thicker letters

Problem 5.46 3µ0 IR2

− (d2 − R2 ).
(a) Find the field (B) as a function of [R2 + (d/2)2 ]7/2
∂  −
z, and show that ∂B/∂z is zero at the ∂2B + (x2 − x1 )2 + (y2 − y1 )2 + (z2 − z1 )2
Set ∂z 2= 0. This holds when d = R. ∂z2
point midway between them (z = 0).
Now, if you pick d just right the sec- Substituting d = R:
ond derivative of B will also vanish
at the midpoint. This arrangement is (
1 1 (x2 − x1 ) (y2 − y1 ) (z2 − z1 )
known as a Helmholtz coil; it’s a con- µ0 IR2 [R2 +(R/2)2 ]3/2 =
[R2 +(R/2)2 ]3/2 − x̂− ŷ− ẑ
B(0) = 1 1 z3 z3 z3
venient way of producing relatively uni- 2 [R2 +(R/2)2 ]3/2 [R2 +(R/2)2 ]3/2
form fields in the laboratory.
(b) Determine d such that ∂ 2 B/∂z 2 = 0 I I  
at the midpoint, and find the resulting  σy  z x̂ x̂ 1
2 1 5R2 =− =− 2 ⇒ ·dΩ = − ∇2 ·d
magnetic
√ field at the center. [Answer: = µ0 IR = √ . z 3 z ẑ 2 z
(5R2 /4)3/2 5 3 2R
8µ0 I/5 5R]
A steady current I; we want to calcu-
Solution Final Answer late the magnetic field at a point in the
8µ0 I plane. We might as well take that point
(a) √ to be the origin (it could be inside or
Use superposition to add the two fields 5 5R outside the loop). The shape of the wire
together: is given, in polar coordinates, by a spec-
Problem 5.49 ified function r(θ).
( ”recipient current” Show, starting with (a)
µ0 IR2 1 the Biot-Savart law (5.32) and the
B= + Show that the magnitude of the field is:
2 [R2 + (d/2 + z)2 ]
3/2 Lorentz force law (5.16), that the force
on loop 2 due to loop 1 (Fig. 5.62) can I
be written as µ0 I dθ
1 B=
3/2
4π r
[R2 + (d/2 − z)2 ]
∂B µ0
I I
ĥ Conceptual Input ¿
Compute the first derivative ∂z : F2 = − I1 I2 dl1 · dl2 .
h2 dl × r̂ dl × r̂
I I
4π µ0 I µ0 I
B= 2
=−
4π r̂ 4π r2
(5.88)
(
∂B µ0 IR2 (−3/2)2(d/2 + z)
= In this form it is clear that F2 = −F1 , µ0 I
I
dθ
∂z 2 5/2
[R2 + (d/2 + z)2 ] since ĥ changes direction when the roles =
4π r
of 1 and 2 are interchanged. (If you
(−3/2)2(d/2 − z)(−1) seem to be getting an ”extra” term, it
+ 5/2
(b)
[R2 + (d/2 − z)2 ] will help to note that dl2 · ĥ = dl1 .) For a circular loop, r = R is constant.
( Solution µ0 I
I
dθ µ0 I 2π
Z
3µ0 IR2 −(d/2 + z) From the Biot-Savart law, the field of B= = dθ
= 5/2 4π R 4π 0
2 [R2 + (d/2 + z)2 ] loop 1 is:
µ0 I1
I
dl1 × x̂ µ0 I µ0 I
(d/2 − z) B= , = [2π̂] = .
+ 4π 1 z 2 4πR 2R
5/2
[R2 + (d/2 − z)2 ]
the force on loop 2 is then: A steady current I; we want to calcu-
Now evaluate this at the midpoint (z = late the magnetic field at a point in the
0): plane. We might as well take that point
dl2 × (dl1 ×tox̂)be the origin (it could be inside or
I I I
µ0
F = I2 dl2 ×B = I1 I2
(
2 4π 1 2 z 2 outside the loop). The shape of the wire
∂B 3µ0 IR2 −d/2 is given, in polar coordinates, by a spec-
= 5/2 ified function r(θ).
∂z z=0 2 [R2 + (d/2)2 ]
⇒ dl2 ×(dl1 ×x̂) = dl1 (dl2 ·x̂)−x̂(dl1 ·dl2 ) (a)
d/2 Show that the magnitude of the field is:
+ 5/2
=0
[R2 + (d/2)2 ] I
µ0
I I
x̂ µ0 I dθ
F = − I1 I2 (dl · dl ) B =
(b) 4π z2
1 2 4π r
2 2
∂2B
Differentiate again to find ∂z 2 : Conceptual Input ¿
(dl2 · x̂)
I I
− dl1 µ0 I
I
dl × r̂ µ0 I
I
dl × r̂
2 2 z2 B= =−
∂2B 3µ0 IR2

−1 −(d/2 + z)(−5/2)2(d/2 + z) 4π r̂2 4π r2
=
∂z 2 2 [R2 + (d/2 + z)2 ]5/2 [R2 that the+second
+ (d/2 z)2 ]7/2term is zero: I
µ0 I dθ
−1 =
+ z = (x2 −x1 )x̂+(y2 −y1 )ŷ +(z2 −z1 )ẑ = 4π r
[R2 + (d/2 − z)2 ]5/2
(b)
(d/2 − z)(−5/2)2(d/2 − z)(−1) γ2 (1/z)
+ . For a circular loop, r = R is constant.
[R2 + (d/2 − z)2 ]7/2
∂  −1/2
(x2 − x1 )2 + (y2 − y1 )2 + (z2 − z1 )2

x̂ µ0 I 2π
I Z
Evaluate at z = 0: µ0 I dθ
∂x2 B= = dθ
4π R 4π 0
∂2B 3µ0 IR2 d2 5d2∂ 
 
= −R2 − ++ (x2 − x1 )2 + (y2 − y1 )2 + (z2 − z1 )2
−1/2
ŷ µ0 I µ0 I
∂z 2 [R2 + (d/2)2 ]7/2 4 4∂y2 = [2π̂] = .
z=0 4πR 2R
Problem 5.54 Solution Solution
Prove the following uniqueness theo- To compute V , we need to know E; to This is reminiscent of Ex. 3.8, in which
rem: If the current density J is specified find E, we might first try to locate the an uncharged conducting sphere was in-
throughout a volume V, and either the bound charge; we could get the bound troduced into a uniform field. In that
potential A or the magnetic field B is charge from P , but we can’t calculate P case the field of the induced charge com-
specified on the surface S bounding V, unless we already know E (Eq. 4.30). pletely canceled E0 within the sphere; in
then the magnetic field itself is uniquely We seem to be in a bind. What we a dielectric, the cancellation (from the
determined throughout V. [Hint: First do know is the free charge Q, and for- bound charge) is only partial.
use the divergence theorem to show that tunately the arrangement is spherically Our problem is to solve Laplace’s equa-
symmetric, so let’s begin by calculating tion, for Vin (r, θ) when r ≤ R, and
Z D, using Eq. 4.23: Vout (r, θ) when r ≥ R, subject to the
{(∇×U )·(∇×V )−U ·[∇×(∇×V )]} dτ boundary conditions
Q
I D= r̂, for all points r > a. (i) Vin = Vout , at r = R,
= [U × (∇ × V )] · da, 4πr2
(ii) ϵ ∂V ∂Vout
∂r = ϵ0 ∂r ,
in
at r = R,
(Inside the metal sphere, of course, E = (iii) Vout → −E0 r cos θ, for r ≫ R.
for arbitrary vector functions U and V . P = D = 0.) Once we know D, it is
a trivial matter to obtain E, using Eq. (4.43)
Solution 4.32: (The second of these follows from Eq.
For arbitrary vector functions U and V . 4.41, since there is no free charge at the
Apply the divergence theorem to the surface.) Inside the sphere Eq. 3.65 says
 Q
function [U × (∇ × V )], 4πϵr 2 r̂, for a < r < b,
noting (from the product rule) that E= Q ∞
4πϵ0 r 2 r̂, for r > b. X
Vin (r, θ) = Al rl Pl (cos θ);
l=0
The )]potential at the center is therefore
∇·[U ×(∇×V )] = (∇×V )·(∇×U )−U ·[∇×(∇×V
Z 0 (4.44)
outside the sphere, in view of (iii), we
V =− E · dl have
Z Z −∞
∇·[U ×(∇×V )]dτ = {(∇×V )·(∇×U ) Z b   Z a  Z 0
Q Q ∞
=− dr− dr− (0)dr X Bl
−∞ 4πϵ 0 r 2
b 4πϵr 2 V
a out (r, θ) = −E 0 r cos θ+ Pl (cos θ).
−U · [∇ × (∇ × V )]}dτ rl+1
  l=0
Q 1 1 1
= + − . (4.45)
Z I 4π ϵ0 b ϵa ϵb
⇒ ∇·[U ×(∇×V )]dτ = [U ×(∇×V )]·da. Boundary condition (i) requires that
As it turns out, it was not necessary for
us to compute the polarization or the ∞
To exit full screen, press Exc bound charge explicitly, though this can X
A1 and B2 (and A2 ). Define B3 = easily be done: Al Rl Pl (cos θ) = −E0 R cos θ
B2 − B1 (and l=0
ϵ0 χ e Q
A3 ≡ A2 − A1 P = ϵ0 χe E = r̂, ∞
Bl
4πϵr2
X
+ P (cos θ),
l+1 l
in the dielectric, and hence R
, so that ∇ × A3 = B3 and l=0

∇×B3 = ∇×B1 −∇×B2 = µ0 J−µ0 J = 0. ρb = −∇ · P = 0, so10

while Bl
Set U = V = A3 in the above identity: Al Rl = l+1 , for l ̸= 1,
Z R
{(∇×A3 )·(∇×A3 )−A3 ·[∇×(∇×A3 )]} dτ B1
 ϵ 0 χe Q
σb = P·n̂ = −ϵ 4πϵb2 , at the outer surface, A1 R = −E0 R + 2 .
0 χe Q R
Z 4πϵa2 , at the inner surface.
=

(B3 ) · (B3 ) − A3 · [∇ × B3 ]−0
dτ Meanwhile, condition (ii) yields
Notice that the surface bound charge at
∞
a is negative (n̂ points outward with X
lAl Rl−1 Pl (cos θ)
Z
2 respect to the dielectric, which is +r̂ ϵr
= (B3 ) dτ
at b but −r̂ at a). This is natural, l=0
I I since the charge on the metal sphere ∞
= [A3 ×(∇×A3 )]·da = (A3 ×B3 )·da. attracts its opposite in all the dielec-
X (l + 1)Bl
= −E0 cos θ − Pl (cos θ),
tric molecules. It is this layer of nega- Rl+2
l=0
But either A is specified (in which case tive charge that reduces the field, within
the dielectric, from 1/4πϵ0 (Q/r2 )r̂ to so
A3 = 0), or else B is specified (in which 2
case B3 = 0), at the surface. In either 1/4πϵ(Q/r )r̂. In this respect a dielec-
case tric is rather like an imperfect conduc- (l + 1)Bl
tor: on a conducting shell the induced ϵr lAl Rl−1 = − , for l ̸= 1,
I Z
surface charge would be such as to can- Rl+2
2
(A3 × B3 ) · da = 0 ⇒ (B3 ) dτ = 0.
cel the field of Q completely in the re-
gion a < r < b; the dielectric does the 2B1
ϵr A1 = −E0 − 3 .
and hence B1 = B2 . qed best it can, but the cancellation is only R
partial. It follows that
Problem
A metal sphere of radius a carries a Problem Al = Bl = 0, for l ̸= 1,
charge Q (Fig. 4.20). It is surrounded, A sphere of homogeneous linear dielec-
out to radius b, by linear dielectric mate- tric material is placed in an otherwise
rial of permittivity ϵ. Find the potential uniform electric field E0 (Fig. 4.27). 3 ϵr − 1 3
at the center (relative to infinity). Find the electric field inside the sphere. A1 = − E0 B1 = R E0 .
ϵr + 2 ϵr + 2
 qb
Evidently +p , (4.52)
x2 + y2 + (z + d)2 Bl Bl
⇒ l+1
= Al bl + l+1 ⇒
b b
3E0 3E0 in the region z > 0. Meanwhile, a
Bl = Al b2l+1 −a2l+1 Al = Al b2l+1 − a2l+1


Vin (r, θ) = − r cos θ = − z, charge (q + qb ) at (0, 0, d) yields the po-
ϵr + 2 ϵr + 2
tential
(ii)
and hence the field inside the sphere is
−a2l+1 Al
 
(surprisingly) uniform: " # Bl
1 q + qb ϵr lAl bl−1 − (l + 1) = (l+1) l+2
3 V = p , (4.53) bl+2 b
E= E0 . 4π ϵ0 x2 + y 2 + (z − d)2
ϵr + 2  
for the region z < 0. Taken together, b 2l+1
Problem ⇒ Bl = −ϵr Al b + a2l+1
Eqs. 4.52 and 4.53 constitute a func- l+1
Suppose the entire region below the tion which satisfies Poisson’s equation
plane z = 0 in Fig. 4.28 is filled with with a point charge q at (0, 0, d), which ⇒ Al = Bl = 0.
uniform linear dielectric material of sus- goes to zero at infinity, which is contin- For l = 1:
ceptibility χe . Calculate the force on uous at the boundary z = 0, and whose
a point charge q situated a distance d normal derivative exhibits the disconti-
above the origin. nuity appropriate to a surface charge σb B1
(i) − E0 bP1 (cos θ) + 2 P1 (cos θ)
at z = 0:
Solution  b
−a3 A1

 
∂V ∂V = A1 b + P1 (cos θ)
The surface bound charge on the xy −ϵ0 − b2
plane is of opposite sign to q, so the force ∂z z=0+ ∂z z=0−
will be attractive. (In view of Eq. 4.39, B1 a3 A1
1

χe

qd ⇒ −E0 b + 2 = A1 b − 2
there is no volume bound charge.) Let =− . b b
us first calculate σb , using Eqs. 4.11 and 2π χe + 2 (x + y + d2 )3/2
2 2
⇒ −E0 b3 + B1 = A1 b3 − a3 A1
4.30.
Accordingly, this is the correct potential ⇒ B1 − E0 b3 = A1 (b3 − a3 )
σb = P · n̂ = Pz = ϵ0 χe Ez , for our problem. In particular, the force h i
3
on q is: (ii) ϵr A1 + 2 a bA
3
1
= −E0 − 2 Bb31
where Ez is the z-component of the total
field just inside the dielectric, at z = 0. 1 qqb
F= ẑ
⇒ ϵr A1 b3 + 2a3 = −E0 b3 − 2B1
 
This field is due in part to q and in 4π ϵ0 (2d)2
part to the bound charge itself. From   2
Coulomb’s law, the former contribution 1 χe q 2) Eliminate B1 by adding these to-
is =− ẑ. (4.54) gether as: 2(i) + (ii)
4π ϵ0 χe + 2 4d2

Problem (B1 − E0 b3 ) + (−E0 b3 − 2B1 )

1 q 1 qd
− cos θ = − = 2(A1 (b3 − a3 )) + (ϵr A1 b3 + 2a3 )
 
,
4πϵ0 (r2 + d2 ) 4πϵ0 (r2 + d2 )3/2An uncharged conducting sphere of ra-
dius a is coated with a thick insulating
p
2 2
shell (dielectric constant ϵr ) out to ra- ⇒ 2B1 − 2E0 b3 − E0 b3 − 2B1
where r = x + y is the distance dius b. This object is now placed in
= 2A1 (b3 − a3 ) + ϵr A1 b3 + 2a3
 
from the origin. The z component of the an otherwise uniform electric field E0 .
field of the bound charge, meanwhile, is Find the electric field in the insulator.
−σb /2ϵ0 (see footnote 6, p. 89). Thus
Solution
r cos θ +
rB
   P l
1 qd σb Vout (r, θ) = −E P0 l+1 Pl (cos θ), (r > b);
σ b = ϵ0 χ e − − ,
4πϵ0 (r2 + d2 )3/2 2ϵ0 Vmed (r, θ) = Al rl + rB l
l+1 Pl (cos θ), (a < r < b);
Vin (r, θ) = 0, (r < a).

which we can solve for σb :
Boundary Conditions: ⇒ −3E0 b3 = A1 2(b3 − a3 ) + ϵr (b3 + 2a3 )
 

1

χe

qd −3E0
σb = − . (4.50)
 ⇒ A1 =
(i) Vout = Vmed , (r = b);
h i h i
a

3 a 3


2π χe + 2 (r2 + d2 )3/2 2 1− + ϵr 1 + 2
(ii) ϵ ∂V∂r
med
= ϵ0 ∂V∂rout , (r = b); b b

Apart from the factor χe /(χe +2), this is So the potential is:
exactly the same as the induced charge (iii) Vmed = 0, (r = a).
on an infinite conducting plane under 
a3 A1

similar circumstances (Eq. 3.10).[11] Applying the boundary conditions Vmed (r, θ) = A1 r − 2 P1 (cos θ) =
Evidently the total bound charge is yields: P Bl r
i) −E0 b cos θ + P (cos θ)

χe
 P l Bl

bl+1 l
qb = − q. (4.51) = Al b + bl+1 Pl (cos θ) −3E0

a3

χe + 2 (ii) ϵr
P
lAl bl−1 − (l + 1) bB l


Pl (cos θ)
 h
i h
i r− cos
l+2
2 1− a 3
+ ϵ 1+2 a 3 r2
r
= −E0 cos θ − (l + 1) bB
 P  b b
l+2 Pl (cos θ)
l
We could, of course, obtain the field of
σb by direct integration. But as in the (iii) Al al + bBl+1 = 0 ⇒ Bl = −a
l 2l+1
Al The electric field is then:
case of the conducting plane, there is a Remember: cos θ = P1 (cos θ), so we will
nicer solution by the method of images. have cases for l ̸= 1 and l = 1. For the E(r, θ) = −∇Vmed (r, θ)
Indeed, if we replace the dielectric by a case where l ̸= 1 we have:
single point charge qb at the image po- 3E0
(i) =
sition (0, 0, −d), we have X Bl
h
a 3

i h
a 3

i
2 1− b + ϵr 1 + 2 b
Pl (cos θ)
" bl+1
1 q X Bl
 
2a3
 
a3
 
V = p = Al bl + l+1 Pl (cos θ) · 1 + 3 cos θr̂ − 1 − 3 sin θθ̂
4π ϵ0 x2 + y 2 + (z − d)2 b r r
Question Solving for Coefficients: Capacitance as a Function of
A very long cylinder of linear dielectric From (i): h
material is placed in an otherwise uni- First find the capacitance, as a function
form electric field E0 . Find the resulting a1 [a1 cos(1ϕ) + (0) sin(kϕ)] = −E0 a cos ϕ
of h:
field within the cylinder. (The radius is Air Part:
a, the susceptibility χe , and the axis is +a−1 [c1 cos(1ϕ) + (0) sin(1ϕ)]  
perpendicular to E0 .) 2λ 2λ b
E = ⇒ V = ln
⇒ aa1 cos(ϕ) = −E0 a cos(ϕ)+a−1 c1 cos(ϕ) 4πϵ0 s 4πϵ0 a
Solution
⇒ a1 = −E0 + a−2 c1 Oil Part:
Potential - Cylindrical:
V (s, ϕ) = a0 + b0 ln s From (ii):
2λ′ 2λ′ 2λ′
 
b
∞ D= ⇒E= ⇒V = ln .
X  k  ϵ(1)a1−1 a1 cos(ϕ) 4πs 4πϵs 4πϵ a
+ s (ak cos(kϕ) + bk sin(kϕ))
k=1
= ϵ0 −E0 cos(ϕ) − a−1−1 c1 cos(ϕ)


Charge Relationship
−k
+s (ck cos(kϕ) + dk sin(kϕ))
ϵ λ λ′ ϵ
−2 −2
⇒ a1 = −E0 −a c1 ⇒ ϵr a1 = −E0 −a c1 . = ⇒ λ′ = λ = ϵr λ.
Boundary Conditions: ϵ0 ϵ 0 ϵ ϵ 0
For the constants, set V = 0 on the yz
Combining Equations: The total charge: Q = λ′ h + λ(l − h)
plane. To find the coefficients, apply the
boundary conditions: Adding these equations together yields: = ϵr λh + λl − λh = λ[ϵr h + l − h]
−2

(i) Vin = Vout at s = a, a1 = −E0 + a c1 = λ[(ϵr − 1)h + l] = λ[χe h + l]
(ii) ϵ ∂V
∂s
in
= ϵ ∂Vout
0 ∂s at s = a, −2

(iii) Vout → −E0 s cos ϕ for s >> a. +ϵr a1 = −E0 − a c1 where l is the total height.
Simplifying yields: Capacitance Expression
Inside s < R: The capacitance is:
In this region ln s and s−k blow up at a1 +ϵr a1 = −E0 −E0 ⇒ a1 (1+ϵr ) = −2E0
s = 0, and that isn’t allowable. Thus: Q λ[χe h + l] 2πϵ0 [χe h + l]
−2E0 −2E0 −2E0 C= = 2λ b

=
ln ab


∞ ⇒ a1 = = = V 4πϵ0 ln a
X (1 + ϵ r ) 1 + (1 + χ e ) 2 + χe
Vin (s, ϕ) = a0 + sk [ak cos(kϕ) + bk sin(kϕ)] .
k=1 Final Potential Inside: Equilibrium Condition
Therefore the potential inside is: Equilibrium force is the upward force of
Outside s > R: the field on the dielectric and the down-
In this region ln s and sk blow up as

2E0
 ward force of gravity:
s → ∞, and that isn’t allowable. Thus: Vin (s, ϕ) = s1 − cos(1ϕ)
2 + χe X
∞
F = 0 =⇒ FE −Fg = 0 =⇒ FE = Fg
s−k [ck cos(kϕ) + dk sin(kϕ)]=. − 2E0 s cos(ϕ) = − 2E0 x
X
Vout (s, ϕ) = ā0 +
k=1 2 + χe 2 + χe Solving for Height h
1 2 dC
Invoking Condition (iii): Resulting Electric Field In- V = mg
side: 2 dh
∞
X
Vin (s, ϕ) = sk [ak cos(kϕ)+bk sin(kϕ)] and hence 1 2 2πϵ0 χe (1)

= ρπ(b2 − a2 )hg
V
k=1 2 ln ab
∂Vin 2E0
Vout (s, ϕ) = −E0 s cos ϕ Ein (s, ϕ) = − x̂ = πϵ0 χe
∂x 2 + χe V2
=h
∞ b
X
−k
ρπ(b2 − a2 )g ln a
+ s [ck cos(kϕ) + dk sin(kϕ)]. Question
k=1 ϵ0 χ e V 2
Two long coaxial cylindrical metal tubes h= b


Invoking Condition (i): (inner radius a, outer radius b) stand ρ(b2 − a2 )g ln a
∞ vertically in a tank of dielectric oil (sus-
Question
X
ak [ak cos(kϕ)+bk sin(kϕ)] = −E0 a cos ϕ ceptibility χe , mass density ρ). The in-
k=1
ner one is maintained at potential V , A sphere of radius R is filled with mate-
∞
and the outer one is grounded. To what rial of dielectric constant εr and uniform
X
−k height (h) does the oil rise, in the space embedded free charge ρf . What is the
+ a [ck cos(kϕ) + dk sin(kϕ)]. between the tubes?
k=1
energy of this configuration?

Invoking Condition (ii): Solution Solution

∞
X Force Calculation Electric Displacement and
ϵ kak−1 [ak cos(kϕ) + bk sin(kϕ)] Know: Force: Field
k=1
From Gauss’s law (in the form of Eq.
∞
!
X 4.23), the displacement is:
= ϵ0 −E0 cos ϕ + −a−k−1 [ck cos(kϕ) + dk sin(kϕ)] .
k=1
 ρf
3 r (r < R),
Remember, P1 (cos ϕ) = cos ϕ, and the D(r) = ρf R 3
dW 1 dC 3 r 2 r̂ (r > R).
coefficients must be equal for each k, F = = V2
as you could prove by multiplying by dx 2 dx So the electric field is:
Pk (cos ϕ) sin ϕ, integrating from 0 to π, Our situation:
and invoking the orthogonality of the ( ρ
f
r (r < R),
Legendre polynomials yields bk = dk = 1 2 dC E(r) = 3ϵ 0 εr
ρf R3
0, ∀k& cos(ϕ) ⇒ k = 1. FE = V 3ϵ0 r 2 r̂ (r > R).
2 dh
Purely Electrostatic Energy Explicit Model Verification 2.2 Step 2: Electrostatic
(Eq. 4.55) Picture the dielectric as a collection of Equations for Dielectric
W1 =
tiny proto-dipoles, each consisting of +q Matter
and −q attached to a spring of constant The two key electrostatic equations are:
" 2 Z R 2 k and equilibrium
# length 0. 1. Gauss’s law for the electric field:
 Z ∞
ϵ0 ρf ρ f 1
r2 4πr2 dr + R6 4
4πr2 dr ρfree + ρbound
2 3ϵ0 εr 0 3ϵ0 R r ⃗ =
∇·E ,
ϵ0
2π 2 5

1
 The electric force on the free end is bal- where ρfree is the free charge density and
= ρ R + 1 . anced by the spring force: ρbound = −∇ · P⃗ is the bound charge
9ϵ0 f 5ϵ2r
density.
qE = kd. 2. The curl of the electric field is zero
Total Energy (Eq. 4.58) in electrostatics:
" Z R Z ∞ #
ρf R3 ρf R 3
  
1  ρf  ρf 1 ⃗ = 0.
W2 = 3
r dr + dr ∇×E
2 3 3ϵ0 ϵr 3 3ϵ0 4
0 R r
Substituting ρbound into Gauss’s law
gives:
In our case:
 
2π 2 5 1
= ρ R +1 . ρfree − ∇ · P⃗
9ϵ0 f 5ϵr ρf ⃗ =
∇·E .
E= r.
3ϵ0 ϵr ϵ0
Notice that W1 < W2 - that’s because
W1 does not include the energy involved The resulting dipole moment is p = qd, Rearranging, we obtain:
in stretching the molecules. and the polarization is P = p/dτ , so: ⃗ + P⃗ ) = ρfree .
∇ · (ϵ0 E
Verification of W2 as Work ρf
Done k= P rdτ. ⃗ this simplifies
From the definition of D,
3ϵ0 ϵr d2
We start with the (uncharged, unpolar- to:
ized) dielectric sphere, and bring in the The energy of this particular spring is:
⃗ = ρfree .
∇·D
free charge in infinitesimal installments
1 ρf
(dq), filling out the sphere layer by layer. dWspring = kd2 = P rdτ, 2.3 Step 3: Boundary Con-
When we have reached radius r′ , the 2 6ϵ0 ϵr
electric field is: ⃗
ditions for E
and hence the total is:
 ρf The boundary conditions for the elec-
 3ϵ0 ϵr r ′3 (r < r′ ), ⃗ at an interface between two
Z
 ρf tric field E
ρf r Wspring = P rdτ.
E(r) = 3ϵ0 ϵr r2 r̂ (r′ < r < R), 6ϵ0 ϵr media are:
 ρf r′3 r̂
 1. The tangential component of E ⃗ is
3ϵ0 r 2 (r > R). Now: continuous:

The work required to bring the next dq ρf (ϵr − 1)ρf ⃗ 1,tan = E

E ⃗ 2,tan .
P = ϵ0 χe E = ϵ0 χe r= r,
in from infinity to r′ is: 3ϵ0 ϵr 3ϵr
so: ⃗ satisfies:
2. The normal component of E
r′
"Z #
R Z
dW = −dq E · dl + E · dl R
ρf (ϵr − 1)ρf ϵ0 (E1,norm − E2,norm ) = σfree + σbound ,
Z
∞ R Wspring = 4π r4 dr
6ϵ0 ϵr 3ϵr 0
where σfree is the free surface charge
"
R r′
# 2π 2 5 density and σbound = −P⃗1,norm + P⃗2,norm
ρf r′3 ρf r′3 ρ R (ϵr − 1),
Z Z
1 1 = is the bound surface charge density.
= −dq 2
dr + dr 45ϵ0 ϵ2r f
3ϵ0 ∞ r 3ϵ0 ϵr r2
R
and it works out perfectly. 2.4 Step 4: Derive Bound-
⃗
ary Conditions for D
ρf r′3 1 1 Question (a)
  
1 1 1
= + − dq. We now derive the boundary conditions
3ϵ0 R ϵr r′ R ⃗ in terms of E
D ⃗ and P⃗ , the electro- ⃗ using the above results.
for D
static eqn’s of dielectric matter, and the ⃗
This increases the radius (r′ ): ⃗ derive the
1. Tangential Component of D:
boundary conditions for E, From the definition D = ϵ0 E + P⃗ , the
⃗ ⃗
boundary conditions that D ⃗ satisfies for tangential component is:
dq = ρf 4πr′2 dr′ ,
an arbitrary dielectric surface.Do not
assume a simple dielectric. ⃗ tan = ϵ0 E
D ⃗ tan + P⃗tan .
so the total work done, in going from
r′ = 0 to r′ = R, is:
Since E ⃗ tan is continuous across the
" #
boundary, the discontinuity in D ⃗ tan
4πρ2f
 Z R Z R
1 1 1
W = 1− r′5 dr′ + r′4 dr′ ⃗
arises from Ptan . Thus:
3ϵ0 R ϵr 0 ϵr 0
2 Answer (a) ⃗ 1,tan − D
D ⃗ 2,tan = P⃗1,tan − P⃗2,tan .
⃗
 
2π 2 5 1 2.1 Step 1: Definition of D
= ρ R +1 = W2 . For most dielectrics, P⃗tan is continuous,
9ϵ0 f 5ϵr ⃗ is defined as:
The displacement field D so D⃗ tan is also continuous. However,
Spring Energy Calculation ⃗ = ϵ0 E
⃗ + P⃗ , in general, if there is a discontinuity in
D
Evidently the energy ”stored in the P⃗tan , it will reflect in D
⃗ tan .
springs” is: where: 2. Normal Component of D: ⃗
• ϵ0 is the permittivity of free space, ⃗ is:
The normal component of D
2π 2 5 • E
⃗ is the electric field,
Wspring = W2 −W1 = ρ R (ϵr −1). • P is the polarization density.
⃗ Dnorm = ϵ0 Enorm + Pnorm
45ϵ0 ϵ2r f
Using the boundary condition for Step 2: Electric Field Due to the - If p⃗ is permanent (fixed), F ∝ r14 . -
Enorm : Point Charge ⃗ F ∝ 15 .
p ∝ E),
If p⃗ is induced (⃗ r
⃗ at the location of
The electric field E The problem does not specify, but the
wording suggests p⃗ is intrinsic (perma-
ϵ0 (E1,norm − E2,norm ) = σfree + σbound , the dipole (distance r away) due to the
nent), so we assume:
point charge Q is:
and substituting σbound = −P1,norm + 1
P2,norm , we get: ⃗ = 1 Q F ∝ .
E r̂, r4
4πϵ0 r2
where r̂ is the unit vector pointing from Part (b)(ii): Attractive or
ϵ0 E1,norm −ϵ0 E2,norm = σfree −P1,norm +P2,norm .
Q to the dipole. Repulsive?
Rearranging: The dielectric is uncharged but polar-
Step 3: Force on a Dipole in an ized. In the presence of the external
External Field ⃗ from Q, the dielectric develops
field E
(ϵ0 E1,norm +P1,norm )−(ϵ0 E2,norm +P2,norm ) = σfree . a dipole moment p⃗. For a linear dielec-
The force F⃗ on a dipole p⃗ in an external tric:
Recognizing that Dnorm = ϵ0 Enorm + electric field E⃗ is given by:
Pnorm , this becomes: ⃗
p⃗ = αE,
F⃗ = (⃗ ⃗
p · ∇)E.
D1,norm − D2,norm = σfree . where α is the polarizability. The direc-
For a spherically symmetric field E ⃗ = tion of p⃗ is:
2.5 Final Boundary Condi- E(r)r̂, the gradient in spherical coordi- - Aligned with E ⃗ if α > 0 (most di-
⃗ nates gives: electrics). - Anti-aligned if α < 0 (un-
tions for D
usual).
1. Tangential Component: dE
⃗ =
∇E r̂ ⊗ r̂.
dr
Step 2: Force Direction
⃗ 1,tan − D
D ⃗ 2,tan = P⃗1,tan − P⃗2,tan . Thus, the force becomes:
From (i), the force is:
dE
If P⃗tan is continuous, then: F⃗ = (⃗
p · r̂) r̂.
dr F⃗ ∝ −(⃗
p · r̂)r̂.
⃗ 1,tan = D
D ⃗ 2,tan . Substitute E(r) = 1 Q
4πϵ0 r 3 : If p⃗ is parallel to r̂ (i.e., p⃗ · r̂ > 0), then:
2. Normal Component: dE 1 −2Q
= . F⃗ ∝ −r̂,
dr 4πϵ0 r3
D1,norm − D2,norm = σfree . which is attractive (toward Q).
So the force is: If p⃗ is anti-parallel to r̂ (i.e., p⃗ · r̂ < 0),
These are the general boundary condi-
⃗ at an arbitrary dielectric sur- then:
tions for D  
face, without assuming a simple dielec- −2Q −2Q(⃗
p · r̂)
F⃗ = (⃗
p · r̂) r̂ = r̂. F⃗ ∝ +r̂,
tric. 4πϵ0 r3 4πϵ0 r3
which is repulsive (away from Q).
3 Question
Step 4: Magnitude and r-
Part (b) Dependence Physical Explanation:
Imagine a charge density distribution of
total charge Q distributed over small The magnitude of the force depends on - For α > 0, the dielectric polarizes
volume VQ . Similarly, an uncharged the orientation of p⃗ relative to r̂: such that the side closer to Q develops a
simple linear dielectric of total volume - If p⃗ is aligned with r̂ (parallel), p⃗·r̂ = p. bound charge opposite in sign to Q, and
- If p⃗ is anti-aligned (opposite), p⃗ · r̂ = the far side develops a like charge. The
Vd has a polarization distribution P⃗ (⃗r); −p. attraction to the closer bound charge
⃗r ∈ Vd with a total dipole moment p⃗. In general, the r-dependence of the force dominates, resulting in a net attractive
They are placed at a distance r from is: force. - For α < 0 (unusual), the polar-
3 3
each other. Assume that r ≫ VQ2 , Vd2 . ization is reversed, leading to repulsion.
Qp Since most simple dielectrics have α >
F ∝ .
r4 0, the force is typically attractive.
i.
This is because: 5 Question
Find the r-dependence of the force that
the dielectric feels from the far-away p · r̂|
2Q|⃗ Qp A certain coaxial cable consists of a
charged object. |F⃗ | = ∝ 3, copper wire, radius a, surrounded by a
4πϵ0 r3 r
concentric copper tube of inner radius
but the dipole moment p⃗ itself is induced c. The space between is partially filled
ii. ⃗ which scales as Q2 . For (from b out to c) with material of di-
by the field E, r
Is this force attractive or repulsive? Ex- ⃗ so p ∝ Q2 .
a linear dielectric, p⃗ ∝ E, electric constant ϵr , as shown. Find the
plain your answer physically, in terms of r capacitance per unit length of this ca-
Thus: ble.
the polarization of the dielectric.
Q Q Q2 6 Answer
4 Answer F ∝ ·
r3 r2
=
r5
.
Part (b)(i): r-dependence of 6.1 Electric Constant:
However, if p⃗ is fixed (not induced by ϵ
the force ⃗ the force scales as 14 . The prob-
E), ϵr = 1 + χe = ⇒ ϵ = ϵr ϵ0
Step 1: Far-Field Approximation r ϵ0
lem states that the dielectric is ”simple
1/3 1/3 linear,” implying p⃗ is induced by E,⃗ so 6.2 Gauss’s Law & Displace-
Since r ≫ Vq , Vd , the charge dis-
tribution Q can be treated as a point
1
the correct scaling is r5 . But typically, ment:
for a permanent dipole, the force is r14 .
I
charge, and the dielectric can be treated
as a point dipole p⃗. Clarifying: D · da = Qf enc , D = ϵE
6.3 Capacitance per Unit 7.1 Electric Field 8.2 Magnetic Field Calcula-
Length: The bound charges produce an electric tion
C Q field. We can use Gauss’s law for the - Inside the cylinder (ρ ≤ R):
= electric displacement D and then relate The total current enclosed by a circular
l V it to E. loop of radius ρ is:
Q In the absence of free charges, ∇·D = 0,
= , and D = ϵ0 E + P.
Vl ρ ρ
However, it’s simpler to directly com-
Z Z
so let Q be the charge on a length l of pute E from the bound charges. Ienc = Jb ·ϕ̂ dl = (−2αωρ′ )·2πρ′ dρ′
the copper wire. Now we need to find - Inside the cylinder (ρ ≤ R): The 0 0
V. volume bound charge ρb = −2α is uni- ρ
ρ3
Z
form. The surface bound charge σb = = −4παω ρ′2 dρ′ = −4παω .
I αR acts at ρ = R. By symmetry, 0 3
Q
D·da = Qf enc ⇒ |D|2πsl = Q ⇒ D = E ŝ.= E(ρ)ρ̂. Apply Gauss’s law to a Ampere’s law gives:
cylindrical Gaussian surface of radius ρ
2πsl
and length L: I
6.4 Electric Field in Re- I B · dl = µ0 Ienc
Qenc
gions: E · da = ,
ϵ0 
4παωρ3

For a < s < b we are in air: =⇒ B(ρ) · 2πρ = µ0 − ,
where Qenc = ρb πρ2 L = −2απρ2 L. The 3
D 1 Q Q flux is E(ρ) · 2πρL, so:
Eab = = ŝ = ŝ. so:
ϵ0 ϵ0 2πsl 2πϵ0 sl
2µ0 αωρ2
For b < s < c we are in the dielectric: −2απρ2 L αρ B(ρ) = − .
E(ρ)·2πρL = =⇒ E(ρ) = − . 3
ϵ0 ϵ0
D 1 Q Q Thus, inside:
Ebc = = ŝ = ŝ.
ϵ ϵr ϵ0 2πsl 2πϵr ϵ0 sl The local bound charge is:
2µ0 αωρ2
2 Bin = − ẑ.
6.5 Potential Calculation: Qtotal = ρb πR L + σb 2πRL 3
Z a Z b Z a - Outside the cylinder (ρ > R):
V =− E·dl = − Ebc ·dl− Eab ·dl = −2απR2 L + αR · 2πRL = 0.
The total current is the sum of the vol-
c c b ume and surface contributions:
Since the total bound charge is zero, the
Z c
Q Q
Z b electric field outside is zero:
= ds + ds
2πϵr ϵ0 sl a 2πϵ0 sl R3
I
b Eout = 0. Jb ·ϕ̂ dl+ Kb ·ϕ̂ dl = −4παω +αωR2 ·2πR
Q
Z c
1 3
= ds 7.2 Final Electric Field
2πϵr ϵ0 l b s 4παωR3 2παωR3
− αρ

ϵ0 ρ̂, ρ ≤ R,
Z b E(ρ) = =− + 2παωR3 = .
Q 1 0, ρ > R. 3 3
+ ds
2πϵ0 l a s Ampere’s law gives:
8 Magnetic Field Ev-
2παωR3
!  
1 c1
Z b
erywhere
Z
Q 1 B(ρ) · 2πρ = µ0
= ds + ds 3
2πϵ0 l ϵr b s a s When the cylinder rotates, the bound
   charges move, creating a bound current µ0 αωR3
Q 1 c b density. =⇒ B(ρ) = .
= ln + ln . 3ρ
2πϵ0 l ϵr b a 8.1 Bound Current Density
Thus, outside:
6.6 Final Capacitance per The polarization is P = αρρ̂. When the
cylinder rotates with angular velocity ω, µ0 αωR3
Unit Length: Bout = ẑ.
the velocity at a point ρ is v = ωρϕ̂. 3ρ
C Q 2πϵ0 The bound current density is:
= =h
i .
l V 1 c


ϵr ln b + ln
b 8.3 Final Magnetic Field
a ∂P 2
Jb = ∇ × M + ,
(
∂t − 2µ03ρ
αωρ
ẑ, ρ ≤ R,
7 Electric Field Every- B(ρ) = µ0 αωR3
ẑ, ρ > R.
where M is the magnetization (here 3ρ
where zero, since the polarization is fixed and
The polarization F gives rise to bound not due to magnetic effects). The time 9 Total Energy Stored
charges: derivative term is zero because P is not in Fields
- Volume bound charge density: ρb = changing in time (as per the problem
−∇ · P. - Surface bound charge density: statement). 9.1 Before Rotation (ω = 0)
σb = P · n̂. However, the rotation causes the bound Only the electric field is present:
For P = αρρ̂: charges to move, leading to a current.
The volume bound charge vb = −2α
Z
ϵ0
moves with velocity v = ωρϕ̂, giving a UE = |E|2 dV.
1 ∂ 1 ∂ 2
∇·P= (ρP (ρ)) = (αρ2 ) volume current density:
ρ ∂ρ ρ ∂ρ Inside the cylinder:
 2
1 Jb = ρb v = −2α · ωρϕ̂ = −2αωρϕ̂. 2 αρ α 2 ρ2
= (2αρ) = 2α. |E| = = 2 .
ρ ϵ0 ϵ0
There is also a surface bound current
Thus, ρb = −2α (uniform inside the due to σb = αR: For a length L:
cylinder).
ϵ0 R α 2 ρ 2
Z
At the surface ρ = R, σb = P · n̂ =
αRρ̂ · ρ̂ = αR. Kb = σb v = αR · ωRϕ̂ = αωR2 ϕ̂. UE = · 2πρLdρ
2 0 ϵ20
 πα2 L R 3 πα2 LR4 πα2 LR4 10 Final Answers
Z
For both regions:
= ρ dρ = = . - Free current enclosed: If (from the
ϵ0 0 ϵ0 4ϵ0 1. Electric Field:
⃗ is tangential and uni-
thin wire). - H
Outside E = 0, so no contribution.  αρ
Thus: − ϵ0 ρ̂, ρ ≤ R, form along the Amperian loop.
B(ρ) = Region 1 (r < R):
0, ρ > R.
πα2 LR4
Ubefore = . 2. Magnetic Field: If
4ϵ0 H1 · (2πr) = If =⇒ H1 = .
2πr
9.2 After Rotation (ω ̸= 0) 3
(
− 2µ0 αωρ
3 ẑ, ρ ≤ R, Magnetic field:
Now both electric and magnetic fields B(ρ) = µ0 αωR2
3ρ ẑ, ρ > R.
are present. µ1 If
B1 = µ1 H1 = ϕ̂.
3. Total Energy: 2πr
2 4
9.2.1 Electric Energy - Before rotation: Ubefore = πα4ϵLR 0
. - Region 2 (r > R):
2 4
The electric field is unchanged (since P After rotation: Uafter = πα4ϵLR + UB ,
is unchanged), so:
0 If
where UB is the (divergent) magnetic H2 · (2πr) = If =⇒ H2 = .
energy. - The difference is due to the 2πr
πα2 LR4 work done to rotate the cylinder, which
UB = . 12.1 Bound Currents and
4ϵ0 generates the magnetic field.
Verification
9.2.2 Magnetic Energy 11 Field Conditions 12.1.1 Step 1: Bound Currents
Z (a) The field in question must satisfy Bound currents arise from magnetiza-
1 ∇ · E = 0 and ∇ × E = 0. The first
UB = |B|2 dV. tion M⃗:
2µ0 condition gives - Volume bound current: J⃗b = ∇ ×
Inside: ⃗ ⃗
0 = ∂k Ek = Djk δk rj = Djk δkj = Dkk = 0. M . - Surface bound current: Kb =
⃗
M × n̂.
2 Magnetization in Region 1 (r < R):
2µ0 αωρ2 4µ20 α2 ω 2 ρ4 The second condition gives

2
|Bin | = = .
3 9
   
0 = ϵisk ∇s Ek = ϵisk Djk ∇s rj = ϵisk Djk δsj M⃗ 1 = µ1 − 1 H ⃗1 = µ1 − 1 If ϕ̂.
Outside: µ0 µ0 2πr
= ϵijk Djk = ϵijk Djk +ϵkj Dkj = ϵijk (Djk −Dkj ).
2 - Volume bound current: ∇×M ⃗1 = 0
µ0 αωR3 µ20 α2 ω 2 R6 2. (b) We must have Ek = −∂k φ.

2 (since M⃗ 1 is azimuthal and symmetric).
|Bout | = = . Therefore,
3ρ 9ρ2 - Surface bound current at r = R:
1
 
Thus: φ = A − Cm rm − Dsm rs rm . ⃗ ⃗ µ1 If
2 Kb1 = M1 × r̂ = − −1 ẑ.
µ0 2πR
"Z Z ∞ 2 2 2 8 #
R
1 4µ20 α2 ω 2 ρ4 µ0 α ω R Magnetization in Region 2 (r > R):
UB = · 2πρLdρ + 2
· 2πρLdρ .
2µ0 0 9 R 9ρ    
⃗ µ2 ⃗ µ2 If
M2 = − 1 H2 = −1 ϕ̂.
Simplify: µ0 µ0 2πr
" Z
R Z ∞ #12 Question - Volume bound current: ∇ × M ⃗2 =
πµ0 α2 ω 2 L 5 6 dρ 0. - Surface bound current at r = R:
UB = 4 ρ dρ + R (a)
. An infinitely extended solid cylin-
9 0 R ρ der of radius R and magnetic perme-  
⃗ ⃗ µ2 If
ability µ1 has a thin wire along its axis. Kb2 = M2 × (−r̂) = −1 ẑ.
The second integral diverges, which is This wire carries a free current If . This µ0 2πR
unphysical. This suggests that our as- whole assembly is now embedded in an
sumption of an infinitely long cylinder is Total Surface Bound Current:
infinite magnetizable medium of perme-
the issue. For a finite cylinder, the in-
tegral would be finite. Alternatively, we
ability µ2 . ⃗b = K
K ⃗ b2 = (µ2 − µ1 ) If ẑ.
⃗ b1 + K
- Using Ampere’s Law, find the mag- µ0 2πR
can consider the energy per unit length netic field everywhere. - Do the same by
by ignoring the L divergence: explicitly calculating the magnetic fields Total bound current:
separately due to the free currents and
⃗ b · d⃗l = (µ2 − µ1 ) If .
I
UB 2 2
πµ0 α ω
 6
R
 bound currents and verify that the re- Ib = K
= 4 + R6 lim ln ρ . sults are identical. µ0
L 9 6 ρ→∞
Using Ampere’s Law to Find the Mag-
netic Field Everywhere The total current (free + bound) for
The logarithmic divergence implies that r > R:
the total magnetic energy is infinite for Step 1: Define Regions and Symmetry
an infinitely long cylinder. In practice, - Region 1 (inside the cylinder, 
µ2 − µ1

we might consider a finite length or ig- r < R): Magnetic permeability µ1 . Itotal = If + Ib = If 1 + .
- Region 2 (outside the cylinder, µ0
nore the divergence for the purpose of
comparing energies. r > R): Magnetic permeability µ2 . -
The system has cylindrical symmetry, Using Ampere’s Law for B: ⃗
9.3 Energy Difference so the magnetic field B ⃗ is azimuthal (ϕ- I
The electric energy remains the same, direction) and depends only on r. ⃗ · d⃗l = µ0 Itotal .
B
but the magnetic energy adds to the to-
tal energy after rotation. The difference 12.0.1 Step 2: Apply Ampere’s For r > R:
is:
Law for H ⃗ 
µ2 − µ1

∆U = Uafter − Ubefore = UB . B2 · (2πr) = µ0 If 1+
Ampere’s Law states: µ0
The increase in energy comes from the
work done to start the rotation, which
I µ2 If
⃗ · d⃗l = Ifree, enclosed .
H =⇒ B2 = ϕ̂.
induces the magnetic field. 2πr