Polymer-Ceramic Composites for Conformal Multilayer Antenna and RF Systems
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University
By
Yijun Zhou, B.S., M.S.
* * * * *
The Ohio State University
2009
Dissertation Committee: Approved by Professor John L. Volakis, Adviser Adjunct Professor Chi-Chih Chen, Co-adviser Adviser Professor Roberto G. Rojas Professor Derek Hansford Co-adviser Graduate Program in Electrical and Computer Engineering
c Copyright by ° Yijun Zhou
2009 ABSTRACT
Novel engineered materials have drawn considerable interest as they provide new
opportunities for future antenna miniaturization and RF packaging. This dissertation
is dedicated to the development and application of the novel polymer-ceramic com-
posites for future compact multilayer antennas and RF systems. The first half of the
dissertation focuses on polymer-ceramic composites to achieve a new class of material
with superior mechanical (light-weight, flexible and load-bearing) and electric (high
permittivity, low loss) properties. Carbon nanotube (CNT) sheets are introduced for
the first time to overcome the issues of reliable printing on polymers. Compared with
single CNTs, the proposed CNT sheets achieve high conductivity for antennas and
RF applications. Concurrently they exhibit strong adhesion to the polymer surface,
making them attractive as a smart skin for future small unmanned areal vehicles
(UAVs) and body-worn applications.
The second half of the dissertation focuses on the application of the polymer- ceramic composites with carbon nanotube sheet printing. Two practical applica- tion examples are elaborated, namely (1) a compact anti-jamming GPS array and
(2) a cylindrically conformal microstrip array. For these applications, the polymer- based dual-layer GPS antenna is used to demonstrate the suitability of polymer- ceramic composites for multilayer antenna configurations. The pre-cure liquid form
ii of the polymer-ceramic composites and low temperature processing are highly de- sirable properties for three-dimensional (3-D) fabrication and multilayer packaging.
Also, the cylindrically conformal microstrip array demonstrates the application of the polymer-ceramic composites for flexible electronics due to their high flexibility, controllable permittivity and low loss. Overall, this dissertation demonstrates for the
first time that polymer-CNT materials are well suited for light-weight, conformal, multi-functional antenna and RF systems.
iii To my wife, Lili, and my parents
iv ACKNOWLEDGMENTS
I would like to acknowledge all of my committee members, especially my advisers,
Prof. Volakis and Dr. Chen for their guidance and advice. I would like to thank
Prof. Hansford and Elif Apaydin at the Ohio State University, Biomedical Engineer- ing Dept., for their cooperation in developing polymer composites and printing on polymers. Also, I would like to thank Dr. Bayram, Dr. Koulouridis and Dr. Kiziltas for their advice on polymer composites and carbon nanotube applications. Special thanks are due to Prof. Dai and Feng Du from the University of Dayton and Prof. Ko- tov from the University of Michigan for their support on carbon nanotube antennas.
Further, I would also like to thank all the other students in the Volakis’ group for their insightful discussions, challenging questions and help on antenna measurement.
Lastly, I want to thank all the staff and fellow students of the ElectroScience Labo- ratory for their support and help.
v VITA
January 5, 1982 ...... Born - Shanghai, China
July, 2004 ...... B.S. in Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China September, 2004-present ...... Graduate Research Associate, ElectroScience Laboratory, The Ohio State University, Columbus, Ohio December, 2006 ...... M.S. in Electrical and Computer Eng., The Ohio State University, Columbus, Ohio
PUBLICATIONS
Journal Publications
1. Y. Zhou, S. Koulouridis, G. Kiziltas, and J. L. Volakis, “A Novel 1.5-inch Quadru- ple Antenna for Tri-band GPS Applications,” IEEE Antennas and Wireless Propa- gation Letters, no. 5, pp. 224–227, 2006.
2. S. Koulouridis, G. Kiziltas, Y. Zhou, D. Hansford, and J. L. Volakis, “Polymer- Ceramic Composites for Microwave Applications: Fabrication and Performance As- sessment,” IEEE Transaction on Microwave Theory and Techniques, vol. 54, no. 12, pp. 4202–4208, Dec. 2006.
3. Y. Zhou, C.-C. Chen, and J. L. Volakis, “Dual Band Proximity-fed Stacked Patch Antenna for Tri-band GPS Applications,” IEEE Transactions on Antennas and Propagation, vol. 55, no. 1, pp. 220–223, Jan. 2007.
vi 4. Y. Zhou, C.-C. Chen, and J. L. Volakis, “Single-fed Circularly Polarized Antenna Element with Reduced Coupling for GPS Arrays,” IEEE Transactions on Antennas and Propagation, vol. 56, no. 5, pp. 1469–1472, May 2008.
5. Y. Zhou, Y. Bayram, and J. L. Volakis, “Polymer-carbon nanotube sheets for conformal load bearing antennas,” accepted by IEEE Transactions on Antennas and Propagation.
Conference Publications
1. Y. Zhou, S. Koulouridis, G. Kiziltas, and J. L. Volakis, “A Miniature Four-Arm Antenna for Tri-band GPS Applications,” IEEE APS/URSI International Sympo- sium, vol.3A, pp. 872–875, Washington, DC, July 2005.
2. Y. Zhou, C.-C. Chen, and J. L. Volakis, “Investigation of Several Miniature Antenna Design for Tri-Band GPS Applications,” Antenna Measurement Techniques Association (AMTA) Symposium, Newport, RI, November 2005.
3. Y. Zhou, C.-C. Chen, and J. L. Volakis, “Proximity-Coupled Stacked Patch Antenna for Tri-band GPS Applications,” IEEE APS/URSI International Sympo- sium, pp. 2683–2686, Albuquerque, NM, July 2006.
4. Y. Zhou, C.-C. Chen, and J. L. Volakis, “Tri-band Miniature GPS Array with a Single-fed CP Antenna Element,” IEEE APS/URSI International Symposium, pp. 3049–3052, Honolulu, HI, June 2007.
5. Y. Zhou, C.-C. Chen, and J. L. Volakis, “A Single-fed Element Antenna for Tri- band Anti-jamming GPS Arrays,” IEEE APS/URSI International Symposium, San Diego, CA, July 2008.
6. Y. Zhou, E. Apaydin, S. Koulouridis, Y. Bayram, D. Hansford, and J. L. Volakis, “High Conductivity Printing on Polymer-Ceramic Composites”, IEEE APS/URSI International Symposium, San Diego, CA, July 2008.
7. E. Apaydin, Y. Zhou, D. Hansford, S. Koulouridis, and J. L. Volakis, “Patterned metal printing on pliable composites for RF design,” IEEE APS/URSI International Symposium, San Diego, CA, July 2008.
8. Y. Bayram, Y. Zhou, J. L. Volakis, B.-S. Shim, and N. A. Kotov, “Textile Conductors and Polymer-Ceramic Composites for Load Bearing Antennas,” IEEE APS/URSI International Symposium, San Diego, CA, July 2008.
vii 9. Y. Zhou, C.-C. Chen, and J. L. Volakis, “A Compact 6-Element Tri-band GPS Array,” Antenna Measurement Techniques Association (AMTA) Symposium, Boston, MA, November 2008.
10. Y. Zhou, Y. Bayram, L. Dai, and J. L. Volakis, “Conductive polymer-carbon nanotube sheets for conformal load bearing antennas,” URSI Radio Science Meeting, Boulder, CO, January 2009.
11. Y. Zhou, C.-C. Chen, and J. L. Volakis, “A Miniature Dual-band GPS Antenna with Slot Loading,” Applied Computational Electromagnetics Society (ACES) Con- ference, Monterey, CA, March 8-12, 2009.
12. E. Apaydin, Y. Zhou, S. Koulouridis, J. L. Volakis, and D. Hansford, “Multilayer Printing on PDMS-Ceramic Composites for RF Integration and Packaging,” IEEE APS/URSI International Symposium, Charleston, SC, June 2009.
13. Y. Zhou, C-C. Chen, and J. L. Volakis, “A Compact 4-Element Dual-Band GPS Array,” IEEE APS/URSI International Symposium, Charleston, SC, June 2009.
14. Y. Zhou, Y. Bayram, J. L. Volakis, and L. Dai, “Conformal Load-Bearing Polymer-Carbon Nanotube Antennas and RF Front-Ends,” IEEE APS/URSI Inter- national Symposium, Charleston, SC, June 2009.
FIELDS OF STUDY
Major Field: Electrical Engineering
Studies in: Electromagnetics Communications and Signal Processing
viii TABLE OF CONTENTS
Page
Abstract...... ii
Dedication...... iv
Acknowledgments...... v
Vita ...... vi
ListofTables...... xii
ListofFigures ...... xiii
Chapters:
1. Introduction...... 1
1.1 Motivation, Challenges and Objective ...... 1 1.2 Organization of the Dissertation ...... 5
2. Polymer-Ceramic Composites for Antennas and Microwave Applications 7
2.1 Introduction ...... 7 2.2 FabricationProcess...... 9 2.3 Dielectric Properties of Polymer-Ceramic Composites ...... 14 2.3.1 MeasurementMethods...... 14 2.3.2 Dielectric Properties of PDMS-Ceramic Composites . . . . . 18 2.3.3 Dielectric Properties of RTV6166-Ceramic Composites . . . 22 2.4 ApplicationExamples ...... 30 2.4.1 Flexible Microstrip Circuits ...... 30 2.4.2 Polymer-based Antennas ...... 33 2.4.3 Two-tone Textured Dielectric Substrate ...... 34
ix 2.4.4 Three-layer Dielectric Rod Antenna ...... 36 2.5 Summary ...... 38
3. Polymer-Carbon Nanotube Sheet for Conformal Antennas and RF Circuits 39
3.1 Introduction ...... 39 3.2 SingleCarbonNanotubeProperty ...... 40 3.2.1 MechanicalProperty ...... 41 3.2.2 ElectricalProperty...... 43 3.3 E-textileCarbonNanotubeSheet ...... 45 3.3.1 FabricationProcess ...... 45 3.3.2 E-textilePatchAntenna ...... 46 3.4 Vertically-aligned Carbon Nanotube Sheet ...... 50 3.4.1 FabricationProcess ...... 51 3.4.2 Polymer-CNT Patch Antenna ...... 54 3.4.3 Polymer-CNT Microstrip Line ...... 66 3.5 Approaches to Increase Carbon Nanotube Sheet Conductivity . . . 68 3.5.1 Vertically-aligned CNT Sheet ...... 68 3.5.2 Horizontally-aligned CNT Sheet ...... 72 3.6 Summary ...... 75
4. CompactPolymer-basedGPSArray ...... 77
4.1 Introduction ...... 77 4.2 Proximity-fed Stacked Patch Antenna with Quadrature Feedings . 79 4.2.1 DesignConcept...... 80 4.2.2 Design Parameters ...... 83 4.2.3 MeasurementResults ...... 84 4.2.4 MutualCouplinginGPSArray ...... 90 4.3 Proximity-fed Stacked Patch Antenna with Branch-line Hybrid Feeding 91 4.3.1 Circular Stacked Patch Antenna ...... 92 4.3.2 Circular Branch-line Hybrid ...... 92 4.3.3 MeasurementResults ...... 96 4.3.4 MutualCouplinginGPSArray ...... 98 4.4 Proximity-fed Stacked Patch Antenna with Quadrature-phase Splitter100 4.4.1 Quadrature-phaseSplitter ...... 103 4.4.2 Measurement Results ...... 104 4.4.3 Compact 6-element Anti-jamming GPS Array ...... 107 4.5 Polymer-based GPS Antenna with CNT Printing ...... 112 4.5.1 Fabrication Process ...... 114 4.5.2 Measurement Results ...... 115 4.6 Summary ...... 118
x 5. Cylindrically Conformal Polymer-based Microstrip Array ...... 120
5.1 Introduction ...... 120 5.2 Transmitting and Receiving Antenna Designs ...... 122 5.3 Omni-directionalArrayDesign ...... 128 5.4 Prototype Cylindrically Conformal Array ...... 133 5.4.1 Tuning by Near-field Probing ...... 133 5.4.2 Measured Performance ...... 136 5.5 Polymer-based Circuit for Cylindrically Conformal Array ...... 136 5.6 Summary ...... 141
6. Conclusion...... 142
6.1 Summary ...... 142 6.2 FutureWork ...... 144 6.2.1 Density Control to Improve CNT Sheet Conductivity . . . . 145 6.2.2 Bundled Carbon Nanotube Threads for Flexible RF Circuits 146 6.2.3 Alternative Material for Light-weight, Flexible, and Load- bearingAntennas...... 147 6.2.4 Multilayer Polymer-based Antenna and RF Circuits . . . . . 148
Bibliography ...... 149
xi LIST OF TABLES
Table Page
2.1 Density and dielectric constant of polymers and ceramic powders . . . 12
3.1 Mechanical properties of carbon nanotubes ...... 43
4.1 Design specification of the compact GPS antenna system ...... 79
4.2 Average axial ratio at different elevation angles for the PFSP with hybridfeeding...... 99
4.3 Average axial ratio at different elevation angles for the PFSP with quadrature-phasesplitter...... 109
4.4 Average axial ratio at different elevation angles for the PFSP element in6-elementGPSarray...... 111
5.1 Design specification of the X-band transmitting antenna ...... 122
5.2 Design specification of the X-band receiving antenna ...... 122
5.3 Gain ripple level (dB) vs. different cylinder size and number of elements131
xii LIST OF FIGURES
Figure Page
2.1 Fabrication process for PDMS-ceramic composite ...... 11
2.2 Fabricate polymer-ceramic composites using Thinky@ vacuum mixer . 14
2.3 Reflection method for permittivity measurement ...... 16
2.4 Capacitance method for permittivity measurement ...... 18
2.5 Validation of reflection and capacitance methods by measuring PDMS- BaTiO3 composites...... 19
2.6 Validation of reflection and capacitance methods by measuring PDMS- BBNTcomposites...... 20
2.7 Dielectric properties of PDMS-BaTiO3 composites...... 23
2.8 Dielectric properties of PDMS-BBNT composites ...... 24
2.9 Dielectric properties of PDMS-MCT composites ...... 25
2.10 Dielectric properties of PDMS-D270 composites ...... 26
2.11 Dielectric properties vs. volume percentage of ceramic powders . . . . 27
2.12 Mixing rule for PDMS-D270 composites ...... 28
2.13 Dielectric properties of RTV6166-D270 composites ...... 29
2.14 Mixing rule for RTV6166-D270 composites ...... 30
xiii 2.15 Polymer-based flexible microstrip line ...... 31
2.16 Measured insertion loss and phase delay of the flexible microstrip line underbendingstrains...... 32
2.17 Polymer-based microstrip patch antenna ...... 33
2.18 Measured performances of the polymer-based microstrip patch antenna 35
2.19 Two-tone textured dielectric substrate ...... 36
2.20 Three-layer dielectric rod antenna ...... 37
3.1 Configuration of a single-walled carbon nanotube ...... 42
3.2 Diffusive and ballistic transport of electrons in one-dimensional wires 45
3.3 E-textileCNTsheet ...... 46
3.4 Fabrication process of E-textile CNT sheet ...... 47
3.5 Process of printing E-textile on polymer-ceramic composite...... 48
3.6 Polymer-based E-textile patch antenna ...... 49
3.7 Measured return loss of the E-textile patch antenna ...... 50
3.8 Measured gain of the E-textile patch antenna ...... 51
3.9 Vertically-aligned CNT sheet ...... 52
3.10 Process for growing vertically-aligned CNT sheet ...... 53
3.11 Process for transferring CNT sheet onto polymer-ceramic composites 55
3.12 Polymer-based patch antenna printed via vertically-aligned CNT sheet 57
3.13 Measured return loss of the vertically-aligned CNT patch ...... 58
3.14 Measured gain of the vertically-aligned CNT patch ...... 58
xiv 3.15 Measured E-plane pattern of the vertically-aligned CNT patch . . . . 59
3.16 Bending test for the vertically-aligned CNT sheet ...... 60
3.17 Stretching test for the vertically-aligned CNT sheet ...... 62
3.18 Cylindrically mounted polymer-CNT patch antennas ...... 63
3.19 Measured performance of polymer-CNT patch bent along E-plane . . 64
3.20 Measured performance of polymer-CNT patch bent along H-plane . . 65
3.21 Polymer-based microstrip line printed by vertically-aligned CNT sheet 67
3.22 Circuit model of the vertically-aligned CNT array ...... 69
3.23 Measure DC resistance of the vertically-aligned CNT array ...... 71
3.24 Squashed CNT array to increase CNT entanglement ...... 73
3.25 Circuit model of the horizontally-aligned CNT array ...... 74
4.1 Simulation model of the proximity-fed stacked patch antenna with quadraturefeedings...... 82
4.2 Simulated resonant modes of the PFSP GPS antenna with quadrature feedings ...... 82
4.3 Tuning the top and bottom patches size of the PFSP ...... 85
4.4 Tuning the proximity probe length of the PFSP ...... 86
4.5 Fabricated proximity-fed stacked patch antenna with quadrature feedings 87
4.6 Measured S-parameters of the proximity-fed stacked patch antenna withquadraturefeedings ...... 88
4.7 Measured RHCP gain of the proximity-fed stacked patch antenna with quadraturefeedings...... 88
xv 4.8 Simulated pattern of the proximity-fed stacked patch antenna with quadraturefeedings...... 89
4.9 Simulated electric fields for the 7-element square PFSP array at higher frequencymode ...... 90
4.10 Simulation model of the circular PFSP with single probe feeding . . . 93
4.11 Simulated return loss and realized gain of the circular PFSP with single probefeeding ...... 93
4.12 Circular branch-line hybrid for RHCP excitation ...... 95
4.13 Measured S-parameters of the circular branch-line hybrid...... 95
4.14 Simulated model and fabricated PFSP with branch-line hybrid feeding 97
4.15 Measured return loss of the PFSP with branch-line hybrid feeding . . 97
4.16 Measured gain of the PFSP with branch-line hybrid feeding ..... 98
4.17 Simulated gain and axial ratio pattern of the PFSP with branch-line hybridfeeding...... 99
4.18 Configuration of the 7-element PFPS array with integrated hybrid feed- ings ...... 100
4.19 Simulated pattern of the 7-element PFPS array with integrated hybrid feedings ...... 101
4.20 Modify the branch-line hybrid to reduce mutual coupling ...... 102
4.21 PFSP with quadrature-phase splitter for 6-element anti-jamming GPS array...... 102
4.22 Simulation model and parameters of the quadrature-phase splitter . . 104
4.23 Measured S-parameters of the quadrature-phase splitter ...... 105
4.24 Measured phase delay (between port 2 and port 3) of the quadrature- phasesplitter ...... 105
xvi 4.25 Configuration of the PFSP and 6-element GPS array with quadrature- phasesplitter ...... 106
4.26 Measured return loss of the PFSP with quadrature-phase splitter . . 107
4.27 Measured gain of the PFSP with quadrature-phase splitter ...... 108
4.28 Measured gain and axial ratio pattern of the PFSP with quadrature- phasesplitter ...... 108
4.29 Simulated RHCP and axial ratio patterns for the GPS element in array setting...... 110
4.30 New setup for measuring the GPS array on an infinitely large ground plane...... 111
4.31 Measured radiation patterns of the GPS element in the 6-element array setting...... 113
4.32 Simulation model of the polymer-based dual-layer GPS antenna . . . 115
4.33 Fabrication process of the polymer-based dual-layer GPS antenna . . 116
4.34 Measured performance of the polymer-based dual-layer GPS antenna 117
5.1 Cylindrically conformal microstrip array ...... 123
5.2 Simulation model of the series-fed 5-patch antenna ...... 125
5.3 Simulated return loss of the series-fed multi-patch antenna ...... 126
5.4 Simulated gain of the series-fed multi-patch antenna ...... 126
5.5 Simulated pattern of the series-fed multi-patch antenna ...... 127
5.6 Sidelobe suppression of the series-fed multi-patch antenna ...... 129
5.7 Simulation of a single patch on a PEC cylinder for roll pattern study 131
5.8 Pattern synthesis of a 20-element array on a cylinder (D=6”) . . . . . 132
xvii πD 5.9 Ripple level vs. normalized spacing distance ( Nλ ) ...... 134
5.10 Phasetuningbynear-fieldprobing ...... 135
5.11 Measured return loss and coupling of the prototype transmitting and receivingantennas ...... 137
5.12 Measured radiation pattern of the transmitting array ...... 138
5.13 Measured radiation pattern of the receiving antenna ...... 139
5.14 Flexible polymer-based antenna and feeding network circuit wrapped aroundthecylinder...... 140
6.1 Compress CNT array to improve the conductivity ...... 145
6.2 CNT threads weaved into high conductivity CNT fabric ...... 146
6.3 Nylon-strengthened copper threads and weaved wire mesh ...... 147
6.4 Polymer-based multilayer antennas and RF circuits ...... 148
xviii CHAPTER 1
INTRODUCTION
1.1 Motivation, Challenges and Objective
Future commercial and military communication systems require a new class of antennas and radio frequency (RF) front-ends that are small, light-weight, confor- mal, and multi-functional. Although the advance in semiconductor technology has been successfully reducing the size of transistors and integrated circuit (IC) chips, the antennas and microwave circuits still remain bulky and become one of the bot- tlenecks for such miniature multi-functional systems. For example, modern commer- cial electronic devices such as cellphones have placed new demands on small non- protruding antennas able to cover GSM bands (850 MHz, 900 MHz, 1800MHz, 1900
MHz) [1, 2, 3], GPS bands (1.227 GHz, 1.575 GHz) [4, 5, 6], Bluetooth, and Wi-Fi band (2.4 GHz) [7, 8, 9]. In manufacturing, transportation, and supply chain man- agement, radio-frequency identification (RFID) technology [10, 11] gained its popu- larity due to the advantage of multiple source scanning and contactless operation.
Meanwhile, RFID call for antenna tags that are small, light-weight, environment- friendly and low cost [12]. Further, future military communication systems such as joint tactical radio system (JTRS) [13] and future combat system (FCS) [14] require
1 continuous frequency coverage instead of multiple frequency band coverage. As a result, multi-functional antennas that are able to bear mechanical loadings and fully integrated with the structure are highly attractive for future military vehicles such as unmanned aerial vehicles (UAV) [15]. Also, light-weight, flexible, and robust antennas are necessary for future body-worn applications [16]. In this dissertation, we focus on developing polymer-ceramic composite material to address the above needs for conformal miniature antenna and multilayer RF systems.
There are several challenges in designing the conformal miniature multi-functional antennas. First and foremost, it is very challenging to achieve a miniature antenna
(< λ/10) with good bandwidth and gain performance [17]. Size reduction is usually achieved by loading antenna with high-contrast materials [18, 19, 20] or introducing artificial reactive loading such as coiling [21] or meandering [22]. However, material loading usually results in heavy weight since the high-contrast materials have large density. On the other hand, reactive loading such as coiling or meandering usually requires three-dimensional (3-D) fabrication of antenna configuration. Therefore, light-weight high-contrast material that allows for 3-D fabrication is the key to future antenna miniaturization. To that end, we propose polymer-ceramic composites [23] since they are light weight and can be mixed to have a dielectric constant up to
²r = 20. Further, they have a pre-cured liquid form, which easily allows for 3-D fabrication of coiled and meandered antennas.
Another challenge for conformal miniature antennas is their electrical performance such as radiation efficiency and gain. To achieve high gain and radiation efficiency, we must keep the material loss and conductor loss minimal. Since polymer-ceramic composites have very low dielectric loss tangent (tanδ < 0.02) for frequency up to
2 several GHz, they are suitable as dielectric substrates for microstrip antennas. Fur-
ther, high conductivity carbon nanotube (CNT) sheet [24, 25] has also been proposed
to realize flexible low-loss printing on the polymer substrates. The resulted polymer-
CNT antennas are very attractive for UAVs and body-worn applications due to their
good properties such as light weight, high flexibility, controllable permittivity, and
low loss.
Thirdly, future antenna specs will be demanding in terms of system conformalty
and structure compatibility. As the antennas conform to the platform surface, they
are bend and stretched. Obviously, the antennas must maintain functionality under
strain and stress conditions. For aircrafts and other military vehicles, the antennas
must also withstand mechanical forces such as pressures and vibrations. There-
fore, light-weight, flexible, and load-bearing materials are demanded to address these
mechanical requirements. Further, the antennas should also be insensitive to tem-
perature change. Such system requirements call for new class of material (such as
polymer-ceramic composites) with superior mechanical properties and stable thermal
behaviors.
Last but not least, we need to consider system integration or packaging of antennas
and RF front-ends. The desired system should be capable of fully integrating antennas
and RF front-ends to realize system-on-package (SoP) [26]. Several materials have
been proposed for SoP applications, including liquid crystal polymer (LCP) [27, 28,
29] and low temperature co-fired ceramics (LTCC) [30, 31]. Both LCP and LTCC
systems can achieve multilayer packaging by layer-by-layer printing. However, to
bond different layers they must be heated (300◦ for LCP and 1000◦ for LTCC) at temperatures that may cause failure to some IC components and fragile wire bonds.
3 Therefore, the material system (such as polymer-ceramic composites) that can be
processed at room temperature is highly favorable for purpose of system packaging.
Given the aforementioned challenges, this dissertation focuses on developing a new
class of composite material, namely polymer-ceramic composites, to realize miniature,
conformal, and multilayer antenna/RF systems. Previous study on polymers [32, 33]
and ceramics [34, 35] mainly dealt with either material separately. Here, for the first
time, we study polymer-ceramic composite properties and apply them to conformal
multilayer antenna systems. Further, we introduce a novel polymer printing technique
using conductive carbon nanotube sheets to realize a flexible, stretchable, and load-
bearing polymer-CNT antenna. Specifically our key contributions include:
1. Development of a new class of composite materials, namely polymer-ceramic composites. They are light weight, have high flexibility, and are of controllable per- mittivity and low loss. Demonstration of their application for antenna loading and
RF systems packaging is also included.
2. Investigation of conductive carbon nanotube (CNT) sheets to realize flexible, stretchable, and low-loss printing on polymer substrates. We include a study of different CNT models and parameter effects to increase CNT sheet conductivity.
3. Development of compact dual-layer GPS antennas/arrays that cover all three
GPS bands. In this context, we demonstrate a polymer-based dual-layer GPS antenna to overcome difficulties of antenna fabrication and integration.
4. Design of a cylindrically conformal microstrip array having an omni-directional azimuth pattern. Demonstration of the concept of polymer-based antenna and RF circuits serving as an example of flexible electronics.
4 1.2 Organization of the Dissertation
The remainder of this dissertation is organized as follows.
In Chapter 2, we present the development of polymer-ceramic composites. The
fabrication process and required equipment are discuss in details. We have also carried
out two types of measurements (reflection method and capacitance method) to char-
acterize the dielectric properties of polymer-ceramic composites. Several application
examples are proposed to show their versatility in antenna and RF systems.
Chapter 3 then proceeds to address the challenge of printing on polymer sub- strates. Carbon nanotube sheets are introduced to realize flexible conductive print- ing on polymers. Two configurations of CNT arrays are elaborated, namely E-textile
CNT sheet and vertically-aligned CNT sheet. Measurement results of the polymer-
CNT patch antennas are given, exhibiting good electrical and mechanical properties.
We also study the circuit models of both vertically and horizontally aligned CNT array and propose to use bundled CNT threads to further increase CNT sheet conductivity.
In Chapter 4, we discuss the development of a compact tri-band GPS array. Specif- ically, the dual-layer configuration with proximity probe feeding is proposed to provide two resonant modes that covers all three GPS bands (L1, L2 and L5). Another im- portant design feature is the single probe feeding for right-hand circular polarization
(RHCP) which greatly reduces the complexity of the feeding network and system load. To address the issues of antenna fabrication and system integration using high- contrast ceramic substrates, we also implement the dual-layer tri-band GPS antenna with polymer-ceramic composites. This polymer-based GPS antenna also demon- strates the application of polymer-ceramic composites for multilayer antenna and RF systems.
5 In Chapter 5, we then present another application example of polymer-ceramic composites, namely a cylindrically conformal microstrip array. The series-fed micro- strip array is designed to provide omni-directional azimuth pattern. Prototype array has been built and tuned to meet the design specification. Polymer-based flexible antenna and RF circuits have been proposed to overcome the fabrication difficulty, thus demonstrating the potential use in flexible electronics.
Finally, the dissertation concludes with a summary of major contributions and
discusses future research topics in polymer-ceramic composites and carbon nanotube
applications.
6 CHAPTER 2
POLYMER-CERAMIC COMPOSITES FOR ANTENNAS AND MICROWAVE APPLICATIONS
2.1 Introduction
Many mobile structures (for example, aircrafts, ships and automobiles) require conformal antennas for high data rate connectivity, which brings the need for material compatibility and for structurally reinforced antennas. This requirement can be par- ticularly challenging for small unmanned aerial vehicles (UAVs) since the antennas are part of the mechanical structures. Concurrent requirements for greater bandwidth and multifunctionality imply an even greater need for conformality. Existing confor- mal antennas are still printed on rigid laminate substrates with curved shapes, making them expensive and cumbersome to manufacture and, hence, not desirable for small and lightweight platforms. Further, an increasing demand for integration of antennas with radio-frequency (RF) front-end makes use of such high-contrast substrates very attractive since they allow for system level miniaturization and integration.
Polymers are rapidly becoming important among materials for microwave and electronic applications whether used in pure form or combined with ceramic powders.
For example, in optoelectronics, polymers have been used to produce mechanically
7 flexible “electronic paper” [36] and high-efficiency light-emitting diodes [37]. Among example RF applications already reported, we note the use of an electro-optic poly- mer in [38] to design photonic RF arrays. In [39], polymer composites were proposed as substrate materials for a scanning antenna. In [40, 41], polymerceramic mixtures were used for thin-film capacitors. More recently, liquid crystal polymers (LCPs) have been proposed for system-on-package (SoP) applications [27, 28, 29], which were traditionally done via low temperature co-fired ceramic (LTCC) technique and stere- olithography metalization [30]. LCPs can be used as printed circuit boards (PCBs) by common etching method and achieve three-dimensional (3-D) SoP configuration through layer-by-layer printing. However, commercial LCPs have very low dielec- tric constants (²r = 3), which does not satisfy the need for high contrast dielectric material. In short, of importance about the polymers are that a) they can be doped with other material to make them functional and control their dielectric properties; b) they are “soft” and pliable (unlike crystalline materials); c) thin polymer layers can be printed and then stacked to form packaged 3D electronics and d) they can be reinforced with carbon-based nanotubes or nanofibers to render them structurally compatible for system embedded antennas and smart skins.
In this chapter, we present the development of polymer-ceramic composites [23,
42, 43] to address the need of truly flexible and conformal RF materials. In sec- tion 2.2, the fabrication process of such polymer composites is revealed, which also shows its advantage of room temperature handling. In section 2.3, we discuss the approaches to characterize the dielectric properties of polymer-ceramic composites.
Two different polymer composites are tested, namely PDMS-ceramic composite and
RTV6166-ceramic composite. The measurement results show that polymer-ceramic
8 composites exhibit controllable permittivity between ² = 3 20 with low loss tan- r ∼ gent tanδ < 0.02, depending on the mixing ratio of ceramic powders to the polymer matrix. In section 2.4, we present some application examples of polymer-ceramic composites. As can be seen, polymer-ceramic composites are well suited for confor- mal multilayer antenna and RF systems. Detailed design concepts and procedures for such multilayer antennas and conformal RF circuits will be discussed in Chapter 4 and Chapter 5.
2.2 Fabrication Process
In this thesis, we mainly focus on two different types of polymers, namely poly- dimethylsiloxane (PDMS) from Dow Corning and RTC6166 form GE Plastic. Both materials are silicone-based organic polymers and are known for their favorable prop- erties: they are nonflammable, water- and chemical-resistant (hydrophobic), stable at high temperature (up to 200◦C) and low cost. More importantly, PDMS and
RTV6166 have very low dielectric loss for frequencies up to several GHz, which makes them desirable for microwave applications. To achieve higher dielectric constant, ce- ramic powders are introduced into the polymer matrix via particle dispersion process, which results in a polymer-ceramic mixture [23, 42, 43]. As a result, polymer-ceramic composites inherit most of their mechanical properties (flexible, stable, hydropho- bic and etc.) from the silicone part and electrical properties (high permittivity, low loss) form the ceramic part. In this thesis, various ceramic powders, namely bar- ium titanate (BaTiO3), Mg-Ca-Ti (MCT), strontium titanate (SrTiO3 or D270) from
Trans-Tech Inc., and Bi-Ba-Nd-Titanate (BBNT) from Ferro Corporation, have been tested to mix with the polymer matrix. Specifically, BaTiO3 demonstrates a wide
9 range of attainable dielectric permittivity (from a few tens to a few thousands) val- ues depending on its chemical form, grain size, environment temperature, and added dopants [44], and thus has been widely employed in capacitor technology [45, 46]. The
BBNT BBNT, MCT, D270 powders in this work have dielectric constant of ²r = 95,
MCT D270 ²r = 140, ²r = 270, respectively. Detailed fabrication process and property characterization of polymer composites are described below.
We first elaborate on the fabrication steps to achieve PDMS-ceramic composites.
The process starts with the preparation of PDMS (T2 Silastic from Dow Corning) by adding one part of cross-link agent to ten parts of silicone gel (mass ratio). The resulting silicone gel is mixed thoroughly and placed into a vacuum chamber for around 10 minutes where excessive gas is removed by venting the surface bubbles within the prepared gel. Next, the ceramic powder is added to the degassed silicone gel at certain percentage (volume ratio) and is again mixed thoroughly. The resulting
PDMS-ceramic slurry mixture is poured into a plastic container (of the desired shape).
Degassing of the resulting mixture is then done by placing the containers into a vented vacuum chamber as done for the pure silicone gel. This process is the most tedious step and plays a critical role to achieving homogeneous ceramic-reinforced PDMS substrates. An average degassing time for a dish (of average 6 mm thickness and
30 mm diameter filled with 20% ceramic powders) is approximately 3 hours. The resulting fully degassed mixture is then left for ambient drying and curing which lasts about 24 hours. The procedure of the above fabrication process is displayed in
Fig. 2.1 with a picture for each step.
We remark that the mixing ratio (10:1) for the PDMS preparation is mass-based ratio which means that, for example, 10g of silicone gel is mixed with 1g of cross-link
10 (a) Prepare PDMS silicone gel (b) Add cross-link agent (10:1)
(d) Add certain percentage of ceramic powder (c) Degas for 10 minutes
(e) Mixing (5 minutes) (f) Pour the mixture into a container
(h) Cure for 24 hours (g) Degas for 3 hours
Figure 2.1: Fabrication process for PDMS-ceramic composite
11 Important properties of the polymers and ceramic powders 3 Material name Density (ρ: g/cm ) Permittivity (²r) PDMS 1.15 3 - 3.5 RTV6166 0.98 2.8 - 3.5 BaTiO3 6 > 1000 BBNT 5.9 95 MCT 3.85 140 D270 5.12 270
Table 2.1: Density and dielectric constant of polymers and ceramic powders
agent. On the other hand, the mixing ratio of ceramic powder to PDMS is volume- based ratio since the composite property is determined by the property of the filling particles and their volume. In this work, the mixing percentage of the ceramic powder
(v%) is defined as: mceramic ρceramic v%= mceramic mPDMS 100% (2.1) ρceramic + ρPDMS × where mceramic, mPDMS and ρceramic, ρPDMS are the mass and density of the ceramic powder and PDMS, respectively. Table 2.1 lists the density and dielectric constant of the different polymers and ceramic powders. The properties of the polymers are provided by Dow Corning and GE Plastic, and the properties of the ceramic pow- ders are given by Trans Tech Inc. and Ferro Corporation. Given the parameters in
Table 2.1, we can mix PDMS-ceramic composites (such as PDMS-BBNT, PDMS-
MCT, or PDMS-D270) up to certain volume percentage by following the procedure displayed in Fig. 2.1.
We also mix the ceramic powders with RTV666 silicone gel from GE Plastic by the same process as displayed in Fig. 2.1. RTV6166 silicone dielectric gels are low viscos- ity liquid silicones, which cure to form very soft, gel-like elastomers. They are often
12 used to preserve dielectric integrity and provide outstanding protection to delicate
electronic circuit assemblies operating in harsh environments. We observe that the
RTV6166 has lower viscosity than PDMS and takes longer time to cure. As a result,
the RTV6166-ceramic composites are much easier to degas since the bubbles are more
likely to break even without using the desiccator as shown in Fig. 2.1(g). Another
difference between PDMS and RTV6166 composites is that RTV6166-ceramic com-
posites have lower dielectric loss than PDMS-ceramic composites as will be discussed
in the next section.
One important feature of the above fabrication process is that it is done at the
room temperature, which is one of the advantages of using polymer-ceramic compos-
ites for multilayer packaging compared with LCP and LTCC. Specifically, to bond
LCP and LTCC layers, they must be heated at temperatures (300◦C for LCPs and
1000◦C for LTCCs) that could cause failure to some IC components and fragile wire bonds. On the other hand, we can easily embed the IC components inside polymer- ceramic composite by pouring the liquid form composite over the circuit. We also remark that the above fabrication process by hand is tedious and not suited for mass production. To streamline the fabrication process, currently we are using the Thinky@ vacuum mixer as shown in Fig. 2.2. The material container inside the Thinky@ mixer
rotates and revolves at very high speed so that materials start circulation in the con-
tainer from top to bottom that results in a quick mixing. As mixing is done in high
pressure more than 400 G force, air bubbles in the material are pushed out at the same
time. In our process, we set the rotation per minute (RPM) to 3000 and mix for 1
minute to achieve an air-less homogeneous polymer-ceramic mixture. In the following
13 sections, we will discuss the dielectric properties of the fabricated polymer-ceramic composites and their application for antenna and RF systems.
Figure 2.2: Fabricate polymer-ceramic composites using Thinky@ vacuum mixer
2.3 Dielectric Properties of Polymer-Ceramic Composites
2.3.1 Measurement Methods
In this section, we first discuss the approaches to measure the dielectric properties of polymer-ceramic composites, namely, reflection method [47, 48] and capacitance method [49]. These two methods have wider measurement frequency band and are easier to carry out compared to other resonant methods such as circuit method [50] or resonator method [51]. The principles of each method are reviewed and the procedures are described. Measurement results via reflection method and capacitance method are then compared with each other to validate the accuracy of the measurements.
14 Reflection Method
Reflection method is usually carried out by measuring the reflection from an open-
ended coaxial probe against the dielectric sample, as shown in Fig. 2.3(a). The equiv-
alent circuit model for such coaxial probe is displayed in Fig. 2.3(b). The capacitance
C1 is mainly determined by the structure of the coaxial probe, and is independent
of the material under test. The dielectric sample under test can be modeled as a
capacitance ²rC2. The resistance R = 1/G in parallel counts for the radiation from the coaxial probe. Therefore, the normalized admittance can be calculated as:
Y = jωC1Z0 + jω²rC2Z0 + Z0G(ω, ²r) (2.2) Y0
where Z0 is the characteristic impedance of the coaxial cable, ω is the angular fre-
quency, and ²r is the complex permittivity of the material under test. For small iris
coaxial probe, the radiation conductance can be expressed as [52]:
5 2 G(ω, ²r)= ²r G(ω, ²0) (2.3)
Therefore, we can obtain the normalized admittance:
5 5 Y 2 2 = jωC1Z0 + jω²rC2Z0 + ²r G(ω, ²0)Z0 = K1 + K2²r + K3²r (2.4) Y0
The factors K1, K2 and K3 are usually complex number. To determine these three factors, we need to measure three media with known permittivity values as calibration.
Z−Z0 1−Y/Y0 1 2 From the measured reflection coefficients, Γ = Z+Z0 = 1+Y/Y0 , we can solve for K , K and K3. After the calibration is done, we are ready to obtain the complex permittivity
of the sample under test from the measured reflection coefficient.
In our work, we measured the dielectric permittivity of polymer-ceramic compos-
ites via an Agilent E8233 network analyzer with material characterization probes.
15 Coaxial probe (Z 0)
Dielectric sample ε− ε r'j r '' ε C1 rC2 G
(a) Coaxial probe against the (b) Equivalent circuit model material under measurement
Figure 2.3: Reflection method for permittivity measurement
OPEN, SHORT and a block of Teflon (²r = 2.1, tanδ = 0) were used for calibration.
Particularly for high frequency measurement, we need to change to small coaxial probe since we always want to keep the radiation from the coaxial probe minimum.
The measured dielectric constant and loss tangent of PDMS-BaTiO3 and PDMS-
BBNT composites (up to 20 GHz) are shown in Fig. 2.5 and Fig. 2.6, which agree with the measurement results via capacitance method. However, we also observe that the measurement curve via reflection method is not very stable. This is caused by the soft nature of polymer-ceramic composites. It is very difficulty to control how much pressure we exert when we press the probe against the soft surface of the polymer sample. This is also the major uncertainty in reflection measurement.
Capacitance Method
As shown in Fig. 2.3(b), the dielectric sample under test can be modeled as a capacitor. Therefore, we can measure the permittivity of the dielectric sample if we can somehow measure the capacitance directly. This can be done using an Agilent
16 E4991A impedance analyzer, which measures the impedance or admittance associ-
ated with the dielectric sample. As shown in Fig. 2.4(a), a dielectric sample under
test (DUT) is positioned between the test fixture’s electrodes to form a capacitor.
Fig. 2.4(b) shows its equivalent circuit model. The complex admittance is:
Cp G Y = G + jωCp = jω( j )C0 (2.5) C0 − ωC0 where C0 is capacitance of the air-filled capacitor and G is related to the dielectric
loss. For a parallel-plate capacitor, the admittance can be calculated as:
Y = jω²rC0 (2.6)
Therefore, we can calculate the effective relative permittivity by:
Cp G ²r = j (2.7) C0 − ωC0
²0S C0 = (2.8) d where S is the area of the electrode, d is the thickness of the dielectric sample and
²0 is the permittivity of air. To minimize the error due to edge capacitance and electrode radiation, we also carry out the calibration by measuring OPEN, SHORT and a standard Teflon sample (0.78 mm thick, ²r = 2.1, tanδ = 0).
The measurement results via capacitance method are shown in Fig. 2.5 and
Fig. 2.6. Since E4991A has a knob to adjust the pressure between the two elec- trodes, it has more stable control of the pressure against the dielectric sample than reflection method. Therefore, the measured curves are more stable than those of re-
flection method. The major error in our capacitance method is associated with the sample thickness (d) which is not measured accurately. Although E4991A only works up to 3 GHz, it is suitable for most of our applications in VHF (30 - 300 MHz) and
17 S DUT
d Cp G
(a) DUT between the test (b) Equivalent circuit model fixture’s electrodes
Figure 2.4: Capacitance method for permittivity measurement
UHF (300 - 3000 MHz) bands. In the rest of the dissertation, we mainly measure the dielectric properties of our polymer samples via Agilent E4991A impedance an- alyzer. The measured dielectric properties of different polymer-ceramic composites are discussed in the next section.
2.3.2 Dielectric Properties of PDMS-Ceramic Composites
Here we carried out the material property characterization by using capacitance method. The measured permittivity and loss tangent of PDMS-BaTiO3 composites are shown in Fig. 2.7(a). The permittivity remains almost constant during the mea- surement frequency spectrum from 100 MHz to 1 GHz, except for a slight linear drop as the frequency increases. The maximum permittivity value is ²r = 20 around for a
25% BaTiO3 volume mixture, and similar results were reported in [45, 46] for BaTiO3 mixtures with polymers. It can be seen from Fig. 2.7(b), the loss tangent increases as frequency increases. Specifically, the loss tangent for a 25% BaTiO3 mixture is tanδ = 0.04 at 1 GHz. The measured permittivity and loss tangent of PDMS-BBNT,
18 14
Reflection method 12 Capacitance method
10 ) r ε 8 ) r
6 ε 10 Permittivity(
4 5 Permittivity(
2 0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz) 0 0 5 10 15 20 Frequency (GHz)
(a) Dielectric constant of a 15% PDMS-BaTiO 3 composite
30
25
Reflection method 20 Capacitance method ) r ε
15
Permittivity( 10
5
0 0 5 10 15 20 25 Volume percentage (v%) (b) Dielectric constant vs. volume percentage of
BaTiO 3 powder (measured at 500 MHz)
Figure 2.5: Validation of reflection and capacitance methods by measuring PDMS- BaTiO3 composites
19 10
9 Reflection method 8 Capacitance method 7 ) r
ε 6
5 10 ) r ε 4 Permittivity( 5 3
2 Permittivity ( 0 1 0 0.2 0.4 0.6 0.8 1 Frequency (GHz) 0 0 0.5 1 1.5 2 2.5 3 Frequency (GHz) (a) Dielectric constant of a 15% PDMS-BBNT composite
20
18 Reflection method 16 Capacitance method 14 ) r
ε 12
10
8 Permittivity( 6
4
2
0 0 5 10 15 20 25 Volume percentage (v%) (b) Dielectric constant vs. volume percentage of BBNT powder (measured at 500 MHz)
Figure 2.6: Validation of reflection and capacitance methods by measuring PDMS- BBNT composites
20 PDMS-MCT, PDMS-D270 composites are shown in Fig. 2.8, Fig. 2.9 and Fig. 2.10,
respectively. As can be seen, they have lower dielectric constants than PDMS-BaTiO3
composite. On the other hand, the loss tangents of PDMS-BBNT, PDMS-MCT and
PDMS-D270 composites are smaller than that of PDMS-BaTiO3 composite. This
is desirable for antenna applications as lower dielectric loss usually results in higher
antenna efficiency and gain.
Fig. 2.11 shows the typical dielectric properties vs. different mixing percentage
(measured at 500 MHz via capacitance method) for the aforementioned polymer-
ceramic composites. As can be seen, PDMS-BaTiO3 can achieve highest permittivity.
However, it also has very large loss tangent, which makes it well suited as RF absorber
material. From Fig. 2.11, we also observe that MCT, BBNT and D270 powders have
even lower dielectric loss than PDMS, making them attractive for microwave and
antenna applications. Specifically, PDMS-D270 composite can achieve controllable
dielectric constant ² = 3 20 with loss tangent tanδ < 0.01 for frequency up to 1 r ∼ GHz.
We also examine the measured dielectric constant of the polymer mixture based on different mixing models which predict the composite properties. For example, according to Bruggeman’s formula [53] (Equation 2.9):
²2 ²m ²1 1 − ( ) 3 = 1 v2 (2.9) ²2 ²1 ² − − m where ²1, ²2, ²m denote the complex permittivity of the polymer matrix, the ceramic powder and the composite respectively and v2 is the volume percentage of the ceramic
powder. Other mixing models are based on Looyenga’s formula [53] (Equation 2.10)
and Bottcher’s formula [53] (Equation 2.11):
21 1/3 1/3 1/3 3 ² = [² + v2(² ² )] (2.10) m 1 2 − 1
3v2²m(²2 ²1) ²m = ²1 + − (2.11) 2²m + ²2 Given the dielectric constant of the D270 powder and PDMS, we can predict the composite dielectric constant at different mixing ratio based on the above mixing rules. The resulted curves are shown in Fig. 2.12 together with the measured dielectric constant of PDMS-D270 composites. As shown in Fig. 2.12, the measured data fit closest to Bruggeman’s formula. Therefore, Bruggeman’s formula can be used to predict the dielectric constant of a high volume ratio polymer-ceramic composite, which is not easy to fabricate. We have not predicted the loss tangent of the polymer composites using the above mixing rules since the loss tangent of the ceramic powders are unknown to us.
2.3.3 Dielectric Properties of RTV6166-Ceramic Composites
We also mix the D270 powders with RTV6166 silicone gel from GE Plastic. The fabrication process is similar to that of PDMS-ceramic composites as described in the previous section. We mix the part A and part B of RTV6166 together with 10% of cross-line agent. Ceramic powders are added subsequently to achieve RTV6166- ceramic composites. Particularly, compared to PDMS mixtures, RTV6166 mixtures have very low viscosity and can be poured into the molds easily. And RTV6166- ceramic composite are much easier to degas. Degassing can be done at the normal pressure with some vibration to break the air bubbles. We also notice that by adding more cross-line agent (around 20%) RTV6166 composites become very crispy and no longer rugged.
22 25
25% 20
20% ) r
ε 15
15% 10 Permittivity ( Permittivity 10% 5 5% 0% 0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz)
(a) Dielectric constant of PDMS-BaTiO 3 composites
0.05
0.045
0.04 25%
0.035 )
δ 20% 0.03 15% 0.025
0.02 10%
loss tangent (tan (tan losstangent 0.015 5% 0.01 0% 0.005
0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz)
(b) Loss tangent of PDMS-BaTiO 3 composites
Figure 2.7: Dielectric properties of PDMS-BaTiO3 composites
23 10
9 30% 8
7 25% ) r
ε 6
5 10% 4 Permittivity ( Permittivity 3 5% 2
1
0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz) (a) Dielectric constant of PDMS-BBNT composites
0.015 )
δ 0.01
5% 0.005 loss tangent (tan (tan losstangent 10% 25% 30%
0 0.2 0.4 0.6 0.8 1 Frequency (GHz) (b) Loss tangent of PDMS-BBNT composites
Figure 2.8: Dielectric properties of PDMS-BBNT composites
24 10
9 30% 8 25% 7
) 20% r
ε 6 5 10% 4
Permittivity ( Permittivity 5% 3
2
1
0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz) (a) Dielectric constant of PDMS-MCT composites
0.015
5% 10% 20% 25% )
δ 0.01 30%
0.005 loss tangent (tan (tan losstangent
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Frequency (GHz) (b) Loss tangent of PDMS-MCT composites
Figure 2.9: Dielectric properties of PDMS-MCT composites
25 14 30% 12 20% 10 ) r
ε 15% 8 10% 6
Permittivity ( Permittivity 5% 4 0% 2
0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz) (a) Dielectric constant of PDMS-D270 composites
0.015 )
δ 0.01
0% 0.005 loss tangent (tan (tan losstangent 5% 10% 15% 20% 30% 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Frequency (GHz) (b) Loss tangent of PDMS-D270 composites
Figure 2.10: Dielectric properties of PDMS-D270 composites
26 30
BaTiO 25 3 BBNT MCT 20 D270 ) r ε
15
Permittivity ( Permittivity 10
5
0 0 0.1 0.2 0.3 0.4 0.5 Volume percentage (v%) (a) Dielectric constant vs. volume percentage of ceramic powders (measured at 500 MHz)
0.05
0.045 BaTiO 3 0.04 BBNT MCT 0.035 ) D270 δ 0.03
0.025
0.02
loss tangent (tan (tan losstangent 0.015
0.01
0.005
0 0 0.1 0.2 0.3 0.4 0.5 Volume percentage (v%) (b) Loss tangent vs. volume percentage of ceramic powders (measured at 500 MHz)
Figure 2.11: Dielectric properties vs. volume percentage of ceramic powders
27 40 Measured data 35 Bruggeman formula Looyenga formula 30 Bottcher formula ) r
ε 25
20
Permittivity ( 15
10
5
0 0 0.1 0.2 0.3 0.4 0.5 0.6 volume percentage (v%)
Figure 2.12: Mixing rule for PDMS-D270 composites
Fig. 2.13 shows the measured dielectric properties of RTV6166-D270 composites.
Comparing Fig. 2.13 and Fig. 2.10, it can be seen that RTV6166-D270 composites have similar dielectric constant as PDMS-D270 composites (² = 3 20). However, r ∼ the loss tangent of RTV6166-D270 is tanδ < 0.005 for frequency up to 1 GHz, much
lower than that of PDMS-D270 composites. This low loss property is particular
important for some high frequency applications. Fig. 2.14 also shows the dielectric
constant (measured at 500 MHz) vs. volume percentage of D270 powders. We also
note that the measured results are close to the mixing model based on Bruggeman’s
formula. One disadvantage of RTV6166 is that it takes around 3 days to cure. In
this work, we mainly use PDMS as the polymer matrix for mixing since it is cheaper
and easier to handle than RTV6166.
28 15 30%
10 ) r
ε 20% 15% 10% Permittivity( 5 5% 0%
0 0 0.2 0.4 0.6 0.8 1 Frequency (GHz) (a) Dielectric constant of RTV6166-D270 composites
-3 x 10 5
4.5 0% 5% 4 10% 3.5 15% )
δ 20% 3 30%
2.5
2
losstangent (tan 1.5
1
0.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Frequency (GHz) (b) Loss tangent of RTV6166-D270 composites
Figure 2.13: Dielectric properties of RTV6166-D270 composites
29 40 Measured data 35 Bruggeman formula Looyenga formula 30 Bottcher formula ) r
ε 25
20
Permittivity ( 15
10
5
0 0 0.1 0.2 0.3 0.4 0.5 0.6 volume percentage (v%)
Figure 2.14: Mixing rule for RTV6166-D270 composites
2.4 Application Examples
Future military and commercial communication systems require a new class of
antennas and RF front-ends that are light-weight, conformal, and multi-functional.
Since polymer-ceramic composites have very desirable mechanical and electrical prop-
erties, they provide us with a novel class of materials for antenna loading and RF
systems integration. Below, we present four application examples of polymer-ceramic
composites in antenna and microwave circuit designs.
2.4.1 Flexible Microstrip Circuits
As polymer-ceramic composites are very flexible, they can be used as substrates for flexible microwave circuit. Here, we make a polymer-based microstrip line, which is the simplest microwave circuit, to demonstrate its capability. Fig. 2.15 displays
30 the configuration of such a flexible microstrip line(4 mm wide and 74 mm long)
printed on a PDMS substrate via evaporation and electroplating for 2 hours. We
note that the return loss (S11) is negligibly affected due to bending. The measured
insertion loss at different bending angles are shown in Fig. 2.16. As seen, the insertion
loss of the bent microstrip line is at most 0.2 dB worse than the planar microstrip
up to 2 GHz, effectively an attenuation factor of 0.8 dB/foot, which is very good
for UHF band transmission lines. The phase delay of the flexible microstrip line is
also within 5◦ of difference as we bend the flexible line, which is due to the slightly stretching or shrinking of the microstrip line. In other words, the flexible microstrip line maintains its functionality even when bent at 90◦. Therefore, the polymer-based
flexible microstrip lines can be used to design conformal feeding network circuits for
the cylindrical microstrip array as will be discussed in Chapter 5.
74mm
4mm
- φ
φ
Figure 2.15: Polymer-based flexible microstrip line
31 0
-0.1
-0.2
-0.3
-0.4
° -0.5 φ= 0 ° S21 (dB) S21 -0.6 φ= 30 φ ° -0.7 = 90 ° φ= -30 -0.8 ° φ= -90 -0.9
-1 0 0.5 1 1.5 2 Frequency (GHz) (a) Insertion loss (|S21|)
200
150
100
50
0
° -50 φ= 0 φ ° phase (S21):degree phase = 30 ° -100 φ= 90 ° φ= -30 -150 ° φ= -90 -200 0 0.5 1 1.5 2 Frequency (GHz) (b) Phase delay (angle of S21)
Figure 2.16: Measured insertion loss and phase delay of the flexible microstrip line under bending strains
32 2.4.2 Polymer-based Antennas
Since polymer-ceramic composites have controllable dielectric constant and low loss tangent, they are very suitable to be used as dielectric loading for antenna miniaturization. More importantly, polymer-ceramic composites allow for three- dimensional integration with matching circuit and other IC components by layer- by-layer printing. More results on polymer-based dual-layer and multilayer antennas are discussed in Chapter 4.
Probe feeding 57 mm
6.5% PDMS-D270 composite 16 mm 45 mm 12 mm
Figure 2.17: Polymer-based microstrip patch antenna
Here we show a simple polymer-based microstrip patch antenna as an example.
As displayed in Fig. 2.17, a rectangular patch was printed on a 6.5% PDMS-D270 substrate (12 mm thick, with ²r = 4.0 and tanδ = 0.008 in the 1-2 GHz band).
The patch (45 mm 45 mm) was realized by electroplating for 2 hours at a current × density of 20 mA/cm2. A coaxial probe located 16 mm from the patch center was used to excite the patch. This inner pin of the coaxial probe was soldered onto the patch without melting the polymer substrate as the polymer substrate can withstand
33 300◦C temperature. The return loss and realized gain for this patch (on a 6” wide
ground plane) are given in Fig. 2.18. As seen the measured and calculated levels are
in good agreement. Of importance is that the S11 is -18 dB at resonance and the
realized gain is about 7 dB. The measurement results demonstrate the capability of
polymer-ceramic composites for microstrip antenna applications.
2.4.3 Two-tone Textured Dielectric Substrate
As reported in [54, 55], three-dimensional topology optimization has been pro-
posed to mitigate the narrow bandwidth associated with high contrast dielectric
loading. This topology optimization can be achieved by using textured heteroge-
neous dielectric substrate which allows us to locally modify the dielectric constant
to enhance radiation while keeping the effective dielectric constant for antenna mini-
aturization [56]. Polymer-ceramic composites are well suited to fabricate such tex-
tured dielectric substrates. Fig. 2.19 shows a two-tone dielectric substrate consisting
of two different dielectric material (²r1 = 9.5 and ²r2 = 6). We first built a plastic mold by drilling holes on a chunk of Teflon. Then, the 10% PDMS-D270 compos- ite was poured into the mold. After curing, we took out the polymer substrate and poured the 20% PDMS-D270 composite to fill the holes. After the second curing, we achieved this two-tone textured dielectric substrate. Since polymer-ceramic compos- ites are fluid-like before curing and can be handled at room temperature, they can be easily molded to make textured dielectric substrates for antenna applications.
34 0
-2
-4
-6
-8
-10
S11(dB) -12
-14 Measured -16 Simulated -18
-20 1.3 1.4 1.5 1.6 1.7 1.8 frequency (GHz) (a) Return loss
10
5
0 Realizedgain (dB)
Measured Simulated
-5 1.3 1.4 1.5 1.6 1.7 1.8 frequency (GHz) (b) Realized gain
Figure 2.18: Measured performances of the polymer-based microstrip patch antenna
35 Fabricated two -tone dielectric substrate Mold for the two-tone substrate ε = r1 9.5 ε = r 2 6
Figure 2.19: Two-tone textured dielectric substrate
2.4.4 Three-layer Dielectric Rod Antenna
Two-layer dielectric rod antenna [57] has been proposed for wideband coverage.
Compared with a single dielectric rod waveguide/antenna, the dual-layer configura-
tion can support the hybrid HE11 mode with broader bandwidth. This design concept
can be extended to multilayer concentric dielectric rod antenna, which has the po-
tential to provide ultra-wideband (UWB) coverage. However, one challenge of such
multilayer configuration is associated with the accurate fabrication of the concentric
multilayer dielectric rods. Common ways of drilling and gluing different parts may
results in misalignment and air gap between the dielectric layers. Polymer-ceramic
composites provide us with a more accurate method to fabricate this multilayer di-
electric rod antenna. Fig. 2.20 shows a three-layer dielectric rod antenna. To fabricate
this antenna, we first made a dielectric cylinder as the core layer (²r1 = 9). Then, we built a mold around the core layer and poured the polymer composite (with different mixing ratio for ²r2 = 6) into the mold. After the second layer cured, we took it out
36 of the mold and built another mold for the third layer and pour the polymer com-
posite (²r3 = 4) into it. After the third layer cured, we finally achieved a three-layer dielectric rod configuration. Different polymer layers cross-linked with each other, making sure that there was no air gap at the interfaces. This example demonstrates that we are able to design and fabricate complicated antenna structure with the help of castable polymer-ceramic composites.
K9 K6 K4
ε = r3 4 ε = r 2 6 ε = r1 9
Launcher Waveguide Radiating section section section
Figure 2.20: Three-layer dielectric rod antenna
37 2.5 Summary
In this chapter, we presented novel engineered materials for antenna and RF systems, namely polymer-ceramic composites. We first demonstrate the fabrication process of PDMS-ceramic and RTV6166-ceramic composites. Then, we discuss the methods to characterize their dielectric properties. The measurement results show that PDMS-D270 composites can achieve controllable dielectric constant (² = 3 r ∼ 20) with low loss tangent (tanδ < 0.02) for frequency up to several GHz. We also
propose several application examples using polymer-ceramic composites. As can be
seen, polymer-ceramic composites are particularly suitable for conformal systems and
multilayer configurations. Therefore, they provide us with novel material platforms
for future communication systems and devices.
One difficulty of applying polymer-ceramic composites is associated with the hy-
drophobic property, which make it very difficult to print metals on top of polymer-
ceramic substrates. This is further discussed in Chapter 3 and the carbon nano-
tube sheet is therefore proposed to address the printing issue. More applications
of polymer-ceramic composites such as dual-layer or multilayer antenna design and
flexible microwave circuit can be referred in Chapter 4 and Chapter 5 respectively.
38 CHAPTER 3
POLYMER-CARBON NANOTUBE SHEET FOR CONFORMAL ANTENNAS AND RF CIRCUITS
3.1 Introduction
Conformal light-weight load-bearing materials are the key to future antennas and
RF front-ends for small aircrafts and body-worn applications. There is also interest for multilayer three-dimensional RF architectures with each layer bearing different functionalities. Such conformal antenna and multilayer circuit structure call for a new class of materials with desirable electrical/RF (low loss, and high permittivity), mechanical (flexible, light-weight, strong shear and tensile rating) as well as thermal properties.
As discussed in Chapter 2, polymer-ceramic composites (such as PDMS-D270 composites) are attractive because they are extremely flexible and not as sensitive to large temperature variations. They also have low loss (tanδ < 0.02) up to several GHz and controllable dielectric constants (relative permittivity of ² = 3 20) [23]. Fur- r ∼ ther, as compared to other materials such as those based on liquid crystal polymers
(LCP) [27, 28, 29] and low temperature co-fired ceramics (LTCC) [30, 31], PDMS composites can be processed at room temperature. Specifically, for bonding LCP
39 and LTCC layers, they must be heated (300◦C for LCPs and 1000◦C for LTCCs) at temperatures that could cause failure to some IC components and fragile wire bonds. With these issues in mind, PDMS-ceramic composites are well suited for con- formal load-bearing antennas and RF systems integration. However, metalization or printing on PDMS substrates remains a challenge. Specifically, common lift-off lithog- raphy methods using metal evaporation does not work well for PDMS due to poor metal-polymer adhesion [58]. Further, interface incompatibilities can easily cause detachment of the printed layers under bending or tensile stress.
In this chapter, we propose a novel polymer-printing technology utilizing carbon nanotube (CNT) sheets [24, 25]. In section 3.2, we first discuss the property of a single CNT, especially its conductivity, which is the major reason of low radiation efficiency of a single CNT dipole. To address the issue of low conductivity of a single
CNT, we then present two configurations of CNT sheets, namely E-textile CNT sheet
(section 3.3) [24] and vertically-aligned CNT sheet (section 3.4) [25]. Fabrication processes are discussed in details. The polymer-CNT patch antennas were fabricated and measured to evaluate the mechanical and electrical properties of the CNT sheets.
Moreover, we also study the circuit models of the vertically and horizontally aligned
CNT array in section 3.5. The effective circuit models indicate that the horizontally- aligned CNT array, or the bundled CNT thread [59, 60], has the potential to achieve higher conductivity while maintaining its mechanical properties.
3.2 Single Carbon Nanotube Property
Carbon nanotubes (CNTs) are cylindrical carbon molecules with novel properties that make them potentially useful in a wide variety of applications [61, 62, 63]. Their
40 name is derived from their size, since the diameter of a nanotube is on the order of a few nanometers (approximately 50,000 times smaller than the width of a human hair), while they can be up to several centimeters in length. There are two main types of nanotubes: single-walled nanotubes (SWNTs) and multi-walled nanotubes
(MWNTs). As shown in Fig. 3.1, a SWNT is simply rolled up from a one-atom-thick graphite sheet. If the cylinder axis is the ξ axis in Fig. 3.1(a), the resulting tube is called a zigzag CNT. If the cylinder axis is the η axis in Fig. 3.1(a), the resulting tube is called an armchair CNT. If the cylinder axis is neither the ξ nor the η axis as shown, the resulting nanotube is called a chiral CNT. Therefore, carbon nanotubes are usually characterized by the dual index (m, n) , where (m, 0) for zigzag CNTs,
(m,m) for armchair CNTs, and (m, n), 0 < n = 0 for chiral CNTs. 6 3.2.1 Mechanical Property
Single-walled carbon nanotubes (SWNTs) have very good mechanical properties due to their hollow shapes and sp2 carbon-carbon covalent bonds. SWNT has a density of 1.3-1.4 g/cm3 [64]. Although SWNT is very light-weight, it is among the strongest and stiffest materials with Young’s Modulus larger than 1 TPa [65, 66].
This number is around 10 times higher than that of the high-carbon steel (0.2 TPa).
The light weight and high strength makes CNTs very attractive for small aircrafts where system load and mechanical strength is critical. Further, CNTs are stable at high temperatures up to 750◦C [63], a desirable characteristic for harsh and high temperature environments. And metallic CNTs do not oxidize [63] and are not sus- ceptible to moisture [64]. Also, CNTs are very good thermal conductor [67, 68]. CNTs can transmit heat at greater than 3000 watts per Kelvin per meter (W/K m). As a ·
41 (a) One-atom-thick graphite sheet. (Circles denote the positions of carbon atoms.)
l: can be mm long
D ~ 1.5 nm
(b) Single-wall carbon nanotube (SWNT)
Figure 3.1: Configuration of a single-walled carbon nanotube
42 Density 1.3 - 1.4 g/cm3 Young’s Modulus > 1 TPa Temperature 750◦ Thermal Conductivity 3000 - 6000 W/K m ·
Table 3.1: Mechanical properties of carbon nanotubes
comparison, copper transmits 385 watts per Kelvin per meter. Therefore, CNTs can
better protect the electric system from overheating because of good heat dissipation.
In short, the mechanical properties of CNTs are summarized in Table. 3.1.
3.2.2 Electrical Property
Carbon nanotubes have been proposed as the building blocks for nanoelectronics
due to their favorable electrical properties [64, 69]. Recently, CNTs have also drawn
significant attention in the antenna and RF community. For example, CNTs are
attractive in realizing antennas at millimeter wave, even up to optical frequency
range [70, 71, 72, 73, 74]. It has been reported that a single CNT dipole exhibits
significantly slower wave velocities (v = 0.02 c, where c is the speed of light) above the p · relaxation frequency of around 53 GHz [70, 72]. However, so far, reported antennas have very low radiation efficiencies due to large resistance along the single CNT [70,
72], around 6.45 kΩ/µm.
According to Ohm’s Law, the resistance of a wire is proportional to the length of the wire. However, as the length of the wire is reduced to the mean free path of electrons (l), the electron transport mechanism changes from diffusive to ballistic [75,
76], as shown in Fig. 3.2. Furthermore, when the width of the wire is reduced to the nanometer or Fermi wavelength scale (λF , around 0.74nm), the resistance between
43 the electrodes is quantized in steps of h/(2e2) = 12.9 kΩ, where h = 6.626 10−34Js × is Planck constant and e = 1.602 10−19c is elementary charge. Therefore, for a × single carbon nanotube, it has large resistance long the nanowire as the diameter is
close to Fermi wavelength. And the resistance is not dependent on the length of the
nanowire. Instead, the resistance can be theoretically calculated by the Landauer
formula [75, 76]: h/(2e2) R = (3.1) P Tj where Tj is the electron transmission probability for the j-th conductance channel.
Since Tj = 1 for ballasitic transport, the resistance of a single CNT is therefore:
h/(2e2) 12.9kΩ R = = = 6.45kΩ (3.2) N 2 where N = 2 is the number of conduction channels of a single CNT. Although this
resistance appears to be very high, the conductivity of a single metallic CNT is
actually very high if we consider its small diameter. As shown in [72], given the typical
CNT length of 1 µm and diameter of 1.5 nm, the effective conductivity calculated
according to Ohm’s Law is:
l 1 10−6 σ = = × = 8.78 107S/m (3.3) R πD2/4 6.45 103 3.14 (1.5 10−9)2/4 × · × · · × which is even higher than the conductivity of copper, 5.96 107S/m. Therefore, × CNTs have become very attractive as the building blocks for nanoscale electronics
and semiconductors. However, for antenna and most RF applications, the single CNT
has too high resistance, which reduces the antenna efficiency significantly. In the next
sections, we study the CNT arrays or CNT sheets to overcome the high resistance of
a single CNT so that we can use CNT for antenna and RF systems.
44 L > l
D > F
(a) Diffusive transport
L ~ l
D ~ F
(b) Ballistic transport
Figure 3.2: Diffusive and ballistic transport of electrons in one-dimensional wires
3.3 E-textile Carbon Nanotube Sheet
Since a single CNT has very high resistance, CNT ensembles or arrays have been proposed to improve the surface conductivity [77, 78, 24, 25]. For example, non- aligned CNT ensembles have been reported to reduce sheet resistance down to around
20 Ω/¤ [78]. However, even this lower resistance is still too high to realize efficient
CNT antennas. In this section, we propose the so-called E-textile carbon nanotube sheet [24] to increase its conductivity for antenna and RF applications.
3.3.1 Fabrication Process
As indicated by its name, the E-texile CNT sheet (Fig. 3.3) is a textile coated with SWNTs and metal particles to improve the sheet conductivity. To fabricate such
E-texile CNT sheet, we first prepared the SWNT dye by dispersing SWNTs (from
45 Carbon Nanotechnologies Inc.) in the diluted nafion-ethanol [79]. A commodity
cotton textile was then dipped in the SWNT dye dispersion for 10 seconds and dried
for 1 hour at 60◦C. This dying process was repeated 10 times to increase the sheet
conductivity of the E-textile, as shown in Fig. 3.4. The darker the E-textile was, the
more SWNT dye it had. After 10 times of SWNT dipping, the sheet resistance of
the E-textile is around 10 Ω/¤. To further increase the conductivity of E-textile, silver particles (Ag) were sputtered for 200 seconds, which improved the conductivity but slightly compromised flexible nature of the E-textile. We next treated the CNT- coated fabric in a hot press overnight for 24 hours at 1000◦C to achieve a strong adhesion of SWNTs and Ag particles into the cotton textile so that the sample still preserves its conductivity even after severe bending. The resulted E-textile had a thickness of 150 µm and a sheet resistance of around 1 Ω/¤.
E-textile patch CNT-coated cloth CNT-coated threads CNTs
Figure 3.3: E-textile CNT sheet
3.3.2 E-textile Patch Antenna
We then demonstrate how to utilize the E-textile CNT sheet as conductive printing
on polymer substrate. Fig. 3.5 displays the process of printing an E-textile CNT sheet
on a polymer-ceramic substrate. We first prepared an E-texitle CNT sheet following
46 Commodity cotton fabric After 1 SWNT dipping After 10 SWNT dipping
Figure 3.4: Fabrication process of E-textile CNT sheet
the above procedures. Then we cut the E-textile into desired pattern according to the microstrip patch antenna geometry specifications. We subsequently mixed the polymer-ceramic composite and poured the mixture over the E-textile fabric. After curing, the E-textile adheres strongly to the polymer-ceramic composite because of its rugged surface, thus implying strong mechanical and chemical compatibility.
To evaluate the RF performance of the E-textile CNT sheet, we measured an E- textile patch antenna as shown in Fig. 3.6. As seen, the E-textile CNT patch (35 mm
35 mm) was printed on a polymer substrate of thickness 300 mils and permittivity ×
²r = 4.0. The E-textile patch had a sheet resistance of 2 Ω/¤. The antenna was mounted on a ground plane to carry out measurements in the OSU-ESL anechoic chamber. To compare performance of the sample antenna to that of an ideal patch, we also modeled a PEC patch with lossless substrate of permittivity ²r = 4.0 in HFSS.
Referring to the Fig. 3.7 and Fig. 3.8 displaying return loss and gain of the E- textile patch, respectively, we note that the proposed E-textile patch antenna has a similar performance as a traditional patch antenna except that E-textile sample has
6 dB of gain at 2 GHz, 2 dB less than an ideal patch. The sample patch also has a
47 I. Prepare E -textile II. Cut into desired pattern
IV. After curing III. Pour polymer mixture on top of E-textile
Figure 3.5: Process of printing E-textile on polymer-ceramic composite
48 (a) Fabricated E-textile patch antenna
(b) Simulation model in HFSS
Figure 3.6: Polymer-based E-textile patch antenna
49 slightly wider bandwidth than that of ideal patch and this is due to the resistive loss associated with the E-textile conductor. It is also important to note that resistance of the E-textile patch plays a critical role in the gain of the antenna as expected. For the same dimensions of E-textile patch with 10 Ω resistance leads to 0 dB antenna gain as is the case of copper (Cu) particles sputtering on the E-textile instead of silver
(Ag) particles.
0
−5
−10
−15 S11 (dB) −20
−25 Measured E−textile patch Simulated PEC patch
−30 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Frequency (GHz)
Figure 3.7: Measured return loss of the E-textile patch antenna
3.4 Vertically-aligned Carbon Nanotube Sheet
One disadvantage associated with the E-textile CNT sheet is its low flexibility and elasticity, which is limited by the cotton fabric. Further, as more SWNTs are coated onto the cotton fabric, the adhesion between the E-textile and polymer composite is compromised especially under bending strains. To address the issue of conductivity
50 10
5
0 Realized gain (dB) −5 Measured E−textile patch Simulated PEC patch
−10 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Frequency (GHz)
Figure 3.8: Measured gain of the E-textile patch antenna
and adhesion to polymer substrates concurrently, we propose to grow a vertically-
aligned CNT array inside the polymer substrate in a similar way as body hairs [25],
but in much higher density (about 3 109 nanofibers per cm2) to form a CNT sheet × (as shown in Fig. 3.9).
3.4.1 Fabrication Process
The vertically-aligned CNT array was synthesized via chemical vapor deposition
(CVD) process [80] (see Fig. 3.10). The process is as follows: first we sputtered an
array of ferrous particles on a silicon wafer to serve as catalysts for CNT growth.
Next the silicon substrate was placed inside a tube furnace (Thermolyne 79400) and
methane gases (CH4) were blown into the furnace via a carrier argon flow. At high
temperature (1000◦F), methane gases were decomposed into carbon atoms, aligned
51 Vertically-aligned CNT sheet
31 mm 5 mm MCT-PDMS 56 mm substrate
(a) Printed vertically-aligned CNT sheet on polymer substrate
Cross-section view Top view
100 nm 100 m
(b) SEM photo of the vertically-aligned CNT sheet
Figure 3.9: Vertically-aligned CNT sheet
52 along the catalyst particles into cylinder forms. By controlling the furnace tempera-
ture (1000◦F) and deposition time (2 hours), we can achieve a vertically-aligned CNT array on the silicon wafer as shown in Fig. 3.9(b). The length of the CNT array is around 200 µm and the CNT density is around 3 109 nanofibers per cm2. ×
silicon wafer
I: Prepare the silicon wafer
Catalyst (FeO)
II: Sputter catalysts
Tube Furnace (1000ºF)
C2H2/ catalyst CH 4
CNTs
Si wafer
III: Grow CNTs inside a tube furnace
Figure 3.10: Process for growing vertically-aligned CNT sheet
53 After growing the vertically-aligned CNT array, we then transfered the CNT sheet onto the polymer substrate using a two-stage curing process. First a thin PDMS composite layer was spin-coated onto the CNT sheet as displayed in Fig. 3.11. After curing, the CNTs were implanted inside the thin polymer layer to form a polymer- coated CNT sheet. The polymer-coated CNT sheet was then detached from the silicon wafer by dissolving SiO2 on the Si surface using hydrofluoric (HF) acid. During this process, we observed around 2% shrinkage of the polymer, further increasing the CNTs density and improving conductivity. In the final step, we embedded the polymer-coated CNT sheet into a larger customized polymer-ceramic substrate for antenna loading. During this curing stage, the polymer-ceramic substrate cross-linked with the coated polymer, leading to a strongly bonded CNT sheet on the polymer- ceramic substrate. In the next section, we present the measurement results to evulate the RF and mechenical performances of the polymer-CNT patch antennas.
3.4.2 Polymer-CNT Patch Antenna
Below we evaluate the electrical and mechanical performances of the vertically- aligned CNT sheet. Specifically, a polymer-CNT patch antenna was printed and measured to demonstrate the radiation characteristics of the polymer-CNT antenna.
We also carried out stress and tensile tests to study the electrical performance of the vertically-aligned CNT sheet as they were bent and stretched. Last but not least, we measured two polymer-CNT patches mounted onto a cylinder to demonstrate its application as conformal antennas.
54 PDMS Polymer-coated CNT sheet
I. Spin coating
II. Detach by HF acid
III. Embed inside large polymer composite
Polymer-ceramic substrate
IV. Cure
Polymer-CNT patch
Figure 3.11: Process for transferring CNT sheet onto polymer-ceramic composites
55 Planar Polymer-CNT Patch Antenna
Referring to Fig. 3.12, we show a 31 mm 31 mm CNT sheet printed on a 56 mm × 56 mm PDMS-MCT substrate which has a dielectric constant of ² = 3.8 and loss × r tangent of tanδ = 0.015 at the resonant frequency of 2.25 GHz. The patch was fed by a coaxial probe soldered to the CNT sheet by conductive epoxy. The polymer-CNT patch was then measured on a 150 mm 150 mm ground plane at the OSU-ESL × anechoic chamber. The measurement data are then compared with HFSS simulations for perfectly electric conductor (PEC) patch and a finite conductivity patch (surface resistance: 0.9 Ω/¤). As shown in Fig. 3.14, the measured gain (blue solid curve) agrees well with the finite conductivity patch (green dashed curve), verifying the low sheet resistance (0.9 Ω/¤) of the vertically-aligned CNT sheet. Further, the measured radiation pattern (Fig. 3.15) is that of a typical patch antenna. Of importance is that the measured gain of the CNT patch was 5.6 dB, i.e. only 0.8 dB lower than that of the simulated PEC patch (red curve) of the same dimensions and substrate dielectric properties. Indeed this is a very good RF performance for practical applications.
Strain and Tensile Tests
As load-bearing antennas are subject to vibration and temperature change, they are usually deformed by bending or stretching. Therefore, it is critical to characterize their electrical properties under strain and tensile stresses. Specifically, it is desirable that the CNT sheet conductivity remains stable as the substrate is bent or stretched.
Here, we only measure the DC resistance of the CNT sheet under different stain and tensile stresses (which can be used to predict the RF performance of the CNT sheet).
56 (a) Vertically-aligned CNT patch antenna
MCT-PDMS substrate Probe feeding 31 mm
5 mm
8 mm 56 mm
(b) Simulation model in HFSS
Figure 3.12: Polymer-based patch antenna printed via vertically-aligned CNT sheet
57 0
−5
−10 S11 (dB)
−15 Measured CNT patch Simulated CNT patch Simulated PEC patch −20 1.5 2 2.5 3 Frequency (GHz)
Figure 3.13: Measured return loss of the vertically-aligned CNT patch
10
5
0
−5
Realized gain (dB) Measured CNT patch −10 Simulated CNT patch Simulated PEC patch
−15 1.5 2 2.5 3 Frequency (GHz)
Figure 3.14: Measured gain of the vertically-aligned CNT patch
58 10
0
-10
gain (dB) gain -20
Measured CNT patch -30 Simulated CNT patch Simulated PEC patch -40 -180 -120 -60 0 60 120 180 θ (degree)
Figure 3.15: Measured E-plane pattern of the vertically-aligned CNT patch
Fig. 3.16(a) shows the measurement setup for the bending test. As seen, we clamped a sample polymer-CNT sheet at the two ends and exerted an external force using a universal test machine to deform the polymer substrate. The deformation was evaluated by the degree of bent angle (θ). The DC sheet resistance was subsequently
measured and recorded as the sample was deforming. Fig. 3.16(b) gives the measured
curves for the two samples bent positively and negatively. We observed that the DC
sheet resistance was fairly stable within a large range of angles up to +/ 130◦. As − expected, the resistance decreases when positive strain is applied since the CNTs are
pushed toward each other. In contrast, negative strain further separates the CNT
“hairs”, leading to higher resistance. Since the rate of decrease/increase in resistance
is slow, we expect that RF performance degradation will also be comparatively slow.
We notice that the CNT sheet recovered the conductivity of the unbend case after
we released the CNT sheet from bending (before we broke the sample).
59 positive strain
negative strain
(a) Measurement setup
3
2.5 Positive strain Negative strain
/square) 2 Ω
1.5
1
0.5 Sheet Sheet resistance (
0 130 140 150 160 170 180 Bent angle (degree) (b) DC sheet resistances vs. strain
Figure 3.16: Bending test for the vertically-aligned CNT sheet
60 It is also important to characterize the conductivity of the polymer-CNT sheet un- der stretching condition as well. To do so, we employed the setup shown in Fig. 3.17(a) and measured the DC sheet resistance as the sample was elongated. In this setup, the degree of elongation is defined as ∆L/L, where L is the original length of the sample and ∆L refers to the length after stretching. Obviously, stretching decreases the CNTs density effectively, leading to an increase in resistance as depicted in Fig. 3.17(b). This behavior also agrees with percolation theory [81, 82, 83] that predicts an exponen- tial relationship between conductivity and CNT density. Nevertheless, the sample resistance was fairly stable within 2% of elongation, a typical maximum stretching for most practical applications. We also notice that the CNT sheet had repeatable conductivity as we stretched and released the CNT sample back and forth. In the fu- ture, we expect to improve the CNT sheet conductivity under large strain and tensile stresses by sputtering metal nanoparticles on the nanotubes and by inter-dispersing horizontal CNTs into the vertically aligned CNTs.
Conformal CNT Patch Antennas on a Cylinder
We then examine two conformal CNT patches mounted on a cylinder surface. As shown in Fig. 3.18, we attached two polymer-CNT patches on a metal cylinder (80 mm in diameter and 160 mm in length). Referring to Fig. 3.18(a), the patch is bent along the E-plane, implying that the current flows along the circumferential direction.
Alternatively (see Fig. 3.18(b)), when the patch is bent in the H-plane, the current
flows along the axial direction. We remark that bending led to a 13% stretching in this case. Therefore, the E-plane resonance frequency was decreased from 2.25 GHz to
1.95 GHz. This frequency shift can be justified by taking into consideration Young’s modulus of the polymer substrate and the geometry of the platform.
61 L/2 L L/2
(a) Measurement setup
300 10 250 8 /square) Ω 6
200 4 /square)
Ω 2
150 0
Sheet resistance ( 0 1 2 3 4 5 Elongation (%) 100
Sheet resistance ( 50
0 0 5 10 15 20 25 30 Elongation (%) (b) DC sheet resistances vs. elongation
Figure 3.17: Stretching test for the vertically-aligned CNT sheet
62 80 mm 80 mm
100º 100º
160 mm J 160 mm J
(a) E-plane bending (b) H-plane bending
Figure 3.18: Cylindrically mounted polymer-CNT patch antennas
The measured return loss and radiation patterns of the conformal CNT patches were identical to the simulated PEC patch on the same cylinder surface. As shown in Fig. 3.19, the E-plane bent CNT patch had a gain of 1.7 dB at 1.95 GHz, viz.
3 dB lower than that of a simulated PEC patch in the same bent configuration.
This is because the CNT sheet resistance was increased since the bending (100◦) and
stretching (13% elongation) reduced the nanotube density. The corresponding H-
plane CNT patch had a measured gain of 2.9 dB at 2.25 GHz, viz. 1.5 dB lower than
the simulated PEC patch on the same cylinder (Fig. 3.20). Notcie that the H-plane
bent patch has higher gain than the E-plane patch since the electrical currents flow
along the unbent direction of the CNT patch and thus not subject to the stretching
effect. For most practical applications, the antennas will not be subjected to such
large bending or stretching. Therefore, we expect higher antenna gain as the strain
and bending will be smaller.
63 0
-5
-10
S11(dB) -15
-20 Measured Simulated with PEC -25 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Frequency (GHz)
(a) Return loss
5
0 Gain (dB) Gain -5
Measured Simulated with PEC
-10 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Frequency (GHz) (b) Realized gain
Figure 3.19: Measured performance of polymer-CNT patch bent along E-plane
64 0
-5
-10
S11(dB) -15 Measured Simulated with PEC -20
-25 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Frequency (GHz) (a) Return loss
5
0 Gain (dB) Gain -5
Measured Simulated with PEC
-10 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 Frequency (GHz) (b) Realized gain
Figure 3.20: Measured performance of polymer-CNT patch bent along H-plane
65 3.4.3 Polymer-CNT Microstrip Line
In order to realize flexible circuits by conductive polymer-CNT sheets, below we then studied the RF performances of the polymer-CNT microstrip lines. As shown in
Fig. 3.21(a), a small section of CNT sheet was integrated with the polymer substrate to realize a CNT microstrip line. Three CNT sheets (L = 60 mm and w = 5 mm) were fabricated by the aforementioned process with density of around 3 109 nanofibers × per cm2 and with thickness t = 80µm, t = 50µm, and t = 35µm respectively.
The substrate was the D270-PDMS composite with dielectric constant ²r = 4. A copper tape microstrip line with the same dimension was also attached to the polymer substrate as a reference for performance evaluation.
Fig. 3.21(b) shows the measured insertion loss (S21) of the three CNT microstrip lines compared with the copper tape microstrip line. As seen, the insertion loss of the CNT microstrip lines are at least 1 dB lower than that of the copper tape microstrip line, which indicates additional attenuation of the CNT microstrip lines.
As the microstrip lines are of the same dimension and printed on the same polymer substrate, we conclude that this attenuation is associated with the conductor loss of the CNT sheets. Given the length of the microstrip lines, the attenuation factor is around 1 dB/60 mm = 16.7 dB/m, which is too high for microwave operation. We also notice that with higher thickness, the conductor loss is reduced. The conductivity calculated by the DC resistance based on Ohm’s Law is around 5000 - 10000 S/m.
Compared to copper, this conductivity is at least 1000 times smaller, which is the limiting factor for the use of CNT sheets for microwave circuits. Next, we propose two CNT array models to improve the CNT conductivity.
66 CNT microstrip line Copper tape microstrip line
W J t
L
(a) CNT microstrip lines with different thickness
1
0
-1
-2
-3
-4 S21 (dB) S21
-5 Copper TL -6 CNT TL (t=80 um) CNT TL (t=50 um) -7 CNT TL (t=35 um) -8 0 2 4 6 8 10 frequency (GHz) (b) Measured insertion loss (S21)
Figure 3.21: Polymer-based microstrip line printed by vertically-aligned CNT sheet
67 3.5 Approaches to Increase Carbon Nanotube Sheet Con- ductivity
As shown before, we can achieve a CNT sheet with either E-textile or vertically- aligned CNT array with a sheet resistance around 1 Ω/¤. This conductivity is good
enough for a microstrip patch antenna operation since the wide patch has an equiva-
lent circuit of resistors in parallel. However, for a narrow microstrip line which usually
has a width less than 1 mm, the previous CNT sheets will result in a resistance of
around 20 Ω since the conductivity is only around 10000 S/m. This is unacceptable
for most microwave applications. To further increase the conductivity of the CNT
sheets for RF circuits, we then investigate the equivalent circuit model of the CNT
sheets and studied the parameter effects on CNT sheet conductivity.
3.5.1 Vertically-aligned CNT Sheet
Since the electron transports along the nanotube ballistically, the electric cur-
rent can flow perpendicular to the nanotube stems only when there are entangle-
ments among CNTs as shown in Fig. 3.22(a). This entanglement, as displayed in
Fig. 3.22(b), is the key to high conductivity of the vertically-aligned CNT sheet.
The equivalent circuit model of the vertically-aligned CNT array is thus shown in
Fig. 3.22(c), where NL, NW and Nt are the number of touching per unit length along
the length, width and thickness direction of the CNT array, respectively. R1, Rc de-
notes the resistance of a unit section of single CNT and the contact resistance between
two touching CNTs, respectively.
68 nanotubes entanglement
J J
(a) SEM photo of the vertically- (b) Simplified model of the aligned CNT sheet vertically-aligned CNT array
L·NL (in series)
W·NW (in parallel) R1+R c
J t·Nt layer (# of touching points along nanotubes)
(c) Equivalent circuit model
Figure 3.22: Circuit model of the vertically-aligned CNT array
69 According to this equivalent circuit model, the resistance of the vertically-aligned
CNT array is therefore calculated as:
LNL R =(R1 + R ) (3.4) c WN tN W · t Equation 3.4 was validated by measuring the DC resistance of different CNT lines.
The measurement results are shown in Fig. 3.23. The dots are the measured DC
resistance of the CNT samples with different length (L), width (W) and thickness
(t). The lines are the curves fit into the measurement points. As seen, the DC
resistance is proportional to the CNT line length (L) and inversely proportional to
the CNT line width (W ) and thickness (t). Particularly, the thickness effect has an offset resistance even when t . This is because the entanglement only happens → ∞ around the tip section of the CNTs (as shown in Fig. 3.9(b)). Therefore, increasing the CNT sheet thickness does not necessarily increase the CNT entanglement after certain CNT length.
The resistance of the CNT sheet can be written in terms of sheet resistance by:
L R = R (3.5) s W
From Equation 3.4 and Equation 3.5, assuming CNTs are uniformly entangled (NL =
NW ), we can obtain the sheet resistance (Ω/¤) of the vertically-aligned CNT array:
R1 + Rc Rs = (3.6) tNt
Therefore, in order to reduce the sheet resistance of the CNT array for microwave circuits, we must increase the entanglement along the thickness direction (Nt). In
other words, we must increase the density of CNTs or reduce the spacing distance
between CNTs.
70 t=50 µm 40
35 w=1 mm 30 w=2 mm w=4 mm 25 ) Ω 20
15
Resistance( 10 W J t 5 0
-5 0 10 20 30 40 50 60 L Length: L (mm) (a) DC resistance measurement (b) Length effect
t=35 µm w=2 mm 60 30
L=20 mm 50 25 L=40 mm L=20 mm L=60 mm 40 L=40 mm L=60 mm 20 ) ) Ω Ω 30 15 20
Resistance( Resistance( 10 10
0 5
-10 0 0 0.2 0.4 0.6 0.8 1 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 1/width (mm -1 ) 1/thickness (um -1 )
(c) Width effect (d) Thickness effect
Figure 3.23: Measure DC resistance of the vertically-aligned CNT array
71 The effect of CNT density on conductivity can also be revealed by percolation theory [81, 82, 83], which models the polymer-CNT sheet as a composite, where the fillers being the carbon nanotubes. In accordance with percolation theory, the conductivity of the polymer-CNT sheet, σ, is given by:
σ ∝ (P P )α (3.7) − c where P is the CNT volume density, Pc is the percolation threshold density, and α is an exponent related to the conductivity. Peculation theory also agrees with the observation of tensile test. As shown in Fig. 3.17, stretching effectively reduces the volume density of CNTs, thus leading to an exponential increase of the resistance.
One way to increase the density is to grow CNTs with smaller spacing distance.
However, it is not easily achieved with the current fabrication techniques. The mini- mum spacing distance between CNTs is around 100 nm. Another approach to increase the entanglement of CNTs is to squash the CNT array as shown in Fig. 3.24. We first grew a long CNT array (t = 5 mm) on the silicon wafer. The sheet resistance was around 10-15 Ω/¤. Then, we squashed the CNT array from the top, reducing the thickness to around 1 mm. The sheet resistance was thus reduced to around 1.8-2.8
Ω/¤, which was around 1/5 of the previous value. This shows that we can improve
CNT sheet conductivity by increasing CNT density.
3.5.2 Horizontally-aligned CNT Sheet
Since the electrons transport more easily along the nanotube than perpendicular to the nanotube, we then propose to grow the CNTs horizontally to make a high conductivity CNT sheet. This concept it shown in Fig. 3.25. The CNTs are bundled
72 Long CNT array 10 Ohm 17 Ohm 5mm
5mm 5mm 8mm
(a) Long CNT array (before squashing)
Squash 1.8 Ohm
2.8 Ohm
~1mm 5mm 8mm
(a) Squashed CNT array
Figure 3.24: Squashed CNT array to increase CNT entanglement
together to realize the horizontally-aligned array. The resistance of this horizontally-
aligned CNT array is calculated as:
LNL R =(R1 + R ) (3.8) c WN t/D W · where R1, Rc are the resistance of the unit section of single CNT and contact resis- tance between two CNTs, NL, NW denotes the number of touching per unit length
along the length and width direction, D is the diameter of a single CNT, and t/D is
approximately the number of CNT layers in the thickness direction. Assuming the
CNTs are entangled uniformly, the sheet resistance can be obtained as:
R1 + R R = c (3.9) s t/D
73 We remark that for the horizontally-aligned CNT array, the electric current can still
flow along CNTs even there is no touching along the thickness direction. This is
the major difference between the horizontally-aligned and the vertically-aligned CNT
array since the vertically-aligned CNT array depends on the touching (Nt) to increase
conductivity. Therefore, the horizontally-aligned CNT array has potential to be more
conductive than the vertically-aligned CNT array as long as it is thick enough to allow
more current flows in parallel.
L
W …… J
t …… nanotubes
(a) Simplified model of the horizontally-aligned CNT array
L·NL (in series)
W·NW (in parallel) R1+R c
J t/D layer
(b) Equivalent circuit model
Figure 3.25: Circuit model of the horizontally-aligned CNT array
The horizontally-aligned CNT array can be fabricated by spinning out the CNTs into long CNT fiber or CNT thread [59, 60]. Then the CNT thread can be bundled
74 together to further increase conductivity. Thicker CNT thread can be weaved into
patch, transmission line and other configurations for RF circuits. This is similar to
the fabrication process of E-textile CNT sheet except that we weave the CNT threads
directly instead of aligning CNTs along the cotton fabric. Due to the ruggedness and
porosity of CNT textile, we expect to achieve strong polymer-CNT adhesion and high
CNT conductivity concurrently.
3.6 Summary
In this chapter, we present a flexible, light-weight and conductive polymer-CNT sheet to address the issue of printing on polymer composites. Two approaches, namely
E-textile and vertically-CNT sheet, are proposed to practically form a conducting sheet on the polymer substrate. Further, the proposed CNT printing achieves very strong adhesion between CNT sheet and polymer substrate, leading to structural integrity under stress, strain and bending.
Specifically, the E-textile CNT sheet technology is based on CNT-coated cotton fabric. The porosity and ruggedness of the cotton fabric increases the adhesion be- tween CNT sheet and polymer surface. The CNT coating and Ag particles sputtering contributes to the high conductivity of the E-textile. We discuss the fabrication pro- cess to achieve an E-textile CNT sheet with a sheet resistance of 1-2 Ω/¤. We also
measured an E-textile patch antenna, which exhibited a gain of 6 dB, only 2 dB lower
than that of an ideal PEC patch (63% antenna efficiency). We note that with more
CNT coating and Ag particle sputtering, the conductivity can be further improved.
However, the flexibility of the E-textile and adhesion with polymer substrate are com-
promised. Future researches may include investigation of alternative textile such as
75 silk, nylon and etc., which are light-weight, flexible and stretchable, to achieve more
flexibility.
We also propose to grow a vertically-aligned CNT array within the polymer sub-
strate to realize a high conductivity printing on polymer. We describe the fabrication
process of the vertically-aligned CNT sheet and how to integrate it with polymer sub-
strate. The resulted CNT patch was measured to have a sheet resistivity of 0.9 Ω/¤
(as compared to 20 Ω/¤ reported with other CNT ensembles). A sample polymer-
CNT patch antenna was fabricated and measured to exhibit a gain of 5.6 dB, viz. only
0.8 dB less than that of a simulated ideal patch (83% antenna efficiency). Mechan- ical tests were also carried out to demonstrate the flexibility of the polymer-CNT sheets. Further, two cylindrically conformal CNT patch antennas were fabricated and measured. The results demonstrate an acceptable gain performance for prac- tical applications, making the proposed polymer-CNT sheets suitable for conformal load-bearing antennas.
To further increase CNT sheet conductivity, we study the equivalent circuit models of the vertically and horizontally CNT array. Since the electrons more naturally tend to transport along the CNT stem, the horizontally-aligned CNT array, or the bundled
CNT thread, can achieve higher conductivity than the vertically-aligned CNT array.
This is one of the future area of interests.
76 CHAPTER 4
COMPACT POLYMER-BASED GPS ARRAY
4.1 Introduction
As discussed before, one important application of polymer-ceramic composites and carbon nanotube sheets is the conformal light-weight antennas for small aircrafts.
For example, future military Global Positioning System (GPS) operations demand small compact GPS antenna and array that covers all three GPS bands (L5: 1575
MHz, L2: 1227 MHz and L1: 1176MHz) [84, 85] and can be conveniently mounted on individual solders or vehicles without compromising military missions and positioning operations. The nominal GPS controlled reception pattern antenna (CRPA) has seven elements and extends 14 inches in diameter, making it too large for airborne platforms of interest. Recently, a variety of antennas have been proposed for GPS receivers [86, 87, 88, 89, 90]. However, most of these designs are too large for anti- jamming GPS arrays. In this chapter, we address two main shortcomings in the state-of-art GPS anti-jamming systems. One of them relates to antenna size and the other pertains to the packaging or integration of the radiators and the antenna electronics. In this chapter, we develop a compact GPS antenna system that is only
3-4 inch large, has low profile and maintains multiple antennas for anti-jamming
77 performances. Further, polymer-ceramic composites and carbon nanotube sheets are
employed to realize the light-weight, high conforamlity and high system integrity
antenna structures.
Table 4.1 lists the design specifications for the compact GPS antenna system.
The first challenge of this compact GPS antenna system is the small GPS element size (around λ/10 at its lowest resonant frequency). Several dual band GPS antenna
have been proposed [86, 87, 88]. However, these antennas does not meet the size
requirement. Secondly, inclusion of the new GPS signal (L5 band) also presents us
a concurrent challenge on maintaining bandwidth while reducing antenna size. The
existing dual band GPS antenna does not cover the new L5 signal due to the narrow
bandwidth. Thirdly, the design of the feeding network is also critical as it is related
to array electronics. For a compact GPS array, it is desirable to reduce the number of
feeding cables and power dividers by exciting each element with a single coaxial cable
while keeping the RHCP excitation. Fourthly, for anti-jamming operation, we prefer
more GPS elements to provide more pattern diversity. Therefore, mutual coupling
between elements need to be considered since it usually affects the pattern coverage
(maximum beam direction, beamwidth and etc).
In this chapter, we first discuss the development of the miniature GPS element
antenna in Section 4.2. Specifically, we designed and measured a square proximity-
fed stacked patch (PFSP) antenna excited with coaxial ports of quadrature phase
delay [91]. The resulted element size is 1.7 inches. To reduce the corner couping
between the square elements, we re-designed the proximity-fed stacked patch antenna
using circular aperture (Section 4.3). We also excited this circular PFSP using an
integrated hybrid so that the overall size of this circular GPS element is 1.3 inches [92].
78 Antenna size 1”-1.2” for each GPS element Array aperture size 4.5” with 6 or 7 ports for anti-jamming operation Frequency L5: 1176 MHz, L2: 1227 MHz, L1: 1575 MHz Bandwidth 24 MHz with gain > 0 dB Polarization right-hand circular polarization (RHCP) AR < 5 dB for elevation angle > 40◦ AR < 10 dB for elevation angle > 30◦ Axial ratio (AR) ◦ AR < 15 dB for elevation angle > 20 AR < 20 dB for elevation angle > 10◦ gain > -5.5 dB for elevation angle > 10◦ Beamwidth ◦ gain > -3.5 dB for elevation angle > 40 Feeding single coaxial cable (50 Ω)
Table 4.1: Design specification of the compact GPS antenna system
Further, we designed a 1.2” PFSP GPS antenna excited by a quadrature feeding
network (Section 4.4). This design has less mutual coupling between GPS elements
since the feeding network is optimized and less affects the GPS elements. We also show
the design of a 6-element anti-jamming GPS array using this 1.2” GPS element [93].
The measured performances exhibit good pattern coverage for all three GPS bands.
In Section 4.5, we focus on the design of a polymer-based dual-layer GPS antenna.
Since polymer-ceramic composites have favorable properties (such as light-weight,
flexible and load-bearing), they are well suited for future conformal multilayer GPS
antenna systems.
4.2 Proximity-fed Stacked Patch Antenna with Quadrature Feedings
Several low-profile antennas have been proposed in the literature for dual-band
(the L1 and L2 bands) GPS applications [86, 87, 89]. However, for those designs, the
79 bandwidth at L2 frequency is not broad enough to cover the new L5 band. Another issue is associated with the larger size of the available designs. For example, B.
Rama Rao et. al. [88] presented a 5” 5” (λ/2 λ/2 at L5) LC-integrated GPS × × antenna. To reduce size, we first designed a modified F-shaped antenna [90], utilizing inhomogeneous high permittivity materials, leading to a size of 1.5” 1.5” (λ/7 × × λ/7 at L5). However, this F-shaped antenna employed a rather complex four-port quadrature feeding to realize RHCP, and it relied on soldering the vertical feed probes to the patches, a feature not as desirable for manufacturing reliability.
4.2.1 Design Concept
Here we design a proximity-fed stacked patch antenna to cover all three GPS bands. Since the L2 (1227 MHz) and L5 (1176 MHz) frequencies are close to each other, a single resonant mode centered at 1200 MHz is introduced to cover both
L2 and L5 bands. Another higher resonant mode centered at 1575 MHz is also introduced to cover the L1 band. A stacked patch is well known for such dual-band operation [86, 94]. Existing stacked patch antennas often use internal probes for mode excitation [87, 89] for L1 and L2 GPS operation. Such excitation method is not fabrication friendly and also lacks for future extension to multilayer configuration.
Both designs in [87, 89] reported an aperture size of larger than 2” 2”. Moreover, × their bandwidths of the L2 mode were not broad enough to cover the L5 band.
Therefore, in addition to size reduction, it presents a challenge to also increase the lower mode bandwidth so that the L2 and L5 bands are concurrently covered. Several feeding techniques are studied to excite the stacked patches. Specifically, the direct probe feeding [95] can excite the lower patch strongly. The upper patch will be excited
80 by coupling, thus increasing the bandwidth. However, it is very difficult to cover the
L1 band (1575 MHz) simply by increasing the bandwidth of a small patch (around
λ/8) working at 1200 MHz. Stacked patches can also be excited through coupling from the slot on the ground plane [86, 96]. However, this requires additional feeding circuitry and cavity behind the ground plane. In this work, we excited the stacked patches by coupling from the probes in proximity [97, 98]. In this way, both of the patches can be exited independently. The bandwidth of the lower frequency mode can be achieved by increasing the dielectric layer thickness. To realize right-hand circular polarization (RHCP), two orthogonal modes (TM01 and TM10)were excited within each layer with a 90◦ phase difference via a commercial 90◦ hybrid.
The detailed geometry of the said proximity-fed stacked patch design is shown in Fig. 4.1. It consisted of two metallic patches (with l1 and l2 being their sizes) located on the top of the stacked dielectric substrates (²r1 = 16 being the relative permittivity of the upper substrate and ²r2 = 30 being the corresponding value for the lower substrate). Two probes with inverted-L shapes were connected to the 50
Ω coaxial cables and were strategically positioned at the proximity of patch edge to excite both high and low resonant modes in top and bottom layer, respectively. Other parameters in Fig. 4.1 include lv and lh denoting the vertical and horizontal length of the inverted-L probe. Fig. 4.2 shows the simulated electrical fields at resonant modes (large arrow represents stronger electrical fields). As can be seen, two distinct resonant modes were present in the upper and lower substrate respectively. The proximity-probe coupling also had the advantage of bringing capacitance to balance the inductance associated with the long probe [97]. Detailed parameter effects and study are shown below.
81 30mm
l h 1 ε = 1 r1 16
h2
lv l2 ε = r2 30
lh coaxial port (0º) coaxial port (-90º)
Figure 4.1: Simulation model of the proximity-fed stacked patch antenna with quadra- ture feedings
Lower frequency mode Higher frequency mode
Figure 4.2: Simulated resonant modes of the PFSP GPS antenna with quadrature feedings
82 4.2.2 Design Parameters
A size reduction of the proposed stacked patch antenna down to 1.2” 1.2” (λ/8 × at the L5 band) implies miniaturization by a factor of 3-4 as compared to a nominal
λ/2 size. To achieve this goal, we employed high permittivity dielectric substrates with large thickness. However, care must be exercised to ensure that the bandwidth requirements were satisfied as the antenna size was reduced. We specifically chose the dielectric constant of the lower and upper layers to be ²r2 = 30 and ²r1 = 16, respectively. The ratio of them was kept similar to the ratio of resonant frequencies of the lower and upper modes so that the patch sizes were approximately equal.
This is important to ensure equal excitation of both mode using proximity probes.
A square conducting patch was then placed on top of each dielectric layer. The size of each patch was iteratively tuned using full-wave simulation (Ansoft HFSS) to produce desired resonant frequencies, 1200 MHz and 1575 MHz, at the lower and upper layer, respectively. This is demonstrated in Fig. 4.3(a) where the simulated broadside realized gain curves for top patch size (l1) varying from 20 mm to 26 mm
(with h1 = 6.4 mm, h2 = 6.4 mm, l2 = 20 mm, lh = 5.3 mm, lv = 10 mm, ²r1 = 16 and ²r2 = 30) are plotted. Increasing the patch size (l1) of the top patch decreases higher resonant frequency from 1600 MHz down to 1420 MHz while the lower resonant frequency remains unchanged. Similarly, Fig. 4.3(b) shows the gain as the lower patch size (l2) is changed from 17 mm to 23 mm (with h1 = 6.4 mm, h2 = 6.4 mm, l1 = 23 mm, lh = 5.3 mm, lv = 10 mm, ²r1 = 16 and ²r2 = 30). As expected, a larger l2 value decreases the lower resonant frequency from 1180 MHz down to 1120 MHz. The change in higher resonant frequency is not significantly. Therefore, it is important to first tune the lower patch before adjusting the size of the upper patch. Next,
83 the optimal lengths of the vertical and horizontal segments in the inverted-L probe are also determined via full-wave simulations. Fig. 4.4 illustrates the effect of the horizontal segment length (lh) and the vertical segment length (lv) on the realized gain with h1 = 6.4 mm, h2 = 6.4 mm,l1 = 23 mm, l2 = 20 mm, ²r1 = 16 and ²r2 = 30.
As seen, we can achieve higher gain by properly adjusting the horizontal probe length.
4.2.3 Measurement Results
The final design parameters of the above square PFSP antenna (see Fig. 4.1) were: h1 = 6 mm (upper substrate thickness), h2 = 6 mm (lower substrate thickness), l1 = 22 mm (upper patch size), l2 = 18 mm (lower patch size), lh = 4 mm (horizontal probe length), lv = 10 mm (vertical probe length), ²r1 = 16 (permittivity of the upper substrate) and ²r2 = 30 (permittivity of the lower substrate). A prototype antenna was then fabricated as shown in Fig. 4.5. The fabrication began with machining the dielectric layers to the proper size and thickness. Subsequently, a piece of copper tape was placed on top of the lower substrate to form the lower patch. Epoxy (²r = 3.5 and tanδ = 0.03) was then used to bind the upper and lower layers. Finally, the upper patch was formed and trimmed to tune the resonant frequency. The inverted-L probe were tuned by fixing the vertical arm and adjusting the horizontal one to achieve the desired bandwidth. Then, the horizontal and vertical components of the probe were soldered prior to carrying out the final measurement. This difficulty motivates an improved probe design that contains only the vertical segment (to be discussed in the next section). To achieve RHCP polarization, the PFSP was fed with a commercial packaged 90◦ hybrid.
84 10
0
-10
-20
l =20 mm -30 1 RealizedGain(dB) l =23 mm 1 l =26 mm -40 1
-50 1 1.2 1.4 1.6 1.8 Frequency(GHz)
(a) Adjust top patch size ( l1)
10
0
-10
-20
l =17mm -30 2 RealizedGain(dB) l =20mm 2 l =23mm -40 2
-50 1 1.2 1.4 1.6 1.8 Frequency(GHz)
(b) Adjust bottom patch size ( l2)
Figure 4.3: Tuning the top and bottom patches size of the PFSP
85 10
0
-10
-20
l =3mm h
RealizedGain(dB) -30 l =6mm h l =9mm -40 h
-50 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Frequency(GHz)
(a) Adjust horizontal pin ( lh)
10
0
-10
-20
l =9mm v
RealizedGain(dB) -30 l =10mm v l =11mm -40 v
-50 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Frequency(GHz)
(b) Adjust vertical probe ( lv)
Figure 4.4: Tuning the proximity probe length of the PFSP
86 23 mm 31 mm
ε = 16 r1 6.4 mm 6.4 mm 10 mm 10 mm ε = r 2 30
Figure 4.5: Fabricated proximity-fed stacked patch antenna with quadrature feedings
The simulated and measured S-parameters and broadside gain of the final design are shown in Fig. 4.6 and Fig. 4.7. Close agreement between measurements and simulations is observed. Most importantly, the gain was kept above 2 dB for all bands (L1, L2 and L5 bands). Also, the cross coupling between the two feeding ports
(S12) was less than -10 dB, implying a good CP performance. Fig. 4.8 gives the simulated patterns of RHCP gain and axial ratio of the lower and higher frequency modes. Note that the 3D patterns are plotted in the polar coordinate system with the radial axis denoting elevation angle (0◦ <θ< 80◦) and the angular axis denoting the azimuth angle (φ). As seen, broad pattern coverage of the upper hemisphere was achieved at all GPS frequencies. The axial ratio results were also desirably low in most regions. The axial ratio was increased at low elevation as expected since vertical polarization was dominant due the PEC boundary condition of the ground plane.
87 0 0
-10 -10
-20 -20
S11(dB) -30 -30 S12(dB)
-40 Simulated design -40 Tuned and measured -50 -50 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Frequency (GHz)
Figure 4.6: Measured S-parameters of the proximity-fed stacked patch antenna with quadrature feedings
10
0
-10 L5 L2 -20 L1 RealizedGain(dB) -30 Simulated design Tuned and Measured
-40 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Frequency(GHz)
Figure 4.7: Measured RHCP gain of the proximity-fed stacked patch antenna with quadrature feedings
88 5 10 RHCP: AR:
0 y 5
x -5
0º~80º -10 0
(a) Lower frequency mode (1200 MHz)
5 10 RHCP: AR:
0 5
-5
-10 0
(b) Higher frequency mode (1575 MHz)
Figure 4.8: Simulated pattern of the proximity-fed stacked patch antenna with quadrature feedings
89 4.2.4 Mutual Coupling in GPS Array
We also investigated a compact 7-element GPS array based on the square PFSP
design. As shown in Fig. 4.9, the 7-element square PFSP GPS array consisted of
one reference element at the center and six auxiliary elements around the perimeter.
The seven PFSP antennas were packed within an aperture of 4.6” in diameter. The
distance between some of the corners were inevitably close, giving rise to strong cou-
pling between elements especially for the higher frequency mode. Fig. 4.9 shows the
simulated electric fields in a plane passing through the center of the upper dielectric
layer at the higher frequency mode. As one can see, the electric fields around the
corners were very strong, indicating strong coupling between the two elements. Such
coupling was undesirable since it radiated the opposite sense (LHCP) of polarization
and affected overall radiation patterns. This finding motivates the choice of circular
patch to increase the separation between elements thus reducing mutual coupling.
4.6’’ 4.6’’
(a) The reference element (b) The auxiliary element (center) is excited (perimeter) is excited
Figure 4.9: Simulated electric fields for the 7-element square PFSP array at higher frequency mode
90 4.3 Proximity-fed Stacked Patch Antenna with Branch-line Hybrid Feeding
The previous proximity-fed stacked patch antenna satisfies most design specifi- cations (as listed in Table 4.1) except that it requires two coaxial feeds to generate
RHCP radiation. This feeding scheme is particularly stressing when several elements are used to form an anti-jamming GPS array. That is, for a 7 elements array, 14 feed cables within a 4.6” aperture are required. By designing the element to have a single- port feed, only 7 cables are needed, a significant simplification. Two approaches are typically used in the literature to excite CP modes for patches. One method amounts to exciting two degenerate orthogonal modes (TM01 and TM10 modes) by offsetting the feed position and cropping the patch corners to ensure excitation of the TM01 and
TM10 modes. Using this method, single-fed stacked patch antennas have already been proposed for GPS applications [87, 89]. However, the phase shift is rather sensitive to the operational frequency, leading to narrow CP bandwidth. A second method employs an external 90◦ hybrid [91] or a similar feeding network [86] for CP excita- tion. The feeding network is usually a PCB circuit printed at the back of the ground plane. However, this feed scheme is not attractive as it requires redesign of existing
GPS electronics to accommodate the back plane PCB.
Considering the bandwidth and location requirements, in this section we design a proximity-fed stacked patch antenna with an integrated branch-line hybrid feeding, which is printed on top of the dielectric substrate (see Fig. 4.14) and on the same side as the dual patches. The branch-line hybrid is well understood and widely used in circuit designs [99]. When integrated with the antenna, the hybrid is part of the radi- ating package, and must thus be designed in concert with the dual patches to address
91 the expected coupling issues. Further, as the antenna and hybrid are miniaturized by high dielectric constant materials, the coupling or interference between them is increasing. Therefore, there are compounded challenges in designing a branch-line hybrid that is small, broadband and suitable for compact arrays. Below, we show our design of the circular PFSP with branch-line hybrid feeding step by step.
4.3.1 Circular Stacked Patch Antenna
To address the issue of corner coupling and diffraction associated with the square
PFSP elements, a circular PFSP is designed instead. As shown in previous section, the proximity-fed stacked patch antenna supports two orthogonal modes (TM01 and
TM10) that are combined to deliver RHCP operation. Therefore we began by first designing a linearly-polarized (LP) stacked patch antenna that cover all GPS frequen- cies. The geometry for the single-fed stacked patch is shown in Fig. 4.10. A vertical metal strip was simply used (printed on the outer dielectric layer) to feed the patches.
In preparation for the CP feeding design, we also included a Duroid 6010LM layer
(²r3 = 10.2 and 0.635 mm thick) at the base. This layer will be used to print the integrated 90◦ hybrid. The geometrical parameters found to achieve good gain at all
GPS frequencies are: D = 33 mm, d1 = 26 mm, d2 = 22 mm, h1 = 6 mm, and h2 = 8 mm. Fig. 4.11 gives the gain and return loss (S11) curves. As seen, the gain is 2.5 -
5 dB covering the lower (L5, L2) and higher (L1) GPS bands.
4.3.2 Circular Branch-line Hybrid
Next we proceed to design a branch-line hybrid to achieve equal power splitting with 90◦ phase shift. As shown in Fig. 4.12, the branch-line hybrid contains a four- port ring-shaped microstrip line printed upon a Duroid 6010LM board (²r3 = 10.2,
92 D
d1 ε = r1 16 h1 d 2 ε = r 2 30
h2
Vertical metal strip Duroid 6010LM ε = Coaxial feeding Ground plane r3 10.2
Figure 4.10: Simulation model of the circular PFSP with single probe feeding
5
0
-5 dB -10
-15 Gain S11
-20 1.1 1.2 1.3 1.4 1.5 1.6 frequency (GHz)
Figure 4.11: Simulated return loss and realized gain of the circular PFSP with single probe feeding
93 0.635mm thick). Port 1 is the input port that is connected to a 50 Ω SMA connector
behind the ground plane. Port 2 and port 3 are connected to the proximity probes of
the PFSP with 90◦ phase difference due to additional propagation length to port 3.
A 50 Ω chip resistor is used as the termination port. Such branch-line hybrid is well
understood and widely used in microwave circuit designs. Its working principle can
be explained by even-odd mode analysis [99]. The design parameters here were the
diameter (d) which is related to the center frequency of the hybrid and the widths (w1
and w2) which control the impedance matching. The center frequency was chosen to
be 1350 MHz (around the center of 1200 MHz and 1575 MHz) for dual band operation.
Therefore, the diameter (d) was adjusted according to:
λ πd = λg = (4.1) √²eff where λg is the wavelength in the PCB. The branch-line hybrid was connected to a
50 Ω coaxial cable and a surface mount chip resistor was employed as the termination
for the hybrid. Therefore, the widths w1 and w2 were chosen properly to ensure
the impedance of the microstrip line to be 50 Ω and 50/√2 = 36.36 Ω, respectively.
The design parameters were chosen to be as follows: d = 9.144 mm, w1 = 0.381
mm and w2 = 0.762 mm. Another important thing to notice is that the hybrid is
placed under the patch for maximum compactness of the antenna element. The PCB
thickness is kept very thin (0.635 mm) to reduce the coupling between the hybrid and
patches. The measured S-parameters of our circular branch-line hybrid compared
with the simulated results are given in Fig. 4.13. As seen, the bandwidth of our
circular hybrid design could cover all GPS frequencies (S11 < -10 dB) with a phase
delay between port 2 to port 3 being 85.8◦ at 1200 MHz and 100◦ at 1575 MHz, − − respectively.
94 33 mm Termination (chip resistor) w1
w2 Port 3
d
Port 1
w2 w1 Port 2
Figure 4.12: Circular branch-line hybrid for RHCP excitation
0
−5
−10
−15 dB
Simulated S11 −20 Simulated S21 Simulated S31 Measured S11 −25 Measured S21 Measured S31 −30 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz)
Figure 4.13: Measured S-parameters of the circular branch-line hybrid
95 4.3.3 Measurement Results
The simulation model and the fabricated circular PFSP with integrated branch- line hybrid are shown in Fig. 4.14. It had a very similar configuration and the same working principle as the previous square PFSP antenna. Instead of the L-shaped probes, two vertical copper strips were printed on the side of the substrates to excite the patches via field coupling. To securely bind the two different dielectric layers together without any low dielectric gaps in our prototype, a commercial dielectric
@ paste (ECCOSTOCK ) with a dielectric constant of ²r = 15 was used. Referring to
Fig. 4.14, the final design parameters were: D = 33 mm (overall aperture diameter), d1 = 26.6 mm (upper patch diameter), d2 = 24 mm (lower patch diameter), h1 = 6 mm (upper layer thickness), h2 = 10 mm (lower layer thickness). The simulated and measured return loss and gain performance are shown in Fig. 4.15 and Fig. 4.16, respectively. Slight shifts in resonant modes and unbalanced between the two linear polarization modes were observed and caused undesired increase in cross-polarization levels. These deviations are probably caused by non-uniform thickness of the dielectric layer and accuracy of the vertical strips (probes). As shown in Fig. 4.16, the fabricated
PFSP prototype antenna sufficiently covered all three GPS bands with RHCP gain greater than 0 dB and cross-polarization isolation greater than 10 dB.
The simulated radiation patterns of the circular PFSP antenna with integrated hybrid feeding (over an infinite ground plane) is also shown in Fig. 4.17. Both RHCP gain and axial ratio (AR) are plotted for the lower (1200 MHz) and higher (1575 MHz) frequency modes. Table. 4.2 compares the averaged axial ratio at different elevation angles with the specified values. These predicted gain levels and axial ratios were obvious sufficient for GPS operations. As can be seen, the RHCP gain was above -5.5
96 Figure 4.14: Simulated model and fabricated PFSP with branch-line hybrid feeding
0
-2
-4
-6
-8
-10
S11(dB) -12
-14 L5 L2 Measured -16 Simulated -18 L1
-20 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz)
Figure 4.15: Measured return loss of the PFSP with branch-line hybrid feeding
97 5
0
-5
-10
-15
Gain(dB) -20 L5 -25 L2 Simulated RHCP L1 Simulated LHCP -30 Measured RHCP Measured LHCP -35 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz)
Figure 4.16: Measured gain of the PFSP with branch-line hybrid feeding
dB for more than 85% of the upper hemisphere (θ from 0◦ to 80◦) and the axial ratio performance also satisfied the specified requirements.
4.3.4 Mutual Coupling in GPS Array
We then investigated the performance of a compact PFSP array with integrated hybrid feedings. The initial test configuration was a 4.8” 7-element array as shown in Fig. 4.18. In this case, the center element was excited with all the other elements terminated. The RHCP gain and axial ratio patterns are shown in Fig. 4.19. De- teriorations of pattern coverage and axial ratio performance were observed in this array configuration compared to the stand-alone element configuration due to strong mutual coupling. Further investigation revealed that strong coupling was caused by the outward propagating waves under each proximity probe. An improved design
98 5 20 RHCP: AR:
0 y 10
x -5
-10 0 0º~90º (a) Lower frequency mode (1200 MHz)
5 20 RHCP: AR:
0
10
-5
-10 0 (b) Higher frequency mode (1575 MHz)
Figure 4.17: Simulated gain and axial ratio pattern of the PFSP with branch-line hybrid feeding
Angle(θ) AR (dB): 1.2 GHz AR (dB): 1.575 GHz Specified AR (dB) 50◦ 4.45 3.80 < 5 60◦ 6.00 5.87 < 10 70◦ 8.97 9.10 < 15 80◦ 14.68 14.96 < 20
Table 4.2: Average axial ratio at different elevation angles for the PFSP with hybrid feeding
99 could utilize an external hybrid configuration as illustrated in Fig. 4.20(b) where the
feeding probes radiate toward the center of the patch instead. As shown in Fig. 4.20,
electrical fields of an adjacent element due to mutual coupling (at the higher fre-
quency mode) is significantly reduced after we modify the hybrid configuration. This
result motivates us to design a feeding network outside the PFSP to achieve RHCP
excitation while suppressing the mutual coupling, which is shown in the next section.
4.8 ’’
center element excited
Figure 4.18: Configuration of the 7-element PFPS array with integrated hybrid feed- ings
4.4 Proximity-fed Stacked Patch Antenna with Quadrature- phase Splitter
The previous external branch-line hybrid design requires a microstrip line to go around the patch, thus increasing antenna footprint. This becomes the limiting factor for the minimum size of the compact GPS array. In addition, isolation among the ports in the branch-line hybrid is not great. As discussed before, one of the major
100 1176 MHz 1227 MHz 1575 MHz RHCP: 5
0
-5
-10
AR: 20
10
0
Figure 4.19: Simulated pattern of the 7-element PFPS array with integrated hybrid feedings
challenges to design a compact anti-jamming GPS array is to reduce the mutual couplings between elements. Examining the resonant fields in previous arrays revealed that the center element played key contribution in mutual coupling. In this section, we design a PFSP antenna with quadrature-phase splitter to realize a compact 6-element anti-jamming GPS array as shown in Fig. 4.21. To achieve a trade-off between size and anti-jamming performance, we removed the center element and relied only on the 6 circumferential radiators for null steering and beamforming. Instead, the center region of the array was used to accommodate the feeding circuits. As shown in
Fig. 4.21, the circular PFSP is similar to the stacked patch design in the previous section. Below, we discuss the design of the quadrature-phase splitter and the anti- jamming GPS array.
101 adjacent element excited element excited by coupling
Diffraction at the junction (outward) leading to strong coupling (a) Circular PFSP with integrated hybrid feeding
adjacent element excited element excited by coupling
termination
Coaxial feed Diffraction at the junction (inward) exciting the patches (b) Circular PFSP with external hybrid feeding
Figure 4.20: Modify the branch-line hybrid to reduce mutual coupling
Quadrature 1.2’’ feeding network d 1 ε = 4.5’’ r1 16 h d2 1 ε = r 2 30 h2 feed
Figure 4.21: PFSP with quadrature-phase splitter for 6-element anti-jamming GPS array
102 4.4.1 Quadrature-phase Splitter
The center region of the 6-element GPS array shown in Fig. 4.21 contains six
quadrature-phase splitters based on the Wilkinson power divider [99]. The two output
ports of the Wilkinson divider are connected to two microstrip lines with different
lengths chosen to provide the desired 90◦ phase difference. The Wilkinson power
divider is widely used in microwave circuits for its merits of equal splitting, phase
balance and output port isolation. A more detail configuration of the quadrature-
phase splitter is shown in Fig. 4.22. The design parameters include r1, l1, θ1, w1 and
w2. For a Wilkinson divider, l1 is chosen such that:
l1 √²eff = λ0/4 (4.2) ·
where ²eff is the effective dielectric constant of the PCB (Duroid 6010LM, ²r = 10.2,
0.635 mm thick) and λ0 is the free space wavelength at center frequency (1.3 GHz).
The angle θ1 was adjusted so that:
[r1(π/2 θ1) r1θ1] √²eff = λ0/4 (4.3) − − ·
to achieve quadrature phase delay between the outputs (port 2 and 3). Finally the
widths of the traces, w1 and w2, were tuned for impedance matching. The corre-
sponding characteristic impedances were 50 Ω with w2 = 0.58 mm and 50√2 = 70.7
Ω with w1 = 0.25 mm. The resulting parameters were: r1 = 18 mm, l1 = 21.5 mm
◦ and θ1 = 31 .
The measured performances of the quadrature phase splitter are given in Fig. 4.23 and Fig. 4.24. As seen, the amplitude of port 2 and 3 differs only by about approx- imately 0.2 dB and achieved an excellent insertion loss of about -3 dB. The return loss and the isolation were also measured to be below -15 dB for all three GPS bands.
103 100 w1 Port 2 l1 1
Port 1 r1 Port 3
Wilkinson power divider w2 delay line
Figure 4.22: Simulation model and parameters of the quadrature-phase splitter
As shown in Fig. 4.24, the phase delay between port 2 and 3 was around 80◦ at the lower mode (L5/L2) and 100◦ at the higher frequency mode (L1).
4.4.2 Measurement Results
As shown in the inset of Fig. 4.25, the whole dielectric puck consisting of two di- electric layers, two conducting patches, two conducting strips as the proximity probes and a bottom conducting lamination was fabricated by PicoFarad, a major vendor for dielectric pucks used in dielectric resonance oscillators and chip capacitors. The fabricated dielectric pucks were tested and exhibited good accuracy (+/- 0.3 mm) and stability. Our measurement on the Picofarad puck showed the dielectric con- stants of 28.5 and 15.8, which were very close to desired ²r2 = 30 and ²r1 = 16. The dielectric loss tangent was found to be less than 0.002. The small deviation in dielec- tric constant was easily compensated by slight adjustment in patch sizes. We have tested several stacked patches and finally adjusted the parameters to be (referring to
104 0
−5 Return loss (S11) Isolation (S23) −10 Insertion loss (S21) Insertion loss (S31) −15 dB
−20
−25
−30 1 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz)
Figure 4.23: Measured S-parameters of the quadrature-phase splitter
105
100
95
90
85
degree 80
75 Measured phase delay 70 Simulated phase delay 65
60 1 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz)
Figure 4.24: Measured phase delay (between port 2 and port 3) of the quadrature- phase splitter
105 Fig. 4.21): D = 30 mm (overall aperture diameter), d1 = 26.4 mm (upper patch di- ameter), d2 = 23.6 mm (lower patch diameter), h1 = 7.5 mm (upper layer thickness) and h2 = 8.5 mm (lower layer thickness).
Element #3 Element #2
4.5’’
Element #1
Element #4 y
Element #5 Element #6
x
Figure 4.25: Configuration of the PFSP and 6-element GPS array with quadrature- phase splitter
The final PFSP element with quadrature-phase splitter was measured on an 8” ground plane. The measured return loss and broadside gain of the quadrature phase splitter-fed stacked patch element are shown in Fig. 4.26 and Fig. 4.27. As seen, the measurements are in good agreement with the simulations. More importantly, both lower and higher mode achieved an RHCP gain above 0 dB with cross-pol below -15 dB. We also measured the PFSP element on a 10’ aluminum plate to emulate the working platform of an infinitely large ground plane. The diffractions from the aluminum plate edges were gated in time domain to get rid of the edge diffraction which affects the radiation patterns. The full radiation patterns of the
106 final PFSP element were measured at 8◦ elevation angle increments and 30◦ azimuth angle increments and is plotted in Fig. 4.28. This pattern exhibits approximately
85% angle coverageof the upper hemisphere (elevation angle > 10◦) with RHCP gain larger than -5.5 dB, satisfying the operation requirements. Table. 4.3 gives the averaged axial ratio at different elevation angles compared with the specifications, which indicates that the final PFSP element produces desirable RHCP performance.
0
-5 Measured Simulated -10
-15 S11(dB) -20
-25
-30 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6 Frequency (GHz)
Figure 4.26: Measured return loss of the PFSP with quadrature-phase splitter
4.4.3 Compact 6-element Anti-jamming GPS Array
The configuration of the 6-element anti-jamming GPS array is shown in Fig. 4.25.
Each GPS element is a PFSP antenna of 1.2 inches in diameter. The final array has a diameter of 4.5 inches. Notice that the six elements are arranged along the array
107 10 Measured RHCP 5 Measured LHCP Simulated RHCP 0 Simulated LHCP -5
-10
Gain Gain (dB) -15 L1 L5 -20
-25 L2 -30 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6 Frequency (GHz)
Figure 4.27: Measured gain of the PFSP with quadrature-phase splitter
5 20 RHCP: AR:
y 0 x 10 0º~90º -5
-10 0
(a) Lower frequency mode (1200 MHz)
AR: RHCP: 5 20
0 10
-5
-10 0 (b) Higher frequency mode (1575 MHz)
Figure 4.28: Measured gain and axial ratio pattern of the PFSP with quadrature- phase splitter
108 Angle(θ) AR (dB): 1.2 GHz AR (dB): 1.575 GHz Specified AR (dB) 50◦ 4.6 2.9 < 5 60◦ 5.4 4.1 < 10 70◦ 8.8 8.1 < 15 80◦ 11.6 11.7 < 20
Table 4.3: Average axial ratio at different elevation angles for the PFSP with quadrature-phase splitter
circumference to maximize inter-element spacing and reduce coupling. The optimal
element orientation was determined to reduce mutual coupling by maximize the sep-
aration between probes from adjacent elements. Under anti-jamming operations each
GPS element receives signals in presence of the other elements [100]. To achieve high
anti-jamming agility, it is desirable to have broad pattern coverage for each GPS
element in the array setting. For evaluation, element #1 (Fig. 4.25) was excited
in presence of the other elements terminated with matching loads. The simulated
RHCP gain and axial ratio pattern (on an infinite ground) are shown in Fig. 4.29.
The red numbers in parenthesis denote specified minimum RHCP gain levels at dif-
ferent elevation angles. As one can see, the average RHCP gain levels exceed these
minimum values. The simulated axial ratio at different elevation angles are also com-
pared with specifications in Table 4.4. It is clear that this design meets almost all of
these requirements, especially when close to zenith direction.
To measure the GPS array on an infinitely large ground plane a new measurement setup was developed at OSU-ESL (depicted in Fig. 4.30). The antenna under test
(AUT) was mounted at the center of a 10 feet ground plane. A probe antenna was positioned 4 feet away from the AUT, measuring its far-field responses (2D2/λ = 5”).
109 -2.4dB (-3.5) -5.2dB (-5.5) RHCP: 5 AR: 20
0
y 10
x -5
0º~90º -10 0 (a) L5: 1176 MHz
-1.0 dB (-3.5)
-3.8 dB (-5.5) RHCP: 5 20 AR:
0 10
-5
0 -10
(b) L2: 1227 MHz
0.9 dB (-3.5) -1.9 dB (-5.5) RHCP: 5 AR: 20
0 10
-5
-10 0 (c) L1: 1575 MHz
Figure 4.29: Simulated RHCP and axial ratio patterns for the GPS element in array setting
110 Angle(θ) AR (dB): L5 AR (dB): L2 AR (dB): L1 Specified AR (dB) 50◦ 6.5 4.2 5.7 < 5 60◦ 8.0 6.2 7.7 < 10 70◦ 10.6 9.3 11.0 < 15 80◦ 16.0 15.1 17.0 < 20
Table 4.4: Average axial ratio at different elevation angles for the PFSP element in 6-element GPS array
By gating in time domain, the contributions of diffraction from the finite ground plane edges were suppressed to avoid pattern contamination, thus emulating an infinitely large ground plane.
10 feet AUT
Figure 4.30: New setup for measuring the GPS array on an infinitely large ground plane
Using the above setup the measured patterns of the GPS element #1 in array setting are shown in Fig. 4.31 for different φ cuts. The measured patterns (curves) are also compared to the simulations (dots). In all cases, we observed broad pattern
111 coverage and fairly good agreements between measurements and simulations. The
higher resonant mode shifted from 1.575 GHz to 1.6 GHz and exhibited some differ-
ences with the simulations. The shift of higher mode frequency can be recovered by
tuning the patch size accordingly. The performances of the other elements were very
consistent to those of element #1.
4.5 Polymer-based GPS Antenna with CNT Printing
In previous sections, we present the design of a miniature tri-band GPS antenna for anti-jamming GPS operation. The key feature of this antenna is the dual-layer dielectric configuration which supports two resonant modes, covering all three GPS bands. High contrast ceramic substrates are employed to support the microstrip patches and achieve miniaturization. However, the disadvantages of using ceramic pucks are obvious. First, ceramics are brittle and very difficult to machine. If we want to modify the thickness or diameter of the dielectric substrates, we have to machine a new tooling setup which is very expensive. Secondly, the ceramics are not easily integrated into the whole antenna systems. We must use high-strength epoxy to glue the ceramic substrates onto the ground plane. And we must pay additional attention to ensure alignment between different ceramic substrates when gluing them together. Thirdly, ceramics are heavy and easily broken due to vibration or mechanic strength. To address these issues, we then employ polymer-ceramic composites as the dielectric substrates for the stacked patch configuration. As discussed in Chapter 2, polymer-ceramic composites have the advantage of being light-weight, conformal and load-bearing. They are very attractive for future conformal antenna systems such as the compact anti-jamming GPS array. Further, as the polymer composites have a
112 φ 5 5 5
0 0 0
-5 -5 -5
-10 -10 -10
-15 -15 -15
-20 Simulated RHCP -20 Simulated RHCP -20 Simulated RHCP Simulated LHCP Simulated LHCP Simulated LHCP Measured RHCP Measured RHCP Measured RHCP -25 -25 =30º -25 =60º Measured LHCP =0º Measured LHCP Measured LHCP
-30 -30 -30 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 θ (degree) θ (degree) θ (degree) φ 5 5 5
0 0 0
-5 -5 -5
-10 -10 -10 dB -15 -15 -15
Simulated RHCP Simulated RHCP -20 Simulated RHCP -20 -20 Simulated LHCP Simulated LHCP Simulated LHCP Measured RHCP =90º Measured RHCP Measured RHCP Measured LHCP -25 Measured LHCP -25 =120º Measured LHCP -25 =150º
-30 -30 -30 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 θ (degree) θ (degree) θ (degree)
(a) Lower frequency mode (1.2 GHz)
φ 5 5 5
0 0 0
-5 -5 -5
-10 -10 =30º -10
-15 -15 -15
Simulated RHCP Simulated RHCP -20 -20 Simulated LHCP -20 Simulated RHCP Simulated LHCP Measured RHCP Simulated LHCP Measured RHCP -25 Measured LHCP Measured RHCP -25 =0º Measured LHCP -25 =60º Measured LHCP
-30 -30 -80 -60 -40 -20 0 20 40 60 80 -30 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 θ (degree) θ (degree) θ (degree) φ 5 5 5
0 0 0
-5 -5 -5
-10 =90º -10 =120º -10 dB
-15 -15 -15
Simulated RHCP -20 -20 -20 Simulated LHCP Simulated RHCP Simulated RHCP Measured RHCP Simulated LHCP Simulated LHCP -25 -25 -25 Measured LHCP Measured RHCP Measured RHCP =150º Measured LHCP Measured LHCP -30 -30 -30 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 θ (degree) θ (degree) θ (degree) (b) Higher frequency mode (1.575 GHz)
Figure 4.31: Measured radiation patterns of the GPS element in the 6-element array setting
113 viscous pre-cured liquid form, they also allow for 3-D fabrication of more complicated antenna configuration. In this section, we demonstrate the design of a polymer-based
GPS antenna which utilizes polymer-ceramic composites as dielectric substrates and carbon nanotube sheet as antenna metalization.
4.5.1 Fabrication Process
The simulation model of the polymer-based dual-layer GPS antenna is shown in
Fig. 4.32. The stacked patch configuration is very similar to the circular PFSP with integrated hybrid feeding (Fig. 4.14), except that the probes are bent inside the lower dielectric substrate. This bending introduces another diffraction point in the lower substrate, which enhances the coupling to the lower patch (1.2 GHz). Therefore, we can achieve better impedance matching for the stacked patches using the new feeding probes. Referring to Fig. 4.32, the design parameters are optimized to be: h1 = 8 mm (upper layer thickness), h2 = 11 mm (lower layer thickness), d1 = 44 mm (upper patch diameter), d2 = 40 mm (lower patch diameter), l1 = 6 mm, l2 = 12 mm and l3 = 10 mm. As the probes are not straight and placed inside the dielectric substrate,
3-D fabrication is required to realize this dual-layer GPS antenna, which is easily achieved by using polymer-ceramic composites.
The fabrication process to realize this polymer-based dual-layer GPS antenna is displayed in Fig. 4.33. We first printed a 90◦ hybrid and soldered the coaxial connector, chip resistor and feeding probes onto the PCB. Then, we built a mold
(2” in diameter) so that we could pour the 20% PDMS-D270 mixture(²r2 = 10) into the mold as the lower substrate. After pouring the lower layer, we placed the conductive carbon nanotube sheet (fabricated by the same process in Chapter 3 with
114 2’’ ε = r1 6
d1 h ε = 1 r 2 10
d l3 2 l2 h2 l1
Diffraction points Feeding port 90º hybrid
Figure 4.32: Simulation model of the polymer-based dual-layer GPS antenna
sheet resistance of 1 Ω/¤) on the top to make the lower patch. It took around 1 day for the lower mixture to cure. The resulted carbon nanotube sheet had very strong adheasion to the polymer substrate after curing. Subsequently, we poured the 15%
PDMS-MCT mixture (²r1 = 6) to make the upper substrate. And another carbon nanotube sheet was placed on the top as the upper patch. Finally, after curing we achieved the polymer-based dual-layer GPS antenna as shown in Fig. 4.33.
4.5.2 Measurement Results
As shown in Fig. 4.34, the measurement results exhibited broad gain coverage for all three GPS bands. More importantly, the gain at lower frequency mode (L5 and L2 bands) was around 4 dB as a result of improved impedance matching due to the bent probes. The cross-pol gain (LHCP) is 15 dB lower than the RHCP gain, indicating good axial ratio of the RHCP radiation. Compared with the previous
PFSP fabricated by ceramic substrates, this polymer-based GPS antenna has higher gain, higher efficiency, and larger bandwidth while enjoying favorable properties such as light weight and robustness.
115 (a) Feeding network (b) Mold for polymer substrates
(d) Attach CNT sheet as (c) Pour polymer mixture the lower patch for the lower substrate
(e) Pour polymer mixture (f) Attach CNT sheet as for the upper substrate the upper patch
(g) After curing
Figure 4.33: Fabrication process of the polymer-based dual-layer GPS antenna
116 0
-5
-10
-15 S11 (dB) S11
-20
-25 Measured Simulated
-30 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz) (a) Return loss (S11)
10
5
0 L5 -5 L2 -10
Gain (dB) Gain -15
-20 Measured RHCP Measured LHCP L1 -25 Simulated RHCP Simulated LHCP -30 1.1 1.2 1.3 1.4 1.5 1.6 Frequency (GHz) (b) Realized gain
Figure 4.34: Measured performance of the polymer-based dual-layer GPS antenna
117 4.6 Summary
In this chapter, we developed a compact GPS antenna system for anti-jamming applications. This 4.5” GPS array achieved a size reduction of nearly 90% as com- pared to a typical 14” aperture. Specifically, we utilize the dual-layer dielectric con-
firgruation to realize the stacked patch antenna, covering all three GPS bands. A key feature of this stacked patch antenna is its proximity probe feeding, which miti- gates the large inductance by means of capacitive coupling. The proximity probes are also fabrication-friendly since no drilling or soldering is necessary. Different feeding schemes for the GPS array have been studied to realize RHCP polarization while suppressing the mutual coupling between the GPS eleemnts. Specifically, we have presented three GPS antennas excited with an external 90◦ hybrid, an integrated branch-line hybrid and a quadrature-phase splitter. Measurement results show that the 6-element GPS array with the quadrature-phase splitters has minimal mutual coupling, thus exhibiting excellent performances above specified requirements.
We also demonstrate the application of polymer-ceramic composites in this com- pact GPS antenna systems. Polymer-ceramic composites are light-weight, flexible, load-bearing and moldable. Therefore, they are very desirable for future compact multilayer antenna systems with high system compatibility and integrity. We partic- ularly show the fabrication process of a polymer-CNT stacked patch GPS antenna.
Conductive carbon nanotube sheets are utilized as antenna metalization. The mea- sured performances exhibit very good gain coverage due to the 3D probes embedded inside the polymer substrates. As seen, polymer-ceramic composites also allow for
3-D fabrication of more complicated antenna and feeding configuration.
118 The design concept of this miniature GPS antenna can certainly be extended for other multi-frequency CP/LP antennas. Further, the approach of stacked patches can also be extended for multilayer configuration, which has potential to provide ultra wideband (UWB) coverage. Future work includes the design and fabrication of multilayer polymer-CNT patches and further antenna miniaturization by introducing artificial inductive loading.
119 CHAPTER 5
CYLINDRICALLY CONFORMAL POLYMER-BASED MICROSTRIP ARRAY
5.1 Introduction
The increasing demand for high-performance, highly compact and portable devices along with easy mass production features like roll-to-roll processing has spurred the growth of flexible electronics, flexible interconnects, flexible circuit boards and flexible chip packaging. The current technology to realize such flexible electronics is to mount electronic devices on flexible plastic substrates, such as polyester, polyimide and
PEEK (polyaryletheretherketone) films [101, 102]. However, the flexibility usually compromises as the thickness of these polymer films increase. On the other hand, polymer-ceramic composites in our work are not only flexible but also stretchable even with several mm thickness. Another advantage of our polymer-ceramic composites compared with the common polymer films is that we can easily modify their electrical properties (permittivity and dielectric loss) to reduce the size of the microwave circuits printed on them. Further, polymer-ceramic composites have a viscous pre-cured liquid form, which easily allow for integration of IC components and 3-D fabrication
120 of multilayer circuits. Given these advantages, polymer-ceramic composites are well suited for truly flexible and stretchable electronics.
In this chapter, we present the design of a cylindrically conformal microstrip array as an example of flexible electronics. Cylindrically conformal microstrip array has lots of practical applications such as missile navigation, air traffic control and so on [103, 104, 105]. Flexible electronics are very desirable for cylindrical microstrip array since they can be wrapped around the mounting platform without redesigning the antenna and feeding networks. In Section 5.2, we first design the transmitting and receiving antenna for the conformal cylindrical microstrip array. Specifically, a series- fed multi-patch antenna is proposed to reduce the radiation from feeding network.
In Section 5.3, we focus on the design of the conformal cylindrical microstrip array to achieve an azimuth pattern coverage with +/- 1 dB ripple. In Section 5.4 we discuss the fabrication process of such conformal cylindrical array. As a prototype design, we fabricated the cylindrical array using discrete PCB elements glued onto the cylinder surface. Fabrication challenges and tuning techniques are demonstrated.
For final fabrication and practical applications, we propose to use polymer-CNT sheet wrapped around the cylinder to realize the cylindrical microstrip array (Section 5.5).
The main objective of this cylindrical array design is to develop a pair of transmit- ting and receiving antennas operating in any 0.5 GHz band between 8.5 and 10.5 GHz to be installed within a cylindrical radome. The physical and electrical specifications are given in Table 5.1 and Table 5.2. Below, we show the design procedures in details.
121 Mounting platform metal cylinder (6” in height, 6.85” in diameter) Frequency 8.5 - 10.5 GHz Bandwidth 300 MHz with VSWR < 1.5:1 Gain To be maximized by the aperture size Elevation pattern Determined by available vertical aperture Azimuth pattern Omni-direction with +/- 1 dB
Table 5.1: Design specification of the X-band transmitting antenna
Number 6 Mounting platform metal cylinder (6” in height, 6.85” in diameter) Frequency 8.5 - 10.5 GHz Bandwidth 300 MHz with VSWR < 1.5:1 Gain 7 - 8 dB higher than transmitting antenna Elevation pattern Determined by available vertical aperture Azimuth pattern 65◦ - 70◦ at -3 dB points
Table 5.2: Design specification of the X-band receiving antenna
5.2 Transmitting and Receiving Antenna Designs
Microstrip patch antennas mounted on cylindrical surfaces have been extensively studied, particularly for aerospace applications [103, 104, 105]. To provide omni- directional azimuth pattern for applications like missile navigation and air traffic control, two approaches are usually adopted. The first one is to wrap a microstrip ra- diator around the cylinder surface and feed it at uniformly distributed locations [103].
The other approach is to arrange discrete antennas along the circumference of the cylinder and excite them with equal amplitude and phase [104, 105], as shown in
Fig. 5.1. However, both approaches rely on the corporate feeding along with the
122 Wilkinson power dividers. For high frequency applications (X-band and above), this feeding scheme tend to suffers from ohmic loss and radiation from the feeding net- work [106]. In this chapter, we present a cylindrical array system design (see Fig. 5.1) that includes a transmitting and a receiving array. Both arrays are formed by 5- patch antenna elements. The 5-patch antennas are fed in series configuration from a single coaxial probe located at the center patch (see Fig. 5.2(a)). Such series feed- ing scheme significantly reduces the overall dissipation loss and line radiation of the feeding network [107, 108, 109].
Receiving array pattern: Receiving array 6’’ Transmitting array Transmitting pattern: array
(a) Conformal cylindrical microstrip array (b) Desired pattern for the transmitting and receiving arrays
Figure 5.1: Cylindrically conformal microstrip array
Fig. 5.2(a) shows the configuration of the proposed series-fed 5-patch microstrip antenna. The center patch is fed by a coaxial probe and other patches are excited interconnected microstrip lines whose lengths and widths are strategically designed to realize equal phase excitation. Compared to the traditional corporate feeding with microstrip lines and power splitters, the proposed series feeding scheme has the advantage of less fabrication difficulty and less pattern interference from the feeding
123 network. In order to feed each patch in phase, the spacing distance (d) was carefully chosen to be one-wavelength in the dielectric substrate (FR4, ²r = 4.4):
d = λ0/√²r (5.1)
where λ0 is the free space wavelength. On the other hand, the spacing distance should
also be kept less than half-wavelength in free space to avoid grating lobes:
d λ0/2 (5.2) ≤
In our design, FR4 PCB board was chosen for its substrate dielectric constant of
²r = 4.4 and low fabrication cost. This results in:
d = 2a (5.3)
The widths of the interconnecting lines (w1 and w2) were adjusted so that the center
patch had more power weighting and the side patches had less. Notches (ws wide and ls long) were also made along the sides of the center patch for better impedance matching. Fig. 5.2(b) also shows the electric fields at resonant frequency. As seen, each microstrip patch was excited in phase and the center patch has the strongest power weighting.
Pattern Beamwidth Control
In order to achieve minimum beamwidth in the plane containing the axis of the cylinder and the sub-array, a larger array aperture is needed. As shown in the inset of
Fig. 5.3, we model the series-fed 3-patch, 5-patch and 7-patch antenna respectively.
Referring to Fig. 5.2, the design parameters were chosen as follows: a = b = 6.8 mm, d = 13.6 mm, ls = 1 mm, ws = 1.48 mm, w1 = 0.74 mm, and w3/w2/w1= α = 0.5.
124 PCB: r = 4.5, d d a 1.5875 mm thick l b s ws w2 w1 Probe feeding
(a) Series-fed 5-patch antenna
E-fields: 8.75 6.25 3.75 1.25 10 7.5 5.0 2.5 0.0 ×10 3 V/m
(b) Electric fields at resonant frequency
Figure 5.2: Simulation model of the series-fed 5-patch antenna
The return loss and the gain performance of the 3-patch, 5-patch and 7-patch antenna are shown in Fig. 5.3 and Fig. 5.4. As can be seen, with more patch elements, the antenna has higher gain (12.4 dB for the 3-patch antenna, 14.6 dB for the 5-patch antenna and 15.7 dB for the 7-patch antenna). The radiation patterns of E-plane and H-plane are also shown in Fig. 5.5, which shows that the half power beamwidth of E-plane is 37.4◦, 23.7◦ and 17.8◦ for the 3-patch, 5-patch and 7-patch antenna, respectively. We also notice that the multi-patch antenna has a broad pattern (3 dB beamwidth > 80◦) in the H-plane, thus satisfying the azimuth pattern requirement of the receiving antenna. Given the 6 inches vertical aperture size, we chose the 5-patch antenna as the transmitting and receiving elements.
125 0
-5
-10
-15 34 mm S11(dB) -20 61.2 mm 3-patch -25 5-patch 7-patch 88.4 mm -30 9.5 9.75 10 10.25 10.5 Frequency (GHz)
Figure 5.3: Simulated return loss of the series-fed multi-patch antenna
20
15
10 Realized gain (dB) 5 3−patch 5−patch 7−patch
0 9.5 9.75 10 10.25 10.5 Frequency (GHz)
Figure 5.4: Simulated gain of the series-fed multi-patch antenna
126 20 3-patch 5-patch 10 7-patch
0
-10 Realizedgain(dB) -20
-30 -90 -60 -30 0 30 60 90 θ (degree) (a) E-plane pattern
20
15
10
5
0
Realizedgain(dB) -5 3-patch 5-patch -10 7-patch
-15 -90 -60 -30 0 30 60 90 θ (degree) (b) H-plane pattern
Figure 5.5: Simulated pattern of the series-fed multi-patch antenna
127 Sidelobe Level Reduction
In most applications, including this work, radiation of sidelobes are undesired and need to be suppressed. A common way to suppress the sidelobe level is to taper the array elements [110], which can be easily implemented in this series-fed microstrip patch antenna as demonstrated in Fig. 5.6. Fig. 5.6(a) shows the E-plane pattern
i−1 of the 5-element array with polynomial microstrip line tapering (wi = α w1, with
α = 0.5, 1, 1.5 and 2). We observed that the second sidelobe level was reduced
with more tapering (smaller α), which indicated that the microstrip line tapering
improved the impedance matching between the microstrip line and patches, thus
reducing the end reflection. We can also taper the patch elements by adjusting
their widths. As shown in Fig. 5.6(b), we simulated the 5-element patch with linear
tapering (b1/b2/b3 = 4/2/1) and binomial tapering (b1/b2/b3 = 6/4/1). Compared to
the uniform 5-element array, their first sidelobes disappeared because of the tapering
effect. However, the tapered patch width also affected the impedance matching, which
led to strong end reflection. In order to further reduce the sidelobe level, we may apply
resistive termination or cascade more patch elements to reduce the end reflection. The
final series-fed 5-patch antenna has a uniform patch width of a = b = 6.8 mm and
interconnecting line width tapering of α = 0.5.
5.3 Omni-directional Array Design
In order to achieve an omni-directional azimuth pattern, the transmitting antenna
must include several elements evenly allocated along the circumference of the PEC
cylinder. As shown in Fig. 5.1, each element covers certain part of the azimuth plane
so that the overall transmitting array has an omni-directional azimuth pattern. In
128 20 α = 0.5 α = 1 10 α = 1.5 α = 2 0
-10 Realizedgain(dB) -20
-30 -90 -60 -30 0 30 60 90 θ (degree) (a) Interconnecting microstrip line width tapering
20
10
0
-10 Realizedgain(dB) Uniform -20 Linear Binomial
-30 -90 -60 -30 0 30 60 90 θ (degree) (b) Patch width tapering
Figure 5.6: Sidelobe suppression of the series-fed multi-patch antenna
129 other words, an array of transmitting elements has to be rolled up to deliver the
omni-directional azimuth pattern. To study the minimum number of patch elements
required for omni-directional azimuth pattern, we modeled a simple microstrip patch
antenna mounted on a PEC cylinder (see Fig. 5.7). The pattern of a N-element
cylindrical array can be synthesized by: N N ~ ~ ~ 2π E(φ, θ)= X Ei(φ, θ)= X E1[φ (i 1) , θ] (5.4) − − N i=1 i=1
where E~ (φ, θ) and E~1(φ, θ) are the far-field patterns of the N-element cylindrical array
and the single patch, respectively. The concept of this array pattern synthesis is also
shown in Fig. 5.8. Fig. 5.8(a) shows the pattern of a single patch on a 6” PEC cylinder
(solid line). Because of rotational symmetry of the N-element cylindrical array, we
can obtain the far-field patterns of the other elements evenly distributed around the
cylinder by rotating the element pattern by 2πi/N , i = 1, 2,...,N(see the dotted
lines in Fig. 5.8(a)). After combining of all the element fields (vector sum), we can
synthesize the array pattern as shown in Fig. 5.8(b). This pattern synthesis method
does not include the mutual coupling between elements. However, for loosely packed
πD array ( Nλ > 1) the synthesized array pattern is very close to the full-wave simulation result of the complete array, which takes long time to analyze. We also notice that
πD for tightly packed array ( Nλ < 1) the array pattern is almost uniform with ripple level less than 1 dB. Therefore, by using this pattern synthesis method, we can easily
estimate the minimum number of elements required to achieve an omni-directional
azimuth pattern for different cylinder size (see Table. 5.3).
As can be seen from Table 5.3, we can achieve smaller gain ripple by increasing the number of array elements and/or reducing the cylinder diameter, which effectively
πD decreases the normalized spacing distance between elements ( Nλ ). Fig. 5.9 shows the 130 D
Z
Y 2.5’’ X
Figure 5.7: Simulation of a single patch on a PEC cylinder for roll pattern study
Ripple level (dB) 4-in 5-in 6-in 7-in 8-in 9-in 10-in 6-element 4 7.5 3 6 5 5 10 10-element 5 4.5 8 8.5 8 3.5 5 12-element 1.25 7.5 2 5 6 3.4 10 15-element 1.35 2.35 3.7 7.5 6.3 2.5 3 18-element 0.1 0.13 4 2.8 5.4 3.4 8.25 20-element 0 0.04 1 4 1.25 7 4.5 24-element 0 0 0 0.25 1 1.2 5.9 30-element 0 0 0 0 0 0.05 1.3
Table 5.3: Gain ripple level (dB) vs. different cylinder size and number of elements
131 0
-5
-10 dB -15
-20
-25 -180 -120 -60 0 60 120 180 θ (degree) (a) Azimuth pattern of a single patch on a PEC cylinder (D=6’’)
0
-2 Ripple level (+/- 0.9dB)
-4 dB -6
-8
-10 -180 -120 -60 0 60 120 180 φ (degree) (b) Synthesized azimuth pattern of N-element array
Figure 5.8: Pattern synthesis of a 20-element array on a cylinder (D=6”)
132 ripple levels vs. normalized spacing distance between elements. As seen, when the
πD spacing distance is less than one wavelength , Nλ < 1 , the gain ripple is decreased monotonously as we reduce the spacing distance. However, for larger spacing distance,
grating lobes are resulted, which causes non-monotonous relationship between the
gain ripple and the spacing distance. Therefore, in the prototype design on a 6.85-
inch cylinder, we chose the 24-element array as the transmitting antenna to achieve
an omni-directional azimuth pattern.
5.4 Prototype Cylindrically Conformal Array
5.4.1 Tuning by Near-field Probing
Fig. 5.1 displays the prototype transmitting and receiving cylindrical arrays. For this prototype design, we printed the antenna elements on discrete PCBs and glued them onto an aluminum polygon. The 24 transmitting elements were fed by coaxial cables connected to three 8-way power splitters and one 3-way power splitter. In prototype fabrication and tuning, we noticed that the phases of the coaxial cable outputs were not exactly the same for two reasons. 1) The coaxial cables were not of the same length due to fabrication inaccuracy. A 1 mm difference will mean around
16◦ phase difference at 10 GHz. 2) The outputs of the power splitters have about
10◦ phase differences. Therefore, one major fabrication challenge was to ensure equal
phase excitation for each antenna element since phase difference is very critical to the
ripple level of the array pattern. Given the above difficulties, we tuned the prototype
transmitting array by near-field probing method (Fig. 5.10) and adjusted the phases
by pulling or pushing the SMA connectors. As shown in Fig. 5.10, the transmitting
element was connected to port 1 of the network analyzer. A short monopole connected
133 10
9
8
7
6
5
4 Ripplelevel (dB) 3 D=4in D=5in 2 D=6in 1 D=7in
0 0 0.5 1 1.5 2 2.5 3 3.5 πD/(N λ) (a) Ripple level vs. number of element (N)
10
9 N=15 8 N=18 7 N=20 N=24 6
5
4 Ripplelevel (dB) 3
2
1
0 0 0.5 1 1.5 2 πD/(N λ) (b) Ripple level vs. diameter (D)
πD Figure 5.9: Ripple level vs. normalized spacing distance ( Nλ )
134 to port 2 was used to probe the near fields of each transmitting element. If the phase delay of S21 was smaller than the average one, the SMA connector was pulled outward a little bit to increase the phase delay. On the other hand, if the phase delay of S21 was larger than the average one, the SMA connector was pushed inward a little bit to decrease the phase delay. This process was repeated for all 24 transmitting elements.
In this way, the near-field phase of each transmitting element was tuned to be close to each other, which is necessary for delivering an omni-directional far-field pattern.
connected to port 1 of VNA short monopole
connected to port 2 of VNA
Figure 5.10: Phase tuning by near-field probing
Another fabrication challenge of the cylindrical array is associated with its feeding network. The weight and size of the Wilkinson power dividers and coaxial cables need to be minimized for future practical applications. Therefore, it is desirable to design a
PCB-based microstrip feeding network to excite all the antenna elements. As will be discussed in section 5.5, for this cylindrical array application, conformal microstrip feeding network is highly attractive to provide excitation with accurate amplitude and phase while minimizing the system load.
135 5.4.2 Measured Performance
The prototype antennas were measured in OSU-ESL anechoic chamber. The mea- sured return losses and the isolation between transmitting and receiving antennas are shown in Fig. 5.11. It can be seen that the transmitting and the receiving antennas worked from 10 GHz to 10.5 GHz with S11 < 10 dB and the isolation was as low − as -50 dB. Fig. 5.12 shows the measured performance of the 24-element transmitting array which was fed through coaxial cables connected to the power splitters. The measured gain was around 2.5 dB (Fig. 5.12), which was 2 dB lower than simula- tion due to the insertion losses of the cables and power splitters. Importantly, the
24-element transmitting array exhibited an omni-directional azimuth pattern with
+/-1.5 dB rippling (Fig. 5.12). This gain ripple was mainly caused by amplitude and phase inaccuracy of the coaxial cable outputs. With a more accurate feeding network, we expect to achieve an omni-directional pattern within +/-0.5 dB rippling.
The measured pattern of the prototype receiving antenna is shown in Fig. 5.13. The measured gain was 8.5 dB and the 3 dB beamwidth was around 20◦ and 90◦ for the elevation plane (φ = 0◦) and azimuth plane (θ = 90◦), respectively. Given the above measured performances, the transmitting and receiving antennas are well suited for low-profile conformal radar applications.
5.5 Polymer-based Circuit for Cylindrically Conformal Ar- ray
As discussed in the previous section, the prototype cylindrical microstrip array was fabricated with discrete PCBs, which are not convenient for array assembly and installation. Another fabrication challenge is associated with the feeding network
136 0
−10
−20
−30 dB
S11 (Tx) −40 S11 (Rx) S21 (Coupling)
−50
−60 9 9.5 10 10.5 11 Frequency (GHz)
Figure 5.11: Measured return loss and coupling of the prototype transmitting and receiving antennas
circuit. The power dividers and coaxial cables are too heavy and have poor accuracy in terms of output phases. To address these issues, we propose to utilize polymer-ceramic composites as substrates for the antennas and feeding network circuits. As shown in
Fig. 5.14, the polymer-ceramic sheet are very flexible and can be even rolled around.
The microstrip antennas can thus be printed on this polymer-ceramic sheet and be wrapped around the cylinder, thus make it very easy to assembly the cylindrical array.
Further, the flexibility of the polymer-ceramic sheets can also greatly facilitate the fabrication by allowing for roll-to-roll processing.
Polymer-ceramic composites are also suitable for flexible feeding network circuits.
Compared to the power dividers and coaxial cables, the printed flexible microstrip circuits (Fig. 2.15) has the advantage of light-weight and high system integration.
Further, with high precision printing, the polymer-based feeding network can achieve
137 5
0 20º -5
-10
-15
Realizedgain(dB) -20
-25
-30 0 20 40 60 80 100 120 140 160 180 θ (degree) (a) Elevation pattern ( = 0º)
10
8
6
4 2
0 +/- 1.5 dB -2
Realizedgain(dB) -4 -6
-8
-10 -180 -120 -60 0 60 120 180 φ (degree) (b) Azimuth pattern ( = 90º)
Figure 5.12: Measured radiation pattern of the transmitting array
138 15
10
5 20º 0
-5
-10
-15 Realizedgain(dB) -20
-25
-30 0 20 40 60 80 100 120 140 160 180 θ (degree) (a) Elevation pattern ( = 0º)
15 10 5 90º 0 -5 -10 -15 -20
Realizedgain(dB) -25 -30 -35 -40 -180 -120 -60 0 60 120 180 φ (degree) (b) Azimuth pattern ( = 90º)
Figure 5.13: Measured radiation pattern of the receiving antenna
139 Flexible polymer antenna wrapped around the cylinder Flexible polymer-ceramic sheet
Figure 5.14: Flexible polymer-based antenna and feeding network circuit wrapped around the cylinder
better accuracy of the output phases and amplitudes than the feeding scheme with power dividers and coaxial cables. The challenge of using polymer-ceramic composites for flexible electronics remains in the issue of metalization. Particularly for flexible electronics, the metalization layer must have high conductivity, strong adhesion to polymer surface, and high flexibility. Carbon nanotube sheets are very attractive solutions to address these needs as we discussed in Chapter 3
The cylindrical microstrip array is one of many applications of polymer-ceramic composites. More flexible electronics can be achieved using polymer-ceramic com- posites and carbon nanotube sheets. Due to their flexibility, low loss, controllable permittivity and high conductivity, the polymer-CNT composites provide us a new choice of materials when we design future compact flexible electronics.
140 5.6 Summary
In this chapter, we particularly focus on the design of a cylindrically conformal microstrip array. We present the design challenge and design concept of such cylin- drical array. Specifically, a series-fed multiple patch antenna is proposed to replace the common microstrip patch array with corporate feeding. The series feeding scheme has the benefit of less radiation loss and less coupling to the patch array. In order to achieve omni-directional pattern coverage, we study the minimum number of elements given certain cylinder size. Analysis results show that the normalized spacing distance between elements need to be smaller than 1 in order to produce an omni-directional pattern with +/- 1 dB ripple level.
As a prototype design, we fabricated the 24-element transmitting array using discrete PCBs. The array was excited with coaxial cables connected to commercial power dividers, which makes the system cumbersome. To reduce the system load and achieve higher accuracy of excitation phase and amplitude, we propose to print the feeding network circuit on the polymer-ceramic substrate. Since the polymer composites are highly flexible and stretchable, we can roll the flexible polymer-based circuit around the cylinder to realize the conformal feeding network circuits. Given their advantages of high flexibility, light weight, controllable permittvity and low loss, polymer-ceramic composites can find more applications as flexible electronics for small and light-weight devices. The current challenge of polymer-based flexible electronics is to realize high flexible printing on polymer sheets. Although the carbon nanotube sheet has exhibited very promising performances, more research work need to be done to improve its conductivity under bending and stretching conditions.
141 CHAPTER 6
CONCLUSION
6.1 Summary
Novel engineered materials have been the focus of recent research in electro- magnetics as they provide new electrical and mechanical properties not available in nature. In this dissertation, we focused on the development (Chapter 2 and Chapter 3) and application (Chapter 4 and Chapter 5) of novel polymer-ceramic composites.
Specifically, Chapter 2 presented the fabrication process of polymer-ceramic com- posites which includes three steps at room temperature: mixing, degassing, and cur- ing. By mixing the ceramic powders with the polymer matrix (PDMS or RTV6166) at certain volume ratio, we were able to achieve high permittivity and low loss di- electric composites. Measurement by both reflection and capacitance methods has shown that we are able to achieve controllable composite permittivity (² = 3 20) r ∼ with low loss (tanδ < 0.02). Several application examples were demonstrated, includ- ing polymer-based microstrip patch antenna, flexible polymer-based microstrip lines, textured dielectric substrates, multilayer dielectric rod antennas, and so on. Their superior mechanical, thermal and electrical properties have made polymer-ceramic
142 composites attractive as substrates for flexible antennas and as packaging materials for multilayer RF circuits.
To utilize polymer-ceramic composites for flexible multilayer antenna and RF systems, we must also address the issue of metalization or printing on polymer com- posites. In Chapter 3, we discussed the challenges of printing on polymers. Specif- ically, common lift-off lithography methods by metal evaporation did not work well for PDMS due to poor metal-polymer adhesion. Further, interface incompatibili- ties could easily cause detachment of the printed layers under bending or stretching stress. Carbon nanotube sheets were then introduced for the first time to realize
flexible, stretchable, and high conductivity printing on polymers. Two approaches, namely E-textile CNT sheet and vertically-aligned CNT sheet, were studied to achieve a CNT sheet of resistance around 1 Ω/¤. This provided us with novel polymer-CNT patch antenna with high gain (around 6 dB) and efficiency (around 83%). We also evaluated the conductivity of CNT sheets under bending and stretching conditions.
Performance degradation after bending or stretching can be explained by percolation theory. To further increase the CNT sheet conductivity, we proposed horizontal CNT alignment, giving rise to high conductivity bundled CNT threads. These can also be weaved into CNT fabrics for flexible high conductivity antenna and RF circuits.
Having light-weight, low loss polymer-ceramic composites and high conductivity carbon nanotube printing, we proceeded to realize conformal multilayer antennas and
RF systems. In Chapter 4, we discuss the development of a compact 4.5-inch tri-band
GPS array. This miniature GPS array is unique due to its dual-layer configuration and proximity probe feeding. Design concepts and realized prototypes were presented.
143 Measurements of prototype GPS array were also carried out showing excellent per- formances covering all three GPS bands (L1, L2 and L5). To address fabrication and integration difficulties with the high contrast ceramic substrates, we employed polymer-ceramic composites to realize this dual-layer tri-band GPS antenna. Above all, this polymer-CNT GPS antenna demonstrated the feasibility of using polymer- ceramic composites for the multilayer antenna and RF circuit applications.
In Chapter 5, we demonstrated another potential application of polymer-ceramic composites. Namely, we designed a cylindrically conformal microstrip array. Specifi- cally, to reduce radiation from the common corporate feeding of the microstrip array, a series-fed multiple patch antenna was designed. We also studied the array configu- ration to achieve omni-directional azimuth pattern with +/- 1 dB ripple. A prototype
24-element cylindrically conformal array was fabricated and tuned to meet the design specification. However, the discrete PCBs and the heavy coaxial cables and power dividers were not desirable for practical use. Therefore, we proposed to print the antenna elements and feeding networks on a flexible polymer sheet, which can be easily wrapped around the cylinder to realize the light-weight cylindrically conformal microstrip array. This application also serves as an example of flexible circuits that can be implemented on polymer-ceramic composites.
6.2 Future Work
Polymer-ceramic composites provided for new approaches in developing future antennas and RF systems. Some future research topics on polymer-ceramic and CNT composites may include the following.
144 6.2.1 Density Control to Improve CNT Sheet Conductivity
As we discussed in Chapter 3, the CNT sheet conductivity is determined by the nanotube percolation. In other words, the more CNTs per unit area are entangled or touched, the higher the CNT sheet conductivity is. Therefore, future research should focus on density control of the CNT array. This concept is shown in Fig. 6.1, and we already tried to compress longer CNT array (3mm thickness) to 1/15 of its original width. Indeed, the CNT density increased by 15 times. Specifically, our initial measurement showed that the compressed CNT array had a conductivity around 50 to 100 times higher than the uncompressed one. However, the compressed CNT sheet was too small for our antenna or circuit design. The challenge was to maintain high
CNT density while increasing yield rate of the CNT sheet.
Uncompressed CNT array Compressed CNT array
Figure 6.1: Compress CNT array to improve the conductivity
145 6.2.2 Bundled Carbon Nanotube Threads for Flexible RF Circuits
To further increase the conductivity of CNT sheets, we propose to interleave horizontally-aligned CNTs (bundled CNT threads) to a CNT fabric (see Fig. 6.2).
Currently, bundled CNT threads are being studied by different groups [59, 111]. The reported CNT threads have higher conductivity (σ = 106 S/m) than the vertically- aligned CNT sheet. However, the bundled CNT threads are more stiff and less stretch- able. Therefore, more research needs to be carried out to render the CNT fabric high
flexibility while improving its conductivity. Future work involves characterizing the bundled CNT threads and weaving flexible CNT threads into fabrics.
L
W
S D
(a) Spool of CNT threads (b) Weaved CNT fabric
Figure 6.2: CNT threads weaved into high conductivity CNT fabric
146 6.2.3 Alternative Material for Light-weight, Flexible, and Load-bearing Antennas
Beside carbon nanotubes, another topic of interest is to search for alternative light-weight, flexible, and high conductivity material for load-bearing antennas. One possible solution is the nylon-strengthened copper threads as shown in Fig. 6.3. As the nylon fibers have good flexibility and mechanical strength, the multistranded nylon-copper threads may have desirable properties in terms of strength, flexibility and conductivity. The composite threads can also be weaved into fabric to make light-weight, flexible and load-bearing antenna/RF circuits. Future work will focus on exploring novel composite threads to achieve desirable mechanical, electrical and thermal properties.
copper Nylon fiber
(a) Nylon-strengthened copper thread (b) Copper mesh
Figure 6.3: Nylon-strengthened copper threads and weaved wire mesh
147 6.2.4 Multilayer Polymer-based Antenna and RF Circuits
As demonstrated in Chapter 4, polymer-ceramic composites are well suited for the dual-layer GPS antennas. Indeed, we can employ polymer-ceramic composites for multi-stacked RF printing (see Fig. 6.4(a)), which has potential to offer wide- band coverage. Such multilayer configuration can be realized by casting polymer mixture into a mold with layer-by-layer printing. Moreover, 3-D fabrication with polymers also allows inclusion of IC components, which can eventually be extended into a conformal multilayer RF front-ends as shown in Fig. 6.4(b). Future studies on system integration, example temperature expansion rate of the materials are necessary to realize the potential CNT multilayer printing.
Antenna/Array
Antenna substrate ……
Circuits layer
(a) Multilayer polymer-based antenna (b) Multilayer polymer-based antenna and RF circuits
Figure 6.4: Polymer-based multilayer antennas and RF circuits
148 BIBLIOGRAPHY
[1] B. Bing and N. Jayant, “A cellphone for all standards,” IEEE Spectrum, vol. 39, pp. 34–39, May 2002.
[2] M. Ali, G. J. Hayes, H.-S. Hwang, and R. A. Sadler, “Design of a multiband in- ternal antenna for third generation mobile phone handsets,” IEEE Transactions on Antennas and Propagation, vol. 51, pp. 1452–1461, July 2003.
[3] K. L. Wong, Y.-C. Lin, and B. Chen, “Internal patch antenna with a thin air- layer substrate for GSM/DCS operation in a PDA phone,” IEEE Transactions on Antennas and Propagation, vol. 55, pp. 1165–1172, April 2007.
[4] S.-Y. Lin and K.-C. Huang, “A compact microstrip antenna for GPS and DCS application,” IEEE Transactions on Antennas and Propagation, vol. 53, pp. 1227–1229, March 2005.
[5] K. Yegin, “AMPS/PCS/GPS active antenna for emergency call systems,” IEEE Antennas and Wireless Propagation Letters, vol. 6, pp. 255–258, 2007.
[6] Y. Zhou, C.-C. Chen, and J. L. Volakis, “Single-fed circularly polarized antenna element with reduced coupling for GPS arrays,” IEEE Transactions on Antennas and Propagation, vol. 56, pp. 1469–1472, May 2008.
[7] M. Ali, R. A. Sadler, and G. J. Hayes, “A uniquely packaged internal inverted- F antenna for Bluetooth or wireless LAN application,” IEEE Transactions on Antennas and Propagation, vol. 1, pp. 5–7, January 2002.
[8] B. S. Yildirim, “Low-profile and planar antenna suitable for WLAN/Bluetooth and UWB applications,” IEEE Antennas and Wireless Propagation Letters, vol. 5, pp. 438–441, 2006.
[9] Y.-S. Shin and S.-O. Park, “A compact loop type antenna for Bluetooth, S- DMB, Wibro, WiMax, and WLAN applications,” IEEE Antennas and Wireless Propagation Letters, vol. 6, pp. 320–323, 2007.
[10] R. Bansal, “Coming soon to a Wal-Mart near you,” IEEE Antennas and Prop- agation Magazine, vol. 45, pp. 105–106, December 2003.
149 [11] K. Finkenzeller, RFID Handbook: Radio-Frequency Identification Fundamentals and Applications. Wiley, 2nd ed., 2004.
[12] K. V. S. Rao, P. V. Nikitin, and S. F. Lam, “Antenna design for UHF RFID tags: a review and a practical application,” IEEE Transactions on Antennas and Propagation, vol. 53, December.
[13] J. Melby, “JTRS and the evolution toward software-defined radio,” in Proceed- ings of Military Communications Conference (MILCOM), vol. 2, pp. 1286–1290, October 2002.
[14] J. L. Hillman, S. D. Jones, R. A. Nichols, and I. J. Wang, “Communications network architectures for the army future combat system and objective force,” in Proceedings of Military Communications Conference (MILCOM), vol. 2, pp. 1417–1421, October 2002.
[15] H. Hellsten and L. M. H. Ulander, “Airborne array aperture UWB UHF radar- motivation and system considerations,” IEEE Aerospace and Electronic Systems Magazine, vol. 15, pp. 35–45, May 2000.
[16] T. F. Kennedy, P. W. Fink, A. W. Chu, N. J. Champagne, G. Y. Lin, and M. A. Khayat, “Body-worn E-Textile antennas: The good, the low-mass, and the conformal,” IEEE Transactions on Antennas and Propagation, vol. 57, pp. 910– 918, April 2009.
[17] A. K. Skrivervik, J.-F. Zurcher, O. Staub, and J. R. Mosig, “PCS antenna design: the challenge of miniaturization,” IEEE Antennas and Propagation Magazine, vol. 43, pp. 12–27, Augest 2001.
[18] J. Volakis, G. Mumcu, K. Sertel, C.-C. Chen, M. Lee, B. Kramer, D. Psy- choudakis, and G. Kiziltas, “Antenna miniaturization using magnetic-photonic and degenerate band-edge crystals,” IEEE Antennas and Propagation Maga- zine, vol. 48, pp. 12–28, October 2006.
[19] P. M. T. Ikonen, K. N. Rozanov, A. V. Osipov, P. Alitalo, and S. A. Tretyakov, “Magnetodielectric substrates in antenna miniaturization: Potential and limi- tations,” IEEE Transactions on Antennas and Propagation, vol. 54, pp. 3391– 3399, November 2006.
[20] J. S. Kula, D. Psychoudakis, W.-J. Liao, C.-C. Chen, J. L. Volakis, and J. W. Halloran, “Patch-antenna miniaturization using recently available ceramic sub- strates,” IEEE Antennas and Propagation Magazine, vol. 48, pp. 13–20, De- cember 2006.
150 [21] B. A. Kramer, C.-C. Chen, and J. L. Volakis, “Size reduction of a low-profile spi- ral antenna using inductive and dielectric loading,” IEEE Antennas and Wire- less Propagation Letters, vol. 7, pp. 22–25, 2008.
[22] D. S. Filipovic and J. L. Volakis, “Broadband meanderline slot spiral antenna,” IEE Proceedings on Microwaves, Antennas and Propagation, vol. 149, pp. 98– 105, April 2002.
[23] S. Koulouridis, G. Kiziltas, Y. Zhou, D. Hansford, and J. L. Volakis, “Polymer- ceramic composites for microwave applications: Fabrication and performance assessment,” IEEE Transaction on Microwave Theory and Techniques, vol. 54, pp. 4202–4208, December 2006.
[24] Y. Bayram, Y. Zhou, J. L. Volakis, B.-S. Shim, and N. A. Kotov, “Textile conductors and polymer-ceramic composites for load bearing antennas,” IEEE APS/URSI International Symposium, San Diego, CA, July 2008.
[25] Y. Zhou, Y. Bayram, L. Dai, and J. L. Volakis, “Conductive polymer-carbon nanotube sheets for conformal load bearing antennas,” URSI Radio Science Meeting, Boulder, CO, January 2009.
[26] K. Lim, S. Pinel, M. Davis, A. Sutono, C.-H. Lee, D. Heo, A. Obatoynbo, J. Laskar, E. M. Tantzeris, and R. Tummala, “RF-system-on-package (SOP) for wireless communications,” IEEE Microwave Magazine, vol. 3, pp. 88–99, March 2002.
[27] D. C. Thompson, M. M. Tentzeris, and J. Papapolymerou, “Packaging of MMICs in multilayer LCP substrates,” IEEE Microwave and Wireless Com- ponents Letters, vol. 16, pp. 410–412, July 2006.
[28] N. Kingsley, G. E. Ponchak, and J. Papapolymerou, “Reconfigurable RF MEMS phased array antenna integrated within a liquid crystal polymer (LCP) system-on-package,” IEEE Transactions on Antennas and Propagation, vol. 56, pp. 108–118, January 2008.
[29] G. DeJean, R. Bairavasubramanian, D. Thompson, G. E. Ponchak, M. M. Tentzeris, and J. Papapolymerou, “Reconfigurable RF MEMS phased array antenna integrated within a liquid crystal polymer (LCP) system-on-package,” IEEE Antennas and Wireless Propagation Letters, vol. 4, pp. 22–26, 2005.
[30] L. Katehi, W. Chappell, S. Mohammadi, A. Margomenos, and M. Steer, “Het- erogeneous wafer-scale circuit architectures,” IEEE Microwave, vol. 8, pp. 52– 69, February 2007.
151 [31] C.-C. Weng, C.-F. Chang, and S.-J. Chung, “Development of a compact low- temperature co-fired ceramic antenna front-end module,” IEEE Transactions on Microwave Theory and Techniques, vol. 56, pp. 2483–2492, November 2008.
[32] D. G. Baird and D. I. Collias, Polymer Processing: Principles and Design. Wiley, 1998.
[33] A. Ram, Fundamentals of Polymer Engineering. Springer, 1997.
[34] D. W. Richerson, Modern Ceramic Engineering: Properties, Processing, and Use in Design. CRC Press, 3rd ed., 2006.
[35] R. C. Buchanan, Ceramic Materials for Electronics. CRC Press, 3rd ed., 2004.
[36] J. A. Rogers, Z. Bao, K. Baldwin, A. Dodabalapur, B. Crone, V. R. Raju, V. Kuck, H. Katz, K. Amundson, J. Ewing, and P. Drzaic, “Paperlike elec- tronic displays: Large area, rubber stamped plastic sheets of electronics and electrophoretic inks,” Proceedings of the National Academy of Sciences, vol. 98, no. 9, pp. 4835–4840, 2001.
[37] D.-H. Hwang, S. T. Kim, X.-C. Li, B. S. Chuah, J. C. DeMello, R. H. Friend, S. C. Moratti, and A. B. Holmes, “New luminescent polymers for LEDs and LECs,” Proceedings of Macromolecular Symposium, vol. 125, pp. 111–120, 1998.
[38] J. Han, B.-J. Seo, S. K. H. Zhang, and H. R. Fetterman, “Single-chip integrated electro-optic polymer photonic rf phase shifter array,” Journal of Lightwave Technology, vol. 12, pp. 3257–3261, December 2003.
[39] Y. Yashchyshyn and J. W. Modelski, “Rigorous analysis and investigations of the scan antennas on a ferroelectric substrate,” IEEE Transaction on Microwave Theory and Techniques, vol. 53, pp. 427–438, February 2005.
[40] S. Ramesh, B. A. Shutzberg, C. Huang, J. Gao, and E. P. Giannelis, “Dielectric nanocomposites for integral thin capacitors: Materials design, fabrication and integration issues,” IEEE Transactions on Advanced Packaging, vol. 26, pp. 17– 24, February 2003.
[41] K. Paik, S. Cho, and J. Hyun, “Novel epoxy/BaTiO3 composite embedded capacitor films embedded in organic substrates,” Proc. Int. IEEE Conf. Asian Green Electron, pp. 68–73, 2004.
[42] F. Xiang, H. Wang, and X. Yao, “Dielectric properties of SrTiO3/POE flexi- ble composites for microwave applications,” Journal of the European Ceramic Society, vol. 27, pp. 3093–3097, 2007.
152 [43] F. Xiang, H. Wang, H. Yang, Z. Shen, and X. Yao, “Low loss flexible SrTiO3/POE dielectric composites for microwave application,” Journal of Elec- troceramics, vol. 10, pp. 1385–3449, 2008.
[44] H. M. Al-Allak, J. Illingsworth, A. W. Brinkman, and J. Woods, “Permittivity- temperature behavior of donor-doped positive temperature coefficient of resis- tance BaTiO3,” Journal of Physics D: Applied Physics, vol. 22, pp. 1920–1923, 1989.
[45] C. K. Chiang, R. Popielarz, and L. P. Sung, “Dielectric properties and mor- phology of ferroelectric ceramic-polymer composite films,” in Proc. Mat. Res. Soc. Symp., vol. 682E, pp. N6.9.1–N6.9.6, 2001.
[46] R. Popielarz, C. K. Chiang, R. Nozaki, and J. Obrzut, “Dielectric properties of polymer/ferroelectric ceramic composites from 100 Hz to 10 GHz,” Macro- molecules, vol. 34, pp. 5910–5915, 2001.
[47] C. C. Li and K. M. Chen, “Determination of electromagnetic properties of materials using flanged open-ended coaxial probe – full-wave analysis,” IEEE Transactions on Instrumentation and Measurement, vol. 44, pp. 19–27, 1995.
[48] J. Baker-Jarvis, “Transmission/reflection and short-circuit line permittivity measurements,” tech. rep., NIST Technical Note 1341, National Institute of Standards and Technology, U. S. Department of Commerce, 2000.
[49] S. B. Cohn and K. C. Kelly, “Microwave measurement of high-dielectric- constant materials,” IEEE Transactions on Microwave Theory and Techniques, vol. 14, pp. 406–410, 1966.
[50] P. A. Bernard and J. M. Gautray, “Measurement of dielectric constant using a microstrip ring resonator,” IEEE Transactions on Microwave Theory and Techniques, vol. 28, pp. 592–595, 1991.
[51] Y. Kobayashi and S. Tanaka, “Resonant modes of a dielectric rod resonator short-circuited at both ends by parallel conducting plates,” IEEE Transactions on Microwave Theory and Techniques, vol. 28, pp. 1077–1085, 1980.
[52] A. Deschamps, “Impedance of an antenna in a conducting medium,” IEEE Transactions on Antennas and Propagation, vol. 10, pp. 648–650, 1972.
[53] G. G. Raju, Dielectrics in Electric Fields. Marcel Dekker, 2003.
[54] G. Kiziltas, D. Psychoudakis, J. L. Volakis, and N. Kikuchi, “Topology design optimization of dielectric substrates for bandwidth improvement of a patch antenna,” IEEE Transaction on Antennas and Propagation, vol. 51, pp. 2732– 2743, October 2003.
153 [55] D. Psychoudakis, Y.-H. Koh, J. L. Volakis, and J. H. Halloran, “Design method for aperture-coupled microstrip patch antennas on textured dielectric sub- strates,” IEEE Transaction on Antennas and Propagation, vol. 52, pp. 2763– 2766, October 2004.
[56] C.-C. Chen and J. L. Volakis, “Bandwidth broadening of patch antennas using nonuniform substrates,” Microwave and Optical Technology Letters, vol. 47, pp. 421–423, October 2005.
[57] J.-Y. Chung and C.-C. Chen, “Two-layer dielectric rod antenna,” IEEE Trans- action on Antennas and Propagation, vol. 56, pp. 1541–1547, June 2008.
[58] E. Apaydin, Y. Zhou, D. Hansford, S. Koulouridis, and J. L. Volakis, “Pat- terned metal printing on pliable composites for RF design,” IEEE APS/URSI International Symposium, San Diego, CA, July 2008.
[59] M. Zhang, K. R. Atkinson, and R. H. Baughman, “Multifunctional carbon nano- tube yarns by downsizing an ancient technology,” Science, vol. 306, no. 5700, pp. 1358–1361, 2004.
[60] A. Lobovsky, J. Matrunich, M. Kozlov, R. C. Morris, R. H. Baughman, and A. A. Zakhidov, “Spinning, processing, and applications of carbon nanotube filaments, ribbons, and yarns.” US patent 2004/0096389 A1, May 2004.
[61] R. Saito, G. Dresselhaus, and M. S. Dresselhaus, Physical Properties of Carbon Nanotubes. Imperial College Press, 1998.
[62] M. S. Dresselhaus, G. Dresselhaus, and P. Avouris, Carbon Nanotubes: Synthe- sis, Structure, Properties and Applications. Springer, 1st ed., 2001.
[63] L. Dai, ed., Carbon Nanotechnology : Recent Developments in Chemistry, Physics, Materials Science and Device Applications. Elsevier, 1st ed., 2006.
[64] P. G. Collins and P. Avouris, “Nanotubes for electronics,” Scientific American, vol. 283, no. 6, pp. 62–69, 2000.
[65] B. G. Demczyk, Y. M. Wanga, J. Cumingsa, M. Hetmana, W. Hana, A. Zettla, and R. O. Ritchieb, “Direct mechanical measurement of the tensile strength and elastic modulus of multiwalled carbon nanotubes,” Materials Science and Engineering A, vol. 334, no. 1-2, pp. 173–178, 2002.
[66] M. Meo and M. Rossi, “Prediction of Young’s modulus of single wall carbon nanotubes by molecular-mechanics based finite element modelling,” Composites science and technology, vol. 66, no. 11-12, pp. 1597–1605, 2006.
154 [67] J. Che, T. agin, and W. A. Goddard, “Thermal conductivity of carbon nan- otubes,” Nanotechnology, vol. 11, pp. 65–69, 2000.
[68] J. Hone, M. Whitney, C. Piskoti, and A. Zettl, “Thermal conductivity of single- walled carbon nanotubes,” Physical Review B, vol. 59, no. 4, pp. 2514–2516, 1999.
[69] P. Avouris, “Carbon nanotube electronics,” Chemical Physics, vol. 281, no. 2-3, pp. 429–445, 2002.
[70] G. W. Hanson, “Fundamental transmitting properties of carbon nanotube antennas,” IEEE Transactions on Antennas and Propagation, vol. 53, pp. 3426– 3435, November 2005.
[71] J. Hao and G. W. Hanson, “Infrared and optical properties of carbon nanotube dipole antennas,” IEEE Transactions on Nanotechnology, vol. 5, pp. 766–775, November 2006.
[72] P. J. Burke, S. Li, and Z. Yu, “Quantitative theory of nanowire and nanotube antenna performance,” IEEE Transactions on Nanotechnology, vol. 5, pp. 314– 334, July 2006.
[73] G. Miano and F. Villone, “An integral formulation for the electrodynamics of metallic carbon nanotubes based on a fluid model,” IEEE Transactions on Antennas and Propagation, vol. 54, pp. 2713–2724, October 2006.
[74] Y. Huang, W.-Y. Yin, and Q. H. Liu, “Performance prediction of carbon nano- tube bundle dipole antennas,” IEEE Transactions on Nanotechnology, vol. 7, pp. 331–337, November 2008.
[75] D. A. Wharam, T. J. Thornton, R. Newbury, M. Pepper, H. Ahmed, J. E. F. Frost, D. G. Hasko, D. C. Peacock, D. A. Ritchie, and G. A. C. Jones, “One- dimensional transport and the quantisation of the ballistic resistance,” Journal of Physics C: Solid State Physics, vol. 21, pp. 209–214, March 1988.
[76] D. A. Wharam, M. Pepper, H. Ahmed, J. E. F. Frost, D. G. Hasko, D. C. Pea- cock, D. A. Ritchie, and G. A. C. Jones, “Addition of the one-dimensional quan- tised ballistic resistance,” Journal of Physics C: Solid State Physics, vol. 21, pp. 887–891, August 1988.
[77] A. Yin, H. Chik, and J. Xu, “Postgrowth processing of carbon nanotube arrays– enabling new functionalities and applications,” IEEE Transactions on Nan- otechnology, vol. 3, no. 1, pp. 147–151, 2001.
155 [78] L. Wang, R. Zhou, and H. Xin, “Microwave (8-50 GHz) characterization of mul- tiwalled carbon nanotube papers using rectangular waveguides,” IEEE Trans- actions on Microwave Theory and Techniques, vol. 56, pp. 499–506, February 2008.
[79] J. Wang, M. Musameh, , and Y. Lin, “Solubilization of carbon nanotubes by nafion toward the preparation of amperometric biosensors,” Journal of the American Chemical Society, vol. 125, pp. 2408–2409, February 2003.
[80] L. Dai, A. Patil, X. Gong, Z. Guo, L. Liu, Y. Liu, and D. Zhu, “Aligned nanotubes,” ChemPhysChem, vol. 4, pp. 1150–1169, November 2003.
[81] D. Stauffer and A. Aharony, Introduction to Percolation Theory. Taylor & Francis, 1992.
[82] J. Ma, J. T. W. Yeow, J. C. L. Chow, and R. B. Barnett, “Effect of percolation on electrical conductivity in a carbon nanotube-based film radiation sensor,” in Proc. IEEE conference on nanotechnology, pp. 259–262, August 2008.
[83] A. M. Lepadatu, E. Rusnac, and I. Stavarache, “Percolation phenomena in silicon-based nanocrystalline systems,” vol. 2, pp. 575–578, October 2007.
[84] J. G. McNeff, “The global positioning system,” IEEE Transactions on Micro- wave Theory and Techniques, vol. 50, pp. 645–652, March 2002.
[85] S. Lazar, “Modernization and the move to GPS III,” Crosslink, vol. 3, no. 2, pp. 42–46, 2002.
[86] D. M. Pozar and S. M. Duffy, “A dual-band circularly polarized aperture cou- pled stacked microstrip antenna for global positioning satellite,” IEEE Trans- actions on Antennas and Propagation, vol. 45, pp. 1618–1625, Nov. 1997.
[87] C. M. Su and K. L. Wong, “A dual-band GPS microstrip antenna,” Microwave and Optical Technology Letters, vol. 33, pp. 238–240, Apr. 2002.
[88] B. R. Rao, M. A. Smolinski, C. C. Quach, and E. N. Rosario, “Triple-band GPS trap-loaded inverted l antenna array,” Microwave and Optical Technology Letters, vol. 38, pp. 35–37, May. 2003.
[89] X. F. Peng, S. S. Zhong, S. Q. Xu, and Q. Wu, “Compact dual-band GPS microstrip antenna,” Microwave and Optical Technology Letters, vol. 44, pp. 58– 61, Nov. 2004.
[90] Y. Zhou, S. Koulouridis, G. Kizitas, and J. L. Volakis, “A novel 1.5” quadruple antenna for tri-band GPS applications,” IEEE Antennas and Wireless Propa- gation Letters, vol. 5, pp. 224–227, 2006.
156 [91] Y. Zhou, C.-C. Chen, and J. L. Volakis, “Dual band proximity-fed stacked patch antenna for tri-band GPS applications,” IEEE Transactions on Antennas and Propagation, vol. 55, pp. 220–223, January 2007.
[92] Y. Zhou, C.-C. Chen, and J. L. Volakis, “Single-fed circularly polarized antenna element with reduced coupling for GPS arrays,” IEEE Transactions on Antennas and Propagation, vol. 56, pp. 1469–1472, May 2008.
[93] Y. Zhou, C.-C. Chen, and J. L. Volakis, “A compact 6-element tri-band GPS array,” Antenna Measurement Techniques Association (AMTA) Symposium, Boston, MA, November 2008.
[94] R. B. Waterhouse, “Stacked patches using high and low dielectric constant material combinations,” IEEE Transactions on Antennas and Propagation, vol. 47, pp. 1767–1771, Dec. 1999.
[95] R. Q. Lee, K. F. Lee, and J. Bobinchak, “Characteristics of a two-layer electro- magnetically coupled rectangular patch antenna,” Electronics Letters, vol. 23, no. 20, pp. 1070–1072, 1987.
[96] F. Croq and A. Papiernik, “Stacked slot-coupled printed antenna,” IEEE Micro- wave and Guided Wave Letters, vol. 1, no. 10, pp. 280–290, 1991.
[97] K. M. Luk, C. L. Mak, Y. L. Chow, and K. F. Lee, “Broadband microstrip patch antenna,” Electronics Letters, vol. 34, pp. 1442–1443, Jul. 1998.
[98] A. K. Shackelford, K. F. Lee, and K. M. Luk, “Design of small-size wide- bandwidth microstrip patch antennas,” IEEE Antennas and Propagation Mag- azine, vol. 45, pp. 75–83, Feb. 2003.
[99] D. M. Pozar, Microwave Engineering. Wiley, 2nd ed., 1998.
[100] G. F. Hatke, “Adaptive array processing for wideband nulling in GPS systems,” in Proc. 32th Asilomar Conf. Signals, Syst., Comput., pp. 1332–1336, 1998.
[101] S. R. Forrest, “The path to ubiquitous and low-cost organic electronic appli- ances on plastic,” Nature, vol. 428, pp. 911–918, April 2004.
[102] K. Jain, M. Klosner, M. Zemel, and S. Raghunandan, “Flexible electronics and displays: High-resolution, roll-to-roll, projection lithography and photoab- lation processing technologies for high-throughput production,” Proceedings of the IEEE, vol. 93, pp. 1500–1510, August 2005.
[103] R. E. Munson, “Conformal microstrip antennas and microstrip phased arrays,” IEEE Transaction on Antennas and Propagation, vol. 22, pp. 74–78, 1974.
157 [104] R. E. Post and D. T. Stephenson, “The design of a microstrip antenna array for a UHF space telemetry link,” IEEE Transaction on Antennas and Propagation, vol. 29, pp. 129–134, 1981.
[105] I. Jayakumar, R. Garg, B. K. Sarap, and B. Lal, “A conformal cylindrical micro- strip array for producing omnidirectional radiation pattern,” IEEE Transaction on Antennas and Propagation, vol. 34, pp. 1258–1261, October 1986.
[106] E. Levine, G. Malamud, S. Shtrikman, and D. Treves, “A study of microstrip array antennas with the feed network,” IEEE Transaction on Antennas and Propagation, vol. 37, pp. 426–434, April 1989.
[107] T. Metzler, “Microstrip series arrays,” IEEE Transaction on Antennas and Propagation, vol. 29, pp. 174–178, January 1981.
[108] B. B. Jones, F. Y. M. Chow, and A. W. Seeto, “The synthesis of shaped patterns with series-fed microstrip patch arrays,” IEEE Transaction on Antennas and Propagation, vol. 30, pp. 1206–1212, November 1982.
[109] E. H. Newman and J. E. Tehan, “Analysis of a microstrip array and feed net- work,” IEEE Transaction on Antennas and Propagation, vol. 33, pp. 397–403, April 1985.
[110] C. A. Balanis, Antenna Theory: Analysis and Design. Wiley, 2nd ed., 1996.
[111] M. Zhang, S. Fang, A. A. Zakhidov, S. B. Lee, A. E. Aliev, C. D. Williams, K. R. Atkinson, and R. H. Baughman, “Strong, transparent, multifunctional, carbon nanotube sheets,” Science, vol. 309, no. 5738, pp. 1215–1219, 2005.
158