<<

International Journal of Antennas and Propagation

Metamaterials

Guest Editors: Alejandro Lucas Borja, James R. Kelly, Fuli Zhang, and Eric Lheurette International Journal of Antennas and Propagation

Metamaterials

Guest Editors: Alejandro Lucas Borja, James R. Kelly, Fuli Zhang, and Eric Lheurette Copyright © 2013 Hindawi Publishing Corporation. All rights reserved.

This is a special issue published in “International Journal of Antennas and Propagation.” All articles are open access articles distributed under 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, USA Se-Yun Kim, Republic of Korea S. M. Rao, USA C. Bunting, USA Ahmed A. Kishk, Canada S. R. Rengarajan, USA F. Catedra,´ Spain T. Kundu, USA Ahmad Safaai-Jazi, USA Dau-Chyrh Chang, Taiwan Ju-Hong Lee, Taiwan S. Safavi-Naeini, Canada Deb Chatterjee, USA B. Lee, Republic of Korea M. Salazar-Palma, Spain Z. N. Chen, Singapore L. Li, Singapore S. Selleri, Italy M. Y. W. Chia, Singapore Y. Lu, Singapore K. T. Selvan, India C. Christodoulou, USA Atsushi Mase, Japan Z. Q. Shen, Singapore Shyh-Jong Chung, Taiwan Andrea Massa, Italy John J. Shynk, USA L. Crocco, Italy G. Mazzarella, Italy M. S. J. Singh, Malaysia T. A. Denidni, Canada Derek McNamara, Canada Seong-Youp Suh, USA A. R. Djordjevic, Serbia C. Mecklenbrauker,¨ Austria P. Wahid, USA K. P. Esselle, Australia M. Midrio, Italy Y. Ethan Wang, USA F. Falcone, Spain Mark Mirotznik, USA D. S. Weile, USA Miguel Ferrando, Spain A. S. Mohan, Australia Quan Xue, Hong Kong V. Galdi, Italy P. Mohanan, India Tat Soon Yeo, Singapore Wei Hong, China Pavel Nikitin, USA Y. J. Yoon, Korea Hon Tat Hui, Singapore A. D. Panagopoulos, Greece W. Yu, USA TamerS.Ibrahim,USA M. Pastorino, Italy Jong Won Yu, Republic of Korea Nemai Karmakar, Australia M. Pieraccini, Italy Contents

Metamaterials, Alejandro Lucas Borja, James R. Kelly, Fuli Zhang, and Eric Lheurette Volume 2013, Article ID 516939, 2 pages

Novel Design of Electromagnetic Bandgap Using Fractal Geometry, Huynh Nguyen Bao Phuong, Dao Ngoc Chien, and Tran Minh Tuan Volume 2013, Article ID 162396, 8 pages

Metamaterial , Jing Jing Yang, Ming Huang, Hao Tang, Jia Zeng, and Ling Dong Volume 2013, Article ID 637270, 16 pages

EBG Size Reduction for Low Permittivity Substrates, Gonzalo Exposito-Dom´ ´ınguez, JoseManuelFern´ andez-Gonz´ alez,´ Pablo Padilla, and Manuel Sierra-Castaner˜ Volume 2012, Article ID 106296, 8 pages

Mie Scattering by a Conducting Sphere Coated Uniaxial Single-Negative Medium, You-Lin Geng Volume 2012, Article ID 856476, 6 pages

Metamaterial CRLH Antennas on Silicon Substrate for Millimeter-Wave Integrated Circuits, Gheorghe Ioan Sajin and Iulia Andreea Mocanu Volume 2012, Article ID 593498, 9 pages

Broadband Microstrip Bandpass Filter Based on Open Complementary Split Ring Resonators,P.Velez,´ J. Naqui, M. Duran-Sindreu,´ J. Bonache, and F. Mart´ın Volume 2012, Article ID 174023, 6 pages

A Method for Extending the Bandwidth of Metamaterial Absorber, Hong-Min Lee and Hyung-Sup Lee Volume 2012, Article ID 859429, 7 pages

Broadband Equivalent Circuit Model for a Coplanar Waveguide Line Loaded with Split Ring Resonators, Victor Sanz, Angel Belenguer, Alejandro L. Borja, Joaquin Cascon, Hector Esteban, and Vicente E. Boria Volume 2012, Article ID 613518, 6 pages

FDTD-SPICE for Characterizing Metamaterials Integrated with Electronic Circuits, Zhengwei Hao, Soheil Saadat, and Hossein Mosallaei Volume 2012, Article ID 282159, 7 pages

Optical and Electrical Properties of Magnetron Sputtering Deposited CuAlO Thin Films, Yongjian Zhang, Zhengtang Liu, Duyang Zang, Liping Feng, Xingsen Che, and Yanyan Li Volume 2012, Article ID 823089, 7 pages Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2013, Article ID 516939, 2 pages http://dx.doi.org/10.1155/2013/516939

Editorial Metamaterials

Alejandro Lucas Borja,1 James R. Kelly,2 Fuli Zhang,3 and Eric Lheurette4

1 Departamento de Ingenier´ıa Electrica,´ Electronica,´ Automatica´ y Comuicaciones, Universidad de Castilla-La Mancha, 16071 Cuenca, Spain 2 Facility of Science and Technology, Anglia Ruskin University, East Road, Cambridge, Cambridgeshire CB1 1PT, UK 3 Department of Applied Physics, School of Science, Northwestern Polytechnical University, Xi’an 710072, China 4 Institut d’Electronique, Microelectronique´ et Nanotechnologie, UniversitedeLille1,avenuePoincar´ eBP69,´ 59652 Villeneuve d’Ascq Cedex, France

Correspondence should be addressed to Alejandro Lucas Borja; [email protected]

Received 14 March 2013; Accepted 14 March 2013

Copyright © 2013 Alejandro Lucas Borja 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 recent years, the possibility of taking advantage of the the absorption area of a metasurface based on geometrical unusual properties of the so-called metamaterial technology variations of a unit cell. has led to a great deal of research activity. These structures Z. Hao et al. report on a modeling method to characterize can be engineered to realize novel electromagnetic properties metamaterial structures integrated with active and tunable and to achieve behaviors that are not found in naturally components based on both FDTD and SPICE methods. The occurring materials. For instance, many efforts have been authors remark along the paper the advantages of their hybrid aimedatimprovingtheperformancesofnovelmetamaterial method versus the conventional FDTD method. arrangements in terms of equivalent circuit descriptions, Y.Zhang et al. explain the optical and electrical properties metamaterial modeling, functionality, miniaturization and of different materials via chemical reaction, which has the reconfigurability. potential to metamaterial preparation, such as Mie resonance The principal goal of this special issue on metamaterials type structures. is to provide an in-depth description of the state of the art of Y.-l. Geng proposes an analytical method for computing research and development in this area. In particular, it high- the electromagnetic scattering from a three-dimensional lights contributions intended to stimulate new metamaterial- (3D) conducting sphere coated with a uniaxial anisotropic inspired designs and practical applications. single-negative (SNG) medium. This is achieved by using P. V elez´ et al. successfully present the design of broadband a spherical vector waves. Numerical results were obtained microstrip bandpass filters based on the use of a novel using this technique. Those results have been compared with compact metamaterial-concept based cell, the so-called open Mie theory, and the two were found to be in good agreement. complementary split ring resonator (OCSRR). After present- G. I. Sajin and I. A. Mocanu present two novel antennas. ing the new proposed resonator, a design procedure of two These antennas are based on zeroth-order resonant (ZOR) different bandpass filters is shown, which are then manufac- millimeter wave composite right/left-handed (CRLH) copla- tured and measured. Experimental results fully validate the nar waveguide (CPW) structures. The antennas were fabri- two band pass filter designs. cated on silicon substrates. They were designed, processed H.-M. Lee and H.-S. Lee describe a method to extend the and electrically characterized in order to operate on two bandwidth of metamaterial absorber using multiresonance different frequencies in the mm-wave domain (i.e., 𝑓1 = structure (a periodic arrangement of an electric-LC resonator 27 GHz and 𝑓2 = 38.5 GHz). and a square loop structure). Numerical simulations and J. J. Yang et al. provide an interesting and informative experimental results show the efficiency of the method, review of various different forms of metamaterial . The demonstrating excellent performance. The authors increase article is detailed and well written. It contains an extensive 2 International Journal of Antennas and Propagation list of references and is excellent reference material for any individual wishing to gain more knowledge of the subject area. G. Exposito-Dom´ ´ınguez et al. propose a design in order to suppress the surface wave propagation modes and con- sequently to reduce the mutual coupling between radiating elements in low permittivity substrates. This structure, which is an improvement of the mushroom electromagnetic band gap (EBG) one by means of double metallic layer design and edge via location, leads to a size reduction factor of about 30%. H. N. B. Phuong et al. propose a design based on fractal Sierpinski Gasket patterns in order to synthesize both broad- and dual-band gap media. These designs are experimentally characterized by means of the suspended microstrip method and compared to the performances of conventional mush- room like structures. V. Sanz et al. propose a new equivalent circuit in order to model a coplanar wave guide (CPW) loaded by split ring resonators. This approach, by taking into account the additional capacitive coupling occurring through the CPW, leads to a broadband description that includes both the left- handed and right-handed propagation bands. The papers received for this special issue present a mix- ture of exciting new developments and authoritative reviews. The quality and breadth of topics covered by the papers is impressive. This special issue will be of considerable interest to students, academics, and industry experts alike. We would like to thank all authors and reviewers for making this special issue in Metamaterials possible. Alejandro Lucas Borja James R. Kelly Fuli Zhang Eric Lheurette Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2013, Article ID 162396, 8 pages http://dx.doi.org/10.1155/2013/162396

Research Article Novel Design of Electromagnetic Bandgap Using Fractal Geometry

Huynh Nguyen Bao Phuong,1 Dao Ngoc Chien,2 and Tran Minh Tuan3

1 School of Electronics and Telecommunications, Hanoi University of Science and Technology, Hanoi 10000, Vietnam 2 Department of High Technology, Ministry of Science and Technology, Hanoi 10000, Vietnam 3 National Institute of Information and Communications Strategy, Ministry of Information and Communications, Hanoi 10000, Vietnam

Correspondence should be addressed to Huynh Nguyen Bao Phuong; [email protected]

Received 7 September 2012; Revised 11 January 2013; Accepted 28 January 2013

Academic Editor: Eric Lheurette

Copyright © 2013 Huynh Nguyen Bao Phuong 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 electromagnetic bandgap (EBG) structural design based on Fractal geometry is presented in this paper. These Fractals, ∘ which are the Sierpinski triangles, are arranged to repeat each 60 to produce the hexagonal unit cells. By changing the gap between two adjacent Sierpinski triangles inside EBG unit cell, we can produce two EBG structures separately that have broadband and dual bandgap. By using the suspending microtrip method, two arrays 3 × 4 of EBG unit cells are utilized to investigate the bandgap of the EBG structures. The EBG operation bandwidth of the broadband structure is about 87% and of the dual-band structure is about 40% and 35% at the center bandgap frequencies, respectively. Moreover, a comparison between the broadband EBG and the conventional mushroom-like EBG has been done. Experimental results of the proposed design show good agreement in comparison with simulation results.

1. Introduction property,allofthosestructuresarenarrowbandwhichmakes them not practical for use in broadband applications. It is a Recently, an intensive interest in metamaterials, such as high real challenging task to create an EBG with wide bandwidth. impedance surface (HIS) composed of metallic or dielectric Several broadband EBG structures were found in the unit cells, can be observed due to their unique characteristics literature. Typically, there have been two approaches aiming in antennas and microwave circuits applications [1]. In gener- to obtain EBG structures with wider bandwidth: the use al, HIS has two unique properties included of electromagnetic of EBGs with via-holes [5] (with the inconveniences of bandgap (EBG) and artificial magnetic conductor (AMC). complex and expensive manufacturing process) and the One of the most important properties of EBG structure adoption of multilayered FSSs over a metallic ground plane or is to prohibit the propagation of the surface wave [2, 3]. multiperiod mushroom-like structure [6] (which yields less Practical applications of EBG structure usually have difficulty compact designs and is rather expensive). Recent research in accommodating its physical size, because the period of efforts focus on the development of planar EBG that does EBG lattices has to be a half-wavelength at the bandgap not need vias and that can be integrated antenna to enhance frequency. This problem had not been solved until the the gain and reduce the backward radiation and increasing mushroom-like EBG was proposed by Sievenpiper et al. [4]. efficiency [7–9]. Then, several of novel EBG structures were presented such As [10], several of Fractal shapes can be used for antennas as spiral EBG, uniplanar compact EBG (UC-EBG), and includes: (1) the von Koch curve, (2) the Minkowski curve, fork-like EBG. These structures have several advantages, (3) the Hilbert curve, (4) the Fractal tree, (5) the Sierpinski such as compact size low loss. However, due to their resonant (gasket and carpet) Fractals, and (6) the Cantor set. These 2 International Journal of Antennas and Propagation geometries also can be used in order to create EBG structures. However, only the Fractal shapes that are used for multiband antennas are able to form broader bandgap EBG. Actually, when an EBG structure has multiband, the broadband can be done by overlapping the different single-band to each other. In literature, Sierpinski (gasket and carpet) structures are Step 1 Step 2 widely used to create multi-band antenna [10]. From the idea of this design that is produced a hexagonal EBG with Fractal geometry, due to the characteristic of an equilateral triangle, Sierpinski gasket is chosen as a metal patch on the top surface in order to create the hexagonal EBG. In this paper, the advantage of the novel hexagonal EBG structure was found by investigating in case of changing Step 3 Step 4 the gap between two adjacent Sierpinski triangles inside EBG Figure 1: Four steps to form a mode-2 Sierpinski gasket structure. unit cell. As a result, two structures are proposed; in which one structure introduces broad bandwidth and the others introduce dual bandgap. The design of EBG structures involv- 𝐺1 𝐺1 ing the mode-2 Sierpinski Gasket triangles (see Figure 1)and without via-holes or multilayer substrate is presented. The proposedEBGstructuresweresimulatedandmeasuredusing 𝐺 the method of suspended microstrip, which is recommended 2 in [11] and a comparison between the measured results and simulated results has been carried out. The rest of this paper is organized as follows. In Section 2, (a) (b) the detailed designs of the novel EBG structures are present- 𝑊 ed.Byusingthesuspendingmicrostripmethod,thescatter 1 𝑊 1 parameters of the proposed EBG structure based on Sierpin- ski triangle that is formed by different steps, are characterized and simulated in Section 3. Besides, the bandgap of the 𝑊3 𝑊4 conventional EBG structure is also determined for verifying 𝑊2 the broader bandwidth of the proposed EBG structure. 𝑊1 Next, the measurement of the proposed EBG structures in (c) fabrication and discussions are also utilized in Section 4, while the conclusion is provided in Section 5.

2. Hexagonal EBGs Design

The resonance frequency and the bandwidth of an EBG 𝐻 structure depend on the unit-cell geometry together with 𝑊 substrate’s relative dielectric permittivity and thickness. Each Metallic ground unit cell implements a distributed parallel network having plane one or more resonant frequencies. The resonance frequency (d) of a parallel circuit is defined as follows: Figure 2: Unit cell geometries: (a) BEBG, (b) DEBG, (c) detailed 1 Sierpinski gasket triangles, and (d) slide view of BEBG. Detailed 𝑊 =𝑊/8 𝑊 =𝑊/4 𝑊 =𝑊/2 𝑊 =𝑊−2(𝐺 + 𝑓𝐶 = . (1) dimensions: 4 1 ; 3 1 ; 2 1 ; 1 2 2𝜋√𝐿𝐶 √ 𝐺1)/ 3, 𝐺2 = 0.5 mm, 𝐺1 =1mm. The bandwidth is also an important consideration and is given by parts in this figure represent the metallic periodic structure, which is etched on a dielectric substrate. Unit cell dimension 1 𝐿 is 𝑊=10mm, and the metallization thickness is 18 𝜇m. 𝐵𝑊 = √ , (2) 𝜂 𝐶 The Sierpinski gasket is a well-known Fractal. The way it can be constructed and its main properties can be found where 𝜂 isthefreespaceimpedance,whichis120𝜋. in [12–14]. In [15], it is shown that the Sierpinski gasket is a The designed unit cell geometry exhibits six symmetry special case of a wider class of Fractals that can be derived planes, which makes it polarization-angle independent. The from the well-known Pascal’s triangle. This class of Fractals proposed EBG structures are shown in Figure 2,whichare can be derived in the following way. Consider an equiangular printed on an FR4 dielectric slab with dielectric constant of triangular grid whose rows shall be labelled by 𝑛 = 1, 2, 3, . . 4.4 and thickness of 1.6 mm and loss tangent of 0.02. Dark Each row contains 𝑛 nodes, and to each node a number International Journal of Antennas and Propagation 3

Suspended Suspended microstrip line Mushroom-like microstrip line Proposed EBG EBG

Dielectric support Dielectric support layer layer Dielectric substrate Dielectric substrate (a) (b)

Figure 3: Geometry of 3 × 4 EBG arrays: (a) mushroom-like EBG array, and (b) proposed EBG array.

is attached. This number is the coefficient of the binomial defined with 𝑆11 above −5dBand 𝑆21 below −30 dB at the 𝑛−1 expansion of (𝑥 + 𝑦) . Now delete from this grid those same time. nodes that are attached to numbers that are divisible by 𝑝, where 𝑝 is a prime number. The result is a self-similar Fractal 3.1. EBG Structures at Different Steps. The hexagonal EBG that will be referred to as the mod - 𝑝 Sierpinski gasket [13]. structures with Sierpinski triangles at different steps are In this paper, The Sierpinski gasket triangle is formed by investigated in case the value of 𝑊 is fixed at 10 mm while the following process (see Figure 2). Firstly, create an equi- the value of 𝐺2 is set at 0 mm and 0.5 mm. The simulation lateral triangle patch with length edge of 𝑊1.Thesecond results of the EBG structures based on the Sierpinski triangle step of this Fractal EBG is constructed by etching out the at step 1, step 2, and step 3 are shown in Figures 4, 5,and6, center equilateral triangle of length edge 𝑊2 inside the respectively.Incaseofstep1,ascanbeseeninFigure 4(a),the patch and subtracting it from the patch. The next stage is EBG structure introduces a bandgap, which is ranging from achieved by subtracting additional three equilateral triangles 5.07 GHz to 7.58 GHz, as 𝐺2 is equal to 0.5 mm. However, the of length edge 𝑊3 of the center triangles that follows the same bandgap was not found in case of 𝐺2 value is equal to 0 mm technique adopted to realize the mode-2 Sierpinski gasket (see Figure 4(b)).AscanbeseenfromFigure 5(a),theEBG geometry (see Figure 2(c)). structure based on the Sierpinski triangle at step 2 has a lower From this design, the Sierpinski gasket triangles are ∘ bandgap than the one in the case of step 1. The bandwidth of arranged to repeat each 60 .Theaimofthisworkallows this bandgap covers from 4.22 to 6.88 GHz. Two bandgaps are modifying the resonance frequencies and the EBG operation defined in Figure 5(b) when the value of 𝐺2 issetat0mm. bandwidth by changing the width and the gap 𝐺1 between The lower bandgap covers from 2.25 GHz to 2.96 GHz, and the hexagonal lattices. Moreover, by changing the gap 𝐺2 the higher one spans from 4.14 GHz to 5.34 GHz. In this case, between Sierpinski triangle units, two EBG structures are both bandgaps are defined not clearly, because the ranges formed separately. The first one, which is called broadband of frequencies of 𝑆11,whichareupto−5dB,arenotflat. EBG structure (BEBG), has a broader bandgap with the value AscanbeobversedfromFigure 6(a),thebandgapofEBG of 𝐺2 greater than 0 mm. The second one, which is called structure in step 3 is larger than the ones in the case of step dual-bandEBGstructure(DEBG),hasdualbandgapwhen 2as𝐺2 is set at 0.5 mm. This bandgap spans the frequencies the value of 𝐺2 issetat0mm.Thesestructuresareshownin from 4.32 to 7.92 GHz. Moreover, when 𝐺2 is equal to 0 mm Figures 2(a) and 2(b), respectively. the bandgaps are defined easier than the one in the case of step 2. From Figure 6(b),thecurveofS11isquiteflatintwo 3. The Bandgap Characteristics bandgaps. The first bandgap spans the frequencies from 2.15 to3.02GHz;thesecondonehasthebandwidththatiscovered In this section, the bandgap characteristics of two hexagonal from 3.81 GHz to 5.20 GHz. Next, the bandgap properties of EBG structures, which are formed by the Sierpinski triangle the EBG structures at step 4 are considered by investigating at 𝐺2 =0and 𝐺2 = 0.5 mm, are investigated. Besides, the the effect of parameters such as the size of unit cell 𝑊,and bandgap of the conventional mushroom-like EBG is also the gap between two adjacent unit cells 𝐺1 to the bandgaps in determinedincomparisonwiththeproposedEBGstruc- both cases of DEBG and BEBG. tures. In order to analyze the bandgap properties of these EBG structures with finite unit number, an experiment concerning transmission through the above structures has been carried 3.2. Broadband EBG (BEBG). In general, in order to obtain a out. A 3 × 4conventionalEBGstructureandproposed wider EBG operation bandwidth, it is necessary to increase EBG structures have been simulated using the method of 𝐿 and reduce 𝐶. Equivalent inductance 𝐿 can be increased suspended microstrip, which is proposed by Fan to measure using a thicker dielectric substrate and also including in thebandgapcharacteristicoftheEBGs,asshowninFigure 3 the geometry narrow and long strips (lines). Equivalent [11]. The operating frequency is set at 7 GHz, and the 50 Ω capacitance 𝐶 can be reduced by reducing substrate’s relative microstrip line is placed on a dielectric support layer with dielectric permittivity and increasing the gap between the the thickness of 0.5 mm. The bandgap bandwidth will be metallization edge and the unit cell edge (and so the gap 4 International Journal of Antennas and Propagation

0 0

−10 −10 −20

−30 −20

−40 −30 S-parameter (dB) S-parameter

−50 (dB) S-parameter

−60 −40

−70 −50 345678910 2 3 4 5 6 Frequency (GHz) Frequency (GHz)

𝑆11 𝑆11 𝑆21 𝑆21

(a) (b)

Figure 4: Operation bandwidth of hexagonal EBG structure based on Sierpinski triangle at step 1: (a) 𝐺2 = 0.5 mm and (b) 𝐺2 =0mm.

0 0

−10 −10 −20 −20 −30

−40 −30

−50

S-parameter (dB) S-parameter −40 S-parameter (dB) S-parameter −60 −50 −70

−80 −60 345678910 23456 Frequency (GHz) Frequency (GHz)

𝑆11 𝑆11 𝑆21 𝑆21 (a) (b)

Figure 5: Operation bandwidth of hexagonal EBG structure based on Sierpinski triangle at step 2: (a) 𝐺2 = 0.5 mm and (b) 𝐺2 =0mm. between adjacent unit cells). Moreover, the thickness of the EBG unit cells (which decreases coplanar 𝐶). More substratehastobeadoptedinordertoobtainbothcompact important, in addition, the contribution due to the capac- size and broad EBG operation bandwidth. itance between unit-cell metallization and the unit-cell In the EBG design procedure, if the dielectric material ground plane (parallel plates) is reduced for smaller unit-cell anditsthicknesshavebeenchoseninthisanalysis,the size (𝑊). total equivalent inductance 𝐿 cannot be altered. Therefore, Finally, the wider EBG operation bandwidth for smaller only capacitance 𝐶 can be changed. Regarding the total unit-cell size (𝑊) of this novel design is due to more signif- equivalent 𝐶 of the unit cell, the contribution due to the icantly decreased 𝐶 values of the parallel equivalent circuit. gap between adjacent EBG cells (coplanar 𝐶)variesasthe Moreover, the EBG operation bandwidth is also increased by distance between unit cells reduces (which increases 𝐶) reducing the relative dielectric permittivity for a given unit- and also the distance between the Sierpinski triangles in cell size. International Journal of Antennas and Propagation 5

0 0

−10 −10 −20 −20 −30

−40 −30

−50 S-parameter (dB) S-parameter S-parameter (dB) S-parameter −40 −60 −50 −70

−80 −60 345678910 23456 Frequency (GHz) Frequency (GHz)

𝑆11 𝑆11 𝑆21 𝑆21 (a) (b)

Figure 6: Operation bandwidth of Hexagonal EBG structure based on Sierpinski triangle at step 3: (a) 𝐺2 = 0.5 mm and (b) 𝐺2 =0mm.

0 0

−10 −10

−20 −20 −30 −30

(dB) −40 (dB) 21 21 𝑆 −40 −50 𝑆

−60 −50

−70 −60

−80 −70 4 6 8 10 12 14 16 18 20 22 24 26 28 30 10 12 14 16 18 20 22 Frequency (GHz) Frequency (GHz) 𝑊=2 𝑊=6 mm mm 𝐺1 = 1.0 mm 𝑊=4mm 𝑊=10mm 𝐺1 = 1.5 mm 𝐺 = 2.0 Figure 7: EBG operation bandwidth variations versus EBG unit-cell 1 mm size (𝑊). Figure 8: BEBG operation bandwidth variations versus the gap between adjacent EBG cells (𝐺1) while the unit-cell (𝑊)isfixedat 4 mm. In order to analyses the effect of the parameters of dimension on the total equivalent capacitance 𝐶,theunit- cell (𝑊)andthegapbetweenadjacentEBGunits(𝐺1)are size increasing, and the bandgap becomes narrow. From investigated while other parameters are constant. The unit- Figure 7,wecanseethattheequivalentcapacitancedimin- cell size (𝑊) is very important parameter to take effect on the ished when the patch size increases, so the resonant frequency bandgap, which affects the equivalent capacitance of the LC is decreased. Since the bandgap width is proportional to 𝐿/𝐶, resonant circuit of the EBG structure. Firstly, we investigate the bandgap becomes narrow. The detail of this is shown in the bandgap of the EBG structure with different values of Table 1. 𝑊.Theunitcellsizeisincreasedfrom2mmto10mm Moreover, the effect of distance between adjacent EBG while other parameters of the EBG structure are unchanged. cells (𝐺1) on the width of the bandgap is also investigated, The bandgap moves to lower frequency region with patch and the simulated results are shown in Figure 8.Ascan 6 International Journal of Antennas and Propagation

0 Table 1: Parametric analysis of EBG unit cell. Unit cell Resonance Thickness Bandwidth Reference size 𝜀𝑟 frequency −10 (mm) (%) (mm) (GHz) 2(𝜆/6.4) 1.6 (𝜆/8) 4.4 49 23.50 −20 𝜆 𝜆 This paper 4( /5.2) 1.6 ( /13) 4.4 65 14.05 6(𝜆/4.9) 1.6 (𝜆/5) 4.4 76 10.06 𝜆 𝜆 −30 10 ( /4.0) 1.6 ( /4) 4.4 87 7.20 S-parameter (dB) S-parameter Table 2: Relative operation bandwidth of EBG structure for various −40 𝑊 while 𝐺2 is Fixed at 0 mm and 0.5 mm. 𝐺 𝐺2 = 0.5 mm 2 =0mm Unit cell size −50 Bandgap Bandgap 1 Bandgap 2 4 567891011121314 𝑊 (mm) (GHz) (GHz) (GHz) Frequency (GHz) 2 17.74–29.25 10.77–14.47 18.80–21.95 𝑆11 4 9.48–18.62 5.51–7.73 10.20–12.33 𝑆 21 6 6.23–13.88 3.72–5.07 6.72–8.53

Figure 9: DEBG operation bandwidth with 𝐺2 isequalto0mmand 10 4.05–10.35 2.06–3.08 3.71–5.26 𝑊 is equal to 4 mm while the other parameters are fixed. Table 3: Operation bandwidth of proposed EBG structures with 𝑊=10mm. 0 BEBG 𝐺 𝐺 DEBG ( 2 = 0 mm) −10 ( 2 = 0.5 mm) Methods Bandgap 1 Bandgap 2 Bandgap (GHz) (GHz) (GHz) (bandwidth (%)) −20 (bandwidth (%)) (bandwidth (%)) 4.05–10.35 2.06–3.08 3.71–5.26 Simulated −30 (87) (40) (35) 4.15–10.00 2.17–2.97 3.58–5.32 Measured (83) (32) (39)

S-parameter (dB) S-parameter −40

−50 of this design that is the structure can be transformed the bandwidth of broadband into dual bandgap when 𝐺2 is equal −60 to zero. When the unit cell size 𝑊 is set at 4 mm, the first 3 4567891011 bandgap spans from 5.51 GHz to 7.73 GHz. The bandwidth Frequency (GHz) of the second one is spreading from 10.20 GHz to 12.18 GHz. This can be indicated more clearly in Table 2. 𝑆11 ItcanbeobservedfromTable 2 that these bandgaps 𝑆21 move forward to the lower frequency regions. It is clear that Figure 10: Simulated results of scatter parameters of conventional when the value of 𝐺2 is up to zero, two above bandgaps are mushroom-likeEBGstructure(𝑊=10mm). moving to overlap each other and the broader bandgap can be obtained. be observed from Figure 10,theEBGoperationbandwidth 3.4. Conventional Mushroom-Like EBG. In order to verify 𝐺 becomes wider by increasing the gap 1 between adjacent the bandgap characteristic of the broadband EBG structure EBGcellsandthebandgapmovestothehigherfrequency (BEBG), an array of 3 × 4 unit cells of hexagonal mushroom- 𝐶 area. It is shown that the coplanar is increased with the like EBG is investigated (see Figure 3(a)). This EBG is also 𝐺 higher values of 1. constructed on a dielectric substrate with a relative permit- tivity of 4.4 and thickness of 1.6 mm. The unit cell dimension 3.3. Dual-Band EBG (DEBG). In this section, the gap 𝐺2 is fixed at 10 mm. The simulated result of transmission between two adjacent Sierpinski triangles inside the EBG unit coefficient is shown in Figure 10. The bandgap is center at cell is also investigated while other parameters are constant. 6.77 GHz, and the bandgap bandwidth is from 5.22 GHz to From Figure 9, it can be observes that two bandgaps are deter- 8.32 GHz, achieve the bandgap bandwidth about 46%. As mined by transmission coefficient. The unique characteristic can be observed from Table 1,thebandgapbandwidthof International Journal of Antennas and Propagation 7

(a) (b)

Figure 11: Photos of the proposed structures with suspended microstrip: (a) array of 3 × 4 DEBG cells and (b) array of 3 × 4BEBGcells.

0 0

−10 −10 Operation Operation −20 BW1 BW2 −20 Operation bandwidth −30 −30 −40 −40 −50 S-parameter (dB)

S-parameter (dB) −50 −60 −60 −70 −70 −80 3 4567891011 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 Frequency (GHz) Frequency (GHz) 𝑆 𝑆 𝑆11 (sim) 𝑆21 (sim) 11 (sim) 21 (sim) 𝑆 𝑆21 (mea) 𝑆11 (mea) 𝑆21 (mea) 11 (mea)

(a) (b)

Figure 12: EBG operation bandwidth (𝑊=10mm): (a) BEBG and (b) DEBG. the BEBG structure at 𝑊=10mm is larger than that of the the criteria of −30 dB, from that the electromagnetic wave conventional EBG at the same dimension of the unit cell. cannot propagate. The bandgap bandwidth is about 83%. The reason is that the surface impedance of the EBG structure 4. Results and Discussions becomes very high within the bandgap. From Figure 12(b), two distinct bandgaps of DEBG structure are defined in Both schematics of proposed EBG structures were simulated which the first bandgap is ranging from 2.17 to 2.97 GHz and and taken into comparison with the measurement of the the bandwidth of the second bandgap is spreading from 3.58 fabrication. The microstrip lines are soldered with SMA to 5.32 GHz. The bandgap bandwidths of the DEBG structure connector to measure the scatter parameters. The measured are32%and39%,respectively.ItcanbeobservedfromTable 3 results of scatter parameters of the arrays are performed by that the measured results show good agreement with the Anritsu 37369D network analyzer. Photos of the proposed simulated results. EBG structures are shown in Figure 11.Thesimulatedandthe measured results of the parameters of BEBG structure and 5. Conclusion DEBGstructureareshowninFigures12(a) and 12(b),respec- tively. It is very efficient to determine the bandgap by the A novel EBG design is presented using Sierpinski gas- ∘ transmission curves. From the measurement of 𝑆11 with the ket triangles, which are arranged to repeat 60 to form criteria of −5 dB, the operation bandwidth can be determined the hexagonal EBG unit cell. Two EBG structures, which have easily when the transmission coefficient 𝑆21 is below −30 dB. broadband and dual bandgap, are proposed by setting the AscanbeobservedfromFigure 12(a),theBEBGintroduces value of the gap between two adjacent Sierpinski inside the a bandgap between the frequencies 4.15 GHz–10 GHz with unit cell is equal to 0 mm and greater than 0 mm, respectively. 8 International Journal of Antennas and Propagation

The suspending microstrip method is used to simulate the [14] H. O. Peitgen, H. Jurgens, and D. Saupe, Chaos and Fractals, New scatter coefficients of the EBG structure. The bandwidth of Frontiers in Science, Springer, New York, 1992. BEBG structure is much larger than that of the conventional [15] N. S. Holter, A. Lakhtakia, V. K. Varadan, V. V. Varadan, and R. EBG. The results show a good agreement between simula- Messier, “On a new class of planar fractals: the Pascal-Sierpinski tions and measurements. Due to several advantages of the gaskets,” Journal of Physics A,vol.19,no.9,article047,pp.1753– structures such as using inexpensive dielectric substrate FR4 1759, 1986. and planar structure, the EBG structures are promising for low profile and low cost antennas in broadband or multiband applications.

References

[1] F. R. Yang, K. P. Ma, M. Yongxi Qian, and T. Itoh, “A uni- planar compact photonic-bandgap (UC-PBG) structure and its applications for microwave circuits,” IEEE Transactions on Microwave Theory and Techniques,vol.47,no.8,pp.1509–1514, 1999. [2] J. D. Joannopoulos, R. D. Meade, and J. N. Winn, Photonic Crystals, Princeton University Press, Princeton, NJ, USA, 1995. [3] E. Yablonovitch, “Photonic band-gap structure,” Journal of the Optical Society of America B, vol. 10, no. 2, pp. 283–295, 1993. [4]D.Sievenpiper,L.Zhang,R.F.JimenezBroas,N.G.Al- exopolous,¨ and E. Yablonovitch, “High-impedance electro- magnetic surfaces with a forbidden frequency band,” IEEE Transactions on Microwave Theory and Techniques,vol.47,no. 11, pp. 2059–2074, 1999. [5] A. Stark, S. Prorok, and A. F. Jacob, “Broadband EBG structures with compact unit cell,” in Proceedings of the 38th European Microwave Conference, EuMC 2008, pp. 698–701, Amsterdam, The Netherlands, October 2008. [6]L.Liang,C.H.Liang,X.W.Zhao,andZ.J.Su,“Anovel broadband EBG using multi-period mushroom-like structure,” in Proceedings of the International Conference on Microwave and Millimeter Wave Technology (ICMMT ’08),pp.1609–1612, Nanjing, China, April 2008. [7] R. Coccioli, F. R. Yang, K. P. Ma, and T. Itoh, “Aperture-coupled patch antenna on UC-PBG substrate,” IEEE Transactions on Microwave Theory and Techniques,vol.47,no.11,pp.2123–2130, 1999. [8] M. Rahman and M. A. Stuchly, “Circularly polarised patch antenna with periodic structure,” IEE Proceedings: Microwaves, Antennas and Propagation,vol.149,no.3,pp.141–146,2002. [9] M. F. Abedin, M. Z. Azad, and M. Ali, “Wideband smaller unit-cell planar EBG structure and their application,” IEEE Transactions on Antennas and Propagation,vol.56,no.3,pp. 903–908, 2008. [10] W. J. Krzysztofik, “Modified Sierpinski fractal monopole for ISM-bands handset applications,” IEEE Transactions on Anten- nas and Propagation,vol.57,no.3,pp.606–615,2009. [11] M. Y. Fan, R. Hu, Q. Hao, X. X. Zhang, and Z. H. Feng, “New method for 2D-EBG structures research,” Hongwai Yu Haomibo Xuebao/Journal of Infrared and Millimeter Waves,vol.22,no.2, pp. 127–131, 2003. [12] C. Puente, J. Romeu, and A. Cardama, “Fractal-shaped anten- nas,” in Frontiers in Electromagnetics, D. H. Werner and R. Mittra, Eds., pp. 48–93, IEEE Press, Piscataway, NJ, USA, 2000. [13]C.Puente-Baliarda,J.Romeu,R.Pous,andA.Cardama,“On the behavior of the sierpinski multiband fractal antenna,” IEEE Transactions on Antennas and Propagation,vol.46,no.4,pp. 517–524, 1998. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2013, Article ID 637270, 16 pages http://dx.doi.org/10.1155/2013/637270

Review Article Metamaterial Sensors

Jing Jing Yang,1 Ming Huang,1 Hao Tang,2 Jia Zeng,1 and Ling Dong1

1 Wireless Innovation Lab, School of Information Science and Engineering, Yunnan University, Kunming,Yunnan 650091, China 2 Radio Monitoring Center of Yunnan, Kunming,Yunnan 650228, China

Correspondence should be addressed to Ming Huang; [email protected]

Received 5 October 2012; Revised 11 December 2012; Accepted 15 December 2012

Academic Editor: James R. Kelly

Copyright © 2013 Jing Jing Yang 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.

Metamaterials have attracted a great deal of attention due to their intriguing properties, as well as the large potential applications for designing functional devices. In this paper, we review the current status of metamaterial sensors, with an emphasis on the evanescent wave amplification and the accompanying local field enhancement characteristics. Examples of the sensors are given to illustrate the principle and the performance of the metamaterial sensor. The paper concludes with an optimistic outlook regarding the future of metamaterial sensor.

1. Introduction perfectlensdoesnotchangethedecayingcharacterofeva- nescent wave, the idea of far field perfect imaging through Metamaterials are manmade media with all sorts of unusual converting evanescent waves to propagating waves has been functionalities that can be achieved by artificial structuring proposed and demonstrated experimentally [13, 14]. On the smaller than the length scale of the external stimulus [1]. They otherhand,theamplificationofevanescentwavehasbeen have provided many possibilities for exploring unknown showntobeabletoenhancetheinteractionbetweenwave physical phenomena such as reverse Vavilov-Cherenkov and mater and then increase the sensitivity of sensors [15]. effect [2], negative refraction [3], cloaking [4–7], concen- A large number of researches about metamaterial sensors trator [8], perfect lens [9], and negative compressibility have emerged over the last several years. Schueler et al. [10]. In recent years, sensing applications of metamaterials [16] reviewed the metamaterial inspired composite right/left- have attracted a great deal of attentions. It is well known handed transmission line microwave sensors. Chen et al. that conventional optics suffer from Abbe diffraction limit, [17] reviewed metamaterials application in sensing with an since they are only capable of transmitting the propagating components, and the maximum resolution in the image can emphasis on split ring resonator-based sensors. Our group never be greater than half a wavelength. In a pioneer work, has been dedicated to the study of metamaterial sensors for Pendry [9] demonstrated that phase of propagating waves a long time, and a great portion of works have been done and the amplitude of the evanescent states could be restored [15, 18–29]. Zheludev [30] analyzed the future development by the perfect lens made up of materials with negative index of metamaterials and pointed out that sensing application of refraction and thus giving rise to a resolution below represents a growing area. the diffraction limit. The essence of the perfect lens lies in In this paper, we make a review of the metamaterial the evanescent wave amplification induced by the negative sensors with an emphasis on evanescent wave amplification refraction materials. It has been achieved experimentally by and the concomitant effects. In Section 2, the generation of Grbic and Eleftheriades [11] using transmission line meta- surfacewaveattheboundarybetweennegativeandpositive materials. For the metamaterial slab with thickness of 𝑑 and materials is revisited, and examples of metamaterial planar loss of 𝛿,theresolutionwillbeΔ = 2𝜋𝑑/ ln(2/𝛿) [12]. If the waveguide sensors and surface whispering gallery mode sen- loss approximated zero, infinite resolution could be realized. sors are given to illustrate phenomenon of evanescent wave Although the loss metamaterials cannot be eliminated, and amplification for sensing application. In Section 3,sensors 2 International Journal of Antennas and Propagation based on planar metamaterial arrays with enhanced sensi- tivity for the detection of mechanical deformation, graphene atomic layer, and for label-free image of biochemical samples Cladding are introduced. In Section 4, examples of sensors based on 0 stacked metamaterial structures are given to illustrate the subwavelength imaging characteristic. Section 5 focuses on 1 Guiding layer the metamaterial sensors based on a single metamaterial particle with the advantages of fabrication simplicity and experimental robustness. In Section 6, the other kinds of metamaterial sensors such as the sensors based on squeezing and tuning effect of epsilon near zero materials and the open Substrate cavity are illustrated. In the last section, a conclusion is given.

2. Evanescent Wave Amplification In 2000, Pendry [9] found that negative refraction makes (a) a perfect lens due to the amplification phenomenon of evanescent wave. To give a vivid picture of this phenomenon, Cladding we revisit the slab waveguide model shown in Figure 1(a). Metamaterial Inthecaseoftimeharmonicfieldandlossless,thatis,both 0 𝜀 and 𝜇 areallrealnumbers,supposethatmagneticfield 𝐻 is polarized along the 𝑦-axis and TM wave travels in Guiding layer the 𝑧-direction. When a layer of metamaterials is covered on the surface of the guiding layer as shown in Figure 1(b), evanescent wave at the boundary between the metamaterial layer and the cladding layer will be amplified. This can be proved by deriving the dispersion equation of this structure. Substrate Figure 2 shows the distribution of 𝐻𝑦(𝑥) along the 𝑥-axis of the simulation model. It can be seen that the evanescent wave in cladding layer is enhanced by metamaterial layer [15].

2.1. Planar Waveguide Sensor. Waveguide sensors have found (b) a wide range of applications such as the detection of harmful Figure 1: Simulation model of the traditional planar waveguide sen- gases [31] and chemical analytes [32]. Such sensors are also sor (a) and the planar waveguide sensor covered with a layer of met- knownasevanescentwavesensorsbecauseoftheevanescent amaterials (b). wave entering into the analyte whose refractive index is to be measured. Due to the interaction of the analyte and the evanescent wave, changes can be observed in the absorption dispersion equation of TM mode planar optical waveguide or phase shift of the light propagating through the waveguide, with metamaterial layer and showed that the sensitivity of giving an indication of the concentration of refractive index this waveguide was much higher than that of traditional oftheanalyte.Theamplificationofevanescentwaveforms TM mode planar optical waveguide sensor. Recently, the the basis for increasing the sensitivity. Many researchers have nonlinear planar optical waveguide sensor with metamaterial attemptedtoincreasethepenetrationdepthandstrength layer is proposed by our group [19, 20], and both the TE and of evanescent wave by bending or tapering the optical TM mode dispersion equations are derived and analyzed in fiber [33–37] and altering the light launching angle [38]. detail. We demonstrated that metamaterials combined with Horvath´ et al. [39, 40] have shown that penetration depth nonlinear waveguide will further enhance the sensitivity of of evanescent wave can be increased by using reverse sym- the optical waveguide sensors. metry configuration that the refractive index of the aqueous Figure 3(a) shows the schematic diagram of a four-layered cladding is higher than that of the substrate material. Taya waveguide sensor model. It is supposed to be infinite in both et al. [41] investigated the sensitivity of asymmetrical optical the 𝑧-and𝑦-direction. The waveguide consists of four layers, waveguides with nonlinear cladding and substrate, of which from the top to the bottom, which are semi-infinite nonlinear the permittivity displays a Kerr-type response. They show cladding, metamaterial layer, guiding layer, and semi-infinite that sensitivity of the nonlinear planar asymmetrical optical nonlinear substrate, respectively. Thickness of the guiding waveguidesensorishigherthanthatoftheconventional layer and the metamaterial layer is denoted as 𝑑1 and 𝑑2, 𝜀 𝜀 𝜀 asymmetrical optical waveguide sensor. Through inserting a respectively. Permittivities of the four layers are nlc, 𝑚, 𝑔, 𝜀 layer of metamaterials with negative permittivity and negative and nls.TheTMmodenonlinearcladdingandsubstrate permeability between the cladding and the guiding layer, aresupposedtobeKerrtype.Normalizedmagneticfields they found that the sensitivity of the waveguide sensor of the nonlinear planar waveguide sensor with and without can be dramatically enhanced [42]. We [18]studiedthe the metamaterial layer are simulated and compared in International Journal of Antennas and Propagation 3

shown in Figure 4(d), sensitivity decreases gradually with 𝑑1, and the thinner the guiding layer, the better the performance of the sensor. Therefore, the sensitivity of TM waveguide 1500 sensor in reverse symmetry configuration possesses much higher sensitivity than TE waveguide sensor. 1000 2.2. Surface Whispering Gallery Mode. From the previous Cladding analysis, we can conclude that the sensitivity of the planar 500 waveguide sensor can be greatly enhanced by metamaterials due to the amplification of evanescent wave. Interestingly, we 0 Metamaterial find that when the dielectric slab covered with a layer of meta- Guiding layer materials with negative permittivity and/or permeability is bent over to form a four-layer cylindrical waveguide as shown − 500 in Figure 5(a), the evanescent wave can also be amplified [21]. Figure 5(b) shows the cross-section of the cylindrical − 1000 waveguide structure. Electric field distribution in the cross- Substrate section of the dielectric waveguide coated with a layer of metamaterials at the eigenfrequency of 𝑓𝑒 = 165.664 THz is − 1500 0 0.5 1 1.5 2 plotted in Figure 5(c). Figure 5(d) displays the electric field distribution in the cross-section of a conventional dielectric waveguide. It is clear that for the metamaterial-assisted 𝐻 (𝑥) Figure 2: Magnetic field 𝑦 distribution in the conventional pla- microring, maximum electric field of WGM moves to the sur- nar waveguide sensor (blue dotted line) and metamaterials assisted face of the metamaterial layer. This is named as surface whis- 𝑑 planar waveguide sensor (red line), and 2 is the thickness of the pering gallery mode (SWGM). It exposes a strong evanescent metamaterials (red line) [15]. field to the surrounding, and thus this region will be quite sensitive in dielectric environment. Besides, SWGM can also be generated on the inner surface of a hollow dielectric waveguide and the surface of a circular or elliptical dielectric Figure 3(b). We can clearly observe that there is a sharp cylinder when coated with a metamaterial layer [22–24]. increase of evanescent field at the boundary between the The SWGM sensor has been demonstrated to have much metamaterial layer and the nonlinear cladding. The magnetic higher sensitivity than that of conventional WGM sensor. To field intensity at the surface of the metamaterial layer is about giveaquantitativeillustrationforthesensitivityoftheSWGM 2.46timesthatofthewaveguidewithoutthemetamaterial sensor, resonant frequency, frequency shift, and 𝑄 factor of layer. For the TE mode nonlinear planar waveguide, the theSWGMindielectricsensingaresimulatedandcompared evanescent application phenomenon can also be obtained. with the traditional WGM sensor, as shown in Table 1 [21]. Sensitivity (𝑆𝑚) of the sensor depends on the modal FortheWGMsensor,theaveragefrequencyshifttoan effective index (𝑁) change rate with respect to the change increase of 0.02 in substance permittivity is only 13.25 GHz, of cladding index (𝑛𝑐), that is, 𝑆𝑚 =(𝜕𝑁/𝜕𝑛𝑐).Through and the sensitivity (defined as resonance wavelength shift differentiating the dispersion relation of the nonlinear planar over the refractive index change unit) is about 1.4 nm/RIU. waveguide sensor, the sensitivity can be obtained. Sensi- For the SWGM sensor, the response to an increase of 0.02 tivity versus 𝑑1 for the nonlinear optical waveguide sensor in substance permittivity is a significant resonant frequency 𝑆 𝑆 with metamaterials ( 𝑚) and without metamaterials ( nm) downshift of 187 GHz in average, and the sensitivity is is plotted in Figures 4(a) and 4(b).Inthecaseofnormal about 29 nm/RIU, which is more than 20 times that of the symmetry (𝑎𝑠 >𝑎𝑐), sensitivity increases with guiding layer WGM sensor. This is due to the amplification of evanescent thickness and reaches a maximum value at a point around field which ensures a strong interaction between light and 𝑑1 = 200 nm and then decreases gradually. Here, 𝑎𝑠 =𝜀𝑠/𝜀𝑔 substance. Interestingly, the sensitivity of the SWGM sensor and 𝑎𝑐 =𝜀𝑐/𝜀𝑔 are the asymmetry parameters. In the case can be further enhanced by increasing the thickness of of reverse symmetry (𝑎𝑠 <𝑎𝑐), the sensitivity decreases the metamaterial layer. Figure 6 shows the relation between monotonously with 𝑑1.Comparingthetwofigures,wecan resonant frequency and substance permittivity for different conclude that metamaterials can improve the sensitivity of the metamaterial layer thickness (𝑡). When the thickness of optical waveguide sensor, and the optical waveguide sensors metamaterial layer is 0.05 𝜇m, the average frequency shift with metamaterials in reverse symmetry mode has much in response to a 0.02 increase in substance permittivity higher sensitivity. For TM and TE mode optical waveguide is about 42 GHz, and the sensitivity is 5 nm/RIU. When sensors with metamaterials, the sensitivity as a function of the thickness of metamaterial layer is 0.2 𝜇m, the average 𝑑1 issimulatedandshowninFigures4(c) and 4(d).Inthe frequency shift will be 204 GHz, and the sensitivity can be case of normal symmetry shown in Figure 4(c),anoptimal increased up to 50 nm/RIU. This is because much more guiding layer thickness is observed at around 60 nm for TE power is transferred to SWGM with increasing metamaterial mode, while the optimal guiding layer thickness for TM mode thickness. Recently, we have demonstrated theoretically that sensor is about 220 nm. In the case of reverse symmetry when a layer of Au film is deposited on the microring instead 4 International Journal of Antennas and Propagation

x 1000 Nonlinear cladding

Nonlinear cladding εnlc Metamaterial 500 Guiding layer d2 Metamaterials εm 0

d1 Guiding layer εg 0 z − 500 Nonlinear substrate Nonlinear substrate εnls − 1000

0 0.2 0.4 0.6 0.8 1

(a) (b)

Figure 3: (a) Schematic diagram of a cross-sectional view of the nonlinear planar waveguide sensor with a metamaterial layer. 𝑑1 = 500 nm, 𝜀𝑐 =2, 𝜀𝑠 = 2.2, 𝜀𝑔 =4, 𝜀𝑚 =−2, 𝑑2 =80nm, 𝑛 = −0.6,and𝜆 = 1550 nm. (b) Normalized magnetic field 𝐻𝑦 distribution along the 𝑥-direction of the nonlinear planar waveguide sensor without (dotted blue line) and with the metamaterial layer (solid red line) [20].

Table 1: 𝑄 factor, resonant frequency, and frequency shift of WGM by Papasimakis et al. [49] based on the metamaterial sensor sensor and SWGM sensor [21]. made up of an array of asymmetrically split ring resonators. 𝜀 Liu et al. [51] showed that the metamaterial sensor fabricated Sensor s using gold film may serve as a highly efficient localized 1.02 1.04 1.06 1.08 1.1 surface plasmon resonance sensor in the near-infrared with 𝑄 15718 15661 15777 15718 15658 sensitivity of 588 nm/RIU. 𝑓 WGM 𝑟 (THz) 198.248 198.235 198.221 198.208 198.195 An intracellular plasmonic label-free imaging by excit- Δ𝑓 𝑟 (GHz) 13 14 13 13 ing multimode resonances in spilt-ring resonators is pro- 𝑄 16930 16922 16923 17083 17072 posed by Lai et al. [43]. Figure 7 shows the SEM images SWGM 𝑓𝑟 (THz) 165.474 165.286 165.097 164.911 164.725 of the designed SRR samples which were fabricated by

Δ𝑓𝑟 (GHz) 188 189 186 186 standard e-beam lithographic and lift-off processes. One sample contains 10 × 10 unit cells, and each unit cell consists of 5×5SRRs. All SRR unit cells contain exactly of the metamaterial layer, SWGM can also be excited when identical SRR pattern from cell to cell. To demonstrate the theexcitationfrequencyislowerthanplasmafrequency[45]. performance of the planar metamaterial sensor, bioimage of human bone marrow-derived mesenchymal stem cells (hMSCs) based on the fundamental resonance signal of SRR −1 3. Planar Metamaterial Array atthewavenumberof1850–2400cm was conducted and compared with conventional imaging technology. Figure 8(a) Planar metamaterials consisting of subwavelength resonators shows the conventional optical microscopic image of the have been proposed for thin dielectric film sensing. To hMSCs grown on the SRRs samples. The black part in the achieve higher sensitivity, the sensor needs to have a sharp background refers to the SRRs structure. In this case, any resonance in its frequency response and a high concentration detail of the inner nucleus and organelles cannot be revealed of electric field to enable the detection of small changes in without the labeling process. Figure 8(b) shows the confocal dielectric environment. When analyte is deposited on the fluorescent optical microscopic image of the hMSCs, in which surface of the resonators, the effective permittivity at the the nuclei of the hMSCs can be observed. However, such a gap of each resonator is increased. Then, the gap under- labeling process is typically expensive and time consuming, goes a significant change in the charge distribution and impeding the practical application of real-time diagnosis. capacitance, which can be observed from the transmission Figure 8(c) displays the intracellular image of the hMSCs by resonance. This sensing mechanism has been demonstrated theSRRplatform.Itdoesnotrequirethelabelingprocess experimentally and successfully applied to a range of planar but directly detects the change of plasmonic resonance of the metamaterial sensors [43, 44, 46–51]. For example, Melik SRR fluctuated by the local attachment of the targeting bioa- et al. [48] demonstrated that the metamaterial-based strain gents. In the experiment, transmission and reflection were sensors are highly sensitive to mechanical deformation, due characterized by a Fourier-transform infrared spectrometer to the large transmission dips and high quality factors. The equipped with an infrared microscope in the wavenumber −1 detection of a single atomic layer of graphene was realized range of 400–8600 cm , and the corresponding mid-IR International Journal of Antennas and Propagation 5

0.6 0.4

0.5 0.3 0.4

0.2 0.3

0.1 0.2

0 200 400 600 800 300 500 700 (nm)

(a) (b) 0.7 0.4

0.6 0.3 0.5

0.2 0.4

0.1 0.3

0.2 0 100 300 500 700 100200 300 400 (nm)

(c) (d)

Figure 4: Sensitivity versus the guiding layer thickness 𝑑1 for the proposed sensor (𝑎𝑚 =𝜀𝑚/𝜀𝑔 = −0.5, 𝑛 = −0.6,and𝑑2 =30nm) with a metamaterial layer (red solid line) and without a metamaterial layer (blue dotted line): (a) normal symmetry (𝑎𝑠 = 0.55 and 𝑎𝑐 = 0.5)and 𝑎 = 0.5 𝑎 = 0.55 𝑑 𝑆 (b) reverse symmetry ( 𝑠 and 𝑐 ). Sensitivity versus the guiding layer thickness 1 for TM mode ( mTM )(redsolidline)andTE 𝑆 𝑑 =20 𝑎 =𝜀 /𝜀 = −0.5 𝑛 = −0.6 𝑎 = 0.55 𝑎 = 0.5 mode ( mTE ) (blue dotted line) ( 2 nm, 𝑚 𝑚 𝑔 ,and ): (c) normal symmetry ( 𝑠 and 𝑐 )and(d)reverse symmetry (𝑎𝑠 = 0.5 and 𝑎𝑐 = 0.55)[20]. images were captured by a focal planar array detector. All refractometry of is proposed by Reinhard et al. [44]. measured spectra have been normalized with respect to the The sensor operates in reflection geometry and exhibits a reflection spectra of an aluminum mirror. In Figure 8(c), strong frequency shift of a sharp Fano-type resonance min- the region colored in red represents the nucleus. This is imum in the presence of a dielectric sample. The magnitude corresponding to the greatest shift of Resonant frequencies of this shift depends on both the refractive index and the and the strongest reflection intensity. The other colored thickness of the sample. The unit cell is a square with an edge parts mainly refer to the cytoplasm, corresponding to the length of 140 𝜇m, consisting of four metallic crosses on top smaller shift of resonant frequencies. In short, this study of a 10 𝜇m thick dielectric matrix with a relative permittivity demonstrated the feasibility of using SRRs for constructing 𝜀𝑟 = 2.67. Each of the crosses is titled by an angle of ∘ the refractive index distribution of hMSCs to obtain images of 22.5 . Details about the geometry parameters are shown in the target cells. The SRR platform possesses many advantages Figure 9(a). When excited by a horizontally polarized THz beyond other optical microscopy such as label-free and real- wave, the distribution of currents and charge is shown in time diagnosis. Figure 9(b). Figure 9(c) is a microscope image of the fabri- A metamaterial-based terahertz (THz) sensor for thick- cated metamaterials. In order to experimentally prove the ness measurements of subwavelength thin materials and capability of the sensor of measuring the thickness of thin 6 International Journal of Antennas and Propagation

(a) (b) (c) 1

0.7

0.3

0

−0.3

−0.7

−1 (d)

Figure 5: (a) Cylindrical dielectric waveguide covered with a layer of metamaterials. (b) Cross-section of (a). Analytical results of normalized electric field distribution (mode 27) in the cross-section of the waveguide with (c) and without the metamaterial layer (d)21 [ ].

0

−0.2

−0.4

−0.6

−0.8

−1 Frequency shift (THz) Frequency −1.2 Figure 7: The SEM images of the fabricated planar SRRs. The sam- ple consists of 5×5SRRs as a unit cell through standard e-beam −1.4 lithographic and lift-off processes [43]. 1 1.02 1.04 1.06 1.08 1.1

sample materials, silicon layers (𝑛 = 3.4)withthicknesses between approximately 50 nm and 1 𝜇mwereevaporated on top of the metamaterial sensor, and the reflection spec-

Figure 6: The relation between 𝜀𝑟 and resonant frequency for dif- tra were measured using THz time-domain spectroscopy. ferent metamaterial layer thickness. The insets show electric field Figure 10(a) shows the variation of the Resonant frequency distributions at resonant state for 𝑡 = 0.05 𝜇m, 0.12 𝜇m, and 0.2 𝜇m, of the sensor as a function of silicon thicknesses. The sensi- respectively [21]. tivity of the sensor has a maximum value of approximately International Journal of Antennas and Propagation 7

(a) (b) (c)

Figure 8: (a) The unlabeled optical microscopic image of hMSCs on the SRR substrate. (b) The confocal fluorescent microscopic imageof the hMSCs. (c) The intracellular image of the hMSCs by the SRR platform [43].

BCB Au

(a) (b) (c)

∘ Figure 9: Geometry parameters of the metamaterial unit cell. 𝑎 = 140 𝜇m, 𝑤=15𝜇m, 𝑔=4𝜇m, and 𝜃 = 22.5 . The metamaterial consists of gold (Au) crosses on top of a BCB film. (b) Distribution of charges (+, −) and currents (arrows) at resonance when excited by a horizontally polarized incident THz wave. (c) Microscope image of a fabricated metamaterials [44].

1.7

1.6 No sample 1.55 1.5 1.5 Isopropanol 1.45 1.4 Rapeseed oilEthanol: water 4 : 1 Ethanol 3 : 1 3 : 1 1.4 Cyclohexane 2 : 1 1.3 1 : 1 Resonant frequency (THz) Resonant 1 : 1 1.35 1 : 3 Isopropanol: glycerin

Resonant frequency (THz) Resonant 1.2 1.3 Glycerin 0 200 400 600 800 1000 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Silicon thickness (nm) Refractive index

Experiment Pure liquids Full-wave calculations Full-wave calculations Isopropanol/glycerin mixtures Analytical model Analytical model Ethanol/water mixtures (a) (b)

Figure 10: (a) Resonant frequencies of the sensor in dependence on silicon thickness. The slope of the curve has a maximum value of 0.4 THz/𝜇m. (b) Experimentally measured Resonant frequencies of the metamaterial sensor in the presence of several liquids and mixtures [44]. 8 International Journal of Antennas and Propagation

0.4 THz/𝜇m for very small layer thickness, and the thickness resolution is 12.5 nm, which corresponds to approximately 1/16000 of the THz wavelength. To demonstrate the effec- tiveness of the metamaterial structure as a liquid sensor, resonant frequencies in the presence of different liquid sample substances are measured as shown in Figure 10(b). The refractive index sensitivity is 0.43 THz per refractive index unit. In the previous cases, the sample substances that l depositedonthesurfaceofthemetamaterialgiverisetoa shift in resonant frequency due to the change of capacitance. In a recent work, Kenanakis et al. [46] studied the sensing capability of a metamaterial structure made of a pair of square metal slabs as shown in Figure 11.Thewholestructure contains 10 × 10 unit cells. Each side of the pair is printed on a separate FR-4 boards placed facing each other at a distance 𝑑,withtheprintedmetallicstructuresfacingoutwards.The space of thickness 𝑑 between the two FR-4 boards is filled (or Figure 11: Schematic of the slab pair structure. A magnification partially filled) with the materials of interest. To demonstrate of one unit cell of the slab pair structure (red shadowed area) is the sensing capacity of the metamaterial structure, trans- presented. 𝐸, 𝐻,and𝑘 correspond to electric field, magnetic field, mission spectrum for air, low density polyethylene (LPDE), and the wave vector, respectively [46]. FR-4, and p-type silicon are measured and compared with numerical simulation, as shown in Figure 12.Asidefromthe very good agreement between simulations and experiments, based on transfer matrix method, and the effectiveness of this we can observe in Figure 12 aquitelargeshiftofthemagnetic structure as a sensor was demonstrated. resonance to lower frequencies as the permittivity of the On the other hand, the stack of metamaterial structures dielectric between the slab pair changes from 1.0 (air) to has been demonstrated to be able to realize perfect lens 11.90 (silicon). Besides, the metamaterial structure is also and to enhance the resolution of sensors [57, 60–63]. For applicable for measuring the thickness of thin dielectric layer example, an impedance-matched, low loss negative index placed between the slabs of the pairs. The experimental results metamaterial superlens was reported by Aydin et al. [62]. It is show that a very thin layer of LDPE with thickness of 0.1 mm capable of resolving subwavelength features of a point source canleadtoaResonantfrequencyshiftaslargeas0.5GHz, with a 0.13 wavelength resolution. Yang et al. [63]. studied indicating the large sensitivity of the design. the potential of quantum metamaterial for subwavelength imaging applications in the mid-infrared. FDTD numeri- 4. Stacked Metamaterial Structure cal simulations demonstrated the negative refraction along an air/metamaterial interface, and a sub-diffraction-limited Unit cells of metamaterials are usually stacked together to image of a structure that is ten times smaller than the incident make novel electromagnetic devices [1, 37–39]. Besides, it has wavelength. A novel microwave nondestructive evaluation been demonstrated that the stack of metamaterial particles sensor based on a negative index material lens constituted or particle arrays is capable of increasing the interaction of stacked SRR structures for detection of material defects between wave and matter and then enhancing the sensitivity was proposed and fabricated by Shreiber et al. [57]. A sketch of sensors [52]. Liu et al. [53] experimentally demonstrated of the sensor system is illustrated in Figure 13(a).Thelens a nanoplasmonic analogue of electromagnetically induced is made up of a stack of double split resonators. Simulation transparency using a stacked optical metamaterials. Results models of the 1D and 2D lenses are shown in Figures 13(b) show that plasmonic structures enable large field strengths and 13(c). The author investigated the performance of the within small mode volumes, and this may pave the way metamaterial lenses for the detection of material defects towards ultracompact sensors with extremely high sensitivity. which were small relative to wavelength. In the sensor system, Seo et al. [54] reported the experimental observations of thesampleispositionedatthefocusspotofthelens.The trappedmoderesonancesinstackedsymmetricelectric defect in the sample will induce some significant changes ring resonators separated by dielectric inserts. This opens in the energy density detected by the monopole. Ideally, the an alternative way of forming sharp resonances in a sym- same monopole could be used to detect the reflected signal, metric metamaterial structure which may have potential but because the metamaterial is lossy, an additional monopole applications in sensor design. Han et al. [55]studiedthe was placed on the other side of the lens to detect the reflected transmission properties of stacked SRR arrays, and sharp signal from the sample. Experimental results show that both resonance coupled with field concentration phenomenon, the 1D and 2D lenses are capable of detecting a 3 mm (0.037 𝜆) which satisfy the requirement of an effective senor were diameter through a hole in a fiberglass material sample based observed. Aylo et al. [56] investigated the transmission and on analyzing the receiving power of the monopole, but the reflection spectra of periodic and random stacks comprising image obtained with the 2D lens is much sharper. Figure 14 positive index materials and negative index metamaterials showsacomparisonofthefocusingspotsizeobtainedforthe International Journal of Antennas and Propagation 9

0 0 −10 −10 −20 −20 −30 −30 −40 −40

Transmission (dB) Transmission Exp. CST (dB) Transmission −50 −50 7 8 9 10 11 12 13 14 15 16 7 8 9 10111213141516 Frequency (GHz) Frequency (GHz) Air FR4 Air FR4 LDPE Silicon LDPE Silicon (a) (b) Figure 12: Simulation ((a) solid lines) and experimental results ((b) dotted lines) regarding the transmission of the slab pair structure enclos- ing several materials with a thickness of 0.5 mm [46].

Translation axis Hole Lens

Emitting monopole Sample

Amplifier

Network analyzer Receiving monopole (a) (b)

(c)

Figure 13: (a) Sketch of the nondestructive evaluation sensor system. (b) 2D unit cell models. (c) 1D unit cell models [57].

1D and 2D lenses. The 1D lens yielded a focus spot size of 0.7 𝜆, 5. Sensors Based on a Single whereas the focus spot for the 2D lens is about 0.48 𝜆.On Metamaterial Particle the other hand, due to the influence of metamaterial loss, the transmittanceofthe2Dlensismuchlowerthanthatofthe1D In a pioneering work, Fedotov et al. [64] demonstrated that one, and then the 1D lens allows for a longer sample standoff the “trapped mode” resonances could be excited by crossing distance and higher transmission. Therefore, the choice of the symmetry of double split resonators. Metamaterials of the lens to be used in a sensor is prescribed by the specific this type are known to show Fano antisymmetric resonant requirements of the testing system. lines in transmission with enhanced local fields that appear 10 International Journal of Antennas and Propagation

1 Besides, we investigated the influence of shape and asymmet- 0.9 ric parameters of the aDSR on the performance of the sensor 𝑄 0.8 [26]. Results show that the spectral response and the factor of the sensor can be flexibly tailored to design requirements 0.7 by adjusting the asymmetry parameter or the topological 0.6 structureoftheresonator.Recently,wereportedanoptimized 0.5 design concept for a microwave resonant sensor based on 0.4 stereocomplementary asymmetric split resonator (CASR) [27]. In terms of its special structure characteristic, the stereo- 0.3 Normalized power Normalized CASRcanbeperforatedonacoppersheet,andthusthe 0.2 influence of the substrate losses on 𝑄 factor can be eliminated. 0.1 Figures 17(a) and 17(b) show the electric field distribution 0 and the corresponding power flow on the surface of the 0 50 100 150 200 resonator at the resonant state. We supposed that the sample Position (mm) substance was deposited in the low gap of the structure and the spectrum was simulated, as shown in Figure 17(c). 1D It is seen that the spectrum moves to the lower frequency 2D side with the increase of sample permittivity. The average Figure 14: Power scans at focus spot distance for 1D lens and 2D frequency shift with respect to a small change of 0.1 in sample lens [57]. permittivity is about 123 MHz. The relation between resonant frequency and sample permittivity plotted in Figure 17(d) reveals a good linearity of the sensor. Moreover, the simplicity “trapped” in the gap of the unit cells. Based on this charac- of such sensing structure enables its utilization in a wide teristic, a metamaterial particle-assisted thin film sensor was frequency range by simple rescaling, which opens up avenues proposed and demonstrated experimentally by Al-Naib et al. toward flexible designing of sensors with superior sensitivity. [58]. It consists of a single mode rectangular waveguide and a single asymmetry double split resonator (aDSR), as shown in 6. The Other Kinds of Metamaterial Sensors Figure 15(a). In this device, the excitation of “trapped modes” leads to an extremely sharp resonance. Then, at the resonant The other kinds of metamaterial sensors include sensors state,theelectricfieldconcentratesinthegapofthering(see basedonthesqueezingandtunnelingeffect,combined Figure 15(c)), and this region becomes sensitive in dielectric right/left-handed metamaterial transmission lines (CRLH) environment. To evaluate the performance of the sensor, the [16], metamaterial probe [67],andopenresonator[28]. The aDSR was coated with a 17.8 𝜇mthicklayerofphotoresistin CRLH sensor is a popular concept for the microwave region. four degrees of coverage, and the transmission coefficient was It utilizes metamaterial transmission line unit cells as building measured as shown in Figure 16.ThecircularaDSRfeatures blocks and can be exploited to allow for different sensor a resonance shift of 9, 24, and 48 MHz for the single-covered applications, such as the measurement of mass flow [68], square, the two covered squares, and the full coverage, respec- level [69], permittivity [70], strain [71], and the velocity tively, demonstrating the enhanced sensitivity of the sensor. of granular materials on belt conveyor systems [72]. More Since only a single aDSR is required, the design of such a res- details relating to this kind of sensors can be found in onant sensor is quite flexible. Besides, the aDSR is located in the review article of Schueler et al. [16]. The metamaterial the waveguide and shielded against environmental influence, probe proposed by Boybay and Ramahi [67]consistsofa so that the measurement is robust and highly reproducible. rectangular waveguide with a layer of single negative or Following the work of Al-Naib et al., great efforts have double negative metamaterials covered at one end. Due to been dedicated to study the sensors based on single meta- the application of evanescent wave, it is capable of sensing material particles. For example, He et al. [65]proposedand thetargetobjectsuchascracksonaluminumplateswith demonstrated the implementation of tip-shaped split ring subwavelength resolution. The open resonator was firstly resonator in microwave thin film sensing. An evanescent proposed by Notomi [73] based on the ray theory in 2000. microwave probe composed of a single SRR excited by a It consists of two homogenous double negative metamaterial simple rectangular loop is proposed by Ren et al. [66]. (DNM) squares in the air. Later, the resonating modes in the It has the advantages of simple fabrication and low cost. open resonator were demonstrated numerically by He et al. Numerical and experimental results show that the presence [74] using the finite-difference time-domain method. Open of the single SRR enhances the evanescent field concentration resonator has been widely applied in dielectric sensing and inthecloseproximityoftheprobe,andthenthesensitivityof laser cavity. Recently, we designed and fabricated a novel open the probe is improved substantially. Our group [25]reported resonator using metamaterial transmission line medium a double negative material particle-assisted microwave sensor [28]. The open resonance effect which is generated by and demonstrated that when a pair of Ω-shaped particles is multiple negative reflections is demonstrated experimentally. locatedinarectangularwaveguide,thetransmissioncanbe Figure 18(a) shows the schematic graph of the open res- greatly enhanced, and such a novel sensor possesses much onator. The two DNM squares are colored in blue. Effective higher sensitivity than that of traditional microwave sensor. permittivity and permeability of the background double International Journal of Antennas and Propagation 11

Port 2

Port 1

(a) (b)

15 5

10 0 (V/m)

5 −5

0 −5 0 5

(c)

Figure 15: (a) Schematic view of the aDSR-assisted sensor. (b) Layout of the circular aDSR. (c) Simulated spatial field distribution for the aDSR [58].

0

−5 (dB)

−10 Simulated

−15

7.6 7.65 7.7 7.75 7.8

Without Two spots One spot All

Figure 16: Measured transmission response of the aDSR for four degrees of coverage [58].

negative media (DPM) are equivalent to those of the air. results of note voltage distribution obtained based on the A photograph of the open resonator designed based on LC software ADS are shown in Figure 18(c). The concentration network is depicted in Figure 18(b).TheunitcelloftheDPM of voltage at the junction of the DPM square confirms the region consists of four surface-mounted inductors in series open resonant characteristic of the structure. Figure 18(d) andonecapacitorinshunttotheground,whiletheunit displays the measurement results of note voltage distribution. cell of the DNM region consists of four surface-mounted Compared with Figure 18(c), the discrepancies are most capacitors in series and one inductor in shunt to the ground, likely due to tolerances and components losses introduced asshownintheamplifiedviewofFigure 18(b). Simulation in fabrication. The field concentration phenomenon observed 12 International Journal of Antennas and Propagation

(VA/m2 )

(V/m) 40000 37500 86587 32500 27500 57710 22500 36312 17500 20456 12500 7500 8707 2500 0 0

(a) (b) 1 9.8

9.6 0.8 9.4 0.6 9.2 0.4 9 0.02

Frequency (GHz) Frequency 0 0.2 8.8 − 0.02 1 1.2 1.4 1.6 1.8 0 8.6 8 8.5 9 9.5 10 1 1.2 1.4 1.6 1.8 Frequency (GHz)

Data Linear fitting (c) (d)

Figure 17: (a) Electric field distribution on the surface of the stereo-CASR at resonant state. (b) Corresponding power flow distribution. (c) Reflection spectra of the sensor for a variation of sample permittivity (𝜀𝑟). Curves from right to left correspond to 𝜀𝑟 =1, 𝜀𝑟 = 1.1, 𝜀𝑟 = 1.2, ..., 𝜀𝑟 = 1.8. (d) The relation between resonant frequency and 𝜀𝑟. The inset shows the residuals of the linear fitting curve[27]. at the resonant state reveals the highquality characteristics of along the channel. As a consequence, a small cavity region 𝜀 the metamaterial open resonator, which may provide many with permittivity cav located in the channel will perturb possibilities for designing sensors with superior performance. the tunneling effect clearly and then introduce a significant The squeezing and tunneling effect of electromagnetic shift in ENZ-related tunneling frequency. The transmission waves transmission through narrow channels filled with spectra are shown in Figure 20.Fromlefttoright,thecurves epsilon (𝜀) near zero (ENZ) metamaterials was demonstrated correspond to the materials with relative permittivity varying theoretically by Silveirinha and Engheta [75]. Since the from 1 to 2. The peak at about 1.8 GHz is a Fabry-Perot wavelength of radiation inside the ENZ metamaterials is resonance, which is strongly dependent on the length of the extremely large, waves propagate inside the ENZ material channel. Additionally, the squeezing and tunneling effect is with no relevant reflection losses at abrupt bends or junctions. independent on the shape of the waveguide. We [29]studied The theory of energy squeezing and tunneling through an the tunneling of electromagnetic field through a 3D coaxial ultranarrow rectangular metallic waveguide was experimen- waveguide channel filled with ENZ materials and showed that tally verified by Edwards et al.76 [ ] using a microwave setup the varying of sample permittivity from 1 to 3 will result in a consisting of three distinct regions in a rectangular waveguide frequency shift of 100 MHz on average. with a narrow channel. Aluetal.[` 59, 77] showed that the phenomenon of energy squeezing and tunneling through a 7. Conclusions narrow waveguide channel can be used for accurate dielectric sensing application. Figure 19 shows the simulation model With the development of metamaterial science, sensing appli- of the permittivity sensor, which consists of an ultranar- cation of metamaterials has attracted more and more atten- row rectangular channel connecting the two sections of a tion.Theamplificationofevanescentwaveattheboundary waveguide. In this way, the ENZ-related supercoupling effects between positive and negative refractive index materials not canbeachievedusingconventionalmaterialsfillinginthe only allows for subwavelength resolution in optical imaging, channel. The infinite phase velocity of the mode near its cut- but also increases the sensitivity of the planar waveguide off results in a uniform and a strongly enhanced electric field sensors and the whispering gallery mode sensors due to the International Journal of Antennas and Propagation 13

DPM cell DNM cell

DPM

DNM

Source Source

DNM

PML DPM

(a) (b) 15 1

0.8

10 0.6

0.4 5

0.2

0 0 0 5 10 15

(c) (d)

Figure 18: (a) Schematic structure of the open resonator consisting of two homogeneous negative refractive index metamaterial squares. (b) Photograph of the metamaterial open resonator based on L-C network. (a) Simulation and (b) measurement results of note voltage distribution [28].

Figure 19: Geometry model of the permittivity sensor [59]. 14 International Journal of Antennas and Propagation

Frequency (GHz) [5] S. Guenneau, C. Amra, and D. Veynante, “Transformation ther- 1 modynamics: cloaking and concentrating heat flux,” Optics Express,vol.20,no.7,pp.8207–8218,2012. 0.8 [6]J.J.Yang,M.Huang,C.F.Yang,andJ.Yu,“Reciprocalinvisibil- ity cloak based on complementary media,” European Physical Journal D, vol. 61, no. 3, pp. 731–736, 2011. 0.6 [7]B.I.Popa,L.Zigoneanu,andS.A.Cummer,“Experimental acoustic ground cloak in air,” Physical Review Letters,vol.106, 0.4 no. 25, Article ID 253901, 2011. Transmission [8] J. Yang, M. Huang, C. Yang, Z. Xiao, and J. Peng, “Metamate- 0.2 rial electromagnetic concentrators with arbitrary geometries,” Optics Express,vol.17,no.22,pp.19656–19661,2009. [9] J. B. Pendry, “Negative refraction makes a perfect lens,” Physical 0 Review Letters,vol.85,no.18,pp.3966–3969,2000. 1.2 1.4 1.6 1.8 2 [10] J. N. Grima and R. Caruana-Gauci, “Mechanical metamaterials: Frequency (GHz) materials that push back,” Nature Materials,vol.11,no.7,pp. Figure 20: Transmission coefficient for the ultranarrow channel 565–566, 2012. 𝐿 = 127 𝑏 = 2𝑎 = 101.6 𝜀=2𝜀 𝜀 =𝜀 [11] A. Grbic and G. V. Eleftheriades, “Overcoming the diffraction sensor with mm, mm, 0, ch 0, 𝑎 =𝑎/64,and𝐿 =𝐿/5[59]. limit with a planar left-handed transmission-line lens,” Physical ch cav Review Letters, vol. 92, no. 11, pp. 117403–1, 2004. [12] R. Marques,´ F. Mart´ın, and M. Sorolla, Metamaterials with Neg- ative Parameters: Theory, Design, and Microwave Applications, enhancement of the interaction between wave and matter. Wiley-Interscience, Hoboken, NJ, USA, 2008. Sensors-based metamaterial arrays and particle(s) utilize [13] Z. Jacob, L. V.Alekseyev, and E. Narimanov, “Optical hyperlens: the local field enhancement and resonant characteristics of far-field imaging beyond the diffraction limit,” Optics Express, metamaterials to achieve high sensitive detection. Besides, vol.14,no.18,pp.8247–8256,2006. simply rescaling the size of metamaterial particles allows for [14] Z. W. Liu, H. Lee, Y. Xiong, C. Sun, and X. Zhang, “Far-field the design of sensors from microwave to optics. Although optical hyperlens magnifying sub-diffraction-limited objects,” metamaterial loss cannot be eliminated, reducing the size of Science,vol.315,no.5819,p.1686,2007. metamaterial arrays and using the stereoparticles instead of [15] M. Huang and J. J. Yang, “Microwave sensor using metamateri- the planar ones are all effective measures for reducing the als,” in Wave Propagation,A.Petrin,Ed.,pp.13–36,InTechInc., impact of losses on sensor performance. We believe that with Klagenfurt, Austria, 2011. the development of the research of evanescent wave and the [16] M. Schueler, C. Mandel, M. Puentes, and R. Jakoby, “Metama- terial inspired microwave sensors,” IEEE Microwave Magazine, concomitant effects, sensors with excellent sensitivity and vol. 13, no. 2, pp. 57–68, 2012. subwavelength resolution might be brought by metamaterials [17] T. Chen, S. Li, and H. Sun, “Metamaterials application in sens- in the future. ing,” Sensors,vol.12,no.3,pp.2742–2765,2012. [18] N. Wenwei, H. Ming, X. Zhe, and Y. Jingjing, “Sensitivity en- Acknowledgments hancement in optical waveguide sensors based on TM wave and metamaterials,” in Proceedings of the 9th International Symposi- This work was supported by the National Natural Science um on Antennas Propagation and EM Theory (ISAPE ’10),pp. Foundation of China (Grant nos. 61161007 and 61261002), 697–700, Guangzhou, China, November-December 2010. the Scientific Research Fund Major Project of the Education [19] W. Niu, M. Huang, Z. Xiao, L. Zheng, and J. Yang, “Sensitivity Bureau of Yunnan Province (Grant no. ZD2011003), the enhancement in TE mode nonlinear planar optical waveguide Research Fund for the Doctoral Program of Higher Educa- sensor with metamaterial layer,” Optik, 2011. tion (Grant no. 20125301120009), and the Natural Science [20] W. Niu, M. Huang, Z. Xiao, and J. Yang, “Nonlinear planar Foundation of Yunnan Province (Grant no. 2011FB018). optical waveguide sensor loaded with metamaterials,” Optoelec- tronics and Advanced Materials, vol. 5, no. 10, pp. 1039–1045, 2011. References [21]J.J.Yang,M.Huang,J.Yu,andY.Z.Lan,“Surfacewhispering- [1] N. I. Zheludev, “What diffraction limit,” Nature Materials,vol. gallery mode,” Europhysics Letters,vol.96,no.5,ArticleID 7, no. 6, pp. 420–422, 2008. 57003, 2011. [22] W. W. Niu, M. Huang, J. Sun, J. J. Yang, J. Yang, and J. Yu, [2] V.G. Veselago, “Electrodynamics of substances with simultane- “Microdisk sensor based on double negative metamaterials,” ously negative values of sigma and mu,” Soviet Physics Uspekhi- International Journal of RF and Microwave Computer-Aided Ussr, vol. 10, no. 4, pp. 509–514, 1968. Engineering,vol.22,no.4,pp.512–521,2012. [3] D. R. Smith and N. Kroll, “Negative refractive index in left- [23] J. J. Yang, M. Huang, and J. Sun, “Double negative metamaterial handed materials,” Physical Review Letters,vol.85,no.14,pp. sensor based on microring resonator,” IEEE Sensor Journal,vol. 2933–2936, 2000. 11, no. 10, pp. 2254–2259, 2011. [4] D. Schurig, J. J. Mock, B. J. Justice et al., “Metamaterial electro- [24] M. Huang, J. Yang, S. Jun, S. Mu, and Y. Lan, “Simulation and magnetic cloak at microwave frequencies,” Science,vol.314,no. analysis of a metamaterial sensor based on a microring resona- 5801, pp. 977–980, 2006. tor,” Sensors, vol. 11, no. 6, pp. 5886–5899, 2011. International Journal of Antennas and Propagation 15

[25]M.Huang,J.Yang,J.Sun,J.Shi,andJ.Peng,“Modellingand [43] Y. C. Lai, H. C. Lee, S. W. Kuo et al., “Label-free, coupler-free, analysis of 𝜔-shaped double negative material-assisted micro- scalable and intracellular bio-imaging by multimode plasmonic wave sensor,” Journal of Infrared, Millimeter, and Terahertz resonances in split-ring resonators,” Advanced Materials,vol. Waves, vol. 30, no. 11, pp. 1131–1138, 2009. 24,no.23,pp.OP148–OP152,2012. [26] J.J.Yang,M.Huang,Z.Xiao,andJ.Peng,“Simulationandanal- [44]B.Reinhard,K.M.Schmitt,V.Wollrab,J.Neu,R.Beigang,and ysis of asymmetric metamaterial resonator-assisted microwave M. Rahm, “Metamaterial near-field sensor for deep-subwave- sensor,” Modern Physics Letters B,vol.24,no.12,pp.1207–1215, length thickness measurements and sensitive refractometry in 2010. the terahertz frequency range,” Applied Physics Letters,vol.100, [27] J.J.Yang,M.Huang,Y.Lan,andY.Li,“Microwavesensorbased no.22,pp.221101–221104,2012. on a single stereo-complementary asymmetric split resonator,” [45]J.J.Yang,M.Huang,D.Z.Chen,H.W.Ding,andF.C.Mao, International Journal of RF and Microwave Computer-Aided “Surface WGM sensor based on cylindrical dielectric wave- Engineering,vol.22,no.4,pp.545–551,2012. guide,” Laser Physics Letters,vol.10,no.1,ArticleID015901, [28]J.J.Yang,M.Huang,F.C.Mao,andY.L.Li,“Experimentalver- 2012. ification of a metamaterial open resonator,” Europhysics Letters, [46] G. Kenanakis, N. H. Shen, C. Mavidis et al., “Microwave and vol. 97, no. 3, Article ID 37008, 2012. THz sensing using slab-pair-based metamaterials,” Physica B, [29] Z. Wu, M. Huang, J. Yang, J. Peng, and R. Zong, “Electromag- vol. 407, no. 20, pp. 4070–4074, 2012. netic wave tunnelling and squeezing effects through 3D coaxial [47] J. F. O’Hara, R. Singh, I. Brener et al., “Thin-film sensing with waveguide channel filled with ENZ material,” in Proceedings planar terahertz metamaterials: sensitivity and limitations,” of the 8th International Symposium on Antennas, Propagation Optics Express,vol.16,no.3,pp.1786–1795,2008. and EM Theory (ISAPE ’08), pp. 550–553, Guangzhou, China, [48]R.Melik,E.Unal,N.K.Perkgoz,C.Puttlitz,andH.V.Demir, November 2008. “Metamaterial-based wireless strain sensors,” Applied Physics [30] N. I. Zheludev, “The road ahead for metamaterials,” Science,vol. Letters, vol. 95, no. 1, Article ID 011106, 2009. 328, no. 5978, pp. 582–583, 2010. [49] N. Papasimakis, Z. Luo, Z. X. Shen et al., “Graphene in a photon- [31] K. Tiefenthaler and W.Lukosz, “Integrated optical switches and ic metamaterial,” Optics Express,vol.18,no.8,pp.8353–8359, gas sensors,” Optics Letters,vol.9,no.4,pp.137–139,1984. 2010. [32] K. A. Remley and A. Weisshaar, “Design and analysis of a sili- [50] R. Melik, E. Unal, N. K. Perkgoz et al., “Nested metamaterials con-based antiresonant reflecting optical waveguide chemical for wireless strain sensing,” IEEE Journal on Selected Topics in sensor,” Optics Letters,vol.21,no.16,pp.1241–1243,1996. Quantum Electronics, vol. 16, no. 2, pp. 450–458, 2010. [33] R. Gravina, G. Testa, and R. Bernini, “Perfluorinated plastic [51] N. Liu, T. Weiss, M. Mesch et al., “Planar metamaterial analogue optical fiber tapers for evanescent wave sensing,” Sensors,vol. of electromagnetically induced transparency for plasmonic 9,no.12,pp.10423–10433,2009. sensing,” Nano Letters,vol.10,no.4,pp.1103–1107,2010. [34] S. K. Khijwania and B. D. Gupta, “Maximum achievable sensi- tivity of the fiber optic evanescent field absorption sensor based [52] M. Navarro-Cia, M. Aznabet, M. Beruete et al., “Stacked com- on the U-shaped probe,” Optics Communications,vol.175,no.1, plementary metasurfaces for ultraslow microwave metamateri- pp.135–137,2000. als,” Applied Physics Letters, vol. 96, no. 16, pp. 164103–164103-3, 2010. [35]C.Elosua,I.R.Matias,C.Bariain,andF.J.Arregui,“Volatile organic compound optical fiber sensors: a review,” Sensors,vol. [53] N. Liu, L. Langguth, T. Weiss et al., “Plasmonic analogue of elec- 6, no. 11, pp. 1440–1465, 2006. tromagnetically induced transparency at the Drude damping [36] S. Guo and S. Albin, “Transmission property and evanescent limit,” Nature Materials,vol.8,no.9,pp.758–762,2009. wave absorption of cladded multimode fiber tapers,” Optics [54] B.-J. Seo, K. Kim, S. G. Kim, A. Kim, H. Cho, and E. Choi, Express,vol.11,no.3,pp.215–223,2003. “Observation of trapped-modes excited in double-layered sym- [37] A. Veselov, C. Thur,¨ A. Efimov, M. Guina, H. Lemmetyinen, metric electric ring resonators,” Journal of Applied Physics,vol. andN.Tkachenko,“Aciditysensorbasedonporphyrinself- 111, no. 11, Article ID 113106, 2012. assembled monolayers covalently attached to the surfaces of [55] N. R. Han, Z. C. Chen, C. S. Lim, B. Ng, and M. H. Hon, tapered fibres,” Measurement Science and Technology,vol.21,no. “Broadband multi-layer terahertz metamaterials fabrication 11, Article ID 115205, 2010. and characterization on flexible substrates,” Optics Express,vol. [38] M. Ahmad and L. L. Hench, “Effect of taper geometries and 19, no. 8, pp. 6990–6998, 2011. launch angle on evanescent wave penetration depth in optical [56] R. Aylo, P. P. Banerjee, A. K. Ghosh, and P. Verma, “Design fibers,” Biosensors and Bioelectronics,vol.20,no.7,pp.1312–1319, of metamaterial based sensors for pressure measurement,” in 2005. Proceedings of the 9th International Symposium on Antennas [39] R. Horvath,´ H. C. Pedersen, and N. B. Larsen, “Demonstration Propagation and EM Theory (ISAPE ’10),vol.7604,pp.760412– of reverse symmetry waveguide sensing in aqueous solutions,” 760412-8, San Francisco, Calif, USA, January 2010. Applied Physics Letters,vol.81,no.12,pp.2166–2168,2002. [57] D. Shreiber, M. Gupta, and R. Cravey, “Comparative study of [40] R. Horvath,´ L. R. Lindvold, and N. B. Larsen, “Reverse-symme- 1-D and 2-D metamaterial lens for microwave nondestructive try waveguides: theory and fabrication,” Applied Physics B,vol. evaluation of dielectric materials,” Sensors and Actuators A,vol. 74, no. 4-5, pp. 383–393, 2002. 165, no. 2, pp. 256–260, 2011. [41]S.A.Taya,M.M.Shabat,andH.M.Khalil,“Nonlinearplanar [58] I. A. I. Al-Naib, C. Jansen, and M. Koch, “Thin-film sensing with asymmetrical optical waveguides for sensing applications,” planar asymmetric metamaterial resonators,” Applied Physics Optik,vol.121,no.9,pp.860–865,2010. Letters,vol.93,no.8,ArticleID083507,2008. [42]S.A.Taya,M.M.Shabat,andH.M.Khalil,“Enhancementof [59] A. Alu and N. Engheta, “Dielectric sensing in epsilon-near-zero sensitivity in optical waveguide sensors using left-handed mate- narrow waveguide channels,” Physical Review B,vol.78,no.4, rials,” Optik,vol.120,no.10,pp.504–508,2009. Article ID 045102, 2008. 16 International Journal of Antennas and Propagation

[60] Z.Wei,Y.Cao,J.Han,C.Wu,Y.Fan,andH.Li,“Broadbandneg- [77] A. Alu` and N. Engheta, “Light squeezing through arbitrarily ative refraction in stacked fishnet metamaterial,” Applied Physics shaped plasmonic channels and sharp bends,” Physical Review Letters,vol.97,no.14,ArticleID141901,2010. B,vol.78,no.3,ArticleID035440,2008. [61] C. P.Scarborough, Z. H. Jiang, D. H. Werner, C. Rivero-Baleine, and C. Drake, “Experimental demonstration of an isotropic metamaterialsuperlenswithnegativeunitypermeabilityat8. 5MHz,”Applied Physics Letters, vol. 101, no. 1, Article ID 014101, 2012. [62] K. Aydin, I. Bulu, and E. Ozbay, “Subwavelength resolution with a negative-index metamaterial superlens,” Applied Physics Let- ters, vol. 90, no. 25, Article ID 254102, 2007. [63] K. Y. Yang, V.Giannini, A. O. Bak, H. Amrania, S. A. Maier, and C. C. Phillips, “Subwavelength imaging with quantum metama- terials,” Physical Review B,vol.86,no.7,ArticleID075309,2012. [64] V.A. Fedotov, M. Rose, S. L. Prosvirnin, N. Papasimakis, and N. I. Zheludev, “Sharp trapped-mode resonances in planar meta- materials with a broken structural symmetry,” Physical Review Letters, vol. 99, no. 14, Article ID 147401, 2007. [65] X. J. He, Y. Wang, J. M. Wang, and T. L. Gui, “Thin-film sensor based tip-shaped split ring resonator metamaterial for micro- wave application,” Microsystem Technologies,vol.16,no.10,pp. 1735–1739, 2010. [66] Z. Ren, M. S. Boybay, and O. M. Ramahi, “Near-field subsurface detection using metamaterial inspired probes,” Applied Physics A,vol.103,no.3,pp.839–842,2011. [67] M. S. Boybay and O. M. Ramahi, “Near-field probes using double and single negative media,” Physical Review E,vol.79, no.1,ArticleID016602,2009. [68] A. Penirschke, M. Schussler, and R. Jakoby, “New microwave flow sensor based on a left-handed transmission line resonator,” in Proceedings of the IEEE MTT-S International Microwave Symposium (IMS ’07), pp. 393–396, June 2007. [69] M. Schussler, M. Puentes, C. Mandel, and R. Jakoby, “Capacitive level sensor for layered fillings in tanks and vessles based on metamaterial transmission line,”in Proceedings of IEEE Sensors, pp.1966–1969,2011. [70] M. Puentes, M. Schuler,¨ C. Damm, and R. Jakoby, “Extraction of capacitive profiles with a planar metamaterial sensor,” Applied Physics A,vol.103,no.3,pp.815–819,2011. [71]C.Mandel,M.Schussler,andR.Jakoby,“Awirelesspassive strain sensor,” in Proceedings of IEEE Sensors, pp. 207–210, 2011. [72] M. Puentes, B. Stelling, M. Schußler,¨ A. Penirschke, and R. Jako- by, “Planar sensor for permittivity and velocity detection based on metamaterial transmission line resonator,” in Proceedings of the 39th European Microwave Conference (EuMW ’09),pp.57– 60, October 2009. [73] M. Notomi, “Theory of light propagation in strongly modulated photonic crystals: refractionlike behavior in the vicinity of the photonic band gap,” Physical Review B,vol.62,no.16,pp.10696– 10705, 2000. [74] S. L. He, Y. Jin, Z. C. Ruan, and J. G. Kuang, “On subwavelength and open resonators involving metamaterials of negative refrac- tion index,” New Journal of Physics,vol.7,no.1,p.210,2005. [75] M. Silveirinha and N. Engheta, “Tunneling of electromagnetic energy through subwavelength channels and bends using 𝜀- near-zero materials,” Physical Review Letters,vol.97,no.15,pp. 157–403, 2006. [76] B. Edwards, A. Alu,` M. E. Young, M. Silveirinha, and N. Enghe- ta, “Experimental verification of epsilon-near-zero metamate- rial coupling and energy squeezing using a microwave waveg- uide,” Physical Review Letters, vol. 100, no. 3, Article ID 033903, 2008. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 106296, 8 pages doi:10.1155/2012/106296

Research Article EBG Size Reduction for Low Permittivity Substrates

Gonzalo Exposito-Dom´ ınguez,´ 1 JoseManuelFern´ andez-Gonz´ alez,´ 1 Pablo Padilla,2 and Manuel Sierra-Castaner˜ 1

1 Radiation Group, Department of Signals, Systems and Radio Communications, University of Madrid, 28040 Madrid, Spain 2 Department of Signal Theory, Telematics and Communications, University of Granda, Granda 18071, Spain

Correspondence should be addressed to Gonzalo Exposito-Dom´ ´ınguez, [email protected]

Received 3 September 2012; Accepted 13 December 2012

Academic Editor: Eric Lheurette

Copyright © 2012 Gonzalo Exposito-Dom´ ´ınguez 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.

Double layer and edge-location via techniques are combined for electromagnetic band gap (EBG) size reduction. The study of the required number of elements and their dimensions is carried out in order to suppress the surface wave propagation modes and consequently to reduce the mutual coupling between radiating elements in low-permittivity substrates. By applying these techniques, the size of the EBG mushroom is reduced by 30%; however, the bandwidth operation maintains its value, and these structures can be integrated between radiating elements in broad bandwidth antennas.

1. Introduction Patch antennas are found to have very strong mutual coupling due to the severe surface waves on thick and high- Low-profile integrated antennas are demanded to be small permittivity substrates. In the literature, it can be found a [1]. In order to reduce the size of the antenna or to increase variety of works which apply metamaterials for the reduction the working frequency band, high-permittivity substrates are of this effect. In [7], four rows of EBG mushrooms are used. However, with this type of substrates, surface wave inserted between the patch antennas in r = 10.2 and thick- propagation modes are enhanced, and there is strong mutual ness (h = 2 mm) substrate. With this configuration, 8 dB of coupling [2]. In array antennas, the separation between mutual coupling reduction is obtained. In [8], different elements should not be higher than 0.8 λ0 in order to avoid substrates are combined; radiating elements are suspended grating lobes and to get high directivity. This last condi- over a thick foam layer in order to increase the bandwidth, tion regarding the separation between radiating elements and meanwhile, EBG structures are printed in a thin high- increases the mutual coupling because the elements are closer permittivity substrate for size reduction and surface wave and they have stronger interaction. suppression. In [9], by using edge-located vias, the size The most common mutual coupling reduction tech- of mushroom-type EBG is reduced by 20%. Among other niques are cavity structures [3], nonuniform feeding distri- strategies, in [10], a fork shape is used. The area occupied by bution [4], unequally space distribution [5], and defected the fork-like structure is less than 25% of the mushroom-like ground plane (DGP) [6], but lately, (EBG) structures are structure. Besides, microelectromechanical systems (MEMS) being used. The introduction of several rows of EBG are used, and reconfigurable stop band is obtained. Another structures between two printed antennas has been proved to studied technique is the use of metal strips. Basically, the increase the isolation [7]. However, when low-permittivity idea is to combine the EBG concept with soft surfaces. A substrates are used in order to enhance the radiation effi- comparison of mushroom-type EBG surfaces and corrugated ciency, the size of the printed antenna elements increases, and and strip-type soft surfaces is shown in [11]. This stripe-type the available space between radiating elements is not suffi- is used in [12, 13] to reduce mutual coupling. Finally, dual cient to introduce the necessary number of EBG rows. band planar soft surfaces are developed in [14]; two sizes of 2 International Journal of Antennas and Propagation

C

L

εr

Figure 1: High impedance surface and its model with parallel resonant LC circuit. The substrate is transparent in order to get better visualization of metallic vias.

W L

W W W + g ≪ λ t1 ≪ λ g t2 + t3 ≪ λ

t3 t1 ∼ λ/4 t2 (a) (b) Figure 2: Layer and top views of traditional EBG structures (a) and multilayered F structure (b). strips are mixed in order to get dual forbidden band. Thanks impedance values for vertical and horizontal propagation to its frequency selective feature, these EBG structures can modes at certain frequencies. In Figure 2, on the left hand be used as a filter. By applying three elements as a ground side, traditional mushrooms are shown, and meanwhile, on plane for a microstrip line, in [15], isolation levels higher the right hand side, a multilayered mushroom structure is than 55 dB in the first frequency band (2.1 GHz) and 40 dB presented. in the second one (2.45 GHz) are achieved. The behavior of this structure is similar to an LC The main goal of this work is to combine multilayer tech- circuit in (1). Below the resonance frequency, the surface is nique with edge-location via [16]tofindanewsolutionfor inductive, while above resonance frequency, the surface is EBG size reduction when using thick and low-permittivity capacitive. Consider substrates (in this case mushroom type [17]). By applying multilayered structure, different shapes, and edge-location √1 ω0 = . (1) via, the effective size is reduced by 30%, and EBG structures LC fit in between printed antennas. This paper is organized Nearby ω , the surface impedance (Z ) is much higher as follows. An overview of the EBG materials theory is 0 s than the impedance of free space, as (2) depicts. Therefore, provided in Section 2. Section 3 is devoted to study different no vertical or horizontal propagation modes are allowed. solutions to obtain surface wave suppression characteristics. Consider In Section 4,different EBG topologies, sizes, and number of periods are simulated and constructed in order to reduce the jωL Z = , (2) size of the mushroom to place it between radiating elements. s 1 − ω2LC Finally, in Section 5, the conclusions are drawn. where the capacitance C is provided by the proximity of the metal plates, according to (3)[19] as follows: 2. EBG Theory Fundamentals     wε (ε ff) − 2w C = 0 e cosh 1 ,(3) EBG technique appears as an application of truncated π g frequency selective surfaces (FSSs) [17]. These structures 0 consist of array patches printed in a substrate, which are and the inductance L (4) is related to the thickness of the short circuited to the ground plane with vias and can be structure, because its value is due to the length of the via as visualized as mushrooms protruding from the surface, as follows: shown in Figure 1. When the period is small compared to the wavelength of L = μ0t. (4) interest, it is possible to analyze the material as an effective medium with a surface impedance (metamaterial with elec- Analytically, it can be√ noticed that the band gap operation trical properties [18]). These “mushrooms” present very high band is proportional to L/C. Thus, for a fixed separation of International Journal of Antennas and Propagation 3

g/2 t m g/2 m m

p p x x p g/2 x

w = 3 w = 2.8 w = 2.1

rvia dvia dvia (a) Original mushroom (b) Double layer mushroom (c) Double layer and edge-location via mush- room (F shape)

Figure 3: Unit cell scheme for eigenmode solutions.

Original Brillouin diagram 13

11

9 Band gap 7 Frequency (GHz) Frequency 5 pxmp Phase/360◦

TM1 TE1 TE2 Figure 4: Brillouin diagram for original mushrooms.

the metal plates, the value of the capacitance is set, and the Therefore,fivetraceshavetobeanalyzedtodescribeallthe working bandwidth can be increased by increasing the length possible propagation modes for these boundary conditions. of the via and therefore the thickness of the substrate, which All the studied structures are developed in a low- is adequate for wideband antennas. permittivity substrate εr = 2.17 and thickness of 1.143 mm. Finally, EGB structures can be seen as bend corrugations, For the traditional mushroom, the size of the original patch where the length of the corrugation is the sum of the via and in X band (7.25–8.4 GHz) is 3 mm × 3 mm, while on the the patch multiplied by two, as it is depicted in Figure 2.The other hand, for double layer mushroom 2.8 mm × 2.8 mm band gap operation of the corrugations starts and finishes size is used. For this last configuration, two stacked substrates when the length is similar to λ/4andλ/2, respectively, which of 0.762 mm and 0.381 mm thickness are used. yields a frequency octave [20]. The electric field is described in terms of an eigenvalue equation, which is solved numerically. In Figures 4, 5,and6, 3. Surface Wave Suppression the mode solutions that satisfy the boundary conditions are shown. The abscissa value represents the wave number which Thanks to the high surface impedance, horizontal or vertical fulfils the requirements at a certain frequency. The lowest line modes are not allowed in the mushroom structures at certain is the TM mode, while the second and third lines are TE frequencies. In order to find the allowed frequencies for modes. A frequency band gap (dashed lines), in which the each wave vector, a single unit cell with periodic conditions surface does not support surface wave propagation of either is simulated. The x-axis represents the dimensions inside polarization, horizontal nor vertical, extends from the top of the unit cell in which the boundary conditions are fulfilled the TM band to the point where TE band crosses the light at certain frequencies for the transmission modes of the line. The comparison between original shape and double structure. In Figures 3(a) and 3(b), the unit cell is symmetric layer mushrooms yields the result of 2 GHz bandwidth for with respect to the via; thus, only three traces have to be the original shape and 1.2 GHz for the multilayer solution. computed. However, for the F shape case in Figure 3(c), the In this work, the combination of multilayer structure EBG structure does not have the same dimensions in x and y. [17]andedge-locationvia[16] for mushroom size reduction 4 International Journal of Antennas and Propagation

Double Brillouin diagram 25 20 15 10 Band gap

Frequency (GHz) Frequency 5 0 pxmp Phase/360◦

TM1 TE1 TE2 Figure 5: Brillouin diagram for double layer mushrooms.

F shape Brillouin diagram 25

20

15

10 Band gap

Frequency (GHz) Frequency 5

0 p xmp t Phase/360◦

TM1 TE1 TE2 Figure 6: Brillouin diagram for F shape mushrooms.

n = 7

W g W t

Figure 7: Simulation scheme for S21 analysis.

Single layer Double layer Double layer edge-location via

Figure 8: Samples of single and multilayered EBG mushrooms with different shapes and number of elements. International Journal of Antennas and Propagation 5

W = 3.6

L = 2.1

Edge location = 0.3

= WTL = 0.79 dvia 0.4 t3 = 0.254

t2 = 0.381

t1 = 0.762

∗Dimensions are expressed in mm (a) Schematic (b) Prototype

Figure 9: Multilayered mushroom with rectangular shape, 4 elements and edge-located via (F-shape).

Original transmission parameter (S21) 0 −10 −20 −30 −40

Amplitude (dB) Amplitude −50 −60 6.5 7 7.5 8 8.5 9 Frequency (GHz)

n = 7sim n = 4sim n = 7meas n = 4meas

Figure 10: Transmission parameters (S21) for original mushrooms.

is discussed. In order to keep the frequency working band scheme for S21 analysis is proposed in Figure 7. In this sim- (15%) and radiation efficiency, the same substrate (εr = 2.17 ulation, different topologies (original, double, or F shape), and thickness of 1.143 mm) is used. As long as the inductance size of the mushroom w,gapsizeg, and number of elements depends on the substrate thickness, which is fixed already, n are tested in order to obtain the surface wave suppression the only available parameter is the capacitance C.Inorderto behavior. increase this parameter, a multilayered structure with edge- In order to validate the whole process, prototypes of four location via is presented. The dimensions of the patches are and seven rows are built. Six prototype transmission lines 2.1 mm × 3.6 mm (this value means 30% size reduction in (TLs) with EBG ground plane are mounted in Figure 8.On the critical direction). The Brillouin diagram for this config- the left hand side, the two circuits are single layered, the two uration is presented in Figure 6. In this case, the bandwidth circuits in the middle are double layered, and the last two keeps its value, but significant size reduction is noticed. circuits on the right hand side combine double layer with edge-location via (F shape). = 4. Mutual Coupling Reduction All the substrates used have permittivity of εr 2.17. However, four different thickness values are used. The TLs The size of the mushroom patches and the necessary number impedance is 50 Ω, and those TLs are printed in a 0.254 mm of periods for mutual coupling reduction are higher than thick substrate. Mushroom in single layer case are printed the available space between radiating elements for low- in a 1.143 mm thick substrate, while in double layer case permittivity substrates. For example, for steering arrays is printed in 0.762 mm (bottom layer) and 0.381 mm thick antennas, separation must not be higher than 0.6 λ0 in order (upper layer) substrates. Therefore, total thickness value to avoid grating lobes. To study this effect, a simulation maintains its value as it can be seen in Figures 9(a) and 9(b). 6 International Journal of Antennas and Propagation

Transmission parameter (S21) 0

−10

−20

−30

Amplitude (dB) Amplitude −40

−50 6.5 7 7.5 8 8.5 9 Frequency (GHz)

n = 7sim n = 4sim n = 7meas n = 4meas

Figure 11: Transmission parameters (S21) for double layer mushrooms.

S parameters array 2 × 1 0

−10

−20

−30

Amplitude (dB) Amplitude −40

−50 6.5 7 7.5 8 8.5 9 Frequency (GHz)

n = 7sim n = 4sim n = 7meas n = 4meas

Figure 12: Transmission parameters (S21) for F shape mushrooms.

19000 19000 11394 11394 7325 7325 4686 4686 2973 2973 1862 1862 V/m 1141 V/m 1141 674 674 370 370 174 174 0 0

y y

z x z x

(a) Without EBG structures (b) With EBG structures

Figure 13: |E| field simulation of two round patches with dual circular polarization. International Journal of Antennas and Propagation 7

In Figures 10, 11,and12, comparisons between the sim- [2] J. R. James and P. S. Hall, Handbook of Microstrip Antennas, ulated structures and the measured prototypes are shown. IEEE Waves Series, 1989. Difficulties due to the small size of the circuits (size circuit [3] R. J. Mailloux, “On the use of metallized cavities in printed approximately 1 cm) add some differences between simula- slot arrays with dielectric substrates,” IEEE Transactions on tions and measurements. However, the overall behavior of Antennas and Propagation, vol. AP-35, no. 5, pp. 477–487, an LC filter in a certain frequency band is observed. It can 1987. [4]C.A.Balanis,Antenna Theory, Analysis and Design,Wiley,3rd be noticed, comparing the figures, that traditional mush- edition, 1997. rooms have larger frequency operation band than F shape [5] B. Preetham Kumar, “Design of unequally spaced arrays for mushrooms. However, F shape ones keep fulfilling the band- performance improvement,” IEEE Transactions on Antennas width and isolation requirements. and Propagation, vol. 47, no. 3, pp. 511–523, 1999. Finally, the tradeoff solution between isolation and avail- [6] M. Salehi, A. Motevasselian, A. Tavakoli, and T. Heidari, able space is carried out, being the necessary number of peri- “Mutual coupling reduction of microstrip antennas using odsforsurfacewavesuppression:n = 4. In Figures 9(a) and defected ground structure,” in Proceedings of the 10th IEEE Sin- 9(b), the chosen topology is shown. With final dimensions gapore International Conference on Communications Systems of 2.1 × 3.6 mm, four elements, double-layered structure (ICCS ’06), Singapore, November 2006. and edge-location via, fulfil the requirements of available [7] F. Yang and Y. Rahmat-Samii, “Microstrip antennas integrated with Electromagnetic Band-Gap (EBG) structures: a low space (10.2 mm) between 2 printed antennas separated 0.6 λ0 in ε = 2.17 and 1.143 mm thick substrate. A bandwidth mutual coupling design for array applications,” IEEE Trans- r actions on Antennas and Propagation, vol. 51, no. 10, pp. 2936– operation in X band, from 7.1 GHz to 8.2 GHz (15%) and 2946, 2003. 10 dB of isolation, is obtained. [8] E. Rajo-Iglesias, O. Quevedo-Teruel, and L. Inclan-S´ anchez,´ In order to prove this solutions four rows of double- “Mutual coupling reduction in patch antenna arrays by using layered edge-located via, EBG mushrooms are introduced a planar EBG structure and a multilayer dielectric substrate,” between two round patches with double circular polariza- IEEE Transactions on Antennas and Propagation, vol. 56, no. 6, tion. The radiating elements are integrated in the same pp. 1648–1655, 2008. substrate, and they are circular polarized. The elements are [9] E. Rajo-Iglesias, L. Inclan-Sanchez, J. L. Vazquez-Roy, and fed by a 90◦/3 dB branch line coupler in order to get the E. Garcia- Munoz,˜ “Size reduction of mushroom-type EBG double circular polarization. surfaces by using edge-located vias,” Microwave and Wireless In Figure 13(a), |E| field simulation for the two patches Components Letters, IEEE, vol. 17, no. 9, pp. 670–672, 2007. is shown for left-handed circular polarization (LHCP). In [10] L. Yang, M. Fan, F. Chen, J. She, and Z. Feng, “A novel compact electromagnetic-bandgap (EBG) structure and its applications Figure 13(b), it can be seen graphically how |E| field value for microwave circuits,” IEEE Transactions on Microwave decays quicker when using double layer edge-location via Theory and Techniques, vol. 53, no. 1, pp. 183–189, 2005. EBG structures. [11] E. Rajo-Iglesias, M. Caiazzo, L. Inclan-Sanchez, and P.-S. Kildal, “Comparison of bandgaps of mushroom-type EBG 5. Conclusion surface and corrugated and strip-type soft surfaces,” Micro- waves, Antennas Propagation, IET, vol. 1, no. 1, pp. 184–189, This paper presents and proposes a combination of double 2007. layer and edge-location via techniques for EBG size reduc- [12] E. Rajo-Iglesias, O. Quevedo-Teruel, and L. Inclan-S´ anchez,´ tion. With these techniques, a 30% of size reduction in com- “Planar soft surfaces and their application to mutual coupling parison to the original mushroom shapes is achieved. This F reduction,” IEEE Transactions on Antennas and Propagation, shape mushrooms maintain the broad bandwidth operation, vol. 57, no. 12, pp. 3852–3859, 2009. and these are used in order to suppress the surface wave [13] S. Quevedo-Teruel, L. Inclan-Sanchez, and E. Rajo-Iglesias, modes and consequently to reduce the mutual coupling “Soft surfaces for reducing mutual coupling between loaded between radiating elements in low-permittivity substrates. PIFA antennas,” IEEE Antennas and Wireless Propagation Letters, vol. 9, pp. 91–94, 2010. [14] E. Rajo-Iglesias, J. L. Vazquez-Roy,´ O. Quevedo-Teruel, and Acknowledgments L. Inclan-S´ anchez,´ “Dual band planar soft surfaces,” IET The simulations contained in this work have been carried out Microwaves, Antennas and Propagation, vol. 3, no. 5, pp. 742– using CST Microwave Studio Suite 2011 under a cooperation 748, 2009. agreement between Computer Simulation Technology (CST) [15] L. Inclan-Sanchez, J. L. Vaquez-Roy, and E. Rajo-Iglesias, “High isolation proximity coupled multilayer patch antenna ´ and Universidad Politecnica de Madrid. The authors kindly for dual-frequency operation,” IEEE Transactions on Antennas thank the company NELTEC S.A. for giving samples of the and Propagation, vol. 56, no. 4, pp. 1180–1183, 2008. substrates, in which the prototypes were built. This work has [16] F. Yang and Y. Rahmat-Samii, “Polarization-dependent elec- been supported by a UPM Grant no. CH/003/2011 and the tromagnetic band gap (PDEBG) structures: designs and sicomoro project with Reference no. TEC2011-28789-C02- applications,” Microwave and Optical Technology Letters, vol. 01. 41, no. 6, pp. 439–444, 2004. [17] D. Sievenpiper, L. Zhang, R. F. Jimenez Broas, N. G. References Alexopolous,¨ and E. Yablonovitch, “High-impedance electro- magnetic surfaces with a forbidden frequency band,” IEEE [1] M. A. Salas-Natera, A. Garc´ıa-Aguilar, J. Mora-Cuevas et al., Transactions on Microwave Theory and Techniques, vol. 47, no. Satellite Communications, InTech, 2011. 11, pp. 2059–2074, 1999. 8 International Journal of Antennas and Propagation

[18] C. Caloz and T. Itoh, Electromagnetic Metamaterials: Transmis- sion Line Theory and Microwave Applications: The Engineering Approach, John Wiley and Sons, 2006. [19] F. Yang and Y. Rahmat-Samii, Electromagnetic Band Gap Structures in Antenna Engineering, The Cambridge RF and Microwave Engineering Series, 2008. [20] P.-S. Kildal, “Definition of artificially soft and hard surfaces for electromagnetic waves,” Electronics Letters, vol. 24, no. 3, pp. 168–170, 1988. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 856476, 6 pages doi:10.1155/2012/856476

Research Article Mie Scattering by a Conducting Sphere Coated Uniaxial Single-Negative Medium

You-Lin Geng

Institute of Antenna and Microwaves, Hangzhou Dianzi University, Xiasha, Hangzhou, Zhejiang 310018, China

Correspondence should be addressed to You-Lin Geng, [email protected]

Received 24 June 2012; Accepted 4 November 2012

Academic Editor: James R. Kelly

Copyright © 2012 You-Lin Geng. 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.

We propose an accurate analytical method to compute the electromagnetic scattering from three-dimensional (3D) conducting sphere coated uniaxial anisotropic single-negative (SNG) medium. Based on the spherical vector wave functions (SVWFs) in uniaxial anisotropic medium, the electromagnetic field in homogeneous uniaxial SNG medium and free space can be expressed by the SVWFs in uniaxial SNG medium and free space. The continued boundary conditions of electromagnetic fields between the uniaxial SNG medium and free space are applied, and the tangential electrical field is vanished in the surface of conducting sphere, the coefficients of scattering fields in free space can be derived, and then the character of scattering of conducting sphere coated homogeneous uniaxial SNG medium can be obtained. Some numericals are given in the end.

1. Introduction media [12–15], where both the permittivity and permeability are negative. Recently, the development of metamaterials, which was first However, “single-negative” (SNG) materials in which introduced by Veselago in 1968 [1], has attracted more and only one of the material parameters, not both, has a negative more attention in the physical and engineering communities. real value may also possess interesting properties when they Metamaterials, which are a kind of new artificial media, have are paired in a conjugate manner. These media include the been studied intensively since the concept of perfect lenses epsilon-negative (ENG) media, in which the real part of was proposed by Pendry in 2000 [2] and the negative index of permittivity is negative, but the real permeability is positive, refraction was experimentally verified by Smith et al. in 2001 and the mu-negative (MNG) media, in which the real [3]. Another breakthrough in metamaterials happened in part of permeability is negative, but the real permittivity is 2006, when a method for controlling electromagnetic (EM) positive. For instance, the idea of constructing an effective fields was proposed using inhomogeneous and anisotropic metamaterials by having layers of SNG media has been metamaterials [4] and the reduced invisible cloak was explored by Fredkin and Ron in [16], and some works successfully realized using metamaterials at microwave fre- have been done by AluandEngheta[` 17–19]. It is well quencies [5]. Since then, more and more researchers have known that the negative medium is not isotropic; it is investigated the properties and applications of anisotropic anisotropy; some works have been done between acoustic or metamaterials. electromagnetic waves and anisotropic metamaterials [20– The EM scattering problem for penetrable objects is very 22], and this paper discusses the scattering by a conducting important to study material properties and arbitrary dielec- sphere coated uniaxial anisotropic SNG medium. tric lenses. There are many classical methods for analysing Based on the spherical vector wave functions in uniaxial traditional two-dimensional (2D) and three-dimensional sphere [11] and uniaxial sphere of left-handed material [13], (3D) dielectric scattering problems [6–11]. In recent years, electromagnetic fields in homogeneous uniaxial anisotropic there have also been many theoretical investigations on single-negative medium can be expressed as the addition the study of EM characteristics for double-negative (DNG) of the first and second SVWFs in uniaxial SNG medium. 2 International Journal of Antennas and Propagation

z Someworksweredoneaboutthisnegativemediumin k 0 the past years [21, 22], and in this paper, either of the per- ɛ0μ0 Hi mittivity or permeability is negative; that is, the permittivi- Ei ty  is negative while the permeability μ is positive, or on the ɛμ contrary, the permeability μ is negative while the permittivity a  is positive. Using Fourier transform [6, 11], the expansion a 2 1 of plane wave factors in terms of spherical vector wave functions in isotropic medium [23], and the properties of y spherical Bessel functions [24], the electromagnetic fields in Region 2 the uniaxial anisotropic SNG medium can be obtained as follows:

Region 1 x 2 2 π  (l) e (l) Region 0 E = Fmnq Amnq(θk)Mmn r, kq l=1 q=1 mnn 0 Figure 1: Geometry of a plane wave scattering by a plasma  e (l) anisotropic spherical shell. + Bmnq(θ )Nmn r, kq k (3)  e (l) + Cmnq(θk)Lmn r, kq Applying the continued boundary conditions of electro- magnetic fields, and the tangential electric field vanishing m 2 × Pn (cos θ )kq sin θ dθ , in the interface of the conducting sphere, the scattering k k k ffi coe cients of scattering fields in free space can be derived. 2 2 π  The numerical results between the present method in this (l) h (l) H = Fmnq Amnq(θk)Mmn r, kq paper and Mie theory of scattering by a conducting sphere l=1 q=1 mnn 0 coated isotropic SNG medium are given, a good agreement  is obtained as we expected, and some new numerical results Bh θ (l) + mnq( k)Nmn r, kq (4) are given in this paper.  In the subsequent analysis, an exp(−iωt)timedepen- h (l) + Cmnq(θ )Lmn r, kq dence is assumed for electromagnetic field quantities but k suppressed throughout the treatment. m 2 × Pn (cos θk)kq sin θkdθk, 2. Formulas n n ∞ m Let us consider a conducting sphere coated homogeneous where and are summed up both from 0 to + while uniaxial SNG medium illuminated by an incident plane issummedupfrom−n to n,andr is pointing to the (θ, φ)- (l) wave. As illuminated in Figure 1, the coated sphere with direction in the spherical coordinates. The coefficients, Fmnq, p p p outer radius a1 and inner radius a2 is located at the are unknown, as in [11, 13]. Amnq(θk), Bmnq(θk), Cmnq(θk) coordinate origin. On the surface of the inner conducting (where p = e or h), and kq are functions of θk and they have sphere, the uniaxial SNG medium with permittivity tensor (l) d = been derived in [11, 13]. The vector wave functions, Mmn, ( ) and permeability tensor (μ) is coated with thickness ( (l) (l) Nmn,andLmn are spherical vector wave functions and they a1 − a2). It is assumed that the incident wave propagates in the +z direction, the incident electric field has unity of are also shown in [11, 23]. amplitude, and is polarized in the +x direction. The incident electromagnetic fields (designated by the The electric field vector wave equation in such a source- superscript inc) can be expanded in an infinite series in free uniaxial anisotropic SNG medium can be written in the isotropic spherical vector wave functions [11, 23]: following form [6, 11]: −1 2 ∇× µ ·∇×E(r) − ω  · E(r) = 0, (1) inc x (1) x (1) E = amnMmn(r, k0) + bmnNmn(r, k0) mn (5) where E denotes the electric field, while  and μ represent the   permittivity tensor and the permeability tensor of uniaxial × δm,1 + δm,−1 , SNG medium; the expression are [11, 13] inc k0 x (1) x (1) ⎡ ⎤ ⎡ ⎤ H = amnNmn(r, k0) + bmnMmn(r, k0) t 00 μt 00 iωμ0 mn ⎢ ⎥ ⎢ ⎥ (6)  = ⎣ 0 t 0 ⎦, µ = ⎣ 0 μt 0 ⎦. (2)   00z 00μz × δm,1 + δm,−1 , International Journal of Antennas and Propagation 3

where the expansion coefficients are defined as: and r = a1 and can be obtained the following expression:

2 2 ∞ π ⎧ F(l) Q(l) Pm θ 2 θ dθ n mnq mnq n (cos k)kq sin k k ⎪ n 2 +1  ⎪ +1 m = l=1 q=1 n =0 0 ⎨i n n , 1, (14) x = 2 ( +1) amn ⎪ (7)   i ⎪ n 2n +1 = δ δ x ⎩i +1 , m =−1; m,1 + m,−1 amn 2 , 2 (k0a1) ⎧ n 2 2 ∞ π ⎪ n 2 +1 l ⎨⎪i +1 , m = 1, F( ) R(l) Pm θ 2 θ dθ 2n(n +1) mnq mnq n (cos k)kq sin k k x  bmn = (8) l=1 q=1 n =0 0 ⎪ n (15) ⎩⎪ n+1 2 +1 −i , m =−1;   2 x i = δm,1 + δm,−1 bmn ,  (k a )2 1, s = l, 0 1 δ = s,l (9) x x 0, s =/ l. where expansion coefficients amn and bmn can be expressed (l) in (7)and(8), and zn (where l = 1, 2, 3, and 4) denotes an appropriate kind of spherical Bessel functions, jn, yn, (1) (2) (l) (l) According to the radiation condition of an outgoing wave hn ,andhn ,respectively.Qmnq and Rmnq have the following (attenuating to zero at infinity) and the asymptotic behavior expression: (1) of spherical Bessel functions, only hn should be retained ⎧ ⎨ d   in the radial functions; therefore, the expansion of scattered Q(l) = Ae 1 rh(1) r (l) r s mnq ⎩ mnq n (k0 ) zn kq fields (designated by the superscript )are k0r dr ⎡  ω d − i μ0 ⎣Bh 1 r (l) r s = As (3) Bs (3) mnq r dr zn kq (16) E mnMmn(r, k0) + mnNmn(r, k0) , (10) k0 kq mn  ⎤ ⎫ (l) ⎬ zn kqr s = k0 As (3) Bs (3) h ⎦ (1) H mnNmn(r, k0) + mnMmn(r, k0) , (11) + Cmnq · hn (k0r) , iωμ0 mn r ⎭ r=a1  ω d   R(l) = i μ0 Ah 1 rh(1) r (l) r ffi As Bs n mnq mnq n (k0 ) zn kq where the coe cients, mn and mn ( varies from 0 to k0 k0r dr +∞ while m changes from −n to n), are unknowns to be ⎡ ⎤ (l) (l)  (l) mn mn d zn kqr determined, M (r, k0)andN (r, k0) denote the spherical ⎣ e 1 (l) e ⎦ − Bmnq rzn kqr +Cmnq (17) vector wave functions defined in [11, 23], and kqr dr r = ω 1/2 k0 ( 0μ0) identifies the wave number of free space, respectively. · h(1) r . Applying the boundary conditions at the surface of n (k0 ) r=a1 uniaxial anisotropic SNG medium, for example, when r = a2, the expansion coefficients of electromagnetic fields The integral in (12)to(15) can be easily calculated by in uniaxial anisotropic medium can be obtained by the Gauss integral [24], and from (12)to(15), it shows that following equations: (l) (i) firstly, the unknown coefficients (Fmnq)ofelec- tromagnetic fields in the uniaxial anisotropic SNG 2 2 ∞ π  medium can be obtained; F(l) Ae (l) s s mnq mnqzn kqa2 (ii) secondly, the coefficients (Amn, Bmn)ofscattered l=1 q=1 n=0 0 (12) fields in region 0 (free space) are calculated; and m 2 × Pn (cos θk)kq sin θkdθk = 0, (iii) lastly, the far scattering field of electromagnetic fields ⎧ from a conductor sphere coated uniaxial anisotropic

2 2 ∞ π ⎨ d  SNG medium by a plane wave and the radar cross- (l) e 1 (l) Fmnq Bmnq rzn kqr section are thus obtained. ⎩ qr dr l=1 q=1 n=0 0 k  ⎫ (l) n qr ⎬ (13) 3. Numerical Results and Discussion e z k + Cmnq r ⎭ In the last section, we have presented the necessary theoreti- r=a 2 cal formulation of the electromagnetic fields of a plane wave m 2 × Pn (cos θk)kq sin θkdθk = 0 scattered by a conducting sphere coated uniaxial anisotropic 4 International Journal of Antennas and Propagation

15 30 ɛt =−2ɛ0, ɛz =−3ɛ0, μt = 3μ0, μz = 2μ0 ɛt = ɛz = 2ɛ0, μt = μz =−2μ0 k0a1 = 2.1π, k0a2 = 2π k0a1 = 1.1π, k0a2 = π 10 20 (dB) (dB) 2 2 λ / λ / σ

σ 5 10 E plane H plane 0 0

0 90 180 0 45 90 135 180 Scattering angle (deg) Scattering angle (deg) Figure 2: Radar cross sections (RCSs) versus scattering angle θ (in E plane degrees): results of this paper (solid curve) and of Mie theory (black H plane round). Figure 3: Radar cross-sections (RCSs) versus scattering angle θ (in degrees) in the E-plane (solid curve) and in the H-plane (short- dashed curve).

SNG medium. Togain more physics insight into the problem, wewillprovideinthissectionsomenumericalsolutionsto the problem of electromagnetic scattering by a conducting 30 ɛ = i ɛ ɛ = i ɛ sphere coated uniaxial anisotropic SNG medium. t (1+0.001 ) 0 , z (2 + 0.001 ) 0 , μ = (−2+0.001i)μ , μz = (−1+0.001i)μ Numerical computations have been performed by apply- t 0 0 k a = k a = ing the theoretical formulae derived earlier in the previous 0 1 3π, 0 2 2π sections. Since there are no numerical results of scattering by a conducting sphere coated uniaxial anisotropic SNG 20 medium, in order to check the accuracy of the newly (dB) 2

obtained numerical results, we performed one trial; that λ / is, we calculated the radar cross-sections using the present σ method and the Mie theory (scattering by a conducting sphere coated homogeneous isotropic SNG medium). The 10 results are shown in Figure 2, where electric dimensions of outer and inner spherical surfaces are k0a1 = 1.1π and k0a2 = π , while the permittivity and permeability tensor elements 0 45 90 135 180 = = = =− are t z 2 0, μt μz 2μ0, respectively, (where and Scattering angle (deg) subsequently, 0 and μ0 stand for the free space permittivity E plane and permeability, resp.). H From Figure 2, it is seen apparently that the radar cross plane sections calculated by using the two methods (i.e., the present Figure 4: Radar cross sections (RCSs) versus scattering angle θ (in method in this paper and Mie theory) are in very good degrees) in the E-plane (solid curve) and in the H-plane (short- agreement in both the E-andH-planes, where the maximum dashed curve). number of n used in (12)to(15) is only 6 to achieve the convergence. It partially verifies the correctness and applicability of our theory as well as the Fortran program  codes. k0a1 = 2.1π and k0a2 = 2π. The maximum number n in After this, we obtain some new results unavailable (12)to(15)toachieveagoodconvergenceisfoundtobe16. elsewhere in the literature. Three examples are considered Figure 4 gives a numerical result of scattering by a herein, and their radar cross sections are plotted in Figures conducting sphere coated loss uniaxial SNG medium. The 3, 4,and5. electric size of the uniaxial anisotropic SNG spherical shell is Figure 3 represents radar cross sections of a conducting chosen as k0a1 = 3π and k0a2 = 2π under the illumination sphere coated uniaxial anisotropic SNG medium of more by an incident plane wave. The permittivity and permeability general uniaxial medium, where the permittivity and per- tensor parameters used for this case are t = (1 + 0.001i)0, meability tensor elements are characterized by t =−20, z = (2 + 0.001i)0,andμt = (−2+0.001i)μ0, μz = z =−30,andμt = 3μ0, μz = 2μ0, and the electric size (−1+0.001i)μ0. As the electric dimension of the sphere of the uniaxial anisotropic SNG spherical shell is chosen as is increased, the maximum number of n used in (12)to International Journal of Antennas and Propagation 5

35 and resonance region, are given and are found reducible to those of spacial cases. The present analysis is believed to be 30 useful in antenna and wave propagation.

25 Acknowledgments 20 (dB)

2 This work is partially supported by Grant no. 60971047 of λ /

σ 15 the National Natural Science Foundation of China (NSFC) and Grant no. Y1080730 of the Natural Science Foundation 10 of Zhejiang Province of China.

5 References 0 0.2 0.4 0.6 0.8 1 1.2 [1] V. G. Veselago, “The electrodynamics of substances with The thickness of coated conducting sphere simul-taneously negative values of ε and μ,” Soviet Physics Uspekhi, vol. 10, no. 4, pp. 509–514, 1968. Forward scattering [2] J. B. Pendry, “Negative refraction makes a perfect lens,” Back scattering Physical Review Letters, vol. 85, no. 18, pp. 3966–3969, 2000. Figure 5: Radar cross sections (RCSs) versus the thickness of the [3] D. R. Smith, R. A. Shelby, and S. Schultz, “Experimental uniaxial SNG medium in the forward (solid curve) and in the back verification of a negative index of refraction,” Science, vol. 292, (dashed curve) directions, when inner radius is fixed at a2 = 0.75λ. no. 5514, pp. 77–79, 2001. The parameters of the uniaxial SNG coating of the conducting [4]J.B.Pendry,D.Schurig,andD.R.Smith,“Controlling sphere are assumed to be t =−1.50, t =−30, μt = 3μ0,and electromagnetic fields,” Science, vol. 312, no. 5781, pp. 1780– μz = 1.5μ0,respectively. 1782, 2006. [5] D. Schurig, J. J. Mock, B. J. Justice et al., “Metamaterial electromagnetic cloak at microwave frequencies,” Science, vol. 314, no. 5801, pp. 977–980, 2006. (15) must be significantly increased to 22 to achieve the [6] X. B. Wu and K. Yasumoto, “Three-dimensional scattering by convergence. an infinite homogeneous anisotropic cylinder: an analytical Later in this section, monostatic RCSs are given; for solution,” Journal of Applied Physics, vol. 82, no. 1, pp. 1996– example, the forward RCS (the scattering angle θ = 0) 2003, 1997. and the back RCS (the scattering angle θ = π)aregiven [7] X. Q. Sheng and Z. Peng, “Analysis of scattering by large in Figure 5. The parameters of the coated uniaxial SNG objects with off-diagobally annisotropic material using finite medium are t =−1.50, z =−30,andμt = 3μ0, μz = element-boundary integral-multilevel fast multipole algo- 1.5μ0, and the electric size of conducting sphere is k0a2 = rithm,” IET Microwaves, Antennas and Propagation, vol. 4, pp. 1.5π, and the change of coated uniaxial SNG medium is 492–500, 2010. from 0.1λ to 1.25λ (λ is the wavelength of incident wave in [8]X.Q.ShengandC.Q.Deng,“Asimpleandefficient free space) at the interval of 0.025. From this Figure, the implementation of the well-conditioned electric-field integral forward RCS of the core-shell system almost increases as the equation,” IEEE Transactionson Antennas and Propagation, vol. 57, no. 2, pp. 582–586, 2009. thickness of the uniaxial SNG coating increases. However, [9] T. J. Cui, W. C. Chew, A. A. Aydiner, and Y. H. Zhang, “Fast- it is interesting to note that the back scattering will be forward solvers for the low-frequency detection of buried oscillating. The RCS in the forward direction is always larger dielectric objects,” IEEE Transactions on Geoscience and Remote than that in the back direction. Sensing, vol. 41, no. 9, pp. 2026–2036, 2003. [10]R.S.ChenandE.K.N.Yung,“Anefficient method to 4. Conclusions analyze the H-plane waveguide junction circulator with a ferrite sphere,” IEEE Transactions on Microwave Theory and The spherical vector wave function expansion solution to the Techniques, vol. 49, no. 5, pp. 928–937, 2001. plane wave scattering by a conducting sphere coated uniaxial [11] Y. L. Geng, X. B. Wu, L. W. Li, and B. R. Guan, “Mie scattering anisotropic SNG medium is obtained analytically in this by a uniaxial anisotropic sphere,” Physical Review E, vol. 70, no. 5, Article ID 056609, 2004. paper. The solution has only one-dimensional integral which can be calculated by Gauss integral [24]easily.Numerical [12] H. F. Ma, J. F. Zhang, X. Chen, Q. Cheng, and T. J. Cui, “CG-FFT algorithm for three-dimensional inhomogeneous results are obtained using the present method and compared and biaxial metamaterials,” Waves in Random and Complex with Mie theory and a fairly good agreement is observed. Media, vol. 19, no. 1, pp. 49–64, 2009. It is shown that the obtained solution is stable even for [13] Y. L. Geng and S. L. He, “Analytical solution for electro- almost isotropic scatterers, since the proposed solution is an magnetic scattering from a sphere of uniaxial left-handed analytical one of a conducting sphere coated uniaxial SNG material,” Journal of Zhejiang University, Science A, vol. 7, no. medium, and the result of the Mie theory is a special case of 1, pp. 99–104, 2006. the present method. The general numerical results, including [14] C. W. Qiu, H. Y. Yao, L. W. Li, S. Zouhdi, and T. S. the conducting sphere coated lossy uniaxial SNG medium Yeo, “Backward waves in magnetoelectrically chiral media: 6 International Journal of Antennas and Propagation

propagation, impedance, and negative refraction,” Physical Review B, vol. 75, no. 15, Article ID 155120, 2007. [15] M. Y. Wang, D. B. Ge, J. Xu, and J. Wu, “FDTD study on back scattering of conducting sphere coated with double- negatibe metamaterials,” International Journal of Infrared and Millimeter Waves, vol. 28, no. 2, pp. 199–206, 2007. [16] D. R. Fredkin and A. Ron, “Effective left-handed (negative index) composite material,” Applied Physics Letters, vol. 81, no. 10, pp. 1753–1755, 2002. [17] A. Alu` and N. Engheta, “Achieving transparency with plas- monic and metamaterial coatings,” Physical Review E, vol. 72, no. 1, Article ID 016623, 2005. [18] A. Alu` and N. Engheta, “Guided modes in a waveguide filled with a pair of single-negative (SNG), double-negative (DNG), and/or double-positive (DPS) layers,” IEEE Transactions on Microwave Theory and Techniques, vol. 52, no. 1, pp. 199–210, 2004. [19] A. Alu` and N. Engheta, “Polarizabilities and effective parame- ters for collections of spherical nanoparticles formed by pairs of concentric double-negative, single-negative, and/or double- positive metamaterial layers,” Journal of Applied Physics, vol. 97, no. 6, Article ID 094310, 2005. [20] J. Christensen and F. J. G. de Abajo, “Anisotropic metamateri- als for full control of acoustic waves,” Physical Review Letters, vol. 108, no. 12, Article ID 124301, 2012. [21]D.R.Smith,P.Rye,D.C.Vier,A.F.Starr,J.J.Mock,andT. Perram, “Design and measurement of anisotropic metamate- rials that exhibit negative refraction,” IEICE Transactions on Electronics, vol. 87, no. 3, pp. 359–370, 2004. [22] L. B. Hu and Z. F. Lin, “Imaging properties of uniaxially anisotropic negative refractive index materials,” Physics Letters, vol. 313, no. 4, pp. 316–324, 2003. [23] D. Sarkar and N. J. Halas, “General vector basis function solution of Maxwell’s equations,” Physical Review E, vol. 56, no. 1, pp. 1102–1112, 1997. [24] M. Abramowitz and I. A. Stegun, Handbook of Mathematical Functions With Formulas, Graphs, and Mathematical Tables, Dover Publications, New York, NY, USA, 1972. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 593498, 9 pages doi:10.1155/2012/593498

Application Article Metamaterial CRLH Antennas on Silicon Substrate for Millimeter-Wave Integrated Circuits

Gheorghe Ioan Sajin1 and Iulia Andreea Mocanu1, 2

1 National Institute for Research and Development in Microtechnologies (IMT-Bucharest) Erou Iancu Nicolae 126A, 077190 Bucharest, Romania 2 Faculty of Electronics, Telecommunications and Informations Engineering, Politehnica University Bucharest, Boulevard Iuliu Maniu 1-3, 061071 Bucharest, Romania

Correspondence should be addressed to Gheorghe Ioan Sajin, [email protected]

Received 5 September 2012; Accepted 4 November 2012

Academic Editor: James R. Kelly

Copyright © 2012 G. I. Sajin and I. A. Mocanu. 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 paper presents two composite right/left-handed (CRLH) coplanar waveguide (CPW) zeroth-order resonant (ZOR) antennas which were designed, processed, and electrically characterized for applications in the millimetric wave frequency range. Two CRLH antennas were developed for f = 27 GHz and f = 38.5, GHz, respectively. The CRLH antenna on f = 27 GHz shows a return loss of RL < −18.78 dB at f = 26.88 GHz. The −3 dB radiation characteristic beamwidth was approximately 37◦ and the gain was Gi = 2.82 dBi. The CRLH antenna on f = 38.5 GHz has a return loss of RL < −38.5dBat f = 38.82 GHz and the −3 dB radiation ◦ characteristic beamwidth of approximately 17 . The gains were Gi = 1.08 dBi at f = 38 GHz and Gi = 1.2dBiat f = 38.6GHz. The maximum measured gain was Gi = 1.75 dBi at f = 38.2 GHz. It is, upon the authors’ knowledge, the first report of millimeter wave CRLH antennas on silicon substrate in CPW technique for use in mm-wave monolithic integrated circuit.

1. Introduction This particular frequency characteristic of the CRLH TL has been exploited in the development of many types of In recent years the area of metamaterials has been getting devices such as coupled-line directional couplers, filters and a lot of attention from the scientific community. Although resonators and various types of antennas [4–14]. Veselago enunciated the theory of left-handed (LH) materials A complete description of the most practical leaky wave more than 50 years ago [1], structures mimicking these and ZOR antennas was done in [4]. properties were developed only about 10 years ago [2]. The current trend in electronics shows that there is Considering the transmission line (TL) parameters, a need of devices with reduced size planar topology and metamaterials were introduced as the concept of Composite easily integrable with active components in microwave and Right/Left-Handed (CRLH-TL). The CRLH is an artificial mm-wave integrated circuits. Although hundreds of anten- TL that can be obtained by combining the RH behavior of the nas configurations on metamaterials have been reported classical TL modeled by a transmission line loaded with series in literature, almost all are demonstrators processed on connected inductors and parallel grounded capacitors, with soft materials for frequencies up to 10 GHz–12 GHz. Very the LH behavior modeled by series connected capacitors and few antennas have been manufactured on semiconductor parallel connected grounded inductors. Such a transmission materials, [6–8, 15, 16], in the mm-wave frequency range, line exhibits both right-handed (RH) and left-handed (LH) [8, 15–17], and ever the less have been built for use in behavior. The first presentation of this type of transmission practical microwave and millimeter wave integrated circuits line and of devices made on this basis is done in [3]. [17–21]. In the quest of more efficiency when allocating the 2 International Journal of Antennas and Propagation frequency bandwidths, countries like Australia and Canada processed, and electrically measured. It is considered to be have chosen to use the 27 GHz spectrum and 38 GHz used in broadband wireless systems for the 27 GHz spectrum spectrum, respectively, for broadband wireless services and in Australia. The device was made of three resonant CRLH for a wide variety of fixed service applications. So, there is the cells made on a 5 kΩcm silicon wafer. The values of the need for using antennas designed to work on this frequencies. elements of the CRLH equivalent circuit have been computed CRLH antennas presented in this paper were processed for ZC = 50 Ω. in coplanar waveguide (CPW) configuration on silicon sub- The obtained values for CRLH elementary cell compo- strate and have been designed to be used in practical applica- nents (see Figure 1(b))are:LR = 0.6nH,LL = 0.13 nH, tions. The CPW technology was preferred due to an easier CR = 267 fF, and CL = 26 fF. For these values the computed technological approach and also due to a straightforward resonant frequency is fsh = 27 GHz. measurement procedure using an on-wafer characterization In order to obtain the antenna radiating structure in installation. It is, upon the authors’ knowledge, the first CPW configuration, the following geometrical dimensions of realization of a millimeter wave CRLH antennas on silicon CRLH cell layout (capacitors, inductors, and cell dimension) substrate for effective use in mm-wave integrated circuits. are obtained (see Figure 1(a)):

(i) for the inductive CPW stub: lL = 212 μm; wL = 2. CRLH Antenna Structure 42 μm; sL = 10 μm; (ii) for the interdigital capacitors: w = 10 μm; s = The devices cf. [15, 16] designed in this paper are zeroth- C C 5 μm; l = 250 μmandg = 65 μm; the number of order resonance (ZOR) antennas consisting of an open- C C digits was 10. ended array of CPW CRLH transmission lines cells. Each cell has a T circuit topology made of two series connected CPW The length of the CRLH cell is p = 690 μm fulfilling the interdigital capacitors and two parallel conected short-ended condition p  λg where λg = 4370 μm is the wavelength CPW transmission lines as inductive stubs. Unlike microstrip at f = 27 GHz. The surface occupied by the radiant antenna, using CPW transmission lines allows obtaining antenna structure composed of 3 CLRH cells is approx. ∼ much smaller circuit area because no large patch area and 2mm× 1.2mm = 2.4mm2. To compare, a patch microstrip no vias to the ground are needed. The elementary CRLH antenna processed on silicon using the same technology, cell used in antenna construction is depicted in Figure 1(a) working on the fundamental mode at the same frequency ∼ while in Figure 1(b) equivalent circuit of the CRLH cell is has a surface of approx. 2.2mm × 1.44 mm = 3.17 mm2. presented. One may see that the CRLH antenna allows a surface size Here 2CL and LR/2 are the equivalent capacitance and reduction of approximately 25% that is very important in the the equivalent inductance of the series capacitor, while LL is management of the mm-wave circuit planar dimensions. the equivalent parallel inductance of the two CPW inductive The antenna input is made of a feeding line of 3400 μm Ω stubs and CR is the equivalent parallel capacitance. The length and the geometry computed to match the 50 char- parallel capacitance CR includes, also, the equivalent parallel acterisctic impedance of the mm-wave circuit. This geometry capacitance to the ground of the interdigital capacitors. It allows antenna structure mounting on a dedicated test fixture is important to point out that the LR/2 and CR are strongly for radiation characteristic and gain measurement. related to 2CL and LL values. A similar CRLH topology was presented for a leaky-wave antenna in [4]. 3.2. 38.5 GHz Antenna Design. Using the same design condi- The design was made for obtaining a balanced CRLH tions as for the previous described antenna, a CPW CRLH structure for which 1/ LLCR = 1/ LRCL. The starting point ZOR antenna on the frequency f = 38.5 GHz was designed, in designing the capacitor is to choose: (i) the operating processed, and measured. This type of antenna is considered resonance frequency fsh = 1/2π LLCR; (ii) the length of to be used in the point to point systems in the 38 GHz the capacitor fingers so that the overall length of the CRLH spectrum in Canada. The elementary CRLH CPW cell has cell would be much smaller comparing to the operating the same layout as it was presented in Figure 1. The antenna wavelength; (iii) the geometry of the interdigital capacitor is formed also of three resonant CRLH cells processed on allowed by the technological limitations. the same kind of silicon wafer. The computed values of interdigital capacitors and the inductive CPW lines for a = = = 3. Antenna Design and Processing on frequency f 38.5 GHz are the following: LR 0.4nH,LL 0.08 nH, C = 210 fF, and C = 18 fF. For these values the Silicon Substrate R L resonant frequency computed with (1)give fsh = 38.8 GHz. Two devices working at the frequencies f = 27 GHz The physical dimensions necessary to obtain a CRLH and 38.5 GHz, respectively, were designed, technologically cell with the previous capacitor and inductor values are the processed, and electrically characterized. In the following, following (see Figure 1(a)): antenna at 27 GHz will be termed as Antenna #1 and antenna (i) for the inductive CPW stub: lL = 212 μm; wL = at 38.5 GHz will be termed as Antenna #2. 40 μm; sL = 8 μm;

(ii) for the interdigital capacitor: lC = 250 μm, wC = 3.1. 27 GHz Antenna Design. A CPW CRLH ZOR antenna 7 μm, sC = 8 μm, gC = 65 μm and number of digits: at the frequency f = 27 GHz was designed, technologically 10. International Journal of Antennas and Propagation 3

s L w L

gC w C

sC

2CL LR/2 2CL LR/2 lC IL

CR L p L

(a) (b)

Figure 1: CPW CRLH elementary cell used in antenna construction (a) and the equivalent circuit (b).

Also at this frequency, the length of the CRLH cell is well the removal by a standard photolithographic process of under the wavelength fulfilling the condition p = 690 μm the metal from the large areas, such as the gaps between  λg = 3110 μm. the signal line and the ground planes. For this, a mask The antennas were processed on a 400 μm thick silicon without the interdigital capacitors was designed. The second wafer (εr·Si = 11.9) with a resistivity of 5 kΩcm. On the step consisted of tracing the fine details of the interdigital silicon surface was grown through thermal oxidation a 1 μm capacitors gaps by laser ablation. = thick SiO2 layer (εr·SiO2 3.9). The metallization was A direct laser writing (DLW) method was implemented obtained by evaporation of 4.000 A˚ Au/500 A˚ Cr. By using to microprocess the Au/Cr layers evaporated on silicon. The silicon, the possibility of integration with active devices in a samples were laser ablated by tightly focusing a femtosecond functional circuit is left open. laser with 200 fs pulse duration, 775 nm wavelength, tens Two technological processing methods were tried: (i) of nanojoules pulse energy, and 2 kHz repetition rate. The standard wet etching photolithography and (ii) laser abla- 2D structures were generated according to a computer tion. controlled algorithm by precisely translating the sample with resolution below 1 μm. 3.3. Technological Approach. Standard photolithography as Figure 3(a) shows the antenna layout after photolithog- technology for antenna processing consists of photoresist raphy in the first step. There can be observed the grounded deposition, masking, exposure, developing, wet etching; lines forming the inductive stubs and the areas where the photoresist remains removal. interdigital capacitors are to be processed through laser A quarter of a silicon wafer supporting some CRLH ablation. Figure 3(b) shows the same area after the capacitors antenna structures with a Suss¨ Microtec on-wafer 67 GHz were processed through the laser ablation method. A good probe-tip contacting an antenna feeding line in order to definition of the lines could be observed: the lines were measure the resonant frequency and return loss is presented straight and the rounding at the corners was minimal. The in Figure 2(a).InFigure 2(b), the layout of a CRLH cell geometry was accurately kept. interdigital capacitors and inductor lines is presented. By applying these two steps process, we can benefit ff Concerning photolithography, this technology is able to from the fast and cost-e ective standard photolithographic provide excellent layouts for microwave devices. A drawback process which maintains good accuracy for larger geometries is the tendency of metallic lines over-etching. During our (tens or hundreds of microns) and the highly accurate but experiments there was necessary to apply a correction factor time consuming laser ablation process in order to obtain fine in the device mask design in order to keep the ratio details (micronic and submicronic geometries). 10 μm/5 μm between metallic finger width and space between However, this technology was not retained due to the the metallic fingers following drawback: during the laser ablation a cloud of In order to obtain devices by laser ablation,atwo- removed metallic material is spread all over the interdig- step technological process was applied. The first step was ital capacitor structure, including the space between the 4 International Journal of Antennas and Propagation

(a) (b)

Figure 2: Silicon wafer with antenna structures (a) and optical microscopy photo showing one of the CRLH cell (b).

(a) (b)

Figure 3: The CRLH antenna after the first step (a) and a SEM image of the CRLH cell after the second step of the laser ablation technology (b).

two digits. These metallic particles may be observed in impedance in the CPWG mode is 44.85 Ω. This difference of Figure 3(b) as well as white dots in SEM image in Figure 4. 2.27 Ω does not dramatically affect the device measurement’s This contamination dramatically increases the device’s losses. results. After obtaining the antenna structures and on wafer measurement of the resonant frequency and return loss, the silicon wafer was diced with an abrasive diamond wheel 4. Measurements and Experimental Results tool, thus obtaining separate antenna chips. These discrete structures were mounted on dedicated test fixtures in order 4.1. Measurement Techniques. As electrical characteristics of to measure antenna’s directivity characteristic and gain. Two CRLH antenna, the return loss, the resonant frequency, the such diced CRLH antennas are presented in Figure 5(a) while radiation characteristic, and the gain were measured. in Figure 5(b) two structures mounted on the test fixtures Return loss and the resonant frequency were measured can be seen. on-wafer using a 37397D vector network analyzer from It should be noted that antennas are used in applications Anritsu, equipped with PM5 set-up from Suss¨ Microtec. In requiring absence of metallic layer on the back of the Figure 2(a) it can be seen a Suss¨ Microtec on-wafer 67 GHz silicon wafer. At the same time, mounting the silicon wafer probe-tip contacting a CRLH antenna feeding line in order on the Suss¨ Microtec working chuck as well as mounting to measure the resonant frequency and return loss. the antenna structures on the test fixtures may excite the The radiation characteristics were measured with a fre- CPWG mode instead of CPW, due to the presence of metal quency generator Agilent E8257D and a spectrum analyzer underneath the antenna acting as ground plane. The line Anritsu MS2668SC, both having full capabilities in the impedance in the CPW mode is 47.12 Ω while the line CRLH antennas frequency range. The measuring planes were International Journal of Antennas and Propagation 5

ϕ

− +

θ Figure 6:Thetransverse(θ) and the longitudinal (ϕ) measurement planes.

Figure 4: Technological drawbacks: metallic particles spread in the ablated area.

Figure 7: Measuring setup for CRLH antenna: detail of the meas- uring arrangement.

The antenna gain was obtained using the De Friis relation for two identical antennas (1):

 2 Pr = 2 λ Gi , (1) Pt 4πR (a) where Pt = power transmitted by emitting antenna, Pr = power at the receiving antenna, Gi = antenna gain with respect to isotropic; λ = wavelength, R = distance between emitting and receiving antenna. In this situation the emitting and the receiving antennas are identical so that the gain of both devices are the same, Gi. The gain in (dBi) is expressed by:    = 4πR Pr G(dBi) 10 log10 . (2) (b) λ Pt

Figure 5: Discrete CRLH antenna structures after cutting (a) and The gain measurement setup shown in Figure 8 uses the after the mounting on a specialized test fixture (b). same instruments utilized for the previous characterization, the arrangement of antennas holders being adapted for using the De Friis relation (2). the transversal plane (θ) and the longitudinal plane (ϕ)as 4.2. Experimental Results for 27 GHz CRLH Antenna. The they are defined in Figure 6. simulation of reflection loss of Antenna #1 which was The measuring setup comprises, also, an antenna holder made with the IE3D Zeland software in the 24 GHz–30 GHz as is shown in Figure 7 having the possibility to rotate frequency band on the previously obtained geometry is between −90◦ and +90◦ in transverse (θ) and longitudinal presented in Figure 9. The simulation results indicate a (ϕ) planes. The emitting device was the CRLH antenna and return loss of S11 =−17.94 dB at the frequency f = as receiving device were used two frequency appropriate horn 27.1 GHz. antennas connected to the spectrum analyzer. The distance The measured results for resonance frequency and return between the emitting CRLH antenna and the aperture plane loss of CRLH Antenna #1 sample are presented in Figure 10. of the receiving horn antenna was 100 mm. The measured From Figure 10 it may be seen that the return loss and the poweratreceptionwasaveragedover50measurements. resonant frequencies are S11 =−18.78 dB at f = 26.88 GHz. 6 International Journal of Antennas and Propagation

_ Antenna 1 S11 measurement 0

−5

−10

(dB) −15 11 S −20 26.88 GHz − Figure 8: Measuring setup for CRLH antenna gain. −25 18.78 dB

−30 24 26 28 30 27 GHz antenna simulation 1 0 Frequency (GHz)

−5 dB(|S(11)|)

−10 Figure 10: Measured return loss of CRLH Antenna #1 for frequency (dB)

11 sweep between 24 GHz and 30 GHz.

S −15

−20 27.1 GHz −25 −17.94 dB 1 −30 0.8 24 26 28 30 0.6 Frequency (GHz) 0.4 dB(|S(11)|) Antenna S11 0.2 Figure 9: Simulation of return loss of the CRLH Antenna #1 in the 90 0 −90 24 and 30 GHz frequency domain. 80 −80 70 −70 60 −60 3 dB The frequency is slightly lower than the simulated one and 50 37 deg −50 the return loss is smaller as well. Also, the antenna bandwidth 40 −40 30 −30 is larger than simulated. − 20 −10 20 Radiation characteristic in transversal plane (θ)ofthe 10 0 27 GHz CRLH antenna is shown in Figure 11 where the ff Figure 11: Radiation pattern in transversal plane (θ) of the CRLH received power at di erent angles was normalized to the Antenna #1 sample at f = 26.88 GHz. maximum received power value. From Figure 11 one may see that the −3 dB beam width of the radiation characteristic is between approximately +21◦ and − 16◦ for meaning beam width of 37◦. Thesecondaryradiationlobeshaveamplitudeslower 1 than 0.3 of the main radiation lobe maximum value and 0.8 ◦ occurs at the angles of approximately ± 40 . 0.6 In order to complete the characterization of the antenna’s 0.4 radiating capability, the radiation characteristic in longitu- dinal plane (ϕ) was, also, measured. The results are shown 0.2 − in Figure 12 where the radiated power at different angles in 90 0 90 the (ϕ) plane between −70◦ and +90◦ was plotted. All the −80 80 measured power values P(ϕ) were rated to the value P(0◦) −70 70 that is the measured power in a point located on the line −60 60 = ◦ − 3 dB normal on the antenna center at ϕ 0 . 50 27 deg 50 From Figure 12 one may find that the radiation char- −40 40 −30 30 acteristic in the longitudinal plane is tilted in the forward − 20− 20 direction, the maximum radiated power occurring at an 10 0 10 angle ϕ =∼ +14◦. The radiated power decreases in the Figure 12: Radiation pattern in longitudinal plane (ϕ)ofCRLH backward direction having a secondary radiation lobe at ϕ =∼ Antenna #1 sample at f = 26.88 GHz. −35◦. International Journal of Antennas and Propagation 7

38.5 GHz antenna simulation Antenna 2 S measurement 0 0 11

−10 −5

−20 −10 (dB) (dB) 11 11

S − 41.08 GHz

S 30 36.12 GHz −20 dB −20 dB −15 −40 38.82 GHz 38.77 GHz − −17.86 dB 38.5dB −20 −50 35 40 45 35 40 45 Frequency (GHz) Frequency (GHz)

dB(|S(11)|) Antenna 2 dB(|S(11)|)

Figure 13: Return loss simulation of the CRLH Antenna #2 in the Figure 14: Return losses S11 versus frequency for two CRLH 35 GHz–45 GHz frequency domain. Antenna #2 sample.

1 The CRLH antenna gain was evaluated at f = 27 GHz, 0.8 the measured data being: λ = 11.11 mm, R = 100 mm, 0.6 power at the emitting antenna: Pt = 0.45 mW, power at the 0.4 receiving antenna Pr = 1.29E − 04 mW. By applying relation (2), the results show Gi = 2.82 dBi. It is to worth noting that, 0.2 due to the longitudinal radiation characteristic tilting in the −90 90 0 forward direction, the antenna alignment in measuring setup −80 (see Figure 8) was adjusted to make the maximum radiation 80 −70 of both antenna face each other. 70 −60 60 −50 50 − 4.3. Experimental Results for 38.5 GHz CRLH Antenna. The 40 40 −30 simulation of reflection loss of the CRLH Antenna #2 in 30 −20 −10 the 35 GHz–45 GHz frequency band was made with the CST 20 10 0 dedicated software. The simulation presented in Figure 13 indicate as working frequency f = 38.77 GHz, where the Figure 15: Radiation pattern in transversal plane (θ)forCRLH Antenna #2 sample at f = 38.82 GHz. antenna’s return loss is S11 =−17.86 dB. Measurements of the resonant frequency, return loss, radiation characteristics, and gain were made using the same setups and methods previously presented. The return Measurements were, also, made at f = 38.8 GHz in the loss and the resonant frequency were measured on-wafer longitudinal plane (ϕ) of the CRLH Antenna #2 sample. The and the radiation characteristic and the gain, by mounting results are shown in Figure 16 where the radiated power at the antenna structure on specialized test fixture, as in different angles were plotted. All the received power values Figure 5(b). were rated to the value P(0◦) meaning the measured value of The measured S11 in 35 GHz–45 GHz frequency range is the received power in a point located on the line normal on presented in Figure 14. the antenna center at ϕ = 0◦. It may be observed that the return loss is S11 =−38, 5 dB Antenna #2 gain was obtained as previously using the at resonant frequency 38.82 GHz denoting a very good De Friis relation (2), the measurement setup being the same matching of the radiating structures. Besides, values of the as in Figure 8. The gain was evaluated in the frequency S11 parameter lower than −20dBareobservableinarather domain 38 GHz-39 GHz. In this frequency domain the gain large frequency bandwidth extending from 36.12 GHz to was Gi = 1.08 dBi at 38 GHz and Gi = 1.2 dBi at 38.6 GHz. 41.08 GHz. The maximum obtained value was Gi = 1.75 dBi at f = The radiation characteristics in the transversal (θ)plane 38.2GHzwhere: λ = 7.78 mm, R = 100 mm, power at the is presented in Figure 15 where the received powers at emitting antenna: Pt = 0, 278 mW, power at the receiving different angles were normalized to the maximum power antenna Pr = 2.39E − 05 mW. Comparing with the data value in the domain θ ∈ (−90◦–+90◦). obtained for Antenna #1 (27 GHz) the matching is better, the The −3 dB beamwidth is approx. 17◦ and the side lobes directivity is higher, and the beam-width is smaller. However, appear at approx. ±50◦ with amplitudes lower by ∼6dB the measured gain of Antenna #2 is smaller than of Antenna compared to the main lobe. #1 due to the increased losses at higher frequencies in the 8 International Journal of Antennas and Propagation

1 on-wafer and the gain and the radiation characteristic were 0.8 evaluated with the antenna structures mounted on the test 0.6 fixture. The mechanics and the antenna structure contacting to the test fixture connector generate this slight frequency 0.4 displacement. However, the results to be retained are those 0.2 obtained from on-wafer measurements because the antenna 90 0 −90 will work integrated on silicon substrate and not as separately 80 −80 encapsulated device. 70 −70 60 −60 − References 50 50 −40 40 −30 [1] V. G. Veselago, “The electrodynamics of substances with 30 −20 20 10 −10 simultaneously negative values of e and μ,” Soviet Physics 0 Uspekhi, vol. 10, no. 4, p. 509, 1968. Figure 16: Radiation pattern in longitudinal plane (ϕ)ofCRLH [2]J.B.Pendry,A.J.Holden,D.J.Robbins,andW.J.Stewart, Antenna #2 sample at f = 38.8GHz. “Low frequency plasmons in thin-wire structures,” Journal of Physics Condensed Matter, vol. 10, no. 22, pp. 4785–4809, 1998. [3] C. Caloz and T. Itoh, “Invited—novel microwave devices and structures based on the transmission line approach of meta- emitting as well as in the receiving antenna on the measuring materials,” in Proceedings of the IEEE MTT-S International chain. Microwave Symposium Digest, pp. 195–198, Philadelphia, Pa, USA, June 2003. [4] C. Caloz, T. Itoh, and A. Rennings, “CRLH metamaterial 5. Conclusion and Comments leaky-wave and resonant antennas,” IEEE Antennas and Prop- Two ZOR millimeter wave CRLH CPW antennas on silicon agation Magazine, vol. 50, no. 5, pp. 25–39, 2008. ff [5]A.Sanada,M.Kimura,I.Awai,C.Caloz,andT.Itoh,“A substrate on two di erent frequencies in the mm-wave planar zeroth-order resonator antenna using a left-handed = = domain ( f 27 GHz and f 38.5 GHz) were designed, transmission line,” in Proceedings of the 34th Conference on processed, and measured as part of two integrated circuits. European Microwave Conference, pp. 1341–1344, Amsterdam, The Netherlands, October 2004. 5.1. Results for the 27 GHz CRLH CPW Antennas. The [6] S. Simion, R. Marcelli, and G. Sajin, “Small-size CPW silicon measured return-loss and resonant frequency for Antenna resonating antenna based on transmission-line meta-material approach,” Electronics Letters, vol. 43, no. 17, pp. 908–909, #1 sample were: S11 =−18.78 dB at f = 26.88 GHz demonstrating good agreement with the simulated data. The 2007. ◦ [7] S. Simion, G. Sajin, R. Marcelli, and F. Craciunoiu, “Silicon 3 dB beam width of radiation lobe is approx. 37 . Concerning resonating antenna based on CPW composite Right/Left- the gain, the values computed from the measured data using Handed transmission line,” in Proceedings of the 37th European = = De Friis relation give Gi 2.82 dBi at f 27 GHz. Microwave Conference (EUMC ’07), pp. 478–481, Munchen, Germany, October 2007. 5.2. Results for the 38.5 GHz CRLH CPW Antennas. For the [8] R. Van Dijk, A. Neto, J. A. G. Akkermans, and J. Mills, 38.5 GHz CRLH Antenna #2 sample, the measured return- “EBG-based 60 GHz on-chip antenna in passive silicon,” in Proceedings of the 38th European Microwave Conference (EuMC loss was S11 =−38.5dBat f = 38.82 GHz showing a good conformity with the simulated data and denoting, also, a very ’08), pp. 682–685, Amsterdam, The Netherlands, October 2008. good matching of the radiating structures. Return loss values − [9] C. C. Yu, M. H. Huang, L. K. Lin, and Y. T. Chang, “A compact lower than 20 dB are observable in a frequency bandwidth antenna based on metamaterial for WiMAX,” in Proceedings extending from 36.12 GHz to 41.08 GHz. The −3dBbeam ◦ of the Asia Pacific Microwave Conference (APMC ’08),Hong width of radiation lobe is approx. 17 . Kong, December 2008, Paper J2-05. The gain in the frequency domain is Gi = 1.08 dBi at [10] P. Seongmin, B. Jung-Woo, C. Sang-Hyeok, and K. Young-Sik, f = 38 GHz and Gi = 1.2dBiat f = 38.6 GHz. Also, the “A metamaterial-based symmetrical periodic antenna with gain is approx. constant in the 38 GHz–38.6 GHz frequency efficiency enhancement,” in Proceedings of the Asia Pacific domain. A maximum gain value of Gi = 1.75 dBi was Microwave Conference (APMC ’08), Hong Kong, December obtained at f = 38.2 GHz. 2008, Paper A3-49. For both the CRLH antenna samples there is a small [11] R. W. Ziolkowski, P. Jin, and C. Lin, “Electrically small frequency difference between the computed resonant fre- metamaterial-inspired antennas: the next generation,” in quency and the measured one for the best return loss in Proceedings of the 3rd International Congress on Advanced ff Electromagnetic Materials in Microwaves and Optics, pp. 44– the working frequency band. These di erences are due to 446, London, UK, August 2009. the metallic layer over etching in the device processing step. [12] A. A. Basharin and N. P. Balabha, “The radiation of antennas ff These di erences were eliminated in a second technological based on metamaterial waveguides,” in Proceedings of the 3rd run by appropriately changing the mask layout. Also, some International Congress on Advanced Electromagnetic Materials differences occur due to the fact that the return loss and res- in Microwaves and Optics, pp. 224–2226, London, UK, August onant frequency of CRLH antenna structures were measured 2009. International Journal of Antennas and Propagation 9

[13] S. Eggermont, R. Platteborze, and I. Huynen, “Analysis of radiation in a metamaterial leaky wave antenna based on complementary split ring resonator,” in Proceedings of the 3rd InternationalCongressonAdvancedElectromagneticMaterials in Microwaves and Optics, pp. 102–1104, London, UK, August 2009. [14] S. Simion, R. Marcelli, G. Bartolucci et al., “Composite Right/Left Handed (CRLH) based devices for microwave applications,” in Advanced Microwave and Millimeter Wave Technologies: Semiconductor Devices Circuits and Systems”,pp. 89–112, INTECH, Vienna, Austria, 2010. [15] A. C. Bunea, F. Craciunoiu, and Gh. Sajin, “28 GHz CRLH antenna on silicon substrate,” in Proceedings of the 41st European Microwave Conference (EuMC ’11), pp. 579–5582, Manchester, UK, October 2011. [16] G. Sajin, F. Craciunoiu, A. Dinescu, and I. Mocanu, “38 GHz metamaterial antenna on silicon substrate,” in Proceed- ings of Asia-Pacific Conference on Antennas and Propagation (APCAP ’12), Singapore, August 2012. [17] S. I. Matsuzawa, K. Sato, Y. Inoe, and T. Nomura, “Steer- able composite right/left-handed leaky wave antenna for automotive radar applications,” in Proceedings of the 36th European Microwave Conference (EuMC ’06), pp. 1155–1158, Manchester, UK, September 2006. [18] K. Mori and T. Itoh, “Distributed amplifier with CRLH- transmission line leaky wave antenna,” in Proceedings of the 38th European Microwave Conference (EuMC ’08), pp. 686– 689, Amsterdam, The Netherlands, October 2008. [19] Takahiro Kawakami and Yasushi Horii, “A compact composite right/left-handed (CRLH) leaky wave antenna composed of 2 × 2 coupled-mushrooms for broadband wireless communica- tions,” in Proceedings of the Asia-Pacific Microwave Conference (APMC ’11), pp. 674–6677, Melbourne, Australia, December 2011. [20] K. Mori and T. Itoh, “CRLH metamaterial receiving leaky wave antenna integrated with distributed amplifier,” in Proceedings of the Asia Pacific Microwave Conference (APMC ’08),Hong Kong, December 2008. [21] T. Kodera and C. Caloz, “Integrated leaky-wave antenna front-end using a ferrite-loaded open waveguide structure,” in Proceedings of the 40th European Microwave Conference (EuMC ’10), pp. 469–472, Paris, France, September 2010. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 174023, 6 pages doi:10.1155/2012/174023

Research Article Broadband Microstrip Bandpass Filter Based on Open Complementary Split Ring Resonators

P. V elez,´ J. Naqui, M. Duran-Sindreu,´ J. Bonache, and F. Martın´

GEMMA/CIMITEC, Departament d’Enginyeria Electronica,` Universitat Autonoma` de Barcelona, Barcelona 08193 Bellaterra, Spain

Correspondence should be addressed to P. Velez,´ [email protected]

Received 6 September 2012; Accepted 26 October 2012

Academic Editor: Alejandro Lucas Borja

Copyright © 2012 P. Velez´ 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.

Broadband bandpass filters based on open complementary split ring resonators (OCSRRs) coupled through admittance inverters, and implemented in microstrip technology, are reported. As compared to other broadband filters based on open split ring resonators (OSRRs), ground plane etching is not necessary in the proposed filters. The selectivity of the filters at the upper transition band is improved thanks to the presence of a controllable transmission zero. To demonstrate the potential of this approach, two illustrative prototype devices have been designed and fabricated.

1. Introduction open resonators (OSRRs/OCSRRs) are electrically smaller by afactoroftwo[19, 20, 26]. Moreover, OSRRs and OCSRR Metamaterials have been a subject of increasing interest in are intrinsically wideband resonators (the reason for that is the last decade. Soon after the synthesis of the first left- that, as compared to SRRs or CSRRs, OSRRs and OCSRRs handed metamaterial based on a combination of metallic exhibit high C/L and L/C ratios, resp., due to their topology, posts and split ring resonators (SRRs) in 2000 [1], the and this favors broadband responses); therefore, these open first works devoted to extend the properties and concepts resonators are of interest for the implementation of moderate of metamaterials to the design of microwave circuits in and wideband bandpass filters. It is remarkable that order- planar technology came into the scene [2–5]. In particular, 3[27], order-5 [28], and order-7 [29] Chebyshev bandpass it was demonstrated in [5] that a coplanar waveguide (CPW) filters based on a combination of OSRRs and OCSRRs have transmission line loaded with SRRs [6] and shunt metallic been designed and fabricated. In such filters, the OSRRs strips exhibits a bandpass functionality with backward (or and the OCSRRs are described by means of series and left-handed) wave propagation in the first allowed band. The parallel resonators, respectively. Despite the fact that some CPW-based structure reported in [5] was the planar analog parasitics must be introduced in the model in order to of the bulk structure reported in [1] and was the seed for adequately describe the resonators loading a host line, the further design of microwave filters based on SRRs, on their effects of these parasitics are not meaningful, and the filters complementary counterparts (the complementary split ring exhibit the intended Chebyshev responses to a very good resonator—CSRR [7]), or on other metamaterial resonators approximation. [8–18]. In this paper, OCSRRs are used for the implementation The filters reported in [8–18] are based on closed of broadband bandpass filters in microstrip technology. The resonators (SRRs or CSRRs) coupled to a host line. However, filters are similar to those filters reported in [20], that filters based on electrically small open resonators, like is, shunt connected OCSRRs coupled through admittance the open split ring resonator (OSRR) [19], or the open inverters. However, the filters in [20] are implemented in complementary split ring resonator (OCSRR), [20] have also CPW technology, whereas the filters proposed in this paper been reported [21–25]. As compared to SRRs/CSRRs, the are implemented in microstrip technology. As compared 2 International Journal of Antennas and Propagation

Port 1

d

c L0 c rext

Port 2 Cc

(a) (b)

Figure 1: Typical topology of the OCSRR (a) and equivalent circuit model (b). The two terminals—ports—of the open resonator are indicated. to the filters based on OSRRs and also implemented in region of the particle and the metallic region in contact microstrip technology, the advantage of our approach is to port 2. This introduces an inductance in series with the that ground plane etching is not necessary. Moreover, the parallel resonator, as reported in [29], and the effect of this new proposed filters exhibit a transmission zero above the inductance/strip is the presence of a short between ports passband of interest that can be used to improve the filter 1 and 2 at a finite frequency (i.e., above the resonance selectivity above the pass band. frequency of the OCSRRs). In the bandpass filters based The paper is organized as follows. In Section 2,wewill on OSRRs and OCSRRs reported by the authors, where review the OCSRR for completeness, since this resonator is the OCSRR is shunt connected to the host line, this short the key building block of the proposed filters. The design produces a transmission zero above the filter pass band. of the proposed broadband filters, including the simulation This improves the filter selectivity at the upper band and and measurement of the filter responses, are reported in this strategy will be considered in the filters reported in this Section 3. Finally, the main conclusions are highlighted in paper. However, in a first order approximation, the circuit of Section 4. Figure 1(b) can be used for design purposes. Let us now consider that the OCSRR is shunt connected to a microstrip transmission line. To effectively connect 2. Open Complementary Split Ring the resonator to ground, vias are required, as depicted in Resonators (OCSRRs) Figure 2(a). The structure is properly described by means of a shunt connected parallel resonant tank, as revealed by The OCSRR is the open version of the CSRR. The typical the reflection coefficient, represented in a Smith Chart in topology and circuit model of the OCSRR is depicted in Figure 2(b) (notice that the reflection coefficientisalwayson Figure 1. For the OCSRR, the particle terminals (or ports) the unit conductance circle, as expected from a reactive load are indicated in the figure. Between these ports, there is an in parallel to the 50 Ω impedance of the port). Figure 2(c) electric short through the metal between the inner and outer indicates that the resonance frequency of the OCSRR is slot rings forming the particle, but there is also capacitive roughly 1.5 GHz, where total transmission (reflection zero) connection through the capacitances across the slots. Thus, appears since the shunt branch is opened at this frequency. according to this, the circuit model of the particle is an Notice that for this particular OCSRR topology, the strip open parallel resonant tank, as Figure 1(b) illustrates. The connecting the inner region of the OCSRR and the host line ff inductance of the particle, Lo, is the inductance of the is wide, and the inductive e ect can be neglected, at least metallic strip between the ring slots, and the capacitance, in the frequency region shown in the figure (the structure Cc, is identical to the capacitance of the CSRR, that is, the is adequately modeled by means of a parallel resonant tank capacitance of a disk of radius ro − c/2(wherero = rext − c − in such region). d/2 is the averaged radius) surrounded by a metallic plane at a distance c of its edge [30] (notice that this means that the capacitance is identical to that of a slot ring with slot width 3. Bandpass Filters Based on c and average radius ro,asitisjustifiedin[30]). Since the OCSRRs and Results inductance of the CSRR is Lo/4 [30], it is expected that the resonance frequency of the OCSRR is one half the resonance The proposed bandpass filters are implemented by means of frequency of the CSRR. The detailed calculation of Cc is shunt resonators coupled through admittance inverters (see reported in [30], whereas Lo is the inductance of the circular Figure 3), following the theory reported in [31]. The first CPW structure that results from the OCSRR topology. prototype device filter is an order-3 Chebyshev band pass The previous model can be improved by considering the filter with central frequency fo = 2 GHz, ripple of 0.01 dB, inductive effect of the metallic strip connecting the inner and fractional bandwidth of 30%. For these specifications, International Journal of Antennas and Propagation 3

Vias In Out

(a)

− j1

− j . − j2 0 5

0 − j . − j5 0 2 −10 S21 0.5 2 5 0.2 ∞ 1 0 (dB) −20 11 S

(dB), 30 S11 j j . 21 5 0 2 S −40

j 2 j0.5 −50 0.5 1 1.5 2 2.5 j 1 Frequency (GHz)

S11 Circuit simulation Circuit simulation Circuit simulation S11 EM simulation EM simulation EM simulation (b) (c)

Figure 2: Topology (a), Smith Chart (b), and frequency response (magnitude) (c) of the electromagnetic and circuit simulation of a shunt OCSRR in microstrip technology with rext = 2.7 mm, c = 0.2 mm, and d = 1.2 mm. The considered substrate is the Rogers RO3010 with thickness h = 0.254 mm and dielectric constant εr = 10.2. The element values of the shunt resonator in reference of Figure 1(b) are: Lo = 2.02 nH and Cc = 5.25 pF.

J Jn,n+1 Y Yo J0,1 J1,2 2,3 n+1 C L C L p1 p1 Cp2 Lp2 pn pn

Figure 3: Bandpass filter network with admittance inverters and shunt LC resonators. considering identical resonators with a value of inductance frequency). The filter topology is depicted in Figure 4,where and capacitance of L = 1.4nHandC = 4.5 pF, respectively, the main relevant dimensions are indicated (the filter has the characteristic admittance of the inverters are found to been fabricated following a standard photo/mask etching be J01 = J34 = 0.023 S, and J12 = J23 = 0.021 S. The technique). The filter response, including the circuit simu- considered L and C values are those extracted from the lation, the electromagnetic simulation (inferred by means layout of the isolated resonator, which gives the required of the Agilent Momentum commercial software), and the central frequency. The inverters are implemented by means measured response (obtained by means of the Agilent PNA of λ/4 lines (λ being the wavelength at the filter central N5221A vector network analyzer) are depicted in Figure 5. 4 International Journal of Antennas and Propagation

Vias Vias 45.37 mm 59.82 mm w w2 1 In w2 w1 Out 7.2 mm 7.48 mm

In Out (a) (a)

(b) (b)

Figure 4: Topology (a) and photograph (b) of the proposed Figure 6: Topology (a) and photograph (b) of the proposed bandpass filter of the first example. The substrate is the Rogers bandpass filter of the second example. The substrate is the Rogers RO3010 with thickness h = 0.254 mm and dielectric constant RO3010 with thickness h = 0.635 mm and dielectric constant εr = ε = . r = . c = . r 10 2. For the OCSRRs, ext 2 7 mm, 0 37 mm, and 10.2. For the OCSRRs, r = 3 mm, c = 0.16 mm, and d = 2 mm. d = . w = ext 1 63 mm. The width of the transmission line inverters is 1 The width of the transmission line inverters is w1 = 0.503 mm for . w = . 0 33 mm for the external ones and 2 0 28 mm for the internal the external ones and w2 = 0.335 mm for the internal ones. ones.

0 − −10 stop band reveals that rejection is better than 10 dB up to at least 5 GHz. S21 −20 Notice that the filter (measured) bandwidth is smaller

(dB) than the bandwidth of the ideal response since the invert-

11 −30 S ers strictly exhibit their functionality at the central filter frequency. This aspect is well known and can be solved − (dB), 40 by considering a larger nominal bandwidth. However, this 21

S S −50 11 is not relevant for our purposes, since the main aim is to demonstrate the potential of OCSRRs for the design of −60 broadband microstrip filters based on admittance inverters, as an alternative to OSRR-based microstrip filters. −70 The second example is an order-3 Chebyshev band pass 12345 filter with central frequency fo = 2 GHz, ripple of 0.1 dB, and Frequency (GHz) fractional bandwidth of 30%. For these specifications, L = Ideal simulation EM simulation 1.4nHandC = 4.5 pF, and the characteristic impedance of Circuit simulation Measure the inverters is J01 = J34 = 0.0184 S, and J12 = J23 = 0.0156 S. The topology and photograph of this filter are shown in Figure 5: Frequency response of the filter of Figure 4. The circuit Figure 6 and the frequency response in Figure 7. In this case, simulation has been achieved by means of Agilent ADS, considering lumped resonators coupled through 90◦ at the central frequency. the pass band exhibits three reflection zeros and the return loss is better than −20 dB between 1.77 GHz and 2.25 GHz. Insertion loss is smaller than 0.9 dB within this frequency interval. The agreement between the electromagnetic simulation and measurement is remarkable. Notice that the measured filter selectivity at the upper edge of the band is much better 4. Conclusions than that obtained by the circuit simulation. The reason is the presence of the transmission zero above the passband, In conclusion, novel broadband microstrip bandpass filters tailored by means of the width of the connecting strip based on OCSRRs coupled through admittance inverters between the inner region of the OCSRRs and the host line. have been proposed. The reported filters represent a sig- Since the considered ripple is so small, the filter exhibits a nificant progress as compared to other microstrip filters quasi-Butterworth response, with a single reflection zero in based on OSRRs, where ground plane etching in the region the pass band. Filter performance is good, with a measured occupied by the OSRRs is necessary. The two designed filters insertion loss smaller than 1.7 dB and return loss better than exhibit good performance, and filter selectivity has been −10 dB between 1.68 GHz and 2.21 GHz. The filter upper improved by the transmission zero inherent to the OCSRRs. International Journal of Antennas and Propagation 5

0 [9] J. Garc´ıa-Garc´ıa, F. Mart´ın, F. Falcone et al., “Spurious passband suppression in microstrip coupled line band pass filters by means of split ring resonators,” IEEE Microwave and −10 Wireless Components Letters, vol. 14, no. 9, pp. 416–418, 2004. [10] J. Bonache, F. Martin, J. Garcia-Garcia, I. Gil, R. Marques,´ and −20 (dB) M. Sorolla, “Ultra wide band pass filters (UWBPF) based on S21 11

S complementary split rings resonators,” Microwave and Optical −30 Technology Letters, vol. 46, no. 3, pp. 283–286, 2005.

(dB), [11] J. Bonache, I. Gil, J. Garc´ıa-Garc´ıa, and F. Mart´ın, “Novel 21

S −40 microstrip bandpass filters based on complementary split- ring resonators,” IEEE Transactions on Microwave Theory and Techniques, vol. 54, no. 1, pp. 265–271, 2006. −50 S11 [12] J. Garc´ıa-Garc´ıa,J.Bonache,I.Gil,F.Mart´ın, M. Del Castillo Velazquez-Ahumada,´ and J. Martel, “Miniaturized microstrip −60 12345 and CPW filters using coupled metamaterial resonators,” IEEE Frequency (GHz) Transactions on Microwave Theory and Techniques, vol. 54, no. 6, pp. 2628–2635, 2006. Ideal simulation EM simulation [13] P. Mondal, M. K. Mandal, A. Chaktabarty, and S. Sanyal, Circuit simulation Measure “Compact bandpass filters with wide controllable fractional bandwidth,” IEEE Microwave and Wireless Components Letters, Figure 7: Frequency response of the filter of Figure 6. vol. 16, no. 10, pp. 540–542, 2006. [14] M. Gil, J. Bonache, J. Garc´ıa-Garc´ıa, J. Martel, and F. Mart´ın, “Composite right/left-handed metamaterial transmis- sion lines based on complementary split-rings resonators and Acknowledgments their applications to very wideband and compact filter design,” IEEE Transactions on Microwave Theory and Techniques, vol. This work has been supported by MICIIN-Spain (Contracts 55, no. 6, pp. 1296–1304, 2007. TEC2010-17512 and CSD2008-00066) and by Generalitat de [15] I. Gil, F. Martin, X. Rottenberg, and W. De Raedt, “Tunable Catalunya (Project 2009SGR-421). stop-band filter at Q-band based on RF-MEMS metamateri- als,” Electronics Letters, vol. 43, no. 21, pp. 1153–1154, 2007. [16] M. Gil, J. Bonache, and F. Mart´ın, “Metamaterial filters: a References review,” Metamaterials, vol. 2, no. 4, pp. 186–197, 2008. [1]D.R.Smith,W.J.Padilla,D.C.Vier,S.C.Nemat-Nasser,and [17] A. L. Borja, J. Carbonell, V. E. Boria, J. Cascon,´ and D. S. Schultz, “Composite medium with simultaneously negative Lippens, “Synthesis of compact and highly selective filters via permeability and permittivity,” Physical Review Letters, vol. 84, metamaterial-inspired coplanar waveguide line technologies,” no. 18, pp. 4184–4187, 2000. IET Microwaves, Antennas and Propagation, vol. 4, no. 8, pp. [2] A. K. Iyer and G. V. Eleftheriades, “Negative refractive index 1098–1104, 2010. metamaterials supporting 2-D waves,” vol. 2, pp. 1067–1070. [18] A. L. Borja, J. Carbonell, V. E. Boria, J. Cascon, and D. Lippens, [3] A. A. Oliner, “A periodic-structure negative-refractive-index “A 2% bandwidth C-band filter using cascaded split ring medium without resonant elements,” in Proceedings of the resonators,” IEEE Antennas and Wireless Propagation Letters, USNC-URSI National Radio Science Meeting, p. 41, San vol. 9, pp. 256–259, 2010. Antonio, Tex, USA, June 2002. [19] J. Martel, R. Marques,´ F. Falcone et al., “A new LC series [4] C. Caloz and T. Itoh, “Application of the transmission line element for compact bandpass filter design,” IEEE Microwave theory of left-handed (LH) materials to the realization of a and Wireless Components Letters, vol. 14, no. 5, pp. 210–212, microstrip ‘LH line’,” in Proceedings of the IEEE Antennas and 2004. Propagation Society International Symposium, pp. 412–415, [20] A. Velez,F.Aznar,J.Bonache,M.C.Vel´ azquez-Ahumada,´ San Antonio, Tex, USA, June 2002. J. Martel, and F. Mart´ın, “Open complementary split ring [5] F. Mart´ın, J. Bonache, F. Falcone, M. Sorolla, and R. Marques,´ resonators (OCSRRs) and their application to wideband cpw “Split ring resonator-based left-handed coplanar waveguide,” band pass filters,” IEEE Microwave and Wireless Components Applied Physics Letters, vol. 83, no. 22, pp. 4652–4654, 2003. Letters, vol. 19, no. 4, pp. 197–199, 2009. [6] J. B. Pendry, A. J. Holden, D. J. Robbins, and W. J. Stew- [21] J. Martel, J. Bonache, R. Marques,´ F. Mart´ın, and F. Medina, art, “Magnetism from conductors and enhanced nonlinear “Design of wide-band semi-lumped bandpass filters using phenomena,” IEEE Transactions on Microwave Theory and open split ring resonators,” IEEE Microwave and Wireless Techniques, vol. 47, no. 11, pp. 2075–2084, 1999. Components Letters, vol. 17, no. 1, pp. 28–30, 2007. [7] F. Falcone, T. Lopetegi, J. D. Baena, R. Marques,´ F. Mart´ın, [22] F. Aznar, A. Velez,´ J. Bonache, J. Menes,´ and F. Mart´ın, “Com- and M. Sorolla, “Effective negative-ε stopband microstrip pact lowpass filters with very sharp transition bands based on lines based on complementary split ring resonators,” IEEE open complementary split ring resonators,” Electronics Letters, Microwave and Wireless Components Letters, vol. 14, no. 6, pp. vol. 45, no. 6, pp. 316–317, 2009. 280–282, 2004. [23] F. Aznar, A. Velez,´ M. Duran-Sindreu,´ J. Bonache, and F. [8] F. Mart´ın, F. Falcone, J. Bonache, R. Marques,´ and M. Sorolla, Martin, “Elliptic-function CPW low-pass filters implemented “Miniaturized coplanar waveguide stop band filters based on by means of open complementary split ring resonators multiple tuned split ring resonators,” IEEE Microwave and (OCSRRs),” IEEE Microwave and Wireless Components Letters, Wireless Components Letters, vol. 13, no. 12, pp. 511–513, 2003. vol. 19, no. 11, pp. 689–691, 2009. 6 International Journal of Antennas and Propagation

[24] A. Velez,´ F. Aznar, M. Duran-Sindreu,´ J. Bonache, and F. Mart´ın, “Stop-band and band-pass filters in coplanar waveg- uide technology implemented by means of electrically small metamaterial-inspired open resonators,” IET Microwaves, Antennas and Propagation, vol. 4, no. 6, pp. 712–716, 2010. [25] A. Velez,´ F. Aznar, M. Duran-Sindreu,´ J. Bonache, and F. Mart´ın, “Tunable coplanar waveguide band-stop and band- pass filters based on open split ring resonators and open com- plementary split ring resonators,” IET Microwaves, Antennas and Propagation, vol. 5, no. 3, pp. 277–281, 2011. [26] F. Aznar, A. Velez,´ M. Duran-Sindreu,´ J. Bonache, and F. Mart´ın, “Open complementary split ring resonators: physics, modelling, and analysis,” Microwave and Optical Technology Letters, vol. 52, no. 7, pp. 1520–1526, 2010. [27] M. Duran-Sindreu,´ A. Velez,´ F. Aznar, G. Siso,´ J. Bonache, and F. Mart´ın, “Applications of open split ring resonators and open complementary split ring resonators to the synthesis of artifi- cial transmission lines and microwave passive components,” IEEE Transactions on Microwave Theory and Techniques, vol. 57, no. 12, pp. 3395–3403, 2009. [28] M. Duran-Sindreu,´ A. Velez,´ G. Siso´ et al., “Recent advances in metamaterial transmission lines based on split rings,” Proceedings of the IEEE, vol. 99, no. 10, pp. 1701–1710, 2011. [29] M. Duran-Sindreu,´ P. Velez,´ J. Bonache, and F. Mart´ın, “Broadband microwave filters based on Open Split Ring Resonators (OSRRs) and Open Complementary Split Ring Resonators (OCSRRs): improved models and design opti- mization,” Radioengineering, vol. 20, no. 4, pp. 775–783, 2011. [30] J. D. Baena, J. Bonache, F. Mart´ın et al., “Equivalent-circuit models for split-ring resonators and complementary split- ring resonators coupled to planar transmission lines,” IEEE Transactions on Microwave Theory and Techniques, vol. 53, no. 4, pp. 1451–1460, 2005. [31] J.-S. Hong and M. J. Lancaster, Microstrip Filters For RF/Mi- crowave Applications, John Wiley & Sons, New York, NY, USA, 2001. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 859429, 7 pages doi:10.1155/2012/859429

Research Article A Method for Extending the Bandwidth of Metamaterial Absorber

Hong-Min Lee and Hyung-Sup Lee

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

Correspondence should be addressed to Hong-Min Lee, [email protected]

Received 13 September 2012; Accepted 28 October 2012

Academic Editor: Alejandro Lucas Borja

Copyright © 2012 H.-M. Lee and H.-S. Lee. 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.

We present a method for extending the bandwidth of a metamaterial absorber (MMA) with multiresonance structure. The basic unit cell of the MMAs consists of a periodic arrangement of an electric-LC (ELC) resonator and a square loop structure. The absorption bandwidth of an MMA is effectively extended by combining five unit cell structures with different geometric dimen- sions into a coplanar unit cell. Experimental results show that the fabricated MMA is 0.8 mm (0.0026 λ0) thick with a peak absorption rate of 93% at 10 GHz and a full width at half maximum (FWHM) bandwidth of 970 MHz.

1. Introduction are low cost and simple. However, most of the MMA struc- tures are based on strong electromagnetic resonances to Nowadays, the microwave absorbers are interested in mil- effectively absorb the incident electromagnetic wave; the itary and civil application to reduce the electromagnetic absorbance bandwidth of MMA is quite narrow. This narrow interference among microwave components or electronic bandwidth characteristic of MMA limits absorber applica- circuits mounted on the same platform. One of conventional tions, so the method for extending the absorbance band- absorbers is Salisbury screen, and its mechanism is explained width of MMA is needed. In recent years, MMAs with dual- well in [1]. This type of absorber involves the use of a resistive band [7] and multifrequency resonator [8, 9]havebeen sheet and a metallic ground plane to cancel out reflections reported, but these types of MMAs show multiple absorption from the screen. Recently, absorber technology has seen peaks at the different frequency bands. In addition, the several advancements with the use of artificially structured absorption bandwidth of these designed MMAs is narrow, metamaterials (MTMs) for creating a high-performance typically no more than 10%. This narrowband width feature absorber for the microwave, terahertz, and near infrared limits the device applications of the resonant type of MMA. frequency regimes [2–5]. According to the effective medium In this paper, we propose an efficient method to extend theory, MTMs can be represented by the complex values of the absorbance bandwidth of an MMA. By combining five = ff electric permittivity εeff( εeff + jεeff) and magnetic perme- unit cell structures with di erent geometric dimensions into = ability μeff( μeff + jμeff). This permittivity and permeability a coplanar unit cell, we obtain the superposition of the can be independently controlled by varying the dimensions five different absorption peaks. Various array configurations of the electric and magnetic components. Additionally, by of five unit cells with different geometric dimensions are tuning the electric and magnetic resonances, an MTM can investigated in order to expend the bandwidth of MMA. The be impedance-matched to free space. As a result, 100% 3D field simulation tool, CST MWS (Micro Wave studio), is absorbance is theoretically possible. Compared to the tradi- used for the design of the proposed MMA structure. tional microwave absorber, metamaterial structure can give some solutions to improve traditional absorber. First, the 2. The Design of an MTM Absorber Unit Cell thickness of resonant metamaterial absorber (MMA) is much thinner than the traditional absorber which has quarter Geometry of the proposed basic MMA unit cell is shown in wavelength thickness [6]. Second, the fabrications of MMA Figure 1. The unit cell is designed on the front and back side 2 International Journal of Antennas and Propagation

a1 respectively. The gap length of ELC structure is 0.9 mm and the width of two metallic patterns is 0.3 mm. w1 We use frequency domain solver Microwave Studio by CST, and the program simulated a single unit cell with appro- priate periodic boundary conditions (PBCs). A unit cell of the structure is placed inside a waveguide with PBC walls and a vertically polarized transverse electromagnetic (TEM) wave impinges upon this structure. The perfect electric a2 1 conductor (PEC) boundary conditions are applied to the top w and bottom walls of the waveguide, whereas perfect magnetic conductor (PMC) boundary conditions are applied to the t1 side walls of the waveguide. The other two opposite sides of the waveguide are assigned as waveguide ports. The simu- lated frequency range is from 9.0 GHz to 11 GHz. This struc- E ture generates antiparallel circulating current between the vertical arms of the ELC and the loop resulting in magnetic response, as shown in Figure 2(d). By tuning each of the H k resonances it is possible to match the impedance to space and minimize the reflectance. Figure 2(e) shows the simulated (a) power loss densities in the ELC resonator structure at a1 10.12 GHz. The strongest electromagnetic energy absorption can take place where the electric field is confined in the vicin- w2 ity of the dielectric substrate. This result indicates that the absorption does not homogeneously occur in the MTM unit cell. Figure 3 shows the simulation results for the proposed MTM absorber, through the plotting of absorption rate A(ω), reflectance R(ω), and transmission T(ω). The scatter-

2 ing parameters of this MTM unit cell were simulated, and w the absorption was calculated by the equation A = 1 − g 2 2 |S11| −|S21| = 1−R(ω)−T(ω). It can be observed that the t2 reflectance and transmission of the absorber sharply reduce to a minimum at 10.12 GHz, which also shows a peak absorption rate of 92%. Based on absorption rate of 80%, it has bandwidth of 220 MHz and the full width at half maximum (FWHM) bandwidth of 425 MHz. (b) In Figure 2, we present the distributions of the simulated electric field for both TE (Figure 2(a)) and TM polarizations Figure 1: Geometry of the basic absorber unit cell (a1 = 5.8 mm, (Figure 2(b)), and magnetic field (Figure 2(c)), along with a2 = 5.8 mm, w1 = 4.1 mm, w2 = 4.3 mm, g = 0.9 mm, t1 = the surface current (Figure 2(d)) for the proposed MTM 0.3 mm, t2 = 0.4 mm); (a) front-view; (b) bottom-view. absorber unit cell in normal incidence at resonance fre- quency of 10.12 GHz. From the electric field distribution shown in Figures 2(a) and 2(b), one can observe that the four orthogonal gaps of the ELC element can confine the electric field owing to accumulated opposite charges both of FR-4 substrate (εr = 4.6, loss tangent = 0.025), and the TE and TM polarizations. Hence, capacitance is produced in thickness of FR-4 substrate is 0.8 mm. The unit cell consists these gaps and the ELC element acts as an electric resonator. of an electric-LC (ELC) resonator and a square loop element. Meanwhile, the induced magnetic field for TE polarization is Many kinds of ELC structure are reported in [10]. A polar- stronger in the regions close to the vertical arms of the ELC, ization insensitive ELC structure, as shown in Figure 1(b), as shown in Figure 2(c). is selected in this design. ELC resonator exhibits a resonant electric response when the electric field is polarized vertically or horizontally along 3. Arrangement Configuration for the gaps of an ELC, and the circulating displacement currents Extending Bandwidth of MMA Cells which are generated between ELC and loop arms exhibit a resonant magnetic response. Overall size of the basic MMA The feasibility of a broad bandwidth MMA can be achieved unit cell is 5.8 mm × 5.8 mm × 0.8 mm including FR-4 sub- by overlapping the absorption peaks of the five cells when strate. The metallic pattern size of an ELC and a square their peaks closed to each other. To extend the bandwidth loop structure are 4.3 mm × 4.3 mm and 4.1 mm × 4.1 mm, of MMA five MTM unit cells with different geometric International Journal of Antennas and Propagation 3

50000 50000 42188 42188 35938 35938 29688 29688 23438 23438 17188 17188 10938 10938

/m 0 0 /m V −7813 V −7813 −14063 −14063 −20313 −20313 −26563 −26563 E −32813 −32813 −39063 k −39063 H E k −50000 H −50000

(a) (b)

400 338 288 238 188 138 87.5 0 /m

A −62.5 −113 −163 − E 213 −263 − H 313 k −400

(c)

400 400 192 192 106 106 58 58 30.8 30.8 15.5 15.5 6.98 6.98 0 0

/m − − 4.23 /m 4.23 A −10.6 A −10.6 −22.1 −22.1 −42.4 −42.4 −78.7 −78.7 −143 −143 −400 −400

(d)

6e86e8 4.17e8 4.17e8 3.11e8 3.11e8 2.31e8 2.31e8 1.72e8 1.72e8 1.27e8 1.27e8 3 9.39e7 3 9.39e7

/m 6.89e7 /m 6.89e7

W 5.01e7 W 5.01e7 3.6e7 3.6e7 2.55e7 2.55e7 1.76e7 1.76e7 1.17e7 1.17e7 7.26e6 7.26e6 3.94e6 3.94e6 0 0 (e)

Figure 2: Simulated results of the proposed unit cell at 10.12 GHz; (a) electric field distribution for TE incidence; (b) electric field distribution for TM incidence; (c) magnetic field distribution for TE incidence; (d) surface current distribution for TE incidence; (e) the simulated average power loss densities of the absorber unit cell at two resonant frequencies. 4 International Journal of Antennas and Propagation

1 10.12 GHz 1 92%

0.8 0.8

0.6 0.6 A , T ,

R 0.4

0.4 Absorption

0.2 0.2

0 0 9 9.5 10 10.5 11 9 9.5 10 10.5 11 Frequency (GHz) Frequency (GHz) Case 1 R Case 2 T Case 3 A (a) Figure 3: Calculated transmission (T), reflection (R), and absorp- tion (A) of the basic MMA unit cell from simulated S-parameters. 1

0.8 Table 1: Geometrical parameters of five different unit cells.

Geometrical parameters (mm) 0.6 Cell type f0 (GHz) a1 a2 w1 w2 gt1 t2 0.4

K1 5.8 5.8 4.0 4.1 1.5 0.7 0.4 10.50 Absorption K2 5.8 5.8 4.0 4.2 1.1 0.4 0.4 10.47 K3 5.8 5.8 4.1 4.3 0.9 0.3 0.4 10.12 0.2 K4 5.8 5.8 4.2 4.4 0.6 0.2 0.4 9.96

K5 5.8 5.8 4.2 4.4 0.7 0.2 0.4 9.69 0 9 9.5 10 10.5 11 Frequency (GHz) dimensions were designed. We select five different resonant Case 1 unit cells with different scaling factors as shown in Table 1 Case 2 and simulate various array configurations of five unit cells Case 3 with a different arrangement, as shown in Table 2.Wedefine (b) the type of K3 as the basic unit cell sample, and some Figure 4: Comparison of the calculated absorption characteristics geometric dimensions of other four units are scaled up or with five different single unit cells with different arrangement and down, as shown in Table 1. In addition, we also simulate the array directions: (a) horizontal array; (b) vertical array. five single unit cells by changing the periodic array direction of the unit cells. Figure 4 shows the calculated absorption characteristics Table 2: Three different types of arrangement with five different comparison of five different single unit cells with different unit cells. arrangement and array directions. To overlap the single res- Type Arrangement onant absorption peak of five single unit cells with different dimensions, the five single unit cells are periodically arrayed Case 1 K1-K2-K3-K4-K5 in order of K1-K2-K3-K4-K5 (Case 1), K2-K5-K3-K1-K4 Case 2 K2-K5-K3-K1-K4 (Case 2), and K5-K2-K1-K3-K4 (Case 3), respectively. Case 3 K5-K2-K1-K3-K4 The simulated results show that the absorption band- width of the horizontal array configurations (cells are arrayed in x-axis) is narrow compared to vertical array configurations bandwidth of vertical array configuration (Case 1) extended (cells are arrayed in y-axis) except in Case 2. Although peak up to 770 MHz. absorption value of vertical arrayed unit cells in Case 1 is As the differences of scaling factors between cells are slightly lower compared to that of in Case 3 configurations, decreased, the five single resonant absorption peaks will it shows the widest absorption bandwidth. The FWHM overlap together. This is mainly due to the combination of International Journal of Antennas and Propagation 5

five different geometric dimensions of unit cell results to shift of the resonance frequency to lower or higher values, depending on the capacitive or inductive nature of the cou- pling. When the five single unit cells are periodically arrayed in order of K1-K2-K3-K4-K5 (Case 1), it shows the smallest differences of scaling factors between cells. In this mixture of five unit cells with different geometric dimensions, the five K1 single resonant absorption peaks will overlap together, and K2 the bandwidth of an MMA is extended. On the other hand, K3 increasing the differences of scaling factors between cells K4 results to splitting of the five single resonant absorption E, y^ K5 peaks will single resonance into five distinct resonances. As H, x^ k, z^ a result, the five single unit cells are periodically arrayed in different order (Case 2 and Case 3), it shows the lowered absorption bandwidth compare to Case 1. Further broaden- (a) ing of the absorption bandwidth is also possible by increasing 1 92% 91.7% the number of different geometric dimensions of unit cell with proper combination. Figure 5 shows the perspective view of the proposed 0.8 MMA structure and the comparison of the calculated absorption for periodic array of single unit cell (K3 type) 0.6 and five different single unit cells arrayed in vertical direction with different arrangement (Case 1). It is observed from 0.4 Figure 5 that the absorption bandwidth of five different sin- Absorption gle unit cells array configuration increases compared to the 9.9∼10.37 GHz single unit cell array configuration. The FWHM absorption (0.47 GHz) 0.2 9.7∼10.47 GHz bandwidth is extended from 470 MHz to 770 MHz by using (0.77 GHz) five unit cell structures with different geometric dimensions. This extended absorption bandwidth of the MMA is mainly 0 due to the electromagnetic field coupling between the neigh- 9 9.5 10 10.5 11 bour unit cells [11]. Figure 6 shows the simulated electric Frequency (GHz) field distributions of the five different absorber unit cells Single cell arrayed in vertical direction (Case 1). 5 arrayed cells At different resonant frequencies, two or three neigh- (b) bouring unit cells exhibit strong electric field in the area between ELCs, indicating a capacitive coupling. As a result, Figure 5: (a) Perspective view of the proposed MMA structure; (b) the proposed MMA structure no longer shows single- comparison of the calculated absorption for periodic array of single unit cell (K3 type) and five different unit cells arrayed in vertical resonant characteristics. By combining five unit cell struc- direction. tureswithdifferent geometric dimensions, some of resonant peaks closely positioned absorber unit cells are overlapped. Thus, the bandwidth of an MMA can be expended due to the multiresonance effect. absorber sample are shown in Figure 8. The sample was = Thenatureofthisabsorptioncanbeunderstoodfrom etched on an FR-4 substrate (relative dielectric constant εr = = Figure 7, which shows the simulated power loss densities of 4.6, loss tangent δ 0.025, and thickness t 0.8 mm) using ff standard photolithography techniques. In order to verify the the absorber unit cell at three di erent resonant frequencies; ff 9.69, 10.12, and 10.5 GHz. At different resonant frequencies, e ectiveness of the MTM absorber cells, a planar array of × the power loss is concentrated strongly in the vicinity of the absorber unit cells (40 45) etched on both sides of an open gaps and two or three neighbouring ELCs unit cells. In FR-4 substrate was mounted on an acryl substrate frame the proposed MMA structure, the strongest electromagnetic for measurement. The interelement spacing between the two energy absorption, as shown in Figure 7,takesplacewhere absorber cells was set to 0.2 mm, and the total size of the × the electric field is confined in the vicinity of the dielectric planar absorber was 243.6 mm 249 mm. We experimen- substrate, as shown in Figure 6. tally verified the behaviour of the absorber by measuring the S-parameters of a planar array of unit cells. Figure 9 shows the schematic of the test setup for reflection and transmission 4. The Fabricated Absorber measurement of the fabricated absorber sample. Measurement Results We used a vector network analyzer and two X-band microwave rectangular horn antennas. Measurements were We fabricated a prototype absorber for experimentation. taken over a frequency range from 9 to 11.0 GHz. Before Photographs of the fabricated two-layer metallization MTM starting the characterization measurements, we performed 6 International Journal of Antennas and Propagation

50000 32813 K1 23020 15763 10386 249 mm K2 6401 3448

K3 /m 0 V −2273 −4814 243.6 mm K4 −8245 (a) −12874 −19120 K5 −27551

−50000 9.6 GHz 10.12 GHz 10.5 GHz Figure 6: Simulated electric field distributions of the absorber unit 249 mm cell at different resonant frequencies.

7e7 5.68e7 K1 4.8e7 243.6 mm 4.03e7 (b) 3.37e7 K2 2.81e7 Figure 8: Photographs of the fabricated prototype absorber. 2.32e7

3 1.9e7 K3 /m 1.54e7 W to eliminate unwanted edge scattering. To obtain the S11 1.23e7 parameters, two horns were focused on the sample sheet 9.57e6 on the same side, as shown in Figure 9(b). The height of K4 7.26e6 the horn antennas is maintained at 1.2 m and the distance 5.28e6 between horns and sample sheet is 0.6 m to eliminate near- ff 3.57e6 field coupling e ects. Figure 10 shows a comparison of the simulated and measured absorption of the proposed MMA K5 2.1e6 using measured magnitudes of the S11 and S21 parameters. A 0 slight increment in the measured bandwidth in comparison 9.69 GHz 10.12 GHz 10.5 GHz with that of the simulated result is attributed to the fabrica- Figure 7: Simulated average power loss densities of the absorber tion tolerances and experiment setup for the proposed unit cell at three different resonant frequencies. MMA. In addition, the experimental absorption frequency band is shifted approximately 60 MHz higher when com- pared to the simulated results. The measured FWHM band- width is extended from 470 MHz to 940 MHz compared a through-reflect-line (TRL) calibration. The reference plane to the single unit cell array configuration. The proposed is located at the surface of the sample. The reflection meas- absorber shows a peak absorption rate of 93% at 10 GHz. urement has been calibrated by replacing the sample with an aluminum plate of same size at the reference plane. 5. Conclusion The transmission measurement has been calibrated with no sample. This paper presents an efficient approach for achieving the As shown in Figure 9(a), a pair of horn antennas were extended absorbance bandwidth of an MMA. The proposed used to transmit the EM wave on the sample absorber sheet MMA unit cells, which consist of an ELC resonator and a and receive transmitted signals to obtain the S21 parameters square loop structure, are etched on both sides of the FR-4 for normal incidence. Microwave absorbing material is substrate. The simulated results of the basic unit cell show placed between two horns and surrounding the sample sheet a peak absorption rate of 92% at 10.12 GHz, and FWHM International Journal of Antennas and Propagation 7

Transmitting Acknowledgment horn antenna Receiving horn antenna This research was supported by Basic Science Research Program through National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (no. 2010-0011646).

Absorber sample References

(a) [1] R. L. Fante and M. T. McCormack, “Reflection properties of the Salisbury screen,” IEEE Transactions on Antennas and Receiving horn antenna Propagation, vol. 36, no. 10, pp. 1443–1454, 1988. [2] N. I. Landy, S. Sajuyigbe, J. J. Mock, D. R. Smith, and W. J. Padilla, “Perfect metamaterial absorber,” Physical Review Letters, vol. 100, no. 20, Article ID 207402, 4 pages, 2008. [3] H. Tao, N. I. Landy, C. M. Bingham, X. Zhang, R. D. Averitt, Transmitting horn antenna and W. J. Padilla, “A metamaterial absorber for the tera- hertz regime: design, fabrication and characterization,” Optics Express, vol. 16, no. 10, pp. 7181–7188, 2008. [4] Y. Cheng and H. Yang, “Design, simulation, and measurement (b) of metamaterial absorber,” Journal of Applied Physics, vol. 108, Figure 9: (a) Schematics of reflection and transmission measure- no. 3, Article ID 034906, 4 pages, 2010. ment test setup for normal incidence and (b) oblique incidence. [5] K. Boratay Alici, A. Burak Turhan, C. M. Soukoulis, and E. Ozbay, “Optically thin composite resonant absorber at the near-infrared band: a polarization independent and spectrally broadband configuration,” Optics Express, vol. 19, no. 15, pp. 1 93% 14260–14267, 2011. [6] J. Lee, Y. J. Yoon, and S. Lim, “Ultra-thin polarization inde- pendent absorber using hexagonal interdigital metamaterials,” 0.8 92% ETRI Journal, vol. 34, no. 1, pp. 126–129, 2012. [7] H. Tao, C. M. Bingham, D. Pilon et al., “A dual band terahertz 0.6 metamaterial absorber,” Journal of Physics D, vol. 43, no. 22, Article ID 225102, 2010. [8]X.Shen,T.J.Cui,J.Zhao,H.F.Ma,W.X.Jiang,andH.Li, 0.4 “Polarization-independent wide-angle triple-band metamate- ∼ Absorption 9.7 10.47 GHz rial absorber,” Optics Express, vol. 19, no. 10, pp. 9401–9407, (0.77 GHz) 2011. 0.2 9.76∼10.7 GHz [9] H. Li, L. H. Yuan, B. Zhou, X. P. Shen, Q. Cheng, and T. J. (0.94 GHz) Cui, “Ultrathin multiband gigahertz metamaterial absorbers,” 0 Journal of Applied Physics, vol. 110, no. 1, Article ID 014909, 8 9 9.5 10 10.5 11 pages, 2011. Frequency (GHz) [10]W.J.Padilla,M.T.Aronsson,C.Highstrete,M.Lee,A.J. Taylor, and R. D. Averitt, “Electrically resonant terahertz met- Simulation amaterials: theoretical and experimental investigations,” Phys- Measurement ical Review B, vol. 75, no. 4, Article ID 041102, 4 pages, 2007. [11] Y. Qiu Xu, P. Heng Zhou, H. Bin Zhang, L. Chen, and L. Jiang Figure 10: Comparison of simulated and experimented absorption Deng, “A wide-angle planar metamaterial absorber based on of five unit cells with different dimensions arrayed in vertical split ring resonator coupling,” Journal of Applied Physics, vol. direction (Case 1). 110, no. 4, Article ID 044102, 5 pages, 2011. bandwidth of 470 MHz. By combining five unit cell struc- tureswithdifferent geometric dimensions, their resonant peaks are overlapped. In addition, by placing the arrange- ment direction of the cells in vertically, the FWHM of the proposed MMA is extended up to 100% experimentally. This proposed configuration is an efficient method for extending the bandwidth of a MMA. The proposed layout can be easily extended to work for more compact, thinner backplane- less planar absorber designs for millimetre and terahertz frequency applications and can be applied for EM wave absorbing materials and microwave devices. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 613518, 6 pages doi:10.1155/2012/613518

Research Article Broadband Equivalent Circuit Model for a Coplanar Waveguide Line Loaded with Split Ring Resonators

Victor Sanz,1 Angel Belenguer,1 Alejandro L. Borja,1 Joaquin Cascon,1 Hector Esteban,2 and Vicente E. Boria2

1 Departamento de Ingenier´ıa El´ectrica, Electronica,´ Automatica´ y Comunicaciones, Universidad de Castilla-La Mancha, Escuela Polit´ecnica de Cuenca, Campus Universitario, 16071 Cuenca, Spain 2 Departamento de Comunicaciones, Universidad Polit´ecnica de Valencia, 46022 Valencia, Spain

Correspondence should be addressed to Victor Sanz, [email protected]

Received 17 September 2012; Accepted 16 October 2012

Academic Editor: Eric Lheurette

Copyright © 2012 Victor Sanz 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 new equivalent circuit for a coplanar waveguide loaded with split ring resonators is presented. The traditional circuits that model these devices are only able to characterize the left-handed propagation band, and their response is very similar to the real one within a very limited bandwidth. In contrast, this proposed broadband equivalent circuit is able to portray not only the left- handed propagation band, but also the right-handed one that occurs at higher frequencies. Besides, the response of this kind of basic cells can be adjusted with the proposed circuit model in a bandwidth close to a decade.

1. Introduction of the equivalent circuit of [5–7] is proposed. This new equivalent circuit preserves the rigorous description of the Many publications, regarding the design of very compact physical mechanisms of the real circuit and, at the same time, filters based on Coplanar Waveguide (CPW) lines loaded is able to simultaneously model the left- and right-handed with Split Ring Resonators (SRRs), can be found in scientific transmission over a bandwidth close to a decade. literature. Different examples of such filter designs are [1– The paper is organized as follows. Section 2 discusses 5]. These filters are designed based on an equivalent circuit the equivalent circuit derivation. Section 3 presents a specific where the value of the different components can be linked in design and the associated results. The performance is also a very direct way to the physical dimensions of the rings and compared with the traditional equivalent circuit. Finally, in the line [5–7]. Therefore, using this circuit model—that is Section 4, the main conclusions of the study are outlined. relatively easy to calculate—, it is possible to design complex filters formed by the concatenation of various basic cells [8, 9]. This equivalent circuit provides a very good starting 2. Equivalent Circuit point, so that an appropriately configured optimization The basic cell used for the filters of [1, 2, 4, 8]canbeseenin algorithm can converge to an acceptable solution. Figure 1 and its equivalent circuit in Figure 2. Initially, the equivalent circuit of [5] was able to In order to simplify the circuit, the symmetry along the characterize the upper right-handed bands. However, in a CPW longitudinal axis has been applied, so in this equivalent later publication the authors modified their own equivalent circuit, only half cell needs to be portrayed. circuit [6]. The new proposal provided a more accurate Since the electrical size of the cell is quite small, the characterization of the coupling and propagation physical CPW line can be modeled using a series inductor and a phenomena of the real circuit made with CPW and SRRs, but parallel capacitor. If symmetry had not been applied, these unfortunately, this modified circuit was not able to model inductance and capacitance values could have been com- the upper right-handed bands. In this paper a modification puted simply by multiplying the inductance and capacitance 2 International Journal of Antennas and Propagation

series-connected inductors, and every half of the ring has to

wr gr be coupled to the appropriate half of the CPW. This ring model is valid for an isolated ring [10]. Even wl rr other more complicated circuit models like, for example, the model presented in [11], presume an isolated ring as wr well. However, in the basic cell of Figure 1, one can see that gc wc gc these rings cannot be considered in isolation. The line acts Figure 1: Basic cell geometry. Black is metal on bottom layer; gray as a ground plane that, in this case, is very close to the is metal on top layer. rings. This ground plane produces an additional capacitive coupling between the rings. This coupling is clearly different from the direct capacitive coupling which is modeled by Ca/2 (see Figure 2). Ca/2 In order to take this new capacitive coupling into consideration, it is necessary to slightly modify the circuit La/2 La/2 that models the ring. This modification can be explained from the physical process responsible for the ring resonance. Within the SRR, first, the current only flows along one kk of the rings. As this current flows alongside this first ring it progressively crosses to the second one due to their mutual

Ll Ll capacitive coupling. When the current reaches the slot in the first ring, it entirely flows along the second ring so that it is not interrupted. From this moment onwards the process inverts itself. The current flows along the second ring Cl Cl and then, due to the capacitive coupling, it is progressively 2Lp 2Z0 4 4 2Z0 transferred to the first ring. Finally, when the current reaches the second ring slot, it flows, again, entirely along the first ring, just as it did at the beginning of the process. A circuit that literally models this phenomenon can be Figure 2: Equivalent circuit. seen in Figure 3. (i) The first ring can simply be replaced by an inductor. (ii) Then, a series capacitor would replace the capacitive per unit length of the CPW, Lline and Cline, by the line coupling that transfers the current to the second ring. length, ll. However, since a symmetry simplification has (iii) Of course the second ring is also modeled by been considered, the inductance must be multiplied by 2. = an inductor. Since the second ring is the current Therefore, the line inductor will become 2Ll 2Llinell. destination, it must be connected to the previous Correspondingly, the capacitance must be divided by the capacitor. same factor. So the line capacitor will be now Cl/2 = Clinell/2. Furthermore, as can be clearly seen in Figure 1, the CPW (iv) Next, another capacitor can model the capacitive is longitudinally divided by a direct connection of the central coupling that brings the current back to the first ring. line and the ground planes. This connection can be modeled (v) Finally, the ring inductances are separated into two with an inductor, Lp. Since only half a circuit is considered series inductors in order to perform an appropriate because of the symmetry applied, this inductor is, indeed, coupling to the line divided by Lp, just as explained equal to 2Lp. before. The central inductor requires that the line be divided into 2 halves. Each half can be modeled with a serial inductor, Obviously, this ring model is exactly the same as Figure 2. whose value is Ll = Llinell, and a parallel capacitor equal to However, when the rings are explicitly separated, it is Cl/4 = Clinell/4. possible to incorporate to the model the capacitive coupling Finally, it should not be forgotten that the equivalent through the line. In order to do that, an additional capacitor, circuit must be excited with an adapted port. The char- Cg , connecting the rings, is added to the circuit, as can be acteristic port impedance must be equal to 2Z0,because seen in Figure 4. the complete coplanar line has been designed with a characteristic impedance equal to Z0. 3. Results As far as the ring is concerned, it is simply a resonator. Initially [5], it was modeled as an isolated parallel LC tank. In order to test the validity of the proposed new equivalent But in this model, the inductor is magnetically coupled with circuit a specific prototype cell has been designed. the CPW and that coupling excites the ring, and, since the This cell has been implemented over a Rogers 4003C line is broken into two parts by Lp, the ring must be divided substrate of 1.524 mm thick and designed to show a left- as well. Therefore, the ring inductance is divided into two handed transmission band of around 3 GHz and an input International Journal of Antennas and Propagation 3

0

La/4 La/4

Ca Ca −10 La/4 La/4

Figure 3: Ring equivalent circuit. −20 Scattering parameters (dB) Scattering parameters √ √ k/ 2 k/ 2

−30 2468 La/4 La/4 Frequency (GHz) Ca Cg Ca HFSS—s11 HFSS—s21 La/4 La/4 Circuit—s11 Circuit—s21 (a) √ √ 0 k/ 2 k/ 2

Ll Ll

−10 C C l 2Lp l 4 4 2Z0 2Z0

−20

Figure 4: Final equivalent circuit. (dB) Scattering parameters

= Ω −30 impedance of Z0 50 . The physical dimensions of the 2468 circuit have been calculated by applying the expressions Frequency (GHz) found in [10]. Then, this basic cell has been simulated by using the commercial software Ansoft HFSS and was slightly HFSS—s11 HFSS—s21 modified in order to tune the response exactly at 3 GHz. Circuit—s11 Circuit—s21 These modified dimensions are shown in Table 1. (b) Similarly, following the recommendations of [10], the components and the coupling coefficient of the original Figure 5: (a) Original equivalent circuit (Figure 2) versus HFSS and equivalent circuit (Figure 2) have been computed. These (b) equivalent circuit without considering frequency dependence values can be seen in the first row of Table 2. versus HFSS. Finally, the values have been modified by optimization in order to make the HFSS response and the equivalent circuit match around the passband. The optimized values are very of the original equivalent circuit (compare Figure 5(b) with similar to the original ones. This fact clearly validates this Figure 5(a)). equivalent circuit. In fact, the response of this new model and the real In Figure 5(a), the HFSS and the equivalent circuit response are almost identical under 6 GHz. Unfortunately, responses are compared and, as mentioned in the introduc- the equivalent circuit response progressively degrades, as tion, both responses are virtually identical around 3 GHz. the frequency is increased. Even so, the right-handed band, Unfortunately, the right-handed band that clearly appears in originally missing, now clearly manifests itself. The afore- the HFSS response around 7 GHz is completely missing in mentioned progressive discrepancy is due to the fact that the equivalent circuit response. the dispersive behavior of the CPW line has been completely In Figure 5(b) the new equivalent circuit response ignored. Therefore, it is necessary to consider the frequency (Figure 4) and the HFSS-simulated response are compared. dependence of the CPW per-unit-length inductance and The results, at this stage, are clearly better than the results capacitance if a more accurate response is desired. On 4 International Journal of Antennas and Propagation

Table 1: Basic cell dimensions. wc (mm) gc (mm) rr (mm) wr (mm) gr (mm) wl (mm) 10 0.64 3.83 0.4 0.4 1

Table 2: Equivalent circuit elements.

Ll (nH) Cl (pF) ll Lp (nH) La (nH) Ca (pF) k

1.793 0.65 2rr 0.2 18.45 0.299 0.271

1.793 0.65 2rr 0.2 18.96 0.299 0.246

0

−10

−20 Scattering parameters (dB) Scattering parameters

−30 2468 Figure 7: Fabricated prototype. Frequency (GHz)

HFSS—s11 HFSS—s21 Circuit—s11 Circuit—s21 (a) 8 0

6

−10 4 Frequency (GHz) Frequency

−20 2 Scattering parameters (dB) Scattering parameters 0 123 Phase (rad)

−30 HFSS 2468 Circuit Frequency (GHz) Measured

HFSS—s11 HFSS—s21 Figure 8: Dispersion diagrams. Measured—s11 Measured—s21 (b) Toachieve a better characterization of the actual behavior Figure 6: (a) Optimized equivalent circuit (Figure 4) versus HFSS, (b) Results versus HFSS. of this kind of cells, it has been decided to take into account the frequency dependency of these parameters. To simplify the other hand, since the coupling factor depends on the the characterization of the above-mentioned frequency per-unit-length inductance of the line [10], this makes it dependency, it has been assumed that it is basically linear. frequency dependent too. Therefore, to fully characterize this frequency dependencies International Journal of Antennas and Propagation 5

Table 3: Frequency dependency of some elements of circuit of Figure 4.

3 GHz Slope Optimized slope

Ll 1.793 nH 0.0531 nH/GHz 0.115 nH/GHz

Cl 0.65 pF −0.003 pF/GHz −0.003 pF/GHz k 0.246 0.0039 GHz−1 0.033 GHz−1 it is sufficient to establish the value of these parameters for 4. Conclusions a specific frequency, in this case 3 GHz, and determine the slope. The traditional equivalent circuit that models an SRR- These parameters have been estimated by using the HFSS loaded CPW has been accordingly modified in order to port solver. The resulting values are reflected in Table 3 take into account the additional capacitive coupling that (columns “3 GHz” and “slope”). happens through the CPW. When this additional coupling The capacitor modeling the capacitive coupling of the is considered, the circuit is able to characterize not only the left-handed propagation throughout the cell, but also rings through the CPW, Cg , has been considered a design parameter. By performing a parametric analysis, an initial the right-handed propagation. This right-handed band could not be modeled with the previous equivalent circuit. If the value for this capacitor has been computed. Specifically Cg = 0.125 pF has been selected as the starting value. right-handed propagation can be modeled by the equivalent Once the initial values have been calculated, the response circuit, the design of balanced cells with this structure will be of the circuit has been optimized. In this case only the slope notably easier. Wider bandwidths will thus be achieved, and of the frequency dependences of for the different parameters the application range of these devices will be improved. and Cg has been optimized. The rest of values are identical to those obtained after optimizing the original equivalent Acknowledgments circuit (Table 2). The optimized slopes can be seen in the last column of Table 3 and, finally, Cg = 0.14 pF. This work was supported by the Ministerio de Ciencia e The results obtained with this new equivalent circuit are Innovacion,´ Spanish Goverment, under Research Projects quite similar to the HFSS results (see Figure 6(a)). This fact TEC2010-21520-C04-03 and -01, and by the Autonomous validates the proposed broadband equivalent circuit. Government of Castilla-La Mancha under Research Projects Finally, a prototype of this cell has been built. In order PPII10-0047-0220 and PPII10-0027-1277. to be able to measure the cell, a taper and a connector have been added to the input and output ports. A photograph of this prototype (top and bottom) can be seen in Figure 7. References Next, in Figure 6(a) a comparison between the scattering [1]J.D.Baena,J.Bonache,F.Mart´ın et al., “Equivalent-circuit parameters given by HFSS and the measured ones is shown. models for split-ring resonators and complementary split- In the figure, it can be seen how the experimental results ring resonators coupled to planar transmission lines,” IEEE are slightly displaced upwards in frequency. This is basically Transactions on Microwave Theory and Techniques, vol. 53, no. a consequence of the fabrication process. The prototype has 4, pp. 1451–1460, 2005. been made using a milling machine. It is impossible to adjust [2]A.L.Borja,J.Carbonell,V.E.Boria,andD.Lippens,“Sym- the milling depth with total accuracy. Therefore, prototypes metrical frequency response in a split ring resonator based are always slightly overmilled. The main consequence of this transmission line,” Applied Physics Letters, vol. 93, no. 20, overmilling is this upward frequency displacement that can Article ID 203505, 2008. be noticed in the measurements. [3]A.L.Borja,A.Belenguer,J.Cascon,H.Esteban,andV.E. Boria, “Wide- band passband transmission line based on met- Finally, in Figure 8 one can see a comparison of the amaterial-inspired CPW bal- anced cells,” IEEE Antennas and dispersion diagrams of the equivalent circuit, HFSS results, Wireless Propagation Letters, vol. 10, no. 12, pp. 1421–1424, and experimental data. The dispersion diagrams have been 2011. calculated from (1) as explained in [12]: [4] A. L. Borja, J. Carbonell, V. E. Boria, and D. Lippens, “Highly selective left-handed transmission line loaded with split ring resonators and wires,” Applied Physics Letters, vol. 94, no. 14,   − Article ID 143503, 2009. −1 1 s11s22 + s12s21 βll = cos . (1) [5] F. Mart´ın, J. Bonache, F. Falcone, M. Sorolla, and R. Marques,´ 2s21 “Split ring resonator-based left-handed coplanar waveguide,” Applied Physics Letters, vol. 83, no. 22, pp. 4652–4654, 2003. [6]F.Aznar,J.Bonache,andF.Mart´ın, “Improved circuit model for left-handed lines loaded with split ring resonators,” Applied If the frequency displacement of HFSS results is not taken Physics Letters, vol. 92, no. 4, Article ID 043512, 2008. into account, one can say that the dispersion diagrams [7]L.J.Rogla,J.Carbonell,andV.E.Boria,“Studyofequivalent´ really resemble each other. It can be noticed that a left- circuits for open-ring and split-ring resonators in coplanar handed transmission band appears around 3 GHz and a waveguide technology,” IET Microwaves, Antennas and Prop- right-handed transmission band around 7 GHz. agation, vol. 1, no. 1, pp. 170–176, 2007. 6 International Journal of Antennas and Propagation

[8] A. L. Borja, J. Carbonell, V. E. Boria, J. Cascon, and D. Lippens, “A 2% bandwidth C-band filter using cascaded split ring resonators,” IEEE Antennas and Wireless Propagation Letters, vol. 9, pp. 256–259, 2010. [9]M.Gil,J.Bonache,J.Garc´ıa-Garc´ıa, J. Martel, and F. Mart´ın, “Composite right/left-handed metamaterial transmis- sion lines based on complementary split-rings resonators and their applications to very wideband and compact filter design,” IEEE Transactions on Microwave Theory and Techniques, vol. 55, no. 6, pp. 1296–1303, 2007. [10] R. Marques,F.Mart´ ´ın, and M. Sorolla, Metamaterials with Negative Parameters: Theory, Design and Microwave Applica- tions, John Wiley & Sons, 2008. [11] M. Shamonin, E. Shamonina, V. Kalinin, and L. Solymar, “Resonant frequencies of A split-ring resonator: analytical solutions and numerical simulations,” Microwave and Optical Technology Letters, vol. 44, no. 2, pp. 133–136, 2005. [12] C. Caloz and T. Itoh, Electromagnetic Metamaterials: Transmis- sion Line Theory and Microwave Applications, John Wiley & Sons, 2005. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 282159, 7 pages doi:10.1155/2012/282159

Research Article FDTD-SPICE for Characterizing Metamaterials Integrated with Electronic Circuits

Zhengwei Hao, Soheil Saadat, and Hossein Mosallaei

Electrical and Computer Engineering Department, Northeastern University, Boston, MA 02115, USA

Correspondence should be addressed to Zhengwei Hao, [email protected]

Received 24 April 2012; Revised 20 July 2012; Accepted 25 July 2012

Academic Editor: Alejandro Lucas Borja

Copyright © 2012 Zhengwei Hao 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 powerful time-domain FDTD-SPICE simulator is implemented and applied to the broadband analysis of metamaterials integrated with active and tunable circuit elements. First, the FDTD-SPICE modeling theory is studied and details of interprocess communication and hybridization of the two techniques are discussed. To verify the model, some simple cases are simulated with results in both time domain and frequency domain. Then, simulation of a metamaterial structure constructed from periodic resonant loops integrated with lumped capacitor elements is studied, which demonstrates tuning resonance frequency of medium by changing the capacitance of the integrated elements. To increase the bandwidth of the metamaterial, non-Foster transistor configurations are integrated with the loops and FDTD-SPICE is applied to successfully bridge the physics of electromagnetic and circuit topologies and to model the whole composite structure. Our model is also applied to the design and simulation of a metasurface integrated with nonlinear varactors featuring tunable reflection phase characteristic.

1. Introduction paper, the Finite Difference Time Domain (FDTD) technique is combined with SPICE-like software to provide a capable In recent years, the subject of metamaterials integrated with numerical technique for time-domain full wave analysis electronic circuits, nonlinear elements, and non-Foster tran- of the composite structure. The technique is applied to sistor configurations have been drawing significant atten- the design and simulation of metamaterials integrated with tion in research communities. Metamaterials are typically circuit elements. Broadband performance is obtained with periodic structures created by array of resonant elements the focus on bandwidth and tunable behavior engineering. which are small and stationary compared to the wavelength Section 2 provides the details of FDTD-SPICE hybridiza- of electromagnetic wave. The combination of metamaterials tion and shows the validation through some examples. In with circuit elements can bring novel and tunable charac- Section 3, the following three applications are explored: (a) teristics to these materials. For instance combining loop- loop-based metamaterial integrated with lumped capacitors based metamaterials with varactors provides a means to to tune the resonance performance, (b) metamaterial inte- electronically tune the performance by applying a DC voltage grated with non-Foster transistor configuration for band- or a high-power signal without modifying the structure [1– width manipulation, and (c) a metasurface combined with 4]. Alternatively, a non-Foster transistor configuration can varactors featuring tunable reflection phase. be introduced to enhance the metamaterial bandwidth [5]. Active components can be also integrated with transmission line metamaterials to compensate the loss [6], and finally 2. FDTD-SPICE Computational Model metasurface can be integrated with circuit elements to tailor their reflection phase performance [7]. 2.1. Hybridization of FDTD with SPICE. As a major compu- The complexity of the structure requires a powerful tational electromagnetic modeling scheme, Finite Difference computational technique to bridge the physics of electro- Time Domain technique can tackle problems by providing magnetic metamaterial with circuit configuration. In this a full wave solution in time. It is a capable method for 2 International Journal of Antennas and Propagation

to simulate any complex integrated configuration as long as + dV IIC (V) it is provided in SPICE’s library. Since conventional FDTD dt can only model very simple circuit elements, SPICE function adds a big advantage to FDTD. V SPICE element The SPICE software used here is SPICE3f5. It has an interactive mode that can receive commands from and send − results back to the interface. It also has “stop” and “resume” commands that can set a breakpoint after running some time Figure 1: Circuit model for Ampere’s Law equation that connects steps in a transient simulation and then resume it. These FDTD and SPICE. features make it very suitable to be connected to FDTD. To exchange data between FDTD and SPICE, interpro- cess communication should be developed. Figure 2 presents modeling a variety of scientific and engineering problems [8, a complete picture of our FDTD-SPICE model. When a 9]. On the other hand, the general-purpose analog electronic simulation begins, FDTD main program calls fork function circuit simulator, SPICE [10], can solve complex circuit which spawns two processes: the son process that will run topologies in time domain. It makes great sense to combine SPICE, and the parent process which is the original FDTD FDTD with SPICE to feature a powerful numerical engine for process itself. Two sockets: Socket[0] and Socket[1], are solving metamaterials integrated with circuit configurations. created in the main program before it forks so that both of FDTD uses Maxwell’s equations (1)and(2)intime the two children have access to the sockets. Here these sockets domain to recursively calculate the electromagnetic fields at act like two-way pipe, where FDTD process sends SPICE each position in three-dimensional space: commands to Socket[0] and reads results from Socket[1], ∂E while SPICE process reads message from Socket[0] and sends ε + J(E) =∇×H, (1) ∂t results back to Socket[1]. FDTD process will be waiting for SPICE outputs while SPICE process is running and vice ∂H μ + M(H) =∇×E. (2) versa. These two processes then communicate through the ∂t two sockets at each time step until the end of the simulation. While both equations can provide the coupling between While DC solution is needed in SPICE for many cases FDTD and SPICE according to reference [8], the first equa- but not very trivial to be obtained with FDTD, literature tion is used in this paper. After some simple manipulations, [12, 13] addressed DC solution problem by separating the Ampere’s Law of (1)canbewrittenas[11] AC and DC components, which is helpful for the metasurface simulations of diode varactors with bias that presented in dV Section 3. C + I(V) = I. (3) dt Our FDTD code has been written in a very powerful manner and it has many features such as modeling periodic Equation (3) describes a circuit model as shown in Figure 1, structures and linear dispersive metamaterials [14]. By = εA/ l where C d is a capacitance that abstracted from the applying auxiliary differential equation (ADE) method [15], model (note that here C is not the intrinsic lattice capacitance it can also model nonlinear dispersive materials as well as l A of FDTD cell), d is the cell size, and is the area of one gain materials. The absorbing wall is based on CPML [16] = A∗J E FDTD unit cell. I(V) ( ) is the current across the where one can truncate the computational domain very close = E∗ l SPICE element, and V d is the voltage across the to the structure. The code has been applied successfully to a = A∗∇×H element, I is the whole current of the current variety of metamaterials composites [9, 17–19]. source. Thus, for the cell containing SPICE element, instead of using (1) to update the electric field, SPICE can carry out (3) and obtain the voltage across the element and then pass 2.2. Test Results. Two cases are studied to validate the FDTD- it to FDTD program. From the voltage, it is easy to obtain E SPICE model. All simulations in this paper use uniform field afterward. gridding and are run on Sun Ultra45 workstation with 2 GHz To determine the voltage across the corresponding CPU and 4 GB memory. element, SPICE software needs to run a transient simulation In the first case as depicted in Figure 3(a), one parallel for a period of Δt, which should be the same as FDTD time RLC circuit is connected to a Gaussian pulse voltage source step. Then, in the next time step, regular FDTD method will in order to obtain the impedance. There are about 8000 time update all the fields except the cell with SPICE element. For steps in the calculation with each time step Δt = 0.2ps, the SPICE element, SPICE program needs to run another Δt, cell size Δx = 0.15 mm. Including CPML boundary, the starting from where it stopped last time. Before each run of whole computational domain is modeled by 26 × 26 × 26 SPICE, FDTD sends data that contains circuit information Yee’s cells. The FDTD process takes 456 seconds, while, since to SPICE. After each time step of SPICE simulation, it the circuit is simple, SPICE process takes only 10 seconds. passes results back to FDTD. In this way, the FDTD and Figure 3(b) shows both the FDTD-SPICE results and the SPICE engines run alternatively until the end of the whole analytical calculation for the real and imaginary parts of the simulation time. With the help of SPICE, FDTD is ready impedance Z,whichagreewellwitheachother. International Journal of Antennas and Propagation 3

FDTD engine starts, creates sockets Main program forks 2 processes

FDTD process SPICE process Initialize FDTD calculation Initialize SPICE calculation

No t

Yes Suspend while listening to Socket [0] Run to next Δt for regular FDTD part

Socket [0] Send SPICE commands to Socket [0] Hear from Socket [0]

Suspend, while listening to Socket [1]

Socket [1] Run to next Δt, Hear from Socket [1] send results to Socket [1]

Update field for SPICE elements

Figure 2: Flowchart for FDTD-SPICE simulation.

In the second case as shown in Figure 4(a),one structure in x direction where each layer has loops periodic BFG403W transistor from Discrete Semiconductor is utilized in y-z directions and the loops are terminated to capacitor in a single frequency simulation with a sinusoidal voltage Cp. The unit loop dimension is 4.8 mm and 1.8 mm in source. The BJT has a saturation current IS = 5.554 aA x and z directions, respectively. The excitation is a plane and the forward active current gain is BF = 145. The wave that has Ez/Hy components and propagates in −x calculation has 18000 time steps with each time step Δt = direction with a Gaussian signature. A transmission line 54.82 fs and cell size Δx = 30 μm. Because of the complex analogy can be obtained for plane wave propagation through property of BJT circuit, the cell size and time step are much the medium [19], which is embedded with a loop that has a smaller than those used in previous simulation. The whole self-inductance Lp and is terminated to a lumped capacitor computational domain has a size of 25 × 37 × 32 Yee’s Cp. By changing the circuit elements (Lp and/or Cp) in the cells. FDTD and SPICE processes take 1734 seconds and 41 equivalent transmission line model, the effective parameter seconds, respectively. Figure 4(b) shows the comparison for of the medium can be tuned, as well as the property of the the waveform of voltage across resistor R2 in time domain, metamaterial. between FDTD-SPICE and complete SPICE simulation. An FDTD-SPICE method is applied to model the structure excellent agreement is demonstrated. and obtain the transmission coefficientasshowninFig- ure 5(b). The loops are modeled in FDTD and the capacitors in SPICE. There are 15700 time steps in the calculation with 3. Applications on Metamaterials each time step Δt = 0.2 ps, cell size Δx = 0.15 mm. The whole computational domain has a size of 216 × 8 × 20 This section presents applications of FDTD-SPICE method Yee’s cells. FDTD and SPICE processes take 1669 seconds to some recent metamaterial designs, namely, loop-based and 1319 seconds, respectively. Because of the multiple-port composite metamaterial with capacitors, metamaterial inte- interprocess communication, SPICE process needs consider- grated with non-Foster circuits, and nonlinear elements able simulation time, which may be improved by introducing loaded metasurface. In all of the cases, uniform gridding, batch data communication between processes. Figure 5(b) perfectly matched layers, and periodic boundary conditions shows 4 simulation results. The red dots represent results are used. Since they all deal with plan wave incidence, total from pure FDTD simulation with Cp = 2.1pF(Cp is also field/scattered field method is also utilized [17]. modeled in FDTD). Then the blue line comes from FDTD- SPICE simulation with the same parameter, and it matches 3.1. Metamaterial Loops Integrated with Passive Elements. A well with red dots, which show the accuracy of our model. periodic structure made from metallic loops can artificially Changing the capacitor Cp allows tuning the resonance and create a medium with permeability characteristics (posi- as well as the equivalent permeability of the metamaterial. tive and negative) [19]. The metamaterial structure to be Figure 5(b) also shows the transmission coefficients for simulated here is shown in Figure 5(a). It is a four-layer the metamaterial loops terminated to Cp = 4.2pF and 4 International Journal of Antennas and Propagation

By FDTD By SPICE VCC = 10 V RC By SPICE V R L 5 k 1 1 1 C1 By FDTD R 2.5 k 1 Q 0.15 mV 20 Ω 2 nH 1 pF 1

V = R R 1 2V R2 E L (a) 25 10 k 5 k 10 k

20 f = 2 GHz 15

10 (a) 5 2

0

Impedance (Ohms) Impedance −5 1

−10

−15 0 246810 Voltage (V) Voltage Frequency (Hz) ×109 −1 Real(Z), analytical Real(Z), FDTD-SPICE Imag(Z), analytical Imag(Z), FDTD-SPICE

(b) −2 0 0.2 0.4 0.6 0.8 1 Figure 3: (a) RLC circuit configuration and (b) its impedance Z − behavior. Great comparison between FDTD results and analytical Time (s) ×10 9 model is illustrated. FDTD-SPICE SPICE (b) Cp = 8.4 pF lumped capacitors, depicted in black and magenta, respectively. Doubling capacitance would half the Figure 4: (a) Transistor circuit configuration. (b) Output voltage waveform in time. The FDTD-SPICE simulation matched very well resonance frequency. As can be seen from Figure 5(b), with pure SPICE simulation. the resonance frequency is shifted from 1.1 GHz down to 0.78 GHz and 0.55 GHz, respectively, according the value of capacitors. checked with the help of pole-zero analysis to ensure that the 3.2. Integration with Non-Foster Circuit Elements. Non- system has absolutely no pole in the Right-Half-Plane (RHP) Foster elements have recently attracted significant interests of the complex S-plane. due to their novel properties in offering negative impedance Here we explore the integration of NIC circuit with performance and successfully engineering the frequency dis- metamaterial loops [5] and its FDTD-SPICE simulation persion characteristics of antennas, transmission lines, and to investigate the bandwidth enhancement of the struc- metamaterials towards the goals of interests [20–23]. A non- ture. A two-layer loop-based composite metamaterial with Foster element can be realized using a Negative Impedance capacitors as shown in Figure 5(a) is utilized. The loops Converter (NIC) circuit by translating a given positive load are terminated to capacitors Cp and inductors L(4 nH). into a negative one. NIC circuit is made from transistor By integrating NIC designs offering −2 nH inductance, we elements and one would need a comprehensive modeling reduce the effect of the positive inductance and enhance the scheme to characterize them while they are integrated with bandwidth. The NIC circuit used in this paper is based on metamaterials. It is also worth mentioning that the design of Linvill’s model [24], and it is an example of Voltage Inversion NIC circuit involves several challenging issues such as noise, NIC (VINIC). As shown in Figure 6(a), the circuit consists of biasing, and stability and special care must be taken to obtain two N-type BJTs and a lumped passive inductor L1 = 2nH a wideband performance for non-Foster circuit. The NIC as the load. After adding the non-Foster circuit to the loop, stability (which is not the subject of this paper) has also been the equivalent circuit model [19] for the transmission line International Journal of Antennas and Propagation 5

R 1 L1 = 2nH R2 + V 1 k− 1 1 k I L 20 V 1

Cp Ip C Lp R p C ··· 8 10 k R7 10 k 1 C NIC Q 1 C2 1 Q2 Ln 3 mm 100 nF 100 nF z Output R3 R4 R5 R6 y 7505 k 5 k 750 1.2 mm 6.6 mm x (a) (a) 0 0

−5 −10 −10

− −15 20

−20 −30

−25 amplitude (dB) Transmssion Transmission amplitude (dB) Transmission

− −30 40 0 0.5 1 1.5 2 2.5 3 02468 Frequency (Hz) × 9 Frequency (Hz) ×109 10 C = L = Cp = 2.1pF, FDTD p 0.13 pF, lumped inductor 4 nH C = L = Cp = 2.1pF, FDTD-SPICE p 0.2 pF, NIC circuit total 2nH Cp = 4.2pF, FDTD-SPICE Cp = 0.2 pF, lumped inductor L = 2 nH Cp = 8.4pF, FDTD-SPICE (b) (b) Figure 6: (a) Basic non-Foster element: NIC configuration based Figure 5: (a) Metamaterial constructed from array of loops on Linvill’s design and equivalent circuit model for the loop- terminated to capacitors Cp and (b) its transmission coefficient. based metamaterial integrated with NIC negative inductor. (b) Transmission coefficient for metamaterial of loops terminated with (i) Cp = 0.13 pF, L = 4 nH; (ii) Cp = 0.2pF,L = 4 nH, NIC circuit Ln = −2 nH; (iii) Cp = 0.2pF,L = 2nH. is depicted in Figure 6(a), where the NIC configuration is included in a negative inductor Ln. The simulation has 11600 time steps with each time −L step Δt = 0.274 ps, cell size Δx = 0.15 mm. The whole of n can enable a wider bandwidth performance (as has computational domain has a size of 128 × 8 × 20 Yee’s been comprehensively detailed in [5]). Lastly, to validate the L = − cells. FDTD and SPICE processes take 686 seconds and 133 results of the second case, NIC circuit n 2 nH together L = seconds, respectively. Because of the complexity of the NIC with inductor 4 nH is replaced by one lumped inductor L = circuit model, SPICE needs more time than the test examples 2 nH. As shown in Figure 6(b), the black curve follows from previous section. Here three cases are performed. the behavior of the second case. Firstly, metamaterial with loop terminated to capacitor Cp = At the end we want to mention that the reason we have − 0.13 pF and lumped inductor L = 4nH is simulated. As considered a positive 4 nH inductor in series with the 2nH depicted by the blue curve in Figure 6(b),ithasaresonance non-Foster element is to make sure the time-domain SPICE around 2.4 GHz. Secondly, to broaden the bandwidth, a modeling is stable (the equivalent total inductance is positive, NIC circuit Ln = −2 nH is added to inductor L = 4nH, +2 nH). making the whole inductance 2 nH. In order to make a fair comparison, the capacitor is made Cp = 0.2 pF so that 3.3. Metasurface Integrated with Varactor Diodes. Metasur- the two simulations have the same resonance. As can be faces have great potential for making low profile antennas observed from the red curve in Figure 6(b), the metamaterial [25], and achieving an electronically tunable reflection phase integrated with non-Foster element has a wider bandwidth is of remarkable interest in this regard. This section presents than the case without it (blue curve). Increasing the value the simulation for a structure of metasurface integrated with 6 International Journal of Antennas and Propagation

b = 10.5mm 4.5

4 Unit cell 0.5 mm ε = 12 μ = 1 3.5 Diode 3 Z PEC 2.5 E k

d = (pF) Capacitance 2 a = 6.5mm 2mm H Y X 1.5

(a) 1 012345 180 Reverse voltage (V) 135 Figure 8: Performance data for SMV1232 Diode Varactor. 90

45

0 response, represents the one without varactor diode, while the others are for results that have diode with different − 45 bias. The varactor diode model used here is Skywork’s −90 SMV1232-079 Hyperabrupt Tuning Varactors [26]. When Reflection phase (deg) Reflection the diode is added to the structure, it increases the total −135 capacitance therefore decreases the resonance frequency of −180 the metasurface. As can been seen from Figure 8, typical 01234capacitance values of the varactor diode decrease when the Frequency (Hz) ×109 reverse bias voltage increases. The capacitance is a maximum No diode Bias = 2V while there is no reverse bias voltage. Thus for reverse bias Bias = 4V Bias = 1V with smaller voltage it should give larger shift in resonance Bias = 3V No bias frequency, which agrees well with the simulation results.

(b) 4. Conclusion Figure 7: (a) Metasurface structure with periodic square metal patches and its unit cell that loaded with varactor diodes. (b) This paper develops a hybridized FDTD-SPICE model and Reflection phase for the metasurface. Controlling the bias of the applies it to the simulations of some novel metamaterials diode tunes the resonance frequency. integrated with electronic circuits. The Maxwell’s equations are linked with SPICE through the interchange between electromagnetic fields and voltages across the SPICE element varactors. The metasurface used in this paper is similar to the nodes. Interprocess communication technique with socket one in [7], which is designed to act as an artificial magnetic is used as it is crucial for FDTD and SPICE to exchange conductor (AMC) and uses active NIC circuits as negative data at proper running time. Diagrammatic flowchart is elements to increase the bandwidth. Here varactor diodes discussed that demonstrates the detail of the execution are integrated in the metamaterial and make the metasurface process as well as the data exchanging with the help of tunable in resonance frequency. sockets. To show the validity of the model, two simple Figure 7(a) shows a metasurface structure consisting of cases that have analytical solution or pure SPICE results are periodic square metal patches and PEC ground plane with investigated. Then, three kinds of metamaterial are studied dielectric slab in between. It also presents the unit cell of and modeled by the FDTD-SPICE engine. First, loop- the metasurface loaded with varactor diodes. The structure based metamaterial with lumped capacitor is investigated is periodic in y and z directions and the incidence plane where changing the capacitance can vary the resonance wave is −x direction. The FDTD-SPICE method is applied frequency. Then, integration of metamaterial with non- to simulate the structure. There are 7000 time steps in the Foster circuit element, which provides negative impedance calculation with each time step Δt = 0.953 ps, cell size and increases the bandwidth, is studied and characterized. Δx = 0.5 mm. The whole computational domain has a size Lastly, metasurface simulation with integrated nonlinear of 45 × 21 × 21 Yee’s cells. FDTD and SPICE processes varactor elements shows a tunable reflection phase property take 386 seconds and 8 seconds, respectively. Figure 7(b) for this metamaterial. The technique is very powerful and its shows the reflection coefficient phase for the metasurface accuracy is validated well. We expect our scheme to be of without varactor diodes, as well as those with diodes that great engine for comprehensive simulation of metamaterials have different biases. The green line, as a reference frequency integrated with active and nonlinear circuits and allow International Journal of Antennas and Propagation 7 exploring new artificial materials combined with a lot of [13] N. Orhanovic, R. Raghuram, and N. Matsui, “Full wave anal- potential complex circuits. ysis of planar interconnect structures using FDTD-SPICE,” in Proceedings of the 51st Electronic Components and Technology Conference, pp. 489–494, June 2001. Acknowledgments [14] H. Mosallaei, “FDTD-PLRC technique for modeling of anisotropic-dispersive media and metamaterial devices,” IEEE This work was supported in part by the US Office of Transactions on Electromagnetic Compatibility, vol. 49, no. 3, Naval Research (ONR) Grant no. N00014-10-1-0264. The pp. 649–660, 2007. [15] M. Fujii, M. Tahara, I. Sakagami, W. Freude, and P. Russer, authors would like to thank Allen Taflove from Northwestern “High-order FDTD and auxiliary differential equation for- University for his kind discussion on the FDTD-SPICE mulation of optical pulse propagation in 2-D Kerr and model theory and acknowledge Vincent Thomas from Los Raman nonlinear dispersive media,” IEEE Journal of Quantum Alamos National Laboratory for his helpful information on Electronics, vol. 40, no. 2, pp. 175–182, 2004. SPICE and sockets programming. [16] J. A. Roden and S. D. Gedney, “Convolutional PML (CPML): an efficient FDTD implementation of the CFS-PML for arbitrary media,” Microwave and Optical Technology Letters, References vol. 27, pp. 334–339, 2000. [17] H. Mosallaei and Y. Rahmat-Samii, “Periodic bandgap and [1]D.A.Powell,I.V.Shadrivov,Y.S.Kivshar,andM.V. effective dielectric materials in electromagnetics: characteriza- Gorkunov, “Self-tuning mechanisms of nonlinear split-ring tion and applications in nanocavities and waveguides,” IEEE resonators,” Applied Physics Letters, vol. 91, no. 14, Article ID Transactions on Antennas and Propagation,vol.51,no.3,pp. 144107, 3 pages, 2007. 549–563, 2003. [2]I.V.Shadrivov,A.B.Kozyrev,D.W.VanDerWeide,andY.S. [18] H. Mosallaei and K. Sarabandi, “Magneto-dielectrics in elec- Kivshar, “Tunable transmission and harmonic generation in tromagnetics: concept and applications,” IEEE Transactions on nonlinear metamaterials,” Applied Physics Letters, vol. 93, no. Antennas and Propagation, vol. 52, no. 6, pp. 1558–1567, 2004. 16, Article ID 161903, 3 pages, 2008. [19] H. Mosallaei and K. Sarabandi, “Design and modeling of [3] D. Huang, E. Poutrina, and D. R. Smith, “Analysis of the patch antenna printed on magneto-dielectric embedded- power dependent tuning of a varactor-loaded metamaterial at circuit metasubstrate,” IEEE Transactions on Antennas and microwave frequencies,” Applied Physics Letters, vol. 96, no. 10, Propagation, vol. 55, no. 1, pp. 45–52, 2007. Article ID 104104, 3 pages, 2010. [20] W. Jendernalik, “A low-voltage CMOS negative impedance [4]I.V.Shadrivov,S.K.Morrison,andY.S.Kivshar,“Tunable converter for analogue filtering applications,” Bulletin of the split-ring resonators for nonlinear negative-index metamate- Polish Academy of Sciences, vol. 55, no. 4, pp. 419–423, 2007. rials,” Optics Express, vol. 14, no. 20, pp. 9344–9349, 2006. [21] R. M. Rudish and S. E. Sussman-Fort, “Progress in use of non- [5] S. Saadat, M. Adnan, H. Mosallaei, and E. Afshari, “Composite Foster impedances to. match electrically-small antennas and Metamaterial and Metasurface Integrated with Non-Foster arrays,” in Proceedings of the Antenna Applications Symposium, Active Circuit Elements: A Bandwidth-Enhancement Investi- pp. 89–108, Allerton Park, Ill, USA, September 2005. gation,” IEEE Transactions on Antennas and Propagation.In [22] P. Jin and R. W. Ziolkowski, “Broadband, efficient, electrically press. small metamaterial-inspired antennas facilitated by active [6] T. Jiang, K. Chang, L.-M. Si, L. Ran, and H. Xin, “Active near-field resonant parasitic elements,” IEEE Transactions on microwave negative-index metamaterial transmission line Antennas and Propagation, vol. 58, no. 2, pp. 318–327, 2010. with gain,” Physical Review Letters, vol. 107, no. 20, Article ID [23] S. Hrabar, I. Krois, and A. Kiricenko, “Towards active 205503, 5 pages, 2011. dispersionless ENZ metamaterial for cloaking applications,” [7] D. J. Kern, D. H. Werner, and M. J. Wilhelm, “Active negative Metamaterials, vol. 4, no. 2-3, pp. 89–97, 2010. impedance loaded EBG structures for the realization of ultra- [24] J. G. Linvill, “Transistor negative-impedance converters,” wideband artificial magnetic conductors,” in Proceedings of the Proceedings of the IRE, vol. 41, pp. 725–729, 1953. IEEE International Antennas and Propagation Symposium and [25] H. Mosallaei and K. Sarabandi, “Antenna miniaturization USNC/CNC/URSI North American Radio Science Meeting, vol. and bandwidth enhancement using a reactive impedance 2, pp. 427–430, June 2003. substrate,” IEEE Transactions on Antennas and Propagation, [8] A. Taflove and S. C. Hagness, Computational Electrodynamics: vol. 52, no. 9, pp. 2403–2414, 2004. The Finite-Difference Time-Domain Method, House, Norwood, [26] Skywork’s SMV1232-079 datasheet. Mass, USA, 3rd edition, 2005. [9] H. Mosallaei and Y. Rahmat-Samii, “Broadband characteri- zation of complex periodic EBG structures: an FDTD/Prony technique based on the split-field approach,” Electromagnetics, vol. 23, no. 2, pp. 135–151, 2003. [10] http://bwrc.eecs.berkeley.edu/classes/icbook/spice/. [11] V. A. Thomas, M. E. Jones, M. Piket-May, A. Taflove, and E. Harrigan, “Use of SPICE lumped circuits as sub-grid models for FDTD analysis,” IEEE Microwave and Guided Wave Letters, vol. 4, no. 5, pp. 141–143, 1994. [12] G. Kobidze, A. Nishizawa, and S. Tanabe, “Ground bouncing in PCB with integrated circuits,” in Proceedings of the IEEE International Symposium on Electromagneti Compatibility, vol. 1, pp. 349–352, August 2000. Hindawi Publishing Corporation International Journal of Antennas and Propagation Volume 2012, Article ID 823089, 7 pages doi:10.1155/2012/823089

Research Article Optical and Electrical Properties of Magnetron Sputtering Deposited Cu–Al–O Thin Films

Yongjian Zhang,1 Zhengtang Liu,1 Duyang Zang,2 Liping Feng,1 Xingsen Che,1 and Yanyan Li1

1 State Key Laboratory of Solidification Processing School of Materials Science and Engineering, Northwestern Polytechnical University, Xi’an 710072, China 2 Key Laboratory of Space Applied Physics and of Ministry of Education, School of Science, Northwestern Polytechnical University, Xi’an 710129, China

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

Received 27 May 2012; Accepted 9 July 2012

Academic Editor: Fuli Zhang

Copyright © 2012 Yongjian Zhang 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.

We have successfully prepared Cu–Al–O films on silicon (100) and quartz substrates with copper and aluminum composite target by using radio frequency (RF) magnetron sputtering method. We have related the structural and optical-electrical properties of the films to the sputtering area ratio of Cu/Al for the target (rCu/Al). The deposition rate of the film and rCu/Al can be fitted by an exponential function. rCu/Al plays a critical role in the final phase constitution and the preferred growth orientation of the CuAlO2 phase, thus affecting the film surface morphology significantly. The film with main phase of CuAlO2 has been obtained with rCu/Al of 45%. The films show p-type conductivity. With the increase of rCu/Al, the electrical resistivity decreases first and afterwards increases again. With rCu/Al of 45%, the optimum electrical resistivity of 80 Ω·cm is obtained, with the optical transmittance being 72%–79% in the visible region (400–760 nm). The corresponding direct band gap and indirect band gap are estimated to be 3.6 eV and 1.7 eV, respectively.

1. Introduction had fabricated p-type transparent CuAlO2 thin films by spin- on technique and reported the film had a conductivity of Transparent conducting oxide (TCO) films have been widely 2.4 Scm−1 with the optical band gap being 3.75 eV. Due to the used in the fields of flat panel displays, solar cells, touch excellent optical-electrical properties, CuAlO2 film is attracts panels, and other optoelectronic devices owing to their increasing research interest for the potential applications high electrical conductivity and optical transmittance in ranging from p-n junction to invisible circuits. visible region [1–3]. Up to now, however, most of the TCOs So far, various deposition techniques have been obtained are characterized by n-type conductivity. The lack employed to fabricate highly transparent conductive of p-type TCOs restricts the development of p-n junction CuAlO2 thin films, including chemical vapor deposition based device. Therefore, developing stable p-type TCOs (CVD) [8], pulsed laser deposition (PLD) [9, 10], sol-gel becomes the hot research topic [4, 5]. Kawazoe et al. [6] [11], and sputtering [12, 13]. Among these techniques, investigated delafossite-structured CuAlO2 and successfully RF magnetron sputtering takes the advantage of strong prepared CuAlO2 films using pulse laser deposition (PLD) adhesion between film and substrate, large area deposition, method in 1997. The obtained films are good p-type TCO low substrate , and good compatibility with materials, with room temperature electrical conductivity current microelectronics. However, various deposition being 0.095 Scm−1, optical transmittance being 80% and parameters such as oxygen partial pressure, the variety of direct band gap being 3.5 eV. Alternatively, Gao et al. [7] sputtering target, and sputtering power may influence the 2 International Journal of Antennas and Propagation

properties of the films. Furthermore, most of the CuAlO2 Figure 1 illustrates the effect of rCu/Al on the deposition rate films deposited by RF sputtering method are always using RD of the Cu–Al–O films on Si (100) substrate. RD increases −1 −1 high-cost CuAlO2 ceramic target. In this work, we simplify from 1.13 nm·min to 1.46 nm·min with the increase of the preparation process by using the low-cost copper and rCu/Al from 20% to 55%. The results can be fitted by an aluminum composite target instead of CuAlO2 ceramic exponential function as target. We investigate the influence of sputtering area ratio = − · of Cu/Al for the target (rCu/Al) on the properties of obtained RD 1.5-2.3exp( 0.1 rCu/Al). (1) films. We elucidate the underlying mechanisms between the film structure and the optical band gap. The deposition rate RD increases with the increase of rCu/Al is mainly due to that the sputtering yield of Cu is higher than that of Al. In addition, the sputtered Cu atom possesses 2. Experimental more energy than Al, thus favoring the formation of defect and nucleation center on the substrate. This also contributes Cu–Al–O films were deposited on silicon (100) and quartz to the increase of RD. substrates, respectively, by RF magnetron sputtering method Figure 2 plots the X-ray diffraction spectra of Cu–Al– ∼ ◦ at room temperature ( 22 C). High purity of 99.999% O films deposited with different rCu/Al. When rCu/Al is 20%, copper and aluminum composite target was used as the the diffraction peaks corresponding to CuAlO2 (104), (015), sputtering materials. The composition of the film was (009), (116) and Al2O3 (113), (306) are observed, indicating controlled by changing the sputtering area ratio rCu/Al of Cu an excess of Al element exists in the film. When rCu/Al and Al for the target. Pure argon and oxygen were used as increases to 45%, the CuAlO2 (018) peak grows remarkably sputtering gas and reactive gas, respectively. The substrates while Al2O3 peaks tend to reduce. CuAlO2 becomes the main were cleaned ultrasonically in 5% (volume content) HF, phase of the film. The change may be due to the following acetone and ethanol for silicon, in acetone, and ethanol reaction [16]: for quartz before being loaded into the chamber. The HF solution was stored in closed plastic container, and Al2O3 +Cu2O −→ CuAlO2 (2) it was used following the safety rules [14, 15], such as ◦ wearing special respirator and gloves to prevent the HF When the rCu/Al reaches 55%, a new peak at 36.4 which from contacting our skin. Before deposition, base pressure is identified to Cu2O (111) emerges, suggesting the surplus of the chamber was evacuated to 4 × 10−4 Pa by rotary and of Cu element in the film. molecular pump. During deposition process, the working The rCu/Al also plays an important role in the preferred pressure was maintained at 0.3 Pa and sputtering power was growth orientation of CuAlO2 diffraction peaks. As seen fixed at 80 W. We varied the rCu/Al over the range 20%–55% from Figure 2,withrCu/Al of 20%, CuAlO2 phase shows to ensure the composition changed from Al being excessive to strong peak along (104) and (015) crystal planes, while the Cu being excessive. The thickness of the films was controlled X-ray diffraction peak of (018) is weak. With rCu/Al increases being 300 ± 10 nm via the deposition duration time. Before to 45%, the peak of CuAlO2 (018) increases significantly characterized the properties, the samples were annealed in and becomes the strongest, suggesting that the preferred GSL-1400X tubular furnace with argon ambience for 3 h. growth orientation of CuAlO2 is (018) with this rCu/Al. The thickness of the films was measured by UVISEL When the rCu/Al is 55%, (018) peak of CuAlO2 weakens and ER wide spectral range Ellipse leaning meter. The structural the preferential growth changes into (104). Although the character was identified by using X’Pert Pro MPD X-ray surface energy of (001) crystal plane might be the lowest in diffractometer with Cu Kα (λ = 0.15406 nm) radiation. The delfossite structure CuAlO2 crystal, the kinetic parameters, surface morphology and chemical compositions were char- for instance, annealing treatment, may also play a role in the acterized by ZEISS-SUPRA-55 scanning electron microscope selection of the preferred growth orientation. (SEM) and OXFORD INCA PentaFET×3 energy dispersive The grain size can be estimated from the full-width half- spectrometer (EDS). An X-ray photoelectron spectroscopy maximum intensity of XRD peak by using Scherrer’s relation (XPS) apparatus (PHI-5400) was employed to determine [17]: the chemical valence of the elements. The conductivity type kλ was identified by HMS-7077 measurement system. Room d = , (3) temperature resistivity of the films was investigated by the β cos θ four-probe method in Agilent 4155c measurement system. where k is a constant of 0.89 for Cu target, λ = 0.15406 nm, UV-3150 spectrophotometer was used to measure the optical θ and β are the Bragg diffraction angle and half intensity transmittance of the films. width. The calculated grain sizes of the films are estimated to be 12.6 nm, 14.1 nm, 17.4 nm, and 15.2 nm for rCu/Al of 3. Results and Discussion 20%, 30%, 45%, and 55%, respectively. Figure 3 displays the typical SEM images and the The deposition rate RD is one of the most important parame- corresponding EDS spectra of the films deposited with ters of the deposition process, which plays an important role different rCu/Al on Si (100) substrate. With rCu/Al being 20%, in the structure and the properties of the films. It can be a large amount of globular precipitation phases have been obtained through dividing the thickness by deposition time. observed, as shown in Figure 3(a). Figure 3(b) illustrates the International Journal of Antennas and Propagation 3

1.5 (104) 1.4 Cu/Al = 55% (015) (009) (111) (116) (018)

1.3 = (nm/min) Cu/Al 45% D R Intensity (a.u.) Intensity 1.2 = Cu/Al 20% (113) (306)

1.1 20 30 40 50 60 20 30 40 50 60 70 80 rCu/Al (%) 2θ deg. Fitt. CuAlO Exp. 2 Al2O3 Figure 1: Deposition rate RD of the films as a function of sputtering Cu2O area ratio of Cu/Al for sputtering target (rCu Al). / Figure 2: XRD patterns of Cu–Al–O thin films deposited with different rCu/Al on Si (100) substrate.

EDS spectrum of the globular phases, showing the atomic ratio of Al:O is around 2 : 3. This suggests that the globular The conductivity type of the films deposited on quartz phase is Al2O3. Figure 3(c) demonstrates the image of the substrate was measured by Hall effect measurement and the film deposited with rCu/Al being 45%. The film shows a electrical resistivity (ρ) at room temperature was studied uniform microstructure with well-defined grain boundaries, by four-probe method. Prior to the investigation, four Au no impurity is observed. EDS spectrum of the film signifies electrodes were deposited on the film surface. the atomic ratio of Cu : Al : O is about 1 : 1 : 2, confirming Figure 5 shows the electrical resistivity (ρ) of the films the XRD analysis that CuAlO2 is the main phase of the formed with different rCu/Al and the inset demonstrates the film. When the rCu/Al increases to 55%, a nonfaceted phase relation between current and voltage for the film deposited is observed. EDS analysis of this phase shows that the with rCu/Al of 45%. From the inset I-V curve, it can be atomic ratio of Cu : Al : O is about 12 : 1 : 5, indicating the seen the linear dependence is obtained, which indicates precipitation phase is mainly composed of copper oxide, ohmic contact has been achieved between Au electrode which is consistent with the XRD result. and the film. With rCu/Al being 20%, the sample shows To further identify the chemical compositions and a high electrical resistivity due to the existence of large valences of the elements, we performed XPS analysis to the amount of insulating Al2O3 in the film [18]. When rCu/Al films deposited on Si (100) substrate. Figures 4(a)–4(c) show increases from 20% to 45%, the electrical resistivity (ρ) the typical XPS spectra of the Cu–Al–O film obtained with decreases from 243 Ω·cm to 80 Ω·cm. The reason may be rCu/Al = 45% after the calibration using C 1s position of that the improvement of crystallization quality reduces the carbon. As shown in Figure 4(a), the “shake-up” peak of scattering and trapping of charge carriers, leading to the 2+ the Cu 2p3/2 at around 943 eV is not observed, indicating enhancement of Hall mobility. Furthermore, the increment 2+ that no Cu presents in the film. Figure 4(b) shows the of CuAlO2 increases the carrier concentration of the film. Cu 2p3/2 peak together with the two separated peaks by With rCu/Al being 55%, the electrical resistivity (ρ) increases using the multipeaks fitting. The peak at the low binding to 156 Ω·cm. In this case, surplus copper element exists in + energy of 931.7 eV is corresponding to Cu in CuAlO2, the film and the copper vacancy which can produce hole while the high binding energy 932.8 eV is corresponding carrier concentration decreases. In addition, the emergence to Cu2O. The intensity of the low-energy peak (931.7 eV) of Cu2O impurity strengthens the scattering and trapping of is remarkably higher than that of the high-energy peak charge carriers, decreasing the Hall mobility. + (932.8 eV), suggesting Cu mainly exists in CuAlO2 phase. Figure 6 presents the optical transmittance spectra of The Al 2p peak region, shown in Figure 4(c), consists of Al the Cu–Al–O thin films deposited with different rCu/Al on 3+ 2p peak of Al (around 74.2 eV), Cu 3p3/2 (around 75.3 eV), quartz substrate. As can be seen, the film deposited with + and Cu 3p1/2 (77.1 eV) peaks of Cu , which is similar to the rCu/Al of 20% exhibits the highest transmittance (77%–84%) result reported by Cai et al. [16]. in the visible region (400–760 nm). It may be due to the The Cu 2p spectra of the other films are similar to large amount of Al2O3 precipitation phase, which has quite that shown in Figure 4(a) where no Cu2+ peaks have been high transmittance in the visible range, existing in the film. observed. This is consistent with the XRD results: no CuO or With rCu/Al being 30%, a decrease (58%–76%) in the film CuAl2O4 diffraction peak is observed in the XRD patterns. transmittance was observed. When rCu/Al increases to 45%, 4 International Journal of Antennas and Propagation

Si Spectrum 2

O Al

0.5 1 1.5 2 2.5 3 (keV) Full scale 21289 cts cursor: 0.000 (a) (b)

Si Spectrum 1

O Cu Al

0.5 1 1.5 2 2.5 3 (keV) Full scale 19129 cts cursor: 0.000 (c) (d)

Spectrum 2

Cu

Si O Al

0.5 1 1.5 2 2.5 3 (keV) Full scale 19129 cts cursor: 0.000 (e) (f)

Figure 3: SEM micrographs and EDS spectra of Cu–Al–O thin films obtained with different rCu/Al: (a) and (b) rCu/Al = 20%; (c) and (d) rCu/Al = 45%; (e) and (f) rCu/Al = 55%. the transmittance of the film increases to 72%–79% in the To further investigate the optical properties, we evaluated visible region (400–760 nm) due to that the CuAlO2 becomes the optical band gap (Eg ) of the Cu–Al–O thin films. the predominant phase. In addition, the decrease of defect The optical absorption coefficient (α) of the films can be density and crystallization improvement of the films also calculated using the following equation: contribute to the improvement of optical transmittance. When rCu/Al reaches 55%, the transmittance decreases again,   mainly because the coexistence of Cu2O phase strengthen the 1 1 scattering effect, lowering the optical transmittance [18]. α = ln ,(4) d T International Journal of Antennas and Propagation 5

0.06 r = 45% Cu 2p3/2 300 Cu/Al 0.03

0 Cu 2p1/2 240 − Voltage (V) Voltage 0.03 Intensity (a.u.) Intensity

cm) −0.06 · 180 −8 −40 4 8 Ω ( Applied current (nA) ρ 930 940 950 960 970 Binding energy (eV) 120 (a)

Cu 2p (CuAlO ) 60 3/2 2 20 30 40 50 60 931.7 eV rCu/Al (%)

Figure 5: Variation of electrical resistivity with different rCu/Al.Inset Cu 2p3/2 (Cu2O) 932.8 eV shows the I-V relation for the sample deposited with rCu/Al of 45%. Intensity (a.u.) Intensity

926 928 930 932 934 936 938 100 Binding energy (eV) (b) 80 Al 2p 74.2 eV 60

40 Cu 3p3/2 Cu 3p1/2 Intensity (a.u.) Intensity 75.3 eV 77.1 eV Optical transmittance (%) 20 66 69 72 75 78 81 84 Binding energy (eV) 200 400 600 800 1000 (c) Wavelength (nm)

Figure 4: XPS spectra of Cu–Al–O thin films obtained with rCu/Al 20% rCu/Al 45% rCu/Al of 45%: (a) Cu 2p spectra; (b) fitting spectra of Cu 2p3/2; (c) fitting 30% rCu/Al 55% rCu/Al spectraofAl2pandCu3p. Figure 6: Optical transmittance spectra of Cu–Al–O films deposited with different rCu/Al. where d is the film thickness and T is the transmittance of the film. The relation between optical absorption coefficient (α) and optical band gap (Eg )canbewrittenas   1/n (αhν) = A hν − Eg ,(5)with rCu/Al reaching 55%. Egi varies in the range of 1.6–1.9 eV and achieves the minimum with r of 45%. ν Cu/Al where A is the absorption edge width parameter and h Egd and Egi may be influenced by the phase constitution means the incident photon energy. The exponential n is 1/2 of the films. With rCu/Al of 20%, the film is composed of or 2 for direct allowed transition (Egd) or indirect allowed Al2O3 and CuAlO2 phases, hence, the optical band gap of transition (Egi). the film can be evaluated by the superposition of pure Al2O3 Figure 7 shows a typical linear fitting process of Eg for and CuAlO2, whose Egd are 9.0 eV [19]and3.5eV[6, 20], = the Cu–Al–O thin film deposited at rCu/Al 45%.Egd and respectively. The direct band gap of the film which consists ν Egi are obtained from the intercept on h axis in the plots of of Al2O3 and CuAlO2 is assumed to be in the range of 3.5– (αhν)2-hν and (αhν)1/2-hν, respectively. Figure 8 compares 9.0 eV. This is in agreement with our result 5.3 eV. For the the Egd and Egi values of the films deposited with different film deposited with rCu/Al of 45%, the main crystal phase of rCu/Al.TheEgd decreases from 5.3 eV to 3.6 eV with increase the film is CuAlO2 and the estimated Egd (3.6 eV) is close to of rCu/Al from 20% to 45%, afterwards, it increases to 4.7 eV the Egd of pure CuAlO2 (3.5 eV) [6, 20]. 6 International Journal of Antennas and Propagation

5.3 eV, indicating our results is consistent with the model. This suggests that quantum size effect resulted from the nano size grain structure may play a role in the optical band gap of (a.u.)

2 the film. / 1 ) A αh (

(a.u.) = 4. Conclusions

2 Egi 1.7eV ) A

αh 123456 Cu–Al–O thin films have been deposited on Si (100) and ( hA (eV) quartz substrates by RF magnetron sputtering technique. The sputtering area ratio of Cu/Al for the sputtering target (rCu/Al) plays an important role in the structure, optical- = Egd 3.6eV electrical properties and optical band gaps of the films. The deposition rate RD increases with the increase of 123456 rCu/Al mainly because of the higher sputtering yield of Cu hA (eV) than Al. With rCu/Al of 20%, CuAlO2 and Al2O3 phases coexist in the film due to the surplus Al element. CuAlO2 Figure 7: Plots of (αhν)2 versus hν for the determination of direct becomes the main phase of the film when rCu/Al reaches band gap (Egd) for the film deposited with rCu/Al of 45% (inset: 45%. Whereas when rCu/Al increasesto55%,aswellasthe determination of indirect band gap Egi). + CuAlO2,Cu2Odiffraction peak also be detected. Cu in the films deposited with different rCu/Al exists in the form of 2+ CuAlO2 or Cu2O, no Cu has been observed. The films show stable p-type conductivity. With the increase of r , 6 Cu/Al the electrical resistivity first decreases afterwards increases. With rCu/Al of 45%, the film shows the optimum optical- electrical properties. The electrical resistivity is measured Ω· 4 to be 80 cm with the transmittance being 72%–79% in the visible region(400–760 nm). The estimated Egd is in the range of 3.6–5.3 eV and Egi in the range of 1.6–1.9 eV which

Optical band gap (eV) depends on rCu/Al. 2 Acknowledgments This work was supported by the Shaanxi Provincial Natural 20 30 40 50 60 Science Foundation of China (Grant no. 2012JQ1016), the rCu/Al (%) Research Fund of the State Key Laboratory of Solidifi- cation Processing (Contract nos. 58-TZ-2011 and SKLSP Egd Egi 201217) and the Northwestern Polytechnical University (NPU) Foundation for Fundamental Research (Contract ff Figure 8: E ect of rCu/Al on the optical band gap of the film: (a) nos. JC20100242 and JC20110245). The authors are grateful direct band gap Egd; (b) indirect band gap Egi. to Dr. Y. P. Li and Mr. H. Yuan, for their help with the experiments and analysis.

Moreover, quantum size effect may also affect the band gap, which can be described by the following equation [20]: References [1] M. H. Ahn, E.-S. Cho, and S. J. Kwon, “Effect of the duty ratio   ∞   π22 1 1 1.8e2 e2  S 2n on the indium tin oxide (ITO) film deposited by in-line pulsed = − (6) Eg 2 + + αn , DC magnetron sputtering method for resistive touch panel,” 2R me mh ε2R R = R n 1 Applied Surface Science, vol. 258, no. 3, pp. 1242–1248, 2011. ff R is the radius of the semiconductor particle and the first [2] M. F. Chen, K. M. Lin, and Y. S. Ho, “E ects of laser-induced term is the quantum energy of localization for both electron recovery process on conductive property of SnO2 : F thin films,” Materials Science and Engineering B, vol. 176, no. 2, pp. and hole. The second term is the Coulomb attraction 127–131, 2011. and the third term represents the band gap of the bulk [3] K. Tonooka, K. Shimokawa, and O. Nishimura, “Properties of semiconductor. As is shown in the model, the change copper-aluminum oxide films prepared by solution methods,” tendency of Eg and R is reverse, that is, Eg should be smaller Thin Solid Films, vol. 411, no. 1, pp. 129–133, 2002. for larger R. The estimated results show that with the largest [4] K. Nomura, H. Ohta, K. Ueda, T. Kamiya, M. Hirano, and H. = grain size 17.4 nm (rCu/Al 45%), the Egd achieves the Hosono, “Thin-film transistor fabricated in single-crystalline minimum value 3.6 eV, while with the minimum grain size transparent oxide semiconductor,” Science, vol. 300, no. 5623, 12.6 nm (rCu/Al = 20%), the Egd obtains the maximum value pp. 1269–1272, 2003. International Journal of Antennas and Propagation 7

[5] G. Thomas, “Invisible circuits,” Nature, vol. 389, no. 6654, pp. 907–908, 1997. [6] H. Kawazoe, M. Yasukawa, H. Hyodo, M. Kurita, H. Yanagi, and H. Hosono, “P-type electrical conduction in transparent thin films of CuAlO2,” Nature, vol. 389, no. 6654, pp. 939–942, 1997. [7] S. M. Gao, Y. Zhao, P. Gou, N. Chen, and Y. Xie, “Preparation of CuAlO2 nanocrystalline transparent thin films with high conductivity,” Nanotechnology, vol. 14, no. 5, pp. 538–541, 2003. [8] H. Gong, Y. Wang, and Y. Luo, “Nanocrystalline p-type trans- parent Cu–Al–O semiconductor prepared by chemical-vapor deposition with Cu(acac)2 and Al(acac)3 precursors,” Applied Physics Letters, vol. 76, no. 26, pp. 3959–3961, 2000. [9] M. Neumann-Spallart, S. P. Pai, and R. Pinto, “PLD growth of CuAlO2,” Thin Solid Films, vol. 515, no. 24, pp. 8641–8644, 2007. [10]J.C.Lee,S.Y.Um,Y.W.Heo,J.H.Lee,andJ.J.Kim, “Phase development and crystallization of CuAlO2 thin films prepared by pulsed laser deposition,” Journal of the European Ceramic Society, vol. 30, no. 2, pp. 509–512, 2010. [11] S. Gotzend¨ orfer,¨ C. Polenzky, S. Ulrich, and P. Lobmann,¨ “Preparation of CuAlO2 and CuCrO2 thin films by sol-gel processing,” Thin Solid Films, vol. 518, no. 4, pp. 1153–1156, 2009. [12] W. Lan, M. Zhang, G. Dong, P. Dong, Y. Wang, and H. Yan, “The effect of oxygen on the properties of transparent conducting Cu–Al–O thin films deposited by rf magnetron sputtering,” Materials Science and Engineering B, vol. 139, no. 2-3, pp. 155–159, 2007. [13] H. Y. Chen and M. W. Tsai, “Delafossite-CuAlO2 films prepared by annealing of amorphous Cu–Al–O films at high temperature under controlled atmosphere,” Thin Solid Films, vol. 519, no. 18, pp. 5966–5970, 2011. [14] D. Peters and R. Miethchen, “Symptoms and treatment of hydrogen fluoride injuries,” Journal of Fluorine Chemistry, vol. 79, no. 2, pp. 161–165, 1996. [15] E. B. Segal, “First aid for a unique acid, HF: a sequel,” Chemical Health and Safety, vol. 7, no. 1, pp. 18–23, 2000. [16] J. Cai and H. Gong, “The influence of Cu/Al ratio on properties of chemical-vapor-deposition-grown p-type Cu– Al–O transparent semiconducting films,” Journal of Applied Physics, vol. 98, no. 3, Article ID 033707, 5 pages, 2005. [17] B. D. Cullity, Elements of X-Ray Diffraction, Addison Wesley, London, UK, 2nd edition, 1978. [18] A. S. Reddy, P. S. Reddy, S. Uthanna, and G. M. Rao, “Characterization of CuAlO2 films prepared by dc reactive magnetron sputtering,” Journal of Materials Science, vol. 17, no. 8, pp. 615–620, 2006. [19] M. Henyk, D. Wolfframm, and J. Reif, “Ultra short laser pulse induced charged particle emission from wide bandgap crystals,” Applied Surface Science, vol. 168, no. 1–4, pp. 263– 266, 2000. [20] L. E. Brus, “Electron-electron and electron-hole interactions in small semiconductor crystallites: the size dependence of the lowest excited electronic state,” The Journal of Chemical Physics, vol. 80, no. 9, pp. 4403–4409, 1984.