international journal of antennas and...

Microstrip Antennas

International Journal of Antennas and Propagation

Microstrip Antennas

International Journal of Antennas and Propagation

Microstrip Antennas

Copyright © 2012 Hindawi Publishing Corporation. All rights reserved.

This is a focus issue published in “International Journal of Antennas and Propagation.” All articles are open access articles distributedunder the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, pro-vided the original work is properly cited.

Editorial Board

M. Ali, USACharles Bunting, USAFelipe Catedra, SpainDau-Chyrh Chang, TaiwanDeb Chatterjee, USAZ. N. Chen, SingaporeMichael Yan Wah Chia, SingaporeChristos Christodoulou, USAShyh-Jong Chung, TaiwanLorenzo Crocco, ItalyTayeb A. Denidni, CanadaAntonije R. Djordjevic, SerbiaKaru P. Esselle, AustraliaFrancisco Falcone, SpainMiguel Ferrando, SpainVincenzo Galdi, ItalyWei Hong, ChinaHon Tat Hui, SingaporeTamer S. Ibrahim, USANemai Karmakar, Australia

Se-Yun Kim, Republic of KoreaAhmed A. Kishk, CanadaTribikram Kundu, USAByungje Lee, Republic of KoreaJu-Hong Lee, TaiwanL. Li, SingaporeYilong Lu, SingaporeAtsushi Mase, JapanAndrea Massa, ItalyGiuseppe Mazzarella, ItalyDerek McNamara, CanadaC. F. Mecklenbrauker, AustriaMichele Midrio, ItalyMark Mirotznik, USAAnanda S. Mohan, AustraliaP. Mohanan, IndiaPavel Nikitin, USAA. D. Panagopoulos, GreeceMatteo Pastorino, ItalyMassimiliano Pieraccini, Italy

Sadasiva M. Rao, USASembiam R. Rengarajan, USAAhmad Safaai-Jazi, USASafieddin Safavi-Naeini, CanadaMagdalena Salazar-Palma, SpainStefano Selleri, ItalyKrishnasamy T. Selvan, IndiaZhongxiang Q. Shen, SingaporeJohn J. Shynk, USAM. Singh Jit Singh, MalaysiaSeong-Youp Suh, USAParveen Wahid, USAYuanxun Ethan Wang, USADaniel S. Weile, USAQuan Xue, Hong KongTat Soon Yeo, SingaporeYoung Joong Yoon, KoreaWenhua Yu, USAJong Won Yu, Republic of KoreaAnping Zhao, China

Contents

Modal Resonant Frequencies and Radiation Quality Factors of Microstrip Antennas, Jan Eichler,Pavel Hazdra, Miloslav Capek, and Milos MazanekVolume 2012, Article ID 490327, 9 pages

Tunable Compact UHF RFID Metal Tag Based on CPW Open Stub Feed PIFA Antenna, Lingfei Mo andChunfang QinVolume 2012, Article ID 167658, 8 pages

Some Recent Developments of Microstrip Antenna, Yong Liu, Li-Ming Si, Meng Wei, Pixian Yan,Pengfei Yang, Hongda Lu, Chao Zheng, Yong Yuan, Jinchao Mou, Xin Lv, and Housjun SunVolume 2012, Article ID 428284, 10 pages

New Configurations of Low-Cost Dual-Polarized Printed Antennas for UWB Arrays, Guido Valerio,Simona Mazzocchi, Alessandro Galli, Matteo Ciattaglia, and Marco ZuccaVolume 2012, Article ID 786791, 10 pages

Design and Analysis of Wideband Nonuniform Branch Line Coupler and Its Application in a WidebandButler Matrix, Yuli K. Ningsih, M. Asvial, and E. T. RahardjoVolume 2012, Article ID 853651, 7 pages

Isolation Improvement of a Microstrip Patch Array Antenna for WCDMA Indoor Repeater Applications,Hongmin Lee and Jinwon ParkVolume 2012, Article ID 264618, 8 pages

Series-Fed Microstrip Array Antenna with Circular Polarization, Tuan-Yung HanVolume 2012, Article ID 681431, 5 pages

Vertical Meandering Approach for Antenna Size Reduction, Li Deng, Shu-Fang Li, Ka-Leung Lau,and Quan XueVolume 2012, Article ID 980252, 5 pages

Microstrip Patch Antenna Bandwidth Enhancement Using AMC/EBG Structures, R. C. Hadarig,M. E. de Cos, and F. Las-HerasVolume 2012, Article ID 843754, 6 pages

High-Performance Computational Electromagnetic Methods Applied to the Design of Patch Antennawith EBG Structure, R. C. Hadarig, M. E. de Cos, and F. Las-HerasVolume 2012, Article ID 435890, 5 pages

A Wideband High-Gain Dual-Polarized Slot Array Patch Antenna for WiMAX Applications in 5.8 GHz,Amir Reza Dastkhosh and Hamid Reza Dalili OskoueiVolume 2012, Article ID 595290, 6 pages

Hindawi Publishing CorporationInternational Journal of Antennas and PropagationVolume 2012, Article ID 490327, 9 pagesdoi:10.1155/2012/490327

Research Article

Modal Resonant Frequencies and Radiation Quality Factors ofMicrostrip Antennas

Jan Eichler, Pavel Hazdra, Miloslav Capek, and Milos Mazanek

Department of Electromagnetic Field, Faculty of Electrical Engineering, Czech Technical University in Prague, Technicka 2,166 27, Prague, Czech Republic

Correspondence should be addressed to Pavel Hazdra, [email protected]

Received 9 August 2011; Revised 10 January 2012; Accepted 13 January 2012

Academic Editor: Charles Bunting

Copyright © 2012 Jan Eichler et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The chosen rectangular and fractal microstrip patch antennas above an infinite ground plane are analyzed by the theory ofcharacteristic modes. The resonant frequencies and radiation Q are evaluated. A novel method by Vandenbosch for rigorousevaluation of the radiation Q is employed for modal currents on a Rao-Wilton-Glisson (RWG) mesh. It is found that theresonant frequency of a rectangular patch antenna with a dominant mode presents quite complicated behaviour including havinga minimum at a specific height. Similarly, as predicted from the simple wire model, the radiation Q exhibits a minimum too. It isobserved that the presence of out-of-phase currents flowing along the patch antenna leads to a significant increase of the Q factor.

1. Introduction

Evaluation of the basic properties of microstrip patchantennas (MPA) has been numerously discussed in literature,see, for example, [1–3]. The two main MPA attributes areresonant frequency (or frequencies but we will deal mostlywith the dominant mode) and the radiation Q factor. So far,only approximate results and semianalytic equations havebeen published. To our knowledge, this is the first timethat these important characteristics have been studied in arigorous way. The antennas are treated by using a modalapproach (hence we do not a priori consider any feedingto be connected), namely, by the theory of characteristicmodes (TCM), [4, 5]. Evaluation of the radiation Q isperformed both by the TCM from the eigenvalues slope andby novel rigorous equations derived by Vandenbosch [6] andVandenbosch and Volski [7].

2. The Theory of Characteristic Modes

For completeness, let us formulate the basics of the charac-teristic modes for perfectly conducting bodies of area S. Thescattered field Es is related to the electric surface currents J bythe electric field integral equation (EFIE) [8][

L(J)− Ei]

tan= 0. (1)

Equation (1) is usually treated within the method ofmoments (MoMs) [8] framework and, due to the structurediscretization, the L operator is known as the “compleximpedance matrix” [Z] = [R] + j[X].

Then the associated Euler’s equation to be solved is

XJn = λnRJn. (2)

Equation (2) is a standard weighted eigenvalue equationleading to a set of real characteristic eigencurrents Jn andassociated eigenvalues λn. Properties of eigenvalues aredescribed in [9], at this moment it is important to notethat λn reflects the amount of net reactive power (thus λn =0 means resonance). Instead of eigenvalues, the so-calledcharacteristic angles αn are introduced to show more visiblebehavior with frequency [9]. Characteristic currents form acomplete orthogonal set, and hence the total current on aconducting body may be expressed as a linear combinationof these mode currents [10].

2.1. Implementation of the Characteristic Modes Theory.Implementation of the modal decomposition process hasbeen done in the MATLAB [11] environment using MakarovEFIE codes [12] with the RWG basis functions [13]. Thisusage is restricted to arbitrary 3D PEC structures with air

2 International Journal of Antennas and Propagation

dielectrics. Our developed TCM tool [14] has the followingmain advantages:

(i) Comsol Multiphysics [15]/MATLAB’s PDE TooLboxmesh import,

(ii) Optional Green’s function for infinite ground planesimulations,

(iii) Single solver/multicore solver/distributed solver(within a computer network with installedMATLAB).

3. The Radiation Q Factor

In [6] a novel theory able to rigorously calculate radiatedpower and stored energies directly from currents flowingalong the antenna has been presented. The radiation Q factoris then readily evaluated by the definition [16]:

Q = 2ωmax

(Wm, We

)Pr

. (3)

The equations for radiated power Pr and stored electric andmagnetic energies We, Wm are

Pr =(

18πωε0

)∫Ω1

∫Ω2

[k2J(r1)J(r2)−∇ · J(r1)∇ · J(r2)

]× sin(kr21)

r21dΩ1dΩ2,

(4)

We = 116πω2ε0

(Ie − IR), (5)

Wm = 116πω2ε0

(Im − IR), (6)

where

IR = k

2

∫Ω1

∫Ω2

[k2J(r1)J(r2)−∇ · J(r1)∇ · J(r2)

]× sin(kr21)dΩ1dΩ2,

(7)

Ie =∫Ω1

∫Ω2∇ · J(r1)∇ · J(r2)

cos(kr21)r21

dΩ1dΩ2, (8)

Im = k2∫Ω1

∫Ω2

J(r1)J(r2)cos(kr21)

r21dΩ1dΩ2, (9)

where k is a free-space wavenumber, J is the surface currentdensity, and r21 is the distance between interacting currentelements. The tilde denotes that the radiation contribution IRhas been subtracted from the stored energies at every pointin space [17]. It is assumed that the currents are flowing in avacuum.

3.1. The Modal Radiation Q Factor. The modal radiation Qfactor may be evaluated from the slope of modal eigenvalues[18]:

Qeig = ω0

2dλ

dω. (10)

y

T (x,y,z)1

1

12

T (x,y,z)x

(0,0)

| |

rr

J

J2

2

r21 = r2 − r1

Ω

Figure 1: Distance between nonoverlapping current elements [23].

In [18], (10) is supposed to be an approximation of theradiation Q, but in resonance it is actually exact.

Since characteristic modes are normalized to radiate unitpower Pr = 1 [4], (3) reduces to

Q = 2ωmax(We, Wm

). (11)

For parallel or series RLC circuit (hence, for one mode), the“impedance QZ” equals the exact “current Q” [6]:

Q = QZ = ω0

2

∣∣∣∣∂Z∂ω∣∣∣∣ = ω0

2

∣∣∣∣ ∂R∂ω +∂X

∂ω

∣∣∣∣. (12)

Inserting

Z = R + jX = 1|I2|

[Pr + j2ω

(Wm − We

)](13)

valid for lossless antennas [19] and using the fact that Pr = 1,(12) results in

Q = QX = ω0

2

∣∣∣∣∂X∂ω∣∣∣∣ = ω0

2∂

∂ω

[2ω(Wm − We

)]= ω0

2∂λ

∂ω= Qeig,

(14)

providing that

λ = 2ω(Wm − We

). (15)

It is therefore concluded that the modal Qeig equals the QX bydefinition, and it can be proven (using the reactance theorem[20, 21]) that in resonance it also equals the radiation Qdefined from energies by (11).

3.2. Software Implementation. The above equations wereimplemented in MATLAB for the RWG triangular meshwhere two different interaction situations occur:

(a) Distant Elements. When the triangular elements are notoverlapping, current density on triangles may be simplyapproximated as point sources located at the centre oftriangles [22], see Figure 1. No actual integration is thenneeded. This centroid approach is very fast with satisfactoryaccuracy as will be shown later (however it may fail forpatches located very close to the ground plane).

International Journal of Antennas and Propagation 3

P3 (x3, y3)

P1 (x

x

1, y

y

1)

P2

P0

P0

(x2, y2)

(0,0)

(0,1)

(0,0) (1,0)

β

α

T1 ( β,α ,γ)T1 (x,y ,z)

h13

h12

h23

A

(a) (b)

r1

r2

Figure 2: Self-term evaluation. (a) Original problem, (b) simplexcoordinates transformation [23].

(b) Overlapping (Self) Elements. As known from the methodof moments, the so-called “self” contributions are of greatimportance when dealing with calculations on discreteelements (meshes).

Here, the self-interaction occurs when two triangles areoverlapping each other. Due to the behavior of integralkernels, only rapidly varying term cos(kr21)/r21 has to becarefully treated. Since k0R21 → 0 (R21 being the longest sideof the triangle T) is satisfied, one needs only to use the firstterm in the Taylor series expansion. The dominant singularstatic part is 1/r21 and the integral to be worked out is

I =∫T

∫T′

1√(x − x′)2 +

(y − y′

)2dx dy dx′ dy′, (16)

where T = T′ is a triangular area. Using simplex coordinatestransformation (Figure 2), the result is [23, 24]

I=−43A2[

ln(1−2h12/L)h12

+ln(1−2h13/L)

h13+

ln(1−2h23/L)h23

],

(17)

where A is the triangle area, hi j are the edge lengths (seeFigure 2), and L is the perimeter of the triangle.

4. Applications: Rectangular Patch Antenna

Let us first concentrate on a rectangular patch antenna ofdimensions L = 50 mm and W = 30 mm (further noted asR50× 30) placed in air at a heightH above an infinite groundplane. Only the dominant TM01 mode will be studied. Thereason for choosing a patch with L/W /= 1 is that we do nothave to deal with degenerated modes.

Using the image theory, the radiator in the XY planeat height z = H above an infinite electric ground plane ismodelled as two patches separated by 2H . The total numberof triangular elements is 676. In the TCM analyser, a properout-of-phase mode is selected (Figure 3).

The resonant frequency of the dominant mode is shownas a function of height H , see Figure 4. It has been evaluatedfrom a modal resonant condition for eigenvalues λ =2ω(Wm − We) = 0 employing an adaptive frequencysweep for each height. The behaviour is quite peculiar,especially for greater heights. For low heights (H < 10 mm or

2H

x

y

z

0.01

0.005

0

−0.005

−0.01

0.015

0.01

0.005

0−0.005

−0.01

−0.015−0.03

−0.02−0.01

00.01

0.02

Figure 3: Model of MPA above infinite ground plane for H =10 mm, dominant mode TM01 shown.

0 5 10 15 20 25 30 35 40 45 50 55 601

1.5

2

2.5

3

3.5

4

4.5

H (mm)

f res

(G

Hz)

fr

r

(TCM)

f (analytic equation)

Regular behaviour

Min. f res

Min. Q

Figure 4: R50 × 30 resonant frequency of the dominant TM01

mode. The dashed red curve is a quasianalytical equation from [1].

H/λres < 0.08), the resonant frequency decreases “regularly,”and quasianalytical formulas (see, e.g., [1, 3]) based onthe fringing field concept are valid below this range. ForH ∼= 25 mm (H/λres

∼= 0.188) there is absolute minimumof the TM01 resonant frequency. Further on, the resonantfrequency rises to reach its maximum for H ∼= 40 mm(H/λres

∼= 0.51). Around this specific height the patch alsoshows the minimum of the radiation Q. The above describedprocess repeats periodically. It is yet unclear to the authorsas what is the physical background to the resonant frequencydiscontinuity around H/λres

∼= 0.5.The terms 2ωWm, 2ωWe, and 2ω(Wm − We) obtained

from (5)–(9) and eigenvalues λ are plotted at Figure 5 for


1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3

0

5

10

15

20

f (GHz)

Rea

ctiv

e en

ergi

es

f res

−5

−10

λ

2ω (Wm)

2ω (We)

2ω(Wm-We)

Figure 5: Reactive energies and their differences for an R50 × 30patch at height of 25 mm.

0 5 10 15 20 25 30 35 40 45 50 55 600

5

10

15

20

25

30

35

40

45

50

H (mm)

Qeig

Q J

Q (

–)

Figure 6: The radiation Q for dominant mode of an R50× 30 patchas a function of height H .

H = 25 mm as a function of frequency. There is excellentagreement between the difference in stored energies and theeigenvalues, both obtained in a completely different manner.

There is also very good agreement between the exact QJ

and Qeig confirming the validity of the proposed algorithmvia (14), see Figure 6. Note that Qeig in (10) does not requirethe currents to be calculated on the structure while QJ isevaluated in a rigorous way from modal currents (11).

From Figure 6 it is seen that the radiation Q has aminimum for a specific height. It is deduced that the reasonlies in the cancelling of the radiated power between the twoout-of-phase currents. Similar behaviour has been observedin the case of two half-wave thin-wire dipoles with oppositesinusoidal currents, separated by d = 2H , see [25] for details.Actually these two out-of-phase dipoles may serve as a verysimple model for a patch antenna with a dominant mode.When the dipoles are reduced to elementary (Hertzian) ones,

an approximate analytical solution is available and in [25] weshowed that the Q is led by the function

fQ(H) ∼= 2Hk2H − sin(k2H)

. (18)

After deriving (18), the condition is worked-out

tan(k2H) = k2H , (19)

and the first nontrivial root of (19) could be approximatedas [25] (

H

λ

)min

∼= 38− 1

6π2= 0.358. (20)

For sinusoidal currents on dipoles the minimum (evaluatednumerically) occurs for H = 0.36λ.

The minimum of the patch under study is obtained atH ∼= 0.4λ, a value that is remarkably close to the simpledipole model.

4.1. Algorithm Convergence. Since no other methods forcalculating modal Q are available, Qeig is taken as a reference,and the relative error percentage is defined as:

relative error =∣∣∣QJ −Qeig

∣∣∣Qeig

· 100, (21)

where QJ is calculated from the currents using (11). Fourdifferent heights H were chosen, H = 1 mm (0.01λ), H =2 mm (0.0185λ), H = 10 mm (0.0803λ), and H = 20 mm(0.151λ), and the relative error was evaluated as a function oftotal triangular elements (including the mirror), see Figure 7.All quality factors were evaluated at the resonant frequencyof the dominant mode for the R50 × 30 patch. As discussedearlier, the centroid approximation became more inaccuratewith low heights H . However, even for the lowest analyzedvalue H = 0.01λ, the relative error is in the order ofa few percent for reasonable mesh density (hundreds ofelements). Further improvements to the integration routineare considered for the future.

4.2. Fractional Bandwidth of the R50 × 30 Patch Antenna. Itis known that the fractional bandwidth (FBW) is related tothe unloaded Q factor and the desired matching VSWR level.For VSWR < s we have [26]

FBW ∼= s− 1Q√s

[%]. (22)

Using a full-wave simulator CST-MWS [27], an R50 × 30patch has been simulated and the FBWCST for VSWR < 2 wascalculated as:

FBWCST = f2 − f1f0

, (23)

where f2 and f1 are margins for VSWR < 2 and f0 isthe centre frequency. Only very low heights were studiedsince we used a simple probe feed which introduces aninductance component to the total input impedance. Thecomparison in Figure 8 shows good agreement of bothfractional bandwidths.


0 200 400 600 800 1000 1200 14000

1

2

3

4

5

6

7

8

9

10

Number of triangle elements

Rel

ativ

e er

ror

(%)

H = 1 mm (0.01 )H = 2 mm (0.0185 )

H = 10 mm (0.0803 )H = 20 mm (0.151 )λ λ

λ λ

Figure 7: Relative error of the Q factor as a function of triangularelements (mesh density).

1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

H (mm)

FBW

(%

)

FBW (CST)

FBW (Q eig )

Figure 8: Fractional bandwidth FBW (VSWR < 2) for a R50 × 30patch.

5. Applications: Fractal Antennas

In this section, a bit more complex structures will be studied.The first one (the “Self Affine U” fractal, SAU), has beendescribed in [28] and further analyzed in [29]. This kindof radiating motif is employed as a dual-band radiatorwith mutually orthogonal radiation patterns at both bands.Therefore we are analyzing the first two modes, where thecurrents are orthogonal. These are depicted in Figures 9and 10 for first (SAU1) and second (SAU2) fractal iteration,respectively. The current of the first (lower) mode J1 hastwo out-of-phase components (see Figure 11 for schematiccurrent paths) while the second mode comprises inphasecurrents only. As we know from previous studies, oppositecurrents contribute to a rapid increase of the radiation Q,and it is expected that J1 will have a much higher Q than J2.

J1 J2

σ1 σ2

Figure 9: The first two characteristic modes (currents and charges)for the SAU1 structure.

J

W

L

0 0

2J1

σ1 σ2

Figure 10: The first two characteristic modes (currents andcharges) for the SAU2 structure.

J1 J2

Original

Infinite ground plane

Image

Figure 11: The main current paths for the first two modes of theSAU1/2 structure.


1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25 3.5 3.75 4100

120

140

160

180

200

220

240

260

280

f (Hz)

Mode no.1

Mode no.2 Mode no.3

×109

αn

(a)

10 12 14 16 18 20 22 24 26 28 300

10

20

30

40

50

60

70

80

90

100

H (mm)

Q

Q2

Q1

(b)

Figure 12: Characteristic angles (left) and radiation Q for the SAU2.

0.030.020.010−0.01−0.02−0.03

0.03

0.02

0.01

0

J1

(a)

0.030.020.010−0.01−0.02−0.03

0.03

0.02

0.01

0

J2

(b)

0.030.020.010−0.01−0.02−0.03

0.03

0.02

0.01

0

−0.01

−

−

0.02

0.03

σ1

(c)

0.030.020.010−0.01−0.02−0.03

0.03

0.02

0.01

0

−0.01

−

−

0.02

0.03

σ2

(d)

Figure 13: Degenerated dominant mode J1, and J2 of the FCL2 antenna (currents and charges).

Figure 11 presents a very simple concept showing themain current paths for the J1 and J2 modes discussed aboveincluding the mirroring effect of the infinite ground plane. Itcould be simply stated that more opposing current paths leadto significant increase in Q.

We show detailed behaviour only for SAU2 (the situationis similar for SAU1)—see Figure 12 that confirms high Q forthe J1 mode. Characteristic angles are calculated for H =29 mm, the actual height for which the dual-band antennawas designed [29].

5.1. The FCL-2 Fractal Antenna. The second presented struc-ture is the so-called fractal clover leaf (FCL) of the seconditeration, [14]. The antenna is fed by an L-probe [30] thatexcites its dominant mode and is located at height H =36 mm. Actually, the dominant mode is composed of twodegenerated modes J1 and J2 (Figure 13). The second highermode J3 is shown at Figure 14 for completeness.

Figure 15 shows the main current paths of these modes,and we can again deduce that the dominant mode will exhibit


0.030.020.010−0.01−0.02−0.03

0.03

0.02

0.01

0

J3

(a)

0.030.020.010−0.01−0.02−0.03

0.03

0.02

0.01

0

−0.01

−

−

0.02

0.03

σ3

(b)

Figure 14: Second higher mode J3 (currents and charges).

Mode 3

Mode 1, 2

=

=

J1 + J2

J3

8.256.875.54.122.751.370−3 .62−7 .24−10 .87−14 .49−18 .12−21 .74

7.326.14.883.662.441.220−3 .77−7 .55−11 .33−15 .11−18 .89−22 .67

Figure 15: Schematic depiction of the dominant current paths for the dominant (J1 + J2) and the second higher J3 modes together with theirmodal radiation patterns.

lower Q compared to J3. This is confirmed by Figure 16—J3

has more than 200x higher radiation Q.

6. Resonant Properties of Studied Antennas

The properties of studied antennas are summarized in thissection. At first we observed that microstrip antenna couldsupport different kinds of modes regarding their Q factors(see Figure 17):

(a) low Q modes with the current flowing in one direc-tion and not changing its phase (dominant modes ofsimple shapes like rectangular, circular patch, and soforth.)

(b) high Q modes with part of the currents flowing in theopposite direction. These modes exist even on simple“U” shaped patch (Figure 9 left) and on complex(fractal) geometries.

Secondly, it has been observed that resonant frequency isquite a complicated function of height. Unfortunately we donot yet have any physical explanation as to why some modespresent minimum values of fr .

Looking at Figure 18, it is clear (and interesting) thatthe resonant frequency behaves quite differently for low-Q and high-Q modes. The resonant frequency of low-Qmodes is much more sensitive to the height, whereas high-Qmodes exhibit almost constant fr when the height is varied.The proposed explanation is that the opposite currents(responsible for high Q) keep reactive fields very close to theradiating structure so the effect of a fringing field coupled tothe ground plane becomes almost negligible.

7. Conclusions

Modal resonant properties of selected microstrip patchantennas have been studied with the help of characteristic


1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 280

100

120

140

160

180

200

220

240

260

280

f (GHz)

Mode 1Mode 2Mode 3

Qeig3 = 226.3

Qeig1= Qeig2 = 10.5

×109

αn

Figure 16: Characteristic angles for the FCL2 structure at H =29 mm.

0 5 10 15 20 25 30 350

10

20

30

40

50

60

70

80

90

100

H (mm)

FCL2SAU1 mode 1SAU1 mode 2SAU2 mode 1

SAU2 mode 2

Qeig

High-Qmodes

Low-Qmodes

R50×50R50×30

Figure 17: Radiation Qs for different antennas/modes.

modes and the novel theory published by Vandenbosch. Ithas been found that the resonant frequency of a simplerectangular patch antenna is quite a complicated functionof its height above the infinite ground. Moreover, thedependency of resonant frequency is also found to be afunction of the radiation Q factor (which is now possible tocalculate in a rigorous way). Due to the complexity of theproblem, no physical explanation for the resonant frequencybehaviour has yet been found.

It is observed that the radiation Q factor decreasesfor “standard” heights (<∼0.1λ), however there exists anabsolute minimum value ofQ that has already been predictedby simple modeling of two elementary out-of-phase dipoles.

0 5 10 15 20 25 30 351

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

3

H (mm)

FCL2SAU1 mode 1SAU1 mode 2SAU2 mode 1

SAU2 mode 2

f r(G

Hz)

High-Qmodes

Low-Qmodes

R50×50R50×30

Figure 18: Resonant frequencies for different antennas/modes.

Using proper feeding techniques (like the L-probe) allows usto design wideband compact antennas.

The theory now puts current distribution and the radi-ation Q factor into objective context. Whenever the currentmode exhibits opposite components, high Q may appear.

Future work is needed to connect the presented theorywith parameter sweeps or even optimization, so we will beable to design novel wideband/multimode compact anten-nas.

Acknowledgments

This work was supported by the Grant Agency of theCzech Technical University in Prague, grant no. SGS11/065/OHK3/1T/13 and by the Project COST 1102. The authorswould like to thank professor Vandenbosch for fruitfuldiscussions, Neil Bell for his comments, and two anonymousreviewers who suggested some improvements to the paper.

References

[1] P. Bhartia, I. Bahl, R. Garg, and A. Ittipiboon, MicrostripAntenna Design Handbook, Artech House, Norwood, Mass,USA, 2000.

[2] K. F. Lee and W. Chen, Advances in Microstrip and PrintedAntennas, John Wiley & Sons, New York, NY, USA, 1997.

[3] J. R. James, P. S. Hall, and C. Wood, Microstrip Antenna Theoryand Design, Peter Peregrinus, London, UK, 1986.

[4] R. F. Harrington and J. Mautz, “Theory of characteristicmodes for conducting bodies,” IEEE Transactions on Antennasand Propagation, vol. 19, no. 5, pp. 622–628, 1971.

[5] E. A. Daviu, Analysis and design of antennas for wirelesscommunications using modal methods, Ph.D. dissertation,Universidad Politecnica de Valencia, Valencia, Spain, 2008.

[6] G. A. E. Vandenbosch, “Reactive energies, impedance, and Qfactor of radiating structures,” IEEE Transactions on Antennas


and Propagation, vol. 58, no. 4, Article ID 5398856, pp. 1112–1127, 2010.

[7] G. A. E. Vandenbosch and V. Volski, “Lower bounds forradiation Q of very small antennas of arbitrary topology,” inProceedings of the 4th European Conference on Antennas andPropagation (EuCAP ’10), pp. 1–4, Barcelona, Spain, April2010.

[8] R. F. Harrington, Field Calculation by the Method of Moments,IEEE Press, New York, NY, USA, 1993.

[9] M. Cabedo-Fabres, E. Antonino-Daviu, A. Valero-Nogueira,and M. F. Bataller, “The theory of characteristic modesrevisited: a contribution to the design of antennas for modernapplications,” IEEE Antennas and Propagation Magazine, vol.49, no. 5, pp. 52–68, 2007.

[10] P. Hazdra and P. Hamouz, “On the modal superposition lyingunder the MoM matrix equations,” Radioengineering, vol. 17,no. 3, pp. 42–46, 2008.

[11] Mathworks, http://www.mathworks.com/.[12] S. N. Makarov, Antenna and EM Modeling with Matlab, John

Wiley & Sons, New York, NY, USA, 2002.[13] S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic

scattering by surfaces of arbitrary shape ,” IEEE Transactions onAntennas and Propagation, vol. 30, no. 3, pp. 409–418, 1982.

[14] M. Capek, P. Hazdra, P. Hamouz, and M. Mazanek, “Softwaretools for efficient generation, modelling and optimisation offractal radiating structures,” IET Microwaves, Antennas andPropagation, vol. 5, no. 8, pp. 1002–1007, 2011.

[15] Comsol Multiphysics, http://www.comsol.com/.[16] J. L. Volakis, Ch. Chen, and K. Fujimoto, Small Anten-

nas: Miniaturization Techniques & Applications, chapter 2,McGraw-Hill, New York, NY, USA, 1st edition, 2010.

[17] J. S. McLean, “A re-examination of the fundamental limitson the radiation q of electrically small antennas,” IEEETransactions on Antennas and Propagation, vol. 44, no. 5, pp.672–676, 1996.

[18] R. F. Harrington and J. R. Mautz, “Control of radar scatteringby reactive loading,” IEEE Transactions on Antennas andPropagation, vol. 20, no. 4, pp. 446–454, 1972.

[19] R. F. Harrington, Time-Harmonic Electromagnetic Fields, IEEEPress, New York, NY, USA, 2001.

[20] D. R. Rhodes, “Observable stored energies of electromagneticsystems,” Journal of the Franklin Institute, vol. 302, no. 3, pp.225–237, 1976.

[21] D. R. Rhodes, “Reactance Theorem,” Proceedings of the RoyalSociety A, vol. 353, no. 1672, pp. 1–10, 1977.

[22] J. Shaeffer, “MOM3D method of moments code theorymanual,” Research report 189594, Lockheed Advanced Devel-opment Company, 1992.

[23] M. Capek, P. Hazdra, and J. Eichler, “A method for theevaluation of radiation Q based on modal approach,” IEEETransactions on Antennas and Propagation. In press.

[24] P. Arcioni, M. Bressan, and L. Perregrini, “On the evaluationof the double surface integrals arising in the applicationof the boundary integral method to 3-D problems,” IEEETransactions on Microwave Theory and Techniques, vol. 45, no.3, pp. 436–439, 1997.

[25] P. Hazdra, M. Capek, and J. Eichler, “Radiation Q-factors ofthin-wire dipole arrangements,” IEEE Antennas and WirelessPropagation Letters, vol. 10, pp. 556–560, 2011.

[26] A. D. Yaghjian and S. R. Best, “Impedance, bandwidth, and Qof antennas,” IEEE Transactions on Antennas and Propagation,vol. 53, no. 4, pp. 1298–1324, 2005.

[27] CST Gmbh, http://www.cst.com/.

[28] S. N. Sinha and M. Jain, “A self-affine fractal multibandantenna,” IEEE Antennas and Wireless Propagation Letters, vol.6, Article ID 891519, pp. 110–112, 2007.

[29] J. Eichler, P. Hazdra, M. Capek, T. Korınek, and P. Hamouz,“Design of a dual-band orthogonally polarized l-probe-fedfractal patch antenna using modal methods,” IEEE Antennasand Wireless Propagation Letters, vol. 10, pp. 1389–1392, 2011.

[30] C. L. Mak, K. M. Luk, K. F. Lee, and Y. L. Chow, “Experimentalstudy of a microstrip patch antenna with an L-shaped probe,”IEEE Transactions on Antennas and Propagation, vol. 48, no. 5,pp. 777–783, 2000.


Research Article

Tunable Compact UHF RFID Metal Tag Based onCPW Open Stub Feed PIFA Antenna

Lingfei Mo and Chunfang Qin

State Key Laboratory of Industrial Control Technology, Department of Control Science and Engineering, Zhejiang University,Hangzhou 310027, China

Correspondence should be addressed to Lingfei Mo, [email protected]

Received 15 August 2011; Revised 9 December 2011; Accepted 28 December 2011

Academic Editor: Seong-Youp Suh

Copyright © 2012 L. Mo and C. Qin. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

For the ultrahigh frequency radio frequency identification (UHF RFID) metal tag, it always has the difficulties of compactdesigning, especially for the conjugate impedance matching, low antenna gain, and fabrication or environmental detuning. Inthis paper, a tunable compact UHF RFID metal tag is designed based on CPW open stub feed PIFA antenna. By changing thelength of the open stub, the impedance of the PIFA antenna could be tuned in a large scale for conjugate impedance matching. Theopen stub makes it easy to tune the resonant frequency to alleviate the fabrication detuning or the environmental detuning, evenafter the manufacture. Moreover, the CPW structure of the open stub feed can resist the effects of the metallic surface and increasethe antenna gain for the compact PIFA antenna. Modeling analysis and simulation are in good agreement with the measurementresults. It showed that the UHF RFID metal tag could be designed compact with good performance based on the CPW open stubfeed PIFA antenna.

1. Introduction

Ultra high frequency radio frequency identification (UHFRFID) is a long-range noncontact automatic identificationtechnology being widely used around the world recently. AUHF RFID system is generally composed of a reader anda tag. The reader reads the information of the tag throughradio frequency (RF) wave. The tag is composed of a chipand an antenna, with no internal battery. All the energy itneeds is obtained from the RF wave transmitted by the reader[1]. Compared with the traditional bar code tag, the UHFRFID tag can be read and written over a long distance witha very high data rate, no matter whether the tag is soiledor dirty. So, UHF RFID technology has been adopted inthe logistics supply, automatic manufacture, traffic manage-ment, property security, and so forth. In some applications,the UHF RFID tags need to be attached on the surface ofmetallic objects, such as steel plates and steel containers.However, when UHF RFID tags, especially those with dipole-like antennas, are placed on the metallic surface, the readingdistance is reduced quickly, or even cannot be read. This is

because the metallic surface boundary changes the radiationefficiency, impedance matching, resonant frequency, andradiation pattern [2, 3]. For this reason, some special tagswhich can be applied to the metallic surface were designed,called the metal mountable tags or metal tags. Impedancematching is very important for the UHF RFID metal tagantenna design. As the input impedance of the UHF RFIDtag chip is generally complex, the antenna impedance shouldbe designed to be the conjugate impedance of the chip forimpedance matching [1]. Therefore, the input impedanceof the antenna should be adjusted flexibly to achieve goodconjugate impedance matching. Microstrip antenna, patchantenna, and PIFA antenna, which can take the metallicsurface as ground plane, are designed for RFID metal tagswith different impedance matching methods [4–8].

Compact UHF RFID metal tag is widely used in practicalapplications. However, for compact antenna design, theimpedance matching of these antennas is limited becauseof the limited antenna structure. The impedance of thecompact antenna should be able to be tuned flexibly in a largeadjusting scale to match different tag chips. Furthermore,


for commercial UHF RFID tags manufacture, the impedancemay not be matched well between the antenna and the chipbecause of the simulation error and the fabrication variabil-ity, such as the substrate permittivity difference, manufactur-ing difference, and chip impedance difference. To get goodimpedance matching and performance, some manufacturersuse automatic laser or mill machine to adjust the antennaphysical structure to tune the impedance of the antenna.For this purpose, the tag antenna should be designed tohave an easy tuning structure, which is easy for laser millingmachine to adjust and achieving good impedance matching.Impedance tuning is also very useful to alleviate the detuningeffects due to the different metallic application environments[2]. Besides the impedance matching and tuning, antennagain is another challenge for the compact RFID metal tag.Compact RFID metal tags always have lower antenna gainand shorter read range. PIFA antenna is widely used forthe compact UHF RFID tag mountable on metallic objects[9–11]. However, with radiation patch size reduction, theantenna gain decreases and the impedance matching becomedifficult. Using two PIFAs can get better antenna gains [12,13] and flexible impedance matching [5, 14]. But the sizewould increase obviously. Therefore, all of these previousworks cannot fit the requirement of compact profile, easytuning, and satisfied antenna gain.

In this paper, a tunable compact UHF RFID metal tagantenna with a CPW open stub feed is proposed. ThroughPIFA antenna, the tag was compact designed. The impedanceof the antenna could be tuned freely by changing thelength of the open stub [15]. Because of the open stubdesign, the impedance matching can be tuned even afterthe fabrication of the tag. Together with the open stub, theCPW is used to resist the effects of the metallic objects[16] and improve the performance of antenna gain [17]. InSection 2, the considerations of the UHF RFID tag design arediscussed first. Based on these considerations, the proposedantenna is described in Section 3, with antenna structure,theoretical modeling and simulation results. In Section 4,measurement results of a prototype based on this designare also provided. Finally, discussion and conclusions arepresented in Section 5.

2. UHF RFID Tag Design Considerations

A typical passive RFID tag is composed of a chip and anantenna, with no internal battery. All the energy it needsis obtained from the electromagnetic wave transmitted by aRFID reader. In a passive back-scattered UHF RFID system,the reader transmits a modulated signal with periods of un-modulated carry wave, which is received by the antennaof the tag. When the chip of the tag is activated by thepower from the antenna, it will send back its identificationinformation by modulating the backscattered signal. Thebackscattered signal is modulated by switching the loadimpedance of the chip between two states [1]. Figure 1illustrates the operation of back-scattered passive UHF RFIDsystem.

Reader

Chip impedance

Chip

Backscattered wave

Antenna impedance

Antenna

Tag

Power + query

Za Zc

Figure 1: Principle of back-scattered UHF RFID system.

For UHF RFID tag, one of the most important criteria ofperformance is the read range. The maximum read range ofthe tag can be calculated as [18]

r = λ

4π

√PtGtGrτ

Pth, (1)

where λ is the free space wavelength, Pt is the powertransmitted by the reader, Gt is the gain of the antenna ofthe reader, Gr is the gain of the antenna of the tag, τ is thepower transmission coefficient between the tag antenna andthe chip, and Pth is the threshold power of the chip. When thereader and the chip of the tag keep the same, the maximumread range of the UHF RFID tag is mainly determined by thedesign of the tag antenna, especially the gain of the antennaof the tag (Gr) and the power transmission coefficient (τ)[2]. The power transmission coefficient τ is determined bythe impedance matching of the chip and the antenna, whichcan be calculated as follows:

τ = 4RcRa

|Zc + Za|2, 0 ≤ τ ≤ 1, (2)

where Zc = Rc + jXc is the impedance of the tag chip,Za = Ra + jXa is the impedance of the tag antenna. Whenthe impedances of the antenna and the chip are conjugatematching, the transmission coefficient τ could get the maxi-mum value 1 and the most energy will be transmitted fromthe antenna to the chip when the tag is being enquired by thereader.

Besides the gain and the impedance matching, the band-width and the radiation pattern are also important consid-erations for UHF RFID tag antenna design. Wide bandwidthmakes the tag to be read in a required bandwidth and thebroadside radiation pattern makes the tag to be read in awide direction scale [19, 20]. In the realistic application, thesize and shape of the tag must be designed to be embedded orattached to the target objects and have a reliable performance[21–23]. And for a commercial RFID tag, the cost is alsoan important requirement to be considered [24, 25]. Thecost of the RFID tag is a critical factor for this technologyto be widely used around the world. Another requirementfor RFID tag antenna design is the easiness for the massproduction [26]. This includes the antenna manufacture,the chip bonding, the tag package, the performance testing,and the frequency tuning. The frequency tuning is usefulfor reducing the differences of chips and substrate materials,


keeping the tags with the same performance before going tobe used.

The requirements for designing a UHF RFID tag antennaare concluded in Table 1. The proposed tunable compact tagantenna is designed according to these requirements.

3. Proposed Antenna Design

A tunable compact UHF RFID metal tag antenna is proposedin this paper. The PIFA antenna design makes the antennacompact than normal microstrip antenna. With an open stubfeed, the antenna can be conjugate impedance matched withthe chip easily by tuning the inset depth and the open stublength [15]. This metal tag antenna can be fabricated cheaplywith normal PCB (Printed Circuit Board) technology. More-over, with the open stub as tuning structure, the workingfrequency of the tag can be tuned even after the tag has beenmanufactured. The antenna bandwidth, radiation pattern,and metal stability also keep with good performance.

3.1. Antenna Structure. The structure and dimensions of theproposed antenna are illustrated in Figure 2. It is a planarinverted F antenna with a shorting wall to reduce size. Theradiation patch has dimensions of W (20 mm) × L (38 mm)and is printed on a FR4 substrate (εr = 4.4, tan δ = 0.02).The dimension of the substrate is (W + 2 mm) × (L +2 mm) × h (3 mm). The open stub feed line is inset into thepatch to decrease the input impedance of the patch [27]. Theinset structure has a length of Linset and a width of Winset

(8 mm). The open stub feed line has a length of Ls and awidth of Ws (3 mm). The chip is attached on the feed portcomposed by the open stub line and the radiation patch. Inorder to enhance the gain of the compact PIFA antenna, aCPW structure is designed for the open stub feed line. Theantenna is attached on a 200 mm × 200 mm metal plate. Theparameters Linset and Ls are used as variables for impedancematching.

3.2. Theoretical Modeling Analysis. The transmission linemodel of the antenna is shown in Figure 3. From the antennamodel, it is easy to know that the radiation patch and theCPW open stub feed lines are in series. Therefore, the inputimpedance of the feed port of the antenna can be calculatedas

Zin = Z1in + Z2

in, (3)

where Z1in is the input impedance of the radiation patch of the

PIFA antenna, Z2in is the input impedance of the CPW open

stub feed line.According to the basic RF circuit theory [28], the input

impedance of the open stub can be simplified as

Z2in = − jZ2

01

tan(βLs

) = − jZ20

1tan(2πLs/λ)

, (4)

where Z20 is the characteristic impedance of the CPW open

stub feed line, β is the wave number, Ls is the length of theCPW open stub feed line.

Table 1: Requirements for designing a UHF RFID tag antenna.

Requirements of UHF RFIDtag antenna design

Effects of improvement

Read range Cost Reliability

(1) Good impedance matching√

(2) High antenna gain√

(3) Wide bandwidth√ √

(4) Broadside radiation pattern√ √

(5) Compact size and shape√

(6) Low manufacture cost√

(7) Easy for testing and tuning√ √ √

(8) Stable performance for use√ √

Shorting wall

Chip

Radiation patch

CPW open stub

Z

Y

X

h

Ground

Ls

WL

Winset

Linset

Figure 2: The open stub feed PIFA antenna structure. The dimen-sions of the radiation patch are (L,W) while the open stub hasdimensions (Ls,Ws) and the height of the substrate (FR4) betweenthe patch and the metallic surface is h.

The input impedance of the open stub only has imagi-nary part and its function of line length is shown in Figure 4.It shows that the reactance of the CPW open stub feed line iscapacitive when the length is less than 0.25 wavelength andis inductive when the length is between the 0.25 wavelengthand 0.5 wavelength. The reactance of the CPW open stub isa function of cotangent, which means that when the lengthof the open stub changes from 0 to 0.5 λ, the imaginary partof the input impedance of the open stub changes from −∞to +∞. Therefore, the imaginary part of the input impedanceof the antenna can be tuned freely by the length of the openstub in a large scale.

3.3. Simulation and Optimization. In order to get a bet-ter impedance matching for the antenna, Finite-Element-Method (FEM) based computational simulation softwareHFSS 12 is used for the simulation and optimization. Forthe UHF RFID tags, the chips generally have compleximpedance, whose imaginary part is large and negativebecause of the rectifier and energy storage capacitor. Inorder to achieve the maximum energy transfer between theantenna and the chip, the input impedance of the antennaand the chip should be conjugate matching. That is, thereal part is equal, and the imaginary part is opposite. Asthe imaginary part is much larger than the real part of the


Feed

PIFA patchCPW open stub

Ls,β

Z20

Z2in Z1

in

Z1in

Z2L = ∞

Figure 3: Transmission line model of the CPW open stub feed PIFAantenna.

−5

−2.5

0

2.5

5

0 0.1 0.2 0.3 0.4 0.5

Inductive

Capacitive

L/λ

Z2 in/(jZ

0)

Figure 4: The input impedance of the CPW open stub feed line.

impedance, the impedance matching is mainly determinedby the imaginary part matching. So, the antenna shouldbe designed to have a structure easy for impedance tuning,especially for imaginary part tuning. As the proposed tagantenna in this paper is designed for the North America UHFRFID bandwidth (902 MHz∼928 MHz), the tag antennashould have good impedance matching at this bandwidth.The chip used for the tag is the RI UHF 00001 01 UHFRFID chip of TI (Texas Instruments), whose impedance is9.9-j60.3Ω at the frequency of 915 MHz. The structure ofthe antenna is shown in Figure 2. In order to simulate thetag antenna on the surface of metal, the tag is simulated onthe surface of a reference metallic plate of 200 × 200 mm2.According to the relative permittivity of the substrate, thelength of the radiation patch (L) of the PIFA antenna ischosen as 37 mm, which makes the PIFA antenna resonantnear the frequency of 915 MHz. The impedance matchingbetween the antenna and the chip is tuned by Linset andLs.

For patch antennas with the inset feed structure, increas-ing the depth of the inset could decrease the input impedanceof the antenna [27]. Therefore, the length of the inset Linset

can be used to tune the real part of the antenna impedance.Figure 5(a) shows the resistance tuning of the proposedantenna with different inset depths (Linset). The resistance ofthe antenna decreases with the increase of the inset depthLinset. As we analyze above, the CPW open stub feed line can

be used to tune the imaginary part of the input impedanceof the antenna, which is shown in Figure 5(b). The reactanceof the antenna increases with the increases of the CPW openstub length Ls. The imaginary part of the input impedanceof the antenna could be tuned freely from −∞ to +∞ bychanging the length of the CPW open stub from 0 to 0.5 λ.

Therefore, for conjugate impedance matching of theproposal antenna, the resistance and the reactance couldbe tuned freely by the depth of the inset (Linset) and thelength of the CPW open stub (Ls), respectively. Throughsimulation and optimization, the parameters of the antennaare finalized as L = 37 mm, W = 20 mm, h = 3 mm,Linset = 14 mm and Ls = 30 mm. With this dimension, theantenna input impedance and the reflection coefficient S11

are calculated as shown in Figures 6(a) and 6(b), respectively.The imaginary parts of the impedance of the antenna andthe chip are matched well at the frequency of 914 MHz.And the real parts of the impedance are matched at thefrequency of 918 MHz. However, as the imaginary part of theimpedance is much larger than the real part, the impedancematching is dominated by the imaginary part. Under thismatching condition, the reflection coefficient S11 is locatedat the 914 MHz with a value of −26 dB. The 3 dB bandwidthof the antenna is 37 MHz (895 MHz∼932 MHz), covering thebandwidth of the UHF RFID of North American. Moreover,with the decreases of the length of the open stub, the workingfrequency of the antenna is tuned from low to high.

4. Measurement Results

Based on the above-optimized parameters, the antennasample was produced with an FR4 dielectric plate, as shownin Figure 7. The chip was attached to the antenna feed portwith the traditional bonding technology. In order to test andcompare the performance with the simulation results, thetag was also mounted on a 200 × 200 mm2 copper plate.A commercial RFID reader, CSL-461 4-Port EPC Class1Gen2 UHF RFID Reader [29], was used to test the tag. Thebandwidth of the reader is 902 MHz∼928 MHz. The outputpower of the reader can be tuned from 15 dBm to 30 dBm,with a step of 0.25 dBm. The antenna of the reader is CS-771-2-R with a gain of 6 dBi. Combining the output power of thereader and the reader antenna gain, the maximum radiationpower is 36 dBm (4 W EIRP). The reader and the tag aremanufactured with the protocol of EPC Class1 Gen2 andISO 18000-6C. According to the tag performance parametersand test methods of EPCglobal, the performance of the tagwas measured based on the back-scattering method [30].The maximum read range, power bandwidth, and radiationpattern were measured with the same method.

Through the back-scattering method, the best impedancematching frequency (resonant frequency) of the antennacould be measured. Because of the fabrication variability,the best impedance matching frequency is a little lower than915 MHz. Then, as shown in Figures 5(b) and 6(b), throughcutting the length of the open stub, the imaginary part of theimpedance can be decreased and the resonant frequency canbe increased to the target working frequency. In this way, thefabricated tag prototype is optimized by tuning the length


0.86 0.88 0.9 0.92 0.94 0.960

5

10

15

20

Chip resistanceRes

ista

nce

(oh

m)

Linset = 12 mmLinset = 13 mmLinset = 14 mm

Linset = 15 mmLinset = 16 mm

Frequency (GHz)

(a)

0

30

60

90

120

150

Rea

ctan

ce (

ohm

)

Chip reactanceconjugate

= 28 mm= 29 mm= 30 mm

= 31 mm= 32 mm

LsLs

Ls

Ls

Ls

0.86 0.88 0.9 0.92 0.94 0.96

Frequency (GHz)

(b)

Figure 5: Antenna input impedance tuning. (a) Input resistance curves of the antenna with different inset depths Linset (L = 37 mm, W =20 mm, h = 3 mm, Ls = 30 mm). (b) Input reactance curves of the antenna with different open stub length Ls (L = 37 mm, W = 20 mm,h = 3 mm, Linset = 14 mm).

0.86 0.88 0.9 0.92 0.94 0.960

20

40

60

80

100

Chip reactance conjugateAntenna reactance

Reactance matching point

Antenna resistanceResistance matching pointChip resistance

Antenna impedanceChip conjugate impedance

Impe

dan

ce (

ohm

)

Frequency (GHz)

(a)

0.86 0.88 0.9 0.92 0.94 0.96

= 28 mm= 29 mm= 30 mm

= 31 mm= 32 mm

Frequency (GHz)

LsLs

Ls

Ls

Ls

−40

−35

−30

−25

−20

−15

−10

−5

0

Refl

ecti

on c

oeffi

cien

tS 1

1(d

B)

37 MHz

932 MHz895 MHz

914 MHz

(b)

Figure 6: The input impedance and reflection coefficient S11 of the antenna with optimized parameters: (a) impedance, (b) reflectioncoefficient S11.

of the open stub to alleviate the detuning effects due to thefabrication process.

The maximum read range of the tag in North Americabandwidth is plotted in the Figure 8. The antenna has a stableread range in the whole North America UHF RFID bandwith a max value of 4.7 meters at the frequency of 915 MHz.The tested results and the theoretical values are almost the

same. The power bandwidth of the tag was measured inFigure 9. The output power of the reader needed to readthe tags at different frequencies was normalized with theminimum value. The minimum is 0 dB. From Figure 9, it iseasy to calculate that the 3 dB power bandwidth is 903 MHz∼927 MHz, which covers most of the North America UHFRFID bandwidth. Compared with the 3 dB bandwidth of the


Figure 7: Photograph of the fabricated tag antenna.

904 908 912 916 920 924 9280

1

2

3

4

5

6

Testing resultsSimulation results

Frequency (GHz)

Red

ran

ge (

m)

Figure 8: Theoretical and experimental read ranges for the openstub feed patch antenna (EIRP = 4 W).

reflection coefficient S11, the 3 dB power bandwidth is a littlenarrow, but the central bandwidth is almost the same. Themeasured radiation patterns of the tag at the frequency of915 MHz are shown in Figure 10. The antenna has nearlybroadside hemisphere radiation pattern performance at bothE plane and H plane. Tested results agree well with thesimulation results. The tag was attached on the metallicplates of different size to test its metal performance. Thetesting results are plotted in Table 2, which shows that themetal tag has stable read range when it is placed on thesurface of different metallic objects.

5. Discussion and Conclusion

A tunable compact UHF RFID metal tag based on CPW openstub feed PIFA antenna is designed in this paper. Using CPWopen stub feed line, the impedance matching and antennagain can be well designed. Moreover, because of the PIFAand the CPW structure, the antenna has stable performancefor attaching on the surface of metallic objects. The workingfrequency of the antenna can be tuned by milling the length

904 908 912 916 920 924 9280

1

2

3

4

5

6

Frequency (GHz)

24 MHz

Pow

er o

f th

e re

ader

(dB

)

Figure 9: Power bandwidth of the tag antenna.

Table 2: Tag performance testing results with the tag attached ondifferent metal plate.

Size of the metal plate Read ranges

200 mm × 200 mm 4.7 m

500 mm × 500 mm 4.8 m

700 mm × 700 mm 4.3 m

of the open stub even after the tag has been fabricated.This can be used to alleviate the detuning effects of thefabrication error and the metallic application environments.With deceasing the length of the open stub, the imaginarypart of the antenna can be reduced and the workingfrequency can be increased. With increasing the length of theopen stub, the imaginary part of the antenna can be increasedand the working frequency can be decreased.

The testing results were in good agreement with thesimulation. This antenna has stable performance on differentsizes of metallic objects. Four features can be concluded forthis antenna design as follows.

(1) By PIFA antenna design, the size of the tag can beeffectively reduced. The length of the PIFA antennais only one half of that of microstrip antenna.

(2) An open stub feed is used to realize the impedancematching for this compact PIFA antenna. The impe-dance matching between the antenna and the chipcould be achieved easily by tuning the length ofthe CPW open stub feed line. And this impedancematching method could be used with different chipsand input impedances.

(3) With the CPW open stub feed line, the impedancematching of the tag could be tuned even after themanufacture of the tag. This makes it suitable foraccurate impedance matching of the UHF RFID tagfor manufacture and different application environ-ments.


210

180

150

120

90

60

30

0

330

300

270

240

0

−10

−20

−30

−40

−30

−20

−10

0

SimulationTesting

(dB

)

(a)

210

180

150

120

90

60

30

0

330

300

270

240

0

−10

−20

−30

−40

−30

−20

−10

0

SimulationTesting

(dB

)(b)

Figure 10: Radiation patterns of the open stub feed PIFA antenna: (a) E plane, (b) H plane.

(4) The CPW structure of the open stub feed canresist the effects of the metallic surface and increasethe antenna gain for the PIFA antenna, which willkeep the metal tag having a stable performance forattaching on the surface of different metallic objects.

Acknowledgment

This work was supported by the major projects of the Edu-cation Administration (Y200907699), Zhejiang province,China.

References

[1] K. Finkenzeller, RFID Handbook, John Wiley & Sons, NewYork, NY, USA, 2nd edition, 2003.

[2] L. F. Mo, H. J. Zhang, and H. L. Zhou, “Analysis of dipole-like ultra high frequency RFID tags close to metallic surfaces,”Journal of Zhejiang University A, vol. 10, no. 8, pp. 1217–1222,2009.

[3] K. Penttila, M. Keskilammi, L. Sydanheimo, and M. Kivikoski,“Radio frequency technology for automated manufacturingand logistics control. Part 2: RFID antenna utilisation inindustrial applications,” International Journal of AdvancedManufacturing Technology, vol. 31, no. 1-2, pp. 116–124, 2006.

[4] S. J. Kim, B. Yu, Y. S. Chung, F. J. Harackiewicz, and B.Lee, “Patch-type radio frequency identification tag antennamountable on metallic platforms,” Microwave and OpticalTechnology Letters, vol. 48, no. 12, pp. 2446–2448, 2006.

[5] B. Yu, S. J. Kim, B. Jung, F. J. Harackiewicz, and B. Lee, “RFIDTAG antenna using two-shorted microstrip patches mount-able on metallic objects,” Microwave and Optical TechnologyLetters, vol. 49, no. 2, pp. 414–416, 2007.

[6] B. Lee and B. Yu, “Compact structure of UHF band RFIDtag antenna mountable on metallic objects,” Microwave andOptical Technology Letters, vol. 50, no. 1, pp. 232–234, 2008.

[7] K. H. Kim, J. G. Song, D. H. Kim, H. S. Hu, and J. H.Park, “Fork-shaped RFID tag antenna mountable on metallicsurfaces,” Electronics Letters, vol. 43, no. 25, pp. 1400–1402,2007.

[8] H.-W. Son and S.-H. Jeong, “Wideband RFID tag antennafor metallic surfaces using proximity-coupled feed,” IEEEAntennas and Wireless Propagation Letters, vol. 10, pp. 377–380, 2011.

[9] M. Hirvonen, P. Pursula, K. Jaakkola, and K. Laukkanen,“Planar inverted-F antenna for radio frequency identification,”Electronics Letters, vol. 40, no. 14, pp. 848–850, 2004.

[10] H. Kwon and B. Lee, “Compact slotted planar inverted-F RFIDtag mountable on metallic objects,” Electronics Letters, vol. 41,no. 24, pp. 1308–1310, 2005.

[11] W. Choi, H. W. Son, J. H. Bae, G. Y. Choi, C. S. Pyo, andJ. S. Chae, “An RFID tag using a planar inverted-f antennacapable of being stuck to metallic objects,” Electronics andTelecommunications Research Institute Journal, vol. 28, no. 2,pp. 216–218, 2006.

[12] J. S. Kim, W. Choi, and G. Y. Choi, “UHF RFID tag antennausing two PIFAs embedded in metallic objects,” ElectronicsLetters, vol. 44, no. 20, pp. 1181–1182, 2008.

[13] S. L. Chen and K. H. Lin, “A slim RFID tag antenna designfor metallic object applications,” IEEE Antennas and WirelessPropagation Letters, vol. 7, Article ID 2009473, pp. 729–732,2008.

[14] L. Xu, B. J. Hu, and J. Wang, “UHF RFID tag antenna withbroadband characteristic,” Electronics Letters, vol. 44, no. 2, pp.79–81, 2008.

[15] L. Mo and C. Qin, “Planar UHF RFID tag antenna with openstub feed for metallic objects,” IEEE Transactions on Antennasand Propagation, vol. 58, no. 9, Article ID 5484680, pp. 3037–3043, 2010.

[16] C. H. Ku, H. W. Liu, and P. J. Wang, “Novel CPW-fed slotantenna for UHF RFID metal tag applications,” IEICE Elec-tronics Express, vol. 8, pp. 410–415, 2011.


[17] Y. Um, U. Kim, and J. Choi, “Design of a compact CPW-fedUHF RFID tag antenna for metallic objects,” Microwave andOptical Technology Letters, vol. 50, no. 5, pp. 1439–1443, 2008.

[18] K. V. S. Rao, P. V. Nikitin, and S. F. Lam, “Antenna design forUHF RFID tags: a review and a practical application,” IEEETransactions on Antennas and Propagation, vol. 53, no. 12, pp.3870–3876, 2005.

[19] L. Mo, H. Zhang, and H. Zhou, “Broadband UHF RFID tagantenna with a pair of U slots mountable on metallic objects,”Electronics Letters, vol. 44, no. 20, pp. 1173–1174, 2008.

[20] M. Lai and R. Li, “Broadband UHF RFID tag antenna withparasitic patches for metallic objects,” Microwave and OpticalTechnology Letters, vol. 53, no. 7, pp. 1467–1470, 2011.

[21] D. Kim and J. Yeo, “A passive RFID tag antenna installed in arecessed cavity in a metallic platform,” IEEE Transactions onAntennas and Propagation, vol. 58, no. 12, Article ID 5582265,pp. 3814–3820, 2010.

[22] H. D. Chen and Y. H. Tsao, “Low-profile meandered patchantennas for RFID tags mountable on metallic objects,” IEEEAntennas and Wireless Propagation Letters, vol. 9, Article ID5419020, pp. 118–121, 2010.

[23] K. Ide, S. Ijiguchi, and T. Fukusako, “Gain enhancement oflow-profile, electrically small capacitive feed antennas usingstacked meander lines,” International Journal of Antennas andPropagation, vol. 2010, Article ID 606717, 8 pages, 2010.

[24] S. Merilampi, L. Ukkonen, L. Sydanheimo, P. Ruuskanen, andM. Kivikoski, “Analysis of silver Ink bow-Tie RFID tag anten-nas printed on paper substrates,” International Journal ofAntennas and Propagation, vol. 2007, Article ID 90762, 9 pages,2007.

[25] L. Yang, A. Rida, R. Vyas, and M. M. Tentzeris, “Novel“Enhanced-Cognition” RFID architectures on organic/paperlow-cost substrates utilizing inkjet technologies,” InternationalJournal of Antennas and Propagation, vol. 2007, Article ID68385, 7 pages, 2007.

[26] G. Orecchini, F. Alimenti, V. Palazzari, A. Rida, M. M.Tentzeris, and L. Roselli, “Design and fabrication of ultra-lowcost radio frequency identification antennas and tags exploit-ing paper substrates and inkjet printing technology,” IETMicrowaves, Antennas and Propagation, vol. 5, no. 8, pp. 993–1001, 2011.

[27] R. Garg, P. Bhartia, L. Bahl, and A. Ittipiboon, MicrostripAntenna Design Handbook, Artech house publishers, London,UK, 2001.

[28] R. Ludwig and P. Bretchko, RF Circuit Design: Theory andApplications, Prentice-Hall, 2000.

[29] CSL. CSL-461 4-Port EPC Class1 Gen2 UHF RFID Reader,http://www.convergence.com.hk/.

[30] Tag Performance Parameters and Test Methods, EPCglobal Inc.,2008.


Review Article

Some Recent Developments of Microstrip Antenna

Yong Liu, Li-Ming Si, Meng Wei, Pixian Yan, Pengfei Yang, Hongda Lu, Chao Zheng,Yong Yuan, Jinchao Mou, Xin Lv, and Housjun Sun

Department of Electronic Engineering, School of Information and Electronics, Beijing Institute of Technology, Beijing 100081, China

Correspondence should be addressed to Yong Liu, [email protected]

Received 12 August 2011; Revised 23 November 2011; Accepted 2 January 2012

Academic Editor: Zhongxiang Q. Shen

Copyright © 2012 Yong Liu et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Although the microstrip antenna has been extensively studied in the past few decades as one of the standard planar antennas, it stillhas a huge potential for further developments. The paper suggests three areas for further research based on our previous workson microstrip antenna elements and arrays. One is exploring the variety of microstrip antenna topologies to meet the desiredrequirement such as ultrawide band (UWB), high gain, miniaturization, circular polarization, multipolarized, and so on. Anotheris to apply microstrip antenna to form composite antenna which is more potent than the individual antenna. The last is growingtowards highly integration of antenna/array and feeding network or operating at relatively high frequencies, like sub-millimeterwave or terahertz (THz) wave regime, by using the advanced machining techniques. To support our points of view, some examplesof antennas developed in our group are presented and discussed.

1. Introduction

The concept of microstrip antenna was first introduced in the1950s [1]. However, this idea had to wait nearly 20 years tobe realized after the development of the printed circuit board(PCB) technology in the 1970s [2, 3]. Since then, microstripantennas are considered as the most common types ofantennas due to their obvious advantages of light weight,low cost, low profile, planar configuration, easy of conformal,superior portability, suitable for arrays, easy for fabrication,and easy integration with microwave monolithic integratecircuits (MMICs) [4–7]. They have been widely employedfor the civilian and military applications such as television,broadcast radio, mobile systems, global positioning system(GPS), radio-frequency identification (RFID), multiple-input multiple-output (MIMO) systems, vehicle collisionavoidance system, satellite communications, surveillance sys-tems, direction founding, radar systems, remote sensing, bi-ological imaging, missile guidance, and so on [8].

Despite the many advantages of typical microstrip anten-nas, they also have three basic disadvantages: narrow band-width, low gain, and relatively large size. The narrow band-width is one of the main drawbacks of these types ofantennas. A straightforward method of improving the band-width is increasing the substrate thickness. However, surface

wave power increases and radiation power decreases withthe increasing substrate thickness [7], which leads to poorradiation efficiency. Thus, various other techniques are pre-sented to provide wide-impedance bandwidths of microstripantennas, including impedance matching networks usingstub [9, 10] and negative capacitor/inductor [11], microstripslot antennas using the U, L, T, and inverted T slots in theground plane (sometimes termed defected ground structures(DGSs)) [12, 13], surface wave suppressing using magneto-dielectric substrate [14] and electromagnetic bandgap (EBG)structures [15], and composite-resonator microstrip anten-nas using metamaterial resonators [16, 17]. Another problemto be solved is the low gain for conventional microstripantenna element. Cavity backing has been used to eliminatethe bidirectional radiation, thereby providing higher gaincompared with conventional microstrip antenna [18]. Lenscovering is an alternative way to achieve gain enhancement.The lens with canonical profile, like elliptical, hemielliptical,hyper-hemispherical, extended hemispherical, used to focusthe radiation beam from the radiator elements. The inte-grated lens microstrip antenna can be treated as compositeantenna combined by microstrip radiator elements anddielectric lens, which is very useful for high frequencies (mm,sub-mm, terahertz (THz), and optical waves) applications


[19]. It is also well known that antenna array is an effectivemeans for improving the gain [20–25].

The last limitation of conventional microstrip antennasis the relatively large size, particularly at lower microwavefrequencies, since their operation frequencies are related tothe electrical size of antenna. In general, the size of the rec-tangular microstrip antenna should be of order of a half-guided wavelength. This limitation was mathematically stud-ied by Wheeler [26] and Chu [27]. There have been nu-merous efforts to minimize the antenna size and obtainthe electrically small microstrip antenna with the raised de-mand towards smaller and smaller wireless devices. Inductiveor capacitive loading are effective ways to reduce the sizeof microstrip antennas [28]. In the former work, we dem-onstrated that the size of microstrip antenna can be minia-turized using composite metamaterial resonators [16, 17].Magneto-dielectric substrates have been widely used tominiaturize microstrip antennas due to magnetic substratesand could provide wider bandwidths than dielectric sub-strates [29–32]. Fractal geometries, which are composed byself-similar structures, have opened an alternative way forantenna miniaturization [33].

From the above discussions, we see that many methodsand materials are used to improve the properties of mi-crostrip antennas. However, there should be a relationshipamong bandwidth, gain, and size of the microstrip antennas.Antenna engineers have recognized that the improvement inone antenna property is frequently accompanied by declinein its other performances. For example, the antenna size isreduced usually at the expense of its bandwidth and gain.Therefore, a more comprehensive consideration must begiven on further developments of microstrip antennas.

In this paper, we will suggest three areas for furtherresearch based on our previous works on microstrip antennaelements and arrays [16–25, 34–41]. We first note that novelmicrostrip antenna topologies are proposed to meet thedesired requirement of variety of potential wireless appli-cations, such as ultrawide band (UWB), high gain, minia-turization, circular polarization, multipolarized, and so on.Next, we discuss the composite antennas based on microstripantennas which have more potent than each individualantenna. Finally, with the development of micro-/nano-machining techniques, antennas/arrays with highly integra-tion and with highly operating frequencies are discussed. Wepresent some examples of antennas developed in our groupto support our points of view.

2. Variety of Microstrip Antenna Topologies

Microstrip antennas have extensively used in commercial andmilitary applications due to their attractive advantages. How-ever, the traditional microstrip antennas have the impedancebandwidth of only a few percent and radiation pattern withomnidirection, which obviously does not meet the require-ments of various wireless applications. To this end, a widevariety of microstrip antenna topologies, including differentmicrostrip antenna element structures and different micros-trip array arrangements, have been studied to meet thedesired requirement such as ultrawide band (UWB), high

Monopole

Reflector

Feeding line

Ground plane

Director

(a) The structure of the quasi-Yagi antenna

(b) The photograph of the quasi-Yagi antenna

Figure 1: Compact broad-band quasi-Yagi antenna.

gain, miniaturization, circular polarization, multipolarized,and so forth.

As we know, microstrip antennas inherently have nar-rower bandwidth and lower gain compared to conventionalbulky antennas. Some microstrip antennas with specialtopologies, like quasi-Yagi, planar reflector antenna, areproposed to replace the conventional bulky antennas. Here,we will take a quai-Yagi antenna as an example to show howto design a planar microstrip antenna with Yagi-Uda end-fireradiation pattern. In addition, a microstrip array with specialarray topology is designed to get dual-polarized property.

2.1. Compact Broad-Band Quasi-Yagi Antenna. A novel S-band compact quasi-Yagi antenna has been designed, fabri-cated and measured by our group, as shown in Figure 1. Thisantenna is composed of a printed monopole-driven element,a printed reflector element, and six printed director elements.

To explain the end-fire radiation behavior of the quasi-Yagi antenna, a comparison of radiation patterns, among(1) microstrip monopole only, (2) microstrip monopole anda reflector, (3) microstrip monopole and a director, (4)microstrip monopole and a reflector with one director, and(5) microstrip monopole and a reflector with six director, isshown in Figure 2. We can observe that both the reflectorand the director can increase the end-fire radiation, and itcould be substantially improved by increasing the number ofdirectors.

The measured VSWR results are shown in Table 1. Abandwidth of 14% for VSWR less than 1.5 is achieved. Thegain of the antenna is above 7.5 dBi, as shown in Table 2. Inthis design, we see that the microstrip antenna with specialtopology could be conveniently used to replace the bulkyYagi-Uda antenna.


Figure 2: Radiation patterns of microstrip monopole only, mi-crostrip monopole and a reflector, microstrip monopole and adirector, microstrip monopole and a reflector with one director, andmicrostrip monopole and a reflector with six director.

Table 1: The measured VSWR of the quasi-Yagi antenna.

No.Frequency (GHz)

3.25 3.5 3.75 Inband

1 1.36 1.34 1.47 <1.5

2 1.37 1.26 1.49 <1.5

3 1.36 1.25 1.48 <1.5

Table 2: The measured gain of the quasi-Yagi antenna (unit: dBi).

No.Frequency (GHz)

3.25 3.5 3.75

1 7.57 8.73 8.35

2 7.58 8.55 8.37

3 7.56 8.77 8.51

2.2. Dual-Polarized Microstrip Antenna Array. The dual-polarized antenna is highly required for the radar, electroniccountermeasure, and aerospace systems. It is known that themicrostrip antenna can easily be integrated with microwavecircuits and feeding network. Here, a novel Ku-band dual-polarization microstrip antenna array with a mixed feedingnetwork, that is, the slot coupled feeding (V-port) and the co-plane feeding (H-port), is designed by our group, as shownin Figure 3. It is a three layers structure: top microstrip patchlayer, middle stripline feeding network layer, and bottomcoplane microstrip feeding network layer. Through properarray arrangement, very good isolation can be obtained.

V

H

0 25 50

(mm)

Stripline feeding network

Patch

Stripline-microstrip transition

Microstrip feeding network

H-shape coupled slot

(a) The structure of the dual-polarized microstrip antenna array

(b) The photograph of the dual-polarized microstripantenna array

Figure 3: Dual-polarized microstrip antenna array.

15 15.2 15.4 15.6 15.8

Frequency (GHz)

H-portV-port

3

2.5

2

1.5

1

VSW

R

Figure 4: The VSWR of the dual-polarized microstrip antennaarray.

The VSWR, radiation patterns, and the isolation betweentwo polarizations of the proposed dual-polarized microstripantenna array are shown in Figures 4, 5, and 6, respectively.The results indicate that this microstrip antenna array has agood impedance matching, good radiation performance, aswell as very high isolation (less than −25 dB), which can bean idea candidate for the dual-polarized wireless systems.


Figure 5: The radiation patterns of the dual-polarized microstrip antenna array at the center frequency.

Figure 6: The isolation of the dual-polarized microstrip antennaarray.

3. Microstrip-Antenna-BasedComposite Antenna

As many antenna designers have found, it is not easy todesign an antenna to meet the user-defined stringent per-formance requirements demanded by special wireless appli-cations like military radars, surveillances, and missile guid-ance, if only one type of antenna is considered. This dif-ficulty may require the use of two more different types orstructures of antenna elements with different characteristics.Composite antenna formed by two more types or structuresof antennas is particularly suitable for these applications dueto more advantages offered by different types or structures ofantennas. For example, it is a challenging task to use singletype of antenna to design a dual-band dual-polarizationantenna for satellite digital multimedia broadcast (S-DMB)application [36]. A composite antenna composed with aleft-handed circularly polarized (LHCP) microstrip antennaand a linear polarized omnidirectional biconical antenna

Slot

Dielectric

Groove guide

Figure 7: The structure of the DCWS.

is proposed by our group to meet this requirement [36].Another example of composite antenna is comprised of adielectric lens and microstrip log-period antenna, which hasbeen widely applied to THz systems (this type of antennawill be further discussed in Section 4.2). Here, we will givean example of composite antenna with “structure composite”method.

3.1. Monopulse Circular-Polarized Dielectric Complex Waveg-uide Slot Antenna Array. Waveguide slot antenna array hasbeen widely used for wireless system, due to its advantagesof high radiation efficiency, high power capacity, and highreliability. However, it is hard to overcome the disadvantageof high cost of fabrication.

One composite antenna with waveguide slot antennaarray property, termed dielectric complex waveguide slot(DCWS), is composed with slot microstrip line and grooveguide, as shown in Figure 7. The slot microstrip line isformed by a metal clad dielectric substrate and slots etchedin the metal. This composite antenna not only maintains theadvantages of the traditional waveguide slot antenna arraybut also has the characteristics of high consistence, easy forfabrication, and low cost.


Circular polarization gridSlot array

Groove guide

Feeding network

X Y

Z

(a) The structure of the monopulse circular-polarized DCWS antennaarray (separating view)

(b) The photograph of the monopulse circular-polarizedDCWS antenna array.

Figure 8: Ka-band monopulse circular-polarized dielectric com-plex waveguide slot (DCWS) antenna array.

A Ka-band monopulse circular-polarized dielectric com-plex waveguide slot (DCWS) antenna array is designed,fabricated, and measured by our group, as shown in Figure 8.It consists of a circular polarization grid, a slot microstriparray, and a groove guide and feeding network. The slotmicrostrip array is fabricated on a Rogers 5880 film withdielectric constant of 2.2 and the thickness of 0.254 mm. Themeasured results of VSWR of sum and different port areshown in Figure 9. Figure 10 shows the measured radiationpattern at the center frequency. Some important arrayperformance parameters such as gain, null depth and axialratio (AR) are also given in Table 3. As shown in themeasured results, very good performance can be obtainedwith the DCWS antenna array. The radiating efficiency ofthe DCWS antenna array is 80%, which is almost the sameas the traditional waveguide slot antenna array. Moreover,the DCWS antenna array has 40% larger bandwidth than thetraditional waveguide slot antenna array.

4. Highly Integration and Highly OperatingFrequency Antennas Based on AdvancedMachining Techniques

It is known that the microstrip antenna was first fabricatedusing PCB technology in 1970s, nearly 20 years after its

Figure 9: The VSWR of sum and difference port of the monopulsecircular-polarized DCWS antenna array.

Figure 10: The radiation pattern of the monopulse circular-polarized DCWS antenna array at the center frequency.

concept was first presented in 1950s [1–3]. Clearly, thedevelopment of microstrip antennas is closely related withthe machining techniques. Recently, various machining tech-niques, including multilayer printed circuit board (MPCB),complementary metal oxide semiconductor (CMOS), low-temperature cofired ceramics (LTCC), and micro-electro-mechanical systems (MEMS), are highly developed, openingopportunities for innovative antennas, such as active anten-nas, reconfigurable antennas, metamaterial-based antennas,THz antennas, and so forth. With the availability of high-precision and high-speed advanced machining techniques,microstrip antennas are growing towards highly integrationof antenna/array and feed network and operating at relativelyhigh frequencies. Since they are all based on the advanced


Parasitic patch

Driven patch

Upper-ground coupled slot

Feeding stripline

Lower ground

Layer 1

Layer 2

Layer 3

Layer 4

Layer 5

h1, εr1

h2, εr2

h3, εr3

h4, εr4

h5, εr5

(a) Schematic side view of the structure of the highintegrate broadband microstrip antenna array

(b) The photograph of the high integrate broadbandmicrostrip antenna array

Figure 11: Ku-band high integrate broadband microstrip antennaarray using MPCB technology.

2

1.8

1.6

1.4

1.2

1

Frequency (GHz)

VSW

R

15.4 15.6 15.8 16 16.2 16.4 16.6 16.8 17

1#2#

Figure 12: The VSWR of the high integrate broad-band microstripantenna array using MPCB technology.

machining techniques, we suggest that a third research areaof microstrip antennas is constantly introducing novel ad-vanced machining techniques. In the following, two exam-ples will be presented to show how important the advancedmachining technique is to fabricate microstrip antennas. Oneis the highly integrate broad-band microstrip antenna arrayfabricated using MPCB technology. Another is THz waveplanar integrated active microstrip antenna using MEMStechnology.

Frequency (GHz)

15.4 15.6 15.8 16 16.2 16.4 16.6 16.8 17

1#2#

22

21

20

19

18

Gai

n (

dBi)

Figure 13: The gain of the high integrate broad-band microstripantenna array using MPCB technology.

Table 3: The measured data of the monopulse circular-polarizedDCWS antenna array.

Fre. (GHz) Gain (dBi) Null depth (dB) AR (dB)

f0 − 0.2 22.8 −37.3 3.8

f0 21.9 −29.9 2.9

f0 + 0.2 22.1 −26 4.1

4.1. High Integrate Broad-Band Microstrip Antenna ArrayUsing Multilayer Printed Circuit Board (MPCB) Technology.Recently, with the development of the multilayer printedcircuit board (MPCB) technology, the microstrip antennascan be designed and fabricated from one-dimensional (1D)to 2D and even 3D structures.

Based on the MPCB technology, a high integrated broad-band Ku-band microstrip antenna array is designed, fabri-cated, and measured by our group, as shown in Figure 11.This antenna consists of a parasitic patch, a driven patch,a stripline feeding network, a broad-band coaxial line tostripline transition, some buried screw holes, and some viaholes. The feeding network is integrated in the bottom of thesubstrate of the antenna. As all of the structures fabricated atonce, the accuracy and the uniformity can be assured. Twoantennas of this type are measured. The measured VSWR,gain, and radiation pattern at the center frequency are shownin Figures 12, 13, and 14, respectively. The measured resultsshow that this antenna maintains good radiation and match-ing performances with relative bandwidth of 13%. They havealso shown good uniformity by using MPCB technology.

4.2. THz Wave Planar Integrated Active Microstrip AntennaUsing Micro-Electromechanical Systems (MEMSs) Technology.THz waves typically include frequencies between 0.1 THzand 10 THz. THz technology is now becoming a promisingtechnology which has potential applications in many fields,such as short-range communication, biosensor, imaging,


(a) E-plane

−80 −60 −40 −20 0 20 40 60 80

θ (◦)

0

−10

−20

−30

−40

Gai

n (

dB)

(b) H-plane

Figure 14: The radiation pattern of the high integrate broad-band microstrip antenna array using MPCB technology at the centerfrequency.

(a) The photograph of the THz monolithic antenna

(b) The photograph of the THz monolithic antennacovered by a dielectric lens

Figure 15: THz wave planar integrated active microstrip antennausing micro-electromechanical systems (MEMSs).

national security, space exploration and communication,and so forth [39–46]. To realize THz transceiver system,antenna is an essential component. We often use horn anten-na, lens antenna, and dielectric parabolic antenna, for THzsystems. However, they are not easy to integrate with mono-lithic integrate circuits. Although the microstrip antennahas the merits of small volume, light weight, and easy

integration with circuit, it is difficult to be processed insuch high-frequency regions. MEMS technology opens theway to design of THz antennas, circuits, and systems. THzmonolithic antenna fabricated using MEMS technology andcovered by a dielectric lens, which can be considered acomposite antenna, are designed, fabricated, and measuredby our group, as shown in Figure 15.

Diodes have the functions of mixing and/or modulatingthe carrier-wave signal. It is an effective way to reduce thepropagation path for detectors application by integratingthe diode and microstrip antenna. The extended hyper-hemispherical dielectric lens is used to increase the gainof the microstrip antenna. An antenna-coupled detectorintegrated with a dielectric lens is designed and fabricated upto THz range by our group. The planar microstrip log-spiralantenna and log-period antenna have been fabricated usingmicro-electromechanical systems (MEMSs) technology. Thephotographs of the antennas are demonstrated in Figure 15.The measured responses of the antenna-coupled detectorworking at different frequency bands are shown in Figure 16,which can be considered to determine the effective operatingfrequencies [19, 40]. This detector gave a valid responsefrom 12 GHz to 110 GHz frequencies. The results prove thevalidity and feasibility of the THz antenna designed usingmicro-electromechanical systems (MEMSs ) technology.

5. Conclusion

The advantages and disadvantages of microstrip antennas arediscussed in this paper. In particular, three areas for furtherdevelopment of microstrip antennas are presented basedon our previous works on microstrip antenna elements andarrays. Variety of microstrip antenna topologies and micros-trip-antenna-based composite antenna are discussed, and


12 14 16 18 20 22

1400

1200

1000

800

600

Res

pon

sivi

ty (

mV

/mW

)

Frequency (GHz)

(a) Ku-band

1200

1000

800

600

40026 28 30 32 34 36 38 40 42

Res

pon

sivi

ty (

mV

/mW

)

Frequency (GHz)

(b) Ka-band

500

400

300

200

100

050 55 60 65 70 75

Res

pon

sivi

ty (

mV

/mW

)

Frequency (GHz)

(c) V-band

200

150

100

50

075 80 85 90 95 100 105 110 115

Res

pon

sivi

ty (

mV

/mW

)

Frequency (GHz)

(d) W-band

Figure 16: Frequency responses test results of the THz wave planar integrated active microstrip antenna covered by a dielectric lens.

the advanced machining techniques pushing the microstripantennas towards the highly integration of antenna/array andfeeding network and the highly operating frequencies aredescribed. To demonstrate the distinctive features of novelmicrostrip antennas, various antenna elements and arrays fordifferent applications are presented. This paper has shownthat the microstrip antennas are still very promising para-digm for civilian and military wireless applications.

References

[1] G. A. Deschamps, “Microstrip microwave antennas,” in Pro-ceedings of the 3rd USAF Symposium on Antennas, 1953.

[2] E. V. Byron, “A new flush-mounted antenna element forphased array application,” in Proceedings of the Phased ArrayAntenna Symposium, pp. 187–192, 1970.

[3] J. Q. Howel, “Microstrip antennas,” in Proceedings of the Digestof the International Symposium of the Antennas and propaga-tion Society, pp. 177–180, 1972.

[4] I. J. Bahl and P. Bhartia, Microstrip Antennas, Artech House,1980.

[5] J. R. James, P. S. Hall, and C. Wood, Microstrip Antenna Theoryand Design, Peter Peregrinus, 1981.

[6] K. R. Carver and J. W. Mink, “Microstrip antenna technology,”IEEE Transactions on Antennas and Propagation, vol. 1, no. 1,pp. 2–24, 1981.

[7] D. M. Pozar, “Microstrip antennas,” Proceedings of the IEEE,vol. 80, no. 1, pp. 79–91, 1992.

[8] R. Garg, I. Bhartia, I. Bahl, and A. Ittipiboon, Micro-Strip An-tenna Design Handbook, Artech House, Boston, Mass, USA,2001.

[9] D. M. Pozar and B. Kaufman, “Increasing the bandwidth of amicrostrip antenna by proximity coupling,” Electronics Letters,vol. 23, no. 8, pp. 368–369, 1987.

[10] H. F. Pues and A. R. Van de Capelle, “Impedance-matchingtechnique for increasing the bandwidth of microstrip anten-nas,” IEEE Transactions on Antennas and Propagation, vol. 37,no. 11, pp. 1345–1354, 1989.

[11] A. Kaya and E. Y. Yuksel, “Investigation of a compensatedrectangular microstrip antenna with negative capacitor andnegative inductor for bandwidth enhancement,” IEEE Trans-actions on Antennas and Propagation, vol. 55, no. 5, pp. 1275–1282, 2007.

[12] K. L. Wong and N. H. Hsu, “A broad-band rectangular patchantenna with a pair of wide slits,” IEEE Transactions on Anten-nas and Propagation, vol. 49, no. 9, pp. 1345–1347, 2001.


[13] S. I. Latif, L. Shafai, and S. K. Sharma, “Bandwidth enhance-ment and size reduction of microstrip slot antennas,” IEEETransactions on Antennas and Propagation, vol. 53, no. 3, pp.994–1003, 2005.

[14] P. M. T. Ikonen, S. I. Maslovski, C. R. Simovski, and S. A.Tretyakov, “On artificial magnetodielectric loading for im-proving the impedance bandwidth properties of microstripantennas,” IEEE Transactions on Antennas and Propagation,vol. 54, no. 6, pp. 1654–1662, 2006.

[15] R. Cocooli, F. R. Yang, T. P. Ma, and T. Itoh, “Apertu-recoupledpatch antenna on UCPBG substrate,” IEEE Transactions onMicrowave Theory and Techniaues, vol. 1, no. 47, pp. 2123–2130, 1999.

[16] L.-M. Si and X. Lv, “CPW-FED multi-band omni-directionalplanar microstrip antenna using composite metamaterial res-onators for wireless communications,” Progress in Electromag-netics Research, vol. 83, pp. 133–146, 2008.

[17] L. M. Si, H. Sun, Y. Yuan, and X. Lv, “CPW-fed com-pact pla-nar UWB antenna with circular disc and spiral split ring res-onators,” in Progress in Electromagnetics Research Symposium,pp. 502–505, 2009.

[18] Y. Yuan, L.-M. Si, Y. Liu, and X. Lv, “Integrated log-periodic antenna for Terahertz applications,” in InternationalConference on Microwave Technology and Computational Elec-tromagnetics, pp. 276–279, 2009.

[19] L. Xin, M. Jinchao, Y. Yong, Y. Weihua, N. Hongbin, andG. Yafen, “Integrated active antennas in terahertz focal planeimaging system,” in Proceedings of the International Workshopon Antenna Technology, pp. 157–163, 2011.

[20] Q. Xu, W. Zhang, H. Sun, and X. Lu, “Millimetre wave multi-polarised microstrip antenna array and application example,”IET Microwaves, Antennas and Propagation, vol. 4, no. 10, pp.1525–1530, 2010.

[21] M. Wei, H. Deng, H. Sun, and Y. Liu, “Design of an X/Kadual-band co-aperture broadband microstrip antenna array,”in Proceedings IEEE International Conference on MicrowaveTechnology and Computational Electromagnetics, pp. 217–220,2011.

[22] Y. Liu, X. Lu, Y. Yuan, Y. Chen, and L. Shi, “Integrated designand research of Ka-band electronically large mono-pulseplanar antenna and feed system,” in Proceedings of the 7thInternational Symposium on Antennas, Propagation and EMTheory (ISAPE ’06), pp. 150–153, October 2006.

[23] Y. Liu, H. J. Sun, X. Lu et al., “Study of the millimetermicrostrip monopulse antenna array in the dual-sensorsystem,” in Proceedings of the 7th International Symposium onAntennas, Propagation and EM Theory (ISAPE ’06), pp. 157–160, October 2006.

[24] Y. An, X. Lv, and B. Gao, “Developing a kind of mi-crostriparray antenna with beam squint,” in Proceedings of the 5thInternational Symposium on Antennas, Propagation and EMTheory, pp. 443–446, 2000.

[25] L. Yong, L. Xin, H. Zhao, and Y. Yuan, “Integrated design andresearch of novel KA-band circular polarized monopulse inter-rogator array antenna,” in Proceedings of the IET InternationalRadar Conference 2009, pp. 1–4, April 2009.

[26] H. A. Wheeler, “Fundamental limitations of small an-tennas,”Proceedings of the IRE, vol. 35, pp. 1479–1484, 1947.

[27] L. J. Chu, “Physical limitations of omni-directional antennas,”Journal of Applied Physics, vol. 19, no. 12, pp. 1163–1175, 1948.

[28] R. Azadegan and K. Sarabandi, “A novel approach for minia-turization of slot antennas,” IEEE Transactions on Antennasand Propagation, vol. 51, no. 3, pp. 421–429, 2003.

[29] H. Mosallaei and K. Sarabandi, “Magneto-dielectrics in elec-tromagnetics: concept and applications,” IEEE Transactions onAntennas and Propagation, vol. 52, no. 6, pp. 1558–1567, 2004.

[30] P. M. T. Ikonen, S. I. Maslovski, C. R. Simovski, and S. A.Tretyakov, “On artificial magnetodielectric loading for im-proving the impedance bandwidth properties of microstripantennas,” IEEE Transactions on Antennas and Propagation,vol. 54, no. 6, pp. 1654–1662, 2006.

[31] P. M. T. Ikonen, K. N. Rozanov, A. V. Osipov, P. Alitalo, andS. A. Tretyakov, “Magnetodielectric substrates in antenna min-iaturization: potential and limitations,” IEEE Transactions onAntennas and Propagation, vol. 54, no. 11, pp. 3391–3399,2006.

[32] P. Ikonen and S. Tretyakov, “On the advantages of magneticmaterials in microstrip antenna miniaturization,” Microwaveand Optical Technology Letters, vol. 50, no. 12, pp. 3131–3134,2008.

[33] J. Anguera, C. Puente, C. Borja, and J. Soler, “Frac-tal-shaped antennas: a review,” in Wiley: Encyclopedia of RF andMicrowave Engineering, K. Chang, Ed., vol. 2, pp. 1620–1635,2005.

[34] L. Shi, H. Sun, X. Lu, C. Deng, and Y. Liu, “Study of theimproved MM-Wave omni-directional microstrip antenna,”in Proceedings of the 7th International Symposium on Antennas,Propagation and EM Theory (ISAPE ’06), pp. 238–241, Octo-ber 2006.

[35] L. Shi, H. Sun, and X. Lu, “A composite antenna withwide circularly polarized beamwidth,” Microwave and OpticalTechnology Letters, vol. 51, no. 10, pp. 2461–2464, 2009.

[36] L. Shi, H. Sun, W. Dong, and X. Lv, “A dual-band multi-function carborne hybrid antenna for satellite communicationrelay system,” Progress in Electromagnetics Research, vol. 95, pp.329–340, 2009.

[37] Y. Chen, H. Sun, and X. Lv, “Novel design of dual-polarizationbroad-band printed L-shaped probe fed microstrip patchantenna for SAR applications,” in IET International RadarConference 2009, pp. 1–4, April 2009.

[38] L. M. Si, H. J. Sun, and X. Lv, “Numerical simulationsof backward-to-forward leaky-wave antenna with compositeright/left-handed coplanar waveguide,” Chinese Physics Letters,vol. 27, no. 3, Article ID 034106, 2010.

[39] L.-M. Si, H.-J. Sun, and X. Lv, “THz leaky-wave antenna withhigh-directivity and beam-steering using CPW CRLH meta-material resonators,” in Proceedings of the 3rd InternationalSymposium on Photoelectronic Detection and Imaging, Beijing,China, 2009.

[40] Y. Liu, L. M. Si, S. Zhu, and H. Xin, “Experimental realizationof an integrated THz electromagnetic crys-tals (EMXT) H-plane horn antenna,” Electronics Letters, vol. 47, no. 2, pp. 80–82, 2011.

[41] L. M. Si, Y. Liu, S. H. Zhu, and H. Xin, “Integrated THzhorn antenna using EBG structures,” in Proceedings of the IEEEInternational Symposium on Antennas and Propagation andUSNC/CNC/URSI National Radio Science Meeting, Universityof Colorado at Boulder, 2011.

[42] Y. Liu, Y. Yuan, X. Lv, and L.-M. Si, “Design of 0.5 THz2D square lattice EBG waveguide transmission line andpower-divider using MEMS technology,” in Proceedings 3rdInternational Symposium on Photoelectronic Detection andImaging, vol. 7385, Beijing, China, 2009.

[43] L.-M. Si, Y. Yong, H.-J. Sun, and L. Xin, “Characterization andapplication of planar terahertz narrow bandpass filter withmetamaterial resonators,” in Proceedings of the International


Workshop on Metamaterials, pp. 351–354, Nanjing, China,2008.

[44] L. M. Si and X. Lv, “Terahertz waves hairpin microstrip band-pass filter and its application to overlaid dielectric materialdetection,” Modern Physics Letters B, vol. 22, no. 29, pp. 2843–2848, 2008.

[45] L.-M. Si, H.-J. Sun, and X. Lv, “Theoretical investigation ofterahertz amplifier by carbon nanotubes within transmissionline metamaterials,” Microwave and Optical Technology Letters,vol. 53, no. 4, pp. 815–818, 2011.

[46] W. Hong, “Millimeter wave and THz communications inChina,” in Proceedings of the IEEE International MicrowaveSymposium Digest (MTT ’11), pp. 1–4, 2011.


Research Article

New Configurations of Low-Cost Dual-Polarized PrintedAntennas for UWB Arrays

Guido Valerio,1 Simona Mazzocchi,2 Alessandro Galli,2

Matteo Ciattaglia,3 and Marco Zucca3

1 Institut d’Electronique et des Telecommunications de Rennes (IETR), UMR CNRS 6164, Universite de Rennes 1,35042 Rennes Cedex, France

2 Department of Information Engineering, Electronics and Telecommunications, Sapienza University of Rome, 00184 Rome, Italy3 SELEX S.I. S.p.A., Sistemi Radianti, 00131 Rome, Italy

Correspondence should be addressed to Guido Valerio, [email protected]

Received 15 August 2011; Revised 26 November 2011; Accepted 6 December 2011

Academic Editor: Athanasios Panagopoulos

Copyright © 2012 Guido Valerio et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A novel class of structures is proposed to realize ultra-wide-band radiating elements for large arrays, providing dual polarization,beam scanning, and compact and inexpensive realization based on suitable rhombic arrangements of dipoles printed on low-costlayered substrates. In a first implementation, four rhombic shapes, orthogonally placed on the same layer, provide two orthogonalpolarizations. In a second implementation, the two polarizations are excited by two rhombic shapes printed on two different layersin a stacked-patch-like arrangement. This latter structure leads to a better lateral shielding of the single radiating element, in orderto reduce mutual interactions among adjacent elements in array environment. The behavioral features of these antennas havebeen tested with various parametric analyses. Practical aspects have been addressed such as the choice of appropriate feeding andof commercially available dielectric layers. The resulting antennas are matched at the input ports in an extremely wide range offrequencies (5–25 GHz), covering various microwave applications, such as aircraft surveillance, weather polarimetric radars, andcontrol and communications systems. Good radiating features, in terms of pattern shape and gain, are observed in a large band offrequencies. The basic scanning performance of large and small array configurations is finally investigated.

1. Introduction

The development of ultra-wide-band (UWB) antennas isrecently providing new solutions for the design of innovative,versatile, and economical radar and communication systems.This class of structures is in fact gaining increasing attractive-ness for the enhanced features of the wideband signals, forinstance in applications to imaging, surveillance, and high-capacity communications [1–4].

In impulse radar techniques, UWB antennas allow totransmit and receive short-pulsed signal waveforms withoutsignificant distortions. In the field of wireless communica-tions, the demand for wideband radiating elements growsup as well. Furthermore, the dramatic reduction of costsrelated to a single equipment for multiple applications(communications, surveillance, weather, etc.) is possible if asuitably wide band of frequency can be used and efficientlytreated by the same radiating element [5–9].

Many UWB antennas based on printed configurationshave been proposed in the past years [3, 10]. They are cur-rently of great interest due to various advantages [11, 12]: lowcost, light weight, compactness and conformability, compat-ibility with printed-circuit-board (PCB) technologies, andhigh reproducibility in the realization of large array radars.The most common implementation is based on monopolestructures, having bidirectional radiating properties. Whilea reflecting ground plane could be introduced to maximizethe gain in the required half plane, this usually introducesresonant behaviors in the currents on the metalization, thussignificantly reducing the useful operational bandwidth ofthe structure.

More complex configurations should be designed ifgrounded structures are of interest: for example, the useof parasitic elements can induce resonances at differ-ent frequencies thus enlarging the total available band


[13–15]. In this paper, novel configurations of printeddipoles arranged in rhombic shapes are proposed to radiateover a very wide band of frequency with applications toarray radars and UWB communication systems, followinga preliminary design presented in [16]. The frequencyranges of interest here are mainly for the C, X, Ku, andK bands (approximately between 5 and 25 GHz). Suitablescaling of such structures anyway allows for covering otherpossible microwave ranges for UWB applications. The basicprinciples for the new types of radiators are introduced inSection 2. Specific solutions are considered here for dual-polarization applications. In Section 3, a first topology isdescribed, composed by four coplanar rhombic elements(see Figure 1). In Section 4, a laterally shielded structure,feasible for implementation in an array environment, ispresented: two rhombic elements are printed on differentlayers, in a simple stacked configuration, each providing alinear polarization orthogonal to each other (see Figure 2).Extensive results are shown for both the structures, to showthe UWB matching features through the relevant inputparameters, and the radiating characteristics of the singleradiating elements and of various array configurations.

2. Basic Structures

The two structures presented here are based on variousarrangements of printed dipoles of different lengths formingapproximately rhombic-shaped elements (see Figures 1 and2). The central patch is the longest and can be fed at its endswith one or two probes; the side patches, being six or eightdepending on the structure, have smaller dimensions and areparasitic. To further increase the bandwidth, the dipoles areprinted on a three-layer structure, designed with low-costcommercially available dielectric substrates. In the presentanalysis, the lower and the upper layers are duroid RT 5880Rogers substrates with relative permittivity εr,1 = 2.2 anddielectric loss tangent tan δ1 = 9.0 · 10−4, the middle layeris an FR4 epoxy substrate with εr,2 = 4.4 and dielectricloss tangent tan δ2 = 200.0 · 10−4. It should be pointedout that, even if loss effects can become nonnegligible(particularly in higher-frequency ranges); in this case thechoice of such inexpensive commercial substrates is mainlyrelated to the strong reduction of costs in the implementationof large arrays of elements, exploiting also well-establishedmanufacturing printed-circuit techniques of PCB [17]. Anumber of other materials having similar electric parametersbut reduced losses can be chosen if higher efficiency is desired(e.g., for the internal layer a dielectric such as TMM4, havinga loss tangent one order of magnitude lower than FR4 can beemployed). Some compared results concerning the influenceof such losses on the gain and efficiency of these antennas willbe presented next. In the following analyses, also nonidealeffects of the metalization are taken into account, consideringfinite-conductivity strips made of copper (σ = 5.8 · 107 S/m,μr = 0.999991) having nonzero thickness (10−6 m). A singlerhombic shape mainly provides a field linearly polarized,with components related to the direction of the relevantdipoles. The dual polarizations can thus be obtained using

zy

x

(a)

z

y

x

Port3

Port2

Port1

Port4

1

23

45

(b)

Figure 1: Geometry of the radiating element under analysis inSection 3, based on four rhombic coplanar elements printed onthe top of a three-layer configuration. (a) 3D view of the antenna.(b) Upper view of the structure. Lowest substrate: thickness h =1.5 mm, relative dielectric constant εr = 2.2, loss tangent tan δ =9 · 10−4 (duroid RT 5880 Rogers). Middle substrate thickness h =7.5 mm, relative dielectric constant εr = 4.4, loss tangent tan δ =200× 10−4 (FR4 epoxy). Uppermost substrate thickness h = 2 mm,relative dielectric constant εr = 2.2, loss tangent tan δ = 9·10−4 (RT5880 Roger). Overall unit-cell dimension 43 × 43 mm. Geometricalparameters of the patches: w1 = 1.9 mm, l1 = 17.3 mm, w2 = 2 mm,l2 = 7 mm, w3 = 1 mm, l3 = 11.3 mm, w4 = 0.6 mm, l4 = 4 mm,w5 = 0.4 mm, and l5 = 2 mm (wi and li are the width and the lengthof the ith patch, resp.). Diameter of the probe d = 0.86 mm.

a couple of rhombic elements, mutually rotated of 90◦. Therelevant features of two possible rhombic arrangements areanalyzed in detail in the next two sections.

3. Unshielded Antenna with CoplanarDual Rhombic Elements

3.1. Structure. A fully dual-polarized radiating element canbe obtained by arranging four rhombic shapes, printed onthe top of the three-layer structure described in the previous


zy

x

(a)

z

y

x

Port3

Port2

Port1

Port4

12 3 42

3

4

1

(b)

Figure 2: Geometry of the radiating element under analysis inSection 4, based on another rhombic element printed on the toplayer, and a rhombic element printed on the second layer. (a) 3Dview of the antenna. (b) Upper view of the structure. Same three-layer configuration as in Figure 1. Geometrical parameters of thepatches of the upper element: w1 = 1 mm, l1 = 29.6 mm, w2 =0.8 mm, l2 = 10 mm, w3 = 0.6 mm, l3 = 8 mm, w4 = 0.6 mm, andl4 = 7 mm. The diameter of the probes feeding the upper elementis d = 0.24 mm. Geometrical parameters of the patches of the lowerelement: w1 = 0.6 mm, l1 = 26.8 mm, w2 = 0.6 mm, l2 = 15.6 mm,w3 = 0.6 mm, l3 = 6 mm, w4 = 0.6 mm, and l4 = 6 mm. (wi and liare the width and the length of the ith patch, resp.) The diameter ofthe probes feeding the lower element is d = 0.4 mm. The size of thebox is 40 mm.

section, as in Figure 1. The central patch of each rhombicelement is fed with a probe at its outer end, and oppositepatches are fed with signals having the same magnitudeand opposite sign. With this geometrical arrangement, onecouple of opposite rhombic elements provides one linearpolarization, and the other couple of opposite elements,rotated of 90◦, provides the orthogonal polarization. If thesame signal is radiated through both the polarizations, ageneral elliptical polarization can be obtained, by suitablytuning the relevant phase shifts among the input probes.In particular, a circular polarization can be obtained byimposing a 90◦ phase shift between adjacent elements.

The antenna is designed to work in a very wide frequencyrange, possibly greater than C-X bands (4–12 GHz) in thecase of interest. Its transverse dimensions of the elementare approximately 4 cm, leading to a possible compactimplementation in array at these frequencies. Different

operational requirements are given at lower and at higherfrequencies. Specifically, at lower frequencies both a goodmatching and a regular radiation pattern are required, whilstat higher frequencies, because of a different utilization ofthe control system, a good matching only can be sufficient.The dimensions of the various patches have been designed inorder to reach an optimum input matching in the workingwide band 5–25 GHz, as shown in next subsection. Furthersimulations, here omitted for brevity, would show that theinput matching is preserved also at frequencies higher than30 GHz, even though the radiation patterns are less regularas the frequency increases.

Beyond these advantages in terms of extremely wideband, low-profile and low-cost realization, and radiationpattern regularity, a key promising feature of this structureis also the dynamic conformability of its radiation pattern,due to the use of a pair of probes for the feeding of eachpolarization. For instance, if an array of elements is placed ona conformable surface, a fine tuning of the pointing angle canbe easily obtained by an appropriate tailoring of the feedingnetwork.

3.2. Analysis and Results. A throughout optimization pro-cedure, involving the dimensions of the patches and thepositions of the feeding probes, has been carried out withthe software package “ModeFrontier” [18], with the aim ofachieving the best input matching in the desired frequencyband. The optimization has been performed at first on asingle rhombic shape and then refined with the full four-rhombus structure. The algorithm used is a multiobjectivegenetic algorithm with multisearch elitism for enhancedrobustness [19]. Its objectives were chosen as the conditions|Si j| < −10 dB at 5 GHz for any i, j, and the variables were allthe dimensions and positions of the various patches.

The results shown here have been computed with thetime-domain solver of the high-frequency electromagneticCAD “CST Studio 2010” [20]. The layered substrate isassumed laterally unbounded, and a single radiating element,composed by four rhombic shapes, is considered as inFigure 1. The input ports are modeled as coaxial cables,fed with proper phase shifts. The feeding network providingthese shifts among the different ports is not simulatedhere and will be object of future work. The CST model isdiscretized with hexahedral cells of average dimension λ/10;open conditions are placed at the side boundary of the cell,thus assuming a laterally unbounded substrate; extra space isadded in the top half plane in order to accurately estimate theradiation patterns. Waveguide ports are defined at the coaxialcables and excited with the fundamental TEM mode.

Based on the data sheets of the electromagnetic param-eters of these materials as a function of the frequency (ifavailable), the dispersive effects can be taken properly intoaccount in the simulations. The relevant analysis showsanyway that these dispersive effects do not change sensitivelythe performance of our antennas in the ranges of interest.

Due to the geometrical symmetries, the input featuresof the antenna are described by three different scatteringparameters. In Figure 3, the magnitude in dB of these three


−60

−50

−40

−30

−20

−10

0

0 5 10 15 20 25

| S11| (dB)| S21| (dB)| S31| (dB)

f (GHz)

Figure 3: Antenna of Figure 1. Magnitude of the scatteringcoefficients of the radiating element fed by the four probes.Reflection coefficient |S11| at the input port 1 (black line) andtransmission coefficients |S21| from port 1 to port 2 (gray line) andfrom port 1 to port 3 (dashed black line).

parameters is shown in the frequency band considered, fromdc to 25 GHz. The reflection coefficient at one input cableis shown in black solid line (due to the symmetry of thestructure, the cable considered is arbitrary, since S11 = S22 =S33 = S44). The coupling between opposite ports (e.g.,|S21|) is shown in gray solid line, and the coupling betweenadjacent ports (e.g., |S31|) is shown in black dashed line. Theoperational frequency band where the parameters are underthe conventional threshold of −10 dB is extremely wide: inthe figure, the band 5–25 GHz is well matched, but furthersimulations would show a good matching also beyond30 GHz, even though, as said, the radiation patterns tend todeteriorate as the frequency increases. In Figure 4(a), the gainis shown on the principal plane ϕ = 0◦ at different frequen-cies from 5 GHz to 25 GHz if the ports 1 and 2 are fed withcommon amplitude and opposite phase: by symmetry, theelectric field has only a θ-component. In Figure 4(b), the gainis shown on the same plane if the ports 3 and 4 are fed withcommon amplitude and opposite phase: in this case, by sym-metry, the electric field has only a ϕ-component. Good radi-ation features are found in the lower band of frequency, whilethe pattern tends to deform at higher frequencies as expected.

The features of the maximum gain are resumed inTable 1, considering their values with respect to the isotropicradiator (dBi and the relevant beam angular locations onthe different planes). In the same table, maximum gain andradiation efficiencies are also presented with reference tothe alternative choice of the TMM4 for the middle layer(with reduced losses, see Section 2 for details). As shown,acceptable results are achieved even with the lossy substrate,while significant improvements in the efficiency can beobtained at higher frequencies with the use of a lower-losssubstrate.

−40

−30

−20

−10

0

1030

0

30

60

90

(dB

i)

(a)

−30

−40

−20

−10

0

1030

0

30

60

90

(dB

i)(b)

Figure 4: Antenna in Figure 1. Gain in polar form on the principalelevation plane φ = 0◦ of the single element at different frequenciesin the operational bandwidth 5–25 GHz: 5 GHz (solid black line),10 GHz (solid gray line), 20 GHz (dashed black line). (a) Ports 1 and2 are fed, exciting a field along the θ direction. (b) Ports 3 and 4 arefed, exciting a field along the ϕ direction.

4. Shielded Antenna with StackedDual Rhombic Elements

4.1. Structure. Once the antenna described in the previoussection is introduced in an array environment, the UWBinput performance of the array could deteriorate dueto strong mutual coupling among adjacent elements. Toovercome such a problem, a lateral shielding of the singleradiating element can be implemented with metallic walls.Such a configuration also enables us to reduce adverse effectsrelated to possible launching of surface and leaky waves in thelayered structure [11, 12]. As an example of this approach,an alternative shielded radiating element is presented here,based on the same rhombic printed shapes described above,but with elements placed on different layer interfaces.

In this case, shown in Figure 2, each couple of oppositerhombic elements is here replaced by a single rhombus,whose central larger patch is fed by two different probes at itsends, carrying signals with common amplitude and oppositephase. With this original feeding structure involving twoprobe elements on the opposite sides of each main strip, it isnoted that, by properly choosing the phase shift between thefeeders, it is also possible to achieve a straightforward controlof the current configuration excited on the strips, thussuitably influencing the directional features of the radiatedbeam. In this structure, the two orthogonal polarizations arethen provided by two different rhombic elements, rotatedby 90◦ and printed on two different layers: the dipoles of


Table 1: Maximum gain: on ϕ = 0◦.

f (GHz) Direction (ϑ) Max gain Efficiency Max gain (low loss) Efficiency (low loss)

5 ϑ = 0◦ 6.6 dBi 92.9% 7.1 dBi 99%

10 ϑ = 90◦, ϕ = 54◦ 3.8 dBi 78.7% 4.5 dBi 96.6%

15 ϑ = 45◦, ϕ = 22◦ 3.2 dBi 80.0% 3.7 dBi 96.8%

20 ϑ = 38◦, ϕ = 24◦ 3.7 dBi 64.1% 4.9 dBi 92.7%

25 ϑ = 90◦, ϕ = 9◦ 5.5 dBi 68.4% 6.2 dBi 93.3%

one rhombus are printed on the uppermost layer and thedipoles of the other rhombus on the middle layer. Also inthis case the box has transverse dimensions of 4 cm, leadingto good performance in a possible array implementation. Onthe other hand, the distance between the structure (i.e., stripsand cables) and the lateral boundaries is greater in this newconfiguration than in the previous one, leading to an easierinput matching also in the shielded configuration.

Again, the dimensions of the various patches have beenoptimized in order to yield the optimum input matching inthe frequency band 5–25 GHz. Of course, considering such adifferent arrangement of the rhombic shapes, rather differentgeometrical parameters have been reached in this secondstructure as optimum values.

4.2. Analysis and Results. The results shown here havebeen obtained through full-wave simulations of the finaloptimized stacked and shielded structure again with thetime-domain solver of CST Studio.

In Figure 5, the magnitude in dB of the three scatteringparameters of the shielded configuration is shown; also withthis structure, a good input matching is reached in the verywide frequency band 5–25 GHz. The reflection coefficientat port 1 (i.e., feeding the upper rhombic shape) is shownin solid black line, the reflection coefficient at port 3 (i.e.,feeding the lower rhombic shape) is shown in solid gray line,while the coupling between the rhombic shapes is shown indashed black line.

The radiation patterns on the principal planes are shownagain for various values of the frequency. In Figure 6(a),the gain is shown on the principal plane ϕ = 0◦, whenprobes 1 and 2 feed the upper element with signals withcommon magnitude and 180◦ phase shift. The electricfield is in this case polarized along the θ direction bysymmetry. In Figure 6(b), the gain is shown on the sameplane when the lower element is fed through probes 3 and4, with common magnitude and 180◦ phase shift. A dualpolarization is radiated with respect to the previous result,the electric field being polarized along the ϕ direction. Asseen, fairly regular radiation patterns are observed for bothpolarizations, in particular in the lower part of frequency.Since this configuration is studied with an “open add space”lateral boundary in CST, an estimation of backlobes is alsopresent in the results.

The gain on the other principal plane ϕ = 90◦ is shownin Figure 7. The upper rhombic shape is fed in Figure 7(a),while the lower rhombic shape is fed in Figure 7(b).

−60

−50

−40

−30

−20

−10

0

0 5 10 15 20 25

(dB

)

| S11|| S33|| S13|

f (GHz)

Figure 5: Antenna of Figure 2. Magnitude of the scatteringcoefficients of the radiating element fed by the four probes.Reflection coefficients |S11| at the input port 1 feeding the upperstructure (black line), |S33| at the input port 3 feeding the lowerstructure (gray line), and transmission coefficient |S31| from port1 to port 3 (dashed black line).

In Tables 2 and 3, a summary of the maximum-gainvalues, locations, and radiation efficiencies is also givenfor different frequencies, with reference to the feedingof the upper and lower element, respectively. As for theprevious unshielded antenna, maximum gain and efficiencyis also presented with reference to the alternative choice ofthe TMM4 for the middle layer (with reduced losses, seeSection 2 for details). In particular, while the lower elementhas a reduced efficiency at higher frequencies, its behaviorcan be substantially improved with the use of a lower-losssubstrate.

4.3. Array Behavior. In order to test further this typeof structure, a first simple analysis has been led whichgives basic information on the scanning-beam directionalfeatures for large arrays. To this aim, the radiation patternof an array of 140 × 140 elements has been computedwith an array-factor approximation, for different values ofthe pointing angle, depending on the selected phase shiftbetween adjacent elements of the array. In Figure 8(a), thepattern on the plane ϕ = 0◦ at 5 GHz is shown for a


−40

−30

−20

−10

0

10

2030

030

60

90

120

150

180

150

120

(dB

i)

(a)

−40

−30

−20

−10

0

10

2030

030

60

90

120

150

180

150

120

(dB

i)

(b)

Figure 6: Antenna in Figure 2. Gain in polar form on the principal elevation plane φ = 0◦ of the single element at different frequenciesin the operational bandwidth 5–25 GHz: 5 GHz (solid black line), 10 GHz (gray line), 20 GHz (dashed black line). (a) Ports 1 and 2 feed theupper element, exciting a field along the θ direction. (b) Ports 3 and 4 feed the lower element, exciting a field along the ϕ direction.

−40

−30

−20

−10

0

10

2030

0

30

60

90

120

150

180

150

120

(dB

i)

(a)

−40

−30

−20

−10

0

10

2030

0

30

60

90

120

150

180150

120

(dB

i)

(b)

Figure 7: Antenna in Figure 2. Gain in polar form on the principal elevation plane φ = 90◦ of the single element at different frequenciesin the operational bandwidth 5–25 GHz: 5 GHz (solid black line), 10 GHz (gray line), 20 GHz (dashed black line). (a) Ports 1 and 2 feed theupper element, exciting a field along the θ direction. (b) Ports 3 and 4 feed the lower element, exciting a field along the ϕ direction.

broadside radiation when the four probes of each elementare fed with common amplitude and a 90◦ shift in orderto radiate a circular polarized field. The two componentsalong θ and ϕ are shown in solid black line and in dashedgray line, respectively. In Figure 8(b), the same quantitiesare computed for a beam pointing at the elevation θ = 30◦

and azimuth ϕ = 0◦. The gain at the main lobe direction isfairly regular, around 50 dBi; the side-lobe levels are ratherreduced (about 20 dB below the main lobe); the consideredphase-scanned pencil beams have half-power beamwidth ofabout 0.4◦. Effects related to grating lobes can be present

in connection of the element spacing as phase shift andfrequency are varied.

Further results have been obtained considering a lineararray made of a small number of cells (3 × 1) along the xdirection in the reference system of Figure 2. In this case, toaccurately predict the array performance, a nonapproximatefull-wave analysis has been necessary with a proper CADimplementation of the overall physical structure of the threeantennas. This rigorous approach allows us to to verifythe actual scanning features of the radiated beam as afunction of the relevant phase shift. Representative behaviors


Table 2: Maximum gain if ports 1 and 2 are fed (Upper Element).


5 ϑ = 0◦ 6.5 dBi 92.9% 6.9 dBi 99.3%

10 ϑ = 0◦ 6.5 dBi 56.1% 10.67 dBi 96.8%

15 ϑ = 55◦, ϕ = 70◦ 4.2 dBi 47.3% 7.01 dBi 88.5%

20 ϑ = 50◦, ϕ = 67◦ 2.5 dBi 40.6% 9.04 dBi 75.0%

25 ϑ = 21◦, ϕ = 90◦ 4.2 dBi 41.5% 7.63 dBi 86.9%

Table 3: Maximum gain if ports 3 and 4 are fed (lower element).


5 ϑ = 0◦ 7.5 dBi 79.0% 8.4 dBi 95.8%

10 ϑ = 0◦ 8.2 dBi 37.0% 10.3 dBi 88.1%

15 ϑ = 61◦, ϑ = 20◦ 2.5 dBi 28.0% 6.8 dBi 80.0%

20 ϑ = 0◦ 2.3 dBi 28.2% 7.8 dBi 88.6%

25 ϑ = 45◦, ϑ = 20◦ 3.4 dBi 29.1% 7.5 dBi 80.8%

−40

−20

0

20

40

60

−90 −45 0 45 90

θ (◦)

(dB

i)

Abs (θ) [dBi]Abs (ϕ) [dBi]

(a)

−40

−20

0

20

40

60

−90 −45 0 45 90

θ (◦)

(dB

i)

Abs (θ) [dBi]Abs (ϕ) [dBi]

(b)

Figure 8: Array of shielded elements as in Figure 2. Gain on the principal elevation plane φ = 0◦ of a 140 × 140 array at 5 GHz. All theports are fed, exciting a circularly polarized field (its θ component in solid black line, its ϕ component in dashed gray line). (a) Beam pointingat broadside (θ = 0◦). (b) Beam pointing at θ = 30◦.

of radiation patterns for this small array are reported inFigure 9. In Figure 9(a), the gain on the two principal planesis shown when the three upper sets of patches are fed in phase(black curves) and when three lower sets of patches are fed inphase (gray curves). In Figure 9(b), the scanning capabilitiesof this small array are presented, showing the gain on theprincipal plane ϕ = when the lower sets of patches are fed,pointing their main beam along the ϕ = 0◦ plane, at θ = 10◦,20◦, 30◦, 40◦. As expected, a fan beam is obtained, as typicalof linear arrays, having maximum gain between 12 and 9 dBi.

In Figure 10, couplings among cells are analyzed, againwith reference to the feeding of the lower patches, only for thesake of brevity. In Figure 10(a), an input cable feeds the lowerpatch only in the central cell. The magnitude of the reflectioncoefficient at this port and of the coupling coefficients with

the other cables are shown, avoiding coefficients having equalvalues due to evident geometrical symmetries. As expected,in the band of frequency investigated, a low level of couplingis found also in this 3 × 1 array. In Figure 10(b), the threelower sets of patches of the array are fed in phase throughone of the two ports; the magnitude of the active reflectioncoefficient at the middle cell and at one side cell is reported,proving a good input matching also in this active arrayconfiguration.

As a last result, in Figure 11 the radiated patterns ofan array made of 3 × 3 elements is shown on both theprincipal planes, ϕ = 0◦, 90◦ for both the upper (Figure 11(a))and lower (Figure 11(b)) patches. Again, a full-wave analysishas been performed, without recurring to simplified array-factor formulations. All the cells of the array are fed in phase


−30

−20

−10

0

10

20

θ (◦)

(dB

i)

−90 −60 −30 0 30 60 90

(a)

−40

−20

0

20

40

60

θ (◦)

(dB

i)

−90 −60 −30 0 30 60 90

(b)

Figure 9: Array of 3 adjacent shielded cells as in Figure 2 at 5 GHz. (a) Gain on the principal elevation planes φ = 0◦ (solid lines) and φ = 90◦

(dashed lines). Beam pointing at broadside (θ = 0◦, all the cells are excited in phase) by the upper patches (black lines) and by the lowerpatches (gray lines). (b) The three lower sets of patches are phased in order to point at θ = 10◦ (black line), θ = 20◦ (gray line), θ = 30◦ (blackline with squares), and θ = 40◦ (gray line with squares).

−60

−50

−40

−30

−20

−10

0

5 10 15 20 25

(dB

)

f (GHz)

(a)

5 10 15 20 25

0

−5

−10

−15

−20

−25

−30

−35

−40

(dB

)

f (GHz)

(b)

Figure 10: Array of 3 adjacent shielded cells as in Figure 2 at 5 GHz. (a) Only the input port 3 feeds the lower patches in the central cell.Magnitude of reflection coefficient (thick black line), of the coupling coefficient with a cable connected to the upper patch in the same cell(thick gray line), and of coupling coefficients with cables in an adjacent cell (thin black lines). (b) All the three lower sets of patches as fed inphase through port 3 (see Figure 2) and the respective translated ports in the other two cells. Magnitude of the active reflection coefficientversus frequencies at the middle cell of the 3 × 1 array (black line) and at one side cell (gray line).

thus radiating a single beam at broadside, with gain around17 dBi, angular width 20◦, and sidelobe level −15 dB withrespect to the maximum gain. As shown in the pictures, aremarkable symmetry of the pattern is obtained on the prin-cipal planes, also if the different polarizations are compared.

5. Conclusion

A new class of UWB low-cost printed antennas has been pre-sented and optimized for dual-polarized radar applicationsin the microwave range 5–25 GHz. The basic single elementis composed by properly fed printed dipoles arranged inrhombic configurations. Different elements can providealmost orthogonal polarizations, depending on their mutual

orientation. Attention was paid to practical implementa-tion with inexpensive dielectrics commonly used in PCBtechnology. These structures are low profile and lightweightand are characterized by high modularity/scalability, whichmakes them suitable to implement low-profile phased arrayantennas of various shapes and sizes.

Suitable extensive parametric analyses have been carriedout by means of advanced numerical tools as concerns themost efficient choice of the strip geometry configurations.A first design of the antenna element is made with fourcoplanar rhombic elements, each fed by a probe reaching oneend of the central dipole. Relevant results are shown for asingle element printed on a laterally unbounded substrate.An advanced alternative design has been proposed, leading


θ (◦)

(dB

i)

−90 −60 −30 0 30 60 90

20

10

0

−10

−20

−30

(a)

θ (◦)

(dB

i)

−90 −60 −30 0 30 60 90

20

10

0

−10

−20

−30

(b)

Figure 11: Array of 3 × 3 adjacent shielded cells as in Figure 2. Gainon the principal elevation planes φ = 0◦ (solid lines) and φ = 90◦

(dashed lines) at 5 GHz. Beam pointing at broadside (θ = 0◦, all thecells are excited in phase). (a) The upper patches are fed. (b) Thelower patches are fed.

to a shielded radiating element which can be suitably used inarray configurations, where the rhombic elements providingthe two polarizations are stacked on different substrates. Inthis structure, lateral metallic walls prevent mutual-couplingeffects among elements and reduce possible undesired effectsof power leakage in the substrate. A good input matchingis obtained in the whole frequency band required, withvery promising performance in terms of radiation patternregularity and gain. The basic performance achievable inarray configurations has been finally addressed. On thisground, manufacturing and experimental test of prototypesof 2D arrays is programmed next.

References

[1] B. Allen, M. Dohle, E. Okon, W. Malik, A. Brown, and D.Edwards, Eds., Ultra Wideband Antennas and Propagation forCommunications, Radar and Imaging, Wiley, 2006.

[2] W. Wiesbeck, G. Adamiuk, and C. Sturm, “Principles of UWBantennas basic properties and design,” Proceedings of the IEEE,vol. 97, no. 2, pp. 372–385, 2009.

[3] C.-C. Chen, “Ultrawide bandwidth antenna design,” inAntenna Engineering Handbook, J. L. Volakis, Ed., McGraw-Hill, New York, NY, USA, 2007.

[4] J. J. Lee, “Ultra wideband arrays,” in Antenna EngineeringHandbook, J. L. Volakis, Ed., McGraw-Hill, New York, NY,USA, 2007.

[5] J. Herd, S. Duffy, M. Weber, G. Brigham, C. Weigand, and D.Cursio, “Advanced architecture for a low cost MultifunctionPhased Array Radar,” in Proceedings of the IEEE MTT-SInternational Microwave Symposium Digest (MTT ’10), pp.676–679, Anaheim, Calif, USA, May 2010.

[6] M. Weber, J. Cho, J. Flavin, J. Herd, and M. Vai, “Multi-function phased array radar for U.S. Civil-sector surveillanceneeds,” in Proceedings of the 32nd Conference on RadarMeteorology, pp. 1977–1987, Albuquerque, NM, USA, October2005.

[7] M. Weber, J. Cho, J. Flavin, and J. Herd, “Multifunctionphased array radar: technical synopsis, cost implications andoperational capabilities,” in Proceedings of the 23rd Conferenceon International Interactive Information and Processing Systems(IIPS) for Meteorology, Oceanography and Hydrology, SanAntonio, Tex, USA, January 2007.

[8] R. Waterhouse, Ed., Printed Antennas for Wireless Communi-cations, Wiley, New York, NY, USA, 2007.

[9] V. N. Bringi and V. Chandrasekar, Polarimetric DopplerWeather Radar: Principles and Applications, Cambridge Uni-versity Press, Cambridge, UK, 2001.

[10] D. R. Jackson, “Microstrip antenna,” in Antenna EngineeringHandbook, J. L. Volakis, Ed., McGraw-Hill, New York, NY,USA, 2007.

[11] C. D. Nallo, F. Mesa, and D. R. Jackson, “Excitation of leakymodes on multilayer stripline structures,” IEEE Transactionson Microwave Theory and Techniques, vol. 46, no. 8, pp. 1062–1071, 1998.

[12] P. Baccarelli, P. Burghignoli, F. Frezza, A. Galli, G. Lovat,and D. R. Jackson, “Approximate analytical evaluation ofthe continuous spectrum in a substrate-superstrate dielectricwaveguide,” IEEE Transactions on Microwave Theory andTechniques, vol. 50, no. 12, pp. 2690–2701, 2002.

[13] K. Ghorbani and R. B. Waterhouse, “Dual polarized wide-band aperture stacked patch antennas,” IEEE Transactions onAntennas and Propagation, vol. 52, no. 8, pp. 2171–2174, 2004.

[14] D. Tallini, A. Galli, M. Ciattaglia, L. Infante, A. De Luca, andM. Cicolani, “A new low-profile wide-scan phased array forUWB applications,” in Proceedings of the European Conferenceon Antennas and Propagation, pp. 1–5, Edinburgh, UK,November 2007.

[15] A. Galli, G. Valerio, D. Tallini, A. De Luca, and M. Cicolani,“Optimization of multifunctional UWB planar phased arrays,”in Proceedings of the European Conference on Antennas andPropagation, pp. 571–574, Berlin, Germany, March 2009.

[16] S. Mazzocchi, G. Valerio, M. Zucca, M. Ciattaglia, A. DeLuca, and A. Galli, “New multifunctional UWB planarantenna based on printed dipoles in rhombic configuration,”in Proceedings of the European Conference on Antennas andPropagation, Barcelona, Spain, April 2010.

[17] A. A. Qureshi, M. U. Afzal, T. Taqueer, and M. A. Tarar,“Performance analysis of FR-4 substrate for high frequencymicrostrip antennas,” in Proceedings of the China-Japan JointMicrowave Conference (CJMW ’11), pp. 159–162, April 2011.

[18] http://www.modefrontier.com.


[19] C. Poloni and V. Pediroda, “GA coupled with computationallyexpensive simulations: tools to improve efficiency,” in Algo-rithms and Evolution Strategies in Engineering and ComputerScience, pp. 267–288, Wiley, 1997.

[20] CST Microwave Studio Manual, CST, Germany, 2002.


Research Article

Design and Analysis of Wideband Nonuniform Branch LineCoupler and Its Application in a Wideband Butler Matrix

Yuli K. Ningsih,1, 2 M. Asvial,1 and E. T. Rahardjo1

1 Antenna Propagation and Microwave Research Group (AMRG), Department of Electrical Engineering,Universitas Indonesia, New Campus UI, West Java, Depok 16424, Indonesia

2 Department of Electrical Engineering, Trisakti University, Kyai Tapa, Grogol, West Jakarta 11440, Indonesia

Correspondence should be addressed to Yuli K. Ningsih, yuli [email protected]

Received 10 August 2011; Accepted 2 December 2011

Academic Editor: Tayeb A. Denidni

Copyright © 2012 Yuli K. Ningsih et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper presents a novel wideband nonuniform branch line coupler. An exponential impedance taper is inserted, at the seriesarms of the branch line coupler, to enhance the bandwidth. The behavior of the nonuniform coupler was mathematically analyzed,and its design of scattering matrix was derived. For a return loss better than 10 dB, it achieved 61.1% bandwidth centeredat 9 GHz. Measured coupling magnitudes and phase exhibit good dispersive characteristic. For the 1 dB magnitude differenceand phase error within 3◦, it achieved 22.2% bandwidth centered at 9 GHz. Furthermore, the novel branch line coupler wasimplemented for a wideband crossover. Crossover was constructed by cascading two wideband nonuniform branch line couplers.These components were employed to design a wideband Butler Matrix working at 9.4 GHz. The measurement results show thatthe reflection coefficient between the output ports is better than 18 dB across 8.0 GHz–9.6 GHz, and the overall phase error is lessthan 7◦.

1. Introduction

Recently, a switched-beam antenna system has been widelyused in numerous applications, such as in mobile commu-nication system, satellite system, and modern multifunctionradar. This is due to the ability of the switched-beam antennato decrease the interference and to improve the quality oftransmission [1, 2] and also to increase gain and diversity[3].

The switched-beam system consists of a multibeamswitching network and antenna array. The principle of aswitched-beam is based on feeding a signal into an array ofantenna with equal power and phase difference. Differentstructures of multibeam switching networks have beenproposed, such as the Blass Matrix, the Nolen Matrix, theRotman Lens, and the Butler Matrix [4]. One of the mostwidely known multibeam switching networks with a linearantenna is the Butler Matrix. Indeed, it seems to be the mostattractive option due to its design simplicity and low powerloss [5–15].

In general, the Butler Matrix is an N × N passive feedingnetwork, composed of branch line coupler, crossover, andphase shifter. The bandwidth of the Butler Matrix is greatlydependent on the performance of the components. However,the Butler Matrix has a narrow bandwidth characteristicdue to branch line coupler and crossover has a limitedbandwidth.

As there is an increased demand to provide high datathroughput [16], it is essential that the Butler Matrix hasto operate over a wide frequency band when used for anglediversity. Therefore, many papers have reported for thebandwidth enhancement of branch line coupler [17–20]. Inreference [17, 18], design and realization of branch line cou-pler on multilayer microstrip structure was reported. Thesedesigns can achieve a wideband characteristic. However, thedisadvantages of these designs are large in dimension andbulk.

Reference [19] introduces a compact coupler in an N-section tandem-connected structure. The design resulted ina wide bandwidth. Another design, two elliptically shaped


P1

8

4.75

12.75 14.5

3P2

P3

P4

Figure 1: Geometry structure of a new nonuniform branch linecoupler design with exponential impedance taper at the series arm.

microstrip lines which are broadside coupled through anelliptically shaped slot, was employed in [20]. This designwas used in a UWB coupler with high return loss andisolation. However, these designs require a more complexmanufacturing.

In this paper, nonuniform branch line coupler usingexponential impedance taper is proposed which can enhancebandwidth and can be implemented for Butler Matrix, asshown in Figure 1. Moreover, it is a simple design withoutneeds of using multilayer technology. This will lead in costreduction and in design simplification.

To design the new branch line coupler, firstly, theseries arm’s impedance is modified. The shunt arm remainsunchanged. Reduced of the width of the transmission lineat this arm is desired by modifying the series arm. Next, byexponential impedance taper at the series arm, a good matchover a high frequency can be achieved.

2. Mathematical Analysis of NonuniformBranch Line Coupler

The proposed nonuniform branch line coupler use λ/4branches with impedance of 50Ω at the shunt arms anduse the exponential impedance taper at the series arms, asshown in Figure 1. Since branch line coupler has a symmetricstructure, the even-odd mode theory can be employed toanalyze the nonuniform characteristics. The four ports canbe simplified to a two-port problem in which the evenand odd mode signals are fed to two collinear inputs [22].Figure 2 shows the schematic of circuit the nonuniformbranch line coupler.

The circuit of Figure 2 can be decomposed into thesuperposition of an even-mode excitation and an odd-modeexcitation is shown in Figures 3(a) and 3(b).

The ABCD matrices of each mode can be expressedfollowing [22]. In the case of nonuniform branch linecoupler, the matrices for the even and odd modes become:[

A BC D

]e

= Z(z)√2

(−1 jj −1

), (1)

[A BC D

]o

= Z(z)√2

(1 jj 1

). (2)

Zo

A = 1

S11

S21 S41

S31

Zo

P1

P2

P

4P

3Z(z)/

√2

Figure 2: Circuit of the nonuniform branch line coupler.

1 1

1 1

1

1

1

1

Open circuit stub

(2 separate 2-ports)

+1/2

+1/2

Se12Se11

Z(z)/√

2

(a) Even mode (e)

1 1

1 1

1

1

1

1

Short circuit stub(2 separate 2-ports)

+1/2

+1/2

So12So11

Z(z)/√

2

(b) Odd mode (o)

Figure 3: Decomposition of the nonuniform branch line couplerinto even and odd modes of excitation.

A branch line coupler has been designed based on thetheory of small reflection, by the continuously tapered linewith exponential tapers [23, 24], as indicated in Figure 1,where

Z(z) ={Z0e−az, for 0 < z ≤ L,

Z0eaz, for L ≤ z < 2L,(3)

which determines the constant a as:

a = 1L

ln(ZL

Zo

), (4)

L = discrete steps length, Zo = Z (0), and ZL= Z (L).


Figure 4: Photograph of a proposed nonuniform branch linecoupler.

Useful conversions for two-port network parameters forthe even and odd modes of S11 and S21 can be defined asfollows [22]:

Se11 =(A + B − C −D)e

ΔYe, (5)

So11 =(A + B − C −D)o

ΔYo, (6)

Se21 =2

ΔYe, (7)

So21 =2

ΔYe, (8)

where

ΔYe = (A + B + C + D)e, (9)

ΔYo = (A + B + C + D)o. (10)

Since the amplitude of the incident waves for these twoports are ±1/2, the amplitudes of the emerging wave at eachport of the nonuniform branch line coupler can be expressedas [22]:

S11 = 12

(Se11 + So11

), (11)

S21 = 12

(Se11 − So11

), (12)

S31 = 12

(Se21 + So21

), (13)

S41 = 12

(Se21 − So21

). (14)

Parameters even and odd modes of S11 nonuniformbranch line coupler can be expressed as (15) and (16) asfollows:

Se11 =Z(z)√

2

(−1 + j − j + 1−1 + j + j − 1

)= 0, (15)

So11 =Z(z)√

2

(1 + j − j − 11 + j + j + 1

)= 0. (16)

An ideal branch line coupler is designed to have zeroreflection power and splits the input power in port 1 (P1)

into equal powers in port 3 (P3) and port 4 (P4). Considering(1) to (16), a number of properties of the ideal branch linecoupler maybe deduced from the symmetry and unitaryproperties of its scattering matrix. If the series and shunt armare one-quarter wavelength, by using (11), resulted in S11 =0.

As both the even and odd modes of S11 are 0, the valuesof S11 and S21 are also 0. The magnitude of the signal at thecoupled port is then the same as that of the input port.

Calculating (7) and (8) under the same o, the even andodd modes of S21 nonuniform branch line coupler will beexpressed as follows in(17)

Se21 = −Z(z)√

2

(1 + j

),

So21 =Z(z)√

2

(1− j

).

(17)

Based on (13), S31 can be expressed as follows

S31 = −12Z(z)√

2

((1 + j

)− (1− j))

= −12Z(z)√

2

(1 + j − 1 + j

)= − j

Z(z)√2.

(18)

Following (14), S41 nonuniform branch line coupler can becalculating as follows

S41 = −12Z(z)√

2

((1 + j

)+(1− j

))= −1

2Z(z)√

2

(1 + j + 1− j

)= −Z(z)√

2.

(19)

From this result, both S31 and S41 nonuniform branchline couplers have equal magnitudes of−3 dB. Therefore, dueto symmetry property, we also have that S11 = S22 = S33 =S44 = 0, S13 = S31, S14 = S41, and S21 = S34. Therefore, thenonuniform branch line coupler has the following scatteringmatrix in (20):

S = −Z(z)√2

⎡⎢⎢⎢⎣0 0 j 10 0 1 jj 1 0 01 j 0 0

⎤⎥⎥⎥⎦. (20)

3. Fabrication and MeasurementResult of Wideband Nonuniform BranchLine Coupler

To verify the equation, the nonuniform branch line couplerwas implemented and its S-parameter was measured. Itwas integrated on TLY substrate, which has a thicknessof 1.57 mm. Figure 4 shows a photograph of a widebandnonuniform branch line coupler. Each branch at the seriesarm comprises an exponentially tapered microstrip line

4 International Journal of Antennas and PropagationS

-par

amet

ers

mag

nit

ude

(dB

)

Frequency (GHz)

6 6.5 7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 12 12.50

−5

−10

−15

−20

−25

−30

S (1, 1)S (2, 1)

S (3, 1)S (4, 1)

Figure 5: Measurement result for nonuniform branch line coupler.

0

50

100

150

200

6 7 8 9 10 11 12

Ph

ase

char

acte

rist

ic (

deg)

Frequency (GHz)−200

−150

−100

−50

S (4, 1)S (3, 1)

Figure 6: Phase characteristic of nonuniform branch line coupler.

Phase shifter

3dB hybrid couplerCross over

45◦ 45◦

P1 P2 P3 P4

P5 P6 P7 P8

Figure 7: Basic schematic of the 4× 4 Butler Matrix [21].

which transforms the impedance from Zo = 50 ohms toZL = 50.6 ohms. This impedance transformation has beendesigned across a discrete step length L = 6.75 mm.

Figure 5 shows the measured result frequency response ofthe novel nonuniform branch line coupler. For a return lossand isolation better than 10 dB, it has a bandwidth of about61.1%; it extends from 7 to 12.5 GHz. In this bandwidth,the coupling ratio varies between 2.6 dB up to 5.1 dB. If

Figure 8: Photograph of microstrip nonuniform crossover.

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10

S-p

aram

eter

s m

agn

itu

de (

dB)

Frequency (GHz)

0

−5

−10

−15

−20

−25

−30

S (1, 1)S (2, 1)

S (3, 1)S (4, 1)

Figure 9: Measurement result for nonuniform crossover.

Figure 10: Final layout of the proposed wideband Butler Matrix4× 4.

the coupling ratio is supposed approximately 3 ± 1 dB, thebandwidth of about 22.2% centered at 9 GHz.

As expected, the phase difference between port 3 (P3) andport 4 (P4) is 90◦. At 9 GHz, the phases of S31 and S41 are85.54◦ and 171◦, respectively. These values differ from idealvalue by 4.54◦. The average phase error or phase unbalancebetween two branch line coupler outputs is about 3◦. Buteven the phase varies with frequency; the phase difference isalmost constant and very close to ideal value of 90◦ as shownin Figure 6.


1 2 3 4 5 6 7 8 9 10 11

Frequency(GHz)

Inse

rtio

n lo

ss (

dB)

0

−2

−4

−6

−8

−10

−12

S (5, 1) simulatedS (8, 1) simulatedS (7, 1) measuredS (7, 1) simulated

S (5, 1) measuredS (8, 1) measuredS (6, 1) simulatedS (6, 1) measured

(a) Input port 1 excitation

Inse

rtio

n lo

ss (

dB)

Frequency (GHz)0

−2

−4

−6

−8

−10

−12



8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10

(b) Input port 2 excitations

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10

Inse

rtio

n lo

ss (

dB)

Frequency (GHz)



0

−2

−4

−6

−8

−10

−12

(c) Input port 3 excitations

1 2 3 4 5 6 7 8 9 10 11

Inse

rtio

n lo

ss (

dB)

Frequency (GHz)0

−2

−4

−6

−8

−10

−12

S (5, 4) simulatedS (5, 4) measuredS (7, 4) simulatedS (7, 4) measured

S (6, 4) simulatedS (6, 4) measuredS (8, 4) simulatedS (8, 4) measured

(d) Input port 4 excitations

Figure 11: Insertion loss of the proposed Butler Matrix whendifferent ports are fed.

Ret

urn

loss

(dB

)

Frequency (GHz)

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10

S (1, 1) simulatedS (4, 4) simulated

S (1, 1) measuredS (4, 4) measured

0−5−10−15−20−25−30−35−40−45

(a) Input port 1 or 4 are excited

Ret

urn

loss

(dB

)

Frequency (GHz)

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10

S (2, 2) simulatedS (3, 3) simulated

S (2, 2) measuredS (3, 3) measured

0

−5

−10

−15

−20

−25

−30

−35

(b) Input port 2 or 3 are excited

Figure 12: Return loss of the proposed Butler Matrix when differentports are fed.

4. Design and Fabrication of the WidebandButler Matrix

Figure 7 shows the basic schematic of the 4× 4 Butler Matrix[21]. Crossover also known as 0 dB couplers is a four-portdevice and must provide for a very good matching andisolation, while the transmitted signal should not be affected.In order to achieve wideband characteristic crossover, thispaper proposes the cascade of two nonuniform branch linecouplers.

Figure 8 shows the microstrip layout of the optimizedcrossover. The crossover has a frequency bandwidth of1.3 GHz with VSWR = 2, which is about 22.2% of its centrefrequency at 9 GHz. Thus, it is clear from these resultsthat a nonuniform crossover fulfills most of the requiredspecifications, as shown in Figure 9.

Figure 10 shows the layout of the proposed widebandButler Matrix. This matrix uses wideband nonuniformbranch line coupler, wideband nonuniform crossover, andphase-shift transmission lines.

The wideband Butler Matrix was measured using Net-work Analyzer. Figure 11 shows the simulation and measure-ment results of insertion loss when a signal was fed intoport 1, port 2, port 3, and port 4, respectively. The insertionloss are varies between 5 dB up to 10 dB. For the idealButler matrix, it should be better than 6 dB. Imperfection


0102030405060

Ph

ase

diff

eren

t (d

eg)

Frequency (GHz)

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10

Phase (port 3)-phase (port 1)Phase (port 2)-phase (port 3)Phase (port 4)-phase (port 2)

(a) Input port 1 excitation

Ph

ase

diff

eren

t (d

eg)

Frequency (GHz)


8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10−115−120−125−130−135−140−145−150

(b) Input port 2 excitations

115120125130135140145

Ph

ase

diff

eren

t (d

eg)

Frequency (GHz)

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10


(c) Input port 3 excitations

Ph

ase

diff

eren

t (d

eg)

Frequency (GHz)

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10


0

−10

−20

−30

−40

−50

−60

(d) Input port 4 excitations

Figure 13: Phase difference of the proposed Butler Matrix whendifferent ports are fed.

Table 1: Output phase difference and estimated direction ofgenerated beam.

P5 (◦) P7 (◦) P6 (◦) P8 (◦) β (◦) θ (◦)

P1 45 90 135 180 45 14.4 [1L]

P2 135 0 225 90 −135 −48.6 [2R]

P3 90 225 0 135 135 48.6 [2L]

P4 180 135 90 45 −45 −14.4 [1R]

of fabrication could contribute to reduction of the insertionloss.

The simulated and measured results of the return lossat each port of the wideband 4 × 4 Butler Matrix is shownin Figure 12. For a return loss better than 10 dB, it has abandwidth about 17% centered at 9.4 GHz.

Figure 13 shows the phase difference of measured resultswhen a signal was fed into port 1, port 2, port 3, and port 4,respectively. The overall phase error was less than 7◦. Thereare several possible reasons for this phase error. A lot of bendsin high frequency can produce phase error. Moreover, theimperfection of soldering, etching, alignment, and fasteningalso could contribute to deviation of the phase error.

Table 1 shows that each input port was resulted a specificlinear phase at the output ports. The phase differences eachbetween the output ports are of the same value. The phasedifference can generate a different beam (θ). If port 1 (P1)is excited, the phase difference was 45◦, the direction ofgenerated beam (θ) will be 14.4◦ for 1L. It is summarizedin Table 1.

5. Conclusion

A novel nonuniform branch line coupler has beenemployed to achieve a wideband characteristic by expo-nential impedance taper technique. It is a simple designwithout needs of using multilayer technology and thiswill lead to cost reduction and design simplification. Thescattering matrix of the nonuniform branch line couplerwas derived and it was proved that the nonuniform branchline coupler has equal magnitude of −3 dB. Moreover, thenovel nonuniform branch line coupler has been employedto achieve a wideband 0 dB crossover. Furthermore, thesecomponents have been implemented in the Butler Matrixand that achieves wideband characteristics.

References

[1] T. A. Denidni and T. E. Libar, “Wide band four-port butlermatrix for switched multibeam antenna arrays,” in Proceedingsof the IEEE International Symposium on Personal, Indoor andMobile Radio Communications (PIMRC ’03), vol. 3, pp. 2461–2464, 2003.

[2] E. Siachalou, E. Vafiadis, S. S. Goudos, T. Samaras, C. S.Koukourlis, and S. Panas, “On the design of switched-beamwideband base stations,” IEEE Antennas and PropagationMagazine, vol. 46, no. 1, pp. 158–167, 2004.

[3] P. S. Hall and S. J. Vetterlein, “Review of radio frequencybeamforming techniques for scanned and multiple beam


antennas,” IEE Proceedings H, vol. 137, no. 5, pp. 293–303,1990.

[4] W-D. Wirth, Radar Techniques Using Array Antennas, IEEPublishers, Stevenage, UK, 2001.

[5] S. Y. Zheng, S. H. Yeung, W. S. Chan, and K. F. Man, “Broad-band butler matrix optimized using jumping genes evolu-tionary algorithm,” in Proceedings of the IEEE InternationalConference on Industrial Technology (IEEE ICIT ’08), HongKong, April 2008.

[6] K. Wincza and K. Sachse, “Broadband Butler matrix in mi-crostrip multilayer technology designed with the use ofthree-section directional couplers and phase correction Net-works,” in Proceedings of the 18th International Conference onMicrowave Radar and Wireless Communications(MIKON ’10),Cracow, Poland, June 2010.

[7] A. M. El Tager and M. A. Eleiwa, “Design and implementationof a smart antenna using butler matrix for ISM band,”in Proceedings of the Progress in Electromagnetics ResearchSymposium (PIERS ’09), pp. 571–575, Beijing, China, March2009.

[8] Y. S. Jeong and T. W. Kim, “Design and analysis of swappedport coupler and its application in a miniaturized butlermatrix,” IEEE Transactions on Microwave Theory and Tech-niques, vol. 58, no. 4, pp. 764–770, 2010.

[9] C. Collado, A. Grau, and F. De Flaviis, “Dual-band butlermatrix for WLAN systems,” in IEEE MTT-S InternationalMicrowave Symposium Digest, vol. 2005, pp. 2247–2250, 2009.

[10] K. Wincza, S. Gruszczynski, and K. Sachse, “Integrated four-beam dual-band antenna array fed by broadband Butlermatrix,” Electronics Letters, vol. 43, no. 1, pp. 7–8, 2007.

[11] T. N. Kaifas and J. N. Sahalos, “On the design of a single-layerwideband Butler matrix for switched-beam UMTS systemapplications,” IEEE Antennas and Propagation Magazine, vol.48, no. 6, pp. 193–204, 2006.

[12] K. Wincza and S. Gruszczynski, “A broadband 4 × 4 butlermatrix for modern-day antennas,” in Proceedings of the 35thEuropean Microwave Conference, pp. 1331–1334, Paris, France,October 2005.

[13] S. Gruszczynski, K. Wincza, and K. Sachse, “Reduced sidelobefour-beam N-element antenna arrays fed by 4 × 4 N butlermatrices,” IEEE Antennas and Wireless Propagation Letters, vol.5, no. 1, pp. 430–434, 2006.

[14] Y. C. Su, M. E. Bialkowski, F. C. E. Tsai, and K. H. Cheng,“UWB switched-beam array antenna employing UWB butlermatrix,” in Proceedings of the IEEE International Workshop onAntenna Technology: Small Antennas and Novel Metamaterials(iWAT ’08), pp. 199–202, Hsinchu, Taiwan, March 2008.

[15] J. He, B. Z. Wang, Q. Q. He, Y. X. Xing, and Z. L. Yin,“Wideband x-band microstrip Butler matrix,” Progress inElectromagnetics Research, vol. 74, pp. 131–140, 2007.

[16] Y. Liuqing and G. B. Giannakis, “Ultra wideband communi-cations,” IEEE Signal Processing Magazine, vol. 21, no. 6, pp.26–54, 2004.

[17] S. Banba and H. Ogawa, “Multilayer MMIC directional coup-lers using thin dielectric layers,” IEEE Transactions on Mi-crowave Theory and Techniques, vol. 43, no. 6, pp. 1270–1275,1995.

[18] J. Sebastien and G. Y. Delisle, “Microstrip EHF butler matrixdesign and realization,” ETRI Journal, vol. 27, no. 6, pp. 788–797, 2005.

[19] J. H. Cho, H. Y. Hwang, and S. W. Yun, “A design of wideband3-dB coupler with N-section microstrip tandem structure,”IEEE Microwave and Wireless Components Letters, vol. 15, no.2, pp. 113–115, 2005.

[20] M. E. Bialkowski, N. Seman, and M. S. Leong, “Design of acompact ultra wideband 3 db microstrip-slot coupler withhigh return losses and isolation,” in Asia Pacific MicrowaveConference (APMC ’09), pp. 1334–1337, St. Lucia, Australia,December 2009.

[21] R. P. Hecken, “A near-optimum matching section withoutdiscontinuities,” IEEE Transactions on Microwave Theory andTechniques, vol. 20, no. 11, pp. 734–739, 1972.

[22] D. M. Pozar, Microwave Engineering, John Wiley& Sons, NewYork, NY, USA, 2nd edition, 1998.

[23] M. Bona, L. Manholm, J. P. Starski, and B. Svensson, “Lowloss compact butler matrix for a microstrip antenna,” IEEETransactions on Microwave Theory and Techniques, vol. 50, no.9, pp. 2069–2075, 2002.

[24] M. Kobayashi and N. Sawada, “Analysis and synthesis oftapered microstrip transmission lines,” IEEE Transactions onMicrowave Theory and Techniques, vol. 40, no. 8, pp. 1642–1646, 1992.


Research Article

Isolation Improvement of a Microstrip Patch Array Antenna forWCDMA Indoor Repeater Applications

Hongmin Lee and Jinwon Park

Department of Electronic Engineering, Kyonggi University, Suwon 443-760, Republic of Korea

Correspondence should be addressed to Hongmin Lee, [email protected]

Received 15 July 2011; Revised 5 November 2011; Accepted 17 November 2011

Academic Editor: Byungje Lee

Copyright © 2012 H. Lee and J. Park. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper presents the isolation improvement techniques of a microstrip patch array antenna for the indoor wideband codedivision multiple access (WCDMA) repeater applications. One approach is to construct the single-feed switchable feed networkstructure with an MS/NRI coupled-line coupler in order to reduce the mutual coupling level between antennas. Another approachis to insert the soft surface unit cells near the edges of the microstrip patch elements in order to reduce backward radiationwaves. In order to further improve the isolation level, the server antenna and donor antenna are installedin orthogonal direction.The fabricated antenna exhibits a gain over 7 dBi and higher isolation level between server and donor antennas below −70 dB atWCDMA band.

1. Introduction

Microstrip patch antennas are used in modern communica-tion systems due to their low cost, lightweight, and planarstructure. However, a microstrip patch antenna has somedisadvantages.

In the design of a microstrip patch array antenna, theradiating patch elements have to be placed in close prox-imity. It produces significant mutual coupling effect thatdeteriorates the antenna performance. The simplest solutionto improve the antenna isolation is to increase the physicalseparation between the antennas. Although the isolationof the antennas becomes higher as the separation lengthis increased, the size of the antenna will subsequently belarger. However, the size of the antenna ground plane islimited in the practical microstrip patch antenna design.Particularly when the patch antenna is printed on highdielectric substrates, its radiation pattern is considerablyaffected by surface waves. On a finite ground plane, surfacewaves propagate until they reach an edge where they arereflected back and diffracted by the edges. These causea significant amount of wasted power in the backwardhemisphere below the ground plane. Various solutions forthe improvement of the radiation performance of patch

antennas on a substrate are available. One approach is to con-struct a periodic structure, such as photonic bandgap (PBG)or electromagnetic bandgap (EBG) [1, 2], surrounding thepatch antenna. Another approach is to use the concept ofartificial soft and hard surfaces [3]. These surfaces have ledto a wide range of applications in antennas other microwavesystems [4]. Both EBG and soft surfaces can be used tosuppress surface wave propagation. The main difference isthat soft surfaces exhibit bandgaps in only one direction,but they offer the best performance in most applicationsof antennas. However, the conventional soft surfaces aremade on the same ground plane of a patch antenna andthe presence of via near the patch may raise the resonantfrequency of the patch if the soft surface strip or cell isvery close to the patch. On the other hand, it requires aconsiderable area to form a bandgap structure.

The purpose of this paper is to present a new method forenhancing an isolation of a microstrip patch array antenna,whose Tx/Rx band is located very near. In order to reducethe mutual coupling between antennas that are located in thesame plane, a passive switching feed network is used [5]. Anew isolated soft surfaces structure that does not share theground plane of a patch is also used for the reduction of thesurface wave. The proposed new soft surface consists of a


number of metal mushroom-type structures that is locatednear the edges of substrate, and it is not connected with aground plane [6, 7]. A two-by-two microstrip patch arrayantenna for the indoor wideband code division multipleaccess (WCDMA) repeater applications is designed andtested experimentally.

2. Switchable Feed Network for Antenna

The schematic of a single-feed switchable feed networkis shown in Figure 1. It basically consists of two quarter-wavelength branch lines with characteristic impedances ofZ1 and Z2 lengths of l1 and l2. Two different rectangularmicrostrip patch antennas of different sizes are connected tooutput ports 2 and 3. The resonant frequencies of these twoantennas are f1 and f2, and the input impedances of eachantenna are ZL1 and ZL2. Lengths of l1 and l2 are chosen asa quarter wavelength of λ1/4 and λ2/4, respectively. When asignal with a frequency of f = f1 is applied to port 1, theinput impedance of the upper branch Zin1 is

Zin1 = Z21

ZL1. (1)

If the resonant frequencies of two patches are very closelylocated ( f1 ≈ f2), the two quarter wavelength branch linelengths become similar (l1 ≈ l2). Since the patch antennaconnected to port 3 is off-resonant, the input impedance ofthe antenna is almost reactive (ZL2 ≈ jX). Hence, the inputimpedance of the lower branch Zin is

Zin2 = Z2ZL2 + jZ2 tanβ2l2Z2 + jZL2 tanβ2l2

≈ − jZ2

2

x. (2)

When the patch antenna was off-resonant, its input reactanceneared zero (x → 0). Using x → 0 in (2), the inputimpedance of port 2 neared infinity (Zin2 → ∞). Therefore,it acts as an open circuit. As a result, the patch at port 2would be at resonance with the patch at the off-resonantport 3. If a signal with a frequency of f = f2 is appliedto port 1, the patch at port 3 will be at resonance with theoff-resonant patch at port 2. This single-feed switchable feednetwork combines two ports of a microstrip antenna witha quarter wavelength feed line and each resonant frequencyof the antennas very closely located; it acts as an idealsingle-pole double-throw (SPDT) switch. The operationof the single-feed switchable feed network when Z1 =Z2 = ZL1 = ZL2 = 50 ohm is presented in Table 1. Inorder to further enhance the isolation between the ports,microstrip/negative-refractive-index (MS/NRI) coupled linecoupler is used with the switchable feed network. An MS/NRIcoupler has advantages, compared to conventional edge-coupled microstrip couplers in terms of coupled power andport isolation [8]. Figure 2 shows the geometry of the 3 dBMS/NRI coupled-line coupler for WCDMA frequency band(1.92–2.17 GHz) applications. It consists of a conventionalMS right-handed transmission line edge-coupled with aNRI transmission line used in its left-handed range. A3 dB implementation of the MS/NRI coupled-line couplerrequires the equality of the magnitude at port 2 and port

Zin

Port 1

Z0

(Input)

Zin 1

Zin 2

l1

l2

Z1

Z2

Port 2

ZL1

(Output)

Port 3

ZL2

(Output)

Figure 1: Schematic of a single-feed switchable feed network.

11.5 mm

Port 3

Port 1

15.7 mm

Port 4

Port 2

0.85 mm

Figure 2: Geometry of 3 dB MS/NRI coupled-line coupler.

Table 1: The operation of the single-feed switchable feed network.

f [GHz] Port Zin ZL Switch state

f = f1Port 2 50Ω 50Ω “On”

Port 3 ≈ ∞Ω jx ≈ 0Ω “Off”

f = f2Port 2 ≈ ∞Ω jx ≈ 0Ω “Off”

Port 3 50Ω 50Ω “On”

3 (|S21| = |S31|). Usually the total length of the couplerand the spacing between the lines should be adjusted until a3 dB coupling level is acquired. A 3 dB MS/NRI coupled-linecoupler is designed on a Rogers RO3210 substrate (relativedielectric constant = 10.2, thickness = 2.54 mm).

The NRI transmission line was 15.7 mm long and theMS-line was 11.5 mm long. The spacing between MS/NRIlines is 1 mm. Figure 3 presents the simulated scatteringparameter results for the MS/NRI coupled-line coupler ofFigure 2. Quasi-3 dB backward coupling is achieved overthe range from 1.7 to 2.7 GHz. The magnitudes of thesimulated S-parameter are listed in Table 2. Figure 4 showsthe geometry of the feed network for two-by-two microstrippatch array antenna. It is fed by a 50Ω coaxial probe anddesigned on a Rogers RO3210 substrate to allow the feedingof a four-element microstrip patch antenna array. The overalldimension of the ground plane is 120 mm × 120 mm. Theaimed two switching frequencies of the proposed switchablefeed network are 1.95 and 2.14 GHz, respectively.

These frequencies are the center of the Tx and Rx bandin the WCDMA system. In Figure 4, all the ends of eachport of the proposed feed network for two-by-two microstrippatch antenna are made with 50Ω microstrip lines. Thecharacteristic impedances of an NRI transmission line andthe MS line were Z1 = 90Ω and Z2 = 75Ω, respectively.


−70

−60

−50

−40

−30

−20

−10

0

(dB)

Frequency (GHz)

S11

S21 S41

S31

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

S-pa

ram

eter

Figure 3: Simulated S-parameter results for the MS/NRI coupled-line coupler of Figure 2.

Table 2: Simulated S-parameters for the MS/NRI coupled-linecoupler.

S-parameter 1.95 GHz [dB] 2.14 GHz [dB]

S11 −12.5 −13.4

S21 −3.3 −3.6

S31 −3.6 −3.1

S41 −27.5 −32.5

Table 3: Simulated S-parameters of the feed network for a four-element microstrip patch array.

Frequency[GHz]

1.94 2.15

Port isolation [dB]

S-parameterWithoutMS/NRIcoupler

WithMS/NRIcoupler

WithoutMS/NRIcoupler

WithMS/NRIcoupler

S23 −15.18 −18.81 −11.18 −18.78

S24 −14.66 −24.73 −12.20 −17.12

S25 −7.03 −23.25 −7.8 −19.57

S32 −15.18 −19.81 −11.18 −23.09

S34 −8.50 −24.01 −8.04 −30.99

S35 −16.88 −26.69 −10.43 −18.91

S42 −14.66 −24.73 −12.20 −17.10

S43 −8.50 −24.01 −8.04 −30.99

S45 −16.31 −18.26 −11.67 −20.54

S52 −7.03 −23.25 −7.8 −19.57

S53 −16.88 −26.69 −10.43 −18.92

S54 −16.04 −18.26 −11.67 −20.54

In order to match these impedances with 50Ω port, a singlesection of quarter wavelength impedance transformer wasadded. The proposed feed network was designed by using

120

mm

120 mm

1

23

4 5

Figure 4: Geometry of the feed network for two-by-two microstrippatch array antenna.

1.27 mm2.54 mm

120 mm

120

mm

1 mm

34 mm 30 mm

Rx

Rx Tx

Tx

z x

y

34 mm 30 mm

Rx

Rx Tx

Tx

Figure 5: Geometry of the proposed switchable feed network fortwo-by-two microstrip patch array antenna.

switched feeder network and a signal was excited at port 1as shown in Figure 4. When the antennas connected at port2 and port 4 were in an on state, the antennas connected atport 3 and port 5 were in an off state. However, the proposedswitchable feed network shown in Figure 1 can operate whenthe antennas are connected at each port.

For the simulation of the port isolation, all the endsof each port were terminated with microstrip patch anten-nas. The simulated scattering parameters of the proposedswitchable feed network compared to those of the switchablefeed network with/without MS/NRI coupled-line coupler arelisted in Table 3. The switchable feed network with MS/NRIcoupled-line coupler exhibits much higher isolation thanthe switchable feed network without MS/NRI coupled-line


(a) 1.95 GHz (b) 2.15 GHz

Figure 6: Simulated surface current distributions.

180 mm

180

mm

(a) Perspective view

19 mm

19 mm

20 mm

(b) Side view

Figure 7: Geometry of the proposed indoor WCDMA repeater antenna.

coupler. When the MS/NRI coupled-line coupler was notused, the scattering parameters (S25, S34, S52, S43) betweentwo antennas showed lower isolation. Since the patchantennas were designed to be excited at their fundamentalresonant mode (TM10), strong coupling is produced betweentwo antennas that are placed parallel to the radiating edgedirection. It can be seen that the minimum isolation levelsbetween the output ports exceed −18 dB at the aimedtwo switching frequencies of the proposed switchable feednetwork.

3. Antenna Design

Figure 5 shows the geometry of the two-by-two microstrippatch array antenna using the proposed switchable feednetwork and proximity coupled square microstrip patchelements.

The antenna consists of three-layer structure: an air layerhaving a thickness of 1 mm, a dielectric substrate layer,and a stacked microstrip patch layer. The square microstrippatches are fed by microstrip lines from perpendiculardirections using the proximity coupled method. This switch-able feed network was designed to achieve beam patternreconfigurable array antenna, which generates ±45◦ linearlypolarized slanted beam patterns at the Tx/Rx frequencybands. In order to dual slant beam, two Rx and Tx microstrippatch elements were placed orthogonally. Figure 6 showsthe simulated surface current distribution at the resonantfrequency of 1.95 GHz (Tx) and 2.15 GHz (Rx). At thefrequency of 1.95 GHz, two patch antennas for Tx bandare resonant, and two patch antennas for Rx-band are off-resonant. As a result, most of the surface currents flowthrough the feed line for Tx-band antennas. On the otherhand, at the frequency of 2.15 GHz, two patch antennas for


12 mm20 mm

3 mm3 mm

1.27 mm

w

l

(a) Parallel plate waveguide model with a cell

l

w

h

(b) Geometry of unit cell

−80

−70

−60

−50

−40

−30

S-pa

ram

eter

(dB)

1.9 1.95 2 2.05 2.1 2.15 2.2

Frequency (GHz)

w = 6.6 mmw = 6.8 mmw = 7 mm

w = 7.2 mmw = 7.4 mm

(c) The simulated transmission characteristics with different sizes of unitcells

Figure 8: Parallel plate waveguide model with different sizes of unit cells.

Rx band are resonant. The perspective view of a microstrippatch array antenna system for the indoor WCDMA repeateris shown in Figure 7. The proposed repeater antenna consistsof a server antenna, a donor antenna, and alumina housing.It occupies a volume of 180 mm × 180 mm × 20 mm.In order to reduce the surface waves radiation from aserver antenna and a donor antenna, corner-edged viamushroom-type unit cells are formed near the edges of theupper dielectric substrate. Figure 8 shows the parallel platewaveguide model with different sizes of unit cells and theresults of the simulated transmission characteristics as afunction of frequency. As shown in Figure 8(b), the unit cellconsists of two parallel rectangular plates with the same size(w = l) and a corner-edged via. The height of the via is1.27 mm. Inside two rectangular plates, dielectric material(relative dielectric constant = 10.2, thickness = 1.27 mm) isplaced. The transmission coefficient S21 of the parallel platewaveguide ports without cell exhibits near −43 dB, as shownin Figure 8(c). When the unit cell is inserted between twoparallel plate waveguides, a stop band occurs at a certain fre-quency.

It depends on the physical size of the unit cell However,the bandwidth of each stop band shows very narrowcharacteristics due to the resonant nature of the unit celland a high dielectric constant of the substrate. In order tocover the bandwidth within the WCDMA frequency band,three different sizes of unit cell array configuration structure(w = l = 6.6 mm, 7.0 mm, 7.4 mm) were used in this work.In addition, a server antenna and a donor antenna backed bythe alumina housings are arranged in orthogonal directionin order to get higher isolation between two antennas.

4. Experimental Result

The photographs of the fabricated two-by-two microstrippatch array antenna structure are shown in Figure 9. Theswitchable feeder layer and a stacked microstrip patch layerare etched on a Rogers RO3210 substrate (relative dielectricconstant = 10.2) having different thickness of 2.54 mmand 1.27 mm, respectively. The characteristic impedanceof each of the two branch feed lines from the coaxial


(a) Top view (b) Switchable feed network

(c) Ground plane (d) Antenna with housing

Figure 9: Photographs of the fabricated two-by-two microstrip patch array antenna.

−60

−50

−40

−20

−10

0

S-pa

ram

eter

(dB

)

1.5 2 2.5

Frequency (GHz)

S11 W/O cellS11 with cell

S21 W/O cell

S21 with cell

Figure 10: Measured return loss and isolation characteristics oftwo-by-two microstrip patch array antenna.

probe was 100Ω and the input impedance of each of thepatch elements with a proximity coupled microstrip feed is50Ω. A comparison of the measured return loss (S11) andisolation (S21) between a server antenna and a donor antennawith/without isolated soft surface unit cells is shown inFigure 10. It is noted that the resonant frequencies are barelychanged.

In the higher and lower bands, the measured −10 dBreturn loss bandwidths are about 84 MHz and 96 MHz,respectively. It meets the bandwidth requirement forWCDMA (1.92–2.17 GHz) applications. Compared to thefabricated antenna without isolated soft surface unit cells,the fabricated antenna with isolated soft surface unit cellsexhibits higher isolation level. When the isolated soft surfaceunit cells are used, the maximum isolation level at thefrequency of 1.94 and 2.15 GHz is −92 dB and −70 dB,respectively. The measured far-field radiation patterns inthe x-y plane (θ = 0◦) and y-z plane (φ = 0◦) at thefrequency of 1.94 and 2.15 GHz are shown, respectively, inFigure 11. It shows linear polarized radiation patterns, andthe main direction of the radiated power was changed due


10

5

0

−5

−10

−15

−20

−25

−30

−25

−20

−15

−10

−5

0

510

0

30

60

90

120

150

180

210

240

270

300

330

(a) x-y plane (1.94 GHz)

10

5

0

−5

−10

−15

−20

−25

−30

−25

−20

−15

−10

−5

0

510

0

30

60

90

120

150

180

210

240

270

300

330

(b) y-z plane (1.94 GHz)

10

5

0

−5−10−15−20−25−30−25−20−15−10−5

0

5

10 0

30

60

90

120

150

180

210

240

270

300

330

Co-polCross-polPower sum

(c) x-y plane (2.15 GHz)

105

0

−5

−10

−15

−20

−25

−30

−25

−20

−15

−10

−5

0

5

100

30

60

90

120

150

180

210

240

270

300

330

Co-polCross-polPower sum

(d) y-z plane (2.15 GHz)

Figure 11: Measured far-field radiation patterns of the fabricated antenna.

to the placement of two patches. The main beam directionwas slanted about −45◦ for the lower band (Tx) and +45◦

for the higher band (Rx). The measured peak gain/radiationefficiency for Tx/Rx band was 7.1 dBi/77% and 8.9 dBi/80%,respectively.

5. Conclusion

The new techniques for the isolation improvement of amicrostrip patch array antenna have been presented. Thetwo main techniques presented here are (1) the single-feed

switchable feed network structure with MS/NRI coupled-line coupler for higher isolation level and (2) the isolatedsoft surface unit cells structure for reducing the surfacewaves. Both structures have been discussed in the paperthrough proper numerical simulation. In order to improvethe isolation further, the server antenna and donor antennafor an indoor repeater system were placed orthogonally.As a result, the fabricated server and donor antennas havesmall separation of 20 mm and exhibit higher isolation level.Experimental results shows that the maximum isolation levelat the frequency of 1.94 and 2.15 GHz is −92 dB and −70 dB,respectively. The proposed techniques can be easily used for


the design of the microstrip patch array antenna with higherisolation level.

Acknowledgment

This research was supported by the Basic Science ResearchProgram through the National Research Foundation of Korea(NRF) funded by the Ministry of Education, Science andTechnology (no. 2010-0011646).

References

[1] Y. J. Park, A. Herschlein, and W. Wiesbeck, “A photonicbandgap (PBG) structure for guiding and suppressing surfacewaves in millimeter-wave antennas,” IEEE Transactions onMicrowave Theory and Techniques, vol. 49, no. 10, pp. 1854–1859, 2001.

[2] D. Sievenpiper, L. Zhang, R. F. J. Broas, N. G. Alexopolous, andE. Yablonovitch, “High-impedance electromagnetic surfaceswith a forbidden frequency band,” IEEE Transactions onMicrowave Theory and Techniques, vol. 47, no. 11, pp. 2059–2074, 1999.

[3] P. S. Kildal and A. Kishk, “EM modeling of surfaces with stop orgo characteristics—artificial magnetic conductors and soft andhard surfaces,” Applied Computational Electromagnetics SocietyJournal, vol. 18, no. 1, pp. 32–40, 2003.

[4] P. S. Kildal, A. A. Kishk, and S. Maci, “Special issue on artificialmagnetic conductors, soft/hard surfaces, and other complexsurfaces,” IEEE Transactions on Antennas and Propagation, vol.53, no. 1, part 1, pp. 2–7, 2005.

[5] H. M. Lee, “Pattern reconfigurable microstrip patch arrayantenna using switchable feed-network,” in Proceedings of theAsia-Pacific Microwave Conference (APMC ’10), pp. 2017–2020,December 2010.

[6] H. M. Lee and J. K. Kim, “Front-to-back ratio improvementof a microstrip patch antenna using an isolated soft surfacestructure,” in Proceedings of the European Microwave Conference(EuMC ’09), pp. 385–388, October 2009.

[7] J. H. Kim and H. M. Lee, “Backward wave reduction of amicrostrip patch antenna using dual-band isolated soft surfacestructures,” in Proceedings of the IEEE International Symposiumon Antennas and Propagation Society (AP-S ’10), pp. 1–4, July2010.

[8] R. Islam and G. V. Eleftheriades, “A planar metamaterial co-directional coupler that couples power backwards,” in Proceed-ings of the IEEE MTT-S International Microwave SymposiumDigest, vol. 1, pp. 321–324, June 2003.


Research Article

Series-Fed Microstrip Array Antenna with Circular Polarization

Tuan-Yung Han

Department of Computer and Communication Engineering, Chienkuo Technology University, Chang-Hua City 500, Taiwan

Correspondence should be addressed to Tuan-Yung Han, [email protected]

Received 21 July 2011; Revised 3 November 2011; Accepted 5 November 2011

Academic Editor: Miguel Ferrando

Copyright © 2012 Tuan-Yung Han. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This study proposes a novel 2 × 2 array antenna design with broadband and circularly-polarized (CP) operation. The proposeddesign uses a simple series-fed network to increase the CP bandwidth without requiring one-by-one adjustment of each arrayelement or a complex feed network. Selecting the appropriate spacing between each array element allows the proposed arrayantenna to generate CP radiation with a low axial ratio. Experimental results based on a prototype show that this 2 × 2 microstriparray antenna achieves a wide 3 dB axial ratio bandwidth of more than 10%. Simulated data are also provided to confirm themeasured results.

1. Introduction

Due to their attractive features, such as low profile, lightweight, and ease of manufacturing using printed circuit tech-niques, microstrip array antennas have come into highdemand for satellite communication applications. To gen-erate circularly-polarized (CP) radiation from a microstriparray antenna, previous research recommends a simpledesign method that uses a corporate-fed network [1] to exciteeach array element simultaneously. To allow each array ele-ment to exhibit equal power amplitude and phase distribu-tion, most corporate-fed networks employ transmission lines(of the same length) and power dividers. However, the CPbandwidth of an array antenna design with a corporate-fed network is usually limited to that of a single array ele-ment. Thus, previous studies propose the method of apply-ing sequential rotation techniques to the feed network toimprove the CP bandwidth of an array antenna effectively[2–4]. Due to the sequentially rotated structure of the feednetwork (with a single feed point), the phases of the fourarray elements (2 × 2 array) are usually orientated sequen-tially at 0◦, 90◦, 180◦, and 270◦. This offers an impedancebandwidth approximately three times wider than that of asingle array element. In this design, the designated arrayelement can be either linearly polarized (LP) [5] or CP type.If an LP array element is used, the design process of the arrayantenna (with sequential rotation technique) will be easier

than that for the CP array element type. This is becausethe process of adjusting the CP performance of each arrayelement can be avoided. However, the resulting compoundarray gain is 3 dB lower than the array using the CP element.Sequential rotation of radiating elements can increase theinput impedance bandwidth and polarization purity andachieve a good symmetrical radiation pattern [6]. However, asequential rotation feed network must provide a delay line toallow various feed point with different phases to connect toeach array element. This results in a relatively complicatedcircuit layout compared to the traditional corporate- orseries-fed network. Several studies indicate that a corporate-fed array antenna using sequential rotation techniques [7, 8]can improve the bandwidth of a corporate-fed array an-tenna and enhance the purity of polarization. However, dis-advantages such as a complex feed network still exist, creatingthe possibility of producing multiple reflections between theelement and the feed.

This study proposes a novel 2 × 2 CP microstrip arrayantenna design using series-fed lines. Experimental resultsshow that the proposed array design and one using sequen-tial rotation-fed network achieve similar levels of CP perfor-mances, including CP bandwidth and peak gain. Since theproposed array is fed by a simple series-fed network, thereis no need for an elaborate phase shifting or power dividingcircuit. This study presents both simulated and measured re-sults for the proposed design.


y

16

5

11

10.5

3

13.51

80

80

(a) (b)

150

1503

1

11

13.5

d

x

1.6 mm FR4

Circular patch

Unit: mm

Ring slot

Ground plane

Foam

0◦

270◦ 180◦

90◦

Open stub

Figure 1: Geometry of the studied microstrip antennas. (a) Single ring-slot-coupled microstrip antenna. (b) 2 × 2 CP microstrip arrayantenna.

2. Array Element Structure

Figure 1(a) illustrates the geometry of a single aperture-coupled microstrip antenna. Two orthogonal modes of theantenna can be excited in series by a microstrip feed linethrough the coupling of the annular-ring slot in the groundplane. Choosing appropriate length (optimum at 11 mm) forthe open stub in the feed line allows the antenna to generategood CP radiation. A previous study analyzes the parametersand design procedure of this aperture-coupled microstripantenna [9]. A prototype was first fabricated according tothe dimensions revealed in Figure 1(a). Figure 2 shows themeasured results along with the simulated results, which

were generated using IE3D software. Experimental resultsindicate that the CP operating bandwidth of the prototype(also defined as 3 dB axial ratio) is approximately 150 MHz(5.1%) with respect to the center frequency measured at2950 MHz. Since the prototype is microstrip-fed, it can bearranged in an array element manner.

3. New Array Design and Results

Figure 1(b) depicts the proposed CP microstrip array anten-na. Four ring-slot-coupled microstrip antenna elements wereexcited through a series-fed network. Simulation results sug-gest that the element spacing, d, is the dominant parameter


0

5

10

15

20

25

30

35

402.6 2.7 2.8 2.9 3 3.1 3.2 3.3

Ret

urn

loss

(dB

)

Frequency (GHz)

(a)

0

1

2

3

4

5

6

7

8

9

10

2.6 2.7 2.8 2.9 3 3.1 3.2 3.3

Axi

al r

atio

(dB

)

Frequency (GHz)

measuredsimulated

(b)

Figure 2: Measured and simulated results for the single CP ring-slot-coupled microstrip antenna. (a) Return loss. (b) Axial ratio.

point point

(a) (b)

Feed Feed0◦

90◦

180◦

270◦

Figure 3: Structure of the reference antennas. (a) Corporate feed network, Array B. (b) Sequential rotation feed networks, Array C.

to determine the CP performance of the proposed array, andthe optimum value is approximately 73 mm for the studiedstructure (d ∼ 0.76λ0, where λ0 is referred to the center fre-quency of the CP operating bandwidth). To confirm thesimulated results, a prototype (Array A) was constructedbased on the dimensions presented in Figure 1(b). Two otherprototypes, Arrays B and C, using the corporate and sequen-tial rotation feed networks, respectively, were also con-structed as references (Figure 3). Except for different feedingnetworks, the three array prototypes were designed with thesame structure, antenna dimensions, and element spacing.Due to the simplicity of series-fed network design, antennaengineers can ignore the process of adjusting the CP char-acteristics of each array element. However, the complex feednetwork designs in sequential rotational-fed and corporate-fed array antennas necessitate tuning the CP performance ofeach array element.

Figures 4(a) and 4(b) present the measured return lossand axial ratio against frequency for the respective array pro-

totypes. Figure also presents simulated results for Array A,indicating that they agree with the experimental results.These results confirm that all of the tested array prototypeshave good impedance matching within their CP operatingbandwidths. In addition, the CP bandwidth (4.7%) of ArrayB is almost the same as that of its array element, and theCP bandwidth (13.8%) of Array C is about three times thatof Array B. As for Array A, the CP bandwidth centered at3110 MHz is approximately 11.5% which is slightly less thanthat of Array C.

Figure 5 plots the measured radiation patterns of ArrayA at 3110 MHz, revealing good left-hand CP (LHCP) radia-tions in the broadside direction. The main beam tilts slightly(about 3 degrees) to the left side in the x-z plane. This mightbe because the array elements are not fed with an exactlyequal power level. Nevertheless, the measured peak gain ofArray A is approximately 12 dBi, which is only 0.3 dB lowerthan Array C. This small difference might be due to the feed-ing phase errors of each element. Table 1 summarizes the ex-


0

5

10

15

20

25

30

35

402.6 2.7 2.8 2.9 3 3.1 3.2 3.3

Frequency (GHz)

Array AArray B

Array CArray A, simulated

Ret

urn

loss

(dB

)

(a)

0

1

2

3

4

5

6

7

8

9

10

2.6 2.7 2.8 2.9 3 3.1 3.2 3.3

Frequency (GHz)

Axi

al r

atio

(dB

)

Array AArray B

Array CArray A, simulated

(b)

Figure 4: Measured results for the three array prototypes; d = 73 mm. (a) Return loss. (b) Axial ratio.

0

10

20

0306090120150180

Mag

nit

ude

(dB

)

Angle (deg)

−30

−40

−20

−10

−180−150−120−90−60−30

(a)

0306090120150180

Mag

nit

ude

(dB

)

Angle (deg)

LHCP

RHCP

0

10

20

−180−150−120−90−60−30

−30

−40

−20

−10

(b)

Figure 5: Radiation patterns of Array A measured at 3110 MHz. (a) x-z plane. (b) y-z plane.

Table 1: Experimental results for the studied single element andarray prototypes.

3 dB axial-ratiobandwidth (MHz, %)

CP centerfrequency (MHz)

Peak gain(dBi)

Single element 150, 5.1 2950 7

Array A 360, 11.5 3110 12

Array B 140, 4.7 2980 12.4

Array C 410, 13.8 2965 12.3

perimental results of CP performance for all three array pro-totypes. These results show that the CP bandwidth of ArrayA is approximately 6.8% larger than that of Array B, and2.3% smaller than that of Array C. Although array A demon-strates a slightly narrower bandwidth than array C, its peakgain measured in the boresight direction is almost similar(compared to Arrays B and C) at approximately 12 dBi. ArrayA also possesses a very simple feed network structure thatdoes not require other circuit features, such as phase shiftingor power dividing circuits. Thus, Array A is a preferred can-didate compared to Array C if a simple feed network struc-ture is required.

4. Conclusions

This study presents a 2 × 2 circularly polarized microstriparray antenna using a series-fed network. Experimental re-sults indicate that this array has a broad CP operating band-width and an acceptable antenna gain. Moreover, the pro-posed design is relatively simple compared to the traditionalarray antenna, which uses a corporate or sequential rotationfeed network.

References

[1] P. S. Hall and C. M. Hall, “Coplanar corporate feed effects inmicrostrip patch array design,” IEE Proceedings H, vol. 135, no.3, pp. 180–186, 1988.

[2] J. W. Baik, K. J. Lee, W. S. Yoon, T. H. Lee, and Y. S. Kim,“Circularly polarised printed crossed dipole antennas withbroadband axial ratio,” Electronics Letters, vol. 44, no. 13, pp.785–786, 2008.

[3] M. Elhefnawy and W. Ismail, “A microstrip antenna array forindoor wireless dynamic environments,” IEEE Transactions onAntennas and Propagation, vol. 57, no. 12, pp. 3998–4002, 2009.

[4] R. Caso, A. Buffi, M. Rodriguez Pino, P. Nepa, and G.Manara, “A novel dual-feed slot-coupling feeding technique forcircularly polarized patch arrays,” IEEE Antennas and WirelessPropagation Letters, vol. 9, pp. 183–186, 2010.


[5] J. Huang, “A technique for an array to generate circular polar-ization with linearly polarized elements,” IEEE Transactions onAntennas and Propagation, vol. AP-34, pp. 1113–1124, 1986.

[6] H. Evans, P. Gale, and A. Sambell, “Performance of 4 × 4sequentially rotated patch antenna array using series feed,”Electronics Letters, vol. 39, no. 6, pp. 493–494, 2003.

[7] D. C. Chung, S. Y. Choi, Y. H. Ko, J. H. Lee, and M. H.Kwak, “Circularly polarized HTS microstrip antenna array,”IEEE Transactions on Applied Superconductivity, vol. 13, no. 2,pp. 301–304, 2003.

[8] M. N. Jazi and M. N. Azarmanesh, “Design and implementa-tion of circularly polarised microstrip antenna array using anew serial feed sequentially rotated technique,” IEE Proceedings:Microwaves, Antennas and Propagation, vol. 153, no. 2, pp. 133–140, 2006.

[9] J. S. Row, “Design of aperture-coupled annular-ring microstripantennas for circular polarization,” IEEE Transactions on Anten-nas and Propagation, vol. 53, no. 5, pp. 1779–1784, 2005.


Research Article

Vertical Meandering Approach for Antenna Size Reduction

Li Deng,1, 2 Shu-Fang Li,1 Ka-Leung Lau,2 and Quan Xue2, 3

1 Key Laboratory of Universal Wireless Communication, Beijing University of Posts and Telecommunications,Beijing 100876, China

2 Shenzhen Research Institute, City University of Hong Kong, Shenzhen 518057, China3 State Key Laboratory of Millimeter Waves, City University of Hong Kong, Kowloon, Hong Kong

Correspondence should be addressed to Li Deng, [email protected]

Received 15 July 2011; Accepted 30 September 2011

Academic Editor: Ahmed A. Kishk

Copyright © 2012 Li Deng et al. This is an open access article distributed under the Creative Commons Attribution License, whichpermits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A novel vertical meandering technique to reduce the lateral size of a planar printed antenna is presented. It is implemented bydividing a conventional spiral patch into a different number of segments and placing them on different sides of the microwavesubstrate with vias as the connections. To confirm the validity of this technique, measured electrical performance and radiationcharacteristics of five antennas with different numbers of segments are compared. The smallest antenna is reduced in size by 84%when compared with the conventional printed spiral antenna.

1. Introduction

The topic of antenna miniaturization has been a subject ofinterest for more than half a century, but in recent years, ithas attained significant attention because of an exorbitantdemand for mobile wireless communication systems. Theneed for antenna miniaturization stems from the fact thatmost mobile platforms have a limited space for all ofthe required antennas in ever-increasing wireless systems.Miniature antennas are in high demand, since the antennasize often imposes a significant limitation on the overall sizeof a portable wireless system.

An example of a miniaturized antenna is a meanderedantenna where a half wavelength dipole is made compact bymeandering the wire [1]. A similar approach can be appliedto design a meander-type slot antenna [2]. Meanderingthe excited patch surface current paths in the microstripantenna’s radiating patch is also an effective method forachieving a lowered fundamental resonant frequency for themicrostrip antenna [3–7]. On the other hand, meanderedantennas are very hard to match to a 50 Ohm line. Thisdifficulty is due to the fact that the radiation of almostin-phase electric currents flowing in opposite directionson closely spaced wires or patches tends to cancel each

other in the far-field region. This cancellation renders aconsiderable portion of opposing currents ineffective as faras radiation efficiency is concerned and leads to a very lowradiation resistance that might have increased using a two-strip meandered line [8]. Consequently, these antennas aredifficult to match, and yet require a very low temperature ofoperation to control material losses [9].

The technique for lengthening the current path men-tioned above is almost based on a coplanar or single-layermicrostrip structure. The current path lengthening for afixed patch antenna can also be obtained by using verticalmeandering technique. By vertical meandering, we can usethe area of substrate more efficiently. Furthermore, we canreduce the cancellation in the above-mentioned conventionalmeandering techniques effectively by rational design of thevertical meandering lines.

In this paper, a novel vertical meandering techniqueaimed at reducing the size of a printed spiral patch antennais proposed. By utilizing this technique on a single turnconventional printed spiral patch antenna, its resonantfrequency can maximally drop from 1.53 GHz to 0.61 GHz,which is 60%. Therefore, the antenna size can be reduced by84%. The antenna finds a reasonable impedance bandwidth(S11 ≤ −10 dB) of 2.6%.


Top sideBottom side

a

b

L2

L1

y

xFeeding probe

Ground plane

r2 r1

wpatch

50 Ohm microstrip line

Spiral-shaped

(a)

Top sideBottom side

a

b

L2

L1

y

x

Via

Ground plane

Feeding probe

r2

r1

w

(b)

Via Top side

Bottom side

Ground planeFeeding probe

z

x

y

(c)

Figure 1: (a) Top view of the conventional printed spiral antenna. (b) Top view of the novel printed spiral antenna. (c) 3D view of the novelstructure.

2. Antenna Structure

Figure 1(a) shows the geometry of a conventional printedspiral antenna. It consists of a small rectangular ground planeetched on the bottom side of a F4B substrate, which hasthickness of 2 mm and dielectric constant of 2.65. A singlespiral-shaped patch and a 50Ω microstrip line are printedon the top side of the same substrate. This line is coaxialfed by a 50Ω SMA (SubMiniature version A) connectorunderneath the ground plane. The feeding probe has radiusof 0.5 mm. In order to reduce the lateral size of this antenna,a novel technique is applied. This technique is implementedby separating the conventional spiral patch in Figure 1(a)into a different number of segments (denoted by N). Then,half of these segments are moved to the bottom side of

the microwave substrate. Finally, vertical conducting vias areused to connect the corners of the segments on the topand bottom layers together (as is well known, the vias areused widely in PCB for connections between layers, it is easyto fabricate). For maximization of the size reduction, thevertical vias should be staggered arranged. Staggered vias atthe cross-corners can make the distance of two adjacent viaslarger than center arranged vias. Thus, staggered arrangedvias have smaller coupling capacitances than center arrangedvias; then the effect of cancellation to the inductances issmaller than center arranged vias too. Therefore, for thewhole antenna with the same length, the staggered vias atthe cross-corners can have larger inductances than centerarranged vias, and larger inductance leads to larger sizereduction. The proposed antenna structure is depicted in


Figures 1(b) and 1(c). It is identical to the conventional spiralantenna in Figure 1(a) when N is equal to 1. Each via hasthe same radius of 0.3 mm. For the proposed antenna, thephysical length of the current flowing path is increased withthe number of segments. Therefore, its resonant frequencyis decreased with the value N . As a result, a larger N canachieve a smaller antenna size. To prove the effectivenessof the proposed size reduction technique, five printed spiralantennas withN equals to 1, 16, 32, 64, and 128 are fabricatedand tested. The other parameters of the antennas are: r1 =15 mm, r2 = 25 mm, w = 5 mm, L1 = 7 mm, L2 = 86 mm,a = 70 mm, b = 92 mm.

3. Results and Analysis

In this paper, we make a fully simulation through a MoM-(Method of Moments-) based software, IE3D. Figure 2 showsthe simulated input return loss of these antennas. It is clearlyseen that the antenna is resonant at 1.49 GHz when N =1. If N is increased to 16, 32, 64, and 128, the resonancefrequency is reduced to 1.31 GHz, 1.14 GHz, 0.89 GHz,and 0.61 GHz, respectively. Therefore, the suppression inresonant frequency is 12.1%, 23.5%, 40.3%, and 59.1%,respectively. The reduction in patch size is 22.7%, 41.5%,64.3%, and 85.5%. The impedance bandwidths (S11 ≤−10 dB) of these five antennas are 12.3%, 11.6%, 8.7%,5.08%, and 2.3%, respectively.

Figure 3 shows the photograph of these five antennas;they have the same structures as the simulated structures.The antenna’s performance is measured by the E5071CNetwork Analyzer and the Near-Field Antenna MeasurementSystem, Satimo. Figure 4 shows the measured input returnloss of these antennas. It can be seen that the antenna isresonant at 1.53 GHz when N = 1. If N is increased to 16, 32,64, and 128, the resonance frequency is reduced to 1.37 GHz,1.2 GHz, 0.92 GHz, and 0.61 GHz, respectively. Therefore,the reduction in resonant frequency is 10%, 22%, 40%,and 60%, respectively. The reduction in patch size is 19%,39%, 64%, and 84%. The impedance bandwidths (S11 ≤−10 dB) of these five antennas are 14%, 12%, 9%, 5.1%, and2.6%, respectively. In summary, the percentage size reductionfor the proposed antenna is increased with the number ofsegments used. More reduction in size is achieved at thecost of narrower bandwidth. This is expected because thebandwidth of an antenna is related to the electrical volumeof its radiation.

Figure 5(a) to Figure 5(d) show the measured radiationpatterns of the proposed antennas with N = 1 to 64 in planeφ = 0◦ and φ = 90◦. Because of the small ground plane,the radiation patterns for the antennas are bidirectional.Patterns for the antenna with N = 128 are not providedsince the lowest measurement frequency of our Near-FieldAntenna Measurement System (0.8 GHz) is higher than theresonant frequency (0.61 GHz) of this antenna. The gainat the resonance frequency 1.53 GHz is 1.2 dBi. When N isincreased to 16, 32, and 64, the gain is reduced to 0.4 dBi,−0.45 dBi, and −1.5 dBi. Therefore, increasing the numberof segments will decrease the antenna gain. This is also

0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2−30

−25

−20

−15

−10

−5

0

N = 128

N = 64

N = 32

N = 16 N = 1

Figure 2: Simulated input return loss of these antennas.

Figure 3: Photograph of these five antennas.

0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

×109

−40

−35

−30

−25

−20

−15

−10

−5

0

N = 128

N = 64

N = 32N = 16 N = 1

Figure 4: Measured input return loss of these antennas.

expected because the radiation efficiency of an antennais dropped with its electrical radiation area [10]. We cansee from Figure 5(b) to Figure 5(d) that copolarization andcross-polarization components are very close to each other.The reason is that the increasing number of vias leads tothe inconsistencies in the structure borders and makes morecross-polarization components which cannot be canceledout. This will be solved in future research.

4. Conclusion

A novel vertical meandering technique to reduce the lateralsize of a printed spiral antenna is proposed. Five printed


−40

−30

−20

−10

010

30

210

60

240

90270

120

300

150

330

180

0

Co-pol φ = 90

X-pol φ = 90

X-pol φ = 0

Co-pol φ = 0

(a)

−40

−30

−20

−10

0

1030

210

60

240 120

300

150

330

180

0

Co-pol φ = 0 Co-pol φ = 90

X-pol φ = 0

X-pol φ = 90

90270

(b)

−40

−30

−20

−10

010

30

210

60

240

90270

120

300

150

330

180

0

X-pol φ = 90

Co-pol φ = 90

Co-pol φ = 0

X-pol φ = 0

(c)

−40

−30

−20

−10

010

30

210

60

240

90270

120

300

150

330

180

0

Co-pol φ = 90

X-pol φ = 0

X-pol φ = 90

Co-pol φ = 0

(d)

Figure 5: (a) Measured radiation patterns with N = 1. (b) Measured radiation patterns with N = 16. (c) Measured radiation patterns withN = 32. (d) Measured radiation patterns with N = 64.

spiral antennas with a different number of segments areconstructed, tested, and analyzed. According to the resultsachieved, it is evident that the size of the antenna can besignificantly reduced by increasing its number of segments.

Acknowledgments

The authors wish to thank Dr. Ji Li and Weijun Hong fortheir helpful comments on the work presented in this paper.This work was supported in part by Shenzhen Science andTechnology Planning Project for the Establishment of KeyLaboratory in 2009 (CXB 200903090021A) and Hi-Tech

Research and Development Program of China (863 Program,no. 2006AA04A106).

References

[1] J. Rashed and C. T. Tai, “A new class of resonant antennas,”IEEE Transactions on Antennas and Propagation, vol. 39, no. 9,pp. 1428–1430, 1991.

[2] J. M. Kim and J. G. Yook, “Compact mender-type slotantennas,” in Proceedings of the IEEE Antennas and PropagationSociety International Symposium, vol. 2, pp. 724–727, Boston,Mass, USA, July 2001.


[3] S. Dey and R. Mittra, “Compact microstrip patch antenna,”Microwave and Optical Technology Letters, vol. 13, no. 1, pp.12–14, 1996.

[4] K. L. Wong, C. L. Tang, and H. T. Chen, “A compactmeandered circular microstrip antenna with a shorting pin,”Microwave and Optical Technology Letters, vol. 15, no. 3, pp.147–149, 1997.

[5] C. K. Wu, K. L. Wong, and W. S. Chen, “Slot-coupledmeandered microstrip antenna for compact dual-frequencyoperation,” Electronics Letters, vol. 34, no. 11, pp. 1047–1048,1998.

[6] J. H. Lu and K. L. Wong, “Slot-loaded, meandered rectangularmicrostrip antenna with compact dual-frequency operation,”Electronics Letters, vol. 34, no. 11, pp. 1048–1050, 1998.

[7] J. George, M. Deepukumar, C. K. Aanandan, P. Mohanan,and K. G. Nair, “New compact microstrip antenna,” ElectronicsLetters, vol. 32, no. 6, pp. 508–509, 1996.

[8] K. Noguchi, N. Yasui, M. Mizusawa, S. I. Betsudan, and T.Katagi, “Increasing the bandwidth of a two-strip meander-line antenna mounted on a conducting box,” in Proceedingsof the IEEE Antennas and Propagation Society InternationalSymposium, pp. 112–115, July 2001.

[9] H. Chaloupka, N. Klein, M. Peiniger, H. Piel, A. Pischke, andG. Splitt, “Miniaturized high-temperature superconductormicrostrip patch antenna,” IEEE Transactions on MicrowaveTheory and Techniques, vol. 39, pp. 1513–1521, 1991.

[10] L. J. Chu, “Physical limitations of omni-directional antennas,”Journal of Applied Physics, vol. 19, no. 12, pp. 1163–1175, 1948.


Research Article

Microstrip Patch Antenna Bandwidth Enhancement UsingAMC/EBG Structures

R. C. Hadarig, M. E. de Cos, and F. Las-Heras

Area de Teorıa de la Senal y Comunicaciones, Departamento de Ingenierıa Electrica, Universidad de Oviedo, Edificio Polivalente,Modulo 8, Campus Universitario de Gijon, Asturias, 33203 Gijon, Spain

Correspondence should be addressed to M. E. de Cos, [email protected]

Received 11 July 2011; Revised 19 September 2011; Accepted 22 September 2011

Academic Editor: Zhongxiang Q. Shen

Copyright © 2012 R. C. Hadarig et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A microstrip patch antenna with bandwidth enhancement by means of artificial magnetic conductor (AMC)/electromagneticband-gap structure (EGB) is presented. The electrical characteristics of the embedded structure are evaluated using MoMsimulations. The manufactured prototypes are characterized in terms of return loss, gain, and radiation pattern measurementsin an anechoic chamber.

1. Introduction

Microstrip patch antennas offer an attractive solution tocompact and ease-low-cost design of modern wireless com-munication systems due to their many advantages as lightweight and low volume, low profile, planar configurationwhich can be easily made conformal to host surface, low fab-rication cost, and the capability of obtaining dual and triplefrequency operations. When mounted on rigid surfaces mi-crostrip patch antennas are mechanically robust and can beeasily integrated with microwave integrated circuits (MICs).

However, microstrip patch antennas suffer from a num-ber of disadvantages as compared to conventional nonprint-ed antennas. Some of their major drawbacks are the narrowbandwidth, low gain, and surface wave excitation that re-duce radiation efficiency. To overcome one of their morecritical restrictions, narrow bandwidth, several techniquescan be used [1]. First of all, a thicker substrate with a low die-lectric constant or a ferrite composition provides a widerbandwidth but the first approach leads to no low-profiledesigns and increased in size, whereas the second solution isexpensive. Secondly, noncontacting feeding methods such asproximity/aperture coupled can be used to improve the im-pedance bandwidth, but this is difficult to fabricate. Anotherpossibility is multiresonator stack configuration with theinconvenient of resulting large thickness prototype [2, 3].

The surface waves can be minimized using electromagneticband-gap structures whereas for obtaining a high gain anten-na an array configuration for the patch elements is needed.

The research in the field of electromagnetic band-gapstructures has become attractive in the antenna communityand is considered to be a key technology for improving mi-crostrip patch antenna performances [4–6]. The use of meta-materials, such as the frequency selective surfaces (FSS) [7–9] is an alternative to face antennas and microwave circuitproblems and can provide either EBG or AMC behavior.In previous works, several narrow band antennas have beenmounted on EBG/AMC structures [10–18].

Depending on the intended application, the 2.4 GHzfrequency band can be used, for example, for wireless com-munications at 2.45 GHz or for RFID systems at 2.48 GHz.In this paper, the main goal is to improve the bandwidth andthe radiation properties of a microstrip patch antenna in the2.48 GHz band using two different approaches: combinationof the patch antenna with an EBG structure in the samelayer and combination of the patch antenna with an AMCstructure in two different layers. The aim of this work ischallenging because two resonant structures are involved andwhen integrated together their resonant behavior is mutuallyinfluenced. Firstly, the design of a microstrip patch antenna,henceforth referred as patch antenna, is shown followedby an adaptation of an AMC design recently presented by


35.44 mm

26.1

5 m

m(a)

Wp = 65.52 mm

Lp=

81.9

mm

(b) (c)

Figure 1: Manufactured prototypes: (a) Patch antenna, (b) Patch antenna-EBG, (c) Planar AMC.

the authors to operate at 2.48 GHz. Then, the patch anten-na is placed above the AMC. This combination will be hence-forth referred as Patch antenna-AMC. Secondly, the AMCstructure is modified to act as an EBG at a frequency closeto the patch antenna resonance frequency. Finally the EBG iscombined with the patch antenna on the same layer, resultingin a design with a uniplanar feature and reduced cost. Thiscombination will be henceforth referred as patch antenna-EBG. Return loss, gain, and radiation patterns of the threeprototypes (all having the same dimensions) are analyzedbased on measurements in an anechoic chamber.

2. Microstrip Antenna Design

The microstrip patch antenna is a narrow band design. Inthis work, the patch antenna suitable for RFID applicationsat 2.48 GHz is designed using ROGER3010 substrate witha thickness of 1.27 mm, relative dielectric permittivity εr =10.2, and loss tangent of 0.0023.The geometry of the patchantenna with its dimensions is shown in Figure 1(a). Thecharacteristic impedance of the transmission line is 50Ω.The antenna design has been carried out by a set of method-of-moments (MoM) simulations with commercial software[19]. From Figure 4, it can be extracted that the simulatedoperating bandwidth of the patch antenna is 20 MHz.

2.1. AMC Characterization. An adaptation of the AMCpreviously designed by the authors [20] is carried out shiftingthe resonant frequency to 2.48 GHz. Based on the Bloch-Floquet theory and on the finite element method (FEM), asingle cell of the lattice with periodic boundary conditions(PBCs) on its four sides is simulated in order to obtain thefrequency band where the periodic structure acts as an AMC.The phase of the reflection coefficient on the AMC sur-face is computed using a uniform incident plane wave (seeFigure 2(a)). Depending on the unit cell geometry togetherwith the substrate’s thickness and relative dielectric permit-tivity, the resonant frequency and the bandwidth of the struc-ture can be tuned. The unit cell dimensions are W ×W =16.93 × 16.93 mm2 and its geometry exhibits four sym-metry planes. The simulated reflection phase of normallyincident plane wave on the AMC surface versus frequency

is represented in Figure 2. The AMC resonant frequency is2.48 GHz, and the AMC operation bandwidth is approxi-mately 130 MHz (5.24%) (see Figure 2(b)). The structure ex-hibits several advantages such as uniplanar feature since nei-ther multilayer substrate no via holes are required, simpli-fying the implementation and reducing its costs.

2.2. EBG Characterization. The periodic structure can becharacterized as EBG using the suspended strip method [21,22] (See Figure 3). A suspended strip line over the 4 × 4 cellarrangement is used to test the transmission response of theelectromagnetic waves. The strip height is 0.02 λ. The struc-ture will block the transmission of power along the strip linefor frequencies within the band gap region and a notice-able reduction in S21 can be observed at a certain frequencyband. The band-gap of the EBG lattice is designed tobe adjacent to the frequency band of the patch antenna,so that when integrating the two structures on the samelayer, their resonances couple each other, and, as a result, awider bandwidth will be generated without disturbing othercharacteristics of the patch antenna such as the radiationpattern. The dimension of the unit cell in the case of EBGcharacterization is W2 = 16.38 mm. The simulated band gapof the EBG structure closer to the patch antenna bandwidthis 45 MHz around 2.5 GHz (see Figure 4).

3. Microstrip Patch Antenna Combinedwith AMC/EBG

3.1. Patch Antenna Placed above the AMC Structure. A 4 ×5 cells planar AMC structure is placed as patch antennaground plane [23] (see Figure 5) in order to analyze if theantenna’s bandwidth and the radiation properties can be im-proved. The antenna is fixed to the AMC structure (seeFigure 1(c)) by a 0.1 mm double-sided nonconducting ad-hesive tape. The microstrip patch antenna bandwidth is20 MHz whereas the AMC operation bandwidth is 130 MHz,having each one the same resonance frequency, 2.48 GHz,and the same dimensions. However, for combining the twostructures, the antenna’s ground plane has been removed andis placed above the AMC. As a consequence the antenna’s


h

PBC

PBC

Unit cell

Wave port

W

W

Conductorground plane

(a)

AMC BW

2.48 GHz SHF bandRO3010

2 2.2 2.4 2.6 2.8 3

Frequency (GHz)

180

135

90

45

0

−45

−90

−135

−180R

eflec

tion

ph

ase

(deg

)

(b)

Figure 2: AMC unit cell: (a) reflection phase simulation setup, (b) phase of the reflection coefficient on the AMC surface.

Metal(cooper)

Suspendedline

Figure 3: Schematic of suspended line above EBG surface (topview).

resonance frequency decreases due to capacitive effects forthose frequencies within the AMC bandwidth. A resonanceis obtained in the AMC bandwidth and outside this bandthe antenna behaves as if its substrate thickness had doubled.Merging both effects, the combined structure resonates in abandwidth wider than the microstrip patch antenna alone,but narrower than the AMC bandwidth. As disadvantage,the thickness of the combined structure is increased. If thedielectric substrate thickness of the patch antenna doubles,the resulting bandwidth (30 MHz) is narrower compared tothe bandwidth of patch antenna-EBG and patch antenna-AMC prototypes.

S11 patch antenna (MoM simulation)S21 EBG-suspended line structure (MoM simulation)

2.4 2.44 2.48 2.52 2.56 2.6 2.64

Frequency (GHz)

0

−5

−10

−15

−20

−25

−30

|S ij|

(dB

)

Figure 4: Resonances to be coupled in order to achieve bandwidthenhancement.

3.2. Patch Antenna Surrounded by the EBG Structure. Inorder to suppress the surface waves and to increase the band-width by means of coupled resonators effect, the EBG latticeis arranged around the patch, forming a uniplanar design[24]. As it has been already mentioned in Section 2.2, theresonance frequency of both structures (patch antenna andEBG structure) is mutually influenced, and depending on thefrequency difference between them and the unit-cell arrange-ment around the patch antenna, the resulting resonance fre-quency changes. The frequencies included on the patchantenna’s bandwidth are adjacent to the ones included onthe lower band gap. The selected EBG arrangement with


ROGER3010

ROGER3010

Figure 5: Patch antenna-AMC prototype layout.

Patch antenna (measurement)Patch antenna (MoM simulation)Patch antenna-EBG (measurement)Patch antenna-EBG (MoM simulation)

2.4 2.42 2.44 2.46 2.48 2.5 2.52 2.54

Frequency (GHz)

0

−5

−10

−15

−20

−25

−30

−35

−40

|S 11|(

dB)

Figure 6: Simulation and measurement comparison between theprototypes: patch antenna and patch antenna-EBG.

respect to the antenna is a tradeoff between performance andsize. The dimensions of the final structure (Figure 1(b)) areW p = 65.52 mm and Lp = 81.90 mm.

4. Results

Prototypes of the patch antenna, patch antenna placed abovethe AMC surface, and patch antenna surrounded by the EBGcells have been manufactured using laser micromachining.The return losses of each manufactured prototype have beenmeasured. As it can be observed in Figure 6 the measuredoperating bandwidth of the patch antenna is 23 MHz. Thedifference in bandwidth between simulations (20 MHz) andmeasurement (23 MHz) results could be due to the fact thatthe commercial MoM software considers infinite extension

|S 11|(

dB)

Patch antennaPatch antenna-EBGPatch antenna-AMC

2.35 2.4 2.42 2.45 2.5 2.55

Frequency (GHz)

0

−5

−10

−15

−20

−25

−30

−35AMC BW

Figure 7: Measurement comparison between the prototypes: patchantenna, patch antenna-EBG, and patch antenna-AMC.

for the dielectric substrate, or even more likely due to manu-facturing tolerances.

In the case of placing the antenna above the AMC surfacethe antenna resonance frequency is shifted downwards to2.43 GHz (see Figure 7) due to the capacitive effects that aregenerated between the two combined structures. Also, as theAMC structure has wider bandwidth than the patch anten-na, the resulting prototype bandwidth increases to 46 MHz,meaning a 100% broader bandwidth (see Figure 7).

When the patch antenna is surrounded by one row ofEBG cells the bandwidth increases 50% (see Figure 7) dueto the property of coupling the frequency bands of the twostructures composing the prototype. It is remarkable thatthis 50% bandwidth improvement is achieved increasingneither the prototype size nor the thickness. The percentagebandwidth comparison of the three prototypes is presentedin Table 1.

Measured radiation pattern cuts in the E and H planesof each manufactured prototype are presented in Figure 8.The patch antenna prototype exhibits copolarization-cross-polarization (CP-XP) ratio better than 25 dB (see Table 2),whereas for the patch antenna-EBG prototype the (CP-XP)ratio and the directivity are even increased. In measurementsthe gain of the patch antenna (4.59 dB) is preserved whenthe antenna is surrounded by one row of EBG cells (Table 1).From the simulation results, using EGB structures aroundthe patch antenna its radiation efficiency increases, due tosurface wave suppression property. However from measure-ment results it can be concluded that for this specific ar-rangement, the radiation efficiency is preserved (while im-proving bandwidth). The difference between simulations andmeasurements relies on the fact that the simulation methodimplemented by Momentum considers infinite dielectricunder the finite EBG metallization but also the differencecould be attributable to misalignments in the anechoicchamber. Radiation pattern properties of the patch antenna-AMC prototype show a (CP-XP) ratio inferior to the other


E plane. Normalized amplitude (dB)

90

60

30

0

330

300

270

240

120

150

180

210

−10 dB

−20 dB

−30 dB

0 dB

Patch antenna, CP f = 2.48 GHzPatch antenna, XP f = 2.48 GHzPatch antenna-EBG, CP f = 2.49 GHzPatch antenna-EBG, XP f = 2.49 GHzPatch antenna-AMC, CP f = 2.43 GHzPatch antenna-AMC, XP f = 2.43 GHz

(a)

H plane. Normalized amplitude (dB)

90

60

30

0

330

300

270

240

120

150

180

210

−10 dB

−20 dB

−30 dB

0 dB

Patch antenna, CP f = 2.48 GHzPatch antenna, XP f = 2.48 GHzPatch antenna-EBG, CP f = 2.49 GHzPatch antenna-EBG, XP f = 2.49 GHzPatch antenna-AMC, CP f = 2.43 GHzPatch antenna-AMC, XP f = 2.43 GHz

(b)

Figure 8: Measured radiation patterns of the prototypes: Patch antenna, patch antenna-EBG, and patch antenna-AMC.

Table 1: Comparison between the three designs.

Prototype Bandwidth (MHz) Directivity (dB) Gain (dB) Radiation efficiency (%)

Meas. Meas. Sim. Meas. Sim. Meas. Sim.

Patch antenna 23 (0.93%) 7.33 5.95 4.59 4.29 53.21 68.23

Patch antenna-EBG 34 (1.37%) 7.50 6.84 4.61 5.56 51.40 74.47

Patch antenna-AMC 46 (1.90%) 6.72 8.52 0 0.79 21.28 16.86

two prototypes and a gain close to 0 dB. As the AMC does nothave the ability to suppress the surface waves and the fact thatthe thicker the substrate, the stronger are the surface waves,the gain of the patch antenna-AMC prototype does notimprove. Also as the CP-XP ratio is worst for patch antenna-AMC than for the other prototypes, part of the energy couldbe radiated in other polarizations and backwards (as it isshown in Figure 8). In order to improve the gain, a gap be-tween the antenna and the AMC surface could be used butthis is technologically less advantageous. In addition, themicrostrip patch antenna’s gain and directivity can be in-creased when more rows or/and columns surround theprototype, so a trade-off between performance and size mustbe taken (the higher the number of unit cells in a periodic ar-rangement, the closer its behavior to an infinite EBG struc-ture).

5. Conclusions

Bandwidth enhancement of microstrip patch antenna in theRFID SHF 2.48 GHz band has been presented. Two differentstructures (AMC/EBG) have been combined with the same

Table 2: CP-XP ratio comparison.

PrototypeCP-XP ratio(E plane, dB)

CP-XP ratio(H plane, dB)

Patch antenna 25.81 25.04

Patch antenna-EBG 30.43 28.79

Patch antenna-AMC 13.85 9.43

microstrip patch antenna in order to characterize their jointperformance. The prototypes have been manufactured andcharacterized based on measurements in an anechoic cham-ber. From the measurements of the two resulting prototypes,patch antenna-AMC and patch antenna-EBG, it can beconcluded that both prototypes improve the bandwidth ofthe patch antenna. Due to the surface wave effect of the EBGstructure, the patch antenna-EBG prototype shows betterradiation properties increasing neither the prototype sizenor the thickness. All the prototypes presented are robust,compact and do not require via holes, being compatible withplanar fabrication technology.


Acknowledgments

This work has been supported by the Ministerio de Cienciae Innovacion of Spain/FEDER under Projects TEC2008-01638/TEC (INVEMTA), and CONSOLIDER-INGENIOCSD2008-00068 (TERASENSE), by the Gobierno del Prin-cipado de Asturias (PCTI)/FEDER-FSE under ProjectsEQUIP08-06, FC09-COF09-12, EQUIP10-31, and PC10-06(FLEXANT), by Grant BP10-039, and by Catedra Telefonica,Universidad de Oviedo.

References

[1] R. Garg, I. Bhartia, I. Bahl, and A. Ittipiboon, Microstrip An-tenna Design Handbook, Artech House, Boston, Mass, USA,2001.

[2] G. Kumar and K. C. Gupta, “Directly coupled multiple reso-nator wide-band microstrip antenna,” IEEE Transactions onAntennas and Propagation, vol. 33, no. 6, pp. 588–593, 1985.

[3] F. Yang, X. X. Zhang, X. Ye, and Y. Rahmat-Samii, “Wide-bandE-shaped patch antennas for wireless communications,” IEEETransactions on Antennas and Propagation, vol. 49, no. 7, pp.1094–1100, 2001.

[4] A. Pirhadi, F. Keshmiri, M. Hakkak, and M. Tayarani, “Analysisand design of dual band high directivity EBG resonator anten-na using square loop FSS AS superstrate layer,” Progress inElectromagnetics Research, vol. 70, pp. 1–20, 2007.

[5] E. Rajo-Iglesias, L. Inclan-Sanchez, and O. Quevedo-Teruel,“Back radiation reduction in patch antennas using planar softsurfaces,” Progress In Electromagnetics Research Letters, vol. 6,pp. 123–130, 2009.

[6] Z. Duan, S. Qu, and Y. Hou, “Electrically small antenna in-spired by spired split ring resonator,” Progress In Electromag-netics Research Letters, vol. 7, pp. 47–57, 2009.

[7] F. Yang and Y. Rahmat-Samii, Electromagnetic Band-GapStructures in Antenna Engineering, The Cambridge RF andMicrowave Engineering Series, Cambridge University Press,Cambridge, Mass, USA, 2008.

[8] M. E. De Cos, F. L. Heras, and M. Franco, “Design of planarartificial magnetic conductor ground plane using frequency-selective surfaces for frequencies below 1 GHz,” IEEE Antennasand Wireless Propagation Letters, vol. 8, pp. 951–954, 2009.

[9] O. Luukkonen, C. R. Simovski, and S. A. Tretyakov, “Ground-ed uniaxial material slabs as magnetic conductors,” Progress inElectromagnetics Research B, no. 15, pp. 267–283, 2009.

[10] H. Shaban, H. Elmikaty, and A. A. Shaalan, “Study the effectsof electromagnetic band-gap (EBG) substrate on two patchmicrostrip antenna,” Progress in Electromagnetics Research B,vol. 10, pp. 55–74, 2008.

[11] F. Yang and Y. Rahmat-Samii, “Reflection phase characteriza-tions of the EBG ground plane for low profile wire antenna,”IEEE Transactions on Antennas and Propagation, vol. 51, no.10, pp. 2691–2703, 2003.

[12] J. R. Sohn, K. Y. Kim, H. S. Tae, and J. H. Lee, “Comparativestudy on various artificial magnetic conductors for low-profileantenna,” Progress in Electromagnetics Research, vol. 61, pp. 27–37, 2006.

[13] S. Chaimool, K. L. Chung, and P. Akkaraekthalin, “Bandwidthand gain enhancement of microstrip patch antennas usingreflective metasurface,” IEICE Transactions on Communica-tions, vol. E93-B, no. 10, pp. 2496–2503, 2010.

[14] J. Liang and H. Y. D. Yang, “Radiation characteristics of amicrostrip patch over an electromagnetic bandgap surface,”

IEEE Transactions on Antennas and Propagation, vol. 55, no.6, pp. 1691–1697, 2007.

[15] G. Poilasne, “Antennas on High impedance ground planes: onthe importance of the antenna isolation,” Progress in Electro-magnetics Research, vol. 41, pp. 237–255, 2003.

[16] S. Zhu and R. Langley, “Dual-band wearable textile antennaon an EBG substrate,” IEEE Transactions on Antennas andPropagation, vol. 57, no. 4, pp. 926–935, 2009.

[17] M. Mantash, A. C. Tarot, S. Collardey, and K. Mahdjoubi,“Dual-band antenna for WLAN application with EBG,” inthe 4th International Congress on Advanced ElectromagneticMaterials in Microwaves and Optics, pp. 794–796, Karlsruhe,Germany, September 2010.

[18] M. Mantash, A. C. Tarot, S. Collardey, and K. Mahdjoubi,“Dual-band CPW-fed G-antenna using an EBG structure,” inAntennas and Propagation Conference (LAPC), pp. 453–456,Loughborough, UK, 2010.

[19] “ADS Momentum EM simulator tool,” http://www.home.agilent.com/agilent/product.jspx?cc=EG&lc=eng&ckey=1475688&nid=-34360.0.00&id=1475688.

[20] M. E. De Cos, Y. Alvarez, R. C. Hadarig, and F. Las-Heras,“Novel SHF-band uniplanar artificial magnetic conductor,”IEEE Antennas and Wireless Propagation Letters, vol. 9, pp. 44–47, 2010.

[21] L. Yang, M. Fan, F. Chen, J. She, and Z. Feng, “A novel com-pact electromagnetic-bandgap (EBG) structure and its ap-plications for microwave circuits,” IEEE Transactions on Micro-wave Theory and Techniques, vol. 53, no. 1, pp. 183–189, 2005.

[22] A. Aminian, F. Yang, and Y. Rahmat-Samii, “In-phase reflec-tion and EM wave suppression characteristics of electromag-netic band gap ground planes,” in IEEE International Antennasand Propagation Symposium and USNC/CNC/URSI NorthAmerican Radio Science Meeting, pp. 430–433, June 2003.

[23] R. C. Hadarig, M. E. De Cos, Y. ALvarez, and F. Las-Heras,“Novel bow-tie antenna on artificial magnetic conductorfor 5.8 GHz radio frequency identification tags usable withmetallic objects,” IET Microwaves, Antennas and Propagation,vol. 5, no. 9, pp. 1097–1102, 2011.

[24] M. E. De Cos, Y. Alvarez, and F. Las-Heras, “Enhancingpatch antenna bandwidth by means of uniplanar EBG-AMC,”Microwave and Optical Technology Letters, vol. 53, no. 6, pp.1372–1377, 2011.


Research Article

High-Performance Computational Electromagnetic MethodsApplied to the Design of Patch Antenna with EBG Structure

R. C. Hadarig, M. E. de Cos, and F. Las-Heras

Area de Teorıa de la Senal y Comunicaciones, Departamento de Ingenierıa Electrica, Universidad de Oviedo, Edificio Polivalente,Modulo 8, Campus Universitario de Gijon, 33203 Gijon, Spain

Correspondence should be addressed to M. E. de Cos, [email protected]

Received 14 June 2011; Revised 19 September 2011; Accepted 20 September 2011

Academic Editor: Shyh-Kang Jeng

Copyright © 2012 R. C. Hadarig et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

In this contribution High-Performance Computing electromagnetic methods are applied to the design of a patch antennacombined with EBG structure in order to obtain bandwidth enhancement. The electrical characteristics of the embedded structure(patch antenna surrounded by EBG unit cells) are evaluated by means of method of moment technique (MoM) whereas fordesigning the unit cell, the finite element method (FEM) together with the Bloch-Floquet theory is used. The manufacturedprototypes are characterized in terms of return loss and radiation pattern in an anechoic chamber.

1. Introduction

Microstrip patch antennas, which are used for both defenseand commercial applications, are replacing many conven-tional notprinted antennas. The most important propertiesof microstrip patch antennas are the lightweight, small-volume and the mass production at low cost whereasintrinsic disadvantages that limit their applications are lowgain, narrow bandwidth, and excitation of surface waves[1]. In order to design antennas with better efficiency andgain and lower backlobe and sidelobe levels, EBG structurescan be used [2–7]. In previous works [8–16] several narrowband antennas have been mounted on EBG structures. Inthis contribution, for obtaining bandwidth enhancementin the 2.48 GHz band, the band-gap of the EBG lattice isdesigned to be adjacent to the frequency band of the patchantenna, so that when integrating together both structurestheir resonances couple each other and as a result a widerbandwidth will be generated. The final design is uniplanarand in addition there is no need for via holes.

The aim of this work is challenging because two resonantstructures are involved and when integrated together theirresonant behavior is mutually influenced. For this reasonthe precision in the simulation results is a key point toachieve bandwidth enhancement. At this point is where high-performance computational electromagnetics play a funda-

mental roll in order to get a proper design in a reasonabletime.

2. Microstrip Antenna

The antenna design has been carried out using Method ofMoments (Momentum) electromagnetic simulator [17] andits geometry (Figure 1(a)) was optimized varying the valuesin a continuous range in order to obtain the frequencyband of interest (2.48 GHz) and the minimum bandwidthrequested (the process is an iterative one in which there willbe a tradeoff between the variation of antenna’s geometryand its effects on the performance). Using parameter sweepwith MoM in a 2 core Intel Xeon X5560 48 GB RAM(equivalent to 16 execution threads) server, there can beobtained the solution for 81 different values of the sweptparameter in just one minute. This is due to a mesh definitionso that 20 cells per wavelength at 3 GHz are taken whichleads to 249 rectangular cells and 171 triangular cells witha matrix size of 740. After applying mesh reduction a matrixsize of 610 results. A dielectric substrate, ROGER3010 having1.27 mm thickness, εr = 10.2 relative dielectric permittivity,and loss tangent of 0.0023, has been used. The measuredoperating bandwidth of the patch antenna is 23 MHz(Figure 4). The difference in bandwidth between the sim-ulated (20 MHz) and measured (23 MHz) results could be


(a) (b)

65.52 mm

35.44 mm

26.1

5 m

m

81.9

mm

Figure 1: Manufactured prototypes: (a) patch antenna (b) patchantenna-EBG.

due to the fact that the commercial MoM software considersinfinite extension for the dielectric substrate, or even morelikely due to manufacturing tolerances.

3. EBG Characterization

The unit cell of an EBG lattice consists of metal pads andsometimes narrow lines that implement a distributed LCcircuit having a resonant frequency [18]. In [19] the sameunit cell but in different frequency band is characterized fromthe point of view of an AMC, computing the phase of thereflection coefficient on its surface, from an uniform incidentplane wave. To search for the frequency band in whichthe periodic structure shows the AMC behavior, the finiteelement method (FEM) together with the Bloch-Floquettheory is used. A single cell of the lattice (with periodicboundary conditions (PBCs) on its four sides) is simulatedto model an infinite structure. The unit cell dimensions areW ×W = 16.38× 16.38 mm2. The simulation is carried outusing a server with 2 processors Intel Xeon E5620 and 64 GBRAM in a configuration equivalent to 2 execution threats and32 GB RAM. The air-box size is λ/2 at the lower frequencyconsidered in the simulation (which is 1 GHz). The solutionfrequency is 3 GHz and a frequency sweep is carried out from1 GHz to 3 GHz with a 0.01 GHz frequency step. The mesh isestablished to take at least 10 tetrahedra per wavelength atthe solution frequency (3 GHz). Mixed-order basis functionsand 30% lambda refinement are used. Maximum Delta S isfixed to 0.02 for the S-parameter calculations. All this leads toa mesh with 3642 tetrahedra (2889 for the air-box and 753 forthe substrate), a matrix size of 16101, and a computationaltime of 8 minutes and 44 seconds for the aforementionedfrequency sweep.

However, for the intended application, the structureshould show EBG behavior. Even though the finite elementmethod (FEM) uses specific boundary conditions such assimulating just half the volume under study (using PerfectMagnetic Conductor (PMC) boundary condition in one ofthe volume walls, the one that would cut the volume intwo identical parts, the size of the electromagnetic problem

Conductor ground plane

Unitcell

w

w

h

Metal(cooper)

Suspended

line

Figure 2: Schematic of suspended line above EBG surface (topview).

that needs to be simulated is reduced), for the currentscenario the method of moments (MoM) could yield afaster computational time (using MoM only the conductivestructure is meshed whereas using FEM, the whole volumearound the structure including the air box is meshed). Theperiodic structure can be characterized as EBG using thesuspended strip method [20, 21] and so the transmissioncoefficient (S21) of a suspended strip line over the periodicstructure is simulated (Figure 2). The structure will blockthe transmission of power along the strip line for frequencieswithin the band-gap region and a noticeable reduction of S21

can be observed at a certain frequency band. To accomplishthis simulation the mesh has been defined so that 20 cells perwavelength at 3 GHz are taken leading to 2085 rectangularcells and 5008 triangular cells with a matrix size of 9706. Afterapplying mesh reduction results a matrix size of 2352. Thesimulation time for 25 frequency steps is 50 minutes in a 2-core Intel Xeon X5560 48 GB RAM (equivalent to 16 threadsexecution) server.

The band-gap of the EBG structure (45 MHz) is adjacentto the bandwidth of the patch antenna (see Figure 3); thuswhen combining the two structures together, bandwidthenhancement is obtained without disturbing other charac-teristics of the patch antenna, such as the radiation pattern.

4. Patch Antenna with EBG Structure

Once the antenna and the EBG structures have beendesigned, the next step is the integration of both resonantstructures together in the same layer forming a uniplanardesign [22]. As it has been already mentioned in the intro-duction, the resonance frequency of both structures is mutu-ally influenced, and so depending on the frequency differencebetween them, and how the unit-cell arrangement aroundthe patch antenna is, the resulting resonance frequency willchange. The design of the patch antenna surrounded by


−5

−10

−15

−20

−30

−25

0

|S ij|

(dB

)

S11 patch antenna (MoM simulation)S21 EBG-suspended line structure (MoM simulation)

Frequency (GHz)

2.4 2.44 2.48 2.52 2.56 2.6 2.64

Figure 3: Resonances to be coupled in order to achieve bandwidth enhancement.

−5

−10

−15

−20

−35

−30

−25

0

Frequency (GHz)

2.42 2.44 2.46 2.48 2.5 2.52 2.54 2.56

|S 11|(

dB)

Patch antenna (MoM simulation)Patch antenna (measurement)

Patch antenna-EBG (measurement)Patch antenna-EBG (MoM simulation)

Figure 4: Simulation and measurement comparison between the patch antenna and the patch antenna-EBG prototypes.

one row EBG lattice has been carried out (Figure 1(b)). InFEM, due to the air-box size when small frequencies areinvolved (as in this case), the electric size of the problem tobe solved is rather big. A proper mesh should take at least10 (generally 20) tetrahedra per wavelength. Depending onthe prototype’s physical size, this could make the matrix ofthe linear equation system to become dense, which is notdesirable in FEM and generally leads to longer computationaltimes and increased memory requirements. However thematrix in MoM is dense, so this is not a problem, and thusthis could be a better choice in general. The disadvantageof MoM is related to dielectric managements as they areconsidered infinite sized. Using MoM the mesh has beendefined so that 20 cells per wavelength at 3 GHz are takenwhich leads to 2460 rectangular cells and 6246 triangular cellswith a matrix size of 12365. After applying mesh reductiona matrix size of 5832 results. The simulation time for 81

frequency steps is 63 minutes in a 2-core Intel Xeon X556048 GB RAM (equivalent to 16 threads execution) server.

A prototype of the Patch antenna-EBG has been manu-factured using laser micromachining. The return losses of themanufactured prototype have been measured and comparedto those of the microstrip patch antenna (Figure 4). Theprinciple of achieving bandwidth enhancement is basedon coupling the resonances of the structures involved. Asthe patch antenna resonates at adjacent frequency bandcompared to the band-gap of the EBG lattice, a signifi-cant bandwidth enhancement of the prototype combiningthe two structures (Patch antenna-EBG) is obtained. Asshown in Figure 4 the resulting bandwidth (34 MHz) ofthe Patch antenna-EBG is 50% wider than the microstrippatch antenna’s bandwidth (23 MHz). Radiation patternmeasurements of the prototypes have been carried outin anechoic chamber to complete their characterization.


90

60

0

120

150

180

210

240

270

300

330

0

30

E plane. Normalized amplitude (dB)

Patch antenna, CPPatch antenna, XP

Patch antenna-EBG, CPPatch antenna-EBG, XP

−10

−20

−30

(a)

H plane. Normalized amplitude (dB)90

60

0

120

150

180

210

240

270

300

330

0

30

Patch antenna, CPPatch antenna, XP

Patch antenna-EBG, CPPatch antenna-EBG, XP

−10

−20

−30

(b)

Figure 5: Patch antenna and patch antenna-EBG-measured radiation patterns.

Table 1: Comparison between the two prototypes.

Prototype Bandwidth (MHz) Directivity (dB) Gain (dB)

Patch antenna 23 7.33 4.59

Patch antenna-EBG 34 7.50 4.61

Radiation pattern cuts in the E and H planes of eachmanufactured prototype are plotted in Figure 5. Using anEBG structure to surround the patch antenna, the directivityincreases due to the surface wave suppression (Table 1).In the case of placing the patch antenna in a frequencyrange outside the band-gap of the EBG structure, MoMsimulations show that the unit cells have no influence in theradiation properties or in the bandwidth.

5. Conclusions

Bandwidth enhancement of microstrip patch antenna bymeans of EBG structure for RFID SHF 2.48 GHz bandhas been presented. Using High-Performance computingelectromagnetic methods the electrical characteristics of theresonant unit cell and the patch antenna have been evaluatedboth separated as well as combined in the same layer. Thesimulated and measured results are in good agreement dueto precision of the methods (MoM and FEM) used insimulations. Both frequency domain techniques, MoM andFEM, can be used once the AMC is designed using FEM withPBCs. This is just an example of MoM and FEM techniquesapplication to the design of antennas with metamaterials, but

there are also other possible approaches, time domain basedsuch as (Finite-difference time-domain) FDTD which couldalso be used.

There was reported a 50% increase of the initial band-width. The patch antenna-EBG prototype presented inthis paper has several advantages such as planar feature,compact size, and low dielectric losses. Neither via holesnor multilayer substrates are required, simplifying practicalimplementation and reducing its cost.

Acknowledgments

This work has been supported by the Ministerio de Cienciae Innovacion of Spain/FEDER under projects TEC2008-01638/TEC (INVEMTA) and CONSOLIDER-INGENIOCSD2008-00068 (TERASENSE), by the Gobierno del Prin-cipado de Asturias (PCTI)/FEDER-FSE under projectsEQUIP08-06, FC09-COF09-12, EQUIP10-31, and PC10-06(FLEXANT), by grant BP10-039, and by Catedra Telefonica-Universidad de Oviedo.

References

[1] R. Garg, I. Bhartia, I. Bahl, and A. Ittipiboon, MicrostripAntenna Design Handbook, Artech House, Boston, Mass, USA,2001.

[2] A. Pirhadi, F. Keshmiri, M. Hakkak, and M. Tayarani, “Analysisand design of dual band high directivity EBG resonatorantenna using square loop FSS AS superstrate layer,” Progressin Electromagnetics Research, vol. 70, pp. 1–20, 2007.


[3] E. Rajo-Iglesias, L. Inclan-Sanchez, and O. Quevedo-Teruel,“Back radiation reduction in patch antennas using planar softsurfaces,” Progress in Electromagnetics Research Letters, vol. 6,pp. 123–130, 2009.

[4] Z. Duan, S. Qu, and Y. Hou, “Electrically small antennainspired by spired split ring resonator,” Progress in Electromag-netics Research Letters, vol. 7, pp. 47–57, 2009.

[5] F. Yang and Y. Rahmat-Samii, Electromagnetic Band-GapStructures in Antenna Engineering, The Cambridge RF andMicrowave Engineering Series, Cambridge, University Press,2008.

[6] M. E. de Cos, F. Las-Heras, and M. Franco, “Design of planarartificial magnetic conductor ground plane using frequency-selective surfaces for frequencies below 1 GHz,” IEEE Antennasand Wireless Propagation Letters, vol. 8, pp. 951–954, 2009.

[7] O. Luukkonen, C. R. Simovski, and S. A. Tretyakov,“Grounded uniaxial material slabs as magnetic conductors,”Progress in Electromagnetics Research B, no. 15, pp. 267–283,2009.

[8] H. Shaban, H. Elmikaty, and A. Shaalan, “Study the effectsof electromagnetic band-gap (EBG) substrate on two patchmicrostrip antenna,” Progress in Electromagnetics Research B,vol. 10, pp. 55–74, 2008.

[9] A. P. Feresidis, G. Goussetis, S. Wang, and J. C. Vardaxoglou,“Artificial magnetic conductor surfaces and their applicationto low-profile high-gain planar antennas,” IEEE Transactionson Antennas and Propagation, vol. 53, no. 1, pp. 209–215, 2005.

[10] H. Mosallaei and K. Sarabandi, “Antenna miniaturizationand bandwidth enhancement using a reactive impedancesubstrate,” IEEE Transactions on Antennas and Propagation,vol. 52, no. 9, pp. 2403–2414, 2004.

[11] L. Akhoondzadeh-Asl, D. J. Kern, P. S. Hall, and D. H. Werner,“Wideband dipoles on electromagnetic bandgap groundplanes,” IEEE Transactions on Antennas and Propagation, vol.55, no. 9, pp. 2426–2434, 2007.

[12] J. Liang and H. Y. D. Yang, “Radiation characteristics of amicrostrip patch over an electromagnetic bandgap surface,”IEEE Transactions on Antennas and Propagation, vol. 55, no.6, pp. 1691–1697, 2007.

[13] F. Yang and Y. Rahmat-Samii, “Reflection phase characteriza-tions of the EBG ground plane for low profile wire antennaapplications,” IEEE Transactions on Antennas and Propagation,vol. 51, no. 10, pp. 2691–2703, 2003.

[14] J. R. Sohn, K. Y. Kim, H. S. Tae, and J. H. Lee, “Comparativestudy on various artificial magnetic conductors for low-profileantenna,” Progress in Electromagnetics Research, vol. 61, pp. 27–37, 2006.

[15] S. Chaimool, K. L. Chung, and P. Akkaraekthalin, “Bandwidthand gain enhancement of microstrip patch antennas usingreflective metasurface,” IEICE Transactions on Communica-tions, vol. E93-B, no. 10, pp. 2496–2503, 2010.

[16] D. Nashaat, H. A. Elsadek, E. A. Abdallah, M. F. Iskander,and H. M. Elhenawy, “Ultrawide bandwidth 2 × 2 microstrippatch array antenna using electromagnetic band-gap structure(EBG),” IEEE Transactions on Antennas and Propagation, vol.59, no. 5, pp. 1528–1534, 2011.

[17] “ADS Momentum EM simulator tool,” http://www.agilent.com/find/eesof.

[18] D. Sievenpiper, L. Zhang, R. F. J. Broas, N. G. Alexopolous, andE. Yablonovitch, “High-impedance electromagnetic surfaceswith a forbidden frequency band,” IEEE Transactions onMicrowave Theory and Techniques, vol. 47, no. 11, pp. 2059–2074, 1999.

[19] M. E. de Cos, Y. Alvarez, R. C. Hadarig, and F. Las-Heras,“Novel SHF-band uniplanar artificial magnetic conductor,”IEEE Antennas and Wireless Propagation Letters, vol. 9, pp. 44–47, 2010.

[20] A. Aminian, F. Yang, and Y. Rahmat-Samii, “In-phase reflec-tion and EM wave suppression characteristics of electromag-netic band gap ground planes,” in IEEE International Antennasand Propagation Symposium, vol. 4, pp. 430–433, June 2003.

[21] L. Yang, M. Fan, F. Chen, J. She, and Z. Feng, “A novelcompact electromagnetic-bandgap (EBG) structure and itsapplications for microwave circuits,” IEEE Transactions onMicrowave Theory and Techniques, vol. 53, no. 1, pp. 183–190,2005.

[22] M. E. de Cos, Y. Alvarez, and F. Las-Heras, “Enhancingpatch antenna bandwidth by means of uniplanar EBG-AMC,”Microwave and Optical Technology Letters, vol. 53, no. 6, pp.1372–1377, 2011.


Application Article

A Wideband High-Gain Dual-Polarized Slot Array Patch Antennafor WiMAX Applications in 5.8 GHz

Amir Reza Dastkhosh1 and Hamid Reza Dalili Oskouei2

1 Sahand University of Technology, P.O. Box 51335-1996, Tabriz, Iran2 University of Aeronautical Science & Technology (Shahid Sattari), P.O. Box 13846-63113, Tehran, Iran

Correspondence should be addressed to Amir Reza Dastkhosh, amir reza [email protected]

Received 9 April 2011; Revised 12 May 2011; Accepted 20 July 2011

Academic Editor: Dau-Chyrh Chang

Copyright © 2012 A. R. Dastkhosh and H. Dalili Oskouei. This is an open access article distributed under the Creative CommonsAttribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work isproperly cited.

A low-cost, easy-to-fabricate, wideband and high-gain dual-polarized array antenna employing an innovative microstrip slot patchantenna element is designed and fabricated. The design parameters of the antenna are optimized using commercial softwares(Microwave Office and Zeland IE3D) to get the suitable S-parameters and radiation patterns. Finally, the simulation results arecompared to the experimental ones and a good agreement is demonstrated. The antenna has an approximately bandwidth of 14%(5.15–5.9 GHz) which covers Worldwide Interoperability Microwave Access (WiMAX)/5.8. It also has the peak gain of 26 dBi forboth polarizations and high isolation between two ports over a wide bandwidth.

1. Introduction

Recently, microstrip patch antennas are one of the mostcommonly used antenna types due to many advantages suchas light weight, low fabrication costs, planar configuration,and capability to integrate with microwave integrated cir-cuits. Thus, the patch antennas are very suitable for vari-ous applications such as wireless communication systems,cellular phones, satellite communication systems, and radarsystems [1–6]. Due to their inherent features they are foundattractive for applications in broadband networks. WiMAXis a standard-based wireless technology for broadbandnetworks providing high data rate communication by usinglow-cost equipment. The access points in this network areusually built with large physical spacing. Therefore, thehigh-gain antenna is necessary to execute the long distancetransmission with a lower power. WiMAX has three allocatedfrequency bands called low band (2.5 GHz to 2.8 GHz),middle band (3.2 GHz to 3.8 GHz), and high band (5.2 GHzto 5.8 GHz). In this work, the low-cost microstrip slot arrayantenna (8 × 8) is designed, simulated, and fabricated foroperation in the frequency band of 5.15 GHz to 5.9 GHz.In each antenna element, two rectangular slots are used forcoupling the microstrip feed lines to the radiating patch.

This antenna has high isolation between the two ports overa wide bandwidth more than 14%. Furthermore, this high-gain (25.5 dBi) array antenna has dual polarization witha minimum half-power beamwidth (HPBW) (vertical: 7◦;horizontal: 6◦). The impedance characteristics, radiationpattern, return loss, and isolation between two ports for thedesigned array are analyzed, simulated, and optimized usingMicrowave Office and Zeland IE3D softwares. Also, S11, S21,and radiation pattern are measured and compared to thesimulated ones.

2. Configuration of Element Antenna

Microstrip patch antennas can be excited by different typesof feeds. In order to achieve the desired performances ofWiMAX antenna, an aperture coupled feed is used because ofits good characteristics such as wide operational bandwidthand shielding of the radiation patches. Moreover, an aperturecoupled feed yields better gain and radiation pattern fora dual-polarized antenna aimed for wireless applications[7–12]. An exploded view of the dual-polarized microstripantenna and a simplified equivalent circuit model for anaperture coupled microstrip antenna are shown in Figure 1.


Slots ingroundplane

Circular patch

Rohacell

Substrate

Hor. feed

Ver. feedWa

h f

Ls

(a)

1

1

Ls

Open stub

Ypatch

n1

n2

Yap

Zin

(b)

Ground

Circular patch Rohacell

DielectricSpacer

hh r

(c)

Figure 1: Configuration of the proposed dual-polarized aperture coupled circular patch antenna; (a) 3D view, (b) simplified equivalentcircuit model of an aperture coupled microstrip antenna, and (c) 2D view Rohacell: εr = 1.06, hr = 12 mm; substrate: εr f = 4.5, h f =0.762 mm, h = 5.9 mm; vertical and horizontal apertures’ dimensions or feed slot (La ×Wa): 14× 2 mm.

Metalplate

Rohacell(radome)

(a)

Horizontal feed(port 2)

Vertical feed(port 1)

(b)

λ/4

50Ω50Ω

50Ω70.7Ω

70.7Ω

(c)

λ0/4

λ0/4

ρ0

ρ1

ρ2 ρN

ρN

Z0 Z2Z1

RL

ZNΓIN

(d)

Figure 2: (a) 3D view of 8 × 8 array antenna with its ground plane. Rohacell (bottom: circular patches): εr = 1.06, hr = 12 mm; substrate(top: slots, bottom: feed lines): εr f = 4.5, h f = 0.762 mm. (b) Feed structure of array antenna. (c) Quarter-wave matching transformer.(d) N-section λ/4 transformer.


5 5.2 5.4 5.6 5.8 6−40

−35

−30

−25

−20

−15

−10

−5

Frequency (GHz)

(dB

)S

S11

S21

S12

S22

(a)

5 5.2 5.4 5.6 5.8 67

7.5

8

8.5

9

Gai

n(d

B)

Vertical

Frequency (GHz)

Horizontal

(b)

5 5.2 5.4 5.6 5.8 6−70

−60

−50

−40

−30

−20

Frequency (GHz)

(dB

)S

and (Measurement)S21S12

(Simulation)and S21S12

(c)

5 5.2 5.4 5.6 5.8 6−35

−30

−25

−20

−15

−10

−5

0

Frequency (GHz)

(dB

)S

Measurement ( )S11

Simulation ( )S11

Simulation ( )S22Measurement ( )S22

(d)

Figure 3: (a) Return loss and isolation versus frequency of one element of dual-polarized antenna element. (b) Gain versus frequency of oneelement of dual polarized antenna element. (c) Isolation. (d) Return loss versus frequency of 8× 8 array antenna.

Table 1: Wideband dual-polarized patch antenna specification.

Frequency range 5.15–5.9 GHz

Peak gain 26 dBi

Horizontal beamwidth 6◦

Vertical beamwidth 7◦

Front/back ratio Better than 28 dB

SLLVertical: −11 dB (center frequency)

Horizontal: −14 dB (center frequency)

Polarization (Dual) vertical or horizontal

VSWR 1.9 : 1 (max)

Impedance 50 Ohms

Mechanical Specification Length = width: 44 cm; depth: 4 cm

The antenna consists of only one substrate (Rogers TMM 4with dielectric constant εr = 4.5), an air layer for enhancing

the bandwidth, and a radome. The input impedance of theantenna at the center of the slot is given by [13, 14]

Zin = n22

n21Ypatch + Yap

− jZ0mcot(βmLs

), (1)

where Ypatch is the patch admittance and Yap is the apertureadmittance (Figure 1(b)). Z0m, βm, and Ls are the microstripline parameters in this equation. Also the coupling of thepatch to the microstrip line is described by a transformer[14]. The dimensions of the element antenna such as slots,feed lines, circular patch, and spaces between them areoptimized with the use of IE3D to achieve best radiationcharacteristics, wide impedance bandwidth, and high iso-lation between two ports. The optimized element antennahas a circular patch with 11.89 mm radius positioned atthe bottom side of Rohacell. Furthermore, two 50 ohmsmicrostrip feed lines (W = 1.5 mm, LV = 15 mm, and LH =23 mm) at the bottom side of the substrate (Rogers TMM 4


5 5.2 5.4 5.6 5.8 65

10

15

20

25

Frequency (GHz)

Fron

tto

back

rati

o(d

B)

VerticalHorizontal

(a)

5 5.2 5.4 5.6 5.8 6

Frequency (GHz)

28

29

30

31

32

33

Fron

tto

back

rati

o(d

B)

VerticalHorizontal

(b)

Figure 4: (a) Simulated front-to-back ratio versus frequency of 8 × 8 array antenna without plate at the back of antenna. (b) Measuredfront-to-back ratio versus frequency of 8× 8 array antenna with metal plate at the back of antenna.

Gai

n(d

B)

5 5.2 5.4 5.6 5.8 6

Frequency (GHz)

VerticalHorizontal

22

23

24

25

26

27

(a)

21

22

23

24

25

26

5 5.2 5.4 5.6 5.8 6

Frequency (GHz)

Gai

n(d

B)

VerticalHorizontal

(b)

Figure 5: Gain versus frequency of 8× 8 array antenna: (a) simulated and (b) measured.

with h f = 0.762 mm, εr f = 4.5) are electromagneticallycoupled to circular patch through two rectangular slotapertures in the common ground plane. As shown inFigure 2, in order to reduce the antenna back lobes, a metallicplate is located at the back of antenna, for example, 22 mmfrom the bottom of the antenna structure. Additionally,Figure 3(a) shows the simulated return loss for two ports(S11 and S22) versus frequency for one element antenna andFigure 3(b) shows simulated gain against frequency for oneelement. As depicted in Figure 3, in the desired bandwidth(5.1–5.9) return loss for both polarizations is more than15 dB and the isolation between two ports (S12) is better than35 dB.

3. Array Antenna

To obtain the desired radiation pattern characteristics, an8 × 8 planar microstrip slot array antenna is designed(Figure 2(a)). The bottom side of substrate consists of thefeeding network which is designed to give equal amplitudeand phase to each element (Figure 2(b)). Additionally,by using T-junction design and a quarter-wave matchingtransformer (Figure 2(c)), the feeds are matched to 50 ohmsfeed line [15, 16]. To provide a match, the transformercharacteristic impedance Z1 should be Z1 =

√RinZ0, where

Z0 is the characteristic impedance of the input transmissionline and Rin is the input impedance of the antenna. The


−60

−50

−40

−30

−20

−10

0N

orm

aliz

edga

in(d

B)

planeplane

−50 0 50 90−90

θHE

(a)

−60

−50

−40

−30

−20

−10

0

Nor

mal

ized

gain

(dB

)

−50 0 50 90−90

θplaneH

planeE

(b)

Figure 6: Simulated antenna far-field radiation pattern at 5.5 GHz: (a) vertical and (b) horizontal.

−80 20 40 60 80−60 −40 −20−60

−50

−40

−30

−20

−10

0

Nor

mal

ized

gain

(dB

)

planeplane

0 90−90

θHE

(a)

−60

−50

−40

−30

−20

−10

0

Nor

mal

ized

gain

(dB

)

planeplane

−50 0 50 90−90

θHE

(b)

Figure 7: Measured antenna far-field radiation pattern at 5.5 GHz: (a) vertical and (b) horizontal.

transformer is usually another transmission line with thedesired characteristic impedance (Figure 2(d)). The spacesbetween elements are set at 50 mm for better radiationcharacteristics. The simulated and measured return loss (S11)and isolation (S21) of 8 × 8 dual-polarized microstrip patchslot array antenna are illustrated in Figures 3(c) and 3(d).Furthermore, the metal plate at the back of array antennareduces the front-to-back ratio about −20 dB, as can beseen in Figure 4. Likewise, the gain of the array antenna indifferent frequencies is demonstrated in Figure 5. Moreover,the simulated and measured E andH plane far-field radiationpatterns of the array antenna at center frequency are shownin Figures 6 and 7. Finally, all vital parameters such as

antenna size, its gain, beamwidth, side lobe level, and front-to-back ratio are summarized in Table 1.

4. Conclusions

This paper has reported the design of a low-cost high-gaindual-polarized patch array antenna for WiMAX applicationsin the 5.15–5.9 GHz frequency band. The antenna has anapproximately bandwidth of 14% and the peak gain of 26 dBifor both polarizations. The design has been achieved withthe use of commercial software packages AWR MicrowaveOffice and Zeland IE3D. The design process aimed at bestreturn losses and fine quality radiation characteristics over


the assumed frequency band. The designed antenna has animpedance bandwidth of approximately 14% and the peakgain of approximately 26 dBi for both polarizations. Thisperformance has been confirmed experimentally.

References

[1] C. A. Balanis, Antenna Theory, Analysis and Design, John Wiley& Sons, 3rd edition, 2005.

[2] K. L. Wong, Planar Antennas for Wireless Communications,John Wiley & Sons, 2003.

[3] S. B. Chen, Y. C. Jiao, W. Wang, and F. S. Zhang, “Modified T-shaped planar monopole antennas for multiband operation,”IEEE Transactions on Microwave Theory and Techniques, vol.54, no. 8, pp. 3267–3270, 2006.

[4] W. C. Liu and C. F. Hsu, “Dual-band CPW-fed Y-shapedmonopole antenna for PCS/WLAN application,” ElectronicsLetters, vol. 41, no. 7, pp. 390–391, 2005.

[5] W. C. Liu, “Broadband dual-frequency meandered CPW-fedmonopole antenna,” Electronics Letters, vol. 40, no. 21, pp.1319–1320, 2004.

[6] J. Y. Li, J. L. Guo, Y. B. Gan, and Q. Z. Liu, “The tri-band performance of sleeve dipole antenna,” Journal ofElectromagnetic Waves and Applications, vol. 19, no. 15, pp.2081–2092, 2005.

[7] K. M. Z. Shams, M. Ali, and H. S. Hwang, “A planar induc-tively coupled bow-tie slot antenna for WLAN application,”Journal of Electromagnetic Waves and Applications, vol. 20, no.7, pp. 861–871, 2006.

[8] C. Y. Wu, S. H. Yeh, and T. H. Lu, “Novel high gainmetamaterial antenna radome for WiMAX operation in the5.8-GHz band,” in IEEE Antennas and Propagation SocietyInternational Symposium (AP-S ’07), pp. 3488–3491, June2007.

[9] L. N. Zhangl, S. S. Zhongl, and X. L. Liang, “Dual-band dual-polarized hybrid antenna array,” in PIERS Proceedings, Xi’an,China, March 2010.

[10] S. D. Targonski, R. B. Waterhouse, and D. M. Pozar, “Design ofwide-band aperture-stacked patch microstrip antennas,” IEEETransactions on Antennas and Propagation, vol. 46, no. 9, pp.1245–1251, 1998.

[11] D. M. Pozar and S. D. Targonski, “A shared-aperture dual-band dual-polarized microstrip array,” IEEE Transactions onAntennas and Propagation, vol. 49, no. 2, pp. 150–157, 2001.

[12] H. J. Song and M. E. Bialkowski, “Ku-band 16 × 16 planararray with aperture-coupled microstrip-patch elements,” IEEEAntennas and Propagation Magazine, vol. 40, no. 5, pp. 25–29,1998.

[13] G. Vetharatnam, C. B. Kuan, and C. H. Teik, “Combinedfeed network for a shared-aperture dual-band dual-polarizedarray,” IEEE Antennas and Wireless Propagation Letters, vol. 4,no. 1, pp. 297–299, 2005.

[14] R. Garg and P. Bhartia, Microstrip Antenna Design Handbook,Artech House, 2001.

[15] B. N. Das and K. K. Joshi, “Impedance of a radiating slot inthe ground plane of a microstripline,” IEEE Transactions onAntennas and Propagation, vol. 30, no. 5, pp. 922–926, 1982.

[16] J. R. James and P. S. Hall, Handbook of Microstrip Antennas,IEE Electromagnetic Waves, Peter Peregrinus, 1989.

international journal of antennas and...

Documents