You are on page 1of 26

# Module I: Electromagnetic waves

Lecture 9: EM radiation
Amol Dighe

Outline

Coming up...

## Radiation from antennas

~ and B
~ from A
~ and
E
Now that we know
Z
1
(~x0 , t ) 3 0
d x ,
(~x, t) =
40
|~x ~x0 |

Z ~ 0
J(~x , t ) 3 0
d x
|~x ~x0 |
(1)
~ = A/t
~
~ =A
~ to obtain
We can calculate E
and B
~
~
E(~x, t) and B(~x, t). However note the subtle point:

=x
=
,
,
(2)
x i
t
t
t

~ ~x, t) = 0
A(
4

~x

## whereas (x , t ) and ~J(~x0 , t ) are specified in terms of the

retarded time t .
~0

(~x0 , t )

x i
t
0
~
J(~x , t )

x i
t

=
=

(~x0 , t ) t

t
x i t

~J(~x0 , t ) t

t
x i

(3)

## Intermediate derivatives: gradient of

Defining ~r ~x ~x0 = (r1 , r2 , r3 ) and r |~r|, we have

 

1
1
(~x0 , t )
0
0

~
~
=
(
x
,
t
)
+
(
x
,
t
)

(4)

i
i
i
x
r
x
r t
r x
t
t
Further using
 
1

x i r t

0
(~x , t )
i
x

=
=

ri
,
r3

t r
(~x0 , t ) 1 ri
=

,
t r dx i t
t
cr

(5)

we get

=
=
=


 

1

1 i

+ x
(~x, t ) d 3 x 0
(x , t )x

i
i
x
r t
r
x
t

Z 
0
0
~
~

1
(x , t )
r (x , t ) 3 0
r

d x
40
r2
cr
t

Z 
1
[(~x0 )]
[(
~x0 )] 3 0
2 r
d x
(6)
40
r
cr
1
40

Z 

~0

~
Intermediate derivatives: curl of A
Similar to the case of derivatives of , we have

~J t r

~ 0
~J(~x0 , t ) 1 ri

~x , t ) =
,
J(
=

x i
t r dx i
t
cr
t

(7)

And then
~
A

=
=

 

1
1 i
i
0
~
~
~
~

x
J(x , t ) + x
J(x, t ) d 3 x 0
x i r
r dx i
t
#
Z "~ 0
0
~
0
J(~x , t )
1 J(~x , t )
r +
r d 3 x 0
4
r2
cr
t
#
Z "~ 0

0
[J(~x )] r [~J(~x0 )] r
+
d 3x 0
(8)
4
r2
cr
0
4

Z 

~ ~x, t)
E(
~ ~x, t)
E(

=
=

1
40

~
A
t
[(~x0 )]
[(
~x0 )]
r
r +
2
r
cr

0 !
[~J(~
x )]
d 3 x 0 (9)
c2r

## The magnetic field is

~ ~x, t)
B(
~ ~x, t)
B(

~
= A
!
Z
0
0
[~J(~x0 )]
[~J(~
x )]
=
r +
r d 3 x 0
4
r2
cr

(10)

~ ~x, t) and B(
~ ~x, t): behaviour at large |~x|
E(
~ ~x, t) and B(
~ ~x, t), there are terms that behave as 1/r 2
In both, E(
and there are terms that behave as 1/r . The former are
proportional to the sources, the latter are proportional to the rate
of change of sources.
When the sources are confined to a small region |~x0 | < d, then
for |~x| >> d, the 1/r terms dominate over the others. These are
the radiative components of the fields.
~ and B
~ fields are then
The radiative E
~ rad (~x, t)
E
~ rad (~x, t)
B

1
40

0
4

[(
~x0 )]
[~J(~x0 )]
r
d 3x 0
cr
c2r
!

[~J(~x0 )] r
d 3x 0
cr

(11)

(12)

## For monochromatic sources

Monochromatic sources correspond to
(~x0 , t) = 0 (~x0 )eit and ~J(~x0 , t) = ~J0 (~x0 )eit .
(Real part always implicit)
This dependence implies
 0 
(~x ) = (~x0 , t ) = 0 (~x0 )ei(kr t) ,
h
i
~J(~x0 ) = ~J(~x0 , t ) = ~J0 (~x0 )ei(kr t) .

(13)
(14)

And hence
(~x, t)

~ ~x, t)
A(

Z
1
ei(kr t) 3 0
0 (~x0 )
d x ,
40
r
Z
i(kr t)
0
~J0 (~x0 ) e
d 3x 0
4
r

## The potentials also then have the same frequency .

(15)

~ rad and B
~ rad in terms of ~J only
E
Given the continuity equation
0 ~J(~x0 , t) + (
~x0 , t) = 0 ,

(16)

## it is clear that if [~J(~x0 )] is known everywhere, so is [(

~x0 )], and
~ and B
~ fields can be written in terms of ~J only.
the radiative E
For monochromatic sources, Some algebraic manipulation using
the above result yields (See Panofsky-Phillips)
~ rad (~x, t)
B

~ rad (~x, t)
E

Z ~ 0
[J(~x )] r 3 0
d x
r
Z ~ 0
1
([J(~x )] r) r 3 0
d x
2
40 c
r
1
40 c 3

(17)
(18)

## Frequency components of radiation fields

~ rad (~x, t) and E
~ rad (~x, t) give the radiation
Fourier components of B
fields in terms of their frequency components:

Brad (~x)

Erad (~x)

Z 
 ikr
i
~J (~x0 ) ~k e d 3 x 0
4c 2
r
Z 
 eikr
i
(~J (~x0 ) ~k) r
d 3x 0
40 c
r

(19)
(20)

## If the sources are monochromatic, the above directly give the

~ and B
~ fields. We shall use this in calculating
corresponding E
the power radiated by periodically time-varying charges and
currents.

Coming up...

## Energy from a radiation pulse

~ =E
~ H,
~ which gives the power
The Poynting vector is N
radiated per unit area along it.
~ is
Total energy radiated per unit area normal to N
Z
Z
~ ~x, t)dt =
~ ~x, t) H(
~ ~x, t)dt
N(
E(
(21)

Z
~ rad eit d H
~ rad0 ei0 t d 0 dt
E
=

, 0 ,t=
Z
~ rad
E
2

~ rad d + 2
H

~ rad H
~ rad d
E

(22)
~ ~x, t) and H(
~ ~x, t) are real, E
~ = E
~ and H
~ = H
~.
But since E(

Then
Z

~ ~x, t)dt = 2
N(

Z
0

~ rad (H
~ rad ) d + c.c.
E

(23)

## Calculating radiated energy for a pulse

~ rad and H
~ rad = B
~ rad /0 obtained
Substituting the expressions for E

## earlier, after a bit of algebra, gives

~ rad (H
~ rad ) =
E

1
(4)2

Z
2
 eikr

0 ~
3 0
0
~
~
d
x
J
(
x
)

r
0
r

~ = r 2 r is
Total radiated energy across a surface d S
Z
~ rad (H
~ rad ) d r 2 d r + c.c.
U = 2 E

(24)

(25)

2
r Z Z 
 ikr

0
1
~J (~x0 ) ~k e d 3 x 0 dr 2 d (26)

4 0
r

In other words,
dU
1
=
d
4

Z
2


0 ~
0
ikr 3 0
~
~
J
(
x
)

k
e
d
x

0

(27)

## Average power radiated by a monochromatic source

Here we calculate the average radiated power /area over a cycle:
1
T

Z
0

~ ~x, t)dt = 1
N(
T

~ ~x, t) H(
~ ~x, t)dt
E(

(28)

## where T is the periodicity of the wave.

~ rad (~x, t) = E
~ rad eit and H
~ rad (~x, t) = H
~ rad eit , the
Since E
0
0
averaging gives
hNi =
=

1 rad
E H0rad
2 0
r
1 1
0
2 (4)2 0

Z 
2
 ikr

~J0 (~x0 ) ~k e d 3 x 0 r

r

(29)

## The average power radiated is then

dhPi
1 1
=
d
2 (4)2

Z
2


0 ~ 0
ikr
3
0
~
J0 (~x ) k e d x

0

(30)

Long-distance approximation
When |~x| >> |~x0 |, then we have
kr = k |~x ~x0 | k |~x| ~k ~x0

(31)

dU
1
=
d
4

Z
2


0 ~
0
i~k~x0 3 0
~
~
J (x ) k e
d x

0

2
r Z 


dhPi
1 1
0
0
i~k~x0 3 0
~
~
~
=
J0 (x ) k e
d x
d
2 (4)2 0

(32)

(33)

## This can be used in many instances, for example for radiation

from antennas, as will be seen in later chapters.

Coming up...

## Dipole antenna from a coaxial cable

Feeding through a coaxial cable:

## An antenna may be represented as a monochromatic sinusoidal

current with frequency :
 

I0
L
~J0 (~x0 ) =
z
(x 0 )(y 0 ) sin k
|z 0 |
(34)
sin(kL/2)
2
(Note: direction of current is the same for z > 0 and z < 0)

## Calculating power radiated

We have ~x0 = (0, 0, z 0 ). Using azimuthal symmetry, we choose
r = (sin , 0, cos ), so that ~k = k (sin , 0, cos ), so that
 

~J0 (~x0 ) ~k = I0 k sin (x 0 )(y 0 ) sin k L |z 0 |

y
sin(kL/2)
2
~k ~x0 = kz 0 cos
(35)
Then we get
Z

(~J0 (~x0 ) ~k)ei~k~x0 d 3 x 0




 
2I0
kL cos
kL
=
cos
cos
(36)
sin sin(kL/2)
2
2
This can be used to calculate the radiation pattern.

## Dependence of radiation pattern on kL

In the large wavelength limit, kL  1
Z

(~J(~x0 ) ~k)ei~k~x0 d 3 x 0 = I0 kL sin

2

(37)

## The average power radiated is then

dhPi
d

=
=

2
r Z 


1 1
0
~J0 (~x0 ) ~k ei~k~x0 d 3 x 0

2 (4)2 0
r
1 1
0
(I0 kL)2 sin2
(38)
8 (4)2 0

## Dipole radiation pattern !

As kL increases, stronger angular dependences appear.
Check CDF demo at
http://demonstrations.wolfram.com/DipoleAntennaRadiationPattern/

## Antenna patterns: dipole

http://www.cisco.com/en/US/prod/collateral/wireless/
ps7183/ps469/prod_white_paper0900aecd806a1a3e.html

## (Note: scale in dB)

http://www.cisco.com/en/US/prod/collateral/wireless/
ps7183/ps469/prod_white_paper0900aecd806a1a3e.html

## (Note: scale in dB)

http://www.cisco.com/en/US/prod/collateral/wireless/
ps7183/ps469/prod_white_paper0900aecd806a1a3e.html

## (Note: scale in dB)

http://www.cisco.com/en/US/prod/collateral/wireless/
ps7183/ps469/prod_white_paper0900aecd806a1a3e.html

## (Note: scale in dB)

http://www.antenna-theory.com/antennas/main.php

## Recap of topics covered in this lecture

~ and B
~ fields in the presence of moving sources
E
~ and B:
~ the 1/r behaviour that
Radiative components of E
dominates at large distances
Poynting vector and power radiated by EM waves
Long distance approximation for radiated power
Radiation patterns in antennas