electromagnetic materials

Proceedings of the Symposium P ELECTROMAGN ETlC MATERIALS

This page intentionally left blankThis page intentionally left blank

ICMAT 2007 International Conference on Materials for

Advanced Technologies

Proceedings of Symposium P

E lectromagnetc Materials

SUNTEC, Singapore 1 - 6 July 2007

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 * LONDON * SINGAPORE BElJlNG * S H A N G H A I * HONG KONG * TA IPEI - C H E N N A I

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-Publication Data 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-1 3 978-981-270-953-0 (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:

Metamaterials (Negative Index Materials)

Frequency Selective Periodic Structures 0

0

Dielectric and Magnetic Composites (with micro- or nano-sized inclusions)

Smart Materials (includes thin films, tunable dielectrics, etc)

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

ix

CONTENTS

Session P1: Opening Session Chair: S.A. Nikitov

P-1 -IN1 Electromagnetic Energy Absorption within Extensive Impedance Structures 3

P-1 -IN2 Electro-Optic Structurally Chiral Materials 11

A. Lugarkov and V. Kisel

A. Lukhtakia and J.A. Reyes*

Session P2: Dielectric Composites Chair: L.R. Arnaut

P - 2 -IN 3

P-2-OR1

P-2-OR2

P-2-OR3

P-2-OR4

P-2-OR5

P-2-OR6

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

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

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

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

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

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

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

21

29

33

38

44

48

52

Session P3: Magnetic Composites (1)

P-3-IN4

Chair: 0. Acher

Reconstruction of Intrinsic Permeability of Inclusions from the Measured

K.N. Rozanov*, A.V. Osipov, D.A. Petrov, S.N. Starostenko and E.P. Elsukov Permeability of a Composite 59

P-3-OR7 Interface Magnetism G. Kopnov, Z. Vager and Ron. Naaman*

67

P-3-OR8 Characteristics of Effective Permeability and Resonance Frequency for Barium-Ferritempoxy Composites 71 Z. W. Li, Y. B. Can, X . Xu and G. Q. Lin

X

P-3-OR9

P-3-OR10

P-3-OR11

Microwave Absorbing Properties of Amorphous FeCuNbSiB Microwires

M. Han*. D. Liang, L. Chen, J. Xie and L. Deng Multilayer Composites 75

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

79

EM Properties of Composites with Glass-Coated Amorphous Ferromagnetic Wires 83 L. Liu*, S. Matitsine, L.B. Kong, G.Q. Lin, C.R. Deng, Y.B. Can and K.N. Roxanov

Session P4: Magnetic Composites (2) Chair: L.J. Deng

P-4-IN5

P-4-OR12

P-4-OR 13

P-4-OR 14

P-4-OR 15

P-4-OR 16

P-4-OR 17

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.P. Wu, Z. W. Li, C. Q. Lin and C. K. Ong Y-Type Barium Hexaferrites 93

Carbonyl Iron Composite Materials for High-Frequency Applications 97 M.A. Abshinova, A. V. Lopatin, N.E. Kazantseva, J. Vilda'kova' and P. Sa'ha

EM Properties in Composites with the Filler of Electroless-Synthesized Ni-P Powder 101 G.Q. Lin and 2. W. Li

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

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

109

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

113

Session P6: Metamaterials (1) Chair: A.N. Lagarkov

P-6-IN6 hlagnetic and Acoustic Metamaterials 119 S.A. Nikitov, S. E. Bankov. Yu.A. Filimonov, A. V. Crigorievskiy, V.I. Grigorievski and S. L. Vysotskii

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

P-6-IN7 127

P-6-IN8 Metamaterials: A New Route to Microwave Magnetism 0. Acher

133

Session P7: Materials Processing

P-7-IN9

P-7-OR18

P-7-OR19

P-7-OR20

P-7-OR21

P-7-OR22

P-7-OR23

P-7-OR24

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

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

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

xi

Chair: X. Yao

141

149

153

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

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

161

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

Annealing Effect on Magnetostatic and Dynamic Properties of Fehl.4Ni,6Cr3,zSiz,,Nb7.8MnJ.6B18 Amorphous Ribbons 169 E.E. Shalyguina*, M.A. Kornarova. V. V. Molokanov and A.N. Shalygin

165

XANES Investigations of Interatomic Interactions in (CoFeZr),(SiOz)l., Nanocomposi tes 173 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

Chair: J.A. Kong

Why Periodic Structures May Not Be Able to Synthesize Negative Indices of Refraction 179 B.A. Munk

On Negative Refractive Metamaterials: Characterization, Bianisotropy, and Applications 186 J.A. Kong, H.S. Chen, X.X. Cheng, J.J. Zhang, D. W. Wang and B . 4 Wu

P-9-IN11

P-9-OR25 Ultra-Thin Radar Absorbing Structures Based on Short Strip Pairs 191 X.S. Rao. S. Matitsine and H. Lim

P-9-OR26 Electromagnetic Characterisation of Conductive Helixes 195 C. Deng, L. Liu and Y. Zhang

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

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

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

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

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

20 1

209

212

216

218

Author Index 223

Session P l

Chair: S.A. Nikitov

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 Gi is 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 - iy normal at y = 0 : +

where q = 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

0.002

0

\ ,

0

-0 5

0 0 5 1 Re(ZIv)’s Fig. 4

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 ’ Y o .

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 = y0/2 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 half- space 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 +I

2 ’ kz -2 -1 0

Fig. 9 bj Fig. 10

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. 1 lc, d. Thus, the structures shown in Fig. 8 and Fig. 1 lc 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 Fig. 11 c)

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 semi- infinite 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.

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. 10- 13,2002. 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. 3966- 3969,2000. A.N. Lagarkov, V.N Kissel, “Near-Perfect Imaging in a Focusing System Based on a Left- Handed-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- 1 Ith 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

94-99, 1968.

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 electro- optic 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 xdc + Uz sin xdc) , xdc E [0,./2] . (6)

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

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(z cos I$ + y sin I $ ) ] H(z, y, z ) = h(z) exp [ir;(z cos 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< o o 1 -sin< cos< 0

o -sin< cosc )

it can be shown that $ ( z ) satisfies the 4x4 matrix ordinary differential equation [4]

(12) d -

dz -#(z) = iA’(z) . 7J’(z), o < z < L .

Detailed expressions of the 16 elements of the cumbersome matrix A’(z) are available elsewhere [4], correct to the first order in Ed“.

By virtue of linearity, the solution of the 4x4 matrix ordinary differential equation (12) must be of the form

$(zz) = U’(zz - ~ 1 ) . @(z1) , o 5 ze 5 L , e = 1 , 2 , (13)

whence q ( L ) can be found in relation to G(0). The matrix u’(L) has to be determined numerically [2].

2.3 Reflection and transmission The incident plane wave is delineated by the electric field phasor

where UL and uR are the amplitudes of the LCP and RCP components, respectively. The electric field phasors associated with the reflected and transmitted plane waves, respectively, are given as

(15) 2s - p- is + p- -ik,zcos8

+rR-) Jz e Z 5 0 ,

14

and

The amplitudes TL,R and tL ,R indicate the as-yet unknown strengths of the LCP and RCP components of the reflected and transmitted plane waves, both of which are elliptically polarized in general.

The propagation vector of the incident plane wave makes an angle 8 E [0, x / 2 ) with respect to the +z axis, and is inclined to the x axis in the xy plane by an angle + E [O, 2x1; accordingly, the transverse wavenumber K. = k, sine, where k , = w- is the wavenumber in free space. The free-space wavelength is denoted by A, = 27r/k0. The vectors

s = -&sin4 + Q, cosq5, pi = =F (uz cos q5 + uy sinq5) cos8 + u, sin8

are of unit magnitude. The reflection-transmission problem amounts to four simultaneous, linear algebraic equation

[2], which can be solved by standard matrix manipulations. It is usually convenient to define reflection and transmission coefficients, which appear as the elements of the 2x2 matrixes in the followine: relations:

Co-polarized coefficients have both subscripts identical, but cross-polarized coefficients do not. The square of the magnitude of a reflection or transmission coefficient is the corresponding reflectance or transmittance; thus, RLR = lrLR12 is the reflectance corresponding to the reflection coefficient TLR, and so on.

3 Numerical results and discussion The number of variables for a comprehensive parametric study of electro-optic SCMs is large. These variables include the local isotropy, uniaxiality, or biaxiality, as determined by the relative values of E&; the local point group symmetry of which there are 20 classes, as determined by the relative values of T J K ; the two angles of incidence 8 and 4; the angle x of the tilt dyadic, the half-pitch R, and the normalized thickness L/R; and the angle X d e . Several numerical studies have been published recently [3, 4, 6, 71.

Sample results are provided here for a locally biaxial SCM, since such materials can offer high electro-optic coefficients which would lower the magnitude of the applied dc electric field. The chosen point group symmetry is orthorhombic mm2, and the relative permittivity scalars and the electro-optic coefficients are the same as for potassium niobate [S]. Furthermore, normal incidence being the most common condition for using planar optical devices, we set 8 = 0". Finally, the effect of q5 not being qualitatively significant on the exhibition of the CBP, we set q5 = 0".

When Ed' is small enough in magnitude, it has no noticeable effect on the remittance spectrums. But that conclusions changes for high values of \Edcl [ 3 , 4 , 6 , 71. Figure 1 shows the

15

Figure 1: Principal reflectances and transmittances of a locally biaxial SCM slab of thickness L = 20R as functions of the free-space wavelength A, and the orientation angle x d c of the applied dc electric field, when Ed' = 0.67 x lo9 V m-' and 6 = 4 = 0". The local crystallographic class of the SCM is orthorhombic mm2. Other parameters are: c r ) = 4.72, cf) = 5.20, $) = 5.43, TI3 = 34 x m V-', ~~3 = 63.4 x m V-l, r42 = 450 x IO-" m V-l, T51 = 120 x m V-', all other T J K = 0, h = 1, a= 150 nm, and x = 90".

m V-', r23 = 6 x

principal remittance spectrums of a structurally right-handed SCM with half-pitch = 150 nm and tilt angle x = go', when Ed" = 0.67 x lo9 V m-' and x& E [0", 90'1. The chosen magnitude of Ed" is high enough to have an effect on the CBP, which also means that the reflectance and the transmittance spectrums change with x&. The center-wavelength of the Bragg regime is 646 nm and the full-width-at-half-maximum bandwidth is 69 nm for x d c = go', but the corresponding quantities are 667 nm and 40 nm for x d c = 0". In addition, the peak value of RRR diminishes by about 10% as x& changes from 90" to 0".

Figure 2: Same as Fig. 1, except that Ed" = -0.67 x lo9 V m-I.

The situation changes significantly when the sign of Ed" is altered, as exemplified by Fig. 2

16

for EdC = -0.67 x lo9 V m-l. The center-wavelength of the Bragg regime is 688 nm and the full-width-at-half-maximum bandwidth is 15 nm for X d c = go", but the corresponding quantities remain at 667 nm and 40 nm for x d c = 0". In addition, the peak value of RRR increases by about 600% as X d c changes from 90" to 0".

Thus, the exhibition of the CBP is affected dramatically in the center-wavelength, the bandwidth, and the peak co-handed and co-polarized reflectance by the sign of Ed' as well as the orientation angle X d c ; in addition, there is a thresholding effect with respect to IEdcl. Although numerical results are presented here only for normal incidence, several numerical studies confirm that these conclusions also apply to oblique incidence. Thus, the possibility of electrical control of circular-polarization filters has been established. Furthermore, periodically perturbed electro-optic SCMs to produce polarization-universal rejection filters are also possible [9]. Theoretical studies on particulate composite materials with eledro-optic inclusions [lo, 111 suggest the attractive possibility of fabricating porous, electro-optic SCMs with sculptured-thin-film technology [2].

The conclusions presented thus far are expected to apply not only for SCMs with continuously changing ~sc"(z ) defined by (l), but also for piecewise homogeneous versions thereof [12]. Such piecewise homogeneous versions are classified into ambichiral, equichiral, and finely chiral types [13]. Finely chiral types are very close to continuously varying SCMs; and, whereas equichiral types do not distinguish between LCP and RCP light, ambichiral types do [14]. Elec- trically controllable ultranarrowband filters are possible by inserting a central twist defect in an electro-optic SCM [15, 161.

To conclude, the possibilities of thinner filters and electrical manipulation of the CBP, depending on the local point group symmetry as well as the constitutive parameters of the SCM, having been established, this presentation of theoretical research is expected to provide impetus to experimental research.

References [l] P.G. de Gennes and J. Prost, The Physics of Liquid Crystals, 2nd ed., Clarendon Press,

Oxford, UK, 1993, Chap. 6.

[2] A. Lakhtakia and R. Messier, Sculptured Thin Films: Nanoengineered Morphology and Optics, SPIE Press, Bellingham, WA, USA, 2005, Chap. 9.

[3] A. Lakhtakia and J.A. Reyes, Theory of electrically controlled exhibition of circular Bragg phenomenon by an obliquely excited structurally chiral material. Part 1: axial dc electric field, Optik (accepted for publication); http://www.arxiv.org/physics/O6lOO73.

[4] A. Lakhtakia and J.A. Reyes, Theory of electrically controlled exhibition of circular Bragg phenomenon by an obliquely excited structurally chiral material. Part 2: arbitrary dc electric field, Optik (accepted for publication); http://www.arxiv.org/physics/O6l1015.

[5] R.W. Boyd, Nonlinear Optics, Academic Press, London, UK, 1992, Chap. 10.

[6] J.A. Reyes and A. Lakhtakia, Electrically controlled optical bandgap in a structurally chiral material, Opt Commun. 259 (2006), 164-173.

17

[7] J.A. Reyes and A. Lakhtakia, Electrically controlled reflection and transmission of obliquely incident light by structurally chiral materials, Opt. Commun. 266 (2006), 565- 573.

[8] M. Zgonik, R. Schlesser, I. Biaggio, E. Volt, J. Tscherry, and P. Gunter, Material constants of KNb03 relevant for electro- and acousto-optics, J. Appl. Phys. 74 (1993), 1287-1297.

[9] J.A. Reyes and A. Lakhtakia, Optics of electrically controlled structurally chiral material with periodic transverse perturbation for polarization-universal bandgaps, Opt. Commun. 270 (2007), 51-57.

[lo] A. Lakhtakia and T.G. Mackay, Electrical control of the linear optical properties of particulate composite materials, Proc. R. SOC. Lond. A 463 (2006), 583-592.

[ll] T.G. Mackay and A. Lakhtakia, Scattering loss in electro-optic particulate composite materials, http://www.arxiv.org/physics/O702039.

[12] E. Reusch, Untersuchung uber Glimmercombinationen, Ann. Phys. Chem. Lpz. 138 (1869), 628-638.

[13] I.J. Hodgkinson, A. Lakhtakia, Q.h. Wu, L. De Silva, and M.W. McCall, Ambichiral, equichiral and finely chiral layered structures, Opt. Commun. 239 (2004), 353-358.

[14] A. Lakhtakia, Ambichiral, electro-optic, circular-polarization rejection filters: theory, Phys. Lett. A 354 (2006), 330-334.

[15] A. Lakhtakia, Narrowband and ultranarrowband filters with electro-optic structurally chiral materials, Asian J . Phys. 15 (2006), 275-282.

[16] A. Lakhtakia, Electrically tunable, ultranarrowband, circular-polarization rejection filters with electro-optic structurally chiral materials, J . Eur. Opt. SOC.: Rapid Pub. 1 (2006), 06006; http://www.jeos.org/index.php/jeos~rp/article/viewFile/O6006/218.

Session P2

Chair: L.R. Arnaut

Modelling and Measurement of Advanced Carbon Fibre Reinforced Plastic Composites for EM1 Control of Spacecraft

L. R. Arnaut, J. F. Blackburn, K. Lees, A. R. Bridge, R. N. Clarke, A. P . Gregory

Enabling Metrology Division National Physical Laboratory

Teddington TWl 1 OLW, United Kingdom

Abstract This paper reports on modelling and measurement activities for determining the shielding effectiveness of carbon fibre plastics at radio and microwave frequencies, based on a collaborative project.

Introduction

Over the past three decades, there has been an increasing use of conductive composite materials such as carbon fibre reinforced plastics (CFRPs) in aircraft structures. Although primarily chosen for their high tensile strength and light weight, their high conductivity also makes them attractive for shielding of external EM radiation and for conduction and grounding of currents. While extensive electromagnetic modelling and experimental characterisation studies have been performed, most activities relate to simplified CFRP models with a few layers, because of computational or tractability issues, to the extent that results cannot accurately estimate the performance of realistic CFRPs used in practice.

Composite material structures differ significantly from metallic structures, particularly at frequencies below 30 MHz, where their magnetic shielding efficiency and conductivity are about three orders of magnitude smaller than those of metals. This difference is much reduced at higher frequencies. Nevertheless, the requirements for CFRPs for spacecraft are not the same as for aircraft. While lightning hazard and “low”-frequency radio communications (from VLF up to HF) were initially considered as the most critical issues for the use of composite materials on aircraft, for spacecraft this is not the case: the latter are not supposed to withstand direct lightning strikes, and they do not use communication frequencies lower than VHF (with the exception of Search & Rescue missions (1 20 MHz)). Instead, low-frequency issues on spacecraft focus on low level analogue signals, as used in scientific and Earth observation spacecraft (for sensitive signals) and switching power supplies and electrical motors (for sources of noise).

The method of joining CFRP panels, e.g., when constructing shielding boxes, is a critical factor as far as electrical bonding and EM shielding are concerned. One of the most adequate and convenient tools to characterize the shielding properties of conductive composite materials is the surface transfer impedance. So far, CFRP laminated composites have been used in spacecraft structures, mainly as skins of sandwich panels with an aluminium honeycomb. Spacecraft construction has traditionally relied on aluminium structures that provide good shielding performance and low impedance grounding paths. When aluminium honeycomb structures with an CFRP skin are used, the issue of the bonding methodology becomes an important factor. Often, equipment bonding relies on a network of grounding rails made of semi-rigid aluminium foils. For some future spacecraft applications, CFRP structures without metal honeycomb may be used, either for structural panels or for equipment unit cases. Their shielding and bonding point of properties have a priori an inferior performance to aluminium. This deterioration requires electromagnetic characterisation. To assess the electrical and EM performance of composite materials, however, the existing research material from aircraft industry forms a good starting point.

In this paper, we report on results obtained from a research project carried out for the European Space Agency (ESA). The overall project was a joint project between four German space companies and two British research laboratories. This paper focuses on the work performed by one of the partners (”L).

21

22

Other aspects of the work will be covered in a companion paper. The objectives and deliverables of the project were to provide experimental and theoretical characterisation of the shielding effectiveness and the transfer impedance of a set of specimens of different CFRP materials and jointing techniques, to develop measurement techniques, and to develop necessary analytical tools for EMC performance evaluation. Specifically, it involved a literature study, the manufacturing and waveguide measurement of a set of CFRP samples including different jointing techniques, an overview of suitable measurements and modeling techniques, the development of semi-analytical and full-wave numerical tools for modeling of realistic CFRP laminates, and a comparison of theoretical with experimental data.

Measurements

At the start of the measurement programme, a survey of the current state-of-the-art for measurement methods of shielding materials was undertaken. Of the twenty techniques surveyed, the waveguide technique was considered to be the most appropriate for this study. Waveguide measurements allow for the measurement of anisotropy, exhibit low uncertainties, enable the characterisation of joints, and require only simple a geometry of the specimen that is calculable. A major drawback of the technique is the limited frequency coverage. The ASTM D 4935-99 standard for shielding effectiveness measurements [ASTM, 19991 was found to be not suitable, because it is a coaxial technique that cannot be employed to characterise anisotropy or jointing. Furthermore, its requires a more complex specimen geometry which makes machining more expensive.

Guided-wave measurements were performed of the transmission (&I) scattering parameter to estimate the shielding effectiveness (SE) of specimens of CFRP materials in the frequency ranges 0.75-1.12 GHz (IEC R9) and 1.7-2.6 GHz (IEC R22). To enable a full constitutive characterization of the instrinsic (bulk) CFRP material, three different uniaxial specimens were constructed with fibre orientations in the x-, y- and z-directions, respectively. A sparse grid of nylon fibres in the perpendicular direction maintained an approximately constant position and spacing between the carbon fibres in longitudinal direction. Measurement results on each material in waveguides IEC R9 and IEC R22 agreed to within the expected systematic uncertainties for S21, except for 2 out of 11 specimens. Leakage due to small gaps between the CFRP specimens and the waveguide walls was found to be a major potential source of uncertainty that may dominate measurement results, unless it is properly remedied or accounted for. A systematic study of different gap reduction methods, involving precision machining, taping, silver painting, and clamping of specimens was undertaken on a rekrence aluminium specimen and on an IEC R22 CFRP specimen (figure 1).

Figure 1: Aluminium reference plate, waveguide-mounted CFRP and honeycomb specimens.

Figure 2(a) compares results using the different methods for an IEC R22 aluminium reference plate and a CFRP specimen. The results indicate a superior effect of painting on reducing the leakage through air gaps compared to taping. The order of magnitude difference between standard and painted specimens is about 40 dB for both aluminimum and CFRP specimens (figure 2), indicating consistency in the reduction of gap effects when using painting, irrespective of the type of material

23

used as the specimen. Silver painting resulted in measured values of the order of -65 dB, which is in excess of the - 100 dB noise level of the network analyser that defines the sensitivity of the method, The data show almost no frequency dispersion (ripple) within the waveguide band. Clamping exerts pressure on the specimen, which results in changes of the CFRP resistivity and causes frequency dispersion. Combined clamping and taping causes extra leakage due to the bridging effect of the tape. It was found that the best results were obtained by using conductive paint, to ensure good electrical contact between the specimen and its holder. Therefore, all subsequent CFRP measurements were performed by applying conductive paint in a standardized manner.

Fmequncy GHz Or -

l o 1 1,' I 9 2 1 2 3 2 5 2 7

20

a 0

crrc .I" ~

=?::$ i zzj 2;40: +nnP.mm@ .I0 i

:iA0i m-

/- --a ,/-

,.- -*--. ,.-. ,'

-_..r --.-I :I ,,.C

.I0

.w -

4" 70 1

(b)

Figure 2: Comparison of different methods for mitigation of gap effects on S2, for IEC R22 specimen: (a) reference aluminium plate specimen; (b) CFRP specimen no. 2.

Figure 3 compares transmission measurement results for the various specimens fabricated and tested as part of this study. Superior shielding effectiveness was found for sample no. 4 because it is characterized by a conductive host medium. Good repeatability was found when comparing with results for the IEC R22 waveguide [Various, 20061.

Frequency GHz

Figure 3: Shielding effectiveness of specimens measured in IEC R9 waveguide.

24

Joints

After an initial measurement as a single-piece material sample, CFRP specimen no. 10 was cut up and jointed, and the effect of this joint on the shielding effectiveness was studied by re-measuring the jointed specimen in the IEC R22 waveguide.

The jointing consists of an assembly of the two specimen pieces using bolts, washers and nuts (figure 4). All surfaces were sanded so that surface conductivity was present to improve the joint. The preferred measurement method for assessing the effect of the joint is to test the specimen with an open gap closed by the lid but without sealing the residual air gap between the lid and specimen, and secondly with additional silver paint on the contact so as to seal this air gap.

(a) (b)

Figure 4: IEC R22 specimen with (a) slot (front view) and (b) joint (rear view).

The measurement results are shown in Table 1, where the reference numbers refer to the following scenarios: joint 0 is without the lid in place, joint 1 is with the lid screwed to the specimen, joint 2 is with the lid screwed to the specimen and including a silver paint connection between lid and specimen, and joint 3 is with the lid screwed to the specimen, including a silver paint connection between lid and specimen and a silver paint connection between the screws and the specimen.

The values of the uncut specimen provide a benchmark for the transmission loss, which should be subtracted from the values in the other columns to quantify the quality of the joint on the SE.

Table 1: Results of jointing on shielding effectiveness of IEC R22 CFRP specimen no. 10.

The Type B (systematic) uncertainty, due to calibration in the magnitude of and physical differences between nominally identical materials, is estimated to be of the order of 5 ~ 1 0 ~ . This uncertainty always acts in the same direction for a given calibration. Therefore, it will not affect the

25

relative uncertainties when comparing specimens measured in the same size of waveguide. However, this uncertainty applies when comparing results between waveguides. For example, one particular CFRP material gave -62.2 (+4.3/ -9.0) dB at 0.75 GHz and -57.8 (+2.8/-4.3) dB at 1.7 GHz. Based on repeatability tests, it was found that Type A (random) errors in the measurements are significantly smaller.

Due to the high reflection and low transmission of the CFRP specimens, the extracted values of the intrinsic complex permittivity from the S-parameter waveguide measurements exhibit uncertainties of the same order of magnitude as the estimated value of the permittivity itself [Various, 20061.

Modelling and Simulation

Both semi-analytical and full-wave numerical modelling methods were developed and their results compared. The simulated structure consists of P plies distinguished by different orientations of the fibres in the plane (rotated with 45 degree increments between plies) (figure 5(a)), whereas each ply contains L layers of parallel fibres, mimicking a realistic CFRP. In actual materials, the fibres are randomly spaced from each other [figure 5(b)] . The resulting small contact points between adjacent fibres cause a certain conductivity perpendicular to the fibre direction, although it is more than three orders of magnitude lower than the effective conductivity along the fibre direction. The typical fibre spacing in realistic CFRP sheets is about 1.2 times the wire diameter. Fibres in adjacent layers are spaced by a distance t; fibres in adjacent plies are spaced by 2t.

Figure 5: (a) Cross-sectional view of multilayer CFFW specimen of width Wconsisting of P plies containing L layers of parallel fibres. (b) Illustration of actual fibre spacing (cross-sectional view).

A semi-analytical model of CFRPs uses second-order impedance boundary conditions for representing the fibres, being a special case of higher-order impedance boundary condition (HOIBC) modelling [Yatsenko et al., 20001. This model is based on the following assumptions: the wire radius is much smaller than the wire spacing (r << d); the concept of average boundary conditions is valid, implying kd << 1; and the host material is homogeneous, isotropic and lossless. The array formed by a single planar CFRP sheet of infinite extent in both directions defines a homogenized current sheet and an average array impedance ZA given by

Z , ( k ) = Z,d + j a a 2

where r /

26

L J

7, = ,/=, k, = m G , k , = k, sin B cos(4 - 4f ), k, = sin Bsin(4 - q5/)

where 4 and 0 are standard spherical angles (azimuth and elevation) for the incident wave vector kinc with respect to a global reference frame oxyz, and with E,,, and p,,, representing the permittvity and permeability of the CFW material, respectively. The reference plane of incidence (4 = &) is spanned by the normal to the CFRP sheet and the direction & of the wires.

n = l , ~ = 8 , i = m . d = s . ~ m , ~ . 5 1 1 m . t = ( t . 7 ~ ~ ~ @ , ~ ~ ( ~ , ~ ~ ; ~ ~ / ~ = ~ .

M), I 80,

.................... ....................

...................

.................... ...................

1 0' 10"

Frequency (MHz) Fresueney ( M Z i

..............

...................

108 Id lorn Frequency(hPHz1

.................... ...

..............

....................

Id ld 10" Fwueney W z )

I a=l P==, k2O. d=57&m. r=Z.Sprn, 1=0.72vm; s,bf=(O.Oide4;r , f~=2.77~~ yS=20,~.125x1o6sIm

7 , m I j

....................

-m .................. ..I.. ....... , .........

iri za -4 ................... ;. ..................

D

-60 ................... ; ..................

-20 ................ ...! ..................

f .................... ! .................

2 .................... :.. ..............

.M

......

I 00 $0' 1o'n FreqrnrY IWd

(b)

Figure 6: HOIBC modeling of multi-ply CFRP composite: frequency responses of shielding effectiveness (SE) and reflectivity @) 2x2 matrices.

27

The overall shielding effectiveness can then be expressed as SE = -20 loglo(T) using the transmission coefficient

Because of the anisotropic properties of the multilayer, a more complete characterization is given by a shielding matrix defined in an analogous manner from the transmission matrix 1 specified by E' = . El"". The reflectivity is defined in analogous manner, by replacing Tap with corresponding Rap. Reflectivity is an important parameter when using CFRPs for internal shielding. Sample results for material type M55J (specimen 10) are shown in figure 6. Comparison with results for a single layer (P=L=l) has shown that such a single layer already provides a good anisotropic shielding performance [SETEViO) = 8.5x10-" dB E 1, S E T M V i O ) = 32 dB]. Increasing the distance between wires within a plane decreases the co-polarized shielding effectiveness, but increases the cross-polar shielding.

As a validation, a different approach based on the period methodic of moments (PMM) was used to model CFRPs numerically using fill-wave simulation. Based on the Floquet-Bloch theorem, a rigorous calculation of wire interactions within a plane is made possible, limited to a finite-size k- space interaction shell (Blackbum et al., 2005). As with HOIBCs, the interaction between layers is accounted for by an algorithm for propagation with multiple reflections. Modelling of a CFRP as a wire-type frequency-selective surface (FSS) permits inclusion of 2-D features ( e g , bends). Connections between wires in adjacent cells were modeled by using overlapping triangular basis functions [figure 7(a)]. Interlayer gaps of width t were simulated as vacuum spacers. The embedding within a dielectric medium was modeled by placing additional dielectric slabs on both sides of the FSS. Reasonably good agreement was found between results obtained for quasi-static excitation of a single-layer CFRP when comparing HOIBC and PMM [figure 7(b)], as well as with other numerical techniques [Parise et al., 20031, [Holloway et al., 20051, [Various, 20061.

Figure 7: PMM modeling of CFRP composite: (a) intercell wire connections for bent wires; (b) PMM frequency response of SE for single-layer straight-wire CFRP (inset: results obtained with HOIB).

Woven CFRP fabrics (T300 material) were also simulated, by introducing strands forming orthogonal braids that were connected by two C-shaped rings to provide the galvanic connections (figure 8). Compared to a geometrically equivalent but nonwoven fabric (M559, the effect of weaving was found to be a lower cut-off frequency and a steeper roll-off.

28

front

. . ! . I .

40 layers = 1 ply

Total of 4 plies = 160 layers

I -5e05 -4e05 -3e05 .2e-05 -1e45 0 le-05 2e~05 3%05 4e~05 -5

Figure 8: Modelling of woven fabric CFRPs.

Conclusions

The theoretical and experimental results of this study of shielding properties of carbon fibre reinforced plastics for EMC in space applications has shown that good correlation between waveguide-measured and predicted performance can be obtained, even for the idealized and simplified theoretical model. More advanced CFRP materials with superior matrix compounds are currently under development, with the aim of increasing the shielding performance still further.

Acknowledgements

This work was supported by the European Space Agency under contract no. 18985/05/NLIJA, with HPS GmbH as the responsible entity.

References

[l] ASTM, Standard Test Method for Measuring the Electromagnetic Shielding Effectiveness of Planar Materials, D 4935-99 (1999) [2] Blackbum, J.F. and Amaut, L.R., “Numerical convergence in periodic method of moments analysis of frequency selective surfaces based on wire elements”, IEEE Trans. Anennas Propag., vol. 5 3 , no. 10, pp. 3308-3315 (Oct. 2005). [3] Holloway, C.L., Sarto M.S., and Johansson, M., “Analyzing Carbon-Fiber Composite materials with equivalent- layer models”, IEEE Trans. Electromagn. Compat., vol. 47, no. 4, pp. 833-844 (Nov. 2005). [4] Parise M, Sarto MS, “Efficient formulation of high-order boundary conditions for the high-frequency modelling of multilayer composite slab’’, Proceedings 2003 IEEE Symposium on EMC, pp. 753-758 (Aug. 2003) [ 5 ] (Various authors) ‘‘Spacecraft EMI control in the presence of composite materials: CFRP measurements results, survey of modelling techniques methods, and development of analytical & numerical tools for multilayer CFRP modeling” (4 parts), Report Series to European Space Agency(ESA), contract no. 18985/05/NL/JA (“EM in the Presence of Spacecraft”) (May 2006). [6] Yatsenko, V.V., Tretyakov, S.A., Maslovski, S.I., and Sochava, A.A.,“Higher order impedance boundary conditions for sparse wire grids”, IEEE Trans. Antennas Propag., vol. 48, no. 5, pp. 72C727 (May 2000).

Electrical Properties of Graphite Filled Cement Composites for Device Application

S Bhattacharya', V. K. Sachdeva', R Chatterjee' ' Department of Physics, Indian Institute of Technology, Hauz Khas, New Delhi-110016, India Department of Applied Sciences, M S Institute of Technology, New Delhi - 110 058, India Department of Physics, Indian Institute of Technology, Hauz Khas, New Delhi-110016, India

2

1

1. Introduction As the concerns have increasingly grown about electromagnetic interference (EMI), the

electrical properties of several materials such as polymer composites. have been investigated for the purpose of shielding [l-31. Cement excels over other materials due to its inherent property of hardness and durability [4]. Compared with other polymeric composites, cement is less expensive, corrosion resistant, and has the flexibility for complex designs and is also an excellent binder. However, cement matrix as such lacks the ability to shield electromagnetic (EM) radiation. Hence, in order to obtain a durable and hard EM1 shielding material, cement filled with conducting graphite filler has been found to be useful.

Cement in its pure form is electrically non-conducting and transparent to EM radiations. In order to provide conductivity and shielding effectiveness, graphite filler is dispersed into the insulating cement matrix. The performance of such a composite system consisting of conducting particle embedded in an insulating matrix has been found to depend strongly upon its interfacial property, behavior of the effective conductivity and dielectric constant [5, 9-1 11. The white cement composite prepared in this work has improved over the cement composites studied earlier [4] and also over its polymeric counterparts. The incorporation of conductive filler to this material does not affect significantly the inherited mechanical properties of cement. The decisive properties for product application have been improved considerably through alteration of material and processing parameters.

2. Experimental The white cement matrix used in this work was procured from Birla Plus India. The

conductive filler was a graphite powder with particle size ranging from 10-20 pm, supplied by Graphite India Limited. The conductivity of graphite with density 1.75g/cm3 is 1.33 x lo5 S/cm. Cement and graphite were tumble mixed for 200 minutes to ensure the proper dispersion of graphite into cement powder. The resulting cemenvgraphite composite powder was compacted in a piston- cylinder assembly at 75 MPa for 15 minutes and subsequently treated with water. The disc shaped samples were of diameter 1.3 cm and thickness 2-3 mm. These pellets were cured in presence of 100% relative humidity for 50 hrs. Cement has an excellent property of hardening when treated with water. Its components react with water to form complex chemical compounds and settle on, which transforms the smooth powdery material into tough hard pellets. These specimen pellets were dried at atmospheric temperature (- 25 OC) for several days. A series of specimen samples were produced by varying the filler concentration from 0 wt% to 20 wt%. Five pellets were prepared for each composition. The data reported here is the mean value. Maximum departure of the observed data from the mean value was found to be less than 2%. This ensures a uniform dispersion of graphite into the cement powder.

The surfaces of the specimens were polished with sand paper to remove graphite rich layer and surface irregularities. To avoid atmospheric and humidity effects, the specimens were sealed in air-free polythene bags, prior to testing. For electrical measurements, a film of conductive silver was painted on both surfaces of the sample. The electrical measurements were made at room temperature by varying frequencies from 10 Hz to 10 MHz using impedance analyzer (HP 4 192 A). For composites containing more than 5 wt% graphite, this impedance analyzer could not be used due to their high conductivities. The density of compression-molded cement sample was found to be 2.42 g/cm3. There was a small decrease in density with increase in graphite concentration.

29

30

loo L 5 10 15 20

c

._

10'

Graphite (wPh) Frequency Hz

Fig. 1 Resistivity of cemenvgraphite composites as a fbnction of filler loading

Fig. 2 Dielectric constant of cemenvgraphite as a function of frequency

3. Results and Discussion The variation of direct current resistivity at room temperature as a function of graphite

concentration is illustrated in Fig. 1. The resistivity decreases rapidly by several orders for the sample with 2 wt% graphite content, suggesting this as the percolation threshold. At this filler content conducting pathways of graphite particles across the composite are close enough to allow the electrons to hopltunnel across the thin gaps between them [ 1,6]. The drastic change in electrical resistivity is followed by saturation at - 1Owt% of graphite contents. For composition higher than 10 wt %, the resistivity flattens when almost all graphite particles are assumed to become part of network of conductive pathways. At 20 wt% concentration, the graphite forms an unbroken conductive near perfect network through which electrons are moved [6, 71. This composite has exhibited electrical resistivity of 3.65 Q-cm.

The dependence of dielectric constant in 10 Hz - 10 MHz frequency range for cement composites with 0, 2, 4 and 5 wt% graphite are shown in Fig. 2. The sample with 5 wt% graphite exhibits a dielectric constant of lo9 at low frequency and decreases with increasing frequency. The dielectric constant of the samples comprising of 2 wt% and 4 wt% graphite content are close to each other and the two curves overlap after 100 KHz. The extremely high dielectric constant of 5wt% as compared to other compositions is a direct result of interfacial polarization [S]. There is extremely small or almost negligible variation in the dielectric constant of the material with 0% graphite. This shows that the frequency dependence of the dielectric constant is generated primarily due to the incorporation of the conductive filler, which in turn causes structural inhomogeneities causing interfacial polarization [S].

The real part of the complex permittivity of cementlgraphite composite as a function of frequency is shown in Fig. 3. As the frequency is increased, the relative difference between the dielectric constants of various compositions are lowered, i.e., their frequency dependence is decreased. The reason for this decrease in E' and also a frequency independent behavior at high frequency range can be accounted for by Maxwell-Wagner-Sillars theory [12-141. MWS theory is a classical theory used to explain the dielectric behavior due to the interfacial polarization of two-phase systems when the volume fraction of dispersion is small [ 151.

31

- _ - 10' t ' """" ' """" '"'"'i ' """" ' """" ' """" ' ,-"I

l o o 10' l o z 10' 10' 105 l o 6 10'

F req u en cyH z Graphite conc (wt%)

Fig. 3. E' of different compositions of cementlgraphite Fig. 4. Variation of D Shore hardness with graphite concentration as function of fiequency

The frequency dependence of the dielectric constant is generated here exclusively by the addition of conductive fillers [ 5 ] . The larger numbers of interfaces between graphite and cement particles contribute to the improvement of the dielectric properties of the composite, thereby increasing their dielectric constant. But this phenomenon is applicable more at lower frequencies compared to the higher frequencies. As the frequency is increased, the interfacial charges are no longer able to polarize sufficiently rapidly so that their oscillations begin to lag behind those of the field. As a result, the frequency dependence of E' is lowered.

At about 10 KHz, a cross-over point between the curves of 2 wt% and 4 wt% of graphite content is observed, i.e., after a certain frequency, 2 wt% shows higher dielectric constant as compared to 4 wt%. This can be ascribed to the reason that the E' could be occurred due to two independent phenomenon in the cemenvgraphite composite, firstly due to dipole polarization in the dielectric cement, secondly due to interfacial polarization at the interface between the graphite and the cement powder [5]. The effect of the interfacial polarization by graphite decreases with increasing frequency, while the dipole polarization by the dielectric powder is not changed. Hence at higher frequency, the E'of 2 wtY0 becomes higher because of its larger content of dielectric material.

An investigation on the hardness of the samples was carried out using Shore-D hardness tester. Fig. 4 illustrates the variation of hardness with filler loading. A random change in hardness, with slight decrease at high filler concentration has been noted. Such a small decrease in hardness with the increase in graphite content indicates that the graphite has not prevented much the inter- particle bonding of cement. A random variation in density from 2.42 g/cm3 to 2.3 g/cm3 was also noticed with increasing graphite content.

4. Conclusion The dielectric constant of cemenvgraphite composites increases with the increase of graphite

concentration and a maximum value of lo9 has been obtained for 5 wt% composite. Dielectric constant decreases with increasing frequency. E' behavior as a function of frequency has been attributed for by Maxwell-Wagner-Sillars theory. Graphite (1 0-20 pm) is a capable admixture for lowering the resistivity of white cement. Resistivity decreases by several orders as the 2 wtY0 graphite concentration is reached. At graphite concentration above 10 wtY0 resistivity of - 4 ohm- cm has been obtained.

References:

[l] V. K. Sachdev, Varij Panwar; H. Singh, N.C. Mehra and R. M. Mehra, Phys. Stat. Sol. (A), 203, 386 (2006),.

32

[2]

[3]

[4] [5]

[6] [7] [8] [9] [lo] M. E. Achour, M. El Malhi, J. L. Miane and F. Carmona, J. Appl. PoZym. Sci., 61, 2009

[l I] N. C. Das, D. Khastgir, T. K. Chaki and A. Chakroborti, Part A: Appl. Sci. Manufacturing,

[ 121 J. C. Maxwell, Electricity and Magnetism (Oxford University, London, 1892), Vol. 1. [13] K. W. Wagner, Arch.,J. Electrotech., 2,371 (1914). [14] R. W. Sillars, J. Inst Electr Eng., 80, 378 (1937). [ 151 L. K. H. Van Beek, Prog. Dielectr., 7,69 (1 967).

V. K. Sachdev, N. K. Shrivastav, Varij Panwar, H. Singh, N.C. Mehra and R. M. Mehra, Phys. Stat. Sol. (A), 203,3754 (2006),. V. K. Sachdev, N. K. Shrivastav, Kamlesh Kumar and R. M. Mehra, Mat. Sci-Poland, 23, 269 (2005). V.K. Sachdeva , R. Chatterjee, and Rajendra Singh [ ICMAT 20071 Kyoung Sik Moon, Hyung Do Choi, Ae Kyoung Lee, Kwang Yun Cho, Ho Gyu Yoon and Kwang S. Suh, J. Appl. Polym. Sci., 77, 1294 (2000). R. Strumpler and J. Glatz-Reichenbach, J. Electroceramics, 3, (4), 329-346 (1999). K. T. Chung; ,A. Sabo and A. P. Pica, J. Appl. Phys., 53,6867 (1982). A. Paul and S. Thomas, J. App. Polym. Sci., 63,247 (1997). Christen Brosseau, Patrick Queffelec and Philippe Talbot, J. Appl. Phys., 89,4532 (2001).

(1 996).

31, 1069 (2000).

Electromagnetic Interference Shielding of Graphite Filled Cement Composites in Relationship to Resistivity and Morphology

V.K. Sachdev ', R. chatterjee and Rajendra singh 'Department of Applied Sciences, MS Institute of Technology, Janakpuri, New Delhi - I I0 058,

India 2Department of Physics, Indian Institute of Technology, Hauz Khas, New Delhi, India

'Defence Research Development Organization Bhawwan, Rajaji Marg, New Delhi, India

1. Introduction The environment is increasingly sensitive to electronic pollution by way of rapidly growing

applications of advanced electrical and electronic devices which emit electromagnetic (EM) radiations. Human body can be adversely affected by these radiations. EM pulse may disrupt neighboring equipments such as computers and devices on board. Electromagnetic interference (EMI) of such unwanted waves is of concern even up to GHz frequency. Hence, in present scenario, the ability of a building to shield EM radiation is of greater significance particularly for housing electronics and electric power plants. Besides, it is essentially required to avert electromagnetic forms of spying.

Reflection and absorption are the ways for solving EM1 problems. In reflection dominating shielding, metals are used due to their high conductivity. The conducting shielding can cause malfunction by acting as an antenna. Also there is a likelihood of electric leakage due to miniaturization of equipments and circuits. Thus absorption dominated shielding materials are considered to be safer for solving EM1 issues. Study of absorbing materials has been on increase [ 1-41 due to development of radar, microwave communication technology, stealth (self concealing) technology and demand for microwave darkrooms and anti-EM1 coatings. Cement has been used as a structural material in buildings in large volumes since it is inexpensive and excellent binder. Tough hard state can be easily achieved by proper curing with water. But cement matrix lacks in ability to shield EM radiation. Using cement paste with solid graphite admixture, shielding effectiveness (SE) of 22 dB at 1 GHz has been reported in literature [5] . Through this method, graphite in cement matrix appears to be less effective for SE compared to cement paste containing stainless steel fibers [6] or coke powder [7]. Moreover these shielding measurements are confined only to 1 and 1.5 GHz frequencies and primarily by reflection. Work presented here is related to absorption dominated EM1 SE for cement/graphite composites in the X-band frequency range (8-12 GHz) beside measurement of return loss (RL) parameter. SE up to 52.5 dB has been achieved. Though this method of processing is unconventional but has the added advantage for achieving the desired rise in values of dielectric properties as a result of Maxwell-Wagner effect.

2. Experimental Graphite powder with 10-20 pm particle size was supplied by Graphite India Ltd. Its

resistivity and density were 7 5 ~ 1 0 - ~ ohm m. and 1.75 x103kg/m3 respectively. Particles of graphite were of flakes shape [S]. The process of preparation of cemendgraphite composites is almost the same as used earlier [8, 91. The powders of graphite and cement were tumble mixed for 150 minutes, compacted at 105 MPa for 15 minutes in a die and then treated with water. The samples were cured in 100% relative humidity for 100 hours and dried at room temperature (-25 OC) for several days. The setting period stabilizes the harden state. A series of cemendgraphite conductive composites specimens with graphite contents of 0-20 wt % were processed. For each composition five samples were prepared. The data in this work represent the mean values. The thickness of all the composites samples were between 0.18 - 0.24 cm.

For electrical conductivity measurements of cemedgraphite composites the procedure and method are the same as used in past [8, 91. Measurements of EM1 SE and RL were made in the frequency range of 8.0-12.0 GHz using WILTRON vector network analyzer of 40 MHz - 20 GHz frequency range. Surface morphologies of these composites were examined by scanning electron

33

34

1 O ’ O . 1 o8 .

h

E

.- E

.z l o 4

2 l o 2 .

$ l o 6 . v

-w cn cn .-

c . . I 0 5 10 15 20

Graphite conc (wt%)

I U

*

4 Gr20wt%

8 9 10 11 12

Frequency (GHz)

Fig. 1 Variation of resistivity in cement composites Fig. 2 SKfin cement/graphite composites as a function of graphite content as a function of frequency

microscope (SEM) JEOL JSM 840. Samples Freeze fractured in liquid nitrogen were investigated.

3. Results and Discussion Fig. 1 shows variation of electrical resistivity of cementlgraphite composites versus graphite

content at room temperature. The nature of the curve is similar to that quoted in the literature for other such composites [8]. Near percolation threshold (-5 wt%) the graphite particles are close enough to allow electrons to hop across gaps between them and the drastic change in the electrical resistivity of several orders believe to be originated due to formation of conducting pathways of graphite particles across the composite. For composition higher than that of the critical graphite fraction (-10 wt %), the resistivity flattens when almost all graphite particles are assumed to become part of network of conductive pathways. At 20 wt% concentration, the graphite forms a continuous conductive near perfect network through which electrons are moved. This composite has exhibited electrical resistivity of 4.2 R-cm.

The measured values of SE for cement composites (contain 0-20 wt% graphite contents) increases with increasing graphite. At 15 wt% graphite an increase of 52.5 dB at 11.5 GHz was registered. A little frequency dependency of SE has been noticed in the range of X band frequency for different graphite contents except in sample with 20 wt%. The weak frequency dependence indicated that the graphite was well distributed in the cement matrix.

Parameter RL is related to the conductivity of composites. Contribution due to reflection of electromagnetic energy in shielding effectiveness can be calculated using measured values of RL. Composites having high value of RL used to show low reflection coefficient (10). For composites with 3-20 wt% graphite, frequency dependency is found to be less than - 3 dB. In composite with 0 wt% graphite concentration RL varies from 3-9 dB over 8-12 GHz frequency range. Maximum value of RL - 3.5 dB at 11.5 GHz for composite with 5 wt% graphite has been observed.

The radiated electromagnetic energy mainly attenuated by absorption and reflection. Hence

And the relation of SE,f with reflectance is expressed as [ 1 13

R stands for reflectance and is the ratio of reflected power density to incident power density in

SE = SE,f + SEabs in (dB) (1)

SE,,f= lologlo (1-R) in (dB) (2)

R = P, / Pi = Anti log (-RL/lO) (3) case of normal incidence using Fresnel equation

35

100 50

t G r OWL 1 95 +Gr 3wt%

Gr 5wt% 40- '- t / t Gr low% I

30- 4 t G r 2 0 w t % F a

4 6 ~ r i ~ w t % i n 9 0 -

v p ---+ 5 8 5 -

W m - p 20 -

u)

2 y 80 2

75 lo- * * * d. & * &

o - " " ' " " " " 8 9 10 11 12 13 14 70

SEref is displayed as a function of frequency in Fig. 2. Curves exhibit frequency dependency of less than 3 dB for different filler concentrations except in case of 20 wtYo graphite composite. In 20 wt% graphite composite maximum variation in SEr,f is noticed only at higher frequency range of 10 - 12 GHz. Fig. 3 illustrates the calculated values for SEab, for these composites as a function of frequency for various graphite concentrations. An abrupt increase can be seen on inclusion more than 5 wt% graphite content. Maximum SEab, of - 42 dB at 9.5 GHz for 15 wt% graphite composite has been noticed. A very small frequency dependency can be seen in of all composites except in 20 wt% graphite. The confused dependency of SEab, in 20 wt% graphite indicates the poor reliability of sample of this composition.

This investigation clearly suggest that SE of these cemendgraphite composites is absorption dominated. During processing graphite get encapsulated with cement matrix (an insulating material) assumed to show dielectric property as reported in case of carbon black (12, 13). Microwave absorption here is related to interfacial or Maxwell-Wagner type of polarization used to occur in heterogeneous dielectrics where one component has a higher conductivity than the other, in which dipole can be induced by an electric field (14).

Figure 4 illustrates the Shore-D hardness of cementlgraphite composites as a function of graphite content. On inclusion of 10 wt % graphite - 10% decrease in hardness has been noticed. Such a decrease in hardness with the increase in graphite content indicates that the graphite has not prevented much the inter-particle bonding of cement. The anticipated decrease in mechanical properties may be sorted out by optimization of material and processing parameters. Recently by means of this we have been able to process cementlgraphite composites with - 74 dB SE with a negligible loss in mechanical properties. These results will be published soon elsewhere.

Cementlgraphite composite samples were fractured at cryogenic temperature. Fractured surfaces were examined using SEM. This study reveals that the increasing graphite content is important in the process of compaction at least up to the levels of present investigation. As seen in Fig.5 (a) - 5(c) addition of 3 to 5 wt% graphite to cement matrix reduces the sizes of slate like masses followed by stone like structure in the form of lumps of particles due to agglomeration. Fig. 5(d) suggests that the stone shape agglomerates are composed of cement matrix intermix with graphite. Their fine contacts have believed to flatten the resistivity in higher compositions as noticed in Fig. 1. Sharp increase in SE in the region above 5 wt% graphite filler understood to be due to a rise in the value of dielectric properties as a result of increase in Maxwell - Wagner effect. This is as a result of direct consequences of structural properties and complexity of agglomerates noticed in SEM micrographs in Fig. 5(c) and 5(d)

~~ - ~

I

- 4

-

m

-

' " " . ' . ' "

.

Fig. 3 SE,b, in cemedgraphite composites as function of frequency content

Fig. 4 Shore-D hardness as a function of graphite

36

. .

Fig. 5 SEM micrographs at magnifications (1000 X) for cemenugraphite composites (a) 0 wtY00; (b) 3 wt%; (c) 5 wtY0; (d) 10 wt??

4. Conclusion An absorption dominated shielding effectiveness of 52.5 dB (at 11.5 GHz) with a cement

composite (- 2 mm thickness) that contains a solid graphite content of 20 wt% has been achieved compared to reported [S] reflection dominated 22 dB (1 GHz) of - 4 mm thickness cement composite at a solid graphite content of 0.92 vol.%. Graphite (10-20 pm) is quite effective admixture for lowering the resistivity of cement. Resistivity decreases by several orders as the 5 wt% graphite concentration is reached. Morphological changes in SEM micrographs with adding graphite reveal its importance in process of compaction. Sharp increase in SE in the region above percolation concentration is due to a rise in value of dielectric properties as a result of increase in Maxwell-Wagner effect.

Acknowledgements This research was financially supported under UGC project No. F. 10-67 / 2001 (SR-I), Govt.

of India. The author would like to thank Graphite India Ltd. Bangalore for providing graphite powder. Useful discussions with Prof. DDL Chung, State University of New York at Buffalo, USA is gratefully acknowledged.

References [l] [2]

[3] [4]

[5] [6] [7] [8]

K. Y. Kim, W. S. Kim, Y. D. Ju , and H. J. Jung, J. Muter. Sci., 27,4741 (1992). S. P. Rum, B. K. Xu, H. Suo, F. Q. Wu, S. Q. Xiang , and M.Y. Zhao, J. Magn. Magn. Mater., 212, 175 (2000). G. Li, G. -G. Hu, H. -D. Zuo, X. -J. Fan, andX. 4. Li, J. Appl. Phys., 90, 5512 (2001). M. R. Anantharaman, S. Sindhu, S Jagatheesan, K. A. Malini, and P. Kurian, J. Phys. D: Appl. Phys., 32, 1801 (1999). J. Cao, and D.D.L. Chung, Cem. Conc. Res., 33, 1737 (2003). S. Wen, and D.D.L. Chung, Cem. Conc. Res., 34,329 (2004). J. Cao, snd D.D.L. Chung, Carbon, 41, 2433 (2003). V. K. Sachdev, R. M.Mehra, andN.C. Mehra, Phys. Stat. Sol., (a) 201 2089 (2004).

37

[9] V. K. Sachdev, N. K. Srivastava, and R. M. Kumar, Mater. Sci. Poland, 23,269 (2005). [ 101 N.C. Das, D. khastgir, T.K. Chaki, and A. Chakraborty, J. Elast. Plast., 34, 199 (2002). [ 111 Y. K. Hong, C. Y. Lee, C. K. Jeong, D. E. Lee, and K. Kim, J. Joo, Rev. Sci. Ins., 74, 1098

(2003). [12] K. S. Moon, H. D. Choi, A. K. Lee, K. Y. Cho, H. G. Yoon, and K. S Suth, J. Appl. Polym.

Sci., 77, 1294 (2000). [13] S. K. Kwon, J. M. Ahn, G. H. Kim, C.H. Chun, J. S. Hwang, and A.H. Lee, Polym. Eng. Sci.,

42,2165(2002). [ 141 A. Paul, and S. Thomas Suth, J. Appl. Polym. Sci., .63, 247 (1 997).

38

Preparation and Optical Characters of Polymer Composite DR13PMMA Films

Y Gao ', Q Ren I , F J Zhang I , X F Cheng2, J Sun2, H L Yang', X B Sun', L Feng' 'Department of Optics, School of Information Science and Engineering, Shandong University,

Jinan 250100, People's Republic of China State Key Laboratoiy of Crystal Material, Shandong University, Jinan 2501 00, People's Republic

of China

Abstract: The guest-host polymer DR13/PMMA thin films at three different concentrations (1 Owt%,

15wt% and 20wt%) were studied in the present work. The films were prepared by the spin-coating method and some optical characteristics of the films such as refraction index, thickness, absorption spectrum, and transmission losses were investigated.. Key words: polymer; optical constants; transmission losses

1. Introduction Polymeric second-order nonlinear optical @LO) materials have been studied for more than 20

years. During this period, a wide variety of polymeric systems have been formulated and investigated due to their promising applications in information processing and fiber communications [ 11. For practical devices, NLO polymers must possess large molecular nonlinearity and optical transparency at the telecommunication wave band, as well as good thermal and chemical stability. The intrinsic low dielectric constant, quick response, good processibility into thins films and compatibility with microelectronic processes are some additional advantages of poled polymers.

Disperse Red 13(DR13)/poly-(methylmethacrylate) (PMMA) is a guest-host polymer system.

The linear electro-optic coefficient y33 of poled guest-host polymer DRl3/PEK-c thin films was

found to be about 25.5 pm/v [2]. We can see that guest material DR13 molecule has relatively large second-order nonlinearity and host material PEK-c has little effect on the electro-optical coefficient, so we choose DR13/PMMA as our research object (because PMMA is purer and more commonly used than PEK-c) to investigate some optical properties. The guest chromophore molecule DR13 is a highly conjugated molecule and a widely developed azobenzene dye, which has advantages such as good dissolvability, ease of composition with polymer, and large light-induced anisotropy. The molecular structure of DR13 is shown in Fig.1 (a) [2]. The host polymer PMMA is a highly transparent polymer whose glass transition temperature (Tg-l OSOC). It has also other advantages, such as, excellent compatibility with guest NLO organic molecules and no strong interaction with organic molecules. The molecular structure of PMMA is shown in Fig.1 (b) [3].

2. Procedure First, DR13 and PMMA were dissolved in 1, 2-dichlorethane, respectively. After the guest

DR13 and the host PMMA were completely dissolved, the two solutions were mixed together. Ultrasonic stirring was used to ensure uniform incorporation of DR13 into the mixture. Three samples with different weight ratios of guest-host DR13/PMMA solutions of 10, 15 and 20wt%) were prepared. Then the mixed solutions were filtered through a syringe with a 0.45 pm filter in order to remove the undissolved materials and impurities. The final solution was spin-coated onto the indium-tin oxide (ITO) glass substrates (n=1.51637 at h489.3nm) to form thin films. The films

39

were left in the atmosphere for one day and then baked at 60°C for one hour to remove the residual solvent.

The solution was also spin-coated on silicon flakes in order to measure the refractive indexes and the thickness of the films. The m-line quasi-waveguide method and SPA4000 Prism-Film Coupler were used to measure these parameters. The absorption spectrum of the film was studied by a Spectrophotometer U-3500. Transmission losses of the films were investigated by photographic technique. The results were processed by an original computer program.

3. Results and Discussions 3.1 Refractive index and thickness of the films

The m-line quasi-waveguide method [4-71 was used to measure the refractive indexes and thickness of the films. The quasi-waveguide method has advantages such as convenience, accuracy and it only refers to angular measurement. The film was spin-coated on to silicon to form a one- side leaky-waveguide (the surface of the silicon was oxidized, the refractive indices of SiOz are 1.4569 at 632.81~11, 1.4466 at 1310.0nm and 1.4437 at 1550.0nm, which is lower than that of DR13PMMA film). The theoretical setup of the m-line quasi-waveguide method is shown in Fig.2. Fig.3 shows the m-line in the present study.

In order to obtain accurate results, these parameters were also investigated with SPA4000 prism-film coupler. The results are shown in Table I . It can be seen that at the same wavelength, the refractive index becomes greater as the weight ratio increases, while the thickness of the film decreases. Because a greater weight ratio means that there is more DR13 or less PMMA in the film, PMMA plays a more important role in determining the thickness of the film than DR13. However, when the wavelength of the input laser is increased, the refractive index of the same weight ratio is reduced, and the mode numbers in the films are also reduced.

The spectral dependence of the refractive index n of DR13PMMA is shown in Fig.4. The dispersion of the refractive index was fitted by the Cauchy dispersion formula [S-101

B C n = A + - + -

l2 /I4 where il is the wavelength (nm); A , B and C are Cauchy coefficients. The values of A , B and C were obtained by fitting the refractive indices of the 1Owt% film at different wavelengths. It was found that A =1.45396, B =1.1429x104 and C =-2.4221x109.

3.2 Absorption of DR13/PMMA films The absorption spectrum shows the possibility of practical application. The absorption spectra

(200 nm-2000 nm) of the three films were measured. Fig.5 shows the absorption spectrum of 20wt% film. The other two samples have similar results to the 20wt% film. It can be seen from the figure that there are two large absorption peaks at about 300 nm and 500 nm and no absorption peak in the telecommunication wave band, so this material can be used potentially in telecommunication devices. 3.3 Transmission losses

Due to scattering, absorption of the materials and the surface quality of the film, transmission loss in the film waveguide is one of the key elements in determining electro-optical film materials for practical applications. The precise evaluation of the propagation characteristics of film optical waveguides will provide reliable basis for the design of electro-optic devices. An experimental apparatus for measuring transmission loss in thin film using an imaging technique was set up [ll-141. This experimental setup is shown in Fig.6.

40

Fig.7 shows the streak image taken by a digital camera. It can be seen from the figure that the laser intensity decreases when the light propagates in the film, which can be described in the form of exponential attenuation:

('g 'X - 'g '0 ) (dB / cm) x-xo

where x is the distance between the coupling pot and investigating pot, Z, is the light intensity at

the input coupling point while I, is the light intensity at the investigating pot, a is the

attenuation coefficient and L is the transmission loss. Digital photos were then transmitted to a computer. A computer program with Borland C++ to process the images was produced, which was first been processed by Photoshop, and then the optical intensity of each point in the image was obtained. The three-dimensional profile of the scattered intensity versus the pels is plotted in Fig.8.

The x-axis as the transmission direction of laser in the film, y-axis as the direction perpendicular to the transmission direction of laser in the film and z-axis as the direction perpendicular to the surface of the film were chosen. The two-dimensional scattered intensity profile along x-axis was obtained by accumulating the scattered intensity of y-axis. Fig.9 shows the profile of Logarithm scattered intensity versus propagation length. Using the paper scale adhering to the glass substrate when taking photos, the pels can be converted into real length. In order to get the

values of transmission losses, Fig.9 was processed by linear fitting, the fitting formula is y = A + Bx . From Eq. (2), the value of the transmission loss was obtained. The transmission losses of the three DR13/PMMA samples are 1.5269dB/cm, 2.7601dB/cm, and 3.6291dB/cm, respectively. It can be seen that the transmission loss increases when more DR13 is added into the polymer PMMA.

4. Conclusions Because DR13/PMMA thin film has a relatively large electro-optical coefficient, some

characteristics of this kind of polymer thin films, such as refractive indices and thickness of film, absorption spectrum in the range from 200 nm to 2000 nm and the transmission losses of the films, were studied. Refractive indices of the three films at three different wavelengths (632.8nm, 1310.0nm and 1550.0nm) were investigated. The refractive index increases with increasing weights ratio ofDR13/PMNA. The thickness of the spin-coated film is slightly larger than 1 pm. There is no distinct absorption peak, except for the two peaks at 300 nm and 500 nm. Because there is no absorption in the telecommunication waveband, this kind of polymer thin film waveguide has potential applications in telecommunications. The program used to process the transmission loss is simple, fast and accurate. From the data, it can be concluded that DR13PMMA is a promising material for industrial application such as an all-optical waveguide switch.

Acknowledgements

(Grant No. 60377016,60476020) and the 863 National Plan (Grant No.2002AA313070) of China. The authors acknowledge the financial support of the National Natural Science Foundation

41

References [ 1 ] D. M. Burland, R. D. Miller, and C. A. Walsh, Chem. Rev., 94, (l), 3 1 (1 994) [2] W. Shi, X. Yin, Changshui Fang, et al., Opt. andLus. Eng., 35, (2), 121 (2001) [3] K.D.Singer, M.G.Kuzyk, W.R.Holland, et al. Appl. Phys. Lett., 53, (19), 1800(1988) [4] R. Ulrich, J. Opt. Soc. Am., 60 (lo), 1337 (1970) [5] R. Ulrich and R. Torge, Appl. Opt., 12 (12), 2901 (1973) [6] Tie-Nan Ding and E. Garmire, Appl. Opt., 22 (20), 1 (1983) [7] P. K. Tien, R. Ulrich and R. J. Martin, Appl. Phys. Lett., 14 (9), 1 (1 969) [8] M. Born, and E. Wolf, Principle of optics, Elmsford: Pergamon, 1983, p. 84 [9] Y. Kim, J. Mater. Sci., 35, 873 (2000) [ 101 S. A. Khodier, Opt. Las. Tech., 34, 125 (2002) [ l l ] Y. Okamura, S. Yoshinaka, and S. Yamamoto, Appl. Opt., 22 (23) 3892 (1983) [12] E. A. Arutunyan, S. K. Galoyan, Opt. Comm., 57 (6), 391 (1986) [13] Y. Okamura, S. Sato, and S. Yamamoto, Appl. Opt., 24 (l), 57 (1985) [14] H.L. Y., Q.Ren, S.Y. G.uo, and G.H. Zhang. Opt. Las. Tech., 35 (4),291(2003)

(a) (b) Figure 1 Molecular structure of DR13 (a) and PMMA (b)

He-Ne Laser

Substrate 3

Figure 2 Theoretic setup of the m-line quasi-waveguide method

Figure 3 The m-line of the film of DR13FMMA

42

_.

2.0

1.5

s 1.0

B .

- : . c

e a

0.5

0.0

Figure 4 Variation in refractive index with wavelength for 1O%wt DR13PMMA film

-

-

-

-

-

I

0 m 400 6MJ 800 1oM3 1200 14w 1m 1800 MOO 2ZM

wavelength(nm)

Figure 5 Absorption spectrum of 20wt% films

He- ubstrate IT0

rn Computer Digital Camera

Figure 6 Experimental setup of imaging technique

Film

Figure 7 Image of scattered streak

43

pmg%a@J n length @

Figure 8 Three-dimensional profile of scattered intensity

transmission loss - linear fit of y=A+Bx

3.6

.. A

$ 3.0 - m - - - . - g 2.8 - -

2.6 -

2.4 -

0 50 1W 150 2w propagation length(image cell)

Figure 9 .Transmission loss curve of 20%wt DR13PMMA thin film

Table I Refractive indexes and thickness of films With different weight ratio at three

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

Electrical and Computer Engineering, James Cook Universiq, Townsville, Australia Email: Mohan.Jacob@,icu.edu.au

Abstract The competence of integrating circuits within LTCC materials is an added advantage to

implement the material in electronic circuits. Since the circuits are buried inside the material, it is not easy to tune the circuit if the required performances are not achieved. Therefore precise knowledge of microwave properties of Low Temperature Co-fired Ceramic (LTCC) materials is crucial before designing the circuit for the efficient design of microwave systems, especially for design of communication filters. Many communication devices operate at low temperatures. In order to implement LTCC at low temperature communication circuits and devices, engineers need precise values of the complex permittivity. The aim of this paper is to characterize LTCC material as a function of temperature at microwave frequencies using a very precise measurement technique, Split Post Dielectric Resonator. DuPont Green Tapes were characterized in the temperature range of 20 K - 300K at a frequency of 9.5 GHz.

Keywords: Dielectric Properties; LTCC; Substrates

1. Introduction The last decade witnessed a dramatic increase in mobile and wireless communication systems.

The multiple functions available in devices become multifold however the size of the devices becomes smaller. Customer’s desirable requirements are typically pocket size, light weight, low power and low cost. These requirements triggered a revolution in the development of wireless communication systems which demand better performance than earlier systems in a multipath environment [ 11. The development in this area assisted customers to be in touch with the business, family or world events when ‘on the run’ [2,3]. The demand for accessing new applications at high data transmission rate introduces new challenges to wireless systems and hand held transceiver designers. As the radio equipment has to operate under varying ambient environment, this creates a challenge to material scientists and engineers to develop materials with low loss, relatively high permittivity and stable performance.

I I

Heraeus CT2000

Motorola T2000

Emca Tat300

Dupont 951

Dupont 943

Fro A6

0 2 4 6 a 10

Fig. 1 : Permittivity of some planar dielectric materials used in microwave components [2,3].

44

45

Emca T880B

Dupont 951

Heraeus CT2000

Motorola T2000

Dupont 943

Fro A6

0 0.002 0.004 0.006 0.008

Fig. 2 : Loss tangent of some planar dielectric materials used in microwave components [2,3].

Figures 1 and 2 show room temperature dielectric properties of the commercially available LTCC materials [3]. LTCC materials are seen as very promising materials for the next generation of handsets due to the mutilayer capability and prospects for a significantly decrease in size of passive components. Low Temperature Co-fired Ceramic (LTCC) technology allows integration of these passive components in electronic systems, as a result the size reduction. Three dimensional circuits that can be developed by the integration of passive components and layers required for the device can be now processed in parallel, reducing the production cost and time. Lower temperature firing of ceramic blocks allows the utilization of highly conductive metals such as gold or silver and decrease in line-loss [ 1,4-61. As a result of this progress, a rapid growth of applications of LTCC in wireless communications has been observed during the last few years. The parallel processing, stable performance and high density of interconnects are advantageous [5], but the poor mechanical strength prevents it from use in mobile phones [6].

We have developed a Split Post Dielectric Resonator [7], which can be cooled down to temperature of 20 K. In this paper we have characterized four LTCC materials of relative permittivity between 6.8 and 9.2 as a function of temperatures (from 25 K to 290 K) at 9.5 GHz.

2. Experiments

Split post dielectric resonator is an accurate measurement technique for the characterisation of planar materials. More details of this technique are available in the literature [7, 81. In our measurements, TEols mode was used since this mode is insensitive to the presence of air gaps perpendicular to z-axis of the resonator. The complex permittivity was calculated based on the rigorous electromagnetic modeling of the split post resonant structure using the Rayleigh-Ritz technique [7, 81. A computer program was developed for the calculation of complex permittivity. The real part of the sample’s complex permittivity is computed from measured resonant frequencies of the resonator using the following equation [8]:

where: h is thickness of the sample under test, f, is the resonant frequency of the empty SPDR, f, is the resonant frequency of the resonator with the dielectric sample. KE is a function of E: and h, and has been evaluated for a number of of E; and h using Rayleigh-Ritz technique. Iterative procedure is then used to evaluate subsequent values of K, and E: from equation (1).

The loss tangent of the tested substrate is calculated from the measured unloaded Q,-factors of

46

the SPDR with and without the sample based on:

where pes is electric energy filling factor of the sample, QDR-' and Qi ' denote losses of the metallic and dielectric parts of the resonator respectively.

The measurement system used for the microwave characterisation of the LTCC materials sample is shown in Fig. 2. The system consisted of Network Analyser (HP 8722C), closed cycle refrigerator (APD DE-204), temperature controller (LTC-lo), vacuum Dewar, a PC and the Split- Post dielectric resonator.

For variable temperature measurements, S-parameter data sets were measured first for the empty resonator and then for the resonator with a given LTCC sample. To obtain precise values of the Qo-factor of the split-post resonator and hence accurate values of tan6 of LTCC substrates, we have measured many points around the resonance and processed measured data sets using the Transmission Mode Q-Factor Technique [9]. The TMQF technique was then used to obtain f, and Q,, values of the empty split post resonator and the resonator with the LTCC sample, at exactly the same temperatures. The microwave parameters cr and tan6 were computed from the resonant frequencies and unloaded Q,-factors respectively using a software developed.

3. Results and Discussion

We have used two types of LTCC DuPontTM GreenTapeTM. The Split Post Dielectric Resonator containing the dielectric material is cooled down from room temperature to 15 K. The resonant frequency of the resonator is identified (TEoll) to be about 9.5 GHz. The resonant frequency and Q-factor is estimated as a function of temperature from the measured S-parameters. The complex permittivity is calculated as described in the previous section. Figure 3 shows the real part of complex permittivity of these two samples. The change in relative permittivity is 3.3% and 4.6% in the temperature range 15 K - 290 K. The uncertainty in measurement is less than 0.5%.

7.7

x .- > 7.6 t:

c. .- .- € a" 7.5

s 5

X a, -

7.4

.4-

O 7.3

Q - 2 7.2 of

7.1 0 50 100 150 200 250 300

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

47

1 o-z

c C a, m

I- v) v) 0 1

S 10‘~

1 o4

I I I I I

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

I I I I I

0 50 100 150 200 250 300

Temperature (K) Fig. 4: The loss tangent of the GreenTapesTM at 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 1 O* 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.55Fe204 and 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 co- existence 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 I 30 40 50 28 60

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

50

I

Ferrite--------,

-0.20 3 400 600 800 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

0 1000 >

f ir 100

10 100 1000 1 OOO(

Freauencv (kHz)

100

0) v )

g10 P

1 100 1000 1 OOO(

Freauencv (kHz) 100 1000 1 OOO(

Freauencv (kHz)

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

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

1000

- d

c

5. 1. .- ;loo a,

!i a, a

10 100 1000 1 oooc

Frequency (kHz)

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

~

-A- XzO.5 - XzO.75

-1

- -2 5.

m P t m -3 0 -I

100 1000 1000 Frequency (kHz)

I

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

[3] A. Bush, Y.K. Fetisov, K.E. Kamentsev, V.F. Meshcheryakov, and G. Srinivasan, J. Mug. Mug.

[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

[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).

(2002).

Muter., 258-259,4547 (2003).

(2006).

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

v) 45 ’ 40 e In

2 35 30

Figure 1 Hardness test

60

50 h 5 c 40 0 .- .- 8 30

2 z 20 c

10

0 0 1 2 3 4 5

1- BF +BC - - - - W F +WC +- TF +TC I Filler content (phr)

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

54

0 1 , I

0 1 2 3 4 L+- TF -A-TC - C - WF - 0 - W C BF +BC killer content (phr) -. ~. ~-

Figure 3 Tensile strength at 100% elongation

12 1 I

0 1 2 14- TF--O- -WF+WC-BC/

3 4 Filler content (phr)

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

F v

z a, m m 4- - P

:: m

Q a,

f!

m E ?! a

20

19

18

17

16

15 1 I I , 0 1 2 3 4

Figure 5 Average breakdown voltages of the samples

55

0 1 2 3 4

TC TF +WC - f WF +BC 4 BF I Fillercontent(phr) _. __ I '-

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,

[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.

pg8 12-825.

Session P3

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, 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 Hashin- Shtrikman (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,

(4) PO - P )

0 0 3 '

1 1

IB(n)dn = p and JnB(n)dn = ___

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 ) = {LJY 4m n , < x < n 2 with = ' ( 1 3 + p T 2 4 m ) (6) otherwise

The EMT implies that a spread in the demagnetization factors appears as a result of interactions between inclusions.

For the Bergman-Milton theory to be applied to actual composites, data on the spectral function are needed. Several approaches to determining the spectral finction are known. First, the spectral function may be calculated directly from known microscopic structure of composite. However, both determination of the microstructure of an actual 3d composite sample and accurate calculation of the spectral function from these data are difficult problems, and there is no reports on successll application of this approach in the literature. Second, the spectral function may be derived from known material constants of the inclusions and measured data on the effective material constant of the composite. This technique has been applied recently to the case of the IR permittivity of dielectric mixtures [8]. And finally, a representation of B(n) as a known function of several unknown parameters may be suggested with further search for these parameters by the best fit of the predictions of the theory to the measured data. This approach is suitable for occasions when the data on bulk material constants of inclusions are not available. It has been attempted to both IR permittivity of dielectric mixtures [9] and microwave permeability of magnetic composites [ lo] .

For inclusions having an irregular shape and distributed randomly in a composite, such representation of the spectral function must evidently have a single peak. A spread in shapes of inclusions and the cooperative phenomena would affect the width of B(n) but would not result in additional peaks, due to random morphology of the composite. In contrast, if most inclusions in a composite have either an elongated or flake-like shape, then the spectral function would have two peaks, which correspond to averaged shape factors along longer and shorter axes of the inclusions. Such form of B(n) has been observed in [S].

61

Ghosh and Fuchs [ l 11 proposed a parametrical approximation of the spectral function as:

0, n < n , or n > n 2 where two of three parameters A, nl, and nz are related by Eqs. (4), leaving one parameter to be fitted. Equation (7) describes a single peak of B(n) and is a generalization of the EMT, as it retains the same form of the spectral function but allows the width of the spectral function to be varied independently of the volume fraction. Therefore, the Ghosh-Fuchs theory (GF) may consider not only cooperative phenomena at high volume fractions of inclusions, as EMT does, but also possible spread in the shape of inclusions. On the other hand, the MG approximation may also be included in the GF theory as the limiting case at nl-+1/3, n2-+1/3. The theory seems to provide a good compromise between the classical mixing rules, which are explicit and simple but do not allow accurate prediction of macroscopic response in general, and the Bergman-Milton theory, which is accurate but difficult to apply. However, the GF theory has never been compared to the data on material constants of actual composites.

The paper is aimed at the analysis of frequency dependencies of permeability observed in composites filled with ferromagnetic powder. The consideration is made making use the concept of the spectral function. In the analysis, the volume fraction dependencies are not employed, because otherwise a particular law for the variation of the spectral function with the volume fraction must be assumed, though implicitly, which is unknown and hard to be validated. No preliminary assumptions are made on the properties of spectral function, with the exception of the general functional form, the validity of sum rules that relates the spectral function with the volume fraction, and the form of peak of spectral function with a single peak corresponding to the irregular shape of inclusions under study.

In the analysis, the intrinsic permeability of magnetic powders is considered unknown. The reason is that it frequently depends on the techniques, by which the powder has been prepared and treated. The intrinsic permeability may therefore differ greatly in powders having the same composition but prepared by different methods. It is difficult to find the intrinsic permeability from the measurement of the samples with low volume fractions, where the interactions between inclusions are negligible. The reason is that in many magnetic composites, the permeability is influenced greatly by agglomeration of the inclusions, even at very low concentrations [lo]. A possible spread in inclusion shape also affects the properties of composite and does not allow the intrinsic permeability to be determined correctly from the effective permeability of diluted composite.

Instead, we use the data on the permittivity in the analysis of the effective material parameters of the composite. Ferromagnetic materials are usually conductors, therefore the intrinsic permittivity of these approaches infinity at low frequencies. If follows from (3) that the static permittivity of composite with conducting inclusions related to the spectral function as:

Another condition that is used in the reconstruction of the spectral function is that the intrinsic permeability of inclusions is the same independently of the volume fraction.

The purpose of the paper is to obtain data on the intrinsic permeability of inclusions and the morphology of composite from the measured frequency dependencies of microwave permeability of the composite. We try various approximations for the spectral function and compare the results obtained with these approximations.

2. Experiment The study deals with composites filled with iron powders prepared by mechanical milling in

an inert atmosphere. a-Fe powder produced industrially by the carbonyl technology is used as a raw material. The raw powder particles are 300 pm or less in size and round in shape. Milling is performed with a Pulverizette-7 planetary mill. The starting powder is loaded together with 10 nun

62

dianieter stainless steel grinding balls into a hardened stainless steel vial. The vial is kept in argon atmosphere to prevent oxidation of the powder. The ball to powder mass ratio is 10: 1.

SEM image of milled powder study is shown in Fig. 1. It is seen that the powder particles are of irregular, stone-like shape. The average size of the particles is about 60 pn.

Composite samples for microwave study are prepared by manual mixing of the powder under study and warmed-up paraffin wax to achieve random distribution of the powder in the sample. The permittivity of wax is 2.25 and dielectric loss is negligible within the frequency range of interest. Then, the samples are shaped to fit the Fig. 1. SEM image of the powder under study

coaxial line, with the sample thickness being typically between 1 and 5 mm. After that, the volume fraction of the powder is recalculated based on measured specific weight of the sample to account for appearance of air pores.

The measurement of microwave permittivity and permeability is made with HP8753A vector network analyzer and a 713 coaxial sample holder in the frequency range of 0.01 to 3 GHz. The permittivity and permeability is determined from two measurements of reflection coefficient of the sample, with a short circuit located immediately behind the sample and at a certain distance behind the sample [ 121. The measured frequency dependency of microwave material parameters is shown in Fig. 2 for four different volume fractions of the milled iron powder.

3. Results and discussion The study is aimed at obtaining data on the spectral function from measured data on the

frequency-dependent behavior of the permittivity and permeability of an individual sample of

15.0 % frequency, GH 1

0.01 0 1 1 4

3

2

1

0.01 0 1 1

10

1

0.1

1

0.8

0.6

0.4

0.2

0

I I I 30.3 ‘ j imaginary permittivity /-

4 ”

A

4-

Ah *I freauencv. GHz

0.01 0.1 1 - .LI“ . .d .” - . . l - J - . .L-&”-

imaainarv Dermeabilitv

0.01 0.1 1

Fig. 2. The measured frequency dependencies of permittivity and permeability of composites filled with milled iron powder

63

certain volume fraction, without employing volume fraction dependencies of the material parameters. The spectral function is approximated making use of several unknown parameters that are to be found from the measured data. The larger number of the parameters is involved, the better accuracy may be obtained in approximating the spectral function.

The problem of importance is how many independent parameters may be extracted from the data. With known volume fraction, two equations for two unknown parameters involved in the spectral function are obtained from sum rules (4). The static permittivity is related to the spectral function by Eq. (8), which provides one more parameter, with the total of three. Notice that the permittivity exhibits frequency dependence over whole measured frequency range, when the volume fraction of inclusions is high enough. To determine the static permeability from the measured data, the data are extrapolated by polynomials t o p 0 .

Further, the condition is used that the inherent permeability of inclusions is the same independently of the volume fractions of inclusions. To quantify this condition, let us prescribe a certain value to the static intrinsic permeability. Substitution of this value to Eq. (3) must produce correct static permeability of composite, which is obtained by polynomial extrapolation of the measured data too. This procedure provides an additional equation for the parameters of the spectral function and therefore yields another parameter, with the total of four. However, this parameter depends on the assumption on the static intrinsic permeability.

If the representation of the spectral fhction involves four parameters, the above procedure produces unambiguous spectral function, which, however, depends on the prescribed value The spectral fimction allows the effective permeability to be calculated for all frequencies of interest. It follows from the above discussion that the same static effective permeability values, which are equal to will be calculated for all volume fractions involved. At higher frequencies, a disagreement may appear between the intrinsic permeability values calculated from the data for different volume fractions, provided that frequency dispersion of the permeability is pronounced enough and the static permeability is not too low. The random mean square difference characterizing this disagreement estimates the disagreement between the theory and the actual properties of the sample. The value of pi,st minimizing this rms difference may be considered as actual intrinsic static permeability of inclusions, and the rms difference in its minimum provides a characterization of whether the accepted form of spectral function is adequate to properties of the composites under study or not.

It is worth noting again that this approach allows the volume fraction dependencies of material parameters to be beyond the consideration. This is an important point that for a volume fraction dependence to be fitted to the measured data, a law must be known governing variations of the morphology of the composite with the volume fraction. This is evidently more difficult problem than determination of the morphology of an individual sample. For this reason, we suggest the approach that above may be applied separately to each sample. Comparison of the data obtained for different samples is used for the check for validity of the approach.

Let us investigate first if an adequate description of the measured data is possible without an assumption of distinguishing shape factors of inclusions, i. e., a finite width of the spectral function. This would be made with the use of the MG approximation (1) that corresponds to a delta-function- like spectral function. However, this representation does not provide sufficient number of parameters to be found. Therefore, we use a sum of two delta functions as a representation of B(n):

B(n)= A, 6(n-nl)+ A, 6(n-n,), (9)

the same as is used in the Hashin-Shtrikman (HS) approach. Equation (9) involves four parameters, A1, A2, nl, and n2, and is therefore consistent with the above approach. Notice that there is no simple physical picture for the microstructure of the composite that corresponds to spectral function (9). However, by the analogy to the HS bounds, we use this model for estimation for the range of intrinsic static permeability values that do not contradict to the data on volume fraction and measured permittivity of the composite.

With the representation of the spectral function given by Eq. (9), the equations for the parameters of spectral function may be solved numerically at ,ui,st 218.5 for all four volume fractions under

64

40 consideration. Lower values of result in n P l , which has no physical meaning. By the

30 analogy to the HS limits, this value may be assumed to be the lowest value that is achievable at any choice of the spectral function. For any higher value of the equations for the parameters of spectral function have a solution: for any prescribed ,uliSt, the value of nl can be found that is small enough to retain the measured difference between static permittivity and permeability.

0.01 0 1 1 Figure 3 plots the frequency Fig. 3. Averaged values of the intrinsic dependence of intrinsic permeability

reconstructed from the HS model (9) with p1,st =I9 and k s t =27. For every curve, two lines are shown that show error bars obtained by comparing the results for different volume

fractions. The discrepancy in the intrinsic permeability is rather low for both the PI,,, values. To make the reconstruction uncertainty of the intrinsic permeability more evident, we average the

reconstructed intrinsic permeability over all volume fractions and recalculate the averaged intrinsic permeability to the effective permeability based on the spectral function obtained for the corresponding sample. The rms difference between the calculated and measured effective permeability characterizes the accuracy of prediction of frequency dependence of permeability. The result of the recalculation is shown in Fig. 4 for =27, where the rms difference is the least. Even in this case, disagreement between the calculated and measured frequency dependencies of the permeability is clearly seen in the figure. Spectral function (9) provides much faster low-frequency shift of the magnetic loss peak than that observed in experiment. For other values of pl,st, this disagreement is even larger. Therefore, it may be concluded that if a theory does not allow for a spread in the shape factors, then it is not able to provide an adequate description of frequency dependencies of permeability.

To account for a finite width of the spectral function, we perform the same procedure with the GF theory given by Eq. (7). This equation involves three free parameters. Therefore, we need not introduce an assumption on in this case. After finding three parameters of the spectral function by Eqs. (4) and (S), we just check the agreement between calculated and measured permeability values, in the same manner as above but within the whole frequency range of measurements.

The results are shown in Figs. 5-7. In Fig. 5, the error bars are given for the intrinsic permeability of inclusion reconstructed by the GF theory, in the same way as those in Fig. 3. It is worth noting that the deviation between the intrinsic permeability obtained from different volume fractions is very small again. Even though the static permeability values are not prescribed in this

20

10

0

permeability obtained by the HS theory with =19 and p,,st =27. Doubled lines show the rms

difference of the results.

1

0.8

0.6

0.4

0.2

0

0.01 0.1 1 0.01 0.1 1

Fig. 4. Comparison between the measured permeability of composites (black lines) and that predicted by the HS approximation (gray lines) at pi,st =27

65

40

30

20

10

0

i t

0.01 0.1 1 0.001 0.01 0.1 1

Fig. 6. The spectral function reconstructed by GF approximation for various volume fractions.

calculation, these are calculated to be very close. This is an evidence for correct representation of the composite morphology by the GF theory for the case of samples under study. The permeability values and the frequency dependence of permeability are very close to those obtained by the HS theory with the optimal choice of pi,st. This may be an evidence that the reconstructed intrinsic permeability is not very sensitive to the representation of the spectral function provided the optimal selection of parameters of these representations.

Figure 6 shows the spectral functions obtained with the GF theory for different volume fractions. With the volume fraction increasing, the spectral functions grow wider, which is an evidence of strong cooperative phenomena, such as cluster formation, in the composites under study. On the other hand, the spectral function has a significant width even at the lowest volume fraction under study. This is an indication of significant contribution to the measured material parameters from inhomogeneities of the composite morphology, such as a spread in the shapes of powder particles.

Figure 7 plots the frequency dependencies of effective permeability calculated from the averaged intrinsic permeability of inclusions, in the same way as those in Fig. 4. It is seen from the figure that the measured and calculated curves agree perfectly. Therefore, the Ghosh-Fuchs theory provides an excellent agreement with the whole set of measured data on the microwave material parameters of the composites under study.

An additional check for the validity of the obtained results is performed based on Snoek‘s constant of the inclusions, S. This value is known to depend on the composition only and to be S=40 GHz for iron. From the microwave performance, Snoek‘s constant is estimated as the product of the static permeability and the ferromagnetic resonance frequency. The conventional approach is to assume that the averaged ferromagnetic resonance frequency is equal to the frequency of magnetic loss peak. However, from the observed shape of the frequency dependencies, and from large mean size of the powder particles it is concluded that eddy currents has a significant effect on the parameters of the ferromagnetic resonance. It is known that the effect of eddy currents results in a low-frequency shift of the loss peak frequency and therefore diminish Snoek‘s constant derived from microwave data. For this reason, a frequency characterizing the ferromagnetic resonance is the frequency where the real part of permeability is unity, because this featue is not affected by the effect of eddy currents [ 131. Two frequency dependencies of the intrinsic permeability shown in Fig. 3 yield S=30 GHz for pi,,t =19 and S=40 GHz for pi,st =27. The latter is in a good agreement with the reference value of Snoek‘s constant for iron.

Notice that the form of the spectral function obtained by the GF theory seems to be in contradiction with the low-frequency dispersion of permittivity. The presence of such frequency dispersion indicates that non-zero values of the spectral function may appear at very low arguments n, which would affect greatly the obtained results. However, the most likely reason for arising of this low-frequency dispersion is related to the conductivity of clusters. As volume fraction of inclusions approaches the percolation threshold, large clusters of conducting inclusions appear in

Fig. 5. Averaged intrinsic permeability obtained by the GF theory. Doubled lines show the rms

uncertainty of the result.

66

1

0.8

0.6

0.4

0.2

0

0.01 0.1 1 0.01 0.1 1

Fig. 7. Comparison between the measured permeability of composites (black lines) and the prediction of GF theory (grey lines)

the composite. The location of the dispersive frequency range associated to a cluster depends on its effective capacitance and resistance. The resistance of an actual cluster depends mostly on the resistance of non-perfect electrical contacts between inclusions comprising the cluster. This resistance typically exceeds greatly the resistance of inclusions. For this reason, the dispersive frequency range exhibits a significant shift towards lower frequencies compared to the case of a cluster with the same geometry and perfect contacts. As the shift is related to the contact resistance, it is beyond the scope of the theories considering two-component composites, in particular, the Bergman theory. On the other hand, non-perfect contacts does not affect the static properties of a composite provided that the volume fraction is below the percolation threshold. Therefore, the above conclusions on the spectral function and the intrinsic permeability still hold.

4. Conclusion 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.

Acknowledgement

Research, Agreement no. 06-08-00788.

References [ 11 A. Sihvola, Electromagnetic mixing formulas and applications. IEEE, 1999. [2] Z. Hashin and S. Shtrikman, J. Appl. Phys., 33,3 12.5 (1962). [3] M. R. Anantharaman et al., Bull. Muter. Sci., 24,623 (2001). [4] J. H. Paterson, R. Devine, A. D. R. Phelps, J. Mum. Magn. Muter., 196197,394 (1999). [5] P. Chen, R. X. Wu, T. Zhao, F. Yang, J. Q Xiao, J. Phys. D: Appl. Phys., 38,2302 (2005). [6] J. V. Mantese et al., J Appl. Phys., 79, 1656 (1996). [7] D. J. Bergman and D. Stroud, Solid. State Phys, 46,147 (1992). [8] A. R. Day, A. R. Grant, A. J. Sievers, and M. F. Thorpe, Phys. Rev. Lett., 84, 1978 (2000). 191 A. Ileilmann, J. Werner, D. Schwarzenberg, S. Henkel, P. Grosse, and W. Theiss, Thin Solid

Films, 270, 103 (1995). [lo] A. V. Osipov, K. N. Rozanov, N. A. Simonov, S. N. Starostenko, J. Phys. Condens. Matter, 14,

9507 (2002). [ l l ] K. Ghosh and R. Fuchs, Phys. Rev. B, 38,5222 (1988). [I21 K. N. Rozanov, N. A Simonov, A. V. Osipov, J. Communicat. Technol. Electron., 47,210 (2002). [13] A. N. Lagarkov, A. V. Osipov, K. N. Rozanov, and S. N. Starostenko, ZCMT 2005: Proc.

Authors appreciate the partial support of the study from the Russian Foundation for Basic

Symp. R: Electromagnetic Materials, July 3-8,2005, Singapore, p. 74.

Interface Magnetism

Gregory Kopnova, Zeev Vagerb, Ron Naamana* 'Department of Chemical Physics bDepartment of Particle Physics

The Weizmann Institute of Sciences, Rehovot 76100, Israel

Most magnetic phenomena are related to spins of unpaired electrons. However, another type of magnetism results from the orbital motion of the electrons. It is typically associated with very weak magnetic moments observed in materials already having magnetic properties.' Recently, a new type of magnetism has been identified that is related to interfaces of diamagnetic materials, "interface magnetism". This magnetism was found in organic molecules adsorbed on metal^,^'^^^ on carbon surface^,^ and in HfO2-coated silicon or sapphire.6 This interface magnetism is characterized by being temperature-independent in the range of 0 to 400K, by having high anisotropy, and by having very large magnetic signals per atom on the surface.

Several mechanisms were suggested for explaining the new phenomenon; some of them assume charge transfer from the bulk to the thin adsorbed layer. The large magnetic moment observed is attributed, according to these models, to collective orbital magnetism resulting from the response of the transferred charge to an applied magnetic field.738 Other model assumes an impurity band due to bond defects. It suggests spin-magnetism that results in ferromagnetic beha~ io r .~

In the present study, we report on an interface magnetism observed for silicon coated with its native oxide. By applying various etching procedures, the surface roughness can be modified. Following the measurements of the surface roughness by atomic force microscope, the surface morphology can be characterized by a structure function S (Ar ) which is defined as:

S ( A r ) = , / ( \ h ( r ) - h ( r + Ar)I2) Ar

where h ( r + AT) is the surface height at a distance Ar from the center h ( r ) , and (....)b means averaging over distinct centers inner the micrograph. For the self-affine surface the structure function scales with Ar as S(Ar) 0 Ara where a is the roughness exponent.

Several types of silicon wafers have been investigated (see supplement material). The wafers The etching were etched using various procedures with either HF," KOH,">I2 or N2H5(OH).

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.

eld

I

4 . 0 ~ 1 0 ~

2 .0~10-~ - 2 0.0

5 g -2.OX10"

-4.0~10-~

-10000 -5000 0 5000 10000 Field (Oe) 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 contamination would 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:

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.

We thank Prof. Itamar Procaccia for helpful discussion and suggestions regarding the

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, 1602- 1605 (1 995); P. Poulopoulos, K. Baberschke, J. Phys.: Condens. Mutter, 11,9495-95 15 (1 999) 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). P. Crespo et al, Phys. Rev. Lett. 93,087204 (2004). Y . Yamamoto et al., Phys. Rev. Lett., 93,116801 (2004); H. Hori, et al., Phys. Lett. A, 263,

[2]

[3] [4]

406-410 (1999).

70

[5] P. Esquinazi, et al., Phys. Rev. Lett., 91,227201 (2003); S. Moehlecke, , P-C. Ho, M.B. Maple, hylosophical Mug. 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;

however, its size was found to be on the order of corresponds to the thickness of the sample after etching.

emu for a gap of 300pm, which

[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 static permeability ~ 0 , ~ and effective resonance frequencyh,, for the composites. The two equations predict the dependence of M,~ and fR .e on p , which is consistent with experimental results. In addition, an important correlation (p~,~ -1) f R , e =pC , which is similar to the Snoek's law, is also 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,e and 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,e is 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 ~ 0 , ~ andfR,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 frequency fR,e of composites.

2. I Effective static permeability PO,,

matrix. Under an applied magnetic field Ho, the initial permeability pee of composite is given by Consider particles with a single domain embedded in an infinite and homogeneous epoxy

where fi is the angle of the iih magnetization Mi with respect to the Ho direction, M is the averaged magnetization along HO direction, 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,

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

ApFNd (2)

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

The interaction energy Ei between the neighboring particles is given by, E, = K, sin2 6 + K2 sin4 6 -I- K, sin6 6cos64 ( 5 )

(6) 1 1 2 . 2 1 1 2 2 2 2

E, =-[-A M sin 6cos2 41 = - [ -F(~)N,M~ sin2 ecos2 41 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

36K 1 In the derivation, NdM << HB , HB =- y'+2K;', H , =a, h 6 , b - 1 = - ~ ~ ~ d and f,,, = ( y l 2 n ) m are M Ms 2

used.

2.3 Snoek-like law

resonance frequency fR,e is obtained

where P0,UO.b andfR.6 are constant parameters for a given bulk material, leading to a constant C. It is known that the Snoek-like law is independent of F@). As compared to Snoek's law

Multiplying Eq.(3) by Eq.(7), a simple relationship between the static permeability ~ 0 , ~ and

(&,e -')fi,e = & % , b -')fj,b = P c (8)

73

there are two significant differences: (1) in Eq.(8), (pa,= -I)f;,, is constant, while (p,,, -l)fR,b is constant in the Snoek's law; (2) Eq.(8) shows the relationship between dynamic parameters for composites, while Snoek's law is applicable to bulk material.

2.4 Correction function F@) For a given particle, it is assumed that the demagnetizing interaction is due to only two

neighboring particles aligned along the applied field direction. The effects of other particles are ignored. In this approximation, the magnetic field distribution inside the given spherical particle is simulated using the commercial software Maxwell-2D-SV. The effective demagnetizing factor Ad can be obtained from the averaged magnetic field inside the given particle, as shown in Fig. 1. An almost perfect linear relationship between Ad and 1-p is observed: A ~ 0 . 3 3 6 ( l-p)-Nd(l-p). Therefore, it is reasonable to assume that F(p)=l-p. Also, the p D Z O

function satisfies the boundary condition: F(p)=O for D I ~

p=l and F(p)=l forp=O.

4"

D O 0 2 0 4 0 6 0 8 1 0 Finally, the effective ~ 0 , ~ Eq.(3) and effective f R , e

Eq.(7) for composites are given by 1 -P

Fig. 1 Calculated A d with variousp (9)

f R , e = f R , b J 1 + Nd (l - P)@O,b - '1 (10) 4 0

35 30

for spherical particles with c-plane anisotropy.

h 2 5 3. Results and Conclusions 20

Z-type barium ferrite composites, , o

Ba&o2Fe2404l/epoxy, were prepared via mixing the fine ii

ferrite powders with epoxy resin. The volume concentration p of the powders varies from 0.05 to 0.6. The complex 1s

permeability is shown in Fig.2. For better accuracy in the ,(i

determination of f ~ , ~ , the permeability spectra are curve- fitted. For example, the curve-fitted results for p=OS are

15

05

shown in the inset of Fig.2, where the grey and black lines 0 0 2 4 6 B 18 I2 11 16

are experimental data and curve-fit results, respectively. f (GHtt

~ 0 , ~ and fR,e for W-type barium ferrite composites, Fk.2 Complex Permeability with BaCoXZn2-,Fel6027 with x=0.7 and 1.0 are taken from [3]. ~ 0 , ~ for NiZn spinel ferrite composites are taken from [ 1,2].

The effective ~ 0 , ~ with various concentration volumes p is curve-fitted using Eq.(9). The curve-fit results are shown in Figs.3 (a) and 4 (a) for Z- and W-type barium ferrite composites, respectively. The symbols are the experimental data and the solid lines are the curve-fit results. The curve-fit lines are in good consistency with experimental values for two series of the composites. The curve-fitted ,&b, 8.95 and 7.2, are in good agreement with the experimental values of 9.5 and 8.0, respectively, for Z- and W-type (x=l .O) bulk materials.

The effective fR ,e with various concentration volumes p is curve-fitted using Eq.(lO). The curve-fit results are shown in Figs.3(b) and 4(b) for Z- and W-type barium femte composites, respectively. The curve-fit lines are in good consistency with experimental values for two series of the composites. The curve-fitted resonance frequency fR ,b is 2.1 GHz for Z-type bulk material, which is slightly higher than the experimental value of 1.5-2.0 GHz.

variOusfJ for Z-tfle ferrite

74

Fig.5 shows the dependence of the effective m,o.e onp for NiZn spinel ferrite composites. The curve- fit line, based on Eq.(9), is in good agreement with the experimental data.

2 0 ' . , c , , , , , , , . . . ' . . , , 0 0 02 I > 4 0 I , ,I x I ,I

\'"lllrn. FO"cc"*T.lln",,

Fig.3 The dependence of (a) p0,=and (b& on p , respectively, for Z-type composites; the solid lines are the results curve-fitted based on Eq.(9) and (10)

,

, , I , 1, I I, 2 1, 1 0 1 $, ' ', n \ o i u m mnmnlr.tbnp

Fig.4 The dependence of (a) po and (b)& on p , respectively, for W-type composites; the solid lines are the results curve-fitted based on Eq.(9) and (10)

Fig.6 shows that the product of (,u0,~-1) and fj3, is proportional to the volume concentrationp for both Z- and W-type composites. The proportionality factor is predicted by Eq.(8) . For Z-type composites, the constant C=35.5 GHz' can be also derived by extending the straight line top=l. On other hand, the constant can also be calculated from C=(pn,, -l)fRzb. p0,b=8-12 andh,b, of 1.5-2.0 for Z-type bulk material give C=18-44.0 GHz'. The averaged value is 31 GHz'. Both methods almost identical values for C. Similar results are also obtained for W-type ferrite composites.

11,111

0 C02L " I l - h p r I , , 7 .? I I ,,pL I 1 "

2 I(", i

1 IMI 3

I no " 2 0 4 O h ,ox , , I 0 '1 l l 2 I, 1 (I b 0 I

\ "IllrnC co"IC.ImPII"", \uiu-ruilonp

Fig.5 The dependence of p0.e on p for NiZn Fig.6 The correlation between (,u~,~-l> 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,e and 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,e andfR,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, 61 0054, 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 XRD 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)min is - 19 dB, tm is 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

[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,

[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

[ 101 M. R. Meshram, Nawall K. Agrawal, Bharoti Sinha, P. S. Misra, J. Magn. Magn. Mater. 27,

(1 990).

330 (2002).

(2004).

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

n sub-spectra LOG ql) with intrinsic resonance frequencies FRJ (j from 1 to n). The experimental complex permeability ye, (i=l to m) is considered to be the weighted sum of

where .zi is the error between the experimental and theoretical data, Lii @ q$ is the theoretic function of jth complex 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

where PO=PI=P,=P,+I=O. Therefore, the objective function CD can be expressed as d j = Pj_, - 2Pj + Pj+, j = 1,2 ... n (4)

79

80

0 0 0 '

0 0 0 0 0 0

0 0 0 0 0 0 . . . . . . . . . 6 - 4 1

- 4 5 - 2 1 - 2 1 ,

where g is the smoothing factor with O<g<l, w,=llA2 is the weight factor, and A is the measured error of p. Therefore, for given initial values ql, the minimum @ is determined by

D =

f l - 2 1 0 0 0 0 ... 0 0 - 2 5 - 4 1 0 0 0 ... 0 0 1 - 4 6 - 4 1 0 O . . . O 0 0 1 - 4 6 - 4 1 O . . . O 0 0 0 1 - 4 6 - 4 1 ... 0 0

0 0 0 0 0 0 0 ... 1 - 4 0 0 0 0 0 0 o . . . o 1

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

\ o 0 0 0 0 0 o . . . o 0

(7)

The probability P can be obtained by solving Eq.(7):

Then, the parameters ql in Eq.(l) can be obtained from iterative calculations based on the non-linear least squares method.

P = (L' . W .L +gD)-' .(LT .WY) (9)

Flow chart of the program is shown in Fig. 1 . The curve-fitting criterion is

(10) x= = C L I E 2

(m - n)A2 The static permeability , d o is obtained from Evaluate 9 from linear equations

n

p; =cp, + 1 (1 1) j=l

,dmm is given by the maximum in the curve- fitted imaginary permeability, and the effective resonance frequency fR is defined as the frequency corresponding ,drmar.

An effective Aef is introduced by

Calculate fitted p(f) and 2

A eff = [(*)2 -1]-"* (12) PO

Aef characterizes the dispersion of the complex permeability. The permeability is of resonant and relaxation dispersions as A e f + 0 and ile$ >>l, respectively.

Fig. 1 Flow chart of the program

81

3. Experiment Composite samples were prepared via mixing fine barium ferrite powders with epoxy resin;

the volume concentration of the powders is 50 %. The mixture was fabricated into annular disks with an outer diameter of 6.9 mm, inner diameter of 3 mm and thickness of about 2 mm for microwave measurement.

The real and imaginary permeability were measured over 0.5 to 16.5 GHz using a HP8722D Vector Network Analyzer with TRL calibration. The measurement fixture is a 7 mm coaxial air-line, with length of 59.96 mm. In addition, the complex permeability over 0.01-1.8 GHz was measured using HP429 1 B RF Impedance/Material Analyzer with 16454(S) Permeability Measurement Fixture.

4. Applications The symbols in Fig.2 are the experimental complex permeability spectra for CozZ

ferrite/polymer composites. As compared to traditional model (without the distribution of resonance frequency), much better curve-fitting (black solid curves) can be obtained using the current model. The curve-fitted lines show two characteristics of p ’ o : (a) near resonance region, p’cf) increases slightly and then decreases rapidly with frequency, (b) above resonance region p ’ m decreases to small than 1 and then increases slowly to 1.

4. I . Smoothing factor 5

0.1 1 l o f (EHA

1 2

0 8 a 21 ... 3 0 4 in n g 0 0

-04

6 a 10 12 14 16 f (GHz)

Fig.2: Comparison between the traditional and Fig.3: The effect of smoothing factor g on the curve- current curve-fit models fitting.

To obtain good curve-fitting, two factors are important: (i) number of resonance frequencies, and (ii) smoothing factor. In principle, a better curve-fitting can be obtained for more number of resonance frequencies. However, this may lead to the oscillated distribution of resonance frequency and even to negative probabilities Pj. These unreasonable characteristics can be eliminated by the smooth factor g, as shown in Fig.3. In general, g varies from 0 to 0.3. For g 2 0.20, the resonance frequencies approach identical distribution. Meanwhile, as g varies from 0 to 0.3, the curve-fitted quality does not change considerably.

4.2. Natural and wall resonances [3] The components of a complicated permeability spectrum can be obviously decomposed

into natural and wall resonances. Complex permeability spectra are shown in Fig.4 (a) and (b) for composites filled with powders of BaCoZnFel6027 and BaCoZnFe16027 doped with 1 % V2O5, respectively. The curve-fitting is performed using two components: the dot lines are gyromagnetic permeability for natural resonance and the dashed lines are wall permeability for wall resonance. For composites doped with V205, the gyromagnetic permeability p’co of 2.0 remains almost unchanged, while the wall permeability p’,,,o increases greatly from 1.4 to 3.0, as compared to that

82

of the un-doped composites. Therefore, total p’o and p”max increase from 2.6 to 4.2 and from 1.2 to 1.7, respectively, due to the significant contribution of wall permeability.

4 [ (a) . 3 0 5 2 3.

1

0

1 f (GHz)

Fig.4 Complex permeability spectra and the curve- fitted results; the dot and dashed lines are for natural and wall resonances, respectively.

f (GHz)

Fig.5 Complex permeability spectra and the curve- fitted results for Ba~CoZ+xTixFe24-2x04, composites with x=O and 1.0, respectively.

4.3. Damping coejicient [4] The effective damping coefficient A& can also improve the high-frequency magnetic

properties: imaginary permeability and resonance frequency. The complex permeability spectra of Ba~C0~+~Ti,Fe~4.2,04~ with x=O and 1.0 are shown in Fig.5. Both have almost identical static permeability p’o. However, the maximum imaginary permeability p”max and resonance frequency f R are significantly different. For x=O, prrmu andfR are 1.2 and about 3 GHz, respectively. However, p fmu increases to 1.8, andfR is shifted to 4.5 GHz for x=l .O.

pfmM andfR are strongly dependent on the damping coefficient Aer by

where F,,r is the intrinsic resonance frequency of natural resonance without damping. From Eq.( 13), small A e ~ can increase prImax and shiftfR to higher frequency. The curve-fitted results show that Aer is 1.33 for x=O, but only 0.78 for x=l.O, as shown in Fig.5. Based on Eq.(13), the calculated pfmu are 1.2 and 1.6, respectively, for x=O and 1 .O, which is in good agreement with the experimental data.

References: [l] C. Kittel, J. Phys. Radium, 12,332 (1951). [2] G. T. Rado,Rev. Mod. Phys., 25, 81 (1953). [3] Z. W. Li, Y. P. Wu and G. Q. Lin, will be published on IEEE Trans. Magn. 2007 [4] Z. W. Li, G. Q. Lin L. F. Chen and C. K. Ong, J. Appl. Phys., 99,63905 (2006)

EM Properties of Composites with Glass-coated Amorphous Ferromagnetic Wires

L. Liu*, S. Matitsine, L. B. Kong, G. Q. Lin, C. R. Deng and Y. B. Gan Temasek Laboratories, National University of Singapore, Singapore

K. N. Rozanov Institute for Theoretical and Applied Electromagnetics, Moscow, Russia

Abstract Composites with regularly or randomly distributed Co-based amorphous ferromagnetic

microwires were fabricated and characterized. Electromagnetic properties of such composites were investigated with impedance and coaxial line methods. High magnetic permeability and loss along the wire axial direction are found with low volume fraction of alloy in VHF band, which can be explained by materials and shape anisotropy. The reflectivity of regular and random composite calculated from measured EM parameters shows potential for high attenuation applications.

1. Introduction Composites filled with glass-coated ferromagnetic microwires attract great interest recently due to

high magnetic permeability and loss for frequency from MHz to GHz. Since the conductive microwires are coated with a layer of glass which prevents current flow across microwires in contact, the composites are expected to have low effective permittivity as compared to that with bare conductive wires. Such feature makes glass-coated ferromagnetic microwires a suitable candidate for high loss applications.

The Taylor-Ulitovsky method was used to fabricate glass-coated microwires with different components and structures of metallic core. The microwires comprised an inner metallic core covered by a Pyrex-like coating. The composition of alloy and process parameters (which include casting rate and cooling rate) determine the microstructure and geometrical characteristics, as well as the static and dynamic magnetic behaviour [l]. Microwires with core diameter of 0.8 to 30 pm and glass coating thickness of 2 to 15 pm can be produced. This method can also control and adjust the geometrical parameters during fabrication, and is capable of producing microwires with repeatable properties in mass production [ 11. High frequency permeability of composite with parallel amorphous microwire was investigated by coaxial line method and method used in thin film measurement [2, 31. Co-based wires exhibit a lower coercive force than Fe-based wires, which is attributed to their different domain structures. Permeability at lOMHz was significantly higher for Co-based wires than for the Fe-based materials [2]. A simple model based on effective medium theory which accounts for skin effect was proposed to determine the apparent pa permeability of wire composites [4], as follow

where q is the ferromagnetic volume fraction in the sample, pliand pm are the measured parallel and matrix permeability, respectively. Equation (1) allows the intrinsic permeability of wire to be extracted from measured data of the composite with long wires oriented along the incident magnetic field. It is found that the permeability of Co-based wire composites depends not only on the composition but also on the dimensional characteristics, such as metallic core diameter and glass-cover thickness [4]. The complex permittivity and permeability of composites filled with regular and random short Fe-based microwire were measured with Q-meter and coaxial line method from 1 MHz to 10GHz. However, the permeability of randomly distributed wire composite is almost 1, which can be explained by the high coercive force of Fe-based wires [5].

P// = qPa + (1 - q)Pm (1)

* Corresponding author, email: [email protected]

83

84

Name of wires I w 1

The aim of this study is to explore the possibility to use composites with randomly and regularly filled microwires for high attenuation applications at VHF band. Anisotropic static magnetic properties are first measured. Next, the parallel and perpendicular permeability of composites with regularly

distributed microwires are obtained by the impedance method. Subsequently, the permeability of composite with randomly distributed short microwires is measured using both impedance and coaxial line method. Finally, the reflectivity of the composites for incident plane rvave is calculated based on transmission line theory.

w 2 I w3

Inner diameter(Fm) I -5 I I

-9 I -10 Outer -10 -14 I -21

Fig. 1 SEM images of (a) Collection ofwires (WI), (b) W1, (c) W2.

diameter(pm)

H, (Oe) (0°/90 ") M, (emW

2. Experiment Co-based CoFeNiSiB amorphous ferromagnetic

microwire was provided by Microfir Tehnologii Industriale. The inner and outer diameters of the microwire are listed in Table I. The geometry and dimension of the microwire were examined with JEOL JSM-6340F field emission scanning electronic microscope (SEM). The microstructure and composition was analyzed using a Philips PW1729 type X- ray diffractometer (XRD). The static magnetic properties were measured with SMVMEDIA 880 vibrating sample magnetometer (VSM). For measurement along the parallel direction, the wire of 130mm was cut from the bobbin and wound on a pin of 3mm in diameter. Next, it was placed into a mold of 7mm in diameter before filling up with epoxy and hardener (9:l in weight) of volume fraction 10 to 20%. Microwire bundles with length of 5mm were used to measure the perpendicular permeability [51. Random

86.2 52.2 ---I

0.22/0.96 0.2710.53 ----

- - - _ _ samples were made of short microwires with length of about lmm. The volume fraction of the microwires vaned from 0.1 to 1% over seven different concentrations. The microwires were then added to the crucible with the epoxy and hardener, and manually stirred for 10 minutes. The mixture of a specific concentration of microwires was then poured into the mold and screwed on tightly. This ensured that the mixture distributed evenly within the mold. The mold was then cured in an oven at 68OC for 24 hours.

Both Anritsu vector network analyzer (VNA) with APC-7 coaxial transmission line and Agilent E499 1 A impedance analyzer were employed to measure the permeability and permittivity. The coaxial line sample has an inner diameter of 3mm, outer diameter of 7mm, and a thickness of about 2mm. The same samples were use in the impedance measurement with 16454A magnetic materials test fixture. Since the air gap has significant effect on the permittivity measured using the coaxial line method, and the wounded wire may not be perpendicular to the electric field, the accuracy of the permittivity is not

85

good. To obtain the perpendicular permittivity, samples containing multi-layer aligned microwires were measured with the impedance method and the 16453A dielectric materials test fixture.

3. Results and discussions Long wires with uniform diameters are shown in Fig.

la. In Fig. l b and lc, the alloy core and glass coating are clearly seen from the broken wires. The thickness W1 of the glass coating is larger than that given by the supplier, which can be caused by errors in controlling the processing parameters or measurement methods. Other parameters agree well with the values listed in Table I provided by the supplier. As shown in Fig. 2, wires of different diameters have similar XRD spectrum, confirming that they are of the same alloy. The broad peaks are due to cobalt and nickel in which no crystalline structure can be found. Hysteresis loops are plotted in Fig. 3 with external magnetic field applied to the direction parallel or perpendicular to the axis of the microwire bundles. Both W1 and W2 have low coercive force along the two directions as given in Table I. Different magnetization curves can be found from the parallel and perpendicular directions, which can be attributed to the anisotropic properties of the alloy or the orientation of domains inside microwires.

2800,

F(GHz)

Fig. 4 Apparent parallel permeability

F(GHz)

Fig. 5 Perpendicular permeability of W 1 (1 3~01%) and W2 (26~01%)

Fig. 4 shows the apparent

parallel permeability

o fWlandW2 obtained from

200

175 CoFeNiBSi

0

20 (degree)

Fig. 2: XRD spectrum of amorphous microwires

- e 50- - E

-100 -5000 -2500 0 2500 5000

Applied field (Oe)

Fig. 3 Hysteresis loop impedance measurement with Eq. (1). Different domain structures in W1 and W2 may result in different intrinsic p, The peak value of the imaginary part depends on the skin effect and gyromagnetic resonance frequencies. Thinner wires have higher resonance frequency as well as larger p' at high frequency. Fig. 5 shows the perpendicular permeability, also obtained by the impedance method. The relatively low permeability at high frequency in the perpendicular direction is consistent with the hysteresis loops in Fig. 3, since this direction has a stronger coercive force and is difficult to magnetize to saturation. However, the low permeability is also due to the demagnetization factor of the circular section of microwires. According to Maxwell-Garnett equation, the maximum effective p of composite with spherical magnetic inclusions is less than 4, regardless of the intrinsic p of the inclusion. In contrast, the demagnetization factor for elongated inclusion with extreme aspect ratio is zero. Hence, the effective p i s

86

proportional to the intrinsic p and volume fraction. Fig. 6 shows the permeability of composites with randomly distributed wires. The static p' and maximum p" depends linearly on the volume fraction up to 1%, which is in agreement with Eq. (1). Even the shape of the dispersion curves in Fig. 6 is similar to the intrinsic properties of the composites with regularly distributed microwires (see Fig. 5) . However, the value is somewhat larger than that predicted by Eq. (l), assuming that the short wires are randomly distributed in 3D. Possible reasons for this discrepancy include a range of sizes of wires in random composites which results in a broader loss peak, orientation of wires along certain directions due to the mixing or filling process which results in a higher p' value.

Reflectivity of a layer of metal-backed composite is calculated using transmission line theory for 30 - 300MHz, as shown in Fig. 7. An isotropic composite of thickness 30mm with 1% of randomly distributed wires has an absorption peak of about -1OdB at 15OMHz. In comparison, an anisotropic composite of thickness 6mm with 10% wires oriented along the magnetic field achieves more than 27dB attenuation at 2OOMHz. The thickness of this composite is about 0.4% of the wavelength at 200MHz, with a density of about 2.4g/cm3. However, only incident wave with H- field along the wire axis can be attenuated. The cross polarized component is totally reflected due to high conductivity of microwire.

F(GHz)

Fig. 6 Permeability of random composites (Wl)

F(MHz)

Fig. 7 Reflectivity of microwire composites at VHF

4. Conclusions The high permeability and magnetic loss with low volume fraction can be attributed to

anisotropic material properties and high aspect ratio of microwires. The thin and low-weight composites with magnetic microwires are promising for high attenuation applications at VHF.

References [ 11 V. S. Larin, A. V. Torcunov, A. Zhukov, J. Gonzalez, M. Vazquez and L. Panina, JMMM, Vol. 249,

[2] 0. Acher, P. M. Jacquart and C. Boscher, IEEE Trans. on Magnetics, 30(6), 4542(1994) [3] P. M. Jacquart and 0. Acher, IEEE Trans. on MTT, Vol. 44(1 l), 21 16(1996) [4] S. Dprot, A. L. Adenot, F. Bertin and 0. Acher, JMMM, Vol. 242-245,247(2002) [5] S. N. Starostenko, K. N. Rozanov and A. V. Osipov, JMMM, Vol. 298,56(2006)

39(2002)

Session P4

Chair: L.J. Deng

Experiments on Electromagnetic Characterization of Ferromagnetic Nanocrystalline Alloy Flake Composites

P. H. Zhou, L. J. Deng State Key Laboratory of Electronic Thin Films & Integrated Devices, China; Engineering Research

Center of Electromagnetic Wave Absorbing Materials, Ministry of Education, University of Electronic Science and Technology of China, Chengdu, 610054, China

Abstract: Ferromagnetic nanocrystalline alloy flake composites can be useful in microwave absorbers, since flakes exhibit excellent electromagnetic properties. This paper reviews and updates our experiments and understanding of nanostructure flakes. Structural and morphological influence of flakes on microwave properties are discussed, with respect to the exchange coupling and surface effect of nanocrystals, particle shape dependence of dielectric constant and percolation threshold. It’s found that percolation threshold for FeSiB nanocrystalline flake composite is around 35%. Keywords: nanocrystalline; flake; electromagnetic properties; composite

The combination of high-saturation magnetization and high-relative complex permeability makes ferromagnetic nanostructure composites attractive for applications in microwave absorption. [I-31 However, the higher the frequency, the thinner must be the particles in order to allow penetration of electromagnetic wave. Consequently, nanocrystalline alloy flakes are promising for thinner absorber filler. [4, 51 Material in this type is characterized by a microstructure well-known for the nanosize crystalline grains with highly specific surface and a flaky morphology for each particle. Nanostructure features in excellent soft magnetic property leading by the exchange interactions well described in the framework of the random anisotropy [6], while the flaky morphology optimizes the microwave permeability in line with the predictions from the modified Snoek limitation [7]. Although extensive literature exists on the properties of nanophase materials [ 11 and particle size dependence of microwave performance in metallic powder/insulator composites [8, 91, only a limited number of systematic studies have been reported on electromagnetic properties of nanocrystalline flakes composites. The difficulty in making precise statements on this issue stems from a lack of better understanding about how nanostructure and flaky morphology of flakes affects the dielectric and dynamic magnetic properties of composites. In this paper, we report our experiments and main results in the study of micro-structural and morphological influence on complex permeability and permittivity of nanocrystalline alloy flake composites in 0.5-18GHz.

Fe-based nanocrystalline alloy powders with various composition, microstructure, and particle morphology (shape and size) have been synthesized by high energy milling of the mixture of elemental powders or the rapid-quenched ribbons. Microstructure and composition of powders were characterized by XRD. Mean grain size of crystallites was estimated by Scherrer analysis. SEM and VSM were conducted to characterize the morphologic and magnetic properties. As-milled powders were randomly dispersed in paraffin at given weight ratios, shaped into toroidal ring. Volume fraction of the alloy powders was determined by density measurements. The microwave properties [&(a), p(a)] of flakes-parafin composites were measured in the 0.5-1 8 GHz range with a APC7 coaxial line associated with a Agilent 8720 ET vector network analyzer. Measurements were conducted in the absence of an external magnetic field. 1. Structural influence

Two important structural features of nanocrystalline alloy for electromagnetic properties are the coupling of nanocrystals and the surface effect in the vicinity of an interface. We have

89

90

conducted experiments to investigate such effects since structural modification are approached by elemental doping as well as adjusting of milling condition [lo].

The magnetic comparison between high energy milling (or called mechanical alloying) produced FeNi nanocrystalline alloyed and unalloyed powder in the same nominal composition evidenced the advantage of nanostructure alloys in improving saturation magnetization and microwave permeability while coercivity decreases. [8] Further researches of the influence of structural parameters such as phase composition, crystal lattice constant, and grain size were conducted on FeCuNbSiB and FeSiB nanocrystalline alloy powers prepared by thermal and successive mechanical treatment of amorphous ribbons. [ l l , 121 By the virtue of extensive exchange coupling between nanocrystals, emerging hard magnetic phase Fe3B in some samples lowers complex permeability by coercivity enhancement and is so unexpected, [ 121 similarly, coercivity decreases with diminishing grain size at the nanometer scale [l l] . Moreover, G. Herzer et al. recently reported their discussion on anisotropy of nanocrystalline alloy, denoting the competition between the averaged anisotropy from coupling effect and uniaxial anisotropy from sources like additional phase, magneto-elastic or field induced anisotropy. [13] It might be the domination of uniaxial anisotropy instead of averaged one causes the high value of coercivity in Fe74Si13B13 nanocrystalline sample in Tab.1, as a relationship between magnetic anisotropy K and grain size D changes from K cc D6 to K cc 0’. However, M e r researches are needed in this field. Tab. 1 Structural and magnetic properties of Fe~.1.,Si,B~~(x=l,5,9,13 )nanocrystalline alloy samples: weight ratio of Si element; grain size (D); saturation magnetization (Ms); coercivity (Hc).

sample Fe74Si13B13 Fe7&sB13 FeszSi~B13 FessSilB13 msi/(msi+mF,) (“h) 8.07 5.45 2.95 0.58 M, (emu/g) 128.72 138.38 144.96 142.48 Hc ( O e ) 152.44 142.50 144.76 150.21 D W ) 9.56 9.85 10.02 10.27

We have also studied the magnetic resonance of nanocrystalline Fe-based alloys, while eddy current loss was suppressed by flaky morphology. [ 141 Simulation was conducted according Aharoni’s formula with particle size changed into grain size, and results were close to experimental. It’s indicated that exchange energy from coupling of nanocrystals plays the dominant role in microwave magnetization, comparing to static magnetic energy.

In particles composed of dozens of small grains with diameters of a few nanometers, surface effect becomes very important as the properties of material inevitably depend on the interfacial properties of the constituents due to the high fraction of the overall material which is in the vicinity of an interface. Saturation magnetization is reported to be dependent on gain size due to the effect of specific surface state both experimentally and theoretically. It’s believed that addictive contribution to anisotropy from surface region of nanocrystals exists [15], but systematic study is unseen. 2. Morphological influence

Particle shape (the axial ratio a=& of a spheroid is often used to describe this feature) and size illustrate the morphological characterization of particle, besides some added factors such as size distribution, agglomeration and so on. Our previous work on mechanically alloyed FeNi and FeCo nanocrystalline particle composites have discussed the optimum particle size around 1 pm and the advantage in improving permeability as shaping into flakes. [8, 161 Particle shape influence on permittivity is explained by the enhancement of depolarization of particles. However, the other way of such an influence on permittivity of metallwax composite is approached in determining the

91

percolation threshold.Ultrafme FeSiB nanocrystalline flakes in the size range of l~5um are successfully prepared,

and the thickness of flakes is about several tenth of a micrometer. [12] Microwave properties ofcomposites are illustrated in Fig.l and Fig.2 [17]. The increase of the dielectric constant with therise of flakes concentration can be ascribed to the increase in the volume fraction of the charges(electric dipoles) in the interface between wax and the metal particles. The large increase of boththe real and imaginary part of the complex permittivity, for volume fraction of flakes larger than35% (Fig.l), can be related with the percolation threshold. Percolation occurs at a critical volumefraction pc where the particles contact with each other and, as a consequence, a continuouselectrical path of the metal particles is built throughout the wax matrix and the composite ismetallized. The high permittivity level is caused by the increase of total particle-wax interface andthe high conductivity resulting from percolation. Considering the increase of permittivity caused byflaky shape [16], it's indicated that flaky shape creates a higherpc, which can be understood by theenhancement of electrical dipoles, considering the shape dependence of bulk's electricdepolarization [16] and larger surface area of flake than sphere in per unit weight of material.

-n-15%—O— 30%-A-45%-v-62%

600 -

4 5 0 -

300-

150 -

10 12 14 16 10 12 14 16 18

f(GHz) f (GHz)

Fig. 1 Complex permittivity versus frequency spectra for composite with different volume fractionof FeSiB nanocrystalline flakes. [17]

Also noteworthy is the distortion of permeability curve for volume fraction near percolationthreshold in Fig.2. Permeability increases with increasing ferromagnetic inclusion concentration, asthe magnetic effect is gradually enhanced. The curve becomes distorted while the interaction ofparticles creates additional magnetic influence such as an increasing anisotropy [18] or eddycurrent.

f (GHz) f (GHz)

Fig.2 Complex permeability versus frequency spectra for composite with different volume fractionof FeSiB nanocrystalline flakes. [17]

92

In summary, electromagnetic properties of ferromagnetic nanocrystalline flakes composite are discussed through two approaches-structural and morphological characterizations. Exchange coupling and surface effect, especially the former, tailor the soft magnetic properties and promise high permeability. However, the surface effect on both magnetic and dielectric properties and the competition between the averaging and the more uniform anisotropy contributions are unclear, while their existence is generally recognized. Morphological influence is studied on the value of microwave permeability/permittivity and the percolation threshold of composites. Percolation threshold of FeSiB nanocrystalline flakes composite is observed near 35%. The enhancement of electrical dipoles with flakes concentration explains the behaviour of dielectric constant. Moreover, particle interactions in the vicinity of percolation threshold extensively transform the microwave spectra of electromagnetic parameters, causing errors in measurement.

References: [l] C. Brosseau, J. B. Youssef, P. Talbot and A. Konn, J. Appl. Phys., vo1.93, no.11, 2003,

[2] V. B. Bregar, IEEE Trans. Magn., vo1.40, no.3,2004, p.1679-1684. [3] S. Sugimoto, T. Maeda, D. Book, T. Kagotani, K. Inomata, M. Homma, H. Ota, Y. Houjou, R.

[4] S. S. Kim, S. T. Kim, Y. C. Yooh, K. S. Lee, J. Appl. Phys., 97, 10F905(2005). [5] S. Yoshida, S. Ando, Y. Shimada, K. Suzuki, K. Nomura, K. Fukamichi, J. Appl. Phys., vo1.93,

[6] G. Herzer, Nanocrystalline soft magnetic alloys, in: K. H. J. Buschow(Ed.), Handbook of

[7] R. M. Walser, W. Win, P. M. Valanju, IEEE Trans. Magn., ~01.34, no.4, 1998, p.1390-1392. [8] P. H. Zhou, L. J. Long, J. L. Xie, D. F. Liang, L. Chen, and X. Q. Zhao, J. Magn. Magn. Mater.,

[9] L. Z. Wu, J. Ding, H. B. Jiang, L. F. Chen, C. K. Ong, J. Magn. Magn. Mater.,

[lO]P. H. Zhou, L. J. Long, J. L. Xie, D. F. Liang, L. Chen, Journal of Electronic Science and

[11]P. H. Zhou, L. J. Deng, J. L. Xie, Y. Q. Liu, IEEE conference on nanoelectronics, p. 365-366. [12]Y. Q. Liu, P. H. Zhou, L. J. Deng, Journal of University of Electronic and Science of China, in

[13] G. Herzer, J. Magn. Magn. Mater,, 294(2005)99-106. [ 141 L. J. Deng, P. H. Zhou, J. Appl. Phys., in revision. [l5] J. Gonzalez, Appl. Phys. Lett., ~01.85, no. 24,2004, p.5944-5946. [16]P. H. Zhou, L. J. Deng, J. L. Xie, D. F. Liang, J. Alloys Compds., in publish. [17]L. J. Deng, P. H. Zhou, Journal of University of Electronic and Science of China, vo1.35, no.4,

[18]L. C. Costa, M. Valente, M. A. Si, F, Henry, Polymer Bulletin 57, 881-887(2006).

p.9243-9256.

Sato, J. Alloys Comp., 330-332(2002)301-306.

no. 10,2003, p.6659-6661.

Magnetic Materials, vol. 10, Elsevier, Amsterdam, 1997, p.4 15.

292(2005)325.

285(2005)233-239.

Technology of China, vo1.3, no.2,2005, p.164-167.

publish.

2006.

High-frequency Magnetic Properties for Composites of ZnNi-substituted Y-type Barium Hexaferrites

Y . P . W U ~ ~ ~ , Z . W . L ~ ~ , G . Q . L ~ ~ ~ , ~ ~ ~ C . K . O ~ ~ ~ a Temasek Laboratories, National University of Singapore, 5 Sports Drive 2, Singapore, 11 7508

Centre for Superconducting and Magnetic Materials, Department of Physics, National University of Singapore, Singapore, 11 7542

b

Abstract High-frequency magnetic and absorbing properties for composites of ZnNi-substituted Y-type

barium hexagonal ferrite, Ba2ZnxNi2.,Fe12022, are investigated. Three magnetic resonance peaks are observed for ferrites with Zn substitution. Based on the measured electromagnetic properties, low reflection loss at microwave frequency is predicted for a composite slab of Ba2ZnxNi2.,Fe~2022.

1. Introduction Microwaves are increasingly used in wireless communication tools, local area networks, etc.,

over the frequency span of 1-20 GHz. Thus, electromagnetic interference (EMI) is becoming a serious problem and much attention has been paid to find suitable microwave absorbing materials. Barium ferrite powders are ideal candidates due to their low cost, low density, high stability, high microwave magnetic loss, and more importantly, large electrical resistivity. This work investigates the high-frequency magnetic properties of composites mixed by Ba~Zn,Ni~.,Fe12022 ferrite powders. The effects of Zn concentration x on high-frequency magnetic and absorbing properties will be reported.

2. Experimental Y-type hexagonal ferrites, Ba2ZnxNi2.,Fe12022 with x varying from 0 to 2.0 in steps of 0.4,

were prepared by the conventional ceramic process. The composites were prepared by mixing crushed ferrite powders with epoxy resin; the volume concentration of ferrite powders was fixed at 50 %. Microstructure was observed using a field-emission gun scanning election microscope (FEG- SEM) for crushed and ball-milled powders of the sintered samples. The complex permeability and permittivity values were simultaneously obtained by measuring the Sll and S2l parameters over 0.5 to 16.5 GHz using an HP8722D vector network analyzer (VNA). The measured p and E data were used to estimate the absorbing characteristics of composites based on the metal-backed single-layer model. In addition, the complex permeability over 0.01-1.8 GHz was measured with an HP4291B impedance analyzer.

3. Results and discussion The dielectric properties of the composites were characterized from 0.5 to 16.5 GHz. The real

and imaginary permittivity, E’ and E”, do not show noticeable dependence on frequency and the composition of ferrites. The real permittivity is about 7.0-7.4, while the imaginary part is about 0.2- 0.3 for all composites over the frequency range studied. Typical complex permittivity spectra are shown in Fig. 1.

The dynamic magnetic properties for all composites were examined from 0.01 to 16.5 GHz. The complex permeability spectra for all samples are shown in Fig. 2.

For composites without Zn substitution, two obvious resonance peaks, namely P1 and P2, can be observed. As the Zn concentration x is increased, a third peak P3 appears at about 0.1 GHz. Further, the intensity of this peak is enhanced as the concentration of Zn ions is increased. Except for the sample of Ni2Y, there are three resonance peaks in the permeability spectra of all other samples. However, it is well known that for polycrystalline ferrites under ac field, there are only two kinds of resonance mechanisms, namely natural and domain wall resonances.

To identify the resonance mechanism, single domain particles for ferrites with x=O.8 and 2.0 were prepared by high energy ball-mill method. Based on SEM morphologies, the average size of

’

93

94

the original powders without ball-milling is about 3 pm, which is much larger than the critical diameter of magnetic domain (about 0.6-1 .O pn) for hexaferrite~.~ However, for the ball-milled powders, most particles are less than 1 pm in diameter. Hence, for composites filled with these ferrite particles, domain wall resonance should not occur due to the absence of domain wall. Fig. 3 shows the permeability spectra for composites with 50 vol% of ball-milled ferrite powders. There is only one obvious resonance peak in the permeability spectra for both composites. As compared to the spectra of samples without ball-milling, the resonance peaks at low frequency (P2 and P3) vanish, while the peak at high frequency (PI) remains observable. Therefore, it is concluded that PI is induced by natural resonance, while the other two peaks (P2 and P3) originate from domain wall resonance.

In addition, the resonance frequency of PI is tightly dependent on the substitution x. For sample with x=O, resonance occurs at about 6.1 GHz. As x increases, resonance is shifted to low frequency. For samples with xL1.2, only two obvious peaks are observed due to the overlapping of PI and P2. The variation should be attributed to the change in the anisotropy field. Braden et al.4 systematically investigated the magnetoclystalline anisotropy of various ions substituted Y-type ferrites, and found that the anisotropy field decreases from 19 kOe for Ni2Y to 8.4 kOe for Zn2Y. Our previous studies5 showed that natural resonance frequency varies almost linearly with the anisotropy field for composites of ferrite with c-plane anisotropy. Therefore, the decrease in natural resonance frequency with Zn2' substitution is expected in composites of NiZn-substituted Y-type ferrites, as shown in Fig. 2.

Electromagnetic wave absorption can be determined from the reflection loss (RL). In the case of the metal-backed single layer, RL is given by

I z,, =z , U2V' - - a h j---JI..

where Zi, is the impedance at the air-material interface looking into the metal-backed single layer, Zo=1 is the normalized impedance of free space, c is the light velocity, t is the thickness of composite, p and E are the intrinsic complex permeability and permittivity of the composite, respectively. The optimum thickness to is determined by the maximum relative bandwidth Wmm=fplJjow. f p and JOw are upper- and lower-frequency limits of the bandwidth for absorption of 10 dB or more, respectively. Fig. 4 shows the frequency response of the reflection loss at to for samples with x=O and 1.2. The values offup,Jow, to and the corres onding W,,,, are listed in Table I for all composites studied. With the substitution of Ni2+ by Zn', the frequency band for 10 dB absorption shifts to low frequency, from 5.4-12.1 GHz for Ni2Y to 3.9-10.7 GHz for Zn2Y. These are potentially good EM materials with low reflectivity and broad bandwidth for microwave applications.

4. Conclusions High-frequency magnetic and absorbing properties have been investigated for composites of

Ba2ZnxNi2-,Fe12022. Three magnetic resonance peaks for composites with Zn substitution were observed. The resonance mechanism for each peak has been identified. In addition, the predicted reflection loss shows that the composites of Ba2ZnxNi2-,Fe12022 are promising candidates for electromagnetic materials with low reflectivity at microwave frequency.

95

Table I: The optimum thickness to, the upper- and lower-frequency limits,f, andf;,,, for absorption of more than 10 dB, and the relative bandwidth of W=fudJ;ow for all composites.

X 0 0.4 0.8 1.2 1.6 2.0

to (mm) 3.1 3.2 3.2 3.3 3.4 3.7

f O w (GHz) 5.4 4.5 4.3 3.9 3.9 3.9

fup(GHz) 12.1 12.3 12.4 11.8 11.0 10.7

W=fuPplJi,, 2.24 2.73 2.88 3.03 2.82 2.74

Fre. (GHz)

5

4

3

2

1

0

-5

,

0.01 0.1 1 10 Fre. (GHz)

Fig. 1 : Typical complex permittivity spectra in the frequency of 0.5-16.5 GHz for composites of Ba;?ZnxNi2.,Fe 12022. of Ba~Zn,Ni2-,Fel202~.

Fig.2: Complex permeability spectra in the frequency of 0.01-16.5 GHz for composites

96

Fre. (GHz)

0

-10 s s 2 -20

f 0 4 8 12 16

Fre. (GHz)

Fig.3: Complex permeability spectra in the Fig.4: Absorbing characteristics for frequency of 0.01-16.5 GHz for composites of composites of Ba2ZnxNi~.,Fe12022 at the ball-milled powers: (a).x=0.8, and (b). x=2.0. optimum thickness: (a). x=O, and (b).x=1.2.

References

[l] Z.W. Li, L.F. Chen and C.K. Ong, J. Appl. Phys. 92, 3902 (2002). [2] H.-S. Cho and S.-S. Kim, IEEE Trans. Magn. 35, 3151 (1999). [3] H. Kojima, Ferromagnetic Materials, Vol. 3, Ed. E.P. Wohlfarth (North-Holland, Amsterdam),

[4] R.A. Braden, I. Gordon and R.L. Harvey, IEEE Trans. Magn. MAG-2,43 (1966). [ 5 ] Z.W. Li, L.F. Chen and C.K. Ong, J. Appl. Phys. 94,5918 (2003).

357 (1982).

Carbonyl Iron Composite Materials for High-frequency Applications

Madina A. Abshinova'*, Alexander V. Lopatin',2, Natalia E. Kazantseva',2, Jarmila VilEikova', Petr Sfia'

Polymer Centre, Faculty of Technology, Tomas Bata University in Zlin, Czech Republic 'Institute of Radio-Engineering and Electronics, Russian Academy of Sciences, Russia

*e-mail: ahshinova@,fi. utb. cz

I

Carbonyl iron powders (CI) are produced by thermal decomposition of iron pentacarbonyl (FeC05) and widely used as filler in polymeric composites. High saturation magnetization of CI and quality factor at high frequencies make them attractive for a wide range of applications, like plastic-encapsulated inductor cores [ 1-21 or for microwave absorbing materials [3-41. For the last- mentioned applications, the complex permeability p*fl=pqp" and permittivity ~ * f l = E' - jd' spectra are important factors.

Composite materials filled with CI were studied from the view point of particle microstructure, size, shape, and concentration effects to reach required electromagnetic properties. The present work deals with CI used for electronic components (ES, EW, HQ, SL, SQ, R-10, R-20, RCI) and microwave absorbers (EA, EN, ES, EW, MCI). Chosen CI particle differ in microstructure and morphology, namely hard grades (ES, EW, HQ, R-10, R-20, MCI) and soft grades (SL, SQ, RCI). The main characteristics of CI are summarized in the Table I.

Hard grades are produced from the primary decomposition process without further chemical treatment. Thus, particles have onion skin structure; tension between onion skin layers causes mechanical hardness. The content of Fe is up to 97.8 %, while the content of other elements (C, N, 0) is about 1.2 %. Soft grades are manufactured by the reduction of hard grades with hydrogen. In reduction process, the onion structure is lost; the content of Fe is increasing up to 99.5 %.

Due to the synthesis of CI, a unique onionlike structure is formed. It consists of a-iron crystalline phase separated by layers of amorphous cementite, namely iron carbides (Fe3C) and nitrides (Fe4N) [5, 61. Therefore, CI particle on one hand is a multicomponent system, and on the other hand is nano-object with alternating nano-layers of a-iron and cementite. Commonly, CI particles are spherical with an average particle size between 4-7pm (Fig. l(a)) and characterized by onionlike structure (Fig. 2(a)). Mechanical or physico-chemical treatment of primary CI powders leads to the disruption of particle shape (Fig. l(b)) as well as the microstructure (Fig. 2 (b)).

Fig. 1. SEM photography of the CI particle morphology: (a) spherical particles (enlargement - 1:5000), (b) flaky like fragments (enlargement - 1:2000).

97

98

Fig. 2. Microstructure of CI particles: (a) onionlike structure; (b) disrupted structure.

CI filled composite materials were prepared by moulding of CI with appropriate amount of silicone elastomer (SYLGARD 184, Dow Corning, USA) and polyurethane (AXSON UR 3420, Axson, France).

Table I. Main characteristics of CI, pressed toroids, and CI filled composite materials.

TypeofCI I Chemical Composition (at %) I Particle size distribution, pm I Initial permeability*

* Initial permeability obtained by measuring pressed CI powders (97 wt. %) with polyvinyl alcohol (3 wt. %) at the pressure 1GPa.

Samples on the base of silicone elastomer were cured between 65"C-15OoC for 4-1 h, depending,on the type and concentration of CI, while samples on the base of polyurethane were cured at 80°C for 4 h in a vacuum chamber. Thus, composites containing 10-52 vol. % of CI were prepared.

99

The complex permeability and permittivity spectra were measured by two methods. In the frequency range from 1 MHz to 3 GHz, p* and E* were investigated by an RF ImpedanceMaterial Analyzer (Agilent E4991A) and in the frequency range from 2 GHz to 10 GHz by resonance method.

The range of frequency dispersion of complex permeability in CI polymer composites is shifted to 109-10’0 Hz in comparison with bulk sample where it occupies frequency range from lo6 to 10’ Hz (Fig. 3). The shift of the frequency region of p* in the composites first of all is attributed to polarization of CI particles separated by a nonmagnetic polymer layer.

Comparative analysis of the identically filled CI composites educed the differences in absolute values as well as in the character of magnetic and dielectric spectra due to the variation of the microstructure of CI particles (Fig. 3 , Fig.4).

Mechanical treatment leads to the structural changes, namely to flaky shape of the particles. It reduces demagnetization factor and resulted in a denser packing of particles. Destruction of amorphous cementite layers intensifies the shift of the domain walls, thus hampering structural relaxation processes and increases the losses associated with eddy currents.

1 0’ 108 10’ 10‘0

Frequency,

11

10

11

10

9 s 9 a 8

5 8 $ 7

2 6 E : 2 5 .& 4 $ 4 ; 3

E 3

2 7

t ’ t i

1 0

0 1

n 2

10’ 108 10s 10‘0 10’ 1on 1 0’ 10’0 (b)

Frequency. Hz Frequency, Hz

Fig. 3 . Complex permeability spectra of net SL sample (a) and CI-rich composites (b).

100

100, , . , . . . . , , . -, , . . . . . . . , 1

10 3 10' 10' 1 on 10'0

Frequency, Hz Frequency, Hz

Fig. 4. Comparative analysis of the complex permittivity spectra of composites with different types of CI and the same content of CI 50 vol. %.

Physico-chemical treatment leads to greater values of p' and p" which is likely to be associated with changes in microstructure of CI particles: (i) the disruption of cementite layers, which suppress the growth of a-iron grains, (ii) considerable reduction of the concentration of carbon and nitrogen in crystalline a-iron lattice, (iii) growth of a-iron domains with the formation of magnetic texture of a-iron grains. Moreover, high magnetic losses are observed in a wide frequency range 10s-lO'o Hz, which results from the large size dispersion, shape anisotropy, and multidomain structure of a-iron grains.

The dielectric properties of CI composites are determined by the layered onionlike structure, namely by amorphous layers of cementite on the boundaries of a-iron blocks and on the surface of particles. This leads to high electric resistivity due to the absence of direct electric contact between particles. Mechanical and physico-chemical treatment destroys dielectric layers, form conducting clusters of contacting particles, and increase conductivity and complex permittivity.

Thus, in spite of negligible difference in chemical composition and size distribution of different types of CI, there is significant difference in electromagnetic properties of CI-filled composites in the broad frequency range.

From the practical point of view, the variation of the type of CI in a composite, namely, the microstructure and the shape of particles, makes it possible to control the absolute values of electromagnetic parameters and to monitor the frequency dependence of,u* and E*.

References [I] A. Tailor and A. Richards, Soft Magnetic Materials for Telecommunications, Press. L., 1953. [2] Handbook of Powder Metal Technologies and Applications, vol. 7, American Society of Metals,

[3] R.N. Johnson and C. Loret, US Patent 20050139282,2005. [4] N.E. Kazantseva, I.A. Tchumit, and N.G. Ryvkina, Journal of Communication Technology and

[5 ] V.S. Volkov, V. G. Syrkin, and I. S. Tolmasskii, Carbonyl Iron, Metallurgy, Moscow, 1969. [6] J. S. Umanskii, X-Ray Diffraction Analysis of Metals, Metallurgy, Moscow, 1967.

1998.

Electronics, 48 173 (2003).

EM Properties in Composites with the Filler of Electroless-synthesized Ni-P Powder

G.Q. Lin''2 and Z.W. Li' I Temasek Laboratories, National University of Singapore, 1 17508; ' Department of Physics,

National University of Singapore, I1 7542

Abstract: Nickel-phosphorus (Ni-P) powders of different sizes, from nanometer to micrometer, were successfully synthesized using electroless plating method. The parameter determining the particle size is also studied. Composites are fabricated by mixing metallic powder with epoxy at certain volume fraction. The dispersions of permeability and permittivity for the prepared composites are measured from 10 MHz to 1 GHz. Based on the experimental results, its application as electromagnetic (EM) materials is discussed.

1. Introduction Although magnetic metallic materials have better magnetic properties than ferrites, their low

resistivity always leads to high eddy current at high frequency, which limits their applications as EM materials. Therefore, it is interesting to study the physical behavior of EM wave for composites containing magnetic metallic powder. [ 1-41 Their potential applications as EM materials at gigahertz are expected.

The technology of electroless nickel (EN) deposition is credited to Brenner and Riddell in 1946 [ 5 ] where Ni, together with the inevitable P (3- 15 percentage), is coated on metallic substrates from aqueous solution. It has been widely investigated and accepted in surface treatment because of its unique qualities, such as uniform thickness, excellent corrosion, wear resistance and hardness. Usually, a good EN bath should be sufficiently stable for long-term use by continuous replenishment with the consumed component. However, the stability can be destroyed when the concentration of reaction reagents is increased and metallic particle in nano-size can be obtained.[6,7] In this paper, a systematic investigation on the electroless-synthesized Ni-P powder is conducted and its application as EM materials is discussed.

2. Experiments Ni-P powder is fabricated using modified electroless method. During the synthesis, 100 ml

NiS04 with complexant, NaOH and NaH2P02 (0.12 M) aqueous solution were first heated to 80 "C. The concentrations of NiS04 and NaOH solutions were slightly changed for different electroless- synthesized Ni-P samples. Then, the NaOH and NaH2P02 solution were subsequently poured into NiS04 solution very rapidly. The color of the solution was changed from green to light green and finally, grey with black particles deposited in solution. During the reaction, a lot of gas is generated from solution. To complete the reaction, the solution was continuously stirred for about 1 h until no gas was observed. All powder was then collected by washing the mixture using distilled water several times and put into oven for 2 hours at 70 "C. The phase structure was characterized using X- ray diffraction (XRD) and the morphology was observed using scanning electron microscopy (SEM) .

Composites were prepared by mixing Ni-P powder with epoxy. One cylindrical sample with diameter of -16 mm and thickness of -2 mm was prepared for the measurement of complex permittivity. Another toroidal sample with inner diameter of 3.0 mm, outer diameter of 6.9 mm and thickness of - 2 mm was prepared for the measurement of complex permeability. All microwave measurement is conducted using Agilent E4991A RF impedance/materials analyzer over 10 MHz- 1 .O GHz.

During the experiment, a commercially available nickel powder from Merck is also studied for comparison purposes.

101

102

3. Results and Discussions When conducting the chemical synthesis in a solution, the concentration of precursor reagents

is always an important factor affecting the properties of final products. In our work, the effect of OH- and Ni2+ concentration in solution is systematically studied. The ratio of OH-/Ni2+ is changed from 1 to 2 while Ni2+ or O H concentration is kept constant at 0.4 M. The XRD results for all Ni-P powders and commercial Ni powder are shown in Figure I, together with the Ni reference data from ICDD. As in commercial nickel metal, all electroless-synthesized Ni-P powders have the cubic structure, except that synthesized at OH-/Ni2+=2 with Ni2+=0.4 M which is oxidized immediately after exposing to air at room temperature. The peaks for all electroless-synthesized Ni-P powders are much broader than those of commercial Ni powder because the powder exhibits a fine dispersion of microcrystalline Ni in an amorphous matrix. This structure was also observed in annealed Ni-P thin film fabricated by electroless plating method.[8-91

The morphologies of all powders are observed using SEM and shown in Figure I1 (a)-(i). The insets in Figure I1 (a) and (i) are magnified image and detailed surface structure for the corresponding metallic powder, respectively. All powders, including electroless-synthesized Ni-P powder and commercial Ni powder, are spherical in shape. However, as compared to the electroless-synthesized Ni-P powder, the surface of commercial Ni powder is much coarser, as shown in the inset of Figure I1 (i). The particle size for electroless-synthesized Ni-P powder is measured from SEM imagines and plotted in Figure I1 fi) as a function of OK/Ni2+. Due to oxidization in air, the particle size for the powder synthesized at OH-/Ni2+=2 with Ni2+=0.4 M is not available. However, its size is believed to be smaller than that synthesized at OH-/Ni2+=2 with OH' =0.4 M, which is not oxidized in air at room temperature. In general, the average particle size is found to be accurately controlled from nanometer to micrometer by adjusting the ratio of OB/Ni2+ in aqueous solution. With the increase in the ratio of OH-/Ni2+, the particle size is decreased. The size is well approximated by the curve shown in Figure I1 6). However, when the concentration of the reaction reagents in aqueous solution increases, abnormal behavior is observed as the OK/Ni2+ ratio increases, such as the powder synthesized at OH-/Ni2+=1 .75 or 2 with Ni2+=0.4 M.

OH~INi2'=0.4 M10.2 M

OHINf+=0.4 MI425 M

OH7hJzi2*=0.5 ML0.4 M

Commercial Ni

I L

20 40 60 80 100 2 Theta

Figure I: XRD results for all studied Ni-P and commercial Ni powder.

103

2.8

2.4

2.0

1.6

1.2

0.8

0.4

1.0 1.2 1.4 1.6 1.8 2.0 OH-lNi"

Figure 11: SEM morphologies for (a)-(h) electroless-synthesized Ni-P powder and (i) commercial Ni powder. The scale bar in figures denotes 10 pm while 2 pm for the inset of (a) and (i).The article

sizes are obtained from SEM picture and listed in (j) according to the ratio of 0 H - N . R

Composites were prepared by mixing the electroless-synthesized Ni-P powder with epoxy at different volume fraction. The Ni-P powders studied here are synthesized at Ni2+=0.4 M with O K /Ni2+=1 and 1.75, i.e. Powders e and h in Figure 11. Composites filled with 10 vol%, 20 vol% and 30 vol% commercial Ni powder are also prepared. The complex permeability and permittivity are measured from 10 MHz to 1.0 GHz for all these samples and all results are shown in Figure III. For electroless-synthesized NbP powder, the density is unknown since it is determined by many factors, such as phosphorus concentration in the final powder or the detailed inner structure of the synthesized metallic particle. Hence, the weight concentration of Ni-P powder in composites is chosen to be same as that with 20 vol% commercial Ni powder.

500

400

% .- 300 ._ z

200

100

x

Q

0

1.2 x .- g 0.9

E 0.6 m L 2 0.3

0.0 0.01 0.1

Frequency (GHz)

-- I

15

' O I 5

0

1.2

0.9

0.6

0.3

0.0 0.01 0.1

Frequency (GHz)

100

10

3 2 1 0 3

2

1

0 0.01 0.1 1

Frequency (GHz)

Figure 111: Dispersions of permittivity and permeability for composites with the filler of electroless- synthesized Ni-P powder synthesized at Ni2+=0.4 M with (a) OB/Ni2+=1 and (b) OH-/Ni2+=l .75,

and (c> Commercial Ni powder at different volume fractions.

1 04

As expected, for composites with the commercial Ni powder as fillers, both permeability and permittivity are enhanced as the volume fraction of Ni metal in composites is increased from 10% to 30%. However, the permittivity is much higher than permeability in all cases, which degrades the impedance match between composites and free space. As we know, a good impedance match is necessary for good EM materials. Hence, nickel powder can serve as EM materials only when its permittivity is greatly reduced. Same results are observed in the composites with the electroless- synthesized Ni-P powder as fillers. For composites with Ni-P powder synthesized at OH-/Ni2+=1, its permittivity is much higher than that with 30 vol% commercial Ni powder, as shown in Figure I11 (a). Moreover, a very poor magnetic res onse is observed in both composites. For the composites with Ni-P powder synthesized at OH-/NiR=1.75, the real permeability is almost 1 over all measured microwave frequency, as shown in Figure I11 (b).

The results are attributed to the fact that the density of the synthesized Ni-P powder is much smaller than that of the commercial powder. When fabricating Ni-P powder, Ni(OH)2 is first deposited after mixing NaOH with NiS04 solution. The particle size of Ni(OH)2 is determined by the ratio of OH'/Ni2+ in the solution. Following redox reaction would happen at the Ni(OH)2 surface, and while the reaction is in progress, the Ni ion in Ni(OH)2 core will diffuse from the inner to the surface. As a result, the core becomes much more porous than the surface, which leads to a low density in Ni-P powder. In other words, the porous inner core can be regarded as an air bubble enclosed by Ni-P metallic layer, which increases the total volume fraction of metallic powder while decreases the real volume fraction of the magnetic phase. As a result, for composites with the same weight concentration of metallic powder, the permittivity should be higher in composites with Ni-P powder than that with commercial Ni powder, while the permeability is much lower, as shown in Figure I11 (a). For Ni-P powder synthesized at OH-/Ni2+=1.75, the average particle size is about 1.34 pm. The inner core is much smaller. As a result, the volume fraction is almost the same as the theoretical value, i.e. 20 ~01%. Therefore, its permittivity is almost the same as the composites with 20 vol% commercial Ni powder. The low permeability observed in this composite may be attributed to the poor magnetic properties in the electroless-synthesized Ni-P powder.

4. Conclusions Ni-P powder with size varying from nanometer to micrometer is successfully fabricated using

modified electroless plating synthesis. The particle size is simply controlled by adjusting the ratio of OH/Ni2+ in aqueous solution. EM results indicate that both commercial Ni powder and electroless- synthesized Ni-P powder are not suitable for application as EM materials because of its poor magnetic properties, high permittivity at microwave. The porous core structure in the electroless- synthesized Ni-P powder also degrades the EM performances due to impedance mismatch between composites and free space. For application as EM materials, it is necessary to reduce greatly the permittivity and simultaneously, increase the permeability in composites with metallic particles as fillers.

Reference [ l ] G. Viau, F. Ravel and 0. Acher, J. Appl. Phys. 76,6570 (1994). [2] Y.B. Feng, T Qiu, C.Y. Shen and X.Y. Li, IEEE Trans. Mum. 42,363 (2006). [3] S. Yoshida, M. Sato, E. Sugawara and Y. Shimada, J. Appl. Phys. 85,4636 (1999). [4] Y. Nie, H.H. He, R.Z. Gong and X.C. Zhang, J. Magn. Magn. Mater. 310,33 (2007). [5] A. Brenner and G.E. Riddell, J. Res. Nutl. Bur. Standards 37,3 1 (1946). [6] S.F. Moustafa and W.M. Daoush, J. Muter. Process. Tech. 181, 59 (2007). [7] A. Roy, V. Srinivas, S. Ram, J.A. De Toro, and J.P. Goff, J. Appl. Phys. 100,094307 (2006). [S] P. Sampath Kumar and P. Kesavan Nair, J. Mater. Process. Tech. 56, 5 1 1 (1 996). [9] H.X. Li, W.J. Wang, H. Li, and J.F. Deng,J. Catal. 194,211 (2000).

Nanosized Ferrite Ceramics Derived from High-energy Milled Powders with Promising Magneto-dielectric Properties over 30-90 MHz

L. B. Kong, Z. W. Li, G. Q. Lin and Y. B. Gan Temasek Laboratories, National University of Singapore, 10 Kent Ridge Crescent, Singapore

119260

Abstract Nanosized Ni0.70Zn0.2sCoo.osFe1 .90Mno.0204 ferrite ceramics, with average grain size of -200

nm, were derived from high-energy ball milled mixtures with various ratios of FezO3/Fe as starting materials. The ferrite ceramics were obtained by sintering the milled powders at 800°C for 8 h. Linear expansion of the green pellets is less than 4%, which is thus a near net-shape processing. DC resistivity, dielectric and magnetic properties of the ferrite ceramics were investigated and compared with their counterpart (with the same composition) prepared via the conventional ceramic process. Magnetic properties of the ceramics were well explained by the magnetic circuit model and Snoek-like law. In addition, the sample derived from the mixture of 50%-Fe203 possessed promising magneto-dielectric properties, with matching values of real permeability and permittivity, as well as relatively low magnetic and dielectric loss tangents, over 30-90 MHz (VHF band). This material could be a potential candidate for miniaturization of VHF antennas.

1. Introduction Miniaturization of antennas has been a challenge to designers, especially for HF (3-30 MHz)

and VHF (30-90 MHz and 100-300 MHz) bands where conventional antennas have rather large physical sizes. Magneto-dielectric materials, with matching permeability and permittivity and sufficiently low magnetic and dielectric loss tangents, could be potential candidates to reduce the physical dimensions of antennas, with their electrical dimensions maintained [ 11. In this paper, we demonstrated that nanosized ferrite ceramics, with promising magneto-dielectric properties over 30- 90 MHz, can be achieved by proper material processing. One of the advantages is that the method is a near net-shape processing, where the final products have almost unchanged dimensions as compared to their corresponding green compacts, making it very useful in practical applications.

2. Experimental The ferrite composition was Nio.70Zn0.2&oo.o~Fe1 90Mno.0204. Commercially available Fe

(99+% purity, Aldrich Chemical Company Inc., USA), Fez03 (99% purity, Aldrich Chemical Company Inc., USA), MnO2 (98% purity, Aldrich Chemical Company Inc., USA) and C03O4 (99+% purity, Aldrich Chemical Company Inc., USA) powders, were used as starting materials. Various compositions of Fe and Fe203 were studied. The percentages of Fez03 [lOO%Fe203/(Fe2+Fe203)] were 10, 20, 30 and 50%. The starting materials were mixed by high- energy ball milling for 12 h. The high-energy milling was conducted using a Retsch PM400 type planetary ball milling system. A 250 ml tungsten carbide vial and 100 tungsten carbide balls with diameter of 10 mm were used as a milling medium. The milling speed was set at 200 rpm. The milled powders were then compacted and sintered at 800°C for 8 h. For comparison, a sample with the same composition was prepared via the conventional ceramic process using all-oxide precursors. Desired amount of oxides were mixed and calcined at 1000°C for 2 h. The calcined mixture powder was then sintered at 1250°C for 2 h. A comparison micro-sized and nano-sized samples is conducted.

Two types of samples, namely disk (diameter of -10 mm and thickness of -1.5 mm) and coaxial cylinder (outer diameter of -20 mm, inner diameter of -10 mm and thickness of -24 mm), were prepared. Disk samples were used in the measurement of permittivity and DC resistivity, while cylinder samples were used in the measurement of permeability. Reaction and densification behaviors of the samples were monitored using a Setaram Setsys 16/18 type dilatometer at a heating rate of 10"C/min in air. Phase compositions of the mixed, milled, calcined and sintered samples

105

106

were analyzed using a Philips PW 1729 type X-ray diffractometer (XRD) with Cu K, radiation. Grain size and grain morphology of the sintered samples were examined using a JEOL JSM-6340F type field emission scanning electronic microscope (FESEM). Densities of the ferrite ceramics were derived from the masses and dimensions of the samples. DC resistances of the sintered samples were measured using a multimeter. DC resistivities of the samples were then calculated based the resistances and sample dimensions. The complex relative permeability and permittivity of the ceramics were measured using the Agilent E4991A RF impedance/materials analyzer over 1 MHz - 1 GHz.

3. Results and Discussion The reaction and densification behavior of the sample with 10% Fez03 is shown in Fig. 1. The

shrinkage, which maximizs at -5OO"C, is attributed to the burning of polymer binder used in the compact of the green pellets. Above this temperature, the sample starts to expand and the expansion becomes slow after about 800°C. At about 1050"C, shrinkage is observed, and the linear dimension remains larger than the initial size even above 1200"C, because it is a dynamic experiment. The expansion is caused by oxidation of the Fe component used in the precursor. Similar behavior was also observed in other samples, with the maximum expansion decreasing with increasing percentage of FezO3. Based on this result, 800°C was chosen as the sintering temperature to obtain our ferrite ceramics. XRD patterns showed that phase-pure spinel ferrites were achieved, except for the sample with 10% Fe203, in which a trace of secondary phases is detected (results not shown). This result

Fig. 1. Linear shrinkage of the lO%-Fe2O3 sample

Fig. 2. Cross-sectional SEM images of the samples: (a) lo%, (b) 30%, (c) 50% and (d)

micro-sized sample.

10'

1 o2 ._ 0 .t 10'

E f: ._

2 l o o

10-l

l o 2 1 oB 1 0' 1 o8 1 og 1 o6 1 o7 1 o8 1 og

Frequency (Hz) Frequency (Hz)

Fig. 3. Complex relative permittivity curves of the ferrite ceramics: (a) lo%, (b) 20%, (c) 30%

Fig. 4. Complex permeability curves of the ferrite ceramics: (a) lo%, (b) 20%, (c) 30% and

and (d) 50%. (d) 50%.

107

The cross-sectional SEM images of representative nano-sized samples, as well as the micro- sized sample, are shown in Fig. 2. The sample derived from the mixture with 10% Fez03 comprises relatively large grains, with a wide grain size distribution (from 0.5 to 1.5 pm). The samples with higher levels of Fe2O3 exhibit a similar microstructure, with an average grain size of 100-200 nm. Nano-sized pores are distributed uniformly in the samples. As the level of Fe203 increases pores appear to be more uniform. The measured densities and the final linear expansion (calculated from the diameters before and after sintering) of the samples indicated that the density decreases monotonically with increasing level of Fez03 powder in the starting mixture. The sintered samples have almost unchanged dimensions as compared to their green pellets, which means that our method is like a near net-shape process. Samples prepared in this way will not suffer dimensional shrinkage and deformation, which could be an advantage for practical applications.

ln

8 J

105

1 00 30 45 60 75 90

1.00 30 45 60 75 90

Frequency (MHz)

0 025

0 020

0015

0 010

0 005

0 000

30 45 60 75 90

Frequency (MHz)

Fig. 5. Magneto-dielectric properties of the 50%-Fez03 sample over 30-90 MHz.

1 o6 10' 1 o8 10' Frequency (Hz)

Fig. 6 . Complex relative permeability curve of the micro-sized sample.

16

14 I"

12

10

---- Calculated . o Experiment Q .

b.

0.05 0.06 0.07 0.08

6ID

Fig. 7. Static permeability versus 6/D of the samples.

Complex relative permittivity and permeability curves of the sintered samples are shown in Fig. 3 and Fig. 4, respectively. The extremely high permittivity of the 10% and 20% Fez03 samples is probably because Fe is not oxidized completely to Fe3+. A trace of Fez+ ions could greatly increase the permittivity, because Fez+ has a larger polarization than Fe3+. This explanation is supported by permeability measurement, especially for the 10% sample, whose complex permeability curve is very similar to composites based on metal powders and polymers [2]. The sample with 50% Fez03 possesses good magneto-dielectric properties over 30-90 MHz, as shown in Fig. 5. This result means that the physical size of antenna can be reduced by a factor of -10 by when loaded with this kind of materials. It is noted that the dielectric loss tangent is essentially below1 O-', as required by practical applications, while the magnetic loss is slightly higher. This property, as well as the impedance, can be further improved by adjusting the processing parameters and material compositions.

108

Due to the presence of porosity, magnetic poles are created on the surface of ferrite grains (or particles) under an applied magnetic field, thus producing a demagnetizing field, which leads to a decrease in permeability. The magnitude of the demagnetizing field is closely related to the grain size and grain boundary characteristics. According to the magnetic circuit model [3], the static permeability po is given by:

6

6 P, (1 + 5’ 1 + P , 5

(1) Po =

where p, is the intrinsic static permeability of materials without any defects, D and 6 are the grain size and the thickness of grain boundaries (including porosity and other non-magnetic phases), respectively. With this formulation, the variation of po with the composition of starting material can be easily understood. As the concentration of Fez03 is increased, the decreased grain size and density (p) lead to an increased ratio of 6/D, and thus reducing static permeability. The ratio of 6/D can be approximately calculated from the measured density based on the following formula [4]:

where pi is the theoretical density of materials. Here, pi is the density of the microsized sample. By applying least-square method to Eq; (l), we can calculate the static permeability po for samples with various concentration of Fe203. The simulated intrinsic static permeability is very close to the value of the microsized sample, as shown in Fig. 6. The curve-fitted results, as shown by the dashed-lines in Fig. 7, are consistent with the experimental values for the ferrites.

4. Conclusions Nanosized Ni~.~0Z~.2~Coo.o~Fe1.90Mno.o204 ferrite ceramics, with average grain size of -200

nm, were derived from high-energy ball milled mixtures with various ratios of FezOJFe as starting materials. The ferrite ceramics were obtained by sintering the milled powders at 800°C for 8 h. Linear expansion of the green pellets is less than 4%, which is thus a near net-shape processing. Dielectric and magnetic properties of the ferrite ceramics were investigated and compared with their counterpart (with the same composition) prepared via the conventional ceramic process. Magnetic properties of the ceramics were well explained by the magnetic circuit model. In addition, the sample derived from mixture of 50%-Fe203 possessed promising magneto-dielectric properties, with matching values of real permeability and permittivity, as well as relatively low magnetic and dielectric loss tangents, over 30-90 MHz (VHF band). This material could be a potential candidate for the miniaturization of VHF antennas.

Acknowledgements One of the authors (L. B. K.) would like to thank Dr T. S. Zhang (School of Materials Science

and Engineering, Nanyang Technological University, Singapore) for his assistance in the measurement of dilatometer.

References [ 11 [2] [3] [4]

H. Mosallaei and K. Sarabandi, IEEE Trans Antennas Propagat. 52, 1558 (2004). L. B. Kong, Z. W. Li, G. Q. Lin and Y. B. Gan, unpublished work. M. T. Johnson and E. G. Visser, IEEE Trans. Mug., 26 [ 5 ] , 1987 (1990). T. Nakamura, T. Tsutaoka and K. Hatakeyama, J. Mug. Mag. Muter., 138,319 (1994).

Thermostable Magnetic Elastomers filled with Carbonyl Iron

Madina A. Abshinovall, Ivo Kuritka', Natalia E. Kazantseva'.', Jarmila VilEtikovB', Petr Sahal 'Polymer Centre, Faculty of Technologi., Tomas Bata University in Zlin, Czech Republic

Institute of Radio-Engineering and Electronics, Russian Academy of Sciences, Russia *email: abshinova@,fi. uth. cz

2

Among different magnetic fillers, carbonyl iron (CI) is a promising material for components of electronic devices and electromagnetic wave absorbers [l-31. However, CI composites do not always possess the temporal and thermal stability, the main factors determining performance reliability of electronics at continuous exploitation and storage. A thorough analysis about the causes of aging of the composites under exposure to temperature is necessary to enhance stability of CI filled polymer composites.

The frequency dispersion of the complex permeability p*@ of the CI composites is determined by several parameters, mainly by microstructure, shape and size of CI particles and volume loading of CI. Variation of these parameters enables to change the absolute values of permeability as well as the frequency dispersion of p*. The decrease of the CI concentration in the composite reduces the absolute value of p* and leads to the resonance frequency shift associated with the magnetic discontinuity in the composites. From the practical point of view, highly filled composites are of special interest [l-41. In this case, the microstructure of CI particles plays a significant role in the frequency variation of p* [4].

One of the main features of CI filled composites is thermal stability of p* due to the significant demagnetization fields in the isolated particles and CI microstructure features [ 5 , 61. Thermal stability of physicomechanical characteristics of CI filled composites is connected with the surface chemistry of CI [7]. Commercially available CI powders are modified by different types of agents in order to isolate each particle. Depending on the insulating layer composition, it could be inert or act as inhibitor or initiator of the polymer matrix destruction. Therefore, it is necessary to consider the surface chemistry of fillers for the development of thermostable CI filled polymer composites, like polysiloxane based elastomers.

Durability of the composites under thermal load is of prime importance for any prospective application where the material is exposed to high temperatures during its manufacture or use. Polysiloxanes degrade in several steps including depolymerisation via cyclooligomerization, and crosslinking of polymeric chains [S]. There are several approaches to thermal stabilization of polysiloxane based materials, particularly use of certain micro-fillers which are able to block the end of silanole groups and macroradicals on the surface and, thus, process a non-chain inhibition of degradation [5].

In order to obtain frequency-stable magnetic materials, we have concentrated our attention on the studies of the structural changes responsible for the temperature variations of electromagnetic and physicomechanical properties in the CI filled polymeric composites. Thus, we consider CI powders from BASF (Germany) and SINTEZ (Russia), differ in morphology, microstructure, and physico-chemistry of surface: EW, SQ coated by SiOz and KM-2 coated by A1203; ES, HQ, SL, MCI without surface modification. Silicone elastomer (SYLGARD 184, Dow Coming, USA) with an operating temperature range from -55°C to 200°C was chosen as a polymer matrix material.

Elastomers containing 10 and 50 vol. % of CI were prepared by mixing of the components and curing in the vacuum at 65°C-100°C degrees for 4h-lh. In order to measure complex permeability, toroidal specimens of inner diameter of 3.1 111111, outer diameter of 8 111111, and a thickness of 3 mm were cut from the composite sheets.

The temperature-frequency dependences of p* of CI composites were measured in the range from 1 MHz to 3 GHz by an RF Impedancehlaterial Analyzer (Agilent E4991A) and in the range from 2 GHz to 10 GHz by resonance method in combination with ESPEC SU-241 temperature chamber within the temperature range was from -30°C to +15O"C.

109

110

The sam les were examined under inert atmosphere of He (5.5 purity, SIAD TP), the gas flow was 0.5 C ~ ~ S ~ ' at normal conditions (30 sccm) by a Thermogravimeter SETARAM SETSYS Evolution 1200. Temperature growth rate was 20"C/min in the range from ambient temperature to 1200°C.

Fig. 1 shows the temperature variation of the real 01') and imaginary part 01") of permeability spectra for composites filled with 50 vol. % of SL and MCI. Remarkable shift of the permeability dispersion region to the microwave range is observed for all composites with increasing temperature up to 150 "C (maximum operating temperature of electronics). Ferromagnetic resonance frequency (f~) for SL composites shifts from 2.7 GHz to 3.2 GHz, however, absolute values of magnetic losses in resonance (p'&) increases from 4 to 5.5; whereas in the frequency range from lo7 to lo9 Hz magnetic losses are decreased. For the composites based on MCI, changes within 2.3-2.7 GHz. However, when compared with SL-type composite, p'La slightly decreases from 10 to 9 with increasing temperature. The largest changes in frequency dispersion character of permeability are observed for composites filled with KM-2 (Fig. 2). In this case, both the real part of complex permeability p' and magnetic losses in resonance p'La increase with temperature in the microwave range. The heat sensibility of magnetic properties, according to X-ray analyses, is connected with the structural changes of CI particles, specifically with a-iron lattice stresses, a-iron crystallites growth, carbides and nitrides impurities coagulation, and carbon and nitrogen content alternation in a-iron [6].

Thermostability of magnetic elastomers is determined by thermal coefficient of permeability (TC,), which is calculated by variation of initial permeability in the temperature range of (T2 - TI):

1 2 ,

10' 1 o8 109 10'0

Frequency, Hz

( 4

10' 108 1 o g 10'0

Frequency, Hz

(b)

Fig. 1. Temperature variation of complex permeability spectra for composites with 50 vol. % of SL (a) and 50 vol. % of MCI (b).

111

Fig. 2. Temperature variation of complex permeability spectra for composites with 50 vol. % of KM-2.

According to this approach, TC, of the investigated composites at lo7 Hz has small values, varying slightly from 0 . 3 4 ~ 1 0 - ~ to 7.5 x ~ O - ~ (l/"C) for KM-2 and MCI respectively. In the frequency range from 1 MHz to 10 GHz composites with KM-2 have positive value of T C , while composites based on MCI have negative values of TC,. For SL-type composites TC, changes from negative to positive approximately at 400 MHz. On the whole, it confirms that investigated composites are thermostable. However, it should be mentioned, that such estimation of thermostability is not completely correct, as far as it does not consider the behaviour of the imaginary part of complex permeability (magnetic losses). In particular, it concerns the resonance frequency shift with rise in temperature. This is the prime important parameter for the development of frequency-stable electronic devices. From this view point, only KM-2 provides required stability of frequency with temperature.

The results of non-isothermal TG experiments are shown in Fig. 3. Graph (a) shows the weight loss in dependence on temperature, graphs (b), (c) and (d) show the weight loss rate (dTG) as a function of temperature for materials SL, MCI, KM-2, in 10 and 50 vol. % concentrations. The weight loss rate curve of pure matrix is shown in all three graph windows to make the comparison easier. Inflexion points on integral TG curves correspond with minima in graphs (b), (c), and (d).

The elastomeric matrix shows no significant TG effects below 200°C (in Fig. 3, graph (a)) as declared by the producer. The pure elastomeric material degrades in two steps. The first one is well defined single step with corresponding single peak on dTG curve (see any of graphs (b), (c), (d)). Thus, prevailing degradation mechanism is the cyclooligomerization. The second step is connected with two competing processes of cyclooligomerization and crosslinking degradation. Similar shape of dTG curves can be observed for all 10 vol. % samples, i.e. SL-10, MCI-10 and KM2-10. The unspecific stabilization manifested as decrease in degradation rates and small shift of the corresponding dTG curves towards higher temperatures can be interpreted as a general effect of particle filler influence on transport phenomena in composites.

Dramatic enhancement of the stabilization effect is observed for 50 vol. 'YO composites of SL and MCI. The weight loss rate is reduced ten times or more at this high filler load. The onset temperature corresponding to the initial 2.5% (TO.025) weight loss is shifted from 355°C for pure matrix to 600°C for SL and 536°C for MCI. It can be assumed that an additional stabilization mechanism is present in SL composite in compare to MCI. On the other hand, the 50 vol. % KM-2 material shows different degradation pattern. The first degradation step proceeds faster than it is in material with lower content of filler. The first peak is similar to that of pure matrix and its onset temperature (T0.025) is only 474°C. Hence, a mechanism accelerating depolymerization is present in the first degradation step, most likely due to water and hydroxyl groups adsorbed on A1203 coating of KM-2 filler. The disappearance of the second stage degradation step in the dTG testifies for hindrance of already running unzipping reaction with most likely forced crosslinking reactions between polysiloxane chains and A1203 coating on the filler surface. These reactivity properties of

112

KM-2 open up the way to further improvements of composite stability by filler surface activation by drying or chemical modification.

Composites filled with KM-2 characterized by onion structure exhibit higher thermomagnetic stability in comparison with SL and MCI composites having disrupted particle structure. Whereas, the absolute value of the TC,of composites is the highest for the latter one. Among the examined CI filled composites, the highest frequency-stable property was shown by elastomer with KM-2. However, as to thermal stability of investigated composites, a destabilization effect of KM-2 filler reactive surface was observed at the low temperature step of degradation. A stabilization effect based on chemical interaction of surface and polysiloxane chains is concluded for the SL and at higher temperatures also for KM-2 filler. MCI filler shows moderate stabilization effect. Consequently, by variation of the microstructure and surface modification of CI particles it is possible to take under control thermomagnetic instability of CI filled elastomers.

-20

' -40 \ siliconeelastomer I b a t r i x material) -50 - graph (a)

(matnx material)

W .- .5 : graph (b)

t . 1 . t . t . t . t . t l

o 200 400 600 am 1000 1200

Temperature, ' C

0 200 400 600 800 1000 1200

Temperature, "C

Fig. 3. Thermogravimetric analysis of polysiloxane elastomer matrix and various CI composites.

References [ l ] S. Picos, Latvian Journal of Physics and Technical Sciences, 2, 33 (1999). [2] Y. Nie, H. He, 2. Zhao, R. Gong, and H. Yu., J. Magn. Magn. Mater., 306, 125 (2006). [3] Y.-B. Feng, T. Qiu, C.-Y. Shen, X.-Y. Li, IEEE Trans Magn., 42,363 (2006). [4] V.S. Volkov, V.G. Syrkin, and I S Tolmasskii, Carbonyl Iron, Metallurgy, Moscow, 1970. [ 5 ] A. T. Ponomarenko, C. Klason, N.E. Kazantseva, M.I. Buzin, M. Alexandre, Ph. Dubois, I.A.

Tchmutin, V.G. Shevchenko, and R. Jerome, Journal of Thermal Analysis and Calorimetry, 55, 537 (1999).

[6] K.H. Chung, C.S. Wu, and E.G. Malawer, Thennochimica Acta, 154, 195 (1989). [7] T.H. Thomas, T.C. Kendrick, Journal of Polymer Science Part A-2: Polymer Physics, 7(3), 537

[S] G. Camino, S.M. Lomakin, and M. Lageard, Polymer, 43(7), 201 1 (2002). (1 969).

Development of Magneto-dielectric Materials based on Lithium Ferrite Ceramics for Miniaturization of Antennas

M. L. S. Teoa3 b, L. B. Kongb, Z. W. Lib, G. Q. Linb and Y. B. Ganb "Hwa Chong Institution (College), 661 Bukit Timah Rd, Singapore 269734.

'Temasek Laboratories, National University of Singapore, I0 Kent Ridge Crescent, Singapore 119260

Abstract This paper presents our study on the densification, grain growth, complex permeability and

permittivity of Li-ferrite (Li0.50Fe2.5004) ceramics. Our objective is to obtain magneto-dielectric materials for miniaturization of HF (3-30 MHz) antennas. Biz03 was employed as sintering aid to reduce the sintering temperature of Li0.50Fe2.5004, while Co was used to further modify the permeability of the ceramics. Desired magneto-dielectric properties over 3-30 MHz were achieved in samples with Co concentrations of x=0.030-0.035, which makes them potential candidates for miniaturization of HF antennas.

1. Introduction Conventional antennas for the frequency band of 3-30 MHz (HF) and 30-300 MHz (VHF) are

physically large, and therefore not suitable for portable applications. A pertinent challenge is to reduce the physical dimensions without affecting its electrical performances. From the laws of physics, it is potentially possible to use material loading to scale down the antenna's physical dimension by a factor of n (refractive index of material), with its electrical dimension unchanged. A class of materials that serves this purpose is the magneto-dielectric materials with high refractive index ( n = m, where p ' is relative permeability and E ' is relative permittivity) [ 11, and almost

matching p'and E' ( Z = , / m =vom =q,,, where ~0 is impedance of free space). The latter condition is particularly important for matching the antenna impedance to free space environment. Moreover, the materials must have sufficiently low magnetic and dielectric loss tangent (< 1 0-2) to minimize losses in the antennas.

Lithium ferrite (Lio.soFez.5004) was selected in this study, because it has high resistivity and good magnetic and dielectric properties. However, Li0.50Fe2.5004 can only be fully sintered at high temperatures, resulting in high dielectric loss. To reduce sintering temperatures, Biz03 is added as sintering aid [2]. It is found that although 1% Biz03 is sufficient for full densification of Li0.50Fez.5004 over the temperature range studied, a higher concentration of 3% is necessary to result in low dielectric loss tangent required for practical applications. Furthermore, Co was used to further modify the magnetic properties of Lio.5oFe2.5004. Good impedance matching, together with refraction index of 13-15 and sufficiently low magnetic and dielectric loss tangents, can be attained in the samples with Co concentration of about 0.030-0.035.

2. Experimental Procedure The solid-state reaction process was employed to synthesize all ferrite samples. We used

commercially available chemical powders, Li2CO3, FezO3, C03O4 and Biz03 as starting materials. In this project we studied two groups of samples: Lio.soFe2.5004 + Biz03 (with weight concentrations 0%, 1%, 3% and 5%) and Lio.50.o.~oxCoxFe2.5~-~,~~~O4 (with x=O.Ol-0.07). All samples were made with 1% Fe deficiency to reduce the formation of Fez+ ions, in order to maintain high DC resistivity and low dielectric loss tangents. First, the oxides of desired amounts were mixed using high-energy milling for 2 h and calcined at 900°C for 2 h in air. The sintering aid, Bi203, was then added to the powders and milled for 2 h before sintering at temperatures ranging from 850°C to 1100°C (at 50°C intervals) for 2 h in air. The heating and cooling rates were set to 5"CIm for both calcination and sintering. Disc samples (diameter -10 mm and thickness -1.8 mm) were made to measure DC resistivity and permittivity, while coaxial cylinders (outer diameter -20 mm, inner diameter -10 mm and thickness -3 mm) were prepared for measurement of permeability.

113

114

A Philips PW 1729 X-ray diffractometer (XRD) with Cu K, radiation was used to analyze the phase compositions of the samples. Densification behavior was studied using a Setaram Setsys 16/18 dilatometer at a heating rate of lO"C/minute in air. Observations of the samples' microstructure and morphology of grain sizes were made using a JEOL JSM-6340F field emission scanning electron microscope (FESEM). The densities of the samples were calculated from their masses and volumes. DC resistivity was calculated from the DC resistance of the disc samples, which was measured using a multimeter at room temperature. An Agilent E4991A RF impedance/materials analyzer was used to measure the complex permittivity and permeability of the ceramics over 1 MHz - 1 GHz.

3. Results and Discussion 3.1. Phase composition, densification and grain growth

Fig. 1 shows the XRD patterns of the Lio.5oFez.5004 ceramics with various concentrations of Biz03 sintered at 850°C for 2 h. No obvious phases other than spinel structure can be found in the XRD patterns. Similar results have also been attained for samples sintered at higher temperatures. It is also found that substitution of LiFe with Co has no effect on phase the formation of the ferrite ceramics in the temperature range used in the present study.

-3

Y

-1.2 : (I)

Linear shrinkage

-1.5 600 750 900 1050 1200

Temperature ("C)

Fig. 1 : XRD patterns of the Li0.50Fe2.5004 Fig. 2: Sintering behavior of the Lio.soFe2.5o04 ceramics sintered at 850°C. powders with/without BizO3.

As anticipated, the densification behavior of Li0.50Fe2.5004 was greatly improved by Bi2O3. Representative sintering curves of the powders with and without Bi2O3 are shown in Fig. 2. Without Bi2O3, the sample started to shrink at -9OO"C, with a maximum shrinkage rate of 0.36%/"C and a final shrinkage of 7.4%. The addition of 3% Biz03 reduced the initial shrinking temperature (-760°C) and enhanced shrinkage. However, further increase in Biz03 content does not lead to additional improvement in the sintering characteristics of Li0.50Fe2.5004.

Representative SEM images of the Li0.50Fe2,~~O~ ceramics are illustrated in Fig. 3. The presence of Biz03 not only improved the sintering property but also enhanced the grain growth of Lio.5oFe2.5o04. For example, pure Li0.50Fe2.5004 ceramics (sintered at 1 000°C) possess spherical grains, with obviously porous microstructure and an average grain size of -1.8 pm, whereas samples with 1% Biz03 have fewer pores and much larger grain sizes (-12.9 pm). It is noted that the variations in grain size with sintering temperature for the samples with Biz03 of 1-5% are very similar, but the grain size decreases slightly with increasing amount of Bi2O3. In addition, the measured density of the samples further confirms the effectiveness of Biz03 in improving the densification property of Li0.50Fe2.5004 ceramics.

The good densification behavior of Li0.50Fe2.5004+Bi203 samples is because Bi2O3 has a low melting point. During sintering at high temperatures, a liquid phase was formed, which facilitates easier rearrangement of the Lio.5oFe2.5004 grains as compared to those in pure Lio.SoFe2 5 0 0 4 . This rearrangement of grains caused the pores of air to coalesce and escape, which improved densification. The presence of a liquid phase also promoted mass transportation and thus facilitated rapid grain growth. Above critical concentration, an increase in the amount of Biz03 causes an increase in the thickness of the liquid phase layer. During sintering, smaller grains move through the liquid phase layer to form larger grains. Due to the presence of BizO3, the separation between

grains increased, resulting in a larger diffusion path for small grains to combine into larger ones. This explains the slight decrease in grain sizes of samples with higher concentrations of Bi203.

*

Fig. 3: SEM images of the Lio 5 o F ~ 5 0 0 4 ceramics sintered at 1000°C for 2 h: (a) without and (b) with 1% of Bi203.

The above results indicate that 1% Biz03 is sufficient for the densification of Lio.5oFe2.5004 ceramics, but higher concentration of Biz03 is required to attain low dielectric loss tangent. This is the reason that 3% Biz03 was used in the Co group. It is found that the effect of Co on the microstructure and grain growth is much less significant than that of BizO3, which therefore is not discussed in this paper.

3.2. Complex relative permittivity Typical complex relative permittivity curves, over 1 MHz - 1 GHz, are shown in Fig. 4. The

addition of 3% Biz03 has effectively reduced the loss tangent. At the same time, the doping also led to high real permittivity. The reduced dielectric loss tangent and increased real permittivity are attributed to the improved densification. The high loss tangent of the samples sintered at 1000°C and above is probably due to the formation of Fez+ ions. It is observed that L ~ o . ~ o - o . ~ o ~0 . 5 0 ~ 0 4 ceramics have similar permittivity properties to the Li0.50Fe2.5004 samples, because the amount of Co is not sufficiently high (50.07). This allows promising magneto-dielectric properties to be achieved by adjusting the permeability.

15

12

1 o6 1 0’ 1 o8 1 o9 Frequency (Hz)

Fig. 4: Representative complex permittivity curve of Lio.~0-o.~o~Co~Fez,~0-0.~0~04 ceramics (x=0.035) sintered at different temperatures.

3.3. Complex relative vermeabilitv

1 o6 10’ 1 o8 10’ Frequency (Hz)

Fig. 5: Complex permeability curve of Lio 50-

0 5oxCoxFez 50-0 5 0 ~ 0 4 ceramics, with different concentrations of x, sintered at 900°C for 2 h.

Alfcomplex pekeability Eurves show that the real permeability remains constant up to a certain frequency, where it peaks and then decreases. The imaginary permeability is usually much lower before the resonance that occurs at higher frequencies, as demonstrated in Fig. 5. Static permeability is defined as the permeability at frequency far from the resonance. As the static permeability increases, resonance is shifted to lower frequencies. This is given by Snoek’s Law [3]. Detailed discussion is not within the scope of our present study. The dependence of permeability on Biz03 and sintering temperature is similar to that of permittivity.

116

Fig. 6 shows the static permeability of L ~ o . ~ o - o . ~ o ~ C O ~ F ~ ~ . ~ O ~ O . ~ Oceramics as a function of Co concentration. For a given sintering temperature, the static permeability increases as the amount of Co is increased from x=O to x=O.Ol, maximizes at x=O.Ol, and then decreases almost exponentially. For example, the static permeability of the sample sintered at 950°C increased from 48.3 (x=O) to 65.3 (x=O.Ol) before dropping to 16.2 (x=0.03). A possible reason is that Co ferrite has positive magnetocrystalline anisotropy while that of Li ferrite is slightly negative [4]. The addition of Co initially neutralizes the negative magnetocrystalline anisotropy; on further addition, it becomes positive. Since permeability is inversely proportional to the absolute value of magnetocrystalline anisotropy, the peak in permeability corresponds to the lowest absolute value of magnetocrystalline anisotropy .

100

80

-3 3 60 E ._ -

E 40 & a 20

0

05 006 007 - 850°C + 900% - 950°C - 1000°C 1O5O0C --f.-llOO°C

J I I I

.OO 0.02 0.04 0.06 Concentration of Co

E 8

8 4 f b 4 0

5 10 15 20 25 30 5 10 15 20 25 30 0.020, , 1.M,

-3 I 101

Fig. 6: Real part of permeability of L~o.so-o.~o~CO~F~~.~O-O,~O~O~ ceramics with 3% of

Bi2O3 as a function of x.

Fig. 7: Magneto-dielectric properties of Li0.50- 0.50~Co~Fe2.50-0.50~04 ceramics (x=0.030)

sintered at 850°C.

3.4. Magneto-dielectric properties It is found that samples with Co concentration of x=0.030-0.035 are potential candidates for

magneto-dielectric materials. Fig. 7 shows the magneto-dielectric properties of 850°C-sintered sample with x=0.030. It possesses highly matching permeability and permittivity over 3-30 MHz. On top of this, its refractive index is as high as n-14, while the dielectric and magnetic loss tangents are both less than or very close to Hence, this class of materials is potential candidate for miniaturization of HF antennas.

4. Conclusions The densification behavior and grain growth of Li0.50Fe2.5004 ceramics were improved

considerably on addition of Bi2O3, because Bi2O3 forms a liquid phase at low temperatures. Although 1% Bi2O3 is sufficient for full densification of Lio.5oFe2.5004 ceramics at temperature as low as 850"C, 3% is required for low dielectric and magnetic loss tangents.

Li0.~0-0.~0~Co~Fe~.50-0.50x04 samples with Co concentrations of x=0.030-0.035 displayed matching values of real permittivity and permeability over 3-30 MHz. With good impedance matching, high refractive index and low dielectric and magnetic loss tangents of less than or close to these materials are believed to be useful for reduction of the physical dimensions of HF antennas.

Acknowledgement One of the authors (M. L. S. T.) would like to thank the Defence Science and Technology

Agency (DSTA), Singapore for providing the opportunity to participate the Young Defence Scientist Programme (YDSP).

References [ l ] H. Mosallaei and K. Sarabandi, ZEEE Trans. Antennas Propagat. 52, 1558 (2004). [2] M. Drofenik, A. Znidarsic and D. Makovec, J. Am. Ceram. Soc., 81,2841 (1998). [3] J. L. Snoek, Physica, 14,203 (1948). [4] T. Y. Byun, S. C. Byeon, K. S. Hong and C. K. Kim, ZEEE Trans. Mag., 35,3445 (1999).

Session P6

Chair: A.N. Lagarkov

Magnetic and Acoustic Metamaterials

S.A. Nikitov, S.E. Bankov, Yu.A. Filimonov, A.V. Grigorievskiy, V.I. Grigorievski, S.L. Vysotskii

Institute of Radioengineering and Electronics, Russian Academy of Sciences Mokhovaya str. 11, Bld. 7, Moscow 125009, Russia

Abstract In this paper we present the results of theoretical and experimental investigation of

propagation of magnetostatic spin wave (MSW) in magnetic films with periodic two-dimensional arrays of slots and strips in adjusting metal layer or in films with periodic structures of etched holes. Magnetostatic spin waves (both, volume and surface) propagating in thin magnetic films with different magnetization are considered. In first case, investigated structure consists of several layers including magnetic film, film with two-dimensional planar array and dielectric layer between them. Dielectric substrate with metal screen also may be included. Boundary problem for the structure is formulated and solved in a quasi-static approximation. The boundary problem is reduced to a solution of the set of integral equations reduced to the electric or magnetic currents flowing along strips and slots respectively. It is shown that the two-dimensional array may be effectively used for control of MSW parameters including stop bands formation etc. In second case forbidden gaps are found in the MSW spectrum and specific surface domain states are described due to magnonic structure of the spectrum. Along with MSW propagation in magnetic film we studied also elastic waves propagation in phononic crystals. Band gap structure of such crystals is calculated and reflection and transmission pecularities in wave spectrum are found.

119

120

1. Introduction During the last decade considerable efforts have been made in the sciences and technology for controlling or engineering the optical properties of the materials. For example, a number of artificially arranged materials were engineered to facilitate light propagation in particular direction or in specific regions only. Such materials also enable light to be localized in chosen channels or zones, or even prohibit the propagation of light completely. They are now known as photonic crystals [l]. Generally speaking, the photonic crystal (PC) is a material that possesses periodic index of refraction. A simple example of photonic crystals, also known as one-dimensional (1 - D) PC is a multilayered periodic structure. In such structures there exist a range of frequencies for which the light (photon) propagation is prohibited. It was also demonstrated that such crystals can be made in two and three dimensions. Such structures can have a complete photonic band gap, meaning that light is prohibited to propagate in any direction inside such a crystal. To realize a PC with a complete photonic band gap, the material must have both high refractive index and proper three dimensional structure state. Similar to PC, another class of crystals known as phononic crystals (PhC) [2] was also reported. These crystals possess the properties of PC but for acoustic waves (phonons) instead of light. There exists, however, still another possibility to control properties of PC by using the magnetic materials for magneto-photonic crystals. It is possible also to engineer magnetic materials where instead of light (or electromagnetic waves) spin waves (SW) are used as the carriers of information. Drawing an analogy from photonic and phononic crystals they may be called magnonic crystals (MC) (because magnons are the quasiparticles of spin waves). Propagation of spin waves in ferromagnetic films with periodically and weakly varied parameters has been studied extensively.

In this work we treat the problem of spin wave propagation in periodic structures from the point of view different from those previously employed, namely, in the approach used in investigating the properties of photonic and phononic crystals. Therefore, we first review and calculate the spectra of MSW propagation in the 2-D ferromagnetic film periodic structure, and then implement the 2-D MCs and measure its propagation characteristics. We also re-consider the propagation of elastic waves in phononic crystals.

2. Magnonic crystals Magnetic films with MSW traditionally are considered as a perspective transmission medium

that may be used for design of a wide class of miniature microwave electronic components. Application of periodic arrays combined with magnetic films opens new opportunities including creation of miniature bandgap structures so-called magnonic bandgap (MBG) crystals that have unique electromagnetic properties. Magnetic film with one-dimensional array was considered by several authors. They demonstrated that the array of metal strips placed over the film surface strongly affects the properties of a MSW which may have anomalous dispersion and negative equivalent material parameters. The aim of this work is an electromagnetic analysis of structures with arrays of a finite length. Structures which were analyzed are shown in Fig. 1. The main ideas and approximations of the analysis are presented below. Thickness of metal screen with slots or thickness of strips is considered as equal to zero. Also we suppose that conductivity of these metal elements is infinite. Width of strips and slots W is much smaller than the period of an array. Thus we may take into account only longitudinal currents flowing along strips and slots (electric and magnetic, respectively). All slots and strips in an array are considered as identical.

Electromagnetic field is a superposition of electric ( H , = 0 ) and magnetic ( E , = 0 ) waves. Due to the short wavelength of the MSW a,,, (a,,, << &, , &, - free space wavelength) periods of an array, dimensions d, h, a are much smaller than 4. It corresponds to the quasi-static situation in which the fields of electric and magnetic waves are almost uncoupled. We are neglecting by this coupling. We also describe magnetic material in a conventional way applying tensor of magnetic permeability. Loss in magnetic material can be taken into account. Slots in the two-dimensional array are tilted along the Oy axis. The angle of incidence of the MSW p has an arbitrary value.

121

In accordance with mentioned above approximations we reduce boundary problem to a set of integral equations for electric or magnetic currents. A key factor in this process is a Green's function of a multi-layer structure (see Fig. 1).

Fig. la. Periodic slot line. Fig.lb. Periodic stripe line. Green's functions for structures shown in fig. 1 were obtained in a spectral domain. To obtain the set of integral equations we replace slots or strips with magnetic and electric currents and then applying Green's function we find the field excited by these currents. Solving the boundary problem we obtain the set if integral equations. Periodicity of the structure along the Oy axis allows consider only the one period along the Oy axis. Thus the number of unknown functions is equal to N (see Fig. 1). Galerkin's technique was applied for integral equations solution. Basis functions for the currents on the strips and slots were taken in a conventional form which takes into account current peculiarities near the metal edges. Good accuracy gives the first order approximation when the only one function is used. The set of linear algebraic equations (SLAE) was obtained as a result of Galerkin's method application. Solving this set one may find unknown currents and then with the help of Green's function the coefficients of reflection and transmission Ri and q. Index i means that the array may excite several waves of different type.

The coefficients of the SLAE were presented as the integrals of double series. Convergence of these series determines numerical efficiency of the SLAE solution. Original series have very slow convergence and it was improved with the help of an analytical approach. Also the terms containing poles and branch points of sub-integral function were selected and the corresponding integrals were analytically calculated. Thus the only continuous functions in form of rapidly converging series were numerically integrated. These approaches increased the efficiency of the numerical solution of the SLAE and what is important that this efficiency almost does not depend on the array length (number N). Three structures were numerically studied: slot array with the magnetization along the Oz axis (volume MSW), slot array with the magnetization along the Oy axis and strip array with the magnetization along the Oy axis (surface MSW). Some results are shown in Figs. 2-3.

Fig.2. Scattering parameters of slot array line. Fig.3. Same for strip array line. Plots in Fig. 2 present scattering parameters of the slot array and the volume MSW (d=lO-*

mm, h=0.8d, W, t W, = 1, W, is the frequency of the ferromagnetic resonance, Px=5.5d, PY=l.5d,

122

w=0.3d, L=4d, N=60, 6 = goo, v, = 0 ). Curves shown in Fig. 3 correspond to the strip array and the surface MSW (d=10-2 mm, h=O, PX=6.2d, w=0.5d, N=40).

The MCs used in the experimental study was the ferromagnetic YIG film grown epitaxially on a nonmagnetic gadolinium gallium garnet (GGG) substrate. MSW can be easily excited in such films using microstrip transducers. Owing to the high quality of the films, the propagation loss should be fairly small, and the MSW should propagate in these films without any significant attenuation within distances of many wavelengths. The 2-D periodic structure studied was facilitated with holes formed by etching. The diameter of the holes and their periodicity were taken to be close to a half-wavelength in order to satisfy the Bragg reflection condition. The area of the film was 1.5cm.0.5 cm, and the film thickness was varied from 5 to 16 pm. The 2-D periodic holes structures were formed in the films by the photolithographic technique involving the following procedure. A silicon dioxide layer was first deposited on the YIG film. The structure was then covered by a 1.3- pm -thick photoresist layer, which was insolated through a chromium mask with a periodic holes pattern. Two types of holes patterns were used: a square lattice and a hexagonal lattice. After exposure, silicon dioxide was removed by a mixture of hydrofluoric acid and ammonium fluoride (solution 1). Then, the remaining silicon dioxide lattice was used as a mask for etching of the YIG film in an aqueous solution of phosphoric acid and iron chloride (with a molar proportion of 49.4 : 49.4 : 1.2; solution 2) [3]. The etching time and temperature were chosen so as to etch the material through a thickness from 1.0 to 4.5 pm. After the etching step, silicon dioxide was completely removed by solution 1. The holes depth and the surface structure were studied by a three-dimensional optical rugosimeter and an atomic force microscope (D300 Digital Instruments and Solver P47H NT-MDT). In order to investigate the MSW spectrum in the resulting 2-D MCs delay line devices were made. The amplitude-frequency characteristics curves (AFC) of propagating MSW were measured. Now, let us examine the experimental data for the case of MSW propagation in the cubic structure. Fig. 4 shows the measured AFC and the phase-frequency characteristic (PFC) plots at the bias magnetic field H O = 800 Oe under different orientations between the antennas and the etched structure. The anglea designates the angle between the MSW wavevector k' and the lattice axis. It can be seen from Fig. 4a that when the pair of antennas was located outside the YIG film with the 2-D lattice neither AFC nor PFC showed any distinct features. The dispersion obtained in this case is typical for the MSW. However, when one or both antennas were located within the YIG film the band gaps appeared in the spectra (see, e.g. Figs. 4c, d). A few observations are now in order with respect to the second band. 1) This band appeared at some value of the angle a for some range of the ratio G between the areas of the etched and unetched parts of the YIG film located between the antennas. Increase in the portion of the etched area led to increase in the MSW losses within the whole frequency spectrum. This increase in propagation loss has made it impossible to register the band observed previously. 2) When the band was registered at a given angle a and the etched area between the pair of antennas was increased the depth of the band increased as well. This dispersion is most important because it shows that the band B2 is not caused by the interference between the incident and reflected MSWs from the boundary of the YIG film or the 2-D periodic structure. It should also be noted that the shapes of the AFC and PFC plots were practically unchanged as the propagation direction of the MSW or the direction of the bias magnetic field was varied. At the placement of the microstrip antennas shown in Fig. 4b, the band gap appeared in both the AFC and the PFC plots. This band corresponds to the 55th interference peak of PFC. Based on the first minimum the distance between the antennas 4 mm, and the MSW wavelength, the wavenumber was calculated to be k=15.7 CM-I. Then for the 55-th minimum the calculated wavenumber was k=860 CM-I. As the angle a was varied ( a < 15' ) the band was seen to move to the longer wavelength or the lower frequency of the MSW. As the axis of the 2-D structure was rotated with respect to the MSW propagation direction at the angle a =15" the lower frequency edge of the minimum in the PFC plot corresponding to the 28-th interference minimum or the wavenumber of 440 CM-I. The measured half-width of the minimum peak is equal to 16 MHz. The location of the minimum and its frequency width changed continuously as the angle a increased.

123

For example, the minima edges at the angles a= 5’ and 10’ occurred at the 46-th (k=720 CM-’) and the 39-th (k= 610 CM-’) minima in the PFC plot and the half-width of the minima was increased to 20 MHz. The above experimental results can be interpreted most simply for the angles a = 0 and a = 45’. In these cases due to the symmetry involved the MSW propagates normally to the 2-D periodic structure and the Bragg condition was fulfilled. The periodicity of the cubic lattice is

ncuboo = 37 pn at the angle a=O. The MSW Bragg wavenumber is k,,b = 849 CM-’. This wavenumber corresponds to the 55-th minimum in the PFC plot (Fig. 4b). Similar results are obtained with the case in which the MSW propagates at the angle a =45’. In view of the fact that the existence of the band was caused by the Bragg reflection we may conclude that in the vicinity of the Bragg resonance condition frequency bands that forbid propagation of the MSWs can be engineered. Reflection of the microwave power from the input antenna increases within these forbidden bands.

Fig.4. Measured amplitude-frequency and phase-frequency characteristic (AFC and PFC) plots for the YIG film with the cubic lattice structure at different angular orientations between the microstrip antennas and the etched portion of the film (the bias magnetic field was set at 800 Oe).

3.9 Frequency T((iHr) %.?

3.9 4.3 4.7 b)

7n O 1

4 .O

4.3 4.7

d)

3. Phononic crystals Materials with periodic acoustic mismatch called as “phononic” or “sonic” crystals have also

received significant research interest, because acoustic waves have their own specific features, and due to existence of similar localization and superlens effects. At the present time the complete band gaps for bulk and surface acoustic waves have been reported, propagating acoustic modes in phononic crystals with channel of defects in periodic lattice have been studied and such special type of surface acoustic wave as Bleustein-Gulyaev-Shmizu wave has been predicted in a two- dimensional piezoelectric composite material.

In calculations of characteristics of acoustic waves in sonic crystals numerical methods are most often used. Among them the method of plane wave expansion (PWE) and the finite difference

124

time domain (FDTD) technique should be mentioned. In the present paper the finite element method (FEM) is applied to calculate dispersion curves of a sonic crystal consisting of void circular cylinders arranged in accordance with quadratic and hexagonal symmetry in the substrate of fused quartz. It is shown that in agreement with previously published results full band gaps exist in the sonic crystal with quadratic symmetry. In the case of horizontal polarization a backward wave in the first band exists.

We deal with stationary state condition, so it is assumed all unknown values are dependent on time as exp(-j@), where o is an angular frequency. In FEM calculations a discrete set of points is introduced into the elementary cell, and the points are connected to each other in such a way that a triangular mesh over the whole region is formed. The unknown wave field in each triangular element is represented by a sum over approximating functions with weighting coefficients for each function. As for approximating functions the polynomials of special type are used, namely, which are not equal to zero only in one nodal point at the boundary of a finite element. So the weighting coefficients in the polynomial expansion can be considered as unknown values of wave fields in the nodal points. The polynomial approximation is substituted into equations of motion of elastic medium, and after some integrations and manipulations a system of linear equations in unknown elastic displacements in nodal points is derived. The right part of this system contains integrals of normal stress components over the boundaries of a finite element. Then, taking into account conditions of continuity of elastic displacements and normal stresses at the boundaries between finite elements and by combination of equations for separate elements the system of linear equations in the whole finite region is obtained. The unknowns of this system are the elastic displacements in all nodal points, and the right part includes integrals of normal stress components over the boundaries of the unit cell.

We consider two-dimensional sonic crystals made of fused quartz with void circular holes arranged in accordance with quadratic and hexagonal symmetry. Figure 5 shows elementary cells of the sonic crystals with triangular mesh of finite elements.

a Y

Fig.5. Elementary cells for quadratic and hexagonal sonic crystals It is assumed that acoustic waves propagate in direction perpendicular to the axes of holes that

is in XY plane of the coordinate system. In 2D-sonic crystals constructed of isotropic materials all bulk acoustic modes can be subdivided into two sets relating to their polarization. One set includes normal modes with X- and Y- components of elastic displacements that are mixed longitudinal and shear modes, we call them as waves with horizontal (H) polarization. Normal modes with shear elastic displacements along Z-axis are separated from previous modes, and called as SV-modes (shear vertical).

Figure 6 shows dispersion curves for SV and H modes calculated over the contour T-X-M-T- points in the first Brillouin zone for the sonic crystal with quadratic lattice. The hole diameter to cell size ratio is equal to d/a=0.9.

125

r X M r Normalized wave vector, k a h

Fig.6. Dispersion curves for sonic crystal with quadratic geometry, d/a=0.9. Solid lines represent mixed modes with horizontal polarization, Dashed lines refer to shear vertical polarization. Dotted

line shows imaginary part of wave number at the end of Brillouin zone in T-X direction

It is seen that in the normalized frequency range from 3.2 to 4.5 there is a full band gap that means that all bulk waves can not propagate. There are no real and positive eigenvalues of the problem (2) at any real wave number, but instead the determinant in (2) can be equal to zero at some complex values of wave number. The complex valued wave number corresponds to attenuation of the wave due to reflection. Dotted line in Fig.6 shows imaginary part of the normalized wave number at the end of Brillouin zone in T-X direction.

Basically, the dispersion curves in Figure 1 are similar to that reported in the paper [4] for a sonic crystal of void holes in aluminum substrate. Note that there is a quasi-shear basic band mode that has negative group velocity starting from about the middle point of wave vector variation in the T-M direction. Additionally there exist modes with negative group velocity in the second band.

Figure 7 shows dispersion curves for SV and H modes calculated over the contour T-M-K-T points in the first Brillouin zone for the sonic crystal with hexagonal symmetry at the same parameter d/a=0.9. Contrary to the case of quadratic symmetry, there is no full frequency gap, however at least three band gaps for SV modes can be seen in Figure 7, one of which is approximately from 2.25 to 3.45, the second is from 5.0 to 7.0, and the third is from 8.5 to 9.7 unit of normalized frequency.

Normalized wave vector, kah

Fig.7. Dispersion curves for sonic crystal with hexagonal symmetry, d/a=0.9. Solid lines represent mixed modes with horizontal polarization. Dashed lines refer to shear vertical polarization.

4. Conclusion

126

The propagation characteristics of the MSWs in the YIG film with 2-D periodic structures have been studied both theoretically and experimentally. The 2-D periodic structures serve to alter the MSW propagation conditions. The spectra of MSW propagating in the YIG film with 2-D periodic structures of surface non-uniformities are calculated. The degree of alteration in the propagation conditions is determined by the parameters of the 2-D structure: etching depth of the holes, structure type and holes density. The forbidden gaps found in the frequency spectra are attributable to Bragg reflection of the MSW from the surface periodic structures. Such ferromagnetic films with 2-D structures for control and processing of microwave spin waves are direct analog of the conventional photonic crystals for optical waves. They are thus called magnonic crystals. As an another result of this work, dispersion curves of normal bulk acoustic modes propagating in a sonic crystal made of fused quartz substrate with void circular holes are calculated using finite element method. In the case of quadratic lattice the results are similar to those previously published and show full frequency band gaps at sufficiently large d/a-parameter. The imaginary part of wave number related to the reflection of acoustic wave at the end of the Brillouin zone has been calculated. There are no full frequency gaps in the sonic crystal with hexagonal symmetry, however a series of gaps exists for modes with shear horizontal polarization.

This work was supported by the Russian Foundation for Basic Research, projects no. 05-02- 17361.

References

[ l ] J.D. Joannopoulos, R.D. Meade, J.N. Winn, Photonic Crystals: Molding the Flow of Light

[2] I.E. Psarobas, N. Stefanou, and A. Modinos, Phys. Rev. B Vol. 62 (2000), p. 5536. [3] Yu. V. Gulyaev, S. A. Nikitov, et al. JETP Lett., Vol. 77 (2003), p. 567.

(Princetone University Press, Princeton, 1995).

Distinctive feature of 1D anisotropic and gyrotropic photonic crystals

Alexey P. Vinogradov', Alexander M. Merzlikin', Alexander V. Dorofeenko', M. Inoue 2,

Alexander A. Lisyansky3

'Institute for Theoretical and Applied Electromagnetism Russian Academy of Sciences, RF Department of Electrical and Electronic Engineering, Toyohashi University of Technology, Japan

Department of Physics, Queens College, the University of NY, US

Abstract. The magneto-optical properties of 1D photonic crystals (PC) are considered. The consideration is focused on the distinctive features of 1D PCs because any devices made on the base of 1D PC are more robust for losses than those employing 2D or 3D PC. To make better off the properties of comparative simple 1D geometry we suggest the usage of anisotropic and gyrotropic materials. Firstly, such PCs exhibit new physical phenomena, namely, formation of the Yeh band gaps, magnetooptical and birefringence effects. Secondly, since the anisotropy and gyrotropy are easy caused by the external electric and magnetic fields the new properties of the PCs are tunable. The functioning of switchable filter, magnetic superlens and other devices are considered.

Recently, the optical properties of photonic crystals (PCs), which are structures with periodically modulated in space constitutive properties, have attracted a great deal of attention (see for example [l]). The new effects are connected with resonant Brillouin scattering of waves. As a consequence, the losses in materials, of which the PC are made, hinder observation of the effects. The influence of losses depends on dimensionality of PC. Since it was found out [2] that 1D PCs are more robust to losses we confine our consideration to 1D PC.

The properties of PCs are chiefly determined by the symmetry of the primitive cell. Below we consider some consequences of the symmetry violation. In particular, the 1D PC made of isotropic materials is invariant under rotation about the z-axis. A PC with bi-layer primitive cell is invariant under reflection of z-axis. Below R, (cp) , C, T and are used to designate a rotation of cp radians about the z-axis, reflection of z-axis and time reversal.

Application of anisotropic and/or gyrotropic materials can change the crystallographic point symmetry of PC and result in opening of a new type of band gap located inside the Brillouin zone [3-51. P. Yeh [3] was the first to point to the existence of such a band gap in 1D PC. Below we refer these band gaps as the Yeh band gaps.

Imagine 1D PC with bi-layer primitive cell. Below, for simplicity, we assume all the layers have identical thickness d. One of the layers, say for distinctness the first one, is made of uniaxial crystal with the axis of anisotropy being parallel to the interface, whereas the second layer is made of isotropic magneto-optical material. Below, we assume that before switching on a static magnetic field the magnetization of the MO material is equal to zero.

Note, that such a PC is invariant instead of R, (cp) under R, (x) only. Following [3], let us consider the mechanism of formation of new band gaps. In the case of

zero magnetization, the MO layers are isotropic and the eigensolutions in these layers can have any polarization In the uniaxial crystal layers the eigensolutions correspond to polarizations of ordinary and extraordinary waves. As a consequence, at zero magnetization the ordinary/extraordinary wave incident on MO layer excites in this layer the wave with the same polarization and vice versa, the latter wave excites in the next uniaxial crystal layer the corresponded ordinaqdextraordinary wave only. It is important to notice that the reflected waves have the same polarization. Thus, there is no coupling between the waves of ordinsuy and extraordinary polarizations: they build up two independent Bloch waves of different polarization. These Bloch waves each can be expended into series of plane waves of fixed polarization, namely these plane waves have polarization of ordinary or extraordinary waves with the vector of electrical field perpendicular or parallel to the axis of anisotropy. The Bloch waves have different wavenumbers k,, and kEx respectively. (The sketch of

127

128

an algorithm for calculation of the Bloch wave numbers see in [6]). These waves have band gaps on the boundaries of the Brillouin zones: kor,Ex = z /2d (see Fig. 1).

Fig. 1 : The scheme of formation of Yeh’s band gap (reduced and extended band patterns). The solid curves correspond to the magnetized MO layer. The dotted curve presents the dispersion of the extraordinary mode. The dashed curve presents the dispersion of the ordinary mode. The PC’s primitive cell consists of a layer of anisotropic material ( E, = 2.0, E,,,, = 8.0) and of a MO layer ( E~~~~ = 3.0, E ~ ~ - ~ ~ ~ ~ = ia = 0.5i), the thickness of each layers equals to d

After application of a static magnetic field perpendicular to the layer’s interface there appears a magnetization of the magneto-optical material, which results in changing the eigensolutions in the previously isotropic and now gyrotropic (magneto-optical) layers.

In the gyrotropic medium an eigensolution is left or right polarized circular wave exciting in the uniaxial crystal layers both ordinary and extraordinary waves. Because the wave numbers of the ordinary and extraordinary waves are different, in the general case, the polarization of the wave on opposite interface of the uniaxial crystal layer is elliptical. Hence inside the next MO layer both right and left polarized wave are excited. Thus, the waves of different polarization become mixed at boundaries. A plane-wave expansion of the corresponding Bloch waves consists of elliptically polarized plane waves. If we are far from the condition kor + k,, = z I d then we observe a small disturbance of the wave numbers and polarization of the main harmonic is about linear one. At fulfillment of the condition kor + k, = z I d there appears strong coupling of the opposite traveling waves and the solutions, at first approximation, may be considered as a superposition of these opposite traveling waves. They look like standing waves (they have knots but their polarization is not linear). These new solutions transfer no energy and correspond to the edge of a new band gap (the Yeh band gap), formed at the point of crossing of the dispersion curves. It is necessary to note that the intersection is possible if the opposite traveling waves belong to different pass bands (see Fig. 1).

Since a system is invariant under Rz(n), and C (like the aforementioned system is) it is a

spectral reciprocal system. Indeed, CRz(n)*o(kx,ky,kz) = Co(-k,,-k,, k,) = w(kx,k, , -k , ) . If

there is no symmetry transformation S{kx,k , ,kz} = {kx ,ky , -kz} , under which the system is

invariant, than the system may be nonreciprocal w ( k , , k, , -kz ) + w ( k , , k, , k, ) [7]. An example is an 1D PC with three-layer primitive cell where one layer is gyrotropic with magnetization along z- axis and two others layers are uniaxial crystals with non-parallel axes of anisotropy.

129

In bound PCs the spatial periodicity (discrete translational symmetry) is broken. If the bound PC is contiguous with a uniform medium characterized by a negative permittivity or permeability the solutions to Maxwell's equations inside the band gap, which were previously nonphysical, may acquire physical sense, forming surface waves that exponentially decay away from the boundary (see Ref. 2 and references therein). If these solutions are localized, following the usage in solid state physics, they are called the Tamm states [2]. Tamm states may also appear at the boundary between different PCs [2]. This state may arise at frequencies where the band gaps of both crystals overlap.

The existence of such states can be observed by studying the transmission or reflection spectra of the finite system. Due to resonant tunneling of the electromagnetic wave through the Tamm state, a sharp narrow peak appears in the transmission spectrum (or a sharp minimum in the reflection spectrum) [2].

In anisotropic-gyrotropic system a Tamm state may be observed that appears due to the formation of the Yeh band gaps in an external magnetic field. If the PC considered in the first example joints a second PC with band gap, then application of a static magnetic field may create the condition for excitation of a Tamm state at the interface. The values of permittivity of the layers in the second PC are E, = 3.1 , = 3 .4 . All the layers are of identical thickness d. Although the parameters are chosen to emphasize the involved effects they are close to realistic ones. After switching on an external magnetic field there appears a Yeh bad gap in the first PC. The corresponding imaginary part of wave vector is shown in Fig.2 (bottom figure, solid line). Note that, first, at zero magnetization the imaginary part of wave vector of the first PC is equal to zero and, second, the low transmittance of the system at zero magnetization is due to existence of the band gap in the second PC. Thus, we can see that after switching on magnetic field and formation of the band gap in the first PC the system becomes more transparent at the very frequency of the Tamm state. (Fig 2). This state is located in the intersection of the band gaps of both PC.

10

un

'm(k/IS .....* ( ~ ~ ~ I , I . A +); OD1 .

i

0 iK) k,,d

Fig. 2: In the upper figure, the solid line shows the calculated transmission coefficients of the system on application of magnetic field. The doted line corresponds to the case of zero magnetization. In the bottom figure the solid line presents the imaginary part of wave vector of the first PC at nonzero magnetization (at zero magnetization it is equal to zero) and the doted line presents the imaginary part of wave vector of the second PC.

0 86 0 87 o nx

It is well known that in gyrotropic system we can observe rotation of plane of polarization of a linearly polarized wave. A linearly polarized wave can be presented as a sum of two circular polarized waves, which are eigensolutions of a gyrotropic medium. Since the phase velocities of

130

these eigensolutions are different the plane of polarization of the linearly polarized wave rotates in space with growth of z-coordinate (magnetooptic Faraday effect). If at zero magnetization the system has a transmittance resonance, like Fabry-Perot filter has, the resonance line will split into resonance lines for left and right circular polarization (Fig. 3b). At frequency where these lines are crossing the transmitted wave is linearly polarized and we can observe the rotation of the plane of polarization. The angle of rotation has a scale of K , because it is connected with phase change at resonance rather than with weak Faraday rotation. A well-known realization of this resonant structure is a gyrotropic defect in photonic crystal [8], at which a localized mode appears. Another realization is a Tamm surface state existing at a boundary of two photonic crystals [2]. Gyrotropic (magnetooptical or MO) material is then included into one of the crystals. The presence of eigenstate (defect-mode or Tamm state) is exhibited as a sharp peak in the transmission coefficient inside the intersecting band gaps of crystals (Fig. 3a, solid line, peak at A = 1.15,urn ). After magnetization of the MO material the difference in phase velocities for two circular polarized waves appears, and the peak splits into two ones, the same being for the phase dependence of transition wave. For the transmitted wave to be of plane polarization the wavelength should be taken at which the transfer functions of both circular polarizations have equal amplitudes (intersection of thick curves at Fig. 3b). The appearance of rotation angle B is a consequence of different phase changes for the polarizations, and it is equal to the half of the difference between Arg(T) for both the waves (Fig. 3b).

If q (A) and T, (A) are transfer functions of the circularly polarized waves, one can

approximately write T2 (A) = T (A + AA) , where AA is the shift of the resonance maximums, which is proportional to the off-diagonal element a of the permittivity tensor. Then

The formula (1) is valid for small AA , but it can be used at estimations. As it is seen in the Fig. l a (dashed line), the largest value of the derivative corresponds to the Tamm state. All these peaks correspond to the poles of the transfer function. Since we deal with open resonator, the eigenfrequencies are complex (Fig. 4). The imaginary part of the frequency w" is inversely proportional to the lifetime of the resonance and to the quality factor. The resonance at A = 1.1 5 ,urn , w"/ c = -0.0002 ,urn-' is the Tamm state, others are more complicated resonances.

a b Fig. 3: (a) The square of modal (solid line) and frequency derivative of phase of transfer function for a system of two PCs. A Tamm state is seen as a peak inside band gaps of the crystals. Vertical lines correspond to the positions of poles of the transfer function. (b) Peaks of the transfer function (thick lines) and corresponding phases (thin lines) for two circular polarizations at non-zero magnetization.

131

-0.004

-0.006

-0.00,s

-0.01

, ' . . .

I , ) "/dfl!n-~ )

Fig. 4: Poles at the plane of complex frequency.

The difference between resonances is in their quality factor. Besides this, close position of several resonances prevents the phase from changing quickly and, consequently, leads to small values of phase derivative and small Faraday angle. Of course, there are no other differences between the resonances, appearing in different structures. The problem is in separating of one eigenstate from the others and obtaining a large enough quality factor. The resonance, corresponding to the Tamm state (or defect-mode), differs from the Fabry-Perot resonances in these parameters, and provides larger rotation angles. Knowing the positions of the poles and corresponding residues, one can present the transfer function as T ( o ) =To + z r e s T ( q ) / ( w - mi).

Then any calculations can be performed analytically, no matter how difficult the structure is. Near some resonance, only one corresponding term (and, if necessary, several neighboring ones) can be taken into account.

It is clear that thick system (large number of layers) provides large rotation angles 8, but small intensity I T 1'. Therefore, an optimum width should exist. To find this optimum, one can choose a quantity that should be maximized. If a transfer function of thick system has a form T - exp(ik&-q&) ( L is a crystal thickness), the decay is determined by the value

K ~ & - -In I T 1 2 , and the decay per unit length is K~ - -In I T 1' / L . We will use a maximization of

the ratio of the rotation angle per unit length 8 / L to the decay per unit length -In I T 1' / L as a criterion of the magnetooptic Q-factor, i.e. QMo = -B/ln I T 1'. Fig. 5 shows dependences of the Q- factor on the frequency interval between the maxima of I T 1' for different schemes. The Tamm resonance in the second band gap is approximately an optimum.

I

I500

1000

500

_. M , , I ___._._.._..,,.., ...-._. l.) ..................... .........

I -

.* 0003 0004 0006 0006 (101

Fig. 5: The MO Q-factor, i.e. QMo = -8/ln I T 1' versus the optical path difference of RC and LC waves at one-pass of MO material. The curves are obtained by variation of off-diagonal terms of permittivity. The triangles correspond to the real values of these terms. Dashed line presents the Fabry-Perot scheme based on the Tamm state in the second band, solid line corresponds to the Tamm state in the fifth band, the dotted line present the Fabry-Perot scheme based on the MO defect mode.

132

Conclusions

to control the properties of PC by application of static electric or magnetic fields. It is shown that anisotropy and gyrotropy may result in observation of new effects that permits

Acknowledgements

05-02-19644 and No. 06-02-81053). This work was supported in part by Russian Foundation for Basic Research (Grants No. No.

References [l] [2]

[3] [4] [5] [6]

[7] [8]

K. Sakoda, Optical Properties of Photonic Crystals. (Berlin: Springer-Verlag, (2001). A. P. Vinogradov, A. V. Dorofeenko, S. G. Erokhin, M. Inoue, A. A. Lisyansky, A. M. Merzlikin, and A. B. Granovsky, Phys. Rev. B 74,045 128 (2006) P. Yeh, J. Opt. SOC. Am. 69 (1979) 742. R. Zengerle, J. Mod. Opt. 34, 1589-1617, (1987) E. Cojocaru, Appl. Opt., 39,4641, (2000) A.M. Merzlikin, A.P. Vinogradov, A.V. Dorofeenko, M. Inoue, M. Levy, A.B. Granovsky Physica B A. Figotin and I. Vitebskiy, Phys. Rev. B 67,0165210 (2003) A. P. Vinogradov, A. B. Granovsky, S. G. Erokhin, M. Inoue, Journal of Communication Technology and Electronics V. 49, Nc1. p. 88-90,2004

Metamaterials: A New Route to Microwave Magnetism

Olivier ACHER CEA Le Ripault,

BP 16, F-37260 France olivier.acher@,cea. f r

Abstract : The metamaterial approach is fruitfd in order to develop microwave magnetic materials. This

paper provides several examples of high permeability metamaterials. A reference sample for permeability measurement is presented. This sample is proven to be immune to magnetic field and temperature drifts. Voltage-tuneable magnetic materials have been manufactured. High permeability magnetic materials have been combined with copper wirings to provide unique frequency-dependent features. Using a Field Summation Method, the effective permeability can be easily extracted from numerical simulation, and field maps provide a better understanding of the effective properties.

1. Introduction: The metamaterial community has established that composites with engineered permeability

could be designed from conductive patterns, without using conventional magnetic materials [ 11. While a few years ago it was the well-established “privilege” of Iron-containing materials to exhibit high frequency permeability features, shall they be ruled out by copper-based composites?

One aim of this paper is to provide some clues on what metamaterials are best for. In some cases, a combination of metamaterial copper wiring and of iron-based material is extremely efficient. We will also outline cases where conventional magnetic materials may remain unchallenged.

Another aim of this paper is to provide an intuitive view of what the microwave permeability stands for in metamaterials. While the notion of permeability in conventional magnetic materials is fairly familiar to many magnetic material scientists, it may not be easily transposed to the case of copper-based materials. Recently, we proposed a technique to compute the effective parameters of a metamaterial, which also provides powerful visual insight on the operation of metamaterials. This technique is expected to be of practical interest for those interested in designing metamaterials.

2. Metamaterials as reference sample for metrolow. In contrast to conventional magnetic materials, copper-based metamaterials are immune to

temperature drifts, external magnetic fields, and aging. For microwave permeability measurement systems [2], they are very attractive to manufacture reference samples and calibration samples. We developed such a reference sample [3]. The sample is based on artificial magnetic inclusions manufactured using CMS components. A tiny solenoid acts as the inductive patterns of the metamaterial. It is connected to a capacitor, in order to tune its resonance frequency Fr=wr/(2x) at a desired value. The resonance condition writes LCo, ’=1, where L stands for the inductance value and C the capacitance value. It was found appropriate to add a resistor, in order to obtain a permeability peak with significant damping. A picture of an artificial magnetic inclusion is given on Figure 1. The permeability of a sample consisting of 2 artificial magnetic inclusions is represented on Fig. 2. It has been checked that the permeability of the sample is immune to magnetic field and to temperature drifts.

133

134

inductor

capacitor

resistor

Figure 1. Artificial magnetic inclusion manufactured using CMS components. The scale is in mm.

I

0 500 1000 1500 2000 2500 300C frequency [MHz]

Figure 2. Imaginary permeability of a sample consisting of 2 artificial magnetic inclusions.

3. Metamaterials as tuneable materials Another attractive feature of the metamaterial route is that it allows a very convenient way to

tune the permeability. Starting from permeability resonance condition LCo, '"1, it is simple to tune the resonance frequency using a varactor. In ref [4], we demonstrated the operation of a tuneable magnetic element using this principle. It has been since used by many other groups [S, 61. The permeability of a voltage-tuneable metamaterial is reported on Fig. 3. The resonance frequency can be tuned over more than one octave. For commercial applications, the electrical control of the permeability appears to be much more attractive than the conventional control through an applied magnetic field. Electronic loads may be used not only to provide tuneability, but also to synthesize completely new properties, as demonstrated in [7]. It may be also mentioned that we demonstrated that the permittivity of a wired-based media could be tuned using an external magnetic field [8]. The underlying mechanism is connected with the so- called "Giant Magneto Impedance" effect. Further works have been conducted on tuneable wire media based on that effect [9].

135

Figure 3. Imaginary permeability of a metamaterial consisting of an inductive loop connected to a load containing a varicap. The permeability is indicated for different voltages applied to the varicap.

4. Magnetic materials with engineered permeability response Metamaterials may exhibit large permeability levels only over a very limited bandwidth. They

can be used as high impedance surfaces, but in this case they provide a large impedance over a very limited bandwidth. This is because Rozanov’s limit applies [lo] to metamaterials as well. Rozanov’s inequation states that their bandwidth time efficiency product is limited by the product of their thickness by their low-frequency permeability. Since the permeability of copper metamaterials is due only to induced currents in the structure, the low frequency permeability is unity. As a consequence, the permeability of metamaterial does not result in an increase of bandwidth, in contrast with conventional magnetic materials.

Figure 4: Permeability of Inductive Textile Composite metamaterials.

However, it is possible to combine conventional magnetic materials and inductive patterns to obtain both significant permeability levels and bandwidth, and engineered permeability features. Many routes are possible to obtain such materials. However, the printed board technology generally used to manufacture metamaterials is not very convenient for that. We have shown that a textile approach was efficient in building such materials. We used bundles of magnetic microwires as high permeability core material, and wound pieces of copper wire around the core [l 11. Figure 4

136

represents the permeability of different Inductive Textile Composites (ITC), made with different number of windings of the copper wire around the magnetic core. The resonance frequency can be easily tuned by one order of magnitude using this geometrical parameter. Tuning the resonance frequency through the material science of conventional magnetic materials would be a much more challenging task! This evidences the interest of developing metamaterials based on conventional magnetic materials. At very high frequencies, however, it can be shown that copper-based metamaterials may outperform conventional iron-based magnetic materials. The cross-over is around 30 GHz [ 121.

5. A visual understandiw of the effective permeability of metamaterials wine the Field Summation Method

Electromagnetic simulation software is commonly used to investigate the properties of metamaterials. The s o h a r e provides the values of the electromagnetic fields in the whole material. It is attractive to use this local information to compute the effective permeability and permittivity. Several approaches have been developed for that [ 13,14,15]. In all these approaches, the effective parameters are defined as a function of local averages of the E, H, D and B fields. We have established that an appropriate expression of the permeability is

where the average of B is taken over the volume of the unit cell, and the average of H is taken on a face of the cell, normal to the electric field. This face should not be crossed by any electrical current. The averaging zones are sketched on Figure 5. This approach has been used to determine the effective properties of metamaterials from electromagnetic simulation software [ 161. it is also very efficient to determine analpcally the effective properties of a variety of composite materials ~171.

Y

Figure 5. Presentation of the Field Summation Technique: averaging zones, and expression for the effective parameters.

in the case no magnetic core is present, B=po.H everywhere, and the expression of the permeability simplifies into :

137

It is then easy to use appropriate field maps to understand the sign and magnitude of p [18]. Figure 6 represents the H field in the unit cell of a metamaterial, at a frequency slightly above its resonance frequency. The field inside the inductive pattern is large, and opposed to the field on the facet normal to E. It follows from (2) that the permeability is negative.

Figure 6. H-Field map within the unit cell of a metamaterial, at a frequency slightly above the resonance frequency.

6. Conclusion The metamaterial approach has brought new approaches to engineer the magnetic microwave

properties of materials. Copper-based metamaterials can be very attractive as microwave magnetic materials. In some other case, the association of copper-based inductive patterns with Fe-based conventional magnetic materials is adequate. A practical approach for the design of metamaterials has been proposed.

Acknowledgements

Reynet, for their major contribution in the vision on metamaterials presented here. I am thankful to Dr A.-L Adenot-Englevin, M. Ledieu, N. Mallejac, J-M Lerat, Ch. Dudek, 0.

References

[l] J. B. Pendry, A. J. Holden, D. J. Robbins, and W. J. Stewart, IEEE Trans. Microwave Theory Tech. 47,2075-2084 (1 999). [2] Yan Liu, Linfeng Chen, C. Y. Tan, H. J. Liu, and C. K. Ong, Rev. Sci. Instrum. 76,06391 1 (2005). [3] M. Ledieu and 0. Acher, French Patent Application FRO5 53223. [4] 0. Reynet and 0. Acher, Appl. Phys. Lett. 84, 1198 (2004). [5] I. V. Shadrinov, S . K. Morrison and Y. S. Kishvar, Optics Express 14,9344 (2006). [6] A. Degiron, J. J. Mock and R.R. Smith, Optic Expres 15, 11 15 (2007). [7] K. N. Rozanov and E. A. Preobrazhenskii, J. Commun. Technol. Electron. 50,858-864 (2005). [8] 0. Reynet, A. -L. Adenot, S. Deprot, 0. Acher, M. Latrach, Phys. Rev. B 66,94412 (2002). [9] D. P. Makhnovskiy, L. V. Panina, C. Garcia, P. Zhukov and J. Gonzalez, Phys. Rev. B 74, 064205 (2006). [lo] K.N. Rozanov, IEEE Trans. Antennas Propagat, 48,1230 (2000). [ 1 11 A.-L Adenot Engelvin, Ch. Dudek and 0. Acher, J. Magn. Magn. Mat. 300,33 (2006).

138

[ 121 0. Acher, 0. Reynet, N. Mallkjac and J.-M. Lerat, proceeding of the SPIE conference vol 5359,76 (2004). [ 131 A.N. Lagarkov, A.K. Sarychev, Y. R. Smychkovich and A. P. Vinogradov, J. Electromagn. Waves. Appl. 6, 1159 (1992). [14] 0. Acher, A. -L. Adenot, F. Duverger, Phys. Rev. B, 62, 13748 (2000). [15] D. R. Smith and J. B. Pendry, J. Opt. SOC. Am. B 23,391-403 (2006). [ 161 J.-M. Lerat, N. Mallkjac and 0. Acher, J. Appl. Phys. 100,084908 (2006). [17] 0. Acher, A. -L. Adenot, F. Lubrano, F. Duverger, J. Appl. Phys., 85,4639-4641 (1999). [ 181 0. Acher, J.M. Lerat and N. Mallkjac, Optics Express 15, 1096 (2007).

Session P7

Chair: X. Yao

Hybrid Processing Technology of Electromagnetic Ceramics and Thick Films

Yao Xi a,

a Functional Materials Research Laboratory Tongji UniversiQ, Shanghai, 200092, China Electronic Materials Research Laboratory

Xian Jiaotong Universiv, Xian, 71 0049, China

Email: xyao@,mail - .tongii.edu.cn xy ao@,mail.xitu.edu.cn -

Introduction Electromagnetic ceramics are one of the major pillars supporting the ultra fast development of

electronics in the past a few decades. Now, the development of electronics has reached such a critical stage that many of the technologies used are close to their technological limit, sometimes, even the basic science behind the scene is reaching its theoretical limit. New strategies and innovative actions for the further development are of vital importance. Requirements on the properties and processing technologies of electromagnetic ceramics are getting more and more complicated and rigorous. New technologies and new materials are needed to meet stringent requirements.

Electronic devices and equipments are getting more and more light and small. The size of discrete electronic components used in devices and equipments is reaching sub-millimeter range. Manufacturing technologies and materials for making such components are very advaned which are comparable to the IC technologies. For example, the multi-layer ceramic capacitors (MLCC) are stacking from tens even hundreds of ceramic films in the thickness of a few micrometers and in dimensions of 8-12 inches. Electrode patterns of capacitor in the dimension of sub millimeters were screen printed onto the green ceramic films using conduction metal paste that is able to cofire with the ceramic films at high temperatures. Tens or hundreds of printed ceramic films were then stacked together forming plate-like thick sheets. Each green sheet was then cut into thousands and thousands of individual components for further high temperature calcination over 1000° C. The stacking of the green ceramic films and the cutting of the green sheets should be very precise to avoid short circuits of the capacitor electrodes. It is truly a very high technological challenge and can be supposed to be an engineering miracle, which is not at all less than that of the semiconductor silicon technology. The further reduction of the size and weight of MLCC requires the film thickness reducing down to around 1 micrometer and cofired with less expensive base electrode metal at lower temperatures. The existing technology can no longer meet such requirements. Many attempts were devoted to develop new technologies. [', 21

For microwave application of electromagnetic ceramics, new technology challenges are quite similar to those discussed above. Many of the microwave devices are fabricated on alumina substrates in the form of thick and thin films. Low temperature cofiring ceramic (LTCC) devices are getting more important and popular, where various ceramic materials and metals are cofired at low temperatures.

To meet these new technological challenges, a new hybrid ceramic processing technology has

141

142

been developed. [3, 41 The key characteristic of this new technology is to graft the sol-gel wet chemistry process to the conventional mixed oxide ceramic process. With this new technology, the advantages of both processes can be superimposed together. The high performance of ceramics prepared by conventional ceramic process under high temperatures can be mostly preserved, while the sintering temperature of the material can be effectively reduced down as inherited from the sol-gel process. The low sintering temperature of this hybrid process enables the technology better compatible to the thick and thin film technology. Low sintering temperature also impedes the chemical reactions among different materials in a cofired system, such as ceramic/metal electrode, ceramic/substrate and ceramics with different structures and performances, making this new technology very useful in developing composite materials from completely different end components.

The potential of this new hybrid processing technology in fabricating microwave devices with electromagnetic ceramics is very promising. Using well-calcined micro-sized or nano-sized dielectric ceramic powder and silica glass sol precursor, it is able to derive high quality glass ceramics with adjustable dielectric constant to meet specific application requirement. Mixing ferroelectric and ferromagnetic ceramic powders together and bonded by sol precursor with same chemical composition as one of the components, it is able to derive multi-ferroic composite with high and adjustable dielectric constant and magnetic susceptibility. Using fine grained ceramic powder or ceramic slurry derived by this hybrid process, it is able to prepare thick film or thin film devices by screen printing or spin coating onto alumina substrates or silicon substrates and densified at much lower temperatures. In this paper, barium-strontium titanate (Ba~,Sr,)Ti03, (BST) as a widely concerned tunable microwave ceramic will be used as an example to introduce this new hybrid ceramic processing technology.

Hybrid Ceramic Processing Technology 13, 41

Figure 1 is a typical flow chat of the hybrid process of BST ceramics. Firstly, BST powder was synthesized and calcined using conventional mixed oxide ceramic process from either BaTi03 and SrTiO3 or BaC03, SrCO3 and Ti02 at high temperatures then subjected to high energy ball mill to form nano sized powder. The size of the powder thus obtained is in the range of tens of nanometers and is quite uniform. To avoid agglomeration of the powder during high energy ball milling, organic surfactant is added. On the other hand, BST sol precursor in the same chemical composition as the powder was synthesized via wet chemistry route, starting from Ba(CH3C00)2,, Sr(CH3C00)2, and Ti(OC2H5)4. Organic solvent was used in the synthesis to adjust the molar concentration and viscosity of the BST sol. The details of both processes are quite common and can be easily found else where. To graft the sol gel process into conventional ceramic process, pre-synthesized BST sol was introduced and mixed with BST nano sized powder by conventional ball milling. Due to the effect of the organic surfactant used, the BST powder can be well dispersed and suspended in the BST sol solution forming very uniform paint-like ceramic slurry. In this hybrid process, the molar ratio of BST introduced in the form of powder and in the form of sol solution is a very important processing parameter. In most of cases, the molar ratio is in the range between 70/30 to 80/20 and should be optimized from case to case. The BST slurry thus obtained can be directly used in spin coating deposition of thin and thick films onto substrates. The viscosity of the slurry can be adjusted by the amount of organic solvent to meet the requirement of spin coating technology. For preparing bulk ceramics by conventional ceramic process or thick films by screen printing process, the solvent of the slurry should be removed by heating at elevated temperatures in the range of 200-400' C.

143

During the heating, hydrolysis and polymerization of the BST sol took place. Liquid BST sol transformed into amorphous solid BST gel. Then the mixture of crystalline powder and amorphous gel can be crushed and ball milled again to form hybrid BST powder which is suitable for the further ceramic and screen printing processing.

Ball milling I I I

Ba(CH,COO)z+

Sr(CH,COO),+

Ti(OC4Hd4

I . BST Sol precursor

d Uniformly suspended slurry

L E I Hybrid BST powder

Figure 1. Flow chat of hybrid process of BST ceramics

Using the hybrid BST powders prepared by this new hybrid processing technology as described above, the sintering temperature of the ceramic can be significantly reduced without substantial deterioration of its physical properties. Figure 2 is the microstructure and temperature spectrum of dielectric constant of (Ba.6sr0.4)TiO3 ceramic prepared by conventional ceramic process. The sintering temperature of the ceramic is about 1400' C, the grain size of the ceramic is in the range of 5-10 pm, the peak dielectric constant is around 6800 (1KHz). Though the dielectric constant in this case is very high due to the well developed very large grains, however, the grain size is too large and the microstructure is not very uniform as required in many cases.

Temperature 'C

Figure 2. Microstructure and temperature spectrum of dielectric constant of (Ba&ro,4)Ti03 ceramic prepared by conventional ceramic process sintered at 1400' C, 2hrs.

144

Figure 3 is the microstructure and temperature spectrum of dielectric constant of (Ba0.6Sr0.4)Ti03 ceramic prepared by hybrid ceramic process using nano-sized original BST powder. The sintering temperature of the ceramic is 1200' C, the grain size of the ceramic is in the range of 1-2 pm. It can be seen from the micrograph that the ceramic is well densified at a temperature 200' C lower than the sintering temperature of the conventional process. The lower sintering temperature of the hybrid process is most inherited from the low crystallization and sintering temperature of the amorphous gel components of the hybrid BST powder. Sol-gel process is usually characterized with much lower sintering temperature because of higher chemical reactivity and shorter diffusion length. The peak dielectric constant of the ceramic is about 3400 (1 KHz), which is lower than that of the conventional process mostly due to the small grain size.

Figure 3. Microstructure and temperature spectrum of dielectric constant of (Ba0,6Sro,4)Ti03 ceramic prepared by hybrid ceramic process using nano-sized original BST powder and sintered at 1200' C, 2 hrs

Figure 4 and Figure 5 are the microstructure and temperature spectrum of dielectric constant of (Ba0.6Sr0.4)Ti03 ceramics prepared by hybrid ceramic process using micro-sized and nano-sized original BST powders respectively. The sintering temperature of the ceramic is 1230' C , which is a little higher than the previous case. The grain size of the ceramics in this case is in the range of 1-5 pm with very small grains.

Figure 4. Microstructure and temperature spectrum of dielectric constant of (Ba0&0.4)Ti03 ceramic prepared by hybrid ceramic process using micro-sized original BST powder and sintered at 1230' C, 2 hrs

145

Figure 5. Microstructure and temperature spectrum of dielectric constant of (Bao.6Sr0.4)TiO3 ceramic prepared by hybrid ceramic process using nano-sized original BST powder and sintered at 123OoC, 2 hrs

The microstructure of ceramics prepared by hybrid processing under a little higher temperatures as shown in Figure 4 and Figure 5 is apparently very inhomogeneous. There are big grains (-5 pm) and small grains (-1 pm and smaller). As a matter of fact, with higher magnification of the microstructure, finer substructure in nanometer range can be observed. Such structure is very similar to concrete structure in civil engineering with big stones, smaller sands and gel substance from cement. We can call the observed ceramic microstructures as concrete-like structure. Though the concrete-like structure is very inhomogeneous, which is against the common pursuit of ceramic processing, however, the material obtained from this hybrid process is well densified and exhibits very good mechanical behaviors. In some special cases, such concrete-like structure will be of benefit to the design and fabrication of ceramic devices. We will not discuss it here in detail.

The peak dielectric constants of BST ceramics derived from the hybrid process under a little elevated temperatures are 5800 and 3400 (1 KHz) respectively for the case of using micro-sized and nano-sized original BST powders as shown in Figure 4 and 5. Considering the peak dielectric constant values of ferroelectric ceramics are very sensitive to their microstructure and apparent density, the dielectric behaviors of BST ceramics derived from this hybrid process and sintered at much lower temperatures are almost as good as those of the materials prepared by conventional process and sintered at very high temperatures. The rather high value of dielectric constants of the ceramics derived from hybrid process are mostly inherited from the well crystallized original BST powders, while the sol derived smaller BST grains stuffing the inter-grainular space making the material very dense. As a matter of fact, to compare two different dielectric materials, the overall dielectric behaviors, specifically the dielectric loss instead of peak dielectric constant are in high priority. From this point of view, BST ceramics derived from hybrid process with nano-sized original powders are better.

Fine Grained BST Ceramics Fine grains and uniform microstructure are still the major objective of many advanced ceramic

materials as required by the developing of advanced ceramic devices. Such “ideal” structure is always symbolized as an “ideal” ceramic. However, to achieve this objective, sometimes, some of their physical behaviors have to be sacrificed. To achieve “better” structure and better performance at the same time of a ceramic material is a challenge to ceramists. The hybrid process suggested in

146

this work provides a very promising way to achieve the objective. The low sintering temperature of the hybrid process prevents the ceramic from grain growth at high temperatures and keeping the grain size small. However, the sintering process of ceramics, by its nature, is a diffusion and mass transfer process, grain boundary migration and grain growth are still there. To further suppress the grain growth process keeping the grain small and uniform, a two step sintering method was used. Figure 6 is the microstructure and temperature spectrum of dielectric constant of (Ba0&0,4)Ti03 ceramic prepared by hybrid ceramic process using nano-sized original BST powder and sintered at 1200' C for 10 minutes and 1150' C for 7 hrs. During the sintering process, the material was heated quickly to 1200' C and kept for 10 minutes to trigger the sintering process, then cold down to 1150' C and kept at that temperature for 7 hours to complete the sintering process while prevent the further growth of grains.

Figure 6. Microstructure and temperature spectrum of dielectric constant of (Ba0&0,4)Ti03 ceramic prepared by hybrid ceramic process using nano-sized original BST powder and sintered at 1200' C for 10 minutes and 1 150' C for 7 hrs.

It can be seen from the microstructure of Figure 6, the grain size is in the range of 100-500 nm and the microstructure of the material is very dense and uniform. The peak dielectric constant of the material is about 4800 (1 KHz) with very low dielectric loss. In comparison with the results of conventional process presented in Figure 1, overall speaking, the BST ceramic derived from hybrid process and two step sintering is much superior to the material derived from conventional process. The new process is characterized with lower sintering temperature, finer and uniform microstructure and lower dielectric loss.

Thick BST Films Thick electromagnetic ceramic films are able to develop better microwave devices. However,

the potential of thick film devices has not yet been adequately explored. One of the reasons is the compatibility between thick film technology and microwave packaging technology. Another limitation is materials to fabricate thick film devices are quite limited. Many of ceramic materials have to be sintered at high temperatures, which would not be able to cofire with substrates and metals as required by the thick film technology. In order to meet the cofiring requirement, glass powder was mixed with the ceramic powder to reduce the sintering temperature, which would deteriorate the performance of the ceramics significantly. With this new hybrid process, this problem can be solved satisfactorily. Figure 7 is the cross section view of a BST thick film screen printed onto alumina substrate and cofired at 1200' C for 2 hrs. The BST printing paste is composed of hybrid BST powder and organic binder and solvent. The thickness of the film is about 15 pm, the

147

peak dielectric constant of the film is 1800. The film can be cofired with Ag-Pde ta l paste to form various microwave devices.

Figure 7. Cross section view and temperature spectrum of dielectric constant of (Bao.&o.,)Ti03 thick film prepared by hybrid ceramic process using a mixed micro-sized and nano-sized original BST powder, sintered at 1200' C, 2 hrs

Discussion The hybrid ceramic processing technology described above is able to effectively reduce the

sintering temperature of most advanced ceramics and ceramic composites, while keeping the physical performance close to the ceramics sintered at much higher temperatures. The low sintering temperature of hybrid process is particularly favorable to prepare fine grained ceramics with highly uniform microstructure, which is required by many of the advanced ceramic devices. In this respect, two step sintering technique is particularly effective and useful. Well densified sub-micron structured ceramics can be achieved. Manipulating the sintering process of the hybrid technology, concrete-like structured ceramics with dense microstructure can be developed. The Physical performance of concrete-like ceramics is almost the same as that of the material prepared by conventional ceramic process and sintered at much higher temperatures. The low sintering temperature of this hybrid ceramic processing technology is particularly useful in preparing ceramic composites or multi-phase ceramics composed of components with completely different chemical compositions and crystal structures such as ferroelectric-ferromagnetic multi-ferroic ceramics. The low sintering temperature of this hybrid process also renders the thick film technology to be able to incorporate many of the advanced ceramics with high sintering temperatures, which prevented their application in the thick film devices before.

The hybrid ceramic processing technology is a very promising new technology and will be able to derive better microwave materials and devices with advanced electromagnetic ceramics.

Acknowledgement The author would like to acknowledge Dr. Wang Zhihong for his initiation work on PZT

ceramics and thick films in developing the idea of hybrid process. Dr. Zhang Hongfang is acknowledged for her comprehensive experimental studies on BST ceramics and thick films. The results of BST ceramics and thick films presented in this work are mostly based on her works. The Chinese National 973 project under contract number: 2002CB613302 and 2002CB613304 are mostly grateful.

148

References [l] D. A. Barrow, T. E. Petroff, R. P. Tandon, M. Sayer, “Thick ceramic coating using a sol gel

based ceramic-ceramic 0-3 composite,” Surface and Coating Tech. 76-77: 113-1 18 (1995). [2] D. A. Barrow, T. E. Petroff, R. P. Tandon, M. Sayer, US Patent, Patent Number 5,585,136

(1996). [3] Weiguang Zhu, Zhihong Wang, Changlei Zhao, et.al. “Low Temperature Processing of

Nanocrystalline Lead Zirconate Titanate (PZT) Thick Films and Ceramics by a Modified Sol-Gel Route,” Jpn. J. Appl. Phys. Vol41: 6969-6975 (2002).

[4] H. F. Zhang, Xi Yao, L. Y. Zhang, “Barium Strontium Titanate Ceramics and Thick-Film Preparation by Low-Temperature Sintering”. J. of Chinese Ceramic Society, 33 (1 1): 1360-1365 (2005).

[5] H. F. Zhang, Xi Yao, L. Y. Zhang, “Microstructure and Dielectric Properties of Barium Strontium Titanate Thick Films and Ceramics with a Concrete-like Structure “, J. of Amer. Ceramic. SOC. (accepted and to be published in 2007)

Low-fire Processing Magnetic+Dielectric Ceramic Composites

Te-Ming Peng, Rung-Tsung Hsu, Chia-Jung Chung and Jau-Ho Jean

Department of Materials Science and Engineering National Tsing Hua University

Hsinchu, Taiwan

Abstract A low-fire Ni-Cu-Zn ferrite+dielectric ceramic composite and its multilayer integrated

laminate, which can be densified at temperatures below 9OO0C, is prepared. To densify Ni-Cu-Zn ferrite at low temperatures, the dielectric needs to contain an effective sintering flux of Bi203 or PbO to the Ni-Cu-Zn ferrite. This enables to densify the resulting ceramic composite or multilayer integrated dielectridferrite laminate at temperatures close to those of pure ferrite and dielectric. The dielectric constant decreases but the initial permeability increases with increasing the amount of ferrite existing in the ferrite+dielectric ceramic composite.

1. Introduction Low-temperature-cofired ceramic (LTCC) with mixed dielectrics or integrated

dielectric/magnetic materials has been identified to be an enabling solution to fabricate miniature filter or electro-magnetic interference (EMI) devices. Undesirable cofiring defects including de-densification, de-bonding, cracks and camber [ 1-31 are always found when different materials are cofired to fabricate monolithic ceramic components. Major root causes to the formation of above cofiring defects are mainly attributed to sintering incompatibility, and densification rate mismatch between different materials [4,5]. The former is mainly chemical, which can be solved by adding common sintering flux, low-softening-point glass or liquid-phase-sintering aid into dielectric and magnetic ceramics. By controlling the amount of sintering fluxes, particle sizes of dielectric and magnetic materials, heating rate or packing density of green tapes, the level of densification rate mismatch could be minimized to reduce the magnitude of tensile stress generated during cofiring. Another approach to solve the cofiring issue of densification rate mismatch is to use ceramic powder mixture, which contains dielectric and magnetic powders. The ceramic mixture is then fired to form dielectric+magnetic composite, which exhibits dielectric and magnetic characteristics in one phase. Although this method can avoid the problem of densification rate mismatch between different ceramic layers, excellent sintering compatibility between dielectric and magnetic powders during cofiring is still required to densify the ceramic composite at temperatures close to those of pure dielectric and magnetic materials. The objective of this study to develop low-fire dielectric+magnetic composite systems, which can be densified at temperatures no greater than 900 "C . Three dielectric materials including Bi2(Zn1/3Nb~/3)~0~ (BZN), Pb(Ni1/3Nb2/3)03-PbZr03-PbTi03 (PNZT) and BaNd2Ti4012+ZnO-B203 (BNTZB) are used. PNZT is a mixture of Pb(Ni1/3Nb2/3)03, PbZrO3 and PbTiO3, exhibiting a pure spinel phase. For BNTZB systems, a 10-15 vol% ZnO-B203 (ZB) glass is mixed with BaNd2Ti4012. Ni-Cu-Zn ferrite has a composition of (Nio.3Cuo.1Zno.60)-(Fe203)o.8, which is prepared by conventional solid-state reaction. The above ceramic powders hve a median size of 0.5-1.0 pm, and can be densified at 875-900°C for 2 h. The densified BZN, PNZT, and BNTZB exhibit a dielectric constant of 85, 2000, and 72 at lMHz, respectively. The densified Ni-Cu-Zn ferrite has an initial permeability of 150 and dielectric constant of 20 at 1MHz. Detailed experimental procedure can be found in the paper published previously [6] , and is not repeated here.

149

150

2. Results and Discussion Figure 1 shows typical densification results of three dielectric+ferrite powder composite

systems fired at 900°C for 120 min. Poor densification is observed in the systems with BNTZB, which is due to the fact that the ZB glass in the BNTZB is not an effective sintering flux to Ni-Cu-Zn ferrite. In contrast, a high relative sintered density greater than 90-95% is achieved in the BZN and PNZT dielectric systems, where Biz03 and PbO are contained in the dielectric systems, respectively. The above results are consistent with those reported previously that either Biz03 or PbO lowers the densification temperature of Ni-Cu-Zn ferrite [7,8]. Similar results are also observed when multilayer dielectric/ferrite laminates are used to measure the linear shrinkage in the X and Z directions. In other words, an effective sintering flux to the ferrite contained in the dielectric ceramics, such as Biz03 or PbO which enhances the sintering compatibility between dielectric and ferrite, is critical to achieve densified ferrite+dielectric ceramic composite and its multilayer integrated laminate during cofiring. In the following, results of the BZN+ferrite composite are given to demonstrate the effect of ferrite content on densification, crystallization, dielectric and magnetic properties of the BZN+ferrite ceramic composite.

Effects of ferrite content on the crystallization behavior of BZN+ferrite composite, fired at 900°C for 120 min are shown in Fig. 2. At 10 vol% ferrite, a new Bi-deficient, crystalline phase of Bil.sZnNbl.507 along with the original crystalline phase of Bi2Zn0.67Nb1.3307 are formed. For the ferrite contents in the range of 30-90 vol%, however, only the new phase of Bil.sZnNbl.507 is detected. This suggests an interfacial reaction taking place between BZN and ferrite. This is also evidenced by SEMEDX results where the Biz03 from BZN diffusing into ferrite, forming a Bi-deficient layer at the interface of BZNlferrite is noted. The above chemical reaction not only changes the crystallization behavior of BZN, but also improves the sintering compatibility between BZN and ferrite because Biz03 was reported to be an effective sintering flux to Ni-Cu-Zn ferrite [7]. This in turn promotes the densification of BZN+ferrite ceramic composite and its multilayer integrated laminate at temperatures required for those of pure BZN and ferrite.

Results in Fig. 3 summarize the dielectric constant and initial permeability at 1 MHz for the composites with different ferrite contents. The dielectric constant initially increases when the ferrite content is less than 30 vol%, but decreases with increasing ferrite content when the ferrite content is above it. This is due to the formation of a new crystalline phase of Bil.sZnNbl.sO7, which has a dielectric constant (E) of 170 [9] much larger than those of Bi2Zn0.67Nb1.3307 ( ~ 8 5 ) and ferrite (~=20). At ferrite contents less than 30 vol%, both BZN phases co-exist. However, only the high-dielectric-constant Bil.sZnNbl.507 phase is formed when the ferrite content is greater than 30 vol%, and that the amount decreases with increasing ferrite content, in agreement with XRD results in Fig. 2. Results in Fig. 3 also show the initial permeability increasing with increasing ferrite content in the BZN+ferrite composite.

3. Conclusions A binary Ni-Cu-Zn ferrite+dielectric powder composite and its multilayer ceramic laminate,

which can be densified at temperatures below 9OO"C, is developed. One of critical factors in designing ferrite+dielectric powder composite or multilayer laminate cofirable at temperatures close to those of pure ferrite and dielectric is to have an effective sintering flux in the ceramic dielectric, such as Bi203 or PbO, to Ni-Cu-Zn ferrite. The dielectric and magnetic properties of the resulting ferrite+dielectric ceramic composite can be fine-tuned by changing the relative composition ratio of ferrite to dielectric powder.

References [l] P. Z. Cai, G. L. Messing and D. L. Green, J. Am. Cerum. Soc., 80, Vol. [8], 1929 (1997). [2] T. Cheng and R. Raj, J. Am. Cerum. Soc., 72, Vol. [9], 1649-55 (1989). [3] S. Ho, C. Hillman, F. F. Lange and Z. Suo, J. Am. Cerum. Soc., 79,2353 (1996). [4] J. H. Jean, C. R. Chang and Z. C. Chen, J. Am. Cerum. Soc., 80, Vol. [9], 2401 (1997).

151

[5] R. T. Hsu and J.H. Jean, J. Am. Ceram. SOC., 88, Vol. [9], 2429 (2005). [6] T. M. Peng, R. T. Hsu and J. H. Jean, J. Am. Ceram. SOC., 89, Vol. [9], 2822 (2006). [7] J. H. Jean and C. H. Lee, Jpn. J. Appl. Phys., 38,3508-3512 (1999). [8] J. H. Jean and C. H. Lee, J. Am. Ceram. Soc., 82, Vol. [2], 343-50 (1999). [9] X. Wan, H. Wang, and X. Yao, J. Am. Ceram. Soc., 80, Vol. [ 101,2745-48 (1 997).

152

8000

= 6000 -

.- x lo

c 8 4000 -

2000

-.--BZN -.- PNZT 4- BNTZB

-Ferrite Content (~01%) mi&wNb?.so, m i d n N b ~ . s o ~ LNiCu-Zn Ferrite

- 100 :& 6 6 5 - - 90 & I -

- 'O L. - 50 4 A

- lo -& N! . . 30 g * , FF E R -

nh

0 20 40 60 80 100

Ni-Cu-Zn Ferrite Content (~01%)

Fig. 1 Relative sintered density as a function of ferrite content for the powder composite of BZN+ferrite. PNZT+ferrite and BNTZB+ferrite fired at 900% for 120 min.

20 30 40 50 60

2 e (degree)

Fig. 2 Effect of ferrite content on the crystallization behavior of BZN+ferrite powder composite fired at SOOT for 120 min.

120, t I60

s . 0

'E u 60 -

40-

, . I . I . I . I

0 20 40 60 60 100

Ni-Cu-Zn Ferrite Content (~01%)

Fig. 3 Effect of ferrite content on dielectric constant and initial permittivity at 1 MHz for BZN+ferrite powder composite fired at 9OO'C for 120 rnin.

Microwave Li-ferrite Material for use in Microstrip Tunable Devices

Rajeev Pourusha2*, N. K. Badolab, Ashoka, P. K. S. Pourush', G. S. Tyagia and G. P. Srivastavad

aDepartment of Physics & Computer Science, Dayalbagh Educational Institute, Dayalbagh , Agra-282005, India

bFerrite Division, Solid State Physics Laboratoy, New Delhi, India 'Department of Physics, Agra College, Agra-282004, India

dDepartment of Electronic Sciences, University of Delhi, South Campus, New Delhi, India

*E-mail: rpourush@,rediffmail.com -

Abstract

The present communication deals with the preparation and characterization of various compositions of microwave Li-ferrite material. These compositions of Li-Ferrite material have been prepared by conventional ceramic method . The characterization of prepared ferrite material involves the magnetic, electrical and surface structural parametric studies like dielectric constant, saturation magnetization (47cMs), Curie temperature (Tc), density measurement and particle size determination. These measured characteristics are suitable for microstrip device development on these prepared samples in microwave frequency region.

1. Introduction Ferrites are ceramic ferromagnetic materials with general chemical compositions MO.Fe203

where M is a divalent metal such as iron, manganese, magnesium, nickel, zinc, cadium, cobalt, copper or a combination of these. The incredible advancement in microwave integrated circuit technology, and as a result, reduction in size and cost of high frequency circuits has given impetus, which has greatly contributed towards the progress in microwave materials. Ferrite materials are the backbone of many microwave devices and systems. [l-8].Unlike a magnetic metal, a ferrite is a magnetic dielectric that allows an electromagnetic wave to penetrate the ferrite, thereby permitting an interaction between the wave and magnetiozation within the ferrite. This interaction has been used to produce a variety of useful devices.

2. Composition The selection of compositional formula plays a vital role in determining magnetic,

structural and electrical properties of the ferrite material. The arrangement of ions in the crystal structure of ferrite affects the various interactions involved in the lattice. Therefore the preparation of ferrite needs careful experimentation keeping various factors in mind. In the present work, Lithium Zinc Titanium (LiZnTi) and Lithium Manganese Titanium (LiMnTi) ferrites have been synthesized from the basic components of lithium ferrites i.e. Lio.s Fe2.5 0 4 . Suitable adjustments of other metal ion concentrations have also been included accordingly in order to maintain the stoichiometry. In addition, a small content of Mn3+ has been included in both the compositions. It not only helps in suppressing the formation of any Fe2+ ions in the ferrite, which promote Fez+ - Fe3+ conduction but also influences magnetostriction being a Jab-Teller ion [l, 21. In order to avoid the escape of lithia at high

153

1 54

temperatures of sintering, Biz03 (0.25 wt %) was added as sintering aid in both compositions [ 3 ] . The composition (ions per formula unit) of the samples are given here

I. For Lithium Zinc Titanium (LiZnTi) Ferrite

11. For Lithium Manganese Titanium (LiMnTi) Ferrites

3. Method of Preparation Preparation methodology adopted in the synthesis of ferrite plays an important role as it involves

various factors, which affect the characteristics of the materials to a great extent. These factors, which influence the course of reactions are homogeneous mixing of starting materials, furnace atmosphere, heating schedule, cooling cycle, nature of impurities and particle size etc. To avoid any variations occurring due to processing parameters, all the samples were prepared and characterized under similar conditions. The ferrite samples selected for the present investigations were prepared by the conventional ceramic technique. The ingredients required for the preparation of these ferrites were calculated on the basis of chemical formula of the samples mentioned above. Analytical grade (AR) LizC03, MnCO3, TiOz, ZnO, Fez03 and Biz03 were used. The stoichiometric ratios of the chemicals were thoroughly mixed in a polypropylene jar containing the zirconia balls. Distilled water was taken as the wetting medium. The mixed paste of the chemicals was then carefully taken out from the jar and dried in oven at 80 OC. The presintering of the mixed powder was carried out at 750 'C in a box furnace and the soaking time was kept four hours for both the samples. The presintered powder was again subjected to ball milling for another four hours in the same manner as used for the mixing in order to reduce the particle size of the powder and to obtain homogeneity in the material.

A small quantity of 2% solution of polyvinyl alcohol (PVA) was added to the mixed powder as binding agent. The binder mixed material was dried and then granulated by passing it through sieve of BSS - 52 mesh size. The sieved material was pressed in disk and toroidal shapes with the help of suitable dies and using hydraulic pressing technique at pressure of 1 todcm'. The pellets and toroids were finally sintered at 1050 'C for four hours. The heating and the cooling cycle of the samples were carried out in air atmosphere in the furnace.

All the samples of a particular composition were sintered together to keep the sintering parameters identical. The sintered samples so obtained were subjected to cutting, grinding, polishing etc. in order to get specific shapes and sizes and were used to measure magnetic and some of the electrical properties. The toroid was used to study the initial permeability in different frequency ranges and B-H loop characteristics. The surfaces of all the samples were grounded to remove any undesired layer formed during sintering.

4. Characterization The following properties of the prepared ferrite material have been studied:

(1) Curie Temperature (T,) ( 2 ) particle size ( 3 ) Saturation magnetization (4) Dielectric constant ( 5 ) Density measurement

155

5. Result and Discussion Values of the above-mentioned parameters of the substrates are measured and listed in table 1.

These values are quiet suitable for device development particularly in L band of microwave frequency band.

Acknowledgment

support to carry out this work. We are thankful to DRDO, Ministry of Defence, Government of India, for providing the financial

References [l] L. G. VanUitert., Proc. IRE- 44, 1294, (1956). 121 G. F. Dionne, IEEE Trans., Mug- 8, pp. 439, (1972). [3] Pran Kishan, D. R. Sagar, S. N. Chatterjee, L. K. Nagpaul, Nitendar Kumar and K. K. Laroia, Adv.

[4] J. Smit and H. P. J. Wijn, “Ferrites” Phillips Tech. Lib. Eindhoven, (1959). [5] F. G. Brockmann, P. H. Dowling and W. G. Steneck, Phys. Rev. - 77, 85, (1950). [6] R. F. Soohoo, “Theory and Applications of ferrite” Prentics Hall Inc., USA, (1960). [7] Dr - Ing Carl Hech, “Magnetic Materials and Their Applications” Butterworth, London, (1974). [8] H. F. Storm, “Magnetic Amplifiers” John-Wiley, New York, (1955).

Ceram - 15, 507, (1986).

156

TABLE 1

Sr. Parameter studied No.

Compositions

I I1

1. Particle size 2.59 pm 13.19 pm

2. Density 4.3466 4.0472

3. Curie Temperature (T,) 261 OC 359 OC

4. Saturation Magnetization 450 Guass 475 Guass

5 . Dielectric constant 14 16

Structural electrical-, magneto-transport and magnetic properties of ZnO embedded nanocrystalline CMR manganites ( L ~ o . ~ S ~ O . ~ M ~ O ~ ) I . , ( Z ~ O ) ,

S. Paul, B Singh, T. K. Nath*

Magnetism & Magnetic Materials Laboratory, Department of Physics & Meteorology Indian Institute of Technology, Kharagpur: 721302, West Bengal, India.

Email: sanioy.phy@,,amail.com

1. Introduction Hole-doped manganites perovskite of type Lal.xSrxMn03 have enormous research interest in

recent past both in fimdamental and applied physics due to their important magneto-transport properties. These materials reveal a large magnetoresistance (MR) [MR= {p (H) - p (O)}/ p (O)], termed as colossal magneto resistance close to room temperature. This effect is known as CMR effect [l-51. This effect can be demonstrated with the help of most popular model “Spin-polarized intergrain tunneling”, proposed by Hwang et a1 [6]. According to Hwang model, adjusting the barrier layer by altering the size of the ferromagnetic grains or diluting with a nonmagnetic compound will significantly influence the tunneling process hence the MR. Recently, enhanced low-field magnetoresistance (LFMR) effect or enhanced room temperature magnetoresistance effect were achieved successfully for manganites composite with non-magnetic or anti-ferromagnetic insulating inorganic oxides, for examples Lao.7Sr0.3MnO3: Ce02, Lao.7Sr0.3MnO3: SrTiO3 [7-81. On the other hand, Yan et a1 [9] also briefly reported enhanced LFMR in a Lao.7Sro.3MnO3: CoFe204 composite, where LSMO is a soft FM metal and CoFe204 a hard FM insulator. Under consideration of the above-mentioned studies, we have prepared a series of (Lw.7Sro 3MnO3) (ZnO) composites and have studied structural, electrical- and magneto- transport and magnetic properties.

2. Experimental Details a) Synthesis Process

(ZnO) (x = 0 to 0.9) were prepared from high-purity LazO3, Mn (CH3COO) 2-4Hz0, Sr (No3) 2 and Zn (No3) y6H2O chemical route “Pyrophoric reaction process. The final products are calcinated at a fixed temperature 650’ C (in air) for 3% h to get nanocrystalline powders. b) Structural Characterizations

The structural analysis has been carried out using XRD, TEM and SEM. XRD pattern in Fig. 1 (a) shows that the prepared precursor powders of (Lao.7Sr0.3Mn03) I - ~ (ZnO) calcinated at 650 “C for 3% h, is combined phase powders, which correspond to a pseudo cubic perovskite (Lao.7Sro.3MnO3) and a hexagonal structure (ZnO). No extra phase is obtained indicating that the reactions between Lao.7Sro.3Mn03 and ZnO grains are negligible. Fig. l(b) shows the TEM micrographs for x = 0 from which the average particle size was estimated in nanometric regime ((p - 15 nm). The nanometric particle size and grain morphology of this nano-composites were directly observed through SEM. Figure l(c) shows the SEM micrographs for x = 0.1 confirming the particle size are in nanometric regime and there is phase segregation, where the Lao.7Sr0.3MnO3 grains are well surrounded by randomly distributed ZnO grains.

Nano-composites (Lao.7Sr0.3Mn03)

3. Results and discussion The basic electrical characterization of composite samples has been done by resistivity

measurement down to T = 80 K. Figure 2 (a) and Fig.2 (b) represent the temperature dependence of normalized zero-field resistance R (T) /R (280 K) as a function of temperature for x=O and 0 5 x 5 0.6. It is evident from the inset of Fig. 2(a) that a broad metal-insulator transition occurs near 250 K for the pure sample with x = 0. The sample with x = 0.1 demonstrates a broad transition at lower temperature where the resistivity value remains almost constant down to lowest attainable temperature (T = 80 K). The metal-insulator transition is understood from a parallel resistor model

157

158

where the sample consists of parallel network of good and bad conduction channels [lo]. In systems where a nonmagnetic phase exists at the grain boundary surrounding the grains comprising of pure magnetic phases, the number of charge carriers and their mobilities, are expected to be determined by tunneling across the grain boundaries. Also it is evidenced that the normalized resistivity ratio increases with x and the percolation threshold of these composites occurs at x = 0.5 and above x = 0.5 sample become insulator.

samples having average particle size of 15 nm for T = 80 K up to 1 T field. The metal-insulator transition has been well understood from a parallel resistor model where the sample consists of parallel network of good and bad conduction channels [lo] as mentioned earlier. It is observable that (Lao.7Sro.3MnO3) (ZnO) composites exhibit enhanced MR at low temperatures compared to the parent compound. At T = 80 K all samples exhibit the usual behavior of polycrystalline sample with a large negative %MR at very low field regime (low field magnetoresistance (LFMR), H < 2 kOe) followed by a slower varying MR at comparatively high field regime, where MR is almost linear in H and p (H) decreases continuously without any sign of saturation up to a field of 10 kOe. Moreover, the LFMR at T = 80 K initially increases with increase of x up to x = 0.2 beyond which it starts to decrease. It seems that both resistance and MR are very sensitive to the ZnO content. The MR results at a field of 2 kOe for all value of x are shown in the inset of the Fig. 3. The origin of LFMR is mainly attributed to the spin-polarized electron tunneling across the ferromagnetic manganites grain boundaries with intermediate insulating ZnO as the tunneling barrier. Nevertheless, the induced grain boundary inside Lao.7Sro.3Mn03 makes additional contribution to LFMR, which is responsible for the enhanced part of LFMR.

with different x values (x= 0, 0.3, 0.5) shown in fig. 4(a), fig. 4(b) and fig. 4(c) at different temperature shown. Result shows that magnetization decreases with increase in weight percent (x) of ZnO. This behavior is expected because non-magnetic behavior of ZnO. From magnetization data we have determined their coercivity which is found to be increasing with the ZnO percentage in the nanocomposites at 20 K and 100 K and samples are found to be super paramagnetic in nature. There is no coercivity at room temperature. It indicates that higher percentage of ZnO nanocomposites become magnetically harder and harder. It is clear from this figure that appearance and disappearance super paramagnetism for all samples as temperature is increased beyond 285 K. These results suggest that magnetic properties of a composite can be tuned in a controlled fashion by changing the ZnO composition and temperature. The presence of single domain regions is responsible for these high values of Hc, where magnetization reversal takes place only by rotation of saturation magnetization vector.

Figure 3 shows the %MR measurements for nanocrystalline (L~.7Sro,3MnO3) lSx (ZnO)

Magnetization study (M-H plots) of the nanocomposite. samples ( L Q . ~ S ~ O . ~ M ~ O ~ ) l-x (ZnO)

4. Conclusion composites by

chemical “Pyrophoric reaction process and investigated in detail their electric, magnetic and magneto-resistive properties. Here ZnO components embedded in the CMR manganites L~.7Sro,3MnO3 matrices. Enhanced MR effects from low temperature up to room temperature are observed in these composites which is strongly depends on the weight fraction of ZnO. Moreover, the magnetic disorder caused by ZnO could provide an additional contribution to the enhanced MR. This effect could be further enhanced by controlling the size, shape and as well as by controlling interfacial disorder by controlling the proper annealing sequence. These materials will be used as high-density storage, sensing devices and next generation digital recording.

We have prepared a series of two-phase-mixed (Lao.7Sro.3MnO3) I - ~ (ZnO)

References [ 11 B.D. Cullity, Introduction to Magnetic Materials, Addison-Wesley Publishing Company,

[2] R. V. Helmolt, J. Weoker, B. Holzapfel and K. Samwer, Phys. Rev. Lett., 71,2331 (1993). [3] S. Jin, T. H. Tiefel, M. Mocormark, R. A. Fastnacht, R. Ramesh and L. H. Chen, Science, 264,

Massachusetts, MA, 1972

159

413 (1994).

914 (1995) [4] H. Y. Hwang, S-W. Cheong, P.G. Radaelli, M. Marezio, and B. Batlogg, Phys. Rev. Lett. 75,

[5] A. P. Ramirez, J. Phys.: Condens. Mutter, 9,8171 (1997) [6] H. Y. Hwang, S. W. Cheong, N. P. Ong and B. Batlogg, Phys. Rev. Lett. 77,2041 (1996). [7] LI. Balcells, A. E. Carrillo, B. Martinez and J. Fontcuberta, Appl. Phys. Lett., 74,4014 (1999). [8] S . Gupta, R. Ranjit, C. Mitra, P. Raychaudhury and R. Pinto, Appl. Phys. Lett. 78,362 (2001). [9] C. H. Yan, Z. G . Xu, T. Zhu, Z. M. Wang, F. X. Cheng, Y. H. Huang, and C. S . Liao, J. Appl.

[lo] P. Raychaudhuri, T. K. Nath, A. K. Nigam and R. Pinto, J. Appl. Phys. 84,2048 (1998) Phys. 87, 5588 (2000).

x = o 5

x=03 x = o 3

~ ~ 0 . 1 5

x=o.1

Fig. l(a) XRD patterns for ( L a 7Sro 3Mn03) with x =0, 0.1, 0.15, 0.2, 0.3, 0.5, 0.6, and 0.9. The asterisks (*) represent the different peaks from ZnO.

(ZnO) Fig. 10) TEM micrograph for (I,% 7Sro 3Mn03) (ZnO), samples with x = 0.

0.6-

20-

o 15-

... 10- o-"

100 IS0 200 250 300 T ( K )

Fig. 2(a) Temperature variation of normalized Fig. 2(b) Temperature variation of normalized resistivity for (La ,Sro 3MnO3) (ZnO) samples with resistivity for ( L a 3Mn03) (ZnO) samples with x = 0. Inset of this figure show the behaviors near metal-insulator transition.

0 5 ~ 5 0 . 6

160

0.00

-0.05

-0.10

.g -0.15

-0.20

-0.25

Fig. 3. The variation of YO MR with magnetic field (H) at T = 80 K. Inset (a) shows the variation of % MR with x.

E -20

-40

-60 I I I -10 -8 -6 -4 -2 0 2 4 6 8 10

H (k Oe)

20 15 10

E 5

v !lo E -5

'3

-10 -15 -20

-10 -8 -6 -4 -2 0 2 4 6 8 Hot*)

3

Fig.rl(a) M-H plots (L~ .~Sro .~Mn03) I,(ZnO) with x = Fig.rl(b) M-H plots (Lao,7Sr,,3Mn03) ,.,(ZnO) with x= 0 at T= 20 K, 100 K and 300 K. 0.3 at T= 20 K, 100 K and 300 K.

30

20

-20 4 I

-10 -8 -6 -4 -2 0 2 4 6 8 H(k Oe)

3

Fig.4(c) M-H plots (Lao.7Sr~.3Mn03) (ZnO) with x =0.5 at T= 20 K. 100 K and 300 K.

Low Cost Synthesis of Nanosized NiFezOd

N K Janjua’ ,S Imtiaz’ , T Hussain’

Corresponding author, Chemistry Department, Quaid-i-Azam University, Islamabad, Pakistan. ’ Chemistry Department, Quaid-idzam University, Islamabad, Pakistan. Email # nkausarjaniua@:vahoo.com and [email protected]

I

ABSTRACT: Pure nickel ferrite nanoparticles have been synthesized by aqueous sol-gel method in air

atmosphere and effect of chelating agent and temperature on the synthesis of nanosized NiFe204 was studied for optimization. Various physical properties were studied using different techniques like XRD, SEM, and FTIR. These particles showed strong ferromagnetism at room temperature in the as- synthesized form. The chelating agents affect the particle size because of their structural nature and number of chelating sites. In the present study, four reducers including citric acid, sucrose, ethylene glycol, and acrylic acid were used. The average crystallite sizes were obtained using Debye-Sherrer’s equation. The sizes of the ferrite particles obtained using sucrose and ethylene glycol were comparable and minimum compared to other reducer:. TemperaEre conditions were also optimized by sintering the as-synthesized samples at 4OO0C, 600 C, and 800 C and the effect of annealing temperature was correlated with the particle size. It was inferred that the nanoparticles are obtained at 400°C. XRD and FTIR studies were carried out in order to reveal the structural properties by changing the temperature and chelating agent. It was concluded that the nanoparticles of nickel ferrite could be synthesized using simple methodology and sucrose as chelating agent which is cost-effective, easily available, and non- toxic.

1. Introduction: Synthesis and characterization of ferrite nanoparticles is an area of current research interest for

new technological and biomedical applications [l-21. Spinel ferrites, MFe204 (M = Ni, Zn, Mn, Co, Mg), are among the most important magnetic materials, for many applications such as ferrofluids, catalysis, microwave devices, magnetic materials, etc. [2, 3 - 61. The properties of ferrite materials are strongly influenced by composition and microstructure of their particles, which can be tailored by varying the preparation methodology. Conventionally, NiFezO4 powders are prepared by solid-state reaction. Other new and revised techniques have also been implied to prepare the ultra fine NiFe204 powders; such as, co-precipitation reaction [7], mechanical milling [8], microemulsion [9], and sol gel method [ 1 01. In this work, NiFezO4 nanosized particles have been synthesized by an alternative facile aqueous sol gellultrasonication route. The size of nanoparticles can be optimized by varying sintering temperature. It was also found that sintering temperature has an obvious effect on the particle size i.e., the size of particle increases with temperature.The nature of chelating agent has a significant effect on the course of sol gel synthesis of nanomaterials for many reasons such as number of chelating sites in the reducer, the toxicity associated with the reducer and the cost effectiveness. XRD and FTIR studies confirmed the formation of cubic spinel NiFezO4 nanoparticles in as- synthesized form with or without sintering and ultrasonication.

2. Experimental: NiFe204 powder was synthesized via the sol gellauto-combustion method by varying chelating

agents using the following procedure. Aqueous solution of Ni(N03).6H20 (Merck) and Fe(N03)3,9H20 (Paureac) were prepared in a 1:2 molar ratio. Four different chelating agents citric acid (Riedel-de Haen), ethylene glycol (Merck), acrylic acid (Acros) and sucrose (BDH) were used. metal nitrates and

161

162

one of the reducer were mixed in 3:4 molar ratio to get a homogeneous solution which was refluxed for 1 hour and then, slowly evaporated on a magnetic stirring hot plate at 70-80 "C to give a viscous gel. During this process the gel swells into fluffy mass. The gel auto-ignited to form a very fine powder of NiFe204 nanoparticles. The powder thus obtained was divided into two parts: one part was ultrasonicated using ultrasonic bath (Elma LC 30M) for 5 hours and other part was annealed at different temperature.(400°C, 600"C,800"C) for 12 hours at heating rate of 5"C/min in Muffle furnace (NEY 2-525). X-Ray Diffractometer (PW3040/xOX, pert PRO Console) with CUKa radiations was used for phase identification and crystallinity. The samples were characterized for their purity using Fourier Transform Infrared Spectrophotometer (BIO-RAD EXCALIBOR FTS 3000/0 10-022 12-2). The spectrums were recorded in 400-4OOOcm-' range using KBr pellet technique. Scanning Electron microscopy (SEM) using LEO440 MODEL was used to examine the morphology of the particles.

3. Results and Discussion: 3.1 FTIR:

In ferrites the metal ions are situated in two different sub-lattices, designated as tetrahedral and octahedral according to geometrical configuration of nearest oxygen neighbours. The absorption bands v1 and v~ around 600 and 400 cm-' are attributed to the stretching vibration of tetrahedral, M04 and octahedral, M06 groups of ferrites, respectively. These bands, v1 and v2 are observed in all the samples of NiFezO4,The FTIR spectra of as-synthesized, ultrasonicated and annealed samples confirmed the formation of NiFezO4 spinels. All the as-synthesized samples showed a few organic impurity peaks besides the two ferrite bands. The annealed and ultrasonicated NiFea04 nanoparticles showed similar bands and no peak for organic impurity. This particular result confirms our view of synthesizing the NiFezO4 nanoparticles via sol-gel / ultrasonication route.

0.10 ' , . , , , , , . I , , . I , , sin i c m ism a m a m mm 1600 ram

Wavenu mberk rn4

Fig.1: The FTIR spectra of NiFezO4 synthesized via sucrose route, NSC = as-synthesized, NUSC = ultrasonicated, NSC400, NSC600, NSCSOO = annealed at 400"C, 600°C and SOO"C, respectively.

3.2 XRD Discussion: XRD patterns of NiFezO4 in sucrose series (NSC) of as synthesized, sonicated and annealed samples at 400,600 and 800°C respectively are shown in Fig. 2. The diffractograms indicate the formation of

163

Fig. 2: XRD pattern of Sucrose series annealed at various temperatures, a) NSC, b) NUSC, c) NSC400, d) NSC600, e)NSC800

spinel NiFezO4 with the signature peaks at 20 "- 18", 30", 35", 43", 53", 57", and 63" having (1 1 l), (220), (3 1 l), (222), (400), (422), (5 1 l), and (440) (hkl) projection values, respectively, exactly matching with the standard ICSD card no. 00-044-1485. For the as-synthesized gel, broadened peaks were observed in the XRD pattern. This indicates the formation of nanoparticles during the initial drying of precursor gel. Ultra sonicated sample show the peaks at same 0 values but these peaks become significantly broader, indicating that crystal size is probably smaller than other samples in the same series (fig. 2b). The phase formation process was monitored with increase in annealing temperature. At 400°C broad, weak bands at 20 - 35" and 62" obtained. The pattern obtained at 600°C show a slight increase of intensity of X-ray peaks. XRD pattern of powder obtained at 8OO"C, however show all signature peaks (ICSD card no. 00-044-1485). The crystallite size was calculated by applying Scherrer's formula, while in the as-synthesized samples obtained from XRD the crystallite size is in the range of 17-22 nm (Table 1). It has been observed that the nickel ferrites crystallize at as low as 4OOOC and even in the as- synthesized form.

Table [l]. Average Crystallite size of NiFezO4 nanoparticles.

3.3 SEM Discussion: The nanostructure of the as-synthesized and sintered samples at various temperature was

observed from SEM images. SEM micrograph of the sucrose sample annealed at 600°C (Fig.4) reveals uniformly distributed fine grain microstructure with less porosity and more agglomeration. The nanoparticles are less agglomerated in the as-synthesized samples (Fig.3).

164

Fig.3: SEM micrograph of NSC as-synthesized sample. at 600

Fig. 4: SEM micrograph of the NSC annealed

3.4 Resistivity Measurements:

(sucrose route) showed conducting behavior in comparison with other samples.

Acknowledgements: Thanks are due to Dr. J. I. Akhter (Deputy Chief Scientist, Head, Physics Division, PINSTECH,

P.O. Nilore, Islamabad, Pakistan), Dr. Rumana Qureshi (Assosiate Professor, Department of Chemistry, QAU Islamabad, Pakistan), Dr. M. Mazhar (Professor, Chairman, Department of Chemistry, QAU Islamabad, Pakistan) and Ms. Samia Saleemi (PhD Scholar, Department of Chemistry, QAU Islamabad, Pakistan) for their cooperation.

The as-synthesized nanoparticles were tested for resistivity measurements. The sample NSC400

References: [l] M. Sugimoto, J. Am. Ceram. SOC., 82,269-280 (1999). [2] L. Hasmonay, J. Depeyrot, M. H. Sousa, F. A. Tourinh, J. C. Bacri, R. Perzynski, Y. L. Raikher

[3] C. G. Ramankutty, S. Suunan, Appl. Catas A: General 218,39-51, (2001). [4] S. K. Kurinec, N. Okeke, S. K. Gupta, H. Zhang, T. D. Xiao, J. Muter Sci, 41, (24), 8 18 1-8 185,

[5] E. Olesen, J. Thonstad,J. Electrochem., 29,293-299 (1999). [6] W. B. Cross, L. Affleck, M. V. Kunetsov, J. Muter. Chem., (9), 2545-2552, (1999). [7] H. Nathani, S. Gubbala and R. D. K. Misra, Mater. Sci. Eng. B, (1 1 l), 95-100, (2004) [8] J. M. Yang, W. J. Tsuo, and F. S. Yen, J. Mugn. Mugn. Muter, (256), 2420, (2003) [9] S. A. Morrison, C. L. Cahill, E. E. Carpenter, S. Calvin, R. Swaminathan, M. E. Mchenery and V.

[lo]

and I. Rsenman, J. Appl Phys, 88, (1 l), (2000).

(2006).

G. Harris,J. Appl. Phys., 95, ( l l ) , 6392-6395, (2004).

and C. Ionescu, J. Opt. Adv. Muter., 5 , (l), 251-256, (2003). M. Popovici, C. Savii, D. Niznansky, J. Subrt, J. Bohacek, D. Becherescu, C. Caizer, C. Enache

Growth and Characterization of Neodymium doped Strontium Hydrogen Phosphate single crystals by Gel method

Valsamma.M.Samuel', Unnikrishnan.N.V2, 1ttya~hen.M.A.~ '"School of Pure & Applied Physics, Mahatma Gandhi University, Kottayam- 686560, Kerala.

India. 'Department of Physics, CUSAT, Cochin-22, Kerala, India.

Abstract The growth of neodymium doped single crystals of strontium hydrogen phosphate in a gel

medium, and their characterization by suitable physical and chemical methods are reported. Silica gel, obtained from SMS solution with a strongly acidic cation exchanger in the H-form, was used for crystal growth. Crystals having different morphologies were obtained. Most of them were spherulites of various dimensions and are opaque. The grown crystals were studied using XRD, FTIR, optical microscopy, EDAX and Thermal analyses.

1. Introduction The gel growth technique appeared quite attractive for growing crystals of group I1

compounds on account of its unique advantages in terms of crystals produced and the simplicity of the process [ 1,2,3]. Rare earth doped phosphates have not been thoroughly investigated. A variety of crystals suitable for research and technology can be grown in silica gels [4, 5 , 61. The alkaline earth phosphates have assumed greater importance in recent years in view of their use as phosphor matrices [7, 81. The present work is therefore directed towards the study of gel growth, and characterization of Nd3+ doped strontium hydrogen phosphate single crystals.

2. Experimental Procedure Crystal growth experiments were carried out in glass test tubes of 2.5 cm diameter and 15 cm

length. The growth process involves the diffusion of a mixture of 0 9M strontium nitrate solution and 0.1M neodymium nitrate solution into a gel in which ortho phosphoric acid is impregnated beforehand. The gel is set up by mixing sodium meta silicate solution of specific gravity 1.03 with aqueous 0.5M phosphoric acid solution in proper ratio such that the pH of the solution is 7. This gel was allowed to set for the desired gelling time. To this reset gel upper reactant is added slowly without disturbing the gel surface. Nucleation starts immediately. The anions of the outer reactants slowly diffused into the gel column containing phosphate ions and react together to form the expected crystals. A light pink precipitate is formed at the instant. This pink precipitation band increases in thickness gradually, as the diffusion proceeds into the gel. The thickness of the precipitation band is found to vary with the concentration of the upper electrolyte. Sphemlites are found below the precipitate after 3 days. The number of spherulites is found to decrease along the length of the gel column with corresponding increase in their size. The diffusion of ions into the gel charged with another ion will be greater near the gel solution interface. This is because of the high concentration of the ions near the gel solution interface.The maximum size observed is about 6mm. in diameter. The growth becomes stagnant after two weeks. Crystals having different morphologies were obtained. Optical microscopy is employed to study the morphology of the grown crystals. The morphology and quality of the gel grown crystals depend on the various gel parameters and the concentration of the reactants.

3. Techniques of characterization The grown crystals were characterized by XRD, FTIR, EDAX and optical microscopy. X-ray

powder diffractograms were recorded using Bruker D8 Advance model with CuKa radiation (h =1.542 An ) with a step of 0.01' and a step time of 0.2s. The XRD pattern is shown in Fig.l.and the data is given in Table-I.

165

166

wave numbe, (cm-') 3432.01

2398.18

1738.25

1132.21 1072.16

1

IntensitJ Assignments

s/b 0-Hsymmetric & asy.stretchin(

S Combination mode

S/b Combination mode

S P=O stretching vs P=O stretching

20 $0 40 50 60

2-Theta(degrees)

25.536

29.311

4000 3500 3000 2500 2000 1500 1000 500

Wavenumber(cm-')

3.48546 3.48546 100 2 0 0

3.04459 3.04458 37 1 2 0

Fig .1 XRD pattern Fig .2 FTIR spectrum

880.90

The FTIR spectrum was recorded using Thermo Nicolet Avatar 370 DTGS spectrophotometer with KBr pellet method over the wave number range 400-4000 cm-'. The FTIR spectrum is shown in Fig.2. The data is given in Table-11.

S I P-o asymmetri

The Thermo gravimetric analysis and Differential thermal analysis were performed on powdered samples using a Perkin Elmer thermal analyzer. The thermo gram was obtained by heating the sample from room temperature to 95OoC in an atmosphere of nitrogen with a heating rate of 15OC per minute. Differential scanning calorimetric analysis was performed on a Mettler, Toledo, DSC 822 system with a scanning speed of 10' C per minute.

In thermo gravimetric analysis, the mass of the sample is recorded continuously as a function of temperature as it is heated in a suitable environment, at a controlled rate .TG curves are characteristic for a given compound, as the changes involved are a result of rupture andor formation of various physical and chemical bonds at elevated temperatures that lead to the evolution of volatile products or the formation of heavier reaction products.

Table-I XRD data I Angle 2 I dA" I dA" I Inte I h k l I

Theta observed calculate nsit I I i d l Y % 1 I Table-I1 FTIR sDectral data

I 995.53 I s I P-~asymmetrie

I ~ 1 stretching 1 567.14 P-P stretching 405.28 0-P-0 bending

167

The TGA & DTG, TGA &DTA, and DSC plots of the sample were shown in Figs.3,4 & 5 respectively. The EDAX spectrum is shown in Fig .6.

4. Results and Discussion The unit cell parameters were calculated with the help of Whole pattern fitting and Rietveld

Refinement program [9,10]. The crystal system is found to be triclinic with a symmetry PI (2). The cell parameters are a = 7.1936 'A, b = 6.80729 'A, c = 7.26058 'A." = 94.673', p =

104.937' and y =88.777'. Volume = 355.543 'A3,

-

The FTIR spectrum shows the identification of 0-H bonding and P-0 bonding. The peaks were identified in comparison with earlier reports [11,12]. The band at 880.90 cm-' is due to P-O- H and of HP04 '-. The HP04 '- band is characteristic of hydrogen phosphates. This indicates that the difference in the electric charge between divalent Sr and trivalent Nd in the crystallized solid is compensated for by simultaneous substitution of divalent HP04 '- for trivalent Po4 3- which should increase the amount of phosphorous in the crystallized solid [13].

The TG curve shows that the compound is stable up to 300'C. It loses weight in the temperature range 300-475'C. The 5.5 % weight loss is due to the decomposition reaction producing compound Nd:SrZP207, which is supported by the endotherm in DTG curve and is confirmed in DTA curve. The broad downward endothermic peak in DSC at the temperature range 344-475'C corresponds to the decomposition of the sample. The peak temperature 424'C indicates the temperature at which the reaction is completed. The broad endotherm indicates a slow change in heat capacity [ 14- 171. The enthalpy value corresponding to the endothermic peak is - 17 1 1.5 1mJ.Based on thermo gravimetric and differential thermal analysis, the thermal decomposition stage of the

crystal is as follows: Nd:2SrH PO4 425' C ' Nd:SrZ PZ 0 7 + HzO

In the EDAX spectrum,the lesser intense peaks at 0.928, 5.033 and 5.72 KeV correspond to M a , L a and L p energies of neodymium giving a clue that the amount of Nd is very small in the crystal. The well defined peak at 0.523 KeV correspond to M a energy of oxygen,, 1.806KeV correspond to L a of strontium and 2.01 3 KeV correspond to M a of phosphorous present in the form of phosphate in the crystal. Thus the presence of Nd can be confirmed by EDAX.

9.5 - ....

0.000 .g

E

9.4 - 2 - - m - 4002%

- c 9.3:

'. 8.2-

.I 2 . b

Q . W G 9 1 -

9.0 -

8.9, , . , . , . , . , . , . , . , . -0.ooe 100 100 MO 400 600 000 100 800 900

Temperature ("C)

Fig 3. TGA & DTG of Nd:SHP

ZW 4w 6W Bw im Temperature ('c)

Fig 4. TGA & DTA of Nd: SHP

168

64 , . , . , . , . , . , . , . , . ~ , , . , . , 50 IW 150 200 250 300 350 400 450 500 550 6W

Temperature CC )

Fig 5 . DSC of Nd: SHP k.V

Fig 6. EDAX of Nd: SHP

5. Conclusions

Gel growth technique is suitable for growing crystals of Nd doped strontium hydrogen phosphate. The crystal system is triclinic with a symmetry P c(2). The P=O and 0 - H bond is identified from FTIR spectrum. It is evident from the TGA that the compound is stable up to 3OO0C and above 475'C it yields a stable residue of strontium pyrophosphate.

The presence of Nd can be confirmed by EDAX.

Acknowledgement

Government of India for a fellowship. One of the authors (Valsamma.M.Samuel) is thankful to the University Grants Commission,

References [ 11 Heinz K.Henisch, Crystal Growth in Gels, Dover Publications inc.Newyork. 1996. [2] Z.Blank, D.M.Speyer, W.Brenner and Y.Okomoto, Nature, 216, 1103 (1967). [3] C.Batra and J. Zemlicka,J. Cryst. Growth, 10, 158 (1971). [4] P.W.Ranby, D.H.Mash and S.T.Henderson, Brit.J.Appl.Phys.Suppl.4,(1954) S18-S25. [5] R. C. Ropp and R .W. Mooney, J. Electrochem.Soc.l07,15 (1960). [6] K. H. Butler, J. Electrochem. Soc.100,25 (1953). [7] H. Koelmans and A. P. M. Cox, J.Electrochem.Soc.104,442 (1957). [8] M. J. B. Thomas, K. H. Butler and J .M. Harris, Illuminating Eng.52,279 (1957). [9] Joint Committee on Powder Diffraction Standards Powder diffraction file, 33, 1335,( 1998). [ 101 R.Masse, J.C.Guite1, Acta CrystallographicaB,24, 1968 (1982). [l 11 K.Nakamoto, IR and Raman Spectra of Inorganic and Coordination Compounds, 4th Edn.

[12] V .S .Joshi and M .J .Joshi, Cryst. Res. Technol. 38, 817 (2003). [13] Kijowska R, J. MaterSci. 39,2017 (2004). [14] Willard.H.H, Merritt.L.L and Dean.J.A, Instrumental Methods 0fAnalysis,5'~ Edn. (1974). [15] Wes1ey.W.M and Wendlandt, ThermalAnaly~is,3'~ Edn.J.Wiley ,NY. (1964). [ 161 Bhadeshia.H.K.D.H, Material Science and Metallurgy, University of Cambridge, (2004). [ 171 Ewing.G.W, Instrumental methods of Chemical Analysis, 5th Edn, McGraw-Hill, NY (1985).

J.Wiley & sons, NY. 168 (1986).

Annealing Effect on Magnetostatic and Dynamic Properties of Fe61.4Ni3.6Cr3.2Si2.4Nb,.~Mn3.6B18 Amorphous Ribbons

E.E. Shalvauina"*, M.A. Komarovaa, V.V. Molokanovb, A.N. Shalygina

"Faculty of Physics, Moscow State University, 1 19992, Moscow, Russia bInstitute of Metallurgy and Metallovedeniya named A.A. Baikova, RAS, Russia

Abstracts The investigation of the magnetostatic and dynamic properties of as-cast and annealed

Fe61.4Ni3.6Cr3.2Si2.4Nb7.sMn3.gB1g amorphous ribbons was carried out. The above amorphous ribbon of 5-mm wide and 35-pm thick was prepared by the melt-spinning technique. The near-surface and bulk magnetic properties of the ribbons were investigated by using Scanning Kerr microscopy and a vibrating sample magnetometer, respectively. A four-probe AC technique was used to measure the magnetoimpedance (MI) ratio as a function of the field H, applied parallel to the ribbon length L and the AC measuring current ( i = 5 mA) of different frequency f (0.1 - 10 MHz). The strong influence of an annealing temperature on static and dynamic properties of the ribbons was discovered.

1. Introduction The development of bulk amorphous materials has attracted considerable interest of

researchers in the last years due to enormous potential of their practical applications [l-31. Recently [4 - 81, multicomponent bulk amorphous alloys with sizes up to a few millimeters were obtained. The above materials were prepared with low critical cooling rates between lo2 and 1 Ws. They exhibited a large glass-forming ability (GFA) and excellent soft magnetic properties. In order to obtain bulk amorphous alloys, the decrease of the critical cooling rate and the existence of a wide supercooled region are needed. The wide supercooled region of these alloys [l], defined as the difference between the crystallization temperature TX and the glass transition temperature TG (ATx = TX - TG), is the reason of their high GFA. The new Fe614Ni~6Cr32Si24Nb7sMn36B18 multicomponent alloy was obtained in [9] by using an idea of the eutectic interaction of glass formation phases. The alloy showed a wide supercooled liquid region (ATx = 60 K [9]) and a good glass-forming ability promising the formation of bulk soft magnetic materials. In this work, the results on the investigation of micromagnetic structure (equilibrium magnetization distribution), the magnetostatic and dynamic properties of the as-cast and annealed Fe614Ni3 6Cr3 $ 3 2 4Nb7 sMn3 6B18 multicomponent amorphous ribbons are presented.

2. Experimental The Fe614Ni3 6Cr3 2% 4Nb7 8Mn3 6B18 amorphous ribbon of 5-mm wide (w> and 35-pm thick

was prepared by the melt-spinning technique. The ribbon was cut into 20-mm length pieces, which were annealed in vacuum for l h at Tan" = 470 - 1080 OC. The microstructure of the studied samples was investigated by X-ray diffraction. The study of the near-surface magnetic properties of the ribbon samples was carried out employing magneto-optical micromagnetometer with a surface sensitivity equal to the 20-nm thickness depth. By scanning the light spot of 20-pm diameter along the ribbon surfaces, the distributions of in-plane magnetization components and local magnetization curves were measured by means of the transverse Kerr effect (TKE), 6. Here 6 = (Z- Zo)/Zo, where Z and lo are the intensities of the light, reflected from the magnetized and nonmagnetized sample, respectively; 6s is TKE at M = Ms; Ms is the saturation magnetization of the sample. Actually, dependences of 6(L, W, H)/& a M(L, W, H)IM, were found. These data allow to obtain the information on the ribbon near-surface micromagnetic structure. The bulk magnetic characteristics of the ribbon sample were measured by a vibrating sample magnetometer (VSM). A four-probe AC technique (HP4192A) was used to measure the magnetoimpedance Z as a function of the magnetic

* 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~AT is the 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

T = 470 OC - H = 2 0 O e -.- H = 20 Oe

O'O 0 200 400 600 800 OVo 0 200 400 600 800 W (urn) 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

T O C 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 [l l] 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 = 1 1 .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

[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)

[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

[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).

(2000).

8, 103 (1998).

T.A. Sviridova, N.P. D'aykova, Rus. J, Perspective Mater. 3, 10 (2003).

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. Voronezh State University, Universitetskaya pl. I , 394006, Voronezh, Russia,

$t@,nhvs. vsu. ru. Voronezh State Technical University, Moskovskii pr. 14, 394026, Voronezh, Russia.

I

2

Berliner Elektronenspeicherring-Gesellschaji fur Synchrotronstrahlungm. B. H., Albert-Einstein- Str. 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-diffraction pattern 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,~-spectra represent 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, Fe203 and 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 IS 7 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 COO I 141.

175

t2g state 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 COO containing 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.

92 i. 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, SiOz and ELNES Si L2,3 in FezSi04 [7].

Fig. 5. XANES 0 K spectra of nanocomposites 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 1 s --f 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 --f 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.

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

[2] L.V. Lutsev, N.T. Kazantseva, LA. Tchmutin, N.G. Ryvkina,Yu.E. Kalinin, A.V. Sitnikoff. J.

[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,

[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

[9] C. Mitterbauer, G. Kothleitner, W. Grogger, H. Zandbergen, B. Freitag, P. Tiemeijer, F. Hofer.

metullovedenie, 91, 21 (2001).

Phys. Condens. Mutter, 15, (22), 3665 (2003).

214422 (2001).

(1 997).

Ultramicroscopy, 96,469 (2003).

Session P9

Chair: J.A. Kong

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. 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

Thus, it seems logical to re-examine Veselago's original paper.

p and E, in particular negative index of refraction.

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

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.

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 =bnld 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 <O. These beliefs are based on references [2,3,4]. Note: No loss is necessary to obtain these features.

1.3. Evanescent waves grow with distance for nl< 0. When propagating waves change into evanescent waves, it

is usually because nl goes imaginary 1.51. Thus, in view of the phase advance postulated above, it should not surprise us that Pendry [6] suggested that evanescent waves in a medium with nl

< 0 would grow and not be attenuated as usual for nl > 0.

Ordinary vese1ag.3 Medium I Medium

for n, o

Refracted for n, c 0 Incident

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

Z-

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).

179

180

1.4. The field vectorsB, @ and the phase vector form a left-handed triplet for n1<0. Also shown by Veselago in his original paper [l]

was the fact that the field vectors E and fi and the direction of phase propagation, s , form a left-handed triplet when n ] < 0, see Figure 3, right. This feature is probably the least observed when performing experiments. However, as we shall see later, it is a theoretical point very powerful in determining whether we have a true Veselago medium or not.

It was not long after Pendry's postulates that a group of physicists at the University of San Diego made a combination of flat wires and split ring resonators, as shown in Figure 4 [7,8,9]. They then performed measurements on a wedge shaped body as shown in Figure 5 at the bottom left. The idea was as illustrated that the refracted field would depend strongly on the sign of the refractive index, q. In fact they measured the refracted field for a Teflon wedge ( E - 2.1) and obtained a refraction curve as indicated by the red curve in Figure 5 top. They also measured the refracted field from a wedge shaped assembly of wires and split ring resonators as shown at the bottom of Figure 5 to the right. Actually the blue measured curves in Figure 5 top were not measured by the San Diego group but were obtained later by a group from Boeing's Phantom work [lo]. They went to great length to obtain the exact refraction both in the far field (NIM 66 cm) as well as

NIM = Negative Index Material

ti

Right-Handed Wave LeR-Handed Wave

Figure 3. In an ordinary medium E , @ and direction of propagation s form a right-handed triplet (shown to the left). In Veselago's medium E , @ and direction of propagation s form a left-handed triplet (shown to the right).

Figure 4. Picture of the original periodic structure used by the San Diego group to demonstrate the presence of negative refraction.

1.2

1 .o

W" 0.6

2 0.6 5 0.4 z

0.2

0 -60 -40 -20 0 20 40 60 60 1C

73

E

Refraction Angle (Deg.) n,co

Horse shoe

. - - - - - - _ _ - -

Note: Blue curve is 20 dB lower than the red curve.

10

Figure 5. Top: Red curves represent the field refracted through a Teflon wedge as shown at the bottom left. Blue curves are perceived as being the field refracted through a wedge made of wires and split ring resonator as shown at bottom middle. Note how the side lobes of the blue curves are similar to the side lobes of surface wave radiation shown at bottom right. (The blue curves are actually -20 dB below (-1% power) the red curves even if they are all shown normalized to loo%.) From [lo].

181

the near field (NIM 33 cm). Note how the sidelobes in the far field pattern are almost gone for the near field case as is typically seen in antenna experiments. However, the most interesting feature is probably that the refracted field for the synthesized material is -20 dB or more below the refracted field from the Teflon case. Such a large loss can not be attributed to either ohmic or dielectric loss for frequencies below 100 GHz. This fact and the presence of sidelobes in the far field suggested to this writer that the refracted field for the synthesized material was actually not a refracted field but merely the radiation pattern for a surface wave that can exist only on a finite periodic structure [11,12,13, and 14 (Figure 4.9)].

2. On the field surrounding an infinite periodic structure of arbitrary wire elements located in one or more arrays

2.1. Single Array of Elements with One Segment Consider a single planar array as shown in

Figure 6. The elements are oriented along g'.' *, where j?',' is arbitrary except that it is contained in the plane of the array. Further, we denote the infinitesimal element length by dZ',', the current by I',' and the reference point of the reference element by El.'. This array with interelement spacings D, and D, is exposed to an incident plane wave with direction of propagation

Denoting the element current in column q and row rn byI$, it follows from Floquet's Theorem that the element currents are given by

i.e., they all have the same amplitude and a phase matching that of the incident plane wave with direction of propagation .? .

It has been shown rigorously that the EM fields from an infinite array are given by a spectrum ?+ of inhomogeneous plane waves [5]:

s = i s , + j s y + 2sz . ( 1 )

(2) I1 . I = I1.1 e-j13qDr&e-jPmDzs, 4.m 0.0

Figure 6. A plane wave with direction of propagation, s , incident upon an infinite array of single segment elements with orientation, I; 'J ; length, db ' ; current, Z',', and reference point, R"'. A plane wave will be scattered in the forward direction, i, (0,O) = s , as well as the

specular direction, k'. (0,O). Note: The total field in the forward direction is the sum of the incident and the scattered fields. Further, there will be an infinite sum of evanescent (exponentially decreasing) waves. They make up the near field associated with the array.

-

The spectrum ?+ denotes the directions of the inhomogeneous plane waves emanating from the array. They are found to be [5]:

?+ = i r x + j r , + 2rz

* In the following the first superscript refers to the array number, the second to the element section.

182

The fields expressed by (3) and (4) depend on ry as follows: For the principal direction k,n = 0,O we see from (6) that ry is always real since (s,(,(sz( 5 1, see (1). This corresponds to a plane wave, ?+ (0,O) , transmitted in the forward direction s and another reflected in the specular direction ?- ( 0 , O ) = i s , - j s , + i s z as illustrated in Figure 6. For Ikl, lnl > O,O, ry may still be real provided the interelement spacings D, and D, are large enough. These directions are denoted grating lobe directions.

However, for higher values of k,n, ry will always be imaginary, i.e. the exponent in the plane

will be real depicting evanescent waves that goes to zero as the point of waveS e - ~ ~ ( ~ - ~ l , l ) .ii

observation R moves away from the array as illustrated in Figure 6.

2.2. Single Array of Elements with an Arbitrary Number of Segments Extension from two to an arbitrary number of element segments is simply done by

induction. Again we conclude that only right handed waves will be emanated from a single array regardless of the shape of the elements.

2.3. Two Arrays of Elements with an Arbitrary Number of Segments So far we have considered only a single array with an arbitrary number of element segments.

We found that the field emanating from such an infinite array consisted of a spectrum ?+ of inhomogeneous plane waves, as given by ( 5 ) and (6):

1. A propagating wave in the forward and specular direction corresponding to k,n = 0,O (also called the principal directions).

2. A finite number of grating lobes if the interelement spacings D, and D, are large enough, corresponding to a finite number of k,n # 0,O.

3. An infinite number of evanescent waves that go to zero as we move away from the array.

As shown earlier, all of these waves are right-handed. We now place another array a certain distance dl to the right of the first array as illustrated in Figure 7. The interelement spacings D, and D, are the same as for array 1 but the number of element segments is arbitrary. Thus, the spectrum Pt is the same for the two arrays.

t_ d,-

A m y 1 RRBY 2

Figure 7. A plane wave with direction of propagation, s^, incident upon two arrays with some interelement spacings Ox and Dz. Each array emanates plane propagating waves in the forward, as well as the specular direction; they are all right-handed and so are their sums, regardless of region. Further there will be an infinite sum of evanescent (exponentially decreasing) waves that represent the near field associated with both arrays. Note: The arrays are located in a medium without dispersion.

We now calculate the currents in all the element segments. Just as the coupling between the segments in one array can be significant as noted above, it will also be significant between the segments in the two arrays. We emphasize that this coupling is always taken rigorously into account both in the theory treated in [5,15] as well as in the PMM program [16,17]. Once we find all the segment currents in both arrays in each other's presence, the determination of the fields emanating from each array is done precisely as was done for the single array case treated earlier and as shown in Figure 7. We define three regions, namely: Region 1 is the semi-infinite space to the left of array 1. Region 2 is the space between array 1 and array 2. Region 3 is the semi-infinite space to the right of array 2.

In Region 1 we observe left going

183

propagating waves radiating from the two arrays. Similarly we have right going waves in Region 3, while we have both left and right going waves in Region 2, as shown. All of these waves are nght- handed. The total field is simply obtained by superposition of the fields from the two arrays. There can be no doubt that in Regions 1 and 3 we will obtain a total right-handed field. Further, in Region 2 we simply obtain a total field of two crossing right-handed waves. Neither one of these waves can ever turn into left-handed waves since that would require the presence of Veselago magic material that everyone agrees does not exist in nature. Remember our support media is assumed to have no dispersion (i.e. linear).

3. On Increasing “Evanescent” Waves: A Fatal Misconception The total evanescent field in Figure 7 is obtained by superposition of the evanescent fields

from each array. However, these will, in general not be in phase and thus the total field cannot be obtained by simple addition of the magnitudes from the individual arrays. In fact, they could be out of phase and actually produce a null somewhere between the two arrays. Whatever the case, it is obvious from inspection of Figure 7 that the total field can only increase when the point of observation moves close to the elements and not all of a sudden because we are in a “Veselago medium”. We are still in a medium without dispersion and straight forward rules prevail. This writer is not aware of any demonstration of increasing evanescent waves except on capacitively loaded transmission line models terminated in a resistive load.

4. Conclusion: Synthesizing Veselago’s medium by a periodic structure is not feasible We have presented what is widely believed about Veselago’s medium. We emphasize that

the negative refraction was conceived by Veselago but some of the other features were suggested by others. Although such materials have never been found in nature, it was suggested by Pendry that such materials could be synthesized by periodic structures with special elements. However, we found some troubling deviation between Veselago’s theoretical material and what can de facto be obtained by using periodic structures. Regardless of element shape the most prevalent was:

1. Negative index of refraction is observed between Veselago’s medium and a medium with E,

,u > 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?

n,* = -1 The concept for Veselago’s flat “lens” is by now well known as shown in Figure 8. It consists of a flat slab where EZ, p2 not only is negative but where also ~2 = -&I and ,u~ = -pl (i.e., 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 with respect to ray , y ~ ~ by section A 1 ~ . Further, Veselago’s medium corresponding to B-43 (see also

section BA3 is inside the metamaterials where the Figure 2). However, if the two rays are to arrive 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].

&+I

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 &and p,” 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.

[ 111 B.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.

[ 141 B.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.

[ 181 G.V. Eleftheriades and K.G. Balmain, Negative-Refraction Metamaterials, Wiley Interscience, 2005.

[ 191 C. 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:

[23] W. Rotman, “Plasma simulation by artificial dielectric and parallel-plate media,” IRE Always positive and very inhomogeneous,” Phys. Rev. Letters, vol. 88, no. 18, May 2002.

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.

2 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]

z = h {[(l+sll)2-s*12] / [(l -sl,)2-s212]}x

e i n ~ s d = ~ & i(l-X2jA (2)

p = n z ; E = ~ Z (3)

where X=(l- S1l2+ S212) / 2S21 . Hence, we can get the permittivity and permeability of the metamaterial from:

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 (e l , e2, e3): . ~ - - ~ o d i u g [ s ~ ~2 ~ 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- k2pl lp 2)’ I ( k: - k?>”

z, = p ~ ( k:- k:)” I (b2&3,u- k:pllp 2)‘

nb = (k:&@- k:p 2/p1)’ 1 ( k: - k:f’

z b = p 2( k: - k:)” I (k:&3p 2- k:p 2Ip1)”

(4)

( 5 )

(6)

(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 ~ = p o diagb l X p l y p lZ1 and p 2=p o diag[p zX 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.

5 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.

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 14-0 1 - 1-071 3.

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),

[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.

2702-2712,2002.

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, 01 6608, 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, 123 11 4, 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 11 513GHz. 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

-4 - -6 m .Z -8 E

0 .; -10

4 -12

cs -14 -1 6 -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) H- field 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.

1 94

Reference [ 11 A. N. Lagarkov and A. K. Sarychev, Phys. Rev. B, 53,63 18 (1 996). [2] V. A. Podolsky, A. K. Sarychev, and V. M. Shalaev, J. Nonlinear Opt. Phys. Mat., 11,65

[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.

(2002).

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 1 1 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 1 oomm

b) Top View Inner Conductor

00 - a) Cross section

Connector c) Front View

I 0

.I 0

' 4 0

Fig. 1: TEM cell measurement set-up

Comparison: Prediction and Measurement (Copper Helix)

, -S11.M Measured x s11.p Predicted in TEM cell

-d- SI~.P-A,~ Predicted in Air

i l

I I 0.0 0.2 0.4 0.6 0.8 1.0 I

L- ~ ~ ~~ I

~~

Frequency, GHz

I

00 3 0 g

l%i?l d) Boundary Condition

-9

-1 2

-1 5

0 0 0.2 0.4 0.6 0.8 1 0

~ _ _ _ Frequency, GHz ~~

I

1S21-M 1 Measured i$ TEMcell , sz1-p 1 Predicted irj TEM cell

Fig. 2: Comparison of measured and predicted results.

198

0

4

ln

I 1 2

I

1 -16

- 15mm-ref

1 0.10 0.11 0.12 0.13 0.14 0.15 0.16' I Freouencv. GHz I

0

~ -5

-1 0

r -15 ln

-20

-25

0.10 0.12 0.14 0.16 0.18 0.20 Frequency, GHr

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

Effect of Conductivity

0 0.2 0.4 0.6 0.8 Frequency, GHr

0

1

-2

m

.-- -3 ln

4

-5

-8

-C"

I - R-100

0 0.2 0.4 0.6 0.8 1 Frequency, GHz

~~

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

, 0 I

-5

m-lo P

r r

ln45

-20

1 -25 1 0.0 0 2 0 4 0 6 0.8 l . o l

Frequency, GHz ~

0

1

2

;;a ln

4

5

4 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, 3rd Edition, 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

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. la . Within a lattice period the distribution of R(x,y) is defined as piecewise-constant function on a grid with small square cells, Fig. lb . 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 exp(- ikR,)

Gs = exp(- il,k,T,)exp(- ilykyTy) i,, i y = 4 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

equipment 4

Fig. 2

204

Both controllable and non-controllable screens may be used to shield an antenna. With a non- controllable 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, on- board 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.

FSS cell T, dB

' p - i - n diode

C=0.2 pF 1 -vertical plane 2 - horiiontal plane T. dB

0

-0.5

-1

-1.5

-2 /

0 -0.1 -2.5 \ I I -0.2

-0.4 0.85 0.9 0.95 1 1.05 1.1 1.15 -0.3 -3

F I F, -0 5 1 - C=O pF

0 10 20 30 40 50 60 70 V, 8, deg. b)

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

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, G0 .25 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

-1 0

-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 mounting Single-sided mounl.

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. Rs is 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

F 3

\ 'in diode

Ri

12mm

Figure 1 : FEM model of smart FSS

Figure 2: Equivalent circuit model of diode

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) 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, . , . I . I . I , I , I ,

2 3 4 5 6 7 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, 14-2 1,

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),

[4] L. Liu, S. M. Matitsine, Y. B. Gan and K. N. Rozanov, J. of Appl. Phys. 98,063512 (2005) 46(2004).

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

where <can be either E or p, A is the static susceptibility, f i is resonant frequency. (3)

<(f)= 1 +* 9 (1)

0) = + ,-(f/f, );+i. ir/ f , ’

Lorentz-permittivity and/or permeability obey Lorentz’s law of dispersion (2)

where il is the damping coefficient. (4) Kittel-permittivity and/or permeability obey Kittel’s law of dispersion

(5) (6)

Real material-permittivity and permeability are obtained from experiments. Mittra’s empirical law [2][3]-permittivity and/or permeability obey the following empirical law

(4) C,(fn) . C ( f n ) <(f)= f”- / , p 2

where Gcfo) and Gcfo) are the real and imaginary part of <at frequencyfo. Reported values of Gcfo), &(fo), a and p are reproduced in Table 1 .

21 2

21 3

E,( 1 GHz) a 1 GHz) 5 0.861 8 8 0.778 10 10 0.778 6

P p r ( 1 GHz) a pi(] GHz) P 0.569 5 0.974 10 0.96 1 0.682 3 1 15 0.957 0.861 7 1 12 1

(7) Mittra's relaxation law [2][3]-material permeability obeys the following empirical law

Reported values of A and f i are reproduced in Table 2. <(f 1 = i& 9 (5)

f i ( GHz) A f,(GHz) A f,(GHz) A f i ( GHz) 35 0.8 30 1 .o 20 1.5 30 35 0.5 18 0.5 30 2.5 25

A 2.0 3.5

2.3 Reflection and Transmission Coefficients

assumed and suppressed) The multilayered medium is assumed to be illuminated by plane wave (time factor elax is

EJnC(,) = EFe-lk'"' 1 , (6) where EF = Eri +Er; is the amplitude of the incident field, Eri ( / / p ) and EY: ( //$fnc) are the parallel and perpendicular components with respect to the plane of incidence. k'"" = k y - i k y is

the propagation vector. k y = -k, p , k y = kh C O S L ~ ' ~ ~ . k r = khsin8" , 6'" = C O S ~ Y ~ + s i n r j . kh is the wave number in the host medium, k,' = w 2 p h ~ h , w is the angular frequency, ph and &h are the permeability and permittivity of the host medium, Bfnc and are the elevation and azimuthal incident angles, respectively.

multilayered medium is available. The reflection coefficient is given by

l"C - m c

Analytical recursive solution [8] for reflection and transmission coefficients of a general

R, = Ri (7)

where the subscriptp stands for polarization which is either // or I,

2.4 Absorbing Band The absorbing band of a multilayered medium is defined as

214

where

B, (8'"") =

2.5 Objective Function for Design of Planar Microwave Absorber In general, there are two approaches. Some researchers consider the planar microwave

absorber as an electromagnetic filter with a desired frequency response. It is therefore straightforward to define the objective function. The other approach adheres to the essence of the problem, i.e., maximizing the absorbing bandwidth for reflectivity below a given threshold. Mathematically, this is defined as

f(x) = max B where (9)

x=[xo XI ... 'N 'N+I]

[xi GI x, = [x: x; t , ] l i n i N

[xk+,

I _ x n real material

x:+~] non - PEC ground only

non - dispersive

[A:'" fit"] Debye

[A:'" f;'"' 4 ~ 1 Lorentz or Kittel

1 x y =

Although the definition of objective function is conceptually very clear, computational implementation is much more challenging and expensive. However, the latter approach will be applied in our study since it is more realistic.

3. Preliminary Numerical Results A two-layer microwave absorber has been designed using real material. Its performance is

shown in Fig. 2. For fabrication purpose, the optimal thicknesses obtained through optimization are not practical. Rounding-off to the nearest realizable values is necessary. Accordingly, the effect of fabrication tolerance should be studied as part of the design problem. The performance of the two- layer microwave absorber with the optimal thicknesses rounded-off is shown in Figs. 3 and 4. The performance remains almost unchanged.

4. Conclusions Design of wideband planar microwave absorber is presented in this paper. A more realistic

mathematical definition of the problem is developed. Differential evolution is applied to solve the problem. A two-layer microwave absorber has been designed. The effect of fabrication tolerance has been numerically studied.

Acknowledgement The author wishes to thank Z. W. Li and G. Q. Lin for providing real material database. Appreciation also goes to Y . B. Gan for inspiring this author to consider material dispersion.

References [I] E. Michielssen, S. Ranjithan, and R. Mittra, Optimal multilayer filter design using real coded

genetic algorithms, IEE Proc. J-Optoelectronics , 139 [6 ] , 413-420 (1992).

21 5

[2] E. Michielssen, J. M. Sajer, S. Ranjithan, and R. Mittra, Design of lightweight, broad-band microwave absorbers using genetic algorithms, IEEE Trans. Microw. Theory Tech., 41 [6],

[3] D. S. Weile, E. Michielssen, and D. E. Goldberg, Genetic algorithm design of Pareto optimal broadband microwave absorbers, IEEE Trans. Electromagn Compability, 38 [3], 5 18-525 (1 996).

[4] S. M. Cui and D. S. Weile, Application of a parallel particle swarm optimization scheme to the design of electromagnetic absorbers, IEEE Trans. Antennas Propagation, 53 [ll], 36 16-3624, (2005).

[ 5 ] S. K. Goudos and J. N. Sahalos, Microwave absorber optimal design using multi-objective particle swarm optimization, Microwave Optical Technology Letters, 48 [8], 1553-1 558 (2006).

[6] K. V. Price, R. M. Storn, and J. A. Lampinen, Differential Evolution: a Practical Approach to Global Optimization, Berlin: Springer, 2005.

[7] A. Qing, Electromagnetic inverse scattering of multiple two-dimensional perfectly conducting objects by the differential evolution strategy, IEEE Trans. Antennas Propagat., 51[6], 125 1- 1262, (2003).

1024-103 1 (1993).

[8] J. A. Kong, Electromagnetic Wave Theory, Cambridge, MA: EMW Publishing, 1999

Figure 1 General Multilayered Medium

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

t,=.7mm,1,=1.7mm (4.6->16)

~-t,=.6mm.l,=1.7mm (4.4-15.4)

1 1,=.7mm.t,= 1.6mm (4.9-s16)

6 ,=.6mm,l,= 1.6mrn (4.7->16)

8 - layer 1: real material 45: ZC0200 I layer 2: real material 70: ZCoTilO ,.$ -10 -

-12

-14

-16 2 4 6 8 10 12 14 16

frequency ( G H r )

Figure 3 Effect of Fabrication Tolerance (Absolute Bandwidth)

'I --absolute -1OdB band: 4.5 - >16.0 :I\ layer 1: ZC0200. .742391 mrn

laver2: ZCoTilO. 1.663238 I!" -6],, \- 1 relative -1OdB band 3.2 - 13.6 I

layer 1 8 2: ZCoZTil. 2.276060 + ,557363 mm I

-16 . , . , . , . , . , . , . , 2 1 6 8 10 12 11 15

frequency (GHz)

Figure 2 A Two-layer Planar Microwave Absorber

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)

- _ _

-15 . , . , . , . , . , . , , , 2 1 6 8 10 12 I4 i s

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 Manna and 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.

21 6

21 7

.-

-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 [ 1 11. 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

21 9

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.,

[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). [ 1 11 W. J. Padilla, A. J. Taylor, C. Highstrete, Mark Lee and R. D. Averitt, Phys. Rev. Lett., 96, 107401

[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).

47,2075 (1999).

(2006).

22 1

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.

5-

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.

20 40 60 80 100 t20

Frequency (THz)

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

Frequency (THz)

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.

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

Huang, Z. 218 Hussain, T. 161

HSU, R.-T. 149

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. Lisyansky, A.A. 127 Liu, L. 83, 195,209 Lopatin, A.V. 97

71, 83.93, 101, 105, 113

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. 173 Tyagi, G.S. 153

Unnikrishnan, N.V. 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. 75 Xu,X. 71

Yang, H.L. 38 Ya0.X. 141

Zhang, D.H. 218 Zhang, F.J. 38 Zhang, J.J. 186 Zhang,Y. 195 Zhou,P.H. 89

electromagnetic materials

Documents