PGF5003: Classical Electrodynamics I

Problem Set 1
Professor: Luis Raul Weber Abramo
Monitor: Natalí Soler Matubaro de Santi
natali.santi@usp.br
(Due to April 13, 2021)

1 Question
~ with f and g scalar functions, prove the first Green identity:
Given the following vector field F = f ∇g,
Z   I  
2 ~ ~
dV f ∇ g + ∇f · ∇g = ~ ~
dS · f ∇g . (1)
V S(V )

What do we need to suppose about f and g to deduce this identity?

1.1 Solution
~ its divergence is given by
Considering the vector field F = f ∇g,
~ ·F=∇
∇ ~ · (f ∇g)
~ = ∇f
~ · ∇g
~ + f ∇2 g. (2)

Integrating the above relation in the volume V we get

Z h i Z Z
~ ~
dV ∇f · ∇g + f ∇ g = 2 ~
dV ∇ · F = ~ · (f ∇g).
dV ∇ ~ (3)
V V V

Then, using the divergence theorem for the field F

Z I
~
dV ∇ · F = dS · F (4)
V S(V )

we have Z   I  
2 ~ ~
dV f ∇ g + ∇f · ∇g = ~
dS · f ∇g . (5)
V S(V )

Basically, f needs to be once continuously differentiable (C 1 ) and g needs to be twice continuously

differentiable (C 2 ).

2 Question
Show that  
2 1
∇ = −4πδ(r). (6)
r

1
2.1 Solution
If we take the gradient of 1/r we have

1 r̂ r
∇ = − 2 = − 3, (7)
r r r
for r 6= 0. Then, we can write  
2 1  r
∇ = ∇ · − 3 = 0, (8)
r r
when r 6= 0. However, integrating the above quantity in all the space (in spherical coordinates) and
using the divergence theorem, we get
Z   Z I
2 1  r r
dV ∇ = dV ∇ · − 3 = − dS · 3
V r V r S(V ) r
Z 2π Z π
r
=− dφ dθ sin θr2 r̂ · 3 = −4π. (9)
0 0 r
Therefore, the value of the Laplacian is zero everywhere except zero and the integral over any
volume containing the origin is equal to −4π. In this way,
 
2 1
∇ = −4πδ(r). (10)
r

3 Question
Using the second Green identity:
Z I  
2 2 ~ − g ∇f
~


dV f ∇ g − g∇ f = dS · f ∇g , (11)
V S(V )

~ and g = 1/R = 1/|x − x0 |, show that

taking f = φ(x0 ) (for the electrostatic potential E = ∇φ)
0
Z I  
3 0 ρ(x ) 1 0 1 ~0 0 0 ~0 1
φ(x) = dx + dS · ∇ φ(x ) − φ(x )∇ , (12)
V R 4π S(V ) R R

~ 0 corresponds to differential operation related to x0 .

where ∇

3.1 Solution
We can substitute f and g into the second Green identity as
Z   I  
0 0 02 1 1 02 0 0 ~0 1 1 ~0 0
dV φ(x )∇ − ∇ φ(x ) = dS · φ(x )∇ − ∇ φ(x ) . (13)
V R R S(V ) R R

Here we need to use that

0 1 0 1
∇2 =∇2 = −4πδ(x − x0 ) (14)
R |x − x0 |

2
and the first Maxwell equation
~ · E = −∇2 φ(x) = 4πρ(x)
∇ (15)

then,
4πρ(x0 )
Z   I  
0 0 0 0 ~0 1 1 ~0 0
dV −4πφ(x )δ(x − x ) + = dS · φ(x )∇ − ∇ φ(x ) . (16)
V R S(V ) R R

Rearranging the above expression, we arrive that

0
Z I  
3 0 ρ(x ) 1 0 1 ~0 0 0 ~0 1
φ(x) = dx + dS · ∇ φ(x ) − φ(x )∇ . (17)
V R 4π S(V ) R R

4 Question
Consider the electric field
Aer/r0
E= r̂. (18)
r
a) Determine the density of charge.

b) Determine the total charge into a radius R.

4.1 Solution
~ · E = 4πρ, we have
a) Using the first Maxwell equation ∇

~ ·E
∇

1 1 ∂
 r/r0

1 Aer/r0 1
 
2 Ae 1
ρ= = r = + . (19)
4π 4π r2 ∂r r 4π r r r0

b) The total charge is given just integrating the previous item in spherical coordinates as
Z Z
Q= dq = dV ρ
V (R) V (R)
2π π R
1 Aer/r0 1
Z Z Z  
2 1
= dφ dθ sin θ drr +
0 0 0 4π r r r0
Z R Z R 
r/r0 r r/r0
=A dre + dr e
0 0 r0
R/r0
− 1 + (R − r0 ) eR/r0 + r0

 
= A r0 e
= AReR/r0 . (20)

3
5 Question
Consider two infinite plates (with zero thickness) with the distributions of charge σ and −σ, respectively.
The plates are orthogonal to each other.

−σ +σ

IV II x

III

Figure 1: Figure for the question 5.

a) Find the electric field in all the regions (I, II, III, IV and total space).

b) Draw a figure, representing the electrical field.

5.1 Solution
a) Considering the Gauss theorem together with the first Maxwell’s equation we have
Z Z I
~
dV ∇ · E = dV 4πρ = dS · E. (21)
V V S(V )

Then, applying this to a cylinder with superficial density σ and superior and inferior area A

4πσA = 2A|E|
E = 2πσn̂. (22)

This means that the plate with −σ will have the electrical field entering on itself and the other, with
σ, have it pointing out of the plane. Considering a vector sum, and the normals written in terms of

4
the angles with the planes and the axes, we just end up with:
√
EI = −2πσ (sin θ + cos θ, sin θ − cos θ, 0)θ=45◦ = −2πσ 2x̂ (23)
√
EII = −2πσ (cos θ − sin θ, sin θ + cos θ, 0)θ=45◦ = −2πσ 2ŷ (24)
√
EIII = 2πσ (cos θ + sin θ, sin θ − cos θ, 0)θ=45◦ = 2πσ 2x̂ (25)
√
EIV = 2πσ (cos θ − sin θ, sin θ + cos θ, 0)θ=45◦ = 2πσ 2ŷ (26)
Etotal = 0. (27)

b)

−σ +σ

IV II x

III

Figure 2: Figure for the solution of item b, question 5.

6 Question
A spherical shell of radius R is made with isolating material and has a surface density of charge σ

2
(that, in principle, we do not know). The electric potential outside the sphere is Vout (r) = V0 Rr cos θ,
where V0 is a constant. The electric field Eins (r) = −VR
0
ẑ. Compute:

a) the electric field outside the sphere Eout (r) and the electric potential inside the sphere Vins ;

b) the superficial density of charge σ;

c) the force per unity of area f over the surface of the sphere;

6.1 Solution
a) To compute the electrical field outside, we can follow its definition according to the gradient of

5
the potential outside the sphere as
"   # "   #
2 2
~ out ∂ R 1 ∂ R
Eout (r) = −∇V =− V0 cos θ r̂ − V0 cos θ θ̂
∂r r r ∂θ r
R2 R2
= 2V0 cos θ r̂ + V0 sin θ θ̂. (28)
r3 r3
The electric potential inside the sphere is giving performing a line integral, “inverting” the pre-
vious relation in the region inside the sphere
R R
−V0
Z Z
0 0
Vout (R) − Vins (r) = − dr · Eins (r ) = − dr0 · ẑ
r r R
Z R
V0 cos θ V0 cos θ
V0 cos θ − Vins (r) = dr0 = (R − r)
r R R
r
Vins (r) = V0 cos θ . (29)
R
Notice that we use V (r = R) to match with the exterior potential on r = R.

b) We know that, if we consider the continuity equation of the electric field we get the superficial
density of charge, then,
1
σ= [Eout (r) − Eins (r)]r=R · n̂
4π 
R2 R2

1 V0
= 2V0 cos θ 3 r̂ + V0 sin θ 3 θ̂ + ẑ · r̂
4π r r R
 
1 1 V0 1 3V0 cos θ
= 2V0 cos θ + 0 + cos θ = . (30)
4π R R 4π R

c) The force could be computed as

σ [Eout (r) + Eins (r)]

f=
2 r=R

3V0 cos θ 1 1 V0
= 2V0 cos θ r̂ + V0 sin θ θ̂ − ẑ
8πR R R R
2
3V cos θ h i
= 0 2 2 cos θr̂ + sin θθ̂ − ẑ (31)
8πR

7 Question
Consider a potential problem in the half-space defined by z ≥ 0, with Dirichlet boundary conditions on
the plane z = 0 (and at infinity).

a) Write down the appropriate Green’s function G(r, r0 ).

b) If the potential on the plane z = 0 is specified to be Φ = V inside a circle of radius a centered

at the origin, and Φ = 0 outside that circle, find an integral expression for the potential at a point P
specified in terms of cylindrical coordinates ρ, φ, z.

6
 c) Show that, along the axis of the circle (ρ = 0), the potential is given by
 
z
Φ=V 1− √ . (32)
a2 + z 2

d) Show that at large distances (ρ2 + z 2  a2 ) the potential can be expanded in a power series in
(ρ + z 2 )−1 , and that the leading terms are
2

V a2 3a2 5(3ρ2 a2 + a4 )
 
z
Φ= 1− + + ... . (33)
2 (ρ2 + z 2 )3/2 4(ρ2 + z 2 ) 8(ρ2 + z 2 )2

Verify that the result of (c) is consistent with this results.

7.1 Solution
a) We need to find a Green function of the form
1
G(r, r0 ) = + H(r − r0 ) (34)
|r − r0 |

that follows ∇2 Φ = −4πρ with Dirichlet boundary conditions Φz=0 = α and Φr=∞ = β.
This problem could be solved by the method of the images, and be thought as a situation where
we have a point of charge q at r0 , which creates a potential of the form 1/|r − r0 | in the presence of a
flat conductor in the plane z = 0. Then, putting the image charge at (x0 , y 0 , −z 0 ) (such as the plane
was a “mirror”), the potential Φ could be written as

q q0
Φ= p +p (35)
(x − x0 )2 + (y − y 0 )2 + (z − z 0 )2 (x − x0 )2 + (y − y 0 )2 + (z + z 0 )2

Taking Φz=0 = 0, leads to the conclusion that q 0 = −q.

In this way, due to the Green method the appropriate Green function is
1 1
G(r, r0 ) = p −p . (36)
(x − x0 )2 + (y − y 0 )2 + (z − z 0 )2 (x − x0 )2 + (y − y 0 )2 + (z + z 0 )2

b) Now, the potential on the plane, i.e., z = 0 could be written as


V, x2 + y 2 ≤ a
Φz=0 = (37)
0, x2 + y 2 > a

and, we are going to use the Green’s theorem

Z Z  
0 0 0 1 0 ∂G ∂Φ
Φ(r) = dV ρ(r )G(r, r ) + dS Φ 0 − G 0 . (38)
V 4π S(V ) ∂n ∂n

Here the surface is composed by the plane (at z = 0) and the infinity. It has zero potential at
infinity and all other regions outside the circle. Then, it only remains the circle in the integration.
It follows that the normal vector points to −ẑ and we have
∂G ∂G 2z
= − = . (39)
∂n0 z0 =0 ∂z 0 z0 =0 [(x − x0 )2 + (y − y 0 )2 + (z − z 0 )2 ]3/2

7
Replacing on Φ (remembering that ρ(r0 ) = 0 and ∂n∂Φ
0 = 0) we obtain
Z
1 2V z
Φ(r) = dS 0 . (40)
4π S(V ) [(x − x ) + (y − y 0 )2 + z 2 ]3/2
0 2

This integral could be changed to cylindrical coordinates: [(x, y, z), (x0 , y 0 , z 0 )] ⇒ [(ρ, φ, z), (ρ0 , φ0 , z 0 )],
using
(x−x0 )2 +(y−y 0 )2 = ρ2 +(ρ0 )2 −2ρρ0 [cos φ cos φ0 + sin φ sin φ0 ] = ρ2 +(ρ0 )2 −2ρρ0 cos(φ−φ0 ) (41)
and we get
a 2π
dφ0
Z Z
zV 0 0
Φ(ρ, φ, z) = dρ ρ . (42)
2π 0 0 [ρ2 + (ρ0 )2 − 2ρρ0 cos(φ − φ0 ) + z 2 ]3/2
c) Here, we can solve the previous integral for ρ = 0. In this way this stays
zV a 0 0 2π dφ0
Z Z
Φ(z) = dρ ρ
2π 0 0 [(ρ0 )2 + z 2 ]3/2
Z ρ0 =a  
zV du z
= =V 1− √ (43)
2 ρ0 =0 u3/2 a2 + z 2
where we have changed the variables using u = [(ρ0 )2 + z 2 ].
d) Once again we use the integral solution, but now we want to do an expansion for power series
in (ρ2 + z 2 )−1 . One way to do this is as follows
zV a 0 0 2π dφ0 (ρ2 + z 2 )3/2
Z Z
Φ(ρ, φ, z) = dρ ρ
2π 0 0 [ρ2 + (ρ0 )2 − 2ρρ0 cos(φ − φ0 ) + z 2 ]3/2 (ρ2 + z 2 )3/2
Z a Z 2π 0 2 0 0 2 −3/2
 2 
zV 0 0 0 ρ + (ρ ) − 2ρρ cos(φ − φ ) + z
= dρ ρ dφ
2π(ρ2 + z 2 )3/2 0 0 (ρ2 + z 2 )
Z a Z 2π −3/2
(ρ0 )2 − 2ρρ0 cos(φ − φ0 )

zV 0 0 0
= dρ ρ dφ 1 + . (44)
2π(ρ2 + z 2 )3/2 0 0 (ρ2 + z 2 )
Taking the expansion
n(n − 1)x2
(1 + x)n = (1 + nx) + + ..., (45)
2
0 2 0 cos(φ−φ0 )
with x = (ρ ) −2ρρ
(ρ2 +z 2 )
, n = −3/2 and n(n − 1) = 15/4 we have
Z a Z 2π
3 (ρ0 )2 − 2ρρ0 cos(φ − φ0 )
  
zV 0 0 0
Φ(ρ, φ, z) = dρ ρ dφ 1 −
2π(ρ2 + z 2 )3/2 0 0 2 (ρ2 + z 2 )
2 )
15 (ρ0 )2 − 2ρρ0 cos(φ − φ0 )

+
8 (ρ2 + z 2 )
V a2 3a2 5(3ρ2 a2 + a4 )
 
z
= 1− + + ... . (46)
2 (ρ2 + z 2 )3/2 4(ρ2 + z 2 ) 8(ρ2 + z 2 )2
Notice that, the above integral is easier if we notice that there is azimuthal symmetry and you could
simply use (φ − φ0 ⇒ φ0 ).
Verifying what happens for ρ = 0, we have (expanding for (1 + x)−1/2 = 1 − x/2 + 3x3 /8 −
5x3 /16 + . . . , with x = (a/z)2 )
V a2 z 3a2 5(a4 )
 
Φ(ρ, φ, z) = 1− + + ...
2 (z 2 )3/2 4(z 2 ) 8(z 2 )2
 
z
=V 1− √ , (47)
a2 + z 2

8
which is a result equivalent to item (c) because the expansion (1 + x)−1/2 for x = a2 /z 2 .

9
8 Question
An infinite metallic plate has a spherical overhang of radius a. This plate is grounded. A charge +q is
placed over the hemisphere of the overhang, with a distance d of the center of the sphere. Show that the
induced charge on the overhang is
(d2 − a2 )
 
0
q = −q 1 − √ . (48)
d d 2 + a2

S2

S1
y q
x

Figure 3: Figure for the question 5.

8.1 Solution
This problem could be solved by the Method of the images, as we have the superposition of a
infinite plane and a sphere of radius a, as indicated in the figure.
Solving the Green function to the plane, considering the charge q on r0 and the image q 0 on
(−x0 , y 0 , z 0 ), we can write the potential
q q0
Φ= p +p . (49)
(x − x0 )2 + (y − y 0 )2 + (z − z 0 )2 (x + x0 )2 + (y − y 0 )2 + (z − z 0 )2

10
 S2

S1
y q
x

Figure 4: Infinite plane with the spherical overhang of radius a and a charge q on dx̂.

Taking Φx=0 = 0, we can say that q 0 = −q.

Now you can think that, the sphere “see” two sources, given by the solution of the plane. Thus,
we have to create two other image charges: one for the real charge: qR0 , on r0RR and other for the first
image qI0 , on r0II . Because it is easier, we can write the potential already in the spherical coordinates
as
q q
Φ= p 0 2 0
−p
2
r + (rR ) − 2rrR cos γR r + (rI ) + 2rrI0 cos γI
2 0 2

qR0 qI0
+p 0 0
+ p
0 2 0
(50)
r2 + (rRR )2 − 2rrRR cos γR r2 + (rII ) − 2rrII cos γI

and take it on r = a, such that Φ(r = a) = 0. Then, one terms equals the other (remembering that
0 0
all the charges will be in the same line, then the angles are the same between: (rR , r) and (rRR , r);
0 0
(rI , r) and (rII , r)), we have:
0 2 0
[a2 + (rR ) − 2a(rR ) cos γR ](qR0 )2 = q 2 [a2 + (rRR
0 0
)2 − 2arRR cos γR ] (51)
2 0 2 0 0 2 2 2 0 2 0
[a + (rI ) + 2arI cos γI ](qI ) = q [a + (rII ) − 2arII cos γI ]. (52)

11
We need to solve the following systems
0 2
[a2 + (rR ) ](qR0 )2 0
= q 2 [a2 + (rRR )2 ] (53)
[a2 + (rI0 )2 ](qI0 )2 0 2
= q 2 [a2 + (rII )] (54)
0
−2a(rR ) cos γR (qR0 )2 2 0
= −q 2a(rRR ) cos γR (55)
2a(rI0 ) cos γI (qI0 )2 = −q 2 2a(rII0
) cos γI (56)

finding that
qR0 = −aq/rR
0 0
, rRR 0
= a2 /rR and qI0 = aq/rI0 , rII
0
= −a2 /rI0 . (57)
Finally, taking the position of the real charge as dx̂, the potential is written as
q q
Φ(r) = p −p
r2 + d2 − 2rd cos γ r2 + d2 + 2rd cos γ
qa/d qa/d
−q +q . (58)
2 2 2 2 2 a2
r2 + ad − 2r ad cos γ r2 + ad + 2r





d
cos γ

Having the electric potential, σ is given, on r = a, due

1 1 ∂Φ
σ = − ∇ΦS1 = −
 4π 4π ∂r r=a
q(r − d cos γ) q(r + d cos γ)
= 2 2 3/2
− 2
(r + d − 2rd cos γ) (r + d2 + 2rd cos γ)3/2

2 2
qa/d(r − a /d cos γ) qa/d(r + a /d cos γ)
−h i3/2 + h

2 2
i3/2 
2 2 2 2
r2 + ad − 2r ad cos γ r2 + ad + 2r ad cos γ





r=a
q 2 2 q 2 2
(a − d ) (a − d )
 
1 a
=− 2 2 3/2
− 2 a2 . (59)
4π (d + a − 2ad cos γ) (d + a + 2ad cos γ)3/2

Thus, we can find the total charge integrating over the surface
Z 2π Z π q
(a2 − d2 )
Z 
a 1
Q = dSσ = − dφ dγ sin γ
0 0 4π (d + a + 2ad cos γ)3/2
2 2

1
− 2
(d + a2 − 2ad cos γ)3/2
(d2 − a2 )
 
= −q 1 − √ . (60)
d d2 + a2

of course after change the variables u = a2 + d2 + 2ad cos γ.

12