Problems 105

If φ satisfies this equation, then the average value of φ over the surface
of any sphere equals the value of φ at the center of the sphere. This
fact (or alternatively Gauss’s law) implies that it is impossible to con-
struct an electrostatic field that will hold a charged particle in stable
equilibrium in empty space.
• In Cartesian coordinates, the curl of a vector function (written as curl F
or ∇ × F) is
¯ ¯
¯ x̂ ŷ ẑ ¯¯
¯
curl F = ¯¯ ∂/∂x ∂/∂y ∂/∂z ¯¯ . (2.111)
¯ Fx Fy Fz ¯
The curl appears in Stokes’ theorem,
Z Z
F · ds = curl F · da. (2.112)
C S
Physically, the curl equals the line integral of F around an area, divided
by the area, in the limit where the area becomes infinitesimal. Since
the line integral of an electrostatic field E around any closed path is
zero, Stokes’ theorem implies that curl E = 0. See Appendix F for
a discussion of the various vector operators in different coordinate
systems.

Problems
2.1 Equivalent statements *
It is arbitrary which of the two boxed statements in Section 2.1 we
regard
R as the corollary of the other. Show that, if the line integral
E · ds is zero around any closed path, it follows that the line
integral between two different points is path-independent.
y
2.2 Combining two shells *
4 (distances are in units of )
We know from Problem 1.32 that the self-energy of a spherical
shell of radius R with charge Q uniformly distributed over it is
3
Q2 /8π ǫ0 R. What if we put two such shells right on top of each
other, to make a shell with charge 2Q? Since we now just have two
2q A
copies of the original system, it seems like the energy should be
twice as large, or Q2 /4π ǫ0 R. However, the above formula gives an
energy of (2Q)2 /8π ǫ0 R = Q2 /2π ǫ0 R. Which answer is correct, 1 B
C
and what is wrong with the reasoning for the wrong answer?
2.3 Equipotentials from four charges * −q 2 3 4 5 x
Two point charges of strength 2q each and two point charges of
strength −q each are symmetrically located in the xy plane as fol-
lows. The two positive charges are at (0, 2ℓ) and (0, −2ℓ), the two
negative charges are at (ℓ, 0) and (−ℓ, 0). Some of the equipo-
tentials in the xy plane have been plotted in Fig. 2.42. (Of course
these curves are really the intersection of some three-dimensional Figure 2.42.
106 The electric potential

equipotential surfaces with the xy plane.) Study this figure until

you understand its general appearance. Now find the value of the
potential φ on each of the curves A, B, and C, as usual taking φ = 0
at infinite distance. Curve A has been arbitrarily chosen to cross the
y axis at y = ℓ, and it can be shown (how?) that curve B crosses
the x axis at x ≈ (3.44)ℓ. Roughly sketch some intermediate equi-
potentials.
2.4 Center vs. corner of a cube **
Consider a charge distribution that has the constant density ρ every-
where inside a cube of edge b and is zero everywhere outside that
cube. Letting the electric potential φ be zero at infinite distance
from the cube of charge, denote by φ0 the potential at the center of
the cube and by φ1 the potential at a corner of the cube. Determine
the ratio φ0 /φ1 . The answer can be found with very little calculation
by combining a dimensional argument with superposition. (Think
about the potential at the center of a cube with the same charge
density and with twice the edge length.)
2.5 Escaping a cube **
Suppose eight protons are permanently fixed at the corners of a
cube. A ninth proton floats freely near the center of the cube. There
are no other charges around, and no gravity. Is the ninth proton
trapped? Can it find an escape route that is all downhill in potential
energy? Feel free to analyze this numerically/graphically.
2.6 Electrons on a basketball *
A sphere the size of a basketball is charged to a potential of −1000
volts. About how many extra electrons are on it, per square
centimeter of surface?
2.7 Shell field via direct integration **
Consider the electric field E due to a spherical shell of radius
R with charge Q uniformly distributed over its surface. In Sec-
tion 1.11 we found E by using Gauss’s law. Find E here (both inside
and outside the shell) by directly calculating the potential at a given
value of r by integrating the contributions from the different parts
of the shell, and then using Er = −dφ/dr. The simplest strategy
R
is to slice the shell into rings, as shown in Fig. 2.43. You will need
to use the law of cosines.
r 2.8 Verifying the inverse square law ****
As mentioned in Section 1.4, Cavendish and Maxwell conducted
experiments to test the inverse-square nature of Coulomb’s law.
This problem gives the theory behind their experiments.
(a) Assume that Coulomb’s law takes the form of kq1 q2 /r2+δ .
Figure 2.43. Given a hollow spherical shell with radius R and uniformly
 Problems 107

distributed charge Q, show that the potential at radius r is (with

f (x) = x1−δ and k ≡ 1/4π ǫ0 )
kQ £ ¤
φ(r) = f (R + r) − f (R − r) (for r < R),
2(1 − δ 2 )rR
kQ £ ¤
φ(r) = 2
f (R + r) − f (r − R) (for r > R).
2(1 − δ )rR
(2.113)
The calculation requires only a slight modification of the
analogous direct calculation (that is, one that doesn’t use the
Gauss’s-law shortcut) of the potential in the case of the standard
Coulomb 1/r2 law; see Problem 2.7.
Note: We are usually concerned with δ ≪ 1, in which case
the (1 − δ 2 ) factor in the denominators in Eq. (2.113) can
be reasonably approximated by 1. We will ignore it for the
remainder of this problem.
(b) Consider two concentric shells with radii a and b (with a > b)
and uniformly distributed charges Qa and Qb . Show that the
potentials on the shells are given by
kQa kQb £ ¤
φa = f (2a) + f (a + b) − f (a − b) ,
2a2 2ab
kQb kQa £ ¤
φb = 2 f (2b) + f (a + b) − f (a − b) . (2.114)
2b 2ab
(c) Show that if the shells are connected, so that they are at the
same potential φ, then the charge on the inner shell is
£ ¤
2bφ bf (2a) − a f (a + b) − f (a − b)
Qb = · ¤2 . (2.115)
k
£
f (2a)f (2b) − f (a + b) − f (a − b)
If δ = 0 so that f (x) = x, then Qb equals zero, as it should. So
if Qb is measured to be nonzero, then δ must be nonzero.
For small δ it is possible to expand Qb to first order in δ
by using the approximation f (x) = xe−δ ln x ≈ x(1 − δ ln x),
but this gets very messy. You are encouraged instead to use a
computer to calculate and plot Qb for various values of a, b,
and δ. You can also trivially expand Qb to first order in δ by
using the Series operation in Mathematica.
2.9 φ from integration **
(a) A solid sphere has radius R and uniform volume charge den-
sity ρ. Find the potential at the center by evaluating the integral
in Eq. (2.18).
(b) A spherical shell has radius R and uniform surface charge den-
sity σ . Find the potential at a point on the surface by evaluating
the integral in Eq. (2.18).
108 The electric potential

(c) When written in terms of the total charge involved, how do the
above two results compare?
2.10 A thick shell **
(a) A spherical shell with charge Q uniformly distributed through-
out its volume has inner radius R1 and outer radius R2 . Calcu-
late (and make a rough plot of) the electric field as a function
of r, for 0 ≤ r ≤ ∞.
(b) What is the potential at the center of the shell? You can let
R2 = 2R1 in this part of the problem, to keep things from
getting too messy. Give your answer in terms of R ≡ R1 .
2.11 E for a line, from a cutoff potential **
Consider the electric field E due to an infinite straight wire with
uniform linear charge density λ. In Section 1.12 we found E by
direct integration of Coulomb’s law, and again by using Gauss’s
law. Find E here by calculating the potential and then taking the
derivative.
You will find that the potential (relative to infinity) due to an
infinite wire diverges. But you can get around this difficulty by
instead finding the potential due to a very long but finite wire of
length 2L, at a point lying on its perpendicular bisecting plane. Use
a Taylor series to simplify your result, and then take the derivative
to find E. Explain why this procedure is valid, even though it cuts
off an infinite amount from the potential.
2.12 E and φ from a ring **
(a) Consider a ring with charge Q and radius R. Let point P be a
distance x from the plane of the ring, along the axis through
its center. By adding up the contributions from all the pieces
of the ring, find the electric field E(x) at point P.
(b) In the same manner, find the potential φ(x) at point P.
(c) Show that E = −dφ/dx.
(d) If a charge −q with mass m is released from rest far away along
the axis, what is its speed when it passes through the center of
the ring? Assume that the ring is fixed in place.
2.13 φ at the center of an N-gon **
Use the technique from the second example in Section 2.6 to cal-
culate the potential at the center of a sheet in the shape of a reg-
ular N-gon with surface charge density σ . Let the distance from
the center to the midpoint of a side be a. Show that your answer
reduces to the result in Eq. (2.27) in the N → ∞ limit.
2.14 Energy of a sphere **
A spherical volume of radius R is filled with charge of uniform
density ρ. Exercise 1.61 and the example in Section 1.15 presented
two methods for calculating the energy stored in the system. Cal-
culate the energy in a third way, by using Eq. (2.32).
 Problems 109

2.15 Crossed dipoles *

Two dipoles, each with dipole moment p, are oriented perpen-
dicularly as shown in Fig. 2.44. What is the dipole moment of the
system?

2.16 Disks and dipoles **

Two parallel disks each have radius R and are separated by a dis-
tance ℓ. The surface charge densities are σ and −σ . What is the
electric field at a large distance r along the axis of the disks? Solve
this in two ways.
(a) Treat the disks like a collection of a large number of dipoles
standing next to each other. Figure 2.44.
(b) Explain why the parts of the two disks that are contained within
the cone in Fig. 2.45 produce canceling fields at point P, and
then find the field due to the uncanceled part of the top disk. P

2.17 Linear quadrupole **

Consider a “linear quadrupole” consisting of two adjacent dipoles
oriented oppositely and placed end to end; see the left quadrupole
in Fig. 2.16. There is effectively a point charge −2q at the center.
By adding up the electric fields from the charges, find the electric
field at a distant point (a) along the axis and (b) along the perpen-
dicular bisector.

2.18 Field lines near the origin **

s
(a) Two equal positive charges q are located at the points (±a, 0, 0).
Write down the potential φ(x, y) for points in the xy plane, and −s
then use a Taylor expansion to find an approximate expression
for φ near the origin. (You can set a = 1 to make things Figure 2.45.
simpler.)
(b) Find the electric field at points near the origin. Then find the
equations for the field lines near the origin by demanding that
the slope dy/dx of a curve at a given point equals the slope
Ey /Ex of the field at that point.
2.19 Equipotentials for a ring ***
(a) A ring with radius R has charge Q uniformly distributed on it.
It lies in the xy plane, with its center at the origin. Find the
electric field at all points on the z axis. For what value of z is
the field maximum?
(b) Make a rough sketch of the equipotential curves everywhere in
space (or rather, everywhere in a plane containing the z axis;
you can represent the ring by two dots where it intersects the
plane). Be sure to indicate what the curves look like very close
to and very far from the ring, and how the transition from close
to far occurs.
110 The electric potential

(c) There is a particular z value (along with its negative) at which

the equipotentials make the transition from concave up to con-
cave down. Explain why this z value equals the z value you
found in part (a). Hint: The divergence of E is zero.
2.20 A one-dimensional charge distribution **
Find (and make rough plots of) the electric field and charge distri-
bution associated with the following potential:

 0
 (for x < 0)
2
φ(x) = ρ0 x /2ǫ0 (for 0 < x < ℓ) (2.116)

2
ρ0 ℓ /2ǫ0 (for ℓ < x).


2.21 A cylindrical charge distribution **

A distribution of charge has cylindrical symmetry. As a function
of the distance r from the symmetry axis, the electric potential is
3ρ0 R2


 (for r ≤ R)
 4ǫ0



φ(r) = ρ0 (2.117)
(4R2 − r2 ) (for R < r < 2R)
4ǫ

0





0 (for 2R ≤ r),
where ρ0 is a quantity with the dimensions of volume charge
density.
(a) Find (and make rough plots of) the electric field and charge
distribution, for all values of r. The derivative operators in
cylindrical coordinates are listed in Appendix F.
(b) From your charge distribution, calculate the total charge per
unit length along the cylinder.
2.22 Discontinuous E and φ **
(a) What kind of charge distribution yields a discontinuous elec-
tric field?
(b) Can you think of a charge distribution (or perhaps the limit of
a charge distribution) that yields a discontinuous potential?
2.23 Field due to a distribution **
Each of the objects described below has uniform volume charge
density ρ. There are no other charges present in addition to the
given object. In each case use ∇ · E = ρ/ǫ0 to show that the elec-
tric field takes the stated form.
(a) A rectangular slab has thickness ℓ in the x direction and infinite
extent in the y and z directions. Show that Ex = ρx/ǫ0 inside
the slab, where x is measured from the midplane of the slab.
(b) An infinitely long cylinder has radius R. Show that Er =
ρr/2ǫ0 inside the cylinder.
 Problems 111

(c) A sphere has radius R. Show that Er = ρr/3ǫ0 inside the

sphere.
(d) Given that the above three setups all involve the same charge
density and the same relation ∇ · E = ρ/ǫ0 , why do they give
different results for the electric field?

2.24 Two expressions for the energy **

(a) Prove the identity

∇ · (φE) = (∇φ) · E + φ ∇ · E (2.118)

by explicitly calculating the various derivatives in Cartesian

coordinates.
(b) This identity holds for any scalar function φ and any vector
function E. In particular, it holds for the electric potential and
field. Use this fact to show that Eqs. (1.53) and (2.32) are
equivalent expressions for the energy stored in a charge distri-
bution of finite extent. You will want to apply the divergence
theorem with a wisely chosen volume.

2.25 Never trapped **

A number of positive point charges, with various magnitudes, are
located at fixed positions in space. Show that no matter where an
additional positive charge q is placed, there exists an escape route
to infinity that is all downhill in potential energy. The “impossi-
bility theorem” discussed in Section 2.12 will be helpful, but that
deals only with small displacements, so you will need to extend
the argument.

2.26 The delta function **

In spherical coordinates, consider the Laplacian of the function
f (r) = 1/r, that is, ∇ 2 (1/r). From Appendix F we have ∇ 2 f =
(1/r2 )(∂/∂r)(r2 ∂f /∂r) for a function that depends only on r. Since
r2 ∂(1/r)/∂r takes on the constant value of −1, we see that ∇ 2 (1/r)
equals zero. Well, almost. It is certainly zero for r 6 = 0, but we must
be careful at the origin, due to the infinite 1/r2 factor out front.
Show that ∇ 2 (1/r) is not equal to zero at r = 0. Do this by
showing that it is large enoughR (or more precisely, infinite enough)
to make the volume integral ∇ 2 (1/r) dv equal to −4π, provided
that the volume contains the origin. The divergence theorem will
be helpful.

2.27 Relations between φ and ρ ***

R gives two relations between φ and ρ, namely φ =
Figure 2.29
(1/4π ǫ0 ) (ρ/r) dv′ and ∇ 2 φ = −ρ/ǫ0 . Show that these relations
are consistent, by operating on the first one with the Laplacian ∇ 2
112 The electric potential

operator. Be careful that there are two types of coordinates in the

equation, primed and unprimed; it can be written more precisely as
ρ(r′ ) dv′
Z
1
φ(r) = . (2.119)
4π ǫ0 |r′ − r|
You will want to solve Problem 2.26 first.
2.28 Zero curl *
Consider the electric field, E = (2xy2 + z3 , 2x2 y, 3xz2 ). We have
ignored a multiplicative factor with units of V/m4 necessary to
make the units correct. Show that curl E = 0, and then find the
associated potential function φ(x, y, z).
2.29 Ends of the lines *
Explain why electrostatic field lines can’t form closed loops, and
why their ends must be located either at charges or at infinity.
2.30 Curl of a gradient **
The electric field equals the negative gradient of the potential, that
is, E = −∇φ. Show that this implies that the curl of E, which we
can write as ∇ × E, is identically zero. Do this by:
(a) calculating ∇ × ∇φ in Cartesian coordinates;
(b) making judicious use of Stokes’ theorem.

Exercises
2.31 Finding the potential *
The following vector function represents a possible electrostatic
field:
Ex = 6xy, Ey = 3x2 − 3y2 , Ez = 0. (2.120)
(We have ignored a multiplicative factor with units of V/m3 nec-
essary to make the units correct.) Calculate the line integral of E
from the point (0, 0, 0) to the point (x1 , y1 , 0) along the path that
runs straight from (0, 0, 0) to (x1 , 0, 0) and thence to (x1 , y1 , 0).
Make a similar calculation for the path that runs along the other
two sides of the rectangle, via the point (0, y1 , 0). You ought to get
the same answer if the assertion above is true. Now you have the
potential function φ(x, y, z). Take the gradient of this function and
see that you get back the components of the given field.
2.32 Line integral the easy way *
Designate the corners of a square, ℓ on a side, in clockwise order,
A, B, C, D. Put charges 2q at A and −3q at B. Determine the value
of the line integral of E, from point C to point D. (No actual inte-
gration needed!) What is the numerical answer if q = 10−9 C and
ℓ = 5 cm?
 Exercises 113

2.33 Plot the potential *

Consider the system of two charges shown in Fig. 2.8. Let z be the
coordinate along the line on which the two charges lie, with z = 0
at the location of the positive charge. Make a plot of the potential
φ (or rather 4π ǫ0 φ, for simplicity) along this line, from z = −5 m
to z = 15 m.
2.34 Extremum of φ *
A charge of 2 C is located at the origin. Two charges of −1 C each
are located at the points (1, 1, 0) and (−1, 1, 0). If the potential φ is
taken to be zero at infinity (as usual), then it is easy to see that φ is
also zero at the point (0, 1, 0). It follows that somewhere on the y
axis beyond (0, 1, 0) the function φ(0, y, 0) must have a minimum
or a maximum. At that point the electric field E must be zero.
Why? Locate the point, at least approximately.
2.35 Center vs. corner of a square **
A square sheet has uniform surface charge density σ . Letting the
electric potential φ be zero at infinite distance from the square,
denote by φ0 the potential at the center of the square and by φ1
the potential at a corner. Determine the ratio φ0 /φ1 . The answer
can be found with very little calculation by combining a dimen-
sional argument with superposition. (Think about the potential at
the center of a square with the same charge density and with twice
the edge length.)
2.36 Escaping a cube, toward an edge **
Consider the setup in Problem 2.5. Will the proton escape if it
moves from the center directly toward the midpoint of an edge?
Feel free to analyze this numerically/graphically.
2.37 Field on the earth *
A sphere the size of the earth has 1 C of charge distributed evenly
over its surface. What is the electric field strength just outside the
surface? What is the potential of the sphere, with zero potential at
infinity?
2.38 Interstellar dust *
An interstellar dust grain, roughly spherical with a radius of
3 · 10−7 m, has acquired a negative charge such that its potential
is −0.15 volt. How many extra electrons has it picked up? What is
the strength of the electric field at its surface?
2.39 Closest approach **
By means of a Van de Graaff generator, protons are accelerated
through a potential difference of 5 · 106 volts. The proton beam
then passes through a thin silver foil. The atomic number of silver
is 47, and you may assume that a silver nucleus is so massive com-
pared with the proton that its motion may be neglected. What is
114 The electric potential

the closest possible distance of approach, of any proton, to a silver

nucleus? What will be the strength of the electric field acting on
the proton at that position? What will be the proton’s acceleration?
2.40 Gold potential **
As a distribution of electric charge, the gold nucleus can be described
as a sphere of radius 6 · 10−15 m with a charge Q = 79e distributed
fairly uniformly through its interior. What is the potential φ0 at
the center of the nucleus, expressed in megavolts? (First derive
a general formula for φ0 for a sphere of charge Q and radius
a. Do this by using Gauss’s law to find the internal and exter-
nal electric field and then integrating to find the potential. You
should redo this here, even though it was done in an example
in the text.)
s 2.41 A sphere between planes **
−s s A spherical shell with radius R and surface charge density σ is
sandwiched between two infinite sheets with surface charge den-
sities −σ and σ , as shown in Fig. 2.46. If the potential far to the
R right at x = +∞ is taken to be zero, what is the potential at the
center of the sphere? At x = −∞?
2.42 E and φ for a cylinder **
For the cylinder of uniform charge density in Fig. 2.26:
(a) show that the expression there given for the field inside the
cylinder follows from Gauss’s law;
(b) find the potential φ as a function of r, both inside and outside
the cylinder, taking φ = 0 at r = 0.
2.43 Potential from a rod **
Figure 2.46.
A thin rod extends along the z axis from z = −d to z = d. The rod
carries a charge uniformly distributed along its length with linear
charge density λ. By integrating over this charge distribution, cal-
culate the potential at a point P1 on the z axis with coordinates
(0, 0, 2d). By another integration find the potential at a general
point P2 on the x axis and locate this point to make the potential
equal to the potential at P1 .
2.44 Ellipse potentials ***
The points P1 and P2 in Exercise 2.43 happen to lie on an ellipse
that has the ends of the rod as its foci, as you can readily verify
by comparing the sums of the distances from P1 and from P2 to
the ends of the rod. This suggests that the whole ellipse might be
an equipotential. Test that conjecture by calculating the potential
at the point (3d/2, 0, d), which lies on the same ellipse. Indeed it
is true, though there is no obvious reason why it should be, that
the equipotential surfaces of this system are a family of confocal
prolate spheroids. See if you can prove that. You will have to derive
 Exercises 115

a formula for the potential at a general point (x, 0, z) in the xz plane.

Then show that, if x and z are related by the equation x2 /(a2 −d2 )+
z2 /a2 = 1, which is the equation for an ellipse with foci at z = ±d,
the potential will depend only on the parameter a (in addition to d),
not on x or z.
2.45 A stick and a point charge **
A stick with length ℓ has charge Q uniformly distributed on it. It Q
Q
lies along the x axis between the points x = −ℓ and x = 0. A point
x=− x=0 x=
charge also with charge Q lies on the x axis at the point x = ℓ; see
Fig. 2.47. Figure 2.47.
(a) Let x = a be the point on the x axis between the two objects
where the electric field is zero. Find a.
(b) There happens to be another point where the electric field is
zero (it’s inside the stick). In addition to this one, are there any
other points in space where the electric field is zero? Why or
why not?
(c) Make a rough sketch of the field lines and equipotential curves
everywhere in the plane of the paper. Be sure to indicate how
the lines and curves make the transition from their shapes close
to the objects to their shapes far from them. (Don’t worry about
what’s going on extremely close to the stick.) What do things y
look like near the point you found in part (a)?

2.46 Right triangle φ **

The right triangle shown in Fig. 2.48 with vertex P at the origin
has base b, altitude a, and uniform density of surface charge σ .
Determine the potential at the vertex P. First find the contribution a
of the vertical strip of width dx at x. Show that the potential at P
can be written as φP = (σ b/4π ǫ0 ) ln[(1 + sin θ )/ cos θ ].
q
2.47 A square and a disk **
P x
Use the result from Exercise 2.46 to answer the following ques-
tion. If a square with surface charge density σ and side s has the b
same potential at its center as a disk with the same surface charge Figure 2.48.
density and diameter d, what must be the ratio s/d? Is your answer
reasonable?
2.48 Field from a hemisphere **
Following the strategy in Problem 2.7, find the electric field at the
center of a hemispherical shell with radius R and uniform surface
charge density σ . That is, find φ as a function of r and then take
the derivative. You might find it easier to Taylor-expand φ before
differentiating. (You already found this electric field in a simpler
manner if you solved Exercise 1.50. The present method is more
involved because we need to do more than calculate φ at just one
116 The electric potential

point; we need to know φ as a function of r so that we can take its

derivative.)
2.49 E for a sheet, from a cutoff potential **
Consider the electric field E due to an infinite sheet with uniform
surface charge density σ . In Section 1.13 we found E by using
Gauss’s law. Find E here by calculating the potential and then tak-
ing the derivative.
You will find that the potential (relative to infinity) due to an
infinite sheet diverges. But in the spirit of Problem 2.11 you can get
around this difficulty by instead finding the potential due to a very
large but finite disk with radius R, at a point lying on the perpen-
dicular line through the center. Use a Taylor series to simplify the
potential, and then take the derivative to find E. Explain why this
procedure is valid, even though it cuts off an infinite amount from
the potential.
2.50 Dividing the charge **
We have two metal spheres, of radii R1 and R2 , quite far apart
from one another compared with these radii. Given a total amount
of charge Q which we have to divide between the spheres, how
should it be divided so as to make the potential energy of the result-
ing charge distribution as small as possible? To answer this, first
calculate the potential energy of the system for an arbitrary divi-
sion of the charge, q on one sphere and Q − q on the other. Then
minimize the energy as a function of q. You may assume that any
charge put on one of these spheres distributes itself uniformly over
the surface of the sphere, the other sphere being far enough away
so that its influence can be neglected. When you have found the
optimum division of the charge, show that with that division the
potential difference between the two spheres is zero. (Hence they
could be connected by a wire, and there would still be no redis-
tribution. This is a special example of a very general principle we
shall meet in Chapter 3: on a conductor, charge distributes itself so
as to minimize the total potential energy of the system.)
2.51 Potentials on the axis **
A hollow circular cylinder, of radius a and length b, with open
ends, has a total charge Q uniformly distributed over its surface.
What is the difference in potential between a point on the axis at
one end and the midpoint of the axis? Show by sketching some
field lines how you think the field of this thing ought to look.
2.52 Spherical cavity in a slab **
Figure 2.49 shows a cross section of a slab with uniform volume
charge density ρ. It has thickness 2R in one dimension and is infi-
nite in the other two dimensions. A spherical cavity with radius
 Exercises 117

R is hollowed out. A few equipotential curves are drawn in the

figure.
(a) Show that an equipotential curve that starts at the center of the
cavity (curve A shown) ends up meeting the surface of the slab
at infinity. Hint: Superpose two oppositely charged objects.
(b) Show that curve A is a straight line inside the cavity, and find B
its slope.
(c) Show that a curve that is tangent to the sphere (curve B shown)
ends up a distance R/3 outside the slab at infinity.

2.53 Field from two shells ** A

One of two nonconducting spherical shells of radius a carries a
charge Q uniformly distributed over its surface, the other carries
a charge −Q, also uniformly distributed. The spheres are brought
together until they touch. What does the electric field look like,
both outside and inside the shells? How much work is needed to
move them far apart?
2.54 An equipotential for a disk *
For the system in Fig. 2.11 sketch the equipotential surface that 2R
touches the rim of the disk. Find the point where it intersects the
Figure 2.49.
symmetry axis.
2.55 Hole in a disk **
A thin disk, radius 3 cm, has a circular hole of radius 1 cm in the
middle. There is a uniform surface charge of −10−5 C/m2 on the
disk.
(a) What is the potential at the center of the hole? (Assume zero
potential at infinite distance.)
(b) An electron, starting from rest at the center of the hole, moves
out along the axis, experiencing no forces except repulsion by
the charges on the disk. What velocity does it ultimately attain?
(Electron mass = 9.1 · 10−31 kg.)
2.56 Energy of a disk **
Use the result stated in Eq. (2.30) to show that the energy stored in
the electric field of the charged disk described in Section 2.6 equals
(2/3π 2 ǫ0 )(Q2 /a). (Hint: Consider the work done in building the
disk of charge out from zero radius to radius a by adding successive
rings of width dr.) Compare this with the energy required to build
up a hollow spherical shell with radius a and uniform charge Q.
2.57 Field near a disk ****
(a) A disk with radius R has uniform surface charge density σ .
Consider a point P a distance ηR from the center of the disk
(where 0 ≤ η < 1) and an infinitesimal distance away from
the plane of the disk. Very close to the disk, the disk looks
118 The electric potential

essentially like an infinite plane, as far as the E component per-

pendicular to the disk is concerned. So we have E⊥ = σ/2ǫ0 .
q Show that the E component parallel to the disk equals
r1
P Z π/2 Ã p !
hR
σ 1 − η2 sin2 θ + η cos θ
Ek = ln p cos θ dθ .
r2 2π ǫ0 0 1 − η2 sin2 θ − η cos θ
(2.121)
Hint: Use the technique from the second example in Section 2.6
(but now with E instead of φ) to find the sum of the fields
from the two wedges shown in Fig. 2.50 (two diverging terms
will cancel). The distances r1 and r2 shown can be obtained
Figure 2.50. with the help of the law of cosines.
(b) Given η, the above integral can be evaluated numerically.
However, if η is very small or very close to 1, it is possible
to make some analytic progress. Show that the leading-order
dependence on η in the limit η → 0 is Ek = σ η/4ǫ0 . And show
that the leading-order dependence on ǫ (where ǫ ≡ 1 − η) in
the limit ǫ → 0 is Ek = −(σ/2π ǫ0 ) ln ǫ. You can verify these
results numerically.
2.58 Energy of a shell *
A hollow spherical shell with radius R has charge Q uniformly dis-
tributed on it. Problem 1.32 presented two methods for calculating
the potential energy of this system. Calculate the energy in a third
way, by using Eq. (2.32).
2.59 Energy of a cylinder ***
Problem 1.24 and Exercise 1.83 presented two methods for calcu-
lating the energy per unit length stored in a cylinder with radius a
and uniform charge density ρ. Calculate the energy in a third way,
by using Eq. (2.32). If you take the φ = 0 point to be at infin-
ity, you will obtain an infinite result. So instead take it to be at a
given radius R outside the cylinder. You will then be calculating
the energy relative to the configuration where the charge is dis-
tributed over a cylinder with radius R. In terms of the total charge
λ per unit length in the final cylinder,
¡ show that the ¢energy per unit
length can be written as (λ2 /4π ǫ0 ) 1/4 + ln(R/a) .
2.60 Horizontal field lines **
Using the r = r0 sin2 θ expression for the dipole E field lines in
Fig. 2.19, find the locations where the curves are horizontal.
2.61 Dipole field on the axes **
A dipole is centered at the origin and has charges q and −q located
at z = ℓ/2 and z = −ℓ/2, respectively. Find the electric field at
position r along the z axis, and also at position r along the x axis (or
anywhere at radius r in the xy plane). Do this by writing down the
 Exercises 119

fields from the two charges and then adding them, making suitable
approximations in the r ≫ ℓ limit. Check that your answers agree
with Eq. (2.36) when θ = 0 and θ = π/2.

2.62 Square quadrupole **

Consider a square quadrupole consisting of two adjacent dipoles
oppositely oriented and placed side by side to form a square, as
shown in Fig. 2.16. If the side length is ℓ, find the electric field
at a large distance r along the diagonal containing the two posi-
tive charges. Be careful to take into account all quantities that are
second order in ℓ/r.

2.63 Two-dimensional dipole ***

Two parallel wires with uniform linear charge densities λ and −λ
are separated by a distance ℓ. Consider the electric field at points in
a given plane perpendicular to the wires. The fields from the wires
fall off like 1/r. So at points far from the wires, we effectively have
a 2D version of a dipole, where the wires act like point charges with
1/r fields instead of the usual 1/r2 Coulomb fields.
Repeat the process in Section 2.7 for this 2D dipole. That is,
find φ(r, θ ) and E(r, θ ), and also find the shapes of the field-line
and constant-potential curves. (The individual potentials relative
to infinity diverge, so you will want to pick a local point as the
reference point. Any choice will work, but you may as well pick
the point midway between the wires.)

2.64 Field lines near the equilibrium point **

(a) Charges 4q and −q are located at the points (−2a, 0, 0) and
(−a, 0, 0), respectively. Write down the potential φ(x, y) for
points in the xy plane, and then use a Taylor expansion to find
an approximate expression for φ near the origin, which you
can quickly show is the equilibrium point. (You can set a = 1
to make things simpler.)
(b) Find the electric field at points near the origin. Then find the
equations for the field lines near the origin by demanding that
the slope dy/dx of a curve at a given point equals the slope
Ey /Ex of the field at that point.

2.65 A theorem on field lines **

If you solved Exercise 2.64, you probably noted that the final result
is the same as for Problem 2.18. This suggests a theorem. Consider
two point charges with arbitrary values q1 and q2 .

(a) First explain why there is exactly one point where E = 0 on

the line containing the charges, except in a couple of special
cases. (What are they?)
120 The electric potential

(b) Show that at the point where E = 0, the closeup views of

the equipotentials and field lines always look like Figs. 12.39
and 12.40, independent of the values of q1 and q2 . In other
words, the constant-φ lines√passing through the equilibrium
point always have slope ± 2. (You will want to look at the
solution to Problem 2.18.)
2.66 Equipotentials for two point charges **
(a) Two point charges Q are located at (±R, 0, 0). Find the electric
field at all points on the z axis. For what value of z is the field
maximum?
(b) Make a rough sketch of the equipotential curves everywhere in
space (or rather, everywhere in the xz plane). Be sure to indi-
cate what the curves look like very close to and very far from
the charges, and how the transition from close to far occurs.
(c) There is a particular point on the z axis (along with its nega-
tive) at which the equipotentials make the transition from con-
cave up to concave down. In Problem 2.19 we saw that in the
analogous setup with a ring, this point coincided with the point
on the z axis where the field was maximum. Does the same
result hold here? Explain why the reasoning we used in Prob-
lem 2.19 (involving the divergence of E) is still valid, or why
it is now invalid.
2.67 Product of ρ and φ **
Consider a charge distribution ρ1 (r) and the potential φ1 (r) due
to it. Consider another charge distribution ρ2 (r) and the potential
φ2 (r) due to it. Both distributions have finite extent, but are other-
wise arbitrary
R and need not R have anything to do with each other.
Show that ρ1 φ2 dv = ρ2 φ1 dv, where the integrals are taken
over all space. Solve this in two different ways as follows.
(a) Consider the two collections of charges to be rigid objects that
can be moved around. Start with them initially located very far
apart, and then bring them together. How much work does this
require? Imagine moving collection 1 toward collection 2, and
then the other way around.
R
(b) Consider the integral E1 E2 dv over all space, where E1 and
E2 are the electric fields due to the two distributions. By using
the vector identity ∇ · (E1 φ2 ) = (∇ · E1 )φ2 + E1 · ∇φ2 (and
similarly
R with the 1s and 2s switched), rewrite the integral
E1 E2 dv in two different ways.
2.68 E and ρ for a sphere **
In the example in Section 2.9, we used spherical coordinates to
verify the relation div E = ρ/ǫ0 for a sphere with radius R and
uniform density ρ. Verify this relation by working in Cartesian
 Exercises 121

coordinates. You will first need to write out the Cartesian compo-
nents of E.
2.69 E and φ for a slab **
A rectangular slab with uniform volume charge density ρ has thick-
ness 2ℓ in the x direction and infinite extent in the y and z direc-
tions. Let the x coordinate be measured relative to the center plane
of the slab. For values of x both inside and outside the slab:
(a) find the electric field E(x) (you can do this by considering the
amount of charge on either side of x, or by using Gauss’s law);
(b) find the potential φ(x), with φ taken to be zero at x = 0; E
(c) verify that ρ(x) = ǫ0 ∇ · E(x) and ρ(x) = −ǫ0 ∇ 2 φ(x).
E0

2.70 Triangular E **
Find the charge density ρ and potential φ associated with the elec-
tric field shown in Fig. 2.51. E is independent of y and z. Assume
that φ = 0 at x = 0.
x
2.71 A one-dimensional charge distribution ** −a a

Find (and make rough plots of) the electric field and charge distri- Figure 2.51.
bution that go with the following potential: φ(x) = B(ℓ2 − x2 ) for
|x| ≤ ℓ, and φ(x) = 0 for |x| > ℓ.
2.72 A spherical charge distribution ***
Find (and make rough plots of) the electric field and charge distri-
bution that go with the following potential:
 ρ0
 (x2 + y2 + z2 ) (for x2 + y2 + z2 < a2 )
 4π ǫ0

φ=
2a3
µ ¶
ρ0 2
(for x2 + y2 + z2 > a2 ),


 −a + 2
4π ǫ0 (x + y2 + z2)1/2
(2.122)

where ρ0 is a quantity with the dimensions of volume charge den-

sity. Note that we are not assuming that φ = 0 at infinity.
2.73 Satisfying Laplace *
Does the function f (x, y) = x2 + y2 satisfy the two-dimensional
Laplace’s equation? Does the function g(x, y) = x2 − y2 ? Sketch
the latter function, calculate the gradient at the points (x, y) =
(0, 1), (1, 0), (0, −1), and (−1, 0), and indicate by little arrows the
directions in which these gradient vectors point.
2.74 Oscillating exponential φ ***
A flat nonconducting sheet lies in the xy plane. The only charges
in the system are on this sheet. In the half-space above the sheet,
122 The electric potential

z > 0, the potential is φ = φ0 e−kz cos kx, where φ0 and k are con-
stants.
(a) Verify that φ satisfies Laplace’s equation in the space above
the sheet.
(b) What do the electric field lines look like?
(c) Describe the charge distribution on the sheet.
2.75 Curls and divergences *
Calculate the curl and the divergence of each of the following vec-
tor fields. If the curl turns out to be zero, try to discover a scalar
function φ of which the vector field is the gradient.
(a) F = (x + y, −x + y, −2z);
(b) G = (2y, 2x + 3z, 3y);
(c) H = (x2 − z2 , 2, 2xz).
2.76 Zero curl *
By explicitly calculating the components of ∇ × E, show that the
vector function specified in Exercise 2.31 is a possible electrostatic
field. (Of course, if you worked that exercise, you have already
proved it in another way by finding a scalar function of which it is
the gradient.) Evaluate the divergence of this field.
2.77 Zero dipole curl *
Verify that the curl of the dipole field in Eq. (2.36) is zero. We
know that it must be zero, of course, because the field is the sum
of the fields from two point charges, but demonstrate this here
by explicitly calculating the curl, using the expression given in
Eq. (F.3) in Appendix F.
2.78 Divergence of the curl **
If A is any vector field with continuous derivatives, div (curl A) =
0 or, using the “del” notation, ∇ · (∇ × A) = 0. We shall need
this theorem later. The problem now is to prove it. Here are two
S
different ways in which that can be done.
(a) (Uninspired straightforward calculation in a particular coor-
dinate system.) Using the formula for ∇ in Cartesian coordi-
nates, work out the string of second partial derivatives that
∇ · (∇ × A) implies.
C (b) (With the divergence theorem and Stokes’ theorem, no coordi-
nates are needed.) Consider the surface S in Fig. 2.52, a bal-
loon almost cut in two which is bounded by the closed curve
C. Think about the line integral, over a curve like C, of any
vector field. Then invoke Stokes and Gauss with suitable argu-
ments. (The reasoning also works if the curve C is a very tiny
Figure 2.52. loop on the surface.)
 Exercises 123

2.79 Vectors and squrl *

To show that it takes more than direction and magnitude to make a
vector, let’s try to define a vector, which we’ll name squrl F, by a
relation like Eq. (2.80) but with the right-hand side squared:
" R #2
Ci F · ds
(squrl F) · n̂ = lim . (2.123)
ai →0 ai
Prove that this does not define a vector. (Hint: Consider reversing
the direction of n̂.)