Microstrip Patch Antennas: Kai Fong Lee and Kin-Fai Tong

Microstrip Patch Antennas
Kai Fong Lee and Kin-Fai Tong
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 788
Modeling Techniques and Basic Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
Brief Description of Modeling Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
Basic Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 792
Broadbanding Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 799
General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 799
Stacked Patches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
Aperture-Coupled Patches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 803
The Wideband U-Slot Patch Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804
The Wideband L-Probe Coupled Patch Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 807
Dual- and Multiband Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 810
Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 810
The Use of U-Slot Patches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 810
Methods of Reducing the Patch Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 816
General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 816
The Use of Shorting Wall: Quarter-Wave Patch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 819
Partially Shorted Patch and Planar Inverted-F Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 820
The Use of Shorting Pin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 821
Wideband U-Slot Patch Antenna with Shorting Pin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 823
Discussion on Finite Ground Plane Size Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 823
Designs for Circular Polarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 826
Basic Principles of Circularly Polarized Patch Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 826
Broadbanding of Single-Feed Circularly Polarized Patch Antennas . . . . . . . . . . . . . . . . . . . . . . . 826
K.F. Lee (*)

Department of Electrical Engineering, The University of Mississippi, Mississippi, USA
e-mail: leek@olemiss.edu
K.-F. Tong
Department of Electronic and Electrical Engineering, University College London, London, UK
e-mail: k.tong@ucl.ac.uk
# Springer Science+Business Media Singapore 2016 787

Z.N. Chen et al. (eds.), Handbook of Antenna Technologies,
DOI 10.1007/978-981-4560-44-3_29
788 K.F. Lee and K.-F. Tong
Frequency-Agile and Polarization-Agile Patch Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 836

Frequency-Agile Microstrip Patch Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 837
Polarization-Agile E-Shaped Patch Antenna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 841
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 846
Cross-References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
Abstract
The basic geometry of a microstrip patch antenna (MPA) consists of a metallic
patch which is either printed on a grounded substrate or suspended above a
ground plane. The antenna is usually fed either by a coaxial probe or a stripline.
In the coaxial case, the center conductor is directly connected to the patch and the
outer conductor to the ground. In the stripline case, energy is coupled to the patch
in several ways: by direct connection, by proximity coupling, and by aperture
coupling. The patch antenna idea appeared to be originated in the early 1950s, but
there was little activity for almost two decades, mainly due to its inherent narrow
bandwidth. It began to attract the serious attention of the antenna community in
the 1970s, as antenna designers began to appreciate the advantages offered by this
type of antennas, which include low profile, conformability to a shaped surface,
ease of fabrication, and compatibility with integrated circuit technology. In the
last three decades, extensive studies have been devoted to improving the band-
width and other performance characteristics. This chapter begins with a brief
description of the modeling techniques and basic characteristics of the MPA.
Methods for broadbanding are then discussed, followed by dual- and multiband
designs, size reduction techniques, circularly polarized patch antennas, and
frequency-agile and polarization-agile designs. The chapter ends with some
concluding remarks.
Keywords
Microstrip antennas • Wideband patch antennas • Patch antennas for circular
polarization • Small-size patch antennas • Dual- and multiband patch antennas
Introduction
The idea of microstrip patch antennas arose from utilizing printed circuit technology
not only for the circuit components and transmission lines but also for the radiating
elements of an electronic system. The basic geometries are shown in Fig. 1. It
consists of a metallic patch which is either printed on a grounded substrate or
suspended above a ground plane and fed against the ground at an appropriate
location. The patch shape can in principle be arbitrary. In practice, the rectangle,
the circle, the equitriangle, and the annular ring are common shapes. Feeding is
usually by a coaxial probe or a stripline. The latter can be directly connected to the
patch, or proximity coupled to the patch, or coupled through an aperture. These are
Microstrip Patch Antennas 789
Fig. 1 The basic structure of the microstrip patch antenna
Fig. 2 Four common feeding methods of microstrip patch antenna
illustrated in Fig. 2 for the rectangular patch. Electromagnetic energy is first guided
or coupled to the region under the patch, which acts like a resonant cavity with open
circuits on the sides. Some of the energy leaks out of the cavity and radiates into
space, resulting in an antenna.
The advantages and disadvantages of the four feeding methods are listed in
Table 1. Of the four, the coaxial feed and the direct stripline feed are the most
common.
Table 1 The comparisons between the four common feeding methods for microstrip patch antenna
Advantages Disadvantages
Coaxial feed Easy to match Large inductance for thick
Low spurious radiation substrate
Soldering required
Microstrip line Monolithic Spurious radiation from feed line,
Easy to fabricate especially for thick substrate when
Easy to match by controlling insert line width is significant
position
Proximity coupled No direct contact between feed and Multilayer fabrication required
patch
Can have large effective thickness
for patch substrate and much thinner
feed substrate
Aperture coupled Use of two substrates avoids Multilayer fabrication required
deleterious effect of a high dielectric Higher back lobe radiation
constant substrate on the bandwidth
and efficiency
No direct contact between feed and
patch avoiding large probe
reactance or wide microstrip line
No radiation from the feed and
active devices since a ground plane
separates them from the radiating
patch
The patch antenna idea appeared to be proposed by Deschamps in the early 1950s
(Deschamps and Sichak 1953), but there was little activity for some two decades. It
was not until the late 1970s that this type of antenna attracted serious attention of the
antenna community, due to the recognition of the many advantages it offered. These
include low profile, conformability to a shaped surface, ease of fabrication, and
compatibility with integrated circuit technology, since active and passive circuit
elements can be etched on the same substrate. Its main drawback is narrow band-
width, typically less than 5 %. Research in the last several decades has resulted not
only in patch antennas with double digit bandwidths but a host of techniques for
designing dual- and multiband, dual- and circular polarization characteristics, size
reduction, and reconfigurable designs. The MPA has found extensive applications in
the wireless industry, both commercial and military, and it is probably not an
exaggeration to say that this type of antennas has become the favorite of antenna
designers. The literature on MPAs is vast, comprising thousands of research articles
in archival journals. There were also several review articles, handbook chapters, and
more than a dozen books, a partial list of which are given in the references [Bahl and
Bhartia (1980); Chen and Chia (2005); Debatosh and Antar (2010); Garg (2000);
Huang (2008); James and Hall (1981, 1989); Kumar and Ray (2003); Lee and Tong
(2012); Lee and Chen (1997); Richards et al. (1981); Pozar (1992); Shafai (2007);
Wong (2002); Zurcher and Gardiol (1995)].
In this chapter, a brief description of the modeling techniques and basic charac-
teristics of the MPA is given in section “Modeling Techniques and Basic Charac-
teristics.” Section “Broadbanding Techniques” presents four major broadbanding
techniques. Dual- and triple-band designs are discussed in section “Dual- and
Multiband Designs.” Methods for size reduction and circular polarization are the
topics of sections “Methods of Reducing the Patch Size” and “Designs for Circular
Polarization,” respectively. Section “Frequency-Agile and Polarization-Agile Patch
Antennas” presents examples of frequency-agile and polarization-agile patch anten-
nas. Concluding remarks are given in section “Conclusion.” To keep the chapter
within the allotted length, dual-polarization designs and the subject of microstrip
antenna array are not addressed in this chapter.
Modeling Techniques and Basic Characteristics
Brief Description of Modeling Techniques
There are two approaches to deducing the performance characteristics of microstrip

patch antennas. One is to devise a physical model based on a number of simplifying
assumptions; the other is to attempt to solve Maxwell’s equations subject to the
boundary conditions.
For the coaxial feed microstrip patch antenna shown in Fig. 1a, a physical model,
known as the cavity model (Lo et al. 1979), has been developed to deduce the
characteristics of the antenna. It is based on a number of simplifying assumptions
which are valid for thin substrates. These assumptions enable the fields between the
patch and the ground plane to be determined analytically for a number of patch
shapes. From these, radiation and impedance characteristics of the basic MPA
geometry can be calculated. Basic characteristics obtained from the cavity model
are presented in section “Basic Characteristics.”
More general methods which aim at solving the full set of Maxwell’s equations
without relying on the thin substrate assumption are known as full-wave methods.
One such method treats the currents on the antenna structure to be the unknowns to
be solved. By requiring the fields produced by the currents to satisfy the boundary
and excitation conditions, integral equations are obtained for the unknown currents,
which are solved by some numerical method, such as the moment method (Mosig
and Gardiol 1985). Another approach of the full-wave method converts Maxwell’s
equations into difference equations, which, together with appropriate models for the
boundary conditions, are solved numerically in the time domain (Reineix and Jecko
1989). This is known as the finite difference time domain (FDTD) method. Still
another method, known as the finite element method, solves Maxwell’s eq. in the
form of the vector wave eq. by the Rayleigh-Ritz variational method (Salon and
Chari 1999).
The cavity model provides physical insight, but works well only for the basic
geometry of the MPA, namely, a conducting patch on a thin, grounded substrate,
Table 2 A list of commercial electromagnetic simulation software

Software name Theoretical model Model dimension Company
Ensemble (DesignerRF) Moment method 2.5 Ansoft (ANSYS)
IE3D Moment method 2.5 Zeland
Momentum Moment method 2.5 Agilent
EM Moment method 2.5 Sonnet
PiCasso Moment method/genetic 2.5 EMAG
FEKO Moment method 3D EMSS
Microwave studio FDTD 3D CST
Fidelity FDTD 3D Zeland
HFSS Finite element 3D Ansoft (ANSYS)
with the patch shape amenable to the method of separation of variables. The full-
wave methods have the advantages of being able to handle thick substrates, various
feeding methods, and thick complicated structures such as multilayers, multiple
patches with or without parasitic elements, shorting walls, and shorting pins, as
well as slots cut in patches. However, they require extensive computation time and
provide little physical insight.
In the last two decades, many commercial simulation software based on full-wave
methods are available. These software have played a key role in the rapid advance-
ment of microstrip antenna research. They have become indispensable tools in the
design of these antennas. Some of these are listed in Table 2. Many of the results to
be presented in this chapter were based on IE3D and HFSS. Some were obtained
using moment method and FDTD codes developed in-house.
Basic Characteristics
The basic characteristics of microstrip patch antennas can be obtained by applying

the cavity model to a coaxially fed microstrip patch antenna. The basic assumption in
the cavity model is that the thickness of the substrate is much smaller than the
wavelength. The region under the patch can then be considered as a cavity bounded
by electric walls on the top and bottom (Et = 0) and, as a consequence of the thin
substrate assumption, bounded by a magnetic wall (Ht = 0) on the sides. The fields
under the patch are transverse magnetic (TM), with the electric field perpendicular to
the patch. These fields are determined by solving the wave equation subject to the
conditions on the boundary and the feed. By virtue of the equivalence principle, the
fields outside the cavity can be calculated from the equivalent sources on the exit
region of the cavity.
Using this approach, the performance characteristics of the probe-fed rectangular,
circular, annular-ring, and equitriangular patch antennas were predicted theoretically
and verified experimentally in the late 1970s and 1980s (Lee and Dahele 1989).
While differing in detail, there are a number of similar features, irrespective of the
Fig. 3 Geometry of a
rectangular patch antenna
shapes of the patch. These are delineated below for the rectangular patch shown in
Fig. 3. The rectangular patch is probably the most commonly used microstrip patch
antenna.
(A) The fields under the cavity are transverse magnetic, with the electric field in the
z direction and independent of z. There are an infinite number of modes, each
characterized by a pair of integers (m, n):
mπx nπy
Ez ¼ E0 cos cos m, n ¼ integers (1)
a b
(B) For the cavity bounded by electric walls on the top and a magnetic wall on the
side, the resonant frequency of each mode is governed by the dimensions of the
patch and the relative permittivity of the substrate er. It is given by
kmn c
f mn ¼ pffiffiffiffi (2)
2π er
where
12
mπ 2 nπ 2
kmn ¼ þ (3)
a b
c = velocity of light in free space.

(C) Because of fringing fields at the edge of the patch, the patch behaves as if it has
a slightly larger dimension. Semiempirical factors are usually introduced in the
cavity model-based design formulas to account for this effect, as well as the
fact that the dielectric above the patch (usually air) is different from the
dielectric under the patch. These factors vary from patch to patch.
Fig. 4 Top view of the

rectangular patch
For the rectangular patch, with a > b, a commonly used formula for the
fundamental mode, accurate to within 3 % of measured values, is
c
fr ¼ pffiffiffiffi (4)
2ð a þ t Þ e e
where
1
ð e r þ 1Þ ð e r 1Þ 10t 2
ee ¼ þ 1þ (5)
2 2 b
is the effective permittivity (Schneider 1969).

(D) The equivalent sources at the exit region (the vertical side walls) are the surface
magnetic current densities, related to the tangential electric fields in those
locations. Referring to Fig. 4, they are:
Along the wall OA,
mπx mπx
Ms ¼ n E ¼ þy z E0 cos ¼ þxE0 : (6)
a a
Along the wall AC,
nπy nπy
Ms ¼ n E ¼ x z E0 cos mπ cos ¼ þyE0 cos mπ cos : (7)
b b
Along the wall CB,
mπx mπx
Ms ¼ n E ¼ y z E0 cos cos nπ ¼ xE0 cos cos nπ: (8)
a a
Along the wall BO,
nπy nπy
Ms ¼ n E ¼ þx z E0 cos ¼ yE0 cos : (9)
b b
The effect of the infinite ground plane is to double the strength of each Ms
source. The electric field and magnetic surface current distributions on the side
walls for the TM01 and TM10 modes are illustrated in Fig. 5. For the TM10
Fig. 5 Illustrations of the magnetic current density distribution on the side walls
mode, the magnetic currents along b are constant and in phase, while those
along a vary sinusoidally and are out of phase. For this reason, the b edge is
known as the radiating edge since it contributes predominately to the radiation.
The a edge is known as the non-radiating edge. Similarly, for the TM01 mode,
the magnetic currents are constant and in phase along a and are out of phase
and vary sinusoidally along b. The a edge is thus the radiating edge for the
TM01 mode.
(E) To satisfy the boundary condition imposed by the feed, the fields under the
patch are expressed as a summation of the various modes. The amplitudes and
phases of these modes are determined by the excitation. The mode with
resonant frequency equal to the excitation frequency will be at resonance and
has the largest amplitude.
(F) Each resonant mode has its own characteristic radiation pattern. For the
rectangular patch, the commonly used modes are TM10 and TM01. However,
the TM03 mode has also received some attention. These three modes all have
broadside radiation patterns. The computed patterns for a = 1.5b and two
values of er are shown in Fig. 6. In the principal planes, the TM01 and TM03
modes have similar linear polarizations, while that of the TM10 mode is
orthogonal to the other two. The patterns do not appear to be sensitive to a/b
or t. However, they change appreciably with er. Typical half-power beamwidths

of the TM10 and TM01 modes are of the order of 100 and the gains are typically
5 dBi. The patterns of most of the other modes have maxima off broadside. For
example, those of the TM11 mode are illustrated in Fig. 6g.
Figure 7 shows the computed and measured radiation patterns of the TM10 and
TM01 modes obtained by Lo et al. (1979) for a rectangular patch with
a = 11.43 cm, b = 7.6 cm, er ¼ 2:62 ,and t = 0.159 cm. Both Eθ and Eϕ
were measured in each of the two cuts, ϕ = 0 and ϕ = 90 . It was found in
their experiment that, in each case, one component polarization was negligible
when compared to the other and is not shown. The dominant component is
called the co-polarization, and the component orthogonal to the dominant
component is called the cross polarization. For thin substrates, cross-
polarization fields are mainly due to the excitation of higher-order modes. A
study of the dependence of the cross-polarized fields on aspect ratio (a/b), feed
position, and relative permittivity can be found in Huynh et al. (1988).
For each mode, there are two orthogonal planes in the far-field region – one
designated as E-plane and the other designated as H-plane. The far-zone
electric field lies in the E-plane and the far-zone magnetic field lies in the
H-plane. The patterns in these planes are referred to as the E- and H-plane
patterns, respectively. For the TM01 mode, the contributions to the far fields are
from the magnetic surface current densities on the side walls containing the
radiating edges. Referring to Figs. 3 and 5, one can show from the directions of
these magnetic currents that the E-plane is the y-z plane (ϕ = 90o) and the
H-plane is the x-z plane (ϕ = 0o).
For the TM10 mode, again referring to Figs. 3 and 5, one can show that the
E-plane is the x-z plane (ϕ = 0o) and the H-plane is the y-z plane (ϕ = 90o).
With appropriate design, circular polarization can be achieved by utilizing
two modes. This will be discussed in detail in section “Designs for Circular
Polarization.”
(G) At resonance, the input reactance is small for thin substrates, while the input
resistance is largest when the feed is near the edge of the patch and decreases as
the feed moves inside the edge. The decrease follows the square of a cosine
function for the TM10 and TM01 modes of a coaxial-fed rectangular patch.
Figure 8 shows the theoretical and measured values of the resonant resistances
of the first two modes of a coaxial-fed rectangular patch.
(H) By choosing the feed location properly, the resonant resistance can be matched
to the feed line resistance, while the use of thin substrates (thickness t 0.03
λ0) will minimize the feed inductance at resonance, resulting in a very near
unity voltage standing wave ratio (VSWR or SWR). As the frequency deviates
from resonance, VSWR increases. For linear polarization, a common defini-
tion of impedance bandwidth is the range of frequencies for which VSWR is
less than or equal to two, corresponding to 10 dB return loss or 10 dB for the
Fig. 6 Relative field patterns for a rectangular patch with a/b = 1.5, fnm = 1 GHz, and
(i) er = 2.32, t = 0.318, 0.159, 0.0795 cm; (ii) er = 9.8, t = 0.127, 0.0635, 0.0254 cm. (a)
TM10, ϕ = 0 . (b) TM10, ϕ = 90 . (c) TM01, ϕ = 90 . (d) TM01, ϕ = 0 . (e) TM03, ϕ = 90 .
(f) TM03, ϕ = 0 . (g) TM11, er = 2.32
reflection coefficient (scattering parameter S11). This is usually also the

antenna bandwidth, as the patterns are much less sensitive to frequency. For
circular polarization, bandwidth is determined by both VSWR 2 and axial
ratio 3 dB.
a c
φ = 90° Eφ φ = 90° Eθ
b d
φ = 0° Eθ φ = 0° Eφ
Fig. 7 Theoretical (x) and measured (solid or dashed line) radiation patterns in ϕ = 0 and
ϕ = 90 planes of a rectangular patch antenna with a = 11.43 cm, b = 7.6 cm, er = 2.62,
t = 0.159 cm. (a) and (b) at resonant frequency 804 MHz of (1, 0) mode; (c) and (d) at resonant
frequency 1,187 MHz of (0, 1) mode (From Lo et al. (1979). # 1979 I.E. Reprinted with
permission)
(I) The losses in the patch antenna comprise radiation, copper, dielectric, and
surface-wave losses. For thin substrates, surface wave can be neglected.
According to Wood (1981), it is required that t/λo < 0.07 for er = 2.3 and
t/λo < 0.023 for er = 9.8, if the antenna is to launch no more than 25 % of the
total radiated power as surface waves. The quality factor Q of a particular mode
is determined by the ratio of the stored to loss energy and determines the
impedance bandwidth of the antenna.
(J) In general, the impedance bandwidth is found to increase with substrate thick-
ness t and inversely proportional to √er. However, the use of low-permittivity
substrates can lead to high levels of radiation from the feed lines, while for higher
permittivities, an increase in substrate thickness can lead to decrease in efficiency
500
measurement
resonant resistance (Ω)
400 theory
300
(1,0) mode
200
100
(0,1) mode
0
0 0.1 0.2 0.3 0.4 0.5
y’/b for x’ = 5.33 cm (0,1) mode
x’/b for y’ = 3.81 cm (1,0) mode
(x’,y’) is location of feed point
feed for (1,0) mode varies rectangular

along this line microstrip antenna
7.62cm
usihng 1/16*
rexolite 2200
feed for (0,1)
3.83cm
mode varies
along this line
5.33cm
11.43cm
Fig. 8 Variation of resonant resistance with feed position of the TM10 and TM01 mode in a
rectangular patch antenna with a = 11.43 cm, b = 7.62 cm, er ¼ 2:62 , t = 0.159 cm (From
Richards et al. (1981). # 1981 I.E. Reprinted with permission)
due to surface-wave generation. Additionally, when the substrate thickness

exceeds about 0.05 λo, where λo is a free-space wavelength, the antenna cannot
be matched to the feed line. As a result, for the basic MPA geometry, the
impedance bandwidth is limited to about 5 %.
Broadbanding Techniques
General Remarks
The operating frequencies of some wireless communication systems are shown in

Table 3.
Table 3 Frequencies of some wireless communication systems

Overall
System Operating frequency bandwidth
Advanced mobile phone service (AMPS) UL: 824–849 MHz 70 MHz (8.1 %)
DL: 869–894 MHz
Global system for mobile communications UL: 880–915 MHz 80 MHz (8.7 %)
(GSM) DL: 925–960 MHz
Global positioning system (GPS) L1: 1,575.42 MHz P-codea: 20.46
L2: 1,227.60 MHz MHz
L5: 1,176.45 MHz C/A codeb: 2.046
MHz
(max. 1.7 %)
Personal communications service (PCS) UL: 1,710–1,785 MHz 170 MHz (9.5 %)
DL: 1,805–1880 MHz
Universal mobile telecommunication UL: 1,920–1,980 MHz 250 MHz
systems (UMTS) DL: 2,110–170 MHz (12.2 %)
Wi-Fi (b, g, n) 2,400–2,497 MHz (a, b, g) 20 MHz
IEEE 802.11(a, b, g, n) (a, n) 5,150–5,350 MHz (n) 20, 40 MHz
IEEE802.11 ac (a, n) 5,725–5,825 MHz (max. 0.8 %)
5,170–5,330 MHz 20, 40, 80,
5,490–5,710 MHz 160 MHz
5,735–5,835 MHz (max. 3 %)
Wireless gigabit 56,160–64,800 MHz 8,640 MHz
(14 %)
WiMAX M2300T-01/02: 100 MHz
2,300–2,400 MHz 194 MHz
M2500T-01: 200 MHz
2,496–2,690 MHz (max. 5.7 %)
M3500T-02/03/05:
3,400–3,600 MHz
Bluetooth 2,400–2,483.5 MHz 83.5 MHz (3.4 %)
a
Precision (P)
b
Coarse-acquisition (C/A)
While the bandwidths of several systems (e.g., GPS) are quite narrow, five of the
systems shown in Table 3 require bandwidths in excess of 8 %. As discussed
previously, the microstrip patch antenna in the basic form of a conducting patch in
a grounded substrate is inherently narrow band and is not able to meet the require-
ments of these systems. While bandwidth can be increased by using lossy substrates,
this is usually not desirable as efficiency will be reduced. In the last three decades, a
number of techniques have been developed to broaden the bandwidths of microstrip
patch antennas, without compromising efficiency. The various designs provide
bandwidths in the range of 10–60 %. Note that a 20 % bandwidth will be able to
cover the first two applications in Table 3, while a 30 % bandwidth will cover the
third, fourth, and fifth applications combined.
The methods developed for efficient wideband patch antenna design are based on
one or more of the following principles:
(A) By means of parasitic elements or slots, additional resonances are introduced so

that, in conjunction with the main resonance, an overall broader band response
is obtained.
(B) Thick substrates of low permittivity are used.
(C) A scheme is devised to reduce the mismatch problem associated with thick
substrates.
Several wideband designs based on the above principles will now be presented.
Stacked Patches
A popular method for increasing the bandwidth of microstrip patch antennas is to use
parasitic elements to create dual or multiple resonances. If these resonances are
suitably arranged, the resulting overall frequency response can be significantly
broader than the case when only a single resonance is present. This is similar to
the situation in coupled tuned circuits. There are two arrangements: the stacked
geometry in which the fed and parasitic patches are on separate layers (Sabban 1983;
Lee et al. 1987, 1995; Barlately et al. 1990) and the coplanar geometry in which all
patches are on the same layer (Kumar and Gupta 1984). Thin substrates are used.
The coplanar geometry increases the lateral area occupied by the patches and is not
as popular as the stacked geometry and will not be discussed further.
One of the first journal papers on two-layer stacked patches is that by Lee
et al. (1987), who studied the configuration shown in Fig. 9.
The substrate has relative permittivities, er = 2.17 and thickness t = 0.254 mm.
They are separated by a distance s by means of spacers. The dimensions of the
patches are identical, with a = 1.5 cm and b = 1.0 cm. It is found that when the
patches are excited at the TM01 mode, with resonant frequency approximately
10 GHz, the impedance bandwidth is about 13 % when the spacing is around
0.0508 cm (0.017 λ0). This is almost an order of magnitude larger than the case
when there is no parasitic patch.
Subsequently, Lee et al. (1995) developed a full-wave moment method analysis
and a computer program for multilayer microstrip antennas. Using this program,
representative design guides for the configuration of Fig. 10, operating at the center
frequency of 5 GHz, are shown in Table 4. In Table 4, design 1 gives the parameters
which achieve a bandwidth of 12 % for the case when there is no superstrate
Fig. 9 Geometry of
rectangular
electromagnetically coupled
stacked patch antenna
Fig. 10 Geometry of the stacked electromagnetically coupled microstrip antenna with superstrate
Table 4 Design examples 12 % 12 % 15 %

for stacked
Bandwidth (BW) Design Design Design
electromagnetically
(VSWR 2) 1 2 3
coupled patch antennas at
the center frequency of e1 1.0 2.2 1.0
5 GHz e2 1.2 1.2 1.2
e3 2.2 2.2 2.2
h1 (mm) 0.0 0.26 0.0
h2 (mm) 3.580 3.630 3.500
h3 (mm) 0.493 0.486 1.200
a1 (cm) 2.296 2.198 2.300
b1 (cm) 1.275 1.465 1.278
a2 (cm) 2.000 2.027 2.000
b2 (cm) 1.111 1.351 1.111
F (cm) 0.928 0.940 0.900
(dielectric cover). When a superstrate of thickness of 0.26 mm and relative permit-

tivity of 2.2 is placed on top of the parasitic patch, the parameters which yield 12 %
impedance bandwidth are given by design 2. Design 3 provides the antenna param-
eters which result in a bandwidth of 15 % when no superstrate is present. If the center
frequency is changed, it is only necessary to scale the length parameters accordingly
(patch dimensions, substrate and superstrate thickness, feed location).
The patterns of stacked patches are stable across the impedance bandwidth.
Typical E- and H-plane half-power bandwidths are 76 and 86 , respectively. This
is to be compared with 92 and 86 for the single patch. The gain of the stacked
patches is about 6.0 dBi and that of the single patch is about 5.2 dBi.
It is obvious that, by extending the stacked patch geometry to multiple parasitic
patches in multiple layers, additional resonances can be introduced, further broad-
ening the bandwidth of the antenna. However, the antenna becomes very thick when
there are more than two layers.
Aperture-Coupled Patches
General Remarks About Aperture-Coupled Feed

While the most popular methods for feeding a patch antenna are directly
connecting a coaxial probe or stripline to the patch, a method which has a number
of advantages is to couple energy from the stripline through an aperture (slot) in the
ground plane, as shown in Fig. 1d. The open-ended microstrip line is located on a
dielectric slab below the ground plane, while the patch antenna is formed on a
separate dielectric slab above the ground plane. The two structures are electro-
magnetically coupled through an aperture in the ground plane between them. In the
original paper by Pozar (1985), the aperture was in the form of a small circular
hole. Subsequently, a more common shape of the aperture was in the form of a
narrow rectangular slot.
As listed in Table 1, one advantage of this feeding method is that the feed network
is isolated from the radiating element by the ground plane, which prevents spurious
radiation. Another advantage is that active devices such as phase shifters and
amplifiers can be fabricated in the feed substrate with high dielectric constant for
size reduction, while the radiating patch can be mounted on low dielectric constant
substrate in order to increase bandwidth and radiation efficiency.
The slot that couples energy from the stripline to the patch can be either resonant
or nonresonant. The bandwidth achievable by nonresonant slot excitation is typically
6–7 % but can be as large as 10–13 % by utilizing thick substrates, since the problem
of probe inductance is not applicable here. If the slot is resonant, it provides another
resonance in addition to the patch resonance, resulting in an antenna with impedance
bandwidth in excess of 20 %. However, the resonant slot has a strong back lobe
radiation which reduces the gain of the antenna and causes interference problems in
certain applications.
As with coaxially fed microstrip antennas, the impedance bandwidth of aperture-
coupled patch antennas can be further enhanced by using stacked or coplanar
elements.
In the next section, an example of a wideband aperture-coupled patch antenna
utilizing a resonant slot is presented.
Example of a Wideband Aperture-Coupled Patch Antenna

By using a resonant slot and relatively thick foam substrate for the patch, Croq and
Papiernik (1990) studied the aperture-coupled patch antenna shown in Fig. 11. Note
that there was a dielectric cover (radome) protecting the patch.
The measured and computed impedances showed that, in the frequency
range 4.85–6.1 GHz, the VSWR was less than 1.5, corresponding to a bandwidth
of about 22 %.
The antenna gain was found to be about 8 dB for the entire bandwidth. The
maximum back to front level was about 14 dB at the frequency of 5.6 GHz and was
about 12 dB over the band. The strong back radiation is a major disadvantage of a
resonant slot aperture-coupled patch antenna.
Fig. 11 Aperture-coupled
patch antenna studied in the
paper by Croq and Papiernik
(1990). (a) Feed: erf = 2.2;
tgδ = 0.001; Hf = 0.762 mm;
Wf = 2.32 mm; Ls = 2.85
mm. (b) Slot: Aw = 0.8 mm;
AL = 15.4 mm. (c) Square
patch: Wp = 17 mm;
Hp = 5.5 mm; erp = 1. (d)
Radome: Hs = 1.6 mm;
ers = 2.2; tgδ = 0.001
The Wideband U-Slot Patch Antenna
The stacked patch geometry with one fed patch and one parasitic patch can provide
up to about 20 % bandwidth, while the aperture-coupled patch utilizing a resonant
slot can yield up to about 25 % bandwidth. The former involves more than one layer
and more than one patch, while the latter also involves more than one layer, as well
as a more complicated feeding arrangement and strong back lobe radiation. A
wideband design which is single layer and single patch is the U-slot patch antenna
shown in Fig. 12, first introduced in 1995 by Huynh and Lee (1995). A number of
studies by Lee and coworkers, as well as by others, followed (Lee et al. 1997, 2010;
Tong et al. 2000; Clenet and Shafai 1999; Weigand et al. 2003). It was firmly
established that the U-slot patch antenna can provide impedance bandwidths in
excess of 30 % for an air substrate of thickness about 0.08 λ0 and in excess of
20 % for material substrates of similar thickness.
Air Substrate
In the original study of Huynh and Lee (1995), the wide-bandwidth characteristics of
the antenna were demonstrated experimentally. It was pointed out in their paper that
the factors contributing to the wideband behavior were (1) the air substrate; (2) a
relative thick substrate (about 0.08 λ0); (3) the capacitance introduced by the U-slot,
which countered the feed inductance; and (4) the additional resonance introduced by
the U-slot, which combined with the patch resonance to produce a broadband
Fig. 12 Geometry of the rectangular U-slot patch antenna
response. Subsequently, Lee et al. (1997) produced a more comprehensive experi-

mental and FDTD modeling of the air-substrate U-slot patch antenna. Results for the
VSWR response of one such antenna are shown in Fig. 13. The impedance band-
width is about 30 %. The measured patterns (not shown here) were stable across the
band. The E-plane and H-plane beamwidths were about 70 and 65 , respectively.
The gain of the antenna was around 7.5 dBi, about 2 dB higher than the traditional
microstrip antenna. The cross polarization was negligible in the E-plane. It was about
12 dB below the co-polarization level in the H-plane at the center of the band, but it
increased to about 8 dB at the outer edges of the band.
While the abovementioned studies, as well as others, have shown that more than
30 % impedance bandwidth can be obtained when an air-substrate thickness of about
0.08 λ0 is used, it should be pointed out that some applications do not need such a
wide bandwidth. For example, referring to Table 3, in Advanced Mobile Phone
Service (AMPS), about an 8.1 % bandwidth is sufficient. Likewise, in Global System
for Mobile Communications (GSM), only an 8.7 % bandwidth is needed. While such
10
9
8
7
6
VSWR
5
4
3
2
1
3.5 3.9 4.3 4.7 5.1 5.5
frequency, GHz
Fig. 13 VSWR of the U-slot patch antenna with dimensions: W = 36 mm, L = 26 mm, F = 13
mm, Ws = 12 mm, Ls = 20 mm, a = 2 mm, b = 4 mm, cx = cy = 2 mm, and h = 5
mm. ( measured, — computed) (From Lee et al. (1997). # 1997 IET, Reprinted with permission)
Table 5 Dimension of the h (mm) W L Ws Ls a b F

U-slot patch antenna with
1 35.5 26 12 10.7 0.1 0.6 15
different air-substrate
thickness (unit: mm) (From 2 35.5 26 11 12.3 0.6 0.8 15
Lee et al. (2010). # 2010 3 35.5 26 11 14.2 0.9 1 15
IEEE, Reprinted with 4 35.5 26 11 16.2 1.8 1.6 15
permission) 5 35.5 26 11 20 3.8 2.1 15
6 35.5 26 11 21 4.3 2.1 15
Table 6 Simulated and measured impedance matching bandwidths of probe-fed U-slot patch
antenna with different air-substrate thickness. The dimensions of the antennas are slightly different
for the different thicknesses. In all cases, the inner radius of the coaxial feed is 0.4 mm (From Lee
et al. (2010). # 2010 IEEE, Reprinted with permission)
Simulation Measurement
h Thickness in lambda Matching freq % BW Matching freq % BW
(mm) (GHz) (GHz)
1 0.018 5.13–5.41 5.31
2 0.033 4.73–5.23 10.04 4.44–5.01 12.1
3 0.048 4.45–5.2 15.54 4.19–4.95 16.6
4 0.063 4.2–5.28 22.78 4.05–5.1 23.0
5 0.078 3.98–5.32 28.81 3.8–4.94 26.1
6 0.089 3.8–5.14 29.97 3.74–4.98 28.4
bandwidths cannot be realized by the traditional patch antenna (Lee and Luk 2010),
it has been demonstrated that these can be realized by a U-slot patch antenna only
0.033 λ0 thick, which has a 12 % bandwidth. The dimensions of the antenna used in
this study are shown in Table 5, and the results are summarized in Table 6.
Table 7 Dimensions of antenna in millimeters

εr W L Ws Ls b F cx cy h
2.33 36.0 26.0 14.0 18.0 4.0 13.0 2.0 2.0 6.4
Table 8 Operating frequencies and bandwidth of the antenna in Table 7

fl(GHz) fo(GHz) fu(GHz) BW (GHz) BW (%)
Computed 2.87 3.28 3.69 0.82 25.0
Measured 2.76 3.16 3.56 0.80 25.3
Material Substrate
Although the first series of investigations of the U-slot patch antenna used an air or
foam substrate, subsequent investigations have confirmed that this wideband design
can also be implemented with material substrates. As expected, the bandwidth of an
antenna on a material substrate is smaller than an antenna on an air or foam substrate.
Tong et al. (2000) presented both experimental study and FDTD analyses of two
antennas on a substrate with relative permittivity er = 2.33. The dimensions of one
of these antennas are shown in Table 7. The operating frequencies and bandwidths of
this antenna are shown in Table 8. The 3-dB gain bandwidths were about the same as
the impedance bandwidths, and the average gains of the antennas were about 7 dBi
across the matching band.
Variations of the U-Slot Patch Antenna and the E-Patch

The use of the U-slot to achieve wideband performance is not limited to a rectangular
patch. Wideband circular patch antennas with a U-shaped slot have been reported
(Bhalla and Shafai 2002). The shape of the slot can also take on other forms, such as
a circular slot and a V-slot (Rafi and Shafai 2004). By letting the width of the
horizontal slot go to zero and extending the two vertical slots to the edge of the patch,
an E-patch is obtained (Yang et al. 2001a). This geometry is shown in Fig. 14. As in
the U-slot, the parallel slots provide an additional path for the currents, giving rise to
a second resonance. The parallel slots can also introduce a capacitance which
compensates for the probe inductance, thus enabling the use of relatively thick
substrates. In Yang et al. (2001a), impedance bandwidths of about 30 % were
obtained for E-patches operating at the center frequency of around 2.4 GHz, using
air substrate of about 0.08 λ0. The antenna parameters for one such antenna are listed
below, in mm: L = 70, W = 30, h = 15, Xf = 35, Yf = 6, Ls = 40, Ws = 6, and
Ps = 10. Ground plane size = 14 21 cm.
The Wideband L-Probe Coupled Patch Antenna
Shortly after the publication of the first paper on the wideband U-slot patch antenna,
another wideband single-layer single-patch antenna was introduced. This design
achieves wideband operation using an L-shaped probe feeding method. The
Fig. 14 Geometry of the

E-patch antenna
Fig. 15 Geometry of the

patch antenna with an
L-probe feed
L-probe feed was first used by Nakano in the electromagnetically coupled curl
antenna (Nakano et al. 1997). Its use in patch antennas was due to Luk
et al. (1998a). The geometry is shown in Fig. 15. This design uses
low-permittivity substrate (air or foam) of thickness about 0.1 free-space wave-
length. The feed is a modified version of the coaxial probe. Instead of the center
conductor extending vertically to the patch and connected to it, a portion of it is bent
in the horizontal direction. The horizontal arm of the probe is approximately a
quarter of a wavelength long. It provides a capacitance to counteract the inductance
due to the vertical part. In the experiment of Luk et al. (1998a), the rectangular patch
had a width Wx = 30 mm and a length Wy = 25 mm. The L-shaped probe,
connected to the inner conductor of a 50 Ω SMA launcher, had a parallel arm
10 7
gain
6
5 5
gain,dBi
SWR
4
0 3
SWR
2
–5 1
3 4 5 6
frequency, GHz
Fig. 16 Measured gain and SWR against frequency (From Luk et al. (1998a). # 1998 IET,
Reprinted with permission)
Fig. 17 Measured radiation pattern at 4.53 GHz (From Luk et al. (1998a). # 1998 IET, Reprinted
with permission)
Lh = 10.5 mm and a perpendicular arm Lv = 4.95 mm. It was excited in the TM01
mode of the patch antenna. The patch was supported by a foam layer of thickness
H = 6.6 mm with a dielectric constant close to unity. Within the matching band, this
thickness corresponded to 0.08–0.12 λ0. The distance between the lower edge of the
patch and the perpendicular arm of the L-probe was D = 2 mm. The probe radius
R = 0.5 mm. The measured VSWR and gain as a function of frequency are shown in
Fig. 16. The bandwidth (VSWR < 2) was 36 % and the average gain was 7.5 dBi.
The measured radiation patterns at 4.53 GHz are shown in Fig. 17. It is seen that the
co-polarization patterns were symmetrical with respect to the broadside direction.
The cross polarization was negligible in the E-plane (y-z plane). It was quite high in
the H-plane (x-z plane) in the direction around θ = 30 . Simulation results of this
antenna using the software IE3D were in good agreement with experimental results.
Subsequent related studies included the use of twin L-probes to increase gain and
reduce cross polarization (Mak et al. 2005) and the T-probe-fed patch antennas (Mak
et al. 2000).
Dual- and Multiband Designs
Introductory Remarks
There are many applications in wireless communications that involve two or more
distinct frequency bands. It is sometimes possible that a broadband microstrip
antenna can cover the frequencies of interest. However, the disadvantage of using
a broadband antenna is that it also receives non-desired frequencies unless some kind
of filtering network is introduced to reject such frequencies. On the other hand, the
advantage of a dual- or multiband design is that it focuses only on frequencies of
interest and is thus more desirable. The frequency bands can be relatively close or
relatively far apart. A variety of designs have been developed to provide the desired
characteristics: by providing multi-resonances, by using multi-modes, or by intro-
ducing band notches in a broadband antenna. Long and Walton (1979) and Dahele
et al. (1987) used stacked circular and stacked annular-ring patches for dual-band
designs. This technique was extended to multibands by Anguera et al. (2003). Dual-
or triple-band characteristics can also be realized by loading a rectangular patch with
one or two stubs (Deshmukh and Ray 2010), by utilizing the TM01 and TM03 modes
of a rectangular patch (Zhong and Lo 1983) or the TM10, TM20, and TM11 modes of
the equilateral-triangular patch (Lee et al. 1988). Recently, it was found that the
U-slot patch geometry, originally introduced for broadband operation, can be suit-
ably modified for dual- and multiband operations (Lee et al. 2008, 2011). This
method is discussed in detail in the next section.
The Use of U-Slot Patches
Most of the studies of the U-slot patch antenna were concerned with its broadband
capabilities. However, recent studies have shown that it can be also used for dual- or
multiband designs. There are two approaches, depending on whether the frequency
ratios are larger than or less than about 1.5. The two cases will be referred to as large
frequency ratio and small frequency ratio, respectively.
Large Frequency Ratios

For the case when the frequency ratio is larger than about 1.5, the approach to dual-
band design is to use a coaxial feed and adjust the U-slot dimensions so that the patch
resonance and the slot resonance do not merge to yield a broadband response.
Figure 18 shows a U-slot patch antenna which operates at a band centered at
Fig. 18 Geometry of dual-band U-slot patch antenna
2.0 GHz and another band centered at 4.8 GHz. The patch dimension L (54 mm)
determines the lower resonance, while the upper resonance is governed by the
dimensions of the U-slot. The simulated reflection coefficient (S11) versus frequency
is shown in Fig. 19. The impedance bandwidths at the lower and upper bands are
3.5 % and 18.2 %, respectively. The radiation patterns are similar to those of the
broadband U-slot patch antenna.
To achieve triple-band operation, two slots are necessary. While it is possible to
use two U-slots, it was found that the second slot in the form of an H-slot offers
better flexibility. Figure 20 shows a U-slot in combination with an H-slot to produce
resonances at 1.94, 4.16, and 5.44 GHz. Figure 21 shows the reflection coefficient
(S11) response curve. The impedance bandwidths at the three bands are 2.6 % for the
lower band, 9.8 % for the middle band, and 10.4 % for the upper band. The radiation
patterns at the three bands are shown in Fig. 22. The co-polarization patterns are
stable. The cross-polarization levels are 20 dB below the co-polarization at the lower
band. At the center frequency of the middle band (4.2 GHz) and high band (5.4
GHz), it becomes very high at oblique angles, as the antenna is electrically thick
(>0.08 λo) and radiation from the vertical coaxial feed becomes significant.
Fig. 19 Simulated reflection

coefficient (S11) of the dual-
band U-slot patch antenna
with dimensions W = 64
mm, L = 54 mm, F = 28.4
mm, Ws = 34 mm, Ls = 22
mm, a = 8.4 mm, b = 11
mm, and h = 6 mm
Fig. 20 Geometry of tri-band U-slot patch antenna

Fig. 21 Simulated reflection

coefficient (S11) of the
tri-band U-slot patch antenna
of Fig. 20 with dimensions
W = 64 mm, L = 54 mm,
F = 28.4 mm, WU = 34 mm,
LU = 22 mm, WH = 28 mm,
LH = 13.5 mm, a = 8.4 mm,
b = 11 mm, c = 1.5 mm,
d = 2 mm, and h = 6 mm
a 0 b 0
330 30 330 30
300 60 300 60
270 90 270 90
–40 –30 –20 –10 –40 –30 –20 –10
0
c 330 30
300 60
270 90
–40 –30 –20 –10
Fig. 22 Simulated radiation pattern of the tri-band U-slot antenna at (a) 1.94 GHz, (b) 4.16 GHz,
and (c) 5.44 GHz (( H-Co, H-x, E-Co)) (From Lee et al. (2010). # 2010 IEEE,
Rather than cutting a second slot in the patch, a triple-band antenna can also be
realized by placing an open circuit stub on the radiating edge of a dual-band U-slot
rectangular patch antenna. Quadruple-band response results if two stubs are used. An
example is shown in Fig. 23a (ground plane not shown). The substrate is glass epoxy
(er = 4.3, h = 0.159 cm, and tanδ = 0.02). The measured VSWR of the triple-band
antenna is shown in Fig. 24. The measured resonant frequencies agree well with the
simulated values of 770, 952, and 1,100 MHz and corresponding bandwidths 16, 18,
Fig. 23 Geometries of the stub-loaded U-slot patch antenna. (a) Triple band; (b) quadruple band
(From Deshmukh and Ray (2010). # 2010 IEEE, Reprinted with permission)
600
VSWR
0.333 0.333
1200
2
1
6000 Frequency (MHz) 1200
Fig. 24 VSWR response of the triple-band antenna in Fig. 23a (From Deshmukh and Ray (2010).
# 2010 IEEE, Reprinted with permission)
and 22 MHz. Finally, quadruple-band response can be obtained with a stub-loaded

rectangular patch with two unequal-length half U-slots cut into it, as illustrated in
Fig. 23b.
Small Frequency Ratios

A simple approach to designing dual- and triple-band patch antenna with frequency
ratios less than about 1.5 was introduced in Lee et al. (2008, 2011). The method
starts with a broadband antenna. When a suitably located U-slot of the appropriate
dimensions is cut in the patch, it introduces a band notch at a frequency within the
original broadband. The broadband antenna is turned into a dual-band antenna.
When a second U-slot is cut in the patch, two band notches can be introduced and
a triple-band antenna results.
The original broadband antennas can be an L-probe-fed patch, an M-probe-fed
patch, coax-fed stacked patches, aperture-coupled stacked patches (Lee et al. 2011),
or a U-slot patch (Mok et al. 2013). In what follows, the use of an L-probe-fed patch
is presented in detail.
Figure 25a shows the wideband L-probe-fed rectangular patch. When a U-slot is
cut in the patch (Fig. 25b), a notch is introduced within the matching band and a
dual-band antenna results. With two U-slots (Fig. 25c), two notches are introduced
and a triple-band antenna results.
Fig. 25 Geometry of three antennas: (a) broadband, (b) dual band, and (c) triple band
Table 9 Dimensions of the proposed antennas (unit: mm) (From Lee et al. (2008). # 2008 IEEE,
Antenna W L H HL LL Ua1 Ud1 Ux1 Uy1 Ua2 Ud2 Ux2 Uy2
a 22 18 5 3.5 8.5
b 22 18 5 3.5 8.5 0.8 2 7.5 10.8
c 22 18 5 3.5 8.5 0.8 2 7.5 10.8 0.8 3.5 4.5 10.8
In Lee et al. (2008), these antennas with the dimensions shown in Table 9 were
fabricated. Air substrate was used. Simulation results, using IE3D software, were
obtained, as well as measured results. The simulated and measured return losses are
shown in Fig. 26, and the impedance bandwidths (S11 < 10 dB) are summarized in
Table 10. Reasonable agreement was obtained between the simulated and measured
results. The discrepancies between the results were due to the fabrication tolerance of
the fabricated antenna prototypes. Figures 27, 28, and 29 show the measured
radiation patterns for the three antennas. The patterns are broadside and the polar-
ization is linear.
For antenna b, the U-slot introduced a notch at around 5.5 GHz. The relationship
between the band notch frequency and the total U-slot length is shown in Table 11.
The total length of the slot is approximately half wavelength for the two proposed
U-slot(s) antennas. The resulting dual-band antenna has a frequency ratio f2/f1 = 1.30,
where f2 and f1 are the center frequencies of the upper and lower bands, respectively.
For antenna c, the two U-slots introduced a notch at around 5.5 GHz and another notch
at around 6.2 GHz. The resulting triple-band antenna has frequency ratios f3/f1 = 1.34
and f2/f1 = 1.16, where f3, f2, and f1 are the center frequencies of the upper, middle,
and lower frequencies. It should be noted that the frequency ratios are determined by
the bandwidth of antenna a, which was 30 % for this case. If one starts with an antenna
with broader bandwidth, e.g., 40 %, the frequency ratios of the dual- and triple-band
antenna will be larger.
A patch antenna with quadruple-band characteristics can be realized by cutting
three half-wavelength long U-slots on the patch. Another method is to use stacked
patches in multiple layers (Anguera et al. 2003).
Fig. 26 Simulated and measured reflection coefficient (S11) of the three antennas: (a) broadband,
(b) dual band, and (c) triple band
Table 10 Simulated and measured impedance bandwidths (From Lee et al. (2008) # 2008 IEEE,
Antenna Simulation (GHz) Measurement (GHz)
a 5.26–7.25 (31.8 %) 5.00–6.80 (30.5 %)
b 4.97–5.22 (4.9 %), 5.94–7.26 (20 %) 4.80–5.18 (7.6 %), 5.80–6.80(15.9 %)
c 4.95–5.20 (4.9 %), 5.74–6.00 (4.4 %), 4.80–5.18 (7.6 %), 5.63–5.95 (5.5 %),
6.41–7.24 (12.2 %) 6.25–6.83 (8.9 %)
Methods of Reducing the Patch Size
General Remarks
In many applications, it is desirable for the dimensions of the patch to be a small

fraction of the free-space wavelength. The resonant length of the microstrip patch
antenna is approximately λ/2, where λ is the wavelength in the dielectric substrate. It
follows that the size of the patch can be reduced by using a substrate material with
high dielectric constant. However, the resulting patch antenna will have narrow
impedance bandwidth. This motivated the search for other size reduction methods.
0
330 30
300 60
270 90
–40 –30 –20 –10
240 120
210 150
180
Fig. 27 Measured radiation patterns of the broadband antenna of Table 9a at 5.9 GHz
(( H-Co, H-x, E-Co, E-x)) (From Lee et al. (2008). # 2008 IEEE,
0 0
330 30 330 30
300 60 300 60
270 90 270 90
–40 –30 –20 –10 –40 –30 –20 –10
240 120 240 120
210 150 210 150

180 180
Fig. 28 Measured radiation patterns of the dual-band antenna of Table 9b: (a) 5.0 GHz and (b)
6.3 GHz (( H-Co, H-x, E-Co, E-x )) (From Lee et al. (2008). # 2008
IEEE, Reprinted with permission)
By placing a shorting wall along the null in the electric field across the center of
the patch, the resonant length can be reduced by a factor of two (Pinhas and
Shtrikman 1988; Chair et al. 1999; Lee et al. 2000). The area occupied by the
patch will be reduced by a factor of four, if the aspect ratio is kept the same. Another
technique to reduce the resonant length is to add a shorting pin in close proximity to
a 0 b 0
330 30 330 30
300 60 300 60
270 90 270 –40 –30 –20 –10 90

–40 –30 –20 –10
240 120 240 120
210 150 210 150

180 180
0
c
330 30
300 60
270 –40 –30 –20 –10 90
240 120
210 150
180
Fig. 29 Measured radiation patterns of the triple-band antenna of Table 9c: (a) 5.0 GHz, (b) 5.8
GHz, and (c) 6.55 GHz (( H-Co, H-x, E-Co, E-x)) (From Lee
Table 11 Relationship between the band notch frequency and the total U-slot length (From Lee
Antenna Freq. of band notch Total U-slot length
b 5.5 GHz 27.5 mm (0.504 λo)
c 5.5 GHz 6.2 GHz 27.5 mm (0.504 λo) 24.5 mm (0.506 λo)
the feed (Waterhouse et al. 1998). The shorting pin is capacitively coupled to the
resonant circuit of the patch, effectively increasing the permittivity of the substrate. It
has been shown that a suitably placed shorting pin can reduce the resonant length of
a circular patch by a factor of three and the area of the patch by a factor of nine.
Broadbanding techniques such as stacked patches, U-slot patch, and L-probe feed
Fig. 30 Geometry of the shorted patch. (a) Top view; (b) side view
can be applied to obtain small-size wideband patch antennas (Shackelford

et al. 2003). All these methods result in radiation patterns with high cross polariza-
tion. This may not be a disadvantage in indoor mobile communication applications.
A low cross-polarization design is that of the folded patch, which, however, is
thicker and more difficult to fabricate (Luk et al. 1998b).
In what follows, the use of shorting wall, shorting pin, and the U-slot patch with
shorting pin is discussed in detail.
The Use of Shorting Wall: Quarter-Wave Patch
Chair et al. (1999) presented experimental results of the quarter-wave patch shown in
Fig. 30. The substrate between the patch and the ground plane is foam, with
thickness h and relative permittivity 1.08. The sides of the patch are
a = b = 3.06 cm long, with one side shorted. The patch is fed by a coaxial probe,
with the feed point at x = 0 and y = d, where d is the distance between the feed
point and the open edge and is adjusted for best match.
Resonant frequency, pattern, and bandwidth (VSWR < 2) measurements were
performed for several thicknesses h, from 2 to 7 mm. Resonant frequencies ranged
from 2.19 to 2.46 GHz, while the bandwidths ranged from 3.59 % for h = 2 mm
(0.017 λo) to 17.66 % for h = 7 mm (0.058 λo). The corresponding bandwidths for a
half-wave regular patch on the same thickness and the same width but double in
length were 1.57 % and 5.55 %, respectively. These results were confirmed quali-
tatively by simulation software. That the quarter-wave shorted patch has a wider
bandwidth than the half-wave patch is due to its smaller volume and therefore less
stored energy, leading to a smaller Q and larger bandwidth. This conclusion was
found to hold only for air/foam substrate and not for material substrate which
supports surface waves. In Lee et al. (2000), it was found that shorted patches on
substrates with er = 2.32 and er = 4.0 had smaller bandwidths than their half-wave
counterparts. This was attributed to the fact that the surface-wave loss was consid-
erably more in a half-wave patch than a quarter-wave patch, leading to a larger
bandwidth for the former.
Table 12 Measured gains of the shorted square patch of Fig. 31

h (mm) fo (GHz) Gain (dBi) at broadside Maximum gain and direction
3 2.06 2.5 2.5 at 0
5 2.17 2.5 3.5 at 30
7 2.46 0.2 2.2 at 45
Fig. 31 Geometry of
partially shorted patch
The measured patterns by Chair et al. (1999) showed large cross polarization in
the E-plane. They also showed that, depending on the thickness, the maximum
radiation could occur off broadside. The measured gains for h = 3, 5, and 7 mm
at the resonant frequencies in the broadside direction as well as in the maximum
direction are summarized in Table 12. It is seen that typical values of the maximum
gain were in the range 2–3.5 dBi. This is about half that of the regular half-wave
patch.
Maximum gain direction is measured from the perpendicular direction in the
E-plane (perpendicular to the shorting plane).
Partially Shorted Patch and Planar Inverted-F Antenna
Figure 31 shows the geometry in which the shorting wall, instead of extending fully
across the width of the patch a, has a width s, where s a.
It was shown in Hirasawa and Haneishi (1992) that the use of a partially shorted
wall had the effect of reducing the resonant frequency of the antenna. Lee
et al. (2000) showed that this was accomplished at the expense of bandwidth.
Their calculated results using IE3D simulation software for an antenna with
a = 3.8 cm, b = 2.5 cm, h = 3.2 cm, and er = 1.0 showed that as s/a decreased
from 1.0 to 0.1, the resonant frequency decreased from 2.69 to 1.61 GHz,
representing a 60 % reduction in frequency or size. However, the bandwidth was
reduced from 7.4 % for s/a = 1.0 to 3.7 % for s/a = 0.1.
The partially shorted patch in the form shown in Fig. 32 is known as the planar
inverted-F antenna (PIFA), because the side view looks like an inverted F. The width
of the shorting wall w is approximately 0.2 L1, while the dimensions of L1 and L2 are
on the order of 1/8 λo.
Fig. 32 Size reduction by using an inverted-F patch
a y b y
(xps1, yps1)
R
x x
W Top view
(xp, yp) (xp, yp)
(xps1, yps1)
L
z
θ
Shorting pin
er d Side view
Fig. 33 Circular (a) and rectangular (b) patches with shorting pin
The Use of Shorting Pin
Another technique for reducing the patch size, very similar to the inverted-F method,
is to use a shorting pin (Waterhouse et al. 1998). This is illustrated in Fig. 33.
The shorting pin causes the fields underneath the patch to bounce back and forth. The
field starts to radiate once the bouncing distance reaches half wavelength. As a result of
the multiple bounces, the physical size of the patch is reduced. Since the bounces are
non-unidirectional, the fields can radiate out from almost all edges of the patch, resulting
in high cross polarization. However, for certain applications such as cellular phone
communication in a multipath environment, high cross-polarized fields are not a concern.
−5
−10
−15
S11 (dB)
−20
−25
1 pin
−30 2 pins
3 pins
−35
1.6 1.7 1.8 1.9 2 2.1 2.2
Frequency (GHz)
Fig. 34 Simulated return loss of the miniature patch antenna with different numbers of
shorting pins
Fig. 35 Simulated radiation 0

0
pattern of the miniature patch
330 30
antenna with one shorting pin
at 1.9 GHz (10 dB/div) −10
−20
300 60
−30
−40
270 90
−50
Eφ (φ = 0°) Eφ (φ = 90°) Eφ (φ = 90°)
If the shorting pin is close to the feed, the resonant circuit of the patch is
capacitively coupled to the pin. This is equivalent to increasing the permittivity of
the substrate, which further contributes to reduction in frequency or size of the patch
(measured in wavelength).
The solid curve of Fig. 34 shows the simulation results for return loss for the
circular patch shown in Fig. 34a with xp = 6.2 mm, xps = 8.3 mm, and yp = yps = 0,
on a foam substrate with er = 1.07 and t = 10 mm. IE3D simulation software was
used. The radius of the patch is reduced by a factor of 3 and the area by a factor of
9 when compared to the case of no shorting pin. The thickness of the foam substrate
is 0.06 λo, and the impedance bandwidth is about 6.3 %. The simulated radiation
patterns at 1.9 GHz are shown in Fig. 35. As noted earlier, the cross polarization of
this type of antenna is very high. The simulation results are consistent with the
experiments reported in Waterhouse et al. (1998).
y
a y
b (xps1, yps1)
(xps1, yps1)
(xps2, yps2) R
R x x
(xp,yp) (xp,yp)
(xps2, yps2)
(xps3, yps3)
Fig. 36 Circular patch with (a) 2 and (b) 3 shorting pins (units in mm and not to scale) (a)
R = 13.2, xp = 4.95, xps1 = xps2 = 11.08, yps1 = 1.95, yps2 = 1.95; (b) R = 15.4, xp = 2.35,
xps1 = 13.3, xps2 = xps3 = 14.1, yps1 = 0, yps2 = 5.13, yps3 = 5.13
The bandwidth can be slightly improved by using multiple shorting pins. A circular
patch with two and three shorting pins is shown in Fig. 36. The reflection coefficient
(S11) for these cases is shown in the broken curves of Fig. 34. The impedance
bandwidths are 7.9 % for the patch with two pins and 10 % for the patch with three pins.
Wideband U-Slot Patch Antenna with Shorting Pin
The shorting pin technique can be applied to the wideband U-slot patch discussed in
section “Aperture-Coupled Patches.” An example of the U-slot patch with shorting
pin is shown in Fig. 37a (Shackelford et al. 2001). The U-slot patch has a length
L = 70 mm (0.2 λo) and a width W = 84 mm (0.24 λo) where λo is the free-space
wavelength at the center frequency 0.86 GHz of the U-slot patch. The patch is
supported by the shorting pin and the probe wire in air. The radii of the shorting pin
and the probe feed are 4.65 and 2 mm, respectively. The shorting pin and the probe
feed are located at the non-radiating edge of the U-slot patch, which is placed on a
1 1 λo ground plane. Figure 38 shows the measured and simulated SWR of this
antenna. The matching frequency range is from 0.75 to 0.97 GHz with an impedance
bandwidth of 25.6 %. The area of the U-slot patch is only 0.055 λo. Figure 39 shows
the simulated x-z and y-z plane radiation patterns at 0.85 GHz.
Studies of the U-slot patch with shorting pin on material substrates have been
reported in Shackelford et al. (2003). It was also possible to further reduce the size of
the patch by removing half of the structure along the plane of symmetry (Chair
et al. 2005). Such an antenna is shown in Fig. 38b and its VSWR response is also
shown in Fig. 39. The corresponding bandwidth is 20 %.
Discussion on Finite Ground Plane Size Effect
The papers from which the above examples were drawn, as well as most of the
papers in the literature on patch size reduction techniques, did not address the ground
Fig. 37 (a) U-slot patch with shorting pin; (b) half U-slot patch with shorting pin
plane size effects. Based on the recent study for the shorted patch on foam substrate
using CST Microwave Studio simulation software, it was found that, when com-
pared to the infinite ground case, the center resonant frequency, bandwidth, and gain
will deviate less than 5 % and 10 % and 1.5 dB, respectively, until the ground plane
is reduced to smaller than 0.5 0.5 λo (Tong et al. 2011).
In general, the ground plane size effects have to be considered with care, as they
are dependent on the particular antenna and the locations of the antenna and the feed
(Best 2009).
5 5
4.5 4.5
4 4
3.5 3.5
SWR
SWR
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.7 0.75 0.8 0.85 0.9 0.95 1 1.05
Frequency (GHz)
Fig. 38 Measured and simulated SWR of the U-slot patch antennas with shorting pin. Half
U-slot (measured). Half U-slot (simulated). Full U-slot (measured). Full U-slot
(simulated) (From Chair et al. (2005). # 2005 IEEE, Reprinted with permission)
a xz-plane yz-plane
0 0
0 0
330 30 330 30
–10 –10
300 –20 60 300 –20 60
–30 –30
270 –40 90 270 –40 90
0 0
b 0 0
330 30 330 30
–10 –10
300 –20 60 300 –20 60
–30 –30
270 –40 90 270 –40 90
Fig. 39 Simulated radiation patterns of the U-slot patch antenna with shorting pin at 0.9 GHz
(10 dB/div). (a) Half U-slot. (b) Full U-slot E θ Eφ
Designs for Circular Polarization
Basic Principles of Circularly Polarized Patch Antennas
The designs described in the previous sections are for linear polarization. There are
many circumstances, such as multipath/fading environments and in communication
with space vehicles above the earth’s ionosphere, in which it is more reliable to use
circular polarization. Circularly polarized patch antennas can be broadly classified
into three types – single feed, dual feed, and sequentially rotated feed. Figure 41
shows four single-feed designs.
Figure 40a and b shows an almost square patch and an almost circular (elliptical)
patch. Figure 40c and d shows a square patch with truncated corners and a circular
patch with indentations. Although coaxial feeds are shown in Fig. 40, the patches
can also be fed by a microstrip line or by coupling through an aperture.
At the appropriate frequency, all four configurations will generate two orthogonal
linearly polarized waves which are in phase quadrature in the broadside direction.
This design is simple, but suffers from very narrow axial ratio bandwidth for thin
substrates. By using the U-slot or L-probe for relatively thick substrates or the use of
stacked patch techniques, the axial ratio bandwidth can be increased to over 10 %.
The basic principle of the single-feed CP patch antenna is that a perturbation of the
dimensions be introduced such that, by feeding the patch at the appropriate location,
two modes with orthogonal polarizations are generated with resonant frequencies
which are slightly different. In the almost square patch and the almost circular patch,
the perturbation is in the form of one of the dimensions being slightly different from
the other. In the square patch with truncated corners and the circular patch with
indentations, the perturbation is in the form of two slightly different diagonal lengths.
In the broadside direction, the amplitudes and phases of the radiated electric fields
of the two orthogonal modes generated are illustrated in Fig. 41. The resonant
frequencies of these modes are fa and fb, respectively.
At the frequency fo, approximately midway between fa and fb, the amplitudes of the
two modes are equal and the phases differ by 90 , resulting in circular polarization. This
condition for perfect CP occurs only at fo and in the broadside direction. As the frequency
deviates from fo and/or the direction is off broadside, elliptical polarization results.
The antenna is only useful for the range of frequencies within which both the
return loss is less than 10 dB (Return Loss Bandwidth, RLBW) and the axial ratio is
less than 3 dB (Axial Ratio Bandwidth, ARBW). This overlapping region will be
referred to as the CP bandwidth.
Broadbanding of Single-Feed Circularly Polarized Patch Antennas
As with the case of the RLBW, the ARBW is found to increase as the substrate
thickness increases. However, this thickness is very limited if the ARBW is to
overlap with the RLBW. For the single-feed patch antennas of Fig. 40, achievable
Fig. 40 Single-feed circularly polarized patches. (a) Almost square patch. (b) Almost circular
(elliptical patch). (c) Square patch with truncated corners. (d) Circular patch with indentations
CP bandwidth is less than 1 %. For thicker substrates, the use of the U-slot patch, the
E-patch, stacked patches, as well as the L-probe feeding techniques is found to be
effective not only in increasing the RLBW and the ARBW but also to enable them to
overlap. The truncated square patch with U-slot, the CP modified U-slot patch, and
the CP modified E-patches will now be discussed in detail.
The Truncated Square Patch with U-Slot

A circularly polarized antenna can be designed using a square patch with truncated
corners. One of the requirements for the CP antenna is that the axial ratio bandwidth
Fig. 41 The amplitudes and

phases of two orthogonal
modes
(ARBW) should be within the return loss bandwidth (RLBW) or, equivalently, the
impedance bandwidth. This overlapping bandwidth is very narrow for the square
patch with truncated corners. Typically, it is about 0.8 %, for an air-substrate
thickness of around 0.02 λo. Yang et al. (2008) have shown that, while increasing
the substrate thickness can increase axial ratio bandwidth and the impedance band-
width individually, this increase is very limited if the ARBW is to be within the
RLBW. On the other hand, by using a U-slot patch with truncated corners and using
substrates of thickness around 0.1 λo, the axial ratio bandwidth can be as wide as
6.1 % and still stay within the impedance bandwidth. The antenna geometry studied
by Yang et al. (2008) is shown in Fig. 42.
The patch has dimensions 28.6 mm (W) 28.6 mm (L) and the ground plane is
100 100 mm. It is fed by a coaxial probe and the substrate is air. Using the IE3D
simulation software, the RLBW and ARBW are studied for five thicknesses. For
each case, the truncation parameter a and the feed position d are adjusted for best
performance and are shown in Table 13. The simulation results are shown in
Table 14.
It is seen that, as the substrate thickness increases from 1 mm (0.016 λo) to 3 mm
(0.046 λo), the RLBW increases from 4.0 % to 7.4 % while the ARBW increases
from 0.82 % to 3.1 %. When the thickness increases further to 4 mm (0.06 λo), the
Fig. 42 Geometry of single W

probe feed square patch
antenna with truncated
corners
y
x d a
z
x H
Table 13 Parameters of Case Substrate thickness (H) a d

antenna as shown in
1 1 mm 0.016 λo 3.3 8.2
Fig. 43. Units of a and d are
in mm 2 1.5 mm 0.024 λo 4.5 7.2
3 2 mm 0.032 λo 4.9 5.5
4 3 mm 0.046 λo 5.9 5.1
5 4 mm 0.06 λo 6.9 4.3
Table 14 Simulation results of the different cases as shown in Table 13

Simulation (GHz)
Case Substrate thickness (H) RLBW ARBW
1 1 mm 0.016 λo 4.86–5.06 (4.0 %) 4.90–4.94 (0.82 %)
2 1.5 mm 0.024 λo 4.88–5.08 (4.0 %) 4.79–4.86 (1.45 %)
3 2 mm 0.032 λo 4.85–5.11 (5.2 %) 4.68–4.78 (2.1 %)
4 3 mm 0.046 λo 4.81–5.18 (7.4 %) 4.52–4.66 (3.1 %)
5 4 mm 0.06 λo 4.89–5.09 (4.0 %) 4.40–4.57 (3.8 %)
RLBW drops to 4.0 % while the ARBW continues to increase to 3.8 %. The
important point to note is that, beginning with case 2, the two frequency bands do
not overlap. Thus, only in case 1 the ARBW lies within the RLBW, and the useful
frequency band is determined by the ARBW, from 4.90 to 4.94 GHz, corresponding
to 0.82 %.
Figure 43 shows the geometry when a U-slot is added to the square patch with
truncated corners. Three thicknesses are studied. The dimensions of the antenna are
shown in Table 15, and the simulation and measurement results are shown in
Table 16.
Fig. 43 Geometry of the W

single-feed U-slot square
patch antenna with truncated
corners Ux Ud
L Uy
y
Ua d
x a
z
x
H
Table 15 Parameters of the antenna as shown in Fig. 44. Units for a, d, Ua, Ud, Ux, and Uy are
in mm
Case Substrate thickness (H) a d Ua Ud Ux Uy
6 4 mm 0.05 λo 5.7 12.6 1 9.8 12 14
7 6 mm 0.08 λo 7.7 9.6 1 9.8 12 14
8 7.5 mm 0.1 λo 8.2 5.6 1 9.8 11 14
Table 16 Results of the different cases as shown in Table 15

Simulation (GHz) Measurement (GHz)
Case Substrate thickness (H) RLBW ARBW RLBW ARBW
6 4 mm 0.05 λo 3.83–4.18 3.96–4.05 – –
(8.7 %) (2.2 %)
7 6 mm 0.08 λo 3.73–4.2 3.96–4.12 3.66–4.16 3.91–4.12
(11.9 %) (4.0 %) (12.8 %) (5.23 %)
8 7.5 mm 0.1 λo 3.84–4.08 3.84–4.09 3.88–4.08 3.82–4.05
(6.1 %) (6.3 %) (5.0 %) (5.84 %)
The following is observed:
1. When the U-slot is added on the patch, the resonant frequency is lowered
compared to the case without the U-slot.
2. By adjusting the dimensions of the U-slot, the RL and AR bandwidths can be
tuned to overlap.
3. The measured overlapping bandwidth is 5.23 % for H = 6 mm (0.08 λo). This is
to be compared to 0.82 % achievable for the case of no U-slot. This is achieved, of
course, at the expense of increasing the thickness of the antenna.
a 0 10
−5 9
−10 8
−15 7
Axial ratio (dB)

−20 6
S11 (dB)
−25 5
−30 4
−35 RL (Meas.) 3
RL (Sim.)
−40 2
AR (Meas.)
−45 1
AR (Sim.)
−50 0
3.5 3.75 4 4.25 4.5 4.75 5
Frequency (GHz)
b 0
330 30
300 60 c
LHCP (Meas.)
LHCP (Sim.)
270 90
–40 –30 –20 –10 RHCP (Meas.)
RHCP (Sim.)
240 120
210 150
180
Fig. 44 Simulated and measured performance of case 7, radiation patterns are at 3.95 GHz
The simulation and measurement results for S11, axial ratio, and radiation patterns
at one frequency for the cases of H = 6 mm and H = 7.5 mm are shown in Figs. 44
and 45.
The CP Modified U-Slot Patch and the CP Modified E-Patch

Instead of truncating the corners of a square patch, circular polarization can also be
generated by the use of an asymmetric U-slot (unequal arms) etched on a square
patch (Tong and Wong 2007) or a modified E-patch in which the two parallel slots
are unequal (Khidre et al. 2010). The geometries of these are shown in Figs. 46 and
47, respectively. By using relatively thick substrates (0.08 λo), the CP bandwidths
achievable are comparable to the truncated U-slot square patch.
a 0 10
−5 9
−10 8
−15 7
Axial ratio (dB)

−20 6
S11 (dB)
−25 5
−30 4
−35 RL (Meas.) 3
−40 RL (Sim.) 2
AR (Meas.)
−45 1
AR (Sim.)
−50 0
3.5 3.75 4 4.25 4.5 4.75 5
Frequency (GHz)
0
b 330 30
300 60 c
LHCP (Meas.)
LHCP (Sim.)
270 90
–40 –30 –20 –10 RHCP (Meas.)
RHCP (Sim.)
240 120
210 150
180
Fig. 45 Simulated and measured performance of case 8, radiation patterns are at 4 GHz
In Khidre et al. (2010), a study was made of the three designs using both air
substrate and a substrate with er = 2.2.
The optimized dimensions of the designs to work at IEEE 802.1 lb/g band
(2.4–2.5 GHz) are tabulated in Tables 17, 18, and 19. The simulated S11 and axial
ratios of the three air-substrate antennas are shown in Figs. 48 and 49, respectively.
From the results, the modified E-patch exhibits 10.1 % impedance bandwidth
(2.35–2.6 GHz), while unequal arms U-slot has 9.25 % (2.37–2.6 GHz) and trun-
cated corner U-slot has 9.65 % (2.27–2.5 GHz). It is clear that all ARBW are aligned
with the corresponding RLBW, resulting in CP bandwidths of 6.5 % (2.38–2.54
GHz) for the modified E-patch, 4 % (2.39–2.49 GHz) for the unequal arms U-slot,
and 4.5 % (2.39–2.5 GHz) for the truncated corner U-slot.
Fig. 46 Geometry of CP Wgd

unequal arms U-slot patch
antenna
Wp
Ws
Lf
L gd
Lp Lul
L ur
Wu Lub
SMA connector hgd
Fig. 47 Geometry of the CP L

modified E-patch antenna
Ls1
Ws
F
W P
x
Ls2
The simulated S11 and axial ratios of the material substrate antennas are shown in
Figs. 50 and 51, respectively. From the results, the modified E-patch exhibits 10.6 %
impedance bandwidth (2.39–2.66 GHz), while unequal arms U-slot has 6.9 %
(2.38–2.65 GHz) and truncated corner U-slot has 13.1 % (2.41–2.75 GHz). Again,
Table 17 Dimensions of CP modified E-shaped patch in mm (From Khidre et al. (2010). # 2010
εr H W L Ws Ls1 Ls2 P F
1 10 77 47.5 7 19 44.5 14 17
2.2 6.7 63 33.5 4 27 6 20 10
Table 18 Dimensions of CP modified U-slot patch in mm (From Khidre et al. (2010). # 2010
εr H Lp Wp Lul Lur Lub Wu Ws Lf
1 10 43.7 43.7 27.3 19.8 10.3 16.9 2.3 12.5
2.2 6.7 32.7 32.7 20 13.5 8.9 11.4 1.5 12
Table 19 Dimensions of CP modified U-slot patch in mm (From Khidre et al. (2010). # 2010
εr H L W a d Ua Ud Ux Uy
1 10 48.2 48.2 12.2 10 2.1 19.8 23 19
2.2 6.7 36 36 7.7 6.5 1.2 19.6 15.5 11.5
Fig. 48 S11 for E-shaped, 0

unequal arms U-slot and
truncated corner U-slot patch −5
antennas. Air substrate (From
Khidre et al. (2010). # 2011
IEEE, Reprinted with −10
permission)
S11(dB)
−15
−20
E-patch
−25 Truncated commer U-slot
Unequal U-slot
−30
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
F(GHz)
all AR bandwidths are aligned with the corresponding S11 bandwidth, resulting in
CP bandwidths of 3.6 % (2.41–2.5 GHz) for the modified E-patch, 2.8 % (2.36–2.55
GHz) for the unequal arms U-slot, and 3.3 % (2.39–2.5 GHz) for the truncated corner
U-slot.
The CP characteristics of the three designs were verified experimentally in Yang
et al. (2008), Tong and Wong (2007), and Khidre et al. (2010).
Fig. 49 Axial ratio for 5

E-shaped, unequal arms
U-slot, and truncated corner
U-slot patch antennas. Air 4
substrate (From Khidre
et al. (2010). # 2011 IEEE,
Axial Ratio(dB)
Reprinted with permission) 3
1 E-patch
Truncated corner U-slot
Unequal U-slot
0
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
F(GHz)
Fig. 50 S11 for E-shaped, 0

E- patch
unequal arms U-slot, and
truncated corner U-slot Unequal U-slot
antennas. er = 2.2 (From −5
Khidre et al. (2010). # 2011
IEEE, Reprinted with
−10
S11(dB)
permission)
−15
−20
−25
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
F(GHz)
Designs for Wider CP Bandwidth

In Yang et al. (2008), a truncated square patch fed by an L-probe with an air substrate
attained a usable bandwidth of 16 %. In Chung and Mohan (2003), truncated corner
stacked square patches achieved a usable bandwidth of 17 %. Both antennas are
relatively thick.
For thin substrates (<0.03 λ0), good CP bandwidth can be obtained by the use of
dual feeds to provide the 90 phase difference for the two orthogonal modes, as
illustrated in Fig. 52. This feeding method can attain about 10 % axial ratio
bandwidth while keeping the substrate thin, although at the expense of more
complexity in the feed circuitry (Hall and Dahele 1997).
The dual-feed design can be considered a special case of multi-feed design. It was
shown that sequentially feeding a patch at suitable multiple locations with
Fig. 51 Axial ratio for 5

E-shaped, unequal arms
U-slot, and truncated corner
U-slot antennas. er = 2.2 4
(From Khidre et al. (2010).
# 2011 IEEE, Reprinted with
Axial Ratio(dB)
permission) 3
2
E- patch
1 Unequal U-slot
0
2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
F(GHz)
Fig. 52 Dual-feed circularly

polarized patch antenna patch
power divider
quarter-wave
feed line transformer
input
appropriate phasing leads to the cancelation of the cross circularly polarized com-
ponents, resulting in improved axial ratio bandwidth (Hall and Dahele 1997).
As in the case of linear polarization, it is desirable in many situations to reduce the
size of CP patch antennas, and a number of methods have been developed. Because
of page limitations, they are not described here. The reader is referred to the article
on “Small Antennas in Wireless Communications” (Wong et al. 2012).
Frequency-Agile and Polarization-Agile Patch Antennas
In some applications, it is desirable to be able to tune the operating frequency of the

patch antenna without having to fabricate a new antenna. In other applications, the
ability to alter the polarization of the antenna will enable more efficient use of the
allocated frequencies. The methods for designing frequency-agile patch antennas
and polarization-agile patch antennas include tuning using varactor diodes, shorting
Fig. 53 Illustrating the use of feed

varactor diodes for tuning
εr Varactor
diodes
Fig. 54 Resonant frequency 1.9

versus bias voltage for a alpha DVM 6733-168-001 varactors
varactor-loaded rectangular
patch antenna 1.8
1.7
f r (GHz)
1.7cm
1.6
3cm
1.5 4.65cm
1.4
0 5 10 15 20 25 30
bias voltage (v)
pins, adjustable air gap, patch antenna with switchable slots, and U-slot and
E-patches with switchable slots. Several of these will be described in this section.
Frequency-Agile Microstrip Patch Antennas
Tuning Using Varactor Diodes

For a given set of patch dimensions, the resonant frequency is primarily governed by
the value of the relative permittivity er of the substrate. If some means is available to
alter er, the resonant frequency will change. One method of achieving this is to
introduce varactor diodes between the patch and the ground plane, as shown in
Fig. 53. The diodes are provided with a bias voltage, which controls the varactor
capacitance and hence the effective permittivity of the substrate. Bhartia and Bahl
(1982) performed an experiment on this method and the results are shown in Fig. 54.
The resonant frequency ( fr) of the lowest mode of the rectangular patch increases
with the bias voltage, which changes the diode capacitance. It is seen that, in this
experiment, a tuning range of some 20 % was achieved with a 10 V bias. The range
increased to about 30 % with a 30 V bias. Note that the curve of frequency versus
bias voltage is not a linear one. Similar results were obtained in Waterhouse and
Shuley (1994) and Kalialakis and Hall (2007).
Fig. 55 Resonant frequency

versus separation of posts for 1.8
6.2 9.0 cm rectangular IE3D
patch antenna with εr = 2.55,
t = 1.6 mm 1.7
1.5cm
feed
Ir (GHz)
b=9.0cm
1.6 Posts
calculated S
1.5 a= 6.2cm
measured e = 2.55
t = 1.6mm
1.4
0 0.25 0.50 0.75 1.00
S/o
Tuning Using Shorting Posts (Pins)

The value of er can also be changed by introducing shorting posts (pins) at various
points between the patch and the ground plane. These shorting posts present an
inductance and therefore affect the effective permittivity of the substrate. The
method appeared to be first introduced by Schaubert et al. (1981), who performed
an experiment on a rectangular patch using two pins, as shown in Fig. 55.
The solid curve in Fig. 55 was measured results, which shows that the resonant
frequency is dependent on the separation of the two posts. A tuning range of some
18 % is obtained as the separation varies between 0 mm and the whole width of the
patch.
It is not possible to analyze the rectangular patch with shorting pins using the
cavity model. Full-wave analysis is necessary and a number of commercial software
can be used to obtain simulation results.
The results using the Zeland IE3D simulation software are shown in the dashed
curve in Fig. 55 for resonant frequency and in Fig. 56 for S11.
Tuning by Means of an Adjustable Air Gap

By introducing an air gap between the substrate and the ground plane in a microstrip
patch antenna, the effective permittivity of the cavity will change. This can be used
to tune the resonant frequency (Dahele and Lee 1985).
The geometry of a microstrip patch antenna with an air gap is shown in Fig. 57.
The feed is not shown and will be discussed shortly.
Consider the cavity under the conducting patch. It is made of two layers: a
substrate of thickness t and an air region of thickness Δ. Compared to the case
with no air gap, the effective permittivity of the cavity is smaller. As a result, the
resonant frequencies of the various modes will increase. Since the effective permit-
tivity decreases as Δ increases, tending toward the free-space value of unity as
Fig. 56 Simulated reflection 0

coefficient (S11) versus
frequency for six values of pin
−5
spacing
−10
S11 (dB)
−15 s/a = 0.2
s/a = 0.4
s/a = 0.5
−20 s/a = 0.6
s/a = 0.8
s/a = 0.9
−25
1.5 1.55 1.6 1.65 1.7 1.75 1.8
Frequency (GHz)
Fig. 57 Geometry of a z Conducting patch

microstrip patch antenna with Substrate
an air gap
et t
Spacer Air gap Δ
Conducting
plane
Fig. 58 A capacitor
consisting of two dielectric
layers
Δ ! 1, it follows that the resonant frequencies can be tuned by adjusting the air gap
with Δ. As a by-product, the bandwidth will also increase partly due to the increase
in the height of the dielectric medium and partly because the effective permittivity is
smaller. It is also to be noted that substrate and etching tolerances can be compen-
sated by adjusting Δ.
A heuristic derivation of the effective permittivity and the resonant frequency of
the two-layer cavity can be obtained by considering the capacitance of a capacitor
with two dielectric layers, as shown in Fig. 58.
The overall capacitance is the two capacitors in series:
e1 e2 A eeff A
C¼ ¼ (10)
e1 d 2 þ e2 d 1 d 1 þ d 2
where
e1 e2 ðd1 þ d2 Þ
eeff ¼ : (11)
e1 d 2 þ e2 d 1
Applying this result to the geometry of Fig. 58,
ee0 ðt þ ΔÞ eðt þ Δ Þ
eeff ¼ ¼ (12)
er e0 Δ þ e0 t ðt þ Δer Þ
Since er > 1,
eeff < e when Δ > 0
eeff ¼ e when Δ ¼ 0
If the effective permittivity is used in the formula for the resonant frequency, we
obtain, for the TMnm mode,
rffiffiffiffiffiffiffi
e
f nm ðΔÞ ¼ f ð0Þ (13)
eeff
where fnm(0) is the resonant frequency when there is no air gap.

The formula for fnm applies to any patch shape. As the air gap width increases, eeff
decreases and the resonant frequency increases. The dependence of fnm on Δ,
however, is not a linear one.
Rectangular, circular, and annular-ring patches with air gaps fed coaxially had
been studied by Lee and coworkers by means of the cavity model as well by full-
wave analysis (Dahele and Lee 1985). The results were validated by an experiment.
Typical tuning range is from 0 % to 20 %.
The disadvantage of applying the adjustable air gap method to coaxially fed
patches is that, every time the air gap width is changed, the coaxial probe has to be
de-soldered and resoldered to the patch. This disadvantage can be avoided if the
patch is fed using coupling through an aperture or stripline, which was demonstrated
in Mao et al. (2011).
Discussion
Compared to the methods of using varactor diodes, or shorting pins, or switchable slot
(Yang and Rahmat-Samii 2001b), the air gap method has the following advantages:
1. It can be applied to patches of any shape.

2. There is no need to add components or etch a slot in the patch.
3. The method is particularly attractive for an array consisting of many elements. If
the elements are fed by apertures or strip lines, the resonant frequencies of all the
elements, and therefore of the array, can be tuned by a single adjustment of the air
gap width.
The disadvantage is that the air gap has to be changed mechanically. However, a
MEMS-based electrostatically tunable circular patch antenna has been developed
(Jackson and Ramadoss 2007).
Polarization-Agile E-Shaped Patch Antenna
Polarization-agile antennas are antennas whose polarization characteristics can be

altered. They are also called polarization reconfigurable antennas and have received
increasing attention in the last decade. Polarization agility is a desirable feature for
wireless applications because it can double the system capacity through frequency
reuse with two orthogonal polarizations without the need of integrating several
individual antennas, thus rendering the wireless devices lighter and more compact.
Reconfigurable circularly polarized antennas also provide a powerful modulation
scheme for microwave tagging systems. They also have potential applications in
multiple-input multiple-output (MIMO) systems. Polarization switching could be
accomplished with the aid of switching devices such as PIN diodes and RF-MEMs.
Using switchable slots, Yang and Rahmat-Samii (2002) demonstrated a
reconfigurable patch antenna for circular polarization diversity. The CP bandwidth
obtained was 3 %. By using PIN diodes in U-slot patches, Qin et al. (2010) built two
antenna prototypes; the first enables switching between linear and circular polariza-
tion, and the second switches between the two circular polarizations, with a CP
bandwidth of 2.8 %. Khidre et al. (2013) designed and built CP polarization-agile
patch antennas incorporating PIN diodes in an E-shaped patch and obtained CP
bandwidth of 7 %. The operation principle, design, and performance of the
polarization-agile antenna studied by Khidre et al. (2013) are presented below.
Operation Principle
When an E-shaped patch antenna is loaded with two switches on the slots as shown
in Fig. 59, the antenna will have four possible states listed in Table 20. At state 1, the
resonant frequency of the original E-shaped patch would be fL and radiate linear
polarized (LP) fields (Yang et al. 2001a). At state 2, the two switches are ON which
allows the surface currents to pass through them, making the current path around the
Fig. 59 Switching states of the reconfigurable E-shaped patch antenna: (a) LP state 1; (b) LP state
2; (c) LHCP state 3; and (d) RHCP state 4
Table 20 Antenna State Switch 1 Switch 2 Frequency Polarization

possible configurations
1 OFF OFF fL LP
# 2013 IEEE, Reprinted 2 ON ON fH LP
with permission) 3 ON OFF f LHCP
4 ON OFF f RHCP
Fig. 60 Current distribution on the E-shaped patch antenna in state 3 during one-time cycle
T (From Khidre et al. (2013). # 2013 IEEE, Reprinted with permission)
slots become shorter. Thus, the electrical length is smaller and the resonant fre-
quency fH would be higher ( fH > fL). Due to the symmetry of the antenna structure,
at state 2, the polarization would also be linear. Consequently, going from state 1 to
state 2 or vice versa is a frequency reconfigurable E-shaped patch antenna with linear
polarization.
Recently, it has been shown that CP fields could be excited by E-shaped patch
antenna with a wide effective bandwidth (9.27 %) (Khidre et al. 2010). Letting one
of its slots shorter than the other would introduce asymmetry and perturb the field
beneath the patch, thus producing CP fields. Therefore, an E-shaped patch antenna at
state 3 effectively has the upper slot shorter than the lower one, and at an appropriate
switch position on the upper slot, LHCP could be attained at resonant frequency
f ( fL < f < fH). In a similar fashion, state 4 gives RHCP. According to the above
mechanism, wideband polarization reconfigurable E-shaped patch antenna is real-
ized. The current distribution of an E-shaped patch antenna in state 3 is shown in
Fig. 60, where the current rotates clockwise which implies LHCP.
Antenna Design
The reconfigurable E-shaped patch antenna geometry with labels is shown in
Fig. 61. It is printed on a thin sheet of RT/duroid 5880 dielectric substrate with
Fig. 61 Geometry of a Wsub

single-feed reconfigurable a
E-shaped patch antenna
with integrated DC biasing L
circuit: (a) top view; (b)
side view: Lsub =140 mm, 2
Wsub = 80 mm, L = 43 mm, S Y
W = 77 mm, Ls = 30 mm, Ws D1
Ws = 7 mm, P = 17 mm,
Yf = 14 mm, Lst =28 mm, 1
Wst = 0.3 mm, S = 0.5 mm, W Yf P Lsub
h = 10 mm, t = 0.787 mm
D2 x
permission)
Ls
Lst y
Wst
b t
h
z
DC
y
RF/DC RF
capacitor PEC substrate (εr = 2.2)
er = 2.2 and thickness t = 0.787 mm. The substrate is mounted 10 mm above the
ground plane by a coaxial single-feed pin. Although the antenna is built over air
substrate for wide bandwidth, the thin dielectric substrate facilitates mounting the
microwave components. The dimensions of the antenna are obtained through the
procedure described in Khidre et al. (2010).
Two RF PIN diodes are inserted on the slots of the E-shaped patch antenna. A
narrow slit of width s = 0.5 mm is incorporated on the E-shaped patch surface to
avoid DC short across the diodes’ terminals. Therefore, the E-shaped patch is
divided into two parts: inner and outer one. Three DC block capacitors are inserted
on the narrow slit to maintain RF continuity. The outer part of the E-shaped patch is
DC grounded through a narrow shorted λ/4 transmission line, and the short circuit is
realized by a vertical via as shown in Fig. 62b. This ensures high impedance at the
edge of the E-shaped patch and hence keeps the RF current unperturbed on the
E-shaped patch surface. The ground plane is common for both the DC and the RF
Fig. 62 Simulated and 0

measured S11 for the proposed
antenna in states 3 and
4 (From Khidre et al. (2013). −10
permission) −20
S11(dB)
−30
HFSS state 3,4
−40 measured state 3
measured state 4
−50
2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
f (GHz)

measured axial ratio for the 9 HFSS
proposed antenna in states
8 measured state 3
3 and 4 (From Khidre
measured state 4
Axial Ratio(dB)
et al. (2013). # 2013 IEEE, 7

Reprinted with permission) 6
5
4
3
2
1
0
2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
f (GHz)
signals. Its size is 200 100 mm. The RF and DC control signals are supplied to the
antenna via the coaxial pin feed and are superimposed together through a tee bias for
RF/DC isolation. Thus, the D2 diode would be ON and the D1 diode would be OFF
when the inner part of the E-shaped patch is positively charged, resulting in RHCP.
For LHCP, the terminals of the DC source have to be reversed; hence, the D1 diode
would be ON and the D2 diode would be OFF.
Antenna Performance
The simulated versus measured results of S11 and axial ratio in both switching states
3 and 4 are shown in Figs. 62 and 63, respectively. Good agreement between
simulated and measured results is observed. Due to the structure symmetry, S11
Table 21 Antenna bandwidth simulated versus measured results at states 3 and 4 (From Khidre
Parameter Simulation Measured
S11 (<–10 dB) 2.39–2.6 GHz (8.4 %) 2.4–2.575 GHz (7 %)
Axial ratio(<3 dB) 2.4–2.6 GHz (8 %) 2.38–2.6 GHz (8.8 %)
Fig. 64 Simulated and measured radiation pattern of the proposed antenna at 2.45 GHz: (a) x-z
plane at state 3; (b) y-z plane at state 3; (c) x-z plane at state 4; and (d) y-z plane at state 4 (From
Khidre et al. (2013). # 2013 IEEE, Reprinted with permission)
and axial ratio are maintained along switching between the two states, which is an
advantage of this design. This can be observed from Fig. 62, where S11 in state 3 is
almost the same as in state 4. Similarly, in Fig. 63 the axial ratio is kept the same in
both switching states. The achieved impedance and axial ratio bandwidths calculated
from simulation in comparison to measurements are tabulated in Table 21. From
Table 21, the measured effective overlapped bandwidth (10 dB S11 and 3 dB axial
ratio) is 7 % (2.4–2.575 GHz), which is wideband for single-feed polarization-agile
microstrip antenna designs.
Antenna radiation pattern at 2.45 GHz is shown in Fig. 64, where good agreement
between the simulated and measured results is observed. The radiation pattern in y-z
plane remains the same after switching, whereas in x-z plane, the radiation pattern is
mirrored due to the symmetry of the structure. Experimental results along with the
simulated one for the antenna gain versus frequency are shown in Fig. 65. Again,
because of the symmetry, the broadside gain is the same in both modes of the

measured gain of the proposed
antenna in either state 3 or 7
state 4 (From Khidre
et al. (2013). # 2013 IEEE, 5
gain (dB)
3
1 HFSS
−1 measured
−3
−5
2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
f (GHz)
antenna. To avoid redundancy, only RHCP mode is shown. Maximum realized gain
is 8.7 dBi at 2.45 GHz according to both simulation and measurement, whereas the
3 dB gain bandwidth is 2.31–2.63 GHz (13 %) from measurement which covers the
S11 and axial ratio bandwidths shown in Table 21.
Conclusion
This chapter began with descriptions of the basic geometry of the microstrip patch
antenna, feeding methods, operation principles, and performance characteristics.
The chapter then proceeded to present its development into a sophisticated and
versatile antenna which is used in numerous military, space, commercial, and
medical applications.
While the advances of the last three decades have been impressive, a number of
challenges remain. For example, in size reduction, most of the methods developed
focused on reducing the patch size, while useful guidelines on the effects of ground
plane size on antenna characteristics have not been adequately developed. In
broadbanding, the efficient methods developed all result in increased volume,
which somewhat mitigate the low-profile advantage of microstrip patch antennas.
The search for wideband, efficient, and low-profile patch antennas remains an area of
active research.
Other than the basic geometry, there is a lack of design guides for most of the
structures discussed in this chapter, such as the wideband U-slot patch, the L-probe-
fed patch, or the patch with shorting pin. Antenna designers have to rely on trial and
error using a simulation software. It will be very helpful to develop useful formulas/
guidelines/procedures which enable the designer to arrive at the first cut parameters,
without relying on trial and error.
Fig. 66 The picture of a shuttle imaging radar antenna consisting of an array of microstrip patch
antennas, (a) in the laboratory and (b) during flight (Courtesy of Jet Propulsion Laboratory, NASA)
Fig. 67 An array of
microstrip antennas are made
conformal to a wing shape
surface (Courtesy of Air Force
Research Lab./Antenna
Technology Branch, Hanscom
AFB, USA)
Fig. 68 A patch antenna in a

cellular phone
It is instructive to end this chapter by showing the photos and a sketch of several
microstrip patch antennas used in space, military, commercial, and medical
applications.
Figure 66 shows an imaging radar antenna used in the Space Shuttle, consisting of
an array of microstrip patch antennas, in the laboratory (a) and during flight (b).
Figure 67 is a picture of an array of microstrip antennas which are made conformal to
a wing shape surface such as the fuselage of an airplane or rocket. Figure 68 shows a
Fig. 69 A base station

antenna array consisting
of L-probe-fed patches
(Courtesy of the Department
of Electronic Engineering,
City University of
Hong Kong)
Fig. 70 A GPS antenna in

the form of a U-slot square
patch with truncated corners
(Courtesy of the Department
of Electronic Engineering,
City University of Hong
Kong)
patch antenna in a cellular phone. A base station antenna array consisting of

L-probe-fed patches is shown in Fig. 69. Figure 70 is a circularly polarized antenna
for GPS applications in the form of a U-slot square patch with truncated corners.
Finally, a sketch of a flexible microstrip patch applicator conformable to a shaped
surface for microwave hyperthermia is given in Fig. 71a (Kobayashi et al. 1989).
A photography of this type of applicator (Ichinoseki-Sekine et al. 2007) in a clinical
environment is shown in Fig. 71b, where the patch antenna is used to impart localized
microwave energy at 434 MHz into the human body to destroy malignant tumors.
Z
a 150 mm
b
X
phantom model Y
of human tissue
Microstrip
patch
Water applicator
bolus
Fig. 71 (a) Flexible microstrip patch mounted on a curved surface for hyperthermia medical
application (From Kobayashi et al. (1989). # 1989 IEEE, Reprinted with permission). (b) The
applicator for the hyperthermia system positioned on a human thigh (From Ichinoseki-Sekine
et al. (2007). # 2007, British Journal of Sports Medicine, Reprinted with Permission)
Cross-References
▶ Antennas in Handheld Devices

▶ Circularly Polarized Antennas
▶ Low-Profile Antennas
▶ Phased Arrays
▶ Reconfigurable Antennas for Wireless Communications
▶ Small Antennas
▶ Wideband Magnetoelectric Dipole Antennas
References
Anguera J, Font G, Puente C, Borja C, Soler J (2003) Multifrequency microstrip patch antenna
using multiple stacked elements. IEEE Microwave Wireless Compon Lett 13(3):123–124
Bahl I, Bhartia P (1980) Microstrip antennas. Artech House, Dedham
Barlately L, Mosig JR, Sphicopoulos T (1990) Analysis of stacked microstrip patches with a mixed
potential integral equation. IEEE Trans Antennas Propag 38(5):608–615
Best SR (2009) The significance of ground-plane size and antenna location in establishing the
performance of ground-plane-dependent antennas. IEEE Antennas Propag Mag 51(6):29–43
Bhalla R, Shafai L (2002) Broadband patch antenna with a circular arc shaped slot. In: IEEE
Antennas and Propagation Society international symposium (IEEE Cat. No.02CH37313), IEEE,
pp 394–397, San Antonio, Texas
Bhartia P, Bahl I (1982) A frequency agile microstrip antenna. In: 1982 Antennas and Propagation
Society international symposium. Institute of Electrical and Electronics Engineers, pp 304–307,
Albuquerque, New Mexico
Chair R, Lee KF, Luk KM (1999) Bandwidth and cross-polarization characteristics of quarter-wave
shorted patch antennas. Microw Opt Technol Lett 22(2):101–103
Chair R, Mak C-L, Kishk AA (2005) Miniature wide-band half U-slot and half E-shaped patch
antennas. IEEE Trans Antennas Propag 53(8):2645–2652
Chen ZN, Chia MYW (2005) Broadband planar antennas: design and applications. Wiley-Chichester
Chung KL, Mohan AS (2003) A systematic design method to obtain broadband characteristics for
singly-fed electromagnetically coupled patch antennas for circular polarization. IEEE Trans
Antennas Propag 51(12):3239–3248
Clenet M, Shafai L (1999) Multiple resonances and polarisation of U-slot patch antenna. Electron
Lett 35(2):101–103
Croq F, Papiernik A (1990) Large bandwidth aperture-coupled microstrip antenna. Electron Lett
26(16):1293–1294
Dahele JS, Lee KF (1985) Theory and experiment on microstrip antennas with airgaps. IEE Proc H
Microwaves Antennas Propag 132(7):455–460
Dahele J, Lee K, Wong D (1987) Dual-frequency stacked annular-ring microstrip antenna. IEEE
Trans Antennas Propag 35(11):1281–1285
Debatosh G, Antar YMM (2010) Microstrip and printed antennas: new trends, techniques and
applications. Wiley, Hoboken
Deschamps GA, Sichak W (1953) Microstrip microwave antennas. In: Third USAF symposium on
antennas, Monticello, Illinois
Deshmukh AA, Ray KP (2010) Multi-band configurations of stub-loaded slotted rectangular
microstrip antennas. IEEE Antennas Propag Mag 52(1):89–103
Garg R et al (2000) Microstrip antenna design handbook. Artech House, Boston
Hall PS, Dahele JS (1997) Dual-band circularly polarized microstrip antenna. In: Lee KF, Chen W
(eds) Advances in microstrip and printed antennas. Wiley Interscience, New York, pp 163–217
Hirasawa K, Haneishi M (1992) Analysis, design, and measurement of small and low-profile
antennas. Artech House Publishers, Boston
Huang J (2008) Microstrip antennas: analysis, design, and application. In: Balanis CA (ed) Modern
antenna handbook. Wiley, Hoboken
Huynh T, Lee KF (1995) Single-layer single-patch wideband microstrip antenna. Electron Lett
31(16):1310–1312
Huynh T, Lee KF, Lee R (1988) Crosspolarisation characteristics of rectangular patch antennas.
Electron Lett 24(8):463–464
Ichinoseki-Sekine N et al (2007) Changes in muscle temperature induced by 434 MHz microwave
hyperthermia. Br J Sports Med 41(7):425–429
Jackson R, Ramadoss R (2007) A MEMS-based electrostatically tunable circular microstrip patch
antenna. J Micromech Microeng 17(1):1–8
James JR, Hall PS (eds) (1989) Handbook of microstrip antennas. Peregrinus, London
James JR, Hall PS, Wood C (1981) Microstrip antenna theory and design. Peregrinus, London
Kalialakis C, Hall PS (2007) Analysis and experiment on harmonic radiation and frequency tuning
of varactor-loaded microstrip antennas. IET Microwaves Antennas Propag 1(2):527–535
Khidre A, Lee KF, Yang F, Elsherbeni A (2010) Wideband circularly polarized E-shaped patch
antenna for wireless applications. IEEE Antennas Propag Mag 52(5):219–229
Khidre A, Lee KF, Elsherbeni A, Yang F (2013) Circular polarization reconfigurable wideband E-
shaped patch antenna for wireless applications. IEEE Trans Antennas Propag 61(2):960–964
Kobayashi H, Nikawa Y, Okada F, Mori S (1989) Flexible microstrip patch applicator for hyper-
thermia. In: Digest on Antennas and Propagation Society international symposium, IEEE,
pp 536–539, San Jose, CA
Kumar G, Gupta K (1984) Broad-band microstrip antennas using additional resonators gap-coupled
to the radiating edges. IEEE Trans Antennas Propag 32(12):1375–1379
Kumar G, Ray KP (2003) Broadband microstrip antennas. Artech House Publishers, Boston
Lee KF, Chen W (1997) Advances in microstrip and printed antennas. Wiley Interscience,
New York
Lee KF, Dahele JS (1989) Characteristics of microstrip patch antennas and some methods of
improving frequency agility and bandwidth. In: James JR, Hall PS (eds) Handbook of microstrip
antennas. Peregrinus, London, pp 111–214
Lee KF, Luk KM (2010) Microstrip patch antennas. Imperial College Press, London
Lee K, Tong K (2012) Microstrip patch antennas – basic characteristics and some recent advances.
Proc IEEE 100(7):2169–2180
Lee RQ, Lee K, Bobinchak J (1987) Characteristics of a two-layer electromagnetically coupled
rectangular patch antenna. Electron Lett 23(20):1070–1072
Lee KF, Luk KM, Dahele JS (1988) Characteristics of the equilateral triangular patch antenna. IEEE
Lee K-F, Chen W, Lee RQ (1995) Studies of stacked electromagnetically coupled patch antennas.
Microw Opt Technol Lett 8(4):212–215
Lee KF, Luk KM, Tong KF, Shum SM, Huynh T, Lee RQ (1997) Experimental and simulation
studies of the coaxially fed U-slot rectangular patch antenna. IEE Proc Microwaves Antennas
Propag 144(5):354–358
Lee KF, Guo YX, Hawkins JA, Chair R, Luk KM (2000) Theory and experiment on microstrip
patch antennas with shorting walls. IEE Proc Microwaves Antennas Propag 147(6):521–525
Lee KF, Yang SLS, Kishk AA (2008) Dual- and multiband U-slot patch antennas. IEEE Antennas
Wirel Propag Lett 7:645–647
Lee KF, Yang SL, Kishk AA, Luk KM (2010) The versatile U-slot patch antenna. IEEE Antennas
Propag Mag 52(1):71–88
Lee KF, Luk KM, Mak KM, Yang SLS (2011) On the use of U-slots in the design of dual-and triple-
band patch antennas. IEEE Antennas Propag Mag 53(3):60–74
Lo Y, Solomon D, Richards W (1979) Theory and experiment on microstrip antennas. IEEE Trans
Long S, Walton M (1979) A dual-frequency stacked circular-disc antenna. IEEE Trans Antennas
Propag 27(2):270–273
Luk KM, Mak CL et al (1998a) Broadband microstrip patch antenna. Electron Lett 34(15):
1442–1443
Luk KM, Chair R, Lee KF (1998b) Small rectangular patch antenna. Electron Lett 34(25):
2366–2367
Mak C, Lee K, Luk K (2000) Broadband patch antenna with a T-shaped probe. IEE Proc
Microwaves Antennas Propag 147(2):73–76
Mak C-L, Wong H, Luk K-M (2005) High-gain and wide-band single-layer patch antenna for
wireless communications. IEEE Trans Veh Technol 54(1):33–40
Mao Y, Padooru Y, Lee KF, Elsherbeni A, Yang F (2011) Air gap tuning of patch antenna resonance.
In: 2011 I.E. international symposium on antennas and propagation (APSURSI), IEEE,
pp 3088–3090, Spokane, Washington
Mok WC, Wong SH, Luk KM, Lee KF (2013) Single-layer single-patch dual-band and triple-band
patch antennas. IEEE Trans Antennas Propag 61(8):4341–4344
Mosig J, Gardiol F (1985) General integral equation formulation for microstrip antennas and
scatterers. IEE Proc H Microwaves Antennas Propag 132(7):424–432
Nakano H, Yamazaki M, Yamauchi J (1997) Electromagnetically coupled curl antenna. Electron
Lett 33(12):1003–1004
Pinhas S, Shtrikman S (1988) Comparison between computed and measured bandwidth of quarter-
wave microstrip radiators. IEEE Trans Antennas Propag 36(11):1615–1616
Pozar DM (1985) Microstrip antenna aperture-coupled to a microstripline. Electron Lett
21(2):49–50
Pozar DM (1992) Microstrip antennas. Proc IEEE 80(1):79–91
Qin P-Y et al (2010) Polarization reconfigurable U-slot patch antenna. IEEE Trans Antennas Propag
58(10):3383–3388
Rafi G, Shafai L (2004) Broadband microstrip patch antenna with V-slot. IEE Proc Microwaves
Reineix A, Jecko B (1989) Analysis of microstrip patch antennas using finite difference time
domain method. IEEE Trans Antennas Propag 37(11):1361–1369
Richards WF, Lo YT, Harrison DD (1981) An improved theory for microstrip antennas and
applications. IEEE Trans Antennas Propag 29(1):38–46
Sabban A (1983) A new broadband stacked two-layer microstrip antenna. In: 1983 Antennas and
Propagation Society international symposium, Institute of Electrical and Electronics Engineers,
pp 63–66, Houston, Texas
Salon S, Chari MVK (1999) Numerical methods in electromagnetism. Academic, San Diego
Schaubert D et al (1981) Microstrip antennas with frequency agility and polarization diversity. IEEE
Schneider MV (1969) Microstrip lines for microwave integrated circuits. Bell Syst Tech J
48(5):1421–1444
Shackelford AK et al (2001) U-slot patch antenna with shorting pin. Electron Lett 37(12):729–730
Shackelford AK, Lee KF, Luk KM (2003) Design of small-size wide-bandwidth microstrip-patch
antennas. IEEE Antennas Propag Mag 45(1):75–83
Shafai L (2007) Wideband microstrip antennas. In: Volakis JL (ed) Antenna engineering handbook.
McGraw Hill, New York
Tong K-F, Wong T-P (2007) Circularly polarized U-slot antenna. IEEE Trans Antennas Propag
55(8):2382–2385
Tong KF et al (2000) A broad-band U-slot rectangular patch antenna on a microwave substrate.
IEEE Trans Antennas Propag 48(6):954–960
Tong KF, Lee KF, Luk KM (2011) On the effect of ground plane size to wideband shorting-wall
probe-fed patch antennas. In: 2011 IEEE-APS topical conference on antennas and propagation
in wireless communications, IEEE, pp 486–486, Torino, Italy
Waterhouse RB, Shuley NV (1994) Full characterisation of varactor-loaded, probe-fed, rectangular,
microstrip patch antennas. IEE Proc Microwaves Antennas Propag 141(5):367–373
Waterhouse RB, Targonski SD, Kokotoff DM (1998) Design and performance of small printed
antennas. IEEE Trans Antennas Propag 46(11):1629–1633
Weigand S et al (2003) Analysis and design of broad-band single-layer rectangular u-slot microstrip
patch antennas. IEEE Trans Antennas Propag 51(3):457–468
Wong KL (2002) Compact and broadband microstrip antennas. Wiley Interscience, New York
Wong H, Luk KM, Chan CH, Xue Q, So KK, Lai HW (2012) Small antennas in wireless
communications. IEEE Proc 100(7):2109–2121
Wood C (1981) Analysis of microstrip circular patch antennas. IEE Proc H Microwaves Opt
Antennas 128(2):69–76
Yang F, Zhang, X-X, Rahmat-Samii (2001a). Wide-band E-shaped patch antennas for wireless
communications. IEEE Trans Antennas Propag 49(7):1094–1100
Yang F, Rahmat-Samii Y (2001b) Patch antenna with switchable slot (PASS): dual-frequency
operation. Microw Opt Technol Lett 31(3):165–168
Yang F, Rahmat-Samii Y (2002) A reconfigurable patch antenna using switchable slots for circular
polarization diversity. IEEE Microwave Wireless Compon Lett 12(3):96–98
Yang S et al (2008) Design and study of wideband single feed circularly polarized microstrip
antennas. Prog Electromagn Res 80:45–61
Zhong SS, Lo YT (1983) Single-element rectangular microstrip antenna for dual-frequency oper-
ation. Electron Lett 19(8):298–300
Zurcher J, Gardiol F (1995) Broadband patch antenna. Artech House Publishers, Boston

Microstrip Patch Antennas: Kai Fong Lee and Kin-Fai Tong

Uploaded by

Copyright:

Available Formats

Microstrip Patch Antennas: Kai Fong Lee and Kin-Fai Tong

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Microstrip Patch Antennas: Kai Fong Lee and Kin-Fai Tong

Uploaded by

Copyright:

Available Formats

Microstrip Patch Antennas

Kai Fong Lee and Kin-Fai Tong

K.F. Lee (*)

# Springer Science+Business Media Singapore 2016 787

Frequency-Agile and Polarization-Agile Patch Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 836

Fig. 1 The basic structure of the microstrip patch antenna

Fig. 2 Four common feeding methods of microstrip patch antenna

Modeling Techniques and Basic Characteristics

Brief Description of Modeling Techniques

There are two approaches to deducing the performance characteristics of microstrip

Table 2 A list of commercial electromagnetic simulation software

The basic characteristics of microstrip patch antennas can be obtained by applying

c = velocity of light in free space.

Fig. 4 Top view of the

is the effective permittivity (Schneider 1969).

reﬂection coefﬁcient (scattering parameter S11). This is usually also the

feed for (1,0) mode varies rectangular

due to surface-wave generation. Additionally, when the substrate thickness

The operating frequencies of some wireless communication systems are shown in

Table 3 Frequencies of some wireless communication systems

(A) By means of parasitic elements or slots, additional resonances are introduced so

Table 4 Design examples 12 % 12 % 15 %

(dielectric cover). When a superstrate of thickness of 0.26 mm and relative permit-

General Remarks About Aperture-Coupled Feed

Example of a Wideband Aperture-Coupled Patch Antenna

The Wideband U-Slot Patch Antenna

Fig. 12 Geometry of the rectangular U-slot patch antenna

response. Subsequently, Lee et al. (1997) produced a more comprehensive experi-

Table 5 Dimension of the h (mm) W L Ws Ls a b F

Table 7 Dimensions of antenna in millimeters

Table 8 Operating frequencies and bandwidth of the antenna in Table 7

Variations of the U-Slot Patch Antenna and the E-Patch

The Wideband L-Probe Coupled Patch Antenna

Fig. 14 Geometry of the

Fig. 15 Geometry of the

Dual- and Multiband Designs

The Use of U-Slot Patches

Large Frequency Ratios

Fig. 18 Geometry of dual-band U-slot patch antenna

Fig. 19 Simulated reﬂection

Fig. 20 Geometry of tri-band U-slot patch antenna

Fig. 21 Simulated reﬂection

and 22 MHz. Finally, quadruple-band response can be obtained with a stub-loaded

Small Frequency Ratios

Methods of Reducing the Patch Size

In many applications, it is desirable for the dimensions of the patch to be a small

240 120 240 120

210 150 210 150

270 90 270 –40 –30 –20 –10 90

240 120 240 120

210 150 210 150

270 –40 –30 –20 –10 90

can be applied to obtain small-size wideband patch antennas (Shackelford

The Use of Shorting Wall: Quarter-Wave Patch

Table 12 Measured gains of the shorted square patch of Fig. 31

Partially Shorted Patch and Planar Inverted-F Antenna

Fig. 32 Size reduction by using an inverted-F patch

The Use of Shorting Pin