LECTURE 22: MICROSTRIP ANTENNAS – PART III

(Circular patch antennas: the cavity model. Radiation field of the circular
patch. Circularly polarized radiation from patches. Arrays and feed
networks.)

1. Circular patch: the cavity model

a y
x

The circular patch cannot be analyzed using the TL method, but can be
accurately described by the cavity method. It is again assumed that only TMz
modes are supported in the cavity. They are fully described by the VP
A = Az zˆ . The Az VP function satisfies the Helmholtz equation,
∇ 2 Az + k 2 Az =
0 (22.1)
which now is solved in cylindrical coordinates:
1 ∂  ∂Az  1 ∂ 2 Az ∂ 2 Az
ρ  + ρ 2 ∂φ 2 + ∂z 2 + k Az =
2
ρ ∂ρ  ∂ρ
0, (22.2)

1 ∂Az ∂ 2 Az 1 ∂ 2 Az ∂ 2 Az
⇒ + + + 2 + k 2 Az =
0. (22.3)
ρ ∂ρ ∂ρ 2 ρ ∂φ
2 2 ∂z
Using the method of separation of variables,
Az = R( ρ ) F (φ ) Z ( z ) , (22.4)
1∂R ∂ 2 R RZ ∂ 2 F ∂2Z
⇒ FZ + FZ 2 + 2 + RF 2 + k 2 RFZ =
0, (22.5)
ρ ∂ρ ∂ρ ρ ∂φ 2 ∂z
Nikolova 2016 1
 1 ∂R 1 ∂ 2 R 1 ∂2 F 1 ∂2Z
⇒ + + 2 + =
−k 2 . (22.6)
ρ R ∂ρ R ∂ρ 2 ρ F ∂φ 2 Z ∂z 2

The 4th term is independent of ρ and φ, and is being separated:

1 ∂2Z
= −k z 2 . (22.7)
Z ∂z 2

Then,
1 ∂R 1 ∂ 2 R 1 ∂2 F
+ + 2 =
−( k 2 − k z 2 ) =
const. (22.8)
R ρ ∂ρ R ∂ρ 2 ρ F ∂φ 2

ρ ∂R ρ 2 ∂ 2 R 1 ∂ 2 F
⇒ + + + (k 2 − k z 2 ) ρ 2 =
0. (22.9)
R ∂ρ R ∂ρ 2 F ∂φ 2

Now, the 3rd term is independent of ρ , and the other terms are independent
of φ. Thus, (22.9) is separated into two equations:
1 ∂2 F
= kφ2 (22.10)
F ∂φ 2

and
ρ ∂R ρ 2 ∂ 2 R
+ + (k 2 − k z 2 ) ρ 2 − kφ 2 =
0. (22.11)
R ∂ρ R ∂ρ 2

We define
k ρ=
2 k 2 − kz 2 . (22.12)
Then (22.11) can be written as [note that (22.11) depends only on ρ]:
∂  ∂R 
ρ  ρ  + (k ρ ρ ) 2 − kφ 2  ⋅ R =
0. (22.13)
∂ρ  ∂ρ 
Thus, equation (22.1) has been separated into three ordinary differential
equations — (22.7), (22.10) and (22.13).

A. The Z-equation
Equation (22.7) is complemented by the Neumann BC at the top patch
and the grounded plane (electric walls):

Nikolova 2016 2
 ∂Az ∂Z
=⇒0 =0. (22.14)
∂z ∂z
Its solution, therefore, is in the form
 π 
Z ( z ) = ∑ c p cos  p z (22.15)
p  h 
with the eigenvalues are k z = pπ / h . Here, p is an integer.

B. The F-equation
The solution of (22.10) is also a harmonic function. We are interested in
real-valued harmonic functions, i.e.,
=F (φ ) ∑ bnc cos(kφnφ ) + bns sin(kφnφ ) . (22.16)
n

Since there are no specific BC’s to be imposed at certain angular positions,

the only requirement for the eigenvalues kφn comes from the condition that
the F (φ ) must be periodic in φ,
(φ ) F (φ + 2π ) .
F= (22.17)
Equation (22.17) is true only if kφn are integers. That is why the usual
construction of a general solution for F (φ ) for a complete cylindrical region
( φ = 0 to 2π ) is in the form
=F (φ ) ∑ bnc cos(nφ ) + bns sin(nφ ) , (22.18)
n

where n is an integer. This is the well-known Fourier-series expansion.

C. The R-equation
Equation (22.13) is a Bessel equation in which kφ is an integer ( kφ = n).
Solutions are of the form of the following special functions:
J n (k ρ ρ ) — Bessel function of the first kind,
N n (k ρ ρ ) — Bessel function of the second kind (Neumann function),
H n(1) (k ρ ρ ) — Hankel function of the first kind,
H n(2) (k ρ ρ ) — Hankel function of the second kind.
Note: H n(1) =
J n + jN n ; H n(2) =
J n − jN n .

Nikolova 2016 3
 Fig. D-1, Harrington, p. 461

Fig. D-2, Harrington, p. 462

Nikolova 2016 4
 The eigenvalues are determined according to the boundary conditions. In
the cavity model, it is required that (magnetic wall)
∂Az ∂R
=⇒
0 =0 , (22.19)
∂ρ ∂ρ ρ =a

and that the field is finite for ρ = a . The Bessel functions of the first kind
J n (k ρ ρ ) are the suitable choice. The eigenvalues k ρ are determined from
(22.19):
∂J n (k ρ ρ ) ′
χ nm
=0 , ⇒ k ρnm = , (22.20)
∂ρ ρ =a a
where χ nm ′ is the mth null of the derivative of the Bessel function of the nth
order J n′ . Thus, the solution of the Helmholtz equation for Az can be given in
a modal form as, see (22.4),
 ′ ρ c  π 
Az(= M mnp J m  χ nm  ⋅ bn cos ( nφ ) + bn sin ( nφ )  ⋅ cos  p
mnp ) s z  . (22.21)
 a  h 
The characteristic equation (22.12) is finally obtained as
k 2 ω 2 µε
= = kρ 2 + kz 2 . (22.22)
From (22.22), the resonant frequencies of the patch can be obtained:
′   π
 χ nm
2 2
µεω
= 2
mnp   + p  , (22.23)
 a   h

′   π
 χ nm
2 2
1
=f r ( mnp )   + p  . (22.24)
2π µε  a   h
Equation (22.24) does not take into account the fringing effect of the circular
patch. To account for the effective increase of the patch size due to fringing,
the actual radius a is replaced by an effective one,
1/2
 2h  πa  
ae =
a 1 +   2h 
ln + 1.7726  . (22.25)
 π aε r    
The first four modes in ascending order are TMz110, TMz210, TMz010,
Nikolova 2016 5
 ′ are
TMz310 where the respective nulls χ nm
′ = 1.8412
χ11 ′ = 3.8318
χ 01
′ = 3.0542
χ 21 ′ = 4.2012
χ31
The resonant frequency of the dominant TMz110 mode can be determined
from (22.25) as
1.8412 ⋅ c
f r (110) = (22.26)
2π ae ε r
where c is the speed of light in vacuum.
The VP of the dominant TMz110 mode is
 ′ ρ c
A=
(110)
z M 110 J1  χ11  ⋅ (bn cos φ + bn sin φ ) .
s (22.27)
 a
Assuming excitation at φ = 0 ( Az has vanishing angular first derivative), we
set bns = 0 . The field components are computed from Az according to the
field-potential relations

j ∂ 2 Az 1 1 ∂Az
Eρ = − Hρ =
ωµε ∂ρ∂z µ ρ ∂φ

j1 ∂ 2 Az 1 ∂Az (22.28)
Eφ = − Hφ = −
ωµε ρ ∂φ∂z µ ∂ρ

j  ∂ 2 Az 2A 
Ez =
− + Hz = 0
ωµε  ∂z 2
k z


For the dominant TMz110 mode,

E0 1
E=
ρ E=
φ 0 Hρ = j ′ ρ / a )sin φ
J1 ( χ11
ωµ0 ρ (22.29)
E0
′ ρ / a )cos φ
Ez = E0 J1 ( χ11 Hφ = j J1′( χ11
′ ρ / a )cos φ
ωµ0

Nikolova 2016 6
From the field components, we can compute the cavity modal impedance for
any feed point specified by ρ and φ . In view of the closed-wall nature of the
BCs, the impedance will be reactive. To obtain the real part of the antenna
impedance, the radiated power has to be computed.

2. Radiated fields and equivalent surface currents of the circular patch

a ρ′
φ′ εr y

Mφ
x

As with the rectangular patch, the field radiated by the circular slot is
determined using the equivalence principle. The circumferential wall of the
cavity is replaced by an equivalent circular sheet of magnetic current density
M sφ = 2 Ez ρ =a
, V/m, (22.30)

radiating in free space. The factor of 2 accounts for the ground plane. Since
the height of the slot h is very small and the slot field is independent of z, we
can substitute the surface magnetic current density over the slot with a
filamentary magnetic current I m = M sφ h :
′ )cos φ , V.
I m = 2hE0 J1 ( χ11 (22.31)

2V0

′ ) is the voltage between ground and the top plate of the

Here, V0 = hE0 J1 ( χ11
patch at the feed ( φ = 0 ).
Using the theory for the radiation field of a circular slot, the following
expressions are obtained for the far field of the circular patch:

Nikolova 2016 7
 Er =
0, Eθ = ′ , Eφ =
−C (r ) ⋅ cos φ ⋅ J 02 C (r ) ⋅ cosθ sin φ ⋅ J 02 , (22.32)
where
k0 aeV0e − jk0 r
C (r ) = j ,
2r
=J 02 J 0 (k0 ae sin θ ) + J 2 (k0 ae sin θ ) ,
= ′
J 02 J 0 (k0 ae sin θ ) − J 2 (k0 ae sin θ ) .

E-plane amplitude pattern:

Eθ (0 ≤ θ ≤ 90
= ,ϕ =
0 ,180) ′ , Eφ = 0
J 02

H-plane amplitude pattern:

Eϕ (0 ≤ θ ≤ 90 ,ϕ= 90 , 270=
) cosθ ⋅ J 02 , Eθ = 0

Fig. 14.23, p. 758, Balanis

Nikolova 2016 8
3. Circular polarization with patch antennas
Circular polarization can be obtained if two orthogonal modes are excited
with a 90° time-phase difference between them. This can be accomplished by
adjusting the physical dimensions of the patch and using either one or two
feed points.

A. Square patch with circularly polarized field

L
 1 
=L W 1 + 
 Qt 
W
(c) Nearly square patch with microstrip-line feed for CP accounting for
losses; Qt = 1 / tan δ eff
y y
Right-hand Left-hand

L (y',z')
L
(y',z')
Feed Feed
Point Point
z z
W W
(d) Coax-feeds for CP

Nikolova 2016 9
 W =L W =L

d d
c c
L L

Right-hand Left-hand

=
(e) CP for square patches with thin slots: c L=
/ 2.72 W / 2.72 , d = c / 10

B. Circular patch with circularly polarized field

Nikolova 2016 10
 FEED-PROBE ANGULAR SPACING OF DIFFERENT MODES FOR CIRCULAR
POLARIZATION

TM110 TM210 TM310 TM410 TM510 TM610

45° 30° 22.5° 18°, 54° 15°, 45°

α 90° or or or or or
135° 90° 67.5° 90° 75°

4. Array and feed networks

Nikolova 2016 11
Nikolova 2016 12