, so
P = 4r2 J = 4r2 B T 4
We an invert this equation, and plug in the numbers to get
100W
1=4
T = = 628K
4(0:03m2 )5:67 10 8
2.2 Surfa e Temperature
(8 pts.) Assume that the Sun radiates like a bla k body at a surfa e temperature of Ts = 6000 K. The planets
an be modeled as smaller spheres, at varying distan es, in thermal equilibrium with the Sun's radiation.
For example, the Earth-Sun distan e is res = 1:50 108 km, while the Venus-Sun distan e is rvs = 1:08 108
km.
Derive an expression for the temperature of a planet surfa e Tp . Su h an expression must involve in some
way the planetary distan e from the sun.
Referen es: This question is Problem 4.5 in Thermal Physi s, Kittel and Kroemer.
Solution: You need to know that a bla k-body in equilibrium re-radiates all the energy it absorbs. First,
let's gure out how mu h power the Sun is radiating, doing the same al ulation as the previous problem
P = 4Rs2 B Ts4 . Now, the fra tion of that power that a tually hits a planet has a 1=4rps
2
be ause of the
separation, and has a Rp be ause of the ross-se tional area of the planet. So the energy absorbed by the
2
planet is
Rp2 Rs2
P = 2 B Ts4
rps
In equilibrium, the bla k body (planet) will re-radiate all of this energy, so we an equate the in oming
power with the outgoing power
Rp2 Rs2
2 B Ts4 = 4Rp2 B Tp4
rps
whi h we an solve to get
Rs2 4
Tp4 = Ts
4rps
2
Note that this answer involves the separation distan e of the planet and the Sun, but not the size of the
planet itself.
2
2.3 Radius of the Sun
(8 pts.) From the fa t that the Sun appears from the Earth to have an angular diameter of = 0:57 degrees,
nd the Earth's temperature, Te . This numeri al answer for the Earth's temperature happens to be quite
good.
Solution: You need to know the ir umferen e of a ir le. The angular size of the Sun is given so that
we an obtain the radius of the Sun. Not a staggeringly diÆ ult question, but here goes. Draw a sli e of a
ir le of radius res and opening angle . Then the ar length of the sli e is 2RS , whi h we an obtain with
some simple geometry, Rs = 360
res = 0:75 106 km. The rest is plugging numbers into the result of the
previous question, to get Te = 300 K.
2.4 Algebra
(8 pts.) Show a simple relationship between the temperature of another planet, Tp , and the Earth's temper-
ature, Te .
Solution: If you got the expression for the temperature of a planet, this is just algebra.
2
res
Tp4 = Te4
rps
2
3 Venus
(40 pts.) Among the planets, Venus presents an interesting hallenge for the simple model suggested in the
previous problem. Some data for Venus and other useful onstants are: Mv = 0:82Me and Rv = 0:95Re.
Also mH gearth =kB = 1:19 10 3 K/meter where mH is the mass of the Hydrogen atom, kB is the Boltzmann
onstant, and gearth is the gravitational a eleration on Earth.
If the formula derived in the previous problem is used to estimate the temperature for Venus, the result
is 352 K. This al ulated value is too low for the Venusian surfa e temperature T (0), but may be taken
as a \ loud temperature" T (h), for a loud of noxious gases at height h = 50 km from the surfa e. The
atmosphere on Venus is 96% CO2 , and there is a pronoun ed \greenhouse ee t." The pressure in the
\ loud", P (h), is about the same as Earth's atmospheri pressure (one Earth atmosphere).
In fa t, both temperature and pressure have now been measured at the planetary surfa e. Given that
T (0) has been measured to be 500Æ C, you are asked to predi t the surfa e pressure P (0). Please explain
what you're doing. The sequen e of questions below, whi h lead to a solution, are supposed to help.
Note: The model suggested here is a stationary CO2 gas on Venus, nowhere near isothermal, with a
measured pressure at altitude h, a measured temperature on the surfa e, and a al ulated (given above)
temperature at altitude h. The task is to al ulate the surfa e pressure. As it happens, P (0) has also been
measured, and the model is not hopeless. Think rst about the ne essary onditions for hydrostati , but not
isothermal, equilibrium.
3.1 Hydrostati Equilibrium
(10 pts.) Explain arefully the onditions for hydrostati equilibrium in a gravitational eld.
Solution: You have to know what hydrostati equilibrium is. In hydrostati equilibrium, the pressure of a
system ountera ts the for e of gravity. Hen e, the system as a whole does not move, hen e the \stati " part
of the name. We an write this ondition as an equation whi h will be useful. Look at a sli e of the system,
of density , ross-se tional area A and height dz . The gravitational for e on this sli e is mg = Agdz .
Pressure is for e per area, so the pressure dierential dP times A must oppose the gravitational for e. We
an move the dierential around to get
dP
= g
dz
The other quantity that we will need is dT =dz . We an get this by remembering a Maxwell identity,
T =P = V =Cp .
dT T dP V
= = ( g)
dz P dz Cp
3
Note that in this system, all these quatities are onstants. So T (z ) is linear in z , and we an write it as
T (z ) = Bz + T (0)where B = gV =Cp .
3.2 Pressure and temperature
(20 pts.) Find the relationship between pressure, density, and temperature as a fun tion of height, and use
it to nd an expression for the pressure ratio P (h)=P (0) as a fun tion of the temperatures T (0) and T (h).
Try to use algebrai expressions { wait to plug in the numbers until the next part.
Solution: You need to have obtained the relations in the previous question, and do some algebra and
integration. First, we make the assumption that the atmosphere is an ideal gas. Remember that the density
is the mass times the number density. Let m be the mass of a CO2 mole ule in the atmosphere, and then
P = m
kB T . Solve this for , and use it in the dP=dz equation to get
1 dP g
=
P dz kB T (z )
This is a logarithmi derivative, and we an integrate to get

P (h)
Z h
mg
log = dz
P (0) 0 kB T (z )
Using the equation for T (z ) we obtained in the previous question, we an do this integral, to get
P (h) mg T (h)
log = log
P (0) kB B T (0)
or mg=kB B
P (h) T (h)
=
P (0) T (0)
3.3 Arithmeti
(10 pts.) With your result of the previous question, together with the given information, al ulate a numeri al
value, in Earth atmospheres, for the Venusian surfa e pressure P (0).
Solution: You have to have gotten the previous part, and know that CO2 is 44 times more massive than
hydrogen. We are given many onstants, whi h we an play with until we get what we want. The temperature
gradient B an be obtained by mat hing the boundary onditions, B = (T (h) T (0))=h. The gravitational
a eleration of Venus an be obtained by the mass and radius relations to Earth gravity, gv = (00::95)82
2 ge . So
we an write the power in the pressure-temperature equation as
mCO2 gv mCO2 0:82 mH ge h
=
kB B mH (0:95)2 kB T (h) T (0)
All these quantities are given, ex ept for the mass ratio, whi h you just have to know. So, with some areful
al ulating, remembering to use meters, we an obtain P (0) = 85 Earth atmospheres.

QualifyingExamProblemPhysicsUnivOfWashington 1992 2000

Uploaded by

Copyright:

Available Formats

QualifyingExamProblemPhysicsUnivOfWashington 1992 2000

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

QualifyingExamProblemPhysicsUnivOfWashington 1992 2000

Uploaded by

Copyright:

Available Formats

Basi

(5 pts.) Show all the for es a ting on the material point.

Lz = m2 _ = mz 2 tan2 _

Now fa tor the quadrati term and an el to get

1.1 Moment of Inertia

1.2 Os illation Frequen y

1.4 Experimental Issues

2.1 Snell's Law

2.2 Spe ial Case

(5 pts.) What is spe ial about n = 2?

3.1 Ro ket Equation

3.2 Relativisti Ro ket Equation

3.3 Classi al Limit

1 Lagrangian and Hamiltonian Dynami s

We an solve this di erential equation to get

Suddenly, we remember that  is not really a variable, so  = 0 and _ =

1.8 Conserved Quantities

1.2 Equation of Motion

1.3 Normal Modes

1.5 Di erent Masses

1.5.2 Big Mass Di eren e

1.6 Weak Coupling

Now use the relation  = R sin   R and the approximation os   1 2 =2 to get

1.1 Label For es

(5 pts.) Write the Lagrangian.

1.3 Equation of Motion

(10 pts.) Derive the equation of motion from the Lagrangian.

mR2  mR2 ! 2 sin  os  + mgR sin  = 0

Divide by ommon fa tors to get the ni er form

1.5 Small Os illations

2 Moment of Inertia Tensor

2.1 About an Axis

(10 pts.) Show that

2.2 Angular Momentum

2.3 Prin iple Axes

3.1 Ellipses and Energy

3.2 Orbit Speed

3.3 Speed Change

1.1 Stati Limit

The expansion is Legendre polynomials is a useful thing to remember,

Now remember that Pl (1) = 1, and that

r = r^ r where k = r^!= . In the radiation limit B

1 Fields, Energy, Momentum, and Torque

(10 pts.) Find the ele tri eld E~ everywhere.

^ 2 =jX~ j, with the origin of the oordinate

In this ase ~r0  X~ . We an write the harge distribution in spheri al oordinates as

2.1 Energy Levels

2.3 Bohr Radius

2.5 Comparing Lifetimes

3.1 Protons and Neutrons

(10 pts.) What are the relative masses of the onstituents?

1 Big Bang Spe trum

1.1 Radiation Spe trum

Consider the energy in a small pie e d of the spe trum

0 (0 ; T )d0 = (; T )d

says  / T 4 . However, this gives T = T=S 5=4 .

1.2 Ba kground Isotropy

2 Spe ial Relativity

Lz = m2 _ = mz 2 tan2 _

We an solve this dierential equation to get

Suddenly, we remember that is not really a variable, so = 0 and _ =

1.5 Dierent Masses

1.5.2 Big Mass Dieren e

Now use the relation = R sin R and the approximation os 1 2 =2 to get

mR2 mR2 ! 2 sin os + mgR sin = 0

r = r^ r where k = r^!= . In the radiation limit B

In this ase ~r0 X~ . We an write the harge distribution in spheri al oordinates as

Consider the energy in a small pie e d of the spe trum

0 (0 ; T )d0 = (; T )d

says / T 4 . However, this gives T = T=S 5=4 .

(5 pts.) Dene the dierential s attering ross se tion d=d

tion (hx~0 j (+) i). So some 's an el, and we get

(10 pts.) Evaluate the dierential ross-se tion, d=d