Escolar Documentos
Profissional Documentos
Cultura Documentos
IN INHOMOGENEOUS ~ffiDIA
by
Dikran Damlamayan
Antenna Laboratory
November 1969
-ii-
ACKNOWLEDGEMENT
Professor Charles Herach Papas, for his continued gUidance and support
during the course of this work, as well as for the unceasing encourage-
I also wish to thank Dr. Henri Hodara of Tetra Tech for his
ABSTRACT
admittance and the radiation field are expressed here, for the first
These quantities are the input admittance of the dielectric slab and the
the slab to the transverse field at the other, both calculated for plane
dielectric slab.
A simple and powerful method has been devised for the computa-
techniques.
TABLE OF CONTENTS
1. INTRODUCTION 1
B. Aperture Admittance 24
C. Radiation Pattern 39
DIELECTRIC SLABS 46
A. Perpendicular Polarization 48
B. Parallel Polarization 56
REFERENCES 84
-1-
1. INTRODUCTION
cation between the vehicle and ground stations. As a result, the study
This report presents a new and simple method for the calcula-
and simplicity of the method stern from the fact that the usual boundary-
situation, is bypassed.
the slab. Similarly the slab transmission coefficients enter into the
a way that it can readily be applied for any aperture shape and corres-
configurations.
along the direction normal to the slab faces. The present method is
particularly useful for slabs which are not homogeneous isotropic media,
field.
-3-
many studies for almost half a century. However since this report deals
for the far field from an aperture in an infinite plane screen on which
coaxial waveguide and radiating into free space. Using similar techniques
homogeneous plasma layer has been calculated by Galejs (5). However his
formulation is quite involved and does not apply to plasma layers whose
electron densities are below the critical density, and hence the relative
permittivity varies between zero and unity. In precisely this range the
plasma does not support surface waves and hence they have not had to
tion has been modified by Croswell, Rudduck and Hatcher (11) to account
dielectric slabs with permittivity greater than one. Fante (12) has
notion. Finally, Bailey and Swift (13) have calculated the admittance
Swift and Cockrell (14) suggested a method for the calculation of the
of the fields in the plasma region, is too involved even for numerical
-5-
geneous plasma slabs have been analyzed by various authors. Tamir and
covered by a plasma layer and have also considered the effect of surface
wave poles on the radiation field. Knop and Cohn (16) have found the
plasma sheaths have been studied by Hodara and Cohn (17) and Hodara (18).
The third chapter is the main body of the report. Here the
plane waves. Using this idea, the aperture admittance and the radiation
pattern are calculated in terms of the plane wave reflection and trans-
-6-
common geometries.
such media. This is done in the fourth chapter. Besides the differential
equations are derived which yield directly the input admittance and the
ratio of the total electric field amplitude at one end of the slab to
nature is well known, however it will be wise to start our analysis from
this point for the sake of presenting a complete treatment of the subject.
z > ° whose parameters may vary along the z-direction (Fig. 2.1).
only this mode will be incident on the aperture. However higher order
the aperture, along with the reflected dominant mode. The transverse
E (x,y)
-0
V e (x,y)
0-0 +2: n
V e (x,y)
n -n
(2.1)
Waveguide
feeding
aperture
Infinitely conducting
ground plane
e (x,y), h (x,y) for the dominant mode, and e (x,y), h (x,y) for
"-1J "-1J -n -n
the higher-order modes, depend on the particular waveguide cross-section
h. u x e. (2.3)
-J. -J.
I y V (2.5)
n n n
relations (2.3), (2.4), (2.5) and (2.1), equation (2.2) can be written as
1!o(x,y) = I o u x e (x,y)
-0
- \'Y u
L n-
x e (x,y)
-n
JE "-1J
(x,y) • e (x,y) dA
-n
(2.6)
n A
tions. The integration in (2.7) is over the aperture area only, since
the tangential electric field vanishes over the conducting ground plane.
I e (x,y)
0-0
= \'
L Y e (x,y)
rrn
fE -0
(x,y).e (x,y) dA
--n
n A
(2.9)
- ~ u x g(x,x',y,y',O,O) x u · -0
E (x' ,y') dA'
A'
Y= I
o
Iv0 (2.10)
-11-
The result is
1
Y
[J ~
A
(x,y) • -e0 (x,y)
(2.11)
-.if
AA'
!o(x,y) • (~x G(x,x' ,y,y' ,0,0) x~) • --0
E (x' ' y' ) dAdA}
Y could readily be found from (2.11). The exact field could only be
imate aperture electric field into (2.11) will still yield a good
geneous in E
o
It is easy to verify that the variation of Y due to small
variations oE of E disappears. Making use of the symmetry property
o 0
e e
-0 -0
(2.13)
. ~ dA - f (~
A'
x .Q. x~) • ~ oE- 0 dA
oY (2.14)
°
The variational expression (2.11) can be put in the more
convenient form
~E -0
(x,y) x -H0 (x,y) • u dA U~(x.y) . e (x,y)
-n d~ 2
Y
A
+ LY A (2.15)
U~(x,y)
A
. e (x,y)
-0 dAJ
2
n n [J~(x.y)
A
. e (x,y)
-0 df
-13-
and the second term in (2.15) involving the summation vanishes because
y = -lz-j
V
E (x,y) x H (x,y) • u dA
-0 -0
(2.17)
o A
in most instances (11, 14, 20). However, for a more exact treatment,
dominant mode and some higher order modes. For simplicity let us
consider only one higher order mode, with mode function e., which would
-~
be the next most highly excited mode, as indicated by the geometry of the
-14-
(2.18)
= c::.o
E (x,y) +
0
a. E .(x,y)
J. - 0 J.
H (x,y)
-0
= -00
H (x,y) + a. H . (x,y)
J. -OJ.
(2.19)
Y = _1
V 2
o
{IE
A
-00
x H
-00
udA+ 2a.
J. fi -00
A
x .!!ai • u dA
+ a.J.
2JE . -OJ.
x H .• U
-OJ.
Y.J.
A
2
= y + 2a. y . + a. (y .. + Y.) (2.20)
00 J. OJ. J. J.J. J.
-15-
where
x H udA (2.21)
on
that dY/da.
~
= O. Hence,
Yoi
a. - - - - - - (2.22)
~
y .. + Y.
~~ ~
Y = Yoo - (2.23)
y ..
~~
+ Y.
~
electric field of the form of the dominant mode, and is the same as
(2.17). It is also evident from equation (2.23) that when higher order
modes are included a mutual coupling exists between the modes. Further-
more, the terms in equation (2.23), Y ' Y ' Y , are of the form of
oo oi ii
(2.17), which suggests that once a method is known for finding the
The results will apply to any aperture shape and to most media of
approach.
on and reflected from the half-space z > 0, in such a way that on the
area, and will vanish everywhere else. Looking at the problem in this
way, the evaluation of the aperture admittance and the radiation pattern
plane wave will enable the computation of the aperture admittance and
E e ikn·r
(3.1)
where k = ~,
o 0
o
n W,
E is the complex wave amplitude, n
to denote the incident fields and the unit vector in the propagation
(n.)
-).
and the normal to the boundary is called the plane of incidence.
It is well known that any plane electromagnetic wave can be resolved into
the plane of incidence, and another for which the electric vector lies
n·
_I
~---4-+I---+-----t--~Y
ikn. 'r
E. (r) = (E ~\jJ + E e ) e -1. -
-1. - 1. 11)(
(3.2)
ikn. 'r
!!i (.E.) neE e
111
E e) e
1.)(
-1. -
where ~i' ~ and ~ are unit vectors in the radial, polar and
n . • r = ux + vy + wz (3.3)
---J..
w cos X
222
The direction cosines obviously satisfy the relation u + v + w = 1.
ik(ux+vy+wz)
E. (r) (E.l. eljJ + Ell cos X e E sin X e ) e
-1 - -p 1\ -z
0.5)
ik(ux+vy+wz)
-1
H. (r) n(EII~ E.l. cos X e
-p
+ E.l. sin X e ) e
-z
0.6)
z > O. The fields of the reflected wave have also the form of (3.1),
but now the subscript r will be used to denote them as well as the
follows from the laws of reflection at a plane boundary that this unit
n = sin X e - cos X e
-r -p -z
0.7)
u e
-x
+ v --y
e w e
-z
n
-r
. r = ux + vy - wz 0.8)
-21-
and H (r)
-r -
can be written from (3.1) and (3.7) as
) ik(ux+vy-wz)
E (r) (E f !:'ljJ + E\lf\l cos X e + E f sin X e e
-r - .L .L -p II II -z
(3.9)
) ik(ux+vy-wz)
H (r)
-r -
e + E.Lf.L sin X e e
n(-E f eljJ + E f cos X -p -z
1111- .L .L
angles X and ljJ). Associating with the waves propagating in the cone
(3.l0a)
00
(3.l0b)
2 ikwz + r e -ikwZ)e~eik(UX + VY)k 2dudv
+ E'JI
.L
- w (e
.L -z
2 2
where w =-Vl - u - v (3.11)
that complex angles for the propagation direction are also included.
z < O.
.J;,(x,y) t{
00
(3.12)
00
E (u,v)
-0
[~'IT]2 JJ ~(x,y) e-ik(ux + VY)dxdy
-00
(3.13)
EJ.(u,v) (1 + fJ.(U'V»~ljJ + Ell (u,v)w(l+ fll(u,v»~
Em/! (u, v)
(3.l4a)
1 + r (u,v)
J.
E (u,v).e E (u,v)
and
A
Ell (u, v)
_-o-=- ::P.L-_ = _-=oJ;.p~ _ (3.l4b)
w(l + f (u,v» w(l + f (u,v)
ll ll
z = 0, the fields in the region z < ° are now given by Eqs. (3.10),
direction.
l{ ~I
00
00
B. Aperture Admittance
if i;<u,v) x
00
y
[~:J H (u,v) • e k dudv
-0 -z
2
(3.18)
_00
~*
E x H
-0 -0
1 - f
.L
(3.19)
1 + f
.L
by (22)
Y. (u,v) (3.20a)
1. 0 .L
1 - fll(u,v)
and Y. (u, v) = (n/w) (3.20b)
1.0 I!
1 + fll(u,v)
aperture antenna with the dominant mode aperture electric field approxi-
2
Y Y.
l.n.L
(u,v) +
~
IE op (u,v)1 Y.
1.n il
(u,v) dUdV) 0.210)
or equivalently,
-26-
y Y. (p) + IEOp(p,lj!) I 2 Y.
A
(p) pdpdlj!} (3.2lb)
lnJ. lnll
where it has been made clear that the input admittances depend only on
relations
E
op
(u,v) = + 'Ju u2 + v 2 Eox (u,v) + ¥ u ~ + v 2 Eoy (u,v) (3.22b)
The expressions (3.21) are valid for antennas with any aperture
preted as the "sum", with proper amplitudes, of all the input admittances
of the plane waves that make up the assumed aperture field. In essence,
the problem is now reduced to calculating the plane wave input admittances
Y.l.n (3.23)
direction, by (22)
(3.25)
2 2 2 2
and - u - v - u - v (3.26)
Making use of equations (3.24) and (3.25), equation (3.23) can be re-
tan kdw
l
Y. (3.27a)
l.n .L tan kdw
l
-Z8-
(Fig. 3.2a). In this case the field quantities are independent of the
-29-
f{IE
00
y = 2nk
V 2 oy
(u)[2 y. (u) +
1.n J.
IE ox (u)l 2
A y.
1.n \I
J
(u) du (3.29)
o _00
V
0
E (x) V e (x) =~e for Ixl:::;; a/2
-0 o -0 -v;.-x
(3.30)
0 for lxl>a/2
00
E (u) = 1-
-0 2n fE (x
-0
)e-
ikux
dx
_00
where a = ka
00
2
sin (au/2)
y
=%J 0
(au/2)
2
y.
lnll
(u) du (3.32)
E (x,y)
-0
V
o
e (x,y)
-0
V~ cos p- ~x' for Ixl s;; a/2, Iyl s;; b/2
(3.33)
=a otherwise
E (u,v)
-0
~ff .J;,(x,y) e-ik(ux + vy) dx dy
_00
V~
o as
8"
= 27Tk
sin (au/2) cos (Sv/2)
e (3.34 )
(au/2) (7T/2)2 _ (Sv/2)2-x
where a = ka , S = kb (3.35)
I I I I Z I I I I I I l~ I I t I I I 2 Z I Z ~
1
-CD
4-
to
~y
CD
--.
I
a
II Z II Z lIlIllllllllllll Z ~
a b
c d
00 00
2
y = "~~~[S~~~/~i2) CO~(SV/2) ZJZ{y (U,V);Z 2 + Y. (u,v) U Z~dUdV
_00_00
(TI/Z) - (Sv/Z) in.l.. u +v l.nll u Z+v')
(3.36a)
or equivalently,
2TI 00
Y = as
8
l~Sin ( ¥cosl/J) cos (¥ sin llJ) ~t2 .
Z l.n (p) cos l/J+Y.l.nll (p)sin2~pdpdl/J
2
2 )
o (¥ cos l/J) (; ) (¥ sin l/J) .1.
0.36b)
conductor radius b (Fig. 3.Zc), the dominant mode (TEM) electric field
Vo 1
~(p,<j» V e (p) = e , for a < p < b
o -1)
-V Z7r -fn(b/a) p -p
=0 otherwise (3.37)
27r
Jf ~(p,<j»
00
o 0
V
i 0
e 0.38)
= 2TI IiZTI -fn(S/a) -p
-33-
Hence, the admittance of an annular aperture fed by a coaxial
Y = in (3.39)
V e (p, <1»
o -0
< a
(3.40)
=0 for p > a
where
o 0
doP
-34-
(3.41)
y
2
= ---=2--
f""@l(ap)j 2 Y. (p) +
~xJ.lxiia)J
2
i (ap)j
2
2
Y.
j
(p) pdp (3.42)
(xIi. - loP l.nll (xl{a) _ p l.nJ.
media.
Most of the results that have just been obtained had been the
has been shown here that all of these results follow quite simply from
expressions (3.21).
The integrals in (3.32), (3.36), (3.39) and (3.42) have the form
"" ""
fo
fll (p) Y.
1. nil
(p) dp and J o
fJ. (p) (3.43)
where the integrations are carried along the real axis of the complex
singularities are poles of order one due to the zeros of the denomi-
nators of Y. ,
1n
in addition to a branch cut due to w =-y 1- p2. (No
(3.44a)
(3.44b)
2
e:-Y p - 1 - -V e: - p
2
tan (3.45a)
2 2
- p tan - p ) (3.45b)
where o = kd (3.46)
The roots of (3.45) determine the eigenvalues for surface wave propagation
along plane dielectric slabs (23). Since the dielectric slab covering
-36-
wave modes that can exist are the even TM modes, for which D (p) = 0,
II
and the odd TE modes, for which D (p)
J..
= 0. In the calculation of the
Res(p ) (3.47a)
n
nTl"
o = -;::===== n = 0, 1, 2, 3, ... (3.48)
2\1 E - 1
where the even integers refer to the even modes, and the odd integers
unity and large negative values. In this case poles may occur, for real
iD (p) (3.49)
11
in the range 0 > E: > - 1.0363 sufficiently thin plasma slabs support
two more surface waves (24), one of which is a backward wave, i.e. a
wave whose phase and group velocities along the interface are in opposite
Whenever poles of the integrand lie on the real axis the path
excursions in the complex p-plane. The choice of the path about each
is shown in Fig. 3.3a, and for a plasma slab with E: < 0 in Fig. 3.3b.
Except for a backward wave pole, marked b, the integration path passes
1m p 1m p
Re p Re p
p= I p=j€ p= I ( b)
a b
Fig. 3.3 Integration path for a loss less dielectric
is ni times the residue at the poles. (-ni times the residue at the
of power confined within the slab. is real and may be called the surface
00
(3.51)
00
ikw (z-d) ik(ux+vy) 2
~~ E"\ft~~~e
W
where the relations between E.l and E II and the transform of the pre-
and Til are the slab transmission coefficients with respect to the
E (r)
-t -
(3.52)
00
We can iden tHy T.1 /[1 + r .1 ] and T \I / [ w(1 + r \I) ] as the ratio of the
(3.53a)
(3.53b)
T iow
.L 2
= e (3.54a)
1 + r
.1.
and
Til
=-- VE2 e
iOw
2
(3.54b)
w(l + r II) w
z
E (r)
-t -
+ E
Til
e e
op w(l+r ) -X
e
J
ikw(z-d) ik(ux+vy) 2
k dudv
ll
(3.55)
H (r)
-t -
00
~ ~E
AOp _-,--T'_I-:-- T
w(l+r ll ) ~"'I' - EO''I'" l;-.Lr .L ~X e
J ikw(z-d) ik(ux+vy) Z
e k dudv
-00
00
g (.E.) = ff g( u , v) e
ikrf(u,v) 2
k dudv (3.56)
_00
becomes
f (u,v)
u
= f v (u,v) = 0, so that the phase is "stationary", is given by,
u
o
= sin e cos ep
(3.58)
v
o
sin e sin ep
(3.59)
pret. We see that out of all the plane waves propagating in directions
point in the region z > d, only the wave that travels in a radial
field.
radiation field only for 8 = n/2, and that for 8 # n/2 the result
we have
E¢ (!)
(3.60)
T I\(8) -iocos8 e ikr
- 2in E (8,¢) e k cos 8
op w(8)(1+r 11 (8» r
(3.61)
For easy reference, the expressions for T/[1 + r J.] and TII/[w(l + r ll )]
. 2
IfE - S1n 8
= 2
(3.62a)
l+r J.(e) -YE-sin 8
-44-
and
£
(3.62b)
w(8) (l+f 11(8»
2 T 11(8) }
E (8 ~¢) w(8) (l+f 11(8» (3.63)
op
fed apertures radiating through a dielectric slab into free space are
slab were present~ then the far-zone fields and the radiation pattern
would still be given by (3.60)~ (3.61) and (3.63)~ but with T/(1+f)=e i6w
as is readily evident from (3.54). Thus we see that the dielectric slab
has the effect of multiplying the components of the far-zone fields with
slab.
we have
(3.64)
J (asin8)
l
E (8,</» sin </>
op k sin 8
The radiation patterns for each of the two principal planes, the
2 TJ.(8)
F(8,0) k
2 A
(3.65)
and
2
F(8,n/2) = k 1E (8,n/2) 1
2 I Til (8) 2
op 1 + f ll (8)
where E 1/! (8,0) and E (8,n/2) can be obtained from (3.64), and
O op
i6w
where T/(l + r) equals e if the aperture radiates directly into
slab are known. Accordingly, the present chapter will be devoted to the
direction normal to its plane faces, then the reflection and transmission
ate y (Fig. 4.1). We then consider a dielectric slab for 0 < z < d,
Reflected wave
E.L
Incident wave
slab.
-48-
the input admittance and the ratio of the total electric field amplitude
at the right face of the slab to the transverse field at the left face.
plane of incidence will be treated, and then the case of the electric
A. Perpendicular Polarization
fashion:
-49-
For z < 0,
ikwz
E (x, z) (e + r.1. e-ikwz) A(x) ,
y
ikwz r.1. e-ikwz) A(x),
H
x
(x,z) = -n wee - (4.1)
For z > d,
ikwZ(z-d)
E (x, z) T.1. e A(x),
y
ikwZ(z-d)
H (x,z) -T .1.n wZe A(x) , (4. Z)
x
ikwZ(z-d)
H (x, z) = T.1.n p e A(x).
z
and letting w(z) ~(Z) - pZ, we can write for 0 < z < d,
E
y
(x, z) = (P.1.(z) +R.1.(z» A(x),
aH aH
-- - x 'kn
- - =-1 z E:( z) E (4.4)
Clz ax Y
aE
and n -.L
Clz
= -ikHx (4.5)
-50-
we find,
2
d [w(z) (P - R )] - ikw (z) (P + R )
dz ~ ~ ~ ~
o (4.6)
and d
dz (P~ + R~) - ikw(z) (P~ - R) o (4.7)
Eliminating first dR ~ /dz and then dP~/dz between equations (4.6) and
(4.7) it is found that P (z) and R (z) satisfy the following pair of
~ ~
coupled equations:
dP 1
--:!: - i kw ( z) P + -,:,-...,.--:- dw(z) (p _ R ) o (4.8)
dz ~ 2w(z) dz ~ ~
1 dw(z) (p _ R )
2w(z) dz ~ ~
o (4.9)
and
(4.11a)
(4.11b)
df1. 1 dw(z) 2
~ = 2w(z) dz (1 - f L) - i2kw(z)f1. (4.12)
w(d) - w
2
f1.(d) (4.13)
w(d) + w
2
which directly follows from (4.11). Also the relation between f 1.(0),
is from (4.10),
(4.15)
dy.
i l.nL 2
- w (z) (4.16)
k dz
-52-
equation (4.8):
dP
1.
d;- = h1. (z) P1. (z) (4.17)
1 dw(z)
where h1. (z) = ikw(z) - 2w(z) dz (1 - r 1. (z» (4.18)
The solution of (4.17) is trivial once f1. (z) is known from equation
(4.12),
~h1. (z) dz
P1. (0) e (4.19)
2w
(4.20)
[w + w(O)] + [w - w(O)]r (0)
1.
(4.1la) or (4.llb) by
2w(d)
T P (d) (4.21)
1. w(d) + w 1.
2
d
2w(d) ~ hJ..(z)dz
T P (0) e (4.22)
J.. wed) + w J..
2
the amplitude of the wave transmitted through the right face of the
z. Then, by definition,
(4.24)
2w(d)
(4.25)
wed) + w
2
Hence,
d
~ hJ.. (z) dz
e (4.26)
The transmission coefficient of the entire slab is given from (4.10) and
(4.23) by
-54-
2w
(4.27)
face of the slab to that at the point z inside the slab is given
interest to find the equation that this quantity satisfies in the in-
T.L
(4.28)
1 + 1.L (z)
d
(4.29)
dz
1 (4.30)
d
T.L(z) ik f y . .L(z)dz
z l.n
----- = e (4.31)
1+1.L(z)
-55-
T (0) T1.
1.
(4.3Z)
1 + r 1. (0) 1 + r 1.
we finally obtain
d
T1. ik J y. (z)dz
o ~n1.
---= e (4.33)
1 + r 1.
case w(z) = VEl - pZ = wI for 0 < z< d, and equation (4.16) can
y(z)
if
y(d)
-1 v(d) -1 v(z)
tan ~
iW
- tan ~
iW
= k (d-z)w l
l l
Letting
we have
du
dz - ikw [y(d) cos k(d-z)w - iW sin k(d-z)w ]
l I I l
then
d
du
-= In
i 1 y(z) dz u
wI cos kdw - iy(d) sin kdw
o i i
T.L
Hence
I + r .L
B. Parallel Polarization
as follows:
-57-
For z < 0,
ikwz
E (x, z) = w (e
x + r lie-ikwz) A(x) ,
ikwz
E (x, z)
z
-p(e - r lie -ikwz) A(x) (4.34)
ikwz r -ikwz) A(x)
H (x,z)
y
n(e - lie
For z > d t
w ikw (z-d)
2 2
E (x, z) =--Til e A(x) ,
x
-F;
ikw (z-d)
2
E (x, z) = - -.E...... Til e
z A(x) , (4.35)
F;
ikw (z-d)
2
H (x,z)
y n~TII e A(x) .
w(z)
E (x,z) = (P lj (z) + R (z)) A(x) ,
x I1
V;W
E (x,z) P (P (z) - R (z)) A(x) , (4.36)
z II 11
~
H (x, z) = n
y ~ (P11 (z) - ~I (z)) A(x) .
-58-
equations
n (aE
az
x _ aE z)
ax
_ ik Hy (4.37)
aH
and -..:L
az
= ikn dz) Ex (4.38)
we find,
(4.39)
and (4.40)
from which i t follows that PII (z) and 11/ (z) satisfy the following
dR II 1 dw(z) 1 ddz)
d;- + ikw(z)R IJ + 2w(z) dz (PIJ + RII ) - 2E (z) dz PI/ o (4.42)
w(O)
w(l + r l\) v;<o) (P II (0) + R11 (0» (4.43a)
-59-
1 - r II (4.43b)
and
(4.44a)
-= (4.45)
dz
(4.46)
r.w(O) .1
+LdO) + wJ r ll (0)
(4.47)
[w(O) J
+[dO) - ":J r ll (0)
-60-
(4.48)
(4.49)
The boundary condition is YI\ (d) = E/W , and the input admittance of
2
the slab is given by Yin (0) = l/w (l-fll /l+fll).
(4.50)
where
1 dw(z) 1 dE(z)
= ikw(z) - 2w(z) dz (1 + f ll (z)) + 2dz) dz f II (z) (4.51)
1 2w (4.52)
v;M [:~~~ + wJ + [:~~~ - wJ f ll (0)
-61-
and
z
~h lI(z) dz
P lI(z) P 11(0) e (4.53)
(4.44a) or (4.44b) by
zw(d)
E (d) (d)
w PII (4.54)
~ + w(d)
E E(d)
2
(4.55)
Zw(d)
Til (d) WeD w
E:(d)
(4.56)
\ft2 ~
E
+ w(d)
E:(d)
2
Hence
d
~hll(z) dz
Til (z) = Til (d) e (4.57)
1 2w
- wJ f ll (0)
(4.58)
r~ (0) Til (0)
+ L:(O)
face of the slab to the transverse field at the point z inside the
that
Til (z)
1 + f lI(z) (4.59)
Then we have
d
dz
(y;w
w(z)
TII(z) J
1 + f ll (z) ) +
ik
2
w (z)
dz) yin/l (z) (
EZi
w(z) 1
Til (z)\
+ f (z); = 0
ll
(4.60)
-v;;w TII(d) fi
-::-----,,....-,- = - -
wed) l+fll(d) (4.61)
d 2
... ,.....-;-: T lI(z) ""\ IE ik J w (z) ( ) d
vdz) -::--_-..,_ = _v_ ~2? e z dz) Yinll z z
w(z) 1 + f lI(z) w (4.62)
2
-63-
(4.64)
with w(z) = wI for 0 < z < d. and boundary condition y(d) = €2/w2.
Writing now y for y inll' we have
.]. --
€l
w
2
J
y(z)
dy
1 y(d)
-1
tan
Letting
we have
then
d Z u(d)
fo
ik -
€l
W
l
y(z)dz
= f
u(O)
du
-= 1
u n
u(d)
u(O)
Hence
only for a few special €(z). In general, numerical methods must be used,
each section of the medium where the permittivity is continuous, and the
even if E(Z) is not, since the former two are the ratio of the tangential
and the latter two are proportional to the reciprocal of the tangential
be one or more points where E(Z) = O. (Then the plasma frequency equals
and, assuming that the first derivative of E(Z) does not vanish at
dE
dz) =
dz Iz=z o (z - Z
0
) a(z - z )
o
(4.65)
-66-
d
E.Y.=
ok 2
1p
2
or
~ _ ikp2 ...".......:d::,::z:.....,- (4.66)
dz a(z-z) y 2 - a (z-z)
o y 0
< z < + ~z
{~
if Z Z
0 0
1 I ok 2
.!.!SL ln
Iz-z 0 I
= (4.67)
y(z) y(z + ~z) a ~z
0
if z - ~z < z < z
0 0
Thus, the solution around the point z can be found from (4.67),
o
provided that the solution at z
o
+ ~z is previously calculated.
In particular,
obtained in an even more direct and simple manner by making use of the
fact that
-67-
The main concern of the present report has been the presentation of
deemed useful that the method should be illustrated with eXfuilples which
waveguide and radiating into an inhomogeneous plasma slab has been cal-
2
where w , the square of the peak plasma frequency, is proportional
p
to the peak electron density in the plasma slab, and the electron density
normalized to its peak value is given by the function fez), while the
frequency v.
fez)
b) a linear profile,
fez) z/d ,
l
2 2
fez) z /d l '
while fez) = 1 for dl < z < d for all three cases (Fig. 5.1). The
homogeneous plasma slab for which fez) = 1 for 0 < z < d, has also
been considered.
f(z)
1~--L.--~_r-----------~
o ---~~z
A homogeneous profi Ie
C linear profile
under consideration.
mode. The aperture radius a has been chosen such that ka = a = 3n/4,
which is just below the cutoff of the next propagating mode (TM )'
Ol
The slab thickness has always been chosen to be kd = ° = 2n.
The aperture admittance, Y, is normalized to the characteristic
We have computed the real as well as the negative of the imaginary part
of (Y/Y)
o
= g - ib, called the normalized conductance, g, and the
from each of the three inhomogeneity profiles and from the homogeneous
slab, when the plasma is lossy with v/w = 0.4. The thickness of the
the homogeneous slab. The least change occurs in the concave parabolic
-72-
1.2---~--r-----,.--r--r--""----,-----r--~-...,
0.8
9\ _._._.-:::-:::..
~lI:I2t~-d~~ . ----
~"o~:~ :E'::'::;.::.:.::':':".:.:."--
~~ ..~..:.:..::.:':-
~--
-0.8
-1.2
(l =31T/4 v/w =0.4
-1.6 8=21T 8 1 =1/20 8
-2.0 homogeneous
convex parabolic
-2.4 linear
----- concave parabolic ......
.....
-2.8 ........
".
".
-3.2L--..L...-.....L.----L---.L--'--~2--4.L.---..Jio......1,6--..a..-:...-~10
0.1 0.3 0.5 0.7 0.9
(wp/w) 2
Fig. 5.2 Admittance of a circular aperture antenna
coated with an inhomogeneous plasma slab
(v/w = 0.4, 01 = 1/20 0)
-73-
1.2 __-~-___r--_.__-~-:___r_-__r--~-__r_-~
0.8
. .".. 1""-
--- . ..
-.--- .
.-
•.---
... ....-........
.......L'.Z!::t~ ::-::
-----
0.4
~
I
I
~
~
-- "....--- .... ,-
-0.4
-0.8
-1.2
Q =37T/4 lI/W =0.4
-1.6 8=27T 8 1=1/ 108
-2.0 homogeneous
convex parabolic
-2.4 linear
----- concave parabolic
-2.8
- 3.2 ""'--_...........
_ - - - I ._ _~_ ___a.._L....__ ____L. '"___~_ _
case where the llair gap" is the least and the greatest in the convex
parabolic case where the "air gap" is the largest. The change in the
2
conductance is appreciable only for (w /w) > 0.7 and it seems to
p
be quite insensitive tofue inhomogeneity profile. Comparing Fig. 5-2
with Fig. 5-3 it is seen that increasing the boundary layer thickness
Figures 5-4 and 5-5 apply for an almost lossless plasma with
The general behaviour of the curves is similar to those for v/w = 0.4.
is due to the fact that at the point where the permittivity vanishes
related by
1 - r
(Y/Y ) = (5.2a)
0 1 + r
1 - (Y/Y )
0
and r (5.2b)
1 + (Y/Y )
0
1.2 ~--r----.,.--,..---.----;----,.---..,.-----,---,---..,
0.8
0.4
g\
~~:E..!E:~·~-::~~0·7:.·~.:.::~·.:.:::·.
----
o - ....::.-.:.:.,...::.:.'7::::- -- _
.--:~ -,
.-._........... I
-0.4
"
. '.'
'.'
':".' I
':~~~
".J\
.\\
~.~\
-0.8 \~\
I \~\
I i\\
I .\\
-1.2 I ~~.~
a = 37T/4 v/w = 0.025 I \ ..~
I \'~
-1.6 8=27T 8, =V20 8 I \.~\
,....'. ,,,
....
-2.0 homogeneous
".
\. .....
\ ....
,,
\ ...... ',
-2.4
convex parabolic
linear
",. '.
""
""
concave parabolic
'.
.
'\
".
'.
.... ,"
"
".
",
-2.8 " ".
" .
" '
9
,e:
\
:::::::::=-..::.:::::-::-:.--------
;.- ::::. :.::.:.~=:::: . .:.::
w· .::.:. :
O~~~;.=.:+==:.:..~ ~
.. ~,~I~~====t======r====F=:::::::J
_ ,I
' 'I
-0.4 '<>..},
"~\'
r,~~
.~\
-O.S ./ '\\
/ \.. .,
b
I \\>,
-1.2 I \... . ". ' ,
a =37T/4 7//W = 0.025 I '\.'. '. ,
8=27T 8,=1/108 I \.\. . . . ..... " . . . . . . . .
-1.6
I ""
\ ....... . . . .
I . ". "-
I \.\. ......... .......-
-2.0 homogeneous
I " .
convex parabolic I ". . .
I , ......
-2.4 linear " ".
I ". ".
concave parabolic I '.
-2.S I '.,
I ".
I ....
-3.20L...I--0......3--0.....l......
5--0......7--0.....L.9--.JL....-2"----......
4----l1o.--'6'----.......
S ----...10
2
(Wp/W)
Fig. 5.5 Admittance of a circular aperture antenna
coated with an inhomogeneous plasma slab
(v/w = 0.025, 0 = 1/10 0)
1
-77-
(0)
o~':::::::..::J---+--+---+--+----f--+---+--+-----1
1.0
I ---- _-----
I ,/ -- --
I
",-
_"""/I (b)
-'"
IfI o - - -- -- -- I
1.0 ----------------
(c)
0~c:::2::+_-_+_-____+_-___f.~f__~-_+_-_+_-_+-____I
LO __ homogeneous
--- concave parabolic -- ---------
(d)
O --a._ _-"-_ _.....-_---I~'""""____L.
_ __ " __ _~_ __...& _
(v/w) and 01' Only the results for the homogeneous slab and the
For the two other profiles, the results fall between the ones shown.
is discussed. Only the principal planes (the xz-plane, and the yz-plane)
F(0,1T/2)
10 loglO F(8,1T/2)
antenna radiating into free space. The yz-plane radiation pattern shows
----,0
-...,...-10
~ ;JV
-20
(b)
Fig. 5.7 Radiation pattern of a circular aperture
is found that even a thin dielectric layer over the aperture drastically
in Fig. 5.9 the plasma is lossy with v/w = 0.4. In both cases we have
o
chosen (w /W)L = 1/2. The yz-plane radiation changes slightly according
p
to the inhomogeneity, while the xz-plane radiation is insensitive of the
shape of the electron density profile. When the antenna radiates into
This fact can be explained simply by remembering Snell's law. Since the
plasma has a real positive permittivity smaller than that of free space,
the free space region. For (w /w)2 = 1/2, this angle is 45°. When
P
the plasma becomes lossy this fact isno longer true, the peaks disappear
case the shape of the radiation patterns would be smooth with no peaks,
the results, should be mentioned here. The time for obtaining numerical
homogeneous
convex parabolic -10
linear
concave parabolic
90°
- -..... 0---
~
--+--10
---20
30°
90°
Fig. 5.8 Radiation pattern of a circular aperture antenna coat-
ed with a lossless inhomogeneous plasma slab (v/w = 0)
(w /w)2 = 1/2 (a) in yz-plane, (b) in xz-plane.
p
-82-
a ---.:;;=:__
---=-;;;;:::::::;~-,
homogeneous
concave parabolic
~-+-- 10
--'-- 20
-30
__~=---, a -..;;;:;::.__
~--t--IO
(b) 90°
Fig. Radiation pattern of a circular aperture antenna coat-
ed with a lossy plasma slab (v/w = 0.4) (w /w)2 = 1/2
(a) in yz-plane, (b) in xz-plane. P
-83-
linear first order equations which require a far less time for solution
Second, the method could be used to apply to media other than the ones
(4) Lewin, L., Advanced Theory of Waveguides, Iliffe and Sons, Ltd.,
(5) Galejs, J., "Slot Antenna Impedance for Plasma Layers," IEEE
(1965).
581-593 (1962).
1941, pp 361-364.
(22) Brekhovskikh, L., Waves in Layered Media, Academic Press, New York
(23) Collin, R. E., Field Theory of Guided Waves, McGraw-Hill, New York,
1960, pp 470-477.
(27) Tamir, T. and Oliner, A. A., "Guided Complex \vaves, Proc. lEE
(29) Bellman, R. and Kalaba, R., "On the Principle of Invariant Im-
(30) Bellman, R. and Kalaba, R., "Invariant Imbedding and Wave Pro-
(31) Brekhovskikh, L., Waves in Layered Media, Academic Press, New York