ieee fractal antenna

3466 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 60, NO. 7, JULY 2012

CommunicationsCircularly Polarized Multiband Microstrip Antenna

Using the Square and Giuseppe Peano Fractals

Homayoon Oraizi and Shahram Hedayati

AbstractBy computer simulation and actual fabrication, it is demon-strated that multiband operation with circular polarization of radiationmay be achieved by the combination of square and Giuseppe Peano fractalgeometries realized on a two layer microstrip antenna. The antenna feedis designed by an electromagnetic coupling system. The proposed antennaconfiguration also achieves some degree of miniaturization, which makesit suitable for wireless applications. The antenna characteristics, such asreturn loss, axial ratio and radiation patterns achieved by the proposedstructure attest to its effectiveness as a mobile radiator.

Index TermsCircular polarization, fractal antenna, microstrip,miniaturization.

I. INTRODUCTION

Mandelbrot first introduced the fractal geometry in 1975 [1], inwhich each sub-section has the characteristics of the whole structurein a smaller scale. This is the basic property of self-similarity. Fractalgeometries have been applied in various science and technologies,such as antennas and radiators. Generally, the utilization of fractalgeometries in antennas tends to reduce their physical sizes and producemultiband response in their radiation characteristics. Since fractalstructures have a repetitive geometry, they can generate long paths in alimited volume. Accordingly, we may refer to fractal geometries, suchas the Koch, Minkowski, Hilbert and tree fractals [2], [5], which havebeen used for dipole and ring antennas.

The property of self-similarity of fractal geometries is usedto achieve multiband operations from fractal antennas and theirspace-filling property is used for the antenna miniaturization. [3], [4],[6]. Fractal geometries are used in radiating systems and even mi-crowave devices to benefit from their interesting properties [6]. Sincethe generation of fractal configurations have an iterative procedure,then they can achieve long linear dimensions and high surface areasin a limited volume [5].

In this communication, a multiband antenna is introduced using thenovel square and Giuseppe Peano fractals. It is designed for operationin the following bands: Global positioning system L1 (GPS 1.575GHz); Hiper-Lan2 (High Performance Radio Local Area NetworkType2) in the band 2.122.32 GHz; IEEE802.11b/g in the bandfrom 2.4 to 2.484 GHz, which is one of the WLAN bands and IMTadvanced system or forth generation (4G) mobile communicationsystem in the band 4.65.2 GHz. We investigate the miniaturizationand multi-banding [6] properties of the square fractal microstrip patchantenna. We also study the radiation properties of the combination

Manuscript received December 22, 2010; revised December 27, 2011; ac-cepted January 24, 2012. Date of publication April 30, 2012; date of currentversion July 02, 2012.

The authors are with the Electrical Engineering Department, Iran Universityof Science and Technolog, Tehran 1684613114, Iran (e-mail: [email protected]).

Color versions of one or more of the figures in this communication are avail-able online at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TAP.2012.2196912

Fig. 1. Configuration of the square fractal geometry.

Fig. 2. The generator of square fractal geometry.

Fig. 3. Reflection coefficient of the initiator and first iteration square fractal.

of square and Peano fractals for the microstrip patch antenna withelectromagnetically coupled feed systems. A prototype sample of theproposed fractal antenna is fabricated and measured. The miniatur-ization, multibanding and circular polarization of the proposed fractalantenna is verified by the simulation results and measurement data.

II. COMBINATION OF THE SQUARE AND GIUSEPPE PEANO FRACTALSConsider the square fractal geometry in Figs. 1 and 2, where the ini-

tiator, first and second iterations are shown. We compare the radiationproperties of the initiator and first iteration [5], where the parametersare selected as mm and

for the resonance fre-quency of about 2.45 GHz. We select the substrate FR4 with dielectricconstant

, height mm and loss tangent .The reflection coefficient of the initiator square fractal and the first

iteration fractal are drawn in Fig. 3. Observe that although the sizeof the two squares are identical, but the resonance frequency of thefirst iteration is less than that of the initiator. The reason for loweringof resonance frequency with the reduction of parameter

is due toincrease of the length of current path on the patch (L), as depicted inFig. 5. Note that the th iteration fractal has n separated regions, whichresonate independently (ignoring the mutual coupling among them),and produce fundamental resonance frequencies. For example,the first iteration fractal has two resonance frequencies, due to the inner

0018-926X/$31.00 2012 IEEE

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 60, NO. 7, JULY 2012 3467

Fig. 4. Reflection coefficient versus frequency for the first iteration squarefractal.

Fig. 5. Surface currents on the patch for three distinct modes, (a) 1st mode( GHz), (b) Spurious mode ( GHz), (c) 2nd mode ( GHz).

Fig. 6. Initiator and generator of the Giuseppe Peano fractal.

square and outer square rings. The surface current on the patch for threedistinct modes are drawn in Fig. 5. Its return loss versus frequency isdrawn in Fig. 4, where three resonance modes are shown. The middleresonance frequency (4.1 GHz) is due to the second resonance of theouter square ring.

Observing the current distributions in Fig. 5, Note that (1) to (6) arederived based on the observations of the surface current distribution onthe patch. This type of reasoning is also followed for the evaluation ofthe patch resonance frequency by measuring the length of current pathbetween its nulls [2]. The resonance frequencies of the first iterationfractal of the square patch may be obtained by the following empiricalrelations:

(1)

(2)

where

and

are the average lengths of the current paths for thefirst and second resonance modes, which may be determined as:

(3)

(4)

These relations may be used for the design of antennas.Consider the initiator and generator of the Peano fractal as shown in

Fig. 6. Application of such a fractal generation to the edges of squarepatch up to the second iteration is drawn in Fig. 7. In this section we in-vestigate the possibilities and properties of the application of Giuseppe

Fig. 7. Implementation of the Peano fractal to the edges of square patch up tothe second iteration.

Fig. 8. Comparison of reflection coefficient of the Giuseppe fractal with othercommon fractals.

Peano fractal geometry for the miniaturization of microstrip patch an-tennas and compare its performance with those of the usual fractals,such as Koch, Minkowski, Sierpinski and Tee-Type. The length of theGiuseppe Peano fractal patch perimeter increases, while its surface arearemains constant without any more space occupation. Consequentlythe antenna miniaturization, maintenance of its gain and increase ofits relative frequency bandwidth are achieved. The frequency responseof S11 for several fractal geometries, such as Koch, Minkowski, Teetype and Giuseppe Peano (with their specified dimensions) are drawnin Fig. 8 for comparison. Observe that proposed Giuseppe Peano fractalgeometry for the microstrip antenna produces comparatively a larger10 dB return loss bandwidth with lower number of iterations, and alsoachieves better antenna miniaturization.

3468 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 60, NO. 7, JULY 2012

Fig. 9. The proposed combination of square and Giuseppe Peano fractal withthe electromagnetic coupling feed.

Fig. 10. A photograph of the fabricated fractal antenna.

III. ANTENNA DESIGN

Novel fractal proposed antenna is shown in Fig. 9. It consists oftwo layers. The lower substrate is Rogers RT/Duroid 5880 (with

mm and

) and the upper substrateis FR4 (with

mm and ). The feedingsystem is by electromagnetic coupling through a microstrip line on thelower substrate and the fractal patch is placed on the upper one. A pho-tograph of the fabricated fractal antenna is shown in Fig. 10.

Fig. 11. Reflection coefficient at the antenna feed point as S11 at three fre-quency bands (a) 1.5 GHz; (b) 2.5 GHZ; (c) 4.9 GHz.

The average lengths of current paths for the first and second reso-nance modes L1 and L2 are derived experimentally:

(5)

(6)

IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 60, NO. 7, JULY 2012 3469

Fig. 12. Axial ratio of the antenna at three frequency bands (a) 1.5 GHz;(b) 2.5 GHZ; (c) 4.9 GHz.

which are used for the antenna design. For the generation of circularpolarization, a perturbation of electrical length is produced on thetwo perpendicular edges of the square patch, which are in the form ofGiuseppe Peano fractals. The aim is to excite two orthogonal modeswith a phase difference of 90. The perturbations on the lengths offractal edges on the outer square, namely S1, S2 and L1 and thoseon the inner square, namely

and

, are made

Fig. 13. Measurement of radiation patterns in the E- and H-planes at three fre-quency bands (a) 1.5 GHz; (b) 2.5 GHZ; (c) 4.9 GHz.

for the generation of circular polarization in the first and also secondand third bands, respectively. These parameters are optimized forthe achievement of axial ratio dB. The other parameters ofstructure are optimized for the desired impedance matching.

Now the effects of variation of main parameters of the antenna struc-ture, such as L1 and a4 are investigated. They should be modified forthe increase of impedance bandwidth for operation in the bands, suchas Hiper-Lan2 and IEEE802.11b/g. For this purpose, the primary an-tenna structure parameters are selected and a parametric study is con-ducted about the optimum values of

mm and

mm. the

3470 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 60, NO. 7, JULY 2012

Fig. 14. The gain of the fractal antenna versus frequency in all applicationbands.

TABLE ICOMPARE GIUSEPPE PEANO PERFORMANCE WITH USUAL FRACTALS

achieved responses of proposed antenna as reflection coefficient versusfrequency are drawn in Fig. 11.

The first and second resonance frequencies determine the length ofouter and inner square sides (

and

), respectively. The optimizedvalues of parameters are given below:

mm,

mm,

mm,

mm,

mm,

mm,

mm. The size of the innerfractal teeth are one third (0.33) that of the outer one. The widthand length of the feed line are 3.4 mm and 38 mm, respectively.A prototype model of the proposed fractal antenna is fabricatedand measured. The simulation results and measurement data arecompared in the following figures (Fig. 11). The reflection coefficient(as S11) at the antenna feed point across the individual S11 forthe three distinct bands are shown in Fig. 11. Note that we haveused substrates Rogers RT/Duroid 5880 and FR4 in the two layers.Substrate FR4 was used for more antenna miniaturization becauseof its higher dielectric constant

. But it has higherlosses, especially at high frequencies. This may accounted for thediscrepancy between the simulation results and experimental data.Observe that the resonance frequency of the first iteration fractalantenna is actually 200 MHz lower than that of the correspondingsimple square patch. Consequently, it is shown that some antennaminiaturization is achievable by the proposed fractal antenna. Thebandwidth at the first resonance frequency (1.5 GHz) is 40 MHz,that at the second one (2.5 GHz) is 900 MHz and that at the thirdone (4.9 GHz) is 310 MHz. The circular polarization of radiationpattern is obtained by different lengths of teeth on the perpendicularsides of the square fractal (namely S1 and S2 in Fig. 8), whichproduce two orthogonal modes with 90 phase difference. The axialratio of the antenna is drawn in Fig. 12. The bandwidth of circularpolarizations at the first, second and third bands are 30, 40 and50 MHz, respectively.

The measurement of radiation patterns in the E- and H-planes forthe first, second and third bands are drawn in Fig. 13. The gain of thefractal antenna versus frequency across the operating bands is drawnin Fig. 14, which is quite good.

IV. CONCLUSION

In this communication, a microstrip antenna is proposed as acombination of square and Giuseppe Peano fractals, which mayproduce three distinct frequency bands of operation with circularpolarization. The antenna achieves some degree of miniaturization.The measured data and simulation results of the fabricated antennafor the return loss, axial ratio and radiation patterns attest to the ef-fectiveness and suitability of the proposed fractal antenna for wirelessapplications. Observe that proposed Giuseppe Peano fractal geometryfor the microstrip antenna produces comparatively a larger 10 dBbandwidth with lower number of iterations, and also achieves betterantenna miniaturization.

REFERENCES[1] D. H. Werner and R. Mittra, Frontiers in Electromagnetics. Piscat-

away, NJ: IEEE Press, 2000, pp. 4881.[2] C. P. Baliarda, J. Romeu, and A. Cardama, The Koch monopole: A

small fractal antenna, IEEE Trans. Antennas Propag., vol. 48, no. 11,Nov. 2000.

[3] C. Puente-Baliaada, J. Romeu, R. Pous, and A. Cardama, On the be-havior of the Sierpinski multiband fractal antenna, IEEE Trans. An-tennas Propag., vol. 46, no. 4, Apr. 1998.

[4] D. H. Werner and S. Ganguly, An overview of fractal antenna engi-neering research, IEEE Antennas Propag. Mag., vol. 45, no. 1, Feb.2003.

[5] J. P. Gianvittori and Y. Rahmat-Samii, Fractal antenna: A novelantenna miniaturization technique, and applications, IEEE AntennasPropag. Mag., vol. 44, no. 1, Feb. 2002.

[6] M. Naghshvarian-Jahromi and N. Komjani, Novel fractal monopolewideband antenna, . Electromagn. Waves Applicat., vol. 22, no. 23,pp. 195205, 2008.