Chapter 5 Magnetostatics The Magnetic Field

5.1 The Lorentz Force Law 5.1.1 Magnetic Fields Outside a magnetic the lines emerge from the north pole
and enter the south pole; within the magnet they are
By analogy with electrostatics, why don’t we study directed from the south pole to the north pole. The dots
magnetostatics first? Due to lack of magnetic monopole. represents the tip of an arrow coming toward you. The
cross represents the tail of an arrow moving away.

If one try to isolate the poles by cutting the magnet, a curious

thing happens: One obtains two magnets. No matter how
thinly the magnet is sliced, each fragment always have two
poles. Even down to the atomic level, no one has found an How a current-carry wire
isolated magnetic pole, called a monopole. Thus magnetic produces a magnetic field?
field lines form closed loops.
1 2

The Magnetic Field of a Bar Magnet Definition of the Magnetic Field

When iron filings are sprinkled around a bar magnet, they When defining of the electric field, the electric field strength
form a characteristic pattern that shows how the influence of can be derived from the following relation: E=F/q. Since an
the magnet spreads to the surrounding space. isolated pole is not available, the definition of the magnetic
field is not as simple.
Instead, we examine how an electric charge is affected by a
magnetic field.

F = qvB sin θ = qv × B
The magnetic field, B, at a point along the tangent to a field
line. The direction of B is that of the force on the north pole of Since F is always perpendicular to v, a magnetic force does
a bar magnet, or the direction in which a compass needle no work on a particle and cannot be used to change its
points. The strength of the field is proportional to the number kinetic energy.
of lines passing through a unit area normal to the field (flux
density). The SI unit of magnetic field is the Tesla (T). 1 T=104 G
3 4
 The Lorentz Force Law Force on a Current-Carrying Conductor
When a particle is subject to both electric and magnetic When a current flows in a magnetic field, the electrons as a
fields in the same region, what is the total force on it? whole acquire a slow drift speed, vd, and experience a
magnetic force, which is then transmitted to the wire.
F = q (E + v × B )
F = qvB sin θ = (nA e)vd B
This is called the Lorentz force law. This axiom is found in
experiments. = (nAevd ) B
=I B
The magnetic force do no work.
F = I ×B
dWmag = Fmag ⋅ dl = q ( v × B) ⋅ vdt = 0 F = I B sin θ
Really? But, how do you explain a magnetic crane lifts a n: is the number of the conductor per unit volume.
container? : is defined to be in the direction in which the current is flowing.
5 6

Example:
Force on a Current-Carrying Conductor The Magnetic Force on a Semicircular Loop
The force on an infinitesimal current element is A wire is bent into a semicircular loop of radius R. It carries a
current I, and its plane is perpendicular to a uniform magnetic
dF = Id × B field B, as shown below. Find the force on the loop.
Solution:
dF = Id × B
dFy = IRB sin θ dθ
π
Fy = IRB ∫ sin θ dθ
0
= 2 IRB = I (2 R) B

The force on a wire is the vector sum (integral) of the forces The x-components of the forces on such elements will
on all current elements. cancel in pairs.
The net force on any close current-carrying loop is zero.
7 8
The Motion of Charged Particles in Magnetic Fields Cyclotron Motion

How does a charged particle move with an initial velocity v What are the frequency and the period? Are they independent
perpendicular to a uniform magnetic field B? of the speed of the particle? Yes.
Since v and B are perpendicular, the particle experiences a The period of the orbit is
force F=qvB of constant magnitude directed perpendicular.
2πr 2πm  m  2π
Under the action of such a force, the particle will move in a T= = = 
circular path at constant speed. From Newton’s second law, v qB  q  B
F=ma, we have 1 qB  q  B
fc = = = 
mv 2 mv T 2πm  m  2π
qvB = ⇒ r=
r qB f c B = 2.8 MHz Gauss
The radius of the orbit is directly proportional to the linear The frequency is called the cyclotron frequency.
momentum of the particle and inversely proportional to the
magnetic field strength. All particles with the same charge-to-mass ratio, q/m, have
9 the same period and cyclotron frequency. 10

Example: Cyclotron Helical Motion

A cyclotron is used to accelerate protons from rest. It has a What happens if the charged particle’s velocity has not only a
radius of 60 cm and a magnetic field of 0.8 T. The potential perpendicular component v⊥ but also a parallel component v//?
difference across the dees is 75 kV. Find: (a) the frequency Helical Motion.
of the alternating potential difference; (b) the maximum
kinetic energy; (c) the number of revolutions made by the
protons.
Solution:
qB
(a) f c = 2πm = 12 MHz
(qrmax B) 2
(b) K max = = 1.76 ×10 −12 J = 11 MeV
2m The perpendicular component v⊥ gives rise to a force qv⊥B
(c) ∆K = 2qV = 150 keV that produces circular motion, but the parallel component v//
is not affected. The result is the superposition of a uniform
K max / ∆K = 11000 / 150 = 73.5 revs.
circular motion normal to the lines and a constant motion
How to determine the maximum kinetic energy? 11 along the lines. 12
 Example: Cycloid Motion Magnetic Bottle/Mirror
Suppose, for instance, that B points in the x-direction, and E What happens if the magnetic field is not uniform? Energy
in the z-direction. A particle initially at rest is released from the transfer between the perpendicular and parallel components.
origin; what path will it follow?

In a nonuniform field, the particle experiences a force that

Solution: points toward the region of week field. As a result, the
1. Write down the equation of motion. component of the velocity along the B lines is not constant.
2. Solve the coupled differential equations. If the particle is moving toward the region of stronger field,
3. Determine the constants using the initial conditions. as some point it may be stopped and made to reverse the
13
direction of its travel. 14

Velocity Selector Mass Spectrometer

A mass spectrometer is a device that separates charged
particles, usually ions, according to their charge-to-mass
ratios.

E
v=
 E = − Ej  E B1
  ⇒ v = i = vi
 B = − Bk  B mv mE
r= =
qB2 qB1 B2
Only those particles with speed v=E/B pass through the
crossed fields undeflected. This provides a convenient way m B1 B2
= r
of either measuring or selecting the velocities of charged q E
particles.
15 16
 Example: Mass Spectrometer Current and Surface Current
In a mass spectrometer shown below, two isotopes of an The current in a wire is the charge per unit time passing a
elements with mass m1 and m2 are accelerated from rest by given point.
a potential difference V. They then enter a uniform B normal Current is measured in coulombs-per-second, or amperes (A).
to the magnetic field lines. What is the ratio of the radii of
their paths? 1 A=1 C/s
Solution:
The surface current density, K, is defined as follows:
2qV consider a ”ribbon” of infinitesimal width d ⊥, running parallel
v= to the flow. Then, dI
m K=
d ⊥
mv 2mV
r= = then r1 / r2 = (m1 / m2 ) In words, K is the current per unit width-perpendicular-to-flow.
qB qB 2
d (σ ⊥ // )
d( )
dI dt d //
Note1: How particle is accelerated by a potential difference? K= ≈ =σ =σv
17 d ⊥ d ⊥ dt 18

Volume Current Density Conservation of Charge

The current crossing a surface S can be written as
The volume current density, J, is defined as follows:
consider a ”tube” of infinitesimal cross section da⊥, running I = ∫ J ⋅ nˆ da
parallel to the flow. Then,
S
dI
J= In particular, the total charge per unit time leaving a volume
da⊥
V is
dQ
In words, J is the current per unit area-perpendicular-to-flow. I= ∫ J ⋅ nˆ da = ∫ (∇ ⋅ J )dτ = − dt
S V

d ( ρ a⊥ dQ d
dt dt V∫
// ) where = ρ dτ .
d( )
dI dt d //
J= ≈ =ρ = ρv
da⊥ da⊥ dt d dρ
∫V (∇ ⋅ J )dτ = − dt V∫ ρ dτ ⇒ ∇ ⋅ J = − dt
19 continuity equation 20
 5.2 The Biot-Savart Law 5.2.1 Steady Currents 5.2.2 The Magnetic Field of a Steady Current
Stationary charges produce electric fields that are constant The Biot-Sarvart law:
in time. Steady currents produce magnetic fields that are µ0 I × rˆ µ I dl ′ × rˆ
also constant in time. B(r ) = ∫
4π r 2
dl ′ = 0 ∫
4π r2
Stationary charges ⇒ constant electric fields; electrostatics. The integration is along the current
Steady currents ⇒ constant magnetic fields; magnetostatics. path, in the direction of the flow.
µ0: the permeability of free space.
Steady current means that a continuous flow that goes on Definition of magnetic field B: newtons per ampere-meter
forever without change and without charge piling up
anywhere. They represent suitable approximations as long
or tesla (T). 1 T=1 N/(A⋅m)
as the fluctuations are reasonably slow.
The Biot-Sarvart law plays a role analogous to Coulomb’s
∇⋅J = 0 law in electrostatics.
21 22

π
Con't
Example 5.5 Find the magnetic field a distance s from a long µ I dl′ × rˆ µ I 2
cos 2 θ s
B(r ) = 0 ∫ 2 = 0 ∫π cos θ dθ
straight wire carrying a steady current I. 4π r 4π s 2 cos 2 θ
−
2
π
µ0 I µI I
= sin θ 2π = 0 (= 2 ×10−7 Tesla)
4π s −
2 2π s s

What is the force between two parallel current-carrying wires?

Sol :
µ0 I × rˆ µ I dl′ × rˆ dF = Idl × B
B(r ) = ∫
4π r 2
dl ′ = 0 ∫
4π r2 µI µII
Then, determine the suitable coordinate : cylindrical coordinate ( s, φ , z ). dF = I 2 0 1 dl = 0 1 2 dl
2π d 2π d
In the diagram, (dl′ × rˆ ) points out of page and have the magnitude dF µ0 I1 I 2
=
dl ′ sin α = dl ′ cos θ dl 2π d
1 cos θ
l ′ = s tan θ ⇒ dl ′ = s sec 2 θdθ and =
r s 23 (attractive force per unit length, why?) 24
Example 5.6 Find the magnetic field a The Bio-Sarvart Law
distance z above the center of a circular for the Surface and Volume Current
loop of radius R, which carries a steady
µ I × rˆ µ I dl ′ × rˆ
current I. The Biot-Sarvart law: B(r ) = 0 ∫ 2 dl ′ = 0 ∫
4π r 4π r2

µ0 I dl′ × rˆ µ0 K × rˆ
4π ∫ r 2
B(r ) = da ′
4π ∫ r 2
Sol : B(r ) = For surface current:

Choose cylindrical coordinate (s, φ , z ). µ0 J × rˆ

4π ∫ r 2
For volume current: B(r ) = dτ ′
In the diagram, (dl′ × rˆ ) sweeps around the z axis,
thus only the z -component survives. For a moving charge: Wrong, why?
z -component of (dl′ × rˆ ) = dl ′ cos θ = R cos θ dφ µ0 J × rˆ µ0 qvδ (r − r′) × rˆ µ0 qv × rˆ
4π ∫ r 4π ∫
B(r ) = dτ ′ = dτ ′ =
1 1 R 2
r 2
4π r 2
= 2 2 and sin θ = 2 2 1/ 2
r 2
(R + z ) (R + z ) A point charge does not constitute a steady current.
25 26

The Magnetic Field of Solenoid Solenoid

Problem 5.11 A solenoid of length L
and radius a has N turns of wire and
carries a current I. Find the field
strength at a point along the axis.

Solution:
Sine the solenoid is a series of closely
packed loops, we may divided into current
loops of width dz, each of which contains
ndz turns, where n =N/L is the number of
turns per unit length.
The current within such a loop is (ndz)I.

27 28
 Solenoid (II)
Con’t Homework #9
z = a tan θ ⇒ dz = a sec θ dθ 2

nIdz = nIa sec 2 θ dθ

Problems: 9, 10, 11, 39, 49
µ0 a 2
dBaxis = nIa sec dθ
2

2(a 2 + a tan 2 θ )3/ 2

1
= µ0 nI cos θ dθ
2
θ2 1
B=∫ µ nI cos θ dθ
θ1 2 0

1
= µ0 nI (sin θ 2 − sin θ1 )
2
B = µ 0 nI (infinite long solenoid)
29 30

5.3 The Divergence and Curl of B The Differential Form of B

5.3.1 Straight-Line Currents
The magnetic field of an infinite straight wire: Suppose we have a bundle of straight wires. Only wires
that pass through the loop contribute µ0I.
The line integration then be
µ0 I ∫ B(r) ⋅ dl = µ I I enc = ∫ J ⋅ da
B(r ) = φˆ 0 enc
2πs
The total current enclosed
by the integration loop.
The integral of B around a circular path of radius s, centered
at the wire, is:
µI
∫ B ⋅ dl = ∫ (∇ × B) ⋅ da = ∫ µ J ⋅ da
0

∇ × B = µ0 J
∫ B(r) ⋅ dl = ∫ 2π0 s φˆ ⋅ φˆ sdφ = µ0 I
In fact for any loop that encloses the wire would give the Does this differential equation apply to any shape of the
same answer. Really? current loop? Yes, to be prove next.
31 32
 5.3.2 The Divergence and Curl of B The Divergence of B

The divergence of B:
The Biot-Sarvart Law for the
µ0 J (r′) × rˆ µ J (r′) × rˆ
general case of a volume
current:
∇ ⋅ B(r ) = ∇ ⋅ (
4π ∫ r 2
dτ ′) = 0 ∫ ∇ ⋅ (
4π r2
)dτ ′
J (r′) × rˆ rˆ rˆ
∇⋅( 2
) = 2 ⋅ (∇ × J ) − J ⋅ (∇ × 2 )
r r r
µ0 J (r′) × rˆ
4π ∫ r 2
B(r ) = dτ ′ ∇ ⋅ ( A × B) = B ⋅ (∇ × A) − A ⋅ (∇ × B)
rˆ 1
2
= −∇( ) (Prob. 1.13)
r r
0 0
J (r′) × rˆ rˆ rˆ
The integration is over the primed coordinates. ∇⋅( 2
) = 2 ⋅ (∇ × J ) − J ⋅ (∇ × 2 )
r r r
The divergence and the curl are to be taken with ∴ ∇ ⋅ B = 0 The divergence of a magnetic field is zero.
respective to the unprimed coordinates.
33 34

The Curl of B A Special Technique

The curl of B:
rˆ Let’s prove that this
µ
∇×B = 0 ∫∇×(
J (r′) × rˆ
)dτ ′
∫ ( J ⋅ ∇) r 2
dτ ′ = 0
integration is zero.
4π 0 r2 0 rˆ rˆ
∇ × ( A × B) = (B ⋅ ∇) A − ( A ⋅ ∇)B + A(∇ ⋅ B) − B(∇ ⋅ A) ( J ⋅ ∇)2
= −(J ⋅ ∇ ′) 2 ,
special technique r r
where r = (r − r ′)
primed
unprimed primed +unprimed ∇ ⋅ ( fA) = ∇f ⋅ A + f (∇ ⋅ A)
0 to be seen next
J (r′) × rˆ rˆ rˆ
∇×( 2
) = J (∇ ⋅ 2 ) − (J ⋅ ∇) 2 ) Using the above rule, the x component is:
r r r
J (r ′) × rˆ rˆ x-x′ x-x′
0, for steady current
x-x′
∇×( 2
) = J (∇ ⋅ 2 ) = J 4πδ 3 (r ) (See 1.5.3) (J ⋅ ∇′) = ∇ ′ ⋅ ( J ) − (∇′ ⋅ J )
r r r3 r3 r3 0, since J(r’@∞)=0
µ0
∴∇ × B = 4π ∫ J (r ′)δ 3 (r )dτ ′ = µ0 J (r ) rˆ x − x′ x − x′
4π ∫ (J ⋅∇) r x2 dτ ′ = ∫ ∇′ ⋅ ( r 3 J)dτ ′ = ∫S ( r 3 J) ⋅ da′ = 0
∇ × B = µ0 J The curl of B equals µ0 times J.
35 What happens if J(r’)≠0 36
 5.3.3 Applications of Ampere’s Law Applications of Ampere’s Law
∇ × B = µ0 J Ampere's law in differential form Like Gauss’s law, ampere’s law is always true (for steady
currents), but is not always useful.
∫ (∇ × B) ⋅ da = ∫ B × dl = ∫ µ J ⋅ da = µ I
amperian loop
0 0 enc
Only when the symmetry of the problem enables you
to pull B outside the integral can you calculate the
∫ B × dl = µ I
amperian loop
0 enc Ampere's law in integral form
magnetic field from the Ampere’s law.
These symmetries are:
Just as the Biot-Savart law plays a role in magnetostatics
that coulomb’s law assumed in electrostatics, so 1. Infinite straight lines
Ampere’s play the role of Gauss’s. 2. Infinite planes (Ex. 5.8)

Electrostatics: Coulomb Æ Gauss, 3. Infinite solenoids (Ex. 5.9)

Magnetostatics: Biot-Savart Æ Ampere. 4. Toroids (Ex. 5.10)

37 38

Infinite Straight Wire Infinite Planes

Example An infinite straight wire of radius R carries a current I. Example 5.8 Find the magnetic field
Find the magnetic field at a distance r from the center of the of an infinite uniform surface
wire for (a) r>R, and (b) r<R. Assume that the current is current K = Kxˆ , flowing over the xy
uniformly distributed across the cross section of the wire. plane.
Solution:
Solution:
(a)
∫B⋅d = B 2πr = µ 0 I
µ 0I
B= (r > R ) ∫B⋅d = B 2l = µ 0 Kl
2πr µ K / 2yˆ for z < 0
µ 0K B= 0
(b) πr 2 B=  − µ 0K / 2yˆ for z > 0
∫ B ⋅ d = B 2πr = µ0 πR 2
I 2

µ 0I
B= r (r < R )
2πR 2
39 40
 Solenoid Toroid
Example 5.9 An ideal infinite solenoid has n turns per unit Example 5.10 A toroidal coil (shaped like a doughnut) is
length and carries a current I. Find its magnetic field inside. tightly wound with N turns and carries a current I. We
assume that it has a rectangular cross section, as shown
Solution:
below. Find the field strength within the toroid.
∫B⋅d Solution:
b c d a
= ∫ B⋅d + ∫ B⋅d + ∫ B⋅d + ∫ B⋅d
a
b
b c d
∫B⋅d = B ∫ d = µ 0 NI
= ∫ B⋅d µ 0NI
B=
a

2πr
BLab = µ 0nLab I
The field is not uniform; it varies
as 1/r. The toroidal fields are
B = µ 0nI
used in research on fusion power.
41 42

5.3.4 Comparison of Magnetostatics and 5.4 Magnetic Vector Potential

Electrostatics 5.4.1 The Vector Potential
 ρ
∇ ⋅ E = Gauss's law ρ ρ
 ε0 ∇ × E = 0 ⇔ E = −∇V and ∇ ⋅ E = ⇒ ∇ 2V = −
∇×E = 0 ε0 ε0
 no name (Faraday's law) ∇ ⋅ B = 0 ⇔ B = ∇ × A and ∇ × B = µ0 J ⇒ ∇(∇ ⋅ A) − ∇ 2 A = µ0 J
 ∇⋅B = 0 Gauss's law for magnetic field Is it possible for us to set ∇ ⋅ A = 0 equals zero? Yes.

∇ × B = µ0 J Ampere's law (Ampere-Maxwell law) Coulomb gauge
F = q (E + v × B ) Lorentz' s force law Proof: If ∇ ⋅ A 0 ≠ 0, let A = A 0 + ∇λ ⇒ B = ∇ × A 0 = ∇ × A
If ∇ ⋅ A = 0, then ∇ 2 λ = −∇ ⋅ A 0 ← similiar to Poisson' s equation
 2 1 ρ
 ∇ V = − ρ / ε 0 V = 4πε ∫ r dτ ′
 0

∇ λ = −∇ ⋅ A 0 λ = 1 ∇ ⋅ A0
∫ dτ ′
2
 4π r
43 It is always possible to make the vector potential divergenceless.44
 The Vector Potential and Scalar Potential Example 5.11 A spherical shell, of radius R, carrying a
uniform surface charge σ, is set spinning at angular velocity
Using the Coulomb gauge, we obtain : ∇ 2 A = − µ0 J ω. Find the vector potential it produce at point r.
µ J (r′)
A= 0 ∫ dτ ′
4π r
For line and surface current,
µ I µ0 K (r′)
4π ∫ r
A = 0 ∫ dl ′ A= da′
4π r Sol :First, let the observer is in the z axis and ω is tilted at an angle ψ
µ K (r′)
What happens when the curl of B vanishes? Vector potential is A(r ) = 0 ∫ da′
4π r
Magnetostatic scalar potential. The surface current density K (r′) = σ v′

∇ × B = 0 ⇒ B = −∇U
⇒ ∇ 2U = 0 (similiar to Laplace's equation)
45 46

µ 0 Rω ( − cosψ sin θ ′ sin φ ′) xˆ 2

4π ∫
A (r ) = R sin θ ′dθ ′d φ ′
r 2 + R 2 − 2 rR cos θ ′ − µ0 R 3σω sinψyˆ ( R 2 + r 2 + Rr ) | R − r | −( R 2 + r 2 − Rr )( R + r )
A(r ) = (− )
µ Rω (cosψ sin θ ′ cos φ ′ − sinψ cos θ ′) yˆ 2 2 3R 2 r 2
+ 0 ∫ R sin θ ′dθ ′d φ ′
4π r 2 + R 2 − 2 rR cos θ ′  µ0 Rσ
µ Rω (sin ψ sin θ ′ sin φ ′) zˆ 2
+ 0 ∫ R sin θ ′dθ ′d φ ′
 2 (ω × r ) inside
4π r 2 + R 2 − 2 rR cos θ ′ A(r ) = 
 µ0 R σ (ω × r ) outside
4

− R 3σω sinψµ0 yˆ cos θ ′  2r 3

A(r ) =
4π ∫ r + R 2 − 2rR cos θ ′
2
sin θ ′dθ ′dφ ′
π Reverting to the “natural” coordinate, we have
− R 3σω sinψµ0 yˆ − cos θ ′
= (2π ) ∫ d cos θ ′
4π 0 r + R 2 − 2rR cos θ ′
2

− µ0 R 3σω sinψ yˆ
1
u
=
2 ∫
−1 r + R − 2rRu
2 2
du

47 Surprisingly, the field inside the spherical shell is uniform. 48

Example 5.12 Find the vector potential
5.4.2 Summary; Magnetostatic Boundary
of an infinite solenoid with n turns per Conditions
unit length, radius R, and current I.

Sol : A cute method that does the job. We have derived five formulas
interrelating three fundamental
∫ B ⋅ da = Φ = ∫ (∇ × A) ⋅ da = ∫ A ⋅ dl quantities: J, A and B.
where Φ is the flux of Β through the loop in question.

∫ B ⋅ dl = µ I
0 enc ⇒ ∫ A ⋅ dl = Φ
Using a circular " amperian loop" at a radius inside the solenoid. Comments:
µ nI •There is one “missing link” in the diagram.
∫ A ⋅ dl = A2πs = ∫ B ⋅ da = µ0 nI (πs ) ⇒ A = 02 sφˆ for s < R
2

•These three variables, J, A, and B, are all vectors. It is

Using a circular " amperian loop" at a radius s outside the solenoid. relatively difficult to deal with.
µ0 nIR 2 ˆ
∫ A ⋅ l = π = ∫ B ⋅ a = µ π ⇒ = φ for s ≥ R
2
d A2 s d 0 nI ( R ) A
2s 49 50

Magnetostatic Boundary Conditions: Normal Magnetostatic Boundary Conditions: Tangential

The magnetic field is not
continuous at a surface with The tangential component of B is
surface density K. discontinuous.
What is the physical picture?

∫B⋅d = µ 0 I enc
Consider a wafer-thin pillbox. Gauss’s law states that Consider a thin rectangular loop. The
curl of the Ampere’s law states that P

∫ B ⋅ da = 0
S
The ends gives nothing (as ε→0), and the sides give
The sides of the pillbox contribute nothing to the flux, in the
limit as the thickness ε goes to zero.
//
( Babove − Bbelow
//
) = µ0 K ⇒ Babove
//
− Bbelow
//
= µ0 K
⊥ ⊥ ⊥ ⊥ In short, B above − B below = µ 0 K × nˆ , where nˆ points " upward."
( Babove − Bbelow ) A = 0 ⇒ Babove = Bbelow
How about the vector potential A?
51 52
Boundary Conditions in Terms of Vector Potential 5.4.3 Multipole Expansion of the Vector Potential

Like the scalar potential in electrostatics, the 1 1

=
vector potential is continuous any boundary: r − r′ (r + r ′ − 2rr ′ cos θ ′)
2 2

1 r′ r′
A above = A below = (1 + ( ) cos θ ′ + ( ) 2 ((3 cos 2 θ ′ − 1) / 2) + …)
r r r
1 ∞
r ′
= ∑ ( ) P (cosθ ′)
r =0 r
⊥ ⊥
∇ ⋅ A = 0 ⇒ Aabove = Abelow The vector potential of a current loop
µI 1 µI ∞ 1
A= 0 ∫ dl ′ = 0 ∑ n +1 ∫ (r ′) n Pn (cos θ ′)dl ′
∇ × A = B ⇒ ∫ A ⋅ dl = ∫ B ⋅ da = Φ ⇒ Aabove = Abelow
// //
4π r- r′ 4π n =0 r
µ I 1 1 1 
= 0  ∫ dl ′ + 2 ∫ r ′ cos θ ′dl ′ + 3 ∫ r ′2 P2 (cos θ ′)dl ′ + 
4π  r r r

53 monopole dipole quadrupole 54

Multipole Expansion A Special Technique

µ I 1 1 1 
Part I Recalling Stokes’ theorem ∫ (∇ × v) ⋅ da = ∫ v ⋅ dl
S
A = 0  ∫ dl′ + 2 ∫ r′ cos θ ′dl′ + r 3 ∫ r′ P (cosθ ′)dl′ +
2
 Let v=cT ∇ × ( fA) = ∇f × A + f (∇ × A) P
4π  r
2
r
magnetic monopole term is always zero. ∫ (∇ × cT ) ⋅ da = ∫ (∇T × c + T (∇ × c)) ⋅ da = −c ⋅ ∫ ∇T × da
S S S

µ0 I 1 µI 1 ∫ cT ⋅ dl = c ⋅ ∫ Tdl ∫ ∇T × da = − ∫ Tdl
2 ∫
A dip = r ′ cosθ ′dl′ = 0 2 ∫ (rˆ ⋅ r′)dl′ S

4π r 4π r
P P P

∫ (rˆ ⋅ r′)dl′ = −rˆ × ∫ da′ (Eq. 1.108, to be shown later) Part II ∫ S

∇′T ′ × da′ = − ∫ T ′dl′, let T ′ = rˆ ⋅ r′
P
Then ∇( A ⋅ B) = A × (∇ × B) + B × (∇ × A ) + ( A ⋅∇ )B + (B ⋅∇ ) A
µ0 1 µ m × rˆ
A dip = − rˆ × ( I ∫ da′) = 0 ∇′(rˆ ⋅ r′) = rˆ × (∇′ × r′) + r′ × (∇′ × rˆ ) + (rˆ ⋅∇′)r′ + (r′ ⋅∇′)rˆ
4π r 2
4π r 2
= (rˆ ⋅∇′)r′ = rˆ
where m = I ∫ da′ is the magnetic dipole moment.
55
∫S
rˆ × da′ = − ∫ (rˆ ⋅ r′)dl′ = rˆ × ∫ da′
P ＃ 56
The Magnetic Field of a Dipole
Homework #10
µ0 mzˆ × rˆ µ0 m sin θ ˆ
A dip = = φ
4π r 2 4π r 2
Problems: 15, 16, 24, 46, 58

µ0 m
B dip = ∇ × A = (2 cos θrˆ + sin θθˆ )
4πr 3

57 58
 Chapter 6 Magnetic Fields in Matter 6.1.2. Torques and Forces on Magnetic Dipoles
6.1 Magnetization
6.1.1 Diamagnets, Paramagnets, Ferromagnets A magnetic dipole experiences a torque in a magnetic field,
just as an electric dipole does in an electric field.
All the magnetic phenomena are due to electric charges in
motion: Any current loop could be built up from infinitesimal
Electrons orbiting around nuclei rectangles, with all the “internal” side canceling. There is
magnetic dipoles
Electrons spinning about their axes no actual loss of generality in using the shape.
When a magnetic field is applied, a net alignment of these
magnetic dipoles occurs, and the medium becomes
magnetically polarized, or magnetized.
The magnetic polarization M, unlike electrical polarization
P, might be parallel to B (paramagnets) or opposite to B
(diamagnets).
Let’s calculate the torque on a rectangular current loop in
A few substances (ferromagnets) retain their magnetization a uniform magnetic field.
even after the external field has been removed. 1 2

Torques and Forces on Magnetic Dipoles Torques and Dipole Moment

Center the loop at the origin, and tilt it an angle θ from the N = IabB sin θxˆ = mB sin θxˆ = m × B
z axis towards the y axis. Let B point in the z direction. where m = Iab is the magnetic dipole moment of the loop.

This equation is identical in form to the electrical analogy.

N = p×E
The torque is again in such a direction as to line the dipole
up parallel to the field (paramagnetism).
Sloping sides: the forces cancel.
Horizontal sides: the forces cancel but they generate a
torque. N = L × F = aF sin θxˆ QM: The Pauli exclusion principle dictates that the
electrons within a given atom lock together in pairs with
The magnitude of the force on each of these segments is: opposite spins, and this effectively neutralizes the torque
F = IbB on the combination.
3 4
 Forces in Nonuniform Magnetic Field Forces on an Infinitesimal Current Loop and Model
In a uniform field, the net force on a current loop is zero: F = ∇(m ⋅ B) (= −∇U , where U = −(m ⋅ B))
=0
F = I ∫ (dl × B) = I ∫ (dl ) × B = 0
Identical to the electrical formula F = ∇(p ⋅ E)
In a nonuniform field this is no longer the case, because
the magnetic field B could not come outside the integral. Does the magnetic dipole consist of a pair of opposite
magnetic monopoles just like an electric dipole?
Fringing field effect : F = 2πIRB cos θ

5 6

6.1.3. Effect of a Magnetic Field on Atomic Orbits Electron Speeds Up or Slows Down
1 e2 v2
Electrons not only spin; they also revolve around the nucleus. = me without the magnetic field.
4πε 0 R 2 R
Let’s assume the orbit is a circle of radius R. The current
looks like steady (really?) The centripetal force comes from two sources:
the electric force and the magnetic force.
e e ev
Current I = = =
T 2πR 2πR The negative charge
v of the electron 1 e2 v′ 2
+ ev ′B = m
4πε 0 R′2 R′
e
evR
Orbital dipole moment m = − zˆ (m = IπR 2 )
2 Assume R′ ≅ R
m (v′2 − v 2 ) me (v′ + v)(v′ − v)
ev′B = e =
R R
eRB
∴ ∆v = (v′ − v) ≈ When B is turn on, the electron speeds up.
7 2me 8
 The Dipole Moment and The Diamagnetism 6.1.4 Magnetization
A change in the orbital speed means a change in the In the present of a magnetic field, matter becomes magnetized.
dipole moment. Upon microscopic examination, it contains many tiny dipoles,
1 e2 R 2
∆m = − e∆vR = − B with a net alignment along some direction.
2 2me
Two mechanisms account for this magnetic polarization:
The change in m is opposite to the direction of B.
1.Paramagnetism: the dipoles associated with the spins of
In the presence of a magnetic field, each atom picks up a unpaired electrons experience a torque tending to line them up
little “extra” dipole moment, and the increments are all parallel to the field.
antiparallel to the field. This is the mechanism responsible
2. Diamagnetism: the orbital speed of the electrons is altered
for diamagnetism.
in such a way as to change the orbital dipole moment in the
This is a universal phenomenon, affecting all atoms, but it is direction opposite to the field.
typically much weaker than paramagnetism.
We describe the state of magnetic polarization by the vector
quantity:
9 M ≡ magnetic dipole moment per unit volume. 10

6.2 The Field of a Magnetized Object Vector potential and Bound Currents
6.2.1 Bound Currents
Can the equation be expressed in a more illuminating form,
Suppose we have a piece of magnetized as in the electrical case? Yes!
material (i.e. M is given). What field does By exploiting the identity, ∂ ∂ ∂ 1
this object produce? (xˆ ′ + yˆ ′ + zˆ ′ )
1 rˆ ∂x′ ∂y′ ∂z′ ( x − x′) 2 + ( y − y′) 2 + ( z − z′) 2
The vector potential of a single dipole m is ∇′ = xˆ ′( x − x′) + yˆ ′( y − y′) + zˆ ′( z − z ′) rˆ
r r2 = =
(( x − x′) 2 + ( y − y′) 2 + ( z − z ′) 2 )3/ 2 r 2
µ0 m × rˆ µ0
A(r ) = 1
4π r 2 The vector potential is A(r ) =
4π ∫ M (r′) × (∇′ )dτ ′
r
In the magnetized object, each volume element carries a Using the product rule ∇ × ( fA) = ∇f × A + f (∇ × A)
dipole moment Mdτ’, so the total vector potential is and integrating by part, we have
µ0  1 M (r′) 
4π ∫ r
µ M (r′) × rˆ A(r ) =  [∇′ × M (r′)]dτ ′ − ∫ ∇′ × [ ]dτ ′
A(r ) = 0 ∫ dτ ′ r 
4π r2 µ0  1  µ 1
how? Prob. 1.60
4π ∫ r ∫ [M(r′) × nˆ ′]da′
=  [∇′ × M (r′)]dτ ′ +
0
11  4π r 12
 Vector potential and Bound Currents
µ0 1 µ 1
A(r ) = ∫
4π r
[∇′ × M (r′)]dτ ′ + 0 ∫ [M (r′) × nˆ ′]da′
4π r
J b = ∇′ × M (r′) K b = M (r′) × nˆ ′
volume current surface current

With these definitions, bound currents

∫ (∇ ⋅ E)dτ = v∫ E ⋅ da µ0 Jb µ Kb
∫ ∫
Gauss's law A(r ) = dτ ′ + 0 da′
v S 4π v r 4π S r
 (∇ ⋅ ( v × c))dτ = c ⋅ (∇ × v )dτ
∫ ∫v
Let E = v × c,  v The electrical analogy
 v∫ ( v × c) ⋅ da = −c ⋅ v∫ v × da volume charge density ρ b = −∇ ⋅ P
S S

Since c is a constant vector, so ∫ (∇ × v )dτ = − v∫ v × da surface charge density σ b = P ⋅ nˆ

13 14
v S

µ 0 Rω ( − cosψ sin θ ′ sin φ ′) xˆ 2

4π ∫
Example 5.11 A spherical shell, of radius R, carrying a A (r ) = R sin θ ′dθ ′d φ ′
uniform surface charge σ, is set spinning at angular velocity r 2 + R 2 − 2 rR cos θ ′
µ Rω (cosψ sin θ ′ cos φ ′ − sinψ cos θ ′) yˆ 2
ω. Find the vector potential it produce at point r. + 0 ∫ R sin θ ′dθ ′d φ ′
4π r 2 + R 2 − 2 rR cos θ ′
µ Rω (sin ψ sin θ ′ sin φ ′) zˆ 2
+ 0 ∫ R sin θ ′dθ ′d φ ′
4π r 2 + R 2 − 2 rR cos θ ′

− R 3σω sinψµ0 yˆ cos θ ′

A(r ) =
4π ∫ r + R 2 − 2rR cos θ ′
2
sin θ ′dθ ′dφ ′
Sol :First, let the observer is in the z axis and ω is tilted at an angle ψ − R 3σω sinψµ0 yˆ
π
− cos θ ′
µ K (r′) = (2π ) ∫ d cos θ ′
Vector potential is A(r ) = 0 ∫ da′ 4π 0 r + R 2 − 2rR cos θ ′
2

4π r − µ0 R 3σω sinψ yˆ
1
u
The surface current density K (r′) = σ v′ =
2 ∫
−1 r + R − 2rRu
2 2
du

15 16
 − µ0 R 3σω sinψyˆ ( R 2 + r 2 + Rr ) | R − r | −( R 2 + r 2 − Rr )( R + r ) Example 6.1 Find the magnetic field of a
A(r ) = (− )
2 3R 2 r 2 uniformly magnetized sphere of radius R.
 µ 0 Rσ
 (ω × r ) inside
A(r ) =  24 Sol : Choosing the z axis along the direction of M,
 µ 0 R 3σ (ω × r ) outside  J ′b = ∇ × M = 0
 2r we have 
K ′b = M × nˆ ′ = M sin θφˆ
Reverting to the “natural” coordinate, we have The surface current density is analogous to that of a spinning
spherical shell with uniform surface current density.

K ′b = M × nˆ ′ = M sin θφˆ ⇔ K ′ = σv′ = σRω sin θφˆ

σRω → M 2
µ 0 M (inside)
B=
3
Surprisingly, the field inside the spherical shell is uniform. 17
Can you find a more direct method? 18

6.2.2 Physical Interpretation of Bound Current Physical Interpretation of Bound Current

Bound surface current Kb: Bound current density Jb:
What is the current in terms of M? What if the magnetization is not uniform?
In terms of the magnetization M, its dipole The adjacent current loops do not completely cancel out.
moment is m=Mat=Ia. So, M=I/t=Kb
∂M z
Consider a thin slab of uniformly magnetized material, with Case (a) I x = [ M z ( y + dy ) − M z ( y )]dz = dydz
the dipoles represented by tiny current loops. ∂y
∂M y
All the “internal” currents cancel. However, at the edge there Case (b) I x = [ M y ( z + dz ) − M y ( z )]dy = dydz
is no adjacent loop to do the canceling. ∂z
∂M z ∂M y
∴(Jb )x = − ⇒ Jb = ∇ × M
∂y ∂z

19 20
 6.3 The Auxiliary Field H
The Role of H in Magnetostatics
6.3.1 Ampere’s Law in Magnetized Materials
What is the difference between bound current and free H plays a role in magnetostatics analogous to D in the
current? J =J +J electrostotics.
b f

Ampere’s law can be written: D allows us to write Gauss’s law in terms of free change alone.
1
(∇ × B) = J = J f + J b = J f + ∇ × M D = ε 0 E + P, ∇ ⋅ D=ρ f
µ0
H permits us to express Ampere’s law in terms of free current
1
⇒ ∇×( B − M) = J f alone.
H=
1
B − M, ∇ × H = J f
µ0 µ0
H What we can control directly.
In terms of H, then the Ampere’s law reads
∇ × H = J f (differential form) Why can’t we turn the bound currents on or off
independently?
∫ H ⋅ dl = I f (integral form)
21 22

Example 6.2 A long copper rod of radius R carries a

uniformly distributed (free) current I. Find H inside and Sol :
outside the rod. Use the Ampere's law in the integral form and
properly choose a suitable Amperian loop.
πs 2
s ≤ R : H (2πs ) = I f enc = I
πR 2
sI
so H = φˆ
2πR 2
I
s > R : H (2πs ) = I , so H = φˆ
2πs

How to determine the magnetic field B?

How to choose a suitable Amperian loop? Symmetry.
23 24
 H and B, D and E 6.3.2 A Deceptive Parallel
∇×H = J f
Which equation is more useful?
In free space ∇ × B = µ0 J
∇⋅D = ρf ∇⋅B = 0
We can easily control the free current I, but not the In matter ∇ × H = J f
free charge. So H can be determined accordingly.
B
On the other hand, the potential difference V can be ∇⋅H = ∇⋅( − M ) = −∇ ⋅ M ≠0
read from the voltmeter, which can be used to µ0
determine E.
At what condition the divergence of H is equal to zero?
The name of H: Some author call H, not B, the “magnetic
field”, but it is not a good choice. Let’s just call it “H”. M // B i.e. M // B // H for uniform material only.

25 26

6.3.3 Boundary Conditions

Homework #11
The magnetostatic boundary conditions can be rewritten in
terms of H and the free surface current Kf.

∇×H = J f ⇒ H above
//
− H below
//
= K f × nˆ
Problems: 6.4, 6.10, 6.13, 6.15
∇ ⋅ H = −∇ ⋅ M ⇒ H above
⊥ ⊥
− H below ⊥
= −( M above ⊥
− M below )

The corresponding boundary condition in terms of B and total

surface current K.
∇ × B = µ0 J ⇒ B above
//
− Bbelow
//
= µ0 (K × nˆ )
∇⋅B = 0 ⇒ ⊥
Babove ⊥
− Bbelow =0
How to express the boundary conditions at metal or
dielectric interface? 27 28
 6.4 Linear and Nonlinear Media
Material Susceptibility
6.4.1 Magnetic susceptibility and Permeability
The magnetization of paramagnetic and diamagnetic
materials is sustained by the field, i.e. when B is removed,
M disappears.
M = χ m H,
where the proportionality constant χ m is called
the magnetic susceptibility.
χm Because M = χ m H ∝ I f
Why not use M = B?
µ0
Materials that obey M = χ m H are called linear media.
Β = µ 0 (H + M ) = µ 0 (1 + χ m )H = µH,
where µ = µ 0 (1 + χ m ) is call the permeabili ty of the material.
29 30

Divergence of the Magnetization

Example 6.3 An infinite solenoid (n turns per unit
length, current I) is filled with linear material of
Does the linear media avoid the defect that the divergence
susceptibility χm: Find the magnetic field inside the
of M is zero? No!
solenoid.
Even though M, H, and B are parallel, the divergence of M
Sol: The problem exhibits solenoidal symmetry. Thus, we is not zero at the boundary. Consider the following example.
can employ the Ampere’s law.

∫ H ⋅ dl = I f (integral form)
∫ M ⋅ da ≠ 0
HA = nAI ∴ H = nIzˆ Gaussian pillbox

⇒ ∇⋅M ≠ 0
B = µ 0 (1 + χ m )nIzˆ
The enhancement of the magnetic field strength depends on
the susceptibility of the material.
and J b = ∇ × M = ∇ × χ m H = χ m J f
Is there a material that the field is significantly enhanced? 31 32
 6.4.2 Ferromagnetism Ferromagnetic Domains

Ferromagnets---which are not linear---require no external

fields to sustain the magnetization unlike paramagnets and
diamagnets.
In a ferromagnet, each dipole “like” to point in the same
direction as its neighbors. All the spins point the same way.

Domain boundaries: Domains parallel to the field grow, and

the others shrink.
Why isn’t every wrench and nail a powerful magnet?
If the field is strong enough, one domain takes over entirely,
Domains. and the iron is said to be “saturated”.
33 34

Hysteresis Loop Curie Temperature and Phase Transitions

Temperature effect: The dipoles within a given domain line
up parallel to one another. However, the random thermal
motions complete with this ordering.

Curie temperature: As the temperature increases, the

alignment is gradually destroyed. At certain temperature the
iron completely turns into paramagnet. This temperature is
called the curie temperature.

Hysteresis: The path we have traced out. phase transitions

In the experiment, we adjust the current I, i.e. control H.

In practice M is huge compared to H.
35 36
 Homework #12

Problems: 6.17, 6.21, 6.23, 6.26

37 38

Supplementary Material (optional): Properties of ferrite materials (I)

Introduction to ferrite materials Nonreciprocal electrical property: the transmission
coefficient through the device is not the same for different
The ferrites are crystals direction of propagation.
鐵磁性 having small electric
conductivity compared to
ferromagnetic materials, Unequal propagation constant: The left and right
反鐵磁性 circularly polarized waves have different propagation
constant along the direction of external magnetic field B0.
Thus they are useful in
high-frequency situations
because of the absence Anisotropic magnetic properties: The permeability of the
of significant eddy ferrite is not a single scalar quantity, but instead is a tensor,
鐵氧體磁性
current losses. which can be represented as a matrix.

39 40
 Classical picture of the magnetization process
Properties of ferrite materials (II)
--- By treating the spinning electron as a gyroscopic top.
Ferrites are ceramiclike materials with specific resistivities
that may be as much as 1014 greater than that of metals and If an electron is located in a uniform static magnetic field
with dielectric constants around 10 to 15 or greater. B0, a torque is given by
m e
T = m × B0 = − p × B0 = B0 × P
p mo
Ferrites are made by sintering a mixture of metal oxides dP e
and have the general chemical composition MO·Fe2O3, =T= B 0 × P = ω0 × P
dt mo
where M is a divalent metal such as Mn, Mg, Fe, Zn, Ni, Cd,
etc. eB0
where ω0 = is called the Larmor fequency;
mo
=
Relative permeabilities of several thousand are common. P= is angular momentum; and
The magnetic properties of ferrites arise mainly from the 2
magnetic dipole moment associated with the electron e=
m= is magnetic dipole moment.
spin. 41 2mo 42

Quantum mechanics’ viewpoint sz=±1/2 Saturation magnetization

In the absence of any damping forces, the actual

precession angle will be determined by the initial position As the strength of the bias field
of the magnetic dipole, and the dipole will precess about H0 is increased, more magnetic
B0 at this angle indefinitely (free precession). dipole moments will align with H0
until all are aligned, and M
reaches an upper limit.
In reality, however, the existence of damping forces will
cause the magnetic dipole to spiral in from its initial angle
until m is aligned with B0. The material is then said to be magnetically saturated, and
Ms is denoted as the saturation magnetization. Ms typically
ranges from 4πMs=300 to 5000 Gauss.
This explains why sz equals ±1/2 in the Quantum
Mechanics. Below saturation, ferrite materials can be very lossy at
But where does the damping force come from? microwave frequencies, and rf interaction is reduced.
43 The ferrites are usually operated in the saturated state. 44
 Curie temperature Properties of some ferrite materials

The saturation magnetization

of a material is a strong
function of temperature,
decreasing as temperature
increases.
This effect can be understood by noting that the vibrational
energy of an atom increases with temperature, making it
more difficult to align all the magnetic dipoles. Why use 4πMs? B = 4πM + H = µH (Gaussian unit)

At a high enough temperature a zero net magnetization The unit of B is gauss; the unit of H is Oersted. They have same
results. This temperature is called the Curie temperature, Tc. dimension.
What does ∆H and Mr mean?
45 46
Ferrite linewidth and remanent magnetization

Anisotropic magnetic properties (I) Anisotropic magnetic properties (II)

If H is the applied ac field, the total magnetic field is Ht = H0 zˆ + H , Omitting higher order terms, the equations can be reduced to
dM x d 2M x dH y
where H << H 0 . The field produced a total magnetization is = −ω0 M y + ωm H y , + ω02 M x = ωm + ω0ωm H x ,
dt dt 2
dt
the ferrite given by M t = M s zˆ + M . dM y
= −(ω0 M x + ωm H x ), d 2M y dHx
Ms is the dc saturation magnetization and M is the dt + ω02 M y = −ωm + ω0ωm H y.
dt 2 dt
additional ac magnetization (in the xy plane) caused by dM z
= 0, whereω0 = µ0γH0 and ωm = µ0γM s
applied field. dt
The component equations of motion: If M and H ∝ e jωt , the above equtions can be reduced to
dM x the phasor equations:
= −µ0γM y (H0 + H z ) + µ0γ (M s + M z )H y
dt  χ xx χ xy 0
(ω02 − ω 2 )M x = ω0ωm H x + jωωm H y ,
dM y M = [χ ]H = χ yx χ yy 0 H ,
= −µ0γM x (H0 + H z ) − µ0γ (M s + M z )H x (ω0 − ω )M y = − jωωm H x + ω0ωm H y .
2 2
dt  0 0 0
dM z ωω jωω
= −µ0γM x H y + µ0γM y H x where χ xx = χ yy = 2 0 m 2 and χ xy = −χ yx = 2 m 2
dt 47
ω0 − ω ω0 − ω 48
 Anisotropic magnetic properties (III) Forced precession of spinning electron (I)
To relate B and H, we have If a small ac magnetic field is superimposed on the static
 µ jκ 0
field H0, the magnetic dipole moment will undergo a forced
B = µ0 (M + H ) = [µ ]H ⇒ [µ ] = µ0 ([U ] + [χ ]) = − jκ µ 0  precession.
 0 0 µ0 
Of particular interest is the case where the ac magnetic
 ω0ωm
µ = µ0 (1 + χ xx ) = µ0 (1 + χ yy ) = µ0 (1 + ω 2 − ω 2 )
field is circularly polarized in the plane perpendicular to H0.
 0

κ = − jµ χ ) = jµ χ = µ ωω m
A right-hand circularly polarized wave can be expressed in
)
 0 xy 0 yx 0
ω02 − ω 2 phasor form as
H + = H + ( xˆ − jyˆ )
A material having a permeability tensor of this form is called
gyrotropic. and in time-domain form as

How to apply this concept to a circularly polarized wave? { }

H + = Re H + e jωt = H + ( xˆ cos ωt + yˆ sin ωt )

49 50

Forced precession of spinning electron (II) Real and imaginary permeability &
propagation and attenuation constant.

propagation constant
β ± = ω εµ ±

RHCP LHCP
ωm 
M x+ = H + ,
ω0 − ω  ωm ωm
+
⇒M = H+ ⇒ µ + = µ0 (1 + ) RHCP
− jω ω − ω ω0 − ω
M y+ = m
H + . 0
ω0 − ω  ω
µ − = µ0 (1 + m ) LHCP
ω0 + ω 51 52
 Faraday rotation --- a nonreciprocal effect Microwave gyrator
Consider linearly polarized
electric field at z=0, represented Gyrator with a
as the sum of a RHCP and a twist section.
LHCP wave:
E0 E
E ( z =0) = xˆE0 = ( xˆ − jyˆ ) + 0 ( xˆ + jyˆ )
2 2
These two polarized waves propagate with different
propagation constants.
E0 E
E ( z =A) = ( xˆ − jyˆ )e− jβ+A + 0 ( xˆ + jyˆ )e− jβ−A Gyrator without
2 2 a twist section.
 β −β β −β 
= E0  xˆ cos( + − )A − yˆ sin( + − )Ae− j ( β+ +β− )A / 2
 2 2 
E β −β
θ = tan−1 y = −( + − )A. This effect is called Faraday rotation.
Ex 2 53 54

Faraday-rotation isolator Resonance isolator

55 56
Four-port circulator Three-port circulator

57 58