Module I: Electromagnetic waves

Lectures 6-7: EM waves in conned spaces

Amol Dighe
TIFR, Mumbai
Outline
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
Coming up...
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
Travelling waves with the same (x, y) prole
We are looking for waves travelling in z direction, while keeping
the same (x, y) prole. I.e. the form

E =

E
0
(x, y)e
i (k
z
zt )
,

B =

B
0
(x, y)e
i (k
z
zt )
(1)
Maxwells (

E) and (

B) equations then become

E
y
x

E
x
y
= i B
z
,
B
y
x

B
x
y
=
i
c
2
E
z
(2)
E
z
y
ik
z
E
y
= i B
x
,
B
z
y
ik
z
B
y
=
i
c
2
E
x
(3)
ik
z
E
x

E
z
x
= i B
y
, ik
z
B
x

B
z
x
=
i
c
2
E
y
(4)
Note that one can factor out the e
i (k
z
zt )
dependence of
E
x
, E
y
, E
z
and B
x
, B
y
, B
z
, so now onwards they have no z- or
t -dependence in this lecture.
Using the last two lines (4 equations), one can write
E
x
, E
y
, B
x
, B
y
in terms of the other two quantities, E
z
and B
z
All components in terms of E
z
and B
z
E
x
=
1
(/c)
2
k
2
z
_
k
z
E
z
x
+
B
z
y
_
(5)
E
y
=
1
(/c)
2
k
2
z
_
k
z
E
z
y

B
z
x
_
(6)
B
x
=
i
(/c)
2
k
2
z
_
k
z
B
z
x


c
2
E
z
y
_
(7)
B
y
=
i
(/c)
2
k
2
z
_
k
z
B
z
y
+

c
2
E
z
x
_
(8)
Note that if E
z
and B
z
both vanish (or are constants), no other
components of

E or

B can survive (unless k
z
= 0, which case
needs to be treated separately.)
However E
z
and B
z
are not free parameters; the above
equations just give four constraints on

E and

B, two more
constraints from the last page are still remaining.
Constraining E
z
, B
z
themselves
E
z
, B
z
themselves must satisfy consistency conditions
E
y
x

E
x
y
= i B
z
(9)
B
y
x

B
x
y
=
i
c
2
E
z
(10)
These correspond to
_

2
x
2
+

2
y
2
k
2
z
+

2
c
2
_
E
z
= 0 (11)
_

2
x
2
+

2
y
2
k
2
z
+

2
c
2
_
B
z
= 0 (12)
If there were no boundary conditions in the x y plane, this would
have a plane wave solution a at x y prole . But conducting
boundaries imply that these elds must have a non-trivial x y prole.
EM wave propagation in waveguides
Let us consider rectangular / circular hollow conducting
cylinders, through which an EM wave will be guided by bending
the boundaries of the cylinders.
A simple solution would have been a plane wave travelling along
z direction, such that

E and

B elds are transverse,
E
z
= B
z
= 0. Such a solution is called as TEM (transverse
electric and magnetic) mode.
Such a mode is not possible in a hollow cylinder, proof given on
the next page
However E
z
and B
z
can individually vanish, such modes are
termed TE (Transverse electric: E
z
= 0) and TM (Transverse
magnetic: B
z
= 0).
Hollow cylinder cannot have both E
z
= 0 and B
z
= 0
Since B
z
= 0, we have (

E)
z
= B
z
/t = 0. Then
E
y
x

E
x
y
= 0 (13)
Since E
x
and E
y
are independent of z, and E
z
= 0, we get

E = 0, i.e.

E can be written as

E = .
In addition, no charges inside the cylinder, so E = 0. That is,

2
= 0.
Now we have a boundary value problem, with
2
= 0 inside
the boundary and =constant on the complete boundary (the
hollow conductor).
This boundary value problem has a solution, =constant
everywhere, and the uniqueness theorem states that this is the
only solution.
Thus, there can be no electric / magnetic elds inside the
waveguide.
TE and TM modes
http://www.allaboutcircuits.com/vol_2/chpt_14/8.html
TEM mode
http://www.allaboutcircuits.com/vol_2/chpt_14/8.html
Coming up...
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
TE modes (E
z
= 0, B
z
= 0) in a rectangular waveguide
Let the walls of the waveguide be at y = 0, b and x = 0, a.
The boundary conditions are then
E
x
= 0 at y = 0, b and E
y
= 0 at x = 0, a
The equations that give E
x,y
in terms of B
z
then imply
B
z
y
= 0 at y = 0, b and
B
z
y
= 0 at x = 0, a
The solution to the differential equation for B
z
, with these
boundary conditions, is
B
z
= Acos(k
x
x) cos(k
y
y) (14)
where k
x
= (m/a) and k
y
= (n/b).
Such a mode is called TE
mn
mode. Note that at least one of m or
n has to be nonzero, else all elds will vanish.
Cutoff frequencies for TE modes
The TE
mn
solution, when substituted in the differential equation
for B
z
, gives

_
m
a
_
2

_
n
b
_
2
k
2
z
+(/c)
2
= 0 (15)
For consistency with the physical situation, k
z
must be real; i.e.
k
2
z
> 0. This gives the condition
> c
_
_
m
a
_
2
+
_
n
b
_
2

mn
(16)
Thus, for a TE mode TE
mn
to propagate, it must have a minimum
frequency
mn
. A waveguide thus acts like a high-pass lter.
TM modes
A similar analysis is possible for TM modes, but this will not be
done here.
Note that the cutoff frequencies
mn
for the TM modes are the
same as those for TE modes.
Coming up...
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
Phase velocity and group velocity
Phase velocity: simply the speed at which the crest of the
wavefront travels in a given direction.
For a plane wave Ae
i (

xt )
, the phase velocity along the
direction

r is
v
ph
=
dr
dt

constant phase
=

|

k.

r|
(17)
If

k is not along

r, typically v
ph
> c. This does not mean that any
signal is travelling faster than light, though.
Group velocity measures the speed at which a signal is
transported. The signal is embedded in the distribution of
frequencies, and group velocity measures how fast the peak of
this distribution shifts. Details on the next page.
Group velocity
The Fourier transform of a wave gives the frequencies the wave
consists of. Consider the situation where the spread in
frequencies is small, which is the only one where we can dene
a group velocity easily. Let the frequencies be conned to the
range =
0
. The corresponding wave vectors are
conned to

k =

k
0

k.
The wave is
(

x, t ) =
_
a(

k)e
i (

xt )
d
3
k , (18)
which may be written as
(

x, t ) = A(

x, t ) e
i (

k
0

x
0
t )
, (19)
where
A(

x, t ) =
_
a(

k)e
i (

xt )
d
3
k (20)
The frequency distribution shifts as a wavepacket, the velocity of
the peak of the distribution is the approximate velocity of the
wavepacket.
Group velocity: continued
Let us consider a one-dimensional case of a wave travelling
along z-axis. At the peak,
0 =
dA
dt
=
A
t
+
A
z
dz
dt
(21)
The group velocity is then
v
g
=
dz
dt
=
A/t
A/z
=

k
=
d
dk

0
(22)
Velocities along z axis for the waveguide
=
_

2
mn
+k
2
z
Phase velocity v
ph
=

k
z
=
c

1(
mn
/)
2
Group velocity v
g
=
d
dk
z
= c
_
1 (
mn
/)
2
Waveguide transports different frequencies at different speeds:
dispersion
Coming up...
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
Circular cylindrical waveguides
No TEM mode, as per the earlier arguments
For TM mode
E
z
= AJ
m
(k

r )e
im
e
i (k
z
zt )
(23)
If the cylinder has radius r
0
,
then the boundary condition is J
m
(k

r
0
) = 0, gives k

(m)
For TE mode
B
z
= AJ
m
(k

r )e
im
e
i (k
z
zt )
(24)
If the cylinder has radius r
0
,
then the boundary condition is J

m
(k

r
0
) = 0, gives k

(m)
k
2
z
= (/c)
2
k
2

cutoff frequency
m,
= ck

(m)
TM and TE modes have different cutoff frequencies, unlike
rectangular waveguides !
Power transmitted by a waveguide
Consider TE mode. i.e. E
z
= 0.
The equations for

E

= (E
x
, E
y
) and

B

= (B
x
, B
y
) become
B
x
=
ik
z
k
2

B
z
x
B
y
=
ik
z
k
2

B
z
y
_


B

=
ik
z
k
2

B
z
(25)
E
x
=
ck
k
2

B
z
y
E
y
=
ck
k
2

B
z
x
_


E

=
ick
k
z

z (26)
The magnitude of Poynting vector (power transmitted per unit
area) is then
|

N| =
|

H
0
|
2
2
=
|

E
0
|
2
2

k
z
ck
0
=
1
2
_

0
k
z
k
|

E
0
|
2
(27)
Comparing with |

N| = (1/2)|E
0
|
2
, this enables us to dene the
conductance of the waveguide as =
_

0
/
0
(k
z
/k). This may
be compared with the conductance of free space,
_

0
/
0
.
Coming up...
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
Coaxial cable
http://en.wikipedia.org/wiki/Coaxial_cable
Propagation through a coaxial cable
TEM Mode is supported (now there are two disjoint boundaries,
so the argument for hollow waveguides does not work.)
TE and TM modes also propagate, but have a threshold
frequency
The TEM mode
Electric and magnetic elds:

E =
E
0

r
r
e
i (k
z
zt )
,

B =
E
0

cr
e
i (k
z
zt )
(28)
Group velocity v
g
= c
Coming up...
1
Waveguides
2
Rectangular waveguide
3
Phase and group velocities
4
Circular cylindrical waveguides
5
Coaxial cable
6
Cavities
Rectangular cavity
Conducting walls at x = 0, a; at y = 0, b and at z = 0, c.
Potential inside the cavity:

mnp
= sin(k
x
x) sin(k
y
y) sin(k
z
z)e
i t
(29)
where k
x
= (m/a), k
y
= (n/b), k
z
= (p/c)
This can be used to obtain

E and

B inside the cavity.
A rectangular cavity supports discrete modes.
LHC accelerator: cavity principle
http://www.lhc-closer.es/php/index.php?i=1&s=4&p=7&e=2
LHC accelerator: bunching cavities
http://www.lhc-closer.es/php/index.php?i=1&s=4&p=7&e=2
Recap of topics covered in this lecture
Propagation in waveguides in terms of E
z
and B
z
TEM, TE and TM modes from Maxwells equations
No TEM modes for hollow waveguides
Waveguides as high-pass lters, as dispersive media
Phase velocity and group velocity
Power transmitted through waveguide
Coaxial cable: TEM propagation, in addition to TE and TM
Cavities for bunching protons together at accelerators