Grechko Sugakov Tomasevich Fedorchenko Problems in Theoretical Physics Mir 1977

MIR PUBLISHERS M O S C OW
.\ llll{'T THE BOO h.
This book is a collection

of problems covering mechanics,
electrodynamics, nonrelativistic
lJ uantum mechanics,
!ltatistical physics
nnd thermodynamics.
Each ~ection opens
with a brief outline
of the main laws
und relationships used
lo solve the problems.
Also information
about the needed
mathematical apparatus is included.
A long with answers there
n 1·o guides to solving
tho more complicated problems.
S l units are used
throughout the book.
JJrofJlems in Theoretical Physics
is intended
for physics majors
at universities
11111l other institutions
of higher learning.
Somo of the problems
n1·o :-~pecifically
for :-~tudonts majoring
i 11 tlmoretical physics.
Ccort.u in onos can Lo used
i 11 tho physics
und llllltlwmntics dopn1·tmonts
111' t.oudwrs collPgos.
Jl. f. fpe'IKO
B. 11. CyraKon
0. <D. ToMaceBM'I
A. M. <Dep;op'leHKO
CBOPI-1111-\
3AAAll
ITO TEOPETI1qECROl1
<DI1311RE
M3JWT8JI:bCTBO «BMcmaH miWJJa»

Problems
•
Ill
Theoretica]
Physics
by
L. G. GRECHKO
V. I. SUGAKOV
0. F. TOMASEVICH
A.M.FEDORCHENKO
Translated
from
the Russian
by
Eugene Yankovsky
MIR PUBLISHERS
MOSCOW
First published 1977
TO THE READER
Mir Publishers would be grateful for your comments

on the content, translation and design of this book.
We would also be pleased to receive any other
suggestions you may wish to make.
Our address is:
USSR, 129820, Moscow 1-110, GSP
Pervy Rizhsky pereulok 2
Mir Publishers
@ English translation, Mir Publishers, 1977

Preface
This text incorporates problems which have been used

for several years at seminars in courses in classical mechanics,
electrodynamics, quantum mechanics, and statistical phy-
sics and thermodynamics at the T. G. Shevchenko State
University in Kiev.
The text draws largely on the Course of Theoretical Physics
by L. D. Landau and E. M. Lifshitz, but also makes use
of other textbooks and handbooks recommended for the
university course in theoretical physics. Some of the pro-
blems have been taken from published problem books listed
at the end of this book, but many are original.
The student will be able to solve the problems if he has
a good knowledge of the fundamentals of theoretical physics,
which are briefly outlined in each section of this book. All
the problems use the International System of Units (SI).
The section on classical mechanics was com piled by
A. M. Fedorchenko, on electrodynamics by V. I. Sugakov,
on quantum mechanics by 0. F. Tomasevich, and on sta-
tistical physics and thermodynamics by L. G. Grechko.
The Authors
Contents
PREFACE 5
Section I. Classical Mechanics 9
Problems 25
Answers 141
Section II. Electrodynamics 50
Problems 61
Answers 160
Section III. Quan lum Mechanics 78
Problems 92
Answers 230
Section IV. Statistical Physics and Thermo-
dynamics 110
Problems 119
Answers 356
APPENDICES 424
1. Basic formulas of vector analysis 424
2. Curvilinear coordinates 425
3. Differential operators in curvilinear
coordtnates 429
4. Mathematical supplement 434
5. Legendre polynomials 441
6. Hermite polynomials 444
7. The confluent hypergeometric junction 446
BOOKS ON THE SUB1ECT 448

SECTION I Classical Mechanics
The mechanics of systems with a finite number of degrees

of freedom. In mechanics a particle is a material body of
mass m whose position in space is determined by three coor-
dinates.
The mechanical state of a system of n particles is charac-
torized by 3n coordinates and the 3n time derivatives of
these coordinates. The law involving changes in the state
of a mechanical system in time is defined by Newton's
nquations
(I-1)
where Fi is the resultant of all the forces acting on the ith
particle; these include both internal forces (those originat-
Ing in the particles of the system) and external forces (those
having a source outside the system and such as considered
given at any instant of time).
From the standpoint of mathematics equation (I-1) is
n system of 3n differential equations. For this reason the
hasic problem of mechanics consists in finding a solution
for this system. We know from the theory of differential
..
oquations that to find an unambiguous solution of the
~ystem we must indicate 6n values r?, r? at a definite instant
of time. In short, the mechanical state of a system at any
.
~nbsequent time is determined by its initial ~mechanical
~., .._- ~·...,._.~--
~tnte r1, rj and by the forces acting on each particle in the

~ystem.
Equations (I-1) are valid only in inertial frames of refe-
rnnce. An inertial frame of reference is one in which a par-
ticle free from forces, i.e. an isolated particle, is in uniform
roctilinear motion. The first law of mechanics states that
11uch frames do exist.
10 PROBLEMS IN THEORETICAL PHYSICS
Forces acting between two particles are represented by

the formula
r·J
Fu=F(ru)-'-
rii
which reflects the following properties (Fig. 1):
(1) Fu = -Fli;
(2) Fu II ru;
(3) the magnitude of the force depends only on the dist-
ance between the two particles.
Classical mechanics rests on the three laws. of Newton,
which were deduced from experiments and observations
of mechanical motion. All
c other assertions and laws
of mechanics, valid for spe-
cific conditions and specific
models, are corollaries of
these three laws.
In a noninertial frame
of reference (one moving
with acceleration) equations
(1-1) do not hold. But we
can preserve the form of
!I equations (1-1) by introdu-
cing what are called forces
of inertia, whose origin
X cannot be explained by the
Flg. 1 action of any specific parti-
cles. The forces are due to
the fact that] the frame of reference moves with 'accelera-
tion. The equation of motion for a particle in a noninertial
frame of reference is
.. + F1ner
mr = F
where .F 1ner =-m dto + [~ X r].. -!- [ffi X [ffi X rJJ -!-
+ 2 [ffi X r]) is the force of inertia. R 0 is the acceleration
of the coordinate origin and ffi is the angular velocity of
this frame [see formulas (I-23) and (1-24)1.
If we proceed from the second law of Newton (1-1) and
the first property of the forces of interaction (see above).
SECTION I. CLASSICAL MECHANICS 11
we can prove that the time derivative of the momentum

vector of a system of particles equals the sum of all the
external forces, Fext:
dp
dt=Fext (1-2)
n •
where p = ~ m;r;; n is the number of particles in the
i=l
system.
If the system is closed, i.e. Fext equals zero, equation
(1-2) gives us the law of conservation of momentum:
p =constant
If we introduce the notion of the centre of mass of a
system
n 1
2J m;ri
R= ...:...i=_l::...,;M-.,..---
n
where M = 2] m 1 , equation (I-2) takes the form
i=l
MR= Fext (I-3)
Jf the system is closed, it follows from equation (I-3) that
.
R=constant
Thus, the velocity of the centre of mass of a closed system
remains constant.
From equation (I-2) we can deduce the law of motion
of a body having variable mass, i.e. the law of jet propul-
sion. In the simplest case, if the main body (of mass m) is
losing or gaining mass, the law of jet propulsion (Meshcher-
skii's formula) takes the following form:
dv
m-= Fext dmt
+--u~---u 2
dm 2
(1-4)
dt dt dt
where m1 is the mass gained, u 1 is its velocity relative to
the main body, and m 2 and u 2 are the respective values
for the lost mass.
Proceeding from the second law of Newton (1-1) and
t.he first two properties of the forces of interaction, we can
i2 PROBLEMS IN THEORETICAL PHYSICS
prove that the time derivative of the angular momentum of

a system of particles equals the sum of the moments of all
the external forces, N:
n
dL •
Tt=L= ~ [r1 X F1]=N (1-5)
i=1
n •
where L = ~ m 1 [r1 X rt].
i=1
We must bear in mind that the radius vectors r 1 of the
particles in the system, which vectors enter into 'the defini-
tions of the angular momentum and the moment of an
external force, must issue from the same point because
both depend on the choice of the coordinate origin.
Newton's third law makes it possible to introduce the
concept of the potential of a force according to the formula
F 1i = -grad 1 U (r1 i) (1-6)
where the potential U (ru) depends only on the distance
between the interacting particles.
We can use the potential concept to prove the following
theorem on the basis of Newton's laws of motion: a change
in the mechanical energy of a system equals the work done
by external forces, i.e.
d (K +; ~ U iJ) = dA (1-7)
ii
n n
where by definition K =-} ~ m ;f and 1 dA =~ (Fext) 1 dr 1•
i=1 i=1
The law of conservation of energy holds for closed systems:
E=K +-} S UiJ=constant (1-8)
ij
If a part of the external forces has a potential V, we can

write formula (1-7) as
n
:e (K+; ~Uil+V)=-~~+~(;t.ft)
u i-1
where f1 is a nonpotential force.

Thus, a closed mechanical system always has seven

integrals of motion (seven functions of coordinates and
velocities), which remain constant upon motion. In the
general case the number of integrals of motion, which do
not depend on time, is 2k - 1 for a closed system, where k
is the number of degrees of freedom. The seven aforemen-
tioned integrals of motion play a special role in physics.
There are two main reasons for this. First, these integrals
of motion always exist regardless of the number of particles
in the system (for a single particle not all are independent).
Second, their existence can also be proved by the funda-
mental properties of space-time. For instance, the law of
conservation of momentum follows from the homogeneity
of space (all points in space have the same status); the law
of conservation of angular momentum follows from the
isotropy of space (all directions in space have the same
status); the law of conservation of energy follows from the
homogeneity of time (all instants of time are equivalent).
The laws of motion have other forms than Newton's.
Using the Lagrangian function (or, simply, the Lagrangian)
and the generalized coordinates, we can write equations
(1-1) in the following form:
~
dt (a.z)-
•
f}qj
fJ.Y!_=
f)q.
I
o~t (1-8')
where Z is the Lagrangian U.efined as Z = K - V (K is

the kinetic energy and V the potential energy of the system);
qi are the generalized coordinates, i.e. any coordinates that
satisfy the sole requirement that the Cartesian coordinates
(used in the system of Newtonian equations) are at any
instant of time uniquely expressed in terms of all the q's:
Qi = ~ ( fs • :~: )
s=l
where fs is a nonpotential force; the subscript k is the num-

ber of degrees of freedom.
If there are nonpotential forces in the system but the

generalized force corresponding to them can be written as
Q 1~ = _ au +..!.... ( au• )
aq.l dt aq;
where U is a function of the coordinates and velocities,

the Lagrange equations of the second kind take the form
..!.... (a~)-
dt aq;
az
aq;
=0 (I-9)
where :£ = K- V + U. For example, the Lorentz force

f = eE + e [; X B)
defined by the equations
E= -grad m - aA
"' at
B=curl A
is a non potential force. It can be written as
· •
fx=eEx+e(yBz-zBy}=- au
ox +ae (au)
d a~
where
U=eq>+e(r·A)
The Lagrange equations can be obtained from the varia-
tional principle, which states that if we introduce the func-
tional S, called action, according to the formula
12
s= J:t; (t, qi, q;) dt, (I-10)
l1
the actual motion will be described by such functions

q1 (t) as ensure a minimum of the functional S provided
that qi (t1 ) and q1 (t 2) are given.
The Lagrange equations are a system of k second-order
differential equations. We know from mathematics that
a system of k second-order differential equations can be
reduced to a system of 2k first-order differential equations.
In mechanics this is done by introducing the Hamiltonian
SECTION I. CLASSICAL MECHANICS f5
function (or, simply, the Hamiltonian), which is a function

of the generalized coordinates and momenta. The general-
ized momenta are defined by the formula
az
P;="'-.- (1-11)
oq;
Since the Lagrangian is a quadratic function of the

generalized velocities, formulas (I-11) give a (linear and
single-valued) relationship between the generalized velo-
cities and the generalized momenta.
The Hamiltonian is related to the Lagrangian in the
following way:
(1-12)
All the generalized velocities in the right-hand side of

(1-12) must be expressed in terms of the generalized momenta
according to (1-11).
The canonical equations of Hamilton are
f)Q!e
P
z ·=- -
oq;
(1-13)
a!IC
q; = op; (1-14)
Equations (1-13) and (1-14) are a system of 2k first-order

differential equations.
In some cases the interaction of bodies is of a peculiar
nature, the nature of a constraint. Constraints impose
certain restrictions on changes in position or velocity.
There is a fairly large class of so-called holonomic constraints,
i.e. restrictions on position that can be expressed by algeb-
raic equations:
fa. (x1 , • • . , Xn, t) = 0, a = 1, 2, ... , s (1-15)
These are the equations of constraints.

To solve problems involving constraints we can use the
I ,ngrange equations of the second kind, if we introduce
Hlll'.h generalized coordinates as satisfy the equations of
l'oJisl.raints automatically, or we can use the Lagrange
equations of the first kind in the following form:

s
m 1 ~~ = F 1 + z; A.a. gradi Ia. (1-16)
Q;=1
which must be solved together with (I-15).

If we define the product -m 1 ~~ as the force of inertia,
we can formulate the d'Alembert principle: a system moves
in such a way that on any virtual displacement the work
of all the forces, including forces of inertia, at any instant
of time equals zero, i.e.
n
~ (Fi -m;;1) llr 1 = 0 (1-17)
i=1
In the absence of constraints this principle gives us the

Newtonian equations (1-1). In the case of ideal constraints
we get the Lagrange equations of the first kind.
If a system of points (particles) rests while the constraints
act on it, the principle (1-17) takes the following form:
n
~ (Fi ·6r1) = 0
i=1
This equation expresses the principle of virtual displace-

ments, which is the basis of statics. If we add to it the
equations of constraints, we can find the condition for the
equilibrium of a system of particles.
Solution of equations (1-1) gives us all the information
about the mechanical state of a system consisting of any
number of particles having an arbitrary law of interaction.
However, even the three-body problem (for instance, the
problem of three particles interacting via the Coulomb
force) poses great mathematical difficulties. For this reason
a variety of approximate methods or models that to one
degree or another reflect the properties of actual system
are used to solve such problems. One is the model of a rigid
body. In mechanics a rigid body is a system of particles
whose distances from each other remain constant in time.
Such a body acts as a single whole while it is in motion.
A rigid body has six degrees of freedom, which can be
chosen in the following way. Let us specify an arbitrary
poiul U of a 1·igid body in an inertial frame of reference

.\'TZ (Fig. 2). This point will be called the pole. The pole
1uay coincide with ~the
eentre of mass, which is z
defmed by the form·ula
~ ~ ~ rp dV
He= M (I-18)
The coordinates of the
pole, X 0 , Y 0 , Z 0 , are the
lhree coordinates that
describe the translation- X x'
al motion of a rigid
body. If we fix the pole
0, the body can revolve
around it. This will chan-
ge the orientation of a X
coordinate system x'y'z' Fig. 2
attached to the body. The
orientation of one coordi-
nate system in relation to another can be described by the
rotation matrix au, which links the sets of basis vectors of
lhese systems of (Cartesian) coordinates:
i = a 11 i'+ a 12j' + a 13k'
j =-·, a i' + a :J' + a k'
21 2 23 (1-19)
k = a 31 i' + a 32 j' + a 33 k'
The rotation matrix has the property
3
~ a,kais = {,ks (I-20)
i=i
The rotation matrix links the components of any vector F
in different coordinate systems:
F x =auF~+ a12F~ + a 13 F;
F 11 = a 21 F~ + a22F~ + a23 F; (I-21)
F z = aat F~ + aa2F~ + aaaF;
Sinco the nine matrix elements au are restricted by six
formulas of type (1-20), we can express the rotation matrix
2 lll49fi
18 PROBLEMS TN THEORETICAL PHYSICS
in terms of three independent parameters. The Euler augles

e, ~'· cp (Fig. ;)) are used for these three parameters. The
possible values of these variables arc
0 < 8 < :rt, 0 ~ ~' ~ 2:rt, 0 ~ cp ~ 2:rt
The rotation matrix is expressed in terms of the angles of
rotation as follows:
(cos 'iJ cos cp - -(sin 1p cos Bcos cp+ sin 'iJ sin 8
-sin 'iJ cos 8 sin cp) + cos 'iJ sin cp)
- (sin 'iJ cos cp + (cos 'iJ cos 8 cos cp-- -cos 'iJ sin 8
+ cos 'iJ cos 8 sin cp) -sin 'iJ sin cp)
sin 8 sin cp cos cp sin 8 cos 0
(1-22)
Thus, the three independent coordinates X 0 , Y 0 , Z 0
of the pole, which characterize the translational motion
of a rigidj body, and the three
z:' Euler angles e, cp, 'Jl, which cha-
racterize the rotational motion
about pole 0, form six variab-
les that fully and uniquely de-
~t-f-rfrf-r-~!J termine tho position of a rigid
body in space. Correspondingly,
the first time derivatives of the
variables describe the velocity
of the rigid body. But to
N characterize the angular velo-
Fig. 3
city in the mechanics of rigid
. ..
bodies wo use not the derivati-
ves 8, cp, 'iJ but the angular velo-
city vector ro, which is introduced by the Poinsot formulas:
di' •I dj' •I dk' 1
cu=(ffiXI], dt=(ffiXJ], dt=[roxk] (1-23)
where i', j', k' are the basis vectors of the coordinate
system that is attached to the body. We can obtain the
components of ro in terms of the Euler angles and their time
•IPI'i v a ti ves:
.
0 cos lp -1- 'I' ~iII () sin lp
.
(!) ;x·' =.c.
.
(J) y' = -.e sin ffJ + "' sin e cos ffJ
(l) z• = ffJ + "' cos e
These are the kinematic equations of Euler.
Using the definition (1-2) of momentum for a system of
pnrticles, we obtain the formula for the momentum of
n rigid body
(1-25)
where M is the body's mass and Rc can be found by (1-18).
If we proceed from the definition (1-5) of angular moment-
•un for a system of particles, we obtain the formula for the
nngular momentum of a rigid body about the origin of an
inertial frame of reference XYZ:
I. = M [R 0 X R +M 0] [Rc X R 0]
+M [Ro X [ro X Rcll + Lrot (I -26)

where Rc is the centre-of-mass vector in relation to the
pole (Rc = 0 if the centre of mass is selected for the pole).
The last member in (1-26) is the angular momentum about
!he pole 0.
If we introduce the inertia tensor, which is defined by
lhe integrals
I xx = ) JJ(y + z
2 2) p dV, I xy = I yx = -- JJJxyp dV
I uu '---= JJJ(x + z
2 2) p dV, I yz = I zu = JJyzp dV
- ) (1-27)
111 "J JJ(x2 +y2)pdV~ I zx = I xz = - JJJzxp dV

lho components of Lrot will be expressed in terms of the
•~om ponents
of ro via the inertia tensor:
3
(Lrot) i = 2J I ijffi j (1-28)
i=t
Proceeding from the definition (1-7) of kinetic energy
l'ut· 11 system of particles, we obtain the formula for the
2*
20 PROBLEMS TN THEORETICAL PHYSICS
kinetir, euergy of a rigid body:

1 . ." • 1
K :f JIR; +MHo [ul X Rei+ 2 (w ·l~rot) (l-2V)
We can always selecl a system of coordinates (attached

to the body and with an origin chosen arbitrarily) in which
the symmetric inertia tensor will have none but diagonal
components. Such a system is called the set of principal
axes. In this system formulas (I-28) become much simpler:
L 1 =, 11 (u 1 , L 2 = I 2 w 2 , L3 = I 3 w 3 (I -30)
The formula for the kinetic energy of rotation is also sim-
plified:
Trot=~ (w·Lrot) =-}(Itw: +Izw: +Iawi) (I-31)
We can obtain the motion equations for a rigid body by

using the corollaries (I -2) and (I -5) of the Newtonian law.
When the centre of mass is taken for the pole, these equa-
tions become
d2X • • ) )
M ( dt2o +w2Zo-wsYo =Ft
d2Y • • )
;1/ ( dtzo+ulaXo-WtZo =F2
d2Z 0
( ~+w • • )
M 1 Yo-WzXo =Fa
} (I-32)
It ~:+(Is- I2) WzWa = Nt
I 2 d:2 +(I 1- I a) WtWa = N z
I 3 d:a +(1 2 - It) WtWz = Na
where the subscripts 1, 2, 3 designate the components of

vectors along the principal axes, and X 0 , Y 0 , Z 0 are the
coordinates of the centre of mass in the same coordinate
system. The moments of forces that enter into the total
moment of forces N are taken relative to the centre of mass
(pole).
The system of six equations (I -32) is a complete system
of differential equations for finding the six functions X (t),
SECTION I. CLASSICAL MECHANICS 2i
Y (t), Z (t), cp (t), 8 (t), \jJ (t). As we see from this system,
the character of motion of a rigid body depends not only
on the mass but also on its distribution, since the moments
of inertia 11 , 1 2 , 1 3 depend on the body's shape and the
distribution of its density.
Another way to formulate the motion equations of a rigid
body is to construct the Lagrangian. The kinetic energy is
expressed by formula (1-29). Given the forces, we can find
the corresponding potentials and construct the Lagrangian.
Then·we formulate the Lagrange equations of the second kind.
The mechanics of continuous media. When we examined
the motion of a rigid body, we assumed that the distances
between the particles constituting the body remain unchang-
ed. But we know from experience that solids have a fairly
wide range of mechanical states (deformation, the pro-
pagation of sound) that cannot be described by the model
of a rigid body. Thus we must consider the internal motion
of the particles of a solid in relation to each other.
A similar situation appears when we examine the motion
of liquids and gases (fluids). The main difference between
the motion of fluids and the motion of elastic solids is
that the particles constituting the liquid or gas can move
a considerable distance from their initial position. In
other words, in contrast to elastic solids liquids and gases
have the property of fluidity. Liquids, gases and clastic
solids are often designated as continuous media.
In studying the motion of liquids, gases and solids we
can do one of two things. We can describe them as a system
consisting of a large number of particles by assigning coor-
dinates to each particle. Or we can consider them as eon-
tinuous media, i.e. media with a continuous rather than
n discrete distribution of particles. This is the approach
in introducing the concept of density.
One of the main concepts in the mechanics of continuous
media is the vector field of velocities. This field is the vector
runction of coordinates and time, v (x, y, z, t), that indi-
rntes the velocity of those particles of the medium that
:•t. time t pass through the point in space with coordinates
:r., y, z. Thns the velocity v (x, y, z, t) does not refer to
nny particle in particular but clwractprizes thP motion
of the medium as a whole.
The equations of motion for a continuous medium are

obtained by applying the law of change of momentum in
time (1-2) to an arbitrary volume of the medium. Forces
of two kinds act on this volume:
(1) the surface forces exerted by the other parts of the
medium and defined by the formula
CFsurtace); =) Piini dS = ~ a::: dV (1-33)
where Pii is the stress tensor and ni are the components of
the outward normal to the surface of the chosen vDlume;
(2) the body forces (e.g., the force of gravity), defined
by the space integral
(Fborty); = J
ftP dV (1-34)
where p is the density of the medium and / 1 is the body
force per unit mass.
The motion equation for a continuous medium has the
form
(1-35)
For the law of change of angular momentum in time to
be valid, it is necessary and sufficient for the stress tensor
to be symmetric:
Pu =Pit
From the law of conservation of mass we can obtain the
relation between the field of velocities and density:
:~ +div pv =' 0 (1-36)
This is the equation of continuity.
If the continuous medium is incompressible, the equation
of continuity gives us the condition for incompressibility
div v = 0 (1-37)
which is oflen used in solving problems involving the flow
of liquid and even gas.
For the system of equations to be complete, we adjoin
the seconrl law of thermodynamics for continnons media
r-de . t- -p-·-
= - d"lVJ- av; (1-38)
dt !] OXj
where e is the internal energy per unit volume and j

-'"' -x grad T is the heat flux vector.
To solve equation (1-35) we must know the stress tensor
P; 1. For viscous fluids we can write it as
Pii = - pt3ii + (11- ~) t)ii div v + 1J ( ax; + ovi )

iJv·
o:r; (1-39)
Substituting this expression into (1-35), we get the Navier-

Stokes equation
r r:; +<v·V)v]=pf-gradp+lldv+sgraddivv (1-40)
where 11 is the coefficient of shear viscosity and ~ the coef-

ficient of bulk viscosity. For incompressible liquids the
system of hydrodynamic equations takes the form
r[ ~: +(v·V) v] = pf -grad P+ lldV}

div v = 0
These are partial differential equations. For this reason
to solve system (1-41) we must fix the initial and boundary
conditions. For boundary conditions we use the "sticking"
condition (vb = 0) on stationary walls.
Equation (1-38) is not connected with system (1-41) for
incompressible liquids. It is used to find the temperature
as a function of coordinates anfl time (see Problems 109
and 111).
The equation of motion (1-35) for a continuous medium
also gives us the motion equation· for an elastically deform-
ed solid. What characterizes the elastically deformed solid
is that even when a fairly strong external force acts upon
it, the particles constituting the solid do not move far from
the equilibrium configuration, and when the force is remov-
nd, the pnrtie]ps return to their initial position. Thus
av {)2u
--at+(v·V) v ~ 012 (I-42)
where u (x, y, z, l) is the displacement vector characteriz-

ing the displacement of the parts of an elastically deformed
:-~olid from the equilibrium position.
Small elastic deformations are characteri1-ed by the

symmetric. strain tensor
U·j=-
1 ( au·
--'
au i )
+- (I-43)
z 2 OXj 8x;
The stress tensor and the strain tensor are related by

Hooke's law
3
Pii = 2] A!J~tlU~tl (I-44)
k, 1=1
where AtJkl is the fourth-order tensor of elastic constants,

which satisfies the symmetry relations
"-zilll = AJtkt
AiJ~tl = AtJlk (1-45)
"-iilll = A~tut
Because of the symmetry relations (I -45) this tensor in
the general case has only 21 independent variables. Thus
we can rewrite Hooke's law as follows:
(1-46)
where the lower index i assumes six values:

one-index numbering 1 2 3 4 5 6
two-index notation xx yy zz yz xz xy
In this manner "-xxxx = Cn, "-xxyy = cl21 Axyxy = c66, etc.
As a consequence of spatial symmetry the number of
independent elastic constants can be further reduced. There
are three independent elastic constants for a crystal with
cubic symmetry, for instance, and matrix cu has the
form
( cu Ctz Ctz 0 0 0 ~);
I
Ctz Cu Ctz 0 0 o r t
Ctz Ctz Ctt 0 0 0 l

0 0 0 (_'44 0 0 .j j
0 0 0 0 0 I :
.: I c,.
l 0 0 0 0 0 c•• ~)~
Isotropic solids are characterized by two elastic constants,

and we can write Hooke's law in the following way:
Pii = Mu div u + 2f.Luu (1-47)
The general equation of motion and equilibrium for an
elastically deformed solid has the form
iJ2u. iJpii
P at 2' =~+Pf; (1-48)
J
where PiJ is defined by (1-44).

Sound-energy flux per unit area (sound intensity) can
be determined by the formula
si = -puui
In isotropic media the time average of the magnitude
of vector S for progressive monochromatic waves is
S- = zP
1 2 2
VsW uo (1-49)
where v8 is the velocity of longitudinal or of transverse

sound waves, p the density of the medium, ul the frequency
and u 0 the sound-wave amplitude.
Formula (1-48) incorporating (I-47) assumes for isotropic
materials the form
iJ2u
p 7it2 = pf +(A.+ f.L) grad div u + f.LL\u (I-50)
In the theory of elasticity the boundary conditions are

(u)I = (u)n, (punJh = (punJ)u (I-51)
where n 1 are the components of the outward normal to the
interface of media I and II. The first condition points to
the continuity of the media and the second to the properties
of the forces.
PROBLEMS
1. A particle of mass m and chargee euters a homogeneous
and stationary electric field E with a velocity v 0 perpendi-
cmlar to the direction of the fielrl. Calculate the particle's
path.
2. A particle of mass m and charge e enters a homogeneous

retarding electric f1eld E with a velocity v0 parallel to the
direction of the field. How much time will it take the par-
ticle to return to its initial position?
3. A particle of mass m and chargee enters a homogeneous
and stationary magnetic field H with a velocity v0 per-
pendicular to the direction of the fwld. Calculate the par-
ticle's path.
4. Consider a homogeneous electric field that changes
according to the law E = E 0 cos wt. A particle. of mass m
)__,___X -----1-- _j
z
Fig. 4
and charge e enters the field with a velocity v 0 perpendi-

cular to the direction of the field. Calculate the particle's
path.
5. Consider homogeneous and stationary electric and
magnetie fields E and H in a certain area of space. The angle
between the two is equal to a. A particle of mass m and
charge e enters this area with a velocity v 0 • Calculate the
particle's path.
6. A cluster of charged particles flies out of point 0,
the particles having different initial velocities v 0 but direct-
ed along the x-axis (Fig. 4). Moving through a region with
an electric and a magnetic field directed along the y-axis,
the particles hit a fluorescent sereon S, which is at a dist-
ance l from point 0. Show that tho track left by tlw par-
ticles on the Rcreen is a parf!bola provided that all the velo-
eities satisfy the condition <•lHllv 0 ~ 1.
7. A homogeneous and stationary magnetic field directed
along the z-axis is created in the region 0 < y < l, -oo <
< x < +oo, -oo < z < +oo (a "magnetic wall"). A par-
ticle of mass m and charge e enters the field with a velocity
v 0 directed at an angle a to the xz-plane. The angle between
the z-axis and the projec-
tion of v 0 on the xz-plane
is ~ (Fig. 5). Find:
(1) the condition for the y
particle's penetration of the
magnetic wall;
(2) the direction the par-
ticle will take after penet-
rating the wall;
(3) the direction of the
particle's reflection from the
wall and determine the con-
v _
ditions in which the laws " - - - - · ·---'--......j'
governing the reflection will Fif{. 5
coincide with optical laws
of reflection.
8. A beam of electrons enters the space between two
pairs of deflecting plates, which have the following voltages:
u X= ul Sin Wt On the vertical plateS and U lj = U 2 COS Wt
on the horizontal ones. Find ·
the path of the beam
on~e scroon S (F~.~ il
all the electrons before
entering have the initial Vo
velocity v0 parallel to the _.:.._-+-t--+---+-t--+-lt---
plates. The length of the "l:::l
plates is l and their distan-
ce from the screen is also l.
9. Consider a harmonic
oscillator with an electric Fi;;. 6
charge e in a homogeneous
and stationary magnetic fwld H (the classical Zeeman
effect). Construct tho motion equation of this oscillator and
!'olve it.
10. Use !.he conditions of Problem 1 and also assume the
particle is acted upon by a force of resistance proportional
lo the first power of tho velocity, R =~ --')'V. Constrncl
!.he motion equation and solve it.
11. Use the conditions of Problem 3 and also assume

the particle is acted upon by a force of resistance propor-
tional to the first power of the velocity, R = -'\'V. Construct
the motion equation and solve it.
12. A mass M with no initial velocity falls from a height
H onto a helical spring (Fig. 7). The mass forces the spring
Fig. 7 Fig. 8
to contract by h. What is the time of contraction if we

neglect the action of forces of resistance and consider the
mass of the spring to be a negligible quantity?
13. A mass m falls in the air with no initial velocity.
Assuming the force of air resistance to be proportional
to the second power of the velocity, R = '\'V2 , find the velo-
city and position of the mass as functions of time. To what
limit does the velocity tend with the passage of time?
14. A cylinder of mass M, radius r, and height h, sus-
pended by a spring whose upper end is fixed, is submerged
in water (Fig. 8). In equilibrium the cylinder sinks to 112
of its height. At a certain moment the cylinder was sub-
merged to 2/3 of its height and then with no initial velocity
started to move vertically. Find the motion equation of
the cylinder in relation to the position of equilibrium if
the stiffness coefficient of the ~pring is c nnd the density
of the water is r.
15. A body falls to the ground from a great height h.
Neglecting the resistance of the air, find the time T that
it will Lake Ll1e body lo reach the ground <md the velocity v
!.hat the body will develop in that time. The earth's:raclius isR.
16. A body of mass m,· thrown at an angl£> a to the hori-
zontal plane with an initial velocity v 0 , moves under the
action of the force of gravity and the force of resistance
of the air R. The force of resistance is proportional to the
first power of the velocity: R = -yv. Solve the motion
equation and determine the maximum height h 11nrt the
.r
Fig. 9
distance s to the base of the maximum height along the

horizontal plane.
17. A particle of mass m moves according to the law
x = a cos wt, y = b sin wt. Determine the force that acts
on the particle at every point of the path.
18. A particle of mass m moves along the path
x2 y2
a2+b2=1
with an acceleration parallel to the y-axis. At t = 0 the
particle is at the point x = 0, y = b and has a velocity v 0 •
Determine the force that acts on the particle at every point
of the path.
19. Find the equation of harmonic oscillations of a par-
ticle that is acted upon by a driving force I = I 0e-a.t cos wt.
The force of friction is proportional to the first power of
the velocity.
20. Consider a particle of mass m that is acted upon by
an elastic force mw~r and a force of friction -myi. Under
what conditions will the motion of the particle be repre-
sented by the diagram in Fig. 9 (i.e. the particle tends to
its position of equilihl'ium, first passing through it)!1 Provl'

that there is only onr maximum rlrviation from Pquilibrium
other than the initial one.
21. A beam of electrons falls on the middle of a vane of
the Crookes radiometer (Fig. 10). Determine the law that
governs the motion of the vane wheel, which consists of
~-~------:c?c
Fig. 10
six vanes, if the electrons are accelerated by a potential V 0

and the electron current is I 0 • The radius of the vane wheel
is R, its width is l, the thickness of each vane is d, and the
K G
v ¢>----'
Fig. 11 Fig. 12
density of the vane material is p. We ignore the friction

in the bearings, the reflection of the electrons, and the
secondary emission.
22. Plane K is a source of charged particles. The particles
are accelerated by a potential difference V between plane K
and grid G (Fig. 11) and then fly through G into space.
(1) What is the reaction force applied to the system
(plane K and grid G) if we take account of the space charge?
(2) What is the power requirement for accelernting the
particles?
(0) Analyze the dependeneo of lhoso two quantities on

the type of partielos used.
23. Consider a eylindrical capacitor (Fig. 12) with a homo-
geneous and stationary magnetic fteld H directed along its
axis (perpendicular to the paper). A particle of mass M
and charge e flies into the entrance slit with a kinetic ener-
gy K. What must the potential difference between the
inner and the outer cylinder be if we want the particle to
fly along the capacitor's midline? How can such a device
be used as a mass-analyzer for ion separation?
Hint. The elrctric field in a cylindrical capacitor is
E'r cc.c V
rln~
l'f
where r 1 is the radius of the inner cylinder, r 2 the radius

of the outer cylinder, and V the potential difference between
the two.
24. What masses should single-charged ions have in
order to fly through the device described in Problem 23
if (1) V = 300 V, r 1 = 6 em, r 2 = 5.4 em, K = 1000 eV,
and H changes from 0 to 10 000 Oe; (2) H = 5000 Oe, r 1 =
= 6 em, r 2 = 5.4 em, K = 1000 oV, and V changes from 0
to 20 000 V?
25. The relative motion of two particles interacting via
the Coulomb law ( V = ~ ) is represented by a conical
section with the parameters
£2
and
P=rfllal
where L is the relative angular momentum, l.t the reduced
mass, and E the energy of relative motion. Prove that
even if E is negative, it always has a lower bound so that
the condition e2 > 0 always holds.
26. Express the period of revolution of an earth satellite
in terms of various parameters.
27. Find the path of a particle of mass m moving in an
external field with a potential V =~ r
.1_
r2
+• Determine
the condition in which the particle will (1) "fall" on the

centre; (2) "pass" to infinity (scatter); (:J) be in alternating

mot ion.
28. Determine the relation between the impact parameter
s and the scattering angle (J) using the conditions of Pro-
blem 27.
29. A particle of mass m and energy E is in a state of
one-dimensional motion in a potential field U (x) (Fig. 13).
U{x)
Determine the particle's pe-
riod of motion.
30. Determine the period of
one-dimensional motion of
a particle of mass m and ener-
gy E in a potential field
U ~~ _ Uo
cosh2 ax
-U0 <E<0.
Fig. 13 31. Determine the period
of one-dimensional motion
of a particle of mass m and energy E in a potential field
U = U 0 tan 2 ax.
32. Find the path of a particle of mass m and energy E
moving in a potential field U = alr 2 (a > 0).
33. Find the scattering angle and the effective cross sec-
tion when a particle of energy E is scattered by a potential
field U = ~ (a > 0).
34. Show that when two particles interact via the Coulomb
law, there is an integral of motion equal to (a vector quan-
tity)
ar
[v X L J+- r
where v is the relative velocity, r the relative radius vector,
L = J.l. [r X v] the relative angular momentum, and a
the constant in the Coulomb law.
35. Consider a rocket on which none but a reaction force
is acting. Find the relationship (Tsiolkovsky's formula)
between time and the rocket's velocity if we know the law
of change of the rocket's mass m (t) and the exhaust velocity

u 1 relative to the rocket.
36. Orbital velocity is the minimum velocity which
a body must attain to establish a permanent orbit, i.e.
to become a satellite. Escape velocity is the minimum
velocity which a body must attain to escape from the gra-
vitational pull of a heavenly body. Find the orbital (v1 )
and escape (v 2 ) velocities for the earth and the moon.
37. Determine the final amplitude (as t-+ oo) of the
harmonic oscillations of a mass m after the action of the
following force (we disregard friction):
t
(1) F = F0 T when 0 <t< T, F = F 0 when T <
< t< oo;
(2) F = F0 --f- when 0 <t< T, F = 0 when T <t <
< oo;
(3) F = F 0 when 0 <t< T, F = 0 when T <t< oo;
(4) F = F 0 sin wt when 0 < t< T = 2(J)n , F = 0 when
T < t < oo. Before the force acted, the oscillator was in
equilibrium.
38. A particle of mass m1 , having a velocity v1 , is scat-
tered by a particle of mass m 2 at an angle El. Find the scat-
tering angle cD in the centre-of-mass frame, the energy
transfer, and the mass ratio for which the energy transfer
is maximal.
39. A particle of mass m is moving under the action of
an external force F = kmr, where r is the particle's radius
vector. Determine the path of the particle if its initial
position is r 0 and its initial velocity v 0 is perpendicular
to r 0 •
40. Solve the motion equation for the cylinder of Problem
14 if the water resistance is proportional to the first power
of the velocity: Fres = -av.
41. Prove that if the linear momentum of a system of n
particles is equal to zero, then the angular momentum of
the system does not depend on the choice of the point relative
to which it is calculated.
42. Prove that the moment of the forces applied to a
system of n particles does not depend on the choice of the
:I -0 149G
point relative to which the moment is calculated if the

resultant of the forces applied to the system is equal to zero.
43. An atom consists of a nucleus of mass A1 and n elec-
trons of mass m each. Eliminate the motion of the centre
of mass and reduce the problem to the motion of n particles.
Find the Lagrangian of the system.
44. In Problem 43 the kinetic energy in the Lagrangian
is not a simple sum of squares. Prove that if the kinetic
energy is expressed in terms of the Jacobi coordinates, it
takes the form of a simple sum of squares. The Jacobi coor-
dinates are
mt
p 1 -=, -
mt
r 1 --r 2 =r 1 - r2
mtl"t +m2r2
Pz = mt + m2 - ra
m1r1 + ... + mjl'j

Pi mt+ ... +mJ
Pn=
mtrt+ ... +mnrn
mt+ ··· mn
= R•
45. Construct the Lagrangian of a dipole whose opposite
charges are of masses m1 and m 2 and which is located in
a homogeneous electric field E.
46. A particle of mass m moves along
the inner surface of a vertical cylinder of
radius r (Fig. 14). Find the pressure that the
/ particle exerts on the cylinder if the sur-
r---...,... ·face of the cylinder is considered to be per-
fectly smooth. The particle's initial velo-
city v 0 forms an angle a with the hori-
zontal plane.
Fig. 14 47. In Problem 46 find the position of
the particle as function of time if the
particle was on the x-axis at the initial instant of time.
48. A pipe AB revolves with a constant angular velocity
w on a vertical axis CD, forming a permanent angle a
with it (Fig. 15). Inside the pipe there is a ball of mass m.
Determine the nature of the ball's motion if its initial
velocity is equal to zero and its initial position is at a dis-
tance a from a point 0. We exclude friction.
49. A thin, straight, and homogeneous rod of length l

and mass 111 revolves with a constant angular velocity ro
about a stationary point 0 describing a r-onical snrface
c B
tr D
A
I ~
n
Fig. 15 Fig. 16
(Fig. 16). Determine the angle of the rod's deviation from

the vertical and the force of reaction at point 0.
50. A homogeneous prism A lies on a horizontal plane.
A prism B (also homogeneous) is placed on prism A
(Fig. 17). The cross sections of both are right trianglE's, and
Fig. 17
the mass of prism A is n times that of B. Determine how

far prism A has moved when prism B, sliding down A,
reaches the horizontal plane. Assume that the prisms and
the plane are perfectly smooth.
51. Two blocks of masses m1 and m 2 connected by a mass-
less, inextensible cord that is passed over a massless pulley A
slide along the smooth sides of a rectangular wedge of
mass m, which rests on a smooth horizontal plane (Fig. 18).
3*
36 J>ROIILEMS IN THEORETICAL PHYSICS
Find the displacement of the wedge 011 tlte horizontal plane

when mass m1 is lowered to a height h.
52. An electric motor of mass M is placed on a smooth
horizontal foundation but not fastened down. A homoge-
neous rod of length 2l and mass M 1 is attached to the motor's
shaft at a right angle. The other end of the rod has a particle
of mass m attached to it (Fig. Hl). The shaft turns at an
angular velocity w. Determine
(1) the equation for the horizontal motion of the motor;
A
m
FiK. 19 FiK. 20
(2) the maximal horizontal stress R which would act

on bolts if the motor were fastened to the foundation;
(3) the required angular velocity of the shaft so that
the motor will bounce on the foundation if it is not bolted
down.
53. Can a complex system of particles whose centre-of-
mass energy is E decay into two systems with energies E 1
and E 2 ?
54. Due to the rotation of the earth about its axis any
freely falling body deviates from the vertical line. Find
this deviation if the initial velocity is zero. Ignore air

resistance.
55. Construct the Lagrangian of a simple pendulum
whose point of suspension moves in the vertical plane accord-
ing to the law y = y (t) and x = x (t). The mass of the
pendulum is m and its length is l.
56. Construct the motion equation for small oscillatioiJS
of a simple pendulum whose point of suspension moves
along the vertical line according to the law x = a cos (!)[.
The mass of the pendulum is m and its length is Z.
57. A mechanical system, depicted in Fig. 20, rotates
about the vertical axis A R with a constant angular velocity roJ.
p{t)
~
Fig. 21 Fig. 22
The body of mass m 2 is able to move along the vertical

axis. Find the Lagrangian of the system and determine the
positions of equilibrium.
58. In Problem 57 determine which of the two positions
of equilibrium is the position of stable equilibrium and
which of unstable equilibrium.
59. Two simple pendulums of equal length l (Fig. 21)
are coupled by a spring with a stiffness coefficient c Ht a
distance a from the suspension points. Determine the fre-
quency of small oscillations and solve the motion equation
of such if at the initial instant of time one pendulum was
detlected from the vertical line by an angle cp 0 •
60. The capacitor microphone consists of a series-circuit
of inductance L, conductance R, and capacitance C of a
eapacitor whose plates (one of which can move) are conpled
by two springs with a general stiffness coefficient c (Fig. 22).
This circuit is connected to a cell with a constant e.m.f. ~

In the state of equilibrium C = C 0 and the distance between
the plates is a. Construct the Lagrangian and write the
Lagrange equation if the mass of the moving plate is m
and a variable force p (t) acts on it.
61. Find the states of equilibrium for the system of
Problem 60. Determine the frequency of small oscillations.
m
Fig. 23 Fi~. 24
62. A ball of mass m hangs at the end of a massless rigid

rod of length l. Two springs with a stiffness coefficient
c are connected to the rod as shown in Fig. 23. Find
, the frequency of small oscilla-
tions.
63. Find the frequency of small
oscillations when the pendulum of
Problem 62 is turned upside down,
i.e. the ball is higher than the for-
mer point of suspension (Fig. 24).
Determine the state of equilibrium
in this case.
m
64. A block of mass M, connected
Fig. 25
to a spring with a stiffness coeffici-
ent c, can move in the horizontal
plane without friction. A simple pendnlum of mass m and
length l is fastened to it (Fig. 25). Find the Lagrangian of
the system and determine the frequency of sma11 oscil-
1ations. (The other end of the spring is fixed.)
65. Construct the Lagrangian of a simple pendulum of
mass m and length l whose point of susprmsion moves in
the horizontal plane according to the law x = x (t).
66. Construct the motion equation for small oscillations

of a simple pendulum of length l whose point of suspension
moves on a horizontal line according to the law x =a cos vt.
67. A homogeneous rod BD leans against a wall (Fig. 26).
The rod's lower end rests on a horizontal plane ancl is held
D
Fig. :!6 Fig. 27
by the line AB. Find the reaction forces at the points of

contact and the tension in the line. The weight of the rod
is P and its length is l.
68. A homogeneous rod AB leans against a vertical- plane
and rests on a horizontal plane (Fig. 27). Two horizontal
lines AD and BC hold the rod in a fixed position; the line BC
and the rod are in the same vertical plane. Find the reaction
forces at the points of contact and the tension in the lines.
The weight of the rod is P.
69. A ball B of mass m is suspended by a string AB and
touches the smooth surface of a sphere of radius r (Fig. 28).
The distance from the surface of the sphere to the static
point A h; d, and the length of the string is l. Find the ten-
sion in the line and the reaction force with which the sphere
acts on the ball. The dimensions of the ball are negligible.
70. Two straight homogeneous rods of length a and b
are rigidly fixed at a right angle whose vertex 0 is connected
l.o a vertical shaft by means of a joint (Fig. 29). The shaft
r·ot.ates with a constant angular velocity (t>. Determine the
r·Plationship between w and the angle cp that is formed by
1he rod with length a and the vertical.
71. Consider a load of weight P at a point F on the plat-

form of a balance (Fig. 30). AB = a, BC = b, CD = c,
I K = d, and the length of the platform EG is l. Determine
A
ill
I
Fig. 2/l Fig. 29
the relationship between b, c, d, and l for the case when

the weight Pw which compensates the weight of the load
does not depend on the position of point F. Also determine
the actual value of Pw in this case.
A 8 CD
Pw
p
f
c
H f' V K
I/\'
'
Fig. 30 Fig . .U
72. A schematic drawing of a device for measuring elastic

constants of solids is depicted in Fig. 31. Find the relation-
ship betwePn the force F that is applied to the sample K
and the distance from the weight P to its zero point 0,
if weight Q balances the device in such a way that all the
arms are in a horizontal position when the weight P is at
its zero point and when there are no stresses in the sample.
73. Two rods AB and OC are connected at a right angle

at a point C (Fig. 32). OC can rotate about a horizontal
axis that passes through 0; A C = CB = a and OC = b.
The points A and B 'are loaded by weights P 1 and P 2 , res-
pectively. Find the angle that AB forms with the horizontal
plane in equilibrium if the weight of both rods is 2p per
unit of length.
74. Three rods of equal length, AB = BC = DC = a,
are connected at right angles at points B and C (Fig. 33).
Fig. 32
The rod AB can rotate about a horizontal axis 0 that divides

the rod in half. Determine the positions of stable and un-
stable equilibrium of the system if the weight of the unit
length of the rod is p.
75. Find the Poisson bracket for the Cartesian components
of the linear momentum p and angular momentum L =
= [r X pl.
76. Find the Poisson bracket for the components of the
n
vector of angular momentum L =-= LJ
i=1
[ri X Pil for a system
of n particles.
77. Show that {cp, Lz} = 0, where cp is an arbitrary
fnnction of the position and momentum of a particle.
78. Find the condition for the linear transformations
of p and q
Q = aq + bp
P = cq +- dp
to be canonical.
79. Show that {f, Lz} = [n X fl. where f is an arbitrary

function of the position and momentum of a particle and
n is a unit vector in the direction of the z-axis.
80. Construct the Lagrangian and the Hamiltonian for
an electrically charged harmonic oscillator that is located
in a homogeneous and stationary magnetic field B (B =
= f.t 0 H, where !-to is the permeability of empty space).
a system of two particles interacting via the Coulomb law.
Express both functions in terms of the centre of mass coor-
dinates and the separation between the particles.
a system of n particles interacting via the Coulomb law.
The system is in an external electromagnetic field.
83. Find the Lagrangian and the Hamiltonian for a sys-
tem of two oppositely charged particles placed in a homo-
geneous magnetic field. Show that by adding to the Lagran-
gian the total time derivative of a specially chosen function
of coordinates, we can make the coordinates of the centre
of mass cyclic (ignorable). What integral of motion corres-
ponds to these cyclic :coordinates?
84. Construct the Hamiltonian for a symmetrical top
with one fixed point in the field of the earth's gravity.
85. Consider a thin disc of mass M sliding along a per-
fectly smooth horizontal plane. A particle of mass m moves
across the disc. In a coordinate system that is attached to
the disc and whose origin lies in the centre of the disc, the
motion of the particle is governed by the law x = x (t)
and y = y (t). Find the angular velocity of the disc as a
function of time if at the initial instant of time the disc
was motionless.
86. On the disc of Problem 85 the same particle of mass m
moves with a velocity at (relative to the disc) along a cir-
cumference of radius R. Find the motion equations.
87. Determine the inertia tensor, relative to the centre
of mass, for the following molecules:
(1) CH 4 ; its structure is represented by the regular tetra-
hedron (a four-faced polyhedron with all of its faces equila-
teral triangles) with the carbon atom Cat its centre and the
four hydrogen atoms H at its vertices. The distance CH is
a = 1.07 A;
(2) H 20; its structure is depicted in Fig. 34a;

(3) NH 3 ; its structure is depicted in Fig. 34b.
88. Prove that for a diatomic molecule the inertia tensor,
calculated in the centre-of-mass reference frame, is deter-
mined only by one quantity, i.e. I = ~-ta 2 , where f.t isthe
reduced mass of this diatomic system and a the distance
between the positions of equilibrium of the atoms.
89. The moment of inertia of a molecule of hydrogen
fluoride HF is 1.37 X 10- 40 g cm 2 if calculated relative to
the centre of mass. Determine the distance between the
atoms of hydrogen and fluorine.
The atomic mass of hydrogen is ff ~------.,.ff
MH = 1.67 X 10-24 g and that of
fluorine is MF= 3.17 x 10-22 g.
90. Determine the ratio between
the moments of inertia of the mo- 0
lecules H 2 , HD, and D 2 and also the (a)
ratio between their frequencies of
vibrations, assuming that the inter-
atomic potentials do not depend
on the isotopic com position of the
molecules.
91. For the following continuous
rigid bodies, each of mass M,
find the principal sets of axes, ba- Fiff. ,'J4
sed at the corresponding centres of
mass, and the principal moments of inertia:
(1) a rod in the form of a right parallelepiped with edges
of length a, b, and c;
(2) a ball of radius R;
(3) a circular cone with an altitude h and the radius of
the base R;
(4) an ellipsoid with the semiaxes a, b, and c;
(5) a spherical shell with an inner diameter d and an
outer diameter D;
(6) a torns with a cross-section radius r and a middle
radius R;
(7) a cylindrical pipe of I ength l with an inner radius r
nnd an outer radius R;
(8) a right triangular prism with an altitude l and the
side of the triangle a;
(9) a right hexagonal prism with an altitude l and the

side of the hexagon a.
92. Consider a symmetrical top along whose axis of
symmetry there acts a constant moment N of external
forces. Construct the motion equation of such a top and
solve it. The resultant of the external forces is zero.
93. Determine the components of angular velocity as
functions of the angle of proper rotation. For a homogeneous
Jl,kn
Fig. 35
ellipsoid that rotates about one of its axis, AB, which in

turn rotates about the axis CD (Fig. 35) that is perpendicular
to AB, find the maximal and minimal values of these com-
ponents in a system of coordinates whose axes are taken
along the principal axes. CD passes through the centre of
the ellipsoid.
94. Solve Problem 93 under the conditions that AB
forms an angle a with CD and the ellipsoid is symmetric
about the AB-axis (Fig. 36).
95. Construct the Lagrangian for a homogeneous cylinder
of radius a that rolls along the inner surface of a cylinder
of radius R (Fig. 37).
96. A right and homogeneous cylinder of mass M, length
l, and radius r rotates about a Yertical axis OZ with a con-
stant angular velocity <•l. ln the process the angle between
the cylinder's axis of symmetry and OZ (OZ passes througl1
the centre of mass of the cylinder) maintains a constant
SECTION T. CL,ASSICAL MECHANICS 45
value a; the distance between the thrust bearing and tho

bearing at the top (see Fig. :iR) is h. Find thP lateral pres-
sure on both bearings.
Fi~. 36
97. Construct the motion equation in the form of Euler's

equations (see the last three formulas of (I-32) on p. 20)
for a symmetrical top in the earth's gravitational field.
il
Fiff. 37 Fig. 38
98. Find the Euler angles as functions of time for the

l'ree rotation of a symmetrical top.
99. Find the elements of the rotation matrix that repre-
sents three successive rotations in the positive (counter-
clockwise) direction: fust, t hrongh an angle fl about the
x-axis; then, through an angle ~, about the new y'-axis;

last, through fP about the new z "-axis.
100. Express the components of angular velociLy through
the angles of Problem 99 and their time derivatives.
101. Figure 39 illustrates the princi pie of a monorail.
C is the body of the coach of mass M 1 , which moves uni-
formly in a straight line, and
R is a frame that can freely
rotate about a horizontal axis
C A 1A 2 • A counterbalance of
mass M 3 is rigidly attached
~ to the frame. A flywheel of
mass M 2 rotates freely about
an axis B 1B 2 in bearings.
The centre of mass of the fly-
wheel lies at the intersection
.t:: point 0 1 of A 1 A 2 and B 1B 2 •
Show that the solution <p =
~ ffiot, e = ljJ = 0 is a stable
solution of the motion equa-
tion of the monorail coach.
Fig. 3!J Here {p is the rotation angle
of the flywheel, ljJ the deflection
angle of the frame H (with the counterbalance) relative
to the body C, 8 the inclination angle between the body's
axis and the vertical line. The flywheel's centre of mass
is at a distance h1 from point 0, the body's centre of mass
is at a distance l from the same point, and the centre of
mass of the counterbalance is h 2 higher than the flywheel's
centre of mass.
102. An elliptic, nonhomogeneous cylinder of length h,
made out of two kinds of materials of densities p1 and p 2 ,
lies on a horizontal surface (Fig. 40). Determine the states
of stable and unstable equilibrium. The lengths of the semi-
major and semiminor axes are a and b, respectively.
103. Find the components of the strain tensor in spherical
and cylindrical coordinates. (See Appendix 2.)
104. Find the deformation, i.e. the displacement vec-
tor u, for a cylinder that is rotating about its axis with
a constant angular velocity ffi. (For the body force f in
equation (1-50) take the centrifugal force per unit mass.)
105. Construct the dispersion equation for elastic waves

in a single crystal with a cubic lattice. Find the phase
velocities of such waves if the waves propagate parallel
to the faces of the crystal, and if the waves propagate per-
pendicular to the faces.
106. A plane, longitudinal, and monochromatic wave
falls at an angle of incidence 8 0 on the plane interface bet-
.X
Fig. 40
ween a vacuum and a solid (Fig. 41). Find the laws of reflec-
tion and calculate the ratio between the perpendicular com-
ponent of the energy flux of the reflected longitudinal wave
!1
Fig. 41 Fig. 42
and the same component of the energy flux of the incident

wave. Calculate the ratio for the perpendicular component
of the reflected transverse wave. In Fig. 41 ui
is the dis-
placement of the longitudinal incident wave, u! the dis-
placement of the longitudinal reflected wave, and u{ the
displacement of the reflected transverse wave.
48 PHOBLEMS IN THEORETICAL PHYSICS
107. Solve Problem 10() for the case when the incident
wave is transverse and its plane of oscillations coincides
with the incident plane (Fig. 42).
108. An incompressible viscous liquid is flowing under
the influence of a pressure drop !1p = p 2 - p 1 between two
infinite, parallel plates that are at a distance d from one
another. Find the field of velocities and pressure distribu-
tion between the plates.
109. Proceeding from the conditions of Problem 108,
find the heat flux and temperature distribution between
the plates when (1) the temperature
.z of the lower and upper plates is
held at a constant value T 0 by
Vacuum means of a thermostat; (2) the lower
plate is adiabatically isolated and
:::~ the upper plate is held at a con-
~w;o;>m7.1~w;o;w;o;~/ y stant temperature.
Liquid 110. Consider two coaxial cylin-
ders with radiuses r 1 and r 2 (r1 >
Fig. 43 > r 2 ). Find the field of velocities
in an incompressible viscous liquid
between the cylinders for the follow-
ing cases: (1) the ouler cylinder rotates about its axis with
an angular velocity ffi 1 ; (2) the inner cylinder rotates with
an angular velocity ffi 2 ; (3) the outer and inner cylinders
rotate in the same direction with angular velocities ffi 1
and ffi 2 , respectively.
111. Under the conditions of Problem 110 fmd the tem-
perature distribution in the liquid if the rotating inner
cylinder is thermally isolated and the motionless outer
cylinder is held at a constant temperature T 0 •
112. Write the right part of the N a vier-Stokes equation
in cylindrical and spherical coordinates.
113. An infinite plate of thickness d is the interface
between a vacumn and a liquid (Fig. 43). Find the natural
frequencies ffi for longitudinal sound vibrations assuming
that all quantities depend on the transverse coordinate x only.
114. Find the dispersion equation for the propagation
of an elastic wave u = f (z} ei(wt-kx) in an isotropic solid
(the Rayleigh wave). Here f (z) is a damped function. Study
the structure of this wave.
115. Determine the natural frequencies of the radial

vibrations of an elastic sphere with a radius R placed in
a vacuum.
116. Determine the ·frequency of the radial vibrations
of a spherical cavity in an infinite elastic medium.
117. Determine the natural frequencies of the sound
vibrations of a gas that fills a closed section of a rectangular
pipe of dimensions a, b, and d. Assume the pipe is rigid.
118. Construct the wave equation and find the disper-
sion equation for sound waves generated by a source
moving with a constant velocity v 0 relative to a liquid
(gas). Find the frequency registered by a receiver that is
motionless with respect .to the liquid (gas).
119. What frequency will be registered by a receiver
that moves in a liquid (gas) relative to a stationary source
of sound (Doppler effect)?
120. Find the field of velocities in a liquid that is flowing
through a ring-shaped pipe (the inner and outer radiuses
are r 1 and r 2 , respectively).
the influence of a pressure drop !!,.p = p 2 - p 1 in a pipe
with an elliptic cross section. The length of the pipe is l.
Determine the field of velocities and the amount of liquid
that flows through the cross section in a unit of time.
the influence of a pressure drop /!,.p = p 2 - p 1 in a pipe
with a circular cross section. Find the temperature distri-
bution if the temperature of the pipe is held at a constant
value T 0 •
4-01496
SECTION II Electrodynamics
The two fundamental quantities that characterize an

electromagnetic field are the electric field vector E and
the vector of magnetic induction B (the magnetic flux den-
sity). The electric field vector is determined from the force
acting on a test charge e, i.e.
F = eE (11-1)
and the magnetic induction is determined from the force
acting on a circuit element dl which carries a current /,
i.e.
dF =I [dl X B) (11-2)
When considering an electromagnetic field in a dielectric
or magnetic material, where the field is created by external
sources or by the polarization and magnetization of the
material, it is convenient to introduce two supplementary
characteristics of the electromagnetic field: the electric
displacement vector
D = E 0E +
P (11-3)
and the magnetic field strength (or intensity)
B
H=--M (II-4)
JlO
where P is the polarization vector representing the electric
dipole moment per unit volume, M is the magnetization
vector representing the magnetic dipole moment per unit
volume, and Eo and flo are the permittivity and permea-
bility of empty space, respectively:
Eo = 8.854 X 10-12 F m - 1 (11-5)
flo = 4n X 10-7 H m - 1 ( II-6)
Eoflo = c-2
where c is the velocity of light in empty space.
SECTION II. ELECTRODYNAMICS 51
The relationships between P and E, M and H are deter-

mined by the properties of the medium. For an anisotropic
medium in the apprc;>ximation linear with respect to the
field and for fields slowly varying in space and time the
following relationships hold true:
Pi = Eoaii,EI< (II-7)
Mi = Xi~tHh (II-8)
The indexes i and k assume the values 1, 2, and 3 and
denote the projections of a vector (P or M) on the x, y,
and z axes. In all formulas where the same index appears
twice we shall automatically sum over that index. In (II-7)
aih is the polarizability tensor of the medium and in (II-8)
Xi~< is the magnetic susceptibility tensor of the medium.
Through the use of (II-3), (11-4), (II-7), and (II-8) we
can obtain the relationships between D and E, B and H:
(II-9)
(II-10)
where
is the permittivity tensor, and

f.lih = {jih + Xi~<
is the permeability tensor.
In an isotropic medium the relationships (II-7)-(II-10)
are simplified thus: c
P = E 0aE (II-11)
M = xH (II-12)
D = ee 0E (II-13)
B = f.tf.toH (II-14)
where E = 1 +a is simply the permittivity of the medium

nnd f.t = 1+ x the permeability.
A set of four equations called the Maxwell equations
are the starting equations of electrodynamics. In the
4*
integral form they are:

~ H 1 dl = ) jn dS + ) ( aa~ LdS (II-15)
~ E 1 dl = -) ( aa~ ) n dS (II-16)
~DndS= JpdV (II-17)
~BndS=O (II-18)
where j is the conduction current density and p the electric

charge density.
In the differential form the Maxwell equations are:
curlH=J· +~
at (II-19)
curiE=-~
at (II-20)
divD=p (II-21)
divB=O (II-22)
At a boundary between two media the electromagnetic
field vectors are subject to the following boundary conditions:
D2n- D1n =a (II-23)

Eu = E2t (II-24)
Bin = B2n (II-25)
[n X (H2 - H 1 )1 = i (II-26)
where n is ...t vector normal to the boundary and directed

from medium 1 to medium 2, a the surface charge density,
i the surface current per unit length flowing along the boun-
dary, E t the tangential component of the electric field
at the boundary, and Dn and Bn are the normal compo-
nents of the electric displacement and the magnetic induc-
tion, respectively, at the boundary. ·
Electrostatics. In this branch of electrodynamics all
quantities are constant in time and electric charges are
stationary. From the Maxwell equations (II-19)-(II-22)
we then have the system of equations of electrostatics:

curl E = 0 (II-27)
div D = p (II-28)
Equation (II-27) is satisfied if we introduce the electro-
static potential by the relation
E =-grad <p (II-29)
In a homogeneous m'edium the potential satisfies the
Poisson equation
~cp = - ...£.__ (II-30)
ee 0
or when there is no charge the Laplace equation
~<p = 0
At an interface of two media with permittivities B1 and
B2 the potential is subject to the following boundary con-
ditions:
C!Jt = <Jlz, (II-31)
Bt ( ~: ) 1 -Bz ( :: ) 2 = e: (II-32)
In electrostatics the electric field does not penetrate a
conductor, and therefore from equation (II -29) it follows
that inside a conductor the potential is constant.
Since :: = 0 inside a conductor, from the boundary con-
dition (II-32) we can find the relationship between the
surface charge density (on the surface of the conductor)
and the potential near the conductor:
0'= - B0B (~)

an s (II-33)
The total charge of the conductor is

e= - ~ B0B ; : dS (II-34)
where the integral is taken over the conductor's surface S.

~-·
For a system of more than one conductor there are usu-
ally two types of problems:
(1) given the potential on every conductor, i.e.

cp (r) Is.= cpi (11-35)
'
find the potential at any point in the system;
(2) given the charge· of every conductor q; i.e.
cp (r) Is.= constant
'arp
- ~ e-dS-=-
q;
(II-36)
an '. eo
si
find the potential at any point in the system.
Many problems in electrostatics that possess cylindrical
symmetry can be solved through the use of functions of
a complex variable. The real and imaginary parts of an
analytic function W (z) = cp +
i'¥ of a complex variable
z = x +iy are solutions of the Laplace equation: ~cp = 0
and ~'¥ = 0. Besides, cp and '¥ are restricted by the con-
dition
~ av +~ av =O
ax ox ay iJy
i.e. the curves cp = constant and'¥ = constant are mutually

orthogonal. Thus cp and '¥ can be the solutions of the elec-
trostatic problem. If cp (x, y) is the potential, the curves
'¥ = constant are the lines of force orthogonal to the equi-
potential surfaces. Let us put cp = 0 on a smooth contour L.
In order to find the equipotential surfaces cp = constant,
we must write the equation for the contour L in the para-
metric form
X = f (P). y = F (P) (11-37)
where f and F are single-valued functions and where the
range of admissible values of P corresponds to the move-
ment of the point (x, y) along the grounded conductor
(cp = 0). We must seek the potential in the form
cp = Im W (11-38)
and W is found from the equation z = f (W) +iF (W).
~; If the distribution of body and surface charges is given,
the potential at point r is
_ 1 f p (r') dV' 1 f ;u (r') dS'
cp (r)- 4neeo J I r-r' I + 4neeo J I r-r' I (ll- 39)
At great distances, i.e. distances that are much bigger

than the dimensions of the system, the potential can be
represented as a series
+ +
cp (r) = cp<O> (r) cp<O (r) cp<2> (r) + ...
(II-40)
where cp< 0 > (r) is the potential produced by a point charge
equal to the total charge of the system:
cp<O>(r)= e (II-41)
4neeor
cp<1 > (r) the potential produced by a dipole with the dipole
moment p = Jp (r) r dV:
cp<1>(r)= (p·r) (II-42)
4:rtee 0r3
and cp< 2 > (r) the potential produced by a quadrupole:
(II-43)
The components of the tensor of quadrupole moment are

defined as follows:
Qik = Jp (r) (3xixk- r 6nJdV 2 (II-44)
The energy of an electrostatic field is
(11-45)
The interaction energy of two electrically charged sys-

tems with charge densities p1 and p 2 is determined thus:
(11-46)
Magnetostatics. Phenomena that take place in a constant

magnetic field, i.e. when the magnetic field strength and
the magnetic induction are independent of time, are gov-
erned by the system of equations~ of magnetostatics:
curl H = j (II-47)
div B = 0 (II-48)
together with the boundary conditions (II-25) and (II-26).

Equation (II-48) is satisfied if we introduce the vector
potential by the relation
B = curl A (II-49)
In a homogeneous and isotropic medium the vector
potential A satisfies the equation
~A = -1-t~-toi
(II-50)
The solution of this equation for a given current density j
has the form
A (r) = flllo i JJ:!:1_ dV' (11-51)
4n J I r-r' I
At distances considerably greater than the dimensions
of the system
A (r) = flflo [m X r) (11-52)
4nr3
where m is the magnetic moment of the system equal to
m= ~ J[r X j (r)] dV (11-53)
The energy of a stationary magnetic field is

W= ~ ) (B·H)dV=+ J(A·j)dV
= f1Jlo \ j (r) j (r') dV dV' (II-S 4 )
8n J I r-r' I
For a system of conductors the energy of the magnetic
field can be rewritten as
W = ~ ~ Linliln (11-55)
i, It
where
L·11. = flf!o i in (rl!.) jdri) dV11dVi (II-S 6)
1 4nfnl i J I r11- r i I
The factors Lilt fori ::f= k are called the mutual inductances,
and for i = k the self-inductances.
Quasi-stationary fields. In the case of fields slowly vary-
ing in time (i.e.
{11-57)
where a* is the electrical conductivity of the medium;

w and 'A are the frequency and the wavelength of the elec-
tromagnetic oscillations, respectively; and l is the charac-
l
teristic length of the· system) the Maxwell equations take
the form
curl H= j
aB
curl E= - - at (11-58)
divD=p J
div B=O
Electromagnetic waves. In the general case of varying
fields we must solve the system of equations (II-19)-(II-22).
When there are no charges and conduction currents, this
system describes an electromagnetic field in free space.
Plane monochromatic waves, in which
E = Eoeikr- irot (II-59)
H = Hoeikr-irot (11-60)
are a partial solution for the Maxwell equations for a free
field. Here w is the frequency of electromagnetic oscillations
and k is the wave vector. In an isotropic medium the direc-
tion of k coincides with the direction of the propagation
of the wave's energy. In magnitude k = w!v, where v =
= c (ef.t)- 112 is the phase velocity of the wave.
The density of the energy flux of an electromagnetic
field is defined by the Poynting vector
S = [E X Hl (II-61)
For varying electromagnetic fields we can express the
relationship between the electric field intensity and the
magnetic induction, on the one hand, and the potentials,
on the other, in this way:
iJA
E=-gradcp-Tt (II-62)
B =curl A (II-63)
If we impose the Lorentz condition
. A+ ej.l. acp o
d lV (II-64)
7£7ft=
the potentials cp and A satisfy the equations

~ llf.l. iJ2<p - p
cp- ~ 8t 2 - - eeo (11-65)
AA llf.l. iJ 2A •
(II-66)
Ll - ~ at2" = - flfloJ
We can write the solutions of (II-65) and (II-66) in the
form of retarded potentials ·
1 (l ( r', t lr-;;-r' I)
cp (r, t) = 4neeo ~ I r-r' I dV' (II-67)
•( , t I r - r' I )
A (r t) = flflo i J r ' - -v- dV' (11-68)
' 4:n: J I r-r' I
In a vacuum and at distances from a system of charges
which are considerably greater than the length of the elec-
tromagnetic wave radiated by this system, r ~'A, we get
1 •
B= c [A X n] (II-69)
E =c [B X n] =[[A X n] X n] (II-70)
where
A(r, t)=__EQ_ i · (r' t - r-(r'·n)) dV' (11-71)
4nr JJ ' c
and n = ..!:.. is a unit vector in the direction of wave pro-
r
pagation.
If, in addition, the wavelength is much greater than the
dimensions of the radiating system of charges, the electro-
magnetic field at great distances can be represented as a
sum of fields generated by a dipole, quadrupole and other
multipoles. The dipole radiation has the maximum inten-
sity. The corresponding field is determined by the relation-
ships
B= !-lo[pXn] }
4:ncr
(II-72)
E= t:r [[p X n] X n]
where p is the dipole moment.

SECTION II. ELgCTRODYNAMICS 59
The intensity of the dipole radiation is determined by

the formula
••)2
. J= ( p (II-73)
6:n:eoc3
Magnetohydrodynamics. This branch of electrodynamics

studies the effect of electromagnetic fields on electrically
conducting fluids. In the hydrodynamic approximation
the motion of a system is described by the variations in
the density, velocity,! and pressure. At low frequencies
the electromagnetic field satisfies the equations (II-58).
The equations of magnetohydrodynamics look like this:
iJpm +d"lVPmV =
-at 0 (II-74)
Pm ~; + Pm (v · V') v =-grad p + [j X B] + 1')/lv-t-pmg (11-75)

iJB
curl E= - -
at (11-76)
curl H =j (II-77)
j =a* (E + [v X Bl) (II-78)
where Pm is the density of the medium, p the pressure, 1')
the viscosity, g the acceleration due to gravity, cr* the
electrical conductivity of the medium, and v the velocity.
Equations (II-74)-(II-78) must be supplemented by the
state equations of the medium.
Let us consider in brief the special theory of relativity.
The fundamental principles of the special theory of rela-
tivity are:
(1) the velocity of light in free space is the same in any
inertial frame of reference and equal to c = 2. 99793 X
X 108 m s-1;
(2) the laws of physics act in all inertial reference frames
in the same way.
These principles have the following corollaries.
Consider an event with coordinates x 1 = x, x 2 = y,
x 3 = z, x4 = ict in a reference frame K and with coordi-
nates x; = x', x; = y', x; = z', x~ = ict' in another refer-
ence frame K'. If K' moves along the x-axis of K with
a velocity v in relation to K, the coordinates of the event
in the two reference frames are linked by the Lorentz trans-

formation
(II-79)
where ~ = vic.
Assume that a body moves with a velocity u in relation
to K. The velocity in relation to K' will then be
u'- Ux-V
U~=
Uy VI=ti2 Uz lfq2 (11-80)
x- vu ' 1- 1-
1--x- VUx VUx
c2 c2 c2
When a force F acts on a particle, the law of motion

in the differential form is
dp -F (11-81)
dt -
where
m0v
(11-82)
P= V1-~2
is the momentum of the relativistic particle, and m 0 is

its rest mass.
A totality of four numbers that transform according
to a Lorentz transformation if we change the frame of
reference is called a 4-vector. This may be the 4-momentum
with the components (p, iE/c), where p is the conventional
momentum and E is the energy; or a wave vector and fre-
quency in a plane electromagnetic wave propagating in
a vacuum (k, iwlc); or the 4-vector of current density
(j, icp); or the 4-vector potential (A, irp/c).
A totality of sixteen numbers that transform according
to a double Lorentz transformation if we change the frame
of reference is called a 4-tensor of rank 2. This may be
the electromagnetic field tensor, which can be represent-
ed as
0 cBz -cBy -iEx)
_ ( -cfJz 0 cBx - iEy
Fa.r.- cBy -cBx 0 - iEz (II-83)
iEx iEu iEz 0
or the energy-momentum tensor
T a.r> = e0 ( F a.14Fr.11 - ! 6ar.F~v) (II-84)
(a, ~. f.!, v=1, 2, 3, 4)
PROBLEMS
Vector analysis
1. Calculate the gradient of a function, f (r), that depends
only on the absolute value of the radius vector r.
2. Calculate div r, curl r, curl <p (r) r.
3. Calculate grad (P·r), grad (Pr~r), (P·V) r, div [Pxr],
curl [r X P], where P is a constant vector.
4. Calculate grad A (r) B (r), div <p (r) A (r), curl (<p (r) X
xA (r)). The functions <p (r), A (r), and B (r) depend only
on the absolute value of the radius vector r.
5. Using Ostrogradski's theorem, calculate the integrals
I=~ r(A·n)dS, I=~(A·r)ndS
if the volume enclosed by the surface is V and if A is a con-
stant vector.
J
6. Show that A dV = 0 if inside the volume divA = 0
v
and on its boundary An = 0.
7. Show that the divergence of vector
A+ _1_ rad \ divA (r') dV'
4n g J I r-r' I
is zero.
8. Find a solution for the Laplace equation that depends
only on the absolute value r of the radius vector.
9. Write the Maxwell equations (II-19) to (II-22) in
(a) cylindrical coordinates; (b) spherical coordinates.
Electrostatics
10. Find the electric field intensity vector inside and
outside a uniformly charged sphere of radius R. The body
charge density of the sphere is p.
11. A sphere of radius R is uniformly charged with
a charge density p. Inside it there is a spherical cavity
of radius R' whose centre is at a distance a from the centre
of the sphere. Find the electric field intensity vector inside
the cavity and inside and outside the sphere.
12. Find the electric field vector inside and outside
a sphere with a body charge density varying as follows:
p = arn
where n > -2. The radius of the sphere is R.
13. Find the electric field vector inside and outside
a uniformly charged solid cylinder of radius R. The electric
charge p_er unit length of the cylinder is x.
14. A layer of nonconducting matter is put between two
parallel planes and is charged to a density p. Find the
electric field vector inside and outside the layer if its thick-
ness is d.
15. Find the capacitances of the following capacitors:
(a) spherical, (b) plane-parallel, and (c) cylindrical. Between
the plates of each capacitor there is a dielectric of permit-
tivity e.
16. Two long, cylindrical conductors are arranged paral-
lel to each other at a distance d. Calculate the capacitance
per unit length of the system provided d ~ R 1 and d ~ R 2 ,
where R 1 and R 2 are the radiuses of the cylinders.
17. Find the equation of the lines of force for a system
of two point charges e and -e with a distance d between
them.
18*. For the case of a homogeneous electric field with
a field vector E write the corresponding complex-valued
potential W. Consider the special case of the electric field
of a charged plane with a surface charge density a.
19. Determine the potential near a grounded angle
formed by two planes x = 0 and y = 0.
* In Problems 18 to 24 we assume the distribution of the potentials
to be two-dimensional.
SECTION II. ELhlCTRODYNAMlCS 63
20. Determine the equipotential surfaces and lines of

for'Ce_;~if theJ pote'iitiil iis <p = Re-'(Vz).~ What grounded
contour has such a potential?
21. Determine the potential and the equipotential sur-
faces if the complex-valued potential is W = In z.
22. Find the potential near a grounded parabola y 2 =
= 4a (x +a).
23. Find the potential near a grounded ellipse a 2
x2
+
+ y2b =
2 1. Consider the limiting case of the potential
near a circle (in the three-dimensional case, a cylinder),
assuming b = a.
24. Find the equipotential surfaces and lines of force
x2 y2
near a grounded hyperbola az - b2 = 1.
25. Find the potential and the electric field intensity
vector on the axis of a flat ring that has a surface charge
density a (the inner radius of the ring is R 1 and the outer
R 2). Consider the following limiting cases: (a) the field
of a flat disc (R 1 -+ 0) and (b) the field of a plane (R 1 -+ 0,
R 2 -+ oo).
26._ Determine the potential of the electric field pro-
duced by the electron of a hydrogen atom, assuming that
the electron charge in the ground state is distributed with
a charge density p = ....!!......3 e_ 2 r1a, where a is a constant.
na
27. Determine the Fourier transform for the potential
of a point charge.
28. Find the potential of the electric field produced
by a charge that is distributed in an infinite medium by
the law p = Po sin ax sin by sin cz.
29. A point charge e is situated at a distance d from
a conducting plane that is grounded. Find the potential
and the electric field intensity vector of the system. Deter-
mine the surface density of the charge that is induced on
the grounded surface. Show that the total induced charge
is equal to -e.
30. Using the method of images, find the potential of
a charge q that is placed inside a right angle formed by
two conducting planes.
31. A point "charge e i~ at a distance d from the centre

of a conducting sphere of radius R. Using the method of
images, determine the electric potential of the sys-
tem. The sphere is grounded.
32. Find the potential of a system consisting of a point
charge e and an insulated conducting sphere of radius R
near the charge.
33. A point charge e is at a distance d from the centre
of a spherical projection of a conducting plane. The centre
of the projection lies on the plane, and the charge is situated
opposite the point on the projection that is farthest from
the plane. Determine the potential of the system if the
radius of the projection is R.
34. At a distance d from the centre of a conducting groun-
ded sphere of radius R an electric dipole p is placed whose
positive charge is closest to the sphere. Find the electric
potential of this system.
35. An electric charge e is placed at a distance d from
the flat surface of an infinite dielectric with a permittivity
8 2 • The permittivity of the medium where the charge is
located is 8 1 • Determine the potential cp and the electric
displacement vector D in the two media.
36. The centre of a conducting sphere of radius R. is on
the flat boundary between two dielectrics with permittivi-
ties 8 1 and 8 2 . The charge of the sphere is e. Find the poten-
tial of this system, the electric displacement vector, and
the charge distribution on the surface of the sphere.
37. Solve Problem 31 by an expansion of the potential
in a series of solutions of the Laplace equation in spherical
coordinates. Determine the surface charge density and the
total charge induced on the sphere.
38. A point charge e is placed at a distance d from a con-
ducting sphere of radius R that has a potential V. Find
the potential outside the sphere and the surface charge
density on the sphere.
39. Determine the potential of a charged sphere of ra-
dius R. The surface charge density varies according to the
law 0' = O'o cos e.
40. Determine the potential and the electric field vector
of a uniformly polarized ball of radius R. The polarization
vector of the ball is P.
SECTION II. ELECTRODYNAMlCS (;;,
41. A conducting sphere of radius R is located in a Iwu-

uniform electrostatic field. Determine the potential around
the sphere.
42. A conducting sphere thai is grounded is located in
a uniform electric field E 0 • Find the potential of this system
and the surface charge density of the sphere.
43. A sphere of radius R made of a dielectric material
is located in a uniform electric field E 0 • Determine the
potential inside and outside the sphere.
44. Consider a sphere that is electrically charged with
a surface density a everywhere except for a spherical seg-
ment near the pole. Determine the potential inside and outside
the sphere if the segment is limited by a circle withe = ex.
45. One face of a rectangular parallelepiped has a po-
tential V. All the rest have a zero potential. Find the poten-
tial inside the parallelepiped.
46. Two opposite faces, z = 0 and z = c, of a rectan-
gular parallelepiped with edges a, b, and c, have potentials
V1 and V2 , respectively. The rest are grounded. Find the
potential inside the parallelepiped.
47. A disc of radius R that has a surface charge density a
is placed coaxially in a hollow cylinder of radius r0 with
conducting walls. Find the potential inside the cylinder.
Consider the limiting case of a point charge inside the
cylinder, i.e. R -+ 0 but the charge of the disc remains
finite, or :rr.R 2 a = e.
48. Find the potential of a field that is produced by
a point charge e and a homogeneous plane-parallel lamina
that is at a distance d from the charge. The thickness of
the lamina is a and its permittivity is e. Consider the parti-
cular case of a point charge on the surface of a semi-infmite
crystal, and compare the obtained solution with that of
Problem 4 7.
49. Find the quadrupole moment of an ellipsoid that
is uniformly charged and has a body charge density p.
50. Determine the potential of an electric field produced
by a point charge e that is placed in a homogeneous and
anisotropic medium with a given permittivity tensor.
51. Find the. electric field intensity vector inside an
anisotropic lamina made of a dielectric and placed in a
uniform field E 0 •
5-01496
52. Calculate the energy of the interaction between the

electron cloud of a hydrogen atom and the proton (which
is the nucleus of the hydrogen atom). The charge density
inside the cloud is p = ~ e- 2 rfa, where a is the Bohr
na~
radius.
53. Calculate the energy of the interaction between two
balls whose distribution of charges e1 and e2 is spherically
symmetric. The distance between the centres of the balls is a.
54. Consider a sphere that sinks into a liquid to a depth
less than half its diameter when it is not charged. What
should be the charge of the sphere so that it sinks to a depth
exactly half its diameter? The mass of the sphere is M,
its radius is R, and the liquid has a density f.! and a permit-
tivity s.
Direct-current electricity. Magnetostatics.

Quasi-stationary phenomena
55. The plates of a spherical capacitor, which are separa-
ted by a conducting medium with a conductivity cr*, have
potentials cp 1 and cp 2 • Calculate the current across the capa-
citor and the resistance of the spherical layer. The radiuses
of the plates are r 1 and r 2 •
56. Find the law of refraction of the lines of current at
a flat boundary between two conducting media with conduc-
tivities cr~ and cr;.
57. The potential difference between two flat electrodes
is V and the distance between them is d. The field-induced
emission of electrons from one of the electrodes continues
until a space charge forms between the electrodes, which
opposes the external field. Find the relationship between
the current density and the potential difference applied to
the electrodes.
58. Find the magnetic field strength inside and outside
a cylindrical conductor with an electric current whose den-
sity j is the same through any section of the conductor. The
radius of the cylinder is R.
59. Find the magnetic f1eld strength inside a cylindrical
cavity in a cylindrical conductor with an electric current
whose density j is the same through any section of the con-
SECTION 11. ELECTRODYNAMICS 67
ductor. The axes of theca vity and the conductor are varallel
and separated by a distance a.
60. An electric current flows through an infinitely loug
conductor of radius R. The current density is alp for p:::;;; R,
where R is the radius of the conductor and p is the distance
from the axis of the conductor. Find the vector potential
and the magnetic field strength inside and outside the con-
ductor.
61. Find the magnetic field strength of a plane with
a surface current of density i that is the same in any point
of the plane.
62. Surface currents with densities i tlow along two paral-
lel planes. Find the magnetic field strength when the cur-
rents flow (a) in the same direction and (b) in opposite
directions.
63. Consider a strip of infmite length and width a made
of a conducting material. A current with a surface density i
flows uniformly through the strip. Find the magnetic field.
Consider the limiting case as the width of the strip tends
to infinity and compare the obtained result with that of
Problem 61.
64. Electric currents I tlow in opposite directions along
two straight, parallel conductors of infinite length placed
apart at a distance d. Determine the vector potential of the
system.
65. Find the vector potential and the magnetic field
strength created by a current I flowing along a ring of
radius R. Examine the special case when the observation
point lies on the axis of the ring.
66. Find the magnetic field strength and the vector of
magnetic induction created by a uniformly magnetized
ball. The radius of the ball is R and the magnetization vec-
tor is M.
67. Determine the magnetic field strength on the axis
of a magnet of cylindrical shape. The radius of the magnet
is R, its length is d, and the magnetization is Jill 0 •
68. Find the magnetic moment of an electrically charged
ball that rotates uniformly with an angular frequency Q.
The charge e is uniformly distributed over the volume of
the; b:~.ll. Show that the gyromagnetic ratio for this system
is e/(2m), where m is the mass of the ball.
5*
68 PROBLEMS TN THRORETTCAL PHYSICS
H9. A sphere of radius R rotates about the z-axis with

au angular velocity &~. Its surface is electrically charged
with a density a. Find the vector potential and the magnetic
field strength inside and outside the sphere.
70. Calculate the force between two straight parallel
wires of infinite length with electric currents / 1 and I 2 if
the distance between the wires is d. The permeability of
the medium between the conductors is I-t·
71. Calculate the force between two coaxial wire loops
of radiuses R 1 and R 2 with electric currents flowing in the
same direction. The distance between the centres of the
loops is d, and both are placed in a medium with a per-
meability 1-t·
72. Find the self-inductance L per unit length of a
transmission line that consists of two coaxial cylinders of
radiuses R 1 and R 2 (R 1 < R 2). The space between the con-
ductors is filled with a substance having a permeability I-t·
73. A solid conductor of radius R 1 with a permeability
r-t1· is placed inside a cylinder of radius R 2 • The space between
the conductor and the cylinder is filled with a substance
having a permeability r-t 2 • Determine the self-inductance
per unit length of this contour.
74. The diamagnetic susceptibility per unit volume is
Ze2N -
X= - fi'm r-torz
where Ze is the electric charge of the atomic nucleus, e
and m are the electron charge and electron mass, N is the
rz
number of atoms per unit volume, and is the mean square
of the radius of the atom, i.e.
r2 = L j r p (r) d't
2
where p (r) is the charge density in the atom. Determine

x for atomic hydrogen. In this case
p (r) = 1(:3 e-2rfa
where a = 0.528 X 10-10 m {the Bohr radius).
75. Show that for a constant and uniform magnetic
f1eld B the vector potential can be chosen in the form A =
= 21 [B X rl.
76. Find the distribution of the electric and magnetic

fields inside a cylindrical conductor which carries an alter-
nating current with frequency w. The electrical conductivity
of the conductor is cr*.
77. Consider a specimen of magnetic material subjected
to a constant magnetic field of intensity H 0 • Under the
influence of this field a magnetization M 0 parallel to the
field appears inside the specimen. In addition to the con-
stant field a variable magnetic field that is perpendicular
to Ho and that rotates with an angular velocity w is imposed
on the specimen. The amplitude h of the intensity of this
variable field satisfies the condition h ~ H 0 • Determine the
additional magnetization caused by the variable field and
find the conditions for magnetic resonance.
78. A ball made of magnetic material is inserted in a
constant magnetic field. The intensity of the field inside
the ball is H 0 • If we assume the size of the ball to be conside-
rably less than the wavelength of natural oscillations of the
magnetic moment, what will be the frequency of these oscil-
lations?
79. A sample of infinite length of nonconducting magne-
tic material is subjected to a constant magnetic field of
intensity H 0 • Find the frequency of natural oscillations if
the wavelength of these oscillations is considerably less
than the wavelength of electromagnetic waves (the magneto-
static approximation).
80. Determine in the magnetostatic approximation the
natural frequency of vibrations of a plate made of magnetic
material and having a metallic covering. The thickness of
the plate is d, the intensity of the constant magnetic field
is H 0 , the plate is magnetized normally to its surface, and
the covering is ideal.
81. Determine the natural frequency of vibrations of a
plate made of a magnetic material and magnetized normally
to its surface. The plate is placed in a vacuum, its thickness
is d, and the intensity of the constant magnetic field is H 0 •
Use the results of Problem 80.
82. Find the natural frequencies of 1wo coupled circuits
if the self-inductances are L 1 and L 2 , the mutual inductance
is £ 12 , the capacitances of the circuits are C1 and C ~· and all
resistances in the circuits are zero.
Propagation of electromagnetic waves.

Wave guides. Resonators.
Magnetohydrodynamics
83. Two plane, monochromatic waves are linearly-polar-

ized in perpendicular directions. Determine the polariza-
tion of the resulting wave if both waves propagate in the
same direction, their frequencies are the same, the ampli-
tudes of the waves are E 01 and E 02 respectively, and the
phase difference for the waves is cp.
84. Determine the damping of electromagnetic radiation
in a medium in total internal reflection.
85. A plane-polarized wave strikes the surface of a nonmag-
netic material normally to this surface. The permittivity
of the material is e and its electrical conductivity is a*. Find
the reflectance R. Consider the limiting case of an ideal
conductor.
86. Determine the amplitudes of a reflected and a refrac-
ted wave in a plane-parallel plate. The thickness of the
plate is d and its permittivity is e. Find the conditions for
the reflection of electromagnetic waves from the plate
to be minimal.
87. A surface wave whose magnetic field strength is per-
pendicular to its line of propagation (a TM wave) travels
along the interface of two dielectrics. The permittivities
of the dielectrics are opposite in sign and equal to e1 and
- I e 2 I· Find the dispersion equation.
88. An electromagnetic wave falls onto a plane surface
of a semi-infmite crystal at an angle e1 . Determine the
directions of propagation of the ordinary and extraordinary
rays in the crystal. The optic axis of the crystal is perpendi-
cular to its surface.
89. Assume that a medium consists of elastically coupled
charged pi1rticles and that the force constants are different
for mutually perpendicular directions. Find the permittivity
tensor of the medium if the volume concentration of the
particles is N.
90. Show that if E (w) is an analytic function in the
upper half of the complex u) plane and approaches unity as
w 1 tends to infinity, the real and imaginary parts of
~~ ((t)) satisfy the following equations:
-00
Ime ()w = - - PJ
00
1 Ree(w 1 )-1 d 1
:n: (J)
1
-(J)
w
-oo
where P means the principal value of the integral.

91. A medium consisting of elastically coupled electrons
has a permittivity equal to
e (w) = 1 + me
Ne2
0
~ fk
- ' w~-w2-i1'kW
k
where N is the volume concentration of the electrons, e

and m are the electron charge and mass respectively, and fk
and wk are constants. Show that this equality satisfies the
dispersion relations in Problem 90.
92. Assume that a substance consisting of quasi-elastic-
ally coupled electrons of volume concentration N is placed
in a homogeneous magnetic field with an induction B 0 • ·
A linearly-polarized monochromatic light wave whose wave
vector is parallel to the magnetic field falls onto the sub-
stance. Find the angle of rotation of the plane of polariza-
tion of the wave when it has travelled a distance l in the
substance.
93. A system of anharmonic charged oscillators is placed
in an alternating electromagnetic field. The potential energy
of an oscillator as a function of displacement is given by the
relation
U= + kr 2 +i S
3
i. j, 1=1
f3uzX;XJXz
Assuming that the factors ~iil are small, find the polariza-
tion vector up to terms linear in ~iil and quadratic in the
electric f1eld vector. Show that the polarization vector has
terms corresponding to oscillations with a frequency twice
the one of the incident wave. The mass of an oscillator ism,
its natural frequency is w0 , its electric charge is e, the volume
concentration of oscillators is N, and the frequency of the

external fi.eld is (J).
94. Find the nonlinear polarization of the system of oscil-
lators of Problem 93 if the external fi.eld consists of two
monochromatic waves with frequencies ffi 1 and ffi 2 •
95. Assume that a medium consists of parallel chains
of evenly distributed oscillators with natural frequencies
ffio. The neighbouring oscillators in a chain interact via
the Hooke's law. Examine the propagation of an electro-
magnetic wave along the chains when the wavelength is
much greater than the distance between the oscillators.
Find the index of refraction for such waves.
96. Examine the propagation of electromagnetic waves
in the space between two conducting plates separated by
a dielectric medium. The distance between the plates is d and
the permittivity of the medium is e.
97. Find the intensity, dispersion equation, and fre-
quency limit for the TE and TM waves in a rectangular
w.ave guide with ideally conducting walls. The dimensions
of the wave guide are a and b.
~98. What is the relation between the tangential compon-
ents of an electric and a magnetic fi.eld near a conductor?
(?'99. Determine the damping of TM waves in a rectangular
wave guide with dimensions a and b. The conductivity of
the walls is a* and their permeability is fl.
100. 'Find the dispersion equation for electromagnetic
waves in a cylindrical wave guide with ideally conducting
walls. The radius of the guide is R.
101. Examine the propagation of electromagnetic waves
along a cylindrical wave guide made of a dielectrical mate-
rial with permittivity e. The radius of the guide is R.
102. Examine the propa~ation of electromagnetic waves
in a dielectric medium that fi.lls the space between two condu-
cting coaxial"cylinders. The cylinders have radiusesR 1 and R 2 •
103. Determine the electric fi.eld vector and the natural
frequencies of TM waves in a rectangular cavity resonator.
The dimensions of the resonator are a, b, and c, and its
walls are ideal conductors.
104. Show that the numher of oscillations in a frequency
L 1L 2 L 3w2f1w
interval 11(•) is n 2 c3 for a rectangular resonator of
dimensions L 1 , L 2 , and L 3 and with ideally conducting

walls.
105. Determine the electromagnetic f1eld and natural
frequencies of electromagnetic waves in a cylindrical reso-
nator of radius R. The distance between the ends of the
resonator is d.
106. A viscous, conducting. and incompressible liquid
moves between two parallel infinite planes. A constant mag-
netic field Ho is directed perpendicularly to the planes.
Determine the distribution of velocities in the liquid if it
is in stationary flow. The distance between the planes is d.
107. A viscous, conducting, and incompressible liquid
is placed between two conducting planes z = 0 and z = d.
The plane z = d moves along the x-axis with a velocity v 0 •
A uniform magnetic field H 0 is'"'Cdirected along the z-axis,
and an electric field E0 along the y-axis. Determine the
distribution of velocities in the liquid.
108. An ionized plasma consists of ions and electrons.
Assume the charge density fluctuations to be small. Exam-
ine the variation of charge density (concentration) and find
the plasma frequency if this frequency is so large that the
ions are not able to catch up with the fwld and stay fixed.
Magnetic interactions and pressure gradients are negligible.
The radiation and scattering

of electromagnetic waves
109. Obtain the equations for the potentials created by
charges and currents in a vacuum provided div A = 0 (-the
Coulomb gauge condition).
110. An electric current of density
j(r, t)= I sin { k: -k 1z I) 8 (x) 8 (y) e 3e-iror
is generated in a thin linear antenna. Find the time average

(over the period) of the antenna's intensity of the radiation
in a unit solid angle. Consider the particular case when the
antenna length dis several Jwlf-waves. Vector e 3 is rlirected
along the antenna.
111. Find the total radiation intensity of a linear antenna
with a wave traveling from point z = -l/2 to point z =
= l/2, where it is absorbed completely, i.e. without reflection.

The electric current in the wave is / 0 cos (rot - kz).
112. Examine the propagation of plane waves inside an
isotropic and homogeneous dielectric, assuming that each
volume element radiates a spherical wave which propagates
with a velocity c. Determine the velocity of the plane waves
in the medium.
113. A plane electromagnetic wave falls on a conducting
cylinder of radius R. The cylinder's axis is perpendicular
to the wave vector and parallel to the magnetic field. Find
the electric and magnetic fields in the scattered wave. Cal-
culate the energy flux scattered by a unit length of the
cylinder.
114. Show that when two identical particles collide,
there is no dipole radiation.
115. Find the intensity of radiation of a particle of mass m
moving in a circular orbit of radius a under Coulomb forces.
Express the answer in terms of the particle's energy.
·116. Determine the time it will take a particle moving
in a circular orbit to "fall" on a charged centre because of
the energy loss through radiation.
117. An atom radiates electromagnetic waves and stays
in an excited state during time T. The time dependence of
the atom's electric field is
- !_ -iwot
E (t) = E 0e -r
Determine the width of the line radiated by the atom.

118. Find the differential cross section for the scattering
of an elliptically polarized wave of frequency ro by an
oscillator. The natural frequency of the oscillator is ro 0 ,
its mass is m, its charge is e, and its damping coefficient
is y.
119. Examine the electromagnetic field created by the
plane z = 0 with a surface eurrent of density
120. Consider a dipole with moment p lhat vibrates

with a frequency <•) in the origin of a coordinate system.
A particle with a polarizability ~ is placed at a p,oint with
a radius vector d (d _l p). Find the radiation intensity of

electromagnetic waves for the system provided d ~ 'A,
where 'A is the radiation's wavelength.
121. A dipole of moment pis located at a distanced from
an infinite conducting plane and vibrates with a frequen-
ey (•J. Determine the damping of the vibration if d ~ 'A,
where 'A is the wavelength of the system's radiation. The
conductivity of the material of the plane is a*.
122. A dipole of moment p vibrates with a frequency ro
in a dielectric medium (of permittivity e1) at a distance d
from the flat interface with another dielectric medium (of
permittivity e 2 ). Find the electric and magnetic fields of
the waves radiated by the dipole. In all calculations d?;> 'A.
123. A particle of charge e moves with a velocity v and
then elastically bounces off a plane. Determine the long-
wave part of the radiation spectrum at the moment of
impact.
Special theory of relativity.

Relativistic electrodynamics
124. Show that two successive Lorentz transformations
in the same direction commute and are equivalent to one
Lorentz transformation.
125. A plane electromagnetic wave propagates in a me-
dium that moves with a velocity v relative to a reference
frame K. Find the velocity of the wave in frame K if the
index of refraction of thf' medium is n. Examine the case
when v~ c.
126. Find the approximate relationship between the ener-
gy of a slow particle and its momentum up to members
proportional to ( :z: f.
2 For slow particles p 2 ~ m2c2 .
127. Find the path of a charged particle moving in a
uniform electric field E 0 • Examine the limiting case of a
slow particle.
128. Find the path of a ehargerl particle moving in n
uniform magnetic field H 0 .
129. Find the path of a relativistic particle of charge e1
nnd mass m in the field of a fixed point charge e2 •
130. A particle of mass M decays into two particles of
masses m1 and m 2 . Find the energies of the oecay products

in the centre-of-mass system.
131. Find the kinetic energies of a f!-meson (rest energy
105.7 MeV) and a neutrino that were produced in the decay
of a fixed :n:-meson (rest energy 139.6 MeV). Use the results
of Problem 130.
132. Find the Lorentz equations for transforming the
energy and the components of momentum of a particle from
one reference frame to another frame moving with a velocity
v relative to the first.
133. A particle moving with a velocity v decays into
two particles whose energies in the centre-of-mass system
are E 1 and E 2 . Fino the relationship between the angle of
emergence and the energies of the products of decay in the
laboratory frame of reference.
134. Find the relationship between the directions of the
velocity of a particle in two reference frames that move
with a relative velocity v.
· 135. Two particles with rest masses m1 and m 2 and ener-
gies E 1 and E 2 collide elastically. If before the collision the
second particle is at rest find the scattering angles in the
laboratory frame, as functions of the energies E; and E~
after the collision.
136. Determine the relationship between the frequency
of a photon scattered by a stationary free electron and
the scattering angle (the Compton effect).
137. Find the frequency of the photon emitted by an
atom in an excited state. The excitation energy of the atom
is !lE, its mass is m, and the atom is stationary.
138. Show that the annihilation of an electron-positron
pair into one photon is prohibited by the conservation law
of 4-momentum.
139. Consider a mirror that moves with a velocity v
in the direction opposite to its normal. A ray of light falls
on the mirror at an angle e. Determine the angle of reflection
and the frequency shift of the reflected wave.
140. Prove that if vectors E and H are perpendicular to
each other in one inertial reference frame, they are perprn-
dicular in any other inertial frame.
141. Find an inertial reference frame in which the electric
ond magnetic fields are parallel.
SECTTON II. RLECTRODYNAMTCS 77
142. Find lhe potentials of a uniformly 1110\·ing L:harge

by using th<' relativistic transforma1ion of the static Cou-
lomb field.
143. Find tho Lorentz equations for transforming the
eomponents of the energy-momentum tensor.
144. Find the Lorentz equations for transforming the
components of the electric and magnetic fields to a reference
frame that moves with a velocity v.
145. Show that the wave equation is not invariaul under
the Galilean transformations but is invariant under the
Lorentz transformations.
146. Show that E 2 - H 2 is invariant under the Lorentz
transformations.
147. Show thal if a magnetic moment I-t moves with a
velocity v (v ~c), this gives rise to an electric dipole
1
moment p = c2. [v X ~-tL
SECTION III Quantum Mechanics
In quantum mechanics a state of a micro particle subjected

to a potential field V (r, t) is described by a complex-valued
wave function '¥ (r, t) which is determined from the time-
dependent Schrodinger equation
ih o'l' ~~· t) = - ~~: ~ '¥ (r, t) + V (r, t) '¥ (r, t) (III-1)
. where m is the particle's mass, and li is the Planck constant h

divided by 2n. If we multiply equation (III-1) by the
complex conjugate function'¥* (r, t) and an equation of type
(III-1) for'¥* (r, t) by'¥ (r, t), and then subtract the second
product from the first, we get the so-called equation of
continuity
:e 1'¥ (r, t) + div [
12 J
~~ ('¥grad'¥*-'¥* grad'¥) = 0 (Ill-2)
This equation is the differential form of the conservation
law for the quantity J
I '¥ 12 dT, which makes it possible
to interpret I '¥ (r, t) 12 as the probability density of finding

the particle at point r at time t, so I '¥ (r, t) 12 dT is the pro-
bability of finding the particle under consideration in the
volume element dT. Then vector
j = ;~ ('¥grad '¥* -- '¥* grad '¥) (III-3)
is the probability current density.

Such an interpretation of I '¥ 12 leads us to the conditions
which '¥ (r, t) must satisfy: the wave function must be
single-valued, finite in every point of space, and continuous
together with all its first derivatives. This last requirement
SECTION III. QUANTUM MECHANICS 79
(t!Je continuity of the ftrst derivatives) is violated only at

points whNP the potenth-.1 has a discontinuity of the spconrl
kind.
If the potential docs not depend explicitly on time, there
exists the so-called stationary solution of the Schrodinger
equation
- ..!_ Et
'I' (r, t) = 'P (r) e h
for which p = J 1¥ 2 and j are time-independent. Substi-

J
tuting this solution into equation (III-1), we come to the

time-independent Schrodinger equation
[- ;~ 11 + V (r) j 'IJ (r) = E'V (r) (III-4)
Generally speaking, the solutions of equation (III-4),
which must satisfy the conditions mentioned above, exist
only for certain values of parameter E, i.e. for certain values
of the energy of the micro particle. Sometimes there is a whole
range of values of E for which 'P is finite, single-valued, and
continuous (the case of a continuous spectrum of E). But
sometimes such a solution exists only for a discrete spectrum
of E. In still other cases the two spectra exist simultaneously.
For a discrete spectrum the wave function 'P can be nor-
malized to unity:
(III-5)
The eigenfunctions for states with different energies are

mutually orthogonal.
For a continuous spectrum the orthonormality condition
can be written with the help of the Dirac delta function:
j 'P* (r, E) 'P (r, E') d,; = c5 (E- E') (III-6)
where the 6-function, as always, is defined from the condi-

tion that for an arbitrary function f (x)
00
j f (x) 6 (x- a) dx = f (a)

-oo
In a variety of problems the solution of the time-indepen-

dent Schrodinger equation is sought in the form 'P = eiB(r)th
8() PROBLRMS TN THEORETICAL PHYSICS
(the Wentzel-Kramers-Hrillouiu approximation, usually cal-

led the WKB approximation), with S (r) satisfying the
equation
2~ (grad S)Z+ d~ 1'1S + V (r) =E (III-7)
By expanding S (r) in a series of powers of in

s = so + inS' + ...
we find, in the zeroth-order approximation, the time-irlde-
pendent Hamilton-Jacobi equation for S 0 , and the firRt- and
higher-order approximations give us small corrections to S 0 •
If the variables are separable, the solution of the Hamil-
ton-J acobi equation is of the form
S0 = 2} SHqi), and SHqi) = ) Pi dqi
where Pi is the canonical momentum conjugate to the gene-

ralized coordinate qi. In particular, for the one-dimensional
case,
S0 = j V2m[E-V(x)] dx (111-8)
If V (x) is a one-dimensional potential barrier, i.e. V (x) > E

in the region a :;:;;;; x :;:;;;; b, inside the barrier S 0 is imaginary,
and so¢ acquires the factor
exp [ - v;m j b
a
VV(x)-Edx]
If we define the transmittance of the barrier as the ratio

of the probability current densities of the transmitted wave
and the incident wave and use the WKB method, we can
get the expression
D=l;~::nsi=D0 exp[ -2 V;m Jyv(x)-Edx] (111-9)

a
where D 0 is a constant.
If a microparticle is in periodic motion in a potential
well, the condition that ljJ be a single-valued function in
SECTION III. QUANTUM MECHANICS £1
the zeroth-order WKB approximation brings liS to the

following formula:
(111-1 0)
where ni = 0, 1, 2,
These have come to be known as the Bohr-Sommerfeld
quantization rule. It was Sommerfeld who in 1916 formulated
the postulates of Bohr's theory of spectra in a way that
made it possible to select allowed orbits from the continuum
of classically possible orbits. The integral in (111-10) is taken
along a closed orbit.
According to Bohr, the electrons move in orbits restricted
by condition (III-10) and do not radiate in spite of their
acceleration. Radiation is emitted or absorbed when an
electron makes a discontinuous transition from one allowed
orbit to another, and the frequency of this radiation is
ffimn = (Em - En)//i (JII-11)
where Em and En are the corresponding energies of the elec-
tron on the orbits.
If we define the expectation value of the particle's momen-
tum via the probability current density as (p) = mj d-r, J
then substitute expression (III-3) for j, and integrate the
second member in the right-hand side by parts, we get
{p) = J'I'* (- in'V) 'I' d-r (III-12)
where we have used the fact that I 'Y 12 = 0 on the boundary
of the integration range. In a similar manner we can defme
the expectation value of the particle's position as (x) =
= J'Y*x'Y d-r.
Operators. In quantum mechanics the momentum of clas-
sical physics is represented by an operator p = -iliV, and
position x by an operator = x, x
(p) = J'Y*p'I' d-r and {x) = J'Y*x'Y d-r
In the case of an operator representing a physical obser-
vable F (p, r) which is a function of both momentum a.Pd
fi-01496
position, we construct the same function but with operators

rand p instead of rand p:
f (p, r) = F ("p, r) = F (-ihV, r) (111-13)
For a series of measurements of an observable F, the
expectation value of this observable in a state '¥ (r, t)
is
(F)= ~ '¥* (r, t) F'¥ (r, t) d-r: (III-14)
where the wave function'¥ (r, t) is a vector in a Hilbert space
and satisfies the orthonormality condition
) w• (r, t) '¥ (r, t) d't= 1
The hermitian conjugate of the operator F, denoted .fr+

is defined by the relation '
.\' 'Y!F'¥ 2 rh= 1'¥z(F+'f 1 )*d-r: (111-15)
where 1!''1 and '¥ 2 are any two functions from a Hilbert space.
The operator representing an observable must be her-
mitian (self-conjugate) and linear, i.e. satisfy the following
conditions
(111-16)
(111-17)
For the case when F is the product of two observables A

and B, F = AB, condition (111-16) is satisfied only if
AB = BA. But if the two operators A and B do not com-
mute, the observable F is represented by the operator F =
1 ~ ~- ~ ~
=2:(AB +BA).
According to definition (111-13), we can introduce the
following energy operators: the kinetic energy operator
~ ~ A
T = Zm, the potential energy operator V, and the total

energy operator (the Hamiltonian) fi. Explicitly,
T= - ~~ !1, fr = v (r), H = T + fr = - ;~ t1 + v (r)
(111-18)
SECTION III. QUANTUM MECHANTCS 83
The three components of the angular momentum and its

square arc represented by tho operators
L:c=YPz-ZPu· .Ly=ziix-;Pz• Lz=xpu-iiiix.
f_z =L~+L~ +L; (III-19)
Let us write Lz and i) in spherical coordinates. We have
~ . a
Lz = -ln-
a<p
L~ 2 -_ ~2A ~2 [ 1 a ( . 8 a ) + 1 az J
_
- " L1S<p- - " -sine as sm ae sin2 e acpz
(III-20)
If we define the variance of the distribution of A about the
expectation value (A) as
(AA2 ) =((A- (A) ) 2) = J'Jl* (L- (A) ) 2 1Jl dT
and then use the relationship (111-15) for L- (A), we

can show that
(AA2) = J
I (L- (A)) 'P 12 d•
Thus, if the variance of the distribution of A in a state with
the wave function 1jJ is zero, i.e. if (A/,2 ) = 0, the condition
(III-21)
holds true.
Hence, the state with the wave function must be described
by an eigenfunction of L which satisfies condition (111-21)
and is finite, single-valued, and continuous.
The scalars An, for which 'i'n satisfy the above mentioned
conditions, are called eigenvalues of L. A specific eigenvalue
An is nondegencrate or a-fold degenerate depending on
whether one or a linearly independent eigenfunctions cor-
respond to it.
The capability of operators to commute has important
physical significance. If LM - M.L = 0, i.e. if the result
of LM- ML acting on an arbitrary function \jJ is zero,
operators L and M have common eigenvalues and, hence,
the corresponding observables A and f.L can simultaneously
6*
84 I'ROHLEMS TN THEORETICAL PHYSICS
assume exact values. Otherwise, if LM - ML =f= 0, we

have the uncertainty relations for 'A and 1-l·
For example, if a particle is in a spherically symmetric
field V (r), the Hamiltonian ii commutes with L2 and Lz,
and its eigenfunctions can be sought in the form
lJl (r, El, cp) = R (r) Y (8, cp) (111-22)
where Y satisfies the following equations:
-1i 2 ~ 8 (j)Y = A.Y and -in aY = aY
acp
(It is easy to see that I) and Lz always commute.) The solu-
tions of these equations that satisfy all the conditions for
eigenfunctions are the spherical harmonics
Yzm (El, cp) = Pzm (cos 8) eimcp (111-23)
From this it follows that 'A = li 2 l (l +
1) and a = lim, where
l = 0, 1, 2, ... and m = 0, +1, +2, ... , +l; Pzm (x)
are the associated Legendre polynomials of degree l and
order m.
For an electron in the Coulomb field of a nucleus with
a charge Ze, the radial part of function (III-22) assumes the
form
n-l- J
lfnl (p) = e-Pfnpl ~ bkph (Ill-24)
k=O
where the coefficients bk satisfy the recurrence relations
2 k+l+i-n
bk+t = bk n (k+l) (k+l+1)-l (l+1) (III-24a)
r f1.2
Here the dimensionless variable p = -,
a
where a = Z~te2'
-
and l assumes the values 0, 1, ... , n - 1.
Z2f.te4
Each eigenvalue of the energy of the electron En = - 2n. 2 n2
has corresponding to it n 2 eigenfunctions of type (111-22),
bearing in mind (111-23) and (111-24), n = 1, 2, . . . .
Integrals of motion. If we find the time derivative of ex-
pression (111-14) and use the time-dependent Schrodinger
equation fir¥ c.= i1i aa;' we can show that the time derivative
of an observable "A has corresponding to it the operator

al = al +_!:_(ilL- Lii) (111-25)
dt . at n.
~
When dl == 0, "A is an integral of motion. If fi and -
L
commute, L does not depend explicitly on time.
Applying equation (111-25), we can find the quantum mecha-
. 1 l aws of mo t.IOn, 1.e.
mea . t .· h e opera t ors Cit
dp; an d dt
d;; .
The eigenfunctions of a hermitian operator form a com-

plete and orthonormalized set of functions, i.e.
~ 1jl~ (r) 1Jln (r) d-r: = «'lmn
for a discrete spectrum of eigenvalues,
) 1Jl* (r, "A) 1jJ (r, "A') d-r = 6 ("A- "A')
for a continuous spectrum, and an arbitrary wave function
can be expanded in terms of the eigenfunctions of L:
A2
1jJ (r) = ~ Cn,~n (r) + \ c ("A) 1jJ (r, "A) d"A (111-26)
n "
At
If we use the above-mentioned orthonormality conditions,

we find that
Cn = ) 1jl~ (r) 1jJ (r) d-r and c ("A)= J1Jl* (r, "A) 1jJ (r) d-r
(III-27)
where
./.2
n 1.1
J
~ I Cn 12 + IC ("A) 12 d"A = 1
From the equation

/.2
(A)= ~ I !
Cn 2 An+ JA IC (lo) 12 dlo (III-28)
" A!
it follows that I en 12 and I c ("A) 12 dA are, respectively, the

probabilities of finding in state lJl (r) the eigenvalue "A = }.n
(in the discrete spectrum) or the eigenvalue in the limits

from "A to "A+ d"A (in the continuous spectrum).
Theory of representations. Matrices. Expansion (III-28)
replaces the wave function '¢ (r) by the coefficients en or
c (/...), and value x by value "A, i.e. there is a transition to a
new representation, the "A-representation. If in the coordi-
nate representation the operator if links functions cn
n
where (k I M I n) = j '¢1;llf'¢n d-r: are the matrix elements
of M. In the transition to the I.-representation a matrix has
replaced the operator.
_An operator L in its own representation reduces to a dia-
gonal matrix since
(k I L I n) = j' 'ljl~L1Jln ill:= An 6hn
with the diagonal elements as the eigenvalues. This holds
true for an operator with a discrete spectrum of eigenvalues.
An operator with a continuous spectrum in its own repre-
sentation reduces to a multiplier equal to the independent
variable (in the momentum representation equals p). p
It is evident that for a hermitian operator M the matrix
elements satisfy the following relationship:
<k I if 1 n > = <n I M I k >*
A particle in an electromagnetic field. In order to gene-
ralize the Schrodinger equation for the motion of a particle
in an electromagnetic field specified by an electric field
vector E and a vector~ of magnetic induction B, which,
in turn, can be defined via potentials A (r, t) and cp (r, t) as
8A
B== curl A, E= --grad m
·r
--
at '
we must replace the momentum operator in the Hamilto-
nian if by the operator p - eA and write the potential ener-
gy as e<p, i.e. in the presence of an electromagnetic field

1
+
A A
H. 2m (p- eA) 2 e<p

(This agrees with the classical case of a particle in a magne-
tic field.) Choosing the vector potential so that divA =0
und neglecting the term e2A 2 as being small, we find the
wave equation of quantum mechanics:
- ;: 11 '¥ -+ itz : (A· grad 'I')+ e<p '¥ = i1i a: (III-29)
Spin. The experiments of Stern anrl Gerlach and also the
rloubling of the energy levels Enz of an electron in an atom
show that apart from having an orbital angular momentum
L = [r X p] (we have mentioned this fact before), it has a
spin angular momentum, or simply spin. The components of
this spin momentum, S "'' S 11 , and S z• can assume only two
distinct values, +1il2 and ---n/2.
If we introduce the dimensionless spin operators Oxa
A A "; hA A nA A /iA
cry, az so that 8x=z-ax, Su=z-ay, Sz='f: az and in

such a way that the operators satisfy the anticommutation
relations crxriu = - cryrix = icr2 , (etc. cyclic), we can represent
these operators by the so-called Pauli matrices. In the crz-
represen ta tion
f1x=
A (01 01) ' .· (0 - 0,i)
f1y = i I
A
az =
( 1
O _
0\
t}
(111-30)
The wave function will then depend on the spin variable

a = +1. This dependence is made explicit by using a two-
component matrix
1jJ (r' a) =
( 'Pt(r))
'P2 (r)
(the spin variable assumes only two values). The two quan-
tities I ljl1 (r) 12 and I ljl 2 (r) 12 are interpreted as the proba-
bility densities of finding thP electron at point r, with
a = -!--1 (called "spin up") or a = -1 (called "spin down"),
i.e. with S z = +1il2 or S z = -n/2.
If it is possible to neglect the coupling between the elec-
tron's orbital angular momentum and the spin, we can
separate the position and spin variables, "ljJ (r, a) =

= "ljlo (r) X (a).
Since the electron has an intrinsic magnetic dipole mo-
ment by virtue of its spin, the operator for this moment
being ~ = ; : o,
the electron interacts with any external
magnetic field, and the energy operator of this interaction
is -(F-t ·B), where B is the vector of magnetic induction of
the field. We thus come to the Pauli equation, which ac-
counts for the electron spin:
1 ~ ~
2m (p- eA) 2 'I' (r, a, t)- (F-t ·B) 'I' (r, a, t)

+ecp"lf(r, a, t)=in a'Y(rata, t) (III-31)
This equation can be made simpler if we write out the first
term explicitly and proceed as we did in evaluating
equation (III-29).
We can write the time-independent equation in the same
way. ·
Approximate methods of solving quantum mechanical
problems. When the equation for the eigenfunctions and
eigenvalues of an operator (for instance, of the Hamilto-
nian) cannot be solved exactly, we can use approximate
methods. One of these is the time-independent perturbation
theory.
Suppose we know the eigenfunctions and the eigenvalues
for a Hamiltonian H0 :
lf01p?, = E~1p~
and we ask for the eigenfunctions and eigenvalues for the
Hamiltonian fi = H0 + W:
H1fh = Eh"ljll<
in the form of an expansion in terms of the eigenfunctions
In thr eq11alion for the coefficients Cnn
chl (E?- Eh) + .L;n {lj WIn} Chn = 0

SECTION III. QUANTUM MECHANICS l:\9
we must express the quantities ckl and Ek as power series in

a small parameter that enters the perturbation energy ope-
rator: W = eV1 , where e is small. Using the method of
successive approximations, we get the wave function and
the energy level:
rh. 'Yk
+ L.J
,h = ,ho ~ (n I WI k) ,ho
o o 'Yn
+•··
E~t-En
notk
Ek = Eg + (k IWI k) + ~ I (n I~ I k~ 12
E~t-En
+ ... (111-32)
notk
provided the corresponding energy level of the unperturbed
system is nondegenerate.;
If the level E~ is a.-fold! degenerate, the zeroth-order
approximation of the wave function of the perturbed system
is expressed as a linear combination of the degenerate
eigenfunctions of if 0 , which correspond to the same energy
level E~:
a.
'Ph = ~ C~'Jlgi3
13=1
where
~ 0 0 0
Ho'Jlw=E~t'Jlki3 (~=1,
2, ... , a)
For the parameters Cfl we get a system of a homogeneous
linear equations
(Eg-E+<kt~JWJk~))cll+ ~ (krdWJk..,>c..,=O
YoFfl
where ~. y = 1, 2, ... , a. Nontrivial solutions for this
system exist if the system determinant is zero, or
I (Eg- E) ~'lily+ (kfll w Ik.,} I= 0(~. y = 1' 2, ... , a.)
(111-33)
Generally speaking, this equation defines a values of E,
and the foregoing equations make it possible to calculate
the coefficients ell and, hence, the wave functions ~'k·
Another approximate method of solving the eigenvalue
equation is the Ritz variational method, based on the varia-
tional principle of quantum mechanics. It can he shown that
for the eigenfun~tion of if for the gronnd state, ~1, where
H~·· = A.¢ anrl ) I t[• 12 d-r = 1, the functional (H) [ tp 1 =
= ~ '¢* H'¢ dT is either maximized or minimized. Whence,

if we calculate (H) [cp] = ~ cp* ilcp dT for a trial function
cp = cp (r, A, a, ... ) that satisfies the normalization con-
dition \ I {p 12 dT = 1, by adjusting parameters A, a, ...
we can ~alculate the upper bound of <H) on E 0 (the ground-
state energy).
To solve the eigenvalue problem for a system of n particles
we must use the Schrodinger equation with a Hamiltonian
n
~
ll= n
--
2 ~ - t'l.h f-V(r 17 r 2 , ••• , rn)
2 mh
k=1
which is the sum of the kinetic energy operators of the par-

ticles and their potential energy operator. For the case of
a system of identical particles we can state that since the
permutation of any two particles leaves the expressions of
type 1'ljJ 12 and ~ 'ljl*L'Ijl dT unchanged, the wave functions
'ljJ (r1 , cr 1 , r 2, cr 2) and 'ljJ (r 2, cr 2, r 1 , cr 1 ) can differ only by
a factor eia:. If we permute these particles a second time, we
find that
'ljJ (r 2 , cr 2 , r 1 , cr 1 ) = ~' (r1 , a 1 , r 2 , a 2 ) eia
= 'ljJ (r2, cr2, rl, al) e2ia:
or eia: = +1. Thus, a system of identical particles may be

described either by symmetric functions (eia: = +1 for
particles with integral spin) or by antisymmetric functions
(eia: = -1 for particlesJ with half-integral spin). This is
valid for both spatial and spin variables.
If we introduce the spin wave functions (a (a) and ~ (cr))
for a single electron
a(+1)=1, a(-1)=0
and
~(+1)=0, ~(-1)=1
that is,
a (a)= ~O ,
11) ~(a)=(~) '
we find that for a system of two electrons whose spins do

not interact, the spin wave functions are composed of such
products as rx 1 rx 2 , rx 1 ~·2 , etc., where rx 1 belongs to one elec-
tron, a 2 to the other, etc.
For the case of an external perturbation TfT (r, t) causing
an unperturbed system (whose Hamiltonian is ii 0) to make
a transition from a stationary state¢~ to another stationary
state ¢Y, we must use the time-dependent perturbation theo-
ry (the theory of quantum transitions).
Denote the eigenfunctions of fl 0 by ¢~. which are the
solutions to ii 0 ¢~ = E~,¢~. The problem is, given the com-
plete set of¢~. to find the solution of the Schrodinger equa-
tion
aqr ~ ~
in 7ft= (H 0 +TV)'¥
in the form of an expansion

'¥ = h Cn (t) 'ljJ~e-iEgt/h
"
As an initial condition at t = 0 we take the system to
be in a particular state ¢t so that 'I' (0) = '~t that is,
Cn (0) = l>nk· The differential equation for Cn (t) is
itt a;: = s
n
Win (t) Cn (t) eiroznt
where W1n (t) = j'

¢Y*W (r, t) \jJ~ do: is a matrix element of
the perturbation energy operator, and ffi In = (EY - E~)(li.
To solve this equation to a first approximation we must
substitute Cn (0) for en (t) in the right-hand side and also
assume that W (r, t) = 0 for t < 0 and t > T, where T
is the time of perturbation. Then
T
C[ (T) = i~ J wlk (t) eiro 1k 1dt

0
Since at l -= 0 the system was in a state ¢t the probability

of transition to a state¢? is I c 1 (T) J2 , i.e.
4:n;2
Pn-+1 = l c1 (T) \2 = ----;[21 Wzh. (ffir~t) 12 (III-34)
00
where W 11, (w 1 ~t) = 2~ ~ W 1" (t) eiwzht dt is the Fourier trans-

-oo
form of a matrix element of the perturbation energy ope-
rator, corresponding to the Bohr frequency w1k.
If tho transition is caused by the radiation of light (either
absorption or emission) whose wavelength is considerably
greater than the dimensions of the system (for an atom,
A~ aB), the perturbation energy can be written in the form
W= -(E (t)·D)
where E (t) is the electric field vector, and D = er is the
dipole moment of the system. And so, in the electric dipole
approximation,
4:rt2
Ph-+l = h21 (DE)t~tl 2 1 E (<Dz~t) 12 (III-35)
From the last relationship we see that under the influence
of radiation the system passes from a kth state into an lth
state only when E (w lk) =I= 0 (that is, if in the radiation
there is a frequency equal to the Bohr frequency of this
transition), and also when
(DE)lh :=/== 0 (III-36)
which defines the so-called selection rules.
When a time-independent perturbation W (r) is "turned
on" over a period of time 0 ::::;::; t ::::;::; -r (the rest of the time it
is "off"), formula (111-34) takes another form:
pk-+l = 2: I wlh 12 X 't X 6 (Eh. -Ez)
PROBLEMS
Bohr"s theory
Using the Bohr-Sommerfeld quantization rule (111-10),
solve the following problems.
1. Determine the energy levels of a particle in a one-
dimensional potential well with walls of infmite height
at x = 0 and x = a.
2. Quantize t.he motion of a one-dimensional harmonic

oscillator.
3. Find the energy levels for a particle that performs small
oscillations in a three-dimensional potential well about its
position of equilibrium (this position being in the origin
of coordinates). The potential energy V (x, y, z) in the
position of equilibrium is zero.
4. Find the energy levels of a particle of mass m that
freely rotates in a plane circle of radius r (a rigid plane
rotator).
5. Find the energy levels of an electron moving in an
elliptical orbit about a nucleus of charge Ze.
6. Determine the energy levels of a hydrogen atom that
moves freely in the limits 0 < x < a, 0 < y < b, 0 <
<z <c.
7. A particle of mass m falls onto a horizontal plane and
elastically bounces up. Quantize the particle's motion,
determine the admissible heights Hn, and calculate the
energy levels of this system.
Operators
8. Find the explicit expressions for the following opera-
tors:
(a) (a: +xr; +-;1 )3 ;
d
(b) ( dx
(c) (x :xf; (d)(:xxf;

(e) [ihV +A (r)J2; (f) (L-M) (L+M).
9. Find the commutation relations for the following ope-
rators:
(a) x and :x ; (b) iltV and A (r); (c) a~ and f (r, e, qJ).
10. Find the translation operators that map (a) 'ljJ (x) into
'ljJ (x +
a); (b) 'ljJ (r) into 'ljJ (r +
a).
Find the operator that rotates space through an angle ex..
11. Find the operators that are hermitian conjugate to
a an
(a ) - . (b) axn .
ax '
94 PROBLEMS IN THEORETICAL PHYS!CR
. ij
12. ChC'rk thn hPrmiticity of the opPrcdm·s :r, lily, and
of the Laplacian operator.
13. Find the operator that is hermitian conjugate to the
operator of space translation by vector a (see Problem 10b).
14. Find the operator that is hermitian conjugate to
• {J
e 'aBQl,
15. Find the hermitian conjugate to the product of opera-

tors A and B.
16. Show that if i and M are hermitian, the operators
~ 1 ~ ~ ~ ~ ~ i ~ ~ ,, ~
F= 2 (LM+ML) and f='i (LM-ML)

are also hermitian.
17. Prove that the expectation value of the square of an
observable is always positive.
18. For operators i and M satisfying the condition
~M -- Mi = 1, find LM 2 - Jlt2£.
19. For operators L and M satisfying lhe condition
LM- Mi = 1 (see Problem 18), find t (L) M- Mf (L).
20. For any two operators A and B that do not commute
prove the validity of the following relationships (provided
that A-l exists):
(a) A-1B2A = (k 1BA)2;
(b) .A-t.8nA=(k 1i1.A)" if n is an integer;
(c) i-1/(B)AA=f(A- 1BA).
21. Prove that the relationship etAfle-r,.~ = fj + C~
holds true provided that AB - RA = C, where C iRa scalar
and ~ is a parameter.
22. Prove that
ei\,plhF (q) e-i't;,p/h = F (q ~) +
where p and q are the momentum and position operators,
respectively.
23. Find the eigenfunctions and eigenvalues for the ope-
t d d .d
ra ors a; an zr!.r.
SECTION Ill. QUANTUM MECHANICS 95
24. Find the eigenfunct.ions and eigenvalues for tl1e Opf'ra-

d
tor x + dx.
25. Find the eigenfunctions and eigenvalues for the ope-
d
rator dcP.
26. Find the eigenfunctions and eigenvalues for sin d~ •
27. Find the eigenfunctions and eigenvalues for cos ( i d~).

. d
ta-
28. Find the eigenfunctions and eigenvalues for e diJ>.
d2
29. Find the eigenfunctions and eigenvalues for dz 2 +
+~!!..
z cl:r •
30. Make the transition from the classical Poisson bracket
to the quantum Poisson bracket assuming that their pro-
perties are the same. For instance, the relation
{f, g(p} '~ g {f, <p} + U, g} cp
(where j, g, <p are functions corresponding to observables,

and { } is the Poisson bracket) holds true for the operators
f~ i. ~.
31. Find the commutation relation for the annihilation
• 1 ~ ~ ~ ~
and creation operators a and a*, where a = Vznw (roq+ip)

~ 1 ~ ~
and a* = V Z.liw (roq - ip).

32. Express the Hamiltonian fi for the simple harmonic
oscillator in terms of the annihilation and creation opera-
tors (see Problem 31).
33. Prove the commutation relation for the operator of
angular momentum L: [L X L] = iliL.
34. Prove that the operator l} = L~ + L~ + i~ com-
mutes with all three components of L (use the results obtain-
ed in Problem 33).
35. Checkj the following commutation relations for the
operator of the dipole moment of a system of N electrons,
96 PROBLEMS IN THEORETICAT, PHYSICS
.'V
<l =~ - e ~ I'i• all(! the operator o[ the system's total angu-
i=1
N
lar momentum L= ~ Li:
i=1
[Lx, dx] = 0, [Lx, d11 ] = iftdZl [Lx, dz] = - ind 11 , etc.,

x, y, z cyclic.
36. Show that
[L2 , d] = 2ft2d + 2ift [d XL]
where L 2 = L~ L~ + +
L; is the operator of the square
of the total angular momentum, and d is as defined in Pro-
blem 35.
Solution of the Schrodinger equation.

Calculation of expectation values
and probability currents
37. Find the, general solution of the one-dimensional

time-dependent Schrodingerd equation for a free particle.
38. At time t = 0 the state of a free particle is specified
by a wave function
x2
"Ill' (x, 0) = Ae -(ii"+ihox
Find the factor A and the region where the particle is
localized. Determine thej probability current density j.
39. Find the Fourier transform of the wave function of
Problem 38. Determine the width of the wave packet in
k-space.
40. A wave packet is represented at the time t = 0 by
a function
x2
--+ihox
'l' (x, 0) = Ae 42
How will it propagate in time, i.e. what will be the form

of the wave function "If (x, t)? Also determine the probabi-
lity density p (x, t) and the probability current density
j (x, t).
41. Find the expectation values of position and momen-

tum for a particle with a wave function \jJ (x) = Ae-x 2 fa 2 +ikox.
42. Calculate (L\x 2 ). and (L\p 2 ) for the particle of Prob-
lem 41 and check the uncertainty relation for these two
quantities.
43. A particle is in a one-dimensional potential well
0 ~ x ~ a, for which V = 0 inside the well, and V = oo
outside. Solve the time-independent Schrodinger equation
for this case.
44. Find the wave function and the allowed energy levels
for a particle in a potential field V (x) of the form
V (x) = 0 for 0 ~ x ~ a, 0 ~ y ~ b, 0 ~ z ~ c
= oo for x < 0, x >a, y < 0, y > b, z < 0, z >c.
45. Find the energy levels and the wave functions for
a particle in a rectangular potential well of finite depth
(the one-dimensional case). The field V (x) is given in the
form
V (x) = 0 for x<-a (in the first region)
= - V 0 for -a ~ x ~ a (in the second region)
= 0 for x >a (in the third region).
46. Find the energy levels of a three-dimensional harmo-

nic oscillator whose potential energy is
V= ktx 2
2
+ k2y 2
2
+ kaz 2
2 .
47. Find the energy levels and the wave functions of a

one-dimensional harmonic oscillator that is located in a
constant electric field E. The electric charge of the oscilla-
tor is e.
48. Consider a one-dimensional harmonic oscillator that
is on its nth energy level. Find (x 2 ) and the expectation
value of potential energy for such a case.
49. Find the average kinetic energy of a one-dimensional
harmonic oscillator whose energy is ~ nw.
7-01496
50. Solve the Schrodinger equation for a particle in

a potential field V = V0 (e-2ax - 2e-ax).
51. Find the energy levels and the wave functions of a
particle in a one-dimensional Coulomb potential well V (x) =
e2
=- jXj"
52. Solve the Schrodinger equation for a three-dimension-
al, spherically symmetric harmonic oscillator with a po-
mw2
tential energy V (r) = 2 r 2 •
53. Solve the two-dimensional Kepler problem, i.e.
find the energy spectrum and the wave functions of a par-
ticle in a potential field V = - zez, where p = Y x 2 +y2 •
p
(All _functions are independent of z.)
54. Solve the Schrodinger equation for a particle in an
infinitely deep, spherically symmetric potential well speci-
fied by the potential
V (r) = 0 for r < a
= oo for r >a.
55. The electron of a hydrogen atom is in its ground
state. Determine (r ), (r2 ), and the most probable value r 0
for this case.
56. Normalize the wave functions of the electron in a
hydrogen atom corresponding to n = 2, where n is the
principal quantum number.
57. Consider a particle in a potential field
e2 C
V(r)= --;:-+r:z
Find the energy levels and the corresvonding wave functions
for the system.
58. Solve the Schrodinger equation for a particle in a po-
tential field V (r) = Ar2 + B2 •
r
59. Consider a particle that in the yz-plane moves freely
in a rectangle 0 :::;;; y :::;;; a, 0 :::;;; z :::;;; b, whereas the rest of
the plane is inaccessible for the particle. Along the x-axis
the particle is acted upon by a quasi-elastic force F ---:- -kx.
Find the energy levels and the corresponding wave functions
for the particle, and calculate the normalization factor.
60. Solve the Schrodinger equation for the electron of a

hydrogen atom in parabolic coordinates.
61. Solve the Sch:r:odinger equation for a particle with
a zero angular momentum (l = 0) in a potential field V =
= -Voe-rfa.
62. What is the energy spectrum of a particle in a perio-
dically changing potential field? The fi.eld is given by the
relations
V (x) = 0 for nl ~
x-< nl +a
(n = 0, ±1, +2, ..• )
= V 0 for nl - b ~ x ~ nl
The quantity l = a + b is the period of the potential.
63. Examine Problem 62 for the case when V = 0 every-
WHere except at points x = nl. In these points V 0 = oo
and b--+ 0 in such a way that
lim m:t =constant
(the Kronig-Penney model). Find the dependence of the

energy E on the wave vector k near the edge of the allowed
energy bands.
64. Examine the case of a semi-infinite crystal with a pe-
riodically changing potential when x > 0, determined as
in Problem 63. In the region x < 0 the potential V = W 0 •
Restrict yourself to the case when E < W 0 (the Tamm levels).
65. Using the uncertainty relations for p and ~. estimate
the energy of the ground state of a one-dimensional harmo-
nic oscillator.
66. Two particles that interact by an elastic force F =
= k (x1 - x 2 ) can freely move along the x-axis (one-dimen-
sional motion only). Find the energy spectrum and the
wave function for this system.
67. Find the energy spectrum and the wave function of
the hydrogen atom, taking account of the participation
of the nucleus in the relative motion about the centre
of mass.
68. Two particles of mass m are able to move along the
x-axis only and interact with each other by an elastic force.
7*
In addition, each of them interacts with the origin x = 0

by another elastic force with a force constant different from
the first. Find the energy levels and the wave functions for
such a system.
Theory of representations. Matrices

69. A particle is in an infinitely deep, one-dimensional
4n21i2
well (see Problem 43) in a state with an energy E 2 = 2ma2 •
Determine the momentum distribution for this particle.
70. Find the operator ;; in the momentum representation.
Determine its eigenvalues and eigenfunctions.
· 71. For a particle in a homogeneous potential field find
the eigenvalues and eigenfunctions of its Hamiltonian in the
momentum representation.
72. Determine the matrix elements of the dipole moment
x p
operator and of 2 and for a particle in an infinitely deep,
one-dimensional well, - ~ ~ x ~ ; .
73. Find the eigenfunctions for the Hamiltonian of a one-
dimensional harmonic oscillator using the momentum repre-
sentation.
74. Find the energy eigenvalues and the matrix elements
of the position and momentum operators in the energy repre-
sentation by using only the commutation relations for p
and q.
75. Use the commutation relations for the components
of angular momentum to find the eigenvalues of the square
of angular momentum, :f.2, the z-component of angular mo-
mentum, Lz, and the matrix elements of Lx and Lu in the
(L2 , Lz)-representation.
76. A wave function 1jJ = Ax (a - x) specifies the state
of a particle in an infinitely deep potential well with walls
at x = 0 and x = a. Find the energy distribution, the
expectation value (E), and the variance of the enemv distri-
bution (11£ 2 ).
77. A plane rotator is in a state with a wave function
1jJ =A sin 2 cp. Determine the probability of finding diffc-
rent values of the z-component of angular momentum, Lz,

and the expectation values <Lz) and <L; ).
78. Find the wave functions in the x- and p-representa-
tions for a particle localized at point x 0 and for a particle
with a definite momentum p 0 •
79. Find the expression for the operator ..!..
r
in the momen-
tum representation.
80. Calculate the angular part of the matrix elements of
the dipole moment for a particle in a spherically symmetric
field.
Time dependence of operators.

Potential barriers. Integrals of motion
81. Check whether the following equation for operators

holds true:
d ....... dA ~ ~dB
Tt(AB)= -;uB+A Tt·
82. Construct the operators dt and dt •

ir dp
83. Find the formula for the expectation value for cur-
rent density if the operator j is defined by the classical
formula, j = p~ (r - r 0 ) v.
84. Determine under what conditions the square of angu-
lar momentum, L2 , and the z-component of angular momen-
tum, Lz, can be integrals of motion.
85. For a particle whose potential energy is ~ construct
an operator that is the quantum analog of the following
quantity in classical mechanics:
K=[vxLJ+~
r
Show that k,, K 11 , and Kz are integrals of motion (compare

tho results with those of Problem 34, Section I).
0~
on. Construct the time derivatives ~ ~
dt and dt for the
oporators ~ and a* of Problem 31.

87. Find the equations of motion for a system with a

Hamiltonian
~ (p-eA)2
H= 2M + ecp (r, t), where A= A (r, t).
88. Find Cita;; an d dt ~

a (p;- eA ;) 1"f the Hamiltonian of
the particle under consideration is of the form
3
fi = ~ cah (Ph- eAh) + m 0c2a 4 + ecp (x1 , x2 , x3 )
h=il
where A =A (x1 , x 2 , x 3 , t) and a,., a 4 are matrices that
satisfy the following conditions: a;a,. +
akai = 26;,..
89. Use the definitions of Problem 88 and its solution
to make sure that (ad;ti) even =Pi
mrel
if Aeven= 1
2 (eA+Ae),
ii
= fEl
A A
where B (E is an eigenvalue of H).

·go. Show that for a system of N particles the total momen-
tum is an integral of motion, provided there are no external
forces acting on the system.
91. A particle' moving in the positive direction of the
x-axis meets a potential step, the form of the potential being
V (x) = 0 for x < 0
= V0 for x > 0
Determine the wave function forE> Vn and for E < V 0 ,
then for both cases calculate the probability current densi-
ties for the' incident, reflected, and transmitted waves and
also the reflectance and the transmittance.
92. Calculate the reflectance and the transmittance when
a flux of particles strikes a rectangular potential barrier
of width a:
V (x) = 0 for x < 0
= V 0 for 0-<. x-<. a
= 0 for x >a
Both quantities are ratios of the corresponding current
densities to the current density of the incident wave.
93. Examine the behaviour of a particle in the following

potential field:
V (x) = oo for x<O
. 0 for O~x~a
= Vo for a~x~b
=0 for b<x
Restrict yourself to the case when E < V 11 • Study the wave
function in the case wn·Pn its amplitude in the innflr region
(0 ~ x ~a) is considerably smaller than in the outer region
(x:;;::: b).
94. C:1lculate the tr:1nsmittance and the electric current
density caused by the emission of electrons from a metal
due to an electric field E. The surface of the metal is in
plane x = 0.
95. Find the dPcay const:1nt 'A if for alpha-decay the
transmittance of the potential barrier D and 'A are related as
'A = nD and if the potential is of the form
V = - V0 for r < r 0
and for r:;;::: r 0 , the alpha-particle interacts with the nucleu8
via the Coulomb law. The electric charge of the nucleus is
Ze, and r 0 ~ 2ZEe 2 • The factor no: v1 , where Vi is the partic-
ro
le's velocity inside the nucleus and r 0 is the radius of the
nucleus, is proportional to the number of collisions with
the "wall" of the nucleus per unit time.
A particle in a magnetic field. Spin

96. Determine the energy levels of a free electron in
a unifUrm magnetic field, with the vector of magnetic induc-
tion B directed along the z-axis.
97. Show that if we substitute the vector potential
A' = A t grad f for A and <p' = <p - :~ for rp in the
time-dependent Schrodinger equation, this leads to a change
in the wave function that has no physical consequences.
98. Construct the probability current density for a par-
ticle in a magnetic field.
99. Prove that the probability current density of Prob-

lem 98 is invariant under the transformations specified in
Problem 97.
100. Perform a canonical transformation 'IJl = Sq;, where
S= exp l~ ~
h
ehrhA (t, 0)] , in the equation
in~=
at
{~
LJ
~
(Ph-ehA(t, rh))
2mh
2
+ v} 'IJl
k
Expand the vector potential in a series

A (t, r) =A (t, 0) + (r·V) A (t, 0) + ...
and obtain an equation for q;, assuming that the dimensions
of the system under consideration (an atom, for instance)
are considerably smaller than the wavelength of the elec-
tromagnetic radiation.
101. Prove that the operators Gx, cry, Gz defined as fol-
lows:
GxCX=~. aya=i~, O'zCX=CX
O:x~=a, cry~= -ia, O:z~= -~
satisfy the same relationships as the Pauli matrices.
102. Construct the operator a;; with the help of the
Hamiltonian ii for a particle with spin 1/2 in a magnetic
field with induction B.
103. Find the eigenfunctions and eigenvectors for the
operators
and ~
ay= i
(0 - i)G ·
104. What is the projection of the square of spin Ti/2 on
a given direction?
105. Check the relationship
(a·A) (o·B) =(A· :B)+ i(hA. x 81
where a is a vector whose components are the Pauli mat-
~ ~
rices O':x:, O'y, O'z.

~ 1 ~ .~ ~ 1 ~ .~
106. Prove that 0'+= 2 (ax+Wy) and 0'_=2 (ax-~O'y}
are conjugate operators. Find the commutation relations for
them, and also the commutation relations between and a±
each of the Pauli matrices Ox, Oy, Oz. Determine a;,.
107. Given the operators ax, cry, Uz (the Pauli matrices),
prove the following relationships:
(a) sin (crxcp) =ax sin cp;
(b) cos(crzcp)=coscp;
·cr ~
(c) e1 Y«P=coscp+iaysin qJ;
(d) eioz!jlcrye_;iJz!jl = e2iaz!jlay.

108. Prove that for the operators a+ and a_ defined in
Problem 106 the following relationship holds true:
(a+a_)n = ;;.+;;.-
where n is an integer.
109. Assume that the transformation to ~ new representa-
~ - i crz q>
tion is carried out by an operator S = e 2 , so that ¢' =
= S¢. Find the operators A' and B' if in the old represen-
tation
A'= ax sin cr + ay cos cr and fJ =ax cos cr- cry sin cp.
110. What are the eigenfunctions of S z and S2 for a system
of two weakly interacting particles n and p of spin fi/2 if
S= an + ap? Determine the corresponding eigenvalues
of Sz and S2 •
111. Calculate the scalar product of the spin vector opera-
tors of two particles in the triplet and singlet states. The
spin of both particles is li/2.
112. Show that the operator (an •Gp)l!. can be linearly
expressed by (an
·Op).
113. Determine the energy levels and the wave functions
of a particle with spin S = 1 (in units of n) if the
Hamiltonian
fi =AS~ +BS~+cs:
where A, B, C are constants.
114. Let an electron pass in y direction through a homo-
geneous magnetic field with induction B parallel to the
z-axis. The spin of the electron points in the positive direc-
tion. When passing the point y = 0 at the time t = 0,
the electron enters an additional homogeneous field with
induction B' parallel to the x-axis. It leaves this auxiliary
field at y = l and t = t 0 • What is the probability of a spin
flip during this time interval?)
Approximate methods
of solving quantum mechanical problems
115. Approximate the energy levels and wave functions of
the anharmonic oscillator
by the time-independent perturbation theory, in the first

and second orders of the approximation.
116. A spinless particle is exposed to a spherically sym-
metric field, and its energy levels are Enz (for the lJDpertur-
bed system). Find the energy levels and the corresponding
wave functions in the first approximation of the time-inde-
pendent perturbation theory if the particle is acted upon
by a magnetic field parallel to the z-axis.
117. A rigid plane rotator is placed in a weak electric
field E directed along the x-axis. The charge of the particle
is e, and it is at a distance a from the centre. Calculate
the energy corrections in the first and second approximations.
118. Find the corrections to the energy and to the wave
function of an electron in a periodic potential V (r), this
being the perturbation. The lattice constant is a. Consider
the case when two energy levels are equal: Ek = Ek+2ng,
where g is the vector of the reciprocal lattice, (ga) = n.
119. A hydrogen atom is placed in a homogeneous electric
field E directed along the z-axis. Find the splitting of the
energy level that corresponds to n = 2 (n is the principal

quantum number).
120. Show that there is no linear Stark effect in the
atoms of group I of the Periodic Table since the energy levels
of these atoms, En 1• are determined only by n and l. (The
Stark effect is the splitting of energy levels in an external
homogeneous electric field.)
121. Analyze the Stark effect in a hydrogen atom (calcu-
late the splitting of the nth energy level in an electric field
E directed along the z-axis) by using the solution of the
unperturbed problem in parabolic coordinates.
122. Consider a hydrogenlike atom (with the charge of
the nucleufl Ze), and aflsume that the nucleus is a uniformly
charged sphere of radiufl r 0 • Calculate the energy shifts
for the n = 1, l = 0 state, and for the n = 2 states, in the
first order of approximation of the time-independent pertur-
bation theory.
123. Suppose a system whose stationary states are '1'0
and '¥1 is in the state "'I' 0 • At the time t = 0 a perturbation
W not depending on time is switched on. Find the time
dependence of "'I' (t) for the perturbed system.
124. Show that if a perturbation W (r) not depending on
time is switched on over a period of time 0 :s;: t :s;: -r and
the rest of the time it is off, the probability of transition
from state n to state k is
21in I W nk 12 -r8 (En- Ek)•
125. Calculate the probability of the ionization of a

hydrogenlike atom by a plane monochromatic wave. The
vector potential of the wave A has the following components:
Ax=Acos(wt-kr), Ay=O, Az=O
Before the perturbation (the wave) ifl switched on, the
electron is near the nucleus Ze and is in the n = 1, l = 0
state. In the final state the electron can be considered free.
126. Two identical particles are located in an external
potential field V (r), and the interaction between them is
H12 • Assume that the solution of the Schrodinger equation is
known for a single particle, and find the solution for the
two particles.
127. Use the solution of Problem 126 to examine the

time dependence of the probability that the two particles
will remain in their original state if at the initial time
(t = 0) one was in the rth state and the other in the sth state
of the single-particle Hamiltonian. Determine the time
that is needed for them to change states.
128. Using the Ritz variational method, find the ground
state of a three-dimensional oscillator. Choose the trial
function in the form
<p = A (1 +
ar) e-ar.
129. Consider the central-force model of the deuteron
using the Ritz method. The neutron-proton interaction is
idealized by the central-force potential V (r) = -Ae-rla.
Choose the trial function in the form
<p = Be-rzr/2a.
130. Using the Ritz method with two trial functions,
'PI = A (1 +
ar)e-rzr and <p 2 = Be-rzr 212, find the energy of
the ground state of the electron in a hydrogen atom and
compare the results.
131. Calculate the ionization potential for a helium atom
using the method of the perturbation theory and varying
the screening constant s. Assume that in the unperturbed
system each electron interacts with the nucleus via a poten-
tial - (2-s) e2 , and include the compensating member
r
-se2rt + -se2r2
in the perturbation.
132. Consider the elastic scattering of particles by a
centre, assuming that the particles interact with the centre
via a potential V (r), which is considered a small perturba-
tion. Find the differential scattering cross section, defining
it as the ratio of the radial flux through an area element dS
far from the scattering centre to the incident flux:
da - liscatdSI
- I itn I '
133. Use the result obtained in Problem 132 to find the
differential cross section for Coulomb scattering {the Ruther-
ford formula).
134. Determine the differential cross section for the elas-

tic scattering of positively charged particles (each with
charge e1 ) by an atom. Regard the atom as a fixed centre
with charge Ze surrounded by a charged uniform sphere of
-ep (r) density. Calculate da for p = p0 e-rla.
135. Calculate the differential and total cross sections
for the scattering of particles by a field V = ~ e-xr (the
r
Yukawa potential).
SECTION IV Statistical Physics
and Thermodynamics
Two methods are used to describe the state of systems that

consist of a great number of particles, the statistical and
the thermodynamic. The first method enables us not only
to obtain the general relationships of the second method
but to calculate the concrete values of the thermodynamic
observables for the given system.
Postulates of classical statistical mechanics. Let us
assume that a system consists of N particles contained in
volume V. The motion of the particles follows the equations
(lB.ws) of classical mechanics. The state of this system is
fully determined by defining generalized coordinate!' q1 ,
q 2 , • • • , q N and momenta p 1 , p 2 , • • • , p N (a point in
phase space of 2N dimensions), and the dynamics by the
Hamiltonian H = H (pi, qi), which makes it possible to
find Pi, qi at any instant of time by means of the equations
iJH iJH
Pi=- aqi, qi= opi (i=1, 2, ... , N) (IV-1)
It is impossible to determine the state of a system with
a large number of particles (N) at any instant of time, and
there is no need to. A study of these many-particle systems
usually focusses on some small number of macroscopic variab-
les. For example, suppose we require that the number of
particles in the system is N, the volume is V, and that the
possible values of energy lie in the interval [E, E +
LlE].
Obviously, many states of the system satisfy these require-
ments. Hence, to calculate for a given system the average
of a macroscopic variable of the type
T
F' =lim+
T-+oo
JF [pi(t), q;(t)] dt
0
(IV-2)
we use the following method.

SECTION IV. STATISTICAL PHYSICS AND THERMODYNAMICS 111
We introduce an infinite number of copies of the given

system [one Hamiltonian H (Pi, qi) but different initial
conditions pt q~] at .some time t. Such an "ensemble" of
systems will be distributed in phase space with a certain
density p (pi, qi, t). If the system is isolated from the
ambient medium, its energy will be a constant value, i.e.
(IV-3)
and the phase points of the ensemble will be distributed
over the hypersurface of constant energy (IV-3).
The probability density is determined in this case by the
following relationship known as the microcanonical distribu-
tion:
P (H)= 6 [H (p;, q;)-E] (IV-4)
Q (E)
where c5 (x) is the Dirac delta function (see Appendix 4), and
Q(E)= ara<:), f(E)= J... J dNpdNq (IV-5)
H(pi' qi)~E
The ensemble average for any physical observable will

now be
- roor NN
P= J :..'OC: J p(H)F(p;. q;)d pd q (IV-6)
Real physical systems are in contact with the ambient

medium. For example,
(a) the mechanical interaction with a source of work that
affects the system. Such a state of the system can be des-
cribed by a Hamiltonian H (p;, q;, as), where as are the
generalized coordinates of external bodies viewed as addi-
tional variables of the system. Then As = - 0°H will stand
as
for the generalized force which the system exerts on the
ambient medium;
(b) the thermal contact between the system described by
a Hamiltonian H (pi, q;, as) and a heat bath (another
system) with a Hamiltonian H 0 (pi, qi). The total Hamil-
tonian is
H(p;, qi, pi, qi, as)=H(p;, qi, as)+Ho(Pi, qi}
+
H' (P;. q;, pi qi)
These systems are in equilibrium if the energy H' is small

and allows the system to exchange energy with the heat
bath fast enough for any state of the total system to be
realized after a long period of time with equal probability;
(c) the material contact between the system and a heat
bath that consists of an exchange of particles. In a state
of equilibrium the energy H' must satisfy the previous
requirements.
In the case of (b) the probability density for the ensemble
has the form (the canonical distribution)
p (H)= Ce-H<P;, qi, as)/9 (IV-7)
where C is the normalization constant, and
1 _ alnQ (E)
e= aE
(IV-8)
Here 8 is a constant common to both systems in equilibrium .

. Denoting C == eF(9, as)/9, we get
F(8, a 8 )= -8ln .\·~ex:) e-H(pi,qi'as119 dNpdNq (IV-9)
The quantity
(IV-10)
has come to be known as the classical partition function. It

is the main quantity when the physical properties of a sys-
tem are evaluated.
From formula (7) we can show that
- = F- 8
E == H (00 aF) 6
- - ( aas
aF) as, As= (IV -11)
and the entropy is defined as

S= -k(.!!...)
ae as
where k is Boltzmann's constant. Since the temperature of
two systems in equilibrium in contact with each other is the
same in both systems, this suggests that 8 is a function of
temperature. We introduce the absolute temperature of a
system, T, assuming that 8 = kT. The function F, defined
by formula (IV-9), can be identified with the Helmholtz free

energy.
So, knowing the Hamiltonian of the system H, we can
fi.nd Z according to (IV-10), and then
F (T, a.) = -kT ln Z (IV-12)
All other quantities we fi.nd by simple differentiation:
E=F-T ( aF)
aT as
A8 = - ( :~ ) T, for instance, p = - ( :~ ) T, (IV-13)
S =- ( :~ L., etc.
Thermodynamic description of a state of a system. From
the relationships just cited we can obtain the basic laws of
thermodynamics.
(1) The first law. The amount of heat obtained by a system
(decrease in the mean energy of the heat bath) goes to change
the internal energy of the system and to perform work
exerted by the system on external bodies:
dQ = dE+ 2j As da 8
s (IV-14)
dQ= -dH0 (pi, qi)
(2) The second law.
dQ -
dS ~ T, T dS =dE+ ~As da 8 (IV-15)
The equality corresponds to the equilibrium state of the

system.
(3) The third law (Nernst's heat theorem). As T-+ 0,
the entropy of a system, S (T, a8 ) , ceases to depend on a.,
i.e. is a permanent quantity for all substances.
The method of thermodynamics rests on these three laws,
using the fact that S and E are state functions of the system
for which
(IV-16)
i.e. dE and dS are total, or exact, differentials.

8-01496
The differential dZ = Xdx + Y dy is an exact differen-

tial if
Here are the relationships for other functions of state

(thermodynamic functions).
(1) The Helmholtz free energy F = E- TS:
dF= -S dT- 2J A 8 da •
•
as )
( iJa ( DA. ) (IV-17)
8 T= iJT a,,
(2) Enthalpy II= E + 2} A.a,:

8
dH =T dS+ ~a. dA,
( ;aT )
iJA a =
( iJa 8
iJS
}
As
(IV-18)
8 11
(3) The Gibbs free energy <1> = F + ~ A,a.:

•
d<l> = - S dT + ~ -z 11 dA,
•
-
( iJAas) 8 T=
(iJa8
iJT
}
Aa
(IV-19)
If in addition we use the equation of state As = A. (T, a.),

we get a number of relationships that can be checked expe-
rimentally. Thermodynamics does not find the numerical
values of the quantities entering into these relationships.
[The line above A. is omitted in formulas (IV-17) to (IV-19).]
Grand canonical distribution. If there are m kinds of
particles in a system and the type of contact is (c), we get
m
( -H(P;• qi' a 8 )+ 2} ILjN j]jkT
P= Ce i=t (IV -20)
where Cis the normalization constant,
1 _ {} ln Q (E, N1, ... , Nm) (IV-21)
w= aE
and
. = kT oln Q ( E, Nt. ... , N m)
~~-- .
oN j
is the chemical potential of the jth type of particlef?.
Bearing in mind a correct calculation of the number of
states (the Gibbs paradox), we find that
m
00 00
:Lj JJ.iN/kT
ei=1
Ntl N 2 ! ... Nml Zo
Nt=O Nm=O
Il(P;• qi, a 8 )
Z0 = j ~~ j e- kT df (N 1) (IV-22)
E(T, a8 , JJ.j)
Defining C e= kT , we get an expression for the
thermodynamic function 8:
8= -kTlnZ (IV-23)
and
s- - -( &3
-}
oT ILJ• a 8 '
-
A 8
=(-&8
-)
oas T, ILJ
(IV-24)
N
- . - - ( __:::_
i}>< )
1- Ofkj T, as
Hence,
m
dB= -S dT- )J A8 da 8 - ~ Ni d~i (IV-25)
• i=t
mi
dS=+ (dE+~ A 8 da 8 - ~· ~ 1 dN 1 ) (IV-26)

j=1
m
dc:D= -S dT + ~ a8 dA.+ ~ ~ 1 dN 1 (IV-28)
s i=i
m
0= -SdT+ lja8 dA8 - ~ N,d~J (IV-29)
' j=i
8*
We see from the last relationship that f-Li> A, and T are

not independent.
The condition for the equilibrium of two systems has
the form
Pt = p 2 (mechanical equilibrium) (IV-30a)
T 1 =T 2 (thermal equilibrium) (IV-30b)
~ (equilibrium with respect to the (IV-30c)
T2 exchange of particles)
Quantum statistics. The stationary state of a quantum
mechanical system of N particles is described by the wave
function 'I' k (q) = 'I' (r1 , r 2 , • • • , r N), which is determined
from the time-independent Schrodinger equation
(IV-31)
where
A fl2 "\:1 1
H = - 2 LJ mk ilk +U (rh ... , rN)
k
The value of the physical observable f in a state k of a sys-

tem (in quantum mechanics an operator f corresponds to
this observable) is found by the rule
(/) = j 'JI'j.j'J!k dq (IV-32)
But in statistical physics each state E k is given a definite

probability of realization depending on the macroscopic
conditions of the system. It follows from this that the avera-
ge (mean) value in quantum statistics is calculated thus
<7> == 7= ~ wk <f>h = ~ wk j 'Phf'i'k dq
k h
=) ) 6 (q- q') j (q) p (q, q') dq dq' (IV-33)

q q'
where
p (q, q') = hk WklYk (q') 'iJk (q) (IV -34)
is the density matrix in the coordinate representation.

By analogy with classical statistical physics wk is chosen

in the form
F(a 8 , T)-Ek
Wk=e kT (IV-35)
Hence,
F = -kT ln Z (IV-36)
where the partition function
Ek Ei
Z=~e-w=~e-WQ(Ei) (IV-37)
k j
and Q (Ej) is the degeneracy multiplicity of the system's

jth energy level. Summing up over j means summing up
over the discrete energy levels and integrating over the
continuous spectrum of the system. All subsequent rela-
tionships are full.y equivalent to the relationships of classi-
cal statistical mechanics.
When there is material contact,
oo m
Z= S ··· ~ ~expr- k~ (EN,n- ~ Nilli)J (IV-38)

N!=O Nm=O n i=1
where EN. n is the energy of the nth quantum state. The

thermodynamic function 8 is determined as in (IV-23) ac-
cording to the formula
8 = -kT ln Z (IV-39)
If the particles are noninteracting, relationship (IV-38)
is greatly simplified because the state of this system is
given by the occupation numbers ni owing to the indistin-
guishability of the particles. Then
En= ~ e;nL> N = ~ ni (IV -40)
i i
where e; is the energy of the ith single-particle state, and

the partition function is
Z= ~ ( e e~~lt ) n;
(IV-41)
where tho summing up is done over the number of particles

that are in the single- particle state with e;.
The general principles of quantum mechanics impose

strict rules on the occupation numbers ni. Only two cases
are possible:
(1) ni = 0 or 1 (particles with a half-integral spin);
(2) ni = 0, 1, 2, . . . (particles with integral spin).
Now, using (IV-41), we get the following relationships
for the average number of particles in a state with energy e:
1
n= 8-IJ.
(Fermi-Dirac statistics) (IV-42)
eliT" +1
1
n= 8-IJ.
(Bose-Einstein statistics) (IV-43)
eliT" -1
The energy and the number of particles are determined by

standard formulas:
(IV-44)
N = LJi n; (the condition for finding !l)
The limit of application of quantum distributions for a

perfect gas is
(IV-45)
Fluctuations and kinetic theory. Statistical mechanics

makes it possible to calculate the fluctuations of physical
quantities because the probability of fluctuation of a macro-
scopic quantity can be expressed in terms of T and a 8 • The
probability of fluctuation of a physical quantity x (the equi-
librium value of which is x 0 ) is
wmln(xo, x)
p(x)=Ce kTo (IV-46)
where W min (x 0 , x) is the minimal amount of work needed
to shift the system from x 0 to x, and T 0 is the equilibrium
value of the temperature in the system.
When the 'deviations from the state of equilibrium are
small, for the probability of fluctuations of thermodynamic
SECTION IV. STATISTICAL PHYSICS AND THERMOD"Yl'iAMICS 119
quantities we get
6.p6. V- 6.T6.s
. p=Ce 2kTo (IV-47)
With this formula we can find the fluctuations of~ any
thermodynamic observable if we choose the independent
variables in the proper way.
In studying various kinetic phenomena we must know
the number of particles that at time t have a radius vector
lying in the interval [r, r +
dr] and a velocity in the inter-
val [v, v + dv], i.e. know
f (r, v, t) dr dv
For the distribution function f we can obtain an integro-
differential equation of the type
[ :t +vt ar-+ ~ 0~1 Jf (r, v 17 t)
= J Jdv2a(Q)lvt-V21
dQ
X [f(r, v;, t)f(r, v;, t)-f(r, v2, t)f(r, v 17 t)] (IV-48)

Here v1 and v 2 are the velocities of the first and second.
particles before collision, v; and v; the same quantities
after collision, a (Q) is the effective cross section of the
collision, and dQ is the solid angle differential.
To get (IV-48) we must (a) consider none but binary colli-
sions, (b) neglect the action of the walls of the vessel
containing the particles, (c) neglect the influence of external
forces on the effective cross section, and (d) consider the
velocities being independent from the position of the par-
ticles in the vessel.
Function f defined in (IV-48) makes it possible to calcu-
late a variety of kinetic coefficients: the tensor of electrical
conductivity, the coefficients of thermal conductivity, vis-
cosity, diffusion, etc.
PROBLEMS
1. A certain system can with equal probability be in
any of its N states. What is the probability of the system
being in one of its states?
2. A simple pendulum performs harmonic oscillations

according to the law
QJ = QJo cos ;,n t ( T = 2n V~)
Determine the probability that in a random measurement
of the angle of deviation its value lies in the interval
[qJ, {jJ + dqJ].
3. The probability that for a certain system the values
of x and y lie in the intervals [x, x + dx] and [y, y + dy]
is
dW (x, y) = Ce-a<x 2+Y 2) dx dy (a> 0)
Assuming that the range of values of x and y is [-oo, oo],

fmd the normalization constant C.
4. Determine for the previous case the probability that
the value of x lies in [x, x + dx].
5. During thermionic emission electrons leave the surface
of a metal or a semiconductor. Assuming that (1) emission
of electrons are statistically independent events and (2)
the probability of emission of an electron in a small time
interval dt is 'Adt ('A being a constant), determine the proba-
bility of emission of n electrons in a time interval t.
6. For the previous case determine
11n2 = (n- n) 2
if on the average n 0 electrons are emitted every second.
7. A perfect gas consisting of N molecules is contained
in a vessel with a volume V. What is the probability that
there will be n molecules at any given time in a volume
V 0 (V 0 is much less than V)? Examine the extreme cases:
(a) n is much less than N, and (b) n is much greater than unity
and 11n = n - ~n n.
8. Prove that for a random variable x the probability
of an event in which x becomes greater than a certain value
a satisfies the Chebyshev inequality
x2
w (x >a):::;; liZ.
9. A particle that at the initial time is at the origin of
coordinates jumps the next instant by one unit to the right
or left with the probability of 1/2. Determine the probabi-

1ity P t (l) that after t steps the particle will be at point l of
the one-dimensional grating (Fig. 44).
10. Determine the probability P t (l) of a similar random
walk of a particle across a two-dimensional (square) and a
three-dimensional (cubic) grating if the particle can move
-It -3 -2 -1
....
0
-- 2 J 4
Fig. 44
to any adjacent point (4 points for the square and 6 for

the cubic) with the probabilities of 1/4 and 116, respectively
(Fig. 45).
11. The following question (first raised by G. Polya)
arises in considering Problem 9: Can a particle always return
Ftg. 45
to its initial point if it moves at random to adjacent points
of the grating? If it cannot, what is the probability of its
not returning to the initial point for the cases examined
in Problems 9 and 10?
12. In space p, q draw the phase trajectory of a particle
that moves with constant velocity in a direction perpendi-
cular Lo the mirror-reflecting walls of a box. The size of the
box in the direction of motion is 2a.
13. Determine the phase trajectory of a body of mass m

that moves with an initial velocity v 0 in a constant gravi-
tational field from point z0 in a vertically upward direction.
Draw the trajectory.
14. Determine and draw the phase trajectory for a par-
ticle of mass m with an electric charge -e that moves under
the influence of a Coulomb force of attraction toward a fixed
charge +e1 • The initial distance between the charges is r 0 ,
and the initial velocity of the negative charge is v 0 = 0.
15. Determine and draw the phase trajectory for a linear
harmonic oscillator described by the equation
~ + y; + ro~x = 0 ( ro0 = V~)
where y ~ ro 0 • Find the change in the phase volume with
the passage of time.
16. Determine and draw the phase trajectory for a com-
pound (physical) pendulum of mass m whose moment of
inertia is I and whose equivalent length is L. Consider
three cases:
(1) H 0 > 2mgL;
(2) H 0 = 2mgL;
(3) H 0 < 2mgL (Hu is the initial energy of the pendulum).
17. Verify whether Liouville's theorem holds for the
following cases:
(1) an elastic collision of two spheres (central impact);
(2) the motion of three particles in a constant gravitatio-
nal field, the initial states of the particles being determined
by the phase points
A (po, Zo), B (po, Zo+ a), C (Po +b, zo)·
18. Determine the normalization divisor of the micro-
canonical distribution for the following systems: (1) N
molecules of perfect~·-monatomic7gas; (2) N independent
linear oscillators.~. · · -
19. Derive the canonical distribution using the examples
in Problem 18 as the model of a heat bath.
20. We can represent the Hamiltonian of a perfect gas
in the form
H=~Hi
i
where Hi is the Hamiltonian of an individual molecule.
Express the classical partition function of the gas in terms

of the partition function of an individual molecule. Deter-
mine the average energy E, the entropy Sand the pressure p.
21. Determine E, S, p, and Cv (Cv is the molar heat
capacity at constant volume) for the following systems
consisting of N noninteracting particles contained . in
volume V:
(1) a monatomic gas;
(2) a diatomic gas with hindered oscillations of the atoms
(a rigid rotator);
(3) a diatomic perfect gas allowing for the vibrations
of atoms in the molecule (consider the case of low tempera-
tures).
22. Determine the energy and pressure for a perfect gas
that consists of N particles and is contained in a vessel
with a volume V when the energy of an individual particle
depends on momentum p in the following way:
(a) H = apl, a > 0 and l > 0;
(b) H = c (m 2c2 + p 2 ) 1 12 , where c is the velocity of
light.
23. The classical partition function of a system is expres-
sed as
A 1
Z (~) = tiN , where ~ == kT
Determine Q (E).
24. A system as a whole revolves with an angular velocity
0. Find the canonical distribution in the revolving sys-
tem of coordinates.
25. A cylinder of height h and base radius R is filled with
a perfect gas. The cylinder rotates with an angular velocity
Q about an axis perpendicular to the base and passing through
its centre. Determine the pressure of the gas on the surface
of the cylinder if the number of particles in the gas is N and
the mass of an individual particle il'l m..
26. Derive the virial theorem for a system of interacting
particles contained in a volume V. The potential energy of
interaction is a homogeneous function of coordinates of
degree n.
27. Determine the average value nn (n > 0) for a mon-
atomic perfect gas consisting of N particles. Use the result
to find the mean square flnctuation in energy /1H 2 =
= (H - H) 2 and the mean square value 6 2 of the fractional

fluctuation (H - H)/H, i.e. <'1 2 = 11H 2 /H 2 •
28. Using the canonical distribution, find the following
distributions (the various forms of the Maxwell distribution):
(1) the probability that the velocity of any particle of
a system lies in the intervals [vx, Vx + dvx], [vy, Vy + dvyl,
[vz, Vz + dvz];
(2) the probability that the speed of any particle lies
in the interval [v, v + dv];
(3) the probability that the kinetic energy of any particle
lies in the interval [e, e + del.
29. Using the results of Problem 28, find the following
values:
(a) the mean of the nth power ofl the velocity vn for
n > -2; -
v
(b) the mean speed and the mean of the square of speed v2 ;
(c) the most probable speed of the particles v 0 •
30. Find the average energy e and the most probable
kinetic energy e 0 of a system of particles of Problem 28.
Explain why they do not coincide.
31. The energy of a particle in a relativistic gas is linked
with the momentum of the particle by the relationship e =
= c (m2c2 + p 2) 112 • Find the Maxwell distribution for this
case.
32. How will the Maxwell distribution change if the
system as a whole moves with a velocity u?
33. Find the probability that two particles will have
the absolute value of the relative velocity v' = v 1 - v 2
in the interval [v', v' + dv']. Also find v'.
34. Find the total number of collisions of one molecule
with the other ones in the system. Consider the molecules
as being elastic spherr r· lf radius R 0 • Determine the mean
free path A..
35. Determine the ratio of the number of particles that
have an energy greater than e1 to that with an energy less
than e1 (e 1 = kT).
36. Determine the total scattering cross section as a
function of temperature if the potential energy of interac-
SECTION IV. STATISTICAL PHYSICS AND THERMODYNAMICS f25
tion is
a
= --;:n. for r>Ro (n>2, a>O).
37. Determine the mean number of collisions experienced
by a single molecule with other molecules in one second in
the two-dimensional case (i.e. for a surface).
38. Each atom of a gas radiates monochromatic light of
wavelength /... 0 and intensity J 0 • Find the radiation intensity
of the gas as a whole as a function of A, the gas consisting
of N atoms.
39. Assuming that the potential energy of an electron
inside a metal is less than its energy outside the metal by
W = ecp, determine the current density of thermionic emis-
sion. The concentration of electrons in the metal is n 0 , and
the electron mass is m.
40. Prove that for a perfect gas with the known law
e = e (p) the pressure is given by the formula
00
p= 4; ~ J f; 1 1PI f(p)djpl
0
3
where f (p) is the momentum distribution function [i.e. f (p)

is the probability that the particle has momentum p].
41. Using the canonical distribution, find for a perfect
gas contained in an external potential force field u (x, y, z)
the probability that the coordinates of any particle of the
gas will lie in the intervals [x, x +
dx], [y, y +
dy],
[z, z + dzl.
42. Find the centre of gravity of a column of perfect gas
in a homogeneous gravitational field if the acceleration of
gravity is g, the mass of one molecule is m, and the tem-
perature is T.
43. A mixture of l perfect gases consisting of equal num-
bers of particles but with different masses of atoms m1 , . • . ,
mk, . . . , m 1 is contained in a cylinder of height hand radius
R and is placed in the earth's gravitational field. Deter-
mine the centre of gravity of the mixture.
44. Let the quantity 4nv2/ (v 2 ) dv represent the probabi-
lity that the velocity of a molecule lies in the interval
[v, v + dv], with f (v2 ) being a differentiable function of

unknown form. Find the Maxwell velocity distribution as-
suming that the probability distributions for the three
Cartesian components of velocity are (1) independent; (2)
identical.
45. Two vessels in which pressures and temperatures are
maintained at PI• TI and p 2 , T 2 , respectively, are connected
by a short pipe with a cross section S. Determine the mass
Fig. 46
of a gas flowing from one vessel into the other if the mass
of the gas molecules ism and if PI = 2p 2 and TI = 2T 2 •
46. A sphere of radiuR R moves with a velocity u in greatly
rarefied gas (i.e. the mean free path of a particle in the gas
is considerably greater than R). The temperature of the gas
is T and its density is n 0 • Assuming that the collisions of
the particles of gas with the sphere are elastic, determine
the force of resistance experienced by the sphere as it moves.
Compare the result with Stokes' law for the force of resist-
ance experienced by a similar sphere moving in viscous
liquid.
47. A small round aperture with a cross section S has
been made in a vessel containing perfect gas. Find the num-
ber of particles that fall on a round disc of radius R situated
at a distance h from the aperture. The plane of the disc is
parallel to the plane of the aperture (Fig. 46). The centres
of S and the disc lie on a straight line perpendicular to the
plane of the aperture. The molecules of gas are governed
by the Maxwell velocity distribution.
48. A rarefied gas is contained in a vessel at a pressure p.
Determine the velocity of outflow v of the gas into a vacuum
through a small aperture S 0 if the molecules of gas are

governed by the Maxwell velocity distribution.
49. Determine the permittivity of a perfect gas consisting
of N molecules with a dipole moment p 0 each. The gas is
located in an external homogeneous electric field E.
50. Make the same calculations as in Problem 49 but
take into account the polarizability a of the molecules,
which does not depend on the magnitude of the external
field.
51. Prove that a system of interacting electric charges
(the problem is considered classical) cannot be in equili-
brium in an external magnetic field.
52. A body with a potential IJlo is placed into a plasma
consisting of electrons (-e) and ions (+e). Determine the
Debye screening distance assuming that the temperature
of the electrons Te differs from the temperature of the ions
Tt and the plasma is quasi-neutral. The number of particles
in unit volume is n 0 •
53. A perfect gas is contained in a vessel that is closed
by a movable piston loaded with mass M. Find the equation
of state of the gas.
54. Derive the Dalton law for a mixture of n perfect gases:
n
P= ~Pt
i=i
where Pt is the pai'tial pressure of the ith component.
55. Prove that for any system with a Hamiltonian H,
1
Cv = kTZ (H -H)Z.
56. Prove that for any physical quantity I (qu ••. , qN,
Pl• ... , PN),
I aH = kT ( li1 ) and I aH = kT ( 81 ) •
aqt aqi api api
57. Particles of a rarefied gas interact according to the
law (Fig. 47)
U = oo for r~r 0
= -U0 ( ~or for r>ro
Find the heat capacity for the gas.
58. Show that the van der Waals equation can be writ-
ten as
(~ + : 2 ) (3w-1) = 8-r
(the reduced van der Waals equation), where
p T V
1t=--, T = - - , W=--
Pcr Tcr Vcr
(Per• Tcr, Vcr are the critical pressure, temperature and
volume, respectively).
59. Determine the average energy and the heat capacity
C v of a perfect gas consisting of N diatomic molecules,
taking account of the anharmo-
u nicity of atomic vibrations in
the molecules. Examine the
case of low temperatures.
60. Atoms in a diatomic mol-
ecule interact according to the
law
A B
U(r)= r12 ---;:s (A, B>O)
(the so-called Lennard-J ones

"12-6" potential). Find the co-
efficient of linear expansion for
Fig. 47 such a molecule.
61. For a system !consisting of
a large number of particles the heat capacity is
Cv = aTn (a > 0, n > 1)
Find Q (E) for such a system.
62. Entropy is sometimes defmed as S = k ln r (E) or
as S = k ln Q (E). Prove the equivalence of these defini-
tions for systems with a large number of particles.
63. Using the general properties of entropy and proba-
bility and assuming that there is a functional dependence
between entropy and probability, prove the Boltzmann
relationship
S = k ln w.
64. Find the work that has to be spent to polarize a unit
volume of an isotropic dielectric.
65. If each of the three variables A, B, C is a differen-

tiable function of the other two, regarded as independent,
prove that
(a) ( :~ )c( :~ t (:~ )B= - 1;
(b) ( :~ )B = I(:~ )
1 B•
66. Establish the connection between the thermal coef-

ficients
1 ( av)
a=y; 7fT P'
where V 0 and p 0 are, respectively, the mean volume and

mean pressure in an arbitrary thermodynamic system.
67. Write the Dieterici equation of state for a real gas
in reduced variables (see Problem 58):
RT ( a )
p= V-b exp - RTV
Do the same for the equation of state
(p + v:/3 )(V -b)= RT.

68. Determine the Boyle temperature for real gases
(i.e. the temperature at which the second virial coefficient
is zero) if a and b are given on the basis of the van der Waals
and Dieterici equations of state.
69. Determine the velocity of a sound wave propagating
in a real gas that satisfies the van der Waals equation of
state.
70. Find the adiabatic equation for an ideal paramag-
netic substance (see Problem 49).
71. Find the adiabatic equation for a real gas.
72. Calculate the difference between the heat capacities
of a dielectric with a constant E (the electric field intensity)
and with a constant D (the electric displacement), i.e.
CE- Cv.
73. Show that ( ~~v )T=T (:; )v·
9-01496
74. Determine the efficiency of two thermodynamic en-

gines working in cyclic processes depicted in Figs. 48 and 49
if the parameters of the cycle e = and p = R ~: are given
and the working fluid is a perfect gas.
p
3
p
-~'
z,
I
I p,
---t-------
1
Pt ---+ + J
' I I
Fig.J48 Fig. 49
75. Using the first and second laws of thermodynamics,

prove the following relationships
(*)T ~' C
( ~) =T, v=
flp s
av
1 ( aT
where a= Vo ) P is the coefficient of volume expan-
1 ( av ) ·
swn,
• ,..,A = -v;;- ap T IS t h e 1sot
· h erma I compress1"b"l"
1 1ty,
and 6 = - J 0 ( ~; )8 is the adiabatic compressibility.

76. Determine the temperature dependence of the e.m.f.
of a galvanic cell in the differential form.
77. Prove the relationship [see formula (25) of Appen-
dix 4)
r ( ap )
ar ) =
( av P- ar v .
E Cv
78. Find the average energy of a rarefied plasma that

occupies a volume V. The plasma is a system of two kinds of
oppositely charged particles (N particles of each kind with

charges +e and -e).
79. Use the result of Problem 78 to _..determine p, S,
and C v for the plasma.
80. Find Cp- Cv in (a) V, T- and (b) p, T-variables.
81. Show that
Cp - CV= [ V- ( )
~! T] ( ;; )V
where H = E +pV is the enthalpy of the system.
82. Prove that for an isotropic dielectric in an external
electric field E,
dF = -S dT - p dV - (P ·dE)
where P is the polarization vector for the dielectric.
83. Find CP - C v for a van der Waals gas.
84. What amount of heat must be transferred to one mole
of a real gas for it to expand from volume vl to volume v2
if pressure p is constant. The equation of state is
(P+ ; 2 ) (V-b)=RT.
85. A process in a system is called polytropic if it
does not change the system's heat capacity C = dQ!dT.
Using this definition, find the equation of such a process for
a monatomic perfect gas.
86. Determine the entropy of a gas whose equations of
state are
V = V 0[1 +a (T -T0)], ( ~~ )T = 0; Cp =constant.
87. Find the adiabatic equation of a gas whose equation

of state is
p =Po (1 +aT- ~V); Cv =constant.
88. Determine the difference between the energy release
of the reaction of formation of one gram-mole of water va-
pour at constant pressure and the energy release of the same
reaction if it takes place without the performance of external
work.
89. Determine the efficiency of a thermodynamic engine
working along the Carnot cycle if the state of the working
9*
fluid is given by the equations

V=V0 [1+a(T-T0 )], ( ~)
ap T =0.
90. Show that when heat is transferred from a heated
body to a less heated one, the entropy increases. Assume
that the temperatures of the bodies equalize and the heat
capacities do not depend on temperature.
91. Find the change in the entropy of a body when it
expands isobarically.
92. Prove that an isotherm cannot intersect an adiabat
twice.
93. In the temperature interval from 0 to 4 °C the coef-
ficient of volume expansion of water is negative. Show that
in this interval water cools under adiabatic compression.
94. Can the relationship Cp = Cv hold for water?
95. The main reason for the fall in temperature in the
atmosphere with growing altitude is the adiabatic expansion
of the upcurrent. Using the adiabatic equation of a perfect
gas, find the change in temperature with altitude.
96. Use the first law of thermodynamics to show that
atmospheric air with a temperature gradient less or greater
than the gradient found in Problem 95 will be, respectively,
stable or unstable in relation to convection.
97. Determine the energy release in the polarization of
a unit volume of a dielectric, neglecting the change in its
specific volume and assuming that
e (T)-1 E
P= 4n ·
98. Show that for a magnetic sample placed in an external
magnetic field H, given the condition that Eo is independent
of H (E 0 is the internal energy of the sample in a vacuum
neglecting the energy of the field there), the following rela-
tion holds for magnetization:
M=f( ~ ).
99. Find the entropy and internal energy of a substance

in an electric field using the following expression for the
Helmholtz free energy

B-1 z
F= F E=o---sn VE.
100. Determine the expressions for Cv, F, S, H, , Cp
for equilibrium radiation.
101. Using the grand canonical distribution, find the
expressions for f.t, p, and S for a monatomic perfect gas.
102. Using the grand canonical distribution, show that
the Poisson distribution describes a system of N noninterac-
ting particles.
103. Find the chemical potential of a perfect gas in an
external potential field U = U (x, y, z).
104. Particles with spin are in an external homogeneous
magnetic field B. What is the ratio of the number of particles
with spins along the field to the number of particles with
spins in the opposite direction?
105. Using the properties of the grand canonical distribu-
tion, show that
pV = kT In Z
where Z is the grand partition function.
106. Prove that N = V ( :P)
It T, V
and f.t = ( !! }
8, V
=
= ( :~ )s, P.
107. Find Cv expressed in terms of T, f.t, and V.
108. Assume that the latent heat 'A is constant and deter-
mine the temperature dependence of the saturated vapour
pressure if the vapour is in equilibrium with the solid
phase.
109. A solution contains N molecules of solvent and
n molecules of the solute (n <t N). Determine the chemical
potentials of the solution f.t and the solute f.t 1 if the chemical
potential of the solvent is f.to·
110. A solution with a concentration c <t 1 of the solute
is in a homogeneous gravitational field. Determine the
change in the concentration c with altitude.
111. Show that the chemical potential of black-body
radiation is zero.
112. Using the conditions of stability of equilibrium,
prove that when a constraint is applied to a dynamic sys-
tern in equilibrium, a change takes place within the sys-

tem, opposing the constraint and tending to restore equi-
librium (the Le Chatelier-Braun principle).
113. Determine the condition for mechanical equilib-
rium of an isolated system consisting of a spherical liquid
drop of radius R and its surrounding vapour.
114. In the following reaction of gases,
H 2 0 +CO =<=± C0 2 + H 2
the equilibrium set in when the temperature was T 0 and the
proportions were: for C0 2 , m1 ; for CO, m 2 ; for H 2 0, m 3 ;
and for H 2 • m~ (the proportions are given in moles). Deter-
mine the affi.nity constant K (p, T).
115. Calculate the critical value of the radius of a drop
of liquid when steam condenses.
116. Prove that a charged drop of liquid will grow even
in an unsaturated vapour.
117. An equilibrium reaction A =<=± J+ + e- (thermal
ionization) takes place in a system. Assuming the tempera-
tures of the ions and electrons to be the same, determine
the degree of single ionization as a function of T and p.
The gases are considered perfect and the energy of ioniza-
tion is E 0 •
118. Calculate the amount of heat given off during a
chemical reaction in which p and T are constant. The af-
finity constant K is known.
119. Determine K and 11Q per particle (see Problem 118)
for the dissociation of diatomic hydrogen
H 2 =<=± H +H
fhe dissociation energy for a hydrogen molecule is 11e =
= 2e~- e~2.
120. Find the osmotic pressure between solutions of
different concentrations that are separated by a semiper-
meable membrane.
121. Prove that the relationship B = - ~ E holds for
free particles with energy e = p 2 /2m that obey the Fermi-
Dirac or Bose-Einstein statistics.
122. A system can be in two quantum states with ener-
gies 0 and e. The degeneracy multiplicities of the states are g1
and g 2 • Find the dependence of S on E and analyze this

dependence.
123. Determine th~ heat capacity of a system of N inde-
pendent. two-dimensional harmonic oscillators each of
which has (n + 1)-fold degenerate energy levels
En = (n + 1) liw (n = 0, 1, ... ).
124. A system has a nondegenerate energy spectrum
e 1 = le (l = 0, 1, 2, ... , n - 1). Determine the average
energy of such a system.
125. Represent the elastic vibrations of a solid in the
Debye model as a phonon gas that obeys the Bose-Einstein
statistics and find its energy. The volume of the body is V,
and the velocity of propagation of transverse and longitud-
inal waves is Ct and c1. respectively. Examine the case of
low temperatures.
126. Determine the heat capacity of an ultrarelativistic
degenerate electron gas (e = cp).
127. Find the thermionic emission current provided the
electrons obey the Fermi-Dirac statistics and the work
function for an electron escaping from the metal is W.
Assume that W- f.t ::» l~:T, where f.t is the chemical poten-
tial level.
128. Find the pressure in an electron gas at T = 0 using
the Fermi distribution.
129. Determine the pressure in a degenerate electron gas
kT
for !1 ~ 1.
130. Using the grand canonical distribution, find the
dependence of entropy Son ;;1 for an ideal gas that obeys (1)
the Fermi-Dirac statistics; (2) the Bose-Einstein statistics.
131. Let g (e) be the single-particle density of states.
Show that the heat capacity of a gas that obeys the Fermi-
Dirac statistics for kT ~ f.to is given as
n2
Cv= 3 k 2Tg (!1- 0).
132. Denoting the energy of a system as E = ~ ni ei,
i
where ei is the energy level of the ith state of the system
and il; is the mean population of the ith state, explain the
meaning of dA and dQ in reversible processes.
133. In the simplest case the spin waves in ferromagnetic

substances satisfy a dispersion equation w = Ak2 , where
k is the wave vector in the spin wave, and A is a constant.
Determine what these excitations contribute to the heat
capacity of crystals at low temperatures.
134. Prove that for crystals the Mie-Griineisen relation
holds true, i.e.
~'\' cv
al= 3V
1 (
where a 1 = 3v av
aT ) P is the coefficient of linear expan-
sion, ~= - t (:; ) T the isothermal compressibility, and

aIn VJ
'V = - aIn v is assumed to have the same value for all
normal modes v 1•
135. Determine Q (E) for a solid at T ~ liw~ax •
136. Find the Fermi level in an intrinsic nondegenerate

semiconductor if the width of the forbidden band has a tem-
perature dependence of the form
Eg = E~- ~T (~ > 0).

137. Determine the concentration of electrons in ger-
manium if the dependence of energy on the wave vector
(the dispersion equation) has the form
Ji2 (kx- k~)2 k~)2 fi2 (kz- k~)2
Ea (k) = Ec + 2mt + fi2 (ky-
2mt + -"'"""2ml---=--
where a = 1, 2, 3, 4 is the number of equivalent minimums
in the conduction band, and mt and m1 are, respectively,
the transverse and longitudinal masses of the electron.
138. In the conduction band of gallium arsenide the
dispersion equation is of the form shown in Fig. 50. Find
the Fermi level for two extreme cases:
(a) a nondegenerate electron gas;
(b) a highly degenerate electron gas.
139. Find the concentration of electrons in a semicon-
ductor if it is known that, when the magnitude of the wave
vector k is small, the dispersion equation has the form

!i2k2
E (k) =Eo+-2- (1-yk2)
mn
where '\' is a constant.
140. Determine the concentration of electrons in a semi-
conductor of the n-type with a narrow forbidden band (in-
----------------------2
Fig. 50
dium antimonide). The dispersion equation for this case is

1 ( -./ 2 2!ik2Eg )
E(k)=Eo+2 + V Ee+ m(O) -Eg
where E is the width of the forbidden band, and m (0)
is the effective mass of the electron near the edge of the
band [m (0) ~ m 0 ].
141. Consider a semiconductor with a width of the for-
bidden band E g and the donor and acceptor levels lying at
a distance of E 1 and E 2 from the lower edge of the conduc-
tion band. Assuming that the electrons in the conduc-
tion band and the holes in the valence band obey classical
statistics, find the chemical potential for this semicon-
ductor. The acceptor concentration is n 2 , and the donor
concentration is n 1 . The effective mass of the electron is mn,
and that of the hole is mpo Consider the case of a donor semi-
conductor (n 2 = 0, E g 2> E 1 ).
142. Prove that for an extrinsic semiconductor the pro-
duct of the hole concentration to the electron concentration
is proportional to the square of the electron (or hole) con-
centration in an intrinsic semiconductor.
143. Determine the heat capacities of electrons and
holes for an intrinsic nondegenerate semiconductor with
the width of the forbidden band being E 0•
144. Find the equation of state for a completely dege-

nerate relativistic electron gas~ with an energy e =
+
= c (m2c2 p2)lf2.1
145. What is the equilibrium ratio of the concentrations
of ortho- and parahydrogen at a temperature of T ~ Tc =
= s!~k , where I is the moment of inertia of a hydrogen
molecule?
146. Derive the Fermi-Dirac distribution from an exami-
nation of particle collision, allowing for the Pauli exclusion
principle.
147. For a crystal in the Debye model determine the
mean square displacement of the atoms of the lattice.
A unit cell of the crystal contains one atom.
148. How do the thermodynamic variables, density and
temperature, for instance, transform in passing from one
inertial frame of reference, K, to another frame, K'?
149. Derive the Planck formula for thermal radiation
in a medium with dispersion n = n (v) (n is the index of
refr::~ction).
150. Find the spin paramagnetic susceptibility of a
system of free electrons if the single-electron density of states
per unit volume is g (E). Consider the case of a weak
f1eld.
151. Determine the magnetic susceptibility of a non-
degenerate electron gas. The energy levels of an electron
moving in a magnetic field B = B z are given by the formula
e= 2 ~n pz+ 2f-I.*B (l + 1/2) + f-tnB (l = 0, 1, ... )
where mn is the effective mass of the electron, and ~tn = 2~(;

is the Bohr magneton.
152. Determine the orbital magnetic susceptibility of
a degenerate electron gas in a weak magnetic field (~tn B ~
~ kT).
153. Determine the sensitivity limit of a mirror-type
galvanometer. The torsion constant of the suspension is a.
154. Find the... fluctuations of pressure in a homogeneous
system that is placed in a heat bath and show that for the
system to be in stable equilibrium the following condition
must hold:
155. Determine the decrease in entropy that develops

during free oscillations of a simple pendulum.
156. Find the correlations 11 T 11p and 11S 11 V.
157. Show that /1S 2 = kCp and 11SI1p = 0.
158. Find the energy fluctuations for a system with two
energy levels.
159. Determine the fluctuations of the number of par-
ticles in perfect gases that obey (a) the Boltzmann distribu-
tion; (b) the Fermi-Dirac distribution; and (c) the Bose-
Einstein distribution. Analyze the results obtained.
160. Show that the following relationship holds for the
grand canonical ensemble:
11HZ= k (!.£)_
iJT
+ f1NZ ( iJN
N
a~ ) 2 •
T
161. Calculate the fluctuations of the position of the

centre of mass for a homogeneous perfect gas contained in
a spherical vessel of radius R 0 •
162. Let q1 and qk be the generalized coordinates of par-
ticles and a 1 and ak the additional generalized forces switch-
ed on at the initial time and acting along the directions qi
and qh. Prove that
-(q-:t--q0-)2 =c 2kT [ i}qt
iJa
J •
a=O
163. Calculate the mean square displacement of a Brow-
nian particle of mass m and ra-dius a, moving in a viscous
medium with a coefficient of viscosity '11·
164. Find the value of the Avogadro number N A if the
mean square displacement for a Brownian particle of mass m
and radius a in time t is 11x2 •
165. Find (z - z0) 2 for a Brownian particle in the gravita-
tional field of the earth.
166. Particles diffuse through a one-dimensional poten-
tial barrier U (x) in a stationary flow. Find the flux density
if the densities of the number of particles in cross sections x 1
and x 2 are known (the Kramers formula).
167. Assuming that the Boltzmann equation for a highly

degenerate gas is
_!j_+v~-eE!L=- f-lo
iJt iJr iJp T
where T is the relaxation time, find the electric conductivity

of metals at low temperatures.
168. For a nondegenerate electron gas find the electrical
and thermal conductivities assuming that along the x-axis
there is a constant temperature gradient, k :~ = constant,
an electric field E is applied to the system, and that T = Av1
(A > 0, l > -7).
169. Determine the tensor of electrical conductivity
for the electrons in a metal that is placed in homogeneous
electric and magnetic fields. Consider the electron gas to be
degenerate.
170. Find the coefficient of viscosity for a gas flow with
a. velocity gradient ~ along the x-axis. The gas obeys classic-
al statistics.
171. Particles of mass m are distributed with a constant
density inside a sphere of radius R. The temperature inside
the sphere is T. At time t = 0 the sphere disappears and
the particles fly in all directions. Neglecting the collisions
between the particles, find the density of the cloud of par-
ticles as a function of time and position.
Answers
SECTION I
mv 2
1. A parabola with the parameter eEO .
2 T _ 2mv0
' - eE '
eJ.tofi
3. A circle of radius v0 /roH, where roH = -m- is the
cyclotron frequency.
4. z= (1-cos ~)cos
eEo [
mffi2 v0
.
rot 0 - (~-sin~) X
v0 vo
X sin rot 0 J;
the z-axis is directed along the field, and t 0 is
the instant of time when the particle enters the field.
5. In a system of coordinates with the y-axis parallel
to H and the x-axis perpendicular to the plane containing
vectors E and H, the path is
X=- voz (1-cosroHt)+ (vox +~sin a) sinroHt
ffiH ffiH mffiH
eEt .
---sma
mffiH
eEt2
Y=~cosa+v 011 t
Z= Voz .
-smroHt-
ffiH
t ( -Vox
ffiH
+-eE- sma
mwH
• ) · (1 -cosroHt )
2
(at t = 0 the particle was in the origin of coordinates).

6. Put a= 0, v 011 = Voz = 0, Vox= v 0 in the solution of
Problem 5, and find x = ...!:Q_ sin roHt. By setting x = l, we
ffiH
can determine T from the relationship ffiHl = sin roHT ~
vo
~ ronT (since ~:l ~ 1) . Substituting T = l/v0 fort in the
expressions for y and z and eliminating v0 , we find that

eE
y=2pz 2 , where p= 2 l2 •
mwH
7. (1) l~ -1 vo 1 Vcos2 asin 2 ~+sin2 a+~cosasin~.
WH WH
Hints to questions (2) and (3): the equation of the projec-
tion of the path on the xy-plane has the form
( y + wHVo •
sm a ) 2 -;-, ( x - Vo
wH cos a sm 1-'
• R) 2
2
= ~ (sin 2 a+ cos 2 a sin2 ~)
wH
Using this equation, we can fmd the points where the

particle comes out of the magnetic wall and construct the
tangents to the path at these points.
8. Hint. First find the coordinates of a specific electron
as a function of time t and of the instant t 0 when the elec-
tron enters the space between the plates. Next find the
instant when the electron hits screen S and the electron's
coordinates at that instant as functions of t 0 • This gives
the equation of the path of the beam of electrons on the
screen in the following form:
x = -~[cos ro ( t 0 +
WVo
_l_) -2 cos rot J
Vo
0
- :~ [sin ro ( t 0 + ~0 ) -sin rot0 J=A cos rot 0 + B sin rot0 •

y= la
wv0
2 [sin ro ( t 0 +_l_)-
v
2 sin rot
0
0]
- :; [cos ffi ( t 0 + : -cos rot J=

0 ) 0 C sin rot 0 + D cos rot0
eVt eV2 d . the distance between the

w here a 1 = md , a 2 = md ; IS
plates.
The path on the screen is the ellipse
(Ay- Dx) 2 +(By- Cx)2 = (AC- BD) 2 •

ANSWERS 143
z =A cos (w 0 t + cp)
where w0 is the natural frequency of the oscillator w1, 2 =
= w;- + -,;- w~ + w: (the z-axis is directed along the mag-
netic field), and wH = e~-tof:l =!!!.... (!lo is the permeability
m m
of empty space).
10. X=C1 + C e -.lt

2
eE
m +-t
y
_.lt
y=C 3 +C4e m
-.lt
Z= C5 +C6 e m
where the constants of integration C1 , • • • , C6 are determin-
ed from the initial conditions. The x-axis is directed along
the electric field.
_.lt
11. x=C 1 +Ae m cos(wHt+a)
_.lt
Y= C2-Ae m sin (wHt+ a)
-.lt
Z=Ca+C,.e m
The z-axis is directed along the magnetic field.

12 T
. =
h
V2g(H+h)
[ 1t
2+arctan2VH(H+h) .
h J
13. V= vr ~g v;: tanh t, X= ~ lncosh v;: t,
V-+ ~ (t-+ oo).
14. x = 6h cos wt, where w __ ,V/ c+nMpr g .

2
15
'V=JI
-,/ 2gRh
R+h'
T=-1
R
v·- R+h
2g
(VRh
R+h R-h)
+ -2- arccos R+h •
16 h = mvo sin a m2g In ( 1 + yvo sin a)

• i' y2 mg
x= mv cos a ( 1 - e-vtfm)
i'
y=_!!!_
i'
(v 0 sincx.+ mg)
i'
(1-e-vt/m)- mgt
i'
v3 sin 2a
S=
2g ( 1 + ;;o sin a )
•
17. F= -mro2r.
mvfib4
Fy = -/i2Y3
19. x = be-a.t cos (rot+ <p)
fo/m
where b = V (w3+a2-w2-2aw)2+4w2 (w-a)2
,
t 2w (w-a)
an <p =- w8-w2+ a2-2aw •
20. The product x 0 X x• 0 must be less than zero, i.e. the

initial velocity ~ 0 must be directed opposite to the initial
displacement x 0 • In addition, the following conditions
must hold:
I=~ I> y- V y2-ro!, '\'~roo.

Rlot2 -. / mVo
21. <p = 4 I V 2e ,
where I = 2pdlR3 is the mo-
ment of inertia of the vanes (for d ~ Z).
22. The space charge can be taken into account by
using the Langmuir-Boguslavski formula (for our case the
result was obtained by C. D. Child in 1911). This formula
is also known as the "three halves power law". (1) The
reaction force is F = ddM
t
v, where M = ..!!!:_
e
Q is the total
ANSWERS 145
mass of the particles, Q the total charge corresponding

2eV ) 1/2 . Then } , = m ( ----;;-
to that mass, and v = ( ----;;- 2eV )1/2 x
7
dQ = e
X dt 2eV )1/2 7··s , w h ere S IS
m ( ---;;- . t h e area of p Iane K .
Substituting the Langmuir-Boguslavski formula j =
=
8 ( e
--geo Zm
)112 ~,
v·1/2
we obtain the final result: F =
= ~ e0V 2Sjd 2 • (2) The power reqnirement is P =IV=

:t )
=jSV= ~ e0 ( 2 112 xV 512 Sjd2.
23. On the midline of the capacitor the following con-
dition must hold:
4T
Tt
_
--,----_
+ r2 ( rt
2e V
+ r 2 ) l n~"2-
+ ef.to }'{ v 2T .
m
Tt
, /-1- , /-m- a312
26. T = 2na3/ 2 V GME = nGmME V 21 E p = 2n REg 112 ,
where ME is the mass of the earth, RE is its radius, and
m is the mass of the satellite.
27. The equation of the path in polar coordinates is
p
p= 1 +e cos aEl
where p = 12 - 26 , e =
a ·v + !~
1
= V-:-1-2::--:~,-/l=2 (l is the ratio of the angular moment1.1m
(F- 2~) , and <J=
to the mass).
(1) The particle will "fall" on the centre if 2~ > l 2 , i.e.
if a is imaginary. The equation of the path is
I PI
P= 1 + e cosh I a I El
(2) The particle will scatter if e ~ 1.
(3) The particle :will be in alternating motion if <J = mjn,
where m and n are integers.
28. <!> = n - ! arc cos(-+) , l = svoo. These formulas
together with the definition of e, e= V 1 + !~ (P-2~),

10-01496
14G PROBLEMS TN THEORETlCAL PHYStCS
give the relation between the impact parameter s and the

scattering angle Cl>. -
29. T = V 2m fJ V E-U
2
dx
(x)
; the motion takes place
XI
between the points x 1 and x 2 ("points of retrogression")
which are the solution of the equation U (x) =E.
:rt -./2m
3o. r=-a v 111 .
31. T= :nV~
aVE+Uo
1 -,/ 2mE -./ a
32. -;:-= V L2+ 2 ma cos crcp, where <J = V 1 + 52 E
(L is the angular momentum and s = L!V 2mE the impact
parameter).
cx:n2
33. e= n (J (8) = -y-
:rt-8
X ez (2:rt-8) sine .
+ ,
35. v = v0 u 1 ln :~t) where m 0 and v0 are respectively
the take-off mass and velocity.
36. For the earth v1 = V gRE ~ 8 km/s, v2 = v 1 V2 =
= V2gRE ~ 11 km/s;
for the moon v 1 = V GmMIRM ~ 1.7 km/s, v2 =
= Vt V2~ 2.4 km/s, where mM = 7.33 X 1025 g is the moon's
mass, RM = 1738 km its radius, and G = 6.67 x 10-s cm 3g- 1s- 2
is the gravitational constant.
2F0 wl\
37. (1) A = mTw3 sm -2-;
•
2F0 .
A =- wT
(2) s m -2- ·'
mw2
(3) A= m~~3 V w2T 2 - 2wT sin wT + 2 (1- cos wT);

(4) A= tr.Fo
mwz '
ANSWERS 147
38. tan e= sin<!>

+ cos <1> ,
m1 1m 2
m 1/m 2 =1.
39. (x/r 0 )Z-k (y/v 0 ) = 1: a hyperbola if the external force
is repulsive (k > 0); an ellipse if the force is attractive
(k<O).
1 1
=-.-+-.
mi+i
1
~ m1
-·· 1=1
MR2 ml 2
2 - +----;r- (qJ 2 +9 2 sin 2 9) -Ed cos e, where e
• •
45. :£ = -
is the angle between the external electric field and the
direction of the dipole moment, d = el, and m = m_;;z .
mv 2 cos2 a .
46. R = 0
r
n, where n IS a unit vector normal
to the surface of the cylinder, and r the radius of the
cylinder.
47. x= rcos ( v0 cos
r
a )
t
. { v0 cos a
y=rsm r
t)
gtZ
z = - 2- + v t sm• IX.
0
48. x = ( a - wg2cos
. 2
. IX) + ·g cos
a ) cosh (ffit sm . 2
a •
Sill a w2 Sill a
148 Pll.OBLEMS IN THEOll.ETICAL PiiYSlCS
3g 3g Mw2l .., I 7g2

49. cp =arc cos 2w2 l , 2w2l ~ 1• R=-2- V 1 + 4l2w4"
50. l= a-b
n-t-1 ·
51 . ~X= _h(mtcosa-t-mzsina).
m-\-mt +mz
l (M -\-2m)
52. (1) x= M-t-Md-m coswt;
(2) R = (M +2m) lw 2 ;
_,/g M-\-Mt-\-m
(3) w;;;:,v Tx M-t-2m ·
53. The decay is possible if E 1 + E 2 does not exceed E.
54. x = 0, y = gw c;s ¢ t 3 , and z = - g~z if the time
of fall satisfies the inequality t X 2w sin 1jJ ~ 1, where 1jJ is
the latitude.
ml2 • •• ••
55. X= - 2- cp 2 + mgl cos cp- mx (t) l cos cp- my (t) l sin cp.
56. q; + w~ (1 +a cos w t) sin cp = 0, where ro~ = f and

aw2 ''
a = g- ; for cp ~ 1, cp + ro~ (1 +a cos wt) cp = 0.
0 0
57. X = m 1a2 (e 2 + w2 sin 2 e) + 2m 2 a 2 e2 sin 2 e + 2 (m 1 +

+ mz) ga cos e, where 0 < e < n/2; the positions of _equilib-
rium are
e0 = 0 an d cos e0 = -g-2 mt + mz ~ 1•
aw mt
58. If cos eo=~ mt + mz > 1' the system is in stable

aw mt .
equilibrium when e 0 =0. If cose 0 <1, the system is in
stable equilibrium when cos eo=~ mt + mz
aw mt
-. l-2-~ -. I 2 ca 2
59. ffit = V ffio + ml2 ' Wz = V ffio- ml2 '
2 g ca2 •
where ro 0 = T + mlZ ,
cp 1 = ~0 (cosro 1t+cosro 2 t), cp 2 = ~0 (cosro 1t-cosro2 t).

ANSWERS 149
60. Hint. The Kirchhoff law for an electric circuit with

an e.m.f. states that
where q is the capacitor's charge.

Since Eq. (1) coincides in form with the equation for
a mechanical oscillator that is under the action of a
constant force and also damped by friction, we can write
the Lagrangian in the following form:
:£ = L~2 _ _qz_ _ re + m~2 _ c (x-l)2

(2)
2 2C (x) 'q 2 2
where l is the length of the spring in the unstretched posi-

tion. We can then obtain an equation of type (1) through
the Lagrangian (2) if we use the Lagrange equation of the
second kind in the presence of dissipative forces Q' = -Rq.
6 1. ffii, 2 = { (C 0 L)- 1 +cfm +J.. [-(-1-+~}2 +
2 2 C0 L m
+ mC4q5l2L J1/2} 1/2 '

2
0
The state of equilibrium ~~ = 0 and ~~ = 0 can be

found if we use Eq. (2) from the solution of Problem
q2 c(x-l)2
60 and set V=2C(X)+ ~q+ 2
62. ffi=v~++·
63. - --. / 2ca2 - J_
(!)- V mlZ l •
· m
Th e ba ll IS · a s t a t e of sta bl e eqm'l'1brmm
· 2caZ
1·f mlZ > Tg .
64. The frequencies can be found from the equation
w•- [-c-+
_M
L
l
M +m] wz+ _c_
M Ml"
.!..= 0
65. In an inertial frame of reference the position of

mass m is
x = x (t) l sin <p, +
y = l cos <p
and the Lagrangian is
ml2 • ••
Z = - 2 - <p 2 + mgl cos <p- mxl sin cp.
66. cp
• •
+ uvp =a cos l't,
2 2 g
where w0 = T, a= - l - .
ay 2
67. Rc= :~ sin2a, Rn=P (t- :h sinasin2a) ,T=

_ Pl •
- 2h cos 2 a sm a.
68. Rn=P, Tn= ~ cota, RA= ~ sin~cota, TA=

p
= 2 cos~ cot a.
l
69. T= mg d+r ,
70 2_ 3g
• (!) - 2 (b3-a3)
b+d b
7 1. -b-=c- Pw=-P.
a
72. F= xl1
el
p
·
73. tan 0.
75. {Lx, Py} = -pz, {Ly, Pz} = -px, {Lz, Px} = -py,
{Lx, Px} = 0, {Lu, Py} = 0, {Lz, Pz} = 0,
{Lx, Pz} = py, {Ly. Px} = Pz• {Lz, Py} = Px•
{Lx, p} = [i X p], {Ly, p} = [j X p],
{Lz, p} = [k X p].
76. {Lx, Ly} = -Lz, {Ly, Lz} = -Lx, {Lz, Lx}
=-LII.
ANSWERS 151
77. Hint. Since cp is a scalar function, it can depend

only on the scalar combinations of r and p, i.e. r 2 , p 2 , and
(p ·r). Thus
acp acp
grad cp = arz 2r + B(p·r) p
acp acp
grad cp = apz 2p + a (p·r) r
The required relation can then be proved by straight-

forward calculations.
78. ad - be = 1.
79. Hint. A vector f (r, p) can be written in the form
f = rcpl + pa
where cp1 , cp 2 , and cp 3 are scalar functions. The required
relation can then be proved by straightforward calcula-
tions.
80. If B does not depend on coordinates, the vector poten-
tial can be written as A = ~ [r X B]; so
r2 ro2r2 e •
X=-z--2 -+ 2m r [r X B],
1 (
Q/8 = 2
e
P- 2m [r x B]
)2
+ -2-.
ro2r2
f.t;z a
81. X= - 2
MR2
-+- 2- - 7
a
and
" p2
o/B = 2M +z;t+-;-'
p2
where M = m1 + m2 is the total mass and !-t = m~t 2 the

reduced mass.
n •2 11 n
~ 1m;ri ~ e;ej ~
82. X = / , - - - - - /, - - 7, e·cp(r· t)
~ 2 8ne0 . ~ r ii ......: ' ''
t=1 1}=1 i~1
n
+~ e;;A (r;, t),
i=1
n n n
[p;-e;A (r;, t)JZ 1 e·e·
&18=~ 2
m;
+-s-"" _'_1 + .......,
ne 0 .L.J ru
'Y, e;cp (r;, t).
i=1 iN i=1
83. Hint. Represent the vector potential in the form

A = [H X r]/2, turn to the coordinates of the centre of
mass and the separation coordinates, and add the total deri-
vative of the function e2~ [r X R]. We then get
X= MR2 + ~;2 +£_-L e (m 2 -mt) (H· [r X r])

2 2 r 1 2Mc
e •
+-(H·[r
c
X Rl).
1 [
84. cfJ8=u Pe+
2 (P¢- Pq, cos 8) 2
sin 2 e
J+ 213p~ +Mgl cos8.
. . . . .
85. [l + 1-t (x 2 + y2)] cp + f..t (xy- yx) = f..t (x 0 y0 - y0 x 0 ), where
I is the moment of inertia of the disc.
86 Rex~ 2
• {jl= 2(I+~R2) t'
t _
<o-
MR [.ex
-Mcos 2R+2(I+~R2) t'
Rex~ J 2
an d 11 = - M R sm
-xr- . [. ex
2R
+ Rex~
2 (I+ ~R 2 )
J
t2 , wh ere cp 1s
· th e
angular displacement of the disc and ~ and 11 are the Carte-

sian coordinates of the disc's centre of mass relative to the
centre of mass of the system.
3
87. (1) It= I 2 = 13 = 8 mHa2 ;
(2) It= 2 ~ 0 a 2 cos2 a, l 2 =2mHa 2 sin 2 a,

I 3 =It+I2;
(3) lt=I2 =
3mNmH
M
( 2
b -3
a2 )
+-2 - ,
mH.a2
l 3 =mHa2 ,
where M is the total mass of the molecule.

88. The inertia tensor will have the form
~-ta2 0 0
lu= 0 ~-ta 0 2
0 0 0
in the coordinate system whose z-axis passes through both
atoms. ·
89. a= 0.8 A..
ANSWERS 153
1 2
90. The moments' of inertia are related as2 : 3 : 1,
and the frequencies <?f vibrations as V312 : 1. V2 :
M 2 M 2 2
91. (1) Izz = 12 (b +c ), Ixx=u(a +c ), Iyy=
2
M
=12 (a2+b2);
- - I zz--5-
(2) Ixx-Iyy- -2M R2·'
(3) I XX= I yy = ~~ ( R 2 + ~2 ) , I zz = 31~ R 2 ;
(4) Ixx= ~ (b2 +c2), Iyy= ~ (a2+c2),

M
Izz = T (a2+ b2);
2M D5-d5
(5) Ixx=fyy=fzz =-5- X D3-d3
(6) I xx =IYY = ~ (R 2 + !r 2 );
M 1 1
(7) Izz=T(R2+r2), Ixx=Iyy=12l2+4 (R2+r2);
(8) I zz -_!!_M
12 a2 ' I xx-
- IYY --12
M l2+!!_ 2.
24 a ,
5 M 2 I
9 I zz=f2 5M 2
M l2 + zra.
() a' xx= I yy=12
92 • I a dwa
dt = N·, ffia = N t+ ffio,
-y;
ffi 2 = A cos ffi [ + 1;o )2 + qJ]
( t
ro 1 = A sin ro [ ( t + 1;o )2 + qJ J
(1 -1) N
where ffi= 32II 3 and ffi 0 , qJ, A are constants of integra-
tion.
93. In securing the body in such a way the angle e equals
n/2. Then, by the kinematic equations of Euler (I-24), w~
have
.
ffi 1 = lJl sin (jl, ffi 2 = ljJ cos (jl, ro 3 = cp
In our case we can write formula (1-31) for the kinetic

energy of rotation:
1 • 1 •
K = 2 (/ 1 sin 2 qH- / 2 cos 2 cp) 1p 2 + -zl cp 3 2 (1)
From the fact that ¢ is a cyclic (ignorable) coordinate it

follows that
az
-.- = -aK. = (I 1 sm
. 2 •
cp +I2 cos 2 cp) 1p = c =constant (2)
aw D¢
. .
Since K is constant, we fmd cp and 1P from (1) and (2) as
functions of angle cp, the parameters of the system, and the
two integrals of motion K and c.
1 • 1 • •
94. K = -z/1P2 sin2 a + 2:/ 3 ((p +¢cos a) 2 •
3 •
95. :£ = 4 M (R - a) 2 cp 2 - M g (R - a) cos cp. The mo-
t_ion is the same as that of a simple pendulum of length l =
3
= 2 (R- a).
96. Since a = 0 and cp = 0, it follows from the kinema-

tic equations of Euler (1-24) that
Wx=O, W1=0
wy = 0, w2 = 1P sin a
Wz = 1p = w, w3 = 1p cos a
The dynamic equaLions of Euler (1-32) yield
N~ =I d;; 1 + (! 3 - I) w2 w3 = (! 3 - I) u;: sin 2a
Using formula (1-21) for the rotation matrix, we get

Nx= L~ cos 'ljJ, Ny= L~ sin 1p
Since the moment of forces is the sum of the moments of
the equal forces of inertia that are applied to points 0
and B, we have
N?i·= -hFy, N 11 = -hFx
ANSWERS 155
Whence
Mw2 . ( zz R2 )
Fx = - ---;g;- sm 2a 12 - T cos wt
M w2 . ( zz
F Y = ---;g;- sm a - R2 ) sm
.
wt
12 4
if for the moments of inertia we use the answer to Prob-
lem 91.
97. I d~t + (I 3 -I) WzW3 = mgl sin e cos cp
I d~z +(I- / 3 ) w1w3 = - mgl sin 0 sin cp

I
3
dw3 =O
dt
where lis the distance from the point of support to the centre
of mass.
98. Using the answer to Problem 97 (we set l = 0), we get
I dd~ 1 + (/ 3- I) w2 w0 = 0, w3 = w0 =constant
I d:z + (/- / 3) W1Wo = 0

whence
w1 =A cos (wt +a)
13-1
w2 =A sin (wt +a), where w= --~- Wo
The angular momentum L is constant since there are no

external forces. To find the Euler angles as functions of time
it is convenient to use a system of coordinates whose z-axis
coincides with vector L. Using formula (I-21), we get
L sin 8 sin cp = L 1 = lw 1 =!A cos (wt +a)
L sin 8 cos cp = L 2 = lw 2 = IA sin (wt +a)
L cos 8 = L 3 = l 3 wo
From the last equality it follows that
e. = arc cos I 3 w0
M
From the first two equalities we find that

£2 sinz 0-
- J2 A2 , A-_!_
- I Y£2- J2ro2
so
qJ =rot+ a, lJl =(roo- ro cos 0) t + 'i'o·

99. x = cos lJl cos qJ x' - sin qJ cos lJl y' + sin lJl z'
y = (cos 0 sin qJ -sin 0 sin lJl cos qJ) x'
+(cos 0 cos qJ +sin 0 sin lJl sin qJ) y' +sin 0 cos lJl z'
z = (-sin 0 sin qJ - cos 0 cos lJl cos (p) x'
+ (-sin 0 cos qJ +cos 0 sin lJl sin qJ) y' +cos 0 cos lJl z'.
100. In the system of coordinates attached to the body
. .
ro x' = "' sin qJ - 0 cos "' cos qJ
. .
ro u' = lJl cos qJ + 0 cos lJl sin qJ
roz• = qJ - 0 sin lJl.

101. The kinetic energy of the body of the coach is
1'62
Kt 2-
=-
where I' is the body's moment of inertia about an axis pas-

sing through point 0 perpendicular to the drawing; the kine-
tic energy of the flywheel
K 2 = M 2 h~0· 2 +-z
I
(tp· 2 + 0· 2 cos 2 1Jl) + -I2~ (qJ-
·
0· sin 1Jl)2
the kinetic energy of the counterbalance
Ka = TM • • •
(x~ + Yi + z;) = ---.J-
M •
[h~t1J + (h + h
2 1 2
•
cos 1Jl) 2 fPJ
where x 3 = h 2 sin lj;, y 3 = (h1 + h 2 cos ljl) sin 0, Z 3 =

+
= (h1 h 2 cos lJl) cos 0 are the coordinates of the counter-
balance's centre of mass in an inertial frame of reference
with 0 as the origin.
The potential energy of the body in the earth's gravita-
tional field is
ANSWEltS 157
The potential energy of the flywheel is

v2 = M2gh1 cos e
The potential energy of the counterbalance is
V3 = M 3 g (h1 +h 2 cos 'ljl) cos 8
If we construct the Lagrangian and then write the Lagrange
equation, we see that cp = w0 t and 0 = 1jJ = 0 are its solu-
tion. To see whether this solution is stable we construct the
Lagrangian of small deviations:
cp
JJ
1
=2a11 e'2 + 21 a221Jl• 2 - b21 e· 1Jl + 2c11
1 82 + 2c221Jl
1 2
where
a11 = I' + M 2h~ + I + M (h + h 3 1 2) 2
a = I + M 3h~, b = I w
22 21 3 0
c11 = g [M l + M h + M (h + h )l, c =
1 2 1 3 1 2 22 M 3 h 2g
The solution cp = Wot, e = "' = 0 will be stable if the
following inequality holds:
I2wo > V c11a22 + V C22a11.
102. The centre of gravity is located at a distance l from
point 0:
l= ~X P1-P2 aa
3:n: PI +P2
The state of equilibrium can be found from the equation
2 e4- cr2
cot cp = cr2 (1- e2)
where e is the eccentricity of the ellipse. The equation has
a solution if e2 ;:;;:::: a. Besides, there is always one stable
state of equilibrium cp = ~ and one unstable state cp = 32:n: •
103. In spherical coordinates (r, e, cp)
1 8UfJ + Ur
Urr = Br '
8ur
UfJfJ =7 ---as -r-
1 8U!fl UfJ Ur
U<l'q; = r sin 8 acp + -r- cot O+ -r-
158 PROBLEMS IN THEORETICAL PltYStCS
1 ( ou
Sill
,-
2u
rS
= ouo _ .!!:.Q_
or r
+_.!._r our
oe '
2u = - - 1 - Bur
<rr r sin e acp + Burp-~
or r
In cylindrical coordinates (r, cp, z)
1 OUcp u
Uq;{j)=---+-r
r acp r
1 OUz OUcp
2u.~z =--a-+-,-
r cp uz
2u
rrp
au ---.!.
'~"·
= __
ar
u
r
+-1r aocp __!!z_
.
104. Into Eq. (1-50) we must introduce an additional

term, the centrifugal force of inertia pw 2 r. This force will
obviously cause the displacement of the particles only in the
radial direction, i.e. along r. Projecting the equation onto
the unit vector er (see formulas (3-4), (3-18), (3-25) and (3-27)
of Appendix 3), we obtain the equation
E (1-a)
(1-f-a) (1-2a) dr
d [ 1 (
r rur
>] -_ -pw r 2
which has a solution that satisfies the condition CJrr = 0

for r = R:
= pw 2 (1-f-a) (1-2a) [( _ ) R 2 - 2]
Ur 8E(1-a) r 3 2cJ r
where E is the Young modulus and a is the Poisson ratio;
both are related to the elastic constants by the equalities
E- t.t(3A.-f-2f.l) _ A.
- A.-f-f.l ,c:J-2(A-f-f.l)
· 105. One transverse wave with a phase velocity of x =

= V c~4 and two transverse-longitudinal waves whose pha-
se velocities satisfy the equation
x 4 - .!_ (cu
p
+C44) x 2
+p2'1 [ CuC44-. (cu+c12l

4
( C12-cH + 2C44) sm· 228] = 0
ANSWERS 159
where Cw c12 , and c44 are the elastic constants of the cubic
lattice and 0 is the angle between the wave vector and the
4-fold axis of symmetry. For e = 0 or e = ~ the waves
propagate parallel to the faces of the cube.
106. el =eo, sin et =_:!._sin eo,
CJ
=I cr sin 28t sin 280 - cf cos2 20t 12

Ri c~sin20tsin28o-f-c[cos 2 20t
l' -1
lt-
2CJCt sin 28 0 cos 28t
+
ct sin 20t sin 200 c12 cos2 28t
12 X ct cos Ot
CJ cos 80
where e1 is the reflection angle of the longitudinal wave,

et is the reflection angle of the transverse wave, c1 and ct are
the respective velocities, and R1 Rt = 1. +
Hint. The boundary conditions are of the form (1-51).
107. 01 =eo, sin el = 2 sin eo,

Ct
= I c~ sin 201 sin 28 0 - c[ cos2 200 12

Rt c~ sin 201 sin 28o+ c12 cos2 28o
where, as in Problem 106, Rd- Rt = 1.

dp tJ.p
108. ay= - -z-= constant and vy = - ZIJl z2 - 4
tJ.p ( d2) ,
where f...pjl is the pressure drop per unit length.
109. (1) j = ( ~P )
2
2~~ , T = T0
+ ( tJ.t ) 2
1ilJx ( ~~ - z4 )
(2) j=(.E!...)
l
2 !-!_,
61]
T=T 0 +(~)
l
2~
96xl]
- ( tJ.p
l
)2 ~z-(
12XT]
tJ.p
l
)2_1_z".
6T]X
f60 PROBLEMS IN THEORETICAL PHYSICS
Hint. Show that in our case de/dt = 0 and use Eq.

(I-38).
dp v~ •
110. (1) dr= p - r ,
(2)
(3)
and Vz = V = 0. 7
r~r~
111. T=T 0 +-x m 1J 2
2 ( 2_
r1 r~
2) 2
r
ln-
r2
1 X
r2
ln--.
r1r2
112. In cylindrical coordinates

(.!!!...)
dt e -
- - .i.p ~
or
r
iJ vr 1 2 iJ 2vr iJ vr 2 1 iJvr 2 iJvcp Vr J

+v [ a;:2+~ iJqJ2 +azz- +-;-a;:--~ ---acp--;:2
( ~) - _.i_.i_~
dt eiP - p r iJqJ
+ v [ --a;2 + -;:2
iJ2vcp 1 iJ2viP
aqJz
iJ2viP 1 iJvlfl
+ 7fi.2+ -r ar + -;:2 aqJ - --;2
2 iJvr Vcp J
( ~) = _ .i_ ~ + V[ iJ2vz + _1_ iJ 2vz + iJ2vz + ...!._ iJvz J
dt ez p oz iJr2 r2 iJqJ2 iJz2 r or
In spherical coordinates
(.!!!...)
dt e 7
= _J._~+v[.i.
p iJr r
iJ2(rv 7 )
iJr2
+-1-
r2
iJ2v 7
iJ62
2 iJvlfl
r2 sin 6 iJqJ
- 2vr - 2ctg6 ve]

r2 r2
( ~) __ .i_ .i_ ~ + v [- ...!._ iJ2 (rve) + _1_ iJ2ve

dt ee - p r ae r iJr2 r2 iJ62
1 iJ2ve + cot e iJve - _1_ cot e iJviP

+ r2 sin2 6 iJqJ2 r2 iJ6 r2 sin 8 iJ8
+2._ r2
iJvr _
iJ8
va
r2 sin2 8
J
ANSWERS 161
( ddvt ) = - _!_ f. ~ + V [_!_ {J2 (rvcp) +_.!._ o2vcp

eq> p r sin 0 orp r or2 r2 ofl2
1 il 2vq> cote ovq> 2 OVr
+ r2sin20 orp2 +~ae-+ r2sin0 arp
2 cos e
+ r2 sin2 0
ove
orp -
vq>
r2 sin2 0
1 '
113. The frequencies can be found from the equation
wd
ta n . cp1
-=l- (1)
c CJP
where p1c is the product of the liquid's density and the

velocity of the waves in it, pc 1 is the same quantity for
the plate. Since usually cp 1 is considerably smaller than
c,p, we can write Eq. (1) approximately in the form
nllCJ j- . cp1 h
Wn=-d-- l pd , w ere n= 1 , 2 , ... , oo
These frequencies are complex quantities due to sound

attenuation in solids. The physical meaning of this is that
the energy of the sound wave is radiated into an unbounded
medium (the liquid).
114. w = xctk, where x is a real root of the equation
c2 c2
~6 - 8~4 + 8~2 ( 3 - 2 cf ) -16 ( 1 - cf ) = 0
The Rayleigh wave has two parts-the longitudinal part
and the transverse part. The ratio between them is
uz 2-s
Ux 2V 1-62 '
115. Since the radial vibrations are longitudinal,
curl u = 0. Thus u can be represented as u = grad <p.
In this case <p satisfies the equation ~ = c~~cp. For radial
monochromatic vibrations the equation takes the form:
1.
r2Tr r dr =
d ( 2 drp)
-cr<p w2
The solution of this equation is

A sin kr i(J)t
<p= -r-e
11-01496
162 PROBLEMS IN THEORETICAL PHYStCS
The boundary condition arr (R) = 0 yields

wR
tan--
CJ 1
wR -1-(!!!!!_)2
2ct
The solutions of this equation are the natural frequencies of

vibrations of the elastic sphere.
116. We seek tho solution in the form of an outgoing
spherical wave
cp=A-e_ _ __
i (WI- :t r)
r
By the boundary condition we got
') . ·• 21/2
l•l=~ + 2ct ( 1 _
rill - /{
_2_}
cr
The vibrations are 1hmpcd since the energy dissipates in
the form of sound wa vos.
+ 1,2 + d 2 , where n, m, and l are
m2
117. w2 = c2n 2 ( az
n2 z2 )
integers that change from 1 to oo, and c is the velocity of

sound in the gas.
118. Hint. We must pass over to the frame of reference
where the sound source is fixed. In this frame the field of
velocities has the form v = v0 + v'. We can now obtain
the wave equation using the motion equation
Dv' + ( ") , _ grad p'
---at vo. v v - - Po
the equation of continuity
h' +(v 0 ·gradp')+p 0 div v'=O
and the thermodynamic equation p = p (p), which relates

the density and tho pressure.
The dispersion equation has the form (w - kv 0 ) 2 = c2 k 2
where c2 = ( ddp} is the square of the velocity of sound.
P P=flo
ANSWERS 163
The frequency registered by the receiver is
v ) -1
ro'=ro ( 1+--f-cose
where e is the angle between the vectors k and v 0 , and {>) is

the frequency of the sound waves generated by the source.
119. ro' = ro ( 1 - v:
cos e) . The notation is the same
as in Problem 118.
120.
121.
where a and b are the semiaxes of the ellipse.
122. T = T0 + ( -l-
1'1. p ) 2 Rl
64Xl]
[ 1- ( 7f
r ) t,
. J
SECTION II
1. grad f (r) = ddt !... .

r r
2. divr=3, curlr=O, curlcp(r)r=O.
3. grad (P·r)=P, grad (P·r) =..!._- 3 (P·r) r (P·V')r=
r3 r3 r5 '
=P, div[Pxr]=O, curl[rxP]=-2P.

r . •
4. grad A (r) B (r) = -r (AB + BA), div cp (r) A (r) =
. .
= ..P..(r·A) +_p_ (r·A), curl cp (r) A (r) =..!. [r X A]+~ [r X
r r r r
' dA ' dB • dcp
xA], where A=ar• B=ar• and cp=Tr·
11*
164 P:il.OBLEMS TN TIIEORETTCAL PHYSlC~
5. We multiply the sought integral by a tonstant vec-

tor p:
(p·l) = p ~ r (A·n) dS = ~ (p·r) An dS
=~ div (p·r) A dV = ~ A grad (p·r) dV
=) (A·p) dV = (A·p) V
Since p is arbitrary,
I= AV
In a similar way we can show that
~ (A·r)ndS=AV.
8. cp=3.-+b.
r
9. In solving this problem make use of Appendix 3.
In cylindrical coordinates the Maxwell equations are:
_!.._ 8Hz _ BH~ _ . + 8Dr
r oq> az - Jr at
8Hr _ 8Hz _ . + aD~

az ar - ] Jz
+ BDz
at
1 BEz BE~ 8Br
-;-a;p----az= -at
BEr BEz IJB~
---az- ar = ----at
_!_~(rE)-_!_ BEr =_ BBz
r ar <r r oq> at
_!_ ~ (rD ) aD~

r ar r + _!_r oq> + aDz
az
-
- p
_!_~(rB)+_!_ aB~
r ar r r oq>
+ BBz
az
=0
ANSWERS 165
In spherical coordina_tes:
1 {
r sine . asa ( sm. SHq; ) - aq;-
aHa } + ----at
= aDr .
lz
_!_ { _1_ aHr _ _!..._ (rH)} _. + aDa
r sine iiq> iir ]fJ
'I' - at
r1 { a
Tr(rHa)-----ae
aHr }
=
. aD!jJ
J!jl+---at
1 { a ( . oE ) aEa } aBr
r sine as .
sm (I q> . - ~ = --at
_!_
r
{-1-
sin 8
aEr _
8q>
_!_ (rE ) } = _ aBa
or !jJ at
r1 { a aEr }
&r(rEa)-----ae = -----;;e
aB!jJ
- 1 _!..._ (r 2D )
r2 or r
+ -r sin1- 8 _!..._
&8
(sin 9D 8 ) + 1
r2 sin2 8 ~
iiD!jJ
= p
~~2 ; (r2Br) + r sfn 8 :e (sin 9Ba) + r2 si~2 8 ~:!jl = 0.

1
3eo pr for r<R
{
10. E=
__1_p~r for r> R.
3e 0r3
11. From the superposition principle for fields it follows

that the sought electric field is equal to the difference be-
tween the electric field intensity of a solid sphere and that
of the charges that in this case are inside the cavity.
The electric field intensity inside the cavity is E =
1.
=-·3-.pa;
eo
1
inside. the sphere but out.side the cavity E = - pr-
3eo
· 1 R'3
+- 3eo p Ir-a 13 (r- a);
-- -· ,_ 1 [ R3 R'3
outside the sphere E=-
3e-0 (:> -3-r~ I r-a 13 (r-a)J'
_ r
where a is the radius vector connecting the centres of the
sphere and cavity.
12. E =
{" ---r
Bo n+3
a
~
Rn+3
for r< R
- - r for r> R.
e0 (n+ 3) r3
X
13. E =
{ 2rre 0RZ
X
2ne 0 r2
r
r
for r<R
for r>R
where r is the distance from the cylinder's axis to the

observation point.
14. Let us direct the z-axis normally to the layer and
select the origin of coordinates in the middle of the layer.
Then
E = £..:.. inside the layer
eo
E = 2pd -1z 1 outside the layer.

e0 z
15. (a) C = 4_j;eR:

2-
2 ,
1
where R 1 and R 2 are the radiuses
of the capacitor plates (R 2 > R 1);
(b) C = 8: , where S is the area of one capacitor pla-
te, and d the distance between the plates;
(c) C = 4 rrel , where l is the length of the capacitor,
l R2
n Rt
and R 1 and R 2 are the radiuses of the plates.
16. Let us assume that the charge per unit length for
the first conductor is x and for the second -x. The poten-
tial of each conductor is composed of the potential q>1
created by the conductor's charges and the potential q> 2
created by the charges of the other conductor. This second
term can be considered to be the same at every point of the
conductor provided the distance between the conductors
is large. For the first conductor we then have
q> 1 = x ln R 1 + -2 x ln d
-2-
m>o ne0
ANSWERS 167
for the second

X X
cpz = -2 -ln R 2 - -2- In d
ne 0 ne 0
The capacitance per unit length of this system is
C= x 2ne 0 ( I d )-1
<rt- <"P2 - d2 = :n:eo n R
ln RtR2
where R = V
R 1 R 2 is the geometric mean of the radiuses
of the conductors.
17. In cylindrical coordinates with the z-axis directed
along the line that connects the two charges, the equation of
the lines of force is
(1)
In our case
E _ er er
r - 4ne 0 [r2 + (z- d/2)2]~/ 2 4ne 0 [r2 + (z + d/2)2] 3 ' 2
Er.p=O (2)
E _ e(z-d/2) e(z-j-d/2)
z- 4:rte 0 [r2-j-(z-d/2)2J 3 ' 2 4ns 0 [r2-j-(z-j-d/2)2] 3 12
Substitute the components of the electric field (2) into (1).
Now go over to new variables u and v instead of r and z
in the following manner:
d
r=-- (3)
u-v
We come to the equation
dv du
(4)
(1 + v2)3/2 (1 + u2)3/2
Integrating Eq. (4), we obtain'
v u
c
or
z-d/2
The pattern of the lines of force is given in Fig. 51.

18. Let the x-axis be directed along the field. Then

W = - I E I (x + iy) = <p + i'l',
where <p = - I E I x and 'I' = - IE I y. Such a field is
created by a uniformly charged surface y = 0, for instance.
In this case I E I = a/(2e 0 ) •
19. Let us use a conformal mapping W = Az2 i to trans-
form the angle into a half-plane, for which we have the solu-
tion from Problem 18. The
complex-valued potential is
W = <p + i'l' = Az2 i
=A (x + iy) 2 i
Whence
<p = -2Axy and
'I' = A (x2 - yz)
Fig. 51 Fig. 52
The equipotential surfaces and the lines of force are de-

picted in Fig. 52.
20. The complex-valued potential W = <p + i'l' = VZ.
Hence
X + iy = <p 2 - '1' 2 + 2i<p'lf
<p2 - '1'2 = x, 2rp'l' = y
Excluding <p and 'I', we obtain the following equations:
y2 -m2. ,.
4qJ2 - 't' -.,.,,
ANSWERS 169
Let us find the lines cp = constant. If cp = C, we come

to the parabola
y 2 = 4C 2 ( C2 - x)
which is turned in the negative direction of the x-axis and
whose vertex is at point x = C. If C = 0, the parabola turns
into a straight line (semiaxis)
y = 0, X< 0
Thus the function cp = Re Vz gives the potential near a
grounded half-line y = 0, x < 0. The equipotential surfaces
are a system of parabolas.
21. Writing z in the form z = rei 8 , we find that cp =In r.
The equipotential surfaces are a system of circles with
radiuses r = constant. In the three-dimensional case such
a potential is created by a charged straight line that lies
along the z-axis.
22. Let us write the equation of the parabola in parame-
tric form
x ---,- ap 2 - a, y = 2ap
where p changes from -oo to +oo. According to Eq. (II-38),
z = aW 2 -a + 2aiW = (W + i) 2 a
Hence
W = Vz;a- i
The potential is determined by the imaginary part of this
function.
23. Writing the equation of the ellipse in parametric
form x = a cos 0, y = b sin 8 and assuming that a =
= C cosh a, b = C sinh a (a 2 - b2 = C2), we obtain
z = a eos W + ib sin W = C cos [1¥ + i (cp - a)l
This suggests that
x = C cos 1f cosh (cp - a)
(1)
y = C sin 'I' sinh (cp - a)
Eliminating 1f and setting cp = constant, we come to the
equation of the equipotential surfaces:
x2 y2
czcoshZ(<p-a} +czsinh2(qJ-a) = 1
This is a system of ellipses whose focuses coincide with the

focuses of the grounded ellipse.
Eliminating cp from (1) and setting "If = constant, we
come to the equation of the lines of force
y2
cz cosz '!' C z·z'¥=1
Sill
This is a system of hyperbolas whose focal distances are the

same as for the ellipses. When a = b,
z =a (cos W + i sin W) = aei<'l'+i!p)
x = ae-!p cos "If
y = ae-!p sin "If
x2+y2
whence e-(j) = az , (p = - 2ln r/a.
25. Using Eq. (11-39) for the potential of charges on a

surface in cylindrical coordinates, we have
2:n: R2
(p = r dcp Jr
J 4:rte 0
cr~r
r2+z2
= _2cr
eo
<V R~ + z2- R~ + z2) v
0 R1
where z is the coordinate of the observation point on the

axis. The components of the electric field intensity vector
are
For the limiting cases we have:

(a) E _ cr ( z z ) ;
z- 2eo TZT- V R~+zZ
cr z
(b) Ez = 2eo TZf ·
26. For a system possessing spherical symmetry the poten-
tial is
r oo
cp (r) = - 1-
eor J\ p (r') r'Z dr' + _i r r' p (r') dr'
eo J
0 r
ANSWERS 171
Substituting the known charge density, we obtain

ffJ = _e_ (1- e-2r/a}- _e_ e-2r/a.
4:n:e 0 r 4:n:e 0a
27. The potential of a point charge is given by the solu-

tion of the equation
L\rp= _ _
e 8(r) (1)
eo
Let us represent rp (r) and 8 (r) in the form of Fourier ex-
pansions:
rp (r) = j qJ (k) eikr dk }
(2)
8 (r) = _1_
(2:n:)3 Jr eikr dk
Substituting integrals (2) into Eq. (1) and equating in the
integrands the coefficients of eikr, we obtain the expression
for the Fourier transform of the potential:
qJ (k} == Eo (2~)3 k2 ·
28• qJ =
Po
Eo (a 2 +b 2 +c 2)
· by sm
· ax sm
sm · cz.
29. The solution which satisfies the Laplace equation

and the boundary condition (ffJ = 0 on the conducting
plane) is
e e
qJ = 4:n:e 0 r - 4:n:e 0 r'
E = _e_r_ _ er'
4:n:e 0 r3 4:n:e 0 r' 3
where r=V(x-d) 2 +y 2 +z 2 and r' = V(x+d) 2 +y2 +z2 •

We assume that the grounded plane is plane x = 0. The
radius vector r is drawn from the charge to the observation
point, and r' from the charge's image, i.e. from point x =
= -d, y = z = 0, to the observation point.
The surface density of the induced charge
ed
(J = - -:::2-:n:':;'lf/7"(y::;;:2:::;:+=z::;;:2:::;:+=d~2);:;;3-
The total induced charge is

00
i i ed dx dy = _e.
- J-ooJ 2n V (y2 + z2 + d2)3
30. qJ 1
= -4neo (..i...
rt
_ _!_+_!_
r r 2
_ _L),
r 3 4
where r 1 is the
the distance from the observation point P to the charge

(Fig. 53).
31. The potential inside the sphere and on it is zero.
Fig. 53
Outside the sphere it must satisfy the equation

i1qJ = - _:__ 6 (r)
eo
The origin of coordinates coincides with the position of the
point charge. The solution of this equation is sought in the
form
e e'
(jJ = 4.ne0 r - 4.ne 0 r' (1)
ANSWERS 173
Vector r' is drawn from the observation point to a point

inside the sphere. The position of the second point is deter-
mined from the boundary condition rp = 0 on the sphere.
Fig. 54
The second member in the right-hand side of (1) may be

thought of as the potential of a charge e' (the image) inside
the sphere. Actually there is no such charge. But the real
charge induced on the surface of the sphere acts in the
same way as a certain charge would without the sphere.
The symmetry of the problem implies that charge e' must
lie on the line that connects the centre of the sphere with
charge e placed at a distance d1 from the centre. From the
boundary conditions it follows that (Fig. 54)
e2. r2 RZ+ d'l- 2dR cos e
72 =72= R2+dr-2d1Rcos e
This condition holds for any angle e if
R 2 = dd 1 , and e' = e V dtfd
where d1 determines the position of e'.
32. rp = -4neo r r
+
1 (.!... - e: .=_:)
r0
, where r and r' have the
same meaning as in Problem 31, and r0 is the distance
from the centre of the sphere to the observation point.
If we consider only the first two members in the parentheses,
the potential on the surface of the sphere is zero. The third
member yields a constant value to the potential on the
surface. The sphere remains neutral. The potential may
be interpreted as having three terms: one from charge e,
174 PHODLEMS IN THEORETICAL PHYSICS
another from charge -e' in the conjugate point, and the

third from e' in the centro of the sphere.
1 ( e e' e' e ) .
33. (jl = Mteo -;:-;---,:;-+r;--~ (Fig. 55); here
-v
e' = e dJid and dt = R 2 /d
In Fig. 55, P is the observation point.
34. Each charge of the dipole induces an image charge.
Since the distances from the charges to the sphere are di-
fferent, the magnitudes of
the image charges will al-
so be different. For this rea-
son we must place a dipole
p', oriented in the same way
as p, and a charge e' at a
point Lhat lies at a distan-
?::_.....,-~~~:gk--4e ce d' = R 2 /d from the spho-
-e re. We must also require
that the potential
pr p'r'
(jl = 4ne 0 r3 + 4ne 0 r'3
e'
+ 4rteor' ( 1)
Fig. 55
of the system be zero on the
surface of the sphere. We
must remember (see Problem 31) that on the surface of the
sphere
r= VR + d 2 2- 2dR cos 8
r' = V R + d'
2 2- 2d' R cos 8
pr = p ( d- R cos 8)
p'r' = p' (d'- R cos 8)
From the condition that cp = 0 when r = R we fmd that
p(d-Rcos8)- Rd33 p' ( Rcos8- 7R2 )
e'd
+y(RZ+d2-2Rdcos 8)=0
ANSWERS 175
This condition must hold for any 8. lienee
pd+ ~ PI+ e~d (R2+ d2) = 0

d3
pR + RZ p 1
+ 2e d 1 2 =0
Solving these equations for e and p we have
1 1
I R d I R3
e = - {j2 p an p = d3 p
These relations together with (1) give the potential of the
system.
35. In the medium with tho charge the solution will be
songht in the form
e e'
fPt "= 4:rte 1e0r - 4:rte 1e 0 r' (1)
Vector r' is drawn from a point that is the mirror image

of the coordinate of the charge in the interface of tho two
media.
In the other medium
e"
(2)
The electric displacement vectors in the two media are

determined by the relations
er e'r' e"r
4nD 1 = --,:3 - ---;:;3 , 4nD 2 = - 3- (3)
r
From the continuity of the potential (rp 1 = rp 2 when
r = r 1 ) and the normal components of the electric displace-
ment vectors (D 111 = Dn 2 when r = r 1 ) we fmd that
e-e ' e"
and e1 + e= e"
Whence
e' = e 2 - 101 e and e" = ~ e
e1+e2 e1+e2
By substituling e' and e" into Eqs. (1)-(3) we can find
the potential and the electric displacement vector in any
point of space.
17fl PROBtEMR IN TIIEORETICA L PIIYRICS
36. We seek the potential in the form

a
qJ=-
r
which satisfies the boundary conditions cp 1 = cp 2 and

e1 ( ~ ) 1 = e 2 ( :~) 2 on the interface of the two dielec-
trics. The constant quantity a can be related to the charge
of the sphere
e= ~ <J dS = - Jee 0 :~ dS
Here e = e1 in medium 1 and e = e 2 in medium 2.

After integrating the last expression we find that
whence
4 D1 2et er 4.,.D 2 = ~...:!.
n = Et + e2 ra ' .. Et +
e2 r3
ete <J _ e2 e
<Jt= 2:n:(et+e 2)R2' z- 2:n:(et+e2)R 2
37. Let us place the coordinate origin in the centre of the

sphere and direct the polar axis along the straight line
that connects the charge and the centre of the sphere. The
equation
dqJ= _ _ e 6 (r-d)
eo
(d is the radius vector of the charge) has a solution that
vanishes at infinity:
00
cp(r, 9)= 4:n:eo ler - dl + ........,

)1, ....!!..!_pi (cos9)
rl+i
(1)
1=0
where P 1 (cos 9) are Legendre polynomials, and b 1 are num-

bers that can be found from the boundary conditions.
We now use the expansion
00
1 ~
r < r'
rl
I r-r' I = 4.1 (r')l+t Pz (cos 9) for (2)
1=0
where e is the angle between r and r'.

ANSWERS 177
From the boundary condition (cp = 0 when r = R) it

follows that
00
~ ( 4~eo ;~1 + R~~1 ) PL(cos 8) = 0

1=0
Since Legendre polynomials are orthogonal functions, the

last expression is valid only if the coefficients of the poly-
nomials are zero, or
e R2l+t
bt=- 4neo ~
00
!p= e _ ___:_!!__"'V(R2)lPL(cos8) ( 3)
4ne 0 I r-d I 4n:e 0d LJ d rl+1
1=0
Using the expansion (2), we can write the last member

in the right-hand side of (3) as the potential of a point
chargee' = -eR!d placed on the line connecting this charge
and the centre of the sphere, at a distance d1 = R 2 /d from
the centre. Thus
e I e' I
!p= 4n:e 0 1r-dl- 4n:eolr-dtl
which coincides with the solution of Problem 31 (found by

the method of images).
The charge density on the surface of the sphere
eR
The total charge induced on the sphere is e' = - 7 •
e VR eR ~ ( R2 )! Pz(cos 8)
38 • !p = 4ne 0 I r-d I + - r - - 4n:eod LJ d rl+t
l=O
00
V e ~ Rl-1
a(R, 8)= 4 R - 4- (2l+1)--Pz{cos8).
1t 1t dl+1
1=0
12-01496
i 78 PROBLEMS IN THEORET!CAt PHYS!Cs
· 39. The potential may be represented as follows:

00
inside the sphere cp 1 = h A 1r 1P 1 (cos e),

1=0
00
outside the sphere <p 2 = 2j B 1r-< 1+11P 1 (cos e)

1=0
. On the sphere the boundary conditions (11-31) and

(11-32) are
IPt = IJlz,
o!J!t o<p 2 a
a,:--ar-=E;
Substituting !Jlt and <p 2 , we obtain
A ao B ao Ra
1 =3e 0 ' 1 =3eo'
A 1 = B 1 = 0 for l ::F 1
Thus
rcos e,
ao R3
CJo
<p 1 = -
3- eo
(jlz = -3---2 cos
e0 r
e.
40. Knowing polarization P, we can find the densities
of the bound surface (a') and bound body (p') charges.
For a uniformly polarized ball p' = 0 and a' = P cos e.
According to the solution of Problem 39,
cp 1 =31- Prcose for r<R
eo
1 R3
3e-
1P2=-
0
Pr-
2 cos8 for r>R
E 1 = -31- P for r<R

eo
E = R3(P·r)r _ R3 _!_ for R
2 eor5 3eo r3 r> .
41. When there is no sphere, the potential of an external
field in a region without charges satisfies the Laplace equa-
tion, i.e.
00 l
IJlext = h 2,; Azmr1Pzm (cos 0) eima
1=0 m=-l
(1)
.A:NS'WEitS 119
where r, 8, a are the spherical coordinates of the observa-

tion point, and P lm (cos 8) are associated Legendre func-
tions.
For a given field the A1m' s are considered known. If we
place a conducting sphere in the field, the potential of the
sphere
1P = lpsphere + IPext (2)
where IPsphere = 2j B 1mr-HP lm (cos 8) eima is the poten-
1, m
tial of the charges induced on the sphere by the external
field. We can determine the B 1m's from the boundary condi-
tions cp = 0 when r = R:
(3)
Equations (1)-(3) fully determine the potential in the
vicinity of the sphere.
42. Inside the sphere and on its surface cp = 0. Outside
the sphere the potential can be sought in the form
cp = 2J b1r-HP 1 (cos 8)- E 0rP 1 (cos 8)
I
Imposing the boundary conditions on the potential, we get

b1 = EoR 3
b1 = 0 for l ::1= 1
Whence
EoR3
cp = - E 0 r cos 8 + - r2 - cos 8, cr = 3e0 E 0 cos 8.
43. Inside the sphere
00
IPt
-v a1r1Pz(cos 8);
=~
1=0
outside the sphere

00
<p 2 =a~ +a~rPdcos 8) + l=O

~ b1r-I-1Pz(cos 8)
The conditions of the problem imply that at large dis-

tances from the sphere
IPzlr-o-oo-+- E 0 z = - E 0 r cos 8 = - E 0 rP 1 (cos 8)
12*
Whence
Imposing the boundary conditions for the surface of the

sphere (r = R). we get
00 00
2.; a1R 1P 1 (cos 0) = L; biR-l-1P 1 (cos 0)- E 0 RP 1 (cos 0)

1=0 1=0
00 00
e 2.;
1=0
a 1lR 1- 1P 1 (cosO)= - 2.;
/·ocO
bt(l + 1) R- 1- 2Pz(cos 0)-
-E 0 P 1 (cosO)
From the two expressions it follows that
3£0 e-1 3
at=- e+2, bt= e+ 2 E 0 R
a 1 =h1 =0 for l :f= 1
Finally
<p 1 = -
3Eo
e+ 2 rcos
O
<p 2 =- E 0 r cos 0 + e-1 E R3 0

e+ 2 -----,:2 cos 0
A polarized sphere creates a potential that can be inter-
preted as the potential of a dipole (11-42) with a moment
4 e-1 3
p= nco e+2 EoR.
44. Inside the sphere
00
outside the sphere

00
<p 2 = 2.; b1r-l-1P 1 (cos 0)

1=0
We can see that <p 1 = <p 2 yields

b1 = a1R 21 +1
ANSWERS 181
From the condition that

a = - e0 ( a<p 2 - a<rt ) for r = R
ar ar
we find the relationship between the charge density a and
the unknown az's:
00
a=e0 ~ a 1 (2Z+1)R 1- 1Pr(cos8)

1=0
We multiply this equation by Pm (cos 8) sin 8 and then
integrate from 0 to rc. Since a = 0 for 0 < 8 < a,
n
a1 = cr f P 1 (cos8)sin8d8
2eoRI-t Ja
or
az= cr (2Z+1t 1 [P 1+1 (cosa)-P!-dcosa)]
2e 0 Rl-1
where PI+ 1 is the Legendre polynomials, and PH (cos a)
= -1 for l = 0.
45. This problem can be solved by separating the vari-
ables in Cartesian coordinates:
cp (r) = X (x) Y (y) Z (z) (1)
Substituting (1) in the Laplace equation. we get
X
1 d2X
dx2 = - a
2 l
_!__ azy = _ ~2 I (2)
y dy2 (
Z
1 azz
dz2 =y
2 I}
where a 2 +~ 2 = y2•
Let a, b, and c be the lengths of the parallelepiped's
edges. We assume that face z = c is the one with the non-
zero potential. Then the particular solution that satisfies
the boundary conditions (i.e. all faces except z = c have
a zero potential) takes the form
Cj) 11 m =sin a 11 X sin ~mY sinh 'VnmZ
nn nm , / n2 m2
where an=--;;:-, ~m = -b-, 'Vnm = Jl V (;Z+b2; n and m
are integers.
The general solution can be found as a linear combination

of particular solutions
00
(jJ (r) = ~ Anm sin <XnX sin ~mY sinh '\'nmZ (3)
n, m=l
The coefficients Anm can be found from the condition

that (jJ (r) = V when z = c:
a b
Jr dx Jr dysinanxsin~mY
A = 4V
nm ab sinh '?nmC
0 0
( 16V
2 • h for n and m odd
= ~ :rt mn sm '?nmC
l
for n or m even 0
Substituting Anm into (3), we can find the potential
at any point inside the parallelepiped.
00
46. (jJ (r) = ~ sin anx sin ~mY [Anm sinh '\'nmZ
n, m=l
+ Bnm cosh '\'nmZ]
where
sinh '?nmC -B nm coth ~·I nmc for n and m odd

16 V2
:rt2nm
0 for n or m even
( 16Vt f or n
Bmn= 1- -
:n;2nm
and m odd
l 0 for n or m even
:rtn :rtm -. / n2 m2
<Xn=a• ~m=-b-, '\'nm=n V a2+b2•
47. In cylindrical coordinates the solution of the La place
equation can be represented in the following form:
(jJ (p, z) = R (p) Z (z)
where R (p) and Z (z) satisfy the equations

azz
dz2
=k2Z
d2R
dp2
+..!_p dR
dp
+k2R=O
ANSWERS 183
A particular solution of these equations is

e"zzJ0 (k 1p) for z<O
O
where J 0 is the Bessel function of the first kind, and k 1 is
determined from the condition that on the surface of the
cylinder the potential is zero, or
l 0 (kzro)=0
The general solution can be found as a linear combina-
tion of particular solutions:
00
<pt = ~ Aze"zz1 0 (kzp) for z < 0

!=1
00
O

!=t
The coefficients A 1 can be found from the boundary

condition on the surface of· the disc. Using the condition
(11-32) and the property of orthogonality of Bessel func-
tions, we come to the relation
R
~ plo (kzp) dp
A,= _a_ ....:.o_....,-_ _
2Bokz ro
l pl~ (k1p) Clp
0
To find the potential of the point charge inside the cy-

linder we use the conditions for transforming to the limiting
case: R --+ 0, nR2 a = e = constant. Then
A, = ___el_,o'-'('--0)'-----
. ro
4neokz l 1: (kzp) Pdp
0
Using the fact that J 0 (0) = 1, we obtain the final ex-
pression for A 1:
48. Let us choose the system of coordinates in such a

way that the point charge is at the origin of coordinates and
the z-axis is directed along the normal to the lamina. The
potential for such a combination of charges and lamina
consists of the potentials created by the charge and lamina
separately. The potential of the polarized lamina satisfies
the Laplace equation. Using its general solution in cylindric-
al coordinates, we can write
00
'Pt = 4nEo (p2e+z2)1/Z +) Bt (k) ehzJo (kp) dk ( - oo < z <d)

0
00
'Pz = ) Az (k) e-kzJ 0 (kp) dk

0
00 (1)
+) B 2 (k) ekzJ0 (kp) dk (d < z < d+ a)
0
00
cp 3 = \ A3(k)e-kzJ 0 (kp)dk (d+a<z<oo)

'6
The first term in cp1 can be represented by an integral of
a Bessel function
00
(p2+1z2)1/2 =) e-klziJo (kp) dk (2)

0
Using the boundary conditions for the potential at z = d

and z = d +
a, we find that
B (k) _ e~ (e-2kcd+a) -e-2kd)
1 - 4nEo (1 - ~2e-2ka)
A (k)- e(1-~)
2 - 4l1Eo (1 -~2e 2ka)
(3)
B k _ e~ (1- ~) e-2kcd+a>
:d ) - 4ne 0 (1- ~2e-2ka)
A (k _ e (1- ~2)
3 )- 4ne 0 (1- ~2e 2ka)
e-1
where ~ = e+1 •
ANSWERS 185
Formulas (1) and (3) give the solution of the problem.

If a charge is on the surface of a semi-infinite crystal, we
must put d--+ 0 and a--+ oo. In this case
B1 - e~ B2 - 0 A 2 -- e
- - 4Jteo ' - ' - 2ne 0 ( e + 1)
Substituting these expressions in (1) and using the ex-
pansion (2), we obtain
(4)
where r = (p 2 +
z2 ) 112 is the distance from the charge to
the observation point. Formula (4) coincides with the respec-
tive expression in Problem 36.
49. The symmetry of the problem implies that the axes
of the ellipsoid are the principal axes of the quadrupole
moment tensor. In the set of principal axes
Changing the variables, x' = x' a, y = y 'b, z = z' c, we

reduce the integration over the volume of the ellipsoid in
Eq. (11-44) to an integration over the volume of the sphere
x'2 + y'2 + z'2 = 1.
Equation (11-44) yields
D XX =~(2a
5
2 -b 2 -c2)
e
Dyy ="5 (2b2 -a2 -c2)
Dzz = ; (2c2- a2- b2)
where e = ~n abcp is the charge of the ellipsoid.

50. Let us solve the equations
div D = e{) (r)

(1)
curl E = 0
assuming that charge e is at the origin of coordinates.
We direct the Cartesian axes along the principal axes of

the permittivity tensor. Then
aq>
Dx = BxeoEx = - BxBo ax
a<p
Dy = eye 0E y = - eye 0 8y (2)
a<p
Dz= BzBoEz= -ezeoaz
Substitute (2) into (1) and get
(3)
Changing the variables, x' = y' = Vex , Veu , z' = Vez ,

we reduce Eq. (3) to the following:
B2<p
8x'2
+ 8y'2
a2<p + B2<p _
8z'2 -
_ e
eo (Ex8y8z)1/2
f, (r') (4)
where we have used a property of <'\-function:

<'l (ax)= ! <'l (x)
Equation (4) has a solution

e 1
<p = 4ne 0 (exeyez)i/2 r'
where r' = (~+~+..::_} 112 •

Ex ey Ez
51. Direct the y-axis along the normal to the lamina.

From the boundary conditions Eu = E 2 t, Dtn = D 211 it
follows that
eoE y- EyxE: r - EyyE y- EyzE z
E= Eo+ Eyy
n
where n is the unit normal vector.
52. W=-~
4ne 0a '
53. W=~.
4neoa
ANSWERS 187
54. If we use the answer to Problem 36, we find that
e=R(1+e)y :m:~ (i nR3 ~g-Mg)

where g is the acceleration due to gravity.
55. 1= 'Pt-CJ? 2 , 1 -(.i.._.i..).
R=-
R 4:n:cr* rt r2
tan- tXt
56 . t *, wh ere• a 1' an d a 21 are the ang1es b e-
crt
an a-2 = cr 2
tween the lines of current and the normal to the boundary
in medium 1 and medium 2, respectively.
57. The potential difference between the electrodes satis-
fies the Poisson equation
~<p= _..£._ (1)
eo
where p = j/v, j is the current density, which in the sta-
tionary case does not depend on x if the x-axis is directed
along the motion of charges. Velocity v can be found from
the law of conservation of energy
mv2
- 2-+e<p=O
We can rewrite Eq. (1) in the following form:

d2q> ;v;n
dx2 = - eoV2!e!cp
Solving this equation under the conditions that <p = 0
for x = 0, <p = - V for] x = d, and ~ = 0 for x = 0 (the
last condition implies that the electric field neat the first
electrode is zero), we get
j=- :~ V2~:'1VIa'2·
58. H = ~ [j x r' for r < R
R2
H = 2r 2 [j X r] for r > R
where r is the!distance from the axis of the cylinder to
the observation point.
59. H = ; [j X a].
60. Let the z-axis be directed along the axis of the con-
ductor. The symmetry of the problem yields
Ax=Ay=O
~A1z = -floi for p <R
~A 2 z = 0 for p >R
The A z-component depends only on the distance from the
axis. In cylindrical coordinates
_!_ _i_ ( P dAtz ) = -flo .!!:..._ for P < R
p ~ ~ p
1 d ( dA2Z)
Pdp p dp =
0 f or p > R
Solving these equations, we get

AlZ = -fl 0ap +C 1 ln p +C 2 (1)
(2)
Considering that a:~z-+ oo as p-+ 0 and that in this

case the magnetic field tends to infinity, we must put cl
equal to zero. The continuity of A z and d~z implies (when
there are no surface currents) that
C3 = -fl 0 aR, C4 = -fl 0 aR (1 - ln R) +C 2 (3)
The relations (1)-(3) defme the vector potential up to
a constant term.
The magnetic field strength is defmed by the following
formulas:
1
H 1x=-asincp, H 1 y=acoscp, H 1z=0 or H 1 =-[axp)
p
where vector a is directed along the axis of the conductor

and in magnitude is equal to a. The absolute value of H 1 is
a and is the same everywhere inside the conductor.
Outside the conductor
H·u=- a: sin cp, H2 y = aR cos cp, Hu = 0 or I H2 1 =a:.
ANSWERS 189
61. Let us choose the system of coordinates so that the

z-axis is directed along the current and the x-axis along the
normal to the plane. From the boundary condition (11-26)
we then have
{
i
--2 for x<O,
Hy='
l 2 for x>O.
62. (a) H = 0 between the planes,
H = i outside the planes;
(b) H = i between the planes,
H = 0 outside the planes.
63. Let us place the origin of coordinates on the midline
of the strip, and direct the z-axis along the strip and the
x-axis normal to its surface. The magnetic fteld does not,
obviously, depend on the z-coordinate. We mentally divide
the strip into strips parallel to the z-axis. The width of
these strips must be so small that in each one the current
can be considered linear. Let the position of any such strip
be y'. and its width dy'. According to Problem 58, at a point
with coordinates x, y this strip will generate a magnetic
field of intensity dH that is determined by the following
expressions:
i (y-y') dy'
dH x = - .,........,~,......,.__'----"-~
2rt [x2+(y-y')2]
dH = ixdy'
Y 2rt [x2+(y-y')2]
dHz=O
Integrating these relations from -a/2 to a/2, we get
H = _i_ln x2+(y-a/2)2
x 4n x2 + (y + a/2)2
Hy= 2~ ( -arc tan Y --.za/ 2 +arc tan Y ~a/ 2 )

iM PROBLEMS IN TiiEoimTICAL PHYSics
As a-+ oo (the limiting case of a conducting plane),

( i
I - 2 for x<O
Hx=O, Hy=~
l 2 for x>O
which coincides with the answer to Problem 61.
64. For linear conductors of finite length l parallel to
the z-axis
I I
Az= 1-tol ( f ~- f ~)
4n J
-l
P1 J
-l
P2
.
Ax=Ay=O
where p1 = (z 2 + r:F 12and P2 = (z 2 + r:) 1' 2 ; r 1 and r 2
Fig. 56
are the distances from the observation point to the first

and second conductors, respectively.
After evaluating the integrals and passing to the limit as
l-+ oo, we get
Az= ~-tol ln....!:!..
2n r1
65. We direct the z-11xis along the normal to the plane

of the ring (Fig. 56). Since the problem is cylindrically
symmetric, we can assume that the xz-plane passes through
ANsw.ER.S HH
the observation point P. The vector potential is directed

along the y-axis, i.e. has only one nonzero component
where p is the distance from the observation point P to

the axis of the ring. If we put q/ = :rt 20, we find that +
'1'1/2
A~p = Jlo~I i (2 sin2 8-1) d8
.. J [(R+p)2+.z2-4Rpsin28]1/2
0
If we introduce the variable k2 = (R+4~;,+.z 2 , by using

straightforward calculations we come to the expression
for A~p:
Alp= ~~ ( : ) 112 [ ( 1 - !k 2 ) K (k) - E (k)]
'1'1/2
where K (k) = J (i-k2~~n2 S) 112 is an elliptic integral of the
0
'lt/2
first kind, and E(k)= J(1-k sin 8) d8anellipticinteg-
2 2 112
o
ral of the second kind. The two integrals are related in
the following manner:
dK E K dE E K
dk= k(i-k2) k' dk=k-k
Using these relationships, we get the final expression for
the magnetic field strength:
H _!__
p-
.z
2:rt p [(R+p)2+.z2]1/2
[ - K (k) + (R-p)2+.z2
R 2 +P 2 +.z 2 E (k)]
H - ...!_ 1
z - 2:rt [(R+p)2+.zzJ112
[K (k) + (R-p)2+.z2
R2-p2-.z2 E (k)]
H~p=O
On the axis of the ring p-+ 0, and so

H 0 H = R2J H 0
p-+ ' z 2(R2+z2)3/2' ,p= .
66. Outside the ball B 2 = ~-toH 2 , curl H 2 = 0, and

div H 2 = 0. This means that the magnetic f1eld strength
can be represented by the gradient of a scalar function (jlm,
which satisfies the Laplace equation ~(jlm = 0, i.e.
H2 = - '\i'(jlm (1)
The general solution for a potential that generates a zero
magnetic field at infmity is then
00
_ ~ Pz (cos8)
(jlm - LJ al rl+i (2)
l=O
Inside the ball H1 , B1 , and M are parallel. Let the z-axis
be in the direction of these vectors. Using the boundary
conditions (the continuity of Br and H 8 at r c-= H.), we
obtain
00
B 1 cos _ LJ
e-flo ~ (l+ 1) azPl(cos 8)
RI+Z
1=0
00
H . e_ '\1
tSln -..:;...)
_a_z_ dPz (cos 8)
Rl+2 d8
1=0
Whence, the nonzero coeffiCients a 1 are those with l = 1.

Substituting Bt - l\1 for H 1 , we find the equations for
!lo
a 1 and M
B 2at M-~=~
t = flo fi3 • !lo R3
Solving them we get
at =J_MR 3 , B 1 = 21-lo M (3)
3 3
From relations (1)-(3) we get the expressions for the
sought quantities outside the ball
Hz= 3
R3 [3r (M·r) --,:3M J'
r5 Bz= floHz
ANSWERS 193
Thus, outside the ball the magnetic field is the f1eld of

a dipole that has a magnetic moment
m=~R
3
3M
Inside the ball

B 1 = 2~0 M, H 1 = --} M.
67. Since no electrical currents are present, we can in-

troduce a scalar potential CJlm which satisfies the relationship
H = - Y'cpm and which can be determined from the Laplace
equation ~CJlm = 0. The general solution of this equation is
( r) = _ _!_ i div M (r') dV' _ _!_.~ M n dS' (1)
CJlm 4n J I r-r' I 4n';l' I r-r' I
where M (r) is the magnetization vector of the magnet, and S'
is any closed surface that contains the magnet.
In our case M = constant and div M = 0, and Mn is
nonzero only on the bases of the magnet. If we choose the
z-axis directed along the cylinder's axis, we get
M o f dx dy Mo f dx dy
CJlm= 4n J [(z+d/2)2+x2+y2]1/2 4n J [(z-d/2)2+x2+y2]1/2
(2)
After integrating (2). we get
CJlm= ~ 0 lV(z+d!2) 2 +R 2 -Iz+dl21

- V(z-d/2)2+R2+ 1z-d/21] (3)
Inside the magnet
ll __ 8<pm _ Mo ( z+d/2 + z-d/2 _ 2}
z- 8z - 2 V(z+d/2)2+R2 V(z-d/2)2+R2 '
eQ
68. Mx= My =0, Mz =sR 2 •
69. The rotary motion of the sphere generates a surface
current in it. In spherical coordinates with the polar axis
directed along the axis of rotation, this current is
icp = aQR sine (1)
13-01496
194 PROBLEMS IN 'rHEORETICAL PHYSICS
Since there are no currents either inside or outside the

sphere, we can introduce a magnetic potential, i.e.
- V'\jl 1 when r < R
H- {
- -· V''i' 2 when r > R
where
00
'i't = ~ A1r 1P 1 (cos 8) (2)

1=0
00
'1'2 = 2.i Blr-<1+ 1>P 1 (cos8) (3)

1=0
The boundary conditions (II-25) and (II-26) for our case

imply that, for r = R,
8\j)t 8¢2 8\jJt 8¢2 0
ar=Tr· r88- rae =~(jJ
Substituting expansions (2) and (3) and formula (1) into

the boundary conditions, and having in mind that
sinS= _d(c;;e) = _ dP 1 ~~osO)
we obtain
At=-; aQR, B1 =+ aQR 4
A1 = B 1 = 0 for l =I= 1
Thus,
2
H= 3 aQR for r< R
(i.e. inside the sphere the field is directed along the z-axis);
H= 3 (m·r)r -~ for r>R
4nr5 4nr3
4n R" a~,
m=y n
i.e. outside the sphere the magnetic field is the field of a

system with a magnetic moment m.
70. The force per unit length is
F= + !A-flo It I 2
- 2:n d
ANSWERS 195
Here the "plus" sign is used when both currents flow in the
same direction, and the "minus" when they flow in opposite
directions.
71. The symmetry of the problem implies that the force
is an attractive force. The radial component of the magnetic
field, which is constant everywhere in the contour, is the
sole contributor to this force. And so
2n
F = flfl 0/ 1Hp (Rt. R 2 , d) ' Rt de= 2nflfloR 1/ 1 Hp (R 11 R 2 , d)
b
where Hp is the radial component of the magnetic field
strength that the current in contour 2 creates in contour 1.
Using the conditions of Problem 65, we come to the final
expression for F:
F- !l!loltl2d [-K(k)+ R~+Ri+d2 E(k)J
- [(R 1 +R 2)2+d2] 1/2 _ (Rt -R2)2+d2
l k z__ 4RtR2
w lere - (Rt +R2)2+d2 .
72. The magnetic field between the coaxial conductors is
I
H = 2rrr
where I is the current in the conductor.

On the one hand the energy of the magnetic field is
W = _.!._ f BH dV = J.L!lo/ 2 ln R 2
2 J 4n Rt
On the other hand,
W =-}Lf2
Comparing the two equations, we find that

L = !l!lo 1 R2
2n n Rt •
73. L = !lt!lo+!l2!lo ln..§...

8rr 2n Rt
74 v = - !loe2a2 N.
' "' 2m ·
13*
76. E = aJ 0 (kr) e-iwt and H = -ai va*i J 1 (kr) e-i<ot,

W/lo
where k= i+i, <T=
a
(+-)
wo flo
112 ,
and J 0 and J 1 are Bessel
functions.
77. The time dependence of the magnetization vector is
given by the following equation:
dM
Tt = f.tog [M X H] (1)
where g is the gyromagnetic ratio.

The magnetic field strength for the given system can be
written in the form
H = Ho + he-iwt
Let us look for the solution of (1) in the form
M = Mo + me-iwt
where m is the additional magnetization induced by the
variable field.
Since h ~ H 0 , we conclude that m ~ M 0 • If we ignore
infinitesimals of order hm, Eq. (1) yields
-iromx = -f.t 0gM 0h 11 + f.t gm
0 11 H 0
-irom 11 = f.t 0 gM 0hx - f.t 0 gmxHo

-iromz = 0
After solving these equations we get
mx = xhx - ivhll
+
m 11 = ivhx Xh 11 (2)
Wo (J)
where X=gM0 f.t 0 W2 -Wo2 , v = gM 0 ~-t 0 W2 -Wo2 , ro0 =-f.t0gH0 •
When ro:::::: ro 0 , the effect of magnetic resonance is observed.
78. If the wavelength is considerably greater than the
size of the sample, we can use the magnetostatic approxima-
tion to describe the oscillations of the magnetic moment.
In this approximation the varying field h and the magnetic
moment m that are generated by the oscillations_satisfy
ANSWERS 197
the following equations:

curl h = 0
div (h + m) = 0
For a sample in the form of a ball these equations are
solved in Problem 66. The solution implies that inside the
bnll
1
h=--gm
On the other hand, h and m are also linked by the relation-

ship (2) of Problem 77. Because of this
hx=- ~ (Xhx-ivh 11 )
h11 = - ! (xh + ivhx)
11
The system of equations has a solution if its determinant

is zero, which places a restriction on the value of the fre-
quency of oscillation w. Non-trivial solutions exist if
W=+g(Ho+!Mo)
where M0 is static magnetization created in the ball by the
field H 0 •
79. In":thej case of a nonconducting! medium and for a
slowly varying magnetic field h the system of Eqs. (II-58)
is of the form
curl h = 0 (1)
div (h + m) = 0 (2)
where m is the varying magnetic moment.
Equation (1) implies that a magnetic potential IJlm can be
introduced, so that h = - 'VCI>m· Substituting the last
equality into (2) and keeping in mind that, according to
Problem 77, m is a function of h, we come to the following
equation for IJlm:
(3)
For an infinite medium we seek the solution of {3) in

the form IJlm = q> 0eikr. We then have f.t (ki + Jc;) + k: = 0
1.98 PROBLEMS IN THEORETICAL PHYSICS
or tan 2sk = fl, where ek is the angle between k and the

z-axis, i.e. between k and the direction of the constant
magnetic field II 0 • Using the value of X from Problem 77,
we get the frequency of natural oscillations as a function of
the direction of wave vecto(k:
w2 = w0 (w 0 + WM cos 2 8ji), where WM = -gflo~t
Thus the spectrum of possible frequencies is
w0 <w<V w0 (w 0 +wM)
We must note that tho obtaine(result is only valid in tho
domain k ~ ~ , i.e. when it is possible to use the magneto-
static approximation and when in Maxwell's equations tho
aE
term at can be neglected.
80. The magnetic potential Qlm for the given system
satisfies Eq. (3) of Problem 79. For an ideally conducting
covering, the boundary condition is of the form Bn = 0, or
az =0 for z=O and z=d

Dcpm (1)
with the z-axis directed along the normal to the surface of

the covering.
Let us examine the waves that propagate along they-axis.
The solution of Eq. (3) of Problem 79 can be written in the
form
where
(2)
From the boundary condition (1) it follows that A = 0

and kz = nnld, where n is an integer. By substituting these
results into (2) and 11sing the rlepenclencc of ~t on w, we get
the dispersion equation for magnet.ostatic waves in the plate
w2 =wo(wo+ 1 +t-6-r )·
81. By analogy with Problem 80 we must also examine the
solution outside the plate. At the boundary this solution
ANSWERS 199
must be fitted to the solution inside the plate:

q>inside
m
= q>outside
m ,
8<J>m
az
I
inside
= 8<1'm
az
I
outside
for z = 0 and z = d.
For the wave travelling along the y-axis the frequency
spectrum can be determined from the conditions that
{1)
The system of equations (1) gives the dependence of ffi

on ky.
82. ffi2 = LtCt + L2C2 =t= [(LtCt-- L2C2J 2 +4CtC2 L~ 2 ] 1 1 2
t2 2CtC 2 (LtLz-Lf2)
83. The head of the resultant electric-field vector describes

an ellipse with the semiaxes
a= V E~1 cos 2 a+ E~ 2 cos2 (a- q>)
b = V E~1 sin 2 a+ E~ 2 sin 21{a- q>)
E~ 2 sin 2<p
w h ere tan 2a= E 2 +E 2 2 . The principal axes of the
01 02 COS !p
ellipse are rotated through an angle a in relation to vec-
tors E 1 and E 2 •
84. The relationship between the angle of incidence, 8~>
and the angle of refraction, 8 2 , is
. 8 sin St
Sill 2=-- {1)
nt2
where n12 = V e 21e1 is the index of refraction of the second
medium relative to the first. If e 2 is less than e1 , relation (1)
holds true if we choose 8 2 complex. The electromagnetic
wave in the medium from which total internal reflection
occurs (when e 2 < e1 ) is given by the formula
Ez = Eozeikr- irot
where k = Ve 2 /c, and ffi is the frequency of the wave.
If we choose the coordinate system in such a way that the
interface of the two media lies in the xy-plane and the wave
vector k in the xz-plane, then

-iw(t x sin El2+z cos 62 )
E2 = Eo2e "2 (2)
where v2 = c!V e 2 is the phase velocity of the wave. Let
us express angle 8 2 in terms of angle 81 using the relation-
ship (1):
cos 82 = + v1- sin 2 82 = + i-. /' sinZ 81 -1
V n12
(3)
where we must choose the "minus" sign for the field to be

finite as z--+ oo. Substituting (3) into (2), we get
sin
. ( t -x -
-l(!) -6t
- ) -KZ
Ez= Eo2e vt
where 'X=~
v2
( sib.: el
n12
-1) 112.
Thus, the wave in the medium from which total internal
reflection occurs is a wave that travels along the interface
and that is also damped in the reflecting media with a loga-
rithmic decrement x. The..depth of§penetration is
x
d = ..!_ = ~ ( sin~ el
w n12
-1) -1/2.
(1-n)2+x2
85. R = (1+n)2+x2 . Here
n = v; rv1- T +----:-(-~=:0--:)"""'2+ 1 /2
and 'X= v; [v + c:::or T'2 1 -1
In the case of an ideal conductor ( e 1;: a* I w),

R = 1-2 ( 2;~w r/2.
86. The electric field can be found froru the equation
rfl.E
dz2
+k E = 2 0 (1)
wh ere k = - c
v-e .
1ns1'd e the plate and k = k 0 = ~
c
outside
the plate. The z-axis is directed along the normal to the
surface of the plate.
ANSWERS 201
The solution of Eq. (1) in the region with the incident and
reflected waves has the form
Et = Eoeikoz + Ae-ikoz
where E 0 is the amplitude of the incident wave. and A that
of the reflected wave.
Inside the plate the solution is
Ez = E+eikz + E_e-ikz
In the region with only the transmitted wave
E3 = Deikoz
where D is the amplitude of the transmitted wave.
The boundary conditions, i.e. the continuity of the tan-
gential component of the electric field vector, bring usl to
the following equations for the electric fields in all three
regions:
E 0 +A= E++E-
E+eikd + E_e-ikd =
Deikod
E 0 -A=nE+-nE_
nE+eik.d- nE_e-ikd = Deikod
where n = Vs is the plate's index of refraction.
Solving this system of equations, we get
A= lfPo(1-e2ikd) Eo
1-poe2ikd
2
E+= . E0
(1+n) (t-p0e2tkd)
2Po
E_=- Eo
(1- n) (e-2ikd- Po)
D= 15oeikd E
. 0
1-poe2tkd
1-n)2 ~ 4n
where Po = ( i+n and u0 = (i+n)Z are, respectively, the
reflectance and transmittance for a semi-infinite medium.
Using the obtained relationships, we find the reflectance
of electromagnetic waves from a plnne-parallel plate:
I A 12 4p 0 sin2 kd
P = I Eo 12 = 153 +4p0 sin2 kd
There is no reflection if
d=~= '
m1c
k 2
where m is an integer, and 'A the wavelength inside the
plate.
- 87. Choose the x-axis directed along the line of propaga-
tion of the wave, the y-axis along the wave's magnetic
field strength, and the xy-plane in the interface of the two
dielectrics. Suppose the medium with permittivity e1 lies
in the region z > 0, and the one with permittivity - le 2 1
in the region z < 0. The solution of the equation
w2
11H y + C"2 eH y = 0
which describes a wave that travels along the interface and
also dissipates far away from the interface has the following
form:
H 1y=H01 eikx-x 1z, where x 1 =Vk2 -ro2 e1/c 2 for z>O
(1)
fl 2 y = II 02 e1kx+x2z, where x 2 = V +
k2 ro 2 je2 l/ c2 for z < 0
The boundary conditions H 1 y = H 2 u, E 1 , = E 2 , (for z = 0)
yield
Hot = H0 ~, X 1fe 1 = X 2/Je 2 l (2)
The last expression holds only if e1 < le 2 1. Eliminating x 1
and x, with the help of formula (1), we obtain
k2 -= w2et I Ez I .
c2 (Jt:ti-Bt)
88. The equality of tbe projections of the wave vectors

on the interface yields (i!) for the ordinary ray
s~n e; = (e )1 /2
sm St 1.
and (b) for the extraordinary ray

B B ) 1/2
kt sin 81 = k2 sine;= kosin e; ( . 2 !r..l+ll
E .l SID 2
2 e:;
Bll COS 2
Here k 0 = role and e; is the angle between the wave vector
in the crystal, k 2 , and the optic axis. Whence, for the ordi-
ANSWERS 203
nary ray
, V Bj"1sin 81
tan e = :-7'=======:=;:::::::;:=-
2 V e J. (e 11 -sin2 8t)
The direction of propagation of the extraordinary ray e;

is connected with H; by the relationship
e
tan e; = _J:._
ell
tan e;
and so
tan e" __ V e J. sin 81
2 V e11 (e 11 - sin2 8t)
89. An electron in a variable electric fipld E 0 e-iwt moves
according to the following equations:
..
mx = - kxx + eE xe-iwt
0
my= - kyy + eEoye-iwt

mz = - kzz + eEoze-iwt
We seek the solution in the form
X=Xoe-iwt, Y=Yoe-iwt, Z=Zoe-iwt
where x 0 , y0 , and z0 can be found from the equations

eEox eF:oy eEoz
Xo= k X -mw2 ' Yo= k y -mw2 ' Zo = -;-k-z_---.::.m=-wn-2
The projections of the polarization vector induced by

the electric field are
Ne2 Ne 2 E
Px = kx-mw2 Ex, Py= ky-mw2 Y•
The components of the permittivity tensor are

e = 1 + -.,.,-N_e_2--;;::- e - 1 + -=-N_e_2_~
xx eo (kx- mw2) ' YY- eo (ky-mw2)
e - 1 + --,-,-N_e_z--::::-
zz- eo (kz- mw 2) '
92. Choose the z-axis in the direction of the magnetk

field. An electron in a variable electric field and a constant
magnetic field moves according to the following equations:

•• e e •
x+ ro:x=-
m Ex+ -yEo
m
y + ffi02 y =m
•• e e •
- E y - - xB0
m
where ro 0 is the natural frequency of oscillations of the
electron. Introducing new variables
r+ = x +
iy, r_ = x - iy,
E+=Ex+iEy. E_=Ex-iEy
and considering that the light wave is monochromatic
(E± ex: ei 001 ), we get
r_=me E_
ffi2-(J)2+-ffiBo
e
o m
Maxwell's equations give the solution for E+ and E_ in

the form of plane waves:
E± = Eo±e-irot+ik±z
where
(J) Ne2 1
k±=-n±,
c
n±=1 + - - - - - -e- - -
e m 0 2 2
ffio -ffi +---:: ffiBo
mJ
Assume that when z = 0, the electric field vector in the
medium is directed along the y-axis and in magnitude is
equal to E 0 • Then
. i k++k- k _-k+
E x= E oe -trot+ - 2- z cos 2 z
. t +. tk++k- k __ k +
E y = E 0e -too -2- z .
SID Z
2
From this solution we see that vector E rotates. Along

a path of length l the vector rotates through an angle IJl =
= k_-k+ l
2 .
93. The electric field vector in a monochromatic wave is
(1)
ANSWERS 205
The law of motion of an anharmonic oscillator subjected

to an electric field is
~ P. E
d2x;
m
2
dt2 + mffio X;+ LJ Piilxixl = e
(2) ;
i. l
Assuming ~iii to be small, we expand x 1 in a series

x 1 = x~01 (t) + x~u (t) + ...
where x -
eE i
-,--,--=---...,.-
t - m (ro3 -ro2)
In the next approximation
x\l> _ _ _!_ ~ e2 ~wEojEoz [ _1_+ cos 2rot J
' - 2 ~ m3 (ro3 -ro2)2 <1)8 ro5 -4ro2
1, l
The polarization vector is

p = p<O> + pw
where
pUl = p(l) (0) + pU> (2ffi}

P (ll (O) __ Ne3 ~ ~iil EoJEol ._
t - 2 LJ m3 ( ro 2
0 -ro2)2 ro02 ' ~- x, y' z
j, l
pll> 2ffi = _ Ne3 ~ ~til E01E01 cos 2rot

1 ( ) 2 LJ m3 (ro 2 -4m2) (ro 2 -ro2)2 •
j, l 0 0
9 4. pu> = P m (ffit - ffit) + pm (2ffit) + pm (ffiz - ffi2} +

+ pu> (2ffi 2} + 2P< > (ffit + ffi + 2p (0) and P< 1 > (2ffi) of
the previous problem, and the projections of P< 1> (ffi 1 ± ffi 2}
206 PlWBLEMS IN TllEORETICAL PllYSICS
are
pill ( -1- ) _ _ Ne3 ""' ~ EoiEoz cos {Wt ± w2) t
• Wt -- (1)2 - 2 L' rn3 w3- (wt ± w2)2
j, l
1
X (w~ -wr) (w3 -w~) •
95. Let the z-axis be directed along the length of the

chains in the direction of propagation of the wave. The nth
oscillator in a chain has the following law of motion
m~: + mw~rn- q [(rn+t -rn) + (rn-1- rn)l = eEn
where q is the elastic constant.
Multiply this equation by e/v0 , where v 0 is the volume
per oscillator. The quantity ernlv 0 = Pn is the polarization
vector. The law of motion then takes the form
•• e2E
mPn + mw~ Pn- q [(Pn+t- Pn) + (Pn-t- Pn)] = _ _
n
vo
If the wavelength is considerably greater than the chain's
period, the polarization vector is a slowly varying function.
Then the equation can be rewritten in the form
•• d2P (z) e2
mP (z) + mw~P (z)-qa 2 -d- 2-=-E
z vo
(z) (1)
where a is the period.
This equation must be solved together with the equation
for the electric field
iJz2 - C2 "ifi2 = eoc2 iJt2

(2)
We seek the solution of (1) and (2) in the form

p = p 0eikz- ioot, E = E 0eikz- ioot (3)
Substituting (3) into (1) and (2), we get
Whence
k= ( q#m-w3 +w2 ± V (q#m-w~ +ol2 )

2
+4a~ ) 112
± 2qa2frn
ANSWERS 207
Thus for a given frequency there are four wave numbers

(and, consequently, four indexes of refraction, which are
connected with each k by a relationship n = : k). Two
waves move in one direction; two others, with the same
indexes of refraction as the f1rst two, move in the opposite
direction. In an isotropic medium it is possible for waves
with different indexes of refraction and the same frequency
to move in one direction. This property is inherent in media
with so-called spatial dispersion, i.e. when permittivity
depends notl only on the frequency of the wave but on the
wave vector as well.
96. Direct the x-axis perpendicular to the plates, and
let the planes where the plates lie be x = 0 and x = d.
The z-axis is then directed along the line of propagation of
the wave. We seek the solution of Maxwell's equations in
the form
E (r, t) = E (x) eikz-iwt }
(1)
H (r, t) = H (x) eikz-iwt
Substitution in Maxwell's equations yields the following

system of equations:
kH x -
kH u = wee 0 E x•
ax-= -zwee
iJHz .
0
E
y,
- kEy = W/loH x
----ax= lW/lo
l"kE x - iJEz . H y
I
)
iJHy • iJEy . (2)

----;;;- = -twee0 E z, ax= ZWfloHz J
iJHx
iJx
+ l"kH z =0 ' aEx
ax
-t- ikE z = 0
The boundary conditions for our case of ideally conducting

plates are
Ey = Ez = H x = 0 when x = 0 and x = dj
The system of equations (2) separates into two independent
systems; one for H x• Hz and E u• and the other for E x• E z
and H u· The first describes the propagation of TE waves,
and the second of TM waves.
For TE waves (Ez = 0),

a2Hz+
ax2 X 2H-0z- (3)
E = _ i roJ.to aH z H = .!:!!._ aHz (4)

y x2 ax 7 X x2 ax
where
x2= :~e -k2 (5)
The boundary conditions are: a:xz = 0 for x = 0 and

x = d. The general solution of (3) is of the form
Hz =Hot sin xx + H 02 cos xx
where the boundary conditions impose the restriction that
H 02 = 0; x = rr.lld, l = 1, 2, ...
Thus for TE waves,

Hz (x) =Hot sin :z x (6)
For each value of l there is a definite type of wave that

can propagate in the system. The magnitudes of the fields
for each type can be found from Eqs. (1), (4) and (6), and
the dispersion equation is simply formula (5).
For TM waves (Hz= 0),
(7)
The boundary conditions are: E z = 0 for x = 0 and x = d.

The solution of Eq. (7) together with the boundary condi-
tions is of the form
:nl
Ez(X)=EoCOSTX, l= 0 , 1, 2, ....
97. Direct the z-axis along the wave guide. For a travel-
ling wave we then have Eo:: eikz, H o:: eikz. The magnetic
field component along the line of propagation of the TE wave
ANSWERS 209
is determined from the equation
azn z
iJxZ
+ azH + "rzn = 0
iJy2
z
z
where "r2 = ~: -k2 • The boundary conditions are: iJ:/nz =0

when x=O, x=a, y=O, and y=b.
The solution of the equation is of the form
H zmn (x, y) = H o COS -nmx

a- COS
nny
-b-
where m and n are integers;

2 2 2 ( m2 n2 }
"r -+"tmn=1t (.i2+b2
For a > b the frequency limit is ffio1 = nc!a, for a < b
it is ffio 1 = nclb.
For TM waves
. nmx • nny
E zmn (x, Y) = E oSln-a-Sln -b-.
98. The electromagnetic field in a conductor decreases

sharply at depths below the surface considerably less than
the wavelength in a vacuum. This is the reason why the wave
vector is directed perpendicular to the surface. The electric
and magnetic fields in the conductor are related to each
other by the equality
V 88o + a;i Et = V f.tf.to [Ht X n]

where E 1 and H1 are the tangential components of the fields,
and n is the normal vector. Since Et and Ht are continuous,
we have the same relation outside the conductor. For an
ideal conductor,
Et = ~ [Ht X n], where ~ = V f.tffif.to/(a*i).
99. Use the conditions of Problem 98. The electric and
magnetic fields decrease like e-az, where ex= 2e ;~ X 0( r/Z
X n 2 ( b :: +a ~: )[ "r~nn ( ~: - "r~nn r/ r2 1
•
14-01496
100. For TM waves,

sin mp
Ezn = EoJ n (yr) {
cosncp
where J n is the nth Bessel function (n an integer), and
y is determined from the condition that Jn (yR) = 0.
For TE waves,
sin ncp
Hzn = HoJ n (yr) {
cosncp
where y is determined from the condition J~ (yR) = 0.
The dispersion equation
2 for TE and TM waves is
( w2 2 }1/2
1 _
tC- -cz-Y .
101. Solving Maxwell's
equations inside and out-
side the guide and fitting
both solutions on the boun-
dary, we come to the con-
ditions
Jt(yR) ...L Ki(BR) _ O
yJo(yR) 1 ~Ko(BR)-
(1)
w2
y2+~2=(e-1)C2 (2)
ew2 w2
where -cz-k2 =y2 , C2-
--k2= -~2, and K 0 and
Fig. 57 K 1 [are modified Hankel
functions.
The graphical solution of Eqs. (1) and (2) is given in
Fig. 57. For frequencies less than a definite value, called
the frequency limit, curves of (1) and (2) do not intersect
and ~ is imaginary. In this case the waves do not propagate
along the guide but are radiated in the surrounding medium.
102. Direct the z-axis along the common axis of the cy-
linders. For a wave propagating along this axis (E, H ex:
ex: eikz-iwt) Maxwell's equations in cylindrical coordinates
ANSWERS 211
are
au z
-r1 ---- "kll"· =
~
. 0E •
-lffiee )
o!jl .. •
l"k/1 -- iJllz =
----;JT" - l·ffiee 0 E 'I'
1 a (. ) 1 aH,
- - 1 11 'l- - - - - = -~ffiee0 'z
. E
r or r alP
(1)
aEz
-1 -a--~"kE~~~ = lffi!!o
. Hz
r !p
·kE ,
~
i1Ez = IC•l;.to
- ---a,:- . H 'I'
a ( E ) aEr . H
r1 Tr r <j, - 71 ---acp = ~ffi!!o z
Also E'l' =Hz= 0 for r = R 1 or r = R 2 • Let us show

that for a wave guide with a multiply connected cross sec-
tion (which is what our system is) there is a possibility of
c Ve. For this
transverse waves with E z = Hz = 0, k = ~
case it follows from the system of equations (1) that
VeEr=Hrp V !!o/f.o )
VeE.,= Hr V !!oleo
_!_ (rE<P)- 8Er = 0
I} (2)
ar alP
_!_(rHrp)- aHr =0
or iJ!p J
Using these equations, we can show that rEr, rEr,;, rHr,
and rHq> must satisfy the two-dimensional Laplace equation
_!_ ( r au )
f}r or
+.!...r a2u
iJ!p2
_ 0
-
(3)
Since rErp and rHr are zero on the boundary of the region
and are harmonic functions inside, this implies that Erp =
= Hr = 0 everywhere. Then (2) yields
8Er =0 iJ(rHrp) =0
iJ!p ' or
14*
Whence, having in mind the dependence on z and t, we get
H = ~ eikz-i(jjt 1 E 7 =-. / ~ eikz-irot

q> r v !-to
BoE 7
For the TM waves we must solve an equation that follows

from Eq. (1):
_!_~(r oEz
r or or
)+_!_
rZ
oZEz +xZE =0
o<pZ Z
(4)
The solution is of the form

sin nq:>
Enz = [AnJ n (xr) +BnY n (xr)] { (5)
cos nq:>
where J n (xr) is the nth Bessel function of the first kind,
and Yn (xr) is the nth Weber's Bessel function of the second
kind.
As distinct from wave guides of a circular cross section,
we have no need to discard solutions that at r = 0 become
infinite, since point r = 0 does not belong to the system.
The boundary conditions yield
Anln (xRl) + BnYn (xRl) = 0} (6)
Anln (xR2) +BnYn (xR2) = 0
This system of equations has nontrivial solutions if the
system determinant is zero, i.e.
Jn (xR 1) Yn (xR 2) - Jn (xR 2) Yn (XR 1 ) = 0 (7)
The last equation defines the values of x for a given n:
Xn 1 , xn2, . . . . For the TE waves Eq. (7) contains J~ and Y~
instead of J n and y n·
103. Ex= Exo cos ax sin ~y sin yze-irot
E 11 = E 110 sin ax cos ~y sin yze- i(jjt
Ez = Ezo sin ax sin ~y cos yze- i(J)t
where a= nm/a, ~ = nnjb, y = nl/c, and w2 = c2n 2 ( ; 2
2
+
+fj2+-;z
n2 l2 ) .
[m, n, l mtegers).
ANSWERS 213
The lowest frequency is

n
(J)mln = C £.
!
where Li is the longer edge of the parallelepiped.

105. From the condition that E 1 = 0 at the ends of the
resonator we come to a solution in the form
Ez = A/ (p, <p) cos kz e-irot
where f (p, <p) satisfies the equation
f a (P Bp
pOp at ) + ""i)2f arp2
8 2/ +~•I2 /-
- 0 t
~·I2 = coc22 _ k2 (1)
and k = nnld (n is an integer). Separation of variables

brings us to the Bessel equation for the radial component.
Thus
Ez =Eo!m'('VtP) eimq> cos kz e-irot
where J m is the mth Bessel function (m is an integer), and
'V 1 is determined from the condition that J m ( 'V 1R) = 0;
l numbers the different roots of a Bessel function.
The natural frequencies of the TM waves can be found
from
co2 a+ n2n2
C2 = 'Vl ---ar
For TE waves the natural frequencies are determined from
the condition that J;,. (-r 1R) = 0.
106. Let the liquid move along the x-axis. The velocity
at any point in the liquid depends only on the position of
that point between the plates (on coordinate z). So does the
magnetic field.
We must now solve the system of equations (II-74) and
(II-75) together'"'with.,.the boundary conditions: v = 0 and
H x = 0 when z = ± d/2. Equation div H = 0 yields
Hz =constant = H 0
For the z-component of Eq. (II-75) we have
floH2
P+-T'=P(x)
dp
an d dx= dx =constant because of statiOnary
dP . fl ow. TlJen
Eqs. (II-75)-(II-77) yield
H
0
.:!!!___+ c2eo rJJ.H x =O
dz a* dx2
rJJ.v
'I'] dz2
+ !toH o ---;JZ"
dHx
=
dp
dX
Solutions of these equations that satisfy the boundary
conditions have the form
d z
coRh -,--cosh--
'2d0 d\)
v = Vo -----'-d-,------'--
co<;h 2do -1
_ 2z.hd .tz
sm --u---sm 1 d
7
H X-- - v04ne0 lfa*n'I °d o
cosh---1
2d0
. 1 r-TJ-
where d 0 =11-
o
V -.-and v0 is the velocity of the liquid
a llo
at z= 0.
For weak fields (d <t d 0), v = v 0 ( 1 - ~22 ) • This is
simply the result of ordinary hydrodynamics.
For strong fields (d ':» d 0 ), v = v 0 ( 1 - e- ( ~ -lz} :o).
107. Equation (11-75) for our case is
flp rJJ.v
-8 =
X
!-1-oOHo (Eo- !-1-oHov) + '11 'd
.. Z
2 (1)
!L=o
By (2)
(3)
Since there is no pressure gradient along the x-axis, we

come to the equation for the distribution of velocities:
rJJ.v 1 1 ·Eo
---V=---
dz2 JloHo
d~ d~
Solving this equation and using the boundary conditions

(v = 0 when z = 0, and v = v 0 when z = d), we come to
ANSWERS 215
the sought distribution

E
. h d-z
Sin -d-
+ SID. h -;r
z -
v (z) = ----'v0"-:-- sinh _z_ + - -
0- 1 - -----'0:.....__--,-_ __,"'0'--
. h- d
sm do floHo . h -d
sm
do do
From Eq. (3) we obtain
cosh _d__ cosh ( _d___ z_)
H x (z) = H 0v0 add0 _ _2d---"-
0 ---"'-.2d_,o:.....___d....::o:..._.:_
2sinh 2: 0
For weak fields (d <S: d 0),
VoZ
V=-
d
For strong fields (d ?;> d0 ),
V=~ (1-e-zldo).
~-toHo
108. We write the continuity equation in terms of the

concentration of the electrons:
!!!:...+div
at nv = 0 (1)
The law of motion of an electron in a plasma with no

pressure gradient is
dv iJv e
at=m+(v·V)v= -;n-E (2)
Maxwell's equations yield

divE=-e (n-no) (3)
eo
where n0 is the ion concentration and also the mean electron
concentration. ·
Let us assume~the motion of charges to be so slow that
the variation of charge density is considerably less than its
mean value. In such an approximation Eqs. (1)-(3) trans-
form into the following:

8/ln +n 0 d"IV v =
---at 0
divE=~ !!..n
eo
~=~E
iJt m
We then find the divergence of both sides of the last equa-

tion. Excluding div v and div E with the help of the first
two equations, we find
iJ2tJ.n
- 8t2 +-- e2
me n0!!..n
0
= 0
!!..n = !!..noe-iropt
where wp = V~
- -
e0 m
0 is the plasma frequency.
109. !!..qJ = - ..£.._ and !!..A--\- a:t~ = - f.loh, where h

e0 c u
is the transverse component of the current density vector:
h• = 41t cur I curI Ji I rj _(r')r' I dV' .

1
1tO. The vector potential is equal to

llo i j (r') eiklr-r'l ,
A(r)=T:t J lr-r' I dV
When the distances from the antenna are great, the vector
potential is
d/2
A (r) = e 3/ e:~~o J sin ( ": - k 1z' 1) e-ikz' cos e dz'
-d/2
where e is the angle between e3 and the direction of propaga-

tion of waves. Evaluating the integral, we get
kd ) kd
~-tofeillr cos ( 2 cos e -cos T
A (r) =e3 ZnrJ sin2 e
B=~A3 sin6
CJlo
ANSWERS 217
The time average radiation intensity per unit of solid

angle is
_ ___£___ I cos ( ..!!t:- cos e) _cos ..!!£-- 2

1
- 8n2ceo sin e
When the antenna length is several half-lengths,
~ ~V~: :;,
-
l
2kl
I cos2
4 cos" (
(-i cos e)
sin2 e
i e)
sin2
cos
e
(kd = n)
(kd = 2n).
111. J = .!1..-. / J:Q_ r r 1- cos u du

4n V e L J 0 u
0
2kl
- 2~l ) (1- cos u) du J.

0
113. The magnetic fi.eld of the incident wave in cylindri-
cal coordinates can be written as follows:
Hz = H ozei(kp cos cp- rot)
The plane wave can be expanded into a Bessel series:
00
m=-oo
The symmetry of the problem implies that the magnetic
fi.eld of the scattered wave will also be directed along the
cylinder's axis.· Let us look for this fi.eld in the form of
a series of cylindrical functions whose asymptotic behaviour
for large p is the same as for the fi.eld
H ,__ p-1/2eikp
These are the Hankel functions Hm (kp). For the scattered
wa v~ we then have
00
Hz= Hoz 2J am.Hw (kp) ei(mcp-mt)

m--oo
The boundary condition is: Hz = 0 on the surface of the

cylinder. For our problem this is equivalent to E'f = 0.
aE
The equation curl H = e 0 7it leads us to the expression
E _ __i_ 8Hz
<~>- eow ap
The boundary conditions yield
• m J'-r,. (kR)
am=(-£) H:n (kR)
Fot· large p,
H (k ) ,..._-. / 2 eikp-(2m+1) ~ i
m p V nkp
Then we fmd the magnetic field for large p:
__ oo ·( 2m+1 )
t mcp---n
H = Ho -. /_2_ ei(kp-rot) ,, a e lo
z zV 1t~ ~ m
m=-XJ
E,~ =, / !-lo H.
V eo
The time average of the energy flux density in the scat-
tered wave is
!S!={Re(E,Jli)= ~ V~:·IH:I 2
The energy flux through a unit lengtn of the cylinder is
2n
J=) !S!pdqJ=2 V , /-
~: IHozl 2 X! ~
"" II' (kR)
H:(kR)
12
•
0 m=-oo
114. The dipole moment of a system of two particles in

their centre-of-mass system is
where m1 and m 2 are the masses of the particles, 1..1. is their

reduced mass, and r is their separation.
For identical particles, e1 = e2 and m1 = m 2 • And so
p = 0, which means that there is no dipole radiation,
i.e. radiation proportional to I pi'·
ANSWERS 219
115. The intensity of dipole radiation can be calculated

using formula (II-73), where p = er. With the help of the
law of motion
e2r
mr=---- 4:ne0r3
we can exclude r. Whence,
J- e6 128:ne0 1E !4
- 96:n~egcsa4m2 3m2c3e2
where E is the particle's energy.

116. If in the course of one period of revolution the par-
ticle's energy changes only slightly, we can use the result
of Problem 115 and find that
dE 128:neo I E 14
dt= 3m2c3e2
whence
3m2c3e2
't - -;--~:;-;--;:;"7.
- 4:ne 0 X 241 E 13
117. The time dependence of the electromagnetic field
can be presented in the form
0 for t <0
E (t) = { t . 1
----troo
E 0 e -r for t> 0
Let us write E (t) in the form of a Fourier integral
co
E (t) = J E (w) e-irot dw

-co
where the Fourier transform

co
J E (t)
E (w) = _f_
2:n -co
e+irot dt =
2:n[ +-i Eo
(roo-ro) J
The radiation intensity is proportional to the squared
modulus of the Fourier transform:
J (w) = 1 Jo
:n-r [ -:t2+ (ro0-ro)2 J
where
foo
J0 = J J (ro) dro
-oo
The half-width of the radiated line is determined from the

relationship
whence
2
Aro=-.
T
118. We can represent an elliptically polarized wave in

the form
E = A cos rot + B sin rot
where A j_ B. Then the differential cross section is
where n is a unit vector in the direction of scatter.

119. Let us seek the electric and magnetic fields rin the
form .
E = Eoteiqxx+iqyy-fkzz-foot }
for z<O
H _ H fqxx+fqyy-ikzz-foot
- ote --
and
· E = E 02eiqxx+fqyll+fkzz-ioot }
for z> ()
H = H o2eiqx x+fqyy+fkzz-ioot
where kz =
-./w 2 2
V C2-- qx- 2
qy ·
Choose the y-axis directed along i. Using the boundary
conditions (II-25) and (11-26) and also the relation between E
ANSWERS 221
and H in a plane wave, we obtain

H Otx = - 2i • H Oty = 0 • H 01z = :--- q:x:
2kz •
k~+q~ . qy .
E 01z= - - 2
Eo1y=- 2 k £, · e £,
ow
eow z
H 02x = 2i • H 02Y = 0• H
02z = -
qx ·
2kz £ •
qxqy . k~+q~ . E qy .
Eozx = 2eoWz
k z, Eozy=- 2eoWz
k z, ozz=-2e
oW
£
Electromagnetic waves are radiated in the ambient space

if ffi 2 /c 2 > q!, +
q~. When this condition is not fulfilled,
an electromagnetic field exists only near the plane together
with the current.
120. The electric field ot the vibrating dipole induces
a dipole moment in the particle. The particle then radiates
electromagnetic waves. Provided d ~ 'A, the electric field
vector at a point with a radius vector d is
E (d)- 3(p·d)d _P_
1 - 4ne0d5 4neod3
When p j_ d,
E1 (d)= - 4~oda
This field induces in the particle a moment
PI = 8o~EI
The net f1eld is created by both dipole moments, p and PI-
At great distances from the system (r ~ 'A),
E = E1 +E 2, H = HI +H 2
where
E w2 ([p X n] X n] HI w2
1 = - 4ne 0c2r ' 1·=- 4ncr [p X n]
w21}
Ez = (4n)2 c2eord3 [[p X n] X n], H 2 -- (4n)2
w21} [
crd3 P X n
]
where n = r/r.
The radiation intensity of the electromagnetic waves is
J= p2w4 (t--1}-)2
6neoea 4nd3 •
121. One of the ways in which the damping manifests

itself is the following. As a consequence of the finiteness
of the electric conductivity, therL exists a flow of electro-
magnetic energy through the plane, which leads to a loss in
the energy of vibration. This energy flux is maximal in the
vicinity of the dipole. Since d <t 'A, we can use the quasi-
stationary approximation to find the electromagnetic f1eld
I
/
/ ...... --- '
I
I
I
\
' ' ___ ...... /
Fig. 58
in tho vicinity of the dipole. In this approximation we can

show that on the plane the dipole induces electric charges
whose field is the field of a dipole that is a mirror image
of the original dipole (Fig. 58). And so in the vicinity of
the dipole (r <t i..),
E=, 3(p·r)r p 3(pt·rtlrt Pt
4:n:e 0 r5 4:n:e0 r3 + 4:n:e rt - 0 4:n:eorr
H = !:!E_ [r X PI + ~ [rt X ptl
4:n: r3 4:n: r?
We can find the density of the energy flux by using the
solution of Problem 98:
S = _f_ , /
2 V !lo
e0
I H t j2 n ( 1)
ANSWERS 223
It is seen from Fig. 58 that

(p·d)d t
r 1 = 2d+ r, Pt = dZ - Pt
II 1 is the projection of the magnetic field on the conducting
plane. That is why on the surface of the plane
H 2iw pcosa
t=J.loc r3
(2)
Substituting formula (2) into (1) and integrating (1) over

the surface of the conducting plane, we get the energy ab-
sorbed by the plane in one second:
.!!. = - r s ds = w2p2 cos2 a
dt J 16ne0cd2
The energy loss of a dipole in a vacuum is less by a factor
of about ('A/d) 2 than that of the dipole on a conducting
plane.
122. Let the dipole be in the origin of coordinates. We
seek the solution in the following form:
in medium 1
H = (pxn) + IPt x nt)
1 4ncr 4ncr 1
Et = [[p X n) X n) + UPt X nt] X n 1]

4nc 0 V E1 c2 r 4neo V f 1 c2r 1
in medium 2
H _ IP2Xn]
2 - 4ncr
Ez=([p2 Xn)Xn]
4ne 0 V c 2 c2r
where r is the distance from the dipole to the observation
point, n a unit vector directed towards the point of obser-
vation, r 1 = r - 2d, n 1 = r1 /r1 , and p1 and p 2 can be found
from the boundary conditions
[p X n] t + [p 1 X nd t = [p 2 X n] t
((p X n] X n] t + [(p1 X Dt] X Dt] t = ([p 2 X n] X n] t
VEt Vet Ve2
where t is a unit vector in the plane of the interface. Let the

latter be the xy-plane. Solving these equations for p1 and p 2 ,
we get
Vet+ e2
V P· P2=
123. In determining the long-wave part of the radiation

spectrum we can assume the charged particle to bounce off
the plane instantly, and so we can use formulas (II-72) and
(II -73).
The time dependence of the particle's velocity is given by
the following relationships:
V = Vxi vyj +
Vzk for t < 0 +
v = Vxi +
Vyj - Vzk for t > 0
Whence
..
p = ev = 2evzll (t) k
.
where k is the unit vector normal to the plane. According to
formula (II-72), we find that for great distances from the
charge
ef> (t)
E (r, t) = 2moe2r Vz [[k X n] X n]
Now, since
00
6 (t) = 2~ J e-iwt dffi,

-oo
we find that
00
E (r, ffi) = 2:
•• Jf
-00
E (r, t) eiwt dt = 2 evz 2 [[k X n] X n]
ne0c r
The energy radiated per unit of solid angle per unit of

frequency interval is
dJ(w) e2v2 • 20
-~'- = - - - Sln u
di."J 4n3c3e 0
where e is the angle between the direction of propagation of

waves and k.
ANSW!l:RS 225
The radiation maximum lies in the plane of reflection.

There is no radiation in the plane normal to the plane of
reflection.
The energy radiated in all directions per unit of frequency
interval is
J (w) = . 2e2v2.
3n2c3e 0
which shows that at low frequencies (long waves) the radiated
energy does not depend on frequency.
125. According to the law of addition of velocities in
special rela ti vi t y,
v+.....:_
n
U=---
1+~
en
when v 1:: c,
u=.....:..+v(t--
n
1 ).
n2
2 p2 1 p<l
126. E = me + 2m -8 m3c2 •
127. Assume that the electric f1eld is directed along the
x-axis, and choose the y-axis in such a way that the particle
moves in the xy-plane when z = 0. The law of motion is
then written in the form
dpx
---cit = e
I E I' dpy
dt
= 0
After integrating we get

Px = e I E I t, P y = Po
The condition Vx = Pxc 2 /E, where E is the energy of
the particle, yields
dx e I E 1 tc2 e I E 1 tc2
dt = [E8 + +
c p8 (ce IE I t) 2] 112 =
2 [E31 + (ce IE I t)2] 112
dy Poc 2
dt = [E81 +(ce IE I t) 2 ) 112
where E 0 = mc E 01 = c (m 2c2 + p 0 2 )11 2 • This gives us the
2,
path of the particle
1 ~2
[E 01 + (ce IE 1t) 2 ]
2
X= efET +x0
15-01496
Y= /1°~1 ln [t+ yt + Ce~~l2 rJ +Yo
226 PROBLEMS IN THEORETICAL PHYSiCS
When E 0 ~ce1Ejt,
x=~t
2m
2 +x 0
'
129. The law of motion of the particle in cylindrical

coordinates is
! (V mr
1-v2jc2 )= mnP2
V 1-v2jc2 (1)
~ (V mr2<p
1-v2jc2
)=0 (2)
! (V mz
1-v2jc2
) e1 ezz
= - 4ne 0r3 (3)
.
! (V mc2
1-•·~;c2
} e1e2r
= - 4ne 0r2 (4)
We can satisfy Eq. (3) if we r:ut z = 0, which corresponds
to the motion of the particle in plane z = 0. The remaining
equations are
d ( mr
. )
.
mr<p2
dt v 1-v2jc2 = v-1-v2jc2
d ( mr2<p ) _ O
dt V 1-v2jc2 -
Equation (2) gives one integral of motion-the angular

momentum-and, hence, the law of conservation of angular
momentum:
mr2<p = L =constant
V 1-v2jc2
Equation (4) gives the law of conservation of energy:
~r=m=c=2 :;;=:;-- e2
1
4eneor = E =constant
V 1- v2jc2
ANSWERS 227
Eliminating t in the first equation, we get

_£___
d<p2
(_.!.}
r
+ (1 - P )_.!.-
2 e1e2E
r - 4na 0£2c2
where e1e2
P = 4naoLc ·
T= p
1 +a cos <V 1-p2 (<p-<p0 )}
£2c2
ere~
16n2a02
where p = E
4 :rteoete2 ' , and e and cp 0 are constants of
integration. We can put cp 0 = 0 by choosing an appropriate

reference point.
The path of the moving particle is not closed. It can
be obtained by slowly rotating an ellipse in its plane.
130 • E t -_ c2 M 2 +mr-mi2M '
E _ 2 M 2 -mi+m~
2- c 2M •
131. TtJ. = 4 MeV, Tv= 29.8 MeV.

132. Let the second frame of reference move along the
x-axis of the first. Since Px• p Y' p z, i ~ are components of
a 4-vector, we can use the law of transformation of 4-vectors
(II-79) to find that
p-...!!_E
x c2
P~ = • P~ = Pu• p; = Pz
V 1-v2jc"-
E' = E-VPx
V 1-v2jc2
133. Let the decaying particle move along the x-axis, and
the particle with the energy E 1 move at an angle 01 to the
same axis. From the solution to the previous problem it
follows that
cos et -
_ -Ptx- -
_ Et-Eot (1-v2jc2)
E2 112
1f2
Pt v ( ct--mfc2}
where Eot is the energy of particle 1 in the laboratory frame

of reference, and m1 is the mass of the particle. A similar
expression can be found for angle 0 2 •
15*
228 PiWBLEMS IN THEORETICAL PIJ:YSlCS
134. tan8'=u'(f-v 2 /c 2 /
v+u' cos8
128 inS
'
where 81 and e are
angles between the directions of the velocities in the two
reference frames and the x- and x' -axes, respectively.
136. w = li wo , where w0 is the frequency

1 + meuJ°2 (1 - cos 8)
of the photon before impact, and n is the Planck constant

h divided by 2n.
137. (J) = ~: ( 1 - :~2 ).
139. Consider a reference frame that moves together with
the mirror. Assume that the x-axis of that system is directed
along the velocity of the mirror (opposite the normal). For
such a system the laws of reflection in a stationary medium
are valid, i.e. the frequency will not change in the course
of reflection. The laws of reflection are:
k~ = - k~x. k~ = k;y, k~ = k~z• w' = w;
where k' and w' are the wave vector and frequency of the
incident wave, and k; and w;
are the same quantities for the
reflected wave. Since the four numbers kx, ky, kz, i ~ are
components of a 4-vector, we can use the law of transforma-
tion of 4-vectors to obtain the relationships valid for a re-
ference frame in which the mirror moves:
kx + ktx = !_
c ( + Ult)
(J)
ANSWERS 22'J
Solving these equations, we find that

1 + ~2- 2~ cos 8
ffit = (!) 1- ~2
cos
e1 = 2~-(1+~2) cosO
1 + ~2- 2~ cos 8
where e and 01 are the angles between the x-axis and the
lines of propagation of the incident and reflected waves,
respectively.
141. If we fllld at least one such frame of reference, any
other frame that moves parallel to E and H will have the
same property. For this reason we must find one frame of
reference that moves with a velocity which is normal to E
or Hand directed, say, along the x-axis. Since in such a frame
the f1elds are still parallel, we have H~ = E~ = 0 and
E~H~ - E~H~ = 0. The law of transformation of fwlds
brings us to the fmal result:
v/c _ [EX H]
1+v2jc2 - eoE2+~toH2 .
143. T
xx = 1-v2jc2
1 ( T' 2iv
xx- -c- 4x- cz
T' v2 T'44 )
T~y=Tyy, Tyz=T~z
T :xy =
(1- v2jc2)
1 1/
2
( T~y-...!:..!!...
c
Tl,.y)
1'4."< = i
c (1 - v21c2)
112 [- icT~x ( 1 + v: )- vT~, + vT:x _.J
c
T•Y = (1- v~fc2)1J2 ( T4y + i: T;y)

1 ( "
1- v2jc2 1 44 + 2 -c-
iv ,
T h - cz T xx
v2 , )
.
E .l -1 [v X ll]
144. El'I=EII• H 1'1=H11• E'-1 = 12
(1-v2jc2)I.
H' _ H.l-e 0 [vXE]
.l - ( 1- v2jc2) IJ2
Here the symbols JJ and .l stand for 'parallel' and 'normal'

to velocity v.
SECTION III
1. As the particle moves, its momentum p remains con-
stant and is equal to -p after the particle is reflected from
the wall. Thus
a
ppdx= 2p) dx= nh
0
whence
nh
Pn=2a and
2. Using the fact that the energy of the oscillator is an

integral of motion
p2 mw2x2 rnw2A2
E= 2m + - 2 - = 2
where w = (klm) 112 is the frequency of oscillations and

A is the amplitude of oscillations, we get
p = V m2w2 (A2- x2)
Substituting x =A cos cp, we find that
~ p dx = mA 2w:n: = nh
Thus the allowed amplitudes and energies are
1 rnA 2 w2
A~, = __.!!:_!:__ and En= ; =nnw.
m:rtw
5. The mo Lion of the electron can be described by two

. .
generalized coordinates, r and <p, and two generalized mo-
menta, Pr = mr and P'l! = mr2 cp. In a central field there are
two integrals of motion: the component of angular momen-
tum along the axis of rotation
P'l! = mr2 cp = L
and the total energy
E = rn;2
2
+· mr2~2
2
- ZeZ
r
ANSWERS 231
The last relationship yields

Pr = V2_m_[-=-_E-+--;;Zr--;;e2:--_-2_,~-::-:-2-]
Since r = 0 at the extremal values of r, it follows that rmax
and rmin are the roots of the radicand. The Bohr-Sommerfeld
quantization rule gives
rmax
~ P<P df[! = L X 2n = n'fh and 2 ) Pr dr = nrh
rmln
The second integral is of the form I= JY

Ti
A+ 2B jr + Cjr 2 dr,
where YA = +i V-2mE (E <0) and yc = -iL =

= -in,.Ji. We can calculate it in the following way. As we
have already mentioned, if for r we substitute r1 or r 2 , the
radicand turns zero; the radicand is positive if r1 < r < r 2 •
Consequently, A < 0 and C < 0, since
A+ 2B +..£.._= A(r-rt)(r-r 2)
r r2 r2
=C(~--1)
r rt
(.i.r __
1 )>0
r 2
Denote by f (z) the function A + 2: + ywhere z &,

is a complex variable. Obviously, z = r1 and z = r 2 are
the branch points of this function.
To calculate I let us cut the complex plane along the
segment r1 - r 2 • We choose f < 0 just above the cut and
f > 0 just below. We then evaluate the integral around
a contour consisting of the upper and lower sides of the cut
(from r1 to r 2 along the lower side and from r 2 to r1 along
the upper). We denote the contour by l and have
r2 rt
) f (z) dz = ) If I dr + j [-I /11 dr = 21

l rt r2
Point z = 0 is the pole of f (z). Hence, l can be distorted

away from the singularities, so tha.t points z = r1 and
z = r2 ar.e .inside .the contour a.n.d z .-::-- ,0 is .outside. If Wtl
consider a contour L in the neighbourhood of z = oo and

a contour 'A around z = 0, according to Cauchy's theorem,
) f (z) dz = ) I (z) dz + ) f (z) dz

L l 1-.
with all the contours by-passed in the positive direction
(counterclockwise).
To calculate th~ integrals along L and 'A, let us expand
f (z) near z = oo and z = 0.
For z = oo
When we pass from points z <r on the lower side of the

cut (where f = +
VA (z - r 1 ) (z - r 2) > and arg I =
to points z > r 2 , the argument of z - r 2 increases by n.
2
0 0)
Consequently, arg f changes from 0 to n/2. Thus in the
above expression we must take = A j, since VA +i IV
arg ( 1 + ~z + . . . ) = 0.
For z = 0
~c (1+ 2~z+At(2 ~~c (1+ z:; + ... )

f(z)=
Now when we pass from points z > r 1 on the lower side of

the cut to points z < r 1 , the argument of z- r 1 decreases
by n, since we pass in the negative direction (clockwise).
Hence, Vc = - i I Vc j.
Using the theory of residues, we have
~ f (z) dz ~ VA ~ ( 1 + ,Z + ... } dz = 2ni ;A

L L
Jf (z) dz ::::_ Vc .\
), ),
~z ( 1 + ~·z + ... } = 2ni
'
JIG
.and finally
l=+rJt.(z)dz- j l(z)dz]~ni (Jl-Vc)

L -?.
ANSWERS 233
If we introduce the quantity n = nq; +nr and substitute the

values of VA, B, and we get Vc,
2:rri (
i
V meZ
2mJEI
+ in'l'1i} = n,h
and
me4
En=- 2h2n2.
6. Let us denote the position and the mass of the nucleus
by r 1 and m1 , and those of the electron by r 2 and m 2 • We
can then separate the variables by introducing r = r 1 - r 2
m 1r1 + m 2r 2
an d R = .
m1+mz
The centre of mass (with a radius vector R) moves like
a free particle of mass M = m1 + m 2 in the given limits,
i.e. 0 < x <a, 0 < y < b, 0 < z <c. As for the relative
motion of the electron, this reduces.to the motion of a parti-
cle with mass fl (the reduced mass of the system fl = m~z, )
mi mz
about a fixed centre (see Problem 5). The centre-of-mass
momenta, P x• P Y• P z• are quantized in the same manner
as in Problem 1. Thus
h2 ( n~ n~ n~
p2
Entn2nsn =2M+ Ere!= 2M --;;2+/)2+ -;;2
} f.!e"
- 2h2n2 ·
7. The particle's position z and its momentum p enter

the total energy, which is an integral of motion, in the
following manner:
E = ;n: + mgz -= mg H
which yields
p= + V2m (E- mgz)
Using the quantization rule
Ii
~pdz=2 ~'V2m(E-mgz)dz=nh
0
.enables us to find I! n and En:

H ·-· ( 3nh ) 2J3 E =(avn;gnh)2J 3
. n- 4mV2g ' n 4V2 •
8. (a) Applying the operator twice to an arbitrary func-

tion 'IJ, we get ( ~ +x ) 2 'IJ= ( ~ +x) ( !!' +x'IJ) =
d2'1J d\jl ( d )2
= dxZ + 2x dx + x2'1J + '\jl and, consequently, dx + x =
d2 d
= dxz+2xdx+xz+ 1;
d
(b) ( CIX+x-
1)3 = dx3+x
d3 3d2
dxz;
{ d )2 d2 d
(c) \X dx =XzdxZ+x d.c ;
d )2- 2 d2
(d) ( dX x - x dxZ + 3x dxd + 1.,
(e) It is obvious that [inV+AJ 2 '1J=(iliV+A)x
x (in grad 'iJ + A'IJ) = - n 2.11jl +in [(V ·A)+ (A· V)] 1jl + A 21jl.
Taking into consideration that (V ·A) 'IJ =(A· grad 'I')+
+ 1jl divA, we come to the final result:
(inV +A)2= - nf1 + 2in (A· V) + in_div A+A2 ;
(f) (1-M) <L+M)=L2 -M 2 + (lM -ML).

9. (a) Applying the operator :x x-x ~ to an arbitrary
function, we get
d
( - x - x d- )
dx dx 'I'=-d d\jl
d.x (x'\jl)-x-
dx ='~
or·
d d
-x-x-=1·
dx dx '
(b) in [(V ·A)- (A· V)) =in divA;

a a at
(c) a'P f- fa'P = a'P ·
10. {a) Let .us defme the sought operator by the .equality
Ta'i' (x) = 1p (x+~)
ANSWERS 235
We then express 'lj) (x +a) as a power series in a:
Noting that ~ -
00
xn =ex, we get Ta =e
~ a-
d
dx;
n.1
n=O
(b) We define Ta as Ta'IJ'(r+a) and get Ta=e(a·V).
~
T a= ~
00
(
a
d
drp
)n • nT1 = ea-d drp.
n=O
11. (a) According to the definition of a hermitian con-

jugate operator,
00
J 'IJ'! d! dx = J
-00
2
-00
00
'\j)z [ ( :X rr 'IJ't dx
00 00
provided that J I 'lj) 1 12 dx and J 'IJ' I 2 12 dx exist, which im-

-00 -00
plies that 'lj) 1 and 'lj) 2 are zero as x-+ + oo. If in the right-
hand side of the equality we integrate by parts:
-00 -00 -oo
we lind that ( __:!:_ ) + - - __:!:_

dx- dx'
(b) In a similar manner we get
(:;nr =(-t)n :xnn·

13. The operator i'~ is defined by the relation
j 'IJ'i (r) Ta"i' 2 (r) d-r= J 'lj) 2 (r) [T;'Ij) 1 (r)]* d-r
236 PROBLEMS IN THEORETICAL PHYSIC.'!
In order to find this operator, we introduce a change in

variables, r +a = r', in the expression defming Ta:
I= ) ljlf{r)Ta ljl 2 ( r) d-r = ) '!'! (r) lJldr +a) d-r
= ) 'l'i (r'- a) 1~ 2 (r') d't'
The last equality stems from the fact that the integration
is carried over the entire space and thus the value of the
integral remains the same.
But since ¢ 1 (r- a) = T-a¢ 1 (r), we can write
I=) ¢ 2 (r) [T -alf1 (r)J* d-r
.I.e. r~+
a=
T~
-a·
14. By definition,
ia
e
~
aq;
"'
=~
( ia f}«:p • According to
n.1
n=-0
Problem 11(b), the hermitian conjugate to is

an
(-1t a~:pn . Consequently,
[ (i :cp rr = (- w(- a~ r = ( a~ r i
is a hermitian operator. Thus the hermiticity of the original
operator follows:
( eia~} + ia!!....
acp =e acp (if a=a*).
15. By _defmition,
) 1ll!AB1Il2 d-r = ~ 'tllz l (ABr111d * d-r
Let us consider the ftrst intrgral. We denole B¢ 2 -"-" ~: 3
and introduce .4+. We can then write
~ ¢!A1113 a.-= ) 'l'a (k1p 1)* dr

Introduce a new function ¢ 4 such that ¢ 4 = A.+¢ 1 and then
return to \jl 2 • We can rewrite the last -equality )n a iorm
ANSWERS 237
convenient for a transformation via fJ+:

~ 'IJla (A +'IJlt) * d'"C = ~ 'IJltB'IJlz d'"C = ~ '1Jl2 (B+'IJl4) * d'"C
=~ '1Jl2 (B+ A+'IJlt)* d-r

Thus
(ABt=B+Ji+.
16. Hint. Use the results obtained in Problem 15.
17. Hint. Use definitions (111-14) and (111-16).
18. Adding + MLM to LiVP - if2L and taking the
common factor i1 out of the differences leftwards and right-
wards, we obtain
LM 2 -M2 L+MLM-MLM= (LM--ML) M
+M (iM-ML)=2M
since LM- ML = 1.
19. Using the results obtained in Problem 18, we prove
that being valid for n = 2 the relationship is valid for
n + 1 as well. Suppose
M£n_ £nif = -n£n-t
Now we form the expression
M£n+t_£n+tj(f =M £n+t_£ (M £n+n£n-t)
=(Mi-LNI)£n_n£n= -(n+1)£n
where we have used the previous·equation. Thus the relation-
ship is proved for any n. By defmition,
00
j (£) = ~ f''~~O) £n
n=O
Consequently,
00
t(L)M-Mt(L)=~ 1 cn~ 1( 0 )(LnJf;r_1Vr£n)

n=O
00
fCn) (0) £n-t

=~ (n-1)1
n=1
238 PROBLEMs IN THEORETiCAL PHYSic§
But if we put n-1 = n 1 , we get
~ J<nt+1) (0) £nt = ~ [/' (O)](nt) £nt = j' (£)

.:.:..J n1! Ll n 11
nt=O nt=O
and in this way
20. In order to prove assertions (a) and (b), we must

present B2 in the form BAA -IiJ. As for (c), let us express
1 (B) as a power series in 13, 1 (B) = ~ :, r> (O) ir, and
n
then use (b).
21. First expand r~A in a series and then, in a way
similar to that of Problem 19, find the commutator
00
n=O
And so
e~A.jje-~A. = e~A. (e-~A: jj + C~e-bA") =B +C~.
22. If we introduce the operators A = eibPI1i and A-1 =
= e-itv;r. and then use the results obtained in Problem
20(c), we can write
ei6pf1iF (q) e-i(p;r. = F (ei6v/1i qe-i~v!ll) (1)
But sinr.e pq-c/p= -in, we have
qe-i6v!n = e-i~p/h (q + ~) (2)
(by analogy with Problem 19).
Substituting (2) into (1), we get the required relation-
ship.
23. We construct the eigenvalue equations for our opera-
tors. For _!:_
dx
~ = A.¢ or ¢ = ef..x
dx
The finiteness of ljl (x) as x--+ ±oo implies that f.. = i~,
where ~ is a real number. In a similar manner, for i :x ,
'i' = e-i;l.x
where f.. is a real number (a continuous spectrum).
24. Separating variables in the standard equation
{x + :x) ljl = f..ljl, we arrive at the equation ~ =('A-x) dx.
After solving this equation, we get
'ljJ = ce;l.x-x2f2
Such a solution satisfies the requirements of being finite,

continuous, and single-valued for any 'A, both real and
complex (a continuous spectrum).
25. We seek the solution of :~ = f..ljl in the form
ljl = ce'Mp
Because our eigenfunction must be single-valued, the solu-

tion must satisfy the condition
'i' (cp) = 'ljJ (cp + 2:rt)
Substituting the original solution, we find f.. from the
conditioneAZn = 1. Thus/..= im, wherem =0, +1, +2, ....
26. In order to find the solution of
sind~ 'ljl=f..'ljl (1)
we express the operator sin d~ in a power series:

1
.
sm-a
d
<p =-a
d
<p --3, -d
.fp 3
d3
+-s,1.fp
d5
-a 5- · · ·
00
( -1)k d2k+1
=~ (2k + 1)! dqJ2k+1
(2)
k~=O
We seek the solution of (1), having in mind (2), in the form

ljJ =earp, where a can be found from the single-valuedness
of the function, a = im (m = 0, +1, +2, ... ; see Prob-
lem 2.')). After substituting this solution into Eq. (1), we
240 PROBLEMS IN THEORETICAL PHYSiCS
get the eigenvalues
A= ~ ( -i)k (im)k =sin (im).

Ll (2k+ 1)!
k=O
27. Similarly to Problem 2G, l(l = eimrp, A = cos m.
28. Similarly to Problem 2G, l(l = eim<P, A = a-"m.
29. I ntrod uci ng a new function U = xl(J, we get
azu =AU
dx2
whose solution is
U I = C'le VIx an d U2 = Cze - Vi:x
For A = -~ 2 < 0 both functions will be fmite as x-+- + oo.
But in order to be fmite at x = 0, l(l (x) must be a linear
combination of U1 and U 2 divided by x,
C eif3x+ C e-if3x
l(J(x)= 1 2
X
and the numerator must turn zero together with the deno-
minator (at x = 0). Thus
sin Bx
C1 +C2 =0 and l(l (x) =C -x-
where ~ is a real number.
30. We consider the left member in the Poisson bracket
as being a product of two operators. Then we do the same
with the right member. In each case we use the relation
given in the problem and get
{ft fz, g} = {it, g} fz + ft {lz, g} (1)
{1, gtgz} = {1, it} iz +gt {/, iz} (2)
Now in equations (1) and (2) we set J = /2! 2 and g = g1g2 •
Using the same relation and keeping to a definite order in
the multiplication of operators, we get the same left-hand
sides, and the right-hand sides can be expressed in two
forms: 1
from Eq. (1),
{ldz. gtgz} = gdft. gz} lz +at. it} izlz + ft idlz. iz}
+ ft {iz, it} gz;
ANSWERS 241
from Eq. (2)

{!I 12, g!g2} = f1 {12, g!} g2 + {h' gI} f2 g2 + g! ~~ {[2, g2}
+ idtt. g2}/2
Subtracting one expression from the other, we get a rela-
~ion~hi p that holds true for any set of four operators [ 1 , f 2 ,
The arbitrariness of the operators implies that for any two

operators 1 and g,
{l' g} = c {/ g- gt}
If we require that {1, gV = {f, g} provided j+ = f and
g+ = g (the hermiticity of the quantum Poisson bracket),
we find that C* = -C, or C = i/li. The dimension of C
must be the one of the reciprocal value of action by analogy
with the dimension of the classical Poisson bracket {f, g} =
n
= ~
/, ( -0of ,.......--
ag Of ag)
0-a- . Tlms, li Is
· an ar b"t
I rary quan t"t
1 y
......., Piuqi qi Pi
i=l
having the dimension of action, and the quantum Poisson
bracket
A - i AA AA
{/, g} =h [f g-g fl.

3t. aa*-a*d=t.
~ fiw A A A A
32. H= 2 (a*a+aa*).
33. We take advantage of the fact that
~if= if~. 'Px'Pu=PuPx, iixil-iiPx= -iMxu

and prove the sought commutation relation for the x-com-
ponent, i.e.
LuLz-LzLy=(zpx-XPz) (xpu-YPx)-(~Pu-1/Px) (~Px- XPz)

=(z Pu- YPz) (PxX-~ Px) =iii (YPz- zPu) = inLx
16-0H96
A cyclic permutation of the indices x, y, z enables us to

obtain the other two components, and in vector form
[LX L] = itzf..
34. To fi.nd the commutation relation for L2 and Lx we
write
J)£x- LxL2 = L~Lx- LxL~ L;Lx- LxL; +
If we add and subtract LyLxf.Jy and LzLxf..Jz, factor out
common multipliers leftwards and rightwards, and use the
results obtained in Problem 33, we get the needed result.
37. We seek the solution of the equation iii iJ"IJ.'~;· t) =
= - ~: ~:~ in the form 1f (x, t) = U (x) cp (t) (this is
a particular solution since the variables may be separated).
Then
in dcp fi2 azu
- -=---
 0. .
Denotmg kZ- 2ma
-~ , we get
-~t
cp (t) = e 2m , U (x) = eihx
where k is a real number. The general solution is

00
1f (x, t) = ~· C (k) eihx-ihh't!2m dk.

-00
38. The normalization constant A is determined from the

00
condition that J
-00
I "'I' (x, 0) 12 dx = 1. Substituting the
given wave function, we arrive at an expression for A:

00
I A lz J
-00
e-x2Ja2 dx =I A j2a Vn =1
whence
ANSWERS 243
In order to determine the region where the particle is

localized we must fmd the probability density p:
p =I 'I' (x, 0) 12 =I A 12e-x2Ja2
The peak of this function lies at x = 0 and the function
rapidly decreases for I x I >a. The width of the wave
packet represented by this function is of the order of a.
The probability current density is
x2
. = ..!:!!._ ( qr a'!'* _'I'* a'¥ ) = I A 12 nko e-ar= liko
j X 2m ax ax m m p
The final expression for j x coincides with the classical for-
mula. The factor p is determined only by the real part of
the exponent in the wave function, and the quantity nlco
m
(the analog of velocity in classical physics) by its imaginary
part.
39. Let us express 'I' (x, 0) in the form of a wave packet
co
(see the solution to Problem 37): 'I' (x, 0) = ) C (k) eikxdk.

-00
Whence
oo oo x2
C (k) = 1
2n Jl ur (
r x,
0) -ikx d
e X=
A
2n Jl ei (ko-k)x-2"ii2 dx
-00 -oo
We then complete squares in the exponent under the integral

sign and get
a2 (ko-k)2 1 .
r
00
C(k)=~e- 2 e-2a2 [x-rcko-kJa2J2 dx

2n J
-oo
a2 (ho-11)2
Aa 2 ·-·
=--e
V2n
The function C (k) is nonzero near k = k 0 , which implies
that
A2a2
IC (k) 12 dk = ~ e-a2 (ko-h)2 dk
is proportional to the probability of fmding the particle's

momentum in the interval (k, k dk). The width of the +
1
wave packet in k-space is !1k ~a
16*
40. If we take the wave function from Problem 37,

00
'I' (x, t) = JC (k) eikx-itzll'tl2m dk and substitute C (k) from

-oo
Problem 39 into it, we obtain
00
Aa f (. . fi.k2t a2k2
'I' (x, t) = V 2n J exp lkx-z 2in--
2-
-oo
By completing squares in the exponent and making use of

~ r-
~ e-ak• dk= y ;
-oo
we obtain
ix)2 a2
[ ( ko+ -;i2 ]
'I'(x, t)= v- exp
Aa
2n t+-'-
'ht 2 - 2 ko
a2 2
ma2
2
00
2 · ix
ko+-2 ) -
~· exp .11+~-
X
-00
_!:._ ( k
2 V ma2 V1+- a
ili.t
ma2 _
dk
A [ x2- 2ia2xk0 + ili.~k3 t]

--r===~exp - .
-. I
V
1 ili.t
+ ma2 2 (1 +....0!_)
ma2
a2
The probability density is, by definition, equal to
p =I 'I' Iz = I A 12
exp
.. I nh2
V 1 + m2a4
ANSWERS 245
which implies that we are dealing with a wave packet

whose peak is traveling with velocity liko and whose width
m
increases in time as
'v
a =a fi2t'2
1+--
m2a4
The probability current density can be represented in

the form
. = .!:.!.!:._ ( 'I' a'l'* __ 'I'* aw )

lx 2rn ax ox
-ik0 -..!... iko--=-)
= .!:.!::._ ( 'I'*ll' a2 -'I'* 'I' a2
2m 1 _-...!:.!!!.._ 1 + ilit
ma2 ma2
htx
ltk0 1 + -m-a-4k;;
= P ---;n ---=-rz..,..zt""z::....
1 +-"-l2-a4-
00 00
41. (x) = ) x / 'iJ /2 dx= 1A j2 ) xe-x 2 fa 2 dx= 0

-oo -oo
00 00
 = I 'il* ( -in ~~ ) ax= )

-00 -00
'il* ( nko + i:: )lJl ax= 1iko.
42. Since (x) = 0 (see Problem 41),
11x = x ·- (x) = x
whence
J x21'1llz f
00 00
(11x2) = dx = IA lz :J.·ze-x2fa2 dx = I~ 12 Vn a3
-00 -00
It follows from the normalization condition that I A 12 =

= (na) -t/ 2 and
a2
(11x2) = 2
To determine (11p 2 ) = n2 <(k - k0 ) 2 ) we can use I C (k) l2dk

from the answer to Problem 39, first normalizing it, i.e. we
proceed as follows:
dW (k) = Be-a 2(k-ho) 2 dk
1= r
-00
Be-a2<k-ko)2 dk= B J!;- a~
Thus
00
(f1p2)=Bfi2) (k-ko)2e-a2(h-ho)2dk
-00
00
= Bfi2 r·. __ a_
a (a2) J
I e-a2<~<-llo)2 dk]
-00
and we immediately arrive at the uncertainty relation

a2 !z2 ft2
(/1.r 2) (/1p 2) = 2 2a2 =4 ·
43. Let us consider the fmt (x < 0) and
third (x > a)
regions, i.e. the regions where V = oo. Taking the Schrodin-
ger equation in these regions in the form
_1_ az'ljlr = ~ (V- E)
'ljlr dx2 lz2
we see that the function \jl 1 must turn zero when V = oo

(this can be proved more rigorously if we use the solution
to Problem 45).
In the second region (0::::;; x ::::;; a) the Schrodinger equa-
tion comes down to
d21jln
dx2
+ ~'lln=
2mE ,1 0
If we denote k 2 = 2mE/fi 2 , we can write its solution as

\jln = A sin (kx +
a). Since tho wave f11nction must be
continuous at all points in space, for one, in the transition
from the first region to the second, x = 0, and from the
ANSWERS 247
second to the third, x = a, we must assume that '¢I (0)

= '¢n (0) and ¢n (a) = '¢m (a), i.e.
A sin a = 0 and A sin (ka a) = 0 +
Hence, a is zero and k can take on none but discrete integ-
nn
ral values: kn = a, where n = 1, 2, . . . . The energy
levels are
E - n2n2f72
n- 2m
The normalization condition
r 00 a a
. ~ I'¢ 12 dx = ~ I '¢n j2 dx =I A 12 ~ sin2 n:.x dx = 1
-00 0 0
yields A= V2;a. The wave function is specified at any

point in space:
'IJli = 0, '¢n = V 2. n:rrx

-Sill - - ,
a a
'¢m = 0.
44. In the regions where V = oo the wave function '¢
is zero (see the solution to Problem 43). But in the regions
where V is zero the wave function satisfies the equation
~'¢ + 2~E 'lJl = O
in which the variables can be separated: '¢ (x, y, z) =
= ¢ 1 (.1:) ¢ 2 (y) ~'a (z). For all three functions we obtain
similar equations as in Problem 43 with analogous boundary
conditions:
'¢1 (0) = '¢1 (a) = 0
¢2 (0) = "'2 (b) = 0
'¢3 (0) = '¢ 3 (c) = 0
When we satisfy these conditions and also the normaliza-
tion condition in the regions where V is zero, the wave
( -v
function takes the form
'1Jln 1n2na X, y, z) =
-8- . n1nx • n2ny • nanz •
-b- S i l l - - Sill -b- S i l l - - ,
a c a c
in the rest of the space '¢ = 0.
The energy corresponding to the wave function is
45. Let us restrict ourselves to the case when E < 0.

In the f1rst and third regions, after we denote x 2 =
= -2mE/Ii2 > 0, the Schrodinger equation takes the form
d2¢I - X2'"I = 0
dx2 'I'
Consequently,
'i'r = Aexx + Be-xx' 'i'ni = Ce-.,.x + Dexx
From the requirement that ¢I and ¢m be finite as x ~ +oo
it follows that D = B = 0. Then
'i'I = Aexx, ¢III= Ce-xx
In the second region we have the equation
dZ'!Jn _I kz,., = 0
d.r2 1 'I'll
where !(!. = 2m (V 0 + E)/li 2• Then

¢n ~co a sin kx + ~ cos kx
At points x = -a and x = a we require that the func-
tion and its derivative be continuous. (This is possible
since the discontinuity of the potential energy at these
points is finite.) Thus,
'1'1 (-a)= 'i'n (-a), 'i'n (a)= 'I'm (a)
~~I lx=-a = d!~I lx=-a' d~:I lx=a = d~~II lx=a

For coefficients A, C, a, ~ we get four homogeneous
linear equations:
Ae-xa+asin ka-~coska=O}
Axe-xa -akcos ka- ~k sin ka= 0
Ce-xa_a sin ka-- ~cos ka = 0 (!)
Cxe-"a + ak cos ka -~k sin ka = 0
ANSWERS 249
This system has a solution if its determinant is zero,

that is, if
x 2 - k2 2kx cot 2ka = 0 + (2)
Solving equation (2) for x, we get two roots:
(a) x = k tan ka. Substituting this into (1), we get
a= 0 , C = A ' A- A
p---
e-xa
cos ka
and
which implies that¢ is an even function: ¢ (x) = ¢ (-x).

(b) x = -le cot lea. System (1) yields
e-xa
a= -A sinka ' ~=0, C= -A
and the solution
•h = Ae= ,~-, = -Ae·-xa sin kx 'Pnr = - Ae-xx
'f'I ' 'I'll sin ka '
is an odd function: ¢ ( -x) = -¢ (x).

Coefficient A is determined by the normalization condi-
tion. For example, the first case gives
-oo
-a n oo
=I A 12 [. .Jr e2xx dx + cos2

e-Zxa r cos
ka .J
2 kx dx+
Jr e-2xx dx]
-oo -a a
-1 A 12e-2xa [_!+_a_
- x cos2 ka + 2ksin2ka J_1
cos2 ka -
If we substitute k t:m lea for x, we find the final expres-

sion for A,
I A l- 2 = ae-2xa ( 1 +__!_+
xa
x2
k2
+....2:___)
k2a
which holds for the second case as well.

To determine the energy levels we make use of the fact

that
where cl is a dimensionless constant that characterizes the

depth of the potential well. Whence,
xa= vc~-(ka) 2
and the roots (a) and (b) are reduced to
V q-kZaz =tanka and ka = -tanka
ka V q-k2a2
Specifying different values for C1 , we can fmd tho solution
by graphical methods. We start by graphing the functions
lf q-k2a2
ft= ka and / 2 =tanka
ka
fa= vq-k2a2 and f,.= -tanka
Now we can see that for C1 < :rc/2 there is only one energy
level, which corresponds to the even wave function. (The
curves y = / 1 and y = f 2 intersect only at one point, and
the curves y = f 3 and y = / 4 do not intersect at all.)
As C1 increases in magnitude, the number of levels will
grow.
46. 'The Schrodinger equation for a three-dimensional
oscillator
_ ;~ ll¢+ ( t
k1 t
+ k + k~Z) ljJ =
2 E¢ ( 1)
permits separating the variables:

¢ (x, y, z) = ¢ 1 (x) ¢ 2 (y) ¢ 3 (z)
If we substitute this function into (1) and divide all
members by it, we obtain, provided we denote rof =
= k;/m (i = 1, 2, 3):
'l't 2m dx2
2
( __1_...!!!.... d \jlt + mwf2 xz) + (__'i'z1_ !!}:_
2m
dZ'lJz + mw~
dyZ 2 Y
2)
+ ( __ 1 ..!!!.... d2\jla + mwJ z2) = E

'l'a 2m dz2 2
(2)
ANSWERS 251
Since x, y, z are independent variables, each parenthesis

must be equal to a constant, which we denote E 1 , E 2 , E 3 ,
respectively. The problem then comes down to three equa-
tions for the one-dimensional oscillator
d21Jlt mw2
-
fl2
2m dxZ +-i-
EtlJlt, etc.
X 2\jlt = (3)
and we bear in mind that E = E 1 +- E 2 +- E 3 • In dimen-
sionless variables £ = x V mw 1 /h, 'I'] = y V mw 21h, ~ =
= z V mw 31n, if we use the known results for the one-
dimensional_oscilla tor, we can get the general solution for
our case:
En1n2113= ( n1 +- ; ) hwt + (nz + ; ) 1iw2 + ("a+ ; ) hwa
~2+'112+~2
lJln1n 2na(£, '1'], ~)=Ce 2 Hn1(£)Hn 2 ('l'J)Hnam

Here Hni are Hermite polynomials, where n; = 0, 1, 2,
(i = 1, 2, 3), and C can be determined from the normaliza-
tion condition.
Let us normalize lJln = Cne-~ 2 1 2 H, (£). If we differentiate
d -£2
the identity ~ + 21;e-~ 2 = 0 (n+ 1) times and introduce
the definition
an -~2
Hn=(-1te; 2 ;Sn (4)
we see that (4) turns Eq. (5) of Appendix 6 into an
identity. Furthermore, it is clear that Hn can be written
in the form of a series: Hn (£) = 2n1;n an_ 2 £n- 2 + + ....
Substituting into the normalization integral
00 00
)l'iJn(£)i 2 d:t=J-/ m:t C~ ~ e-~ 2 H;,(£)d£=1

-00 -00
the definition (4) for one of the Hermite polynomials and

integrating n times by parts, we get
~z H
1= ,I
V
n C2 ( - 1) n
mw1 n Jf ane-
d£n n
(£) d£
-00
ann
But since a~nn = 2nn!, it follows that
C2
n
=v· r--
TnWj
nli
_1_
2nn!
It is evident, now, that the normalized solution for the

three-dimensional oscillator can be obtained in the form
~2+Tl2+t:,2
V m3wlw2w3 e- 2 Hnl (~) H"2 (11) Ilna (~)

'ljJ (x, y, z) = (lin) 3 / 4 X (2" 1 +n 2 +" 3 n!l X nz! X na!) 112
47. For this case the Schrodinger equation
---+
2m dx2
h2 d21p
-2x2 -eiE I x ) 'ljl=E'¢
( mw2
can be reduced to the problem of the harmonic oscillator

by completing squares in the expression for the potential
energy. Introducing new variables
X=X-~
I mw2 '
t
<,=
·(-h
mw
-Xh
E =E+ e21E12
1 2mw2
we arrive at the equation

_ .!!!._
2m dxi
d21p + mw2
2
x 2'¢ =
1
E 'ljJ
1
and we can (see the solution to Problem 46) write the eigen-
functions
and eigenvalues
of the Hamiltonian.
48. According to the solution of Problem 46 the oscillator
in its nth quantum state is characterized by a wave function
'i'n = Cne-6 2 12 H n (~)
ANSWERS 253
By definition
00 00
(x2)n = ~ I lPn 12 x2 dx = c;, ( mftw r:!(2 ) e-£2H~s2 d~

-00 -00
Now we substitute expression (4) from the solution of

Problem 46 for one of the Hermite polynomials in the above
integral and integrate by parts n times. We get
(x2)n=C~ ( :w )3/2(-i)n r d~~:£2 Hn~2d~
-00
Evidently,
dn(Hn~ 2 )
~n
dn (
d~n
tn+2+
an'<>
tn+
an-2'<>
)- (n+2)1t2
. . . - an -2-1- "' + n 1an-z
Here an = 2n and an- 2 = - n (n ~ i) an • Next we substitute
the values of the following integrals into (x2)n:
f e-s ds=Vn, f ~2e-~2 d~= ~1t

-00
2
-00
and get
(x2)n=C; ( :w r/2 [2n (n12)! ~1t -2n n(n;1) n! Vn]
= m~ ( n+ ~)
Hence,
= mw 2 ( 2) =~(
( V) n 2X n 2 n +_!_)-En
2-2.
49. Since for the one-dimensional case the kinetic energy

~ ft2 d2
operator is T = - 2 m dxZ , it follows that
00 00
(T) = - ..!!!:._
2m Jr 1h.
'I'
d21jl
dx2
dx = !!:.._
2m Jr I~
dx
12 dx
-oo -oo
In the lgiven instance n=3 and l\J=l\J 3 =C3 e-; 2 1 2 l/ 3 (~),

where H 3 = 8~3 -12~. We change the variable x to ~. x =
= ~ (mwfnt 1' \ substitute c:
= V mw/(izn) 23 ~ 31 from the
solution of Problem (46), and get
The integrals we have here are done easily if we differentiate

00
with respect to a parameter. Since J e-a' 2 d~= V nja,

-00
00
J
-00
~zne-as 2 ~= (- (}~ rj -00
00
e-as 2 d£
_ 1 X 3 X ... X (2n-1) 1 / Jt
- 2n r a2n+1
and
(T> 3-31Vn
_ ,.l(j) 21
Xz-X
v·-rt=z-nw 1 ( 1)
3+-z.
50. To solve the equation
-:: ::~ +V 0 (e-2ax_2e-ax)'\l=Eljl
it is convenient to introduce a new variable y = e- r:r.x. We

denote
(1)
and
2mE
8= h2a2 (2)
and we get an equation of the form
yz d2¢
dy2
+Y dyd¢ -~z (yz- 2y) l\J + e\11 = 0
ANSWERS 255
Let us ftrst study it at points

y-+ oo (x-+ -oo) and y-+ 0 (x-+ oo)
In the first case the equation takes the form
d2¢oo - ~2\jloo = 0
dy2
This has a solution
\jloo = e-f>y
(The second particular solution, \jloo = e+il11, will tend to
infinity when y tends to infinity.) Near point y = 0 the
substitution of ljJ = y"- gives an equation for k:
k2 +8=0, or k= + V -8
where we leave none but the terms of the lowest order.
For 8 > 0 both solutions will serve No = y±v=e =
= e±aV-8 x remains finite if V
- 8 is imaginary); the
energy spectrum is continuous.
For 8 = -"A2 < 0 only the solution ljJ 0 = y"A ("A > 0)
satisfies the requirement of finiteness.
To find the general solution, we must substitute \jJ =
= y"Ae-f>ll F (y) and then solve the obtained equation for
F (y):
y2 d2F
ayz dF
+ [(1 + 2"A) y-2~y 2 ] Ty + 2~y [ ~-A.- 21 JF= 0
Next we express F (y) as a power series, F (y) = ~ a,..y1\
1<=0
and when this is inserted into the equation, we find that
the coefficients of y"- satisfy the recurrence relation
2~ ( k+J. +! -~)
ak+1=a,.. (k+i)(k+i+21..) (k=O, 1, 2, ... )
a,.. 1 ~ k~ and,
Since for large k's the relation becomes a"-+ 2
hence, F y-+oo ~ e2 fiY, the only solution that will satisfy
the requirement of finiteness as y -+ oo is the one represen-
ted by a cut-off series. We can obtain this if for k = n,
say, ~ - n - 1/2 = "A; "A must be nonnegative. Now we
substitute ~ and "A from (1) and (2) into this condition, and
we get the discrete energy spectrum E, = - V 0 X
X [ 1- -.VI 2f>2rx2
~Vo ( n -t- 21 ) J · with
n = 0, 1, 2, ...
2
The· number of levels is restricted by Lhe condition
n+_!_ < R=-. I 2mV 0

2 t' Jl ft2rx2
and, apparently, depends on the depth V 0 of the well.

For instance, there are no discrete levels for ~ < 112.
112 d21jJ e2
51. The equation - 2m dx2 - Xj ljJ = Elp for E > 0
will have a continuous spectrum. tet us consider the case
2mE me2 .
of E < 0. We denote 1' 2 = -fi.2 and x = ft2.y and mtro-
duce the new variable ~ = 2'\'x. Now we can write the
equation in dimensionless variables:
(1)
.First let us consider the region > 0. s

When s tends to infinity, the equation transforms into
d~t; -! 'l'oo = 0 with a solution that satisfies the condition
of fmiteness, namely lpoo= e-~1 2 •
We introduce a new function f (~) and substitute ljJ =
= ri.1 2f (~} into Eq. (1). We thus get an equation for
f (s): ~ ~~
d~ 2 - Sdf
~
+
xf = 0. To study the behaviOur . of
s
f (s) near = 0 we substitute sa
for f and, retaining the
members of the lowest order, we find that a (a - 1) = 0.
In other words, there are two solutions: (1) f (s) = 1 and
(2) f (s) = s. The first solution cannot he used, however,
since the finiteness of ljJ (s) at point s = 0 will mean that
e21jJ d21jJ •
Vlp = - T -+ oo and, hence, d~2 -+ oo. Thus, ljJ (s) will
also he infinite. For this reason, none hut the second solu-
tion is suitable.
00
The substitution of f (s) = ~ aksk leads in the usual

k=1
way to a recurrence relation for the coefficients:
k-x
ak+ 1=ak (k-1-i)k (k=1, 2, ... )
ANSWERS 257
Next we take the limit of this formula for k }> 1, and we

find that in this case f (s) turns into e~. since their series
expansions coincide. Hence, for 'iJ = e-~1 2 f (s) to be finite,
f (s) must become a polynomial, and this is possible when
its series expansion is cut off at the nth member.
We can see from the recurrence relation that x must
me2
equal n for an+I = 0. If we substitute n = x = ---;(i: X
li
X V -2mE , we get a discrete energy spectrum
me4
En=- 2'2
.! n 2
and the corresponding wave functions are

n-1
'Jln (£) = e-s 12 ~ s k=O
ak,tSk
s
To fmd the solution for < 0 we introduce lJ = -£. The
equation takes on the very same form as for s > 0:
d2,p(l]) +(~-_.!._) ¢=0
dl]2 l] 4
and its solution, which continues 'iJ (s) into the region
s > 0, is
n-i n-1
'Jln ('I'])= - e-11/2'1'] ~ alt+t'I'J" = esl2£ ~ ak+t ( - s)k.
k=O k=O
52. To find the wave function for the spherically sym-

metric oscillator, as for any problem with a central sym-
metry, we must look for it in the form
¢ (r, e, <p) = / (r) P 1m (cos e) eim(j)
where P tm (cos e) is the Legendre polynomiall = 0, 1, 2, ...
and m = 0, +1, ... , +l.
Then f (r) must satisfy the equation
_ .!£_ [ dZJ +~ !:1._ _ z (Z + 1)
2f.l. dr2 r dr r2
!] + f.tcuz
2
rZj = Ef
If we introduce the new function U

solution to Problem 46, proceed to s= r
17-01496
=
vrf and, as in the

fln.W and A= h~'
the equation takes the form

azu +[2A-Z(l+1)- szJ U =0
d~Z 6Z
Obviously, as £tends to infinity, Uoo = e-£"' 2 • And as £
tends to zero, if we substitute U 0 = £a for a, we get the
equation a (a - 1) = l (l +
1). Hence, a 1 = l 1 and +
a 2 = -l, and none but a 1 gives the finite function as s
tends to zero, i.e. U 0 = £1+1.
We use these results to find U (£) in the form
U = e-6zl2£l+iv (£)
After substituting, we get the equation for v (£):
dZv
d~Z
+ 2 !::!._
d~
[ l +1
6 - £j + 2 [A - 2
J
l- ~ v = 0
and we look for the solution in the form of a series v =
00
= 1
· h=ko
ak\;k. If we equate the coefficients of each power
·
of £with zero, we get a recurrence relation for the coefficients
ak+z=ak
2[k l+ +;- A.
(k+ 2)(k+ 2l+ 3)
J~k (as k-+oo)
2ak
(1)
~:(k;;.ko)
Next we determine the lowest exponent k 0 • It follows from

k 0 (k 0 +
1) - 2 (l +
1) k 0 = 0 that none but the solutions
with k 0 = 0 remain finite at zero. From formula (1) it
follows that as £ -+ oo this series turns into e6" and U-+ e~·/ 2 •
To avoid this we must cut off the series, which can only
be done when A = p l ~ If we also consider that
+ + .
the series can only begin with k 0 = 0, then in this expres-
sion p must be even, i.e.
p = 2n, where n = 0, 1, 2, ...
Then
A=2n+l+; and En=nw (2n+Z+;)
The function that corresponds to an energy level with
given n and l is
e-~· 12 6 1 + 1 v (6)
'i'nlm(r, 8, fl!)=C ~ n P 1m(cos8)eim(p
,
ANSWERS 259
Obviously, the level is degenerate: when N = 2n l +

is even, we get l = 0, 2, 4, ... , N; when N is odd, we
get l ~= 1, 3, 5, ... , N. In addition, for given n and l
the level is (2l +
1)-fold degenerate in m.
53. For the two-dimensional problem,
u2 ia 1a2
11 = ap2 +p- -ap+p2 acp2
If V = V (p) and does not depend on angle <p, the variables
can be separated, which means that the wave function can
be sought in the form
'i' (p, <p) = einL(p U (p)
(H and iz commute in such a field.) For U we get the equa-
tion
_ !!!:._ (
2!1
d2U
dp2
+.!_p dU _dp
!!!!.._
p2
u) _ Ze2p U = EU
Now we consider E < 0 (for E > 0 the spectrum is

continuous). Substituting at for p and introducing the new
function f = yi)u, we get the equation
d2f
dt2
+( 2~-LE
h2 a
2 + 2!1Ze2 !!;__ m2-1(4) f= O
h2 t t2
h2
We choose a= ~1Ze 2 and denote
2 2~-LE 2
"Y=-~a (1)
We thereby simplify this equation and get
d2f ( 2 mZ- ~)
""dt2+ T-"Y -
2
tz f= 0
(O~t~oo)
As t-+ oo, we fmd that /oo = e-vt, and we discard the

solution e+'' 1•
For E > 0, it follows from (1) that "Y = i~. Then both
solutions e±vt = e±i(3p/a are finite and the spectrum of E
is continuous.
17*
As t-+ 0, assuming that / 0 = ta, we get a (a - 1) =

=
1
m2- 4' i.e. a 1 = m ~ and a 2 = - ( m - {-).
+
1
Since m = 0, +1, +2, ... , we must choose a = lml +2
and look for the general solution in the form
lml+~
f (t) = e-Y1t zv (t)
As a result, for the function v (t) we get the equation
00
and look for the solution in the form of a series v = ~ aktk.

k=ko
If we equate the coefficient of the lowest power of t with
zero, i.e. the coefficient of tho-I, we get
k 0 (k 0 +2 Im I) = 0
whence
k 0 = 0 or k 0 = -2 Im I
The second solution for t-+ 0 gives a function that tends
to infinity. The coefficients of th, where k = 0, 1, ... ,
being equated with zero, give the recurrence relation
2 [ y ( k+l m I+}) -1 J (2)

ah+i=ak (k+1)(k+1+2lml)
From this for k ~ 1 we get ak+l -+ ~ . (The function eZvt,
ak k
when expanded into a series, satisfies the same relation for
its coefficients.) To prevent v (t) from turning into e2vt
as t grows, we must cut off the series at, say, k = n, i.e. so
that
y[n+lml+ ~J-1 =0
[see formula (2)1. According to (1) we find that for the two-
dimensional Kepler problem the energy of a particle takes
ANSWERS 261
on the following values:

~-te4Z2
En=- 1 2
~~ 2 { n +I m I+ 2)
(n = 0, 1, 2, ... , I m I = 0, 1, 2, ... )
The function corresponding to an energy level with given
n and m is
1 -v.E... lml+~ n k
'i'nm (p, qJ) = eimcp VP" e a( ~) k~O ak ( ~) •
54. For a particle in a central field,

ljJ (r, 8, qJ} = eimcp P lm (cos 8) I (r)
Its radial part I (r) satisfies the equation
d2f +~.!!1..+( 2~-tE l (l+1) )I=O
dr2 r dr ft2 r2
for r < R, and f = 0 for r > R (see the solution to Prob-

lem 44). Hence, for f which is the solution for the above
equation the boundary condition will be f (R) = 0.
If we introduce k2 = 2~: and X (r) = (r) 112 I (r), for X (r)
we get the Bessel equation
azx + -
_!_ ax + 1k2- (l+2
1 )2-1 X = 0
dr2 r dr _ r2 _
±(1+.!.)
Since X (r) = J ( 1) as r-+ 0, the only solu-
(kr)-+ r 2
1+2
±
tion that satisfies the requirement of finiteness is
J 1 (kr)-+ Va _ sin kr (if l = 0)
l+z T
The energy levels corresponding to these functions can

be derived from the continuity conditions for the function
at r = R, i.e. J 1 (kR) = 0. We denote the roots of this
l+z
Bessel function by bU) and write
(t) ;,_z (bU>lz
En = 2J!R2
Evidently, at l = 0,
bn(O) = n"' d E(O) -- f12n2n2
•• an " - 2~-tR2 •
55. When the electron in the hydrogen atom is on the
lowest energy level, then n = 1, l = 0, and m = 0. Accor-
ding to formula (III-24), the function that corresponds
to this level is 'ljl100 = Ce-,.fa. The constant C is determined
from the normalization condition
00
) I'i'foo j2 d-r = 4n:C2) e-2rfar2 dr = n:C2a3 = 1

0
00
The integral is ofthe type .\ e-xxndx = nl. Then I 'ljl100 j2dT

0
is the probability of fwding the electron in dT. We then
calculate
00 00
"(r)=
Jr ri•h 't'fOO
l2 4:rtr 2 dr=_i_ r e-2rfar 3 dr=~1!_a~=~a
a3 J a3 24 2
0 0
00 00
(r2 ) = ) r 2 1 'iJ 100 12 4nr 2 dr ~~ ; ) e- Zr!ar~ dr = 3a2

0 0
We find the most probable value r 0 by equating with
zero the derivative of I 'ljl100 l2 r 2 = C 2e-Zrfar2 , which deter-
mines the probability of finding the electron at a given
distance from the origin. Hence, r 0 = a.
56. To write the required functions, we use formulas
(111-23), (III-24), and (III-24a). At n = 2, l can be equal
to 0 or 1.
If l = 0, the sum in formula (III-24) has two members,
with k = 0 and k = 1, and the angular part Y 00 (El, <p) = 1.
b1 r )
Thus 'iJ 200 (r, El, <p) = b0 e-r!<Za} ( 1 +bo-a , where b1 - 21. ro=
[see formula (III-24a)], and the normalization condition
1= JJJ I 'i'2oo 12 r 2 dr sin ede d !_cos e
a
'ljlzu = b; e-rl( 2a> : sine eiq>;
'ljlzt-1 = b; e-r/(2a) !_sine e-i~

a
The normalization constants are
b1 = - 1
2 V 8na3
and b;· = 8 V1na3a
57. Since the potential energy does not depend on the
angles
'¢ (r, 8, !Jl) = U (r) Pzm (cos 8) eimq>
and the radial part of the function, U (r), satisfies the
equation
-~ ( d2U +~ dU
2/l dr2 r dr
_z (Z+1)
r2
U) _ (.:!..-.£)
r r2
U= EU (1)
Next we turn to the dimensionless coordinate p = r/a,

liz 2En2 •
where a = - 2 , and energy e = - : r = -"\'2 (we w1ll
~e ~e-
examine none but the discrete energy spectrum E < 0).
We collect in formula (1) the"''members of type 1/r2 and
denote s (s +
1) = l (l 1) + + h2.
2p.C
Now we can rewrite
the equation in the form
d2U+_!dU
dp2 pdp
_s(s+1>u+(.!-
p2 p "\'
2 )U=O
We introduce the function 'X= rU and- get the equation

~-+[ :·-"\'2 _ s (spt1) JX. O
Using the usual method, we find Xoo = e-.VP and Xo = p•+I

(the second solution Xn = p-• is not suitable for s > 0).
The substitution of x (p) = e-'~'Pp'+If gives the equation
for f:
p~J+ [2 (s+ 1)-2-vp] ~~ -2 [y(s+ 1)-ir/= 0
and we look for the solution in the form of a power series

00
/= ~ akp"
h.=O
By equating the coefficients of p11 with zero, we get

2[y(k+s+1)-1]
ak+l=ah. (k+i) (k+Zs+Z) (k=O, 1, 2, ... )
To prevent f from turning into e2VP as p-+ oo (the possi-
bility can be seen from the asymptotic relation ak+t = 2ki')
ak 11»1
we must cut off the series, i.e. we must see that
, / 2n2E 1
'V= V -fle4= k+s+1
Hence, the energy levels of the particle are determined
by the formula
)!e~
Eks=- 2n2(k+s+1)2
where k= 0, 1, ... , and s = - ~+ y -(l-+---:~c-:)-:::-+---,-~-{:::- ~

2 2
~ l+ )lC if 2~; ~ l + ~.
/j2 ( l+ ~)
If we introduce the principal quantum number n = k +
+ l + 1, the energy
E - -)le'
nl- 2fi2 [n + )lC ]2
/j2 ( l + ~)
proves dependent on n and l, and the functions that corres-
pond to these levels
k
'IJlnlm (r, 8, <p) = e-VPp' ~ app11P!m (cos 8) eim<P
p=O
depend on the n 1mbers n, l, and m. We now see that

En 1 is (2l+1)-fold degenerate (since m = 0, +1, ... , ± l).
58. As usual we assume that lJl (r, e, <p) = U (r) P 1m(cos 8) X
X eim<P, and for the function F = rU we get a problem
ANSWERS 265
~i milar to Problem 52. Its solution gives us
En,=na[4n+2+ V<2Z+1)2 + kJ
8 A
2
where
?: _ V2~.tB r 2
<:,- 1i
k-n
a,.+t =a,. 1 (k= 0, 1, 2, ...)
(k+1) ( k+2a + 2)
a= 1+ 1V(2l+ 1)2 + sk: .

59. Using the solutions to Problems 43 and 46, we can
write the solution in the form
•h ( )
'Ynn 1 ~ x, y, z =
A e -62f2H n (?:) • -a-
"' Sill nt:rtY • n2:rtz
sm-b-
s=xvll:
for 0 ~ y ~a and 0 ~ z ~b. In the rest of the space
¢= 0. The energy levels that correspond to the function are
1 } . Ji2nf n2 + Ji2nin2
n+ 2 + 8~.ta2
- (
Enn1n2 -nw 8~tb2
The normalization condition yields
AZ 4 -.~ 1
· Qjj V lin X 2nn1 •
60. We denote· the ordinary coordinate as r', and the
Laplacian in relation tor' as 11'. Next we turn to dimension-
less vafiables r = r' Ia and 11 = a2 11' (where a = 1i2/fle2 is
the Bohr radius), and to the dimensionless energy e defi.ned
by the~relation E = ee: e ,
= ~~ and we write the Schro-
dinger equation for the hydrogen atom:
1 1
-211'1'--,:- 'I'= 8'Jl
In parabolic coordinates (see Appendix 3) the equation

takes the form
2
u+v [a( Otjl) a( aw)
_Tu U Tu +Tv Va,;
4 u v a2"'J+-2-~J+8'1Jl=0
+_!_(..!.+_!_) a~p2 u+v
Evidently, we can separate the variables:
'ljJ (u, v, cp) = U (u) V (v) <D (cp)

Here <D(!p) satisfies the equation
1 d2!D
- - = constant= - m2
!D -
d~p2
whence
<D === ei1rnP
where m = 0, +1. +2, . . . .

- We substitute ~~~ = -m2¢ into the equation and multi-
ply it by utv. We thereby get like equations for U(u)
and V (v):
d (
-du udU
- ) --U+-uU+aU=O
du
m2
4u
8
2
(1)
d ( vdV)
-dv - m2
--V+-vV+~V=O
8
(2)
dv 4v 2
with
(3)
We seek the solution of (1) by examining U (u) as u ~ oo
and as u ~ 0. Obviously, Uoo is determined from the equation
d2U 00 8
dU2+2Uoo=0, I.e. Uoo=e
• V- ~
±u
(1.£~0)
For 8 > 0 both solutions are finite and the energy spectrum
is continuous. For 8 < 0 the only finite solution is U,;, =
= exp ( -u V 8/2). Let us find the energy spectrum 8 < 0.
For u ~ 0 we_-assume that 'U0 = uv, and by equating
with zero the coefficient of the lowest power of u, we get
ANSWERS 267
m2
y2 4 = 0 or y = +m/2. We must put y = I m 112 for
-
the finiteness of U 0 •
We look for the general solution in the form
U=ulml1 2 e-uV-fxF(u), where F= I;akuk

k=O
The equation for F (u) in the form
U ~; + (Im I+ 1 - 2u J/r- ~ ) ~~
+fcx-(lm1+1>V -~JF=o
gives us the recurrence relation for ak, which we get by
equating with zero the coefficient of ull. (k = 0, 1, 2, ... ):
~ ak
2v -;k for k "J;-1
For large u's, i.e. when the members with k "J;. 1 play the
dominant role in F, we fmd the asymptotic form F (u)-+
v - 8/2) rSince
00
-+ exp ( 2u ePX = ~ (p~)k and the ratio of

k=O
the coefficients of xk+t and xk in this series is
pk+t pk p -]
(k+f)! :k!~ k . Hence,
U = e -uV- ~ X F X ulmJ/2-+ e+uV- ~ X uJml/2
and tends to infinity as u-+ oo.

Thus, the series for F (u) must be cut off. This is possible
if for, say, k = n1 , the following equality holds:
ex= y- ~ (2nt + I m I+ 1)
nt
Then F (u) = ~ a~~.uk = Fn 1 (u) becomes a polynomial of
k=O .
degree n 1, and U is finite everywhere.
We solve Eq. (2) in a similar manner and we find that

for V (v)
~ = -(- ; (2nz-f- I m I + 1)
and
Vn 2 =e -vV- ~ X vlml/2 X Fn 2 (v)

It follows from condition (3) that
2(n1 +nz+rml+ 1)-(- ~ = 1

Introducing the principal quantum number n = n 1 + n 2 +
+ I m I + 1, which obviously can take on the values
n = 1, 2, . . . , we get
1
En=- 2n2 and 'li'nln2m(U, v, <p)=CUni(u}Vnz(v)eimq>
u+v
= Ce -"'""'2'nulml/2vlm(/2Fn1 (u)Fn 2 (v)eimq>.
6L In a central field for l = 0 the wave function is simply
the radial part, '¢ = R (r), which satisfies the eqlJation
_r
fi2 [ 1 tP
-2; J
dr 2 (rR) -V 0e-rlaR=ER
We will confine ourselves to E < 0. We introduce the new

function X = rR and make a change of variables y = e-rl2 a.
Then X will satisfy the equation
d2X
dy2
+. ..!_y ~+
dy
(cz- _f!_)
y2
X= 0
where ~.
C2 _ 8!J.V0a2 2_ 8!J.Ea2 >O
- !i2 ' q --~
This is Bessel's differential equation whose general solution
is
As .y-+ 0 (i.e. as r-+ oo), X must remain finite. In this

case J_q ex y-q becomesJ infinite. Hence, C2 = 0. As y-+ 1
(as r-+ 0), X must be equal to zero. Therefore, Jq (C) = 0,
ANSWERS 269
i.e. any value of C must be a root of the qth Bessel function.

Since the values of the roots increase as q increases and
the first root of J 0 is approximately 2.4, we can estimate
C > 2.4, i.e.
This is the condition for there to be at least one energy

level in a potential well with a depth of the order of V 0
and a width of the order of a. For r > a and as y-+ 0, we
can obtain a similar form of the wave function
Jq(Cy)=Ay<I=Aexp(- ;~)=Aexp(- rV~ ).
62. First we link the solutions 'ljJ (x) and 'ljJ (x + l). The
Schrodinger equations for these cases are
h 2 d21p (x)
- 2 f! dxZ + V.(x) 'ljJ (x) = E'\jl (x)
h2 'dZ'IjJ (x+l) ,
-2/! d(x+l)Z +V(x+l)'\jl(x+l)=E'ljl(x+l)
d2
We see that because V (x + l) = V (x) and d(x + l)2
d2
= dxz,both functions correspond to one energy level E.
Considering E a simple eigenvalue, we find that these
functions differ only in a constant factor, i.e.
'¢_(x l) = p'ljl (x) +
In the general case we can obviously find that
'ljJ (x + nl) = pn'ljl (x) (n = 0, ±1, ... ) (1)
It follows then from the requirement of the finiteness of
I p I = 1, i.e. p = eikl, where k is any real number;
'ljJ (x) that
'ljJ (x + l) = eikl'¢ (x) (2)
If we solve the Schrodinger equation for -b =:::; x =:::; 0
(the first region), for 0 =:::; x =:::;a (the second region), and
for a =:::; x =:::; b (the third region), and use condition (2)
we obviously will have a solution for all x's.
First region (V = V 0 ). The equation takes the form
- _!!!_ d21p1
2fl. dx2
+ Vo'IJli = E'\jlr
Denoting ').} = 21-l (V 0 - E) , we can write

1i2
'\)Jr = C1e/..x + C2e-J..x

Second region (V = 0). Here the equation is
-..!:!:.... d2'im = E'\jJn
2!! dx2
and the solution is
'\)Jn = Caeix.x + C,.e-ix.x
where x 2 = 2f1E/n2 •
Third region (V = V 0 ). The equation is the same as in
the first region, i.e.
!i2 d2.¢m
- 2/l----;w:---- + Vo'\)Jnr = E'\jJm
and its solution is
'\)Jm = C5 e'J..x + C6 e- )..x
If x lies in the first region, x + l lies in the third, and
according to formula (2) the solutions are linked by the
condition
'\)Jm (x + l) = eikl'\jJr (x)
Hence, C5 =C 1eikl-'J..l and C6 =C2eikl+'J..l, i.e.
'\)Jm = eikl (C,e'J..(x-1) + C2e -'J..(x-ll)
Solutions lJli, lJln, lJlm must be continuous with their
first derivatives in passing from the first region to the
second (x = 0) and from the second to the third (x = a).
This leads us to the following conditions:
'\)JI (0) = '\)Jn (0) and C1 + C2 = C3 + C,.
dkr lx=o= d:f:I lx=O and A.(C1 -C2 )=ix(C3 -C4)
'\)Jn (a)= lJlnr (a) and C3eix.a + C,.e-ix.a = eikz (C 1e-'J..b + C et..b)
2
dljln
dx
Ix=a = dljlrn
dx
Ix=a and ix (Caeix.a- C,.e- ixa)
= A,eiht (C,e-'J..b- C2eJ..b)

ANSWERS 271
Now we have a system of four homogeneous linear equa-

tions for the unknowns C1 , C 2 , C 3 , and C~. For there to be
a nontrivial solution, the system determinant, as we know,
must be zero:
1 1 -1 -1
'). -I. -ix ix
-eikl-?..b _ eikl-t.b eixa e-ixa =0
_ f..eikl-?..b f..eikl-?..b ixeixa - ixe-i><a
This brings us to an equation that determines the energy
of the particle in the periodic field:
J..2-x2
f (E) === cosh 'Ab cos xa + 2A.x sinh Ab sin xa = cos kl (3)
On the left the energy enters through x and l. It is ob-

vious that for this equation to be valid, I f (E) I must
not exceed unity. We can see that at xa = n:n:, this is not
so, since then f (E) = +cosh Ab and I f (E) I > 1. Hence,
the energies
E = !i2n2:n2
2~ta2
prove to be forbidden for a particle in a periodic field (an

electron in a crystal).
63. In the limiting case as V 0 ---+ oo and b---+ 0, so that
sinh Ab
M ---+ 0 and Ab ---+ 1, Eq. (3) of Problem 62 turns into
the identity
sin xl
f (E) =cos xl +P-x-l-=cos kl (1)
where P = lim A.~ab . If we denote tan ~ = :Z , it takes

Vo-+oo
b-+0
the form
cos (xl- tl)
cos kl
cos tl
Obviously, the edges of the allowed energy bands cor-
respond to the conditions cos (xl - ~) = + cos ~. i.e.
xl = n:n: or xl - 2~ = n:n:. Substituting xl = n:n: - e, we
get
(-1)n (cos e- tan~ sin e) =cos kl
For 0 < e ~ l the expression on the left is less than one.

Consequently, xZ = nn are the onsets of the forbidden energy
bands. We will find in a similar manner that xZ = nn + 2~
are the onsets of the allowed energy bands. The nth energy
band is determined by the xl that lie in the limits
(n- 1) n + 2~ ::::;;; xl::::;;; nn
If xl :::2> 1, then ~ = arctan .!__l

x
~.!....
nn
and the width of the
forbidden energy band between the nth and the (n + 1)st
allowed energy bands is
~ (xnl) = 2~=~
nn
If we take the valut x = x 0 , for which kl = 0, and expand
the left-hand and right-hand sides of (1) as I (E) = A (x 0 ) +
+ B (x - x 0) and cos kl = 1 - -(kl)2 2 - , from the equation
(kl)2
A (x0 ) +B (x- x 0 ) = 1 - -2-
we get
X= C +Dk2
Since x = V
2~2E , we have E = E 0 + k 2F. The coefficients
C, D, Eo and F can be expressed in terms of A, B and x 0 •
64. We introduce kl = ~ and write Eq. (1) of the
solution of Problem 63 in the form
I (~) ==cos s+
p sins
s = cos kl
In Problem 63 we have found that the right edge of the
allowed energy band corresponds to ~ = nn and that at
this point I(~) = ( -1t. In the forbidden energy band,
where ~ > nn, the equation holds for complex values of kl
since I I I > 1.
If I(~) > 1, then cos kl = cosh 11. > 1 and kl = if-1.
If I(~)< -1, then cos kl = -cosh 1-1 < -1 and kl =
= if-1 + n.
The function I (s) > 1 for s ~ 2nn and for sin ~ > 0,
and I(~) < -1 when s
~ (2n + 1) n and sin < 0. s
Hence, if we introduce e = +1 so that e sin ~ > 0, then
cos kl = e cosh ll and in the forbidden energy band the
ANSWERS 273
following condition holds:

P sin~
e cosh 1-L = f (s) = ~ + cos ,.E {1)
Next we find the solution for the Schrodinger equation

for a semi-infinite crystal. Now condition (1) of Problem 62
holds for none but n > 0 (since for n < 0 the potential is
not periodic) and, hen~e, I p I can also be less than unity.
If we introduce p = e1" 1, solutions with complex k become
possible provided Im k > 0.
For x < 0 the equation
- !!:...
2[1 d
2
dx2'1J + J;V O't'
'" = E•"'t'
has the solution
Vq2-~2-=-
'l-'x<O = Ae l
where we denote q = l (2!-L W 0 /li 2 ) 1 12 and = xl = s

= l (2,...,E/n2 ) 1 12 • Here (q2 - £2) 112 > 0.
For x > 0 (inside the crystal) we consider region I
(0 ~ x ~ l), where
'!-'I= C1eixx + Cze-ixx {2)
In region II {l ~ x ~ 2l),
(3)
But if x = 0, then x = 0 + l = l lies in region I, and we
can consider C 3 = C1 and C4 = C 2 • By virtue of formula (1)
of the solution of Problem 62, we get the condition
C 1eixl + C 2e-ixl = eikl (C1 + C2)
i.e.
1-ei(~-kl)
Cz= -C11 -e -i<Hkt> (4)
where, as before, s
= xl. Hence, for x > 0, the function
is determined by (2) with condition (4).
Because ¢ and ~~ must be continuous at x = 0,
A = C+c
1 z an d AVq2-~2
l
_. (C
- t'X
C)
1- z
18-01496
274 f>tiOnl,EMS IN 'l'!IEOllETlCAL PHYSlCS
whence
V q2_£2= eikl-:-coss
sms
s (5)
Let ~1s consider this equality for complex values of k.
s
Then e'k 1 = erll (where e sin ~ 0), and condition (5)
gives
V- q2 - - sins
s
£2 -
s +
- cos = ee- ll (G)
More than that, condition (1) must also hold. If we subtract

(6) from (1), we get
(P-Vq2-£2) 5 i~s=esinhf-L (7)
Since sin s and e have the same sign and Vq 2 - s 2 , s,

and f-L are greater than zero, we get
P-V q2-£2>0
i.e.
q2 - p2 < £2 < q2
Only this condition allows for the existence of additional
levels corresponding to complex values of k. We square
both sides, subtract (7) from (1), and get an equation that
determines the energy levels (s):
;; - V q 2- s2 = £cots (8)
Now let us show that the function corresponding to these

energy levels, which lie in the forbidden energy band,
decreases as I x J increases on both sides of the boundary
"crystal-vacuum" (the plane x = 0).
We fmd that for x < 0, the solution 'llx<o =
= A exp ( : V q2 - s 2 ) does have this property. For~ > 0,
the solution satisfies the condition lJl (x + l) = e'k 1lJl (x)
(a periodic field), which we can write in the form
'ljl(x) 'ljl(x+l)
eikx = eik (:.:+1) u (x)
where u (x) is a periodic function. Consequently, for a
complex k,
'¢ (x) = eikxu (x) = ee-llxtlu (x)
ANSWERS 275
Hence, we have found a state with an energy level lying

in the forbidden energy band, and the probability of disco-
vering the particle decreases exponentially on both sides
of x = 0 (the surface of the crystal). The value ~. i.e. the
position of the energy level, can be found by solving Eq.
(8) graphically.
65. Since Px = -in :x and (;; = x, the condition
J(rJ)= J1Px'i'+irJxti'l2 d't~o

yields
(1)
Let us consider the Hamiltonian of the one-dimensional

harmonic oscillator:
fi = pz + f.lroz;z
2~-t 2
. Iy,
Ob vwus (pz)
(H) =~ + -f.lroz2- (x 2)
. Sus
b t"t
1 u t·mg 4 (xZ)
liz f or(p 2 ),
we get
nz 2
(II)~ 8!J. (x2) +f.l~ (xz) = f ((xz))
We choose (x 2 ) that corresponds to the minimum off, i.e.
use the condition
of h !J.ro2
iJ (x2) = - Sf.l ((x~))2+-2- = 0
and substitute it into the expression for f ( (x 2 )):
2 nro liro
f ((xo}) =-z' (H)~T·
66. We introduce the coordinate of the centre of mass X
and the relative coordinate (separation) x = x 1 - x 2 ,
separate the variables, and get
.PX
1-
~) = Ce h e-s2f2 Hn (~)
where s=,VI ~ro x and f.!= m:;' 2 • The energy levels

" mt m2
are
18*
216 Pll.OBLEMS IN THEOll.ETICAL PHYSICS
67. We separate the variables as in the solution to Prob-

lem 66. Then we can look for the function in the form
¢ = F (R) <!> (r)
where F = eiPR/t• and <!> = <'Dnlm (r, e, qJ) = Unl (r) X
X Pzm(cos 8) X eimiJl. Here Unz = e-r/(n4) X~

n-1
k=l
al< ( : r is the
same as in formula (III-24), and

E - p2 Jle4
n, P - 2 (m+ M) 2Ji2n2
(P is continuous and f.L = mm:M , m and M are the masses

of the electron and the nucleus, respectively).
68. Obviously, the total potential energy in this problem is
V (xt, x2) = ~ (x~ + x:) + ~~ {x1 - x 2) 2
where k and k 1 are the elastic constants that characterize
the interaction of the particles with point x = 0 and with
each other. If we introduce the coordinate of the centre of
t
mass X = Xt x2 and the relative coordinate (the separa-
tion) x = x1 - x 2 , we get the equation
1i2 d2¢
-2M dX2
1i2
-Zit
d21jl +Moo2 2
dx2 -2- X ¢ + fL 2oof X ¢ = E¢
2
where M = 2m is the total mass of the system, f.L =' m/2 is

• /k" -. / T d -. /k + 2kt
t h e re d uce d mass, w = V M = V 2m , an W1 = V 2!l
= Vk~kt . We separate the variables and, if we substi-
tute ¢ = f (X) F (x), we get two one-dimensional equations
for harmonic oscillators with natural frequencies w and w1 :
ANSWERS 277
By denoting ~ = .. /Mii X and u = .. /T x, we can by ana-

V ro V ~
logy with Problem 46 write the solution as
'l'ntna = Ce-62f2Hnt (s) e-u212Hna (u)
where Hn is the Hermite polynomial. The energy level that

corresponds to this function is
En 1n2 = ( n1 + {-) 1iro + (nz + +) 1iro1.
69. In the x-representation the particle in such a poten-

tia I we II Wl. th an energy E n = nZ/iz:n:z
.
2maz
h as m
· t h e mterva
· I
0 ~ x ~ a the corresponding function '¢n =V

,-
.! sin n:n:x ,
a a
for x >a or for x < 0 the wave function is zero (see the
solution to Problem 43).
The momentum distribution is determined by the wave
function in the p-representation, to which we can pass
from 'ljl (x) in the usual manner:
00
cp (p) = J
-00
'ljl (x) 'ljl; (x) dx
where 'ljlp(x) = (2:n:!)l/ 2 eipx/h are the momentum eigen-

functions in the x-representation, the functions being
normalized so that
00
J 'ljl; (x) '\jlp• (x) dx = 6 (p- p')

-00
For the given case we substitut8 n = 2 and '\jl2 (x)

and get
00
cp (p) = r '\jlz (x) (2:n:li)l/2

J
i eiPX/h dx
-oo
a
= 1 \ s1n- . /h d x
• 2:n:x e-tpx
(:n:lia) 112 • a
0
2 e-ipa/h_i
-- (:n:lia)I/2 nan 2 ~~--~
4:n:2Ji2- p2a2
Consequently, the probability of fmding the particle with

a momentum in the interval from p to p dp is +
!!!!:_)
32:rran3 sin2 (
dW(p)=!fP(p)j 2 dp= (a2p2-4n21i~~2 dp.
70. In the x-representation a new :i: = x. To pass to
representation we use the expression for the expectation
value of A.:
(A) = J1p* L1p d-t
In a similar expression for (x) we substitute 1p (x) expres-
sed in terms of qJ (p):
00
1P (x) = (2n~)I,2J ) qJ (p) eipX/11 dp

-00
and get
00 00
(X)= f qJ* (x) X 1 lf2 f qJ (p) eipx/h dp dx

J
-oo
(2rt1i} J
-oo
We change the order of integration over p and x, note that

xeipx/11 = ~ ~ eipx/ 11 , integrate over p by parts [qJ (p) = 0
at the limits of integration by virtue of the requirement
that I qJ {p) 12 be integrable] and get
(x) =
r
J 'P* (x) (2nli)I/2 J
1 'f'
qJ (p)
1i
-i
8eip:x;f11
8p dp dx
-oo -oo
00 00
= J
r (2:rtli)
'ljl* r eip:x;/11 in 8pBcp dp dx
1f2 J
-oo -co
Jf 'P*
Since -- 1- eivxtn dx = rp* (p) we have
(2:rth) 1/2 ,
00
(x) = J
-oo
qJ* (p) iii :P qJ (p) dp
i.e.
ANSWERS 279
We look for the eigenfunctions of :i:P going by the usual

rule. These are functions that satisfy both the equation
in acp
ap
= x<p (p)
and the conditions of finiteness, single-valuedness, and
continuity. The solution of this equation
<p (p) = Ce-ixp/h
will satisfy all the conditions for any real value of x; the
spectrum of :i: is continuous.
71. For a particle in a homogeneous potential field,
~ ~ a
V = Ax = A in iJp (see Problem 70). The equation for the
eigenfunctions of the Hamiltonian in the momentum repre-
sentation has the form
L2m <p +inA arp = E<p
ap
We separate the variables and write
E-p'2. i (P 3 )
dcp 2m hA --Ep
(j)= iliA dp, whence <p-Ce 6m
The finiteness of <p (p) is ensured for any real value of E,

i.e. the energy spectrum is continuous. We determine the
constant C from the normalization condition
oo
) <Jlk(P)<JlE•(p)dp=6(E-E')=C 2 J _!_
oo
ehA
(E' -E) p
dp
-oo -oo
:/iA.
00
Since 6(E-E')= 2~ J ei<E-E'>Ydy, we have C2 = 2

-00
72. We determine the wave functions for the type of

potential given in Problem 43.
For x > ; or x <- ; , the wave function 'ljJ (x) = 0.
For - ~ ~x~;,
'ljJ (x) = Aeikx + Be-ikx
where k = (2mE/n 2 ) 1 12 •
The requirement of continuity of the wave function at

x = ± ; yields
'I' ( _ ; ) = Ae-ika/2 + Beiha/2 = 0
(1)
'I' ( ; ) = Aeika/2 + Be-ika/2 = 0
whence ka = nn. Condition (1) reduces to (A +B) cos n; =
= 0 or to (A -B) sin n2:rt = 0, i.e. for n even we have

'i'n (x) = V-T. mrx
a SID - a
and for n odd
'i'n (x) = V-cos--
2
a
n:rtx
a
(the functions are normalized in the usual way).
a/2
The matrix element (n I ex I m) = e
-a/2
J ¢~x¢m dx is
nonzero, obviously, only when n and m have different

parities. Otherwise, the integrand will be an odd function
and the int<'gral will become zero. Let n be even and m odd.
Since the integrand is even, if we introduce y = ~ a
, we
can write
I I = 2e 2a-
(n ex m) J x sm. -
a/2
nnx
a
m:rtx d
cos-
a
x
I)
n/2
= 2: ; J fsin(n+m)y+sin(n-m)y]ydy
0
n+m-1
nf2 • sin (n m) + ~ (-1)_2_
and since J y SID (n + m) y dy = (n+ m)~ (n-j- m)Z
0
we have
n+m-1 n-m-1
2ae
(n I ex I m) =:rt2
[(-t)-2- (-1)_2_ J
- (n+m)2 +~--'----;-;;-
(n-m)2
n-m+i Baemn
=( - i ) _2_n2(n2-m2)2
ANSWERS 281
a/2
By ·analogy the expression (n I x 2 1 m) = J x2.¢:'1jlm dx
-a/2
is nonzero only when n and m have the same parities. If n
and m are even,
a/2
4 r 2 . nnx . m:nx d
(n IX m) =a J X S l l l a Slll a
21
X
0
11/2
=~32 J y [cos(n-m)y-cos(n+m)y]dy
2
0
71-1ll
( f)-2- 8a2nm
= - n2 (n2-m2)2
since
n-m
11/2
i n n n(-1)_2_
J y2 cos (n- m) y dy= (n--m) 2 cos (n- m) 2 = (n-m)2
0
For n and m odd

4a2 n2+m2
(n I x21 m) = 3l2 (n2-m2)2
The matrix element of momentum is
+a/2
(nIp I m) = ~ \ 'ljl~ d~; dx =F 0
-~/2
if n and m have different parities. For instance, with n
even and m odd,
li2
(nlplm)=-:--
z a
J.sm--d
a/2
nnxd( cos--
a x
mnx)dx
a
-a/2
11/2
= ---:--
li 2m
1 a
J [cos (n- m) y- cos (n+ m) y] dy
0
=-~2m [ ( - 1 ) -2--
n-m-1
_
n+m-1
(-1)_2_
J
1 a n-m n+m
= ( -f)n-m+i ~ 4mn
1a n2-m2
If n and m exchange parity, the result for (n I p I m) will

be the same.
73. The equation for the eigenfunctions of a one-dimen-
sional harmonic oscillator in the p-representation has tho
form
cp (p) + - 2- iii
p2 mro2 ( d )2
2m dp cp (p) = Ecp (p)
If we introduce the dimensionless variable 'YJ = pl(mnffi) 1 12

and put li2(t)
A., we arrive at an equation that coincides
E=
with the dimensionless equation for the same problem in
the x-representation. Referring to Problem 46, we can write
cpn (p) = c-TJ212 H nTJ (Hn is the Hermite polynomial)
En= (n++)nffi
Hence, for a simple harmonic oscillator the coordinate
and momentum distributions are similar.
74. For a simple harmonic oscillator
~ p2 mro2q2
H= 2m+ 2
and
(H)=~~+m; 2 (q2)~0
Consequently, if has no negative eigenvalues, i.e. Hn ~ 0.
We introduce
x= iJ+imffiq {1)
.X+= p-imffiq (2)
and bear in mind that pq- qp = -in. Now we can repre-

sent ii either as
1~ 1 ~ ~ 1 ~ ~ ~~ ~~
H = 2m (p 2 + m 2ffi 2q2 ) =2m [XX+ imffi (pq-qp)}
= _1_ XX+ liw (3)

2m 2
or as
ANSWERS 283
We construct the commutator

iiX-Xii=(-2m1-xx+ +!::::_)x-x
2
(-2m1-x+x- nw)=nrox
2
In like manner
frX.+-x+if = -nroX+
When we have written these commutators in the represen-
tation that diagonalizes il, i.e. (n I H I n') = lln{Jnn'• we
get
nro (n I X I n') = (n I H X I n')- (n I XH I n')
= ~ [ (n I HI n") (n" I X I n')- (n I X I n") (n" I HI n')]
n"
= (Hn -Hn•) (nl X In')
It follows from this that
(nIX In') (Hn- Hn• -Tiro)= 0
i.e. (n I X I n') is nonzero only if Hn• = Hn - n(J). By
analogy <n I X+ I n' >=I= 01 if" Hn· = Hn + n(J). We write
the relationship · ·
1-x+x-~
H=-2m 2
in the same representation, and for the diagonal matrix
element we get
(nIH In)= H n = 2~ ~ (nIX+ In') (n' I X In)- li2w
n'
In the sum the member with (n I X+ I n') can differ from
zero only if Hn' = Hn +Tiro is also an eigenvalue of H.
Otherwise, the sum will turn zero and H n will be equal to
- ~ro , which is impossible in view of what has been said.
Consequently, the first assumption must be true, i.e. if Hn
is an eigenvalue of ii, then H n, = H n +
n(J) is also its
eigenvalue.
Now we consider expression (3). In a similar way we
find that
Hn = 2~ ~ (nIX j'n') (n' rx+ In)+ ~(j) (4)
n'
where Hn' = Hn - 1iro, and either Hn• is also an eigenvalue

of H (i.e· H n is not the lowest energy level possible) or,
if this is not so, (n I X I n'} = 0 and Hn = is the !i;
lowest energy level, and the next level lies tiro higher.
Hence, we have found that
(5)
and (n I X I n - 1 } and (n I X+ I n +
1 } differ from zero.
If we bear in mind that ( (n I X I n- 1 })* = (n - 1 I X+l n} 1
we can obtain from (4) and (5) the equation
I (n I X I n - 1) 12 = 2mnliro
Whence
(nIX I n-1} =((nIX+ I n-1))* = V2mhron eicp
Returning to (1) and (2), we get (choosing ei!Jl = i)
(n Iq In -1} = (n -11 q In)=~
(n I PI n-1) = ((n-11 PIn))*= i V m;w n.

75. Let us use the commutation relations for angular
momentum
Lxiy- Lyix = in£: and ixL2 - I)Lx = o
where L2 =L~ +L~+L~. Now we introduce
x =Lx+iLy and X+=Lx-iLy
Obviously, Lz, X, and X+ commute with L2 • We write the
relationship L2 X- XI}= 0 in the representation where L2
and Lz are diagonal, i.e. <ll VIZ'>= M6w, and we get
<Zm IX ll'm' > (Lf- M·) = 0
Here the only matrix elements that differ from zero are the
diagonal in l matrix elements of i, and also of X+ and iz.
For this reason in future relationships between these opera-
tors we can simply consider f.s to be its eigenvalue Lf.
Let us represent L2 in the form
f} = (Lx + iLu) (Lx- iLu) + i (LxLu- LuLx) + L:
~ ~ ( ~ 1i )2 fi2
=XX++ L~--z - 4 (1)
On the other hand,

L2 =X+x + (L~+ ~ )2 - ~2 (2)
If we construct the commutators of Lz with X and X+,

we will see that
LzX -XLz= Lz (Lx+ iLu)- (Lx+iLu) Lz
= i"nLu + i (- inLx) = nX
and, similarly,
Lzx+-it+Lz= -ni+
We write these relationships in the (L 2 , Lz)-representation
and denote <lm I Lz llm') = mMmm' to get
li (lm I X llm') = Sm' [{lm I Lz llm") {lm" I X llm')
- {lm I X llm") {lm" I L~ llm'>] = (lm IX llm') (m- m') n
i.e.
<Zm IX llm' > (m- m' - 1) n= 0
Hence, {lm I X I lm') ::1= 0 only if m' = m - 1. In the
same way {lm I X+ llm') ::1= 0 only if m' = m 1. +
Next we write the diagonal (mmth) matrix element of (1)
and get
M = Sm' {lm I X llm'> {lm' I X+ llm} + (m-.!.)
2
2 ft2-.!!!:...
4
In the sum the member that can differ from zero is the one
for which m' = m - 1. If such an m'n is not an eigenvalue
of iz, the whole sum is zero and
U-
- (m-.!.)
2
2
n2 -!!!:...
4
Otherwise,
286 t>twilLEMS IN '1'liiWRE1'ICAL PHYSICS
Thus, for each M there is a certain minimum eigenvalue

of Lz equal to m 0 n, so that
Lf = ( m0 - ~ ) 2 n2 - ~2
and any other eigenvalue of Lz must differ from m 0 h by
an integral multiple of n.
Now we write the diagonal element of (2):
Ll = ~
m'
(lm I X+ jlm') (lm' I X llm) + (m+ ~ n r 2- ~2
Here (lm I X+ llm') =t= 0 only if there is an eigenvalue of
+
Lz equal to m'n = (m 1) n. Otherwise, i.e. if mn = Tnth
is the greatest value of Lz for the given value of Lf, we have
1 )2 2
Lz2 = ( m1+2 fi2
n -~;
It is evident that
1i -. / 2 fi2
mth= -2+ V Lt +T (3)
is the greatest eigenvalue of Lz, and

li -. /
t;
m0 n=2- V Lt
2
+ 4h2 (4)
is the smallest.
The difference (m1 - m0 ) n = 2 Lf ~2 - n must be V +
equal, obviously, to 2ln, where 2l is an integer (since all
eigenvalues of Lz must be separated from each other by an
interval of n). Hence,
2yM+ ~2 =(2l+1)1i and Lf=l(l+1)n2
where l can be an integer (if 2l is even) or a half-integer.

In either case, as we see from the introduction of 2l, it is
nonnegative, i.e. l = 0, 1, 2, ... or l = 1/2, 3/2,
Substituting V
Ll ~2 in (3) and (4), we get
+
m1 = mmax = l and m 0 = mmln = -l
ANSWERS M7
Since <lm' I X+ I lm) = ( <lm IX llm' ))*, we have
L~=l<lmiXIl, m-1)Jz+(m- ~rn,z_ ~2
Hence,
((l, m-11X+jlm))*=(lmJXJl, m-1)
= n {z (l + 1) + ! - ( m- ~ r
= n v(l +m) (l-m+ 1)
It follows from the definition of X and X+ that
A X+X+ A X-X+
Lx= 2 , Lu= 2i
and, consequently,
(lmJLxjl, m-1)= ~ Y(l-m+1)(l+m)=(l,m--11Lx!lm)
(lm!L 11 jl, m-1)= - 2ih Y(l-m+1)(l+m)

= ((l, m-11 L 11 Jlm))*
where m = -l, -l + 1, ... , l - 1, l.
76. To fmd the energy distribution, (E), and (t).E 2 ) we
normalize the function \jJ = Ax (a - x) and expand it in
a complete set of normalized eigenfunctions of the Hamil-
tonian. These functions are the same as in Problem 72, i.e.
'IJln(x)= V 2
-
a
. mtx
s ma- (n=1, 2, ... )
a
The condition 1 = A2 Jx 2 (a- x) 2 dx yields A2 = !~. Now
0
we determine en:
n _a
en= ) 'ljl (x)
0
'ljl~ (x) dx =A V : Jx (a-x) n:x dx
0
sin
a3 n 2a3
+-(-1)
n:rt
[ ( - 1 ) n -1] }
- -(n:rt)3
=A V-2 - [ 1 - ( - 1 ) l
a (n:rt)3
2a3 n
and we get
W (En)= I Cn 12 = (!!~6 (1- ( -1tJ 2 o#: 0
for n = 1, 3, 5, . . . . For even n, on the other hand, the

functions ljJ and 'Pn have different parity in relation to
x- 2a . For n = 1 ( E 1 = ~~)21la2 we have
W (E 1) = 240~~ 22 ~ 0.999
i.e. the particle in this state with an overwhelming proba-
bility will have an energy equal to E 1 .
We can compute the expectation value of the energy as
(E)=~ EnW (En) or as
n
{E)= f
-oo
'V*H'Vdx=A 2 1x (a-x) ( - ~; ::2 )
0
5n2
X [x(a-x)] dx=-2
f.! a
(V = oo for x < 0 or x >a, but V'P=O).
00
960n2 n2~ 2 .
We will find that (E)= ~ n 6~6 2f.la 2 g1ves the same
n=1, 3, 5
10
result (E)= ~2 Et.
To find {/1E 2 ) we must f1rst find (E 2 ). Since we cannot
consider V21jJ equal to zero, we calculate the value of (E2 )
as follows:
00 00
<E2>= J'V*H2'Vdx= J (HlP)*(H'V)dx

-oo -oo
a
= J{ ----[Ax(a-x)]
n2 d2 }2 dx= (-n2A )z a=--
30n4
2f.l dx2 ll f..1.2a4
0
or
00 00
ANSWERS 289
Whence
(t-..Ez) = (Ez)- ((E) )2 = 5zh: •
j..t.Q
77. We write the given normalized function in the form

1jJ = ____!____ [ 1 _ ..!.. ei2cp _ .i.. e- i2cp]
Van:- 2 2
It is clear from this that in measuring the angular momentum
we can find the values Lz = 0, +2n, -2n with probabilities
2
w (0) = 3' w (2) = w (- 2) = 61
This gives
(L 2 ) = 0, (L~) = ; n2 •
78. For the particle localized at point x 0 we can write the
wave function as an eigenfunction of In the x-representa- x.
tion, as we know,;; = x, and the equation reduces to x¢ =
= x 01jl, i.e. (x - x 0 ) 1jJ = 0 and 1jJ (x) :::/= 0 for x = x 0 •
Thus, 1jJ (x) = A6 (x - x 0 ) and the energy spectrum is
continuous. The normalization condition
i 6(x-x )6(x-x0)dx=6(x -x0)

00
A2 0 0
-oo
yields A = 1, i.e. 1Jlx0 (x) = 6 (x- x 0 ).

In the p-representation, as we know, 'X=in (see the :P
solution to Problem 70), and the eigenfunctions can be
obtained from the equation
·~ acp
tn -ap= X0qJ
from which
IPxo (p) = Ce-ixop/h
The constant C can be determined from the condition that
00
-oo
i (jlxo (p) (j)~ 1 (p) dp = 6 (xt- x 0 )
19-01ft96
whence
1
CZ=-
2nli
00
since 6~(x 1 - x0) = 2~ J

-oo
ei<:xt-zo>u dy.
Similarly, the momentum eigenfunctions are

'ljlp {x) = CeiPo:x/1'1 and <pp (p) = 6 (p- Po)·
79. To write a relationship simiJar to
'ljl {r) = : 1~ 1 (r)
in the p-representation, we must transform both functions
to this representation. Let
<p (p) = 1
(2rtli) 8 / 2 J
r 'ljl (r) e-ipr/11 d-r:
and
(j)f (p) = 1 8/
(2rtli) 2
r 'IJlt {r) e-ipr{lt d-r:
J
We substitute lp1 (r)/r for ljJ (r) in the first integral and
write 1/r in the form of a plane-wave expansion. We as-
sume that
~ = a (k) ei(kr) dk J
To find a (k) we apply the Laplace operator to this equality:
~ ! = Ja (k) ~ei(kr) dk = - Ja (k) k 2ei(kr) dk
On the other hand,

~ rt = - 4n6 (r) = - 4n
(2n)3 J1 ei(kr) dk
1
Hence, a (k) = 2nZk2 ·
We substitute ~~ for k, which yields dk = d:; , and get

the final result
1
- - -1-
r - 2n2Ji
J--e
pf
;..!:!E.. d
1
Pt h
ANSWERS 291
Then
1 r1 ipr
<p (p) = (2:rt1i)8f2 Jr "'• (r) e --~~- dt:
-- -
i
-
J-dpt- ( 1 J•" (r) e_ i(P-Pt)r
II d-r: )
2:rt2n Pl (2:rt1i)s,a 'Yt
nnd since the expression in the parentheses is, by defini-

tion, <p1 (p - p1), we get
·"=-1-.
'Y
r <pt(p-p;)
2:rt2h J pf dp
1
Or, after substituting p' for p-p.,
1
<i' (p) = 2:rt2n J
r IP•(p-p')2
<P'> dp'
The operator 1 r
in the p-representation is an integral
operator.
80. The wave function of a particle in a spherically sym-
metric field has the form
'ljlnzm = Rnz (r) Pzm (cos 8) eimcp
We calculate the matrix element of Dz = ez:
(nlm I Dz l n'l'm') J
= 'ljl~zmeZWn'l'm' dt:
Evidently, we can represent it in the form of a product of
integrals over r, e, and <p:
(nlm I Dz I n'l'm') = Irleicp
Since the angular part of the wave function is the same for
any spherically symmetric field, we can calculate I e and I !Jl
without specifying the potential V (r). Since
z = r cos e. dt: = r 2 sin e dr de d<p
if we assign to I~ all the multipliers depending on <p, we get
21t
I~= J e- imcpeim'cp d<p = 2n!Smm'
0
Consequently,
(nlm I Dz I n'l'm') = 0 for m' :;6 m
19*
At m = m', for Ia we have

n
Ie= JP m(cos8)cos8PI'm (cos8) sine de
1
0
Introducing X= cos e, we write

1
I e= JPlm (x) xPl'm (x) dx

-1
The Legendre polynomials satisfy the relationship

xP 1m= aPr.-t. m + bPz-t, m
where
In view of this
1
Ie = J(aPz+t.
-1
m +bPz-t, m) Pz•mdx= a81+1. I'+ b81', 1-1
and differs from zero only for l' = l + 1.

To calculate the matrix elements of D x and D y it is
convenient to introduce the notation
D± = Dx + iDy = er sine e±icp
Once we have
(nlm I D± I n'l' m') = J'IJ?~1m r sin Se±iq>'IJ?n'l'm' d1: = I;I6I~
we can calculate I~ and Ie. Since
21t
I~= ~ e-imq> e±iq>eim'cp dcp = 2nl5m•, m+1
0
the matrix element of D+ differs from zero only at m' =
= m- 1, and that of D_ at m' = m 1. We calculate Ie +
for D+ assuming that m' = m- 1:
It
I6 = JP 1m (cos 8) sine Pz·, m-1 (cos 8) sine de

0
ANSWERS 293
Substituting x =cos 8 (sin 8 = -v 1 · x2 ), we get

1
fa= JPzm(x)V1-x Pz•,m-1(x)dx
-1
2
But the theory of the Legendre polynomials yields

V 1-x2 Pz•, m-1 = atPt'+1, m + btPl'-1, m
where
_-. /(l'+m) (l'+m+1) b = _ • /(l'-m+ 1) (l'-m)
a1 - V +
4 (l' 1)2-1 • t v 4(l')2-1
and, consequently,
fa= at6t, t'+1 + bt6z, 1'-1
which is not equal to zero at l' = l + 1.
ForD_, if we assume that m' = m + 1, we get
1
fa= JPzm(x)V1-x Pz',m+i(x)dx

-1
2
Substituting V 1- x 2 P 1m = a2P 1+t, mH + b2Pz-1. m+t• where

_ , /(l+m+1) (l+m+2) b = _-. /(l-m) (l-m-1)
a 2 -- V 4(l+1)2-1 ' 2 v 4l2-1
we get
fa= a261+1. I'+ b2{)1-1, I'
The expressions for the matrix elements of Dx and Dy
can be obtained in an obvious way:
1
(nlm I Dx I n'l'm'} = 2 [(nlm ID+ I n'l'm'} + (nlm I D_l n'l'm'}]
1
(nlm I Dy I n'l'm'} = 2i ((nlm I D+ I n'l'm'}- (nlm I D-1 n'l'm')]
It is evident that both differ from zero for l' = l ± 1 and
m' = m + 1.
82. We use the general definition
di. ai. i A A A A
crt=m+r;(HL-LH)
~ p2 ~ ar ~~ ~ ~
Since H = 21-t + V (r), ae= 0, and xy- yx= 0 for any two
components of r, we find that = o and rv- vr
ir , ~~ ~~
Tt= 2p.li (Jh- rp2)
We write r=ix+jy+kZ and ~=.P~+P~+.P: and we use
the basic commutation relations of quantum mechanics,
PxY-YP:r= iMx11 , etc. Thus we get
dr , ~ ~ ~ ~ ~
dt = p.!i [i (p~x-x p:) + j(p~y- y p~) + k (p~z- zp;)]
~ ~ ~ ~ ~ ~ ~
2
But p~x-xp~ = p~x- PxXPx+

PxXPx-xp~. Now if we group
the members in pairs, we can factor out the common mul-
tiplier Px from the first pair to the left and from the
second pair to the right. Then p~; xp~ = -2i1ipx and, -
hence,
dr t ~ ~ ~ p
dt= 2p.!i ( -2i1i){ipx+ jpy + kp 2 }= "it
The operator p does not explicitly depend on time either,

an,d
dp i ~~ ~ ~
Tt=lf [Hp-pH]
Since PxP PyPx = 0,
11 - the expression reduces to
all
dt
= 2...
li
(Vp- pV)
We now compute it in the x-representation. Then = -iliV p
and V = V (x, y, z). Computing
(Vp- p V) \jl (r)=- in [V grad \jl- grad (V¢)1 = ili(grad V) \jl
we get
aP
dt
= -grad V = F
Since the commutation relations do not change in trans-

forming to other representations, this relationship always
holds.
ANSWERS 295
~ ~. iii
83. Since the operators rand r = - - V do not commute
m
(see the solution to Problem 82), the operator j must be
represented in a symmetrized form, i.e.
J=- ~~ [6(r-ro)V-V6(r-ro)l
J
Following the general rule, we compose (j )= 'ljl* (r) Jljl(r)dT.
Then we integrate the second term by parts and, using the
property of the delta function, we get the final expression
for (j ):
m ( '\j) • gra d '\j)-'ljlgra d 'lj) *) r=ro·
(].) = - ilie
2
84. According to the usual rules of commutation,

dLz
--=x--y-
~ av ~ av
dt ay ax
and iz is an integral of motion in a field with a~ axis of
symmetry OZ. In a central force field, for V = V (r), we
have
85. In composing the operators Kx, ku, and kz we must

symmetrize the products of type v/Jz
since the operators vy
and Lz do not commute. Hence,
By analogy we determine k y and K z·

Since the field is spherically symmetric, H will commute
with Lx. Ly, Lz. We must keep this in mind when we com-
pose commutators. To prove that d!x~ = 0 without invol-
•
ved computations we must calculate thAe commutators of [f
with each of the operators .,;Y, v~, and ~ separately. These
are
~ ~ 1 ~ ~ ilia y
[H, Vy]=-[H,
· m Pul= - -
m -
r3
~ ~ ilia z
[H ' v
z
]=-
m - r3
-
[ fl ' r; ]=!!!..[_:__!:_(
m.r3 1i
Px_ x(r·p)
r r3
)]
If we substitute these expressions into

ak. i A ~
at =n[H, Kxl
we see that ak:x
at = 0 .
86. If we use the results of Problems 31 and 32, we can
easily find that
cia
dt = - zwa an
. d a;i* = zwa
---at . A* .
87. We use the commutation relations for coordinates

and momenta, PxY- Px = - iMxu• and first write y ~~ .
For simplicity, let us denote Px= P:x-eAx. Since com- x
mutes with Py and fiz (and, all the more so, with e<p),
we get
a?c i A~ A A i ~2 A A )2
Tt=r;(Hx-xH)= 2lif.t (Pxx-xl x)
~A~~~ i ~~~A p
[Px(PxX-.TPx) + (Pxx-xPx) Pxl = --,l-
A
= 2 /ij.l
ap a ~
Next we find at =dt (Px-eAx). Since A(r, t) can depend
aP aA
ont,wehave a:=-e 8tand
aP aA i A , A A A ~ ~ A
t 2 /i [P~Px-PxP~+P:Px-PxP;]
_ax =-e ax+-
t 11
ie
+n A
(<pP x - P x<p)
A
Obviously,
ANSWERS 297
Denote
D= P~Px-fixfi~= fiu (fiJjx-Pxfiy) + CPufix-fixfiu) Pu

First we find
.
Smce ~ A
Pu ~
x- A xPu ·n aAx · h ave
= - t ay, v.e
~ ~ ~ ~ ( aAy iJA )
P yP x- P xP y = - ilie ---g;;-- iJyx = - ilie (curl A)z
and
D= -ilie(PyBz+BzPy), where B=curlA
Thus,
dP e ~ ~ ~ ~
a/= -e aAat-e ax+2il
acp
[PyBz+BzPy-PzBu-BuPz]
and since - a:-

JA iJ
a~ =Ex, we get the Lorentz force equa-
tion in the operator form:
d:
dP
=
e-;
-eEx+ y(YBz+Bzy-zBy--Byz)
-;-; -;
where E is the electric field vector, and B the vector of

magnetic induction.
88. The operator ::i:1 does not depend explicitly on time
and commutes with d.;, &~o, A, and QJ. If we use the commuta-
tion relations of coordinates and momenta:
PiXk- XkPt = - i1i.6ik
we get
i ~ ~~ ~~ ~
= T C<X.t (PtXt- XtPt) = C<Xt

~ ~ dP
We denote P t = P t - eAt and calculate a/ :
d ~ i.lAt i ~ ~ ~ ~ ~
dt (p,- eAt)= -e Tt+""'h [ca 2 (P 2 Pt -PtP 2 )
~ ~~ ~~ i ~ ~
+ caa (P3Pt- PtP3)] + T e (cpPt -Ptcp)

The operators P1 commute with cii. Besides (see the solu-
tion to Problem 87),
and
Thus,
Since cci 1 = Xtt we again get the Lorentz force equation in

the operator form.
90. For 'a system of N particles the operator of the
N
total momentum is P= ~
i=t
Pi and the Hamiltonian is
N
where ~ Ui (r 1) is the total potential energy in an exter-
i=t
nal field, and ~~ U1k (r 1k) is the energy of interaction
i>k
between the particles of the system.
Applying the commutation relations, we find that
~ N
dP ~
dt=- LJ gradkUk.
k=1
91. Since V (x) has a discontinuity at x = 0, we must
denote the region x < 0 as region I (V = 0) and the region
x > 0 as region II (V = V 0 ). After solving the Schrodin-
ger equation in these regions, we must fit the solutions,
ANSWERS 299
i.e. at point x = 0 equate the functions and equate their

first derivatives.
In region I the Schrodinger equation becomes
d2'1Jli ka,r,
dx2
+- 0
l't'l-
a 2mE
where k 1 = """ii2"", and its solution is
'llr = Cteik!x + Cze-ik!x (1)
In region II the Schrodinger equation is
d2'1Jln
dx2 +k2•" -0
2't'II-
wh ere k 22 = 2m(E-Vo)
/i.Z ;
ka > Ofor E
2 > V0, and k: < 0 for
E < V 0 • Its solution is
'Jln. = Caeik2X + c,.e-ikax (2)
We fit the solutions and their derivatives:
'Jlr (0) = 'Jln (0), Ct +Cz=Ca+C4
d!I lx=O = dd;rl=o ' ikt (Ct- Cz) = ik2 (C 3 - C4)
Four constants must satisfy two equations. Since one of the
constants can be chosen arbitrarily and in region II for
physical reasons we can expect to find none but a particle
moving in the positive direction, i.e. Px = 1ik 2 > 0, we
must put C,. = 0. If E < V0 , then k 2 = ia and e-ikax =
= e+ax-+ oo as x-+ oo. For ¢ 11 to be finite at every point,
C,. must be zero. The equations give us
C2 = kt-k 2 and EJ_ = 2kt
Ct kt+k2 Ct kt+k2
Now we must determine the reflectance and transmittance
as ratios of corresponding current densities:
R= I/~etl I and
]~
D= Iit~ans I
]~
In expression (1) the first member represents the particle

moving in the positive direction of the x-axis (the particle
is falling on the step), and the second represents the reflec-
ted one. With the aid of these functions we compose the

corresponding current densities:
hn = flkt I Ct 12
m
and iretl = - hkt I Czl 2
m
and we get
R=l Z:+Z: 2
1 ' D=l (k:~~22) 2 1
For E < V 0 we find that k 2 = ia and that the function
\jlu = C 3e-ax is a real function and diminishes as x moves
away from zero. Evidently, in this case itrans = 0 and
D = 0. Accordingly,
B = I kt+za
k 1 - ~a 12 = 1.
92. As in Problem 91, R =I i~etl I

/In:
and D= Iit~ans I·
/In
We find the functions that characterize the incoming, ref-
lected, and transmitted particles (waves).
We denote as region I the region x < 0 (V = 0) and
introduce k~ = 2'/:zE As region II we denote the region
0~ x ~= V0 ) and k~ = Zm(Etz-; Vo). As region III
a (V
we denote x >a (V = 0). Next we write the Schrodinger
equation and its solution in each region
~~I+ k~'iJI = 0, 'IJli = Cteik1x + Cze-ikix
~~~I + k:'ljln = 0,
d21Jlrrr
~
+ ks,., 0
i'YIII = •
We assume that C6 = 0 since in region III there is none but

the transmitted wave (see the solution to Problem 91).
Requiring that \jl and ~; be continuous at x= 0 and x= a,
we get four equations for the C's:
'IJlr (0) = 'IJlu (0), Ct + Cz = C3 + C4;
ANSWEll.S 301
'Pn (a)= '¢In (a), Caeih2a+ C4e-ih2a= Cseih!a;
d¢nl
dx
- d¢ni
x·=a- dx
Ix=a'
ik2 (Caeik2a- C~e-ik2a) = iktC seik!a
Now we can determine the ratios of different Ci to C1

iii ( d¢*
(i = 2, 3, 4, 5). Bearing in mind that I. =2m- •IJ--
'I dx
- '¢* : ) , we find that
. likt
lin=--;;;-
l c1 12, lren=
. likt I c2 12 ,
-m I. trans= ---;;;-
likt Ic5 12
and R=lC 2/C 1 j2 , D=jC5/Cd 2 •
R =I ~~ 12' D =I ~: 12
From the four equations we determine c2;cl and C5/CI:
where
1 -1 -1 ()
k2 k2
-1 -k; -;;; 0
~~= e-ih2a _ eik1a
0 eik2a
0 eik2a _ e-ik2a kt t t"ha

--e
k2
-4 ik 1a [ k i(kf+k~)sink 2a].
- e cos za- 2ktk2 '
-1 -1 -1 0
k2 k2
-1 -kt -;;;_ 0
~2= eik2a e-ih2a _ eikta
0
0 eik2a - e-ik2a - -kte t'k 1a
k2
2eik1a
= Tk
Ct 2
i (k;- k~) sin k 2a;
1 -1 -1 -1
kz k2
-1 -kl -1
L\s= kl =4
0 eiksa e-iksa 0
0 eik2a -e-iksa 0
Now we substitute the calculated values of L\1 , L\ 2 , L\ 6 into

the expression for R and D, and, considering E > V0
(k 2 is real) and simplifying, we get
R= (k~-kf)2sin2k 2 a = (k~-kf)2sin2 k 2 a
,j2ktk2 cos k 2 a- (kf+ k~) i sin kza 12 4kfk~ +(k:- kf)2 sin2 k 2 a
It is easy to see that D + R = 1 and that for k 2a = nn

the barrier is transparent, i.e. D = 1 and R = 0. The solu-
tion holds for V 0 > 0 and for V 0 < 0 (when the particle
passes over a potential well).
If we consider the case of V 0 > 0 and E < V 0 , we get an
imaginary k 2 = i~. Then sin k 2a = i sinh ~a, and we find
the expressions for R and D:
(kf + ~2)2 sinh2 ~a

R = 4kf~2 + (kf + ~2) 2 sinh2 ~a
t4kf~2
D= 4kf~2 + (kf + ~2)2 sinh2 ~a
For ~a ~ 1 we get
D ,...., 16kf~2 -2~

_,...., <kf+~l)2 e •
93. Let region I be the region 0 :;:;;:; x :;:;;:; a, region II be

a :;:;;:; x :;:;;:; b, and region III be x ~ b.
We note that V = oo for x < 0. Hence, '¢ = 0. If we
choose '¢I so that continuity is ensured at x = 0, we can
write the solution of the Schrodinger equation in the re-
ANSWERS 303
gions I, II, III:

d::1 +kz,p 1 =0, where k 2 = 2: : and 'li'I=Asinkx;
d2'1Jln
dxZ - x z,.,,'I'll= 0 , wh ere x 2 = h2
2m (V0 - E) an d
'Pn = Btex(x-a) + B2e-)!(x-a>;

d7x~11 + k2'lJIII = 0, 'Pn = eih(x-b) + Ce-ih(x-b)
The solution is normalized in such a way that the ampli-
tude of the wave leaving the potential well is unity. This
does not restrict its general character since all the equa-
tions are homogeneous and determine none but the ratios
of the coefficients.
We write the continuity conditions on the boundaries
x = a and x = b for the function and its first derivative:
A sinka =B 1 +B 2 (1)
kA cos ka= x (B 1 -B2) (2)
B 1exl + B 2e-xl = 1 + C (3)
x (B 1ex1-B 2e-x 1) = ik:(1-C) (4)
Here we have introduced l = b - a as the width of the bar-
rier. If we add and subtract (1) and (2), we find B 1 and B 2 •
Then if we substitute them into conditions (3) and (4) and
repeat the operations of adding and subtracting, we get
equations that define the amplitude A of the wave in the
inner region and the amplitude C of the wave falling on the
barrier:
: { sin ka [cosh xl + ~ sinh xZ J

+ ~ cos ka [sinh xl + i~ cosh xl J} = 1
: { sin ka [cosh xl- t~ sinh xl J
+ ~ cos ka [sinh xl- ~ cosh xl J} = C

Hence,
C =sin ka [cosh xl- T,; sinh xl J+~cos ka [sinh xl- if coshxZ J

sin ka [cosh xl + ~ J
sinh xl + ~ cos ka [sinh xl + ~ cosh xl J
We note that the numerator is complex conjugate to the
denominator. Hence, I C 12 = 1. The incident (falling)
wave is completely reflected at x = 0. We write the expres-
sion for A in the form
: exl { ( 1 + i~ ) (sin ka +: cos_ka)
+(1 - ~ ) {sin ka- ~ cos ka) e- 2x! } = 1

and we see that when x is real (E < V 0 ), the second member
is always much less than the first (e- 2 xz ~ 1). So if we
neglect it, we find that I A 12 ex: e-2 x1, i.e. I A 12 ~ 1 (the
amplitude of the wave is much less in the inner region than
in the outer). This holds for all values of energy except
when
•
sm 7
i£0 a 0 + -Xo
ko cos k a = 0 or tan k a = - ko
0 - Xo
(If we compare this with Problem 45, we find that this con-
dition gives the energy levels in a potential well of limited
depth.) Then the member with e- 2 xz plays a decisive role
and A grows substantially:
A= 2i V~ exol and C = - xo+iko

xo- £k0 x0 - iko
i.e.
I A j2= 4e2xo! and IC 12 = 1
Thus, near the energy eigenvalues for a particle in a
potential well of finite depth the amplitude of the wave
in the inner region changes abruptly from e- 2 xoz to e+ 2 xo 1•
94. To characterize the forces that prevent the electrons
from leaving the metal we place the origin of coordinates
on the interface between the metal and vacuum and assume
that the potential energy of the electrons in the metal is
lower than their energy in the vacuum by V 0 , i.e. we assume
ANSWERS 305
V = 0 for x < 0 (in the metal) and V = V 0 for x > 0 (on

the interface).
Let the x-axis be normal to the surface of the metal.
If we apply the external electric field E in the positive
direction of this axis, then for x > 0 the potential energy
will be V (x) = V 0 - e I E I x, and the probability of the
electron passing through the barrier will be determined
by the transmittance
D=D0 exp[ -2 V;m r]fV(x)-Exdx]

Xi
Now we need only compute the integral in the exponent.

The problem becomes one-dimensional. The only impor-
tant factor is the movement along the x-axis, and Ex =
2
='f~ denotes the energy connected with this movement.
Points x1 and x 2 are determined from the condition V (x1 ) =
= V (x 2) = Ex· For our problem x1 = 0 and x 2 is deter-
mined by the equation
Vo- e I E I x 2 =Ex
Then
X2
Jr 11
y V 0 - Ex- e I E I x dx = 2
- ae IE I (v0 - Ex- e I E I x) 3/2!X2
0
0
2
= 3eiEI ( o-Ex)
v 3/2
and
4 V2m V 3/2]
D=D0 exp [ anelEI ( o-Ex)
i.e. grows as I E I and Ex grow.
If we denote dn as the number of electrons inside the
metal (per unit volume) that possess momenta in the inter-
val (p, p +dp), the density of the electric current leaving
the metal in the positive direction of the x-axis is
j=e JVxDdn, where Vx= ~
The integration is done over all values of Pu and Pz and
over Px > 0. If we assume that the electron gas is extre-
20-01496
mely degenerate (i.e. behaves as it would when T = 0 K),

we get
d = 2 'dp:x;dpy dpz
n h3
(the mean occupation of a state is one) for ~ ~. where ~ I:

is the maximum energy, i.e. the level of the chemical
potential, and
dn=O for L>~
2m
Hence, if we pass to cylindrical coordinates in momentum
space and assuine that Py = p cos q> and Pz = p sin q>,
we can write
V2mt lf 21-p~' 2n
i= !: I dp:x:
0
Pdp V:x:D (vx) dq>
0
I
0
Now we substitute '11 = ~ - E:x:, d'l') = -vxdPx• integrate

over q> and p, and get
t
. 4nem 1
J=h3 J 'I']D('I'J)d'I'J
0
D('I'])=Doexp [ - 3
4 lf2m
he lEI (Vo-~o+TJ)
,. 3/2]
Since D ('I'J) diminishes very quickly as '11 increases, the
members with small '11 will play the dominant role. So if
we expand the exponent in a power series in '11 and denote
2lf2m T)
lie IE I (Vo - ~) 312 = q and Vo _ ~ = ~. we can extend the
limits of integration over ~ to infinity and get
2 00 2
'=D 4nem -3q(V -")2
1 o h3 e o ~o JI e -q6!:At=D~(Vo-~)2
~o ""'o h3 q2 e
--aq
0
-AIEI2
-· exp [ 4V2m
-31Eihe(Vo-~o) ,.3!2] '
95. The alpha-particle in the nucleus lies in a deep poten-
tial well. We may assume appr?ximately that V- -V0
ANSWERS 307
for T ~ T 0 ; To characterizes the range of nuclear forces.

For T ~ T 0 (outside the nucleus), V = 22e2 • The transmit-
'
tance of the particle through a barrier limited by a straight
22 2
line at T = To and by a hyperbola V = re for T >To is
determined as
D=D0 exp ( -
2 ,r-
v 2m
li Jr2v -
ro
2Ze2
7 --EdT
) (1)
where E is the energy of the particle falling on the barrier,

and To and T2 are the points of retrogression, at which V = E,
i.e. T 2 = 2Ze2 /E.
To calculate the integral in (1) we introduce cos2 u = .!_ •
r2
Obviously, at T = r 2 , we have u = 0. Now if we denote
cos2 u 0 = ro , we can write
r2
uo
I= JV
0
cos12 u -1 X 2 sin u cos u du YE z;
2 2
= 2Ze2 [ sin 2uo]

V:E Uo--2-
0n the assumption that-./ 70 ~1 we expand u0 in a series
u 0 =arccos
V
V --
r2
ro
72
~--
n
2
v-r
-0
'2
Then
Hence
I~ 2Z~ [~- 2 -. /
VE 2 v rJ 0
72
and
D ~ Do exp [ - ! (n:~ +V2mZe J
2 2T 0 )
20*
where~voo = V2Eim is the velocity of bthe emitted alpha

particle measured far from the nucleus (where V = 0), and
A=nD 0 exp[- ~ (n:~ 2 +V2mZe2ro)J.
96. If we choose the z-axis in the direction of the vector
of magnetic induction and write the components of the
vector potential in the form Ax= -By, Ay = Az = 0,
we can reduce the equation
~ 2
(p-;:A) 'P = E'IJ
to the form
-~f1•h-!:!:__eBy
2m 't' m
B'ljl +~B2y2 •h=E•h
ax 2m 't' 't'
(1)
Since the coefficients do not depend on x or z, we can look
for 'ljl in the form
'P = eirxxeiflzf (y)
After substituting 'IJ into (1), we get the equation for f:
n2 d2f mw~ 2
-2m dyf+T ytf=ef
where we have denoted ro0 = eB , e = E 0 - 1i22~ 2 , - y 1 = y +
m m ·
na
+eLf·
We have come to the equation for the one-dimensional
harmonic oscillator. If we introduce the dimensionless
coordinate ~ = y1 V m:o ,we can, in accordance with the
solution of Problem 46, write
fn (Yt) = Ce-~ 21 2 Hn (~)
( H n is the Hermite polynomial) and
Bn = ( n + ~ ) firoo n = 0, 1, 2, ...
1
Thus, for an electron in a uniform magnetic field, the comp-

lete wave function and the energy spectrum are
'¢naJ3 = Cnei(rxx+flz>e-~212H n (~)
eli fi2~2
EnJ3= 2m B(2n+1)+~
ANSWERS 309
V-
The normalization condition yields
c _.!. mw 0 1
n- 1i lin V 2nn!
The energy spectrum is continuous for motton along the
z-axis (in the direction of the magnetic field) and discrete
for motion in the plane of vector B.
97. If we compare the time-dependent equation for a
particle in an electromagnetic field, which is characterized
by a vector potential A and a scalar potential qJ,
2~ (- inV- eA) 2'1' = (in ! - eiP) 'I'

with the equation containing the changed potentials A' =
= A + grad f and qJ' = IP - ~: and, hence, the changed
function 'I'',
2~ (-inV-eA-egradf) 2 '1''= (in! -eqJ+e :~) 'f!

we see that the operator V on . 'I' is equivalent to
( V- ~ grad f) on 'I'', and the operator on 'I' is equiva-!
lent to ( ! - :: ) on
~ '¥'. This shows that 'I' differs
from 'I'' by a factor e-ieffli:
ie/
'I' = 'I'' e- T
Indeed, if we differentiate 'I' with respect to, say, t, we get
ief
, ief - h -ief
a'l' = a'l' - "+ '¥' ae = -li- (_!__!!__!L)
'I''
at at e at e at 1i at
For this kind of transformation such expressions as
I 'I' 12 and ('A.) = J
'I'* L'l' d-e= do not change.
98. We write the time-dependent equation and remove
the parentheses (p - eA) 2 , and we get
ft2 /);. qr ine · ine . e2
-2m + m (A-gra~ 'I')+ 2m qr dtv A+ 2m A2'1'
. +e~JJ'I' =in ~;' (1)
The complex conjugate equation is

ft2 !':!. '¥* -
-2m m
ine ine 'I'* d.lV A
(A . grad '¥*) - 2m
+ 2m
e2 A2'Jf* + ecpr
nr•-
- "1i ---at
-L
81¥* (2)
From Eq. (1) multiplied by '¥* we subtract Eq. (2)

multiplied by 'I' and get the expression for I 'I' 12 • :t
Then we divide the difference by in and get
~ I '¥ 12 = ;~ ('¥* !':!. '¥- '¥!':!. '¥*) + ~ ['¥* (A· grad '¥) +
+(grad 'Jf* ·A)'¥+ '¥*'¥divA] (3)
Since
'¥* !':!.'¥ - '¥ !':!.'¥* = div ('¥* grad '¥ - '¥ grad'¥*)
and
'¥* (A·grad '¥)+(grad'¥* ·A)'¥+'¥*'¥ div A=div{A'l'*'l")
we can represent (3) as
al1f j2 +
-at div J" -- 0
where
J• = -2ill ('I' grad UJ
1 *- ur
r • grad ur e A I '¥ I2.
r)--
m m
100. In the new representation we multiply the equation
on the left by §+ and get
(1)
where
Obviously,
(2)
and
ANSWERS 311
Here we have carried out the expansion for A (t, rh) indi-
cated in the problem. After substituting (3) into (2) and
neglecting members of order r~ and (r11 • p11 ), we get in place
of (1) the following equation for \jl:;
~2
in :i = ~ 2~h
k
<p + (V- Ed) \jJ
~ BA (t, 0)
where d= 2.J ehr11 , E=- at (divA=O).
k
101. It follows from Oza =a and Oz~ = -~ that the
operator az has two eigenvalues equal to +1 and -1 and,
hence, corresponds to a component of the spin vector (in
a
units of '/i/2). The eigenfunctions of Z! i.e. a and ~. will
not, obviously, be the eigenfunctions of ax and But if au·
we add and subtract the equations
axa=~. ax~=a, (1)
aya=i~, cry~= -ia (2)
we get
ax (a+~)= (a+~), Oy (a+ i~) =a +i~
~~-~=-~-~. ~~-~=-~-~
These equations show that ax and Cry have the same eigen-
values as a%, equal to + 1. u.:we form the expression
we see that cr xO z -a
zGX = - 2ia
y. Aside from this, by a p-
plying operator cry to (1), we get
~ ~ (a) ~ (~) ( - ia) (. a )
GyGx ~ =Gy a = i~ =-iaz ~.

auax = - ia:
The notation axau (;) = iax (;) means that
axaya= iaza and axay~= iaz~

And since this holds for all the eigenfunctions (there are
only two), it holds for any function. Hence, we can simply
write the equation for the operators.
~ ~
102. Since ax commutes with p and r, if we compose
~ dad/'
following the general rule, we have
da i ~~ ~ ~ e ~ ~
dt = T (Hax-axH) = n:;(ayBz-azBu)·
103. We write the usual eigenvalue equation for the
operator ax: OxX ="-X· Next we represent the sought func-
tion in the form of a matrix X = ( : ) . If we act on this
~
with the matrix ax = (01 01) , .
we arnve at the equation
i.e.
( :) = A ( :) or b = Aa and a= M
and, hence, A2 = 1 or A = +1. Now we substitute the
eigenvalue A1 = 1 into the equation b = A.a, and we get
the eigenfunction x1 =a (!}
that corresponds to this eigen-
value. If we do the same for A2 = -1,rwe get x2 = a(_~)
It follows from the normalization condition
X~Xt=lal2 (1 1) (~) =2lal 2 =1

that a= 2-112 •
Similarly, for au=(~ -~) from the equation
(~ -~)(:)=A(:)
ANSWERS 313
we get
-ib='Aa, ia='Ab
Whence,
1-=+1, b=+ia; x+t=J 2 (;), x_t=J 2 (_~).

104. We express the spin operator S in terms of o:
n. ~
s = 20'
A
where ax, Oy, Oz are the Pauli matrices, which satisfy the
conditions a~= a~= a~= 1 and axay = - a11 ax (etc. cyclic).
Now we project S on a unit vector in the direction of a,
(S ·a) , and, squaring this expression and using the properties
a
of ax, a11 , az, we get
(S·a)2 n.2 ~ ~ ~ ~ ~
(crxax + Oya 11 + Ozaz) (crxax + cr11 ay + Ozaz)
A
~ = 4a 2
li 2 ~2 ~2 2 ~2 2 ~ ~ ~
=
2
4a 2 [Ox ax+ 0 11 a11 + Ozaz + (crxcry + OyOx) A
h2
X axau+ .. . ] = 4·
105. We remove the parentheses in the left-hand side
of the relationship and then make use of the fact that
a~=a~=a~=1 and axa11 = -aiix= -iaz (etc. cyclic).
We find that axAxaxfi:x will reduce to AxBx and
oxAxaufly+ UyAyaxBx = iaz (AJJy-AyBx) = iaz [A X B]z
Q.E.D.
106. If we use the commutation relations
axa 11 - ayax ==[ax, cry]=- ray, crxJ = 2iaz
we can easily show that
[a±, crx1 = + az, [a±, a11 J = iaz, [a±, azJ =+a±.

107. We prove (a), (b), and (c) by expandingsin (crx<p)
and cos (oz<p) in series. The series contain for sin (ox<p)
none but odd powers and for cos (OziJl) none but even powers.
At the same time,
<J~n+l = ax and <J:n = 1
To prove (d) we take what we have proved in (c) and write
aue-icrzcp =au (cos ljl- iaz sin 'P) = (cos 'P + iaz·sin 'P) au
= eicrzcray.
~~ 1~ ~ ~ ~ 1 ~
108. We write a+a-= 4 (ax+ iay} (ax- iay) = 2 (1 +az).
If we square this expression, we find that
~~ 1 ~ 1 ~ ~~
(a+a-) 2 = 4 (1+az) 2 = 2 (1+az)=a+a-
For an arbitrary n the relationship can be proved by mathe-
matical induction.
109. By definition, A' = SAS+, where s+ = iazcp/2,
Since ~z anticommut.es with Ox and ay, if we use the results
of Problem 107 and the properties of the Pauli matrices
(axay = iaz, etc.), we can write
A'= e-iazcp.A =(cos ljl- iaz sin 'P) (ax sin 'P +au cos 'P) =au
In the same way
B' = e-iGzcpjj =Ox.
110. Let a and ~ be functions that are acted upon by
a
operators Ox, y• and Oz in the way indicated in Problem 101.
We denote function a for the neutron (n) by an =a (Sn),
and for the proton (p) by a" =a (Sp). Clearly, for a system
of weakly interacting particles we must look for the func-
tion of the two-particle system in the form of a product of
single-particle functions. Notably, we will look for the
eigenfunction of the operator S z = Onz +
Opz in the form
X (Sn, Sp) = Aanap + Ban~p + C~nap + D~n~p
If we act on it with S z• we get
SzX = +
(Onz +apz) X= Aanap Ban~p- C~nap- ~n~P
+ +
Aanap- Ban~P Cfinap- ~n~p = 2Aanap- 2Df3n~P
ANSWERS 315
Each of the four terms in x therefore is an eigenfunction

of Sz., which correspond, respectively, to eigenvalues 2, 0,
0, -2. In units of 1i this corresponds to 1, 0, 0, -1. (The
two functions an~P and ~nan are degenerate.) Now we
construct S2 and determine its action on X· Obviously,
S2 = (apx + anx) 2 + ... = 6 + 2 (an ·ap). since cr~x= ... = 1
and
S2x = 6x + 2 (a~. Op) (Aanap + Ban~p + c~nap + D~n~p)
We find
A (anxapx+ Or.yapy + anz.Crpz) anap
=A (~n~p- ~nBp + anap) = Aanap
In the same way,
B (anxapx+ anyapy + anza pz) an~p
=B [~nap+ i~n (- iap)- an~p] =B (2~nap- an~p)
C (anxapx +anyapy+ anzapz) ~nap

= C (an~P+ an~p- ~nap)= C (2an~p-~nap)
D (anxapx+ anyapy+ anzapz) ~n~p
= D (anap- anap + ~n~p) = D~n~p
Consequently,
S2 x = 8Aanap + 8D~n~P + 4 (B + C) (an~p + ~nap)
It follows from this that ana, and ~n~p are eigenfunctions
of S2 that correspond to the eigenvalue 8, i.e.
( ~ S)2 anap= 1 (1 + 1) n2ctnap
and
( ~ S) 2
~n~p = 1 X 2X n 2~n~p
Next we write the equation
S2 (Bctn~p + C~nctp) = 4 (B + C) (an~P + ~nap)
where we must look for B and C such that Ban~p + C~nap

be an eigenfunction of S2 • For this we write
4 (B + C) (an~P + ~nap) = 4 A(Ban~p + C~nap)
Then
B + C = AB, B +C= AC
This system of two homogeneous linear equations has a non-
trivial solution if the system determinant
1-A. 1 I
1 1 1 - A = (1- A. )2- 1 = 0
Hence, there are two solutions:
(1) A = 2, B = C. The eigenfunction an~P + ~nap cor-
responds to an eigenvalue of S2 equal to 8 and an eigen-
value of S z equal to 0;
(2) A = 0, B = -C. The eigenfunction proves to be
antisymmetric, i.e. equal to an~.n - ~nap· It corresponds
to the zero eigenvalues of S2 and S z·
Hence, the eigenvalue of ( ~ § r
equal to 2n2 has corres-
ponding to it three symmetric functions anap, ~nap +
+1i ~an~P• ~n~P• which describe states with eigenvalues of
2 S z equal to n, 0, -n. These states form a triplet. The
eigenvalues of S2 and S z equal to zero have corresponding
to them one antisymmetric function an~P - ~nap, which
is a singlet state.
~ li ~ ~ li ~ ~2 ~ 2
111. Let S 1 = 2 a 1 and S 2 = 2 a 2 (where ai = 3, aix =
= ... = 1). Let us consider the square of the sum of these
operators:
S2= §~+ s; +2 (St·S2)
In both the triplet and the singlet state (see the solution to
Problem 110) S\ S~, and S~ have definite eigenvalues: for
S2 this is 1i2s (s + 1) [s = 1 for the triplet state and s = 0
for the singlet state], and
s2 _ s2 -1i2..!. ~- ~ fi2
1- .- 2 2- 4
ANSWERS 317
Hence, we can write the values of the scalar product of

the spin vector operators of two particles:
~ ~
(S 1 • S 2) whose eigenvalues are
{ :z in the singlet state
3
- 4 n2 in the triplet state
In these states, respectively, (o1 ·0 2) = 1 and (a1 ·02) = -3.
112. In:accordance with the solution of Problem 111 we
denote the eigenfunctions of the triplet and singlet states
as Xt and Xs and write
(at·Uz)Xt=1xt; (at·Uz) Xs= -3xs
Since the three triplet and one singlet spin functions form
a complete set for the system of two particles, we must
first investigate the action of 1 ·0 2)k on Xt and Xs· Obvi- (o
ously,
therefore,
In the same way,

(at· az) 2 Xs = (a u
1· 1) ( - 3Xs) = ( - 3) 2 Xs
and
If we assume that.
~ ~ k ~ ~
(af.az) =A+B(at·a2 )
we can choose A and B in such a way that the relation holds
when acting on Xt and Xs and, hence, on any spin function
of the two particles. Since
~ ~ k ~ ~ k
(at•az) Xt=(A+B)Xt and (a1 -a 2) Xt=1Xt
we have
On the other hand,

............. k ... ,..
(af ·az) X,= [A+ B (a1 ·a2 )] X8 =(A- 3B) X8
and
i.e.
A- 3B = ( - 3)"
A- 3+(-3)" B- 1-(-3)"
- 4 ' - 4
Hence,
A A ll
(O't·O'z) =
3+(-3)1<
4
+ 1-(-3)1<
4
A
(O't•O'z)
A
For example, for k = 2

(ar az) 2 = 3-2 (ot · uz).
113. We use the matrix representation to solve this pro-
blem. Let us expand the eigenfunction in a complete set
of states with S = 1:
+1
2J am'i'tm
'I'= m=-1 (1)
The cefficients am satisfy the equations

+1
2J
m'=-1
H 1m, lm'am' = eam (2)
We reorganize the Hamiltonian to produce

ii =!(A-B) (S!+S:)+ (C- AtB) s: +A+B
where §± = Sx ± iSy, and the eigenvalue of S2 is assumed
to be 2.
To calculate the matrix elements of S! and we consi- s:
der the system with spin 1 as a pair of particles whose spins
are equal to 112. Then, as in the solution to Problem 110,
we denote
1
'i'u = GttGtz, 'I' to= VZ (a1~2 + Gtz~t), '1'1-1 = ~1~2
(We use 1 and 2 instead of the symbols n and p.)

ANSWERS 319
Clearly, the spin vector operator of the pair of particles

in units of n will be
1
Sx = 2 ((Jxt + (Jxz) (x =X, y, z)
A A A
Using the rules indicated in Problem 101, we see that

A 1 A t"V2
:Sx'i'tt = VZ 'i'to• Su'i'u = - 2- 'i'to
"V2 t"V2
('i't -t + lJlu), S u'i'to = -
A A
Sx'i'to = - 2- 2- Nt - 1 - 'i'u)
A 1 A -i
Sx'i't -1 = Vi 'i'to• Su'i't -t = Vi 'i'to
Now it is easy to find the result of S1 acting on all three
functions. We see that
S+'i't -1 = V2 'i'to and S!lJlt -t = 2¢u
S-'i'u = V2 'i'to and S~lJlu = 2¢t -t
The action of Sl on the other functions yields zero.
If we substitute the values (S~)mm' into Eqs. (2) and
recall that s:
has none but diagonal elements, we arrive at
the following equations:
+ At B -e) a + A 2 B
(C
A 2 B a,+ ( C
1
(A+B-e)a,~o
+ At B -e) a_t =
a_t = 0 )
0
IJ (3)
We equate the system determinant with zero and get the

energy levels and [using Eqs.' (2) and (3)1 the corres-
ponding spin functions: · ·
1
Bt=C+B, 'i't = Vi ('i't - 1 - 'i'u)
e2 =A+B, lJlz = 'i'to
1
e3 =A+C, 'i'a = y:2 (lJlu + 'i't -t)·
114. Before the electron enters the auxiliary field B',

it is described by the wave function 'I' = ei(ky- wt) (o) a
that satisfies the equation
a'l' n2 ~ ( eli )
(fiat=- "2m ~'l' +f.tBBoz'l'
0
f.lB=- 2mc
n2kz
where nffi=--zm+f.tBB.
If now, at t = 0, the field B' is switched on, the equa-
tion takes the form
o'l' n2 ~ ~
in-at=- 2m ~'l' +f.lB(Boz+ B'ox) 'l'
Now we substitute
'l' = ei(kY-wot) [a (t) a (o) + b (t) ~ (o)]
n2k2
where nffi0 = ~, and find the equations for a (t) and b (t):
in :: = f.tBBa + f.tBB'b
in :~ = - f.tBBb + f.tBB' a
For their solution we assume that
a.(t) = Ae-ietlh, b (t) = Ce-iet/h
and for A and C we get a system of two algebraic equations:
(f.tBB-e) A+ f.tBB'C= 0
f.tBB'A- (f.tBB+e) C = 0
which has a nontrivial solution if the system~determinant
is zero. The determinant vanishes for two different values
of ffi,
Also
C -A -B+VBZ+B'Z and C -A -B-V BZ+B'Z
1- 1 B' z- 2 B'
The general solution is
-B+,B''BZ+B'2 A (n)J
'l' = ei(ky-ro0t) { A 1e-ie11111 [ a (a)+ ----'-"'n.-v---'--p v
+ Aze-ie2tlh [a (o)- B+ V;z+B'Z ~(a) J} (1)

ANSWERS 321
'I' here still remain two integration constants, A 1 and A 2 ,

lo be determined from the initial conditions a (0) = 1 and
/J (0) = 0. This yields A 1 + A 2 = 1 and C1 + C 2 = 0.
We substitute the expressions for A 1 and A 2 into (1), denote
r = f.t:
V B 2 + B' 2 , and get the following equation
1tr = eW<Y- wot) {[cos ' t - i B
V BLj-B'2
sin -r] a (a)
+i -v B'
BZ+B'Z sin 't ~(a)
}
Clearly, the probability of spin flip, i.e. of finding the

particle with spin downward at the time t 0 = .!_,
v
is deter-
mined by the square of the modulus of the coeff1cient of
~ (a) and is equal to
B'Z . 2
B2+B'2 Slll 't
with a maximum at 't= ~ = ~B VB 2 +B' 2 +.

115. Since the solution of the unperturbed problem for
e1 = e 2 = 0 (the simple harmonic oscillator; see Problem
-16) gives us the nondegenerate eigenvalues of if 0 ,
E~ = ( n +-}) nw
to find the eigenvalues and eigenfunctions of operator
if = H0 + W we must use the formulas
En=E~+<n!WJk>+ ~ l<~o~;~1 2 (1)

h.4n n h.
and
.,,'t'n = .~,o
't'n
+ L.J
"\.1 (kI w I n) ohO
£0 -EO '!'h. (2)
k=fn n k
where ljl~ = Cne-~ 21 2Hn (~), Hn is the Hermite polynomial,

n = 0, 1, 2, ... , and ~ = x (mwlli) 112 •
The problem reduces to calculating the matrix elements
<k I w I n > = <k I W1 I n > + <k I w2 I n >
21-01496
'322 PROnLEMS IN THEORETICAL PHYSICS
where wl = elx3 and Wz = ezX4 • If we Use the matrix

elements of the coordinate calculated in Problem 74
<n]xln+1>=V n~m~ 1 ) and <nlxln-1>=V ;~w

(the other (n I x I k) are zero) and write these results in
n compact form
(nIX I k) = v 2!w (Vn6n, k+1 + vn+ 1 6k, n+i)

we can evaluate (n IWi I k) by using the law of matrix
multi plication:
(n I wj I k) = e1 (n I x 3 1k) = e1 ~l (n I x 2 ll) <ll X I k)
In turn,
I I
(n x 2 ]l) = ~ (n I x p) (p I x ll) =
p
(VP 6p, 1+1 + P+ 16~, p+i)

X v
Obviously, in products of type VP" 6p, l+i we can replace
VP' by V l + 1 and 6n. p+i by 6p, n+ Resultantly,
(n I x 2 jl) = 2 ~w [ V n (l + 1) ~ 6n-1. p6p, 1+1
p
+ V nl ~
p
6n-1, p6p, 1-1 + V (n+ 1) l ~
p
6n+1, p6p, 1-1 J
and, for example, the first term will be
V n (l + 1) h 6n-1, p6p, 1+1 = V n (l + 1) (6 · 6)n-1. lti
p
. 6n,l+2Vn(l+1)=6n.I+2Vn(n-1)
Since 6n, is a unit matrix, 62 = 6.
ANSWERS 3~3
When we have performed this operation with all terms,

we get
(njx2 jl>=( Z~w) lVn(n-1)6n.l+2
+ (2n-f- 1) 6nz -f- V (n-f- 1) (n-f- 2) 6n, 1-21
If we substitute this expression into (n I x3 I k) anrl make

similar calculations, we can write
(nlx3 jk>=( Z~w t 2

~
l
!Vn(n-1)6n,Z+2 -f-(2n-f-1)6nz
Hence, (n I WI I n) = 0, and the matrix element

(n I WI I k) for a given n is nonzero in only four cases:
at k = n + 1 and at k = n + 3. Consequently, this term
gives a correction to (1) in none but the second approxima-
tion, and to (2) in the first, and we can limit ourselves
to these. If we bear in mind that the denominators in for-
mulas (1) and (2) for k = n + 1 turn into E?,- E~±1 =
= +liffi, and for k = n + 3 into E~ - m,±3 = +3/'i(t), we
can write
,., = '"o
'Yn 'Yn
+ (-li,-)3f2
2mw
[ V n (n-1) (n-2)
3/iw
•hO
'Y n- 3
+ V(n+1)(n+2)(n+3),h 0 + 3(n-f-1) 312 1h 0

-3fiw '~'n+3 -fiw ~'n+1
3 3J2
+ ~(J) 1p~-1 J
When we calculate the energy, we must consider not only
the correction of the second order in relation to but also w]
the correction of the first order in relation to W2 , i.e. we
must calculate (n I W 2 In)= e 2 (n I x 4 1 n). In a similar
21*
324 PROBLEMS IN THEORETiCAL PHYSiCS
way we get
(n I x• In)=~ (n I x 3 1k) (k I xI n)
r
k
= ( 2~w ~IV n (n-1) (n- 2) 6n, k+3 + 3n3126n, k+1

k
+3 (n+ 1/12 6n, h-1 + V (n+ 1) (n+ 2) (n+3) 6n, k-31

X rVn6n, + vn+ 16n, 1<-tl
k+1
= ( 2~w r [3n2 + 3 (n+ 1)2 ]

(We get results that differ from zero only when we multiply
6n, k+ 1 by 6n, k+ 1 and 6n, k- 1 by 6n, h+ The other products
give zero; for example~ 6n, k+ 3 X 6n, k+1 = 6n-3 , n-1 = 0.)
Finally, "
En= (n+{-) nffi+e 2 (nlx"ln)+e~ [ ((nJx~~:- 3 >) 2
+ ((n 1x3J n+ 3))2 + ((n Jx31 n-1))2 + ((n 1x3J n + 1))2 J.
-3hw hw -hw
and after we substitute the evaluated matrix elements, we

find that
En=(n+{-)nffi+3e2 ( Z~w r(2n2 +2n+1)
- :! ( 2:w r (30n 2 +30n+11).

116. In the absence of a magnetic field the unperturbed
equation
has the solution

1jJ = 1Jln1m (r, El, cp) = Rnl (r) P 1m (cos El) eim fll
In the presence of a magnetic field the Hamiltonian is

ihe ·
= H 0 + - (A ·V') [if we neglect A 2 ); here 1-l is the
~ ~
H
fl
mass of the particle. Considering the second term to be
ANSWERS 325
the energy of perturbation, we can fmd the correction to

Enz in the fi.rst approximation:
E~z = E- Enz = (n IWIn)= J'1\J~!m i~e (A· V) 'Pnlm dT:
If the z-axis is directed along the magnetic induction
vector B, we can choose the vector potential as
1 1
Ax= -z-By, Ay=z-Bx, Az=O
Then
·n (A · n) _ iliB a
( . a )-
_ -
iliB a
~ v - -- 2 XTy-Y{k 2 -arp-
and since a~~m = im1Jlnzm ( '1\Jnzm is an eigenfunction of the

operator Lz = in a~ ) , the correction to the energy is
,
Enz= -Zitnm
eB
J1 l'ljlnzm I2 dT:= ·-Zitnm
eB
To determine the eigenfunction of the perturbed problem

and the corrections to the energy in the second approxima-
tion we must calculate the nondiagonal matrix elements
(n l m I WI n l m') -_ ~
iliBe
J1 •h* a1pnlm' d
't'nlm aqJ T:
= -~
JiBe
m' JI 'Pnzm'\jlnlm'
* dT:
which are equal to zero if m =F m'.

Thus, a magnetic fi.eld does not change the eigenfunctions
of the unperturbed problem (¢' n 1m = lJl~,zm), and the energy
correction in the second approximation is zero. The energy
levels split depending on m:
Enzm=Enz- ~~ nm.
117. Once we have calculated the matrix elements of
the perturbation W = e I E I a cos - J.ta e
2 1E 1
a
[ ei(m-1)q> _ ei(m+1)q> J
m V 2n h2 V 2n 2m- 1 2m+ 1
h2m2 e I E 12 a4 1
Em = 2)la2 + fi2 f..t 4m2- 1 '
118. The solution of the unperturbed problem is 'ljl~ =

1 h2k2
=--=- eikr and E~ = - - . To make the spectrum dis-
V£3 ' 2m
crete we impose the requirement that
1~~ (x+L) = 'ljl~ (x)
To calculate the matrix elements of the energy of pertur-

bation we expand it in the Fourier series:
V (r) = ~ Vge211igr
g
Then
where 2ng 0 = k' - k, and we have the following solution

of the perturbed problem
V ei(k+2ng)r
1_
~
1h - _ eikr + 2m
't'k- V£3 n,2 V£3 g=F 0
k2:_ (k+2ng)2
and
h2k2 2m " 1 I Vg 12
Ek = 2,;"" + Vo + h2 LJ k2- (k+2ng)2
g=f 0
If ER = Ek+ 2 11g, we must seek the solution of the perturbed
problem in the form
'iJ= C1'iJ~+C2'¢~+211g
From the secular equation we get

1 1 2 h2k2
1 ~1,2=-=- [eikr+ei(k+211g)r] Ek' =2'nl+Vo±IVgj
V2L3 -
The energy spectrum of the electron in the periodic field
has forbidden ~bands ne.ar (k·g) = -g 2 , the width of which
is 2 I VgJ.
ANSWERS 327
119. In the absence of the field the unperturbed problem

lthe hydrogen problem; see (111-24) and (III-24a)l has
n 2-fold degenerate eigenvalues. Hence, the value n = 2
has ·corresponding to it four eigenfunctions determined by
the equation
'We write them out in more detail:

'¢~ = 'liJ2to = R21 (r) P to (cos 0) = R 21 cos 0
'ljJ~ = '¢211 = R 2t (r) P 11 (cos 0) ei<P = R 21 sin Oei<P
'ljJ~ = '¢ 21-1 = R 21 sin Oe- irp (1)
Directing the z-axis along the external electric field E,
in which the atom is placed, we can write the Hamiltonian
of the perturbed problem in the form
~ ~ 0 ..
H=H -elEiz, obviously, W= -eiElz
The eigenfunctions of this operator that correspond to
E~ must be sought in the form of a linear combination
4
of functions (1): 'ljJ = ~ C;'Jl?. The energy levels are found
i=-1
from the condition that the system determinant be zero:
I (E~- E) 6;k + (i I WI k) I= 0 (i, k = 1, 2, 3, 4) (2)
Clearly, (i I W I k) can be represented as a product of in-

tegrals over r, 0, and cp:
(iiWlk>= -eiEI) 'iJ?*z'liJRd-r= -eiEIIrlelcp
The integral over = (2 1 w 14>
= {3 J w J 4) = 0
In addition, if m = m' and l = l', we get

n
I a= JP m (cos e) cos eP m (cos e) sine de
1 1
0
1
\ 1Pzm(x)l 2.'t'dx=0
=
.:1
because 1P zm (x) 12 is an even function of x and, hence,
the integrand is odd. For the same reason all the (k I W I k)
are zeros. Only (1 I W I 2) and (2 I W 11) are nonzero:
oo n 2n
(1 I w 12> = - e IE I JR2oR21r dr J 3 cos 2 e sine de Jdq>

0 0 0
4n
= -3eiE Ir I
00
To calculate Ir = JR 20 R 21 3r dr we must write the normal-

o
ized radial functions for hydrogen explicitly. According
to (III-24),
r
R n z = e- r;a ~ a" ( :
n-1
k=l
r
fi2
where a= - 2- is the Bohr radius, and different ah arc
1-te
linked by the recurrence relation
2( ~-1) ah-1
a"= /c(k+1)-l(l+1J
Con~equently, R 21 =a 1e-r/2a: {in tl1is case l=n-1=1

and the entire sum reduces to one term a 1 : ) and R 20 =
=a~e-r!2ax (1- ~ ). (Here n-1=1, l=O, and a 1 =
2(+-1) a~ , )
- x 1 2 = - ~ . The coefficients a; and a 1 are
ANSWERS 329
determined from the normalization conditions
:a rr dr = 1
00
J1'!'zoo 1a.= J1¢~ 1a.= (a~? 4:rt J

2 2
0
e-rta ( 1 - 2
00
J l'iJzto d't = J 'I'~ d't = a:2:n J

12 I 12
0
e-r/a ~:r
:rt
X Jcos 0 sin 0 dO= 1

2
Jr
00
Since 11 e-r!a dr = a11 +1n!, we get

0
e-r/2a r
R2o= 1 e-r/2a(1--r-), Rzt= -
V8na3 2a V8na3 2a
Substituting these into In we can write
00
== -16na4 JI e-rla ( 1- _r_) r~ dr

I 1-
r 2a
0
-1-(4!a 5 - -2a
=16na4 1-5!a6 )= -~
4n
and, hence,
<1 I w I 2 > = <2 I w I 1> = 3e I E I a
Equation (2) yields
E' =E~ + 3e IE I a, E" = E~- 3e IE I a, E"' = E"" = E~
The corresponding funclions are
1p' = c; ('IJ~ + ~J~), 'p" = c~ (~J~- '!'~)
From the normalization condilion we get

' , 1
cl = cl = y2
The splitting of the energy levels of the hydrogen atom
in an electric field proves to be proportional to the magni-
tude of the applied field.
120. In the absence of the f1eld the Schrodinger equation

for the unperturbed problem
[- ;; ~ + V (r) J1jJ = H \jl = Ent'i' 0
has degenerate eigenvalues: for every level Enz there are

2l +1 eigenfunctions 'ilntm that differ in m while retaining
the same n and l. Consequently, we must look for the wave
function of the perturbed problem in the form
l
'ljJ = ~ Cm11Jnlm
m=-l
For an atom in the external electric field E the equation is

(if o - e I E I z) 1Jl = E1!J
and the term -e I E I z (the field is directed along the
z-axis) plays the role of perturbation W. We get the energy
levels by making the determinant zero:
I (Enz - E) 6mm' + (m I W I m') I = 0
Obviously,
(m I WI m') = -e IE I ~ 'IJl~tmZ1JlnLm' do:= - e IE I I,.lel(j)
For m ==1= m',
2:rt
I I}= ) e-imq>eim'q> dqJ =0

0
For m=m',
:rt
I e = ) P 1m (cos 0) cos 8P 1m (cos 8) sin 8 d8

0
1
= ) IP 1m(x) l2 xdx=0
-1
because the integrand is odd.

Hence, the equation for E reduces to
(Enz - E) 21 +1 = 0, i.e. E = EnJ
ANSWERS 331
There is no: splitting (in this approximation) of the energy

levels in the external electric field.
121. As in Problem 60, we denote the ordinary coordinates
and energy as r' and E', and the dimensionless as r = ~ a
2
( a = f!e
/i 2 ) and e = :1' . Now we can write the Schrodinger
. e a ·
equation for a hydrogen atom placed in an electric field E
that is parallel to the z-axis:
--} ~ 'iJ - + 'iJ- gz'iJ = e1~

Here the perturbation is -e I E I z' = -gz, where g =
= !Eiaz is the dimensionless analog of the external electric
e
field.
In parabolic coordinates u = r + z, v = r - z, (jl (see
Appendix 3) this equation takes the form
81jJ ) 81jJ ) 1 ( 1 1 ) [)21jJ
au: uau + av, v av- +-:r -u+-v
f) ( f) (
a(jlz
+ [ 1 +e utv + 1(u -v 2 2) J'i'= 0

which permits separation of variables:
¢ = U (u) V (v) eimqJ (m = 0, +1, +2, ... )
Here U (u) and V (v) satisfy the equations
_!:_
dtt
(u dU }
du
+[at+.!:!:-~-=-+
2 4u
JU = 0
gu2
4
(1)
:v ( v ~~ ) + [ ~~ + e~ - ~L: - g~2 JV 0 = (2)

and a 1 +~~ = 1. If we compare these with the correspond-
ing equations when g = 0, solved in Problem 60, where we
found that
U~ 1 (u) = e- u V- ~ F~~~ (u) ulml/2
a=-v -~(2nt+lml+1)
(for the fll'st equation), we can seek the solution of (1)
using the perturbation theory. The eigenvalue a is non-
degenerate, so we must look for the correction da = a1 - a

in the form
00
A
Lla=- Jl Un0*1 - gu4-2 0
Un 1 du
0
Here
00
Jw~~l 2 dU= 1
0
The polynomial F~~~ satisfies the equation
u=
d2Fiml
nl
du2
+ (I m I+ 1- 2u v-- 2
aplml
- ~) _n_t
du
+ [ a-(lml+1) V- ~ JF~~ 1 =0
If we substitute the value of a and introduce a new variable
x=2uy- ~ (3)
the equation becomes
x
azplml
dx~1 + [j m I+ 1-.x·] a:
dFiml
1
+ n 1 F~~I = 0
nt
Now, writing F~~l = ~ bhx,.., we fmd the recurrence rela-
h=O
tion for the coefficients bh:
k-nt
bh+i=b,.. (k+1)(k+imi+1) (k=O, 1, ... ) (4)
We can verify explicitly that if we choose

b0 =(I m I + 1) (/ m I+ 2) ... (I m I + n1)
we have
F~~l = ex] dnl (e-xxlmJ+nl) = ( - t t l Xnt + . .. (5)
xlml dxnt
The other coefficients can be computed by using formula (4).

ANSWEll.S 333
For an arbitrary function f (x) we evaluate the integral

00
I= ) xlmle-xF~~lf (x) dx
0
Substituting (5) and integrating by parts, we find that

00
eX ant (e-X•Xnt+fml)
I= ) xlmle-x__ f dx
xlml axnt
0
00
= ( -1)nt r e-Xxnt+fmj ant! dx

J0 axnt
Using this result, we calculate the normalization integral

and the matrix element. We introduce the variable (3) and
write
nt (x) ~·lml/2
Unt -- Cntme-x!zplm\ ~
Then
In the same way
Je-xxlmlF~~l (F~~I. x2) dx

2 00
A g Cntm
uCX.= -~~ (V -2e)3
0
2 00
g Cntm (-1)nt f e-xxnt+lml dxanntt (F~m,l.x2) dx

= -T <V -2e)3 ~
Since
ant (F~;Ix2) 2 +b +b 1
bnt (n 1 +2)!
21 X nt-1
:(nt+1)1
il X nt-2nt
axnt
and bn 1 = (- 1t 1 and since formula (4) yields

b _nt(nt+lmi)b b _(nt-1)(nt-1+1ml)b
n1- 1 - _ 1 np n1- 2 - _ 2 n1- I
we get
00
L1a = g I e-x [ (nt + 2)! xn1+lml+2

8en1! (nt +I m J)! J 2!
0
- n!(nt+lmi)(nt+1)! xn!+lml+l
1
+ nt(nt-1)(nt+lml)(nt+lml-1)
21 n1 1.xn 1+1ml]dx
After simplifying we get
L1a= :e [6n~+6nt(l ml + 1)+ (\ m \+ 1) (J m\ +2)]

If we turn to Eq. (2), in which the term - g~z plays
the role of perturbation and the eigenvalue is -~ 1 , we get
00
.:1~ = ~~- ~ = J\V\~ 1 (v) \

0
2 gt dv
= - :e [Gn~ + 6n 2 (/ m / + 1) + (\ m / + 1) (\ m J + 2)]
(since the unperturbed equations (1) and (2) are identical).

The energy s is determined from the condition
1=a+L1a+~+L1~=V- ~ 2(nt+nz+lm\+1)
+ %e [6(n~-n;)+6(\m\+1)(n 1 -n 2 )]
Since n1 + n 2 + I m I + 1 = n and in the correction term
proportional to g we can replace s by its value e 0 = - 2 ~2
from the unperturbed problem, the final condition takes
the form
- 2en 1r1 -
V- 23 gn 3 (n 1 - n 2 ) J
= 1
ANSWBRS 335
whence
1 3
8 = - 2n2 -2gn(n1-nz)
Since for a given n the numbers n 1 and n 2 can assume values

from 0 to n - 1, the maximal value of n 1 -n 2 is equal to
n - 1, and the minimal value is equal to - (n- 1).
Hence, the correction to e 0 can assume 2n - 1 values.
Thus, an energy level in the hydrogen atom under the
influence of a field splits into 2n - 1 levels, and the magni-
tude of the splitting is proportional to the field (g ex I E 1).
122. We determine the electric fteld intensity inside the
nucleus using the Gauss theorem, and bearing in mind that
the potential must be continuous at r = r 0 , we fmd that,
for r < r0 , the potential energy of the electron is
U(r)= - 3-Ze2
- ( 1 -1-r2- )
2 ro 3 r3
i.e. the perturbation is
Ze2
V(r)=-r--U(r) for r~Jo
=0 for r>r 0
The correction to the lowest level (n = 1) is calculated
using '¢ 0 = '¢100 = V ::
3 e-Zr/a and is equal to V 00 • Since
r0 ~ 10-12 em and a ~ 10-s em, when we evaluate the
integral, we can substitute unity for the exponential func-
tion and get
. Em
Even for Z ~ 100, the ratiO E~o> ,..,... 10- 4
0
For the 4-fold degenerate level (n = 2) we find the cor-
rections from the corresponding determinant. They are V 11
and V 22 = V 33 = V 44 • (The calculations are done using the
functions '¢ 1 from Problem 56. None but the diagonal matrix
elements of V (r) are nonzero.) If we neglect terms of the
order of r 0/a compared with unity, the corrections prove
336 PROBLEMS IN THEORETiCAL PltYS1CS
equal since
E<u = _1_ Z4e2 ( ~)2 (l = O)
1 20 a a
Em= _1_ Z2e2 ( Zro )'' (l = 1)
2 1120 a a
Hence, the level with n = 2 splits into Lwo sublevels,
the energy depending on l as well as on n.
123. The functions '1'0 and '1'1 satisfy the equation
H~qr n =·nD'Yn
~ iJt
and
i
'I'n=Un(r)e-TIEnt, J 'l'~'l'md't=8nm (n, m=O, 1)
After the perturbation W is switched on, the Schri:idinger

equation
is to be solved in terms of the stationary states:

'I'= a 0 (t) '¥ 0 a1 (t) '1'1 +
where I a 0 12 + I a 1 12 = 1, and the initial values of the
coefficients are a0 (0) = 1 and a1 (0) = 0. Substituting '¥
into this equation, we find by the usual method that
'
ilia0 =a0W 00 + a 1W 10ej_(Eo-EJ)I
1i
• j_(E!-Eo)l
ilia 1 =a 0 W 01 e~'~ +a 1W 11
where W;,. = J
uHVu~< d-e are the matrix elements of W.
If we introduce a 0 (t) = a 0 (t) and a 1 (t) = ei<Eo-E,lt11ia1 (t),
the equations for a;
+ a 1W1o
ilia0 = a 0TV 00
.
ilia1 = a W 1+a 1 [WH + E1-Eol
0 0
are homogeneous linear equations with constant coefficients,

and a; can be sought in the form a 0 = Ae-mt and a 1 =
ANSWERS 337
= Be-W.t. Coefficients A and B are found from the equations

(W 00 - nQ) A WtoB = 0 +
WotA + [W11 + E 1 - Eo- nQ] B = 0
The determinant of these two equations vanishes for the
two frequencies Q 12 :
"?
nQt2=Woo+--z±
-. /
V !Ww
12
+4
"?2
with
Hence,
ao = Ate-Wtl + A2e-i!J2t, ai = Bte-Wtl +B2e-i~t
where
__ liQi-WooA (._ 1 2)
B ,- w f t- I
10
(1)
The initial conditions permit evaluation of the constants of
integration:
(2)
and, hence,
a 0 = A 1 (e-m,t_ e-i!J2t) e-i02t +
at= Bt (e-Wtt_ e-m2t)
If we substitute (1) into (2), we get
A _ Woo-1i!J2
1- 1i (Qi -!Jz)
(liQt-Woo) (1iQ2-Woo)
li(Q1-Q2)W10
Q.E.D.
Let us now estimate the behaviour of I ai 12 = I ai (t) 12
in time. Calculating the modulus in the usual way, we
find that
+
I ao 12 = 1 4 (A~-At) sin 2 crt
and
I a1 12 = 4B~ sin2 crt
where
1i
22-0U96
Using the expressions for Q1 and Q 2, we can show that

4 (A~-At) = (2At-1)2-1 = [ 2Woo-li (Qt+Qz)
1 !i (Qt-Qz)
]2 -1
_ __r:__ 1 __ I~Wtol 2
-4a2 - a2
since 4cr2- y2= 41 W 10 12 and, hence,

Ia.o'l2=
..
Iao 12= 1- I Wto
a2
12 sin2crt
Ia.,l2 =I at'lz =II Wd~ 12 sin2 crt

The perturbation results in the system oscillating between
the two states 'I' 0 and '1'1 with a frequency
1'-./ y2
cr=-,;: V 1Wtoi2+T·
124. The probability of transition is determined by
formula (III-34). To compute C11. (t) we bear in mind that
Wkn = constant for 0 :::;:;; t:::;:;; T and is equal to zero for all
other t. Hence,
't sin rokn't"
C (t) = Wkn f eiwknt dt = 2 ~kn eiwkn 1:12 2
k in J '!i rokn
0
and
sin2 { Ek-En •)
pk.-+n= _!i~ I Wkn 12 { Ek _ 2; n ) 2 = : 2 I Wkn 12 F(Ek -En}
!i
The function F (Ek -En) has a maximum at En = Ek
and if T is sufficiently large, we can express F using th~
delta function. Let us consider the integral
00
J= J
-00
/(Ek)F(Ek-En)dEk
We su bstitute Ek-En = z and dEk = 21i -:t, and we find

dz
2 1i T
that
00
J= i !(En+
J 2hz)
T
sin2z ~dzx-r
z2 2
-00
ANSWERS 339
Now if we direct -r to infinity, we get

J=-rn ~ f(E,J, i.e. F(Ek-En)=-rn ~ t3(Ek-En)
and, hence,
Pk-+n = 21i11 't I W kn 12 t) (Ek- En)•
125. The probability of transition from the state ljl1
into ljlp under the influence of the perturbation W (r, t)
is determined in the first approximation by the coefficient
1 r T
cp (T) = T!i J W pt(t) e

i
Ep-Et
"
,
dt
0
where Wp 1 = J
ljl~W (r, t) ljl1 d-r, and E 1 and Ep are the
energies of the corresponding states. For the initial state
we must write the wave function of the electron in the
Coulomb field of the nucleus Ze with n = 1, l = 0, m = 20 •
This function is ljl1 = ljl100 = Ce-rzta, where a = ll~2 •
From the normalization condition we have C = 11: : . y

In the final state the electron is described by a plane wave
1
ljlp = lfveiprtli, where V is the volume of the region where
the atom is located. To determine the number of states
with a given momentum we impose, as usual, the require-
ment of periodicity, ljlp (x + L) = ljlp (x) (where £3 = V),
and we get the allowed values of momentum:
hn
Pn=y
It follows from this that the number of states with mo-
menta in the interval [p, p + dp] is equal to p dpxdpydpz =
= 2 dpxtydpz V (the factor 2 indicates the number of
possible spin projections). Now if we use spherical coordi-
nates inlthe momentum space, we can write the number of
states with the energy in the interval [Ep, Ep + dEp]
(where Ep = f;)
and with the direction of the momentum
22*
in the angle dQ as
2mp
p (Ep) d Ep d Q = 11,3dQ dEp
For this problem the vector potential is given as
Ax= A cos (wt- kr), Ay = Az = 0 (k II OZ)
Hence,
~ e ~ eA ~
W= -lt(A·p)= -it cos(wt-kr)XPx
Since Wp 1 = J¢~W¢1 d-r: = Wtp = [ J ¢~W'~Jp d-r:]* and
Px¢p= Px¢p, it follows that the matrix element Wp 1 can
be written in the form
W pt = - e~x J'IJ~ cos ( wt- kr) 'Pt d-r:
Representing cos (wt- kr) in the form·

cos (wt- kr) = ~ [ei(oot-kr) + e-i(oo!-kr)]
we can write
T i
(T) -- - epxA [ \ '"* -ikr,h d \ ii<EP-E!+hw)t dt
cP 2iliJJ. J 't'Pe 't't 't J e
0
+ J'IJ;eikr~'t d-r: Jeh(EP-Et-hw)t J

T, i
dt
0
Since
i
T i -(E -Et±hw)T
\ h(Ep-Et±hw)t d _ eh P -1
J e t- i ,
o (Ep-Et ± liw) h
of the two terms in Cp (T), the one in which the denominator
contains Ep - E 1 - nw ~ 0 plays the main role in the
photoelectric effect (Ep > E 1). We neglect the second term
and substitute the plane wave and the Coulomb wave func-
tion for lpp and ¢ 1 , respectively, and we get
_!:_(E -Et-hW)T
c (T) _ _ eApx .. / Z3 eh P -1 I
P - 2!iJJ. V :rta3V
[Ep-E 1 -/iw] h
i
ANSWERS 341
where
I= Je-Zr/ae 1
• pr
(kr- T) d't = Je-Zrlar2 dr Je-iqr
oo n
cos a
0 0
2l"C
X sin 8 d8 Jdcp
0
Here f- k = q, and the z-axis is directed along q so that

00
(q·r)=qrcose. Then, since \ e-arrdr=1/a}, we have

0
l= ~:rt
~q
r.lr
.
oo z
e-r(-;z-iq)rdr-
oo
J0
z
r e-r(-;z+iq)rdr]= .
B:rt
2
z + q2)
(~
.!_
a 2
The probability of transilion into the interval 11Ep,
dW= Jlcp(T)j p(Ep)dEpdQ

2
t;.Ep
contains along with a smoothly varying function of E the

expression
For this reason the probability of the photoelectron being

emitted into the angle dQ in unit time is
dW -32 ~ A2Z5 P~P ~6(E -E 1 -nro) dQ
T - n J..l. h3a5 ( Z2 ) 4 li P
-+q2
a2
with Ep = E 1 +
nro = nw - J, where J is the energy of
ionization of the atom; E 1 = -J = - 11;~~ 2 • Substituting
the explicit expression for q2 into the denominator, we get

~+qz=~+
a2 a2
..f!_+kz-
ft2
2Pk cos a
n
=
2!l [ftZ2e4
h2
p2
2ft2 + 2/l
J+ k2 --,;-cosa
2pk
=~nro+k
12
2 - Zpk cos a;::::::: 2""w
n n (1- ~cosa)
c
if we neglect k 2 <t p 2/h 2 •

Now if we direct vector k along the z-axis, we get Px =
= p sin a cos cp and finally
dW = _::_ A2 ~ sin2 a cos 2 <p 3 dQ 6 (_f!_ + J _ hro).
T !l 21ta5/l4 w' {1--cos
v a)4 P 2!l
c
126. According to the conditions of the problem the

equations
~
HtUn (r;) =- ft2
Zm l!;Un + V (r;) Un = EnUn (r;) (i = 1, 2)
have been solved, and we must find the eigenfunctions and
eigenvalues of the operator if = il 1 + il 2 + fi 12 assuming
that il 12 = il 21 • We view il12 as the perturbation and first
solve the unperturbed problem
Hol:J'ot (rl, r2) =(fit+ H2) Wot (rl, r2) = Eol:J'ot (rl, r2)
Obviously, we can write
Wo1 (r1, r 2) = Ur (r1) Us (r 2)
The energy level corresponding to this function is E 0 =
= Er + E •. But we can bring another function into con-
formity with this energy level, namely,
l:J'o2 (r1 , r 2) = Ur (r1 ) Ur (r 2)
(The particles have exchanged states. The first is now in the
s state, and the second in the. r state.) ~Since the energy
level E 0 is two-fold degenerate, we look for the eigenfunc-
tion: of the perturbed problem in the form
1:J' = a1:J' o1 + b1:J'02
We substitute this into the equation (il1 + ii 2 + H12 - E) X
X 1:J' = 0 and denote E = Er Es + +
e, and we~·get the
ANSWERS 343
simplified equation
(H12 - e) (a'l'ot + b'l'o2) = 0
Next we multiply this first by '1':1 and then by 'l'ri2 and
integrate over the entire space. We find that we have two
equations for a and b:
(C11 - e) a +C 12 b = 0 (1)
C 21 a + (C 22 - e) b = 0 (2)
Since fi 12 is symmetric, the coefficients C11 = C 22 and
C12 = C 21 . Indeed, if we replace r1 by r2 and r2 by r1 , this
transforms 'I'01 into 'I' 02 and, hence, the coefficient
Cu= JJw:Jitz'l'otd'ttd'tz
= JJu~ (rt) u~ (r2) HtzUr (rt) Us (rz) d'tt d'tz
differs only in the labeling of the variables from
Czz= JJw: fitz'l'ozd-rtd-rz

2
= JJu~ (r2) Us (rt) HtzUr (rz) Us (rt) d-rt d-rz

ln the same way,
Ctz= J w:Jitz'l'o2d-rtd-rz=C21= r 'l'ri/l21'1'otd-rzd-rt

since H12 = H21 • We denote C11 = C 22 = K and C12 =
= C 21 = A and write equations (1) and (2) in the form
(K - e) a + Ab = 0
Aa + (K - e) b = 0
From the condition that the determinant vanishes we find
that
K- e =+A
(1) e = K + A and a = b. Hence, the level E' = Er +
+ E, + K + A has the corresponding function
'I'' =a ('l'ot + 'l'o2) = a lur (r1) u, (r 2) + Us (r1) Ur(r 2)]
(2) e = K- A and a = -b. The level E" = Er +

+ E. + K - A has the corresponding function
'¥" = a ('I'ot - 'I'o2) = a [ur (r1 ) Us (r 2 ) - Us (r1 ) Ur (r 2 )1
1
From the normalization condition we find that a = y- 2
since
and
·' 'fri1 'foz d'tt dTz = ( ~ ui (rt) Us(rt) d'tt r =0
If the particles involved are electrons, the complete func-
tion (J) (r1 , r2 , a 1 , a 2), i.e. the spatial and spin functions,
must be antisymmetric. Hence, tho level
E' = Er + Es + K + A
has one function corresponding to it:
1
<Da = V2" ('fot + 'foz) 'Xa (all O'z)
(There is only one antisymmetric spin two-electron func-
tion; see Problem 110.) Consequently, this is the singlet
state. The second eigenvalue
E" = Er + Es + K- A
corresponds to the triplet state, since we can combine with
1
the antisymmetric spatial function 'I'" = y2 ('I' 01 - 'I'o2)
three symmetric spin functions
?(~=CXtCX2, ?(~=~t~2• 'X;;'= J (ctt~2+~tCX2)•
2
127. The wave function of the system at the initial time
is given in the form (see the solution to Problem 126)
'I' t=o = Ur (r1) Us (r 2) = 'I' 01 = 'I' (0)
To determine its change in time, we represent 'I' (t) as a
result of the action of a certain operator S on 'I' (0): 'I' (t) =
= S'I' (0). We substitute this function into the time-depen-
ANSWERS 345
dent Schrodinger equation:
itt ~ qr (O) = fisw (O)

After a formal integration over t we find that
~ _.!. flt _.i.flt
S= e " and '¥ (t) = e 11 '¥ (0)
If we expand'¥ (0) in a complete set of eigenfunctions of ii,
which satisfy the equation H¢n = Enl!Jn, we get
i ~ i
--Ht ""' --Ent
'¥ (t) = e h ~ Cn'IJn = Ll Cne 11 l!Jn
n n
since the action of any operator function f (H) on the eigen-
functions of ii is
f (H) 'i'n = f (En) 'Pn
Hence, to determine'¥ (t) we must first expand its initial
value'¥ (0) in a complete set of eigenfunctions of H, i.e. in
a series in l!Jn·
In Problem 126 we found the eigenfunctions of H in the
zero approximation.
1
The function '¥' = VZ ('¥01 + '¥02 ) corresponding to the
1
level E' = Er + Es + K +A and '¥" = V2 ('l"ot- '¥ 02 )
corresponding to the level E" = Er Es K - A. + +
From this we find that
'l"ot =J 2 ('¥' + '¥") = '¥ (0)
and according to what was proved earlier
'l"(t) = J2 ( i- e- E't'¥' + e --k-E"t'¥")
If we substitute E' and E" and return to 'l"ot and '¥02 , we

find that
nr () __ 1 -{ (Er+Es+K) 1 [(nr nr ) --k-At (nr nr ) {At]
T t - 2 e Tot+To2 e T 01 - T 02 e +
= ct(t) 'l"ot + Cz (t) 'l"oz
where
[ - ~. (Er+Es+K)t Jcos-;;-
At
*
ct(t) =exp
c2 (t) = exp [ - (Er+Es+ K) t sin J ~t

i.e. the probabilities of finding the system in the 'I' 01 or
'I' 02 states is, respectively,
W 01 = I c1 12 =cos2TAt , w 02 = Ic
2
12 = Sin
. 2 TAt
The time needed for the particles to exchange states, i.e.

for 'I' 01 = Ur (r1 ) Us (r 2 ) to transform into 1f 02 = Ur (r 2 ) X
X Us (r1 ), is determined by the fact that c1 (-r) = 0, i.e.
A -rln = rc/2, whence,
nli
-r= 2A .
128. The parameters A and ex. of the trial function = Jcp*Hcpd-r
calculated via this function will become a minimum provi-
ded the set of trial functions are normalized. This last
requirement gives
f J(1+cx.r) e-2a.rr dr
00
I<JJI 2 d-r=4nA2 2 2
0
4!a2 J 7nA2
= 4 rcA
z[ 21
(2a)3 +2c:x. (2a)4
31
+ (2a)5 = ----aa = 1
i.e.
Az=~
7n
The Hamiltonian of a three-dimensional isotropic oscillator
ii = i' + v= - ~
2m
~ + rmwz
2
r2
Since p is hermitian, we can easily transform (T}:
<T> = 2~ Jcp•j)2cp d-r = 2~ JIpep 12 d-r = 2~ JIgrad cp 12 d-r

ANSWERS 347
and since <p does not depend on the angles, we get

I grad <p 1 =I ~~ I= Aa2re-ar
and
Calculation of (V} gives us
81 mffi2
= 28fi.2'"
and, hence,
(H} = li (a)= 1: n-:2 +~~ :~2
The normalization condition is considered explicitly, and
we must find the minimum of H (a). Obviously, a 0 is deter-
mined from the equation
_!_ (~
aa 14
ft2a2
m
+8128 mffi2)
a2
I
ao
=O
and is equal to
2_3-./3rm(!)
ao- V 2 -n-
and thus
-
H (a0} = (H}m 1n = 7
9-./3
V nw ~ 1.575 nro
2
which is only 5 per cent more than the exact value of the
lowest energy level for this system, Eo = liro. i
129. The normalization condition for the function <p =
= Be-ttrj2a is
00
Jr I <p 1 d't = 4nB2 Jr e-arfa,-2 dr = 4nB2 ~

2 21a
=
3
1
0
and so
B2=~
8:rta3
As in Problem 128, we compose~=

dr
-B.!!:_e-arf2a and
2a
calculate
(T) ~~ JI dc.p 12
-
dr
dT: X f12 ft2
- = - B2
2m 2m
a2
-4n
4a2
J
00
ft2a2
e-arfar2 dr=--
8ma2
0
and
<V>= -A Jr cp*e-rfacpdT= -4nAB Jr e -<a+I)-ar

00 T
2 2 dr
0
a3
= -A(1+a)3
From the condition of the minimum of the function
we get.
ft2a0 _ 3Aa~ =0
4ma2 +
(1 a 0 )4
Only numerical methods are suitable to solve this equation

of the fourth power in a.. For this we must have the concrete
values of the parameters of the problem. The ground state of
the deuteron will then be equal to <H) {a. 0 ).
• n2 e2
130. For the hydrogen atom, H = - 2m -~ - - . As
r
in Problem 128, we compute the sought quantities using
both functions, and we get
me 4
(H) min [ 'Ptl ~- 2.071!2
Clearly, cp1 is a better trial function than cp 2 since

Hmtn [cpl] <Hmtn [q>z].
ANSWERS 349
131. Let us represent the Hamiltonian of the problem of

two electrons near the nucleus
~ n2 !i2 2e2 2e2 62
H= -2m ~ 1 -2m ~z---;;--r;-+~
in the form if= if 1 + ii 2 + H12 , where
Ht = -~~t _ (2-s) e2
2m rt
Hz= -~~ 2 - (2-s)e2

2m r2
~ e2 se2 se2
Htz=-----
rt2 Tt r2
The equation H1 u (r1 ) = E 0u (r1) is the Schrodinger equa-

tion for a hydrogen atom but with a charge of the nucleus
(2 - s) e (we shall determines in such a way that E is mi-
nimal). Consequently,
me4
E 0 = - 2!i2
- (2-s} 2l• ~ .
and the corresponding normalized eigenfunction is

V where ~'=
u=
-vn'\'3 e-vrt
, I
meZ
n2
(2-s)
The ground state of helium in the zero approximation (the

unperturbed problem) is described by the function
'¥ (r1 , r2) = u (r1 } u (r 2)
We must calculate the energy correction 8 = E - 2E 0 as
8= J'f*Htz'fdTtdTz
=e2'\'6
n2
f
J
dTt f
J
dTze-2'\'(rt+r2)(_1__
rt 2
!..__!.._)
r1 r2
(1)
It is easiest to integrate using elliptical coordinates, placing
the nucleus and the first electron in the focuses and denoting
r 1 = 2c, r 2 = c (£ - '1']), r 12 = c (£ +'I'])
dT 2 = C3 (£ 2 - '1'] 2) dsd'l']d<p
The coordinates £, 'I'], <p change within the limits
1 =::;;;; ; =::;;;; oo, -1 =::;;;; 'll =::;;;; 1, 0 =::;;;; <p =::;;;; 2n
Then
00 1
e= e~6 ) d't'te-2Yr1 X 2nc2) ~) d1J(~2-1)2)
1 -1
X e-2yc(;-11) (-1- _.!_ --~~-)

~+11 2 ~-1]
= e~6 ) d't1e- 2Yr12nc2I (r 1) (2)
We evaluate the following integrals over ~ and 1):

oo -2yc 00
I - ) e-2yci,n: __e__
o- .....,- 2yc ' It=) e-2yc'~d~
1 1
e-2yc
=~
(1 +-1
2yc
)
e-2yc [ 2 2 J
Iz-
-
oo
)
1
e-2YC' ~2d ~-
-- - 1 + -2yc
2yc +- -
(2yc)2
1
e2yc e-2yc
Ao = ) e2Yc11 d1J = -
2yc
-1
1
At= r e2'1'CTJ1Jd1)= e2yc
J 2yc
(1--1-)+ e-2yc (1 +-1-)
2yc 2yc 2yc
-1
Az =
re2'1'CT]
J 1)2 d1) = e2yc [
2yc 1-
2
2yc + (2yc)2
'2 J
-1
-2'VC
e-2yc [ 1
+ ~+-2-J
2yc (2yc)2
These integrals enter into I (r1 ) in the following way:

I(rt)=; (IoAz-I2Ao)-(s+1)IoAt+(1-s)ItAo
After substituting the computed Ik and A,., we get
I(rt)= ( 2~) 3 [1-s-2ycs-e-ivc(2yc+1)]
2
= y3rf [1-s-ysr 1 -e-2Yr1(1+yr1)]
ANSWERS 351
and, introducing it into (2), we integrate over r 1:

00
e= e:~6 J4nr: dr e- 2vr1 X 2n ( i )2 I (rt)

0
1
4e21' J[(1- s) e- 2vr1- S'\'r~ 2vr1_

00
= 3 r1 e- r 1e- t.vr1
0
-l'r:e-4Yr1)dr1 =e21'( ~ -2s)
Consequently, since e2y= ";:: (2-s), we have
E= 2E0 + e= -"'"i2
me4 me4
(2-s) 2 +--;:;,2 5
(2-s) ( 8- 2s )
me4
= -¥(2-s) (11
s+s ) (3)
From aE = 0 we find that

as
5
so=w
Hence, the energy of the ground state of helium is
Emln=E(so)=- ( 16
27 )2 Vme4
The ionization potential of the atom is then the difference
between the obtained value I E (s 0 ) I and the energy of a
singly ionized atom of helium
22me4 2me4
IEHe+l=---w-=~
i.e.
J= ~; [ ( ;~ ) 2 - J
2 = 0.8477 ~; = 23.0 eV
(The experimental value is J = 24.46 eV.)
132. According to the conditions of the problem, we must
find the approximate wave ifunction that satisfies the equa-
tion ·
L\lJJ +
k2lJJ - 'AU (r) lJJ = 0
where k 2 = 2mE/n 2 > 0 (the scattering problem), and
'AU = 2m V (r)/n 2 is considered small. For 'A = 0 the equa-
tion ~¢ 0 +
k2¢ 0 = 0 has (solution ¢ 0 = eikr (the incident
particle).
In the first approximation we look for ¢ = ¢ 0 A.¢1. +
Identifying members with A, in the first power, we get the
equation for ¢ 1 :
To find its solution we compare (1) with the equation for

retarded potentials
1 82cp
~c:p-7 8t2 = - 4:rtp(r, t)
whose solution is
p ( r', t _ I r -: r' I1)
c:p (r, t) = J I r-r' I d,;
when p = p0e-icJ>t and c:p 0 (r, t) = c:p0 {r) e-iwt. Then c:p0 (r)
satisfies the equation
{2)
and has a solution
i~lr-r'l
i Po(r')e c
C:Vo {r) = J I r-r' I dT:'
00 Ueikr . . .
Equation (2) for c = k and Po = - 4n comctdes wtth
Eq. (1). So. for the scattered wave we can, obviously,
write
'i'scat = A.¢1 = - 4n J
1 r dT:' eiklr-r't 2m
I r-r' I f/.2 V (r') eikr'
For r:} r' we have the approximate relationship I r - r' I ~

~ r - r' cos ct. We neglect r' compared with r in the deno-
minator of the integrand, denote kr' cos a = (k1 ·r') {k1 po-
ints in the direction of the scattered wave), and introduce
I k - k1 I = K = 2k sin ~ , where 9 is the scattering angle.
ANSWERS 353
As a result we get
'" --
'!'scat- - 4n r
li,2 - , - J d't'Iei(k-k!)r'V ( r I)
1 2m eikr
= - -2m- -
eikr
-
4nfl2 r
JV
00
(r 1) r 12 dr 1 X 2:rt J•
11
etKr
1
cos ij' •
SID 8 I d8 I
0 0
We assume that V (r') is a central potential. Since

11
(" eiKr' cos W sin 8' dO' = 2 sin Kr'
J Kr'
0
00
2m (" sinKr'
if we denote f (8) = - ."ll2 J dr'r' 2V (r') Kr 1 , we can write
0
for the scattered wave
eikr'
'i'scat= - • f (8)
r
Composing the current densities by using the functions

and 'i'scat, we get
'i'o = eikr
dcr= lisc~tdSI =lf(8)j2~=j/(8)12dQ.

I hn I r2
133. We compute dcr for V (r') = ~~ 2 by using the results

obtained in Problem 132. To make the integral converge.
we temporarily introduce a factor e-a.r', i.e. we set V (r') =
Ze2
- , e-a.rl. Then
r
00
2m Ze2 ( -a.r' . K d ,
f (e) = -Ji2 2m Ze2
sm r r = - 7i2 a 2 + K 2
1
K Je
0
and for a= 0, by substituting K, we get

f (e) = _ 2mZe2
ft2K2
d _ 4m2Z2e4 dQ _ m2Z2e4 dQ
cr- 4K4 -
1i ft4k4 . 2
e.
Slll 2
134. We write the energy of interaction between the par-
ticle that is colliding with the atom and the nucleus of that
23-0H96
atom (the charge of the nucleus being Ze) together with tho
electrons distributed around the nucleus with a density
p (r):
V { ') _ Zee1 _ i p (r") d-t''
r - r' ee t J I r' - r" I
Hence, according to Problem 132, we write
where
To compute I (r") = J
eiKr' d•'
I r' -r" I we note that this_ ~integral
satisfies the equation !J.I (r") = - 4neiKr" and, therefore, if
we set I (r") = AeiKr", we get -K 2AeiKrM= -4neiKr", i.e.
4n
A = 1f2 . Thus,
I t = 4n
K2 '
I 2-- 4n Ji p (r ") eiKr" d•"
](2
and
f (8) = - ~~;:i {Z- F (K)}
where
00
F (K) = 4n Jp (r'') sinKr"Kr" r" dr" 2 is the form factor (1)

0
We set p = p0e-rfa, and from the normalization condition

00
Jp dT = 4np0 Je-rfar dr2 = 4np0 X 2a3 = Z

0
we can compute po:

z
Po =8Jtii3
ANSWERS 355
Substituting this into (1), we get
~ e-r/a
00 eiKr e-iKr
F (K) = 4np0 ~ r 2 dr
2tKr
0
1 .
1 z
)2 -, ·- (1 +a2K2)2
( -+tK
a -
and
d _ 4m2e2ef
a- n4K4
zz [ 1 - 1
(1 +a2K2)2
J dQ2
If we introduce K = 2ksin ~ and nk= mv, we find

Z2e2e 2
dcr = ( mt )z esc~ 28 [ 1- 1
(1+a2K2)2 J' dQ
For fast particles that scatter at large angles, aK :J> 1 and
Z2e2ef ~ 8
dcr ~ ( mt ) 2 esc 2 dQ
For small aK we have 1- (1 +!ZKZ)Z ~ 2a2K 2 and

d "' 16m2e2ef
a '"""' li4
zza~ dQ
i.e. the differential cross section is finite.
135. According to the solution of Problem 132, dcr =
= I f (8) l2dQ, where f (8) is the scattering amplitude in
the form
00
f (8) = - ~~ J si~~r V (r) r dr 2 (1)

0
and K =I k- ktl = 2k sin ~ (8 is the scattering angle).

Substituting the Yukawa potential into (1), we get
;i [J e-<x-iK)rdr- Je-<x+iK)rdr]
00 00
f(8)=- ~~:
0 0
mA [
= - n2Ki
1 1
x--iK- x+iK
J= - h2(x2+K2)
2mA
23*
and, hence,
4m2A2 •
da = !i4 (xZ+ KZ)Z 2:rt Sin 9 d9
The total cross section is
( 2mA )2
xh2 •
SECTION IV
1
1. P=f[·
2• dw = 2dt = _!_ dq> .
T n V !J>8 -cp2
3. c =.!::..
1t
4. dw (x) = V~ e-ax 2 dx.

5. We use Pn (t) to denote the probability of em1sswn
of n electrons in time t, and P 0 (t) to denote the probability
of no emission of electrons in the same time. Assuming (1)
and using the rule for calculating the .probability of two
consecutive events, we get
Pn (t + dt) = Pn-1 (t) PI+ Pn (t) (1 -PI) (1}
P 0 (t + dt) = P 0 (t) (1 - P 1 )
where P 1 is the probability of the emission of one electron
in the time dt. According to condition (2), this probability is
PI = 'Adt (2)
Now we expand the left-hand sides of Eqs. (1) in a
power series in dt and then tend dt to zero. We get
dP;/t) =I..!Pn-dt)-Pn(t))
(3)
dPo = - 'APo (t)
dt
ANSWERS 357
To this system of differential equations we must add the

initial conditions for Pn (t), namely
p 0 _ { 1 for n = 0 (4)
n ( )- 0 for n :#= 0
It is easy to find the solution of (3) with initial conditions (4):
P n (t) = <~r e-'A.t (the Poisson distribution).

6. Using the distribution obtained in Problem 5, we have
'oo
tmz = ~ n P n (t)- [ ~ nP n (t)

n=O
2
n=O
oo
r
00
- f..t [A.tl 'Y, (A.t)n-2 e-'A.t

- 1.::.! (n-2)1
+ LJ
~ (A.t)n-t
(n-1)1
e-'A.t]
n=2 n=1
00
- (f..t)2 re-M ~ (A.t)n-1 ]2 = f..t

L LJ (n-1)!
n=1
00 00
"i.l xn _ :t ~ -
since LJ nf- e . But f..t= LJ nPn (t) = n= n0 t, and so
n=O n=O
11n2 = n0 t.
7. The probability that volume V 0 contains one molecule
is expressed as P = ~. Hence, the probability that any n
molecules from the total N will land in V 0 is expressed as
Pn (Vo) = C'NPn (1-P)N-n
)'
where C']v = n.'<::~ n. is the number of ways in which
we can choose any n molecules from the total N. The obtain-
ed formula
PN<Vo)= <N~~)tnt (~or(1- ~or-n (1)
is known as. the binomial distribution.
We now consider the two extreme cases.
(a) Since n = PN, we have

- - N -
P = lim
n
(n)n (
N->oo n!
1 -...!:.)
N
-- (n)n
nl
e-n (2)
This coincides with the ftnal formula of the solution to

Problem 5 if we remember that "At = n.
(b) Using Stirling's approximation ln nl~ nlnn- n, we
ftnd from (2) that
ln Pn = n ln n- n-ln n! = - (n- ~n) ln ( 1 + ~; )+ ~n
Now bearing in mind that f...n {: n, we get the relationship
for Pn,
lnPn~-(~~ 2
2n
correct to terms of the order of 0 [ ( ~nn )2 J. Hence,
_ <n-n)2
Pn=Ce zn
00
Normalizing this probability to unity, JPn dn= 1, we

-oo
ftnd that
C=-t-
V2nn
The expression obtained for Pn is called the Gaussian
distribution.
8. The given inequality follows from the simple properties
of the integral and the function p (x):
w (x >a)= Jp (x) dx~

a
r
a
p (x) :~ ~dx~ r
-oo
p {x) :: dx~ ~
Here iJ (x) is the probability that the value of the random
variable x lies in the interval [x, x dx]. +
9. We consider the following expression:
11 . . h
2;(e11Jl+e-t1Jl), w ere -n~cp~n
ANSWERS 359
Here the coefficient of ei<P can be interpreted as the proba-

bility of moving one step to the right, and of e-i<P one step
to the left. Clearly, the probability that after t steps the
particle will reach point l will be equal to the coefficient
of eiiPl in the binomial expansion
[ 21 ( ett (1)
Now we multiply (1) by 2~ e-iiP 1 and integrate the equation

obtained with respect to <p from - n to n: +
n
;n ~ [~ (ei<P+e-i<P)Je-i<P 1 d<p=Pt(l)
-n
since
n
_1_ r ei<p(n-k) d<p = 6nk = { 1 . for n=k
2n J
-11
0 L for n=t=k
Hence, the sought probability will be
11
P 1 (l) = - 1-
2n Jf cos' <p X e-i<Pl d<p.
-11
10. By analogy with Problem 9 we get the following re-

sults:
(1) for the square grating the probability that after t
steps the particle will reach point I = {l1 , l 2) is
n n
P t (1) = ( 2 ~) 2 ~ ~ [ !(ei~~'t + e- ill't + ei~~'2 + e -i<P2) J
-11 -11
X e -i<<Ptlt +!P212) d<pt d<p2
(2) for the cubic grating this probability is
Pt(l)=Pt(llo l2, Z3)
11 11 n
= (2~)3 ) )
-n -n -n
J[! (cos <p1 +cos <p2 +cos <ps) J
- 11. Let us consider the totality of trajectories ending at

point I and introduce the generating function
00
Us (z, 1) = ~ Pt(l) zt
1=0
where s is the dimension of the space.

Now we show that this function expresses the probability
that the particle will not return to the origin of coordinates
(the initial point). Let A be an event that can be repeated,
! 1 the probability that event A will take place for the first
time in the jth trial, and u 1 the probability that A will
take place in the jth trial regardless of..whether it occurred
earlier.
We set u 0 = 1 and construct the polynomials
00 00
u (z) = 2j u 1z1, F (z) = 2j f1z;

i=O i=1
It is easy to see that
Uj = UofJ + udi-1 + • • • + Ut,.tfl

Multiplying both sides of the relationship and summing up
with respect to j from 1 to oo, we get
u (z) - 1 = F (z) u (z)

or
F (z) = 1 - [u (z)]-1
But F (1) = / 1 2+ / + ...

is the probability that event A
will take place at some time.
!I"" There are thus two possibilities:
"""-(a) if u (1) = oo, then F (1) = 1. and A will certainly
occur;
(b) if u (1) < oo, then F (1) < 1 and there is a positive
probability that event A will not take place.
ANSWERS 361
In our case u (1) (the probability that the particle will

at some time return to the initial point) coincides with
11 11
u,
r(1) =-- _1_
- (2n)•
r r
J ··· J 1
d<pt .•• d<ps
_ 11 _ 11 1 - - (cos q>t + •.. +cos q> 8 )
s
These integrals for the one-dimensional and two-dimen-

!'ional cases diverge for small values of angle cp;. For the
one-dimensional case this is evident. For the two-dimensional
case this can be proved if we pass to polar coordinates. Thus,
in the one-dimensional and two-dimensional cases the part-
icle in a random walk will always return to the initial point.
-a
(Fig. 59 Fig. 60
In the three-dimensional case the integral can be evaluated

numerically:
111 11 11
u (1) = - 1 -
s (2n)3 Jr Jr Jr 1
.. dept dcp2d!Jla -
~
1 52
·
-11 -1l-11 1-3 (cos lpt +cos cp 2 +cos 'fa)
i.e. there is a certain probability (0.34) that the particle

will not return to the initial point.
12. Since p is conserved, the phase trajectory has the
form shown in Fig. 59.
13. The law of energy conservation gives us the equation
for the phase trajectory:
p2 p2
°
-2m + mgz = ...,..--
0 + mgz~m
The trajectory is the parabola shown in Fig. 60.

14. p = ± y 2mee 1 C- r~ ) •
15. We assume that at the initial time the velocity and

the coordinate of the oscillator are equal to v 0 and x 0 , res-
p
Fig. 61
pectively, and we have

yt
x = e- 2 [ x 0 cos rot +: sin rot J
Since y ~ ro0 , we get
wv = e -.t [_00
Vo cos rot - • rot
x0 sm J
Hence,
(x2 + ~: ) = (X~ + ~~ ) e-YI
The equation for the trajectory represents an elliptical helix

(Fig. 61). The change of phase volume with passage of time
follows the law
r (t) = r r dp dx = r r a(p, x) dpo dxo
JJ J J a(Po, xo)
G Go
1 . t
= e-YI JJ -mvsin rot
Go
cos rot -smro
mv
cos rot
dp 0 dx0 =e-Yif(O).
ANSWERS 363
16. It follows from the law of conservation of energy that

p2
2; +mgL(1-coscp)=Ho
where I is the moment of inertia, and prp is the generalized
momentum. Hence,
pq = + V2I (H 0 -mgL) + 2/mg cos q>
Case (1). The phase trajectory is shown in Fig. 62. It
Pep
-V2IH0 -
Fig. 62
corresponds to a rotational motion in two different directions.

Case (2). The phase trajectory is shown in Fig. 63. We
can see that at point cp = :rt the two branches of the phase
pep
V2IH0 ------
Fig. 63
trajectory come into contact with each other. An ambiguity

appears at this point. But if we calculate the time needed
to reach point q> = :n, we find that it is equal to infinity.

Indeed, since pep = I ~~, we have
1t
t=) ld~ =00
0 (4.mgLI cos2 ~;2)1/2

Case (3). The possible values of angle q> will lie in the
interval [ -q> 0 , + q> 0 ], where q> 0 is determined from the con-
dition prp = 0. The phase trajectories will have the form
pq;
Fig. 64
shown in Fig. 64 and will describe the oscillations of the

pendulum.
17. (1) The laws of energy and momentum conservation
give us the connection between the momenta of the particles
before and after collision:
, m1-m2 2mt
Pt = mt+m2 Pt + mt+m2 P2
, ll'lt-m2 2m2
p2=- mt+m2: P2+ mt+m2 Pt
Hence, the Jacobian
a(pi, p~. ql., q~)
1
a(Pt. P2• Qf, Q2)
i.e. the phase volume is conserved.

(2) The triangle's vertices A (p 0 , z 0), B (p 0 , z 0 +d),
C (p 0+ b, z0 ) will after a certain time t take the positions
ANSWElRS 365
A ' (p 1, z1),w herep 1 =p0 -mgt,z1 =z0 + --;nt-

Po
2gt
2
B' (p2 , z2 ), where p 2 =Po- mgt, z2 = z1 +a

C'(p3 , z3), where p 3 =Pt+b, Za=Zt+~t
m
(see Fig. 65).
Obviously, the area of the new triangle will be
ab
S= 2 =So
Thus, the area occupied by the phase points has changed in
form but not in magnitude.
p
c'
C(p0 +b,zo) _...............-~
~B'
,______ " . . . . . . . . . A
I .,...,"
_J~~~~~~~//~~~~=/~-~r
A(p0 ,Z0 ) B(p0 ,r0 +a)
Fig. 65
18. The probability that the energy H of a closed system

lies in the interval [E, E +
dEl follows from the microcanon-
ical distribution given by formula (IV-4)
ll (H-E)
p (H)= (l (E)
Here 6 (H - E) is the Dirac delta function [see formula (1)

in Appendix 4], and Q (E) is the normalizing factor of the
distribution:
Q (E)= ar
an I
H=E
where r is the phase volume limited by the hypersurface
of constant energy H (p;, q;) = F. From this we find that
3N
I 3N I N :l't-2 .E:!....
(1) r (H)= j ... j d p = SN (2mH) 2
r(l2+1)
366 PRbili..EMS IN TH:MRE'l'ICAL PHYSlCS
where the integration was done over the hypersurface de-

termined by the condition Pr + · 2·~ + P~ ~H [see formu-
la (11) of Appendix 4]. This yields
2 r ( 3~ +t) t-~
p (H)= 6 (H-E) 3N 3N 3N E 2
2m_2_:n:_2_
(2) f (H)=) .2~. ) dN pdNq=HN ( 2::n: )N ~I

where the integration was done over the hypersurface defined
p2+w2q2 P~+w2q~
by the condition 1 2m 1 + . . . + 2m ~ H [see
formula (11) of Appendix 41. This yields
(H)=6(H-E) (N-i)! (~)N.
p EN-1 2:n:m
19. If there is a thermal contact between the system with

energy H (pt, qt) (i = 1, ... , n) and the heat bath with
energy,, H 0 (pi, qi) (i = 1, ... , N ~ n), which together
form a closed system with energy li + H 0 = E, we can
easily find the probability density for the ensemble from the
microcanonical distribution. According to the law of addi-
tion of probabilities, we get
p (H)= j p (H, H 0) dpi dqi
But p (H, H 0 ) is given by the microcanonical distribution

(H H)= 6 (H+H 0 -E)
p ' o Q (E)
So
(H)= Q 0 (E-H)
p Q (E)
where Qo (E- H)= aj)Hfo

0 Ho=E-H
I
and r 0 is the volume limited
by the hypersurface of constant energy for the heat bath,
H 0 (pi, qi) = constant. Using the results of Problem 18, we
get for the two models the sought canonical distribution.
3N
- 3N
(1) Qo (E-H)= 3N (2nm) 2 (E-H)-2--1
2 r ( 3; +1)
We tend N to infinity but set
lim ~=~kT
E-+oo N 2
N-+oo
Then
3N 2!:._ 2!:._ -1 3N _ 1
- (2mn) 2 E 2 H ,2
p (H) = 1im 2
N-+oo
g (E) lim ( 1 -
N-+oo ~NkT)
2
The first limit is finite since
ar
Q(E)= o(H+H)0
I
H+Ho=E
~Qo(E)
Therefore,
3NkT
p (H)= constant X)~~ [ ( 1-

2H
3NkT
)2H JH/kT
=constant X e-H/(kTl;
(2) The solution in this case is similar to case (1) but requ-
ires that
The formulas obtained coincide with (IV-7), which can be

derived in general form regardless of the specific model of
the heat bath.
20. For a perfect gas
N N
H= ~ [e(p 1)+U(rt)l= ~ Ht
i=1 i=1
where e (pt) is the kinetic energy of an individual molecule,

and U (r;) is the potential energy of interaction of a mole-
aGS Pit0:13LEMS IN 'l'ltEOitE'l'!CAL PltYStCS
cule with the walls of the vessel:

U(rd=O for r 1 c: V1
= oo for r 1 (/:_ V
Then
N
- ~ H;JkT
Z= J~:.; J e i=i dNpdNr
= VN [ J~: Je- ::r df JN = VN (zi)N 1
Denoting z1 == f (T), we find the Helmholtz free energy

F = - NkT In V - NkT ln f (T)
and then the entropy, pressure and average energy, respec-
tively:
S = Nk In[/ (T) V]-!- NkT a 1 ~~ (T)
= NkT E=NkT 2 olnf(T) 1
P v I arj ·
p~
21. (1) Assuming that Hi= 2 ~ 1 we get [see formula
(13) of Appendix 4]:
s12 NkT
zi = (2nmkT) I p = -v-,
3 3 3
S=Nkln V +"2 Nk[ln (2nmkT)+ 1], E = 2 NkT 1 Cv= 2 kN.
(2) The energy in this case is
H _ P~ _1_ [ ~ Pfcp J
I- 2M + 2f:lr3 p,a + sin2 8
z 1 = (2nM kT) 312 8n2 r~~-t kT =A (kT) 312

1- + - 1- 1 A== 8n2 r~f.t (2nM) 8 \
where M = m1 + m21 - f.l.1- = - mt m2
and r0 is the distance between the atoms in the molecule.
Then
NkT 5 5
p = -v- 1 S=Nln(VA)-!- 2 Nk1nVkT+ 2 Nk 1
5 5
E=-sNkT 1 Cv= 2 Nk.
ANSWEM 369
(3) For the approximation of small vibrations we have

P~
H 1 ~2M 1 f 2 PlQl P~a l U ( ) "Y ( 2
+ 2ill Pir + r2sin2 e+----;:2_J + ro +2 r-ro)
where the potential energy of the atoms is represented as
U (r)- -u (ro)+:rtazul
arz !'=ro(r-ro) z+ ...
azu
{)r2
Ir=ro = V
At low temperatures we can easily estimate the classical
partition function by bearing in mind that the integrand
in (IV-10) has a sharp maximum at point r 0 • Calculating
the energy from U (r 0 ), we find that
Zt = (2:rtM)aJ2 4:rt (2:rtf.t/J2 (kT)3 r~ V 2n"YkT = B (kT) 7'2
where B = (2:rtM) 3/2 4:rt (2:rtf1) a12 r~

V Y2n .
r-
Th en
NkT
p =-v-, S=Nkln V · B+ 27 Nkln(VkT)+:rNk,
7
E= ~ NkT, Cv= ~ kN.

22. In both cases p = N;T since the gases are perfect,
N
andH=LiHt.
i=i
oo apl
-- 4n
r (1.)
l
(a) Zt = 4:rt ) e hT p2 dp = -l- a3f2
0
Hence, E = j 3NkT. In the given case

l
1 ~-~E·
- 3V '
oo c Vm2c2+p2
(b) z 1 = 4:rt ) e- hT p 2 dp. Substituting me sinh t
0
for p, we reduce z1 to
00
z 1 = 4:rtm3 c3 .\ e- zo cosh t sinh 2 t cosh t dt

0
24-01496
370 PiWBLEMS TN THEORETICAL PHYSICS
mc2
where z0 == -;?F . Hence,
·zi = .±:::._ m3c3 [Ko (z 0 ) + ~ K1 (zo)]

zo zo
where K 0 (z 0 ) and K 1 (z 0 ) are modified Hankel functions of
order 0 and 1. The system's average energy is then
- Ko (zo)+ ( z~ +~) Kt (zo)-
E = NkT 1 + 2 2 zo .
Ko (zo) + - Kt (zo)
zo
Using the asymptotic expansions for the modified Hankel
functions, we see that for kT ~ mc2 ,
3
E = N mc 2 + "2 N_kT.
23. Let the minimum value of the system's energy be

equal to zero. Then
""
Z(~)= e-flE Q(E)dE j ( 1)
0
If we assume that E >0 and that for E < 0, Q (E) = 0

we find
<J+ioo
Q (E)= - 1-.
2m JI eflEZ (~) d~
<J- ioo
because it follows from formula (1) that Z and Q (E) are

connected via the Laplace transformation. Now, using the
Cauchy theorem for the derivative, we find
<J+ioo
n (E) __f_ I ~ {3E dR _ AEN- 1 1
~' - 2ni J 13N e p- (N-1)1 '
0-ioo
24. p = C exp { -
II (vi, ri)
kT + 2kT1 ,
~ m [Q X ri) 2
}
,
i
where vi and ri are, respectively, the velocities and radius

vectors of the particles in the revolving system of coordinates.
ANSWERS 371
25. Using the result of Problem 24, we get

2 kT mQ2R2 .
F = F 0 - N kT ln [ m~J 2 R 2 (e ZkT - 1) J
Recalling that dV = 2nhR dR, we find
N kT U (R) e-U/kT m!J.2R2
p = - -V- -----;;r- e-U/kT -1 ' lj =- 2
26. We consider the following total time derivative of the

sum over all the particles in the system:
N N N
d fJT •
dt ~ r;p; = ~ Pi api + ~ r;p; (1)
i=i i=1 i=1
Next we fmd its time average

t
fit= lim+ J
l-+00 0
F (t') dt' (2)
Now, if the particles move in a finite region and with ve-

locities that do not turn into infinity in this region, the
time average of the left-hand side of (1) will give zero
(t-+ oo). For the terms in the right-hand side we use Euler's
theorem on homogeneous functions and get
N ,-.-/
~ fJT ~
LJ P; ap; = 2T = 2T0
i=1
where T 0 is the system's average kinetic energy.

Now let us find the time average of the second term
N N
~ r;p; = ~ r;F;
i=1 i=1
where F; is the force acting on the ith particle. The sum

N
~ r; F; is called the virial in mechanics.
i=1
Let the particles of the system move in volume V. Then
a force -np (t) will act on the particles from each unit area
(n is the external normal to the surface that encloses volume
24*
V and p (t) is tho pressure at the given time that the particles
exert on the unit area). Since, besides this, forces of inter-
action F;i = - Fn act between the particles, we get
N N r----'
,.....J ~· ,-..J .
~ r-F·= - · nrpdS- '51 r- - = -p0 ·3V-nU0
au
L.J ' ' '. ..:::..J ' ar i
i=1 . i=1
since
~~ nrP{t) dS = p(t) ) ) ) div r dV = 3p0V

s v
where p 0 is the average pressure in the system, and U 0 =

1 ,---__.1
= 2 ~- U (r1 , ri) is the average potential energy of inter-
t, 1
action between the particles.
So in the final analysis we find that the pressure in the
system is
where e0 is the average kinetic energy per particle.

r(~+,;)
27. Hn = (kTt 2N [see formula (14) of Appen-
3
r(-2 )
dix 4]. We can easily find now that
28. Let
Then, using the theorem on addition of probabilities, we can

find the probability that the velocity of a particle will be
ANSWERS 373
in the interval [v, v + dvl. So for case (1) we find that
N
~l mvi2
dp (v) =e
---
mv2
r
2kT J
oo
:._·c.: j
re •=1
. -
:...J
--
2kT
d
N-1
v
U(r, ... , rrv)
X J:._: J e- kT • dNr dux duy duz
= Ce-mv 2 i(ZhTldux duy duz

where C = ( 2:kT)
312
is found from the normalization con-
dition
00
J~ J
-oo
dp (v) = 1
Recalling that dux dvy duz = u2 sin 2 8 du d8 dr:p and s = nw2/2,

we get for case (2)
dp(u)=4n ( -m- ) 3/2 e-mv 2 /(2hTlu 2 du
2rtkT
and for case (3)
dp (s) =
V rt (kT)3
2 e-e/kT Vs ds.
29. We follow the general rule of finding the mean and
get
oo mv2
n=
v
(__!!!:_)3/2
.2n kT
4 \ 2- ZhT
n j u e
n+ d =
u
_4_ ( 2kT )n/2 r (n+3)
Vn m 2
0
Now we use the value of the gamma function [see f~rmulas

(5) and (6) of Appendix 4] and find that
v= ..V/ 8kT
nm'
v2 = 3kT
m
A..nd from the condition :u (e-mv 2 /(2kT)u2) = 0 we find that
Uo
· = -.V/ 2k.T .
m
30. It follows from the formulas of Problem 29 that

- mV2 3 mv kT
e=-2-= 2 kT, e0 =-f-=- 2
2-
0bviously, e
=I= en, although the following relationship
holds for Lhe energy of the entire system:
H 0 ~H E =
This relationship can hold only if there are a great number
of particles in the entire system.
31. We write the Maxwell momentum distribution and get
cV~
dp (p) = Ce kT
The normalization condition Jdp (p) = 1 yields
C= { 4n (mc) 3 [ 2 ( ~~2 ) K1 ( ~~~2 ) + !~2 K 0 ( ~~ ) J} -t
where K 0 (z) and K 1 (z) are modified Hankel functions.

m ) 3/2 -~(v-u)2
( ZnkT
32. dp (v, u) = e 2kT dvx dvy dvz.
33. We determine the probability that the first particle
has a velocity v1 , and the second v 2 • The canonical dis-
tribution yields
mtV:+m2v~
dp (vt. v 2 ) = ce 2kT dv 1 dv 2
To find c we use the normalization condition, and we get
m1vi+m2v~
d p (v 1, v2 ) = 16n 2 ( 7ntm2
4112k 2 T 2
) 3/2 e 2kT 2 2
V1 V2
d v1 d v2
Now we change variables, v' = v 1 - v2 and v 0 = mv tm v

1 1
mt
2 2 ,
m2
and using the fact that m 1m 2 = f.tM, we get
U.v'2
dp (v ', ) 4 ( 2nkT
Vo = n
!1 )3/2 - ''kT '2d '
e ~ v v X
4n ( 2nkT
M )3/2
Mv8
X e- 2kT v~ duo
1 1 1 )
( ftf=m 1 +m 2 , -=-+-
ft m~ m~
ANSWERS 375
We integrate this expression with respect to the centre-of-

mass velocity v 0 and thus find the relative velocity distribu-
tion of the particles:
Mv'2
dp (v')=4 :rt (~.
2nkT )
312 -
e 2hT v'2d'
v
v'=V2v
v
where is the mean velocity of the particles of the gas.
34. To characterize the collision of particles we introduce
the total scattering cross section, a, which is the ratio of
the probability of the given collision in unit time to the
flux of the particles per unit area. If the mean number of
particles in unit volume is n, the average number of colli-
sions of one molecule with all the others per unit time will be
v = ) a (v') nv' dp (v')
For our case a= 4:rtR~, then
v=4:rta Y kf
-Jtm
Hence,
A,=~= _ _1_-;-=--
v 4nnR5 V2
Vn
2
err (1)- .i_
e 2
35. , where erf (x) = ~ 1 :;:;. x
1f Jt +.i_- V Jt erf (1) V "
2 e 2
X
X Je- 12 dt is the error integral [see formulas (15)- (18)

0
in Appendix 4]. But erf (1) = 0.84, and so n2 /n 1 = 0.4.
36. The total scattering cross section in this case is
a= :rtD2 [ 1 + D~m )2 J
where D = 2R 0 , and v' is the relative velocity of motion of
the particles. We avera~e this expression with respect to
v' and get the Sutherland relation
cr = nD2 ( 1 + ;)
a
where a = Dnk •
37. v = n0 au V2. Here n~ = : , v is the mean velocity

in the two-dimensional case, N is the total number of
particles on the surface, S is the surface area.
38. Since the atoms move with different velocities, an
observer will see light of all wavelengths due to the Dop-
pler effect. For example, if the atom is moving away from
the observer with a velocity Vz and the observer is on the
z-axis, the light will seem to him to have the following
wavelength:
And so the intensity of light seen by the observer in the

interval from 'A to 'A + d"A will be
Jd"A =a dn ('A)
where dn ('A) is the number of atoms in the entire volume
that radiate light with wavelengths lying in ['A, 'A + d"A],
and a is a constant that is determined by the condition
) J("A)d"A=NJ 0
If we assume the Maxwell velocity distribution to be valid
in our case, we have
dn ('A)= dn (vz) = N ( 2;;;.T )112 e-mv z/(2kT} dvz
2
mc2
2 ) 1/2 ---(A.-A.o> 2
- N ( me 2kTA.2
0
d"A
- 2A8nkT e
since
Vz= { 0 ("A-"Ao)
Then
J ('A) d"A = ·:xN 1 e-<A.-A.o} 2 /0 2 d"A
Vn6 2 ·
ANSWERS 377
where 8= -,!"
2 k~:~3 is the Doppler half-width of the spect-
ral line. We now find a:
00
= I
NJ
0 v 62
aN
:rt
j0
e-('A.-'Ao)2J62df..
00
~ ----===-
aN ~ e-('A.-'A.ol 2/0 2 d'A =aN
Vnoz -00
a=Jo
where the lower limit of integration has been changed to
-oo since the integrand for 'A < 0 is practically zero.
Fig. 66
Thus, the spectral density is represented by the Gaussian

distribution (Fig. 66):
J ('A)= VJo~ e-('A.-'Ao)2f62 •
:rt6
39. If we assume that the energy of a free electron inside

the metal is less than its energy outside the metal by the work
function e
The value of Vox can be determined from the condition that

the electron leaves the metal, i.e.
mv3x
-2-=e<p
Then
• - noev -e<J!/(hT)
Jx- 4 e
which is the classical Richardson equation.

41. Integrating the canonical distribution with respect
to momentum, we find the sought probability:
U(x, y, z)
dp (r) = Ce hT dx dy dz
The number of these particles in unit volume is
dn (r) = n 0 dp (r)
where n 0 is the average number of particles in unit volume.
This is the Boltzmann distribution.
42. By definition we have
00 mgz
~ ze -liTdz
0 kT
Zo = oo mgz mg
~ e-liTdz
0
since the Boltzmann distribution yields

mgz
-liT
dm(z) o: e
43. The centre of gravity for one type of particles is

ANSWERS 379
kT
where z0 11. = - - . The centre of gravity of the entire sys-
mkg
tern will be
l
~ Nmhzh l
11.=1 = lkT __1_ ~ hmn
Zo = - . , - I - - - gM M L.J h
~ Nml!. h=1 e zoh -1
11.=1
(M= ~m~~.).
It
45. M = 0.41 2 •2:::T

46. We introduce spherical coordinates by directing the
z-axis along vector u (Fig. 67). If the particle collides with
the sphere at point A with a rel-
u
ative velocity v, we get I v; I =
= I v' I and y = y' (an elastic col-
lision). It follows from this that
the sphere will experience "resis-
tance" from the impact with one
particle equal to
-2mv' cosy cos a (1)
since the sphere receives momen-
tum equal to 2mv' cos y directed
along OB. But
cos a = cos 8 cos y Fig. 67
+sin 8 sin y cos ~ (2)

(the law of cosines for the spherical triangle ABC).
The Maxwell distribution in the given coordinates is
dn (v') =n ( 2n~T ) 3/2
x exp { - 2;T (v' 2 + u 2 - 2v'u cos 8)} v' 2 sin 8 dv' d8 dcp (3)
and the number of collisions in unit time, with point B (y,
~) lying in the solid angle differential sin y dy d~, is equal
to the number of particles in volume
R 2 sin y dy d~ v' cos y (4)
(an area element multiplied by path per unit time).
Keeping in mind the symmetry of the problem and using

relations (1)-(4), we get the total force of resistance
oo n
F = - 2mn ( 2 :r ) 312
R2 ~ v' 4 dv' ~ 2n sin 8 d8
n/2
X exp [ ·- 2;T (v' 2 + u2 -2v'ucos8) J~ cos y sill'y dy
2
0
2n
X ~ d~ (cos 8 cosy 1- sin 0 sill'\' cos P)
0
00 -m (u-v')2
=2 n 2R2m (.....!!!_)3/2(kT)2
2nkT
\ '2[(!!!.!!.. '-1) e
muz J v . kT v
2hT
0
m (u+v')2
+e 2kT ( :; v' + 1 ) dv'J
Using the relationships (15)-(18), Appendix 4, for the
error integral, we find the final form for F:
F = - nkTnR2 [ :~ b : (1 11 + 2b2) + (2b2+ 2- 2! 2 ) (b)]

b
where b = V u and (b)= V2_ I e-x 2 dx. The order

mj2kT n ~
of magnitude of b is determined by the ratio of the velo-
city of the sphere to the mean velocity of motion of
the particles. We consider the case for b ~ 1. The expansion
2 ( b3
<D(b)= Vn b-3+ ... )
then gives
F= -6naRv
where
CX=~nR .. / mkT
9 V 2n
We compare this expression with Stokes' law for the force
gf resistance experienced by a sphere moving in viscou~
liquid:
F = -6nlJ Rv
where lJ is the coefficient of viscosity, and we find that for
small velocities of the sphere these formulas have a quali-
tative similarity.
eo --
47. N = ~
0
dN (9) = S :v -,; ~! R2~ h2 (see Fig. 46).
48. Using the Maxwell distribution, we find the number

of molecules that pass through the aperture S 0 in time dt:
oo oo mv2
- dN = dt n 0S 0 ~ Vx dvx ~~ e- 2 kT dvy dvz ( Z~T ) 312
V - d t = - 8 ykT
0 -co
kT p
=n0 S 0 2nm
-dt kT
0
2nm
The velocity of outflow of gas will be

dN pS 0 -
V= -dt= 4kT V.
49. Keeping in mind the relationship As=kT( 8 !~sz)T,

we find that the mean dipole moment is
p = _ kT ( aIn z )
iJE T
where
Hi
Z=(z 1)N, z 1 = Je-'kTdr, Ht=Hot-(Po·E)

Here H 01 is the energy of one particle without the field.
Hence (see the solution to Problem 21),
p 0 E cos e
Zt = qJ (T)
l'l
f
e hT sine de
where IJl (T) is a known function of temperature. Then

P = N Po [coth PoE -
kT
_!!!._]
p E
0
= N p L ( PoE
kT
)0
where L (x) = coth x --..!... is the Langevin function.

X
382 PRODLEMS IN THEORETICAL PHYSICS
In the case of high temperatures or weak electric fields

PoE {:: 1
kT --;::,
Then expanding L ( P~: ) in a power series
l (~)
J kT
= _.!_3 PoE +O [ ( PoE ) 3 ]
kT kT
Fig. 68
we got a linear dependence of P on E (Fig. 68):

p
= 31 N po ---;;r
p0E
The polarization is then
p = ~ = ~~ P!: = ~E
where ~= ! ~ :f is, by definition, the polarizahility
of the gas. The permittivity is then
e = 1 + 4n~ = 1 + ~ n:;~ (n= ~ ) .

50. e = 1 + 4n (net + ! ';:} ).
51. Choose the z-axis along the vector of magnetic induc-
tion B of the external field, i.e. B = 0 · i 0 ·j B 0 ·k. + +
The vector potential will then be
A= O·i +B 0 ·x·j + O·k
ANSWERS 383
and the Lagrangian will be

N 2
X(r,v)= "LJ' -:l--U(r,

m;vi " ' e·ek
. .. ,rN)+ LJ ;ik (vi·vk)
i=l "*i
Hence,
oo 3N
Z= ~ . . . ~ oxp { - k~ [ ~ :,~ ;k- L J} dN rdN p

-00 h=l
oo 3N
= Jf .. · J\' exp { - 1 [
kT ~ or"
iJL •
r11 - L
J} Dihd Nrd N v
-00 1!·.~1
where the determinant

D. _, iJPtx
th- OVtx ' · ' ' '
DPtx
dVNz
1-1
-
iJ 2L
OVtx OVJx '
iJ2L
' , , ' 8Vtx iJVNz
I
does not depend on B because in the terms with B, X is linear
in v. The exponent (the expression in braces) does not de-
pend on B either, because terms with B will cancel out.
Hence, the magnetization vector along the f1eld is
Mz= -kT iJlnZ =0
OB
which means that a classical system cannot be magnetized
stably. This appears paradoxical because a "classical" ex-
planation is often given for dia- and paramagnetism. This
"classical" explanation assumes the existence of stable
electrical systems with stable electron orbits. But this is
the very thing that is not explicable from the standpoint
of classical physics .
.. / TtTc
52. X= V 4nnoe2 (T!-Te) '
53. The Hamiltonian in this case has the form
p2
H=H(pi> qi)+ 2; +Mgz
where PM and z are the momentum and coordinate of the

piston. Since the relationship Mg = pS must always hold,
the canonical distribution takes the form
p2
H(pi> q1)--m--pV}
{
p=cexp - kT
We find the classical partition function for this case:

00 pfu- 00 pv ].!!__
Z= J e- 2MkTdPM Je-liTdV(2nmkT) 2
yN
-00 0
= N! (2nm)_2_
3N+1
v! _ 5N+3
(kT)_2_p-<N+1)
and the mean volume of the system is
V = ji V p (V) df = ( oF ) = N + 1 kT
op T p
This relationship determines the equation of state when

the volume varies.
54. Hint. Use the property of additivity for the Hamil-
tonian.
55. We consider an arbitrary mechanical quantity
U (p;, qi, a8 ). Differentiating the mean value of U with res-
pect to T 1 we get
afJ 1
-
aT
= --(U-U)(H-H)
kT2
We set U = H and find that
-- ! -H lq ·=+oo f-H
56. f aH
oq;
=- kTfe-kT I
q;=-oo
df' + kT
Jr e-liT_!1_
oq;
dr
= kT ;;t 1 where dr' = dp 1 ••• dpN dq 1 ••• dq1-1dqt+1 ... X

X dqN. We prove the second equality in a similar way.
ANSWERS 385
57. For small densities the partition function of a real

gas is
Z = Zperfect (1 + ~ ~ ~}
BUovo 1 4:rtr~
where~= -8v 0 +~, and v0 = 8 - 3 - is the volume
of one molecule.
We denote n=N!V and b=4Nv0 , and we get
E-E nbUo
- o- ( uk; }
1+nb -1
( nb U 0 )2 1
Cv=(Cv)perfect-k ~ [
_1+nb
( Uo
--w-1
)J2
As temperature falls, the heat capacity falls too.
58. The critical point is determined by three equations:
(p- ; 2 ) (V -b)=RT
(!~)T=O, !~=0
From this we find Per• Tcr• Vcr and get the sought relation-
ship (the reduced van der Waals equation).
59. The partition function of the system is (see the solution
to Problem 21)
00
zi = 4nV (4n2M~-t) 312 (kT) 3 \ r2e-UikT dr

0
Considering the anharmonicity of atomic vibrations, we get
U (r- ro) = 21 azu I 1 a3U j

iJr2 r=ro (r- ro) + 6 fJr3 r=ro (r- ro)
2 3
+ ''41
""
iJ4Ul
-d 4
r r=ro
(r-ro) 4 + ...
= '\' (r-r 0 )Z+ a (r-r 0 ) 3 +~ (r-roY' + ...
where a, ~. '\' are the respective derivatives. If we consider
the anharmonic addition small, we can calculate the integral
25-01496
MB Pl:WllLEMs IN 'l'l:IEoim'rlCAL PHYSlcS
approximately by expanding the exponential function with

the anharmonic terms in a power series
00 -'\'X~
-A (kT)3 ro2 jf eli:T

Zi-
[t - ax3
kT -
~x4
kT
-t-2 (kT)Z
az
X
s + Jd
••• _ X
-00
=Ar; - ( ; (kTf 12 [1 +BkT]
( A== 4nV (4n2Mf1) 312 ' B= ~ ~ _l_

8 y3
L)
4 yZ
Whence,
E = ~ NkT + Nk 2TZB
Cv= ~ Nk+2Nk 2BT= (Cv)pertect + Cv
So the correction to the heat capacity is Cv = 2Nk 2 TB,
i.e. proportional to temperature, and therefore significant
at high temperatures.
60. The atoms in the molecule are at some mean distance
r 0 from each other. We determine the coefficient of linear
expansion as follows:
X
al=--
r0T
where x is the mean displacement of an atom from the
equilibrium position. But
00
~ xe-UfkT dx
-oo
X=oo
-----
~ e-U/kTdx
-oo
and
) z
lj X -
( )-
A --
(-
rA
2
B )
rg + r33 -
26A- -7B
(-
2
- X
rA r8
where r0 =;Y2B;A is determined from the condition

-au 0 s mce
0. = .
r
· at t he mnumum
· · · t -azu
pom 0 2 r
I
~~
> 0 , th e
ANSWERS 387
. 13A 2B
coeffictent ---ri""--
ro
6
'o
> 0. Then
oo '\'X2
Jr e -kT ( x+ &x3 )
kT dx
~kT_§_
4 1'2
-00
And we get
X
a- zr- - -34 -1'2
0T -
l\ k
- r-
0 •
1
(n+1) anTI ~
61. Q (E) ex exp -'----'---'-:-1-- En+ 1 •
knn+1
63. In accordance with the assumption we write S =
= f (w). Let the system consist of two independent sub-
systems. Then according to the general properties of entropy
and probability we get
s= sl + s 2 = f (wl) + f (w2)
f (wl) + f (w 2) = f (wl • W2)
Hence,
S = f (w) = constant X In w
To determine the constant we must apply this relationship
to a perfect gas. The constant proves equal to the Boltzmann
constant k.
64. dQ =dE+ p dV- (dP·E).
65. Let the functional dependence of A on B and C be
expressed by f (A, B, C) = 0. Then
(_!!_)
oA B, C
dA+ (~)
oB A, C
dB+ (!L)
oC A. B
dC=O
If A is constant, this becomes
( :~ ) A, C ( :~ ) A= - ( :~ ) A. B
25*
388 PlHYBLEMS IN THEORETICAL PHYSICS
i.e.
( :~ ) A= - ( :~ ) A, Bl (:~ )A, C
Similarly
( :~ ) B = - ( :~ )B, cl (;~ ) A, B
( :~ )c=- ( :~ t, c/ ( :~) n, c
Multiplication of these three equations yields
Interchange of A and C in the second of these equations

yields
( ~~ ) B = 1/ ( :~ ) B •
66. Selecting A = p and B = V, as in Problem 65, we get
a= P~'V·
67. Hint. We determine Pen Ter, Ver as in Problem 58
and introduce the variables
:rt = - p - '
T v
't=--'
Ter
ffi=--
Ver
Per
For the Dieterici equation of state we have
2 ( 1 -1- )
n (2ro -1) = 'te nro
For the other case
( :n: + ), 3 } ( 4ro - 1) = 5't.
68. TB= b~ •
69. The speed of propagation of sound waves in a gas is
v=V(~~}s
ANSWERS 389
Considering the relationship
( op ) Cp ( op )
Tv s= Cv W T
(which will be proved in Problem 75) and the relationship

Vp = It (~t is the molecular weight), we get
V= V
But for the van der Waals gas
v- ft~: (!~ ) T
( 8p) 2a RT
W T=-ya- (V-b)Z
Then
70. HM-v=constant andy_~;.

T
71. (V- b)R exp ( - ~ i: dT) =constant.

0
72. CE-Cv=
TEZ ( Be
4ne iJT
)2 >0
Hint. Use the relationships
dQ = ( :: )a dT + [A+ ( ~! ) Jda
T
1
dA 0 = - 4n (E·dD), D=e(T) E.
73. By definition Cv = T ( :: ) v. Hence,
(
0~; )T =T 8~ [ ( :~ ) v JT = T ( :~ ) V
For the van der Waals gas,
iJZp
iJTZ = 0
i.e. Cv does not depend on the volume.
74. For the thermodynamic engine working according to

the cycle shown in Fig. 48,
1']=1--1-
e'l' -1
where '\' = ~: . For the one working according to the cycle

in Fig. 49,
1 1 p'~'-1
1']= - l'ev-1 p-1
Hint. Use the adiabatic equation Tvv- 1 = constant and

8
1 T ,. 2
~
lf,-!J& \ 4 T-dT
• \ J
e
Fig. 69
the expression for the efficiency of a thermodynamic engine

A
'l']=Q
where A is the work done in the cycle, and Q is the heat
liberated in the same cycle.
75. Hint. Use the properties of J acobians [see formula (26)
of Appendix 4] and the relationship (see the solution to
Problem 80)
Cp-Cv= -T,
( av
:~
r)P
(ap T
76. We consider the reversible galvanic cell and we let

it work in the Carnot cycle (Fig. 69). At first the cell works
isothermically (1-2) with a constant e.m.f. ~. then it works
adiabatically (2-3). We then send a current from an external
source through it, and the work is done isothermically (3-4)
with a constant e.m.f. ~ - L\~ and then adiabatically (4-1).
ANSWERS 391
On isotherm 1-2 the cell is heated by an external heat

source with Q1 = E 2 - E 1 +
W 1 , where W 1 = e~ is the
work expended to transfer charges isothermically, and
(E 2 - E 1 ) is the change in the internal energy of the cell.
That change is equal to the energy release -qe, where q is
the energy release ~per passing charge. Thus
Q1 = e1t- qe
On adiabat 2-3 the e.m.f. of the cell will diminish by
~~ (and the temperature by ~T). Therefore,
~A e~~
f]=Qt= (~-q)e
where ~A = e~~ is the work done in the cycle equal to the

area 1-2-3-4-1. But for the Carnot cycle
T-(T-~T) ~T
rt= T =r
Solving the last two relationships simultaneously, we get
the Gibbs-Helmholtz equation for a galvanic cell:
~=q+T ( ~; )v·
77. ( .aT) = a(T, E). a(T, V) = p-T (% )v
av E a (T, V) a (V, E) Cv '
78. The internal energy of the plasma is
E = Epertect + E1
where Epertect = CvT is the; average kinetic energy of the
plasma (the internal energy of a perfect gas), E 1 =
= ; Nerp+ (0) - ; Nerp_ (0) is the average energy of
the electrostatic interaction between the particles of the
plasma, and fP+ (0) and fP- (0) are the potentials created by
all the charges except the given one at its point of location.
Let us determine these potentials. The charge density at a
distance r from the fixed charge will be
p (r) = e (n+ - n_)
3 92 PROBLEMS IN THEORETICAL PHYSICS
The potential <p is determined from the Poisson equation

with consideration for the Boltzmann distribution:
d<p = 4nen 0 [ exp ( ~~ ) - exp ( - ~~ ) J

where n 0 = ~ , and N is the number of particles of one
kind in V.
If temperatures are high, zj: ~1, and thus
8ne2n0 )
d <p = x 2 <p ( x 2 = ---w-
The solution of this equation has the form

. Ct
<p (r) = - e-'X.r
c2 exr
+-
r r
Here <p (r) is the potential at a point that is at a distance
r from the given charge, the potential being· created by all
the charges (including the given one). We require that the
potential at infinity and at points where the charges are
located be finite, and we get C1 = e and C 2 = 0. Hence,
<i'+ (0) = -ex, cp_ (0) =ex
2 • /8ne2N
E=Epertect-Ne V VkT •
79. Using the general relationship E = - T2 :r (: )v,

we find
2 2 -. /8nNe2
F=Fpe..tect-3Ne V VkT
Hence,
_ NkT 1 N 2 -. /8nNe2
P- --y-3 e V kTV3
_ 1 2 -. /8nNe2
S- Sperfect- 3 Ne V kT3V
1
Cv=(Cv)pertect+2Ne 2 • /BnNe'!.
V 3kT3V'
ANSWERS 393
( ap)2
80 () C =T o(S, p) • o(T, V) =C -T 7ff v
'a P iJ(T,V) iJ(T,p) v (!E.)
av r
(b) C
v
=T 0 (S' !2._ • 0 (p, T)
iJ (p, T) iJ (T, V)
=C
P
+T
[see formula (29), Appendix 4].
81. Hint. Use the first law of thermodynamics expressed
in terms of p and T.
82. Hint. Use the results of Problem 64.
R
83. Cp-Cv= 2a(V-b)2.
i - RTV3
Cp-C
85. pVn=constant, where n= Cv-C is the coefficient
of polytropy.
86. S = S 0 + Cp ln T - a V 0p = constant.
87. Cv ln T + ap 0 V = constant.
88. It follows from the first law of thermodynamics,
dQ = dE + p 0 dV, that in the first case dQp = d (E + p 0 V),
and in the second case dQv = dE. Therefore,
dQp- dQv = p 0 dV
Hence,
Qp- Qv = RT (n 2 - n1)
where n1 and n 2 are the numbers of moles of the reacting par-
ticles before and after the reaction. For the reaction H 2 +
+ ~ 0 2 ~ H 20 we have n1 = 3/2, n 2 = 1 and
1
Qp-Qv= -2RT.
T-T 0
89. 1] = T
90. Hint. Use the second law of thermodynamics for re-

versible processes.
Cp
91. (~S)p=___,...(=,...,av,...,-) ~v.
T aT p
92. Let the isotherm intersect the adiabat twice, at points

A and C. If we examine the closed cycle ABC (Fig. 70),
we find that ~ pdV =1= 0. On the other hand, the entropy
Fig. 70
of the system in states A and C is the same, i.e. SA = Sc.

Hence,
A= Q = ~ T dS = T f dS = 0
A~lC
The solution of this contradiction lies in the fact that an
isotherm cannot intersect an adiabat twice. The principle
that follows from this is that thermodynamic states resulting
from an initial state isothermically cannot result from the
same state adiabatically.
93. Hint. Use the first law of thermodynamics.
94. Cp = C v for t = 4 °C.
95. The change in the pressure of a perfect gas with alti-
tude is determined by the relationship
dp = - pg ·dh = - CIJ. ..J!._ dh
R T
ANSWERS 395
where the molecular weight of air 11 = 29 kg/kmol, and

g = 9.8 m/s2 • Since the process under consideration is adia-
batic, we have
dp V dT
-p=y-1 Y
These two equations yield
!!._= _ v-1 ~(
dh '\' R y
=2)
Cv
Since y = 1.4 and R = 8.314 X 103 JK - 1kmol-1 , we

get
:~ ~ -10 Kkm-1
(The actual mean gradient is about -8 Kkm-1 .)
96. The condition for mechanical equilibrium found in
Problem 95 holds when the temperature varies. Now, will
this state of atmospheric air be stable if due to convection
the temperatures of different layers become the same? To
find the answer we will consider two volumes of gas of unit
mass situated at heights h and h +
dh. Let us assume these
two volumes change places. If their total energy increases as
a result, the earlier state of the system will be stable despite
convection.
The change in the energy will be
~E + pLlV
since both temperature and pressure change with altitude.
Hence, the condition of stability will have the form
dE dV
. dh +Piih~o
But
dV R dT V dp dp
P(ih=)tlih-Cih• Th= -pg
pV
E=CvT=--
v-1
From this we get the condition of stability of atmospheric
air in relation to convection:
!!..._";:::;:: _ IlK (y-1)
dh :::-- Ry •
£2 ae
97. Q = Bit T aT •
99. Using the expression for F, we get
S =- (2f...)
ar v =So+ VE2
Bn
(!!:.._)
ar v
4nnp 2
But e = 1 + akT 0 (see the solution to Problem 49). Hence,
S=S _ VE2(e-1)
0 4nT
where S 0 = - ( a;o )v, and
E = F + TS = E 0 - V EZ ~~- 1)
where E 0 = F 0 + TSo.
100. For black-body radiation with an energy density
u the radiation pressure is p = u/3. Now we apply to this
radiation the relationship T dS = dE p dV, which +
easily takes the form
T ( :~ )v= ( :~ )T + p
and we get
u = aT 4
where o is a universal constant that cannot be defined thermo-
dynamically (o=7.56 X 10-16 JK- 4 m- 3 ). All the other
thermodynamic functions are now easily determined. Since
E = aT 4 V, it follows that
S= ~ aT3V, F=-! aT"'V

<D = F + pV = 0 and H = T S
We find the heat capacities:
Cv = 4o T 3 V
and since for black-body radiation an isobaric process is also
1
an isothermic process (p = 3 crP), we get
Cp =+oo
ANSWE:tts 397
which means that v= c:c

= oo, although the adiabatic equa-
tion for black-body radiation is
pV 413 =constant (S is constant).
101. Since
Z= ""
""
1 ~ )' e _ _!!_
e k1'
L V )
k1' df = exp ( e k1' -
.LJ N! h3N A.3
N=O
where 'A= V 3mkT

h is the "thermal" de Broglie wave, we
have
Jit.a N kT
f.l.=kTln-y, p=-v-
3 - - eV
S=-
2 Nk+Nk ln-=- NA.3
(e is the base of natural logarithms).
102. The sought probability is
1 V N :S+~tN
pN = "NT ( 1:3) e k1'
But since
8 = - kT eJJ./k1'
v ,
1:3 - = e~
N
v
kT J.:.3
we have
1 --N
PN=Nfe-NN
which is the Poisson distribution.

103. The dependence of !.1. on U can be found from the
relationship
f.l.= ( !! )s,v=flo+U
Specifically, in a homogeneous gravitational field,
!.1. =flo+ mgz.
104. If the average number of particles with spin "up"
is N1 , and with spin "down" is N 2 , we get (see the solution
398 i>RO:Sl..EMS IN' TliMRETlCAL PHYSlCS
to Problem 103)
J.t1 = kT ln N1 - J.lBB + <p (T)
f-t 2 = kT ln N2 + f.LBB + <p (T)
where J.lB is the Bohr magneton. Now using the condition
of equilibrium, J.t1 = J.t 2 • we have
jij 2 2flBB )
Nt =exp
(
--w- .
105. pV = - 8.
106. Hint. Use the relationships
g = -pV
dE = TdS - pdV + J.l dN
dH = TdS Vdp + + f-t dN.
107 C =kT -(!!__) - (% ):-
• v aT .... (aN )
all T-
108. The Clausius-Clapeyron equation is

dp A.
dT=T(V2-Vt)
Far from the critical point we can assume that V 2 ~ V1
(V1 and V 2 are the molar volumes of the vapour and solid
phases, respectiv~ly), and the gas obeys the equation of
state for perfect gas. Hence,
~
p =constant X e -1fT.
109. Let the concentration of the solution be c = ~ ~ 1
and the Gibbs free energy of the solvent be <l> = N~-to (p, T).
We denote by ~ the change in <l> when one molecule of
the solute is added to the solvent. Then by virtue of the
condition c ~ 1 we can assume that the molecules of the
solute do not interact. Now if we take into account the in-
distinguishability of the molecules of the solute, we get the
following expression for the Gibbs free energy:
fl
 must be
a homogeneous function of degree one in n and N. This
yields
f (p, T)
N
(so that under the logarithm sign is a function of degree

zero_in n and N), and
<l> = NJ.~- 0 + nkT ln ; + ncp (p, T)
where cp (p, T) = kT ln f (p, t). Using the obtained expres-
sion for <l>, we find the respective chemical potentials of the
solution and the solute:
0~ 0~
J.t= oN =J.to-kTc, J.t 1 =a;-=kTlnc+kT+cp(p,T).
110. For the solute and the solvent the conditions of

equilibrium in the gravitational field have the form
(T = constant) ·
kT ln c + q> {p, T) = - mgz +constant
J.to - kTc + M gz = constant
If we differentiate these equations with respect to z and
bear in mind that the volume of the solution is
V = (~) = N ( Of.lo) +n (~)
iJp T iJp T iJp T
we have
.!_~+!!_~= _!L {M ~-m}
c dz V dz kT 0 Vo
where M is the mass of a molecule of the solvent, m the
mass of a molecule of the solute, and V 0 = T and (to)
oCJl)
V1 = ( iJp T =
V- NV0
n are the volumes related to one
molecule of the solvent and the solute, respectively. In
the first-order approximation in c we get the solution
c=c 0 exp [ - :; (m-M ~~}]

400 Pil.OBtEMS IN THEORETICAL PltYS1C~
This is the barometric height formula corrected for Archi-

medes' principle.
111. It is easily shown that for black-body radiation
the Gibbs free energy <D = E - T S + p V = 0 (see the
solution to Problem 100). Hence, f.L = 0.
112. Consider a system whose state is described by the
variables a 1 and a 2 ; and A 1 and A 2 are the generalized forces.
If q> (a 1 , a 2 ) is a certain function of the state of the system,
we have
and
( aat ) ( aa2 )
8A2 At= 8A1 Az
because f = q> - a1 A 1 - a 2A 2 will also be a function of

state. Using these relationships, we fi.nd that
This yields
( aa1 ) < ( aa1 )
8A1 a2 8A1 Az
since (~~ 2)
ua2 At
> 0 because of the condition of stability.
This inequality expresses the following physical fact: an
external force A 1 changes the parameter a 1 and, hence, the
parameters a 2 and A 2 , and the measure of this effect will be
the derivative :~:. At fi.rst, obviously, there will be almost
no change in A 2 • The force exerted will he characterized by
(~~JA 2 . But later new values of a 2 will establish them-
selves in the system and the external force exerted will be
characterized by ( :~1 ) a 2 , which proves to be reduced.
. 1
113. The condition for mechanical equilibrium of the
"vapour-liquid" system in the presence of a surface separat-
ing the two phases is of the form (f.L and p are the same for
both phases)
dF = - p 1 dV1 - p 2 dV 2 +adS
ANSWERS 401
where p 1 is the pressure in the drop, p 2 is the pressure in the

vapour, a is called the surface tension, and dS is the surface
differential. Keeping in mind the constancy of the total
volume vl + v 2 and the sphericity of the interface, we get
2a
Pt=P2+7f·
114. Since 2J v; = 0 in this reaction,
i
i.e. the affinity constant does not depend on p.

115. Assume that a drop of liquid of radius R has formed
in the vapour. The Helmholtz free energy will then change by
!!J.F = (f.t 2 - f.t 1 ) N + aS
where N is the number of particles in the drop, and f.t 1 and
f.tz are the chemical potentials of the vapour and the liquid,
respectively. But
N = 4nR3
3v
where v is the volume related to one particle in the liquid.
If f.tz < f.t 1 , it is easy to prove that f..F reaches its maximum
at point Rcr = ~. From this it follows that if the
~tt -!!2
"nucleus" has R > Rcr. the drop will grow. Otherwise, it

will vaporize.
116. Assume a drop of radius R takes on an ion with
charge q and radius R 0 • The change in the Helmholtz free
energy will be
R
e J•
3-v (f.t 2 - f.t 1) + 4naR 2 + -8n
-4nR3 . . q2
f..F = -er 2 dr
Ro
1 r q2
co
1
00
1 q2
+ 8n J r2. dr- &l J --;:2 dr
R Ro
= 4 nR3
3v
(~-t 2 -~-t 1 )+4naR 2 +:f_(1-_!_)
. 2 ·e
(J...--
R R
1- )
0
where e is the permittivity of the liquid. The last term in

this expression is always negative and grows as R grows, i.e.
26-01496
402 PROBLEMS rN THEORETiCAL PHYSiCS
11F decreases. Hence, a drop can grow even if ·fL 2 > f..t 1 , i.e.
even in vapour that has not yet reached the saturation point.
117. Bearing in mind the results of Problem 101 and
applying the condition of chemical equilibrium, we get
f-LA = f..li + f.le
kT 1 cAph3 o - kT ln ciph3
n (2nmAkT) 312 kT -f- fL A- (2n:mAkT) 312 kT
+ f.llo + kT ln (2nmekT)
ceph3
312 kT
+ f.leo
We denote the initial number of atoms as Nand introduce
the degree of single ionization a. We then get the following
relationships for the number of particles of the reactants and
the respective concentrations:
ne=aN, n1 =aN, nA = (1-a) N,
a a 1-a
Ce·=--, C[= 1+a' CA= i+a
1+a
From this we get
a-r- 1 p (~)3/2e:~J-1/2
- L + (kT)5/2 2nme •
118. When p and T are constant, the f1rst law of thermo-

dynamics yields
11Q = 11E + p 0 11 V = - T 2 a~ L'l.T!D
where r!J is the Gibbs free energy. But the change in the
chemical potential for a reversible chemical reaction with
constant p and T is expressed as
/1$ = ~ f..l;I1N; = ~ f.l;V; = kT ln K (T)
i i
Hence,
!1Q = - kT2ay ln K (T)
a
119. For the reaction H 2 +!: H + H we have

ln K (T) = 2X' (Tl; X (T)
where x(T) = f..lH2- kT In PH2 and x'(T) = Jlli- kT In PH·

ANSWERS 403
The chemical potentiai J.!H was determined in Problem

101, and J.!H 2 we find from tho grand partition function
for a diatomic perfect gas:
00
Z- "" _1_ eJl.N/(kT)zN

- L.J Nl
zN zN
trans! rot vibr
N=O
= ~ _1_ [ Jl./(kT) V ( 2nMkT ) 3/2 1

L.J N! e hZ hv
N=O 1-ehT
where M is the reduced mass of the two atoms in a mol-
ecule, and I is the moment of inertia of a molecule. The
"zero point" for the energy of vibrations En = hv ( n + } )
is the energy of zero-point vibrations, and y is a constant
that takes account of the symmetry of a hydrogen molecule
(for a diatomic gas y = 2, the two-fold degeneracy).
But
oc N
~ ~I =ex and pV=kTlnZ
N=O
then
= Jl.lkT ( 2nM) (kT)7/2 4n2J _ 1
P e h2 h2 1 _ehv/kT
f1 = kT In p- ~ kT ln kT- kT ln f( 2~~ f 12 4~:1 J

+ kT In 1-e~v/kT + J-to
Hence, for kT ~ hv, we have
1 K (T) = - !i!._+ 2f.l~-2f.l~2 + 1 4/ y;t
n 2 kT n vm 1 3 2
(m is the mass of a hydrogen atom), and
K (T) = 41 ~;t _.!.._ el:!.e/kT and f..Q = !!!.._ + f..e

(kT) 3 12 v V kT 2
where f..e = 2~-t~ - 2J-t~ 2 is the dissociation energy for
a hydrogen molecule.
120. f..p= (N2 -;Nt) kT.
26*
121. We consider one quantum mechanical state to be

a thermodynamic system and calculate the function 8 1 :
n;(ll-8;) 11-ei
8;(Fermi)= -kT In~ e hT = ( +!) kT In (1 ± ekT) (1)
Bose n.,po
l
(The energy of such a "system" is n; e;, n 1 is the number of par-

ticles in the "system", summation is done from 0 to 1 for
fermions and from 0 to oo for bosons, and l! is negative for
bosons.)
The number of states with a given momentum (see Prob-
lem 125 from the section "Quantum Mechanics") integrated
over all possible directions of the momentum is
4np2 dp
dg(p)=g 8 h3 V
Hence the number of states with a given energy is
dg(e)=BVYede
Since the spectrum is dense, the total thermodynamic func-
tion 8 is
8(Fermi) =
Bose
~. 8. (Fermi)=
1
1
Bose
( =
'
11 ) BV kT
""
Jr ]n (1 + ekT) Ve de
0
11-e
Integrating by parts, we get
00
= - ; J d(e) = - ; E.
0 ± 1
e hT
122. S=k[ln-8-,+..!.ln( e-E

_ e-E e E
.!:!.)]
gl
..
. This relationship is depicted .in Fig. 71, from which we
can see that in the energy range from e+g2 to e the deri-
gl g2
vative (!!)as, which by definition is equal to 1/T, is nega-
ANSWERS 405
tive, i.e. in this energy range the temperature must be con-

sidered negative. Note that negative temperatures correspond
to higher energies than positive. We will always meet this
s
I
!lz I
klng:
, k
---+--- I
~~----~----~~-E
cg2 E;
g,+g2
Fig. 71
fact when we have quasi-equilibrium systems with an energy

spectrum with an upper limit.
123 .Cv=Nk(liro)2 1
2 kT . h 2 /iro
sm 2kT
8 nB
124. E= - 8- - - ns
ekT -1
125. The average number of phonons with an energy hv

is n = hv 1 . The number of states in the frequency range
ekf -1
[v,v+dv] is
Q (v) i:lv = 4nv 2 dv ~~
where
3 2 0 2
C3=-ca-r-"CT
t I
Renee,
x3d.r:
eX-1
where Vmax is the maximal Debye frequency determined

Vmax
from the condition J Q (v) dv = 3N (conservation of
0
numbers of degrees of freedom of the system).
If tern peratures are low, we can set the upper limit at
infinity. Then [see formula (21), Appendix 4]
4 V k 4T 4
E=sC3~n5 •
126. Cv = N (3n 2 ) 213

3~n ( ~ ) 113 kT.
-00 -00
The factor e- ~; 11 for metals is about e- 10 at T ~ 102 K.

So using the fact that for x {: 1 we have ln (1 + x) ~ x,
we get
w
. _ 4nenm (kT) 2 -w
lx- h3 e
Here w = W- f..t is the effective work function. This is the

quantum Richardson equation.
_ 1 (
12s. P-r; 3n 2
)2;3
. -m v
fl2 ( N )5/3
129.
2 E 2 N
p=3v=r; [ 5n2 ( kT )2] '
vf..to 1+12 ~ where f..to=
tz2 (
= 2m 3n2 vN ) 2/3 •
IS the Fermi• level for electrons in
a metal at T=O.
-k f (n; Inn;+ (1-n;) ln (1-n;)]
130. S = { for fermions
- k f (ni lnni -(1 +ni) ln(1+ni)l for bosons
Here
1
n; = e.-JJ. and
' +1
1(/iT
ANSWERS 407
00
oE ) E= f
( e-•t) ,
eg (e) de
131. Cv= ( 7if v' J where ~t is
o exp ---IT- + 1
found from the condition that
00
1 g (e) de = N
e-j.t) +1
J0 exp ( ~
At low temperatures these integrals can be represented as
r
J.l.
E = J eg (e) de+
nZ
+6 nZ
k2 T 2 ~g' (~) + ...
6 k 2T 2 g (~t)
0
[see formula (20) of Appendix 4]. But for T = 0
J.l.o
Jg(e) de=N
0
and also ~ - ~ 0 is much less than ~ 0 and ~ as T tends to

zero. Therefore, with an accuracy up to terms quadratic in
T we find that
Jr g(e) de+ 6n2 k 2 T 2 g' (~) + ... =0
~=~ 0 -~(kT)
6
2 [~lng(e)J
_de e=J.l.o
Using the obtained value of ~. we get with the same degree

of accuracy
Hence,
n2 2
Cv ~ 3 k g (J.to) T.
132. Let us show that dArev = - h n; de;. By defini-

;
- a<;:/.
tion, for reversible processes n; = - a~~ . If we use formula
(1) from the solution of Problem 121 and note that a-;:=
()';:/.
l
a'B; ~ iJe·
- - -h;- ' and de;= ~ ' da", we construct
.:>;-a
% .
l=1
Thus,
dQ =dE+ dArev = hi e; dn;.
C = { aE ) = _!_ [ V li ( }!_) 5/2 T5/2 f x 1 dx J
3 2
133. V aT V aT - 4n2A 3/2 li J -1 ex
0
= ~~ n ;~ 2 ( :li r' 2
X 1.341T 312 = BT 312 [see formula (21)
of Appendix 4].
134. The state of a solid whose atoms are considered to
be harmonic oscillators is determined by the volume V and
the set of oscillators (n; = 1, 2, ... ). In such an approxi-
mation the energy of the crystal is
3N--6
E = E 0 (V) + ~ (n; +i-) hv;(V)

i=1
where E 0 (V) is the interaction energy of N motionless par-

~icles
of the cr~stal. ·
ANSWERS 409
Knowing E, we can find the pressure p =~ kT ( a----ay-

lnZ) T'
where
3N-6
Z= e-Eo(V)/kT n
i=1
1
h'V
. h i
2 sm kT
Then
3N-6
p = _ ( OE0
oV
)
T
1
__
V
~
LJ
( 1
2 hv; + /'~~/lkT) _ 1
hv; ) (aalnln v;)
V
i=1
=- ~~,o + ~~
Hence,
(.~)
ar v
= ~Cy
v
Now, using the formulas in the solution to Problem 65 and
the definitions for a.. 1 and ~. we find the sought relationship.
~ n4N ( -!.} 3
135. Q (E) o: e8 /k ex: e5 To , where T 0 = hvmaxfk.
136. Let the lower edge of the conduction energy band
be Ec, and the higher edge Ev. Then from the condition of
the electroneutrality of a semiconductor (n = p) for the
simplest law of dispersion,
!z2k2
En=Ec+-2
mn '
we get
Hence,
= E~-sT+~kT l mn.
~ 2 4 n mp
137. n= ~~ (8mfmt) 312 (kT) 312 F 1;z k~ (~t-·Ec)] [see r

formula (22) of Appendix 4]. When integrating with
respect to k1 we took infinite limits, which, obviously, is
possible only if these limits lie outside the area of occupied

states. Practically speaking, this is always the case.
138. The total concentration n is equal to the sum of tho
concentrations in the lower (n1 ) and upper (n 2) bands, i.e.
n= n1 + n2 = (2 !) 3 Jdk { 1 + exp [ k~ (E 1- ~-t) J} -1

+( 2 ~) 3 Jdk{1+exp[k~(E2 -~-t)]}-t
h2\ k 12 . Jt2\ k 12
Here E 1 (k) = 2 and E 2 (k) = 2 + E 0 • The energy
m1 m2
is reckoned from the lower edge of the first baud. Then
= 2(m 1kT)3/2 F 1 2 (L) + 2 (m 2kT)3/2 F 1 (~L-Eo)
n 2nh2 1 kT 2nnz 12 kT
Using the asymptotic form of the Fermi integral [see

formulas (23) and (24) of Appendix 4], we fmd
j.t = 1£T In n
2 ( kT ) 3/2 [m3/2+m3/2e-Eo/(kT)]
2nll2 1 2
n = g~z ( 2h~t ) 3/2 !-t3J2 [ 1 + ( :~ )3/2 ( 1- ~o ) 3/2 e (I-t- Eo) j

where the step function
1 for x>O
ex-
( )-
{
0 for x < 0.
00
139. n = !)
(2 3 Jf (E) dk = ) g (E) f (E) dE, where g (E)=
0
= ~~ :; is the density of states. When the energy is small,

we easily find that
k2 = _1 (
2y
1--. /1-
V
8myE ) '"""' 2mE ( 1
nz '"""' nz
+2myE
nz
)
Hence,
n= 2 (
'
~~::
,.,.
rm [
' r
F 1/z ( kr )+ 1:
' , o.
m~~kT F 3/2 ( :r )l""'•
ANSWERS 411
140. If we choose the edge of the conduction band as

the energy's zero point, we get
g (E)= z~z [2mn~O) r/2 [ El/2 ( 1 + :g) 1/2 ( 1 + ~~) J

In the case of a degenerate semiconductor we have
n
= _1_ [2m ·(o)J3/2
3:n2 - fiZ ~
312 ( 1+ L)
Eg
3/2
.
141. Let the lower edge of the conduction band 1 (Fig. 72)
be the zero point for calculating the energy. The probability
that the quantum state with
energy s is not occupied by f
an electron, i.e. is a hole
by defmition, is expressed 1
as
f' (s) = 1- f (s) L.J I
1
exp ( ft _ k~ ) +1
The energy of an electron
in the conduction band is Fig. 72
n2kz
- , on
Bn = 2 the donor
mn
level it is s = - E 1 , on the acceptor level s =
= - E 2 , and in the valence band 2 it is s = - E 0 - £p,
where cp = 2n2 k 2 is the kinetic energy of a hole. Now we
mp
write the condition for the electroneutrality of the semi-
conductor:
J 2g (e) de +~ 1
band I ex P ( En;;;. ft) + 1 acccp
Assuming that the electrons in the conduction band 1 and

t}fe 4oles in the valence })and 2 obey Boltzmann's statistics 1
we get for x = exp (~lkT) 1he following equation:
(2:rtmnkT) 312 + nz
4:rt3Ji3 X e-E2/hT __!__+ 1
X
(2:rtm kT) 312 1 n

- p
4n3Ji3
e-Eo/kT-
x
+ e-Et/kTx+i
I
where n1 and n 2 are, respectively, the concentrations of do-

nors and acceptors.
Obviously, the concentrations of electrons in the conduc-
tion band n and holes in the valence band p are expressed
as follows:
This is an equation of the fourth degree in x. Therefore, we

consider its solution in a special case: n 2 = 0 for kT ~ E 1 •
Then
= kT In 4n3!i3nl
~ (2:rtmnkT) 312
and n ~ n1 , i.e. all the donors are ionized.
We can examine an acceptor-type semiconductor in a simi-
lar way.
142. Hint. Use the results of Problem 141.
143. The energy of electrons and holes in 1 cm 3 is
00 00
E=2) ej(e)g(e)de+2) (E 0 -e')f'(e')g'(e')de'

0 0
= n(3kT +Eo)
where f (e) and g (e) are the distribution function and the
density of states of the electrons, f' (e') and g' (e') are the
same for holes (see Problem 141), and
n- (
2
:rt
v ,n-;n- n p
kT)3/2
e-Eo/kT
. - ilf21i3 ·. , '
ANSWERS 413
Hence
Cv = kn [ ~ + 3 ( :~ ) + ~ ( :~ )2 J
which is valid for E 0 ~ kT (the electron gas is nondegene-
rate).
144. p = Bn~li 3 [Po ( ; p~- m2c2 ) V p~ + m c + (mc)q X 2 2
X ·
arc smh Po
me_ J
, where Po = (3:n: 2 ) 1/3 h- ( VN ) 1/3 is the maxi-
mum Fermi momentum.

- Tc i (i+i)
LJ (2j+1)e T
145. Nortbo=3i=1,3...
T , where T c =
Npara _ _.......£i{i+1)
~ (2j+1)e T
j=O, 2 ...
ft2
= 2lk" Keeping in mind that T <t Tc, we get
_ 2 Tc
Northo 9e T
Npara
146. Let W (Eh) be the probability that a particle is
in a state with energy Eh. Clearly, if two particles with
energies E 1 and E 2 , respectively, are to pass into states
with the energies E 3 and Eq after colliding, the latter must
not be occupied (here the Pauli exclusion principle comes in).
Now using the hypothesis that for a large number of partic-
les the probabilities of the direct and reverse processes are
the same, we find for the process
E1 +E 2 ::;::: E3 +E 4
the following functional relationship:

W (E 1 ) W (E 2 ) [1 - W (E 3 )J [1 - W (E,)J
= W (E 3 ) W (E,.) [1 - W (E1 )l [1 ~ W (E 2 )l
we introduce the function I (E) = w-l (E) - 1 and get

The solution of this equation for E 1 +E 2 = E3 +E 4 is

given by the function
f (E) =AeaE
Hence
1
W (E)= aE •
Ae +1
147. Consider the crystal as an aggregate of 3N normal
modes (more precisely, 3N - 6, but this is unimportant
when N is great) with frequencies ffij. The average energy
associated with each mode is ( ni + ~) hwi. The energy
related to the jth oscillator will be
NMwjrj = (ni+f) hwi
where M is the mass of an atom, and ri is the contribution of
the jth normal mode to the atom's displacement.
We divide the equality by N M wj and sum up with respect
to j, and we get
( -n·+-
1 )
r2= _n_ ~ J 2
MN . Wj
J
We replace the sum by an integral and, keeping in mind the

Debye nature of the spectrum, we find that
wmax
2
r =
n Jr co th ( 2kT
2M N
nw ) ---c;)
g (w) d
ffi
0
9Nw2
where g (w) = -3- (see the solution to Problem 125);
wmax
ffimax is determined from the condition that the number
of degrees of freedom of the system must be conserved:
00 max
Jg (w) dw = 3N.
0
If we define the Debye temperature as Tv= liwmax, then
k
for T ~Tv we have
- 9n2 ( 2n:2T2 )
r 2 = 2MkTv 1 + 3T;"" ·
ANSWERS 41S
148. p;=p(1-~ 2 t1, T'=~~:=TVt-~ 2 •

149 ( T) _ 8nn 3 (v) h 3 d ln [vn (v)] hv
• u v, - c3 V d ln v ehv/kT -1 •
150. The energy of an electron with a spin magnetic mo-

ment flB in a magnetic f1eld B (directed along the r-axis)
is
p2
8= 2m+ flBB
Hence, the magnetization (magnetic moment density) is

00
M=flB J[g(e+flBB)-g(e-flBB)]f(e)de
0
00
2 1 ag 1
~ 2flBB J de ~ de
o e kT +1
Now we determine the magnetic susceptibility in certain
special cases:
(a) for T = 0 K
M 2
x= 73 = 2flBg (flT=o)
where J.!T=O is the chemical potential of the electrons at ab-
solute zero;
(b) strong degeneracy (kT ~ J-t)
X= 2fl~g (fl) [ 1 + ~2 k2T2 iJ2l;;z + ... J
(fl)
[see formula (20) of Appendix 4];

(c) no degeneracy (kT ~ J-t)
J.L-B
since t = ew
(n is the number of particles per unit
volume).
151. In the case of Boltzmann's statistics the partition
function of the system is Z = z~. If the electron gas is ;!
in a cube with edge L and the fi.eld B is directed along the
416 Pll.OBLEMS IN T1IEO:ll.ET1CAL PtlYSICS
z-axis, we have
00 2 2f,!*H (1+...!_)
L \ - ____£=-.__ - 2 B
z1 = h J·e 2mnkT dp 2 ~ Q (E 1) e kT 2 cosh fl:T
-co l
• en en
where fL = -2 mnc
- 2 me, and Q (E 1) is the degeneracy
, [LB = -
multiplicity of the lth energy level, which is
£2 \ f L2eB
Q(EI) = h,2 JJ dpl. dpy = h 2c
where the integration region is
2+ 2
2fL*Bl < P~
mn
Py < 2[L*B (l + 1)
since all the levels that at B = 0 lie in the integration re-
gion combine for B :=F 0 into one level 2[L*B(l ~). +
In the case of weak fields
M=NA[Lntanh flkBTB -NA[Lncoth (fl*B
kT
_!!!.._}
fl*B
~ ~~B ( [L~- ~ f.l* 2 )

X ~ ~; ( f.lt + {- fL *2 ) •
2:rt (2m) 312 2

152. ')(. = - 3h3 fLB•
Hint. Determine the Helmholtz free energy using the
Euler-MacLaurin summation formula
00 00
~ f(n)= Jf(n)dn+{f(O)- 1
12 /'(0)+ 7 ~ 0 /"'(0)- ...•
n=O 0
153. ~=-v ~.
154. ~p 2 = -kT ( ;~ ) •.
155. ~s = .,...- mg~rp 2 •
ANSWERS 417
156. The solution is similar to that of Problem 154.

In the variables V and T we have
-v ~s = kT (av)
~ aT v.
157. The relationships in the condition of the problem

follow from the solution of Problem 154.
158 • (LlAfi)Z = g1g2aZ
a/hT 2 e
a/hT
, where a= e2 - e1 •
(g2+g1e )
-
159. Using the relationship .1N 2 = kT -a (aN) , we get
It T, as
the following results:
(a) .1N 2 = N, 6= v\v ;
(b) .1n~=nt(1-ni), 6=,/1 _n.i;
V ni
(c) -.1nf=ni(1+nt),
- - u$1=}/1+_7ii.
ni
For the Fermi gas (b), as we can see, the fluctuation of

the number of particles becomes zero for = 0 and ni 1,
though the relative fluctuation is equal to oo for = ;,i 0.
For the Bose gas the fluctuation remains fi.nite (equal to
unity) even for very great Iii.
160. Hint. Use the following relationships:
( ~~ ~IT = ( ~: ) ~
) T [ - ( :~ N]
)
( :; )T = Jl-T ( :~ );;v·
161. (Mc) 2 =~1jr2 dxdydz .. 5~ R~, where R 0 is the
radius of the sphere, and p (r) is the density of the gas.
162. By virtue of the homogeneity of time and the rever-
sibility of the equations of mechanics we have
qitqOk = qOqt qOqO _ qtqt
i k' i k - i k
27-01496
4i8 PROBLEMS IN THEORETICAL PHYStCS
Now we find
(q!- q?) (qk- q2) = q1 (qk- q2)- q? (qk- qg) = 2q~ (qk- q2)
But the meati of a physical quantity F = F (p, q) over
a nonequilibrium ensemble ll 0 (p, q) - aQ will be
F= JF(p, q, a)e 'I' -Ilo+aQ -aQo

h.T ar
where a is an additional force switched on at t = 0 in the
direction of coordinate Q.
After finding the derivative of this expression with respect
to a and making a vanish, we get
( oFt_ iJF) =-1 pt (Qt-QO)

iJcx. iJcx. a=O kT
Hence,
aqt
(q~-q~) (qt- qO) = 2kT (-i)
t t h. h. iJa.h ah.=O
In the special case, for q1= qh. = q and a 1= ak =a,
(l-qo)z=2kT (aqt)
iJcx. a=O
.
163. Let us assume that the equation of motion for a
Brownian particle can be represented in the form
mx
..= - 6na11x
.
+a
Here the first term in the right-hand side is the drag force
(the Stokes flow), and the second term is an additional force
switched on at t = 0. Solving this equation with the initial
conditions
x
.I t=o = 0, X I t=O = 0
we get
ex. am - 6 :Jta1)t
X (t) = tina'l") t+ (6na'l"))2 (e m -f)
For t ~ 6;a'll,
11x2 = 2Dt
ANSWERS 419
which is known as the Einstein diffusion equation (here
D = 6~~TJ).
164. NA= RTt
3na1J(L\x2)
165. If the z-axis is directed along the gravitational f1eld,
we get
(zt -z0) 2 = 2Dt + ( 6na1J
mg ) 2 t2.
166. The equation of diffusion in the presence of an ex-

ternal fteld U = U (x) is of the form
~-D.!_[....!!_ au+ an
at - ax kT ax ax
J
where D is the diffusion coefficient, and n (x) is the con-
centration of particles.
For a stationary process n = n (x),
ojx =0
ox
where the flow of particles along the x-axis is
l. = -D ( oU n
X
--+-
ox
on ) =-De- u /kT_ (neUik)
kT ox
OX
i) T
We integrate this expression from x 1 to x 2 and get
. n (x2) exp [ uk~ 2 ) J-n (xt) exp [ uk~t) J

Jx= X2 {-D).
J exp [uk~)J dx
Xi
a=~
n
= ( 2mn~-t
3n2h3
)3/2 ' 6at3
1 for
= { 0 f or
a =1= ~
We made no assumption in this case about the dependence of
't' on velocity.
27*
420 l>tl.OaLEMS IN TltEORE'I'!CAt t>ltYSICS
168. For this problem we must determine the current den-

sity it and the heat flux j 2 along the x-axis:
jt= Jevxfdv
Jz=
. Jr Tvxfdv
mv2
The distribution function f is found from the kinetic equation

f=fo-'t ( eE 8fo +vx 8fo)
m 8vx ox
We assume that
m
fo=n ( - -
)3/2 e --
2h.T
mv2
2nkT
aT
and that the field E and the temperature gradient ax have
little effect on f 0 • Then
t:eE
f=fo+--w-vxfo- kTZ
. t:Vx [ 3
e-ykT fo ax
J 8T
Hence (see formula (3) of Appendix 4],

h= 4enA
3m 11 n
r(l+5)
2
(2kT)I/2[eE-(_!_+t)k.!!__]
m . 2 ax
. = 4nA r(l+7) (2kT)!/2 kT[eE-(..i..+2) k.!!__]
]z 3m "V n 2 m 2 ox
Now we represent j 1 and j 2 in the form
h = LuE + kLtz ::
j 2 = L21E + kL 22 !!___
ox
It follows from these relationships that the electrical and
thermal conductivities are
cr = ~ ~2VA1t r Ct s ) ( 2~T r'2 (:: = o)
K= [ -£ 22 + L12L 21J
Lu
= 3nA
3m V n
r ( l+7)
2
( 2kT
m
)112 kT
(the thermal conductivity is measured at it = 0).
ANSWERS 421
We consider two special cases.

(1) l = 0 (the relaxation time -r = A = constant). Then
ne2A 5 1/li'
11 = --;:- ' K =2 nf..o ]I 2nm
(2) -r = l..o (/.. 0 is the mean free path of an electron). Then
v
4 ")/ 1 8 -./kT
11 =3 e 2nf..o V ~2nmkT ' K =3 nf..o V ,2nm ·
169. For E and H constant, the stationary Boltzmann
equation is
- e ( E+ [: X H]) :~ = - ~~ fo
If e (p) = ~:. the term with H will vanish when we sub-
stitute f 0 for f in the right-hand side. Hence,
-e (v·E) Bfo -~ [v X H} a(f-fo) = __ f- fo
ae c ap 't
We look for the solution of this equation in the form

f = fo- (v· a) a;~
where vector a (e) is unknown.
Substituting the assumed f into Boltzmann's equation,
we find the following equation for a (e):
( ) ( . H
-e(v·E)+[vxro}a= + v~a ffi= ~c)
Hence,
a
-eE + [ro xa} = -:r
We first find the scalar product of this equation and ro,
and then the vector product of the equation and ro. From
the two products we find a:
a= - 1 +e: 2 -r 2 {E +-r2 (ro ·E) ro +-r [ro x El}
Hence,
1 for a=~. e 1H I
where 6afl= { 0 f A and ffi=--.
or a;/= t' me
In the last expression we use "plus" or "minus" depending
on whether the permutation of a, ~. y is even or odd with
respect to x, y, z.
If the z-axis is directed along the field, we have
170. The Boltzmann equation has the form
(v·Vr)/(r, v)=- f-lo

't
where
fo=n( 2;kT) 312 exp{- 2;T [(vx-~Y} 2 -f-v~+v~J}

since the field of velocities for the given flow is Vx = ~y,
Vy = 0, Vz = 0
Hence,
/=/o-'t'(v·Vr)/o=fo+'t'~v~ a~v,., fo(v~, v~, v;)
where v~ = Vx - ~y, v~ = vy, v; = Vz are the velocities of
molecules in relation to the moving gas. Now we determine
the momentum carried along the y-axis through a perpendi-
cular unit area in unit time (this is the shear stress X y)
Xy=m't'~ ~ dv'v~2/0 =n't'kT ~; =T} ~;

The factor Tl = nkT't' is the coefficient of viscosity.
171. The function f (r, v, t) satisfies the equation
%+ (v·Vr) f =0
At t = 0,
f (ro, Yo, t) =Po (ro) /o (v)
ANSWERS 423
Since the particles move by inertia, r 0 = r - vt, and

f (r, v, t) =Po (r- vt) / 0 (v)
will be the solution to Boltzmann's equation.
Hence,
p (r, t) = ) f (r, v, t) dv
~ m m
- ( m ) 1/2 .!_ f [ 2hTt2 (ro-r)2- - 2kT12 (ro-r)2] ( ) d
- 2nkTt2 r j e e P ro ro ro
0
Now taking into account that
Po for O::;;r0 ::;;a
{
p (ro) = 0 for r 0 >a
we find that
+<D [y 2;;.t2 (r+a)] -<D [V 2k;t2 (r-a)]}

X
where <D (x) = v 2_

:rt •
0
' e-tl dt.
Appendices
1. Basic formulas
of vector analysis
gradqJ=V'qJ= ~: i+ ~: i+ ~~ k (1)
~ qJn dS = ~ grad cp dV (n is an outward normal) (2)

. aa aay aaz
diva=(V·a)= axx +ay+---a;:- (3)
~ (a·n) dS= JdivadV (4)

j k
curl a= [V' X a]= .!_
ax _!_
ay .!._
az
ax ay az
= ( aaz - aay ) i + ( aax - aaz ) • + ( aay - aax ) k (5)
~ fu fu fu J fu ~
J[n x a] dS = ~ curladV (6)
ab ab ab
(a·V) b= ax ax +ay Ty +azaz- (7)
• fJ2 grad f + f grad qJ (13)
APPENDIX 2. CURVILINEAR COORDINATES 425
div (<pa) = <p diva+ (a-grad <p) (14)

curl (<pa) = <p curl a+ [grad <p X a) (15)
div [a X b) =h curl a-a curl b (16)
grad (a-b)= (a· V) b+ (b· V) a
+[a X curl b] + [b X curl a) (17)
1
2 grad a 2 = (a·V) a+ [a X curl a] (18)
curl [a X b] = (b · V) a- (a· V) b 70
+adivb-bdiva (19)
Jcurln adS
~ a 1 dl = (20)
L B
~ <pdl= J[nxgrad<p]dS (21)

L B
2. Curvilinear coordinates
Many problems are solved more easily if instead of Car·
tesian coordinates we use coordinates more naturally
related to the problems. For a problem with axial symmetry,
for example, it is convenient to use cylindrical coordinates;
for a problem with spherical symmetry, spherical coordi-
nates; etc. Such coordinate systems are called curvilinear
coordinate systems (or, simply, curvilinear coordinates).
Since vectors and operations on vectors (div, curl, etc.)
are usually defined in the Cartesian coordinate system, we
must have formulas that express these operations in an arbit-
rary coordinate system.
Let us consider the projections of a vector in a curvilinear
coordinate system; in this system the Cartesian coordinates
of the vector, x, y, z, are functions of the curvilinear coordi-
nates:
X = X (ql, q2, qa)
Y = Y (ql, q2, qa) (1)
z = z_ (ql, q2, qa)
or in vector form
r = X (ql, q2, qa) i + Y (ql, q2, qa) i + z (ql, q2, qa) k
The derivatives
~=~i+~.+~k
aqt aqt aqi J aqt
ar _ ax • + ay • + az k (2)
aq 2 - aq 2 1 aq 2 l aq2
ar _ ax •
---I
+ ay •
-]
+ -
az k
aqa aqa aqa aqa
in the general case form a set of three linearly independent
vectors, since the Jacobian of transformation (1) is not zero.
The absolute values of these vectors are, respectively,
They are called the Lame parameters (also, scale factors).

If the three vectors (2) are divided by (3), i.e.
we get three unit vectors, which can be considered as the

basis. The basis is not, in general, orthogonal, but from
now on we shall use none but curvilinear coordinate systems
in which the basis defined by (4) is orthogonal. The vectors
of the basis of such a curvilinear coordinate system possess
the following properties:
(e 1 • e 2) = (e 1 • e3 ) = (e 2 • e3 ) = 0
(5)
e 1 = [e2 X e3 ], e2 = [e3 X et], e 3 = [et X ez]
The Cartesian basis f, j, k and the basis~. e 11 , e 8 are con-
nected through formulas (2) and (4).
APPENDIX 2. CURVILINEAR COORDINATES 427
Let us consider the special case of cylindrical coordinates.

Formula (1) takes the form
X = p COS q:>
y = p sin q:>
Z=Z
or, if we put q1 = p, q2 = q:>, and q3 = z, then
r = p cos q:>i + p sin cpj + zk
We then calculate the Lame parameters using formulas (3):
H1 =VI ( :; )2 + ( :~ )2 + ( :; )2 = V cos2 q:> + sin2 q:> = 1

H 2 =p, H 3 =1
Formulas (4) give us the relation between the Cartesian
and cylindrical bases (we denote the second by ep, e'l>, ez):
ep = ar = cos q:>I'+ sm
ap . cp].
e'P =
ar =
P1 aq;- .
- sm q:>I
. + cos q:>J. (6)
ar
ez=-=k
az
A direct check verifies the orthogonality of this basis, i.e.
(ep · eq;) = (ep · ez) = (e'P · ez) = 0.
We can easily find the inverse of (6):
i = - e'P sin q:> + ep cos q:>
+
j = e'P cos q:> ep sin q:> (7)
k=ez
The projections of a vector a on a curvilinear basis, i.e.
a= axi + ayj + azk = apep + a'Pe'P + azez (8)
can be found if we substitute (6) into (8); thus
ax = ap cos q:> - afll sin q:>
au = ap sin q:> + aq, cos q:> (9)
42 8 PROBLEMS IN THEORETICAL PHYSICS
In a similar manner the spherical coordinate system

X = r sin 8 COS <p
y = r sin e sin <p
z = r cos e
is related to the Cartesian system through the following for-
mulas, which we give for purposes of reference:
H 1 = 1, H 2 = r, H 3 = r sin 8
Cr =sin 8 COS <pi+ sin 8 sin <pj + cos8k
e 8 =cos e cos <pi+ cos 8 sin cpj- sin Sk (10)
ecp = -sin <pi+ cos cpj
It is easy to prove that this basis is orthogonal, i.e.
(er·e((l) = (er·ea) = (ecp·ea) = 0
The components are related in the following manner:
ax = ar sin e cos qJ + ae cos e cos qJ - acp sin qJ
ay = arsine sin <p + aa cos e sin <p + acp cos <p (11)
az = ar cos e - aa sin e
When r and z are used simultaneously, we turn to the
parabolic coordinates:
U=r+z, V=r-z, <p=arctanJ!..
:r;
The name parabolic originated from the fact that the surfaces
u = constant and v = constant are paraboloids of revolu-
tion. (We can check this by squaring r = u - z and r =
= v + z. We then obtain the following equalities:
which are paraboloids of revolution if we put u and v con-

stant.)
Putting u = qv v = q2 , l>ENDIX 3. l:l!FFEltENT!At. Ol>EltA'l'OM 429
Calculating the Lame parameters via formulas (3), we

find, respectively,
~/~
H2= V 4V• H 3 = v-uv
3. Differential operators
in curvilinear coordinates
grad f in orthogonal coordinate systems. In the Cartesian
coordinate system grad f is determined in the following way:
d (of.+ ay
gra f= a;• M J"+ a;,M k)
_ (.!L
-
oq, + _!L oq2 + _!}_ oqa ) i
oq 1 ax oq 2 ox oq 3 ox
+ (_!!_
oq 1
+ oq, _!!_ oq2 + _!1_ oqa ) •
oy oq2 oy oq 3 oy J
+ (..!L
oq, oz
oq, + _!1_ oq2 +_!I. oqa ) k
oq2 oz oqg oz
= _!!_ ( oq, i + oq, j + oq, k)

oqt OX oy oz
+ _!1_ ( oq2 i + oq2 • + oq2 k)

oq2 ox oy J oz
+ _!!_ ( oqg i + oq 3 • + oqg k)

iJqg OX oy J oz
!1 grad q + vq2
= vqt 1!1 grad q + uq3 !1 grad q
2 3 (1)
Let us consider the vector grad q1 • The scalar products of
this vector by e1 , e 2 , e 9 are, respectively,
( rad
g qt
.e)= _1_ (
1 Ht
oq,
ox
.!=.._+
oqt
oqt
oy
_!]!_+
oqt
8qt
oz
.!!__)
oqt
=-1-
Ht
( rad ·e)= _1_ ( 8qt ~ + 8qt _!jL_ + oq1 .!!__) = 0

g q, 2 H2 ox oq2 oy oq2 oz &q2
(grad q1 · e3 ) = 0
This suggests that grad q1 is directed along e1 and that its
absolute value equals 1/H1 :
1
grad q1 = H;' e 1 (2)
In a similar manner
1
grad q2 = Hz e 2
1
grad qa = Ha ea
Substituting the formulas obtained into (1), we get
1 of 1 of 1 of
grad I= H. oq 1 e1 +Hz oqz ez + Ha oq3 ea (3)
For cylindrical coordinates
H 1 = 1, Hz= p, H 3 = 1
gradf= ~/ ep+_!_ ~~ eq;+ ~f ez (4)
up p u~ uZ
For spherical coordinates

H 1 = 1, H 2 = r, H 3 = r sin 8
of 1 of 1 of
grad I=-;;-
uT
er +- ;;a ee + -.-" -;:;-- eq;
T uu T Slll u u~
(5)
div a in orthogonal coordinate systems. In the Cartesian

coordinate system div a is defined in the following way:
. oax
d lVa=- + -oay + -oaz
ox oy oz
If we expand a along the basis vectors of a curvilinear
coordinate system, we can write [see formula (14) in Append-
ix 1]:
div (a1 e1 + aze 2 + a 3 e 3) = a1 div e1 + a 2 div e 2 +
+ a 3 div ea + (e1 ·grad a 1 ) + (e 2 ·grad a 2 ) +(e 3 ·grad a 3 ) (6)
To complete the calculation we must find div el' div e 2 ,
and div e 3 • For this we take the curl of (2):
curl grad q1 =curl ( ; 1 e1) =0 (7)
Making use of formulas in Appendix 1, we obtain

1 1
H1 curl et- p 2 [grad H 1 X ed = 0
AP~:11:Nbtit 3. b!F'li'E'flEN'r!At OPERATORS 4M
whence
1
curl e 1= Ht [grad H 1 Xed (8)
According to formula (3)

1 iJH1 1 iJH1 1 BHt
gradHt= Ht oqt et+ Hz oqz ez+ H3 oq3 e3 (9)
Substituting (9) into (8) and making use of formula (5),

Appendix 2. we obtain
1 iJHt 1 iJH1
cur l e 1 = H 1H 3 -,-e
vq3
2 --HH -,-e3
1 2 vq2
(10)
In a similar manner
1 oH 2 1 iJH2
cur l e 2 = H 2 H 1 -,-e
uq1
3- H H
3 2
-,-et
uq3
(11)
1 iJH 3 1 iJH 3
cur l e3 = H 3 H 2 -vq2
, - e1- H H -,-e3
3 1 uq1
(12)
Since by formula (5) of Appendix 2 e1 = [e 2

follows from formula (16), Appendix 1, that
div e1 = div [e 2 X e 3 ] = e 8 curl e 2 - e 2 curl e 3 (13)
Substituting (11) and (12) into (13), we obtain '
dive 1 =_1_ iJH2 +-1- iJH3 (14)
HtHz iJqt HtH3 iJq1
In a similar manner
divez=-1- iJH3 +_1_ iJH1 (15)
H2Ha iJq2 H1H2 iJq2
dive3=_1_ iJHt +-1- iJHz (16)
HtH 3 oq3 H 2 Ha iJq3
Substituting (14)-(1G) and grad a1 , grad a 2 and grad a 3
[using formula (3)1 into (6) and carrying out the required
transformations, we get
diva= H 1; 2H3 [ 0: 1 (a 1H 2 H 3)
+..;._
uqz
: - (a3H1Hz)]
(azHtH 3 ) - -uq3 (17)
By substituting the Lame parameters of the cylindrical

coordinate system into (17) we get the final formula for
432 PM:BLEMS IN 'l'liEORET!CAL PltYS!CS
diva:
. 1 a 1 iJarp iJaz
dtv a= pap (pap)+ pacp+ 7ii:" (18)
In the case of spherical coordinates
. _ 1 a 1 a . 1 iJacp
dtv a- --z-a
r r (r ar) +-.-e
r Sill
2
ae (sm eaa) + r2 sm
. 26 -a-
qJ (19)
curl a in orthogonal coordinate systems.
curl a = curl (a1 e1 + a 2e 2 -l· a 3e 3 ) = a 1 curl e1
+ a 2 curl e 2 + a 3 curl e 3 + [grad a 1 X e1 ]
+ [grad a 2 X e 2 ] + [grad a 3 X e 3 ] (20)
Substituting (10)-(12) and grad a 1 , grad a 2 and grad a 3
[using formula (3)1 into (20) and carrying out the required
transformations, we get
(21)
For purposes of reference we give the formula for curl a

in cylindrical coordinates:
1 iJaz iJarp ) ( iJap iJaz )
curla= ( - ---
p iJqJ iJz
ep+ - iJz
--iJp
eq;
+ ( p1 Tpa (pa)- p1 iJap )

iJqJ ez (22)
In the case of spherical coordinates

1
curl a= r sine {a.
as (sm ea.p)- iJaa}
iJIP er
+ _!_r { -.-1 -e
Sill
oar - _!___ (rarp) } ea +.!. { :!._ (raa)- oar } e (23)
O(jl or r or ae
M ~ in · orthogonal coordinate systems. Since M =
= div grad/, it follows that after substituting grad f [see
APPENDIX 3. DIFFERENTIAL OPERATORS 433
formula (3)1 into (17) we obtain
a ( H2Ha .!L)
• A 1 [ H t aqt
dtv grad I= IJ.I = H tH2IJ 3 aqt
+ a ( HaHt 21..) + a ( HtH2

Hz aq2
2l_.) J
Ha aqa
aqz aqa
In the particular case of cylindrical coordinates
~ _ _!_ ~ (
I- p ap P ap
!}_) + __!_ az t fl
pz acpz + azz (24)
in spherical coordinates
~I -- _1
r2
!_ ( r2 .!.1_ )
ar ar +r2 sin21 e a ( .
ae Slll
at )
O00
1
+ r2 sin at
2
8 acpz
(25)
in parabolic coordinates
M-_4_ {~
- u+v au
(u.!..L)
au
+~
av
(v!.L)
av
+ u+v !!:1...}
4uv acpz
(26)
~a in orthogonal coordinate systems. The explicit form of

aza aza aza
~a = axz + ayz + azz
in curvilinear coordinates is found with the help of formula
(12), Appendix 1:
~a = grad di v a - curl curl a

Since this expression in the general case is cumbersome,
we write out ~a for purposes of reference in cylindrical coor-
dinates
(27)
28-01496
in spherical coordinates
2 ( 1 a . 1 aa«P)
~a= [ ~ar- 1=2 ar +sin 6 88 (sm eaa) +sin 6 acp
J er
+ [~a e + ~ 8ar _
r2
( ae aa _ cos 6 aaiP)
2 sin2 6 sin2 6 acp
Je
e
+ [~alP+ 72 s~n 6 (~a;+ cote~~ - 28~: 6 ) JeiP (28)

In formulas (27) and (28) the operator ~ acts on scalar
functions, as in (24) and (25) respectively.
4. Mathematical supplement
The Dirac delta function. Let I (x) and all its derivatives
be continuous functions in the interval (-oo, oo). The
delta function is then defined by the relationship
00
JI (x) 6 (x- a) dx =I (a) (1)

-00
where f (x) -+ 0 as x-+ + oo.

It follows from the definition that
00
(i) J6 (x) dx = 1,
-oo
(ii) 6 (ax)= I! I 6 (x),

(iii) 6n (x) = ( -1)n 6 (x),
J ei~~x
00
(iv) 6 :(x) = 2~ dk (the Fourier expansion),

-oo
k
(v) 6[cp(x)]= ~ f~{:~8 ?, where x 8 are the simple

•=1
roots of cp (x) = 0.
APPENDIX 4. MATHEMATICAL SUPPLEMENT 435
The gamma and (complete) beta functions. These are

defined by the following integral equalities:
00
r (a)= J e-x;ca-1 dx (a> 0) (2)

0
1
B (ex, ...,R) _- Jf (1- x )fl-1 xa:-1 dx = f r(a) f (~)
(a+~) (3)
0
If we integrate (2) by parts, we get

r (ex + 1) = exr (ex) (4)
For ex = 1 or ex = 1/2 the relationship (4) gives
r (1) = 1 (5)
00
r (1/2)=2) e-Y 2 dy= Vn (6)

0
Using these relationships, we can determine r (a) for ex =

--r, where n = 0, 1, 2, . . . . For other va1ues of ex
= n+i
we must use special tables.
The volume of a sphere of radius R in n-dimensjonal space.
The volume bounded by a surface x: x: + + ... +
x; =
= R 2 (the equation of a sphere in n-dimensional space) is
(7)
If we change the variables, using xi = YiR, we get

(8)
where
Furthermore, the integration is carried over the domain

limited by the surface u: y~ + + ... +
y; ~ 1.
28*
Using the relationships (3), (4), (9), we can easily cal-

culate Vn (1): -
1
Vn(1)= J dyl J ... J dy2 . .. dyn

-1 Y~+ ... +y~~1-y~
1 n-1
= J(1-y~)--dyt Vn-d1)
-1
2
=Vn-t(1)B (f, nti)

f(1/2)f (~)
vn-1 (1) (10)
r ( n11)
Now with this recurrence relation we get
(nti) ___
f(1/2)f f(l/2)f ( ; )
V n (1) = _____:__...:..
nn/2
__:__:._ ... Vt(1) = - - -
r(;+t) r(nt 1 ) r(;+t)
since V1 (1) = 1. Whence
(11)
Calculation of integrals of type

00
I (a)= J xme-a.xn dx (12)

-co
m=O, 1, 2 .. .
n= 2, 4, 6 .. .
If m is odd, I (a) is zero. If m is even,
00
I (a)= 2 Jxme-a.xn dx
0
The integral in the last expression reduces to the gamma

function if we change variables, using y = axn:
(13)
-00
a n
We note that the equality
r(m+1)
!
oo
) xme-axn dx = ~ (14)
0 a n
holds for all m ~ 0 and n ~ 0, and m and n are not required

to be integers.
For particular cases expression (14) yields
00
(i) r xe-ax2 dx= _1_

J 2a'
0
rJ x2e-ax2 dx-- _n4a3/2

00
1/2
(ii) __
'
0
00
(iii) .\ x 3e-a:x 2 dx = ~ a 2,
0
00
(iv) Jx e-ax2dx--1 - -
4
- 4
r<3J2>
a3/2 •
0
Error integral. The error integral is defined as

X
CD (x) = erf (x) = V2_

n
Jo e-t dt
2 (15)
When x~ 1, the integrand in (15) can be expanded in a pow-

er series; thus,
<D(x)
x~o
= v-2(
n
X-
x3
3X 11
x5
+521-
X
... ) (16)
If we apply definition (15), we can show that

00
~
V:rt
J e-t dt= 1 =F (x)
±x
1
(17)
(!) (1) = v2_ J0 e-t

1t
2 dt ~ 0.84 (18)
Some integrals of quantum statistics. In evaluating

physical properties of quantum fermion systems we often
meet an integral of the following type:
00
I ( )= f f(e)de
1
1-l Jo exp ;;; +1 (8 ll)
where f (e) is such that the integral converges.
If we change the variables, using y = 8 -;;; , and restrict
ourselves to low temperatures, we reduce the integral to
the asymptotic series
JJ.fkT oo
I (u.)
1 r
= kT f
J
f (f.L-kTy) dy
e-Y+1
+kT Jf I (f.L+kTy)
eY+1
dy
0 0
00 ll
=kT \ f(l.t+kTy)-f(J-t-kTy) dy+ \ f(e)de
J eY+1 J
0 0
Hence
ll 00 00
f "" (kT)2n-t (2n-i) \ y2n-t
It(f,l)= J f(e)de+2kT LJ (2n-i) f (!-l) J ell+i dy (19)
0 n=1 0
But
00
(-f)m
= r (2n) ~ (m+ f)2n
m=O
= r (2n) (1-2 1 - 2n) ~ (2n)
00
where ~ (2n) = ~ m-zn is the Riemann zeta function, which

m=i
for some values of n is
n2
~ (3/2) = 2.612, ~ (5/2) = 1.341, ~ (2) = 6 '
n4
~ (4) =go' q3) = 1.202, ~ (5) = 1.037
And so,
ll
/1 (!1) = J1(erae
0
00
+2kT~ (kT)2n-t l2n-1)'(J.t) f(2n) {1-21-2n] ~ (2n) (20)

LJ (2n-1)1
n=f
In evaluating physical properties of quantum boson sys-
tems we often meet an integral of another type:
yn-1 r
00
/2{11) = J eY-1 dy
0
Calculations similar to the previous one yield
/ 2 (!1)=
r eY-t
00
J
yn-1
dy=f(n)~(n) (21)
0
i
Returning to / 1 (!1), we must note that if I (e)= f(j~i),
the integral is called the Fermi integral:
(22)
Since the Fermi integral is widely used in the theory of

solids, it has been tabulated for a great number of values of
j. In the simplest case of)= 1;2 we give estimates for Fj('r]):
(23)
F e"l
'1']~1.3
112 ('I']) ~ 1 +0.27e"~ for
(24)
The condition for the totality of the differential of a

function of two variables. We call the linear part of the
increment of a function f = f (x, y) the total differential,
i.e.
d I= ( !~ ) dx + ( !~ ) dy = I
11 x x dx + I y dy
The necessary and sufficient condition for the expression
dF (x, y) = Q (x, y) dx +P (x, y) dy
to be a total differential when Q, P, and their derivatives,
( ~~) x and (a:;) 11
, exist and are continuous is the condi-
tion that
(~~ L= ( ~~ )y (25)
If condition (25) holds,
Functional determinants of two functions and their pro-

perties. The functional determinant or Jacobian of two
functions u (x, y) = u and v (x, y) = v is the determinant
a(u, v)
( :: )y ( ;; )y
a(x, - (26)
y)
(:; L (:~ L
APPENDIX 5. LEGENDRE POLYNOMIALS 441
If we use this definition and the properties of determinants,

we can prove the following relationships for J acobians:
{) ( u, v) ( {)u )
{) (x, v) = ox v
(27)
{) ( u, v) ( {)u )
{) (v, x) = - Tx v
(28)
{) (u, v) = {) (u, v) X {) (p, q)
(29)
{) (x, y) {) (p, q) {) (x, y)
If u = u (p, q, l), v = v (p, q, l), and if p = p (x, y),
q = q (x, y), l = l (x, y), then
{) (u, v) {) (u, v) X {) (p, q) + {) (u, v) {) (q, l)
{) (x, y) {) (p, q) {) (x, y) {) (q, l) {) (x, y)
+ {) (u, v) X {) (l, p) (30)
{) (l, p) {) (x, y)
5. Legendre polynomials
The equation
!:-
dx
[<1- x 2 ) dPzdx(x)J + l (l + 1) P l (x) = 0 (1)
has for its solution the Legendre polynomials

p (x)=-1_. dl(x2-f)l (2)
1 2ll! dxl
We can see this if for Z = (x 2 - 1) 1 we composP tlw
identity (x- 1) ~!- 2lxZ = 0 and differentiate il l tim!'s
with respect to x.
According to equality (2)
P 0 (x)=1, Pt(:r) =X,
1 1
P 2 (x) = 2 (3x2 -1), P 3 (x) = 2 (5x 3 -3x)
etc., and we see that P 1 ( -x) = ( - 1) 1 P 1 (x).
The function 1-,- 1- is the generating function for
r -r 1
P 1 (x) when r' =. 1 and r < 1. Denoting the angle between
r and r' by arc cos x, we write
00
1
F (r, x) = -:;r:;===;;;==:;===;;
'V1-2rx+r2
It is evident that for r > 1 by introducing the variable

r 1 = 1/r we get the expansion
00
1 = ~ P 1 (x) r-1-1
V 1-2rx+r2 1=0
P 1 (x) can also be defined in the following way:

p (x) = _1_ r _1_ (z2-f)l dz
(4)
1 2ni JL 21 (-x+z)l+i
where Lis an arbitrarily chosen closed contour encompassing

the point z = x, which is the pole of the integrand. Expand-
ing f (z) = (z 2 - 1)1 into a series in powers of z - x and
using the theory of residues, we see that definition (4) is
consistent with definition (2). Substituting ~/2 for ;2--~
in (4), with z = ~ (1 - V 1 - 2x~ + ~ 2) (we choose z's
that turn into x as ~-+ 0), we get
1 1 t d~
Pz (x) =2nr11 '.r ~l+l V 1-2~x+~z
1 [
=11
{}l
{}xl
1
V1-2~x+~2 t=O
J
This demonstrates the equivalence of (3) and (4).
Using formula (3), we can easily obtain the recurrence
relations between Legendre polynomials and their deriva-
tives.
aF
By composing the product (1 - 2xr r 2) ar , we get +
the equality
00
V 1-2xr+r2
x-r = ~ (x-r) P 1 (x) r 1
1=0
00
= ~ (1-2xr+r 2 )ZPz(x)rl-1 (5)

1=0
Identifying powers of r, we find the first recurrence relation:

+ +
(l 1) Pl+ 1(x)- (2l 1) xP 1 (x) lPz-1 (.1) = 0 + (6)
APPENDIX 5. LEGENDRE POLYNOMIALS 443
oF
We compose the sum F + 2r 7fr and get
(7)
On the other hand

00
which, if added to (5), yields
00 00
_ x (1-r2)
- (V1-2xr+r2) 3
~ ZPz-tr 1 + ~ lP 1r 1- 1
l=O l=O
Comparing the last expression with (7), we get
(2l + 1) xP z (x) = (l + 1) P1+ 1 (x) + lP"i-t (x)

If we compose the product (1 - 2xr + r oF
ax, 2) we find the
second recurrence relation:
P 1 (x) = Pi+1 (x) - 2xPI (x) + Pi-t (x) (9)
Substituting Pi+1 from (6) into (9), multiplying the obtained

expression P 1 (x) = xPi (x) - Pi-t (x) by 2, and adding it
to (9), we find the third recurrence relation:
(2l + 1) P (x) = dPz+t(x) _ dPz_t(x) (10)

l dx dx
We can find the normalization integral for Legendre poly-

nomials if we use defmition (2). Let us suppose thHt l ~ lc.
Then
1 1
\ p (x)P (x)dx= 1 \ dl(x2-f)l a~<(x2-f)l!. dx= ...
J
-1
It 1 2111 211kJ J dxl
-1
dxk
2
= 21+1 l)lll
where we integrate by parts, bearing in mind that on the

boundaries x = + 1 the function (x 2 - 1) turns zero.
6. Hermite polynomials
Hermite polynomials H, (x) can be defmed on the interval

-oo ::::; x::::; oo by means of the generating function
00
F(x, t)=e-t2+2tx=ex"e-<t-x)"= ~ H~l(x) tn (1)

n=O
from which we immediately get
H (x) = ( anF (x, t) ) = ( -1)n ex• an (e-x") (2)
n axn t=O dxn
The equahty
• BF(x t)
---a:}- = 2t
aF(x t)
------iif--- • f ( )
wtth the use o 1 takes
the form
00 00
~ - 1 H' (x) tn= 2 ~ t!!n (x) tn

LJ n! n LJ n!
n=O n=O
whence, by identifying powers of t, we get the relation

dHn (x) _
dx -
2nH n-1 (x •) (3)
In a similar manner from the equality aF~; t) +

+ 2 (t - x) F (x, t) = 0 we obtain the relation
Hn+l (x) - 2xHn (x) + 2nHn-I (x) = 0 (n ~ 1) (4)
APPENDlX 6. HERMITE POLYNOMIALS 445
Combining (3) and (4), we come to the homogeneous linear

equation for Hn (x)
H'; (x) - 2xH~ (x) 2nHn (x) = 0 + (5)
The orthogonality of Hermite polynomials can be found
from the following relation:
r
J Hm(x)Hn(x)e-x 2 dx=(-1t
r
J Hm(x)
dn ( -x2)
d:n dx
-oo -oo
where we use formula (2) and set n > m. Integrating by

parts and noting that for x = + oo all the derivatives of
e-x• turn zero, we obtain
r00
r dn-le-x2
J Hm (x) Hn (x)e-x• dx= ( -1)n-l 2m J Hm-1 (x) dxn
00
1
-oo -oo
rH
00
= ... = ( - 1)n-m 2m X m! J 0 (x)

dn-m (e-x2)
dxn-m dx= 0
-oo
where we have used formula (3). To normalize Hn (x) we put

n equal to m. Then
00
J II; (x) e-x dx 2 = 2n X n! Vn

-co
The orthonormalized functions are the functions

1
fPn (x) = (2n X n! V n;)l12 H n (x) e-x2
(n = 0, 1, 2, ... )
The general expression for Hn (x)
H n (x) = (2xt- n (~-;- 1 ) (2x)n-2
+ n"(n-1) (n-2)
2!
(n-3) ( 2 )n-lo _
X •••
makes it possible to compute any one of them. For instance,

H 0 (x) = 1, H 1 (x) = 2x,
H 2 (x) = 4x2 - 2, H 3 (x) = 8x3 - 12x
In addition Hn (-x) = (-1)" Hn (x).
7. The confluent
hypergeometric function
For all z and a fi.nite and for any value of c not equal to
0, -1, -2, ... this function is defi.ned by the series
• _ a z a (a+ 1) z2 • _ z
F(a, c, z)-1+c-u+ c(c+i) 21 + .. . (F(a, a, z)-e) (1)
The confluent hypergeometric function (often called Kum-

mer's confluent hypergeometric function) is a particular
solution of the equation
dZQ> dQ>
z dzZ +(c-z){h-a<D=O (2)
The second solution of (2) is the confluent hypergeometric

function of the second kind
z1 -cF (a- c + 1, 2 - c; z) (3)
whenever c is not an integer. The function F (a, c; z) satisfi.es

a number of relations, which follow from formulas (1) and
(2):
F (a, c; z) = ezF (c - a, c; z)
(c - a) F (a - 1, c; z) + (2a - c + z) F (a, z; z)
= aF (a + 1, c; z) (4)
a
-d F (a, c; z) = - (a+ 1, c + 1; z)
aF
z c
an . _ r (c) r (a+ n) F .
dzn F(a, c, z)- f(a) X f(c+n) X (a+n, c+n, z)
If a is a negative integer, a = - n, or zero, the function

F (a, c; z) reduces to a polynomial of degree n:
. ) = 1 --;z
F ( -n,c,z n + ... +(- 1)n (c+n-i)lz
(c-1)1 n
z1-cezr (c) an ( C+n-1 X -z) (5)

= f(c+n) azn z e
APPENDIX 7. CONFLUENT HYPERGEOMETRIC FUNCTION 447
The asymptotics of F (a; c; z) are

Rez-+oo: F(a,c;z)= ~~~) za-c.ez·[1+0(lzl-1)]
Rez-+-oo: F(a,c;z)=r~~a)(-ztal1+0(jzj- 1 )] (6)

Re C-+ oo: F (a, c; z) = 1 + 0 (I c l- 1) for finite z and a
Any equation of type
d2q> dq>
(aox +b
0) dxZ +(at X +bt) Tx + (a x+ b
2 2) cp = 0 (7)
can be reduced to (2) by the following substitutions;
cp = evx(!>' X = A.z 11 +
where v, A., 11 are determined from the system of equations
ao11+bo= 0
+ +
a0 A. (2a 0v a 1) = 0 (8)
+ +
a0v 2 a1v a2 = 0
Books on the Subject
Batygin, V. V., and Toptygin, I. N., Problems in Electro-

dynamics, Academic Press, London, 1965.
Cronin, J. A., Greenberg, D. F., and Telegdi, V. L., Uni-
versity of Chicago Graduate Problems in Physics with
Solutions, Addison-Wesley, Reading, Mass., 1967.
Fliigge, S., Practical Quantum Mechanics, Vols. 1-2, Spring-
er-Verlag, New York, 1971.
Gol'dman, I. 1., and Krivchenkov, V. D., Problems in
Quantum Mechanics, Addison-Wesley, Reading, Mass.,
1961.
Kogan, V. I., and Galitskiy, V. M., Problems in Quantum
Mechanics, Prentice-Hall, Englewood Cliffs, N. J.,
1963.
Kotkin, L. G., and Serbo, V. G., Collection of Problems in
Classical Mechanics, Pergamon Press, Oxford, 1971.
Meshcherskii, I. V., A Collection of Problems of Mechanics,
Pergamon Press, Oxford, 1965.
Problems in Thermodynamics and Statistical Physics, (P. T.
Landsberg, editor). Pion, London, 1971.
Tomasevich, 0. F., Zbirnik zadach z teoretichnoi ftziki
(A Collection of Problems in Theoretical Physics),
Kiev University Press, Kiev, 1958 (in Ukrainian).
Printed in the Union of Soviet Socialist Republics

FllllTHCif\ll\t; .mn
TITLES
Introduction
to Plasma Physics
B. M. SMIRNOV
A book for senior college and .gralluato
students interested in the physics
of a weakly ionized gas. It deals
with the main concepts of the physies
of weakly ionized plasmas and de:-~eriht•.'
the characteristics of a practically
realizable plasma. The book exami III'H llw
properties of a weakly ionized gas, 1
the propagation of radiation in sucl1 J{ll~"·'·l

and the elementary interactions hotwonu
radiation and gas. As an example
of a specific plasma it considers
the properties of the weakly ionizetl I{UH
of the earth's upper atmosphere.
A Handbook of
Methods of Approximate
Fourier Transformation
and Inversion
of the Laplace Transformation
V.I. KRYLOV and N. S. SKOBLYA
Harmonic analysis
and Laplace transformation
are very often used in the solution of mauy
theoretical and practical problems.
This text contains most of the known
methods of the approximate invnrsiou
of the Laplace transformation and motlwdH
of calculating Fourier integrals.
It is designed for scientists and eu!{iiiPors
who have to do with the theory
or applications of the Laplace
transformation and the Fourier intogmls.
It can serve as a useful reference hool<
for computing centres and design bureaus.

Grechko Sugakov Tomasevich Fedorchenko Problems in Theoretical Physics Mir 1977

Uploaded by

Copyright:

Available Formats

You might also like

Grechko Sugakov Tomasevich Fedorchenko Problems in Theoretical Physics Mir 1977

Uploaded by

Document Information

Original Description:

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

Grechko Sugakov Tomasevich Fedorchenko Problems in Theoretical Physics Mir 1977

Uploaded by

Copyright:

Available Formats

MIR PUBLISHERS M O S C OW

.\ llll{'T THE BOO h.

This book is a collection

M3JWT8JI:bCTBO «BMcmaH miWJJa»

Mir Publishers would be grateful for your comments

@ English translation, Mir Publishers, 1977

This text incorporates problems which have been used

BOOKS ON THE SUB1ECT 448

The mechanics of systems with a finite number of degrees

~tnte r1, rj and by the forces acting on each particle in the

Forces acting between two particles are represented by

we can prove that the time derivative of the momentum

prove that the time derivative of the angular momentum of

If a part of the external forces has a potential V, we can

where f1 is a nonpotential force.

Thus, a closed mechanical system always has seven

where Z is the Lagrangian U.efined as Z = K - V (K is

where fs is a nonpotential force; the subscript k is the num-

If there are nonpotential forces in the system but the

where U is a function of the coordinates and velocities,

where :£ = K- V + U. For example, the Lorentz force

the actual motion will be described by such functions

function (or, simply, the Hamiltonian), which is a function

Since the Lagrangian is a quadratic function of the

All the generalized velocities in the right-hand side of

Equations (1-13) and (1-14) are a system of 2k first-order

These are the equations of constraints.

equations of the first kind in the following form:

which must be solved together with (I-15).

In the absence of constraints this principle gives us the

This equation expresses the principle of virtual displace-

poiul U of a 1·igid body in an inertial frame of reference

in terms of three independent parameters. The Euler augles

+M [Ro X [ro X Rcll + Lrot (I -26)

111 "J JJ(x2 +y2)pdV~ I zx = I xz = - JJJzxp dV

kinetir, euergy of a rigid body:

We can always selecl a system of coordinates (attached

We can obtain the motion equations for a rigid body by

I 2 d:2 +(I 1- I a) WtWa = N z

I 3 d:a +(1 2 - It) WtWz = Na

where the subscripts 1, 2, 3 designate the components of

The equations of motion for a continuous medium are

where e is the internal energy per unit volume and j

Pii = - pt3ii + (11- ~) t)ii div v + 1J ( ax; + ovi )

Substituting this expression into (1-35), we get the Navier-

r r:; +<v·V)v]=pf-gradp+lldv+sgraddivv (1-40)

where 11 is the coefficient of shear viscosity and ~ the coef-

r[ ~: +(v·V) v] = pf -grad P+ lldV}

where u (x, y, z, l) is the displacement vector characteriz-

Small elastic deformations are characteri1-ed by the

The stress tensor and the strain tensor are related by

where AtJkl is the fourth-order tensor of elastic constants,

where the lower index i assumes six values:

Ctz Ctz Ctt 0 0 0 l

Isotropic solids are characterized by two elastic constants,

where PiJ is defined by (1-44).

where v8 is the velocity of longitudinal or of transverse

In the theory of elasticity the boundary conditions are

This circuit is connected to a cell with a constant e.m.f. ~