Electromagnetic Materials: Proceedings of the Symposium P, ICMAT 2007

Proceedings of the Symposium P

ELECTROMAGNETlC MATERIALS

This page intentionally left blank

ICMAT 2007 International Conference on Materials for Advanced Technologies

Proceedings of Symposium P

E lectromagnetc Materials 1 - 6 July 2007

SUNTEC, Singapore

Edited by

Lim Hock, Serguei Matitsine, Gan Yeow Beng and Kong Ling Bing Temasek Laboratories, National University of Singapore

K World Scientific NEW J E R S E Y * L O N D O N * SINGAPORE

BElJlNG * S H A N G H A I * HONG KONG * T A I P E I

-

CHENNAI

Published by

World Scientific Publishing Co. Re. Ltd. 5 Toh Tuck Link, Singapore 596224 USA ofice: 27 Warren Sbeet, Suite 401-402, Hackensack, NJ 07601 UK ofice: 57 Shelton Sbeet, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-PublicationData A catalogue record for this book is available from the British Library.

ELECTROMAGNETIC MATERIALS Proceedings of the Symposium P, ICMAT 2007 Copyright 0 2007 by World Scientific Publishing Co. Re. Ltd All rights reserved. This book, or parts thereoj may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN-13 978-981-270-953-0(pbk) ISBN-I0 981-270-953-3(pbk)

Printed by FuIsland Offset Printing (S) Pte Ltd, Singapore

V

FORE WORD

International Conference on Materials for Advanced Technologies (ICMAT) is a biannual conference organized by the Materials Research Society, Singapore, in association with National University of Singapore, Institute of Materials Research and Engineering, and Nanyang Technological University. The inaugural ICMAT in 2001 was attended by about 1,500 international delegates, including four Nobel Laureates. Since then, ICMAT has continued to attract the active participation of Nobel Laureates and leading international scientists and engineers, and its attendance has steadily grown to about 2,300 in 2005. We started organizing a symposium on Electromagnetic Materials for ICMAT in 2003. This symposium was dedicated to the studies of materials/structures that exhibit electromagnetic effects. A collection of papers of high quality were presented, on theoretical research, characterization and measurement techniques, design and fabrication methods, and special applications. Encouraged by the success, we continued the symposium in 2005. At ICMAT 2007, we are pleased to run the symposium for the third time. We have made many friends at the two earlier symposia. We look forward to welcoming them, and the many new participants of this and other symposia of ICMAT, in early July 2007. On behalf of the Organizing Committee of this Symposium, I wish to thank the ICMAT 2007 Conference Committee for organizing such a successful international conference. We are very pleased to be able to play a small supporting role with this symposium in a niche area. Our invited speakers have, despite their busy schedules, kindly accepted our invitation. We extend to them and the many authors who submitted papers to the symposium our sincere appreciation. It is their valuable contributions that make this symposium an exciting event. The Scientific Programme Committee must take the credit for the efficient and stringent reviewing process, and for putting up the excellent programme. World Scientific Publishing Co. Pte. Ltd., in its usual professional manner, has published this handsome volume of the proceedings ready for our participants at the opening of ICMAT 2007. I wish all participants a fruitful and stimulating conference, and our guests from overseas a pleasant and enjoyable visit to Singapore.

Professor LIM Hock Chair Symposium P (Electromagnetic Materials) ICMAT 2007

vi

Symposium P: ELECTROMAGNETIC MATERIALS Chair: LIM Hock Temasek Laboratories, National University of Singapore, Singapore Co-Chair: CAN Yeow Beng Temasek Laboratories, National University of Singapore, Singapore Co-Chair: Konstantin N. ROZANOV Institute for Theoretical and Applied Electromagnetics, Russia Co-Chair: LEE Kim Seng DSO National Laboratories, Singapore

SCOPE OF SYMPOSIUM The Symposium deliberates on the electrical and magnetic properties of materials relevant to the design of unconventional antennas, microwave circuitslcomponents, anti-reflection media and coatings, EM1 shielding structures, radomes, etc. Though a classical research topic, some recent advancement in technologies has led to new capabilities to create and control fine-scale structures. This has inspired scientists to develop new materials with exceptionally high permittivity or permeability, as well as metamaterials (or negative index materials) with unusual electromagnetic properties. Novel materials based on the use of active devices to control their electromagnetic performances have also been proposed. The multi-disciplinary nature of these new materials has brought together researchers from materials science, physics and electrical engineering to explore and deepen our current understanding of electromagnetic wave propagation. A wide range of new commerciaVdefence applications of these materials is expected to emerge in the near future. Topics of Interest:

0 0

Dielectric and Magnetic Composites (with micro- or nano-sized inclusions) Metamaterials (Negative Index Materials) Smart Materials (includes thin films, tunable dielectrics, etc) Frequency Selective Periodic Structures Material Processing and Fabrication Techniques Characterization of Electromagnetic Properties of Materials

vii

INVITED SPEAKERS 0 0 0 0 0 0

0 0

Olivier ACHER, CEA Le Ripault, France Luk ARNAUT, National Physical Laboratory, UK DENG Longjiang, University of Electronic Science and Technology of China, China Jin Au KONG, Massachusetts Institute of Technology, USA Andrei N. LAGARKOV, Institute for Theoretical and Applied Electromagnetics, Russia Akhlesh LAKHTAKIA, Pennsylvania State University, USA Benedikt A. MUNK, Ohio State University, USA Sergey A. NIKITOV, Institute of Radioengineering and Electronics, Russia Konstantin N. ROZANOV, Institute for Theoretical and Applied Electromagnetics, Russia YAO Xi, Tongji University, Shanghai, China

TECHNICAL PROGRAMME COMMITTEE 0

0 0

0 0

GAN Yeow Beng, Temasek Laboratories, National University of Singapore KONG Ling Bing, Temasek Laboratories, National University of Singapore LI Zheng-Wen, Temasek Laboratories, National University of Singapore LIU Lie, Temasek Laboratories, National University of Singapore Serguei MATITSINE, Temasek Laboratories, National University of Singapore QING Anyong, Temasek Laboratories, National University of Singapore RAO Xuesong, Temasek Laboratories, National University of Singapore

PUBLICATION AND LIAISON COMMITTEE

0 0

Karrie CHAN, Temasek Laboratories, NUS, Singapore Suhana HANAN, Temasek Laboratories, NUS, Singapore KONG Ling Bing, Temasek Laboratories, NUS, Singapore Irene LEOW, Temasek Laboratories, NUS, Singapore

This page intentionally left blank

ix

CONTENTS Session P1: Opening Session

Chair: S.A. Nikitov

P-1-IN1

Electromagnetic Energy Absorption within Extensive Impedance Structures A. Lugarkov and V. Kisel

P-1-IN2

Electro-Optic Structurally Chiral Materials A. Lukhtakia and J.A. Reyes*

Session P2: Dielectric Composites P-2-IN3

P-2-OR1

P-2-OR2

3 11

Chair: L.R. Arnaut

Modelling and Measurement of Advanced Carbon Fibre Reinforced Plastic Composites for EM1 Control of Spacecraft L.R. Amaut, J.F. Blackbum, K. Lees, A.R. Bridge, R.N. Clarke and A.P. Gregory

21

Electrical Properties of Graphite Filled Cement Composites for Device Application S. Bhattacharya, V.K. Sachdeva and R. Chatterjee

29

Electromagnetic Interference Shielding of Graphite Filled Cement Composites in Relationship to Resistivity and Morphology V.K. Sachdev, R. Chatterjee and R. Singh

33

P-2-OR3

Preparation and Optical Characters of Polymer Composite DR13PMMA Films Y. Gao, Q. Ren, F. J. Zhang, X.F. Cheng, J. Sun, H.L. Yang, X.B. Sun and L. Feng

38

P-2-OR4

Temperature Dependence of the Complex Permittivity of GreentapesTM M.V. Jacob

44

P-2-OR5

Preparation and Electromagnetic Properties of MgCuZn Ferrite-ZSU Dielectric Composites J. Bera* and P.K. Roy

P-2-OR6

A Study on the Effect of Inclusion of Micrometer-Sized and Nanometer-Sized Particles to the Properties of Silicone Rubber

48

52

N.R. Hamzah

Session P3: Magnetic Composites (1) P-3-IN4

Chair: 0. Acher

Reconstruction of Intrinsic Permeability of Inclusions from the Measured Permeability of a Composite K.N. Rozanov*, A.V. Osipov, D.A. Petrov, S.N. Starostenko and E.P. Elsukov

P-3-OR7

Interface Magnetism G. Kopnov, Z. Vager and Ron. Naaman*

P-3-OR8

Characteristics of Effective Permeability and Resonance Frequency for Barium-Ferritempoxy Composites Z. W. Li, Y.B. Can, X . Xu and G.Q. Lin

59 67

71

X

P-3-OR9

Microwave Absorbing Properties of Amorphous FeCuNbSiB Microwires Multilayer Composites M. Han*. D. Liang, L. Chen, J. Xie and L. Deng

P-3-OR10

Curve-Fitting of Complex Permeability and Its Applications Z.W. Li

P-3-OR11

EM Properties of Composites with Glass-Coated Amorphous Ferromagnetic Wires L. Liu*, S.Matitsine, L.B. Kong, G.Q. Lin, C.R. Deng, Y.B. Can and K.N. Roxanov

79

Session P4: Magnetic Composites (2)

P-4-IN5

P-4-OR12

75

83

Chair: L.J. Deng

Experiments on Electromagnetic Characterization of Ferromagnetic Nanocrystalline Alloy Flake Composites P.H. Zhou and L.J. Deng

89

High-Frequency Magnetic Properties for Composites of ZnNi-Substituted Y-Type Barium Hexaferrites Y.P. Wu, Z. W. Li, C.Q. Lin and C.K. Ong

93

97

P-4-OR 13

Carbonyl Iron Composite Materials for High-Frequency Applications M.A. Abshinova, A. V. Lopatin, N.E. Kazantseva, J. Vilda'kova'and P. Sa'ha

P-4-OR 14

EM Properties in Composites with the Filler of Electroless-Synthesized Ni-P Powder G.Q. Lin and 2.W. Li

101

Nanosized Ferrite Ceramics Derived from High-Energy Milled Powders with Promising Magneto-Dielectric Properties over 30-90 MHz L.B. Kong, Z.W. Li, C.Q. Lin and Y.B. Can

105

P-4-OR 15

P-4-OR 16

Thermostable Magnetic Elastomers Filled with Carbonyl Iron M.A. Abshinova, I. Kuritka, N.E. Kazantseva, J. VilZkova' and P. Sa'ha

P-4-OR 17

Development of Magneto-Dielectric Materials Based on Lithium Ferrite Ceramics for Miniaturization of Antennas M.L.S. Teo, L.B. Kong, 2.W. Li, G.Q. Lin and Y.B. Gan

Session P6: Metamaterials (1)

109

113

Chair: A.N. Lagarkov

P-6-IN6

hlagnetic and Acoustic Metamaterials S.A. Nikitov, S.E. Bankov. Yu.A. Filimonov, A. V. Crigorievskiy, V.I. Grigorievski and S.L. Vysotskii

119

P-6-IN7

Distinctive Feature of 1D Anisotropic and Gyrotropic Photonic Crystals A. P. Vinogradov, A.M. Merzlikin, A. V. Dorofeenko, M. Inoue and A.A. Lisyansb

127

P-6-IN8

Metamaterials: A New Route to Microwave Magnetism 0. Acher

133

xi

Chair: X. Yao

Session P7: Materials Processing P-7-IN9

Hybrid Processing Technology of Electromagnetic Ceramics and Thick Films X. Yao

141

P-7-OR18

Low-Fire Processing Magnetic+Dielectric Ceramic Composites T.-M. Peng, R.-T. Hsu, C.-J. Chung and J.-H. Jean

149

P-7-OR19

Microwave Li-Ferrite Material for Use in Microstrip Tunable Devices R. Pourush, N.K. Badola, Ashok, P.K.S. Pourush, G.S. Tyagi and G.P. Srivastava

153

P-7-OR20

Structural Electrical-, Magneto-Transport and Magnetic Properties of ZnO Embedded Nanocrystalline CMR Manganites (Lao.,Sro.3Mn03)1.,(Zn0), S. Paul, B. Singh and T.K. Nath*

157 161

P-7-OR21

Low Cost Synthesis of Nanosized NiFez04 N. K.Janjua, S. Imriaz and T. Hussain

P-7-OR22

Growth and Characterization of Neodymium Doped Strontium Hydrogen Phosphate Single Crystals by Gel Method M.S. Valsamma, N. V. Unnikrishnan and M.A. Ittyachen

165

Annealing Effect on Magnetostatic and Dynamic Properties of Fehl.4Ni,6Cr3,zSiz,,Nb7.8MnJ.6B18 Amorphous Ribbons E.E. Shalyguina*, M.A. Kornarova. V.V. Molokanov and A.N. Shalygin

169

P-7-OR23

P-7-OR24

XANES Investigations of Interatomic Interactions in (CoFeZr),(SiOz)l., Nanocomposites E.P. Domashevskaya, S.A. Storozhilov, S.Yu, Turishchev, V.M. Kashkarov, V.A. Terekhov, O.V. Stognej, Yu.E. Kalinin, A. V. Sitnikov and S.L. Molodtsov

Session P9: Metamaterials (2) P-PIN10

P-9-IN11

173

Chair: J.A. Kong

Why Periodic Structures May Not Be Able to Synthesize Negative Indices of Refraction B.A. Munk

179

On Negative Refractive Metamaterials: Characterization, Bianisotropy, and Applications J.A. Kong, H.S. Chen, X.X. Cheng, J.J. Zhang, D. W. Wang and B . 4 Wu

186

P-9-OR25

Ultra-Thin Radar Absorbing Structures Based on Short Strip Pairs X.S. Rao. S. Matitsine and H. Lim

191

P-9-OR26

Electromagnetic Characterisation of Conductive Helixes C. Deng, L. Liu and Y. Zhang

195

xii

Session P10: Materials Applications

Chair: S. Matitsine

P-10-IN12 Frequency Selective or Controllable Metafilm as a Part of On-Board Antenna Screen V. Kisel

20 1

P- 10-OR27 Smart Frequency Selective Surface with Conductive Fiber Array and Diodes L. Liu*, S. Matitsine, P. K. Tan and Y.B. Gan

209

P-10-OR28 Design of Broadband Planar Microwave Absorber Anyong Qing

212

P-10-OR29 Design of Broad Band Microstrip Patch Antennas using Air Gap in Microwave Frequency P. K.S. Pourush, S. Mann and R. Pourush

216

P- 10-OR30 Infrared Magnetic Response Metamaterials from a Virtual Current Loop Resonator Z. H u n g , D.H. Zhang, Y. Hou and J. Chu

218

Author Index

223

Session P l

Chair: S.A. Nikitov

This page intentionally left blank

Electromagnetic Energy Absorption within Extensive Impedance Structures

Andrey Lagarkov, Vladimir Kisel Institute for Theoretical and Applied Electromagnetics, Moscow, Russia The filamentary-source excitation problems are considered with regards to extensive cavities and plane structures with surface impedance specified. The solutions are based on rigorous approaches. Particular features of the electromagnetic field absorption in such structures are shown. The choice and optimal values of impedance are proposed to ensure the fastest field attenuation when going through a duct with impedance walls. The problem regarding top possible (hopefully, total) field suppression of a filamentary source placed above non-uniform impedance plane is discussed. New designs of the electromagnetic field absorbers and resonators are suggested which may be engineered with the use of metamaterials. Introduction: Electromagnetic modeling a cavity lined with coating Interest to the field propagation along the imperfect surface has about century-old history, the beginnings of which trace back to the Sommerfeld’s solution of the classical problem for the dipole radiating above the plane with finite conductivity. Later, as the radio broadcasting evolved, a lot of publications appeared which dealt with electromagnetic field propagation in the presence of an absorbing half-space. At present, a large number of problems exist which require understanding of the electromagnetic processes peculiar to the multiple interactions (“re-reflections”) of the wave traveling between imperfect surfaces. Corresponding phenomena are rather complicated even if the wave propagates between a pair of parallel plates. In any case, an effective investigation of the corresponding electromagnetic processes is possible only if the deep insight into the simpler problem of the point source excitation of an imperfect plane is reached. Solutions of these problems form the basis of the modern hybrid algorithms to calculate electromagnetic fields within extensive cavities; an important example of such a cavity is the air duct of a jet aircraft intake [l]. Numerous particular features of the electromagnetic excitation of a cavity can be revealed by studying rather not complicated structures, see Fig. l a , [2], [3].

3

4 More sophisticated models and algorithms which account for the complex shape and the presence of absorber coatings on the cavity walls give reliable results close to the measured data (see, for example, [4]-[6]). However, a very important “inverse” problem, namely, how to choose the absorber properties to secure a lowest possible level of the radar backscattering from the cavity, is solved by today mainly through selection of the coatings with proper angular dependencies of reflection coefficient, bearing in mind a ray picture of the field transport along the cavity. Note, the grazing wave incidence onto the walls is of prime interest because of the low efficiency of coatings in this case; that is why the backscattering patterns of intakes show significant peaks around the nose-on directions of the external illumination. At the same time, the geometrical optics considerations do not necessarily result in the optimal choice of coating because of complex diffraction phenomena in a realistic duct. Fig. lb, reproduced from [7], shows an example of the strong discrepancies between the results of the field calculation by the ray (upper picture) and the rigorous (lower picture) techniques even in a simple case of a waveguide formed by a pair of parallel conducting plates. This paper shows another possible way to get near to optimal absorber properties originating from energy considerations applied to a model problem for the point source excitation of an impedance plane. The surface impedance is chosen so as to provide for a maximum power flux density along the normal to the coating. The conclusion is made that the optimized coatings secure lowest backscattering at the typical dimensions and geometry of a duct. Usage of metamaterials for the same purposes is considered. It was found that with a point source radiating in a presence of an impedance plate one can attain even complete field suppression in an outer space provided several relationships are perfectly satisfied.

Choice of RAM to coat the walls of an extended cavity Let’s consider a possible definition of the model problem, Fig. 2. For the sake of simplicity, we shall consider 2D monochromatic case with the field frequency w.Let the point source be placed above the plane y = 0 with the constant surface impedance 2 specified. The source is a filament of x-directed magnetic current, therefore vector g’ of the incident field has a component perpendicular to the impedance plane, and vector G iis parallel to the plane. We should determine the value of impedance which provides for the highest possible power flux density transferred across the plane y = 0 in the given point zo or through the specified area of that plane. Rigorous solution of the boundary problem results in the following expressions for the tangent component of the magnetic field H , and the real part of the Poynting vector in the +

direction of

where q =

-

iy normal at y = 0 :

d p ,k

= 2 z / A , 2,

= Z/Wo,

Wo = 120z (Ohms), R is the wavelength.

5 Consider the illustrative example, when the filament is placed at the altitude of about half a wavelength, kyo = 3 . Fig. 3 shows the results of calculation of the power flux density transferred across the plane y = 0 as the dependence upon the location of observation point zo . Three options for the impedance 2 are tested. Curve 1 refers to the case of the absence of impedance plane, when the filament is located in the unbounded space and the electromagnetic wave freely travels across boundary y = 0 without reflection. Curve 2 corresponds to the case when the impedance of the plane is equal to the intrinsic impedance of the free space 2 = Wo. Note, this particular value of 2 provides for the total transmission (“absorption”) of the plane wave, normally incident onto the plane. When an observation point is located not so far from the filament (at small zo values and rather large angles a, see Fig. 2 ) the curves 1 and 2 are close to each other. However, at the larger zo (i.e., at low, “grazing” angles a) the power flux across the plane y = 0 with 2 = Wo appears to be much less as compared to the case of the free space (see Fig. 3b). Nevertheless, one can create an electromagnetic wave absorbing coating to secure an increase in power flux transition across the media interface and, correspondingly, attenuation of the field energy at grazing incidence. For example, let a conducting plane be coated with 0.65 mm layer of RAM, its permittivity and permeability be chosen as ~ = 1 4 - i O , ,u=1.7-i1.6. At the wavelength of il=3 cm the equivalent impedance of such a structure is almost independent from the angle of plane wave incidence, and its value is about 2 / Wo = 0.29 + i0.21. Curve 3 is drawn for this case. It shows that in a wide region of zo values (at the low angles a, Fig. 3b) a much greater portion of the field energy is transferred across the coating boundary as compared to the case of “matched against normal incidence” (curve 2) or even “perfectly non-reflective” coating (i.e., free space, curve 1). The coating with these properties is suitable to apply onto the air duct walls to achieve RCS reduction of the intake at the incidence directions close to the compressor axis. Varying the of impedance at the given task options (for example, yo and zo), one can define its optimum value, which assures the maximum power flux density (see example shown in Fig. 4). Numerous calculations indicate that when using homogeneous coatings, the best results are likely to be achieved at the inductive surface impedance, if Re(Z/Wo) = 0.2 ...0.5 and Im(2 / Wo)= 0.1...0.3. This conclusion agrees with the published data and physical assumptions that the better conditions for wave absorption are secured by an impedance with inductive component, particularly, due to surface waves excitation and higher field concentration nearby the duct walls. Finally, calculations carried out for the realistic designs of complex intakes demonstrated the superiority of the coatings chosen in the way described above. \

,

0.m

0.004

0

0.002 -05

0 0

05

1

Fig. 4

Re(ZIv)’s

6

Point source energy absorption by a half-space Now it is natural to set few questions about what value of impedance of a plane should be chosen to absorb the maximum portion of energy radiated by a point source (say, filamentary current), how much the amount of the absorbed energy is and how to create such an impedance. Note, in view of the symmetry of the radiation pattern of the filament, at the absence of the plate (in the free space) equal power fluxes are radiated into upper and lower half-spaces, and exactly one half of the radiated energy penetrates through the plane y = 0, see Fig. 5 . Calculations showed that even at some “optimal” but constant value of Z (see, for example, Fig. 4) the integral of the real part of the Poynting vector taken over the surface y = 0 (that is, zo = -a,..+ co) does not exceed a half of radiated power as well. However it is evident that the lower half-space can absorb more than one half of radiated power provided the impedance distribution is inhomogeneous. For example, one can define a function of impedance distribution over the plane to have Z = W o closer to the filament, at large angles a, and choose 2 from considerations of maximum energy absorption (see, for example, Fig. 3 and Fig. 4) while moving away from the source, i.e., at lower values of a. Of course, impedance 2 should vary rather smoothly along the plane to prevent from strong diffraction, which may cause degradation of the coating performance. Another way may be suggested to create a system which would consume more than half a power radiated from the source. Let place a specially designed scatterer in the region y < 0 . Then an asymmetric radiation pattern with respect to the y = 0 plane can be formed with its main lobe directed downward, see Fig. 6 (similar trick is used in the Uda-Yagi dipole antennas). In doing so, the major portion of energy is directed into lower half-space. Further, it may be absorbed in an ordinary way. Once the tangential components of the electric and magnetic fields are calculated in the plane y = 0, one can evaluate the desired distribution of the equivalent surface impedance of such a system. Total transition of the point source radiation into a half-space The system shown in Fig. 6 may be further complicated. Evidently, it is possible to make the field cancellation in the upper half-space more complete and, correspondingly, to increase the portion of the energy absorbed in the lower half-space by increasing a number of auxiliary scatterers. The question arises: what maximum portion of energy emitted by source can be directed into lower half-space without using any additional devices (say, mirrors) in the upper half-space, at Y ’Yo. It will be shown below that one can create even such a passive system which secures total cancellation of the source field in the upper half-space and, correspondingly, transfers the whole of the emitted energy into the lower half-space.

Consider an example of designing such a system, firstly, on a qualitative level.

Far field pattern

Fig. 5

Fig. 6

7

\

,

kd/” Mirror

-yo

source

Fig. 7 Fig. 8 Let a filamentary source with a single x-directed component of the electrical current be placed in the point yo over the conducting half-plane y = 0, Fig. 7. As known, in this case the secondary field can be interpreted as produced by the mirror source, the currents in filament and in its image are of the same magnitude but their phases are opposite to each other. In other words, the sign of the wave phase is reversed when reflection from the conductor occurs. Let a focusing flat plate (Veselago’s lens [8]) with a thickness of d = y 0 / 2 made of the metamaterial with E = -1, p = -1 be inserted between the source and the plane at the altitude h so as 0 < h < y 0 / 2 . Then the focusing point and its mirror image coincide with each other right at the surface of the conducting plate (see, for example, the ray picture in Fig. 8). Once the total phase advance along ray paths is calculated bearing in mind the negative phase velocity of the wave traveling through the plate and the phase reversal of the field due to the reflection from the conducting plane, one can discover that in the region y > yo the incident and secondary fields mutually cancel each other. In an ideal case, when electromagnetic losses in the plate are infinitesimally small, the total field in the upper halfspace tends to zero. Rigorous solution of the corresponding boundary problem results in the same conclusion. This is illustrated by Fig. 9 and Fig. 10, which show the absolute values of the total field in the vicinity of the source (in the plane perpendicular to the filament of electrical current). Contour plots are given in Fig. 9, and corresponding 3D images of the field distribution are shown in Fig. 10. Maxima at plate interfaces

-2

-1

0

Fig. 9

bj

’

kz

2

Fig. 10

+I

8

Upper figures (a) depict the results obtained at kyo = 1, d = 2h = yo /2, E = p = -1 - iO.001 (time dependence is chosen as exp(iwt)), geometry of the problem corresponds to Fig. 8. Lower figures refer to the case of plate absence, when E = p = 1. They are given for reference purposes. Note, in the presence of the metamaterial plate the field in the region y > yo is almost equal to zero in contrast to the second case, when the field of the filamentary source does not attenuate. It is seen most clearly on the cross sections of 3D images of the field distribution, Fig. 11 (a:no plate inserted, b: the case of the geometry shown in Fig. 8). Open resonator The regions with high field concentration due to accumulating reactive energy are worth noting in the figures (see Fig. 9a, Fig. lOa). They arise next to the metamaterial plate faces while field compensation in upper half-space occurs. These maxima reach especially great values in the case of the plate arrangement side-by-side to the conducting surface, h = 0, Figs. 1lc, d. Thus, the structures shown in Fig. 8 and Fig. 1l c may serve as prototypes for designing novel open resonators without usual restrictions on the thickness of the system in terms of wavelength. Note, previously a different idea of a “thin” metamaterial-based resonator of “closed” type was suggested [9](the metamaterial sheet was sandwiched between a pair of conducting plates). Other design of an open resonator is also known [lo], it is based on the negative refraction property of photonic crystal or metamaterial prisms. Correspondence to the “superresolution”phenomenon. Effect of losses. One of the specific features of the Veselago’s lens is the ability to produce an image with extremely fine details as its resolution is not restricted with so called “diffraction limit”. This surprising fact was firstly pointed out by Prof. Pendry [ll]. Later it was shown that the absorption in metamaterial plays a crucial role in view of achieving superresolution in practice. And the smaller the plate thickness (in wavelengths), the higher is the upper level of losses to secure desired resolution (see, e.g., [12]).

i

4 c)

Fig. 11

Similar conclusions can be made regarding the performance of the systems under consideration. Even if one tends to compensate only propagating modes of the far field in upper

9

half-space, rather strict requirements should be placed to the quality of metamaterial. But to attain the nearfield compensation in the vicinity of the source (around yo point), the mirror image should be developed with “superresolution”, which is achievable only with extremely low losses in the plate. Though, at small kyo and kd one may expect rather good results even using existing metamaterials with noticeable absorption, as was in the case of electrically thin focusing plate [12]. Passing on to the greater values of kyo , the near field is much more difficult to compensate, and this is illustrated in Fig. 12 (geometry of the Fig. 8, d = 2h = y0/2, kyo = 14).

Electromagnetic wave absorber with special angular properties Finally, note that metamaterials may be efficiently used to create novel absorbers of the electromagnetic energy of a plane wave. Their special properties may be achieved, particularly, due to arranging a wave path so as to cross the metamaterial structure with the result of phase advance compensation. An example of the RAM design usable under the incidence of perpendicularly polarized (TM) plane wave is shown in Fig. 13. Provided the electromagnetic response of the semiinfinite film, particularly, its transition and reflection coefficients were properly chosen, the wave reflected from the film cancels the wave penetrated into and returned back from the region y < y o . This latter wave got a negative phase correction when propagated in the metamaterial plate and additional phase reversal because of the reflection from the conducting plane. It is interesting that total phase advance of that wave is equal to z independently on the incidence angle. Therefore, it is possible to achieve a very broad angular range in which such an absorber should operate efficiently, in contrast to classical designs, like Salisbury screen [ 11. In fact, only deviations of semi-transparent film properties impose certain limits to the angular performance. Finally, as there are no fundamental physical restrictions on the thickness of the described absorber, it can be made electrically thin (at least, in principle), as well as earlier suggested system of complementary metamaterials [13]. Conclusion Thus, a way to attain nearly optimal absorber properties originating from energy considerations was suggested, the technique is based on a model problem solution for the point source excitation of an impedance plane. Next, it was shown that the metamaterials provide a variety of new opportunities in designing novel absorbers and resonators, the latter may be even open. The paper reports about an important (though not so evident) result of potential total absorption of the radiated field of omnidirectional point source by a flat surface with properly chosen distribution of the impedance. Such a surface may be engineered with the use of the metamaterials.

Fig. 12

Fig. 13

10

References E.F. Knott, J.F. Shaeffer, M.T. Tuley. Radar cross section. Artech House, Boston-London. 1993. H. R. Witt, E. L. Price, “Scattering from hollow conducting cylinders”, Proc. IEE, 115, no. 1, p. 94-99, 1968. H.Ling, S.-W. Lee, R.-C. Chou, “High-frequency RCS of open cavities with rectangular and circular cross sections”, IEEE Trans. Antennas and Propag., 37, no. 5, p. 648-654, 1989. F. Obeleiro-Basteiro, J.L. Rodrigues, R.J.Burkholder, “An iterative physical optics approach for analyzing the electromagnetic scattering by large open-ended cavities”, IEEE Trans. Antennas andPropag., 43, no. 4, p. 356-361, 1995. H.T. Anastassiu, J.L Volakis, D.S. Filipovic, ”Integral equation modeling of cylindrically periodic scatterers in the interior of a cylindrical waveguide”, IEEE Trans. Microwave Theory Tech., 46, no. 11, p. 1713-1720, 1998. V.N. Kisel’, A.I. Fedorenko, “Electromagnetic modeling of the jet aircraft intake with the interior impeller”, Con$ Proc. 2002 Int. ConJ: on Mathematical Methods in Electromagnetic Theory (MMET*02), Kiev, Ukraine, Sept. 10-13, vol. 2, p. 508-510,2002. H. Ling, R.-C. Chou, S.-W. Lee, “Rays versus modes: pictorial display of energy flow in an open-ended waveguide”, IEEE Trans. Antennas andPropag., 35, no. 5, p. 605-607, 1987. V.G. Veselago, “The electrodynamics of substances with simultaneously negative values of E andp”, Sov. Phys. Usp., 10, p. 509, 1968. N. Engheta, “An idea for thin subwavelenrth cavity resonators using metamaterials with negative permittivity and permeability”, IEEE Antennas and Wireless Propag. Lett., 1, p. 1013,2002. S. He, Y. Jin, Z. Ruan, J. Kuang, “On subwavelength and open resonators involving metamaterials of negative refraction index”, New Journal of Physics, 7,p. 210,2005. J.B. Pendry, ‘Wegative refraction makes a perfect lens”, Phys. Rev. Lett., 85, no. 18, p. 39663969,2000. A.N. Lagarkov, V.N Kissel, “Near-Perfect Imaging in a Focusing System Based on a LeftHanded-Material Plate”, Phys. Rev. Lett., 92, 077401,2004. A. Alu, F. Bilotti, N. Engheta, L. Vegni, “A thin absorbing screen using metamaterial complementary pairs” Proc. of joint 9th International Conference on Electromagnetics in Advanced Applications (ICEAA 2005) i-1Ith European Electromagnetic Structures Conference (EESC 2005), 12-16 Sept. 2005, ISBN 88-8202-094-0, ed. Roberto D. Graglia; Politecnico di Torino, Torino, Italy, p. 229-232, 2005

Electro-optic Structurally Chiral Materials Akhlesh Lakhtakial and Juan Adrian Reyes' 'Department of Engineering Science & Mechanics Pennsylvania State University, University Park, PA 16802-6812, USA 'Fisica Quimica, Instituto de Fisica, Universidad Nacional Aut6noma de MCxico, MCxico D. F. 04510, Mexico

Abstract: Numerical examination of the solution of the boundary-value problem of the reflection and transmission of a plane wave due to a slab of an electro-optic structurally chiral material (SCM) indicates that the exhibition of the circular Bragg phenomenon by the SCM can be controlled by the sign and the magnitude of a dc electric field as well as by its orientation in relation to axis of helicoidal nonhomogeneity of the SCM.

1

Introduction

Unidirectional periodic nonhomogeneity arising from structural chirality - i.e., a helicoidal variation of anisotropy along a fixed axis - is a distinctive feature of cholesteric liquid crystals [l]and chiral sculptured thin films [2], which exemplify structurally chiral materials (SCMs). The circular Bragg phenomenon (CBP) is exhibited by SCMs, by virtue of their periodicity and structural chirality. An incident, circularly polarized, electromagnetic plane wave of the same handedness, but not of the opposite handedness, as a SCM is highly reflected in a certain wavelength-regime, provided (i) the SCM contains a sufficiently large number of periods along the fixed axis, and (ii) the angle of incidence with respect to that axis is not very large. Exhibition of the CBP by SCMs underlies their use as circular-polarization rejection filters in optics. Control of the CBP is very desirable for tuning the Bragg regime as well as for switching applications. One way would be to use SCMs that are electro-optic. The theory of electrooptic SCMs is the focus of this presentation, with emphasis on the twin possibilities of thinner filters and electrical control of the CBP, depending on the local crystallographic class as well as the constitutive parameters of the SCM. In the following sections, vectors are denoted in boldface; the Cartesian unit vectors are represented by u,, u,, and u,; symbols for column vectors and matrixes are decorated by an overbar; and an exp( -iwt) timedependence is implicit with w as the angular frequency.

2

Theory in brief

The axis of helicoidal nonhomogeneity of the chosen SCM of thickness L is designated as the z axis, and the SCM is subjected to a uniform dc electric field Edc.The half-spaces z 5 0 and z 2 L are vacuous. As an arbitrarily polarized plane wave is obliquely incident on the SCM from the half-space z 5 0, reflected and transmitted plane waves exist in the half-spaces z 5 0 and z 2 L , respectively.

11

12

2.1

Structurally chiral material

The optical relative permittivity matrix of the chosen SCM may be stated as [3, 41

The matrix E ~ E ( z )incorporates both the Pockels effect [5] and the arbitrarily oriented but uniform Ed",and is given correct to the first order in the components of the dc electric field by

Here.

are the principal relative permittivity scalars in the optical regime; and the electro-optic coefficients are denoted by T J K (with 1 5 J 5 6 and 1 5 K 5 3) [3, 51. The SCM can be locally isotropic, uniaxial, or biaxial; furthermore, it may belong to one of 20 crystallographic classes of local point group symmetry. The tilt matrix -sinX

0

cosx

cosx

0

sinx

(4) involves the angle x E [0,n/2] with respect to the z axis in the zz plane. In (1),the use of the rotation matrix defined by cosc - sine 0 S z K ) = si;c c0;c (5)

;)

(

involves the half-pitch s2 of the SCM along the z axis. In addition, the handedness parameter h = 1 for structural right-handedness and h = -1 for structural left-handedness. Without significant loss of generality, let

Ed"= EdC(U, cos x d c

2.2

+

sin xdc)

Uz

,

xdc

E [0,./2]

Propagation in the SCM

The Maxwell curl postulates for the chosen SCM slab are given by

.

(6)

13 where E , and p, are the permittivity and the permeability of free space (i.e., vacuum). As a plane wave is incident obliquely on the SCM, Vz we set [2]

E(z, y, z ) = e ( z ) exp [in(zcos I$

+ y sin I$)]

H(z,y, z ) = h(z) exp [ir;(zcos 4

+ y sin I$)]

where the wavenumber m and the angle I$ are determined by the incidence conditions. After defining the column vectors (9)

and

@ ( z ) = A2

(%) 7 J ( z ) , *

where the unitary 4x4 matrix

f cos< -sin
t: .-

7.6

€

a" -Xa,

7.5

s

7.4

.4-

O

5Q 2 of

7.3 7.2 7.1 0

50

100

150

200

250

300

Temperature (K) Fig. 3: The real part of complex permittivity of the GreenTapesTMat a frequency of 9.5 GHz.

47

1o-z

I

I

I

I

I

____________________----------------------

c

C

a,

m

S

I-

10‘~

v) v)

0

1

1o4

0

I

I

I

I

I

50

100

150

200

250

300

Temperature (K) Fig. 4: The loss tangent of the GreenTapesTMat a frequency of 9.5 GHz.

Figure 4 shows the loss tangent of the materials as a function of temperature. The change in loss tangent with temperature is very small over the measured range. It is interesting to note that the loss tangent of Sample#l is one order in magnitude less than that of Sample#2.

4. Conclusions The real part of relative permittivity and loss tangent of four types of DuPont Green TapeTM were precisely characterised at 9.5 GHz in the temperature range 15 K - 300 K. The measurements were performed using a cryogenic split post dielectric resonator. The real part of complex permittivity is between 7.2 and 7.5 for both the materials. The change in permittivity is around 0.5% in the temperature region 15 K - 200 K. In contrast to many other dielectric materials, the change in loss tangent over the measured temperature range is very small. Therefore the GreenTapes may be materials to use in conjunction with superconducting materials or in low temperature electronic circuits.

Acknowledgements This work was done under the financial support of ARC Discovery Project DP0449996.

References [l] J. F. Kiang, “Novel Technologies for millimetre wave applications” (pp. 173-190), Kluwer Academic Publications, USA (2003). [2] M. V. Jacob, J. Krupka, J. Mazierska and M. Bialkowski, Proceedings of Asia-Pacific Microwave Conference 2006 [3] [online] www.tdextek.com “LTCC Technology” accessed on 1O* July 2006. [4] H. Mandai, K. Wakino and N. Nakajima, APMC2001 SMMM, Taipei, Taiwan, 142-145 (2001). [5] M. Valant, APMC200I SMMM, 2001,Taipei, Taiwan, 6-1 1 (2001). [6] P. Barnwell, C. Free, and Z. Tian, APMC2001 SMMM, 2001,Taipei, Taiwan, 1-4 (2001). [7] M. V. Jacob, J. Mazierska and M. Bialkowski, Ceram. Eng. Sci. Proc., 26, [5], 209-216 (2005). [8] J. Krupka, S. Gabelich, K. Derzakowski and B. M. Pierce, Meas. Sci. Technol., 10, 1004 (1999). [9] K. Leong and J. Mazierska, IEEE Trans. Microwave Theory and Techniques, 50,2115 -2127 (2002).

Preparation and Electromagnetic Properties of MgCuZn Ferrite-Z5U Dielectric Composites

J. Bera*,P. K. Roy Department of Ceramic Engineering National Institute of Technology, Rourkela-769 008, Orissa, INDIA *Corresponding Author Email: j [email protected] Abstract A new dielectric-ferrite composite ceramics were prepared by mixing ferrite powder with compositions Mgo.25Cuo.2Zno.55Fe204and BaTiO3 based Z5U dielectric powder in different weight fraction. Uniaxially pressed toroids and pellets were sintered at temperatures 5950°C. The co-firing behavior of dielectric-ferrite composite was investigated using thermomechanical analyzer (TMA). The coexistence of spinel ferrite and BaTiO3 phases were confirmed by XRD analysis. The permittivity of composites continuously decreases with increasing ferrite content and the initial permeability decreases with increasing dielectric content. Dielectric and magnetic losses were increased upon composite formation due to the generation of interfacial charges in the ferrite-ferroelectric interface.

1. Introduction Recently, dielectric-ferromagnetic and ferroelectric-ferromagnetic composite materials have stimulated much scientific and technological interest due to their interesting electromagnetic/ magnetoelctric properties [l-31. Even though there are a few single phase materials known to possess simultaneous ferroelectric and magnetic ordering, the coupling between the sublattice of ferroelectric and magnetic orderings is not particularly strong for device application. In view of this, the research focus in multiferroics has been shifted towards synthesizing composites and multilayer structures of the two different phases. The dielectric-ferromagnetic composite can provide capacitance and inductance in single device. So these materials can be used for passive LC filters applications [4]. These filters have high industrial requirements for suppressing electromagnetic/radio frequency interference (EMIRFI) in electronic cixuitry . The performance of filters in composites can be tailored easily by adjusting the capacitive and inductive components through compositional variation. The key issue to prepare capacitive-inductive LC filters composites is to select the right constituents. The most important process in manufacturing of defect-free multilayer chip LC devices involves the cofiring of capacitor and inductor materials at a low temperature. Defects could generate from severe chemical reaction and from mismatched densification kinetics between two different materials. Although, a variety of multiferroic composites have been fabricated with ferroelectric phases like BaTi03, PZT, PMN and ferromagnetic phases including CoFe204, NiFe204, NiZn-, NiCuZn-ferrite [5-81, there has been no prior work on the preparation of composites with MgCuZn ferrite and BaTi03 based Z5U ceramics. MgCuZn ferrite and BaTi03 based Z5U dielectrics have been widely used in manufacturing multilayer inductors and capacitors due to their superior magnetic and dielectric properties. In the present work, ceramic composites of 'x'MgCuZn ferrite-'(I-x) Z5U dielectric with different 'x' were prepared and cofiring behavior and electromagnetic properties were investigated. 2. Experimental The dielectric-ferrite composite ceramics were prepared by mixing MgCuZn ferrite powder with composition Mgo.~~Cuo.~Zno.~Fe~04 and Z5U-type dielectric powders. Mgo.25Cuo.*Zno.s5Fe20~ powder was prepared through nitrate-citrate sol-gel auto combustion process. The detail synthesis process has been reported elsewhere [9]. A commercial Z5U dielectric powder (Manufacturer: Tam Ceramics International, Niagara Falls, NY) was used in this study. The material was based on BaTi03 composition blended with some zirconate, zinc niobate and zinc borate flux. Composites of 'x'ferrite-'(I-x) 'dielectric with x= 0.25, 0.5 and 0.75 were prepared by mixing two respective

48

49

powders. The mixed powders were dried and uniaxially pressed at a pressure of 50 MPa to form green toroidal and pellet specimens. Sintering behavior of composite was investigated using thermal mechanical analysis (TMA, Netzsch DIL 402C, Germany). Specimens were sintered at 900- 950°C for 2 hrs in air. Phase identification of the sintered samples was performed using Philips XRD with Cu-Ka radiation. Silver electrode paste was printed on two surfaces of sintered pellet for dielectric measurements. Enamelled Cu wires were wound on sintered torroid samples to measure the magnetic properties. The inductance was measure using HP-4192A impedance analyzer and initial permeability was calculated as stated in [lo].

3. Results and discussion Fig. 1 shows the XRD patterns of pure ferrite, pure Z5U and 5050 wt% composite sintered pellets. It is clearly seen that two phases, BaTiOs and spinel ferrite, are separately present in the composite. No other phases were observed in the XRD analysis, which suggest that no significant chemical reaction took place between ferrite-dielectric during co-firing of their composite. This is very important for preparation of multiferroic composites so that the ferroelectric and ferromagnetic properties of the individual component could not degrade after sintering. Fig.2 shows the comparison of sintering behavior between composite (5050 wt %) and pure materials. It shows that there are difference in sintering shrinkage between ferrite and Z5U ceramics. The onset temperature of ferrite (-840°C) is slightly higher that of Z5U ceramics. However, the densification of ferrite is completed at a higher densification rate over a narrow temperature range compared with that of Z5U ceramics. Higher densification rate of ferrite is due to the presence of Cu in its composition. Onset temperature of Z5U ceramics (-800°C) is lower due to the presence of zinc borate flux and it sintered slowly compared to ferrite. The composite ceramics have an onset temperature (- 820°C) in between the individual pure materials and the rate of sintering is similar to that of Z5U ceramics. It seems that there is no or very little effect of ferrite on sintering behavior of the composites. The results also indicate that the composite can be sintered at low temperature (5950°C).

* Ferrite # Z5U Ceramic

I

30

40

50

28

I 60

Fig.1 XRD patterns of sintered ferrite (a), Z5U (b) and 5050 wt% composite (c).

50 I

Ferrite--------,

3

-0.20 400

600

800

1000

1201

Temperature (OC)

Fig. 2 Shrinkage curve at 5"C/min heating rate in air for ferrite, Z5U dielectrics and their 50:50 wt% composite (x=0.5). Dielectric and magnetic properties were measured as a function of frequency. Fig. 3 and Fig. 4 show the frequency dependences of permittivity and dielectric loss of different composites in the frequency range of 100 KHz-1OMHz at room temperature. As expected, permittivity decreases with increasing ferrite content due to the increase in the amount of non ferroelectric material, i.e. dilution effect of non-ferroelectric particles. Permittivity remains almost constant upto MHz frequency range. The dispersion of Z5U permittivity occurs near 10 MHz may be due to its relaxation losses. Fig. 4 shows that dielectric losses of ferrite containing ceramics are much higher than that of pure Z5U ceramics in low frequency range. This is due to the generation of space charges at the interface between ferrite and dielectrics with ferrite addition. Interfacial space charge polarization phenomena are predominant in lower frequencies. That is why the dielectric loss of x=l and 0.75 compositions decreases with the increase in frequencies. Dielectric loss of Z5U ceramics increases at higher frequencies may be due to relaxation losses. x=0.75 composite shows a low loss characteristic in the higher frequency range and may be suitable for high frequency LC filter application. 10000

100

0 1000 >

fir

0) v)

g10

P

100

10 100

1 1000 Freauencv (kHz)

1OOO(

Fig. 3. Frequency dependence of permittivity for different composites.

100

1000 Freauencv (kHz)

1OOO(

Fig. 4. Frequency dependence of dielectric losses for different composites.

I ~

-1

1000

-

-A-

- -2

5.

d

1 .

XzO.5

-XzO.75

5.

c .-

Pm

;loo

t

a,

! i a,

m -3 0

-I

a

10

100

1000 Frequency (kHz)

1oooc

Fig. 5. Frequency dependency of permeability for different composites.

100

1000 Frequency(kHz)

1000

Fig. 6. Relative magnetic loss factor as a function of frequency for different composites.

Fig. 5 and Fig. 6 show the frequency dependency of initial permeability and magnetic loss of different composites in the frequency range of 100 KHz-1OMHz at room temperature. As expected, permeability decreases with increasing Z5U dielectric content due to the dilute effect of the ferrite phase. It also shows that the frequency dispersion of permeability shifted to higher frequency with increasing dielectric content according to Snoek’s law that a higher initial permeability is accompanied by a lower gyromagnetic resonance frequency. The fact that the relative magnetic loss factor (RLF) of the composites and pure Z5U dielectric is higher in the low frequency range may be due to extrinsic behavior related to windings. However, their RLFs of them are lower in the high frequency zone (Fig. 6).

4. Conclusions Ferrite-dieletric ceramic composites were prepared by mixing MgCuZn ferrite and BaTiO3 based Z5U dielectric powders. The densification behavior of the composite ceramics mainly depends on the densification behavior of dielectric ceramics. The sintered composites had two phases i.e. BaTi03 perovskite and spinel ferrite phases. The co-existence of the two phases was confirmed by XRD analysis. The electromagnetic properties, such as permittivity and initial permeability change continuously between those of two components containing composites. Thus, the dielectric-ferrite composites ceramics with tunable electromagnetic properties can be prepared by adjusting relative content of the two components. These materials can be used for multilayer chip LC filter applications. References [ l ] C.W. Nan, Y.H. Lin, and J.H. Huang, Ferroelectrics, 280, 153-163 (2002). [2] M.I. Bichurin, V.M. Petrov, Y.V. Kiliba, and G. Srinivasan, Phys. Rev. B, 66 (13), no. 134404 (2002). [3] A. Bush, Y.K. Fetisov, K.E. Kamentsev, V.F. Meshcheryakov, and G. Srinivasan, J. Mug. Mug. Muter., 258-259,4547 (2003). [4] A. Rafferty, Y. Gun’ko, and R. Raghavendra, J. Eur. Cerarn. Soc., 24,2005-2013 (2004). [5] Van den Boomgaard, J. and Born, R. A. J., J. Mater. Sci.,13, 1538-1548 (1978). [6] Bunget, I. and Raetki, V., Phys. Stat. Solidi (a), 63, K55-K57 (1981). [7] Srinivasan, G., Rasmussen, E. T. and Hayes, R., Phys. Rev. B, 67(014418), 1-10 (2003). [8] C Miao, Ji Zhou, X. Cui, X. Wang, Z. Yue, and L. Li, Muter. Sci. Eng. B, 127 (1) 1-5 (2006). [9] P.K.Roy, .and J.Bera, J. Mug. Mug. Muter. 298, 38-42 (2006). [lo] J.Bera, P.K.Roy, Phys. B, 363, 128-132 (2005).

A Study on the Effect of Inclusion of Micrometer-sized and Nanometer-sued Particles to the Properties of Silicone Rubber N.R. Hamzah Universiti Teknologi Mara, Faculty of Electrical Engineering 40300 Shah Alam, Selangor, Malaysia e-mail: ngahramzi@,,vahoo.com.sg 1. Introduction The term “nanodielectric” is coined for dielectrics filled with nanometer-particles [l]. It is currently popular known as “nanocomposites”. Recent studies showed that the particle size of the filler, contributes to the strengthening of the mechanical properties of the matrix polymer, space charge accumulation at the interfaces between the particles and matrix polymer, dielectric breakdown strength, dielectric constant and loss factor. These property enhancement are attributed to the behaviour of the interfacial interaction zone surrounding the particulates where the physical and chemical bonding influence the mobility of the accumulated charges, and a double layer is formed at the interfacial region influencing the local conductivity [2]. At 10% wt loading, the micrometer-sized particles and polymer matrix (micro-filled polymer) exhibits structure related to a-transition (local chain mode) while nanometer-sized particles and polymer matrix (nano-filled polymer) behaves like a foreign inclusion, exhibiting interfacial phenomena [3]. Once the size of the nanoparticles approaches that of the polymer chain length, they start to act cooperatively. In micro-filled polymer, a substantial mitigation of internal charge will accumulate, e.g. at the particulate-matrix interfaces. Maxwell-Wagner effect results in significant interfacial polarisation in micro-filled polymers. Also, the presence of significant number of space charges causes the internal stress to increase to almost 10 times higher. In nano-filled polymer, there is a reduction in the number of the space charges and their densities are small. At lower loadings, it had been postulated that nanoparticles reduce the bulk charge accumulation by introducing a local conducting path through the overlapping of nanometric double layers [2]. For a 10% wt loading, the conductivity of the nano-filled is indistinguishable from those of micro-filled material. When the percolation limits is exceeded (higher loadings), the nano-filled material starts to show appreciable conductivity [3]. Nanometric particles (higher surface area) may contribute to the process of tether chain entanglement. Enhanced tethered zones restrict the polymeric chain movement; thus lowering the permittivity [ 2 ] . The Maxwell-Wagner effect is either not produced or suppressed in nano-filled polymers and thus permittivities of the latter remain constant over a considerable frequency range. Increase in internal stress is negligible and thus leads to a significant increase in the dielectric breakdown strength [2].

2. Sample Preparation and Experimental Setup The objective of this work is to investigate the mechanical and electrical properties of silicone rubber (SIR) upon the addition of micro- and nano-sized particles. Three types of SIR were employed, namely, Toshiba TSE22 1-4U, Wacker Chemie GmbH Elastosil WL-402 and Bayer HRV5/50U, measured to be of hardness 37, 44 and 50, respectively. China clay and French chalk, whose particle sizes are 4 microns and about 100 nm, respectively, were chosen as the fillers and hydrogen peroxide as the vulcanising agent. French chalk could only be added up to 0.1 phr while for china clay it is 1 phr. The fillers were added in proportions of 2, 3, and 4 parts per hundred parts of resin (phr). The compositions were then moulded to the standard thickness. Only up to 4 phr are usable, higher addition makes the compositions to become soft and difficult to be moulded. Hardness and tensile strength tests were selected for mechanical properties, while volume resistivity and breakdown strength tests for electrical properties. Frequency responses of the dielectrics were also investigated. Also, FTIR scans on the samples were made.

52

53

3. Experimental Result Figure 1 shows the results of hardness tests. The hardness of SIR Hardness 37 (H37) in general improves with French chalk although a slight drop at 2 phr. For SIR Hardness 50 (H50), hardness deteriorates with the inclusion of the fillers. Performance of SIR Hardness 44 (H44) series is in between the cases for H40 and H50. FTIR scans in transmission mode had been conducted on the samples. The result shows that the transmission of the samples reduces with increasing filler (see Figure 2) content especially between wavenumber 3600 to 4400 cm-’, which is a “fingerprint” region of SIR. The magnitudes of the stress at 100% elongation of the samples are shown in Figure 3. From the dielectric frequency response profile of the samples, the peak values of the relative permittivity (at 60 Hz) are compared and shown in Figure 4. Figure 5 and Figure 6, respectively, show the average values of the breakdown voltages and the volume resisitivities of the sample. 55 50

’

Iv)n 45

e 40

2 35 30

Figure 1 Hardness test

60 h

50

5 c 40 .-0 .-8 30

z 2

c

20 10 0

1

0

1-

BF +BC

--

2

- - W F +WC

3

+-TF +TC

5 Filler content (phr)

4

I

Figure 2 Comparison of average % transmission (FTIR) at 3800 to 4400 cm-’

54

0 1

,

I

0

1

2

L+-

~-

TF -A-TC

- C - WF

- 0 - W C-.

3 BF +BC

~.

4

killer content (phr)

Figure 3 Tensile strength at 100% elongation 12 1

I

0

1

2

1 4 TF--O--WF+WC-BC/

3

4

Filler content (phr)

Figure 4 Relative permittivity (Peak @ 6OHz) of the samples

F

20

a,

-m

19

P

z

18

:m:

17

a,

16

v

m

4-

f! Q

m

E a?! 15 1 0

I

I

,

1

2

3

Figure 5 Average breakdown voltages of the samples

4

55

I '-

1

0

TC_.

TF +WC

2

_ _- f

WF +BC

4

BF

I

4

3 Fillercontent(phr)

Figure 6 Volume resistivities of the samples

4. Discussion There was no significant change in the pattern of the FTIR spectrum of raw and filler-added SIRS, indicating that there is no change in molecular structure with the filler addition. The maximum loading is 3 phr of French chalk for H37 and 2 phr of china clay for H50. For H44, the result is non-conclusive. It could be due to the fact that the percolation limit had been exceeded and filler acts cooperatively with the matrix. The dielectric breakdown for H37 improves at 3 phr French chalk loading and for H50; it improves at 2 phr china clay addition but deteriorates at 3 phr. The volume resistivities of H37 and H50 improved at those loadings. 5. Conclusion It can be concluded that the percolation limit for H37 is the nanometer particle (French chalk) at 2 phr and for H44 is the micrometer particle (china clay) at 3 phr. For H50, it is micrometer dimension at 3 phr. Higher hardness SIR allows higher particle size for its percolation limit. Tests showed tensile strength increases with higher addition (up to 4 phr). An improvement of up to 20% can be obtained. An improvement of 50% in volume resistivity against the raw SIR was recorded in the H50 series. For the breakdown withstand test, improvement of 4% and 1% for H50 with china clay and H37 with French chalk, respectively, both at 2 phr filler addition has been obtained. The relative permittivity ranges from 2 to 4 with filler addition against 2 to 3 for raw SIR. Correlations between the results obtained from the measurements and tests had been investigated. Hardness index can be used as an indicator for the performance of the tensile strength and dielectric breakdown, with higher indices indicate higher tensile strength and higher breakdown strength. References [l] Lewis, T.J., Nunometric DieZectrics, IEEE Trans. on Diel. and Elect. Insul., Vol. 1, 1994, pg8 12-825. [2] Roy, M., Nelson, J.K., MacCrone R.K., Schadler, L.S., Reed, C.W., Keefe, R., and Zenger, W., Polymer Nanocomposites Dielectrics - The Role of the Interface, IEEE Trans. on Diel. and Elect. Insul., Vol. 12, No. 4,2005, pg629-643 [3] Fothergill, J.C., Dissado, L.A., and Nelson, J.K., Nanocomposite Materials for Dielectric Structures, Final Report for EPSRC Grant GR/R 71788/01, University of Leicester, U.K., 2004.

This page intentionally left blank

Session P3

Chair: 0.Acher

This page intentionally left blank

Reconstruction of Intrinsic Permeability of Inclusions from the Measured Permeability of a Composite

K. N. Rozanov*,A. V. Osipov, D. A. Petrov, S. N. Starostenko, Institute for Theoretical and Applied Electromagnetics, Moscow, Russia, E. P. Elsukov Physical-Technical Institute, Ural Branch of Russian Academy of Sciences, Izhevsk, Russia Abstract. Microwave material parameters of composites filled with iron powder are studied experimentally. The powder is prepared by mechanical milling. From the measured data on the frequency dependence of permittivity and permeability, the intrinsic permeability of the powder is reconstructed. The effective material parameters of the composites are shown to be affected greatly by the shape distribution of the powder particles. It is found that the Ghosh-Fuchs theory provides an excellent agreement with the measured permittivity and permeability of the composites.

1. Introduction The theory of macroscopic material parameters of a two-component composite typically considers simple idealized models, from which a mixing rule can be deduced to predict, among other properties, the macroscopic permittivity and permeability. A number of mixing rules have been suggested, see, e. g., [l]. The most common of these are the Maxwell Garnet equation (MG):

In Eqs. (1, 2), p and no are the volume fraction and the shape factor (demagnetization factor) of magnetic inclusions, and pi, ,uh, and ,ue are the permeability of inclusions, host matrix, and composite, respectively, all being a function of frequency$ Equations (1) and (2) imply that the effective permeability of composite is determined by the averaged demagnetization factor of inclusions. For an isotropic composite, all possible permeability values are within the HashinShtrikman (HS) bounds [2]. Mixing rules have clear physical meaning and provide a powerful tool for understanding the macroscopic response of composites. Comparison of the measured data with the mixing rules is conventionally made by comparing volume fraction dependence of permeability measured at a certain frequency with predictions of the theory. To calculate the theoretical permeability, the values must be known of all parameters involved, i.e., the volume fraction, the permeability of inclusions, which is taken from the data on bulk material of inclusions, the permeability of host matrix, which is typically equal to unity, and the demagnetization factor, which is conventionally assumed to be equal to 1/3 that corresponds to the spherical shape of inclusions. Many examples of such comparisons are found in the literature, with the best fit shown to be provided by the MG [3], EMT [4], or other theories [5]. In many cases, no conventional theory is able to provide a good agreement with the measured data [6]. Therefore, different composites provide data consistent with different theories. The reason for this discrepancy is that the mixing rules do not take into account the detailed microstructure, or morphology of the composite. That is why, in general, the mixing rules do not allow the macroscopic response of actual composites to be predicted quantitatively: the morphology of composites varies greatly, and the macroscopic response will vary accordingly. The morphology can be taken into account in terms of the Bergman theory [7], which provides an unambiguous and universal characterization of a composite. The theory derives the permeability as: * corresponding author, e-mail krozanov0vandex.i-u

59

60

where B(n) is referred to as the spectral function. The spectral function accounts for a spread in the demagnetization factors. The spread may result from the excitation of inhomogeneous fields within inclusions that is due to either non-ellipsoidal shape of inclusions in composite or the effect of the cooperative phenomena between inclusions. Another important reason for the spread in the demagnetization factors may be a deviation in the shapes of inclusions. The sum rules, 1

1

PO - P ) IB(n)dn = p and JnB(n)dn= ___ 0 0 3 '

(4)

relate the spectral function to the volume fraction of inclusions. The second of Eqs. (4) is valid for an isotropic composite only. It is of importance that the permittivity is governed by the same mixing rule as the permeability does, with the same value of demagnetizatioddepolarization factor no and all p substituted by E , or by the same spectral function B(n). All mixing rules may be considered as particular cases of the Bergman theory. For example, the spectral function for the Maxwell Garnet model is a delta function: B ( n ) = p +(1-p)no),

(5)

The second of Eqs. (4) is conventionally neglected in this approach, because it determines the behavior of B(n) at n-1. Features of B(n) appearing at arguments close to unity have a minor effect on the effective material constants of composite provided that the intrinsic material constant of inclusions is high enough, which is the most frequent occasion when the frequency of interest is not very high. The spectral function for the EMT is given by:

B b ) ={ 4m LJY

n,<xI2 or N2H5(OH). The etching followed one of the following procedures, HF etching, N2H5(0H) etching, and KOH etching. The magnetic moment was measured using a MPMS2 SQUID-type magnet~meter.'~ Atomic force microscopy (AFM) measurements were performed using semicontact mode. Typically, large magnetic moment could be measured only if the samples were treated both with HF and KOH. Etching either with HF alone or KOH alone result in samples that are diamagnetic. Figure 1A presents the magnetic moment measured as a function of the applied field for a p-Si(100) sample after treatment with HF following the additional etching with KOH. The data is presented as magnetic moment per weight. Figure 1B presents the magnetic moment per area obtained from (A) after subtracting the diamagnetic contribution of the bulk. Figures 1C and 1D show the structure function, S ( A r ) , for this surface and the AFM from which the correlation was obtained, respectively. The strong magnetic signal shown is observed only if the roughness exponent, a,reaches high values of about 0.9. High exponent (near one) values in some scale, means that the averaged slopes are equal over this scale. Namely, the features responsible to the roughness are of the size of this scale. Indeed in Fig. 1 the slope is maintained below hundreds of 67

68 nanometers. The correlation between high exponents and, large magnetic moment indicates that interface magnetism is global on the scale of hundreds nanometers, rather than atomic property.

4.0~10~

-

2 g5

2.0~10-~ 0.0 -2.OX10"

eld

I Field (Oe)

-4.0~10-~ -10000 -5000

0

5000

10000

Field(0e)

Figure 1: A) The magnetic moment measured as a function of the applied field for a p-Si(100) sample after treatment with HF and KOH. The data is presented as magnetic moment per weight. B) The magnetic moment per area obtained from (A) after subtracting the diamagnetic contribution of the bulk. C) The structure function, ~ ( A Y )as a function of Ar (see equation l), D) The AFM image obtained for the sample after etching with HF and KOH. The effect exists only for doped silicon, independent of the type of doping, n or p. Etching by NzHs(0H) did not cause the appearance of any significant paramagnetic signal. As in all other cases of interface magnetism reported, also here, no significant temperature effects could be detected. The signal shown in Fig. 1 corresponds to 1.3e-5 Bohr magneton for each Si atom in the sample for an applied field of 1000 Oe. However, if we consider only the surface atoms and assume a flat surface, the magnetic moment measured at 1000 Oe corresponds to about 5.5 Bohr magneton per atom. With increasing field, the magnetic moment increases and can reach up to ten Bohr magnetons per surface atom. Saturation in the magnetic signal was not achieved even at a field of 1 Tesla. Although the surfaces are highly corrugated, we did observe anisotropy in the magnetism, indicating that the magnetic moment is larger when the magnetic field is applied perpendicular to the surface. Temperature-dependence studies revealed that the magnetism observed has no significant temperature dependence; nevertheless, a slight decrease in magnetic moment was observed with decreasing temperature. These studies also confirm that paramagnetic impurities do not contribute to the signal, since any such contaminationwould result in the known 1/T dependence. The surface of silicon contains unpaired electrons that exhibit paramagnetic properties when examined by magnetic resonance spectroscopy. I 4 However, the spin associated with these electrons cannot explain the large magnetic moment observed in the current study. Spin magnetism is not

69

consistent with the temperature dependence measured, with the anisotropy, with the large magnetic signal corresponding to several Bohr magnetons per surface atom, and with the strong dependence on roughness exponent. The large magnetic moment, its anisotropy, and the fact that the magnetism is observed only at interfaces, are all consistent with collective orbital magnetism that is initiated by the charge transfer between the substrate and the thin layer. In the present study, we found a procedure that produces samples with a strong paramagnetic response. The magnetic response is so large that it masks the diamagnetic contribution from the bulk silicon. The fact that the large magnetic response is observed only when the sample is treated both with HF and KOH and not when the etching is performed with HF alone or with KOH alone serves as another proof that the measured signal is not associated with any contamination. The magnetism observed is related to the surface roughness in a complex way, not completely understood. Clearly, when the features defining the surface roughness are on the scale of few hundreds of nanometers, the magnetic response is at its maximum. Hence, the orbital magnetism, causing the observed magnetic moment, is most probably associated with motion of electrons in domains of the size of this order. The magnetic response requires oxidation to take place, so an interface Si-SiO, is formed. Therefore, HF etching that dissolves the oxide layer on the surface results in no paramagnetic contribution. KOH etching that follows HF etching is capable of reacting efficiently with the bare silicon surface (since the HF removed the oxides) and therefore KOH etching is an efficient procedure for introducing large corrugation. Although the precise mechanism and theory related to this new magnetic phenomenon is not yet known, accumulating data clearly indicate that a new type of magnetism exists. The data indicate that the effect is the result of a large magnetic moment for each surface atom and therefore cannot be related to spins only. The present work, in which silicon could be made magnetic, may open the way for using room temperature silicon as a magnetic material, a possibility that has enormous technological implications. Acknowledgements: We thank Prof. Itamar Procaccia for helpful discussion and suggestions regarding the structural function. We wish to thank Dr. Hagai Cohen for performing the XPS experiments and Dr. Yishay Feldman for the AFM work. We thank Dr. Sidney Cohen and Dr. Shirley Daube for critical reading of the manuscript and good advice. This study was partially supported by the Schmidt Minerva Center and the Nancy and Stephen Grand Center for Sensors and Security. References: [I] See for example: G. H. 0. Daalderop, P. J. Kelly, & M. F. H. Schuurmans, Phys. Rev. B, 44,12054-12057 (1991); M. Tischer, 0. Hjortstam, D. Arvanitis, H. J. Dunn, F. May, K. Baberschke, J. Trygg, J. M . Wills, B. Johansson, 0. Eriksson, Phys. Rev. Lett., 75, 16021605 (1995); P. Poulopoulos, K. Baberschke, J. Phys.: Condens. Mutter, 11,9495-95 15 (1999) [2] I. Carmeli, G. Leitus, R. Naaman, S. Reich, Z. Vager, J. Chem. Phys., 118, 10372-10375 (2003); I. Carmeli, G. Leitus, R. Naaman, S. Reich, Z. Vager, Isr. J. of Chem., 43,399-405 (2003). [3] P. Crespo et al, Phys. Rev. Lett. 93,087204 (2004). [4] Y . Yamamoto et al., Phys. Rev. Lett., 93,116801 (2004); H. Hori, et al., Phys. Lett. A, 263, 406-410 (1999).

70

[5] P. Esquinazi, et al., Phys. Rev. Lett., 91,227201 (2003); S. Moehlecke, , P-C. Ho, M.B. Maple, hylosophicalMug. B, 82,1335-1347 (2002). [6] M. Venkatesan, C.B. Fitzgerald, J.M.D. Coey, Nature, 430,630 (2004); J.M.D. Coey, Solid State Sc. 7,660-667 (2004). [7] Z. Vager, R.Naaman, Phys. Rev. Lett., 9, 087205 (2004). [8] A. Hernano, P. Crespo, M.A. Garcia, Phys. Rev. Letts., 96,057206 (2006). [9] M.E.R. Dotto, M.U. Kleinke, Physica A 295, 149 (2001). [lo] Y.J. Chabal, et alJ. Vuc. Sci. Technol.,A 7,2104-2109 (1989). [ l l ] Q.-B. Vu, D. A. Sfricker, P. M. Zavrackyt, J. Electrochem. Soc., 143, 1372-1375 (1996). [12] K. Sat0 et al. Sensors andActuators A, 64,87-93 (1988). [ 131 The gap in the straw itself induces a paramagnetic-like response in the magnetometer; emu for a gap of 300pm, which however, its size was found to be on the order of corresponds to the thickness of the sample after etching. [I41 A. Stesmans, V. V. Afanas’ev, Phys. Rev. B 57, 10030-10034(1998).

Characteristics of Effective Permeability and Resonance Frequency for Barium-ferrite/Epoxy Composites Z. W. Li, Y. B. Gan, Xu Xin and G. Q. Lin Temasek Laboratories, National University of Singapore, I 0 Kent Ridge Crescent, Singapore, 119260 Abstract: A model that considers the demagnetizing interaction between particles in composites with c-plane anisotropy is proposed. Based on the model, we obtained equations on the effective , effective ~ resonance frequencyh,, for the composites. The two equations static permeability ~ 0 and and fR.e on p , which is consistent with experimental results. In predict the dependence of M , ~ -1), ~ f R , e =pC , which is similar to the Snoek's law, is also addition, an important correlation ( p ~ obtained from the two equations. 1. Introduction High-frequency magnetic properties of the composites are determined not only by the properties of the ferrite particle itself, but also by the interaction between the ferrite particles. Two important characteristics, which are significantly different from those of the bulk materials, have been observed: ( 1 ) the effective static permeability ~ 0 is, much ~ smaller than M , b of the corresponding bulk materials, with the effective resonance fR.e shifted to higher frequency; (2) effective m,eand j& are not in linear correlation with the volume concentration p . It has been reported that the static permeability P0,b of bulk NiZn spinel ferrite is as large as 840-1400, with the resonance frequency fR,b of about 10 MHz.[l,2] However, at p=0.4, the effective m,eis only 2.5, while the effectivefR,b is as high as 1500 MHz [2]. Therefore, it is worthwhile studying the dependence of , u ~andfR,e ,~ on p in both theory and applications of composites. In this work, a new model is proposed which shows the dependence of ~ 0andfR,e , ~ on p , consistent with experimental data. In addition, a correlation between ~ 0 andh,,, , ~ similar to the Snoek's law, is also obtained.

2. Model Demagnetizing effect, produced by both the shape of particles and the interaction between particles, plays an important role in determining the effective static permeability pee and the effective resonance frequencyfR,e of composites. 2. I Effective static permeability PO,, Consider particles with a single domain embedded in an infinite and homogeneous epoxy matrix. Under an applied magnetic field Ho, the initial permeability pee of composite is given by

where fi is the angle of the iihmagnetization Mi with respect to the Ho direction, M is the averaged magnetization along HOdirection, and p is the volume concentration of particles in the composite. Consider that HeH;+Hd, where H; is the magnetic field inside the particle and Hd is the demagnetizing field The susceptibility of particle ( p 0 , b - I ) is given by (dM/dH;)H+ and the demagnetizing factor Nd is approximately considered as Hd/M. We have

Consider a composite comprising two particles and a homogeneous epoxy matrix. When the distance of the particles is sufficiently large compared to the size of the particles, the two particles can be considered as isolated particles, and there is no interaction between the particles. However,

71

72

for two closely spaced particles, the positive and negative magnetic charges created on the surface of the two particles under the applied magnetic field will partially cancel out, thus decreasing the demagnetizing field within each particle. Therefore, Nd in Eq.(l) should be replaced by an effective demgnetizing Ad involving the interaction of the two particles, ApFNd (2) where F is the correction function of the demagnetizing factor Nd. It is obvious that the demagnetizing interaction between two particles is related to at least two factors: the distance 6 of the two particles and the size d of the particles. Note thatp-[d/@. It is reasonable to assume that F is a function ofp: F= F(p). Hence, Eq.( 1) is re-written as

2.2 Effective resonance frequency The universal resonance frequencyh,, is given by [4]

for the precession of magnetization vector about a stable direction, i.e. the direction of aE 186' = 0 and aE la4 = 0 . In Eq.(4), the gyromagnetic ratio y/2z is about 2.8 GHzkOe, 6 and 4 are the polar and azimuth angles associated with the direction of M and E is the free energy of system. For a given spherical particle with single domain, only two energies are considered: E=E,+Ei. The magnetocrystalline anisotropy energy E,, is given by E, = K, sin26 +K2sin46 -IK , sin66cos64 (5) The interaction energy Ei between the neighboring particles is given by, 2 . 2 1 1 1 1 E, =-[-A M sin 6cos241 = - [ - F( ~ ) N, M~sin2ecos2 41 (6) 2 2 2 2 The factor 112 before the square bracket indicates that the energy is shared by two particles. For spherical particles, the demagnetizing energy produced by the shape of particles is not considered, because the resonance frequency is not related to the demagnetizing factors. Substituting Eq.(5) and (6) into Eq.(4), the resonance frequency at 6=d2 and &O (for c-plane anisotropy ) is obtained as

In the derivation, NdM f&andp spinel ferrite composites; the solid line is the results curve-fitted based on Eq.(9) . In conclusion, a model that considers the demagnetizing interaction between particles is proposed for ferritefepoxy composites with c-plane anisotropy. Based on the model, we obtained Eqs.(9) and (10) that predict the dependence of effective static permeability m,eand effective resonance frequency fR.e on volume concentration p. These are in good agreement with experimental results for barium femte composites. In addition, the correlation, h,,,- l)f,',,, = pC, between m,eandfR,e,similar to Snoek's law, is also obtained from the two equations.

References: [ 11 T. Nakamura, T. Tsutaoka and K. Hatakeyama, J. Magn. Magn. Mater., 138,3 19 (1994). [2] Y. Konishi and H. Komori, Microwave Optical Tech. Lett., 16, 156 (1997). [3] Z. W. Li, L. F. Chen, Y. P. Wu and C. K. Ong, J. Appl. Phys., 96,534 (2004). [4] J. Smit "Magnetic properties of materials", ed. Smit, (McGraw-Hill), 1971, ppl.

Microwave Absorbing Properties of Amorphous FeCuNbSiB Microwires Multilayer Composites Mangui Han', Difei Liang, Liang Chen, Jianliang Xie, Longjiang Deng, State Key Laboratory of Electronic Thin Films and Integrated Devices, University of Electronic Science and Technology of China, Chengdu, 610054, I! R. of China; Engineering Technology Center of Electromagnetic Wave Absorbers, Ministry of Education, Chengdu, 610054, I! R. of China. (* corresponding author: [email protected] ) Abstract Amorphous FeCuNbSiB microwires have been fabricated by the melt extraction method. The amorphous microwires have been heat-treated at T = 573 K, 673 K, 723 K and 773 K. It is found that the microwave permittivity and permeability are dependent on the annealing temperatures. The microwave absorbing performances of multilayer microwire absorbers have been compared. The results show that the absorbing performance can be modified by combination of different microwire layers. Key words: electromagnetic wave absorber; microwires; amorphous magnets. 1. Introduction Fe-based amorphous and nanocrystalline wires have been widely studied due to their giant magnetoimpedance (GMI) effects, and have been mainly employed in sensor applications, such as magnetic field sensors, position sensors [11, etc. Frequently reported methods of fabricating magnetic microwires include melt-extraction method, Taylor-Ulitovsky method and Taylor method [21. Different fabrication methods can result in different domain configurations in wires [31,and the rapid cooling rate during wire fabrication process also can result in a complex stress distribution in microwires [21. There has been an ever increasing interest in studying the electromagnetic (EM) wave absorbers due to their wide applications in anti-EM interference coating and microwave darkroom [4-51. The frequently used magnetic materials for the EM wave absorbers include ferrite (spinnel ferrites and hexaferrites), ferromagnetic particles, polycrystalline Fe fibers, Co fibedepoxy resin composite, etc. FeCuNbSiB nanocomposite is an excellent soft magnetic material with its cutoff frequency in the GHz range. In this paper, firstly we will investigate the microwave properties (permeability and permittivity) of amorphous FeCuNbSiB magnetic microwires annealed under a series of temperatures, which are below its primary crystallization temperature. Secondly, we will evaluate the EM wave absorbing performances of microwire/wax composite absorbers with single, double, three, and four layer structures. 2. Experimental details The FINEMET@ (nominal alloy composition is Fe73.5Cu1Nb3Si13.5Bg) microwires are fabricated by the melt extraction method. A scanning electron microscopy (SEM) is used to investigate the morphology of as-prepared microwires. A differential scanning calorimeter (DSC) curve is obtained on a differential scanning calorimeter to find the crystallization temperatures of as-prepared microwires. The heating rate for the DSC measurements is 10 K/min. X-ray diffraction (XRD) is used to check the crystallization state for the as-prepared microwires. Annealing treatments on magnetic wires under argon atmosphere for 1 hour are carried out at different temperatures,: 573 K, 673 K, 723 K and 773 K separately. The heat-treated microwires are then cut into 1 mm long, and are homogenously mixed with wax (microwires / wax volume ratio is about

75

76

1:2). The mixtures are made into ring shape specimens for measuring the complex relative permittivity ( E , = E'- je") and the complex relative permeability (pr= p' - jp") using an Agilent 8720ET vector network analyzer within the frequency range of 0.5 GHz - 4.0 GHz.

3. Results and discussions The morphology of as-prepared microwires is shown in Fig. 1 (a). The average diameter of wires is about 50 !.I m. The X R D pattern of as-prepared microwires is presented in Fig. l(b). It shows a typical pattern of amorphous alloys, indicating that the as-prepared microwires are in the amorphous state. The broad peak centered around 28 = 45' confirms the characteristics of the short-range order of Fe-based amorphous alloys. In order to find the crystallization temperatures for amorphous microwires, the DSC measurement has been done. The results are shown in Fig. 1(c). As pointed out in ref. 6 , for FeCuNbSiB alloys, there exists two crystallization temperatures, which are also confirmed by us in Fig. l(c). Our DSC measurements reveal that the primary crystallization temperature is 812 K and the secondary crystallization temperature is 952 K. These two crystallization temperatures are close to those found in a similar alloy composition Fe72CulNb&313,5B9 (TI = 815 K and TZ = 961 K) reported in ref. 6. According to ref. 6, at the primary crystallization temperature, nanocrystalline a-Fe (Si) grains are formed in the amorphous matrix. At the secondary crystallization temperature, nanocrystalline Fe- and Nb- borides are formed. Similar results have also been reported in ref. 7. Therefore, our annealing treatments below the primary crystallization temperature would have an effect on the internal stress, which are formed during the preparation process.

Fig. 1 (a) the morphology of as-prepared microwires; (b) X R pattern of as-prepared microwires; (c) the DSC curve of as-prepared microwires. The microwave dispersion spectra of permeability and permittivity are shown in Fig. 2 (a) (d). Fig. 2 (a) and (b) show the complex relative permeability in the selected GHz region. The variations of permeability are possibly due to the effect of annealing treatment on the magnetic properties of FeCuNbSiB microwires. Here, the annealing treatments below the primary temperature (T = 8 12 K) can result in the following effects: (1) it reduces the residual internal stress. According to ref 8, the internal stress for the as-prepared amorphous FINEMET@ (Fe73,5CulNb3Sir6.5B6) ribbon (70 x 2 . 5 ~ 0.034 mm3) is about 15 MPa. While annealing at T = 853 K, the internal stress is reduced to 0.2 MPa. In our case, the FINEMET@alloy is prepared in the form of microwire. Both the magnitude and distribution of internal stress is different from those of ribbon in ref. 8. However, through their findings, we can learn that the internal stress is highly sensitive to annealing treatments for the FINEMETO type amorphous alloys. (2) The annealing treatments can change the saturation magnetostriction coefficient h, of the microwires. 1, of the amorphous FINEMETO alloy is about 8 ppm [61. Therefore, the magnetoelastic energy can play an important role in the stress-induced anisotropy and the permeability of samples studied. Therefore, the dependence of permeability on the annealing temperatures in Fig. 2 (a) and (b) is possibly due to the relief of internal stress and the variation of magnetostriction constant in the amorphous microwire. The dependence of permittivity on annealing temperatures is also shown in Fig. 2. Fig. 2

77 (c) shows the real parts (d), and (d) shows the imaginary parts (E") of the relative permittivity. For the values of microwires annealed at T = 573 K and 723 K, E' values do not vary significantly within the frequency range. At T = 773 K, E' has a maximum value at f = 2 GHz; while for T = 673 K, E' has a minimum value at f = 2.0 GHz and have two peaks at f = 0.8 and 3.6 GHz. At T = 723 K and 773 K, E" decreases with increasing the frequency, showing a relaxation feature. However, for T = 573 K and 673 K, E" increases with increasing the frequency. Also, a close examination on Fig. 2 (d) reveals that in the frequency range of 0.5 - 3.0 GHz, values of microwires annealed at T = 723 K, 773 K are much larger than those of microwires annealed at T = 573 K, T = 673 K. Obviously, the permittivity is strongly dependent on the frequency and the annealing temperature, especially, in the lower frequency region.

NGnrl

Fig. 2. Microwave permeability (a, b) and permittivity (c, d) properties of amorphous microwires heat treated at different temperatures. Based on the measured complex permittivity and permeability, and based on the assumption that a single layer of microwires/wax composite is attached on a metal plate, the electromagnetic wave absorbing performances can be evaluated by the following equation ['I: RL = 20 log((Z,,,- Z,) /(Zin + Z,)I

(1)

where the RL denotes the reflection loss in dB unit. A lower RL value indicates a better EM wave absorbing performance. ZO is the characteristic impedance of free space. Zi, is the input characteristic impedance at the absorber/free space interface, which can be expressed as:

c is the velocity of light, f is the frequency, t is the thickness of an absorber. pr and E~ are the measured data for the relative complex permeability and the relative complex permittivity respectively. The matching thickness (tm)and the matching frequency (fm)are defined as the t value and the f value associated with a minimum RL value. The optimized EM wave absorbing performances of single layer absorbers are compared in Fig. 3(a). Clearly, the microwires heat treated at T = 723 K show the best absorbing properties among four absorbers with single layer of microwires. For T = 723 K, the minimum RL value ((=)fin) is about -22 dB,tm is about 9.5 mm, and f, is 1.2 GHz. For T = 573 K, the ( R L ) ~ n is about -10 dB,and the tm and fm are 7.5 mm and 1.2 GHz respectively. However, the microwires heat treated at T = 773 K, 673 K show poor absorbing performances. In order to meet the growing demands for thinner absorbers, we have studied the absorbing properties of microwire absorbers with multilayer structures to see whether thinner microwires absorbers can be obtained. Theoretically, for multilayer absorbers, the following impedance matching condition should be satisfied to get an acceptable absorbing performance [lo].

Where q = qo(pi/&i)"*,q o = ZO.Zl is the wave impedance of the out-most layer absorber. The RL values for a multilayer absorber can be calculated as: RL = 20104(Z, - Z,) /(Zl + Z,)l.

78

Fig. 3 the absorbing performances of microwires/wax composites, #1, #2, #3 and #4 denote the microwires heat treated at T = 573 K, 673 K, 723 K and 773 K respectively. The absorbing performances of absorbers with multilayer structures (2, 3 and 4 layers of microwires/wax composites) are shown in Fig. 3 (b), (c), (d), (e) and (f). All the t values in Fig. 3 are in mm unit. For the two layers structures, comparing all the possible sequence combinations of microwires treated at different temperatures, we find that the double-layer absorber with the microwires heat treated at T = 573 K and 723 K shows a better absorbing performance, see Fig. 3 (c). (=)fin is about -19 dB and tm is 8.5 mm. In Fig. 3 (b), the RL values of the absorber with microwires heated treated at T = 573 K and 673 K are inferior to those in Fig. 3 (c). The absorbing performances of three-layer microwire/wax composites are shown in Fig. 3(d) and (e). The combination of microwire/wax layers in Fig. 3(d) show that (=)fin is about -15 dB, t, is 5.2 mm, and f, is 2.6 GHz. In Fig. 3(e), (RL)fin is about -18 dB,tm is 8 mm and f, is 1.55 GHz. For the four-layer microwire/wax composites, the combination in Fig. 3(f) shows the best absorbing performance: (RL)minis - 19 dB,tmis 9.1 mm, and f, is 1.2 GHz. As a conclusion, the results in Fig. 3 show that the absorbing performance (the t,, f, and RL values) of microwires/wax composites can be adjusted by selecting a suitable multilayer structure. With their (RL)fin values close to that of the single layer absorber (T = 723 K in Fig. 3(a)), the tm values of multilayer absorbers (Fig. 3 (c)-(f)) are smaller.

References [l] M. Han, D. F. Liang, L. J. Deng, J. Muter Sci., 40,5573 (2005). [2] H. Chiriac, T. Ovari, Prog. Muter. Sci. 40,333 (1996). [3]K. Mohri, F. Humphrey, K. Kawashima, K. Kimura, M. Mizutani, ZEEE Trans. Magn., 26, 1789 (1990). [4] J. R. Liu, M. Itoh, Appl. Phys. Lett. 83,4017 (2003). [5] S. Sugimoto, T. Maeda, D. Book, T. Kagotani, K. Inomata, M. Homma, J. Alloys Compnds. 301, 330 (2002). [6] M. McHenry, M. Willard, D. Laughlin, Prog. Muter Sci. 44,291 (1999). [7] D. M. Lin, H. S. Wang, Y. C. Wu, M. L. Lin, Chin. Phys. 8,455 (1999). [8] M. Carara, M. Baibich, R. Sommer, J.Appl. Phys. 91,8441 (2002). [9] T. Maeda, S. Sugimoto, T. Kagotani, N. Tezuka, K. Inomata, J. Magn. Magn. Mater. 281, 195 (2004). [ 101 M. R. Meshram, Nawall K. Agrawal, Bharoti Sinha, P. S. Misra, J. Magn. Magn. Mater. 27, 207 (2004).

Curve-fitting of Complex Permeability and its Applications

Z. W. Li Temasek Laboratories, National University of Singapore, Singapore I19260 Abstract: The curve-fitting program for complex permeability was made based on a model that considers the distribution of resonance frequency. A smoothing factor is introduced to eliminate unreasonable oscillation of resonance frequencies due to the statistical error in experiment. The program is successfully applied to the analysis of complicated permeability spectra.

1. Introduction High-frequency properties of materials are described mainly by their complex permeability spectra. Typically, the spectrum comprises two or more components of resonances and exhibits complicated characteristics. Therefore, it is important to decompose the spectrum into natural and wall resonance components and acquire the corresponding high-frequency parameters. In general, the natural and wall resonance permeability spectra are described by Kittel [11 and Lorentz equations [2], respectively. For bulk materials, typically one resonance frequency is considered in the equations. Consider the composite filled with randomly distributed barium ferrite particles. First, the ferrite particles are embedded in epoxy. The shape and size of the particles, the porosity, and the distribution of particles in composites will lead to inhomogeneities in demagnetizing effect and magnetic anisotropy. Second, without an applied magnetic field, the unsaturated magnetization vector is randomly distributed along the easy-magnetization direction for each particle. These factors lead to a distribution of resonance frequencies, instead of a single resonance frequency. In this work, a curve-fitting program for complex permeability was developed based on a model with a distribution of resonance frequencies. 2. Method The experimental complex permeability ye, (i=l to m) is considered to be the weighted sum of n sub-spectra LOGql) with intrinsic resonance frequencies FRJ(j from 1 to n).

where .ziis the error between the experimental and theoretical data, Lii @ q$ is the theoretic function of jthcomplex permeability, Pj is the corresponding probability, and ql is the curve-fitting parameters. A possible representation of function L , @ q$ could be the Kittel equation, given by

for natural resonance, where Fn,r is the intrinsic resonance frequency and il is the damping coefficient. Another possible representation of Lii @ q$ could be the Lorentz equation

for wall resonance, where F , , and F w , ~are the relaxation and intrinsic vibration frequencies, respectively. To eliminate the non-physical oscillation of resonance frequencies due to the statistical error in experiment, a smoothing function d is introduced j = 1,2 ...n d j = Pj_,- 2Pj + Pj+, (4) where PO=PI=P,=P,+I=O.Therefore, the objective function CD can be expressed as

79

80

where g is the smoothing factor with O

* Corresponding author: Elena Shalyguina. Tel.: +7 (495) 9392435; fax: +7 (495) 9328820. E-mail address: [email protected]. 169

170

field H, applied parallel to the ribbon length L and the AC measuring current (i = 5 mA). Helmholtz coils created the cyclic magnetic field. The magnetic field dependencies of the magnetoimpedance (MI) ratio for different values of the current frequencyf(0.1 - 10 MHz) were obtained by plotting AZIZ (%) = {Z(H) - ZSAT}/ZSAT x 100% for the cyclic field H. Here Z ~ AisTthe magnetoimpedance Zfor H = 40 Oe. 3. Results and Discussion The strong influence of annealing temperature on the local magnetic properties was revealed. For illustration, Fig. 1 shows the typical the local magnetization curves, observed for the free sides of the as-cast and annealed at T = 470 "C samples. Fig. 2 displays the typical distributions of the magnetization components, parallel to H, observed for the same samples by scanning the light spot of 20-pm diameter along the ribbon width, W. From Fig. 1 one can see that in the as-cast sample, there is the negligible distinction of the near-surface local magnetization curves. The local hysteresis loops were revealed to have analogous behavior. This result is evidence of the high homogeneity of the local magnetic properties, which can be ascribed to the slight dispersion of the magnetic anisotropy of the as-cast sample. It should be pointed out that these data differ from ones obtained by us for amorphous ribbons of other compositions [lo], in which the strong dispersion of the magnetic anisotropy was found. At the same time, from Fig. 1 one can see that in the annealed at T = 470 OC sample, the local magnetization curves are practically identical, i.e., the homogeneity of its local magnetic properties rises. The magnetization distributions, observed for the examined samples, show also the improvement of the homogeneity of their local magnetic properties after annealing. In particular, from Fig. 2 one can see that in the as-cast sample, the distributions of MlMs( w> exhibit a few almost periodically repeating peaks, but in the annealed sample, the values of MIMS change insignificantly. Analogous data were obtained for the samples, annealed at other temperatures. Moreover, the same peculiarities of the near-surface local magnetic characteristics were observed for the wheel ribbon sides.

Fig. 1. The typical local near-surface magnetization curves observed by T E ' for the free sides of the as-cast and annealed at T = 470 "C ribbon samples. The magnetic field H was applied parallel to the sample length L and perpendicular to the plane of incident light.

0,8

-.-= H

O'O 0

T = 470 OC

-

20 Oe

200 400 600 800 W (urn)

OVo 0

H=20Oe

200 400 600 800 W(rm)

Fig. 2. The typical distributions of the magnetization components, parallel to the magnetic field H, observed for the free sides of the as-cast and annealed at T = 470 "C samples by scanning the light spot of 20-pm diameter along the ribbon width, W, at the fixed value of L.

171 200

-

150

6100 50 0 0

200 400 600 8001000 TOC

0

200 400 600 800 1000 T 'C

Fig. 3. Dependences of the near-surface and bulk magnitudes of the coercivity, Hc, (a) and the saturation field, Hs, (b) on the annealing temperature.

It was found that with increasing the annealing temperature, the bulk magnetic properties of the samples change also. Fig. 4 displays the dependences of the near-surface and bulk magnitudes of the coercivity, Hc, and the saturation field, Hs, on the annealing temperature. From Fig. 4 one can see that the bulk near-surface and bulk values of HC and Hs have analogous behavior but the bulk magnitudes of HC and Hs are smaller (about 4-5 times) than the near-surface ones. This fact can be explained by the presence of surface roughness and microstructural changes, in particular, the increase of the concentration of metalloid atoms at surface layers that is characteristic for materials, prepared by the melt-spinning technique. Moreover, the values of HC and Hs for the free ribbon sides are lower than those for the wheel sides. Such difference of the near-surface magnetic properties is typical for the amorphous ribbons and can be explained by both different residual stresses, induced at the wheel and free sides during the quenching and annealing process, and the distinguishing morphology of these sides. One can see also that the dependences of Hc(7) and Hs(T) are complicated. The discovered temperature behavior of Hs and Hc can be explained by microstructural peculiarities of the samples. The table 1 shows phase content of the as-cast and annealed samples, determined by using X-ray diffraction data. According to XRD data, temperature T 580 OC corresponds to the initial stage of isothermal crystallization. So, the as-cast and annealed at T = 470 and 525 "C ribbon samples remain amorphous. The improvement of the magnetic properties of the annealed samples in comparison with the as-cast one can be explained by temperature changing an effective constant of the induced magnetic anisotropy, K$,* oc hso. Here hs and o are the magnetostriction and internal stresses, arising in the process of the ribbon production, respectively. It is known [ll] that a thermal treatment of amorphous ribbons causes the decrease of stresses o. As a result, the magnitude of K3a@decreases that causes the diminution of HS and Hc. Minimal values of HC and HS were found for the sample annealed at T = 525 "C. At T = 580 OC, nanocrystalline a-Fe, NiCr3Si and (Fe,Mn,Cr)zB phases with the dimensions of nanocrystallites d, equaled approximately to 8 nm, arise from amorphous phase. With increasing annealing temperature to T = 820 OC, instead of a-Fe phase, a new Fe3Si phase appears. The magnitude of d for all three phases is about of 20-30 nm. The presence of three high dispersible phases in the alloy causes a high microhardness (HV = 16.5 GPa) and, as consequence of it, the values of Hs and Hc increase abruptly.

-

172

f

(MHz)

Fig. 4. Dependence of MI ratio peaks on frequency f, observed for the as-cast and annealed at T = 525 O C

samples: (a) and (b), respectively. The subsequent increase of the annealing temperature is accompanied by the completion of crystallization processes and, in particular, by the appearance and the growth of (Cr, Fe)4B phase and also the increase of Fe$i phase content that cause the decrease of the microhardness (W = 11.O GPa) of the annealed samples. As a result, the magnitudes of Hs and Hc decrease. It was revealed that the thermal treatment of the samples influences also on their dynamic magnetic characteristics. For illustration, Fig. 4 displays the dependence of peaks of the magnetoimpedance ratio on the frequency f of the AC measuring current observed for the as-cast and annealed at T = 525 "C samples. It was found that in the as-cast and annealed at T = 525 OC samples, the maximum values of the magnetoimpedance ratio AZ/Z (YO) are equal to 25 and 140 %, respectively.The maximum value of AZ/Z ("YO) was observed for the sample annealed at T = 525 "C. 4. Conclusion The study of magnetostatic and dynamic magnetic properties of the Fe614Ni36Cr3 ZSiZ 4Nb7 BMn3 6B18 multicomponent as-cast and annealed ribbons showed that the determined regime of the thermal treatment of the studied alloy allows to obtain the essential improvement of the above characteristics. In particular, the annealed at T = 525 "C sample exhibits soft magnetic properties, the high homogeneity of the near-surface magnetic characteristics and the large magnetoimpedance ratio. It was found that the appearance of high dispersible multiphase structure in the annealed samples on the initial stage of crystallization processes is accompanied by the growth of the coercivity and the saturation field, i.e., the thermal treatment of the above ribbon causes the transition from a soft magnetic state to a hard magnetic one without changing its composition. As a result, a making of different magnetic composites is possible. At last, one can point out that the above ribbon was used as precursor for obtaining bulk alloys. Results of the investigation of magnetic properties of the received bulk alloys will be presented somewhere else. Acknowledgement

The work has been partly sponsored by Russian Fund of Fundamental Investigation (Grant N 05-02-16293). References [ 11 A Inoue, Acta. Mater. 48,279 (2000).

[2] A. Inoue, A. Makino, T. Muzushima, J. Magn. Magn. Mater. 215416,246 (2000). [3] Molokanov V.V., Petrzhik M.I., Mikhailova, T.N., Kuznetsov, I.V. Rus. J. Metals. 5, 112 (2000). [4] A. Inoue, A. Makino, T. Muzushima, J. Appl. Phys. 81,4029 (1997). [5] A. Inoue, T.,, Zhang, H. Koshiba, J. Appl. Phys. 83,6326 (1998). [6] A.Makino, T. Bitoh, I. Murakami, T. Hatanai, A. Inoue, T. Masumoto, J. Phys. France (Part 2) 8, 103 (1998). [7] A. Inoue, T. Zhang, H. Koshiba, T. Itoe, Mat. Res. SOC. Symp. Proc. 554,251 (1999). [8] H. Chiriac, N. Lupu, J. Magn. Magn. Mater. 2 1 5 4 6 , 3 9 4 (2000). [9] V.V. Molokanov, A.N. Shaligin, M.I.. Petrzhik, T.N. Mikhailova, K.S. Filipov, B.I.. Kashin T.A. Sviridova, N.P. D'aykova, Rus. J, Perspective Mater. 3, 10 (2003). [lo] E. Shalyguina, L. Bekoeva, N.Tsidaeva, Sensors h Actuators. 81,216 (2000). [ 111 D. Atkinson, P.T. Squire, M.R.J. Gibbs, S. Atalay, D.G. Lord, J. Appl. Phys. 73,341 1 (1993).

XANES Investigations of Interatomic interactions in (CoFeZr),(SiOz)l, nanocomposites

E.P. Domashevskaya', S.A. Storozhilov', S.Yu. Turishchev', V.M. Kashkarov', V.A. Terekhov', O.V. Stognej2,Yu.E. Kalinin', A V. Sitnikov', S.L. Molodtsov3. I Voronezh State University, Universitetskayapl. I , 394006, Voronezh, Russia, $t@,nhvs. vsu.ru. 2 Voronezh State Technical University, Moskovskii pr. 14, 394026, Voronezh, Russia. Berliner Elektronenspeicherring-Gesellschajifur Synchrotronstrahlungm.B.H., Albert-EinsteinStr. 15, 12489, Berlin, Germany. 1. Introduction Magnetic nanocomposites before percolation threshold represent superparamagnets where ferromagnetic metal grains having size of several nanometers are situated in the bulk of dielectric matrix. Such nanocomposite materials possess a number of physical properties that differ them from the usual materials: a giant magnetoresistance (GMR), a good ability to absorb electromagnetic radiation in HF and UHF ranges, ability of a wide change of the resistivity value and so on [1-31. Granulated nanocomposites can be undoubtedly related to the advanced materials for the application in different electronic devices. The main purpose of the work is performing of the experimental investigations concerning the nature of interatomic interactions and phase composition of amorphous nanocomposites of (Co45Fe45Zr10)~(Si02)1-~ with the use of X-ray absorption near edge structure (XANES) technique that is most sensitive to the chemical environment of the elements in multi-component amorphous compounds.

2. Experimental Composites of (C045Fe45Zrlo),(Si02)1~~ were obtained by ion-beam sputtering of the composite target on glass-ceramic substrate. A composite target represented the plates of Co45Fe45Zr10 alloy with charges of single-crystalline quartz. Composition of the deposited composites was varied by the change of the plates number, i.e. the ratio of metallic and dielectric components in rather wide limits. The samples represented the films of nanocomposite with a thickness of 3-5 pm on glass-ceramic substrate with a content of metallic component of 34 - 64 at.% [3]. In order to perform the investigations of composites structure (CO~~F~~~Z~IO)~ films with different composition were obtained with a thickness of about 50 nm. The structure of obtained amorphous composite materials was investigated using electron microscopy analysis which has demonstrated that formation of amorphous nanogranulated structure occurred in the process of deposition (fig. 1). With an increase of the metallic component content x in nanocomposite the mcan size ofthe grains increased from 2 to 8 nm [3].

Fig. 1. Microstructure and electron-diffractionpattern of (Co45Fe45Zr10)62.4(Si02)37.6 [3].

173

174

The study of magnetoresistance properties of the obtained nanocomposites showed the presence of GMR 3.5% in magnetic field H = 11 kOe. Previously it was found that GMR is due to spin-dependent tunneling of electrons through isolating barrier between the grains [ 11. XANES spectra were obtained at the Russian-German beamline of BESSY I1 synchrotron (Berlin) having the energy range of 30 - 1500 eV. Energy broadening and the relative error in the determination of absorption was of 0,2 eV and 1 %, respectively. In the experiment total electron yield (TEY) was measured using channel electron multiplier. The depth of analysis for XANES in TEY mode is of about 5 nm.

-

3. Results and their discussion Interpretation of XANES spectra was made in the dipole rules approximation applying contemporary ideas on the multiplet structure of the spectra as well. In the dipole approximation XANES Lz,~-spectrarepresent the distribution of the local partial density of states (LPDOS) of s,d symmetry in the conduction band while XANES K-spectra represent LPDOS of p-symmetry. Let us first consider XANES spectra of the elements that form metallic component of nanocomposites, i.e. Fe L2,3, Co L2,3 spectra, and then XANES spectra of dielectric component of nanocomposites, i.e. Si L2,3 and 0 K spectra. 3.1 XANES Fe L2.3 spectra of nanocomposites XANES Fe L2,3 spectrum of nanocomposite consists of two absorption edges L3 and L2 separated by the energy of spin-orbit splitting 13 eV (fig. 2). In addition, we present Fe L2,3 spectrum in fig. 2 obtained by us for a mixture of two phases of iron oxides FeO + Fez03 in equal ratio 1:1, as well as L2,3 spectra of high resolution for Fe and its oxides FeO, Fe203and magnetite Fe304 (FeO.Fe203) [4]. One can see that L2 and L3 absorption edges show multiplet structure. In particular, Fe L3 edge consists of two clearly separated peaks, the so-called white lines at the energies of 708,5 eV and 710 eV, i.e. the value of splitting is lODq = 1.5eV. Their appearance in the spectra of d-metal compounds is due to the effect of crystalline field as well as to p-d and d-d coulomb and exchange interaction [5]. These peaks at L3 edge represent transitions of 2p63d" -+ 2p53d"+'type into low-energy t2g and high-energy eg states. In the compounds involving divalent ions of Fez+,the main maximum of XANES L2,3 spectra is due to the transition into the low-energy

-

if2

nr ne ns m m

IS7 s ils

m m

tw

)46

7%

smm m ec8

Energy lev

Fig. 3. XANES Co L2,3 spectra of nanocomposites and reference XANES L2,3 spectra of Co and COOI141.

175

t2gstate while in the compounds with trivalent ions of Fe3+the main maximum in these spectra is due to the transitions into the high-energy eg state. Fe L1 edge has similar structure. Comparing the obtained spectra with the reference ones one can conclude that nanocomposites involve as trivalent Fe3+,as divalent iron Fe2+.The ratio of the peaks I(Fe3')/r(Fe'2) in L3 and L2 edges changes non-monotonously with an increase of metallic component in nanocomposite that corresponds to the mixed tetrahedron and octahedral symmetry of the nearest neighbour environment for the atoms of iron as, for example, in silicates, minerals or ferrites [6]. 3.2 XANES Co L2.3 spectra Fig. 3 represents XANES L2,3 spectra of cobalt in nanocomposites demonstrating two edges of L3 and L2, separated by the energy gap (- 15 eV) of spin - orbital interaction similar to L2,3 spectra of iron. From comparison with the reference spectra it follows that L3 edge in nanocomposites represents unresolved fine structure that corresponds to multiplet structure of the spectrum of the oxide COOcontaining double-charged ion Coz+in octahedral environment. Thus, metallic component represents cobalt nanoferrite FezO3-CoO (being a semiconductor, ferromagnet), where a half of Fe3+ions is in tetrahedron environment while the rest of Fe3+ions together with Co2+ions are in octahedral environment. Nanoferrite is surrounded by the shell of iron silicate FezSiO4 (Mott-Hubbard type insulator, antiferromagnet), where Fe2+ions are arranged in octahedral environment. Unfortunately, we could not obtain XANES Zr M2,3 spectra of metallic component, possibly, due to the low concentration of Zr atoms (- 5%) of the total composition of nanocomposite. In order to confirm the latter assumption and to analyze dielectric component in (C045Fe45Zr1&(SiO2)1-~nanocomposites XANES Si L2,3 and 0 K spectra were also obtained. 3.3 XANES Si L2.3 spectra XANES Si L2,3 spectra represent transitions of 2p -+ 3s,d into non-occupied states of the conduction band. Fig. 4 gives Si L2,3 spectra of nanocomposites together with the reference spectrum of amorphous Si02, demonstrating a pre-peak in the form of a doublet at 105.3 and 105.9 eV (transitions of 2~112,312-+ alp) and a maximum at 107.8 eV (transitions of 2p -+ tZu), as well as ELNES Si L2,3 spectrum of silicate FezSi04 [7]. Comparing the experimental spectra with the reference spectrum of SiO2 one can see that Si Lz,3 spectra of nanocomposites demonstrate a smeared structure instead of a distinct doublet pre-peak and thus they represent a mixture of the spectra characteristic of silicon oxide and iron silicate.

i.

92

l . 1 . 1 . 1 . 1 . 1 . 1 . 1 . ~ . I . l . t . ~ . , , J

S4 86 98 100 102 104 106 $08 110112 114 118118 120 122

Energy [evl

Fig. 4. XANES Si L2,3 spectra of nanocomposites, Fig. 5. XANES 0 K spectra of nanocomposites SiOz and ELNES Si L2,3 in FezSi04 [7]. and FeO, Fez03 [S], SiOz , ELNES of COO [9].

176

3.4 XANES 0 K spectra XANES 0 K spectra of oxygen in nanocomposites represent the local partial contributions of the density of 0 2p states, participating not only in the formation of the dielectric component of Si02 but metal-containing component as well. 0 K - spectra of absorption edges in Si02 have no any fine structure. However, the spectra of oxides of 3d - metals show a narrow absorption band (pre-peak), that is due to the transitions of 0 1s 0 2p into non-occupied antibonding 0 2p states hybridized with Me 3d states in the conduction band that can be interpreted as t2g, eg symmetrical bands separated by the crystal field of ligand [8]. Second broad band with some specific features in the energy range of 538 - 544 eV is also due to the transition of 0 1s 0 2p, though into unoccupied 0 2p states hybridized with 4s,p states of d-metals which can be related to 3a1, and 4t1, orbitals, respectively [8]. From comparison of the experimental XANES 0 K spectra with the reference ones presented in fig. 5 it follows that the obtained spectra for the samples with x = 0.34 and 0.53 correspond best of all to the summarized spectrum obtained by summation of the reference XANES spectra of oxides in the ratio of 0.2Fe0+0.3Fe203+0.5CoO. The differences of the spectra for the rest of the samples can be due to the formation of silicates of transition metals along with the formation of their oxides. Thus, a complicated shape of XANES 0 K spectra means that oxygen can participate in the formation not only of the dielectric component of nanocomposite but also in the oxides of its metallic component (ferrites), as well as silicates of d - metals that bind together metallic and dielectric components. --f

--f

4. Conclusions Analysis of XANES spectra of (C045Fe45Zrlo),(Si02)1-, nanocomposites means the presence of interatomic interaction between the atoms of metallic and dielectric components in nanocomposite. The most active role in this process belongs to oxygen ions along with the ions of d - metals (Fe3+,Fe+2,Co2+)which form not only nanoferrites of Fe20yCo0, but also silicates of d metals. Thus, local analysis characteristic of XANES technique allowed to find out that (C045Fe~5Zrlo)~(SiO2)1-~ nanocomposites represent more complicated multiphase cluster system than it was assumed previously while explaining their macroscopic properties [1-31.

References [l] O.V. Stogney, Yu.E. Kalinin, A.V. Sitnikov, I.V. Zolotukhin, V.A. Slusarev, Fizika metullov i metullovedenie, 91, 21 (2001). [2]L.V. Lutsev, N.T. Kazantseva, LA. Tchmutin, N.G. Ryvkina,Yu.E. Kalinin, A.V. Sitnikoff. J. Phys. Condens. Mutter, 15, (22), 3665 (2003). [3] Yu.E. Kalinin, A.N. Remizov, A.V. Sitnikov, N.P. Samtsova. A h . Muter. 3,62 (2003). [4] T.J. Regan, H. Ohldag, C. Stamm, F. Nolting, J. Liining, J. Stohr, R.L. White. Phys. Rev. B, 64, 214422 (2001). [ 5 ] C. Colliex, T. Maunobi, C. Ortiz. Phys.Rev. B, 44, 11402 (1991). [6] Laurenca J. Garvie, A.J. Craven, R. Brydson. Am. Mineral. 79,411 (1994). [7] L.A.J. Garvie, P.R. Buseck. Am. Mineral. 84,946 (1999). [8] Z.Y. Wu, S. Gota, F. Jollet, M. Poll&, M. Gautier-Soyer, C.R. Natoli. Phys. Rev. B, 55, 2570 (1997). [9] C. Mitterbauer, G. Kothleitner, W. Grogger, H. Zandbergen, B. Freitag, P. Tiemeijer, F. Hofer. Ultramicroscopy, 96,469 (2003).

Session P9

Chair: J.A. Kong

This page intentionally left blank

Why Periodic Structures May Not Be Able to Synthesize Negative Indices of Refraction Ben A. Munk, Life Fellow Abstract In this paper, we first list some of the features that are widely accepted as being facts regarding materials with simultaneously negative p and E, namely: a) negative index of refraction; b) advance of the phase of a signal as it moves away from the source; c) an increase of the evanescent waves as they get further away from the source; and d) while the E- and H-field in an ordinary material form a right handed triplet with the direction of phase propagation, they will in a material with negative p and E form a left handed triplet. Such materials have never been found in nature. However, numerous researchers have suggested ways to produce them artificially. Periodic structures of elements varying from simple straight wires to very elaborate concoctions have been claimed to produce negative index of refraction. Nevertheless, we shall here show that according to a well known theory based on expansion into inhomogeneous plane waves, it does not seem possible to obtain the features that are characteristic for materials with negative p and E as listed above. Thus, it seems logical to re-examine Veselago's original paper. We find that it is mathematically correct. However, when used in certain practical applications like, for example, the well known flat lens, it may lead to negative time. While such a solution might be mathematically acceptable, it will violate the causality principle from a physical point of view. So it should not surprise us that we so far have encountered difficulties when trying to create materials with negative p and E, in particular negative index of refraction. Ordinary vese1ag.3 Medium

I

1. What currently is assumed about Veselago's medium.

for n,

1.1. Negative Index of Refraction Veselago [ 11 concluded in his original paper that the index of refraction, nl, between an ordinary medium and one with negative E and p would be negative. Thus, as illustrated in Figure 1, the refraction angle, Or, would, according to Snell's Law, have the same sign as the angle of incidence, 4, when nl > 0, while it would be negative for nl < 0.

Medium

o

Refracted for n, c 0 Incident

Figure 1. Snell's Law for an ordinary medium adjacent to Veselago's medium for index of refractionnl> 0 and nl < 0

1.2. Phase advance when nl c 0. If a lossless dielectric slab is placed in front of a groundplane, then the input impedance, Zi,for an ordinary material with nl > 0 will be obtained by a rotation pd = b n l d in the clockwise direction as shown in Figure 2. Similarly, if nl < 0, Zi is obtained by rotation in the counterclockwise direction. In other words, we experience a phase delay for n1 >O and a phase advance when nl 0.

179

Orn""d Plane

Figure 2. Perception of the input impedance Z, of a dielectric slab in front of a groundplane for index of refraction nl > 0 (top) and for nl < 0 (bottom).

180

1.4. The field vectorsB, @ and the phase vector form a left-handed triplet for n1 0. The phase match between the incident and refracted field was explained by the concept “backward traveling waves” as discussed in [18,19,20,21]. However, no trace of such waves was found in a lossless periodic structure. However they can exist on cables terminated in a proper load. Experimental evidence of negative refracted fields in a finite periodic structure is plagued by persistent unexplained loss in excess of - 20 dB [10,14]. This writer has suggested that the field observed is not a refracted field but radiation from a surface wave characteristic for finite periodic surfaces [ 141. Further, we found no evidence that periodic structures with interelement spacings less than 2 2 could change the direction of the incident field as one would expect for index of refraction n # 1. 2. It is widely believed that the input impedance of Veselago’s medium mounted in front of a groundplane can rotated the “wrong” way (counterclockwise) in the Smith chart, See Figure 2 and ref [2,3,4]. We found absolutely no indication of such a phenomenon in lossless periodic structures suspended in a dispersionless medium. 3. Just like propagating waves in Veselago’s medium can rotate the “wrong” way in a Smith chart, it is quite logical that evanescent waves might increase. In fact, it is generally believed that Veselago’s material will support an “evanescent” wave that increases as you move away from the source, see for example ref [18], Figure 3.27 and Section 3.7. We found that a periodic structure could only produce truly evanescent waves that would decrease as you move away from the individual arrays. 4. Veselago claims that a plane wave propagating through his material is left-handed, i.e. E , H and the direction of propagation (phase) forms a left-handed triplet, while E , I? and the Poynting’s vector (energy direction) form a right-handed triplet as usual, regardless of the handedness of the medium. This implies that we will observe a time advance as we move

184

away from the source, see Figure 2 as well as [2]. This concept is alternatively explained by “backward traveling waves” [20,211. (Note that very few of the classical textbooks treat this subject at all.) However, we found from rigorous calculations that the field from an infinite periodic structure regardless of the element shape is always right-handed both inside and outside the periodic structure. Further, there was never any trace of “backward waves” whatsoever. And as all experienced antenna engineers know, nothing ever moves backward in a Smith chart as long as our load impedance is purely imaginary (Foster’s reactance theorem). It should finally be emphasized that all impedance components in the discussion so far have been completely lossless including the termination of the space behind the periodic stmcture. When resistive or dielectric loss is present, the situation changes radically, even if only the termination is lossy. Basically we will, in that case, move inside the rim of the Smith chart such that Foster’s reactance theorem no Ionger holds. This case will be discussed elsewhere where we will illustrate a typical case in the form of a transmission line terminated in a complex load. This is a little easier to comprehend than a periodic structure and it is furthermore discussed already several places [18,19]. Subsequent extension to periodic structures will be facilitated.

5. Appendix: Veselago’s Flat Lens: Is It Really Realistic? The concept for Veselago’s flat “lens” is n,*= -1 by now well known as shown in Figure 8. It consists of a flat slab where EZ, p2 not only is &+I = -pl (i.e., negative but where also ~2 = -&I and ,u~ n12 = -1) such that the refracted angle always is the negative of the angle of incidence. We show two rays emanating from the source point S located to the left. They cross inside the lens at a point denoted Cross 1 and outside to the right at a point denoted Cross 2. Such crossings are often thought to be focal points. However, more is required for such a classification. Foremost of all, we must require that all rays arrive with the figure 8. Veselago’s flat lens with 6 = - E ~and p2= same phase. Inspection Of the two rays in Figure -pl. The longest path ray will be delayed in phase 2 clearly shows that ray SB is delayed in phase corresponding to - 4 8 but be advanced in Veselago’s medium corresponding to B-43 (see also with respect to ray , y by~ section ~ A 1 ~ Further, . Figure 2). However, if the two rays are to arrive section BA3 is inside the metamaterials where the at the same time at Cross 1, it must involve signal is advanced precisely by the same amount negative time in Veseiago’s medium. according to Figure 2 such that the two rays will arrive at Cross 1 in phase. However, we must also require the two rays to arrive at the same time at the crossing. Obviously that would require the time delay A$ to be cancelled by a time advance BA3, i.e. negative time! While negative time does not “offend” mathematicians, it is definitely not an option open to physicists and engineers. For further discussion about causality, see [22,23]. Thus it is no wonder that we have trouble synthesizing Veselago’s medium!

6. References [ 11 V.G. Veselago, “The electrodynamics of substance with simultaneously negative values of

&andp,” Sov. Phys. Uspekhi, vol. 10, no. 4, pp. 509-524, Jan 1968. [2] N. Engheta and R.W. Ziolkowski, Metamaterials, Physics and Engineering Explorations, IEEE Press, Wiley Interscience, 2006.

185

[3] N. Engheta, “Compact cavity resonators using metamaterials with negative permittivity and permeability”, Proceedings on Electromagnetics in Advance Applications (ICEAAOI), Torino, Italy, 2001. [4] N. Engheta, “Is Foster’s reactance theorem satisfied in double-negative and single-negative media?” Microwave and Optical Technology Letters, vol. 39, No. 1, October, 2003. [5] B.A. Munk, Frequency Selective Surfaces, Theory and Design, Wiley, 2000. [6] J. B. Pendry, “Negative refraction makes a perfect lens,” Phys. Rev. Letters, vol. 85, no. 18, pp. 3966-3969, October 2000. [7] R.A. Shelby, D.R. Smith and S. Schultz, “Experimental verification of a negative refractive index of refraction,” Science, vol. 292, pp. 77-79, April 2001. [8] D.R. Smith, W.J. Padilla, D.C. Vier, S.C. Nemat-Nasser and S. Schultz, “Composite medium with simultaneously negative permeability and permittivity,” Phys. Rev. Lett., vol. 84, no. 18, May 2000. [9] D.R. Smith, D.C. Vier, N, Ktoll and S. Schultz, “Direct calculation of permeability and permittivity for a left-handed metamaterials,” Appl. Phys. Lett., vol. 77, no. 14, October 2000. [lo] C.G. Parazzoli, R.B. Gregor, K. Li, B.E.C. Koltenbah and M. Tanielian, “Experimental verification and simulation of negative index of refraction using Snell’s Law,” Phys, Rev. Letters, vol. 90, 2003. [ 111B.A. Munk, D.S. Janning, J.B. Pryor and R.J. Marhefka, “Scattering from surface waves on finite FSS,” IEEE Trans. Ant. and Prop., vol. 49, December 2001. [12]D.S. Janning and B.A. Munk, “Effect of surface waves on the current of truncated periodic arrays,” IEEE Trans. Ant. and Prop., vol. 40, September 2002. [13]B.A. Munk, “A new interpretation of negative p1 and ~1 produced by a finite periodic structure”, Proc. ICEAA 05, p.727-732, Torino, Italy, September, 2005. [ 141B.A. Munk, Finite Antenna Arrays and FSS, Wiley, 2003 [15]B.A. Munk, G.A. Burrell and T.W. Kornbau, “A General Theory of Periodic Surfaces in Stratified Media,” Tec. Report 784346-1, Ohio State University ElectroScience Laboratory, Nov. 1977. [16]L.W. Henderson, The scattering of planar arrays of arbitrary shaped slot and/or wire elements in a stratz5ed dielectric medium, Ph.D. Dissertation, Ohio State University, 1983. [17] L.W. Henderson, “Introduction to PMM,” Tech. Report 715582-5, Ohio State University ElectroScience Laboratory, February 1986. [ 181G.V. Eleftheriades and K.G. Balmain, Negative-Refraction Metamaterials, Wiley Interscience, 2005. [ 191C. Caloz and T. Itoh, Electromagnetic Metamaterials, Transmission Line Theory and Microwave Applications, Wiley Interscience, 2006. [20]S. Ramo, J.R. Whinnery and T. Van Duzer, Fields and Waves in Communication Electronics, Third Edition, John Wiley and Sons, 1994. [21]J. A. Kong, Electromagnetic Wave Theory, Second Edition, EMS Pub., 2000. [22]P.M. Valanju, R. M. Walser and A.P. Valanju, “Wave refraction in negative-index media: Always positive and very inhomogeneous,” Phys. Rev. Letters, vol. 88, no. 18, May 2002. [23]W. Rotman, “Plasma simulation by artificial dielectric and parallel-plate media,” IRE Trans. on Ant. &Prop., January 1962.

On Negative Refractive Metamaterials: Characterization, Bianisotropy, and Applications J. A. Kong, H. S. Chen, X. X. Cheng, J. J. Zhang, D. W. Wang, B.-I. Wu

The Electromagnetics Academy at Zhejiang University, Zhejiang University, Hangzhou, China. Research Laboratory of Electronics, Massachusetts Institute of Technoloa, Cambridge, MA, USA.

1 Introduction Metamaterials that possess negative refraction (NR)properties could be artificially fabricated in the forms of split-ring resonators (SRRs) [l], periodical photonic crystals (PCs) [2], or transmission line (TL) structures [3]. However, the principles of their negative refraction behaviors are quite different. For PCs, the NR behavior is caused by periodicity; for TL structures, it is caused by the high-pass LC network configuration; while for the left-handed metamaterials (LHMs) realized with SRRs and rods, or some modified ring resonators, it is due to the resonance behaviors of the unit cells. In these structures, the SRRs act as magnetic resonators that exhibit negative permeability property [4]while the rods act as electric plasma like media that exhibit negative permittivity property [5]. Since these structures resonate in a wavelength range much larger than their structure size, they exhibit bulk electromagnetic properties, and can be characterized by constitutive parameter tensors. In order to characterize the bulk media properties of the left-handed metamaterials, various retrieval approaches have been proposed [6,7]. The retrieval algorithm is based on the reflection and transmission coefficients of an electromagnetic wave incident onto the slab of the metamaterial, and then the effective permittivity and permeability of the metamaterial structure can be inversely calculated. We shall extend the retrieval approach and apply the method to the experiments, from which we can reveal some intrinsic properties of the LHMs realized with resonant structures [S], and show the difference from the other two kinds of configurations, i.e. the PCs, and the TLs. We also address the fabrication and design issues of artificially manufactured metamaterials and their characterization from the viewpoint of constitutive relations in the electromagnetic theory. Characterization of metamaterial The retrieval algorithm has been well established for the extraction of the effective permittivity and permeability of the metamaterials [6,7]. Based on the retrieval algorithm, we can show experimentally the intrinsic difference between the left-handed metamaterial and PCs. Since the left-handed metamaterial does not require the structure unit cells to be periodic, we can experimentally demonstrate the left-handed properties of the metamaterial with aperiodic structures. With a slab sample of the aperiodic metamaterial [S], it is shown that the structure has negative refractive index in certain frequency range. The retrieval experiment is carried out in a parallel plate waveguide, with the setup similar to that in [9]. The metamaterial sample is set in the center region of the waveguide with microwave absorber surrounded it. After calibration, we can get the transmission and reflection coefficients of a quasi plane wave incident onto the slab sample. Then the impedance and the refractive index of the metamaterial can be inversely calculated by [7] 2

z = h {[(l+sll)2-s*12] / [(l -sl,)2-s212]}x e i n ~ s d = ~ &i(l-X2jA (2) where X=(l- S1l2+ S212) / 2S21 . Hence, we can get the permittivity and permeability of the metamaterial from:

p=nz; E = ~ Z (3) We find that in the frequency range from 9.35 to 10.2 GHz, both the permittivity and permeability are simultaneously negative, revealing that the negative refraction behavior of the left-handed

186

187

metamaterial is not caused by the periodicity, but by the subwavelength resonance behavior of the material, which is different from that of PCs. The demonstration also shows that the retrieval approach is very powerful in the extraction of the bulk material parameters. The only disadvantage is that it requires a plane wave normally incident onto the material. The thickness of the slab sample has to be very uniform along its transverse cross section because the undesirable reflection and diffraction from the incident wave over a big sample of the material will cause the result less accurate. Therefore, we extended the current approach into the rectangular waveguide measurement, where the matching of the test specimen is not so crucial because only slab-shaped samples with small cross sections are required. The only difference is that inside the waveguide, the incident wave is obliquely incident onto the slab, and there need several independent set of measurements with different orientations of the metamaterials to retrieve all the constitutive parameter tensors of the anisotropic metamaterial. In the waveguide-based retrieval approach, we assume the parameter tensors have the following forms in the principal system (el, e2, e3): . ~ - - ~ o d i u~2g [ ~s ~3 1 p=podiag[p , I p2 p31. Compared with the normal incidence of the plane wave, the refractive index and impedance of the material in this case are different, which are not only functions of the material parameters but also functions of the transverse wave numbers. We use the SRR structure [lo] and focus on three parameters: pl,p2 and ~ 3 .In order to retrieve these three parameters, two independent measurements are necessary. In the first measurement (case a), the axes el, e2, e3 of the slab sample are along the direction of i , -2, and 9, respectively. In the second measurement (case b), the axes el, e2, e3 of the slab sample are along with the direction of i ,.?, and 9, respectively. In the two measured cases, n and z are defined by:

n,

= (k:&3pI-

z, = p

k2pllp 2)’

I ( k:

-

k?>”

~ k:( k:)” I (b2&3,uk:pllp

n b = (k:&@-

2)‘

k:p 2/p1)’ 1 ( k: - k:f’

(4) (5)

(6)

z b = p 2( k: - k:)” I (k:&3p 2- k:p 2Ip1)” (7) where k , = z / a is the transverse wave number in the rectangular waveguide. The subscripts ‘a’ and ‘b’ denote that the results are calculated from the measurements of case (a) and case (b), respectively. Therefore, we can inversely calculate the parameters of p1, p 2 and ~3 from Eqs. (4-7).

3 Metamaterial with bianisotropy The retrieval algorithm shows that the constitutive parameter tensors of the metamaterial can be successfully obtained, and from the results we see that the metamaterial can be characterized by a negative permittivity and a negative permeability in some directions. In fact, when the metamaterial shows bianisotropy, it needs more parameter components to characterize them. Here we proposed a possible realization of a biaxial left-handed material with a chirality term located in the diagonal position. The realization is based on a modified S-ring resonator as shown in Fig. 1(a), where the top and bottom arms of the S-ring resonators on both sides of the dielectric substrate are shorted with some vias, so both the effective capacitances between the top and bottom horizontal strips disappear, leaving only the effective capacitance in the middle. Under magnetic induction, the induced current will flow around the top and bottom half loops, and from the equivalent circuit analysis [Fig. 1 (b)] we can see that a net electric dipole moment is induced in the z direction, which we refer it as a bianisotropy term. Therefore, the total electric and magnetic behaviors of the particles are given by:

where a:, a;, azzm , a,:

are functions of the particle dimensions.

188

b

I.(

ibj

Fig. 1. (a) The structure of circular bianisotropic S-ring resonator and (b) its equivalent circuit model. In (a) the top and bottom arms are shorted with vias marked with dashed circles so that a bianisotropy effect is introduced. With the introduction of the bianisotropy, cross-polarization effect is induced in this kind of structure. Detailed theoretical analysis shows that the characteristic propagating waves inside of the structure are two circle polarized waves. Therefore, when a plane wave with an Ey polarization is normally incident onto the slab, both the transmission and reflection waves contain the Ez polarized component, in addition to the Ey polarized wave component.

4 Applications Both the retrieval algorithm and the theoretical analysis show that left-handed metamaterials can be characterized by constitutive parameter tensors. With these macroscopic parameters, metamaterials are more feasible to be implemented in some novel applications, such as subwavelength resonators [ 121, backward couplers [ 131, antennas, etc. The couplers utilizing left-handed metamaterial have shown successfully the backward coupling effect, in these couplers, the two branches are filled with a left-handed material and a right-handed material respectively. Thus the bandwidth of the coupler is dependent on the left-handed bandwidth of the metamaterial. In fact, we show theoretically that by setting the two coupler branches with two kinds of left-handed metamaterial, the bandwidth of the coupler is greatly enhanced [14]. The schematic of the LHM-LHM coupling in the waveguide coupler is shown in Fig.2. The two left-handed metamaterials have a relative permittivities ~1 and ~ 2 respectively, , and their relative permeabilities are biaxially anisotropic with p ~ =opd i a g b l X p l y p lZ1 and p 2=p o diag[pzX p 2y p 2 4 . By expressing the electric and magnetic field in these two regions and applying the boundary condition at the interface of the metamaterial, we can get the guidance conditions of the coupler and the time-averaged Poynting power densities in region 1 and region 2 given by:

189

(a)

(b)

Fig. 2. (a) Schematic of the coupling region. The two regions are filled with two kinds of biaxially anisotropic metamaterials. (b) Experimental realization of the wide backward waveguide coupler. Therefore, when it is in the stop band of LHM in region 1 and the passband of LHM in region 2 at the same time, namely, p,, > 0,p2,,< 0 , and vice versa, we could conclude that the Poynting vectors in the two regions must be contra-directional in the z direction, which indicates a backward coupling. Therefore, the bandwidth of the backward coupling is a union of the two left-handed pass bands. In the experiment, we use the S-ring resonator and the SZ-ring resonator as different LHM samples. The result confirms the theoretical analysis and the interesting phenomenon provides a way of realizing wideband coupling with flexibility and could be applied in the design of many other novel microwave devices.

Conclusion The paper investigated various aspects of left-handed metamaterial, including the characterization of metamaterial, the ring design, and the applications. We have shown a waveguide based retrieval algorithm which can extract the constitutive parameter tensors efficiently. The left-handed metamaterial with bianisotropy is proposed and shows the structure can be characterized by a more complex parameter tensors. Furthermore, with the characterization method, we can design and optimize the microwave components in a more flexible way. Applications in areas of antennas and radiation are also explored. 5

Acknowledgement This work was supported by the Chinese Natural Science Foundation under Grant No. 60531020, by the China Postdoctoral Science Foundation under Grant No. 2006039033 1, in part by the Office of Naval Research under Contract NO00 14-01- 1-0713. References [l] R. Shelby, D. Smith, and S. Schultz, “Experimental verification of a negative index of refraction,” Science, 292, 77-79, 2001. [2] P. V. Parimi, W. T. Lu, P. Vodo, S. Sridhar. “Photonic crystals: Imaging by flat lens using negative refraction. ” Nature .426,2003. [3] G. V. Eleftheriades, A. K. Iyer, and P. C. Kremer, “Planar negative refractive index media using periodically L-C loaded transmission lines, ” IEEE Trans. Microwave Theory Tech. 50(12), 2702-2712,2002. [4] J. Pendry, A. Holden, D. Robbins, and W. J. Stewart, “Magnetism from conductors and enhanced nonlinear phenomena,” IEEE Trans. Microwave Theory Tech. 47,2075-2084, 1999. [5] J. Pendry, A. Holden, W. Stewart, and I. Youngs, “Extremely low frequency plasmons in metallic mesostructures,” Phys. Rev. Lett. 76,4773, 1996.

190

[6] D. R. Smith, S. Shultz, P. Markos, C. M. Soukoulis, “Determination of effective permittivity and permeability of metamaterials from reflection and transmission coefficients,” Phys. Rev. B. 65, 195104, 2002. [7] X. Chen, T. M. Grzegorczyk, B.-I. Wu, J. P. Jr., and J. A. Kong, “Robust method to retrieve the constitutive effective parameters of metamaterials,” Phys. Rev. E 70, 016608, 2004. [8] H. Chen, L. Ran, D. Wang, J. Huangfu, Q. Jang, and J. A. Kong, “Metamaterial with randomized patterns for negative refraction of electromagnetic waves,” Appl. Phys. Lett. 88, 031908, 2006. [9] D. Wang, J. Huangfu, L. Ran, H. Chen, T. M. Grzegorczyk, and J. A. Kong, “Measurement of negative permittivity and permeability from experimental transmission and reflection with effects of cell misalignment,” J. Appl. Phys. 99, 123114, 2006. [lo] H. Chen, J. Zhang, Y. Bai, Y. Luo, L. Ran, Q. Jiang, and J. A. Kong, “Experimental retrieval of the effective parameters of metamaterials based on a waveguide method,” Optics Express 14, 12945, 2006. [ l l ] R. Marques, F. Medina, and R. Rafii-El-Idrissi, “Role of bianisotropy in negative permeability and left-handed metamaterials,” Phys. Rev. B. 65, 144440, 2002. [I21 D. Wang, L. Ran, B.-I. Wu, H. Chen, J. Huangfu, T. M. Grzegorczyk, and J. A. Kong, “Multi-frequency resonator based on dual-band S-shaped left-handed metamaterial,” Optics Express 14, 12288,2006. [13] Y. Yuan, L. Ran, H. Chen, J. Huangfu, T. M. Grzegorczyk, and J. A. Kong, “Backward coupling waveguide coupler using left-handed material,” Appl. Phys. Lett. 88, 21 1903,2006. [14] J. Zhang, H. Chen, Y. Luo, Y. Yuan, L. Shen, L. Ran, and J. A. Kong, “Wideband backward coupling based on anisotropic left-handed metamaterial,” Appl. Phys. Lett. 90,043506,2007.

Ultra-thin radar absorbing structures based on short strip pairs

X. S. Rao, S. Matitsine, and H. Lim Temasek Laboratories, National University of Singapore, Singapore I I7508 Abstract: In this paper, we propose ultra-thin radar absorbing structures based on short strip pairs. The thickness of the structure is about M O O wavelength or less at the working frequency. The absorption is attributed to the electromagnetic resonance of the short strip pairs which work in a similar way as the split-ring resonators. Theoretical work has shown that effective magnetic responses can be generated using nonmagnetic short wire pairs or short strip pairs (SSPs) [l-21. Further investigations [3-61 have demonstrated that the SSPs respond to electromagnetic waves in a similar way as split-ring resonator (SRR) proposed by Pendry et al. [7]. Some groups have combined SSPs in their design of metamaterials and successllly realized both negative permittivity and negative permeability at frequencies from microwave to visible optical frequency [4,6]. Even though the size of the SSP is considerably larger than the split ring resonator at the electromagnetic resonant frequencies, which is not desirable to form a true metamaterial where a large ratio of wavelength to lattice constant is needed, the SSPs do have some advantages over the split ring resonators. For example, the relatively simple design of SSPs can be easily fabricated and experimentally characterized. This is even more crucial as the frequency is pushed up to the infrared and optical regions. In this paper, we apply the SSPs to the design of ultra-thin radar absorbing structures not realizable by conventional methods [8]. Even though the applications are different, the underlying physics should be similar - the electromagnetic responses of the SSPs plays a key role in realizing ultra-thin radar absorbing structures. The structure of OUT design is a two-dimensional periodic array with a square lattice. The unit cell of the structure is shown in the inset of Figure 1. The strip lies above a dielectric substrate. The other side of the substrate is backed by a uniform metallic surface as required in real applications. There is only one strip in the structure. However, since the strip is placed near to a metallic ground plane, an image is formed. The strip and its image thus form a SSP. When the EM wave illuminates on the structure with H-field perpendicular to the long edge of the strip, the SSP supports an electromagnetic resonance at a frequency which is determined mainly by the length of the strip. At resonance, EM energies are strongly dissipated by the structure.

Figure 1. Simulated S11 vs. frequency for the absorbing structures based on the SSPs. The design parameters are listed in the text. Inset: the unit cell of the absorbing structure based on the SSPs.

191

192

Using full-wave electromagnetic field solver HFSS [9], the proposed absorbing structure is numerically studied. Shown in Figure 1 are the simulation results for several absorbing structures based on SSPs with different strip width. The other parameters are kept constant: dielectric permittivity ~=6(1-j0.05),magnetic permeability p=l, thickness of the substrate t=0.2mm, unit cell size axb=5.4mmx5.4mm, strip length 1=5mm. The metallic strip is made of copper with a thickness of 20pm. All structures show absorption dips around 11513GHz. It should be stressed that the free space wavelengths at these frequencies are about 23-26mm) while the thickness of the absorbing structures is only 0.2mm, less than M O O of the wavelength, hence, “ultra-thin”. It is clear that when other design parameters are constant, a wider strip is preferred over the narrower strip for better absorption. In the simulated case, the widest strip with w=5mm (it is actually a square patch) has the strongest absorption of about -18dB, while the narrowest strip with w=lmm achieves only about -5dE3 absorption. We also perform simulations to investigate the effect of loss tangent of the substrate on the absorption. Since the square patch works better than the strips for better absorption, and clearly the square patch is less polarization-dependent, we focused our simulations on the square patch with I=w=5mm. Shown in Figure 2 are the simulation results. The only variable in the simulation is the dielectric loss tangent of the substrate, while the other parameters remain unchanged. We found that the reflection coefficient has a non-monotonic dependence on the loss tangent of the substrate. When the loss tangent increases from a low value (0.01) to a high value (0.07), the absorption increases first and then decreases above a certain value (around 0.03-0.05). However, even for the highest value (0.07) used in the simulation, the substrate is still a low-loss substrate. This is very different from the conventional absorbing structures where normally requires a high-loss substrate. 0 -2

-m .Z

-4 -6 -8

E

;.

-10

4 -12 0

cs -14 -16 -1 8 -20 10

10.5

11

11.5

12

12.5

13

Freq (GHz)

Figure 2. Simulated S11 vs. frequency for the absorbing structures based on the square patch on the substrates with different dielectric loss tangent.

The non-monotonic dependence on the loss tangent of the substrate can be well understood. As we mentioned previously, the absorption is due to the electromagnetic resonance of the SSP which operates like the split ring resonator. If the loss of the substrate is sufficiently large to damp the resonance, the resonance may not occur at all. This is very similar to the over-damped oscillator in mechanics. To verify that the ultra-thin structures based on SSPs can be used for EM wave absorption, we fabricated several prototypes based on the SSPs and measured their performances. Figure 3 shows the experimental results for a strip sample measured with EM waves with different polarizations. The parameters are: ~=3(1-j0.01),p=l, t=0.15mm, axb=l3mmx13mm, Z=w=I0.5mm, w=2mm.

193

Figure 3. Measured S11 vs. frequency for the strip fabricated on a dielectric sheet with a thickness of 0.15mm. Normal incidence with (a) H-field perpendicular to the strip and (b) Hfield parallel to the strip.

Figure 4. Measured S11 vs. frequency for the patch pattern fabricated on the same dielectric sheet as shown in Figure 3 with thickness of (a) 0.15mm and (b) 0.3mm, respectively. Figure 3(a) shows the result for a normally incident plane wave with the H-field perpendicular to the long edge of the strip, while in 3(b), the H-field is parallel to the long edge of the strip. As expected, the results are strongly polarization-dependent. An obvious absorption null, around 3.5dI3, is observed near 9GHz for the first case while no absorption is observed in the second case. Figure 4 shows the measured results for two prototypes of square patches with Z=w=I0..5mm, and substrate of thickness of 0.15mm and 0.3mm. Other parameters are the same as in Figure 3. As compared to Figure 3(a), the absorption null in Figure 4(a) is stronger, about -6dB, which is consistent with our simulation results in Figure 1. The absorption null in Figure 4(b) is even more prominent, about -20dB. This result is very promising for real applications. In conclusion, we have proposed ultra-thin absorbers based on the electromagnetic resonance of the SSPs. The advantage in thickness and weight of this design are very prominent as compared to conventional absorbers. However, we admit that since it is based on resonance effect, the absorber still suffers some disadvantages for practical applications. The biggest problem is that it operates only within a narrow bandwidth due to the resonant nature. Further study shows that the bandwidth can be significantly increased if we can replace the dielectric substrate with certain frequency-dependent materials.

194

Reference [ 11 A. N. Lagarkov and A. K. Sarychev, Phys. Rev. B, 53,63 18 (1996). [2] V. A. Podolsky, A. K. Sarychev, and V. M. Shalaev, J.Nonlinear Opt. Phys. Mat., 11,65 (2002). [3] G. Dolling, C. Enkrich, M. Wegener, J. F. Zhou, C. M. Soukoulis, and S. Linden, Opt. Lett,. 30, 3 198 (2005). [4] V. M. Shalaev, W. Cai, U. K. Chettiar, H. K. Yuan, A. K. Sarychev, V. P. Drachev, and A. V. Kildishev, Opt. Lett., 30, 3356 (2005). [5] A. N. Grigorenko, A. K. Geim, H. F. Gleeson, Y. Zhang, A. A. Firsov, I. Y. Khrushchev, and J. Petrovic, Nature, 438, 335 (2005). [6] J. F. Zhou, L. Zhang, G. Tuttle, T. Koschny, and C. M. Soukoulis, Phys. Rev. B, 73,041 101 (R) (2006). [7] J. R. Pendry, A. J. Holden, D. J. Robbins, and W. J. Stewart, IEEE Trans. Microwave Theory Tech., 47, 2075 (1999). [8] E. F. Knott, J. F. Shaeffer, and M. T. Tuley, Radar Cross Section, 2nd ed. (Artech House, Boston, 1993) chapter 8. [9] HFSS, Ansoft Corporation.

Electromagnetic Characterisation of Conductive Helixes DENG Chaoran*, LIU Lie' and ZHANG Yongjian*

* DSO National Laboratories, # Temasek Laboratories, NUS Abstract A measurement method was proposed to characterise conductive helixes individually. The measured results agreed well with the simulated using finite element method (FEM). Geometrical and conductivity effects of helixes on scattering were investigated. It was found that the resonant phenomenon of helixes depends on the geometrical parameters such as diameter and total extended length, and conductivity of materials made of the helixes. The resonance disappears when impedance was greater than 3 /sq if the helixes are made of resistive tapes. Keywords: TEM cell, microwave measurement, helixes, scattering parameters

1. Introduction Metallic helixes have been widely used and investigated as antennas, meta-materials and electromagnetic absorbing material^'"^), since they respond to electromagnetic field strongly despite their small physical dimensions. As meta-materials and absorbing materials, they are in the form of a periodic array or randomly distributed inclusions. They are normally not characterised as a single element in measurements, because the scattering of a single element at its first resonance is too weak to be measured in free space. The half wavelength of the first resonant frequency is about the extended length of the helix. In this paper, a transmission line technique is attempted to measure the scattering of a pair of helixes in VHF and UHF. The measured results are compared with numerical calculations by FEM using HFSS from ANSOFT. 2. Fabrication of Helixes Helixes were made of thin conductive tapes wound on FR4 tubes of 5cm in diameter. The length and wall thickness of the FR4 tubes are about 18cm and 0.3mm, respectively. They are almost transparent in VHF and UHF bands due to the thin wall thickness and the low dielectric permittivity and loss. As for metallic tapes, a 70pm thick copper foil with pressure-sensitive glue on one side was employed. The tape width was varied from 8mm to 15mm. It took about 6 and half turns to make a helix with an extended length of lOOcm on the 5cm tubes. The resonant frequency is about 160MHz in free space. Similarly one resistive helix was made of a resistive tape with surface resistivity of 30 n/sq This resistive tape was made by a screen printing technology and the printing ink is carbon-based from commercial markets. The surface resistivity value was verified in DC and microwave frequency. Another resistive helix was made of nickel coated graphite fiber tissues, which impedance was about 3 Wsq.

3. Measurement Technique TEM cell 4, was used to characterise the scattering phenomena of the helixes. It is known to be used to study EM1 and EMC. A schematic drawing of the setup is shown in Fig. 1. It is made of cupper plates. The TEM cell is a kind of a coaxial airline, however its field distribution is not as homogeneous as the cylindrical coaxial airline. There is no lower cut-off frequency like a coaxial airline, but the upper cut-off frequency is about lGHz for the present design. The cross section of the measurement cavity is 180x120 mm, however is divided into two partitions at the centre. The dimension for each partition is 180x60 mm which limits the size of helixes to be measured. The upper wall of the TEM cell can be fully opened up for placing samples to be measured. The helixes stand up in the TEM cell and E-field is perpendicular to the axis of the helixes. Calibration of the measurement was up to the N-type connectors to TEM cell. The dynamic range for the system is about 20-25 dB below 0.2GHz and about 15dF3 below 0.7GHz. It was improved to be better than

195

196

20dB below 0.7GHz recently. The OdE3 reference of S11 and S21 was the empty TEM cell without any samples inside.

4. Simulation The FEM code used was HFSS v10. Only one partition of TEM cell was simulated since the TEM cell was symmetrical as shown in Fig. Id. The helixes were simulated as impedance surfaces without any volumes in order to reduce CPU time. The surface conductivity was varied from PEC to lossy resistive surfaces up to 100 Rlsq. PEC boundaries were employed for the faces (side walls of TEM cell) to be parallel to the axis of the helixes and periodic boundaries for the faces (top and bottom walls) to be perpendicular to the axis because this arrangement avoids the lower cut-off frequency of a waveguide. As a comparison, scattering of the helixes in free space was predicted when the boundaries were set to be radiation or perfect absorbing boundaries without reflection. The simulation accuracy was set to be comparable to that of the measurement. 5. Results a) Comparison of Measurement with Prediction Firstly, the results predicted and measured for the copper tape helixes are shown in Fig. 2. As expected, the first resonant peak is predicted at 0.15GHz in free space due to the total length of 100cm. On the contrary the first resonant peak is shifted to a lower frequency at 0.12GHz as predicted and measured in TEM cell due to the coupling of the helixes with the TEM cell wall. The coupling also exists between the helixes in arrays or composites. Although, the S11 peak height predicted in free space is lower than that in the TEM cell, the predicted peaks correspond well to the measured with an expected frequency difference. The S21 measured results agree well with the predicted ones. The minor difference in the S21 results may be caused by difference in the points of the measurement and prediction. The frequency points are 1601 measured from 0.05 to 1.05 GHz, however, the prediction interval was O.01GHz and one order courser than the measurement. From the measured results, the peak width is only about 0.01GHz. To simulate the same interval as the measured (0.000625GHz), CPU time would be tens of hours for each curve. It is not so efficient as compared with few minutes to measure a pair of helixes. Therefore, the TEM cell measurement technique provides a new way to characterise helixes, which can also consider the coupling effect between the helixes.

b) Effect of Helix Geometries The geometrical parameters for helixes are helix diameter, helix pitch, helix turn, and wire diameter or tape width. To make comparison more relevant, the total extended length of the helixes shall be same. Firstly, the tape width was tapered from 15mm to 3mm in about 40mm at both ends and the width was also reduced from 15mm to 8mm. The measured S21 results are shown in Fig. 3a (left). It is obvious that the reduction in the tape width shifts the first resonant peak to a higher frequency due to change of intrinsic helix properties and reduction in coupling with the wall of TEM cell. The diameter of the helix diameter was changed from 50mm to 40mm. Although a total length of 90cm was maintained, the first resonant peak was still shifted to a higher frequency due to reduction in coupling with the wall of TEM cell as shown in Fig. 3b (right). The increase in the total extended length shifts the peaks to a lower frequency (comparison of black lines in the left and right figures). c) Effect of Conductivity of Helixes Conductivity of the helixes is one of the key parameters dominating the resonant phenomena. Fig.4 plots the predicted results which show how the S-parameters change with frequency and conductivity. The peak height decreases with the increasing surface impedance. The peaks are disappeared when the impedance is greater than 3 Rlsq. The resistive helixes are lossy in broad frequency ranges as compared with the metallic, but the difference in low frequency ranges near the first peak is negligible. For highly conductive helixes, the transmission loss (S2 1) and reflection

197

(S1 1) are small except at the resonant peaks. For highly resistive helixes, the S21 is reasonable, but the S 11 can be small for instance when the impedance is 100 R/sq. To confirm the prediction, lossy helixes were fabricated and measured. The results are shown in Fig. 5. The impedance of 3 and 27.5 Rlsq was employed due to availability of the resistive materials. As expected, no any resonant peaks were observed when the impedance is 27.5 R/sq but a small peak was found around 0.2GHz when the impedance is 3 Wsq. The S11 and S21 measured agree with the prediction reasonably well. There are few reasons causing some differences between the measurement and prediction. For example, only the real part of the impedance was considered in the prediction, but the real materials are complex. Also EM field distribution in experimental setup is slightly different from that in the numerical model. 6. Remarks A TEM cell measurement techniques was proposed to measure helixes and the measured results agreed with the prediction reasonably well. This technique fills the gap between coaxial airline for homogeneous materials of small size, and free space for composites and arrays of a few wavelengths. It will be useful to characterise inhomogeneous and anisotropic materials such as meta-materials.

100 mm

500 mm

1oomm

b) Top View

Inner Conductor

g

3 00 0

-

l%i?l

00

a) Cross section Connector

c) Front

d)Boundary Condition

View

Fig. 1: TEM cell measurement set-up I

0

Comparison: Prediction and Measurement (Copper Helix)

il

.I 0

I

-9

'

,

-S11.M Measured x s11.p Predicted in TEM cell -d- SI~.P-A,~ Predicted in Air

40

I 0.0 0.2 0.4 0.6 L- Frequency, GHz ~

~

~~

-12

1S21-M

1.0 I ~~

I

,

-15

I

0.8

00

0.2

sz1-p

11

I

Measured i$ TEMcell Predicted irj TEM cell

0.4 0.6 Frequency, GHz

~~

Fig. 2: Comparison of measured and predicted results.

0.8

10

~___

198

0

4

ln

12

I

-15mm-ref

I

11

-16 0.10

0.11

I

0.12

0.13 0.14 Freouencv. GHz

0.16'

0.15

0.10

0.12

I

0.14 0.16 Frequency, GHr

0.18

0.20

0.8

1

Fig. 3: Effects of geometries on helixes' scattering. Effect of Conductivity

0

0

1

-5

~

-10

-2

m r

.--3

-15

ln

ln

-20

4

-25

-5

-C"

-R-100

I

-8 0

0.2

0.4 0.6 Frequency, GHr

0.8

0

0.2

0.4 0.6 Frequency, GHz

~~

Fig. 4: Effect of conductivity of materials made of helixes.

,

0

I

0 1

-5

2

m-lo P r

;;a

ln45

ln

4

-20

11

5

-25

4 0.0

02

04

06

l.ol

0.8

Frequency, GHz

0.0

0.2

0.4

0.6

0.8

I.o

Frequency, GHr ~

Fig. 5: Measured and predicted results of lossy helixes.

Reference: [l] John D. Kraus and Ronald J. Marhefia, Antennas, 3rdEdition, Mc Graw Hill, 2002 [2] A. N. Lagarkov, V. N. Kisel and V. A. Chistyaev, J. Mug. Mug. Muter., 258-259, 161 (2003). [3] C. F. Bohren, R. Luebbers, H. S. Langdon, and F. Hunsberger, Appl. Opt., 31 (30), 6403 (1992). [4] X.-D. Cai and G. I. Costache, IEEE Trun. Electromug. Comp., 35 (3), 398 (1993).

Session P10

Chair: S. Matitsine

This page intentionally left blank

Frequency Selective or Controllable Metafilm as a Part of On-Board Antenna Screen Vladimir Kisel Institute for Theoretical and Applied Electromagnetics, Moscow, Russia Abstract: A problem of plane wave penetration through thin-layer structure (such called “metafilm”) with variable surface properties is considered. The choice of these properties is defined by suggested application of the metafilm as a part of low-observable antenna screen. The problem is solved by the integral equation method. Numerical and experimental results are discussed which illustrate the utilization of frequency-selective, photo- and electronically controlled metafilms to attain lower radar visibility of slot antenna arrays. In past several decades a lot of attention has been attracted to the properties of complex media. Recently the special class of composite structures, so called metamaterials, has been outlined and has gained considerable interest because of wide range of their potential applications, particularly, in electromagnetics. While metamaterials are generally 3D-extended structures, interesting effects (including superresolution) can be observed in thin-layer metamaterial systems. So it looks natural that a new term, a “metafilm”, has been introduced to mean a surface distribution of electrically small scatterers [ l ] . Metafilms may be promise candidates to use as controllable surfaces, which are of great interest in the field of electromagnetic compatibility or antenna design. In particular, a metafilm may constitute a frequency-selective surface (FSS), i.e. perforated conducting foil, a grating with the cells of special shape or electronically or photo-controlled film. Naturally, the concept of metafilm looks to be best fitted to FSS with cell dimensions and lattice constant much less then a wavelength. Note, this is the case of particular interest in the scope of antenna shielding from undesired external radiation, as grating lobes do not appear in the backscattering pattern of the antenna screen made of such a metafilm. A lot of literature has been published over past 20 to 30 years with the FSS theory, design and application (see, for example, [1-51>. Nevertheless, little data are available with regards to controllable FSS taking account of the features of realistic controllable films and devices and restrictions caused by peculiarities of external conditions. There is a lack of practical recommendations concerning the choice of controllable devices and their properties, the influence of their parasitic capacitance, inductance and losses on the FSS performance and so on.

A suitable structure for modeling a controllable FSS is shown in Fig. 1. This is an infinite plane with 2D-periodical distribution of the complex surface resistance R , Ohms/o, which may be varied over the certain portions of surface thus modeling the operation mode switching, Fig. l a . Within a lattice period the distribution of R(x,y) is defined as piecewise-constant function on a grid with small square cells, Fig. l b . A plane electromagnetic wave is incident from the upper hemisphere. Reflection and transmission coefficients have to be determined. Solution of the problem. Note, that in literature (for example, see [ 5 ] ) FSS excitation problems are solved generally by using surface equivalence principle. Integral equations are constructed with respect to the equivalent magnetic current on slots in a perfectly conducting plane. Thus a simpler and numerically efficient computational model may be created at the price of lacking generality of the problem setting. Particularly, it becomes impossible to take into account material losses of the perforated plane, and this makes difficult to consider variety of structures, particularly those based on photosensitive films. Here we use integral equations set with respect to the components of the surface density of electrical current J , , J,, over the whole structure consisting of supporting perforated surface and a control units (CUs). Resistive boundary conditions [6] (see Fig. l c ) were applied to solve the problem. In particular case of zero surface resistance they correctly describe a perfectly conducting

201

202

surface, so it is possible to treat even metallic FSS with various CUs within the scope of the same technique.

One can consider both thin-film or discrete CUs, in the latter case CU is specified as a square cell with equivalent surface resistance equal to the impedance of CU at given frequency. Another useful feature of the approach is the possibility to take into account the reactive properties of CU (for example, intrinsic capacitance or inductance of photosensitive device or switching diode, presence of dielectric substrate etc.). Corresponding properties of CU are specified by imaginary part of its complex surface resistance. Let's make use of the boundary condition and the relationship between surface current components on the adjacent cells Jlx+l,ly+l = J I J y

exP(- ikxTx)exp(- ikJyL

where k, = -k sin 0, cos qo, k,, = -k sin 0, sin qo, k is the free space wavenumber, T, , Ty are the lattice constants along x and y , I, Zy are cell numbers along x and y directions correspondingly. Then the following integral equation system can be obtained:

203

where i-x

Gs =

exp(- il,k,T,)exp(-

i,, i y = 4

exp(- ikR,) ilykyTy) 4ZRi

,

= Rs/Wo is the specific surface resistance normalized by W, = 1207r Ohms; integral equations are written for a single (zero numbered) FSS cell So.

These integral equations were solved numerically, a system of linear algebraic equations (SLAE) was created by application of moment method with rooftop expansion and weighting functions. To sum the series in SLAE entries the acceleration technique [7] was used. The scattered field and corresponding reflection and transmission coefficients were evaluated by means of the theory of vector potentials using surface currents calculated as a result of SLAE solution. Potential area of application and selected results.

On-board radar antenna is known to be the main source of the secondary (scattered) electromagnetic field of an aircraft in front hemisphere [S]. Antenna bay contributes into the vehicle radar cross section (RCS) due to incident wave scattering from gearing and electronic equipment units installed in the bay and due to scattering from radar antenna (Fig. 2a). Antenna bay equipment may be easily hidden through using radar absorbing coatings. It is much more difficult to decrease radar signature of the antenna plate, which is typically a flat slot antenna array, Fig. 2b. RCS contribution of such an antenna is defined by mirror reflection of the incident wave from conducting disk surface (in the directions close to its normal axis) and by diffraction from the periodic system of slots. When considering typical geometrical relationships of such an antenna array, the slot diffraction becomes especially distinct around the incidence/observation directions deflected by 30°..600 from antenna axis (it is well seen in Fig. 6 discussed later). To achieve low-observable properties from forward-looking antennas and to enhance electromagnetic compatibility of other radiating systems the special screens may be applied either being embedded into a radome or installed onto an antenna plate. RCS reduction is achieved due to the special choice of the screen shape, which ensures lower backscattering (at out-of-band frequencies) as compared to a non-shielded antenna (Fig. 2c). Incident

Radar antenna

4

equipment

Fig. 2

204

Both controllable and non-controllable screens may be used to shield an antenna. With a noncontrollable screen installed, the antenna can operate provided the screen is almost transparent within the radar frequency band. RCS reduction is achievable only at out-of-band frequencies. However, these screens are rather simple and cheap, they have low thickness and weight and may be used under strong mechanical and electrical impacts. Controllable screens are much more multi-functional devices, with their use antenna RCS may be decreased even at the radar operational frequencies provided that the radar is not in active mode. An example of such an engineering solution [9] is a photosensitive film deposition over the internal surface of antenna radome, the resistance of the film being dependent from the level of external controlling optical illumination. In an ideal case the film conductivity should be switchable between almost infinity (“shielding” mode) and almost zero values (screen “absence” mode, onboard radar may be activated). In practice, the surface conductivity should be variable within the range of order 30 Ohms through 300 kOhms per unit square. Unfortunately, there are technological difficulties in creating such a material with the required range of conductivity variation. These difficulties may be smoothed by means of embedding certain conducting structures or inclusions into the photosensitive film. As was briefly mentioned above, an alternative engineering solution is a screen manufactured of FSS with cell resonators tuned so as to pass electromagnetic waves in narrow frequency band of radar operation and to reflect waves with other frequencies [Z, lo]. To make the screen opaque within radar operational frequency band one can incorporate electronically or photo-controllable semiconductor structures into the resonators. In fact, this is another way to design a controllable metafilm. Realistic antenna screen has a rather complicated geometry, but all of its principal dimensions and curvature radii are much greater than the wavelength. The surface structure of such a screen is designed so as to exclude the possibility to generate and sustain traveling surface waves. Therefore it may be possible to calculate field penetration through a screen starting from physical optics approximation and using reflection and transmission coefficients evaluated for each small portion of the screen surface. These coefficients can be defined from the solution of the plane wave excitation problem for the infinite flat FSS tangent to the screen surface in the given point. On the next step one can iteratively account for mutual interactions between antenna and screen, if necessary. So it is possible to combine high-frequency and rigorous approaches to achieve reasonable numerical efficiency of the computational model. The computational model briefly discussed above was used to create a software package which may be used to optimize FSS geometry taking into account properties of CUs and to evaluate the performance of screens made of various metafilms. Of course, engineering a controllable screen leads to a trade-off solution. As a rule, when choosing a higher quality-factor Q of the resonant structure (due to the proper shape of a lattice element or increased lattice constant) the requirements to CUs become soften, but angular and frequency properties of the structure do degrade. For example, the calculations showed that in order to achieve a significant attenuation of transmitted energy (about -10 dB at the normal incidence) the specific surface resistance of the controlled thin film inside a linear slot structure should not exceed, say, 1000 Ohms/o. To achieve the same attenuation with lower Q elements (like circular loop) the surface resistance should come to 300 Ohmsh. But at the absence of resonant elements (that is, in the case of continuous thin resistive film) that value of transmission is achieved only at R=80.. .90 Ohms/o (note, within much broader frequency band), Highest (“dark”, for a photosensitive film) value of CU resistance should be at least of order 3.105 Ohms/o to exclude noticeable losses in transmission mode of the screen. Again, the greater Q, the larger “dark” resistance should be chosen. Many of the FSS features may be enhanced due to decreasing the lattice constant. This is especially important for angular dependencies, which become more uniform. To achieve this goal, loaded elements with dense package in a lattice may be recommended. Dielectric filling of the

205

structure results in the same effect. Intrinsic capacity of discrete CUs (say, p-i-n diodes) may also greatly change both angular and frequency properties of the FSS.

T, dB

FSS cell

0

-0.5

-1 ' p - i - n diode

C=0.2 pF 1 -vertical plane 2 horiiontal plane

-

T. dB

0

-1.5 -2 -2.5

-0.1 -0.2

-3 0.85

-0.3 -0.4 -0 5

0

10

20

30

b)

40

50

60

70

V ,8, deg.

\I I

/

0.9

0.95

1

1.1

1.05

1.15

F I F,

1 - C=O pF 2 - G O . 2 pF, Fo=10GHz 3 - C=0.25 pF, Fo=10GHz

C)

Fig. 3 For example, the connection of CU with self-capacity of C=0.2 pF (which is typical for cheap p-i-n diodes), into one of the investigated slot FSS (see Fig. 3) resulted in significant narrowing the bandpass and lowering its central frequency. To attain initial central frequency value all the geometrical dimensions were to be reduced by a factor of more than 3. Note, that despite rather good angular properties of the resulted structure due to the lower value of lattice constant (Fig. 3b), such a controlled metafilm seems to be expensive in production because of the large number of CUs per unit surface. Finally, subsequent increasing the capacity in this structure by 25% (from 0.2 to 0.25 pF) resulted in further lowering the central frequency, from 10 GHz to 9,l GHz (Fig. 3c, curve 1: frequency response of the transmission coefficient, FSS with no CUs; curve 2: the same structure with CU installed, C=0.2 pF; curve 3: the same, G 0 . 2 5 pF).

As well known, in most cases the angular properties of transmission coefficients are far from required uniformity in the specific plane (as a rule, in H-plane). Significant attenuation of transmitted wave is often observed at the incidence angles of 45 degrees or more. That is why we believe that the best way to create an antenna screen is to choose a separate device design (see Fig. 2c) with a special curvilinear surface (like a bi-cone) so as to provide for optimal angles of the surface illumination, in contrast to a popular concept of the combined bandpass FSS radome [lo], Fig. 4. A FSS-based cap-like structure shown below (see Fig. 5) is an example of the experimental design of antenna screen to decrease RCS signature of a plane slot antenna array with the diameter of 0.34 m. The considered results may be useful to get an understanding about possible degree of the antenna RCS reduction and level of the far field pattern distortions caused by the screen influence.

LAntenm

Fig. 4

206 Antenna

Side view

Fig. 5 For RCS evaluation purposes, the operation mode of this screen was switched manually in the process of measurements. Instead of electronic short-circuiting, a narrow conducting strip was connected as a bridge into the gap of each slot. It was experimentally shown that thus shielded antenna has extremely low backscattering level over wide frequency and angular ranges. Insertion losses in the main lobe, as a rule, did not exceed 0.2.. .0.3 dB. Other antenna properties (particularly, side lobe level) did not deteriorate significantly, as well. At least, the antenna pattern distortions owing to the screen were about the values typical for ordinary nose cone radomes. Examples of measured backscattering and far field patterns of the antenna with FSS screen are shown below, in Fig. 6 and Fig. 7 correspondingly (curve 1: unequipped antenna, i.e. no screen installed; curves 2,3: antenna with FSS screen in “opaque”, Fig. 6, or “transparent”, Fig. 7, mode).

Fig. 6

207

10

0 -10 -20

-30 -40

-50 -60

Fig. 7 Electronic circuits with p-i-n diodes to switch the screen between operation modes were also experimentally tested, as well as the performance of photosensitive films and discrete devices. Some of the experimental prototypes are presented in Figs. 8-10. A photosensitive FSS with discrete CUs (modified photoresistors) is shown in Fig. 8. Electronically controlled FSS with switching p-i-n diodes are presented in Figs. 9-10 together with some design tricks needed to separate low-frequency control circuits from high-frequency resonant structures. Calculations and measurements showed that it is possible to achieve a level of insertion losses of order -10.. .-20 dB (“opaque” screen mode) at the initial level of about -0.5 dB (“transparent” screen mode). Operational frequency band and angular performance are determined in many respects by the properties of the specific CUs used.

Photosensitive FSS

Cross section Double-sided mounting

Diele.ctric oil

Fig. 8

Fig. 9

208 Single-sided mounl. mounting

Ei

Fig. 10

-ud

Results of the measurements of controllable FSS samples together with numerical investigations demonstrated that as a whole, electronic p-i-n diode switches look preferable as compared to the photosensitive devices. Of course, strong demands are often placed to the electronic switches, for example, a minimum of intrinsic losses, minimal capacitance (say, 0.05 pF or less), breakdown strength etc. But in many cases the industry produced p-i-n diodes match the requirements satisfactorily.

References C.L. Holloway, M.A. Mohamed, E.F. Kuester, A. Dienstfrey. Reflection and transmission properties of a metafilm: with an application to a controllable surface composed of resonant particles. IEEE Trans. Electromagn. Compat., vol. 47, no. 4, pp. 853-854, Nov., 2005. M. Hook, J.C. Vardaxoglou, K. Ward. Application of frequency selective surfaces. 27th ESA Antenna Technology Workshop on Innovative Periodic Antennas: Electromagnetic bandgap, left-handed materials, fractal and frequency selective surfaces, 9-1 1 March 2004; ESA Publications Division, Noordwijk, The Netherlands, 2004. R. Mittra, C.H. Chan, T. Cwik. Techniques for analyzing frequency selective surfaces - A review. Proc. IEEE, vol. 76, no. 12, pp. 1593-1615, Dec. 1988. B.A. Munk. Frequency selective surfaces: Theory and design, New York: Wiley, 2000. A.F. Peterson, S.L. Ray, R. Mittra. Computational methods irz electromagnetics, New York, NY, IEEE, Inc., 1998. Senior T.B.A. Scattering by resistive strips. Radio Science. 1979. V.14. N25. P.911-924. Lampe R., Klock P., Mayes P. Integral transforms useful for the accelerated summation of periodic, freespace Green's functions. IEEE Trans. 1985. vol. MTT-33, no. 8, pp.734-736. Ruck G.T., Barrick D.E., Stuart W.D., Kirchbaum C.K. Radar cross section handbook, New York: Plenum Press, 1969. Martin M.T., Duhl M.L. Method and apparatus using photoresistive materials as switchable EM1 barriers and shielding. US Patent No.5278562. Jan.ll.1994. ~~. HOlO 17/00. - - --[ 101 E.F. Knott, J.F. Shaeffer, M.T. Tuley. Radar cross section, Artech House, Boston-London, 1993. \

Smart Frequency Selective Surface with Conductive Fiber Array and Diodes L. Liu*, S. Matitsine, P. K. Tan, and Y. B. Gan Temasek Laboratories, National University of Singapore, Singapore Abstract Smart frequency selective surface (FSS) with conductive fiber array and diodes was investigated. Transmission coefficient was measured with free space method and simulated using finite element method (FEM). The resonance frequency of fiber array changes when the diodes are switched on or off. Tunable transmission coefficient is observed between 3 to 8GH2, which can be useful for antenna applications.

1. Introduction FSS has been used in hybrid radomes, band-stop filters, subreflector and circuit analog absorbers for radar cross section reduction (RCSR) and many other civilian and military applications [l]. The main limitation of FSS is that such structures are not tunable or re-deployable. Once they are fabricated, the properties of the screens, such as frequency and bandwidth, etc, cannot be modified to meet changes in operational requirement. It is therefore appealing to have smart or adaptive screens which are tunable to the incident signal. Poly(ani1ine)-silver-polymer electrolyte composite with tunable resistivity or permittivity under voltage biasing was used in microwave smart screens [ 2 ] .However, it is well known that tunable materials are not stable, while stable materials are hardly tunable. Therefore, conductive polymer based tunable composite are not suitable for challenging environment. Active FSS loaded with pin diodes was integrated into a single layer microwave absorber with reflectivity as a function of diode bias current [3]. As compared to materials with tunable properties, this type of material does not require high biasing voltage and large devices. Since the physics of p-n junction is fully understood, it is also easier to compensate for the environmental effect of diodes. For passive regular fiber array, it is possible to achieve stop band response for transmission coefficient (at resonance) [4]. The aim of this paper is to design smart FSS based on such regular fiber array, with tunable transmission response by controlling the on/off state of the microwave diodes. Numerical simulator is used to optimize the design. 2. Numerical simulation The FEM software High Frequency Structure Simulator (HFSSTM)is employed in the numerical study. The unit cell element is modeled using tetrahedral elements, assuming that the structure is infinite in the transverse directions. Typically, smart FSS resides in an unbounded free space. Following the standard practice in all FEM simulations for unbounded problems, the domain of computation is truncated by defining the perfectly matched layer (PML), a fictitious anisotropic layer, on the surface of a box to include only the object under study and its immediate free space region. This region of free space is typically more than a quarter-wavelength from the object, at the frequency of interest. Adaptive meshing technique automatically refined the mesh at locations where the error in the numerical result is large. A convergence condition is defined (which is the difference in the electric field strength between the current and previous iterations to be less than a prescribed value) to obtain sufficiently accurate results. Upon satisfying this condition, the computation process stops.

* Corresponding author, [email protected]

209

210

A unit cell of the smart FSS sheet comprises two strips of conductors and power lines, and a diode, as shown in Figure 1. The diode is modeled using the lumped RLC boundary with the circuit parameters provided by the supplier in Figure 2. Rsis a 6R series resistor and C, is a 0.2pF junction capacitance. R, is the junction adjustable resistor. The resistance of R, can be changed from a few ohms to tens of thousands ohms, depending on the external bias current. A plane wave with electric field E parallel to the conductors

PBC I

Dipoles

P

F3

\ 'in diode

Ri 12mm

Figure 1 : FEM model of smart FSS and wave vector k perpendicular to the layer surface illuminates the model at normal incidence. The PML boundary conditions are imposed on surfaces that are perpendicular to the wave vector. The periodic or linked boundary conditions (PBC) Figure 2: Equivalent circuit model of diode are applied to the surface parallel to the wave vector. The coherent transmission of the composite sheet is obtained from the ratio of the average transmitted electric field intensity to the incident field intensity [4]. 3. Experiments The surface was fabricated with a copper-coated dielectric substrate of thickness 50pm, with a copper thickness of 25 pm. The permittivity of the substrate is 3.5 and loss tangent is about 0.01. The size of the sample is 20cm by 20cm, populated with 16x5 elements. Surface mounting microwave diode working from 1 to 6GHz is used in this study. All diodes are linked in parallel. The sample is driven with a portable variable power source, which has 6 AA size batteries providing the biasing voltage varying from 0 to 2.66 V. When the positive voltage is applied to the diodes, it is switched on with a low R,. When no bias voltage is applied, is the diodes are switched off with a high Rj. The transmission coefficient was measured with the free space method [4]. The measurement setup includes a vector network analyzer, vertically mounted broadband transmitting and receiving electromagnetic horn antennas. The frequency range of interest is 2 to 9 GHz. To eliminate multiple scattering between the sample and the horns, time-domain gating is applied. Diffraction effects at the edges of the sample are minimized by attaching a high-quality ring-shaped wave absorber of inner diameter 15 cm to the transmit horn.

-20 -25

~

-

-30, 2

. , . 3

I

4

.

I

5

.

I

6

,

I

7

,

I

8

,

9

F(GHz)

Fig. 3 Measured and calculated transmission coefficient

21 1

4. Results and discussions Measured and computed transmission coefficients of the smart FSS are shown in Figure 3, with the diodes switched on and off (solid lines and symbols for measured and simulated data, respectively). The resonance frequency is 6.5GHz when diodes are switched off, and 5.5 GHz when diodes are on. The transmission coefficient of the smart FSS can be adjusted between 3 to 8 GHz. The difference in the transmission coefficient for the on and off states of the diodes is defined as the tunability of the smart surface, which varies from 1dB (3GHz) to more than 20dB (6.5GHz). Numerical results are in good agreement with measured data. The difference between the simulated and measured resonance frequencies is about OSGHz. The difference can be attributed mainly to the circuit parameters of the diode, which may not be accurate, and could be frequency dependent over broad bandwidth. For the FSS without pin diodes, better agreement can be found from the calculated and measured resonance frequency. If the capacitance of C, is reduced to 0.15pF, the resonance frequency shifts up for about OSGHz for both on and off state which results in better agreement. Since the IUC values of the equivalent circuit were obtained at low frequency (lMHz), more accurate parameters obtainable by measurement over the frequency band of interest are expected to improve the agreement between simulation and measurement results. 5. Conclusions Smart FSS based on fiber array and diodes was designed, fabricated, measured and simulated. It is observed that the smart FSS has tunable transmission coefficient between 3 to 8 GHz. Good agreement is obtained between simulation and measurement results.

References: [ 11 B. A. Munk, Frequency Selective Surfaces-Theory and Design, John Wiley & Sons, INC, 14-21, 2000. [2] Barry Chambers, Smart Mater. Strut. No. 9,273(2000) [3] A. Tennant and B. Chambers, IEEE Microwave and Wireless Components Letters, V14(1), 46(2004). [4] L. Liu, S. M. Matitsine, Y. B. Gan and K. N. Rozanov, J. of Appl. Phys. 98,063512 (2005)

Design of Broadband Planar Microwave Absorber Anyong Qing Temasek Laboratories, National University of Singapore 5 Sports Dr 2, Singapore I I7508

1. Introduction Microwave absorber is of great scientific and engineering importance. Extensive studies have been carried out on this multi-disciplinary topic including material synthesis, analysis, design, fabrication, measurement, and applications. The design of broadband planar microwave absorber is a fundamental problem in this topic. A planar microwave absorber is a multilayered medium backed by perfect electric conductor (PEC). Various approaches have been proposed, among which the trial and error method and the graphical method are adopted by most early designers. Unfortunately, such approaches turn out to be rather cumbersome as the configuration and material properties become more complicated. Mathematically, the design of a broadband planar microwave absorber can be cast into an optimization problem. In recent years, application of stochastic optimizers to the design of microwave absorber has attracted increasing interest. Previously, genetic algorithms [I][2][3] and particle swarm optimization [4][5] have been applied to the problem. In this paper, we propose to use the differential evolution [6][7] approach. 2. Formulation 2.1 Configuration A general multilayered medium is shown in Fig. 1. Each layer (including host medium and ground medium) is assumed to be homogeneous, isotropic and dispersive (conductivity is assumed to be non-dispersive). Dispersions of permittivity and permeability are treated separately. 2.2 Material Dispersion All natural and artificial materials exhibit frequency dispersion. To design a practical planar microwave absorber, the dispersive behavior of the materials must be taken into account. Material dispersion may obey one of the following models: (1) Non-dispersive-permittivity and/or permeability remain constant over the entire frequency band of interest. (2) Debye-permittivity andor permeability obey Debye’s law of dispersion 16.0 'I layer 1: ZC0200. .742391 mrn laver2: ZCoTilO. 1.663238 I!" I 1 relative -1OdB b a n d 3.2 13.6 layer 1 8 2: ZCoZTil. 2.276060 + ,557363 mm I

-6],,

.

-16 2

-

,

. , .

,

1

6

8

.

, .

. ,

,

10

,

.

11

12

15

frequency (GHz)

Figure 1 General Multilayered Medium

2

Figure 2 A Two-layer Planar Microwave Absorber

objective: maximum -1OdB absolute bandwidth -t, =.742391mm, I,= 1.683238 (4.5 >16) 1,=.7mm.t,= 1.6mm (4.9-s16)

-

1

t,=.7mm,1,=1.7mm ,=.6mm,l,= 1.6mrn ~-t,=.6mm.l,=1.7mm

6

-

objective: maximum -10dB relative bandwidth I =2.833423mm (3.2 - ,161 _ _ t=2.8mm. (3.4 - 13.6) . . . t=2.9mm, (3.2 - 12.9)

(4.6->16) (4.7->16) (4.4-15.4)

8

layer 1: real material 45: ZC0200 layer 2: real material 70: ZCoTilO

I-10

,.$

-

-12

-14

.

-15

-16 2

4

6

8

10

12

14

16

frequency ( G H r )

Figure 3 Effect of Fabrication Tolerance (Absolute Bandwidth)

2

, 1

.

, 6

.

, 8

.

, 10

.

, 12

.

, I4

,

,

is

frequency (GHz)

Figure 4 Effect of Fabrication Tolerance (Relative Bandwidth)

Design of Broad band Microstrip Patch Antennas using Air Gap in Microwave Frequency P.K.S. Pourush’ , Sandhya Mannaand Rajeev Pourush’

’Microwave Lab, Department of Physics Agra College, Agra-282002, India Department of Physics & Computer Science Dayalbagh Educational Institute, Dayalbagh, Agra-282005, India E-mail: rpourush@red$fmail.com

Abstract In recent years, compact and broad band microstrip antennas have received much attention because of their potential applications in many portable communication systems. Several significant advances in improving the inherent narrow operating bandwidth of microstrip antennas have been reported. Microstrip antennas with air substrate are one of the effective and convenient approaches for improving bandwidth. In this communication an attempt has been made to develop broadband microstrip radiator by creating an airgap between ground plane and dielectric substrate. The analysis has been performed for hexagonal patch microstrip antenna in S and X band of microwave frequency range. Variation in bandwidth and resonant frequency with height of airgap (A) has also been estimated and plotted. It has been realized that by employing this technique, it is convenient to design broad band microstrip array configurations. Keywords: Microstrip Broadband Antenna, Air gap.

1. Introduction Recently, the use of microstrip antennas has become increasingly popular because of many unique and attractive characteristics. However, the major weakness of microstrip patch antenna is its inherently narrow bandwidth that restricts its wide band applications. Recently, much progress has been made to broaden the bandwidth of microstrip antenna. There are several techniques to enhance the bandwidth of microstrip antennas including the use of parasitic elements, aperture coupled, impedance matching network and by creation of air gap etc. [l-51. Among these techniques the air gap method has many advantages such as no addition of costly components, applicability to any configuration of patch radiator and its attraction for array applications. Thus in the present paper, this method is employed for improving the bandwidth by introducing an air gap between the substrate and the ground. Also, by creation of air gap the resonant frequency can be tuned just by varying the air gap width A of the antenna. The air gap has the effect of lowering the effective permittivity of the cavity under the patch, resulting in an upward shift in the resonant frequency. The bandwidth will increase, partly due to the increase in the height of the dielectric medium and partly because the effective permittivity is now smaller. 2. Formulation and computation of the problem The analysis has been performed for hexagonal microstrip patch antenna (HPMA) geometry. The hexagonal element is assumed to be a resonant cavity with perfectly conducting wall. As circular disk is a limiting case of a polygon with a large number of sides, the resonant frequency for the dominant as well as for the higher order mode is calculated ref. [6].The computed values of bandwidth of HPMA with air gap at 3 GHz and 10 GHz with respect to A are presented in Fig. 1. The values of resonant frequencies calculated for various air gap width for both the frequencies are shown in Fig.2.

216

217

.-

-10GM

AD.

Fig. 1

Fig. 2

3. Conclusion This paper presents a study of bandwidth enhancement relative to air gap width for hexagonal patch microstrip antenna at two different frequencies i.e., 3 GHz and 10 GHz. The variation of bandwidth with air gap width for this geometry is shown in Fig. 1. The variation of resonant frequencies with A have been presented in Fig. 2. For this analysis HPMA is designed on double sided copper claded RT duroid substrate (~,=2.33)of thickness, t=0.159 cm and loss tangent, tanA=0.00066. The present results are quite promising and the air gap concept may be applied to array configurations of microstrip antennas for creation of broad band antenna system which has potential applications. References 1 . J. C. Louvigne and A. Sharaiha , “Broad band tapered printed quadrifilar helical antenna”, Electron. Lett., 37 [15] 932-933 (2001). 2. S. C. Gao , L.W. Li, P. Gardner and P.S. Hal, “Wide band dual-polarised microstrip patch antenna”, Electron. Lett., 37 [20], 1213-1214 (2001). 3. J. F. Zurcher and F. Gardiol, Broadband Patch Antennas. Nonvood, MA: Artech House, 1995. 4. J. S. Dahele and K. F. Lee, “ Theory and experiment on microstrip antennas with airgaps”Jnst. Elec. Eng. Proc., 132 [7], 455-460, (1985). 5. K. F. Lee, K. Y. Ho, and J. S. Dahele, “Circular-disk microstrip antenna with an air gap”, IEEE Trans Antennas Propagat., 32, 880-884 (1984). 6. I. J. Bahl and P. Bhartia, Microstrip Antennas, Artech house, 1980.

Infrared Magnetic Response Metamaterials from a Virtual Current Loop Resonator

Zhiming Huang’,’, D. H. Zhang’, Yun Hou’, Junhao Chu’ ‘School of Electrical and Electronic Engineering,Nanyang Technological University, Singapore, 639798 ’National Laboratory for Inpared Physics, Shanghai Institute of Technical Physics, Chinese Academy ofSciences, 500 Yu Tian Road, Shanghai, 200083, People s Republic ofChina

Abstract Infrared magnetic response is achieved by our proposed new metamaterial with a three-layer structure. The metamaterial is formed by a pair of homogeneous parallel plates separated by a thin medium, in which a virtual current loop resonator can be formed upon excitation of an electromagnetic field. Strong magnetic response has been observed and the resonant frequency can be widely tuned by varying the structure dimensions. The easy fabrication and high interfacial quality of the new structure will make the applications of the magnetic response and negative refractive index metamaterials a reality. In 1968, Veselago conceived of a material whose index of refraction could be negative with both a negative permittivity and a negative permeability, which would reverse nearly all known optical phenomena [ 11. Such material has not been realized until recent years when the artificially structured materials, or metamaterials, were reported [2]. These metamaterials open the door to a variety of new physical phenomena and potential applications [3,4]. A negative permittivity is not unusual and occurs in any metal from zero frequency to the plasma frequency. However, a negative permeability, which means a negative magnetic response, at optical frequencies does not occur in natural materials. At present, a magnetic metamaterial has been formed from a periodic array of nonmagnetic, conducting, split-ring resonators (SRRs), achieved in essence just by mimicking a small LC circuit structure of eigenfrequency WLC = (LC)-ln. Each SRR structure consists of a magnetic coil with inductance L and a capacitor with capacitance C [5-81. Since the first demonstration at microwave frequencies [9], the achieved magnetic resonance frequencies have been increased by more than four orders of magnitude over the last few years [5-71, reaching a record of 370 THz (800 nm wavelength)

PI. The structure of the metamaterials currently studied needs to be as fine as possible [4]. A variety of potential applications, including higher resolution optical imaging and nanolithography, will be limited by the complexity of the SRR structures. For example, a superlens, one of the most desirable applications in the negative refractive index, at optical wavelengths requires the structure being extremely smooth with a surface roughness less than 1 nm. Otherwise, the surface imperfections would scatter the incident light and wash out the finer details carried by the evanescent waves [4, lo]. Another issue regarding the practicality of the present metamaterials is their complicated electromagnetic response, which makes their utilization as devices complicated and a full electromagnetic characterization difficult [ 111. Furthermore, to improve the oscillator strength of the magnetic resonance, the number of SRR per unit area should be high enough [S]. This is limited by the capability of manufacture. Therefore, there is a strong demand to explore the new structures of the negatively refracting metamaterials [4, 121.

In this report, we propose a new infrared metamaterial formed by two thin parallel plates separated by a dielectric medium, where the top plate is semitransparent for light. Magnetic resonant response at optical frequencies from the metamaterial is experimentally observed by spectroscopic ellipsometxy, and verified by Faraday’s law and optical transfer matrix methods. The new structure is of significant importance to the applications, especially at optical and terahertz frequencies, because

218

219

perfect interfaces and/or surfaces can be successfully realized in the layered structures using modem growth techniques [13, 141. Let's consider the metamaterials consisting of a pair of parallel plates with nanometer size in z direction and infinite in xy plane as shown in Figure 1. Assume that the thickness of the metamaterials is of subwavelength order (- 119 - 1/10 in our samples, where 1 is the wavelength of the excited field at the resonance frequency), which allows the composite to behave as an effective medium to external THz radiation. A virtual current loop (VCL) will be formed along a contour {Pi,P2,P3, P4} in Fig. 1 thanks to the displacement currents. To verify the above theory, we designed and fabricated three samples A, B, C using the structure of LaNi03/Pb(ZrTi)03/Pt on Si substrate with the optical semitransparent LaNiO3 layer on the top and metal Pt layer at the bottom. The thicknesses of the Pb(ZrTi)03 for samples A, B, and C are 645, 575, and 500 nm, respectively. The thicknesses of LaNiO3 and Pt are the same for the three samples, and they are 45 and 50 nm, respectively. The samples were grown by radio frequency magnetron sputtering under a working pressure of 15 mTorr at a rf power of 80 W. Spectroscopic ellipsometric (SE) measurements were carried out by an improved variable-angle infrared spectroscopic ellipsometer (PhE-104) [15] in the frequency range of 20 - 120 THz. The accuracy is better than 1% for tanp and cosA in the measurements. Figure 2 shows the frequency-dependent ellipsometric measurements for sample A at three different incident angles of 20,60 and 70", respectively. An obvious resonance peak is observed around 43.5 THz, which indicates that the amplitude of the p-polarization light is much stronger than the s-polarization light. The peak shifts slightly and its intensity varies with the angle of the incidence. Beyond this resonant peak, tanp is less than 1 for the three angles of incidence, which is not difficult to understand for the materials with electric response. However, the resonant response can come from either electric response or magnetic response. To clarify it, firstly let's see the dispersion of tanp as a function of the angle of incidence with purely electric response. The inset in Fig. 2 shows the dispersion of a typical uniform CaF2 thin film of 750 nm at 43.5 THz. The curve is smooth and has a minimum tanp at the principal angle (A = 90") of the material. The inset also shows the response peak values of sample A versus the angle of the incidence from 20 to 80'. A smooth evolution of the peak response in the broad angle range strongly suggests that the metamaterial can be viewed as homogeneous slab in the growth direction at the corresponding frequencies. If the magnetic response centered at 43.5 THz in the spectrum for sample A (Fig. 3, solid curve) results from the constituent parallel plates, then this resonant frequency should scale with dimensions in terms of Maxwell equations. In order to justify our findings, two more metamaterials (samples B and C) with different dimensions d were characterized (Fig. 3, dash and dot curves). The two metamaterials both exhibit a similar magnetic mode to sample A, and their resonant frequencies occur at 46.5 and 51.8 THz, respectively. We find an expected monotonic blue shift of resonant frequencies as the dimensions of plates are scaled down, which elucidates that the magnetic response is from the constituent parallel plates. We can use the transmission and reflectance information, which is extracted from SE fitting data, to evaluate the metamaterial effective permeability pefwith transfer matrix method [16, 171. In Fig. 3, we display the simulated real (ur) and imaginary (u,) parts of the effective magnetic permeability that corresponds to the samples A, B, and C, respectively. The magnetic resonant responses are obtained for all three samples with the same centre frequencies as that of ellipsometric data. Obvious negative permeabilities have been achieved when the frequencies are over the resonant ones for all three samples, because the induced dipole moments lag and are completely out of phase with the excitation

220

fields. This is an important precondition for the realization of a homogeneous layered metamaterial with a negative index of refraction. To further verify the magnetic response of the metamaterials, we performed a numerical simulation using the circulating current in the loop illustrated in Fig. 1. For simplification we assumed identical impedance for the top and bottom layers. All the other calculating parameters come from the experimental data. The results are shown in Fig. 4 for sample A. Both the shape and peak position coincide well with those calculated from the transfer matrix method. The narrowed width and increased height by Faraday’s law are mainly due to the substitution the symmetric top metal plate for the semitransparent conducting film. In conclusion, we have demonstrated that infrared magnetic resonance can be realized in the homogeneous layered metamaterials with a resonator consisting of a pair of parallel plates based on the VCL concept. This uniform planar structure we invented first meets the virtues to recover evanescent waves carrying the finest details of the object and be completely compatible with modern mature thin film technology. It is of great significance to the device applications and extension of higher frequencies to deep UV range. The uniform metamaterials offers further opportunity to extend the negative refraction concept. We believe that it will open previously unknown avenues of investigation in this fast-growing subject. This work was supported by A*Star SERC grant (No. 0421010078), National Natural Science Foundation (No. 60407014 and 60527005), National Grant Foundation Project (No. G001CB3095), and Shanghai Grants (06QH14018).

References: [l] V. G. Veselago, Sov. Phys. Usp.,10,509 (1968). [2] J. B. Pendry, A. J. Holden, D. J. Robbins, and W. J. Stewart, IEEE Trans. Microwave Theory Tech., 47,2075 (1999). [3] J. B. Pendry, Phys. Rev. Lett., 85,3966 (2000). [4] D. R. Smith, Science, 308, 502 (2005). [5] T. J. Yen, W. J. Padilla, N. Fang, D. C. Vier, D.R. Smith, J. B. Pendry, D.N. Basov, and X. Zhang, Science, 303,1494 (2004). [6] S. Linden, C. Enkrich, M. Wegener, J. Zhou, T. Koschny, and C. M. Soukoulis, Science, 306, 1351 (2004). [7] S. Zhang, W. Fan, B. K. Minhas, A. Frauenglass, K. J. Malloy, and S. R. J. Brueck, Phys. Rev. Lett., 94,037402 (2005). [8] C. Enkrich, M. Wegener, S. Linden, S. Burger, L. Zschiedrich, F. Schmidt, J. F. Zhou, Th. Koschny, and C. M. Soukoulis, Phys. Rev. Lett., 95,203901 (2005). [9] R. A. Shelby, D. R. Smith, S. Schultz, Science, 292,77 (2001). [lo] N. Fang, H. Lee, C. Sun, X. Zhang, Science, 308 534 (2005). [ 111 W. J. Padilla, A. J. Taylor, C. Highstrete, Mark Lee and R. D. Averitt, Phys. Rev. Lett., 96, 107401 (2006). [12] J. B. Pendry, Science, 306, 1353 (2004). [13] D. D. Fong, G. B. Stephenson, S. K. Streiffer, J. A. Eastman, 0. Auciello, P. H. Fuoss, C. Thompson, Science, 304, 1650 (2004). [14] G. Dehlinger, L. Diehl, U. Gennser, H. Sigg, J. Faist, K. Ensslin, D. Gr”utzmacher, E. Muller, Science, 290,2277 (2000). [151 Z. M. Huang, J. H. Chu, Appl. Opt., 39,6390 (2000). [16] D.R. Smith, S. Schultz, P. Marko, and C. M. Soukoulis, Phys. Rev. B, 65, 195104 (2002). [17] X. Chen, T. M. Grzegorczyk, B. Wu, J. P. Jr., and J. A. Kong, Phys. Rev. E, 70,016608 (2004).

221

20

40

60

80

100

t20

Frequency (THz)

FIG. 1: Illustration of uniform metamaterials with a open resonator consisting of a pair of parallel plates. Loop currents, closed by displacement currents (dashed lines), are excited by external electric and magnetic fields around the contour {PI,P2, P3, P4). The length between PI and P2 is 2a; the distance between the plates is d; and the thickness is dl and d2, respectively, for the plates 1 and 2. The permittivities are E I , ~ 2 and , E for the plates 1 and 2 and the medium between them; and the permeabilities p1, p2, and p are unity for all layers. k is the wave vector of incident light; H is the magnetic field; E is the electric field.

FIG. 2: Amplitude ratio tany of LaNi03/Pb(ZrTi)03/Pt metamaterial (sample A) measured at three different incident angles (square: 20*, circle: 60" and triangle: 70'). Inset: Maximum response ratio of p-polarization to s-polarization for sample A as a function of incident angle. The ratio of p-polarization to s-polarization for normal dispersion material CaF2 (refractive index: 1.36; thickness: 750 nm) with purely electric response is also depicted. The ratio of CaF2 has been multiplied by a factor of 30 to improve visibility.

Faraday's law -2 -4

5-

Frequency (THz) 20

30

40

M

60

70

SO

Frequency (THz)

FIG. 3: (Top) Reflectance ratio of the p-polarization to s-polarization response as a function of frequency for three different artificial magnetic structures, A (solid), B (dashed), and C (dotted) at an incident angle of 200. The resonance frequency is a function of the nanostructure parameters. (Middle and bottom) The real br)and imaginary @,) effective magnetic permeability functions as calculated by transfer matrix method for samples A, B, and C.

FIG. 4: Comparison of the real (top) and imaginary (bottom) effective magnetic permeability functions for sample A calculated by transfer matrix method (solid curve) and Faraday's law method (dash curve), respectively.

This page intentionally left blank

223

AUTHOR INDEX

Abshinova, M.A. 97, 109 Acher, 0. 133 Arnaut, L.R. 21 Ashok 153 Badola, N.K. 153 Bankov, S.E. 119 Bera, J. 48 Bhattacharya, S. 29 Blackbum, J.F. 21 Bridge, A.R. 21 Chatterjee, R. 29, 33 Chen, H.S. 186 Chen, L. 75 Cheng, X.F. 38 Cheng, X.X. 186 Chu,J. 218 Chung, C.-J. 149 Clarke, R.N. 21 Deng,C. 195 Deng,C.R. 83 Deng,L. 75 Deng, L.J. 89 Domashevskaya, E.P. 173 Dorofeenko, A.V. 127 Elsukov, E.P.

59

Feng,L. 38 Filimonov, Yu.A.

119

Gan, Y.B. 71,83, 105, 113,209 Gao,Y. 38 Gregory, A.P. 21 Grigorievski, V.I. I19 Grigorievskiy, A.V. 119 Hamzah, N.R. 52 Han,M. 75 Hou,Y. 218 HSU,R.-T. 149 Huang, Z. 218 Hussain, T. 161 Imtiaz, S.

161

Inoue, M. 127 Ittyachen, M.A. 165 Jacob,M.V. 44 Janjua, N.K. 161 Jean, J.-H. 149 Kalinin, Yu.E. 173 Kashkarov, V.M. 173 Kazantseva, N.E. 97, 109 Kisel, V. 3, 201 Komarova, M.A. 169 Kong, J.A. 186 Kong, L.B. 83, 105, 113 Kopn0v.G. 67 Kuritka, I. 109 Lagarkov, A. 3 Lakhtakia, A. 11 Lees, K. 21 Li,Z.W. 71,79,93, 101, 105, 113 Liang, D. 75 Lim,H. 191 Lin, G.Q. 71, 83.93, 101, 105, 113 Lisyansky, A.A. 127 Liu, L. 83, 195,209 Lopatin, A.V. 97 Mann,S. 216 Matitsine, S. 83, 191,209 Merzlikin, A.M. 127 Molodtsov, S.L. 173 Molokanov, V.V. 169 Munk, B.A. 179 Naaman Ron. 67 Nath, T.K. 157 Nikitov, S.A. 119 Ong,C.K. 93 Osipov, A.V. 59 Paul, S. 157 Peng, T.-M. 149 Petrov, D.A. 59 Pourush, P.K.S. 153, 216 Pourush, R. 153,216

224 Qing Anyong

212

Ra0,X.S. 191 Ren, Q. 38 Reyes, 3.A. 11 Roxanov, K.N. 83 Roy,P.K. 48 Rozanov, K.N. 59 Sachdev, V.K. 33 Sachdeva, V.K. 29 SAha, P. 97, 109 Shalygin, A.N. 169 Shalyguina, E.E. 169 Singh, B. 157 Singh, R. 33 Sitnikov, A.V. 173 Srivastava, G.P. 153 Starostenko, S.N. 59 Stognej, O.V. 173 Storozhilov, S.A. 173 Sun, J. 38 Sun,X.B. 38 Tan,P.K. 209 Teo, M.L.S. 113 Terekhov, V.A. 173

Turishchev, S.Yu. Tyagi, G.S. 153 Unnikrishnan, N.V.

173

165

Vager,Z. 67 Valsamma, M.S. 165 VilEfikovB, J. 97, 109 Vinogradov, A.P. 127 Vysotskii, S.L. 119 Wang,D.W. 186 WU,B.-I. 186 w u , Y.P. 93 Xie, J. Xu,X.

75 71

Yang, H.L. 38 Ya0.X. 141 Zhang, D.H. Zhang, F.J. Zhang, J.J. Zhang,Y. Zhou,P.H.

218 38 186 195 89