Progress In Electromagnetics Research, PIER 86, 87–110, 2008

RADIATION OF A HORIZONTAL DIPOLE IN THE

PRESENCE OF A THREE-LAYERED REGION AND
MICROSTRIP ANTENNA

L. Liu, K. Li, and Y. H. Xu

The Electromagnetics Academy
Zhejiang University
Hangzhou, Zhejiang 310058, China

Abstract—In this paper, we study in detail both the trapped surface

wave and lateral wave excited by a dipole antenna parallel to the
plane boundaries of a three-layered region in spherical coordinate.
An approximate formula is obtained for the solution  of the electric-
type pole equation when a condition is satisfied by k12 − k02 l ≤ 0.6.
Similarly, an approximate formula is obtained for the solution of
the
 magnetic-type pole equation when a condition is satisfied by
k12 − k02 l − π/2 ≤ 1. Furthermore, because of its useful applications
in microstrip antenna, the radiation patterns of a patch antenna with
specific current distributions are treated specifically. Analysis and
computations are carried out in several typical cases. It is seen that,
for the component E0Θ , the total field is determined primarily by the
trapped surface wave of electric type, and, for the component E0Φ ,
the total field is determined primarily by the trapped surface wave of
magnetic type.

1. INTRODUCTION

The problem of the electromagnetic field of a horizontal or vertical

dipole in a multi-layered region has been visited widely by many
investigators [1–28], especially including Sommerfeld, Wait, and King.
In the past decade, several investigators have revisited the old problem
and concluded that the trapped surface wave, which is determined by
the sums of residues of the poles, can be excited efficiently by a dipole
source in the presence of a three-layered region [29–36].
In a recent paper [37], the radiation from a vertical electric dipole
in the presence of a three or four-layered region was investigated. It
88 Liu, Li, and Xu

is found that the far-field patterns on the air-dielectric boundary are

determined primarily the trapped surface, which is not considered in
chapter 8 of [13]. The new development in [37], naturally, rekindled
the interest in the further study on the radiation of a horizontal electric
dipole in the presence of a three-layered region. It is found that it is
more important to study the radiation from a horizontal dipole in the
presence of a three-layered region because of its useful applications in
microstrip antenna.
In what follows, we will attempt to treat in detail both the
radiation of a horizontal dipole in the presence of a three-layered region
and the radiation patterns of a patch antenna with specific current
distributions.
z

(x, y, z)
ρ

Horizontal electric r1
dipole at (0,0,d)

y
Region0 ( air, k 0 )

x
l Region 1 ( dielectric, k 1 )

Region 2 (perfect conductor, k 2 → ∞)

Figure 1. Unit horizontal electric dipole at (0, 0, d) over the air-

dielectric boundary z = 0 in the presence of a three-layered region.

2. RADIATION OF A HORIZONTAL DIPOLE IN THE

PRESENCE OF A THREE-LAYERED REGION

The relevant geometry and Cartesian coordinate system are shown in

Fig. 1, where a horizontal electric dipole in the x̂ direction is located at
(0, 0, d). Region 0 (z ≥ 0) is the space above the dielectric layer with
the air characterized by the permeability µ0 and uniform permittivity
0 , Region 1 (−l ≤ z ≤ 0) is the dielectric layer characterized by the
permeability µ0 , relative permittivity 1r , and Region 2 (z ≤ 0) is the
rest space occupied by a perfect conductor. Then, the wave numbers
Progress In Electromagnetics Research, PIER 86, 2008 89

in the three layers are

√
k0 = ω µ0 ε0
√
k1 = ω µ0 ε0 ε1r
k2 → ∞. (1)
With the time dependence e−iωt , the complete electromagnetic field
generated by a horizontal dipole in the three-layered region in
cylindrical coordinate system can be obtained in [32]. We write
ωµ0 Idl
E0ρ (ρ, φ, z) = − cos φ
4πk02
   2  
2k0 2i z−d ik02 3k0 3i
· − 2 + 3+ − 2 − 3 eik0 r1
r1 r1 r1 r1 r1 r1
    
2k0 2i z+d 2 ik02 3k0 3i
+ 2 + 3+ · − 2 − 3 eik0 r2 + πk02
r2 r2 r2 r2 r2 r2

γ ∗ γ ∗ tan γ ∗ l ∗  
1E iγ0E (z+d) ∗ (1) ∗ (1) ∗
· 0E 1E
e λ H (λ ρ) − H (λ ρ)
j
q  (λ∗jE ) jE 0 jE 2 jE

tan γ ∗ l ∗ (z+d) ∗
 
(1) ∗ (1) ∗
+πk02  ∗
1B
·eiγ0B
λ jB H 0 (λ jB ρ) + H 2 (λ jB ρ)
j
p (λjB )
   
2 π z+d
+2e−i 4 k03 A
π π
· eik0 r2 · −ei 4 + iA
πk0 ρ 2k0 ρ ρ
  2 
π k0 ρ z+d
+ A2 · exp −i − iA
2 2 ρ
 
 2
k0 ρ z+d 1
·erfc  −i − iA  − 2ik03 · eik0 r2
2 ρ πk0 ρ
     2 
−i π π π k0 ρ z+d
· · − √ ei 4 T exp −i + iT
k0 ρ k0 ρ 2 2 ρ
 
 2 
ik0 ρ z+d
·erfc  − + iT  (2)
2 ρ 

ωµ0 Idl
E0φ (ρ, φ, z) = − sin φ
4πk02
    
ik02 k0 i ik02 k0 i
· − − 2− 3 eik0 r1
+ − 3− 3 eik0 r2 + πk02
r1 r1 r1 r2 r2 r2
90 Liu, Li, and Xu

γ ∗ γ ∗ tan γ ∗ l ∗  
1E iγ0E (z+d) ∗ (1) ∗ (1) ∗
· 0E 1E
e λ H (λ ρ) + H (λ ρ)
j
q  (λ∗ ) jE
jE 0 jE 2 jE

tan γ ∗ l ∗ (z+d) ∗
 
(1) ∗ (1) ∗
+πk02  ∗
1B
·eiγ0B
λ jB H 0 (λ jB ρ) − H 2 (λ jB ρ)
j
p (λjB )
    
2 ik0 r2 −i π z+d
+2e−i 4 k03 A
π
i π4
e −e + iA
πk0 ρ k0 ρ 2k0 ρ ρ
  2 
π k0 ρ z+d
+ A2 exp −i − iA
2 2 ρ
 
 2
k0 ρ z+d 1
·erfc  −i + iA  − 2ik03 · eik0 r2
2 ρ πk0 ρ
   2 
π π π k0 ρ z+d
· − √ · ei 4 · T exp −i + iT
k0 ρ 2 2 ρ
 
 2 
k0 ρ z+d
·erfc  −i + iT  (3)
2 ρ 

iωµ0 Idl
E0z (ρ, φ, z) = cos φ
4πk02
     2 
ρ z−d k 3ik0 3
· − · 0
+ 2 − 3 eik0 r1
r1 r1 r1 r1 r1
    2 
ρ z+d 3ik0 k 3
+ · + 2 − 3 eik0 r2 0
r2 r2r2 r2 r2

γ ∗ tan γ ∗ l ∗  2
∗
H1 (λ∗jE ρ)
(1)
+2πk02 1E
 ∗
1E iγ0E (z+d)
e λ jE
j
q (λjE )

1
+2ik03 A · · eik0 r2
πk0 ρ
   2 
π π π k0 ρ z+d
· + √ · ei 4 · A exp −i − iA
k0 ρ 2 2 ρ
 
 2 
k0 ρ z+d
erfc  −i − iA  (4)
2 ρ 

   
µ0 Idl z−d ik0 1
B0ρ (ρ, φ, z) = − sin φ − − 2 eik0 r1
4π r1 r1 r1
Progress In Electromagnetics Research, PIER 86, 2008 91
  
z+d ik0 1
+ − 2 eik0 r2 + πk02
r2 r2 r2

γ ∗ tan γ ∗ l ∗  
1E iγ0E (z+d) ∗ (1) ∗ (1) ∗
· 1E
e λ H (λ ρ) + H (λ ρ)
j
q  (λ∗jE ) jE 0 jE 2 jE

γ ∗ tan γ ∗ l ∗ (z+d)
 
1B ∗ (1) ∗ (1) ∗
+π 0B
λ eiγ0B
H (λ ρ) − H (λ ρ)
j
p (λ∗ ) jB
jB
0 jB 2 jB

 
1 ik0 r2 −i
−2k02 A e ·
πk0 ρ k0 ρ
   2 
π π π k0 ρ z+d
· + √ ei 4 A · exp −i − iA
k0 ρ 2 2 ρ
 
 2 
k0 ρ z+d
·erfc  −i − iA 
2 ρ
 √   
2 i π4 2 ik0 r2 2 iπ z + d π
+2k0 e e − e 4 − iT ·
πk0 ρ 2 2 k0 ρ
  2 
π k0 ρ z+d
+ T 2 exp −i + iT
2 2 ρ
 
 2 
k0 ρ z+d
· erfc  −i + iT  (5)
2 ρ 

   
µ0 Idl z−d ik0 1
B0φ (ρ, φ, z) = − cos φ · − − 2 eik0 r1
4π r1 r1 r1
  
z+d ik0 1
+ − 2 eik0 r2 + πk02
r2 r2 r2

γ ∗ tan γ ∗ l ∗  
1E iγ0E (z+d) ∗ (1) ∗ (1) ∗
· 1E
e λ H (λ ρ) − H (λ ρ)
j
q  (λ∗jE ) jE 0 jE 2 jE

γ ∗ tan γ ∗ l ∗  
0B 1B iγ0B (z+d) ∗ (1) ∗ (1) ∗
+π  ∗
e λ jB H 0 (λ jB ρ) + H 2 (λ jB ρ)
j
P (λjB )

1 ik0 r2 π π π
−2k02 A · e · + √ ei 4 A
πk0 ρ k0 ρ 2
  2 
k0 ρ z+d
· exp −i − iA
2 ρ
92 Liu, Li, and Xu
 
 2 
k0 ρ z + d π 2 ik0 r2
·erfc  −i − iA  + 2k02 ei 4 e
2 ρ πk0 ρ
  √  
−i 2 iπ z + d π
· − e 4 − iT
k0 ρ 2 2 k0 ρ
  2 
π k0 ρ z+d
+ T 2 exp −i + iT
2 2 ρ
 
 2 
k0 ρ z+d
·erfc  −i + iT  (6)
2 ρ 

iµ0 Idl
B0z (ρ, φ, z) = sin φ
   4π    
ρ k0 i ρ k0 i
· − + 2 eik0 r1 + + 2 eik0 r2
r1 r1 r1 r2 r2 r2


λ∗jB 2 tan γ1B
∗ l
∗ (z+d) 1
H1 (λ∗jB ρ)
(1)
+2π e iγ0B
− 2k02
j
p (λ∗jB ) πk0 ρ
   2 
π π π k0 ρ z+d
· − √ ei 4 · T · exp −i + iT
k0 ρ 2 2 ρ
  
 2 
k0 ρ z+d
·erfc  −i + iT  eik0 r2 (7)
2 ρ 

where
 
r1 = ρ2 + (z − d)2 , r2 = ρ2 + (z + d)2 (8)
k12 λ∗jE k02 λ∗jE  
q  (λ∗jE ) = − ∗ + i ∗ tan γ ∗
1E l + γ ∗
1E l sec 2 ∗
γ1E l (9)
γ0E γ1E
∗
 
λ tan γ ∗ l γ ∗ l
p (λ∗jB ) = − ∗ + iλ∗jB 2 ∗
jB 1B
∗ + 0B ∗ sec γ1B l (10)
γ1B γ0B γ1B
 
k0  2
A = k1 − k0 tan
2 k1 − k0 l
2 2 (11)
k1

k12 − k02
T =  (12)
k0 tan( k12 − k02 l)
  √
2e−i 4 F (p∗ )
π
erfc −ip∗ = (13)
Progress In Electromagnetics Research, PIER 86, 2008 93
 2
k0 ρ z+d
p∗1 = + iT (14)
2 ρ
2  
k0 ρ z + d
= p∗2 − iA . (15)
2 ρ
It is noted that λ∗jE and λ∗jB are the j-th poles of the integrals of
electric-type(TM) wave and magnetic-type (TE) wave, respectively.
Correspondingly,

∗ 2 − (λ∗ )2 ;
γmE = km jE m = 0, 1 (16)

∗ 2 − (λ∗ )2 ;
γmB = km jB m = 0, 1. (17)

Here the subscript m refers Regions 0 and 1 and j refers to a number

of poles.
Subjecting a condition z+d  ρ, it is known that r1 ∼ r0 −d cos Θ,
r2 ∼ r0 + d cos Θ in phases and r1 ∼ r2 ∼ r0 in amplitudes. Taking
into account the relations between cylindrical coordinate and spherical
coordinate, and ρ = r0 sin Θ, z = r0 cos Θ, we have
E0r (r0 , Θ, Φ) = E0ρ (ρ, φ, z) sin Θ + E0z (ρ, φ, z) cos Θ (18)
E0Θ (r0 , Θ, Φ) = E0ρ (ρ, φ, z) cos Θ + E0z (ρ, φ, z) cos Θ (19)
E0Φ (r0 , Θ, Φ) = E0φ (ρ, φ, z) (20)
B0r (r0 , Θ, Φ) = B0ρ (ρ, φ, z) sin Θ + B0z (ρ, φ, z) cos Θ (21)
B0Θ (r0 , Θ, Φ) = B0ρ (ρ, φ, z) cos Θ − B0z (ρ, φ, z) sin Θ (22)
B0Φ (r0 , Θ, Φ) = B0φ (ρ, φ, z). (23)
When the far-field conditions |p∗1 | ≥ 4 and |p∗2 | ≥ 4 are satisfied, the
Fresnel-integral terms are written in the forms

π ∗ ik0 ρ iT
k02 · iT · e−ip1 · F (p∗1 ) = ·
k0 ρ ρ2 iT + (z + d)/ρ
 3
iT 1
+ 2 (24)
ρ iT + (z + d)/ρ

π ∗ ik0 ρ −iA
k02 · (−iA) · e−ip2 · F (p∗2 ) = ·
k0 ρ ρ2 −iA + (z + d)/ρ
 3
−iA 1
+ 2 (25)
ρ −iA + (z + d)/ρ
In practical applications the antenna is usually placed on the
surface of the earth, either as a based-insulated dipole or a monopole
94 Liu, Li, and Xu

base-driven against a buried ground system. In this case d/r0 ∼ 0,

k0 d ∼ 0, substitutions (2)–(7) into (18)–(23) lead to the following
formulas.

ωµ0 Idl cos Φ ik0 r0 k0 + ik02 r0 A2 sin Θ
E0r = − e · −
2πk02 r02 sin Θ
iT k0 r0 (cos Θ + iT sin Θ)2 + T sin Θ
+
r03 (cos Θ + iT sin Θ)3
 2 2 
k0 A r0 (−iA sin Θ + cos Θ)2 (i cos Θ + A sin Θ) 



+A2 k0 sin Θ cos Θ − ik0 A3 sin2 Θ
+
r02 (−iA sin Θ + cos Θ)3 

 

ωµ0 Idl cos Φ 
γ0E
∗ γ ∗ sin Θ tan(γ ∗ l)
∗
− · 1E
 (λ∗ )
1E
eiγ0E r0 cos Θ
4  q jE
j
 
·λ∗jE · H0 (λ∗jE r0 sin Θ) − H2 (λ∗jE r0 sin Θ)
(1) (1)

2i cos Θγ ∗ tan(γ ∗ l) ∗
eiγ0E r0 cos Θ λ∗jE 2 H1 (λ∗jE r0 sin Θ)
(1)
− 1E
 ∗
1E

j
q (λjE )

tan(γ1B∗ l) sin Θ
∗
+  ∗ eiγ0B r0 cos Θ
j
p (λjB )
"
 
·λ∗jB H0 (λ∗jB r0 sin Θ) H2 (λ∗jB r0 sin Θ)
(1) (1)
+ (26)

ωµ0 Idl cos Φ ik0 r0

E0Θ = − e
2πk02

k02 Ar0 + k0 cos Θ + ik02 A2 r0 sin Θ cos Θ
· −
r02 sin2 Θ
 
k02 r0 A2 (cos Θ − iA sin Θ)2 (A cos Θ − i sin Θ)
−k0 A2 sin2 Θ − iA3 k0 sin Θ cos Θ
+
r02 (cos Θ − iA sin Θ)3
"
iT k0 r0 cos Θ(cos Θ + iT sin Θ)2 + T sin Θ cos Θ
+
r03 sin Θ(cos Θ + iT sin Θ)3

ωµ0 Idl cos Φ 
γ0E
∗ γ ∗ tan(γ ∗ l)
∗
− 1E
 ∗
1E
eγ0E r0 cos Θ
4  q (λjE )
j
Progress In Electromagnetics Research, PIER 86, 2008 95
 
·λ∗jE cos Θ H0 (λ∗jE r0 sin Θ) − H2 (λ∗jE r0 sin Θ)
(1) (1)

γ ∗ tan(γ ∗ l) ∗
eγ0E r0 cos Θ λ∗jE 2 H1 (λ∗jE r0 sin Θ)
1E 1E (1)
+2i sin Θ  ∗
j
q (λ ) jE

tan(γ ∗ l) ∗
+ cos Θ 1B
eiγ0B r0 cos Θ
j
p (λ∗jB )
"
 
·λ∗jB H0 (λ∗jB r0 sin Θ) H2 (λ∗jB r0 sin Θ)
(1) (1)
+ (27)


ωµ0 Idl sin Φ ik0 r0 iAk0 cos Θ−k0 A2 sin Θ−ik02 r0 sin2 Θ
E0Φ = − e
2πk02 r02 sin3 Θ
ik0 A3 r0 (−iA sin Θ + cos Θ)2 + A3 sin Θ
−
r03 sin Θ(−iA sin Θ + cos Θ)3
"
−k 2 T r0 (iT sin Θ + cos Θ)2 + iT k0 sin Θ
+ 0
r02 (iT sin Θ + cos Θ)3

ωµ0 Idl sin Φ 
γ0E
∗ γ ∗ tan(γ ∗ l)
∗
− 1E
 ∗
1E
eiγ0E r0 cos Θ
4  q (λjE )
j
 
·λ∗jE H0 (λ∗jE r0 sin Θ) + H2 (λ∗jE r0 sin Θ)
(1) (1)

tan(γ ∗ l) ∗
+ 1B
eγ0B r0 cos Θ · λ∗
j
p (λ∗jB ) jB

"
 
H0 (λ∗jB r0 sin Θ) H2 (λ∗jB r0 sin Θ)
(1) (1)
· − (28)

  
µ0 Idl sin Φ ik0 r0  iA−k0 r0 T sin Θ−i cos Θ+
ir0 sin Θ cos Θ
2
B0r = − e
2π  r02 sin Θ
 
k0 T r0 (iT sin Θ + cos Θ)2 (cos Θ + iT sin Θ)
+T sin Θ(T sin Θ − i cos Θ)
+
r02 (iT sin Θ + cos Θ)3
"
iA2 sin Θ − k0 A2 r0 (−iA sin Θ + cos Θ)2
·
k0 r03 (−iA sin Θ + cos Θ)3

µ0 Idl sin Φ 
k02 γ1E
∗ sin Θ tan(γ ∗ l)
∗
− 1E
eiγ0E r0 cos Θ
4  q  (λ∗jE )
j
96 Liu, Li, and Xu
 
·λ∗jE H0 (λ∗jE r0 sin Θ) + H2 (λ∗jE r0 sin Θ)
(1) (1)

γ ∗ sin Θ tan(γ ∗ l) ∗
+ 0B 1B
eiγ0B r0 cos Θ
j
p (λ∗jB )
 
·λ∗jB H0 (λ∗jB r0 sin Θ) − H2 (λ∗jB r0 sin Θ)
(1) (1)



λ∗jB 2 tan(γ1B
∗ l)
∗ r

·H1 (λ∗jB r0 sin Θ) (29)
(1)
−2i cos Θ iγ0B
e 0 cos Θ

j
p (λ∗jB ) 

µ0 Idl sin Φ ik0 r0

B0Θ = − e

2π  
 iA cos Θ − k0 r0 sin Θ T cos Θ + i sin Θ + ir0 cos2 Θ
2
·
 r02 sin2 Θ
 
k0 T r0 (iT sin Θ + cos Θ)2 (iT cos Θ − sin Θ)
+T sin Θ(T cos Θ + i sin Θ)
+
r02 (iT sin Θ + cos Θ)3
"
iA2 sin Θ cos Θ − k0 A2 r0 cos Θ(−iA sin Θ + cos Θ)2
+
k0 r03 sin Θ(−iA sin Θ + cos Θ)3

µ0 Idl sin Φ 
k02 γ1E
∗ cos Θ tan(γ ∗ l)
∗
−  (λ∗ )
1E
eiγ0E r0 cos Θ
4  q jE
j
 
·λ∗jE H0 (λ∗jE r0 sin Θ) + H2 (λ∗jE r0 sin Θ)
(1) (1)

γ ∗ cos Θ tan(γ ∗ l) ∗
+ 0B 1B
eiγ0B r0 cos Θ
j
p (λ∗jB )
 
·λ∗jB H0 (λ∗jB r0 sin Θ) − H2 (λ∗jB r0 sin Θ)
(1) (1)



2iλ∗jB 2 sinΘtan(γ1B
∗ l)
∗ r cosΘ

H1 (λ∗jB r0 sin Θ) (30)
iγ0B (1)
+ e 0

j
p (λ∗jB ) 


µ0 Idl cos Φ ik0 r0 k0 Ar0 sin Θ + r0 cos Θ
− iT
B0Φ = − e − 2 2
2
2π r0 sin Θ
ik0 A2 r0 (−iA sin Θ + cos Θ)2 + A2 sin Θ
−
r02 (−iA sin Θ + cos Θ)3
Progress In Electromagnetics Research, PIER 86, 2008 97
"
k0 T 2 r0 (iT sin Θ + cos Θ)2 − iT 2 sin Θ ik0 d cos Θ
+ e
k0 r03 sin Θ(iT sin Θ + cos Θ)3

µ0 Idl cos Φ  2
γ1E
∗ tan(γ ∗ l)
∗
− k0  ∗
1E
eiγ0E r0 cos Θ · λ∗jE
4  q (λjE )
j
 
· H0 (λ∗jE r0 sin Θ) − H2 (λ∗jE r0 sin Θ)
(1) (1)

γ ∗ tan(γ ∗ l) ∗
+ 0B 1B
· λ∗jB · eiγ0B r0 cos Θ
j
p (λ∗jB )
"
 
H0 (λ∗jB r0 sin Θ) H2 (λ∗jB r0 sin Θ)
(1) (1)
· + . (31)

3. ANALYTICAL TECHNIQUES FOR DETERMINING

THE POLES λE AND λM

In this section we will analyze the pole equations for both electric-type
(TM) and magnetic-type (TE) waves. The pole equation for electric-
type (TM) wave is expressed as follows [32]:

q(λ) = k12 γ0 − ik02 γ1 tan γ1 l = 0 (32)

It follows that
 
# $
k02 k12 − λ2 sec2 k12 − λ2 l
dλ
=     . (33)
dl √ k21 λ + k02 tan k12 − λ2 l + k02 lλ sec2 k12 − λ2 l
λ −k02
 √ 2 2
When tan k12 − λ2 l =
k12
√λ2 −k02 is substituted into (33), and
k0 2 k1 −λ
considering that k1 is a real number, it is seen that all the terms of
the above equation in the right are

positive because of the poles λjE
are between k0 and k1 . So that k12 − λ2 l is always positive, and it
dλ
means that dl> 0. Then, it is known that λjE increase as l in the
interval nπ ≤ k12 − k02 l ≤ (n + 1)π. When a condition is satisfied by

k12 − k02 l < 0.6, we write
  2
√
−k12 2k0 + 2k14 k0 + 8k05 (k12 − k02 )l2
λE = k0 + . (34)
16k06 l2
98 Liu, Li, and Xu

Similarly, the pole equation for magnetic-type (TE) wave is expressed

in the following form [32].
p(λ) = γ1 − iγ0 tan γ1 l = 0. (35)
Obviously, (35) can be re-written as
   
π
k12 − λ2 tan k12 − k02 l − − λ2 − k02 = 0. (36)
2
Then,
   
k12 − λ2 λ2 − k02 sec2 k12 − λ2 l
dλ
= √ #√ $ #√ 2 $. (37)
dl λl λ2 −k02 sec2 k12 −λ2 l λ tan k −λ2 l
√ λ
+ √ − √ 212
k12 −λ2 k12 −λ2 λ −k0
√ 2 2
k −λ
Substituting tan γ1 l = − √ 12 2 into (37), and considering that k1 is a
λ −k0
dλ
real number, it is easily verified that > 0 in the interval k0 < λ < k1 .
 dl
When a condition is satisfied by k12 − k02 l − π/2  1, we have
  
π π
tan k12 − k02 l − ≈ k12 − k02 l − (38)
2 2
From (37), after algebraic manipulation, we write
#√ √ $2
2k0 − 2k0 − 4Ab Bb
λB = k0 + (39)
4Bb2
where
  
π
Ab = k12 − k02 k12 − k02 l − (40)
2
π k0
Bb = · − 2k0 l. (41)
2 k12 − k02

As shown in Fig. 2, it is seen that the approximate values of λE are

agreement
 with the corresponding accurate values under the condition
k1 − k0 l < 0.6 for electric-type (TM) wave. Similarly, from Fig. 3,
2 2

it is seen that the approximate values of λM are agreement

 with the
corresponding accurate values under the condition k1 − k0 l − π/2 
2 2

1 for magnetic-type (TE) wave.

Progress In Electromagnetics Research, PIER 86, 2008 99
21.6
The accurate value of λ
The approximation of λ

21.5

21.4

21.3
λ

21.2

21.1

21

20.9
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
k 21 − k 20 l

Figure 2. The values of λ vs. the thickness l of the dielectric layer

for the electric-type (TM) wave with f = 1 GHz and 1r = 2.65.

28
The accurate value of λB
The approximation of λ B
27

26

25
λB

24

23

22

21

20
1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2

k 21 − k 20 l

Figure 3. The values of λ vs. the thickness l of the dielectric layer

for the magnetic-type (TE) wave with f = 1 GHz and 1r = 2.65.
100 Liu, Li, and Xu

(x', y', z')

r0
r
z'

y'
2s
J (x'', y'')
x
0 2w
x'

air
2h
dielectric
l
substrate

conducting base

Figure 4. Microstrip patch antenna end-driven by microstrip

transmission line.

4. MICROSTRIP ANTENNA

Microstrip consist of strip transmission lines and antenna on the surface

of a thin dielectric layer coating a highly conducting base, which can
be regarded as a perfect conductor. The basic elements of them are
horizontal electric dipoles. In this section, we will treat a rectangular
patch antenna end-driven by transmission lines, as illustrated in Fig. 4.
The current-density distribution of a patch antenna with a width 2w
and length 2h has the following form.

Ix (0) −h ≤ x ≤ h
Jx (x , y  ) = cos(kL x ); (42)
2w −w ≤ y  ≤ w
√
where kL = k0 1r eff is the wave number that would characterize the
patch as a segment of a microstrip transmission line for which 1r eff
is the relative effective permittivity at the operating frequency, Ix (0)
is the total current traversing the center line x = 0 of the patch. In
the present study we assume that the transverse distribution of the
x̂ -directed current is uniform. Actually, there is a ŷ  -directed current
with large peaks at the edges |y  | = w and a minimum at the center
y  = 0. Since the impedance of the patch is very large when the
driving-point current is small compared to Ix (0), the assumed current
(42) should be a good approximation. The approximate formulas for
the characteristic impedance Zc of the microstrip transmission line and
the relative effective permittivity 1r eff were given in (3.20)–(3.22)
by Hoffman [38]. The relative permittivity 1r eff is approximated as
Progress In Electromagnetics Research, PIER 86, 2008 101

follows:
 −1/2
1r + 1 1r − 1 5l
1r eff = + 1+ . (43)
2 2 w
In the far-field zone, r ∼ r0 = (x2 + y 2 + z 2 ) is adequate in
amplitudes. In the phases, it is necessary to use the more accurate
formula.
 1/2  1/2
    x x y  y 
r∼ r02 − 2x x − 2y y = r0 1−2 2 −2 2 . (44)
r0 r0

Following the similar steps addressed in Sec. 15.12 of [12] or Sec. 9.7
of [13], and use is made of the above approximations, the field factor
for the patch antenna is written in the following form.
Ix (0)
P (Θ, Φ) = J(x )J(y  )
2w
 
2kL sin(kL h) cos(k0 h sin Θ cos Φ)
 −2k0 sin Θ cos Φ cos(kL h) sin(k0 h sin Θ cos Φ) 
= Ix (0) 



kL2 − k02 cos2 Φ sin2 Θ
 
sin(k0 w sin Φ sin Θ)
· . (45)
k0 w sin Φ sin Θ
If h is chosen as kL h = π/2, the field factor (45) reduces to (15.12.14)
in the book [12] or (9.159) in [13]. Multiplying (27), (28), (30), and
(31) by the field factor in (45), the far-field components of the patch
antenna are expressed as follows:
r
[E0Θ (r0 , Θ, Φ)]p = B0Θ (r0 , Θ, Φ) · P (Θ, Φ) (46)
r
[B0Θ (r0 , Θ, Φ)]p = B0Φ (r0 , Θ, Φ) · P (Θ, Φ) (47)
r
[E0Φ (r0 , Θ, Φ)]p = B0Θ (r0 , Θ, Φ) · P (Θ, Φ) (48)
r
[B0Φ (r0 , Θ, Φ)]p = B0Φ (r0 , Θ, Φ) · P (Θ, Φ). (49)

It is noted that these are valid in the far-field zone and not valid in the
intermediate zone.

5. COMPUTATIONS AND DISCUSSIONS

Considering the electric field at the high frequency of interest, f =

1 GHz, and the far-field conditions |p∗1 | ≥ 4 and |p∗2 | ≥ 4, with
102 Liu, Li, and Xu

1r = 2.56 and l = 0.04 m or l = 0.1 m, the radial distance ranges

over r0 = 7 m. As shown in Figs. 5 and 6, the total field, both
the trapped surface waves of electric type (TM) and magnetic type
(TE), and the lateral wave are calculated from (27) for E0Θ r (r , Θ, Φ)
0
r
and (28) for E0Φ (r0 , Θ, Φ), respectively. Graphs of the amplitudes of
the components E0Θ r (r , Θ, Φ) and E r (r , Θ, Φ) for a unit horizontal
0 0Φ 0
dipole vs. the angular degrees of Θ are computed and shown in Figs. 7
and 8, respectively. It is noted that the DRL waves include the direct
wave, ideal reflected wave, and lateral wave. From Figs. 5–8, it is seen
that, for the component E0Θ , the total field is determined primarily by
the trapped surface wave of electric type, and, for the component E0Φ ,
the total field is determined primarily by the trapped surface wave of
magnetic type when both the dipole point and observation point are
located on the air-dielectric bound.
3
10

2
10
Total field
The trapped surface wave
Lateral wave

1
10
| in V/m
0Θ
|E

0
10

-1
10

-2
10
20 30 40 50 60 70 80 90 100
ρ in m

Figure 5. Electric field component E0Θ in V/m with f = 1 GHz,

1r = 2.65, l = 0.1 m, d = z = 0 m, and Φ = 0.

In Figs. 9 and 10, graphs of the amplitudes of the components

r (r , Θ, Φ)] as calculated from (46) and [E r (r , Θ, Φ)] from (47)
[E0Θ 0 p 0Φ 0 p
for the patch antenna vs. the angular degrees of Φ are shown with
f = 1 GHz, 1r = 2, r0 = 10 m, kL h = π/2, w = h, l = 0.1 m, d = 0 m,
I0 = 1 A, and Θ = π/2. It is found that the amplitude of the total wave
is approximately equal to that of the trapped surface wave under the
conditions of d = 0 and Θ = π/2. This also demonstrates that total
Progress In Electromagnetics Research, PIER 86, 2008 103
3
10

2
10

1
10
|E0 | in V/m

Total field
Trapped surface wave of magnetic type
Φ

Trapped surface wave of electric type

0 Lateral wave
10

−1
10

−2
10
20 30 40 50 60 70 80 90 100
ρ in m

Figure 6. Electric field component E0Φ in V/m with f = 1 GHz,

1r = 2.65, l = 0.1 m, d = z = 0 m, and Φ = π/2.

2
10
Total field
Trapped surface wave of electric type
DRL waves

1
10
|E0Θ | in V/m

0
10

−1
10

−2
10
86 87 88 89 89.5 89.55 89.6 89.65 89.7 89.75 89.8 89.85 89.9 89.95 90
Θ in degee

Figure 7. Electric field component E0Θ in V/m vs. Θ with f = 1 GHz,

1r = 2.65, r0 = 80 m, l = 0.1 m, d = 0 m, and Φ = 0.
104 Liu, Li, and Xu
3
10

2
10

Total field
Trapped surface wave of magnetic type
1
10 Trapped surface wave of electric type
|E0 | in V/m

Lateral wave
Φ

0
10

−1
10

−2
10
86 87 88 89 89.7 89.75 89.8 89.85 89.9 89.95 90
Θ in degee

Figure 8. Electric field component E0Φ in V/m vs. Θ with f = 1 GHz,

1r = 2.65, r0 = 20 m, l = 0.1 m, d = 0 m, and Φ = π/2.

20
Total field
DRL waves
18 Trapped surface wave

16

14

12

10 Θ=π/2

2 Φ

0
0 1 2 3 4 5 6 7 8 9 10
E0 Θ

Figure 9. Electric field component [E0Θ r (r , Θ, Φ)] radiated by a

0 p
patch antenna vs. Φ with f = 1 GHz, 1r = 2, r0 = 10 m, kL h = π/2,
w = h, l = 0.1 m, d = 0 m,I0 = 1 A, and Θ = π/2.
Progress In Electromagnetics Research, PIER 86, 2008 105
25
Total field
DRL waves
Trapped surface wave

20
Θ=π/2

15

10

0
0 5 10 15 20 25 30 35 40 45
E0 (Θ, Φ )
Φ

Figure 10. [E0Φr (r , Θ, Φ)] radiated by a patch antenna with f =

0 p
1 GHz, 1r = 2, r0 = 10 m, kL h = π/2, w = h, l = 0.1 m, d = 0 m,
I0 = 1 A, and Θ = π/2.

80

60

40

20

Φ
0

20

40

60

80
30 20 10 0 10 20 30
E 0 Θ (Φ , Θ )

Figure 11. Three-dimensional diagram of the component

r (r , Θ, Φ)] radiated by a patch antenna with f = 1 GHz,  = 2,
[E0Θ 0 p 1r
r0 = 10 m, kL h = π/2, w = h, l = 0.1 m, d = 0 m, and I0 = 1 A.
106 Liu, Li, and Xu

100

50

50

100

50 40 30 20 10 0 10 20 30 40 50
E
0Φ
(Θ, Φ )

r (r , Θ, Φ)] of antenna
Figure 12. Three-dimensional diagram of [E0Φ 0 p
with f = 1 GHz, 1r = 2, r0 = 10 m, kL h = π/2, w = h, l = 0.1 m,
d = 0 m, and I0 = 1 A.

field is determined primarily by the trapped surface wave when both

the dipole point and observation point are located on the air-dielectric
boundary.
In Figs. 11 and 12, three-dimensional diagrams of the two
components [E0Θ r (r , Θ, Φ)] and [E r (r , Θ, Φ)] for a patch antenna
0 p 0Φ 0 p
are shown, respectively. It is noted that the inner curves in Figs. 11 and
12 is for Θ = 86◦ and the outer curves for Θ = 90◦ . For the component
r (r , Θ, Φ)] , the radiation is enlarged greatly in the region near
[E0Θ 0 p
Θ = π/2 and Φ = 0, or π. For the component [E0Φ r (r , Θ, Φ)] , the
0 p
radiation are enlarged greatly in the region near Θ = π/2 and Φ = π/2.
The radiation from a patch antenna can be determined in the same
manner for other types of load, such as resonant lines with impedance
load and matched line with travelling-wave currents.

6. CONCLUSIONS

The far-field radiation of a horizontal dipole in the presence of a three-

layered region have been investigated for its possible applications in
microstrip antenna. In the analysis and computations on the radiation
Progress In Electromagnetics Research, PIER 86, 2008 107

of a horizontal dipole and microstrip antenna, which are addressed in

chapter 15 of [12] and chapter 9 of [13], the trapped surface wave is
not considered. In this paper we consider both the trapped surface
wave and lateral wave in detail. The approximate formulas have
been obtained for determining
 the poles of the electric-type waves
under the condition of k1 − k0 l ≤ 0.6 and that of the magnetic-type
2 2

waves under the condition of k12 − k02 l − π/2 ≤ 1. In addition, the
radiation patterns of a patch antenna with specific current distributions
are analyzed. Computations show that, for the component E0Θ , the
total field is determined primarily by the trapped surface wave of
electric type, and, for the component E0Φ , the total field is determined
primarily by the trapped surface wave of magnetic type.

REFERENCES

1. Sommerfeld, A., “Propagation of waves in wireless telegraphy,”

Ann. Phys., Vol. 28, 665–736, Mar. 1909.
2. Sommerfeld, A., “Propagation of waves in wireless telegraphy,”
Ann. Phys., Vol. 81, 1135–1153, Dec. 1926.
3. Wait, J. R., “Radiation from a vertical electric dipole over a
stratified ground,” IRE Trans. Antennas Propagat., Vol. 1, 9–12,
1953; Vol. 2, 144–146, 1954.
4. Wait, J. R., “Radiation from a vertical electric dipole over a curved
stratified ground ground,” Journal of Res. Nat. Bur. Standards,
Vol. 56, 232–239, 1956.
5. Wait, J. R., “Excitation of surface waves on conducting dielectric
clad and corrugated surfaces,” Journal of Res. Nat. Bur.
Standards, Vol. 59, No. 6, 365–377, 1957.
6. Wait, J. R., “Electromagnetic surface waves,” Advances in Radio
Research, 157–219, Academic Press, New York, 1964.
7. Wait, J. R., Electromagnetic Waves in Stratified Media, 2nd
edition, Pergamon, New York, 1970.
8. Wait, J. R., “Radiation from a vertical electric dipole located over
laterally anisotropic groundplane,” Electronics Letters, Vol. 26,
No. 1, 74–76, 1990.
9. Wait, J. R., “Electromagnetic fields of a vertical electric dipole
over a laterally anisotropic surface,” IEEE Trans. Antennas and
Propagat., Vol. 38, No. 10, 1719–1723, 1990.
10. Mahmoud, S. F., “Remarks on ‘The electromagnetic field of
a vertical electric dipole over the earth or sea’,” IEEE Trans.
Antennas Propagat., Vol. 46, No. 12, 1745–1946, 1999.
108 Liu, Li, and Xu

11. Collin, R. E., “Some observations about the near zone electric field
of a hertzian dipole above a lossy earth,” IEEE Trans. Antennas
Propagat., Vol. 52, No. 11, 3133–3137, 2004.
12. King, R. W. P., M. Owens, and T. T. Wu, Lateral Electromagnetic
Waves: Theory and Applications to Communications, Geophysical
Exploration, and Remoting Sensing, Springer-Verlag, 1992.
13. King, R. W. P., G. J. Fikioris, and R. B. Mack, Cylindrical
Antennas and Array, University Press, Cambridge, 2002.
14. King, R. W. P. and S. S. Sandler, “The electromagnetic field
of a vertical electric dipole over the earth or sea,” IEEE Trans.
Antennas Propagat., Vol. 42, No. 3, 382–389, 1994.
15. King, R. W. P., “The electromagnetic field of a horizontal electric
dipole in the presence of a three-layered region,” J. Appl. Phys.,
Vol. 69, No. 12, 7987–7995, 1991
16. King, R. W. P., “The electromagnetic field of a horizontal electric
dipole in the presence of a three-layered region: Supplement,” J.
Appl. Phys., Vol. 74, No. 8, 4845–4548, 1993.
17. King, R. W. P. and S. S. Sandler, “The electromagnetic field of a
vertical electric dipole in the presence of a three-layered region,”
Radio Sci., Vol. 29, No. 1, 97–113, 1994.
18. Wait, J. R., “Comment on ‘The electromagnetic field of a vertical
electric dipole in the presence of a three-layered region’ by Ronold,
W. P. King and S. S. Sandler,” Radio Sci., Vol. 33, No. 2, 251–253,
1998.
19. King, R. W. P. and S. S. Sandler, “Reply,” Radio Sci., Vol. 33,
No. 2, 255–256, 1998.
20. Collin, R. E., “Hertzian dipole radiation over a lossy earth or sea:
Some early and late 20th century controversies,” IEEE Antennas
and Propagation Magazine, Vol. 46, No. 2, 64–79, 2004.
21. Koledintseva, M. Y., V. V. Bodrov, I. V. Sourkova, M. M. Sabirov,
and V. I. Sourkov, “Unified spectral technique application for
study of radiator behavior near planar layered composites,”
Progress In Electromagnetics Research, PIER 66, 317–357, 2006.
22. Xu, X.-B. and Y. Huang, “An efficient analysis of vertical dipole
antennas above a lossy half-space,” Progress In Electromagnetics
Research, PIER 74, 353–377, 2007.
23. Abo-Seida, O. M., “Far-field due to a vertical magnetic dipole in
sea,” Journal of Electromagnetic Waves and Applications, Vol. 20,
No. 6, 707–715, 2006.
24. Zheng, Q.-R., Y.-Q. Fu, and N.-C. Yuan, “Characteristics
of planar PBG structures with a cover layer,” Journal of
Progress In Electromagnetics Research, PIER 86, 2008 109

Electromagnetic Waves and Applications, Vol. 20, No. 11, 1439–

1453, 2006.
25. Chen, H. T., J.-X. Luo, and D.-K. Zhang, “An analytic formula of
the current distribution for the VLF horizontal wire antenna above
lossy half-space,” Progress In Electromagnetics Research Letters,
Vol. 1, 149–158, 2008.
26. Sharma, A. and G. Singh, “Design of single Pin shorted three-
dielectric-layered substrates rectangular patch microstrip antenna
for communication systems,” Progress In Electromagnetics
Research Letters, Vol. 2, 157–165, 2008.
27. Attiya, A. M., “Dyadic Green’s function of an elementary point
source above a periodically-defected-grounded dielectric slab,”
Progress In Electromagnetics Research B, Vol. 4, 127–145, 2008.
28. Zhang, L., Y. H. Zhang, W. Hu, W. Yu, and G.-Q. Zhu,
“Automatic incorporation of surface wave poles in discrete
complex image method,” Progress In Electromagnetics Research,
PIER 80, 161–178, 2008.
29. Zhang, H. Q. and W. Y. Pan, “Electromagnetic field of a vertical
electric dipole on a perfect conductor coated with a dielectric
layer,” Radio Science, Vol. 37, No. 4, 13-1–13-7, 2002.
30. Zhang, H.-Q., K. Li, and W.-Y. Pan, “The electromagnetic field
of a vertical dipole on the dielectric-coated imperfect conductor,”
Journal of Electromagnetic Waves and Applications, Vol. 18,
No. 10, 1305–1320, 2004.
31. Zhang, H.-Q., W.-Y. Pan, K. Li, and K.-X. Shen, “Electromag-
netic field for a horizontal electric dipole buried inside a dielectric
layer coated high Lossy half space,” Progress In Electromagnetics
Research, PIER 50, 163–186, 2005.
32. Li, K. and Y. Lu, “Electromagnetic field generated by a
horizontal electric dipole near the surface of a planar perfect
conductor coated with a uniaxial layer,” IEEE Trans. Antennas
and Propagat., Vol. 53, No. 10, 3191–3200, 2005.
33. Mei, J. P. and K. Li, “Electromagnetic field from a horizontal
electric dipole on the surface of a high lossy medium coated with
a uniaxial layer,” Progress In Electromagnetics Research, PIER 73,
71–91, 2007.
34. Xu, Y. H., W. Ren, L. Liu, and K. Li, “Electromagnetic field of a
horizontal electric dipole in the presence of a four-layered region,”
Progress In Electromagnetics Research, PIER 81, 371–391, 2008.
35. Xu, Y. H., K. Li, and L. Liu, “Electromagnetic field of a horizontal
electric dipole in the presence of a four-layered region,” Progress
110 Liu, Li, and Xu

In Electromagnetics Research, PIER 81, 371–391, 2008

36. Xu, Y. H., W. Ren, L. Liu, and K. Li, “Electromagnetic field of a
horizontal electric dipole in the presence of a four-layered region,”
Progress In Electromagnetics Research, PIER 81, 371–391, 2008.
37. Liu, L. and K. Li, “Radiation from a vertical electric dipole in the
presence of a three-layered region,” IEEE Trans. Antennas and
Propagat., Vol. 55, No. 12, 3469–3475, 2007.
38. Hoffman, R. K., Handbook of Microwave Integrated Circuits,
Sec. 3.2, Artech House, Norwood, MA, 1987.
39. Gradshteyn, I. S. and I. M. Ryzhik, Table of Integrals, Series, and
Products, Academic Press, New York, 1980.