Chapter 9: Radiating Systems, Multipole

Fi ld and
Fields d Radiation
R di ti
An Overview of Chapters on EM Waves : (covered in this course)
source term ini wave equation
i boundary
b d
Ch. 7 none plane wave in space or in
semi- spaces separated
two semi
by the x - y plane
Ch. 8 none conducting walls
Ch. 9 J , e-it outgoing wave to
prescribed, as
in an antenna
Ch. 10 J, e-it outgoing wave to
ind ced bby incident EM waves,
induced a es
as in the case of scattering of a
plane wave by a dielectric object.
Ch. 14 moving charges, outgoing wave to
such as electrons in a synchrotron
 9.6 Spherical Wave Solutions of the Scalar
Wave Equation
Spherical Bessel Functions and Hankel functions : Although
this chapter deals with radiating syatems, here we first solve the scalar
source-free wave equation in the spherical coordinate syatem. The
purpose is to obtain a complete set of spherical Bessel funtions and
Hankel functions, with which we will expand the fields produced by
the sources.
th
The scalar source-free wave equation is [see (6.32)]
2 1 2
( x, t ) 2 2 ( x, t ) = 0 (9.77)
c t

Let (x, t ) = ( x, )e it d (9 78)
(9.78)
Each Fourier component satisfies the Helmholtz wave eq.
( 2 + k 2 ) (x, ) = 0 with k 2 = 2 c 2 (9.79)
 9.6 Spherical Wave Solutions (continued)

In the spherical
p , ( 2 + k 2 ) = 0 is written
coordinate system,
y
1
r 2 r ( r 2
r
+ ) 1
r 2 sin ( sin
) + 1 2
r 2 sin 2 2
+ k 2 = 0
L = U (r ) P( )Q( ),
Let ) we obtain
b i
PQ 12 r
r ( r 2 U
r ) + UQ 2
r sin
1
( sin P
) + UP 2 2 2Q
1
r sin 2
+ k 2UPQ = 0
r 2 sin 2 The -dependence is isolated within this term,
Multiply by
UPQ
Q so this term must be a constant. Let it be m2.
P )] + 1 Q
2
sin 2
[U1 r (r 2 Ur ) + k 2 r 2 + 1
P sin
(sin Q 2
=0
=l (l +1))
=m2
Dividing all terms by sin 2 , we see that the
2l + 1 (l m))! m
Pl (cos )eim
r -dependence is isolated within this term. So
this term must be a constant. Let it be l (l + 1).
4 (l + m)!
Thus, as in Sec. 3.1 of lecture notes,
( )), Qm ((cos )); Q = eim , eim PQ = Ylm ( , )
P = Pm (cos
rejected because of divergence at =
 9.6 Spherical Wave Solutions (continued)

U(r) is a function of l. Rewrite it as fl (r ).

) Then,
Then
d 2 + 2 d + k 2 l (l +1) f ( r ) = 0 (9.81)
dr 2 r dr r2 l

Let fl (r ) = 1/ 2 ul (r ) 2 2 (l +1/ 2)
2
2 + r dr + k ul (r ) = 0 (9.83)
1 d 1 d
r dr r 2

ul (r ) = J l + 1 (kr ), Nl + 1 (kr ) [Bessel functions of fractinal order]
2 2

fl ( r ) = 1
1/ 2 J l + 1 ( kr )
), 1
1/ 2 Nl + 1 (kr )
r 2 r 2

( )
1
= 2 J
l (kr ) = jl (kr ) + inl (kr )
(1)
j
l ( kr ) 2 kr
k l+21 ( kr ) h
Define and
nl (kr ) = ( 2kr ) 2 Nl + 1 (kr )
1
hl(2) (kr ) = jl (kr ) inl (kr )
2

spherical Bessel functions Hankel functions

(x, ) = Alm hl (kr ) + Alm

(1) (1)
hl (kr ) Ylm ( , )
(2) (2)
(9.92)
lm
 9.6 Spherical Wave Solutions (continued)
n0 ( x )
j0 ( x) j0 ( x = 0) = 1
n1 ( x)
j1 ( x) n2 ( x)
j2 ( x )
x

x From G.
G Afken,
Afken
"Mathematical Methods
for Physicists"

jl ( x) 1x sin ( x l2 )
x l

1x cos ( x l2 )
x 1, l l 2 x l
jl ( x)
(2l +1)!! 1 2(2l +3) +
x x nl ( x)

x l l +1 eix
(2l l+1)!!
x 1, l
nl ( x) 1 x + 2
hl ( x) (i ) x
(1)
[ spatial
x 1
2(12l ) dependence of spherical waves
waves.]]
See Jackson pp. 426-427 for further properties of jl , nl , hl(1) , and hl(2) .
 9.6 Spherical Wave Solutions (continued)

Expansion of the Green function : Solution of the Green equation

( 2 + k 2 )G (x, x) = 4 (x x) (6.36)
is g
given byy (derived
( in Sec. 6.4.))
G (x, x) = xx in infinite space and for outgoing- (6.40)
ik x x
e
wave boundary condition.
We may also solve (6.36) in the same way as in Sec. 3.9, i.e. write
G (x, x) = gl (r , r )Ylm *
( , )Ylm ( , ),
lm
solve for gl (r , r ) for r > r and r < r [where (x x) = 0], and then
pp y boundaryy conditions at r = 0,, r = , and r = r . The result is
apply
l
G (x, x) = 4 ik jl ( kr< ) hl (kr> ) Ylm
(1) *
( , )Ylm ( , )
l =0 m =l
Equating the two expressions above for G (x, x), we obtain
l
eik x x = 4 ik jl ( kr< ) hl ( kr> ) Ylm
(1) *
( , )Ylm ( , ), (9.98)
x x
l =0 m =l
where r< and r> are, respectively, the smaller and larger of r and r .
 9.6 Spherical Wave Solutions (continued)

S
Summary off Diff
Differential
ti l E
Equations
ti and
d Solutions
S l ti :
Source-free D.E. Laplace eq. 2 = 0 (
Helmholtz eq. 2 + k 2 = 0 )
Solutions
i x i y 2 + 2 z eik x x , eik y y , eik z z , etc.
Cartesian e , e , e , etc.
(Sec. 2.9) (Sec 8.4)
(Sec. 8 4)
2
J m ( kr ), eim , e kz , etc. J m 2 k z2 r , eim , eik z z , etc.
cylindrical c
(Sec 3.7)
(Sec. 3 7) (Sec. 8.7)

Y
lm ( , ), r l
, etc. Ylm ( , ), jl (kr ), nl (kr ), etc.

spherical
p (Secs.
(Secs 3.1, 3 1 3.2)
3 2) (Sec 9.6)
(Sec. 9 6)
D.E. with a 2G (x, x) = 4 (x x) ( 2 + k 2 )G (x, x) = 4 (x x)
point source b.c. G () = 0 b.c. outgoing wave
Solutions ik x x
G= 1
x x
G= e
x x
[Eq. (6.40)]
(Green functions)
Series expansin Eqs. (3.70), (3.148), (3.168) Eq. (9.98)
of Green function
 9.1 Radiation of a Localized Oscillating Source
Review of Inhomogeneous Wave Equations and Solutions :
2 1 2 = (6.15)
c 2 t 2 0
in free space, and A
2 satisfy Lorenz gauge.
2
=
A 1 A J (6.16)
c t
2 2 0
Basic structure of the inhomogenous wave equation:
2 1
2 2
2
= 4 f ( x, t ) (6.32)
c t
Solution of (6.32) with outgoing-wave b.c.:
(x, t ) = in (x, t ) + d 3 x dt G + (x, t , x, t ) f (x, t ) (6.45)
x x f (x, t ) in (6.45)
t (t c )
where G + (x, t , x, t ) = is evaluated at
x x the retarded time.time ((6.44))
is the solution of
)G + (x, t , x, t ) = 4 (x x) (t t )
2
( 2
2 2
1 (6 41)
(6.41)
c t
with outgoing wave b.c.
 9.1 Radiation of a Localized Oscillating Source (continued)

Using (6.45) (assume ini = 0) on (6.15) & (6.16), we obtain the

gereral solutions for A and , which are valid for arbitrary J and .
x x
A ( x, t ) 1 t (t
) 0 J (x, t )

x dt
c
=
3
d x x (6.48), (9.2)
(x, t ) 4 (x, t ) / 0
In general, the sources, J (x, t ) and (x, t ), contain a static part
and a time dependent part. For static J (x) and (x), (9.2) gives the
static A and in Ch. 5 and Ch. 1, respecticely.
0 3 J (x)
A ( x) = A ( x) = d x (5.32)
4 x x x x
1 (x) source
( x) = ( x) = d x
3
(1.17)
(1 17)
4 0 x x
Question: It is stated on p. 408 that (9.2) is valid provided no
b d
boundary surfaces
f are present. Why?
h [See discussion
di i below
b l (6.47)
( )
in Ch. 6 of lectures notes.]
 9.1 Radiation of a Localized Oscillating Source (continued)

Fields by Harmonic Sources : Only time time-dependent

dependent sources can
radiate. Radiation from moving charges are treated in Ch. 13 and
Ch. 14. Here, we neglect any static or J and specialize to sources
of the form (as in an antenna):
(x, t ) = (x)eit
(9 1)
(9.1)
it
J ( x, t ) = J ( x ) e
Sub. (9.1) into (9.2) and carryy out the t -integration,
g we obtain
0 3 eik xx
A (x, t ) = A (x)e it with A( x) = d x J ( x), (9.3)
4 x x
where k = / c.
Question: Why are the charge distribution and scalar potential
nott required
i d for
f the
th determination
d t i ti of H andd E?
Note: We shall assume that J(x) is independent of A(x), i.e. the
source will not be affected by the fields they radiate.
radiate Otherwise,
Otherwise
(9.3) is an integral equation for A(x).
 9.1 Radiation of a Localized Oscillating Source (continued)

A simpler derivation of (9.3): We specialize to harmonic sources

from the outset. Then, only (6.16) is required.
2
A ( x, t ) 2 2
1 2
A ( x, t ) = 0 J ( x, t ) (6.16)
c t
Let J ( x, t ) = J ( x)e it and A( x, t ) = A( x)e it
( )
2 + k 2 A (x) = 0 J (x) [inhomogeneous Helmholtz wave eq.]
The Green equation
q for the above equation
q is
( )
2 + k 2 Gk (x, x) = 4 (x, x) (6.36)
S l ti off (6
Solution (6.36)
36) with
ith outgoing
t i wave b.c.
b
ik x x
Gk (x, x) = xx
e (6.40)
0 ik x x
A ( x) = d 3
xGk (x, x) 40 J (x) = 4 d x xx J (x),
3 e

which
hi h is
i (9.3).
( )
 9.1 Radiation of a Localized Oscillating Source (continued)

R it (9.3),
Rewrite (9 3)
0 3 eik xx
A ( x) = d x J (x), (9 3)
(9.3)
4 x x
H = 1 A ((everywhere)
y ) ((9.4))
0
Maxwell eqs. give
iZ 0
E = k
H (outside the source) (9.5)
where Z 0 = 0 0 = 377 (impedance of free space, p. 297).
Thus, given the source function J (x), we may in principle evaluate
A (x) from (9.3) and then obtain the fields H and E from (9.4) and
(9.5).
Note that eit dependence has been assumed for J , hence all
other q
quantities which are expressed
p in terms of J.
 9.1 Radiation of a Localized Oscillating Source (continued)

0 3 eik xx
Near - Field Expansion of A(x) = d x J (x) (9.3)
4 x x
Before going into algebraic details, details we may readily observe some
general properties of A (x) near the source (r ).
ik x x
For x outside the source and r ((or kr 1), ), we let e 1
1 l 1 r<l *
and use = 4 +
Ylm ( , )Ylm ( , ). (3.70)
x x l =0 m =l 2l + 1 r>
l 1

with r> = r and r< = r .

l 1 1
A ( x) 0 +
Yl
lm ( , ) d 3
x J ( x ) r l ( , ) (9
l *
Ylm (9.6)
6)
l =0 m =l 2l + 1 r
l 1
kr 1
The integral in (9.6) yields multipole coefficients as in (4.2). Thus,
(9.6) shows that A (x) can be decomposed into multipole fields, which
fall off as l1+1 just as the static multipole fields, but with the e it
r
dependence. We will show later that, far from the source (kr 1), A (x)
behaves as an outgoing spherical wave .
 9.1 Radiation of a Localized Oscillating Source (continued)

Full Expansion of A(x): We may in fact expand A (x), ) without

approximations, by using (9.98). For x outside the source, we have
r> = x = r , r< = x = r . Hence,, ((9.98)) can be written
ik x x l
e
xx
= 4 ik jl (kr ) hl ( kr ) Ylm
(1) *
( , )Ylm ( , )
l =0 m =l
ik x x
0 3 e
Sub. this equation into A( x) = d x J ( x), we obtain
4 x x
A (x) = 0ik hl(1) (kr )Ylm ( , ) d 3 xJ (x) jl ( kr )Ylm
*
( , ), (9.11)
l ,m
eikr (2l 1)!! l
where hl (kr ) =
(1)
a (ikr ) n
x
i ( kr )l +1 n =0 n x
source
( 1)n (2l n )!
with an = (2l 1)!!(2l 2 n )!!n! (a0 = 1, a1 = 1)
(See Abramowitz & Stegun,
Stegun "Handbook of Mathematical Functions,"
Functions "
p. 439.)
 9.1 Radiation of a Localized Oscillating Source (continued)

(9.11) is an exact expression for A(x). We now assume kd 1 (i.e.

source dimension wavelength). Then, kr 1 and jl (kr ) reduces to
( kr)l kd 1

jl (kr ) kr 1 = (2l +1)!! (9.88)
(9 88)
n
(2l 1)!! l x
ikr
e
Sub. hl(1) (kr ) = an (ikr ) n d x
k )l +1
i ( kr n =0 x = (r , , )
J,
and (9.88) into (9.11), we obtain x = ( r , , )
1 Ylm ( , ) elikr+1 [1 + a1 (ikr ) + a2 (ikr ) 2 + + al (ikr )l ]
2l +1
A ( x) = 0 r
(1)
l ,m d 3 xJ ( x) r l Y * ( , )
lm
(1) is the combination of (9.6) and (9.12) in Jackson. It is valid for
kd 1 and anyy x outside the source. The region g outside the source is
commonly divided into 3 zones (by their different physical characters):
The near (static) zone: d r ( kr 1)
The
h iintermediate (i d i ) zone: d r ( kr
di (induction) k 1))
The far (radiation) zone: d r ( kr 1)
 9.2 Electric Dipole Fields and Radiation
Rewrite (1):
1 Y ( , ) eikr [1 + a (ikr ) + a (ikr ) 2 + + a (ikr )l ]
2l +1 lm l +1 1 2 l
A ( x ) = 0 r
((1))
l ,m d 3 xJ ( x) r l Y * ( , )
lm
Take the l = 0 term [Y00 = 1 ] J x dxdydz
4
= dydz xJ x d x xx dx
p d J
and denote it by A (x)
l =0 0 eikr
A (x) = A(x) = 4 r d xJ (x)
p 3
J x J y J z
i 0 eikr
= x( x + y + z )dxdydz
= 4 p r , (9.16)
give no contribution because J
h p = x (x)d x
where 3
(4
(4.8)
8) J y
dy = J y
d
=0
is localized: y
d
(9.16) gives the electric dipole
contribution to the solution.
solution It is = x Jd 3 x
valid for kd 1 and any x outside Jd 3 x = x Jd 3 x
the source. = i x (x)d 3 x = ip
Question: Why is there no monopole
J + t = 0 p
term (see p. 410)?
 9.2 Electric Dipole Fields and Radiation (continued)
i0 eikr
Rewrite (9 16): A (x) =
(9.16): p
4 p r (9 16)
(9.16)
iZ 0
From (9.4), H p = 1 A and from (9.5), E p =
p
k
H p
0
H p = ck 2 ( n p ) e (1 1 )
ikr
4 r ikr

{ ( ) }
(9.18)
k ( n p ) n e r + [3n ( n p ) p ] 13 ik2 e
ikr
E = 4
p 1 2 ikr
0 r r
In the far zone (kr 1), (9.18) reduces to a spherical wave
H p ck 2 ( n p ) eikr p component
4
p r
of source n (9.19)
E Z0H np
x
In (9.19), we see that E p and H p kd 1
p p d
phase and E , H , and n are
are in phase,
mutually perpendicular. This is a general property of EM waves in
unbounded,, uniform space. p Given anyy two of these qquantities,, we
can find the third.
 9.2 Electric Dipole Fields and Radiation (continued)

H p = ck 2 ( n p ) eikr (1 1 )
4 r ik
ikr

{ ) }
p
(
(9.18)
E = 4 0 k ( n p ) n r + [3n ( n p ) p ]
ikr
1 2 e 1 ik2 e ikr
r3 r
In the near zone (kr 1), (9.18) reduces to
H p i ( n p ) 12 p component
4 r of source n
p (9.20)
4 0 [
3n ( n p ) p ] 3
1 1 x
E r kd 1
d
(i) E p and H p are 90o out of phase average power = 0.

(ii) E p has the same spatial pattern as that of the static electric
it
dipole in (4.13), but with e dependence.
(iii) 0 H 2 (kr ) 2 0 E 2 E-field energy B-field energy.
Questions: (i) Why does E p have the static field pattern? (ii) To
obtain (9.20), we have neglected a few terms in (9.18). But some of
the
h neglected
l d terms are still
ill important
i in
i the
h near zone? Whath are
they and in what sense are they important?
 9.2 Electric Dipole Fields and Radiation (continued)
dP
d t
= time
time-averaged
averaged power in the far zone/unit solid angle

(
= 12 Re r 2n E p H p*
) (9.21)
(9 19)
(9.19)
c2Z0
=
32 2 k 4
| ( n p ) n |2
(9.22)
This vector gives the direction of Ep , i.e.
i e the
polarization of the radiation (see figure below.)
c 2 Z0k 4 2
P t = total power radiated = 12
p (9 24)
(9.24)
In general, p = p x ei e x + p y ei e y + pz ei e z . If p H
p n
= = , then p has a fixed direction, p = p 0ei x p
E
with p 0 = p x e x + p y e y + p z e z , and
c2Z0
dP
d t
= k 4
p sin 22
. (9.23)
32 2

Otherwise the direction of p and

Otherwise, dP vary
d t dipole radiation pattern
with time, but P t
is still given by (9.24).
 9.3 Magnetic Dipole and Electric Quadrupole Field
Rewrite (1):
1 Y ( , ) eikr [1 + a (ikr ) + a (ikr ) 2 + + al (ikr )l ]
2l +1 lm l +1 1 2
A ( x) = 0 r
(1)
l ,m d 3 xJ ( x) r l Y * ( , )
lm
Take the l = 1 terms [a1 = 1]]
l =1 0 eikr
A ( x) = 3 r2
(1 ikr ) Y1m ( , ) d 3 xJ (x)r Y1*m ( , )
m =1,0,1

p. 109
Y1m ( , )Y1*m ( , ) = 83 sin sin ei ( )
m =1,0,1 n

+ 43 cos cos + 83 sin sin ei ( ) x
d x
= 43 [sin sin cos( ) + cos cos ]

= 43 cos = 43r n x
set l = 1 in (3.68)
 9.3 Magnetic Dipole and Electric Quadrupole Fields (continued)

Th
Thus,
l =1 0 eikr 1
= ( ik ) xJ (x)(n x)
3
A ( x) 4 r r
d (9.30)
0 eikr 1
= 4 r r
( ik ) { d 3 x 12 [(n x) + (n J )x] + d 3 x 12 (x J ) n}
electric
l t i quadrupole
d l radiation
di ti magnetic
ti di
dipole
l radiation
di ti
= AQ + A m ,
ik 0
( ikr ) x outside the source
for
f kd 1 and d any (9.33)
ik
ikr
where A (x) =
m
4
1
(n m) r
1e

with m = 12 (x J )d 3 x [magnetic dipole moment].

moment] A m gives the
magnetic dipole contribution through (9.4) and (9.5):

4 {
H m = 1 k 2 (n m) n eikr + [3n(n m) m ]
r ( 1
r3 r ) }
ikk2 eikr
ik
(9.35)

Em = Z0 k 2 (n m) e (1 1 )
ikr
4 r ikr (9
(9.36)
36)
 9.3 Magnetic Dipole and Electric Quadrupole Fields (continued)
In the far zone (kr 1),
), we have the spherical
p wave sloution:
H m dP Z0 4
2
( )
k 2 n m n eikr
4 r d t 32 2 k | n m |

2 direction of E
m
Em Z0H n
m
P Z0 4
t 12 k | m |
In the near zone (kr 1) (i) Em andd H m are 90o outt off phase
1), h
average power = 0.
4 [ ( ) ] r3
H m 1 3n n m m 1
(ii) H m
has the same spatial pattern
m Zk
E 0
4 i ( n m ) 1
2
as that of the static magnetic dipole
r it
in (5.56),
( ), but with e dependence.
p
(iii) B-field energy E-field energy.

The electric qquadrupole p radiation,,
discussed in (9.37)-(9.52), is more quadrupole
radiation
complicated. Here, we only illustrate its pattern
radiation pattern by the figure to the right.
 Comparison between Static and Time-dependent Cases
relations r -dependence
p of E
multipole
lti l definition
d fi iti off multipole
lti l
between , and B (d : dimension
expansion moments
J, E, and B of the source)
E, B 1/ r l + 2
spherical
harmonics

q = (x) d 3 x
For r d , all multipole
static ( x ) E( x ) expansion p = x ( x)d 3 x fields can be significant.
[(3.70)] or For r d , multipole
case J ( x) B( x) y series
Taylor Qij = (3 xi xj r 2 ij ) ( x)d 3 x fields are dominated by y
[(4.10)] of the lowest-order
1 m = 1 x J ( x)d 3 x nonvanishing term.
x x 2

There is no time-dependent (a) near zone r d

monopole for an isolated E, B eit / r l + 2
source (see p. 410). Approx. the same field
p, Qij , and m have the pattern and r -dependence
as for the corresponding
( x) E( x ) spherical same expressions as those
static multipole, but with
time- of their static counterparts,
eit dependence (hence
harmonics
dependent B ( x ) expansion but with the eit time
called qquasi-static fields.))
case J ( x) [(9 98)] of
[(9.98)] dependence
dependence.
ik x x
(b) far zone r d
EM waves In time-dependent cases,
E, B eikr it / r
e
x x electric multipoles can
(spherical EM waves)
generate B-fields and
magnetic
ti multipoles
lti l can All multipole fields 1/ r ,
generate E-fields. relative power levels
unchanged with distance.
 9.4 Center-Fed Linear Antenna
AQ
Qualitative
lit ti Look
L k att the
th CCenter
t -FFed
d Li
Linear Antenna
A t :
C Traveling wave

Energy i
L
source Transmission i
line
R Center-fed linear antenna
LC oscillator
i
+

i

near zone far zone
In the near zone, E and B are principally generated by and J ,
respectively ( largely static field patterns).
patterns) In the far zone
zone, E and
B are regenerative through dtd B and dtd E ( EM waves).
 9.4 Center-fed Linear Antenna (continued)
Detailed Analysisy : The center-fed linear antenna is a case of
special interest, because it allows the solution of (9.3) in closed form
for any value of kd, whereas in Secs. 9.2 and 9.3, we assume kd<<1.
0 3 eik xx
A ( x) = d x J (x), (9.3)
4 x x
where J (x) = I sin ( kd2 k z ) ( x) ( y )e z (9.53)
sin ( kd2 k z ) e ik xx

0 I d / 2
A ( x) = e z
4 d / 2
dz
x x
z

Note: (i) J is symmetric about d d d n

2 2 J 2
z = 0. J ( z ) = J ( z ) J
d
z=0
d J d
(ii) I is the peak current 2 2 2
only when kd .
Q
Question: The
h antenna appears to bbe an open circuit.
i i How can
there be current flowing on it?
 9.4 Center-fed Linear Antenna (continued)
1 1
x x = (r 2rr cos + r ) = r[1 ( r 2 )]2
2 2 2 2nx r 2

r n
= r[1 12 ( 2nrx r2 ) 18 ( 2nrx r2 ) + ]
2 2 2

r r
x x
= r n x + 21r [r 2 (n x) 2 ] + r=x (2)
x x r n x if r r r = x
Hence, if r d , we can write x x r z cos .
0 Ieikr d /2 sin ( kd2 k z ) eikz cos
A ( x) e z
4 d / 2 dz r z cos
(9.54)
r d r
0 Ieikr cos ( kd cos ) cos ( 2 )
kd
= ez
2
(9.55)
2 kr sin
2

Note: z cos in r z1cos can be neglected if r d . But z cos
iin eik ( r z cos ) makes
k an important
i contribution
ib i to the
h phase
h
angle even at r d .
 9.4 Center-fed Linear Antenna (continued)

In the far zone,

one k sin A
H = n E / Z0 H = 1 A = ik n A H = 0
0 0

= 12 Re r 2n E H* = 20 r 2 H = A
dP Z 2 Z0
2 2 2 2
d t
k r sin ((3))
2 0
2
2 z
cos( 2 cos ) cos( 2 )
kd kd
, for r d
Z0 I 2
= (9.56)
8 2
i
sin and any kd d n
2
z=0
2 cos ( cos ) / sin , kd =
2 2
Z0 I d
= 2 2 2 (9 57)
(9.57)

8
4 cos ( 2 cos ) / sin , kd = 2
4 2

half-wave antenna full-wave antenna

(kd = ) (kd = 2 )
z z dP superposition of 2 half-wave
d antennas excited in phase
dP
d
d J
2
most coherent
d J d
2 J
less coherent
narrower beam width
 9.4 Center-fed Linear Antenna (continued)

Rewrite (9.56)
cos ( kd cos ) cos ( )
kd 2
, for r d
Z0 I 2
dP = 2 2
(9.56)
d t 8 2 sin and any kd
Limiting case: kd 1 ( d)
cos x 1 x2 (x 1)
2
cos ( kd cos ) 1 k 2d 2 cos 2
2 8

cos ( 2 ) 1 8
2 2
kd k d

2 2 2
2 1 k d cos 1 +
2 k 2d 2
Z0 I
dP
d t
8 8
8 2 sin
Z0 I 2
= ( kd ) 4
sin 2
[valid for kd 1] (4)
512 2

This
hi has
h theh same k andd depedence
d d as in
i (9.23),
( ) which
hi h was
derived by assuming kd 1.
 9.4 Center-fed Linear Antenna (continued)

Radiation Resistance and Equivalent Circuit:

( )
J (x) = I sin ( kd2 k | z |) ( x) ( y )e z kd2 I 1 2|dz| ( x) ( y )e z
kd 1 I 0 (peak current, z d )
current
Z0 I 2 Z 0 I 02
2 ( kd ) sin = (9.28)
dP 4 2 2 2
Thus, from (4), d t
( kd ) sin
512 128 2

2 Z 0 I 02
0 d 1 d cos d t = 48
1
P t dP
d t
d = dP (kd )2 (9.29)
Rrad : radiation
di ti resistance.
it
I 02
= Rrad , Rrad is part of the field definition of
2
impedance,
impedance see 2nd term in (6
(6.137)
137).
Z
where Rrad 240 (kd )2 5(kd ) 2 ohms [See pp. 412-3.]
I 0e it
I 0e it d
2
Equivalent
V0eit V0e it Rrad
circuit for
a center-fed
center fed
d
circuit antenna
2
dimensions
 9.4 Center-fed Linear Antenna (continued)
Problems:
1. The full-wave antenna radiation in (9.57) can be thought of as the
superposition of two half-wave antennas, one above the other,
excited
it d in
i phase.
h D
Demonstrate
t t this
thi by d i i dP/d for
b rederiving f the
th
full-wave antenna [(9.57), kd = 2] by surperposing the fields of
two half
half-wave
wave antennas (each of length d/2,
d/2 see figure below).
below)
2. If the two half-wave antennas in problem 1 are excited 180 out of
phase, derive dP/d again by the method of superposition.
3. Plot the approximate angular distribution of dP/d in problems 1 P
and 2. Explain the difference qualitatively.
z r 1
d antenna 1 of r
n
2 length d2 d
2
z=0 These 3 line are nearly
d r2
2 antenna 2 of parallel when point P is
d
single antenna l h d2
length 2 f ffrom th
far the antenna,
t as
of length d d cos is assumed here.
4
 9.4 Center-fed Linear Antenna (continued)

Solution to problem 1: Principle of superposition requires that we

add the fields (not the powers) of the 2 antennas, each of total length d2 .
z
Rewrite (9.55)

A ( x) = e z
0 Ie

( 2 ) ( 2 )
ikr cos kd cos cos kd

d n
2
z=0 (9.55)
2 kr sin
2

d
2

(9.55) applies to a single antenna of total length d (see fig. above.)

So the field of each of the 2 antennas in this problem can be obtained
from (9.55) by replacing d in (9.55) with d2 and expressing r with r
1
respect to the center of each antenna (i.e. by r1 and r2 ).
r
cos ( kd4 cos ) cos ( kd
4 )
antenna 1 d
ikr
0 Ie 1,2 2
A1,2 = ez , (5) r2
2 kr1,2
12 sin 2 d
antenna 2 2
where r1 = r d4 cos and r2 = r + d4 cos . d cos
4
We may approximate r1,2 in the denominator of (5) by r , but must
use the correct r1,2 for the phase angles in the exponential terms.
 9.4 Center-fed Linear Antenna (continued)

It is assumed that each antenna in this problem is excited in the

half-wave pattern, hence we set k d2 = in (5) and the superposed
field of the 2 antennas (excited in phase) is given by
0 I ikr i 2 cos i 2 cos cos(2 cos )
A = A1 + A 2 = e z e [e +e ] (6)
2 kr
k i
sin 2

0 I ikr cos (2 cos )

2
= ez e antenna 1
d
kr sin 2 2
antenna 2 d2
k r sin A
Z0 2
From (3), dP
d t
= 2 2 2
2 02
same as the full wave
=
Z0 I 2
2 2
(
cos 4 2 cos ) sin
2
solution in ((9.57))
Solution to problem 2:
If the two half-wave antennas in pproblem 1 are excited 180o out
of phase, we simply replace the " + " sign in (6) with a " " sign.
 9.4 Center-fed Linear Antenna (continued)

Thus,
Thus
0 I ikr i 2 cos i2 cos cos(2 cos )
A = A1 A 2 = e z e [e e ]
2 kr sin 2
0 I ikr sin(2 cos ) cos(2 cos )
= ie z e
kr sin 2 antenna 1 d2
A
Z0 2
From (3), dP
d t
= k 2 2
r sin 2
antenna 2 d
2 0
2 2

2 sin ( cos ) cos ( cos ) sin 2 ( cos )

2 2
Z0 I Z0 I 2
= 2 2 =
2 2 sin 2 8 2 sin 2
S l ti to
Solution t problem
bl 3:

phase dipole out of

phase qu
in p p radiation quadrupole
d upo e radiation
d o
Question: How does a phased array antenna work?