USF Libraries
USF Digital Collections

Investigation of a Rectenna element for infrared and millimeter wave application

MISSING IMAGE

Material Information

Title:
Investigation of a Rectenna element for infrared and millimeter wave application
Physical Description:
Book
Language:
English
Creator:
La Rosa, Henrry
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Microstrip slot antenna
Schottky diode
Detector
Coplanar waveguide (CPW)
Wideband
W band
Membrane
Dissertations, Academic -- Electrical Engineering -- Masters -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This thesis presents the rectifying antenna potential for infrared and millimeter wave energy conversion. Infrared imaging is one of the emerging technologies that have attracted considerable attention in the next generation of military, medical, and commercial applications. Moreover, with the ever-increasing congestion of the electromagnetic spectrum at RF and microwave frequencies and the establishment of firm civilian and military requirements best met by millimeter wave systems, the interest in the technology has grown and is now firmly established. During this work a 2.5GHz slot antenna, a 2.5GHz Schottky diode detector, a CPW-to-Microstrip transition, a fully integrated Rectenna element, and a 94GHz slot antenna were designed, fabricated, and tested. Results on the performance of the devices show a great deal of correlation between the simulated and measured data. To perform an initial study, the CPW-fed narrow slot antenna is designed at 2.5GHz and implemented on an FR-4 board. This investigation serves as the basis for the development of the Rectenna element at millimeter wave frequencies. In order to increase the bandwidth of the slot antenna, a 2.5GHz CPW-fed wide slot antenna with U-shaped tuning stub is realized, which provides a 60% increase in bandwidth while keeping the same radiation characteristics. In addition, a set of simulations is performed to show how a reflector plate affects the radiating properties of the slot antenna. A 2.5GHz square-law detector is also designed, fabricated, and tested in order to rectify the RF signal delivered by the antenna. The fabricated detector presents a well matched condition at the design frequency with a dynamic range found to be from --17dBm to --50dBm. The low frequency Rectenna element prototype is then integrated within a single FR-4 board.^ This is accomplished by implementing a compact via-less CPW-to-Microstrip transition. Finally, a 94GHz CPW-fed wide slot antenna is realized on a 10μm high resistivity silicon membrane. This antenna shows a great deal of similarity to the 2.5GHz slot antenna. This low profile antenna presents at least a 10dB return loss over the entire W band frequency window. Simulated antenna efficiencies of up to 99% were achieved assuming a perfect conductor.
Thesis:
Thesis (M.S.E.E.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Henrry La Rosa.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 134 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001966245
oclc - 262835149
usfldc doi - E14-SFE0002221
usfldc handle - e14.2221
System ID:
SFS0026539:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Investigation of the Rectenna Concept for Millimeter Wave Applications by Henrry La Rosa A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Co-Major Professor: El ias Stefanakos, Ph.D. Co-Major Professor: Kenneth A. Buckle, Ph.D. Thomas M. Weller, Ph.D. Shekhar Bhansali, Ph.D. Date of Approval: November 5, 2007 Keywords: Microstrip Slot Antenna, Scho ttky Diode, Detector, Coplanar Waveguide (CPW), Wideband, W band, Membrane Copyright 2007, Henrry La Rosa

PAGE 2

Dedication To my fianc Stephanie Pedersen, my parent s Felix N. La Rosa and Concepsion A. La Rosa, and my sister Arlettys La Rosa

PAGE 3

Acknowledgements I would like to express my gr atitude to the individuals w ho have contributed to this thesis work. I gratefully acknowledge my a dvisor Elias Stefanakos, Ph.D., P.E. and my co-advisor Kenneth A. Buckle Ph.D., P.E. for their excellent guidance, unconditional support, and for providing me with the opportun ity to work in such interesting projects such as the one mentioned in this thesis. I also would like to thank Thomas M. Weller, Ph.D. for his excellent tec hnical support and guidance. In addition, I would like to express my gratitude to Shekhar Bhansali, Ph.D., Mr. Bernard Batson, and the Bridge to the Doctorate program for their financial s upport and supervision. I would also like to recognize Modelithics Inc. and the WAMI program for providing me the necessary equipment needed for the success of this project. My gratitude also goes to all my colleagues in ENB 412 and ENB 245 who made the success of this project possible. Especially I would like to express my gratitude to Subramanian Krishnan for the development and fabrication of the MI M diode as well as the fabrication of the 94GHz sl ot antenna. I would also like to especially thank Suzette Presas for her unconditional friendship, supp ort, and reviewing this thesis work. In addition, I would like to thank Alberto Rodrig uez and Sergio Melais for their technical support and guidance through out the project. My gratitude also goes to Saravana P. Natarajan for assisting with the 94GHz system fabrication process. Finally, I want to thank Bojana Zevanovic for reviewing this thesis work.

PAGE 4

i Table of Contents List of Tables iv List of Figures v Abstract xii Chapter 1 Introduction 1 1.1 Overview 1 1.2 Motivation 4 1.3 Thesis Organization 4 1.4 Contributions 5 Chapter 2 2.5GHz Slot Antenna Design 7 2.1 Introduction 7 2.1.1 Background Theory 7 2.1.2 Input Impedance 9 2.1.3 Radiation Pattern 11 2.2 Motivation for Using a CPW-fed Slot Antenna 14 2.3 CPW-fed Narrow Slot Antenna 15 2.4 Antenna Efficiency and Gain Definitions 24 2.5 CPW-fed Wide Slot Antenna with Tuning Stubs 26 2.6 Antenna Simulations with Reflector in Place 30 2.7 Result and Comparison 34

PAGE 5

ii 2.7.1 2.5GHz Narrow Slot Antenna 34 2.7.2 Radiation Pattern Measurements 36 2.7.3 2.5GHz Wide Slot Antenna with U-shaped Tuning Stubs Results 41 2.8 Chapter Summary and Conclusions 46 Chapter 3 2.5GHz Detector Circuit Design Integration of Rectenna Element 47 3.1 Introduction 47 3.1.1 Background Theory 48 3.2 Detector Circuit Design 51 3.3 Rectenna Element Integration 61 3.3.1 Via-less CPW-to-Microstrip Transition 61 3.4 Results and Comparison 66 3.4.1 2.5GHz Detector Circuit 66 3.4.2 Via-less CPW-to-Microstrip Transition Results 71 3.4.3 Antenna/CPW-to-Microstr ip Integration Results 73 3.4.4 2.5GHz Rectenna Element Integration 75 3.5 Chapter Summary and Conclusions 78 Chapter 4 94GHz Antenna Design 79 4.1 Introduction 79 4.1.1 Millimeter Wave Band Characteristics 79 4.1.2 Advantages and Disadvantages 79 4.1.3 Applications 82 4.2 94GHz CPW-fed Sl ot Antenna Design 84 4.2.1 Surface Wave in Silicon 85

PAGE 6

iii 4.2.2 Substrate Thickness Characterization Using Momentum 86 4.2.3 TRL Standards and S ubstrate Thickness Issues 90 4.3 Results and Comparisons 92 4.4 Chapter Summary and Conclusions 99 Chapter 5 Summary and Recommendations 101 5.1 Conclusions 101 5.2 Future Work and Recommendations 104 References 106 Appendices 109 Appendix A: 2.5GHz Rectenna Elements Measurements' Setup 110 Appendix B: Low Frequency Calib ration Standards and Results 117 Appendix C: W band Calibration Standard Results 125

PAGE 7

iv List of Tables Table 2.1 Comparison Between the Slot and Patch Antenna Characteristics. 14 Table 2.2 Comparison Between the Center-fed and Offset-fed Simulated Antenna Parameters. 24 Table 2.3 Comparison Between the Simulated Antenna Parameters with no Reflector, Finite Reflector, and In finite Reflector Placed at dr = 33 Table 3.1 Step Impedance Filter Dimensions. 53 Table 3.2 Measured Input Power and Output Voltage as a Function of Attenuation Setting. 69 Table 4.1 Radar System Performance in Millimeter Wave, Microwave, and Optical Frequencies. 82 Table 4.2 Millimeter Wave Applicatio ns in the Four Major Areas of Wireless Communication [24]. 83 Table 4.3 Slot Antenna Efficiency, Dir ectivity, and Gain as a Function of Substrate Thickness. 88 Table A.1 Recorded Data for the Rectenna Elements Connected by a M-M Barrel. 115 Table A.2 Recorded Data for the Fully Integrated Rectenna Element. 116

PAGE 8

v List of Figures Figure 1.1 Nahas Model for Microwave-to-DC Energy Conversion [3]. 2 Figure 2.1 (a) Center-fed Slot Antenna Short Circuited Through the Substrate. 8 Figure 2.2 Voltage and Current Distribution Around the Slot. 9 Figure 2.3 Different Feeding Methods Used to Decrease the Slot Antenna Input Impedance. 10 Figure 2.4 (a) Slot Antenna on an Infinite Ground Plane, (b) Complementary Printed Dipole Antenna. 11 Figure 2.5 (a) Radiated Fiel ds of a Slot Antenna on an Infinite Ground Plane, (b) and Radiated Fields of th e Complementary Dipole Antenna. 12 Figure 2.6 Slot Antenna Fed by a Microstrip Line with (a) Vertical Polarization, and (b) Horizontal Polarization. 13 Figure 2.7 CPW-fed Slot Antenna Feeding Mechanisms: (a) Center-fed Slot Antenna, (b) Offset-fed Slot Antenn a with Bond Wire for Stability, (c) Capacitively Coupled, and (d) In ductively Coupled Slot Antenna. 16 Figure 2.8 (a) Current Density Distribution of the Center-fed Slot Antenna, and (b) Radiated Electric Field. 17 Figure 2.9 Center-fed Slot Antenna Results. 18 Figure 2.10 Effect of the Offset Point dS on the Series Resistance of the Slot Antenna. 19 Figure 2.11 Normalized Impedance of th e Slot Referenced at the Edge of the Antenna. 20 Figure 2.12 ADS Schematic Utilized to Fi nd the Impedance of the Slot Antenna. 20 Figure 2.13 Normalized Impedance of th e Slot Referenced at the Slot. 21 Figure 2.14 Phase of the Slot Antenna at ds = 0, 5, 8, 10, and 12mm. 22

PAGE 9

vi Figure 2.15 Linearly Polarized Radia tion Pattern Obtain in Momentum. 23 Figure 2.16 CPW-fed Wide Slot Antenna with U-shaped Tuning Stub Layout. 27 Figure 2.17 (a) Current Density Distri bution of the Center-fed Wide Slot Antenna with U-shaped T uning Stub, and (b) 3D Radiated Electric Field. 28 Figure 2.18 (a) Wide Slot Antenna Return Loss, and (b) Wide Slot Antenna Phase. 29 Figure 2.19 Simulated VSWR of the CP W-fed Wide Slot Antenna with U-shaped Tuning Stubs. 30 Figure 2.20 Microstrip-fed Slot Antenna w ith a Plane Reflector for Unidirectional Radiation. 30 Figure 2.21 Reflector Spacing Effect on the Wide Slot Antenna RL. 31 Figure 2.22 Simulated Radiation Pattern as a Function of Reflector Spacing. 32 Figure 2.23 3D Radiated Field. 33 Figure 2.24 Photograph of the Narrow Slot Antenna Fabricated on a FR-4 Board. 34 Figure 2.25 Narrow Slot Antenna M easured vs. Simulated Data. 35 Figure 2.26 Measure vs. Simulated RL of the Tuned Narrow Slot Antenna. 36 Figure 2.27 Measured vs. Simulated Phase of the Narrow Slot Antenna. 36 Figure 2.28 Radiation Pattern of a Vertical Slot and Coordinate System. 37 Figure 2.29 (a) Photograph of the Narrow Slot Antenna E-plane Measurement Setup. 38 Figure 2.30 (a) Photograph of the Narrow Slot Antenna H-plane Measurement Setup. 39 Figure 2.31 Measured vs. Simulate d Cross-polarized E-plane. 40 Figure 2.32 Measured vs. Simulate d Cross-polarized H-plane. 40 Figure 2.33 Measured vs. Simulated Return Loss of the 2.5GHz Wide Slot Antenna with U-shaped Tuning Stubs. 41

PAGE 10

vii Figure 2.34 Measured vs. Simulated Wide Slot Antenna Impedance. 42 Figure 2.35 Measured vs. Simulated VSWR. 43 Figure 2.36 (a) Photograph of the Wide Slot Antenna E-plane Measurement Setup. 44 Figure 2.37 (a) Photograph of the Wide Slot Antenna H-plane Measurement Setup. 45 Figure 3.1 Typical Forward Current vs. Fo rward Voltage in Logarithmic Scale Representation. 49 Figure 3.2 Conventional RF Detector Circuit Schematic. 50 Figure 3.3 Step Impedance Low Pass Filter Momentum Layout. 53 Figure 3.4 Step Impedance Low Pass Filter Response. 54 Figure 3.5 Shorted Stub IL and RL. 55 Figure 3.6 Harmonics Generated by the R ectification Process at the Input of the Detector. 55 Figure 3.7 Harmonics Generated by the Rect ification Process at the Output of the Detector with Stub in Place. 56 Figure 3.8 Detector Input Port Interface. 57 Figure 3.9 Simulated Output Voltage vs. Input Power. 59 Figure 3.10 Simulated Detector Sensitivity. 59 Figure 3.11 (a) Optimized Detector Circuit Layout. 60 Figure 3.12 Via-less CPW-to-Mic rostrip Transition Sections. 62 Figure 3.13 CPW-to-Microstrip Transition Simulated Data. 63 Figure 3.14 2.5GHz Wide Slot Antenna and CPW-to-Microstrip Transition Integration. 64 Figure 3.15 Radiation Pattern Comparison Between the Simulated Wide Slot Antenna with and without the Transition in Place. 66

PAGE 11

viii Figure 3.16 Photograph of the Fabricated 2.5GHz Detector Circuit with SOLT and TRL Calibration Reference Plane. 67 Figure 3.17 2.5GHz Detector Circuit Si mulated vs. Measured RL Using Different Calibration Techniques. 68 Figure 3.18 Simulated vs. Measur ed Detector Performance. 70 Figure 3.19 Visual of the 2.5GHz Via-le ss CPW-to-Microstrip Transition. 71 Figure 3.20 Simulated vs. Measured CPW-to-Microstrip Transition. 72 Figure 3.21 (a) Visual of the Fabr icated Slot Antenna/Transition. 73 Figure 3.22 Simulated vs. Measured Radiati on Pattern for the Integrated Antenna. 75 Figure 3.23 Photograph of the Fully Integrated Rectenna Element. 76 Figure 3.24 Comparison Between the Measur ed Rectenna Element in a Single FR-4 Board vs. the Rectenna Element Connected by M-M Barrel. 76 Figure 4.1 Millimeter Wave Atmospheric Absorption for Horizontal Propagation [25]. 80 Figure 4.2 Atmospheric Attenuation Du e to Precipitation Rate [25]. 81 Figure 4.3 3D Radiation Pattern for the 94GHz Slot Antenna as a Function of Substrate Thickness. 87 Figure 4.4 94GHz Slot Antenna (a) Layout, (b) Simulated RL, and (c) Simulated Phase. 89 Figure 4.5 94GHz vs. 2.5GHz Simulated 2D Radiation Pattern. 90 Figure 4.6 94GHz Elements on a 10 m Silicon Membrane. 91 Figure 4.7 Visual of th e Fabricated Devices. 93 Figure 4.8 Close up Visual of: (a) the 94GHz Slot Antenna and (b) 20 m Silicon Membrane. 93 Figure 4.9 Measured vs. Simulated Ante nna RL; Obtained with the CS-5 Calibration Kit. 94

PAGE 12

ix Figure 4.10 RL of the Different Slot Antennas Measured with the CS-5 Calibration Kit. 95 Figure 4.11 IL of the Different Slot Antennas Measured with the CS-5 Calibration Kit. 95 Figure 4.12 RL of the Different Slot Antennas Measur ed with the On-wafer Calibration Kit. 96 Figure 4.13 (a) Top View of the Fabricated Substrate and (b) View of the Back Etched Membranes. 96 Figure 4.14 Close up of the Fabri cated Slot Antenna on top of ~5 m Silicon Membrane. 97 Figure 4.15 RL of the Different Slot Antennas Measured with the CS-5 Calibration Kit. 98 Figure 4.16 RL of the Different Slot Antennas Measur ed with the On-wafer Calibration. 98 Figure 4.17 Phase of the Different Slot Antennas Measured with the On-wafer Calibration Standards. 99 Figure A.1 Block Diagram for One-Po rt S-parameter Measurements. 110 Figure A.2 Block Diagram of a Full Tw o-Port S-parameter Measurements. 111 Figure A.3 Typical Block Diagram for Antenna Range Measurements. 112 Figure A.4 Square-law Detector Bl ock Diagram Measurement Setup. 112 Figure A.5 Block Diagram for the Inte grated Rectenna Element Power Measurements. 114 Figure B.1 IL of the SOLT Thru Standard. 117 Figure B.2 Phase Response of the SOLT Thru Standard. 118 Figure B.3 RL of the SOLT Load Standard. 118 Figure B.4 RL of the SOLT Open Standard. 119 Figure B.5 RL of the SOLT Short Standard. 119

PAGE 13

x Figure B.6 (a) Drawing of a Given TRL Standard Kit and Reference Plane. 120 Figure B.7 IL of the TRL Thru Standard. 122 Figure B.8 RL of the TRL Thru Standard. 122 Figure B.9 Phase of the TRL Thru Standard. 123 Figure B.10 Phase of the TRL Delay Standard. 123 Figure B.11 RL of the TRL Delay Standard. 124 Figure B.12 RL of the TRL Open Standard. 124 Figure C.1 Visual of the W band Measurement Setup. 125 Figure C.2 IL of the CS-5 Thru Standa rd (After Reference Plane Shift to Probe Tips). 126 Figure C.3 RL of the CS-5 Thru Standa rd (After Reference Plane Shift to Probe Tips). 126 Figure C.4 Phase of the TRL Thru Standard (After Reference Plane Shift to Probe Tips). 127 Figure C.5 RL of the CS-5 550 m Delay Standard (After Reference Plane Shift to Probe Tips). 127 Figure C.6 Phase of the CS-5 550 m Delay Standard (After Reference Plane Shift to Probe Tips). 128 Figure C.7 RL of the CS-5 Open Standa rd (After Reference Plane Shift to Probe Tips). 128 Figure C.8 Phase of the CS-5 Open Standard (After Reference Plane Shift to Probe Tips). 129 Figure C.9 Drawing of the On-w afer Calibration Standards. 130 Figure C.10 IL of the On-wafer Thru Standard. 130 Figure C.11 Phase of the On-wafer Thru Standard. 131 Figure C.12 RL of the On-wafer Thru Standard. 131

PAGE 14

xi Figure C.13 RL of the On-wafer Delay Standard. 132 Figure C.14 IL of the On-wafer Delay Standard. 132 Figure C.15 Phase of the On-wafer Delay Standard. 133 Figure C.16 RL of the On-wafer Open Standard. 133 Figure C.17 Phase of the On-wafer Open Standard. 134

PAGE 15

xii Investigation of the Rectenna Concept for Millimeter Wave Applications Henrry La Rosa ABSTRACT This thesis presents the rectifying antenna potential for infrared and millimeter wave energy conversion. Infrared imaging is one of the emerging techno logies that have attracted considerable attention in the next generation of military, medical, and commercial applications. Moreover, with the ever-increasing congestion of the electromagnetic spectrum at RF and microwave frequencies and the establishment of firm civilian and military requirements best met by millimeter wave systems, the interest in the technology has grown and is now firmly established. During this work a 2.5GHz slot antenna, a 2.5GHz Schottky diode detector, a CPW-to-Microstrip transition, a fully integrated Rectenna element, and a 94GHz sl ot antenna were desi gned, fabricated, and tested. Results on the perfor mance of the devices show a great deal of correlation between the simulated and measured data. To perform an initial study, the CPW-fed na rrow slot antenna is designed at 2.5GHz and implemented on an FR-4 board. This in vestigation serves as the basis for the development of the Rectenna element at millimet er wave frequencies. In order to increase the bandwidth of the slot antenna, a 2.5GHz CPW-fed wide slot antenna with U-shaped tuning stub is realized, which provides a 60% increase in bandwidth while keeping the same radiation characteristics. In addition, a set of simulations is performed to show how

PAGE 16

xiii a reflector plate affects the ra diating properties of the slot antenna. A 2.5GHz square-law detector is also designed, fabr icated, and tested in order to rectify the RF signal delivered by the antenna. The fabricated detector pres ents a well matched condition at the design frequency with a dynamic range found to be from dBm to dBm. The low frequency Rectenna element prototype is then integrated within a single FR-4 board. This is accomplished by implementing a compact vialess CPW-to-Microstrip transition. Finally, a 94GHz CPW-fed wide slot antenna is realized on a 10m high resistivity silicon membrane. This antenna shows a great deal of similarity to the 2.5GHz slot antenna. This low profile ante nna presents at least a 10dB return loss over the entire W band frequency window. Simulated antenna efficiencies of up to 99% were achieved assuming a perfect conductor.

PAGE 17

1 Chapter 1 Introduction 1.1 Overview The purpose of this thesis work is to aid in the development of an un-cooled antenna coupled Metal-Insulato r-Metal (MIM) diode detector also known as a Rectenna for the purpose of infrared and millimeter wave ener gy conversion. To date, infrared imaging has being accomplished by quantum and therma l detectors. Quantum detectors provide extremely high sensitivity and fast response; however they require cryogenic cooling mechanisms, which increase the detectors complexity and cost. Thermal detectors, on the other hand, use micro-bolometers, which ar e simply structured devices with room temperature operation. The downside of microbolometer is that it suffers from low sensitivity and slower response than quantum detectors. The proposed Rectenna concept presented here upholds all the advantages of the micro-bolometer described above, while providing a much faster response and higher sensitivity. R. Baily first proposed the Rectenna c oncept in 1972 for solar energy conversion [ 1], [ 2]. Baily suggested that by using an ante nna array in conjunction with a high-speed rectifier, solar radiation could be capture d and efficiently conve rted into DC power providing efficiencies greater than 85%. Due to fabrication restrictions and to the limited understanding of material properties at light frequency, focus was placed in the

PAGE 18

2 development of an un-cooled detector for infrared and millimeter wave applications using the same concept proposed by Baily. Nahas [ 3] developed the first Rectenna elemen t model consisting of a dipole antenna, a step im pedance filter, an extremely fast rect ifier, and an output DC filter. This Rectenna element model is presented in Figure 1.1. Low-Pass Filter DC Filter Antenna Rectifier Load Figure 1.1: Nahas Model for Microw ave-to-DC Energy Conversion [3]. In Figure 1.1 it can be observed that the antenna acts as a transducer between free s pace and the detector circuit by intercepting an incident electromagnetic wave within a frequency range and efficiently delivering an RF signal to the detector circuit. The step impedance filter prevents the higher order ha rmonics generated by the detector from reradiating through the antenna, thus increasing the amount of power that is delivered to the load. Nonlinear devices such as Schottk y-barrier diodes are responsible for the rectification process, and the output DC f ilter provides a DC path to the load by separating the high frequency comp onents from the DC signal. A 2.5GHz prototype of the Rectenna element was initially characterized in order to get familiar with the system and find possibl e solutions at millimeter and infrared frequencies. During the course of this thesis work three planar antennas, a Schottky diode

PAGE 19

3 detector, a CPW-to-Microstrip transition, a fu lly integrated Rectenna element, and a 94GHz slot antenna were implemented. Thes e elements presented good agreement with the expected results. This was accomplished by dividing the project into four different sections. First a 2.5GHz CPW-fed narrow slot antenna was optimized so that it provided extremely high efficiencies. A slot antenna desi gn was chosen due to its similarities to the conventional dipole with the slot antennas el ectric field bei ng orthogonal to the dipoles electric field. In addition, integrating tuning stubs within the aperture can easily increase the bandwidth of the antenna while maintaining the same radiation characteristics. Furthermore, slot antennas co uld be used to achieve both unidirectional and bidirectional radiation. Unidirectional radiation is achieved by placing a reflector plate parallel to the slot antenna on the opposite side of the substrate. In order to complete the Rectenna elemen t, a square-law detector circuit using a commercially available Schottky diode was de signed, fabricated, and tested. A squarelaw detector essentially serves as a power-m easuring device with multiple applications in RF circuitry. These detectors are commonly used in radiometers, power-scavenging systems, RFID tags, and in numerous other mobile applications. The 2.5GHz detector circuit was designed using the HSMS-2850 Sc hottky diode to rectify the RF signal delivered by the antenna. In order to impl ement the Rectenna el ement within a single board, a via-less CPW-to-Microstrip transiti on was employed. Transitions are commonly used to match the propagation modes of di fferent transmission li nes or waveguides in order to minimize reflections and power loss at a given interface. After getting acquainted with the problems at lower frequencies, a 94GHz wide slot antenna with tuning stubs was characterized. The antenna was fabricated on a 10m high

PAGE 20

4 resistivity silicon membrane that serves to prevent surface waves from propagating through the substrate, thus incr easing the overall efficiency of the antenna. It will be shown that even though the jump in freque ncy is extensive, the performance of the antenna is comparable to th at of the 2.5GHz antenna. 1.2 Motivation The Rectenna concept has great potential no t only for millimeter wave applications and infrared energy conversion, but also for efficiently collecti ng solar radiation and converting it into DC power. Theoretically, by utilizing the Recte nna concept conversion efficiencies approaching 100% have been predicted [ 4], which surpass the current state of the ar t solar cells effici encies. In literature [ 5], efficiencies of up to 85% have been reported at lower frequ encies. The successful development of a Rectenna element at millimeter wave frequencies can also reduce the overall cost of millimeter wave sensory systems. This device can also provide multiple applications for military and commercial industries such as: concealed weapon imaging, automatic braking, radiometry, and missile guidance systems. 1.3 Thesis Organization This thesis work is divided into five ch apters. The first chapter consists of the introduction, composed of the motivation behind the project, thesis organization, and major contributions. Chapter 2 presents the design of two 2.5GHz CPW-fed slot antennas. An in-depth analysis of slot ante nnas is provided along with a brief theoretical description of slot antennas. The effects of a reflector plate on the impedance and radiation characteristics of the antenna are also analyzed through a set of Agilents Momentum simulations. The optimiza tion of a conventional CPW-fed slot antenna

PAGE 21

5 that achieves efficiencies of up to 99% is also presented. In order to increase the bandwidth of the antenna, a CPW-fed wide sl ot antenna with U-shaped tuning stubs is implemented. Chapter 3 shows the design of a conventiona l Schottky diode dete ctor circuit with distributed elements as the matching network and filtering mechanism. A sensitivity analysis of the detector ci rcuit is also performed in order to demonstrate how the matching network affects the overall performance of the square -law detector. In addition, a CPW-to-Microstrip transition is employed in order to integrate the antenna and the detector circuit within a singl e FR-4 board. The separate Re ctenna elements connected by a Male-to-Male (M-M) connector are also anal yzed and compared to the fully integrated Rectenna element by measuring its square-law characteristics and sensing properties. Finally, an in-depth comparison between the s imulated and measured data is presented. Chapter 4 of this thesis discusses the de sign, fabrication, and characterization of a 94GHz slot antenna. This antenna was designed on top of a 10m high resistivity silicon membrane and presented simulated efficiencies of up to 99%. A brief description of surface waves within the silicon substrate al ong with a parametric study performed in Momentum to study the effects of the subs trate thickness on th e antenna radiation properties is provided. 1.4 Contributions Many contributions were made during the cour se of this thesis work. For instance, a 2.5GHz Rectenna element was successfully in tegrated within a single FR-4 board and tested for its sensing properties. In the pr ocess, a compact via-less CPW-to-Microstrip transition was implemented presenting extrem ely low loss at the design frequency of

PAGE 22

6 2.5GHz. In addition, a 94GHz wide slot antenna was fabricated on a 10m high resistivity silicon membrane and measured to validate its performance over the entire W band frequency window. To the best of the authors knowledge this is the first time this specific antenna has been implement on a 10m high resistivity silicon membrane.

PAGE 23

7 Chapter 2 2.5GHz Slot Antenna Design 2.1 Introduction Aperture antennas have been studied for the past fifty years becoming a separate research area of their own [ 6]. Throughout the years many c onfigurations have been realized such as: square, rect angular, circular, and elliptical apertures just to m ention a few. Aperture antennas are also very practi cal for space applications, since they can be flush-mounted on the metallic surface of a spacecr aft or aircraft. In addition, strip and slot combinations offer an additional degree of freedom in the design of microstrip antennas. This thesis work mainly concentrates on slot antennas and their characteristics. Throughout this chapter, a slot antenna is characterized, which is the first component of the Rectenna element. A brief theoretica l description of slot antennas and the terms used to describe them are presented. In addition, the antenna properties are shown by presenting two low frequency prototype ante nna designs operating at 2.5GHz. Also, some of the advantages of using a CPW-fed slot antenna over patch and microstrip-fed slot antennas are mentioned in Section 2.3. Finall y, the optimized antenna prototypes are fabricated and tested showing a great deal of correlation to the simulated data. 2.1.1 Background Theory A slot antenna consists of a slot cut in a metal surface with an orthogonal feeding mechanism in most cases. The slot dimensi ons are usually smaller than a wavelength and

PAGE 24

8 much smaller than the ground pl ane in which they reside. Figure 2.1 illustrates a conventional center-fed slot antenna presented by Garg et al. [ 7]. The slot antenna can be efficiently excited by physically shorting the m i crostrip line through the substrate into the far edge of the slot as illustrated in Figure 2.1a. The most commonly used feeding topology is presented in Figure 2.1b. In this case, an open-circuited stub is terminated approxim ately at a from the edge of the slot, which pr esents an electrical short circuit at the edge of the antenna. Other fe eding techniques have been utilized [ 6-12] such as the one m entioned subsequently in this chapter. Lm Slot Microstrip Figure 2.1: (a) Center-fed Slot Antenna Short Circuited Through the Substrate. (b) Microstrip Slot Antenna Terminat ed with an Open-circuited Stub. The microstrip conductor excites the slot by coupling a voltage potential across the aperture forming a distributed electric fiel d. Within the aperture, the voltage maximum and current minimum occur at the center of the slot, while the voltage minimum and current maximum take place at the two edges of the slot. The fringing fields, which are R a d i a t i n g S l o t hD i e l e c t r i c S u b s t r a t e S h o r t i n g S t r i p M i c r o s t r i p L i n e G r o u n d P l a n e e r(a) (b)

PAGE 25

9 caused by current crowding at the edges of the slot, increase the electrical length of the antenna [ 6]. Thus, the ph ysica l length L of the slot antenna is denoted by the equation 2.1 (2.1) where S is the wavelength in the slot and l is the electrical length caused by the currents flowing around the edges of the slot [ 6]. The current and vol tage distribution of the s lot antenna is presented in Figure 2.2, which is key to understanding som e of the impedance matching techniques discusse d in the subsequent sections. W L Vmax, IminVmin, Imax Figure 2.2: Voltage and Current Distribution Around the Slot. 2.1.2 Input Impedance According to Bahl and Bhartia [ 6], the input impedance of the antenna consists of a series com bination of the radiation resistance R and the reactive component X At resonance the reactive component of the impeda nce is taken to be zero; leaving a purely real component. Literature has reported that the impedance of the resonant center-fed slot antenna is approximately 550 which is far from being matched to the feeding mechanism. It should be mentioned that th is high impedances depend on the substrate characteristics. Due to the voltage and current relationship presented in Figure 2.2, the im pedance of the antenna decreases towards the edges of the slot, and in most cases, different feeding techniques can be used to match the antennas input impedance to the lLS 2

PAGE 26

10 characteristic impedance of the line, whic h usually increases the bandwidth of the antenna [ 6]. Figure 2.3 presents some of the feeding techniqu es used in [ 8], [ 9] to reduce the antennas resistance. Figure 2.3: Different Feeding Methods Us ed to Decrease the Slot Antenna Input Impedance. (a) Offset Micros trip-fed and (b) Center-fed but Inclined Microstrip-fed. As suggested by Yoshimura [ 8], the radiation resistance of the slot antenna can be reduced by feeding it at a point away from th e center of the slot, which is also known as the offset-fed slot antenna illustrated in Figure 2.3a. For most slot antennas, the im pedance mainly depends on the width of the sl ot, the location of the feed point, the size and width of the tuning stubs, and the diel ectric material used. The other feeding technique is depicted in Figure 2.3b, which consists of incl ining the slot about its axis. Yoshim ura [ 8] also stated that the resonant length of the center-fed slot antenna is greater than the resonant length of the offset-fed slot. As a rule of thumb, the resonant length of the offset-fed slot antenna usually lies between .4 and .5 [ 6]. The other technique used to reduce the im pe dance of the antenna consists of stub tuning the slot antenna as suggested by Pozar [ 10]. This technique is similar to that of Figure 2.1b except that the length of the stub is no longer term inated at a from the (a) (b)

PAGE 27

11 edge of the slot. The microstrip tuning st ubs change the resonant frequency of the antenna so that the input resistance matche s with the feed line impedance at the new resonant frequency [7]. The stub can be tuned to cance l out the reac tive part of the impedance and reduce the radiation resistan ce of the antenna as well. In addition, Yoshimura [ 8] declared that the distortion cr eated at the feeding m echanism has minimum or no effect on the far field radiation pattern of the antenna. 2.1.3 Radiation Pattern The easiest way to demonstrate the radiation characteristics of the slot antenna is to consider it as a dual value pr oblem to the dipole antenna case. As illustrated in Figure 2.4, both antennas are excited at their center which is represented by a voltage source. 2 2 Figure 2.4: (a) Slot Antenna on an Infinite Ground Plane, (b) Complementary Printed Dipole Antenna. Booker [ 11] stated that the radiati on pa ttern of the slot is the same as the complementary half wavelength printed dipole represented in Figure 2.4b, with their fields being orthogonal to each other. The two c ases can be treated as a boundary-value problem by finding the appropriate solution to Maxwells equations while satisfying the proper boundary conditions. Where, E and H are the electric and magnetic fields of the aperture respectively. (a) (b)

PAGE 28

12 (2.2) (2.3) Either of the above wave equations can be solved to satisfy the following boundary conditions. For the dipole depicted in Figure 2.4b, the tangentia l components of the m agnetic field are zero outside the perimeter of the dipole, and the normal components of the magnetic field are zero within the pe rimeter of the dipole. For the slot in Figure 2.4a, the tangential com ponents of the electric field are zero outside the perimeter of the slot and the normal components of the electric field are zero within the perimeter of the slot. From the above boundary conditions it is observed that both problems are similar, differing only by the fact that the electric a nd magnetic fields are interchanged. This is illustrated in Figure 2.5. It should be noted th at if the slot antenna in Figure 2.5a is rotated by 90 abou t the y-axis the electric and magnetic fields vectors are aligned in the same direction. E H E H E E H H Y Z X E H E H E H Y Z X H E E H Figure 2.5: (a) Radiated Fiel ds of a Slot Antenna on an Infinite Ground Plane, (b) and Radiated Fields of the Complementary Dipole Antenna. (a) (b) EE2HH2

PAGE 29

13 The electric field for the slot ante nna has the same doughnut shape of a dipole antenna. Also, since the electric and magnetic fields of both antenna s are interchangeable, the slot antennas electric and magnetic fiel ds can be expressed as a function of the magnetic and electric fields of the di pole antenna. This is given by 2.4 and 2.5, (2.4) (2.5) where ES and HS are the electric and magnetic fields in the slot; and Ed and Hd are the electric and magnetic fields in the dipole respectively [ 6]. For a full derivation of the slot antennas electric and m agnetic fields refer to [6], [ 7], and [ 8]. Slot antennas can present both linear and circular polarization. For the narrow slot antenna depicted in Figure 2.6, the radiation normal to the ground plane of the horizontal slot in F igure 2.6a is vertic ally polarized while the radiation normal to the ground plane of the vertical slot presented in Figure 2.6b is horizontally polarized. Figure 2.6: Slot Antenna Fed by a Microstrip Line with (a) Vertical Polarization, and (b) Horizontal Polarization. (a) (b) d SHkE1 d SEkH2

PAGE 30

14 2.2 Motivation for Using a CPW-fed Slot Antenna CPW-fed slot antennas have many advantag es over other microstrip radiators. For instance, slot antennas have the advantage of providing bidirectional and unidirectional radiation. Unidirectional radiat ion is achieved by placing a re flector plate parallel to the substrate. In Section 2.5 of this chapter a se t of simulations were performed in Agilents electromagnetic simulator Momentum to demonstrate the reflec tor spacing effect on the impedance and radiation characteristics of the antenna. Table 2.1 summarizes some of the characteristics and advantages of slot antennas over patch antennas [ 7]. Table 2.1: Com parison Between the Slot and Patch Antenna Characteristics. Characteristics Patch Slot Analysis and design Easy Easy Fabrication Very easy Very Easy Tolerance in fabrication Critical Not very critical Profile Thin Thin Shape flexibility Any shape Limited Radiation fields Unidirectional Unidirectional and bidirectional Polarization Linear and circular Linear and circular Bandwidth Narrow Wide Dual frequency operation Possible Possible Spurious radiation Moderate Low Isolation between radiating Fair Good elements Frequency scanning Easily possible Possible Cross-polarization level Low Very low End-fire antenna Not possible Possible

PAGE 31

15 The CPW-fed slot antenna was chosen for this project because it provides many advantages over the conventional microstrip-f ed slot antenna. For instance, the CPW-fed slot antenna does not require alignment since th e entire structure is fabricated on one side of the substrate. This feeding technique redu ces the fabrication erro r tolerance found in microstrip and coax fed slot antennas. In addition, CPW lines can be easily integrated with monolithic components within the same surface reducing the overall cost of the device. 2.3 CPW-fed Narrow Slot Antenna In this section an experimental analysis on an offset CPW-fed narrow slot antenna is presented. Some of the feeding topologi es published in the literature [ 7], [ 12] are depicted in Figure 2.7. In the center-fed CPW slot antenna illustrated in Figure 2.7a the two electric field com ponents of the CPW aper ture excite the slot antenna across its width. As mentioned by [7], the CPW apertures radiated fields almost add out of phase in the far field providing extrem ely low cr oss-polarization. The same feeding technique used to reduce the radiation resistance of th e antenna also applies to the CPW-fed slot antenna as shown in Figure 2.7b. The problem with this feeding mechanism is that it requires wire bonding in order to stabili ze the ground plane and suppress higher order modes that can provide cross-polarization. The other feeding m echanisms are depicted in Figure 2.7c and Figure 2.7d. The inductively coupled feeding m ethod is found to be extremely useful for series-f ed array configurations [ 7].

PAGE 32

16 Figure 2.7: CPW-fed Slot Ante nna Feeding Mechanisms: (a) Center-fed Slot Antenna, (b) Offset-fed Slot Antenna with Bond Wire for Stability, (c) Capacitively Coupled, and (d) Inductively Coupled Slot Antenna. The slot antenna design mentioned in this section was characterized using Momentum. First, a 2.5GHz center-CPW-fed slot antenna was simulated to analyze its characteristics and to become familiar with the current distribution around the antenna. As illustrated in Figure 2.8, the electric current density is more intens e at the edges of the slot, which agrees well with literature. In addition, the radiation pattern possesses the sam e doughnut shape as that of the dipole antenna and presents a high degree of symmetry about its axis. (a) (b) (c) (d)

PAGE 33

17 Figure 2.8: (a) Current Density Distribution of the Center-fed Slot Antenna, and (b) Radiated Electric Field. The resonant length of the antenna is found to be approximately 40mm with a complementary width of 4mm. An FR-4 board ( r = 4.3) was chosen for this design. Figure 2.9 illustrates the simulated results of the center-fed slot antenna. The resonance occurs at approxim ately 2.5GHz, but its impe dance is far from being matched to the characteristic impedance of the CPW line. As mentioned before, the radiation resistance of the slot antenna can be matched to a 50 transmission line by feeding it at a certain distance dS away from the center of the slot. This feeding technique is illustrated in Figure 2.3a. A parametric study was performed on Mom entum to find the optimum offset point. The distance of the offset point was varied from dS=0mm to dS=12mm, and the impedance of the antenna was monitored. The simulated results are depicted in Figure 2.10. (a) (b)

PAGE 34

18 0 2 4 6 8 10 12 2.0 2.3 2.5 2.8 3.0 Freq (GHz)Series Resistance -200 -100 0 100 200 2.02.32.52.83.0 Freq (GHz)Phase (degrees) Figure 2.9: Center-fed Slot Antenna Results. (a) Series Resistance and (b) Phase. (a) (b)

PAGE 35

19 0 10 20 30 40 50 60 70 80 2.0 2.3 2.5 2.8 3.0 Freq (GHz)Series Resistance ds=0 ds=5mm ds=8mm ds=10mm ds=12mm Figure 2.10: Effect of the Offset Point dS on the Series Resistance of the Slot Antenna. It is observed that as the feeding point a pproaches the edge of the slot antenna the aperture impedance increases. The plotted impedance illustrated in Figure 2.10 contrad icts the impedances reported in literature [ 7]. This is due to the fact that the sim ulated impedance is referenced at the edge of the antenna, which is approximately a away from the slot. Figure 2.11 presents the normalized impedance of the antenna in a Sm ith Chart. From this plot it is evident th at the impedance at the center of the slot is extremely low and increases as the feeding poi nt moves towards the edge of the antenna.

PAGE 36

20 freq (2.000GHz to 3.000GHz)S(1,1) S(2,2) S(3,3) S(4,4) S(5,5) Figure 2.11: Normalized Impedance of the Slot Referenced at the Edge of the Antenna. In order to find the actual impedance of the slot, the simulated S-parameters were extracted from the Momentum simulation and imported into Agilents Advanced Design System ( ADS ). Then, a CPW line was added to obtain the actual impedance of the aperture. The complete schematic utilized to shift the reference plane to the edge of the slot and the simulated re sults are depicted in Figure 2.12 and Figure 2.13. CPW CPW2 Temp= L=18.5 mm G=.3393 mm W=3.0139 mm S1P SNP5 File="Test2_offset_By_5mm_mom_a.ds" 1R e f Term Term2 Z=50 Ohm Num=2 Figure 2.12: ADS Schematic Utilized to Find the Impedance of the Slot Antenna. dS = 0mm dS = 5mm dS = 8mm dS = 10mm dS = 12mm

PAGE 37

21 freq (2.000GHz to 3.000GHz)S(1,1) S(2,2) S(3,3) S(5,5) S(4,4) Figure 2.13: Normalized Impedance of the Slot Referenced at the Slot. As illustrated in Figure 2.13, the S-parameters are rotated on the Smith Chart by placing the reference plane at the edge of the slot. The results show that the slot impedance is very high when it is center-fed and drops as it approaches the edge of the aperture, which agrees well with literature. Th e actual series resistance of the center-fed slot antenna was found to be approximately 420 and the optimal feeding point was found to be at dS = 8mm. As mentioned before, the re sonant length of the offset slot antenna is smaller than the center-fed slot antenna. Figure 2.14 illustrates this behavior. dS = 12mm dS = 10mm dS = 8mm dS = 5mm dS = 0mm

PAGE 38

22 2.22.42.62.8 2.0 3.0 -100 0 100 -200 200 freq, GHzphase(S(1,1)) phase(S(2,2)) phase(S(3,3)) phase(S(4,4)) phase(S(5,5)) Figure 2.14: Phase of the Slot Antenna at dS = 0, 5, 8, 10, and 12mm. The downside of the offset-CPW-fed narrow slot antenna is that it requires wire bonding in order to balance th e CPWs ground plane. The failu re to balance the ground planes can create other operating mode s causing non-desired cross-polarization. Figure 2.15 presents a com parison between the center a nd the offset-fed slot antennas simulated radiation pattern. It is observed that the simulated radiation pattern of the center-CPWfed slot antenna possesses extremely low cr oss-polarization, while the offset-CPW-fed slot antenna provides a cross-po larization that is 10dB below the co-polarized pattern. It should be mentioned that the offset-CPW-fed slot antenna was simulated without a wire bond. ds = 0mm ds = 5mm ds = 8mm ds = 10mm ds = 12mm

PAGE 39

23 Center-Fed Slot -60 -50 -40 -30 -20 -10 0 10 -180-90 0 90 180 Theta (degrees)Mag (dB) E-co E-Cross Offset-Fed Slot -60 -50 -40 -30 -20 -10 0 -180-90 0 90180 Theta (degrees)Mag (dB) E-co E-cross Figure 2.15: Linearly Polarized Radiation Pa ttern Obtain in Momentum. (a) Center-fed Slot Antenna, (b) Offset-fed Slot Antenna. Table 2.2 also presents a comparison of th e center and offset-f ed slot sim ulated antenna parameters. (a) (b)

PAGE 40

24 Table 2.2: Comparison Between the Center -fed and Offset-fed Simulated Antenna Parameters. Simulated Antenna Parameter Center-fed Offset-fed Power Radiated (mW) 1.2 2.5 Directivity (dB) 4.73 4.68 Gain (dB) 4.71 4.65 Efficiency 99% 99% 2.4 Antenna Efficiency and Gain Definitions According to Balanis [13], the overall efficiency 0 of the antenna is provided by equation 2.6. (2.6) where: r is the reflection (mismatch) efficiency = 21 c is the conduction efficiency d is the dielectric efficiency and is the voltage reflection coefficient at the input terminals of the antenna given by equation 2.7. (2.7) where ZIN is the input impedance of the antenna, and Z0 is the characteristic impedance of the line. Since the dielectric and conduction efficiencies are somewhat cumbersome to calculate and cannot be obta ined separately from meas urements, they are usually combined to give the antenna radiation efficiency cd = cd.. This equation can be used to relate the gain and the directivity of the antenna. Balanis [ 13] stated that according to the IEEE standards, gain does not include losses arising from impedance mismatches (reflection losses) and polarization dcr 0 0 0ZZ ZZIN IN

PAGE 41

25 mismatches (losses). The gain of an antenna is defined as the radiation intensity radiated in a certain direction compared to an isotropic radiator. An isotropic radiator is defined as a hypothetical lossless antenna having equal radiat ion in all directions. In practice, relative gain is used, and is defined as the ratio of the power gain in a given direction to the power gain of a reference antenna in its referenced direction [ 13]. If the direction is not specified, the power gain is m easured in the direction of maximum radiation. Equations 2.8 and 2.10 illustrate how the gain and directivity of the antenna are related by the efficiency of th e antenna. The gain of the antenna G( ) can be expressed as (2.8) where (2.9) and the directivity is given by equation 2.10. (2.10) Now if the equation 2.10 is substituted into eq uation 2.8, it is observed that the gain and the directivity are related by the efficiency of the antenna expre ssed in equation 2.11. (2.11) In addition, the maximum gain of the ante nna is directly related to the maximum directivity by finding the maximum radiation intensity of the antenna. (2.12) rad cdP U G ),( 4),( Pin Pcd rad radP U D ),( 4),( ),(),( D Gcd 0 0),( ),( D D GGcd MAX cd MAX

PAGE 42

26 To calculate the overall gain of the antenna, which is defined as the absolute gain of the antenna G0abs [ 13], the transmission line mismatch and polarization losses have to be included in the ana lysis. If the antenna is perfectly matched to the feeding mechanism, the absolute gain is equal to the relative gain of the antenna. From the above expressions and the simulated antenna parameters, efficien cies of up to 99% were achieved under the assumption that copper is a perfect conductor. 2.5 CPW-fed Wide Slot Antenna with Tuning Stubs For some applications wide operational bandwidths are desired. The proposed antenna presented in this section provides many advantages over the narrow slot antenna presented in the previous section. These incl ude wider bandwidth, lo w cross-polarization, and added tuning capabilities. In addition, wide slot antennas are less sensitive to fabrication errors and alignment issu es found in narrow slot antennas [ 6]. The antenna presented in this section is composed of a CP W-fed wide slot utiliz ing a U-shaped tuning stub to achieve broadband operation. The bandwidth of an antenna can be calculated in one of two ways. Let Uf and Lf be the upper and lower frequency of operation, respectively, in which the impedance or other performance characteristics of the antenna do not change dramatically. The impedance bandwidth is usually measured at a VSWR 2 or a RL 10dB which is equivalent. In most cases the antennas ba ndwidth is represented as a percentage expressed by equation 2.13 (2.13) 100 x f ff BWC LU

PAGE 43

27 where Cf is defined as the center frequency of op eration or as the design frequency, in this case 2.5GHz. According to [ 14], an antenna is said to be broadband if its impedance does not cha nge significantly over about an octave ) 2/( LUff or more. Wide slot antennas inherently present wi der bandwidths than narrow slot antennas, and the addition of tuning stubs creates multiple resonances which increases the overall bandwidth of the antenna. Throughout the literature [ 15], [16], [ 17] many stub tuning techniqu es have been realized providing impedance bandwidths of about 35% to 60%. Chair [ 18] reported a CPW-fed wide slot an tenna with U-shaped tuning stubs in conjunction with a back reflector for ultrawideband operation that presented impedance bandwidths of up to 120% measured at a RL 10dB. Even though this project does not require ultra-wideband operation, the same tuning topology presented in [ 18] was realized. Figure 2.16 illustrates the proposed antenna and slot dim ension relative to a wavelength. Figure 2.16: CPW-fed Wide Slot Ante nna with U-shaped Tuning Stub Layout. Slot L = ~ Metal GND Copper

PAGE 44

28 The antenna depicted in Figure 2.16 presents the same curre nt and voltage distribution as the narrow slot antenna. In addition, this antenna provides the sam e doughnut shaped radiation characteristics found in the narrow slot antenna and the dipole depicted in Figure 2.17b. Figure 2.17: (a) Current Density Distribution of th e Center-fed Wide Slot Antenna with U-shaped Tuning Stub, and (b) 3D Radiated Electric Field. As mentioned before this antenna has many advantages over the narrow slot antenna. For instance, a matched condition is achieved w ithout the need of an offset feed point providing extremely low cross-polarization. Th e antennas resistance is mainly controlled by the width of the slot, the length and width of the tuning stub, the location of the stubs, and the dielectric material. The simulated RL and phase of the antenna are presented in Figure 2.18, and it is important to mention that the reference plane in the simulations was placed at th e edge of the antenna. The antenna is well matched over a large frequency range. It should also be noted that the antenna presents multiple resonances. These resonances are mainly due to the length of the slot, the position of the stubs, and the length of the stubs increasing the overall bandwidth of the antenna. (a) (b)

PAGE 45

29 -50 -40 -30 -20 -10 0 1.52.02.53.03.54.0 Freq (GHz)|S(1,1)| (dB) Simulated (a) -200 -100 0 100 200 1.52.02.53.03.54.0 Freq (GHz)S(1,1) Phase (degrees) Simulated (b) Figure 2.18: (a) Wide Slot Antenna Return Loss, and (b) Wide Slot Antenna Phase. Even though the focus of the antenna design was for it to be well matched at the design frequency rather than an ultra-wideband operation, broadband operation was achieved with bandwidths of up to 70% referenced to a VSWR 2, which is depicted in Figure 2.19.

PAGE 46

30 0 1 2 3 4 1.52.02.53.03.54.0 Freq (GHz)VSWR Simulated Figure 2.19: Simulated VSWR of the CP W-fed Wide Slot Antenna with U-shaped Tuning Stubs. 2.6 Antenna Simulations with Reflector in Place As mentioned before, unidirectional radiat ion can be achieved by placing a reflector plate parallel to the substrate as illustrated in Figure 2.20. Figure 2.20: Microstrip-fed Slot Antenna with a Plane Reflector for Unidirectional Radiation.

PAGE 47

31 In this section the effect of the reflector plate spacing on the impedance and radiation characteristics of the antenna is presented. According to [ 6] and [ 7], the RL of the antenna can change dramatically when the antenna to reflector plate spacing approaches and hardly changes as the reflector plate spacing approaches 1 A parametric study to demonstrate this be havior was performed in Mome ntum. The reflector spacing dr was varied from 1/8 to 1 while monitoring the RL and the radiation pattern of the antenna. Figure 2.21 presents the simulated results of this study. -60 -50 -40 -30 -20 -10 0 1.01.52.02.53.03.54.0 Freq GHzS11 (dB) Figure 2.21: Reflector Spacing Effect on the Wide Slot Antenna RL. The results presented in Figure 2.21 show that the opt im um reflector plate spacing needed to maintain the same return loss as that of the slot ante nna without a reflector plate lies between and 1. On the other hand, the radi ation pattern of the antenna achieves the optimal front-to-back ratio of 10dB and minimum side lobes at a reflector spacing of approximately as illustrated in Figure 2.22. dr =1/8 dr =1/4 dr =1/2 dr =3/4 dr =1 No Reflector

PAGE 48

32 Reflected Radiation Pattern 180 0 270 90 -60 -50 -40 -30 -20 -10 0 No Reflector dr =1/4 dr =1 Figure 2.22: Simulated Radiation Pattern as a Function of Reflector Spacing. The amount of energy that is reflected is directly proportional to the size of the reflector. Figure 2.23 shows a performance comparis on between a slot antenna with a finite and o ne with an infin ite reflector plate placed at away from the antenna. As expected, the 3D radiation plot of the antenn a with the infinite gr ound plane eliminates the back lobe associated with the antenna that has a finite reflector plate and concentrates all the energy toward the top of the antenn a. A trade-off must be made between the reflector size and front-to-back lobe ratio. Once the reflecto r plate design is completed, the antenna dimensions must be optimized in order to obtain the desired bandwidth and return loss.

PAGE 49

33 Figure 2.23: 3D Radiated Fiel d. (a) Slot Antenna with a Finite Reflector Plate, and (b) Slot Antenna with Infinite Reflector Plate. Table 2.3 presents simulated antenna character istics of the slot antenna with different reflector plate sizes placed at a Table 2.3: Comparison Between the Simula ted Antenna Parameters with no Reflector, Finite Reflector, and Infi nite Reflector Placed at dr = Antenna Parameter No Reflector Finite dr =1/4 Infinite dr =1/4 Power Radiated (mW) 2.33 1.95 3 Directivity (dB) 5.29 8.6 8.07 Gain (dB) 5.15 8.2 8.05 Efficiency (%) 96.8 91.2 99 Back lobe Attenuation (dB) 0 10 From Table 2.3, efficiencies of 91.2% or greate r can be achieved in the three cas es with the assumption that the ground plane is a pe rfect conductor. In addition, it is observed that the back lobe attenuation mainly depends on the reflector plate size and spacing. Also, it can be noted that the gain and direc tivity of the antenna almost double with the addition of a reflector plate. (a) (b)

PAGE 50

342.7 Result and Comparison 2.7.1 2.5GHz Narrow Slot Antenna The optimized CPW-fed narrow slot antenna was fabricated on a FR-4 board using the LPKF ProtoMat 91s milling tool and measured with the HP-8753D Vector Network Analyzer (VNA). Conventional SOLT calibration standards were utilized to remove the systematic errors from the measured return loss (RL). For calibration results and standard definitions refer to Appendix C. Figure 2.24 illustrates a phot ograph o f the fabricated antenna, and Figure 2.25 shows a comparison between the measured and simulated RL. The f requency shift between the simulated and measured data is due to variation of dielectric constant r across the FR-4 substrate and the fact that the SMA connector parasitics were not taken into account in the simulation. Figure 2.24: Photograph of the Narrow Sl ot Antenna Fabricated on a FR-4 Board.

PAGE 51

35 -40 -30 -20 -10 0 1.52.02.53.03.54.0 Freq (GHz)S11 (dB) Measured Simulated Figure 2.25: Narrow Slot Antenna Measured vs. Simulated Data. Copper tape was utilized to physically tune the antenna and compensate for fabrication errors, r tolerance within the substrate, and the SMA connector. The re-tuned antenna measurements are illustrated in Figure 2.26 and show a great deal of correlation to the sim ulated data. The measured and simu lated RL were found to be 36.5 and 33dBm at 2.51GHz and 2.48GHz respectively with an impedance bandwidth of 10% measured at a VSWR 2. The gain of the antenna was measur ed to be 4.5dB, which compares to the simulated gain of 4.65dB. The measured an d simulated phase of the tuned narrow slot antenna is presented in Figure 2.27.

PAGE 52

36 -40.00 -30.00 -20.00 -10.00 0.00 1.502.002.503.003.504.00 Freq (GHz)RL (dB) Measured Simulated Figure 2.26: Measure vs. Simulated RL of the Tuned Narrow Slot Antenna. -200 -100 0 100 200 1.52.02.53.03.54.0 Freq (GHz)Phase (degrees) Simulated Measured Figure 2.27: Measured vs. Simulated Phase of the Narrow Slot Antenna. 2.7.2 Radiation Pattern Measurements The coordinate system for the narrow slot antenna radiation pattern measurements is illustrated in Figure 2.28. The feed of the antenna is oriented in the x-direction at =0 degrees, and linearly polarized in th e x-direction. To measure the radiated fields in the electric plane E, the antenna is fixed at =0 degrees and rotated a bout the z-axis in the

PAGE 53

37 direction from 0 degrees to 360 degrees. On the other hand, the magnetic field H is obtained by fixing the antenna at =0 degrees and rotated about the x-axis in the direction from 0 degrees to 360 degrees. Figure 2.28: Radiation Pattern of a Ve rtical Slot and Coordinate System. The radiation pattern of the antenna was obtained with the aid of the anechoic chamber, the HP-8753D VNA, and DAMs computer tool. Figure 2.29 illustrates a visual of the setup utilized to measure the radiati on pattern of the slot antenna as well as a comparison that shows good agreement between the simulated and measured data. The measured narrow slot antennas coordinates utilized in Figure 2.29a are identical to those of Figure 2.28.

PAGE 54

38 (a) Co-Pol E-Field 270 0 90 180 -60 -50 -40 -30 -20 -10 0 Measured Simulated Figure 2.29: (a) Photograph of the Narrow Slot Antenna E-plane Measurement Setup. (b) Measured vs. Simulated Co-polarized Eplane for the Narrow Slot Antenna. On the other hand, the coordinate system for the magnetic plane measurements deviates from the coordinate system described in Figure 2.28. As illustrated in Figure = 0 = 0 90 x z y (b)

PAGE 55

39 2.30a, the antenna is fixed at =90 degrees and rotated about the z-axis in the direction from 0 degrees to 360 degrees. The measured and simulated results of the co-polarized magnetic planes are presented in Figure 2.30b. Furthermore, the cross-polarized electric and m agnetic planes are depicted in Figure 2.31 and Figure 2.32. (a) Co-Pol H-Field 0 180 90 270 -60 -50 -40 -30 -20 -10 0 Measured Simulated Figure 2.30: (a) Photograph of the Narro w Slot Antenna H-plane Measurement Setup. (b) Measured vs. Simulated Co-polarized H-plane for the Narrow Slot Antenna. = 0 = 0 90 x z y (b)

PAGE 56

40 Cross-Pol E-Field 0 90 180 270 -60 -50 -40 -30 -20 -10 0 Measured Simulated Figure 2.31: Measured vs. Simu lated Cross-polarized E-plane. Cross-Pol H-Field 180 -10 270 90 -60 -50 -40 -30 -20 -10 0 Measured Simulated Figure 2.32: Measured vs. Simulated Cross-polarized H-plane

PAGE 57

412.7.3 2.5GHz Wide Slot Antenna with U-shaped Tuning Stubs Results The optimized wide slot antenna was also fabricated on an FR-4 board, and measured using the HP-8753D VNA along with the sa me SOLT calibration standards used to measure the narrow slot antenna. A comparison between the simulated and measured RL is shown in Figure 2.33. The results illustrate that wide slot a ntennas are less sensitive to fabrication and dielectric constant variations within the FR-4 substrate. The discrepancy between the simulated and measured data is mainly due to the CPW-to-SMA connector transition and connector losses, which were not taken into account during the calibration process. The measured and simulated retu rn loss was found to be 28dB and 42dB, respectively, at an operational frequency of 2.5GHz. The measured gain of the antenna was found to be 5.4dBi, which compares well to the simulated gain of 5.1dBi. -50 -40 -30 -20 -10 0 1.52.02.53.03.54.0 Freq (GHz)RL (dB) Measured Simulated Figure 2.33: Measured vs. Simulated Retu rn Loss of the 2.5GHz Wide Slot Antenna with U-shaped Tuning Stubs.

PAGE 58

42 A comparison between the simulated and measured impedance of the wide slot antenna is shown in Figure 2.34. Figure 2.34b reveals that the antenna possesses multiple resonances and is well m atched at the design frequency of 2.5GHz. Antenna Series Resistance 0 25 50 75 100 1.5 2.0 2.5 3.0 3.5 Freq (GHz)Real (ohms) Measured Simulated Antenna Reactance -60 -30 0 30 60 1.5 2.0 2.5 3.0 3.5 Freq (GHz)Imaginary (ohms) Measured Simulated Figure 2.34: Measured vs. Simulated Wide Slot Antenna Impedance. (a) Real and (b) Imaginary. (a) (b)

PAGE 59

43 As illustrated in Figure 2.35 a bandwidth of approxim ately 69% was obtained referenced to a VSW R 2, which agrees with the simulated bandwidth of 70%. 0 1 2 3 4 1.52.02.53.03.54.0 Freq (GHz)VSWR Measured Simulated Figure 2.35: Measured vs. Simulated VSWR. The coordinate system presented in Figure 2.28 remains valid for the wide slot antenna rad iation pattern measurements presented in Figure 2.36. In addition, the wide slot an tenna was measured under the same c onditions previously mentioned in the narrow slot antenna measurements. The antenna setup for the co-polarized electric field is illustrated in Figure 2.36a, and the measurement results are depicted in Figure 2.36b. The com plementary magnetic plane of the wide slot antenna and a visual of the setup are also presented below. The result shows a high de gree of correlation between the simulated and measured radiation patterns.

PAGE 60

44 (a) Co-Pol E-Field 180 270 0 90 -60 -50 -40 -30 -20 -10 0 Measured Simulated Figure 2.36: (a) Photograph of the Wide Slot Antenna E-plane Measurement Setup. (b) Measured vs. Simulated Co-polarized Eplane for the Narrow Slot Antenna. = 0 = 0 90 x z y (b)

PAGE 61

45 (a) Co-Pol H-Field 270 180 90 0 -60 -50 -40 -30 -20 -10 0 Measured Simulated Figure 2.37: (a) Photograph of the Wide Slot Antenna H-plane Measurement Setup. (b) Measured vs. Simulated Co-polarized Hplane for the Narrow Slot Antenna. = 0 = 0 90 x z y (b)

PAGE 62

462.8 Chapter Summary and Conclusions In this chapter an in-depth analysis of sl ot antennas and their characteristics were presented. In the process, a parametric study of the series resistance of the narrow slot antenna as a function of the feed point dS was provided. In addi tion, the effects of a reflector plate on the impedance and radiatio n pattern of the wide slot antenna was studied. The two low frequency antenna prototyp es consisting of an offset-CPW-fed slot and a center-CPW-fed wide slot antenna with U-shaped tuning stubs were developed and presented a high degree of correlation with the simulated data. The wide slot antenna presented many adva ntages over the narrow slot antenna. For instance, the wide slot antenna is less sensitiv e to discrepancy in diel ectric constant of the substrate and the fabrication tolerances. Al so, the wide slot antenna provides a much wider bandwidth than the narrow slot antenna presented in Section 2.3. Furthermore, the added tuning capability eases the design a nd optimization of the antenna, and can be easily altered to obtain a wider range of oper ation. Finally, since th e ground plane of the antenna is symmetric, wire bonding is not required to reduce the higher order modes otherwise found in the offset -fed narrow slot antenna.

PAGE 63

47 Chapter 3 2.5GHz Detector Circuit Design In tegration of Rectenna Element 3.1 Introduction The second part of the Rectenna element co nsists of a detector circuit. These devices are used in many applications such as: power monitoring, automatic gain control circuits, automatic braking circuits, and signal strength indicators. In addition to these applications, diodes can be utilized as enve lope detectors, doublers, mixers, and squarelaw detectors depending on the power levels at which the diode is operated and the treatment of higher order harmonics. This thesis work mainly concentrat es on square-law detectors and their characteristics. This chapter presents a brief background theory of square-law detectors as well as the design and optimizat ion of detector circuits with the aid of Agilents Advanced Design Systems (ADS) CAD software. Also, a theo retical overview of CPWto-Microstrip transitions will be presented along with the development of a via-less CPW-to-Microstrip transition used in this th esis. The implementation of the transition is necessary in order to integrate the CPW-fed slot antenna with the microstrip detector circuit. Finally, a comparison is made be tween the two components of the Rectenna element realized in a single FR-4 board a nd individual Rectenna elements connected by a Male-to-Male (M-M) connector.

PAGE 64

483.1.1 Background Theory Detectors are essential components in wi reless communication and sensory systems. The two main types of diode detectors are e nvelope and square-law detectors. Envelope detectors are largely used in RF circuits to demodulate an amplitude-modulated signal by providing an output voltage that is directly proportional to the envelope of the incoming signal. Square-law detectors, on the other hand, are basically a power measuring device, which provides a DC output voltage that is di rectly proportional to the input RF power or the square of the input voltage as its name implies [ 19]. Rectification is accomplished by using nonlinear devices such as Schottky diodes. Equation 3.1 describes the small-signal voltage to current relationship of a commercially available Schottky diode [ 20]. (3.1) where nkTq / q is the charge of an electron, k is Boltzmanns constant, T is the temperature, n is the ideality factor, and Is is the saturation current. Figure 3.1 illustrates the I-V characteristics of the co mme rcially available HSMS-285X Schottky diode family series. Typical values for the diodes series resistance Rs are found in the SPICE parameter table of th e device data sheet [ 20]. )1()()( IRsV SeIVI

PAGE 65

49 0.01 0.10 1.00 10.00 100.00 0.00.20.40.60.81.01.21.41.61.8 Forward Voltage (V)Forward Current (mA) Figure 3.1: Typical Forwar d Current vs. Forward Voltage in Logarithmic Scale Representation. In a sense, all detectors present square-law characteristics at low power levels. So, if the applied voltage also contains a sm all RF signal represented by equation 3.2 (3.2) where V0 is the applied bias voltage and v0 is the magnitude of the RF signal. The diodes current can be expressed as a Taylors series expansion which provides the first three harmonics generated by the rectification process [ 21]. (3.3) The above Taylors series expansion only rem ains valid for sufficiently low power levels (Pin dBm). If the input power is too large, small-signal behavior will not apply and the output will become saturated. The dynami c range of the device is dictated by the )cos(0 00twvVV )(cos 2 )cos(0 2' 2 0 0 00twG v twGvIId d )2cos( 4 )cos( 40 2 0 0 0 2 0 0twG v twGvG v IId d d

PAGE 66

50 noise floor at the lower end (Pin < -55dBm) and by distortions at the upper end. The higher order harmonics generated by the diod e can be easily suppressed by implementing a simple low pass filter as illustrated in Figure 3.2. RF Input DC Output Figure 3.2: Conventional RF Detector Circuit Schematic. The input inductor serves as a DC return path for the diode, wh ile the output bypass capacitor acts as an RF short that separa tes the RF components from the output DC voltage, which keeps most of the power acros s the diode terminals. Finally, the output load resistance increases the ove rall output voltage of the de tector. All of the components described above play a signif icant role on the out put voltage and the sensitivity of the device. Sensitivity is a figure of merit that describes the quality of the detector, and it is representative of how much output voltage is obtained for a given input RF power. According to Agilent Technologies [ 22], the junction capacitance Cj, load resistance, and mismatch losses also dictate the sensi tivity of the detector. Equations 3.4 3.7 describe the sensitivity of the detector from an ideal case to a more realistic scenario. (3.4) (3.5) L Cb RL W V IS52.0 W mV RRCIjSj S ) 1( 52.022 1

PAGE 67

51 (3.6) (3.7) Equation 3.4 is only valid for an ideal diode with a zero bias current where the parasitic effects, reflections, and load resistor are neglected. A more reasonable approximation is presented by equation 3.5 where one can obser ve that some portion of the RF signal is bypassed by the junction capacitance Cj which decreases the overall sensitivity of the device. Finally, the mismatch loss, which plays a major role in the sensitivity of the device, is given by equation 3.7. In Section 3.2 of this chapter a comparison between the simulated detector circuits with and without a matching ne twork are presented, which demonstrates the significance of the mismatch loss. The detector circuit for the proposed Rectenna element is similar to th e detector circuit presented in Figure 3.2. The difference lies in th at for the Rectenna element, the detector circuit requires some sort of filtering mechanism to prevent higher order harmonics from re-radiating through the antenna. 3.2 Detector Circuit Design In this section the development of a det ector circuit that uses the HP8250 zero-bias Schottky diode is presented. This process is done with the aid of ADS As discussed in Chapter 1, the detector circuit for the Rect enna element presented by Nahas [3] consists of a matching network, a step impedance low pass filter, a fast rectifier, and an output DC filter. The input filter is usually designed to operate at a slightly higher cut-off frequency in order to achieve the lowest possible atte nuation at the design fr equency. Distributed elements are used in this design to reduc e the packaging parasitics found in lumped W mV R R RRCIL j jSj S 1) 1( 52.022 2 W mV )1(2 23

PAGE 68

52 components. This is necessary since packagi ng parasitics are much more significant at millimeter wave and infrared frequencies. For the low frequency Rectenna prototype, a distributed low pass filter was designed to work at a cut-off frequency of 3.2GHz and provide at least a 20dB of insertion loss (IL) at the second harmonic of 5GHz. In orde r to obtain the desired attenuation, a five element distributed filter was realized. The normalized element values for a maximally flat low pass filter obtained using the table presented by Pozar [ 21] are shown below. g1 = 0.618 = L1 g2 = 1.618 = C2 g3 = 2.000 = L3 g4 = 1.618 = C4 g5 = 0.618 = L5 The electrical length ( li) of each section was obtained using the following equations (3.8) (3.9) where L and C are the normalized element values of inductance and capacitance, respectively and R0 is the reference impedance. ZH and ZL represent the sections of very high and very low characteristic impedan ce of the line. Acco rding to Pozar [ 21], the ZH/ZL ratio should be as high as possible; therefore, the actual values of ZH and ZL are usually chosen to be the highest and lowest characteristic impedance that can be fabricated. These types of filters are widely known as step impedance filters due to the dramatic changes in characteristic impeda nce from one section to the other. After obtaining the elec trical length li shown in Table 3.1, the A DS LineCalc computer tool HZ LR0 0R CZL

PAGE 69

53 was utilized to determine the physical dimensions of the filter, which are also tabulated in Table 3.1. Table 3.1: S tep Impedance Filter Dimensions. Section Zi = ZL or ZH li Wi (mm) li (mm) 1 120 14.75 0.368 2.92 2 20 37.08 11.05 6.38 3 120 47.75 0.368 9.42 4 20 37.08 11.05 6.38 5 120 14.75 0.368 2.92 Figure 3.3 shows the layout of the step im pedance filter as drawn in Agilents Momentum, in which the thin and wide secti ons of transmission lin e are more inductive and capacitive, respectively. Figure 3.3: Step Impedance Low Pass Filter Momentum Layout. The simulated results depicted in Figure 3.4 show that the filter has a 3dB cut-off frequency of approxim ately 3.2GHz and presen ts an insertion loss (IL) of 0.5dB at the Rectenna element design frequency of 2.5GHz. L5 Z0 C4 L3 C2 L1 Z0

PAGE 70

54 -50 -40 -30 -20 -10 0 1.02.03.04.05.0 Freq (GHz)Filter Responce Return Loss Insertion Loss Figure 3.4: Step Impedance Low Pass Filter Response. One drawback of the filter presented above is that it requires many sections to achieve the desired attenuation, which in turn in creases the size of the Rectenna element dramatically. In addition, the step impedance low pass filter presents more losses than the shorted stub used in this design. To reduce the size and losses of the input low pass filter, a shorted stub was implemented in order to prevent the higher order harmonics from re-radiating through the antenna. Figure 3.5 shows the ADS sim ulated response for the shorted stub. From this figure, it is observed that the stub looks like an open circ uit at the design frequency of 2.5GHz while shorting the 5GHz harmonic to ground.

PAGE 71

55 -50 -40 -30 -20 -10 0 1.02.03.04.05.0 Freq (GHz)Shorted Stub Responce Return Loss Insertion Loss Figure 3.5: Shorted Stub IL and RL. In addition, utilizing the shorte d stub only presents a simulated IL of 0.07dB at the design frequency and attenuates the second order harmonic by 25dB. Furthermore, it provides a DC path to ground allowing the device to be come self-biased; keeping all the power across the diodes terminals. The harmonic balance simulation presented in Figure 3.6 f urther illustrates this concept. 2468 01 0 -40 -30 -20 -10 0 -50 10 freq, GHzdBm(RFin) 2468 010 -40 -30 -20 -10 0 -50 10 freq, GHzdBm(RFin) Figure 3.6: Harmonics Generated by the R ectification Process at the Input of the Detector. (a) Without Stub, and (b) with a Stub Present. Figure 3.6b shows that the s horted stub provides a DC path to ground and attenuates the higher o rder harmonics preventing them fr om reaching the source; in this case the (a) (b)

PAGE 72

56 antenna. The downside to th is technique is that shorted stub re sonates at odd multiples of the resonant frequency, which is illustrated by the 7GHz harmonic in Figure 3.6, which is not attenuated. The sam e sort of behavior is observed at the output of the detector where the output DC voltage has to be separated fr om the high frequency components. Figure 3.7 shows that all the higher harmonics are attenuated effectively prov iding an isolation of about 40dB at the design frequency. This was accomplished by implem enting a shunt opencircuited stub at the output of the detector, which provides the same susceptance as that of the bypass capacitor. 2468 010 -30 -25 -20 -15 -10 -35 -5 freq, GHzdBm(Vout) 2468 010 -60 -40 -20 -80 0 freq, GHzdBm(Vout) Figure 3.7: Harmonics Generated by the R ectification Process at the Output of the Detector with Stub in Place. (a) Without the Bypass Capacitor and the RF Block Inductor. (b) With the Bypass Capacitor and the RF Block Inductor in Place. As mentioned before, the matching network between the antenna and the detector circuit plays a big role in the output voltage an d sensitivity of the de tector. This behavior is illustrated in Figure 3.8, where S and IN represent the source and input reflection coefficient, respectively, and a1 and b1 represent the incident and reflected wave at the assigned reference plane, respectively. (a) (b)

PAGE 73

57 Figure 3.8: Detector Input Port Interface. The power available from the source can be characterized as the power delivered by the source under a conjugately matched load; in this case, the load is the input impedance of the detector circuit [ 23]. So, the power available from the source is given by equation 3.10. (3.10) The power delivered to the detector is define d as the difference between the incident and the reflected power at the i nput reference plane, and it is expressed by equation 3.11. (3.11) where, (3.12) and, (3.13) where bS is the reflected wave l ooking into the source. 2 21 2 1S S AVSb P 2 2 1 2 1 2 11 2 1 2 1 2 1IN INabaP INS Sb a 11 2 2 21 1 2 1INS IN S INbP

PAGE 74

58 According to [ 23], the power delivered to the detector ( PIN) and the power available from the source ( PAVS) can be related by th e mismatch factor ( MS) at a given reference plane. (3.14) This factor is utilized to quantify the amount of PAVS that is delivered to the input of a given device, and it is ex pressed by equation 3.15. (3.15) so, (3.16) It should be observed that when IN= S *, the mismatch factor is equal to unity, thus under this condition PIN = PAVS. The analysis presented above is valid for any given reference plane. This concept is further reinforced by anal yzing the output voltage and sensitivity of the device by performing ADS simulations for the detector circuit with and without a matching network. The simulated result presented in Figure 3.9 and Figure 3.10 show that the m atching network is a key component of the detector circuit, since it dictates the output voltage and sensitivity of the device. It is observed that a decrease in output voltage and thus sensitivity occurs when th e source is not conjugately matched to the detectors input impedance since some of th e power gets reflected back to the source. This behavior agrees with the literat ure and equations presented previously. S AVS INMPP 2 2 21 )1()1(INS IN S SM 2 2 21 )1()1(INS IN S AVS INPP

PAGE 75

59 0.0 0.1 1.0 10.0 100.0 1000.0 10000.0 -50-45-40-35-30-25-20-15-10-50510 Frequency (GHz)Output Voltage (mV) Matched Mismatched Figure 3.9: Simulated Output Voltage vs. Input Power. Red plot = no Matching Network; Blue Plot = Matc hing Network in Place. 100 1000 10000 100000 -50-45-40-35-30-25-20-15-10-50510 Frequency (GHz)Sensitivity (mV/mW) Matched Mismatched Figure 3.10: Simulated Detector Sensitivity Red = no Matching Network, Blue Plot = Matching Network in Place.

PAGE 76

60 An optimized detector layout along with the simulated return loss of the detector is illustrated in Figure 3.11. Figure 3.11b shows that the dete ctor is well m atched at the design frequency providing ma ximum power transfer, and therefore, increasing the detectors output voltage and sensitivity. Detector RL -50 -40 -30 -20 -10 0 2.02.32.52.83.0 Freq (GHz)RL (dB) Simulated Figure 3.11: (a) Optimized Detector Circui t Layout. (b) Simulated 2.5GHz Detector Return Loss. (a) (b)

PAGE 77

613.3 Rectenna Element Integration As previously mentioned, a Rectenna elem ent is composed of an antenna section integrated with a detector circuit. In or der to assemble the low frequency Rectenna elements prototype within a single FR-4 board, a CPW-to-Microstrip transition is required since the feeding mechanism of the antenna differs from that of the detector circuit. Even though the CPW and microstrip lines operate in a quasi-TEM mode, their electric and magnetic field dist ribution are not the same. If th e transition is not properly designed, some of the incident power at the tr ansition can be reflected, or radiated into free space, which can interfere with the antennas radiation charac teristics. In this section, a brief description of a via-less CPW-to-Micro strip transition is described along with its integration to the CPW-fe d wide slot antenna. 3.3.1 Via-less CPW-to-Microstrip Transition Historically the demand for high density and high performance microwave and millimeter wave circuits has increased dramatically making RF devices more compact and highly integrated. These devices are of ten fabricated using a combination of transmission lines such as the low frequency pr ototype presented in this chapter. For this design, a low loss, well matched, via-less CPW-to-Microstrip transition was realized to ease the field distribution from the CPW-fed slot antenna to the microstrip detector circuit. A cross-section representation of each segment of the transition is depicted in Figure 3.12. Here, it is obser ved that the transition consists of two segments: a CPW-togrounded coplanar waveguide (GCPW) and a GCPW-to-Microstrip segment.

PAGE 78

62 Figure 3.12: Via-less CPW-to-Microstrip Tr ansition Sections. (a) CPW, (b) GCPW, (c) Microstrip Cross-sections, (d) Transition Top View. The GCPW section depicted in Figure 3.12b is responsible for electrically shorting the CPW and M icrostrip grounds together. This is accomplished by making the width of the GCPWs ground be a wide, which becomes a short circuit at the design frequency of 2.5GHz. In addition, it should be noted that the GCPW section eases the field distribution of the CPW-to-Microstrip transition, thus reducing the power loss due to radiation and reflection present at the interface. The width of the midsection center conducto r and tapering angle of the transition presented in Figure 3.12d were optimized in Momentum while monitoring its insertion loss (IL), RL, and radiation ch aracteristics. The optimized transition presented an IL better than 0.4dB over a freque ncy range of 1GHz to 3.6GHz, which is illustrated in (a) (b) (c) (d)

PAGE 79

63 Figure 3.13. In addition, a RL better than 25dB was achieve d at the design frequency and an S11 15dB over a frequency ran ge of 1GHz to 4GHz, also illustrated in Figure 3.13. CPW-to-Microstrip Transition -1.4 -1.2 -0.9 -0.7 -0.4 -0.2 1.01.52.02.53.03.54.0 Freq (GHz)IL (dB) Simulated CPW-to-Microstrip Transition -35 -30 -25 -20 -15 -10 -5 0 1.01.52.02.53.03.54.0 Freq (GHz)S11 (dB) Simulated Figure 3.13: CPW-to-Microstrip Transition Simulated Data. (a) Insertion Loss and (b) Return Loss. (a) (b)

PAGE 80

64 The slot antenna with the transition incorpor ated was re-simulated to study its effects on the antennas RL and radiation character istics. This data was extracted from Momentum and compared to the wide slot ante nna characteristics previously presented in Chapter 2. Figure 3.14 shows the integrated layout in Mom entum along with the simulated RL comparison between the wide slot antenna with and without the transition. The simulated results presented in Figure 3.14b shows that the transition does not have a significan t effect on the RL of the antenna at the design frequency of 2.5GHz. The main difference between the two simulations is the f act that the bandwidth of the antenna with the transition decreased by 8%. Figure 3.14: 2.5GHz Wide Slot Antenna a nd CPW-to-Microstrip Transition Integration. (a) Momentum Layout and (b) RL Comparison Between the Simulated Antenna with and without the Transition. (a)

PAGE 81

65 2.5GHz Wide Slot Antenna with and W/O Transition -50 -40 -30 -20 -10 0 1.01.52.02.53.03.54.0 Freq (GHz)RL (dB) Sim. Ant Sim. Ant/transition Figure 3.14: (Continued) Figure 3.15 illustrates a comparison between the simulated radiation pattern for the antenna with and without the transition. These sim ulated results show that integrating the transition presents little or no effect on th e radiation characteristics of the antenna. (b)

PAGE 82

66 Simulated Radiation Pattern 270 0 90 180 -60 -50 -40 -30 -20 -10 0 With Transition W/O Transition Figure 3.15: Radiation Pattern Comparison Between the Simulated Wide Slot Antenna with and without the Transition in Place. 3.4 Results and Comparison 3.4.1 2.5GHz Detector Circuit The optimized detector circuit was fabri cated on an FR-4 board using the same techniques previously mentioned in Section 2.6.1. To measure the RL of the fabricated detector two types of calibrations were perf ormed in order to remove the systematic errors from the measured data. The RL of the detector was first measured using the same SOLT calibration technique mentioned in Section 2.6.1. This technique does not take into account the connector losses sinc e the calibrated reference plan e is at the edge of the connector as depicted in Figure 3.16. Previously, it was me ntioned that these connector losses were responsible for th e losses found in the m easured antennas RL. The second calibration technique used to measure the dete ctor circuit consiste d of a Thru-Reflect-

PAGE 83

67 Line (TRL) calibration. The TRLs design and results are described in Appendix C. This calibration technique takes into account the connector losses by putting the reference plane at the edge of the matching network. Figure 3.16: Photograph of the Fabricated 2.5GHz Detector Circuit with SOLT and TRL Calibration Reference Plane. The RL of the detector circuit wa s measured using an HP-8753D VNA. Measurements performed with the SOLT calib ration show that the SMA connector used limits the RL of the detector circuit to approximately 28dB, which is depicted in Figure 3.17. This result agrees well with the discrepancy found in Chapter 2 between the sim ulated and measured antenna RL. It shou ld also be noted that the measurement performed with the TRL calibration agrees very well with the simulated data since more losses are taken into account. As mentioned in Ch apter 2, the shift in frequency is mainly due to the inaccuracy of th e FR-4 dielectric constant. SOLT Reference Plane TRL Reference Plane

PAGE 84

68 Detector RL -50 -40 -30 -20 -10 0 2.0 2.3 2.5 2.8 3.0 Freq (GHz)RL (dBm) Simulated SOLT CAL TRL CAL Figure 3.17: 2.5GHz Detector Circuit Si mulated vs. Measured RL Using Different Calibration Techniques. Now that the detector circuit is well matched to the source impedance of 50 it was tested for its performanc e by measuring the square-law region and sensitivity. A diagram of the square-law measurement setup along w ith a brief description on how to make square-law measurements is presented in Appendi x B. The equipment utilized for this measurement consisted of an RF source, in this case the internal source of the HP-8753D VNA, the Anritsu ML-2438A power meter, a variable step attenuator, and the HP-3478A digital multimeter. The accuracy of the meas urement is limited by the ability of the variable step attenuator to accurately provide the same attenuation between measurements, and by the digital multimeters ability to detect small voltages. Table 3.2 presents the m easured input power and output DC voltage as a function of attenuation setting. Here, it is observed that the variable step attenuator presents up to 0.2dB error tolerance between input power measurements.

PAGE 85

69 Table 3.2: Measured Input Power and Output Voltage as a Function of Attenuation Setting. Attenuator Setting Input Power in (dBm) 0.2dBm Output Voltage (mV) Input Power in (mW) 1mW Sensitivity in (mV/mW) 0 8.06 2223 6.397 347.488 2 6.05 2014 4.027 500.103 4 3.97 1793 2.495 718.754 6 2.02 1587 1.592 996.729 8 -0.05 1351 0.989 1366.644 10 -2.18 1153 0.605 1904.712 12 -4.18 956 0.382 2502.983 14 -6.27 772 0.236 3270.524 16 -8.18 619 0.152 4070.902 18 -10.25 485 0.0944 5137.381 20 -11.90 394 0.0646 6102.337 22 -13.90 303 0.0407 7437.768 24 -15.98 229 0.0252 9074.767 26 -17.95 172.3 0.0160 10746.95 28 -20.02 126.70 9.954e-3 12728.48 30 -21.90 93.12 6.457e-3 14422.58 32 -23.93 66.74 4.046e-3 16496.29 34 -26.05 46.43 2.483e-3 18698.15 36 -28.07 32.00 1.56e-3 20518.71 38 -30.14 21.3 9.68e-4 21997.82 40 -32.11 14.00 6.15e-4 22757.68 42 -34.09 9.29 3.9e-4 23824.06 44 -36.18 5.95 2.41e-4 24689.77 46 -38.15 3.93 1.53e-4 25668.03 48 -40.23 2.56 9.48e-5 26992.3 50 -42.40 1.6 5.75e-5 27804.81 52 -44.39 1.1 3.64e-5 30226.84 54 -46.5 0.7 2.24e-4 31267.85 56 -48.49 0.45 1.42e-5 31784.29 58 -50.65 0.3 8.61e-6 34843.46 60 -52.45 0.2 5.69e-6 35158.47 From the data presented above, the square-law region of the detector can be identified by plotting the output DC voltage as a function of the input power. The square-law region is the section of the plot that follows a linear relationship between the input power and output DC voltage. The dynamic range of this detector was found to be approximately

PAGE 86

70 from -17dBm to -50dBm. This range is dictat ed by distortions at high power levels and by the noise floor at low power levels. This behavior along with the sensitivity of the device is illustrated in Figure 3.18. The sensitivity of the device was obtained by plotting the ratio of output voltage to inpu t power (W) vs. the input power (dBm). Detector Square Law 0 1 10 100 1000 10000 -50-40-30-20-10010 Input Power (dBm)Output Voltage (mV) Simulated Data Measured Data Detector Sensitivity 100.00 1000.00 10000.00 100000.00 -50.0-40.0-30.0-20.0-10.00.010.0 Pin (dBm)Sensitivity (mV/mW) Simulated Data Measured Data Figure 3.18: Simulated vs. Measured Detector Performance. (a) Square -Law Region and (b) Sensitivity. Square-law Region (a) (b)

PAGE 87

713.4.2 Via-less CPW-to-Microst rip Transition Results The CPW-to-Microstrip transition was also fabricated on a FR-4 board and measured using the HP-8753D VNA. A visual of the fabricated transition is depicted in Figure 3.19. Figure 3.19: Visual of the 2.5GHz Vi a-less CPW-to-Microstrip Transition. It is observed that the transition is not bi gger then a regular SM A M-M connector. From the measured results it is observed that the CPW-to-Microstrip transition presents an IL better than 0.4dB over a frequency span of 1G Hz 3.6GHz and an IL of 0.2dB at the design frequency of 2.5GHz. The discrepancie s between the simulated and measured data are mainly due to the connector losses and diel ectric constant variations within the FR-4 board.

PAGE 88

72 CPW-to-Microstrip Transition -1.4 -1.2 -0.9 -0.7 -0.4 -0.2 1.01.52.02.53.03.54.0 Freq (GHz)IL (dB) Simulated Measured CPW-to-Microstrip Transition -35 -30 -25 -20 -15 -10 -5 0 1.01.52.02.53.03.54.0 Freq GHzS11 (dB) Simulated Measured Figure 3.20: Simulated vs. Measured CPW-to -Microstrip Transition. (a) Insertion Loss and (b) Return Loss. (b) (a)

PAGE 89

733.4.3 Antenna/CPW-to-Microstrip Integration Results The antenna was integrated with the CPW-to-Microstrip transition on a single FR-4 board in order to check its performance. An image of the fabricated antenna as well as a comparison between simulated and measured RL is presented in Figure 3.21. The m easurement correlates well to the simulate d slot antenna. The an tenna and transition integration still sustain a 70% bandwidth measured at a RL 10dB. In addition, it provides a RL = 22dB at the design frequency of 2.5GHz. As mentioned before, the SMA connector adds losses to the measurements which will not be present in the full Rectenna element integration consisting of the slot an tenna, the CPW-to-Microstrip transition, and the detector circuit. Figure 3.21: (a) Visual of the Fabricated Slot Antenna/Transition. (b) Comparison Between the Simulated Antenna with and without Transition and Measured Antenna with Transition. (a)

PAGE 90

74 2.5GHz Wide Slot Antenna with and W/O Transition -50 -40 -30 -20 -10 0 1.01.52.02.53.03.54.0 Freq (GHz)RL (dB) Sim Ant Sim Ant/transition Meas Ant/Transition Figure 3.21: (Continued) Finally the radia tion pattern of the integr ated antenna was captured using the same techniques presented in Chapter 2 and compar ed to the simulated radiation pattern. The results depicted in Figure 3.22 shows good agreement to the simulated radiation pattern. From these results it is concluded that the tr ansition presents little or no affect on the antenna return loss and radiation ch aracteristics of the slot antenna. (b)

PAGE 91

75 Radiation Pattern After Integration 0 90 180 270 -60 -50 -40 -30 -20 -10 0 Meas Antenna/Transition Simulated Antenna/Transition Figure 3.22: Simulated vs. Measured Radi ation Pattern for the Integrated Antenna. 3.4.4 2.5GHz Rectenna Element Integration After characterizing each of th e Rectenna elements individually, they were fabricated on a single FR-4 board, and measured for its sens ing characteristics. A visual of the fully integrated Rectenna el ement is depicted in Figure 3.23. The Rectenna elements connected by a M-M connector were m easured and compared to the fully integrated Rectenna. The setup of the measurement is depicted in Appendix B. Figure 3.24 illustrates a comparison of the output voltage from both the Rect enna element and the individual devices connected by a M-M barrel. The output DC volta ge is plotted as a function of the input power delivered to the detector circuit by the receiving antenna. Both measurements were taken at a transmitting receiving distance of 4.5 feet.

PAGE 92

76 Figure 3.23: Photograph of the Fully Integrated Rectenna Element. 0.1 1 10 100 1000 10000 -56-46-36-26-16-6 Input Power (dBm)Output Voltage (mV) Antenna_MM_Detector Integrated Rectenna Detector Figure 3.24: Comparison Between the Measur ed Rectenna Element in a Single FR-4 Board vs. the Rectenna Element Connected by M-M Barrel. The Rectenna element fabricated within a si ngle FR-4 board provides a higher output voltage than the Rectenna elements connect ed by a M-M connector. This is due to the fact that the integrated Rect enna is better matched at the design frequency, thus reducing the Ms which translates into maximum power transfer.

PAGE 93

77 The input power presented in Figure 3.24 was obtained by m easuring the input power incident to the transmitting antenna and then using Friis transmission equation 3.17 to calculate the power received by the slot antenna. (3.17) If both the transmitting and receiving antennas are well matched and aligned for maximum directional radiation and reception, gi ven that there are no polarization losses, equation 3.17 reduces to its simple st form presented by equation 3.18. (3.18) where 24 R is the free-space loss factor, which take s into account the losses due to the spherical spreading of the energy by the antenna. Pt and Pr represent the transmitting and receiving power, respectively, and G0t and G0r represent the maximum gain of the transmitting and receiving antenna, respectively. For these calculations it was assumed that both antennas were well matched at the de sign frequency and were aligned to remove the polarization losse s. Under these conditions, the antennas Pr is the summation of Pt, G0t, G0r, and the free-space loss factor as long as the terms are expressed in dB. Appendix B presents the measurement set up along with the data obtained for the Rectenna element with a nd without a M-M connector. 2 2 2 2 ),(),( 4 11rtrrrttt r t cdrcdt t rD D R ee P P rt t rGG RP P00 24

PAGE 94

783.5 Chapter Summary and Conclusions In this chapter a brief desc ription of the background theo ry, design, and testing of a detector circuit was presented. Measured data correlated very well to the simulated data. The detector circuit offers a dynamic range of -17dBm to -50dBm, which agrees well with the literature. In addition, by comparing the measured RL using two types of calibration techniques, it was concluded that the discrepancies between the simulated and measured RL is mainly caused by the SMA connect or parasitics. Also in this chapter, the sketch of the design of a via-less CPW-to-Mic rostrip was presented showing a great deal of correlation to the simulated data. This tran sition presents an IL better than 0.4dB over a frequency range of 2.6GHz and a RL better than 15dB over a frequency span of 4GHz. Finally, the Rectenna elements were fabricated within a single FR-4 board and compared to the Rectenna elements connected by a M-M barrel. The measured result shows that the fully integrated Rectenna element provides a higher output DC voltage than the Rectenna element connected by a M-M barrel.

PAGE 95

79 Chapter 4 94GHz Antenna Design 4.1 Introduction 4.1.1 Millimeter Wave Band Characteristics In the past twenty years the interest in millimeter wave devices has increased dramatically due to the realizat ion that there are some limitations to what can be achieved with microwave, infrared, and optical systems. Infrared and optical systems provide extremely high resolution, but they suffer fr om many disadvantages such as dust, smoke, and fog conditions which degr ade the overall performance of the system. According to Bhartia and Bahl [ 24], the resolution of millimeter wave sensors is not significantly lower than that o btained with infrared and optic al sensors, and in conjunction with other techniques such as injection locking, phase locking, and power combining of multiple diodes, sensory system performance ca n be improved. The short wavelength characteristic of millimeter waves in combin ation with its interact ion with atmospheric constituents and wide operating bandwidths provide many advantages as well as some disadvantages. 4.1.2 Advantages and Disadvantages One of the main advantages at millimeter wave frequencies come from its short wavelengths characteristics which translate into a reduction of co mponent size. This significant reduction in the com ponent size allows for the production of compact systems,

PAGE 96

80 which are desirable for missiles, satellites, a nd aircraft applications. In addition, compact antenna arrays can be reali zed to provide narrow beamwidths, which in turn provide greater resolution and precision in target tracking and discrimination applications. Devices that operate at millimeter wave frequencies also present extremely wide bandwidth. The millimeter wave band ranges anywhere from 10mm to 1mm in wavelength. Figure 4.1 illustrates that the main windows in the m illimeter wave range exist at 8.6, 3.2, 2.1, and 1.4mm wavelengt h corresponding to frequencies of 35, 94, 140, and 220GHz, respectively. As seen in Figure 4.1, the bandwidths available at each of these windo ws are extremely large, and ar e approximately equal to 16, 23, 26, and 70GHz, respectively [ 25]. Figure 4.1: Millimeter Wave Atmospheri c Absorption for Horizontal Propagation [ 25]. These large bandwidths allow for high info rm ation rate capability, which enables obtaining fine target signature, very high rang e resolution for precision tracking, target identification in radars, and high sensi tivity radiometers. Finally, environmental interaction characteristics of millimeter wave devices such as atmospheric attenuation

PAGE 97

81 and losses due to aerosol, dust, smoke, and ba ttlefield contaminants are lower than in infrared and optical frequencies; making millimeter wave systems ideal for military applications. Even though millimeter wave frequenc ies provide multiple advantages over microwave, infrared, and optical frequenc ies, there are some obvious limitations. For instance, smaller component size increases the need for higher precision in manufacturing, thus increasing the overall cost of the device. Furthermore, millimeter wave component production quantities are lowe r which subsequently increases the cost of the devices needed for system design, in tegration, and characterization. Furthermore, the inherently narrow beamwidth of millimeter wave antennas is not suitable for large volume searches. Another disadvantage is that th e range capability of radars at millimeter wave is reduced in bad weather, and the backscattering in rain, which can imitate targets, is also a limitation of millimeter wave systems. The atmospheric attenuation due to precipitation rate is illustrated in Figure 4.2 [ 25]. Figure 4.2: Atmospheric Attenuatio n Due to Precipitation Rate [ 25].

PAGE 98

82 Wicker and Webb [ 26] provided some of the advantages of millimeter waves over m icrowave and optical frequencies for rada r applications, which are illustrated in Table 4.1. Here it is obse rved that millimeter wave frequencies present at least fair operation under any of the mentioned circumstances. Table 4.1: Radar System Performance in Millimeter Wave, Microwave, and Optical Frequencies. Radar Characteristics Microwave Optical Millimeter Wave Tracking accuracy Poor Good Fair Classification/ identification Poor Good Fair Covertness Poor Good Fair Volume search Good Poor Fair Adverse weather performance Good Poor Fair Performance in smoke, aerosol, dust, etc. Good Fair Good Based on the brief description of millimeter waves mentioned above and the frequency scalability of microwave components, it was decided to scale the slot antenna presented in Chapter 2 up in frequency, and study its characteristics at 94GHz. The 94GHz frequency window was chosen since it is one of the military frequency bands, and provides multiple commercial applications as well. 4.1.3 Applications Even though millimeter wave systems cannot replace microwave or optical systems entirely, they do provide a wide range of applic ations in the four major areas of wireless communications. Table 4.2 presented by Bhartia and Bahl [ 24] summarizes some of the m ain applications of millimeter waves. The successful design and characterization of the

PAGE 99

83 Rectenna element mentioned in the preceding chapters at millimeter wave frequencies could significantly lower the overall cost of millimeter wave systems. Table 4.2: Millimeter Wave Applicatio ns in the Four Major Areas of Wireless Communication [ 24]. Radar Low angle tracking Remote sensing of environment Secure military radar Surveillance Interference free radar Target acquisition High resolution radar Navigation Imaging radar Obstacle detection Ground Mapping Missile guidance Space object identification Fuses Harbor surveillance radar Ai rport surface detection radar Target characteristics Target designators Hand-held radar Range finders Radar cross-section measurements Active missile seekers Communications Secure military communications Satellite to satellite communications Point to point extremely wideband Earth to space communications communications LPJ communications Interference free communications Railroad communications Radiometry Remote sensing of the environment Ground target detection Radio astronomy Missile detection Ship detection Missile guidance Space-based radiometers Clear air turbulence sensor Instrumentation Plasma diagnostics Automatic braking Rocket exhaust plume measurements Spectroscopy Remote vibration sensor Prediction of blast focusing Model radar cross-section measurements

PAGE 100

844.2 94GHz CPW-fed Slot Antenna Design In the past three decades millimeter wave sources, circuits, and systems have received a great deal of atten tion. Antennas, on the other hand, ha ve not been extensively studied at millimeter wave frequencies. This is probably due to the fact that microwave antennas can be scaled up in frequency and still present sim ilar radiation characteristics. According to Bhatia and Bhal [ 24] the recent commercial and military applications in areas such as high-resolution radars, m issile guidance, short-range communication systems, and radiometric sensors have emphasized the need for the development of a wide variety of antennas that satisfy each systems requirements. Many antenna configurations such as leaky wave antennas, dielectric rod antennas, and microstrip antennas have been studied and developed. In order to reduce the cost, size, and weight of the millimeter wave system, a low profile planar antenna is needed. The slot antenna design presented in Chapter 2 fits this description and can be easily integrated with IC and CMOS technology. In addition, slot antennas can be conformed to different surfaces in aircraft, missiles, rocket s, space shuttles and so forth. In order to successfully implement the wide slot antenna presented in Chapter 2 at millimeter wave frequencies some key considerations have to be taken into account. For instance, at microwave frequencies substrates are usually mu ch thinner than the dielectric wavelength ( d). This is not the case at millimeter wave frequencies, where the commercially available substrate thicknesses are us ually comparable or thicker than d. The likelihood of surface waves propagating through the substr ate increases with the increase in the dielectric constant and th ickness of the substrate.

PAGE 101

854.2.1 Surface Wave in Silicon A high resistive silicon substrate was chos en for this design since it can be easily integrated with IC and CM OS technologies. In order to design a high performance antenna on a silicon substrate at millimeter wave frequencies some limitations need to be overcome. The high dielectric constant of the silicon substrate (r = ~11.7), and the discontinuities presen t within the radiati ng element imply that surface waves can be easily generated and propagated through the subs trate; potentially in creasing side lobe levels, coupling between radiating elements and reducing the antennas efficiency. At 94GHz the dielectric wavelength of silicon is approximately 100 m, which is thinner than some of the commerci ally available silicon wafers (400 m, 250 m, and 100 m). According to Luy et al. [ 27], the two possible surface wave modes for a grounded silicon substrate are designated as the transverse m agnetic (TM) and the transverse electric (TE) modes. The cut-off frequencies for a given Nth order TE or TM mode are provided by equations 4.1 and 4.2. (4.1) (4.2) where c0 is the speed of light in free space, h is the substrate thickness, and r is the dielectric constant. Peter [ 28] also stated that the cut-o ff frequency of the first order m icrostrip mode is given by equation 4.3. (4.3) 120 r TMh Nc fc 14 )12(0 r TEh Nc fc h Zc fcHE 0 00 12 ,

PAGE 102

86 where 0 = 377 is the free space characteristic impedance and Z0 is the characteristic impedance of the line. From equations 4.1 a nd 4.2 it can be observed that the thickness and the dielectric constant of the substrate ha s a major effect on the cut-off frequency of the higher order modes. It shoul d also be noted that the TM0 mode has a zero cut-off frequency and is always presen t in the substrate. As the substrate thickness increases, more surface waves can exist, which makes coupling of the lower order modes stronger. According to Gauthier et al. [ 29], the higher order modes can be m ade non-propagating by minimizing the substrate thickness. The recommended substrate thickness is on the order of d/10. For this reason, it was decided to design the 94GHz wide slot antenna with U-shaped tuning stubs on a 10 m high resistivity silicon membrane. 4.2.2 Substrate Thickness Characterization Using Momentum The 94GHz slot antenna design started by in vestigating the effects of the substrate thickness on the radiation characteristics of the antenna. A parametric study was performed in Momentum by varying the substrate thickness to 400 m, 250 m, 100 m, and 10 m while monitoring the antennas efficiency and 3D radiation pattern. The Momentum 3D plot presented in Figure 4.3 illustrates how the substrate thickness affects the rad iation characteristics of the antenna. Figure 4.3a through 4.3c show that a large portion of the radiated fields goes into the s ubstrate. In addition, dom inating side lobes are observed when the substrate thickness is comparable to the dielectric wavelength. Momentum, however, assumes an infinite substrate, which can be misleading when it comes to surface waves. In a finite substrate, the field can propagate through the substrate and radiate at the edges presenting endfire radiation. As illustrated in Figure 4.3, the

PAGE 103

87 10 m substrate thickness closely resembles the radiation pattern pres ented by the 2.5GHz antenna prototype. Figure 4.3: 3D Radiation Pattern for the 94GHz Slot Antenna as a Function of Substrate Thickness. (a) 400 m, (b) 250 m, (c) 100 m, and (d) 10 m. The parametric study presented above also il lustrates that the ante nnas efficiency is strongly dependent on the substrate thickness, which is presented in Table 4.3. The sim ulated result shows that th e antenna efficiencies at 400 m, 250 m, 100 m, and 10 m are 19, 19.6, 46.9, and 99%, respectively. As in the 2.5GHz prototype, a perfect conductor was assumed in the simulation process. (a) (b) (c) (d)

PAGE 104

88 Table 4.3: Slot Antenna Efficiency, Directivity, and Gain as a Function of Substrate Thickness. Substrate Thickness ( m) Directivity (dB) Gain (dB) Efficiency (%) 400 4.11 -3.115 19 250 5.74 -1.33 19.6 100 4.90 1.55 46.9 50 5.07 2.297 52.86 10 4.628 4.61 99 The Momentum layout of the 94GHz slot antenna on a 10 m silicon membrane is presented in Figure 4.4. Also in Figure 4.4, the simulated RL along with the phase of the 94GHz slot antenna is p resented. The simulated result illustrates that the antenna is well matched at the design frequency of 94GHz and presents an impedance bandwidth of approximately 55% referenced to a RL 10dB. The phase plot also shows that the antenna goes through multiple resonances, which agrees well with the low frequency prototype presented in Chapter 2. Since large bandwidths are not required for this thesis work, the antenna was optimized to present a perfect match at th e design frequency. If necessary, the bandwidth of the antenna can be easily increased by optimizing the tuning stub placement, length, and width.

PAGE 105

89 -60 -50 -40 -30 -20 -10 0 707580859095100105110115120 Freq (GHz)RL (dB) Simulated RL -200 -150 -100 -50 0 50 100 150 200 708090100110120 Freq (GHz)Phase (degrees) Phase Figure 4.4: 94GHz Slot Ante nna (a) Layout, (b) Simulated RL, and (c) Simulated Phase. (a) (b) (c)

PAGE 106

90 A comparison between the simulated radiat ion pattern of the 2.5GHz prototype and the 94GHz antenna is also presented in Figure 4.5. 270 180 90 0 -60 -50 -40 -30 -20 -10 0 94GHz Slot Antenna 2.5GHz Wide Slot Antenna Figure 4.5: 94GHz vs. 2.5GHz Si mulated 2D Radiation Pattern. In Figure 4.5, it can be observed that the 94GHz antenna has the sam e radiation pattern as the low frequency prototype presented in Chapter 2. In addition, the simulated directivity and gain of the 94GHz antenna were found to be 4.628dB and 4.61dB. This translates to an efficiency of 99% under the assumption th at gold is a perfect conductor. From these results, it can be concluded that the low frequency prototype slot antenna could be successfully scaled up in frequency while maintain ing the same radiation characteristics. 4.2.3 TRL Standards and Substrate Thickness Issues As mentioned before, in order to obtain th e best efficiency and radiation pattern, the antenna had to be fabricated on a 10 m silicon membrane. The problem with this substrate thickness is that it cannot be probed without da maging the silicon membrane. Therefore, a set of TRL standards were designe d in order to probe at a thicker section of

PAGE 107

91 the silicon substrate, in this case 250 m. Figure 4.6 illustrates the layout of som e of the elements including a cross-section of the 10 m silicon membrane. Figure 4.6: 94GHz Elements on a 10 m Silicon Membrane. (a) TRL Thru Standard, (b) Silicon Membrane Cross-Section, and (c) Slot Antenna. The CPW slots width decreases as the CPW line changes its thickness from 250 m to 10 m silicon membrane. This is the case sinc e the dimensions of the CPW line depend on the substrate thickness. In order to account for the 54.7 slope at the edges of the silicon membrane, the ground planes of the CPW line were tapered to minimize the reflections present at the discontinuity. Thes e reflections can give birth to higher order modes, which are undesirable at the measurements reference plane. Thus, the length of Ref Plane Ref Plane 250m 10m 54.7 Si 10m Membrane (a) (b) (c)

PAGE 108

92 the thru was designed to be as large as possi ble to ensure that the higher order modes die out by the time they reach the measurement s reference plane. According to Agilent Technologies [ 30], if enough separation between the tw o probes, and betw een the probes and the DUT, is not provided, coupling of the higher order modes could produce unwanted variations during th e error correction process, which would translate into inaccurate measurements. A separation equal to two wavelengths between the two probe tips is usually recommended [ 30]. 4.3 Results and Comparisons The optimized 94GHz slot antenna along with the TRL calibration standards were fabricated on a bulk micromachined silico n membrane using standard lithography techniques, and measured using the Wiltron 360B Network Analyzer. A visual of the measurement setup along with the modules required for the W band measurements are presented in Appendix C. The TRL calibration standa rds along with the calibration results are also provided in Appendix C. Twelve slot antennas and two sets of TR L calibration standards were realized on each 2 silicon wafer. Four of the antennas were de signed to operate at a substrate thickness of 10 m. The other antenna sets were designed to operate at substrate thickness of 8 m, 9 m, 11 m, and 12 m in order to account for any minor discrepancies that would arise from fabrication. A photograph of the fabricated devices along with a visual of the back etched silicon membranes are depicted in Figure 4.7.

PAGE 109

93 Figure 4.7: Visual of the Fabricated Devices. Right a nd Left Photograph Show the Fabricated Membranes.The Center Picture is the Top View of the Silicon Substrate. For this project, a total of two silicon wafers were fabricated. The first successfully fabricated substrate, which is illustrated in Figure 4.7, contains silicon membranes varying in thickness from 15 m t o 25 m. These thicknesses were measured using a profilometer. Figure 4.8 also illustrates a close up of the fabricated slot antenn a and corresponding membrane, whose membrane thic kness was measured to be approximately 20 m. Figure 4.8: Close up Visual of: (a ) the 94GHz Slot Antenna and (b) 20 m Silicon Membrane. Each of the antennas was measured using two types of calibrations standards. The best return loss obtained with the CS-5 calibration kit is illustrated in Figure 4.9. It can be seen that th e antenna presents a well matc hed condition at the design frequency of 94GHz. The discrepancies between the simulated and measured data are due in part to the (a) (b)

PAGE 110

94 thickness variation within the membrane, and the fact that there is an error tolerance of 2 m within the mask. -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RL (dB) Simulated Measured Figure 4.9: Measured vs. Simulated RL; Obtained with the CS-5 Calibration Kit. Figure 4.10 also illustrates the measured re su lts for the rest of the antennas, which were optimized to operate at substrate thickness of 8 m, 9 m, 10 m, 11 m, and 12 m respectively. The simulated results show that all of the measured antennas fall within the design frequency range. In addition, all of the antennas pres ent a RL of 10dB or better over the entire W band frequency window. Furthermore, Figure 4.11 illustrates the mutual coupling between each antenna set. The results s how that each antenna set presents at least a 35dB IL when fabricated on a 25 m high resistivity silicon membrane with a 2500 m separation between the antennas.

PAGE 111

95 -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RL (dB) Simulated 8 m Ant 9 m Ant 10 m Ant 11 m Ant 12 m Ant Figure 4.10: RL of the Different Slot Ante nnas Measured with the CS-5 Calibration Kit. -70 -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)IL (dB) 12 m Ant 11 m Ant 9 m Ant 8 m Ant Figure 4.11: IL of the Different Slot Antennas Meas ured with the CS-5 Calibration Kit. After examining the devices performance w ith the CS-5 calibration kit, an on-wafer calibration was performed in order to de-embed the systematic errors that are present within the feeding line. The re-measured slot antennas with the on-wafer calibration

PAGE 112

96 results are presented in Figure 4.12. As expected this results closely resem ble those obtained with the CS-5 calibration kit. -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RL (dB) Simulated 8 m Ant 9 m Ant 10 m Ant 11 m Ant 12 m Ant Figure 4.12: RL of the Different Slot Ante nnas Measured with the On-wafer Calibration Kit. A visual of the second s ubstrate is presented in Figure 4.13. This substrate was overetched resu lting in membranes thicknesses ranging from a 1 m oxide layer to approximately 8 m silicon membrane. Figure 4.13: (a) Top View of the Fabricated Substrate and (b) View of the Back Etched Membranes. (a) (b)

PAGE 113

97 It should also be noted that in Figure 4.13 the silicon membra nes are more tran sparent than the membranes presented in Figure 4.7. The transparency of the membrane is an indication of the membrane thickness. A cl ose up of the fabricated antenna and its corresponding membrane is also presented in Figure 4.14. Figure 4.14: Close up of the Fabricated Slot Antenna on Top of ~ 5 m Silicon Membrane. These antennas were also measured us ing both the CS-5 TRL probe tip and the onwafer calibration method. A comparison betw een the simulated and measured slot antennas, both using the CS-5 and the onwafer calibration ki t are presented in Figure 4.15 and Figure 4.16 respectively. These results show that th e worst case frequency shifts obtained with the CS-5 and on-wafer calibratio n kit were 8% and 5%, respectively. The phase of some of the measured slot antenna s obtained using the on-wafer calibration is also presented in Figure 4.17.

PAGE 114

98 -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RF (dB) Simulated 9 m Ant 10 m Ant 11 m Ant 12 m Ant Figure 4.15: RL of the Different Slot Antennas Meas ured with the CS-5 Calibration Kit. -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RL (dB) Simulated 9 m Ant 10 m Ant 11 m Ant 12 m Ant Figure 4.16: RL of the Different Slot Ante nnas Measured with the On-wafer Calibration. 5% 8%

PAGE 115

99 -200 -150 -100 -50 0 50 100 150 200 7580859095100105110 Freq (GHz)Phase (degrees) Simulated 9 m Ant 10 m Ant 11 m Ant 12 m Ant Figure 4.17: Phase of the Different Slot Antennas Measured with the On-wafer Calibration Standards. 4.4 Chapter Summary and Conclusions This chapter provides a brief description of the millimeter wave frequency band along with some of the advantages and disadvantages that millimeter wave frequencies have over microwave, infrared, and optical freque ncies. In addition, it was demonstrated through a set of Momentum simulations that the substrate thickness plays a major role on the antennas radiation characteristics. Fina lly, the optimized antennas were successfully fabricated on a bulk micromachined silico n membrane using standard lithography techniques, and measured using the Wiltron 360B Network Analyzer. The measured return loss using both calib ration techniques shows th at the majority of the fabricated antennas are well matched at the design frequency of 94GHz, and present at least a 10dB RL or better over the entire W band frequency window. The simulated radiation pattern of the 94GHz slot antenna closely resembles the radiation pattern of the 2.5GHz slot antenna described in Chapter 2. Finally, the simulated gain and directivity of

PAGE 116

100 the 94GHz antenna closely rese mbles the gain and directivit y obtained with the 2.5GHz antenna. Based on these results it was concluded that the 2.5GHz antenna can be scaled up in frequency without compromising its radiation characteristics.

PAGE 117

101 Chapter 5 Summary and Recommendations 5.1 Conclusions In this thesis work, the design of two CPW-fed slot antenn as, a Schottky diode detector circuit, and a CPW-to-Microstrip tran sition operating at 2.5GHz were presented. These low frequency devices were fabricated on an FR-4 board and tested individually to validate their performance. In addition, a low frequency Rectenna element was successfully integrated within a single FR-4 board and tested for its sensing characteristics. In the process, a set of simulations were performed to determine the dominating factors controlling the performance of each device. It was demonstrated that an off-center feeding method of the na rrow slot antenna can be utilized to decrease the high impedance of the resonant aperture and successfully match the antenna to the feeding line. The downs ide of this feeding technique is that it requires wire bonding in order to stabilize the ground planes of the CP W line. The failure to do so can lead to higher order modes, whic h are believed to be re sponsible for the high cross-polarization levels found in the narrow slot antenna. In order to overcome these shortcomings, a CPW-fed wide slot an tenna with U-shaped tuning stubs was implemented. Since this antenna is center-fed and the ground structure of the CPW line is symmetric about its center conductor, it doe s not require wire bonding. This antenna provides extremely low cross-po larization, and presents the same radiation characteristic

PAGE 118

102 of the narrow slot antenna. Moreover, a 60% increase in bandwidth over the narrow slot antenna was observed in bot h simulated and measured data. The wider bandwidth response of the CPW-fed wide sl ot antenna is mainly dictat ed by the length, width, and location of the tuning stubs within the aperture. Finally, it was shown that by placing a reflector plate at the reverse side of the substrate from the antenna, with an optimum reflector plate spaci ng of approximately unidirectional radiation is achieved. The reflector plate minimizes the back radiation and concentrates that energy towards the top of the antenna, which provides approximately a 3dB improvement in the gain of the antenna. The detector circuit of the Rectenna elemen t was also designed, fabricated, and tested in order to study its sensing properties. It was found that the junction capacitance of the diode, the load resistor, and the matching network have a c onsiderable effect on the output voltage and sensitivity of the detector. In addition, the filtering systems at the input and output of the detector have to be properly designed in order to prevent the higher order harmonics from re-radiating through the receiving antenna, and to isolate the input RF signal from the output DC voltage. In both measured and simulated data it was shown that the dynamic range of the detector circuit was found to be approximately from -17dBm to -50dBm, with a corresponding sens itivity of approximately 28,000mV/mW. In order to completely integrate the low fr equency prototype Rectenna element within a single FR-4 board, a via-less CPW-to-Mic rostrip transition was realized. It was demonstrated that the integration of the transition does not change the radiation characteristics of the antenna. This compact transition presents a well match condition at the design frequency and provides an IL better than 0.4dB over a 3.5GHz frequency span.

PAGE 119

103 Finally, both of the Rectenna components we re integrated within a single FR-4 board and tested for its sensing properties. It was demonstrated through a series of measurements that the fully integrated R ectenna element has the same square-law characteristics of the detector circuit. The fully integrated Rectenna is better matched at the design frequency then the separate Rect enna elements connected by M-M connector, which is responsible for the output vo ltage discrepancy between these two. Once the low frequency prototypes were va lidated through simulated and measured data, the 2.5GHz wide slot antenna was scaled up in fre quency to operate at 94GHz. A high resistivity silicon substrate was chos en for this design since it can be easily integrated with IC and CMOS technology. It was demonstrated through a set of Momentum simulations that the antenna pe rformance heavily depe nds on the substrate thickness. It was shown that wh en the substrate thickness is co mparable to the dielectric wavelength, higher order modes can be presen t that lead to high side lob levels and surface waves, both of which degrade the ove rall efficiency of the antenna. The 94GHz antenna fabricated on a 10 m silicon membrane was then optimized to provide simulated efficiencies of up to 99%. The optimized antennas were then fabricated on a bulk micromachined 10 m silicon membrane using standa rd lithography techniques. The measured results show that the fabricated an tenna presents at least a 10dB RL over the entire W band frequency window. In addition, the antenna presents a well matched condition at the design frequency of 94GHz. To the best of th e author's knowledge this is the first time that this specific antenna has been fabricated and tested for the W band frequency band.

PAGE 120

1045.2 Future Work and Recommendations Throughout the course of this thesis work many devices were fabr icated and tested, which lead to interesting conclusions and ideas for future work. For instance, further research needs to be done in order to miniaturize the low frequency Rectenna prototype. The miniaturization of the low frequency prototype is desirable since it could reduce the cost and size of many mobile applic ations, but the gain and directivity of the antenna are likely to decrease as well. Thus new miniaturization techniques are needed in order to reduce th e size of the low frequency Rectenna element without degrading the overall performance of the device. Moreover, a great deal of research stil l needs to be done in order to accurately complete a Rectenna element at 94GHz. First, the mutual coupling between antennas needs to be studied further. From the measur ement results presented in Chapter 4, it was inconclusive if the high isolation between the two antennas is due to the etched membranes, or the distance between the antenn as. In future research, a set of antennas placed at different distances from each other should be fabricated and measured to study the mutual coupling between the antennas. This is a key study needed in order to develop a high performance antenna arrays. Furthermore, the radiation pattern for the fabricated slot antenna has to be measured in order to validate the simulated radiation pattern. Finally, a high frequency model for the Meta l-Insulator-Metal (MIM) diode has to be developed in order to accurately design the dete ctor circuit at 94GHz. The ability to scale MIM diodes up to 30THz makes it a more desirable choice when compared to the HSCH 9161 GaAs detector diode. Additionally, the MIM diode can be fabricated within the Rectenna element using standard IC processi ng techniques. Lastly, an array of Rectenna

PAGE 121

105 elements has to be developed in order to improve the overall performance of the device. The successful completion of this research plan could lead to the next state of the art imaging, spectroscopy, or sensory systems.

PAGE 122

106 References [1] R. L. Bailey, Journal of Engineering Power (1972). [2] W. C. Brown, The History of Power Transmission by Radio Waves Microwave Theory and Techniques, IEEE Trans actions on, Vol.32, Iss.9, Sep 1984 pp: 1230-1242. [3] J. J. Nahas, Modeling and Comput er Simulation of a Microwave-to-DC Energy Conversion Element Microwave Theory and Techniques, IEEE Transactions on, Vol.23, Iss.12, Dec 1975, pp: 1030-1035. [4] I. H. On, J. S. Rice. and D. C. Thorn, "A Theoretical Study of Microwave Beam Absorption by a Rectenna," NASA Lyndon B. Johnson Space Center, Jan 1981, NAS916055. [5] W. C. Brown, "The Receiving Antenna and Microwave Power Rectification," Journal of Microwave Power 5, 1970, pp: 279 [6] I. J. Bahl, P. Bhartia, Microstrip Antenna, Artech House, Dedham, MA, 1980, Ch 6. [7] R. Garg, P. Bhartia, I. Bahl, A. Ittipiboon, Microstrip Antenna Design Handbook Artech House, Norwood, MA, 2001, Ch 7. [8] Y. Yoshimura, A Microstripline Slot Antenna, Microwave Theory and Techniques, IEEE Transactions on, Vol.20, Iss.11, Nov 1972, pp: 760-762. [9] Jeong Phill Kim; Wee Sang Park, Network Modeling of an Inclined and OffCenter Microstrip-Fed Slot Antenna , Antennas and Propagation, IEEE Transactions on, Vol.46, Iss.8, Aug 1998, pp: 1182-1188. [10] D. Pozar, A Reciprocity Method of Analysis for Printed Slot and Slot-Coupled Microstrip AntennasAntennas and Propagation, IEEE Transactions on [legacy, pre -1988], Vol.34, Iss.12, Dec 1986, pp: 1439-1446. [11] H. G. Booker, Slot Aerials and Thei r Relation to Complementary Wire Aerials (Babinets Principle), J. I. E.E., Vol.IIIA, 1946, pp: 620-226.

PAGE 123

107 [12] S. Sierra-Garcia, J. J. Laurin, Study of a CPW Induc tively Coupled Slot Antenna Antennas and Propagation, IEEE Tran sactions on, Vol.47, Iss.1, Jan 1999, pp: 58-64. [13] Constantine A. Balanis, Antenna Theory Analysis and Design, 3rd ed., John Wiley & Sons, Inc., Hoboken, New Jersey, 2005. [14] Warren L. Stutzman and Gary A. Thiele, Antenna Theory and Design, John Wiley & Sons, Inc., New York, 1981. [15] Jyh-Ying Chiou; JiaYi Sze; Kin-Lu Wong A Broad-Band CPW-fed StripLoaded Square Slot Antenna Antennas and Propagation, IEEE Transactions on, Vol.51, Iss.4, April 2003, pp: 719-721. [16] X. Ding, A.F. Jacob, CPW-fed Slot Antenna with Wide Radiating Apertures Microwaves, Antennas and Propagation, I EEE Proceedings -, Vol.145, Iss.1, Feb 1998, pp: 104-108. [17] Horng-Dean Chen Broadband CPW-fed Square Slot Antennas with a Widened Tuning Stub Antennas and Propagation, IEEE Transactions on, Vol.51, Iss.8, Aug 2003, pp: 1982-1986. [18] R. Chair, A. A. Kishk, K. F. Lee, C. E. Smith, and D. Kajfez, Microstrip Line and CPW-fed Ultra Wideband Slot Ante nnas with U-shaped Tuning Stub and Reflector, Progress In Electromagne tics Research, PIER 56, 2006, pp: 163-182. [19] Stephen A. Maas, The RF and Microwave Circuit Design Cookbook, Artech House, Inc., Norwood, MA, 1998, Ch 5. [20] Avago Technologies Data Sheet, Surface Mount Zero Bias Schottky Detector Diodes, HSMS-285X Series. Copyright 2006 Avago Technologies, Limited. [21] David M. Pozar, Microwave Engineering, 3rd ed., John Wiley & Sons, Inc., Hoboken, New Jersey, 2005, Ch 10. [22] Agilent Technologies Application Note 969, The Zero Bias Schottky Detector Diode, HSMS-285X Series. Copyright 1999 Agilent Technologies, Inc. [23] Guillermo Gonzalez, Microwave Transistor Amplifiers Analysis and Design 2nd ed., Prentice-Hall, Inc., Upper Saddl e River, New Jersey, 1997, Ch 2. [24] P. Bhartia, I. J. Bahl, Millimeter Wave Engineering and Applications, John Wiley & Sons, Inc., New York, NY, 1984.

PAGE 124

108 [25] http://www.phys.hawaii.edu/~anita/web /paperwork/currently%20organizing/Milit ary%20EW%20. [26] L. R. Wicker, and D. C. Webb, The Potential Military Application of Millimeter Waves, AGARD Conf. Proc ., CP 225, on Millimeter and Submillimeter Wave Propagation and Circuits, 1978, pp: 11-16. [27] J. F. Luy, K.M. Strohm, J. Buechler, P. Russer, "Silicon Monolithic Millimetre-Wave Integrated Circuits," Microwaves, Antennas and Propagation, IEEE Proceedings H Vol.139, no.3, Jun 1992 ,pp: 209-216. [28] P. Russer, "Si and SiGe Millimeterwa ve Integrated Circuits for Sensor and Communications Applications," Micr owaves and Radar, 1998. MIKON '98., 12th International Conference on Vol.4, pp: 330-344 Vol.4, May 1998, pp: 20-22. [29] G. P. Gauthier, J. P. Raskin, L. P. B. Katehi, G. M. Rebeiz, "A 94-GHz ApertureCoupled Micromachined Microstrip An tenna," Antennas and Propagation, IEEE Transactions on Vol.47, no.12, Dec 1999, pp: 1761-1766. [30] http://cp.literature.agilent .com/litweb/pdf/5091-3645E.pdf.

PAGE 125

109Appendices

PAGE 126

110Appendix A: 2.5GHz Rectenna Elements Measurements Setup The reflection coefficient (S11) of the Rectenna elements was obtained using the measurement setup presented in Figure A.1. The reflection co efficient m easurement is a one-port measurement in which only P1 of the RF Vector Netw ork Analyzer (VNA) has to be calibrated. The accuracy of the measured reflection coefficient is determined by the precision of the standards utilized in the e rror correction processes, which is widely known as the calibration standards. In this measurement, a set of pre-determined calibration standards consisting of an open, s hort, and load (SOL) were utilized during the calibration process. See Appendi x B for calibrated results. Figure A.1: Block Diagram for On e-Port S-parameter Measurements. In order to obtain both re flection and transmission coeffi cients of a given device under test (DUT) a full two-port calibration is required in order to de-embed the systematic errors from the measuremen ts. A block diagram of the full two-port calibration setup is presented in Figure A.2. This calibration technique is very sim ilar to the SOL calibration presented above. The onl y difference is that it requires one more known standard in order to complete the error model. This calibration

PAGE 127

111Appendix A (Continued) technique, known as the SOLT calibration, consis ts of measuring a short, open, load, and thru standards. Figure A.2: Block Diagram of a Full Two-Port S-parameter Measurements. The radiation pattern of the 2.5GHz slot antenna was measured using the measurement setup presented in Figure A.3. Here, the source is used to set the transm itting RF power at a pre-determined po wer level. Then, the DAMs computer tool controls the turn table, which rotates the receiving antenna in the azimuth and elevation direction. At each angle the pattern record er saves the transmission coefficient (S21) value until the azymuthal angle reaches 360.

PAGE 128

112Appendix A (Continued) Figure A.3: Typical Block Diagram for Antenna Range Measurements. The 2.5GHz detector circuit square-law and sensitivity measurements were performed using the measurement setup presented in Figure A.4. Here, the VNA acts as the RF source, and the dig ital multimeter is used to measure the rectified output voltage. The power meter is calibrated prior to making the m easurement, and it is utilized to measure the actual power available from the source. The variable step attenuator is than varied and the corresponding DC voltage at the output of the detector circuit is recorded. Figure A.4: Square-law Detector Block Diagram Measurement Setup.

PAGE 129

113Appendix A (Continued) The block diagram presented in Figure A.5 was utilized to measure the sensing properties o f the fully integrated 2.5GHz Rectenna element. This measurement setup closely resembles that of the square-law de tector, with some added features. The power amplifier is needed to increase the power incident to the transmitting antenna. The power meter is responsible for accurate ly reading the power available at the output of the step attenuator, and the digital multimeter read s the output DC voltage delivered by the Rectenna element. The transmitting antenna and the Rectenna element must be aligned to reduce any polarization losses. Finally, the in put power to the transmitting antenna and the output DC voltage are recorded as a functi on of attenuation setting. The recorded data for the fully integrated Rectenna element and the individual Rectenna elements connected by a M-M connector are presented at the end of this Appendix.

PAGE 130

114Appendix A (Continued) VNA Variable Step Attenuator Power Meter Power Sensor Digital Multimeter PIN Zero Length Cables 4.5 Ft Transmitting Antenna Rectenna Element Power Amplifier Figure A.5: Block Diagram fo r the Integrated Rectenna Element Power Measurements.

PAGE 131

115Appendix A (Continued) Table A.1: Recorded Data for the Rect enna Elements Connected by a M-M Barrel. Attenuator Setting Transmitting Antenna PIN (dBm) Receiving Antenna PR (dBm) Rectenna Output VDC (mV) 0 22.93 -2.94 1175.5 2 21.1 -4.63 960.3 4 19.19 -6.54 776.96 6 17.12 -8.61 610.00 8 15.09 -10.64 476.85 10 12.78 -12.95 352.20 12 10.76 -14.97 270.50 14 8.84 -16.89 205.60 16 6.87 -18.86 150.70 18 4.91 -20.82 110.80 20 2.87 -22.86 80.59 22 0.89 -2484 57.25 24 -1.05 -26.78 39.96 26 -3.06 -28.78 27.18 28 -5.03 -30.86 18.19 30 -7.13 -32.86 11.68 32 -9.13 -34.86 7.64 34 -11.09 -36.82 5.0 36 -13.08 -38.81 3.25 38 -15.06 -40.79 2.14 40 -17.07 -42.80 1.42 42 -19.09 -44.82 0.95 44 -21.08 -46.81 0.65 46 -23.12 -48.85 0.45 48 -25.16 -50.89 0.3 50 -27.34 -53.07 0.21

PAGE 132

116Appendix A (Continued) Table A.2: Recorded Data for the Fully Integrated Rectenna Element. Attenuator Setting Transmitting Antenna PIN (dBm) Receiving Antenna PR (dBm) Rectenna Output VDC (mV) 0 22.95 -2.92 1250.7 2 21.13 -4.60 1023.2 4 19.24 -6.49 826.98 6 17.17 -8.56 652.67 8 15.12 -10.61 513.89 10 12.88 -12.85 388.89 12 10.97 -14.76 298.50 14 9.00 -16.73 229.17 16 6.98 -18.75 173.04 18 4.99 -20.74 129.60 20 2.88 -22.85 93.99 22 0.89 -24.84 67.96 24 -1.07 -26.80 48.47 26 -3.08 -28.81 33.53 28 -5.07 -30.80 22.83 30 -7.17 -32.90 14.86 32 -9.19 -34.92 9.81 34 -11.16 -36.89 6.47 36 -13.15 -38.88 4.22 38 -15.14 -40.87 2.78 40 -17.15 -42.88 1.85 42 -19.17 -44.90 1.26 44 -21.15 -46.88 0.84 46 -23.18 -48.91 0.58 48 -25.20 -50.93 0.41 50 -27.36 -53.09 0.28

PAGE 133

117Appendix B: Low Frequency Calib ration Standards and Results In Chapter 3, the CPW-to-Microstrip tran sition results show a significant amount of ripple, which in some cases degrades the IL (S21) of the device. The origin of these errors mainly comes from the calibration process in which the cable movement, SMA connector, and transition from one medium to another (coaxial to CPW) have a significant effect. The cable movement changes the phase of the calibrated data and adds uncertainties to the calibration correction proce ss. In addition, the transition from coaxial line to CPW and microstrip line represents discontinuities, which further contribute to the ripple effect present in the measured IL of the CPW-to-Microstrip transition. The calibration results for the 2.5GHz antennas, detector circuit, and CPW-to-Microstrip transition S-parameter measurements are pres ented below. It should be mentioned that these results were obtained usi ng the HP-8753D network analyzer. -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 1.01.52.02.53.03.54.0 Freq (GHz)dB S(2,1) Thru S(1,2) Thru Figure B.1: IL of the SOLT Thru Standard.

PAGE 134

118Appendix B (Continued) -1.00 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75 1.00 1.01.52.02.53.03.54.0 Freq (GHz)degrees S(2,1) Thru S(1,2) Thru Figure B.2: Phase Response of the SOLT Thru Standard. -100 -80 -60 -40 -20 0 1.01.52.02.53.03.54.0 Freq (GHz)dB S(1,1) Load S(2,2) Load Figure B.3: RL of the SOLT Load Standard.

PAGE 135

119Appendix B (Continued) -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 1.01.52.02.53.03.54.0 Freq (GHz)dB S(1,1) Open S(2,2) Open Figure B.4: RL of the SOLT Open Standard. -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 1.01.52.02.53.03.54.0 Freq (GHz)dB S(1,1) Short S(2,2) Short Figure B.5: RL of the SOLT Short Standard.

PAGE 136

120Appendix B (Continued) As mentioned before a Thru-Reflect-Line (T RL) calibration kit was designed in order to measure the actual input impedance of the de tector circuit. The a dvantage of this type of calibration is that it takes into account the SMA connector losses, and gives an extra degree of freedom, which allows one to cont rol the measurements reference plane. As depicted in Figure B.6, the length of th e th ru standard is arbitrar ily chosen, and it sets the reference plane of the measurement. The length of the open is of the thru standard, and the length of the delay is chosen to be at the design frequency of 2.5GHz. The useable bandwidth of a single delay line is limited to an 8:1 (frequency span /start frequency) ratio. If a larger frequency range is needed, multiple line pairs with the same characteristic impedance can be utilized. It should be mentioned that the frequency span of each delay line should overlap with one another. Detector Circuit Ref-Plane Thru Open Open Delay Ref-PlaneRef-Plane Ref-PlaneRef-Plane L1 Figure B.6: (a) Drawing of a Given TRL Standard Kit and Refe rence Plane. (b) Visual of the Fabricated TRL Standards. (a) (b)

PAGE 137

121Appendix B (Continued) The key terms and formulas needed to cal culate the TRL standard definitions are presented bellow. It should be mentioned that the thru and reflect standards delays were defined to be zero picoseconds. L1 = LTHRU + L2.5GHz_ DELAY 2THRU OPENL L eff = 3.29 LTHRU = 4 cm LOPEN = 2 cm L2.5GHz_ DELAY = 1.6531 cm s cm x s m x c veff p 10 810653.110653.1 T2.5GHz_DELAY = ps v Lp DELAY GHz8.995.2 ps v L Tp DELAY GHz DELAY GHz42.49_5 _5 Finally, after calculating the standard definitio ns, a TRL calibration kit was fabricated by on a FR-4 board and implemented for erro r correction using the HP-8753D VNA. The calibration results are presented below.

PAGE 138

122Appendix B (Continued) -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 1.01.52.02.53.03.54.0 Freq (GHz)dB S(2,1) Thru S(1,2) Thru Figure B.7: IL of the TRL Thru Standard. -100 -80 -60 -40 -20 0 1.01.52.02.53.03.54.0 Freq (GHz)dB S(1,1) Thru S(2,2) Thru Figure B.8: RL of the TRL Thru Standard.

PAGE 139

123Appendix B (Continued) -1.0 -0.8 -0.5 -0.3 0.0 0.3 0.5 0.8 1.0 1.01.52.02.53.03.54.0 Freq (GHz)degrees S(2,1) Phase S(1,2) Phase Figure B.9: Phase of the TRL Thru Standard. -180 -150 -120 -90 -60 -30 0 1.01.52.02.53.03.54.0 Freq (GHz)degrees S(2,1) Phase S(1,2) Phase Figure B.10: Phase of the TRL Delay Standard.

PAGE 140

124Appendix B (Continued) -100 -80 -60 -40 -20 0 1.01.52.02.53.03.54.0 Freq (GHz)dB S(1,1) Delay S(2,2) Delay Figure B.11: RL of the TRL Delay Standard. -0.50 -0.30 -0.10 0.10 0.30 0.50 1.01.52.02.53.03.54.0 Freq (GHz)dB S(1,1) Open S(2,2) Open Figure B.12: RL of the TRL Open Standard.

PAGE 141

125Appendix C: W band Calibration Standard Results The W band measurements were perfor med using the Wiltron 360B Network Analyzer along with the Wiltron SM5184 tran smission and reflection modules. As the name implies the transmission and reflection modules expands the frequency range of the network analyzer to the 65GHz 110GHz W band frequency window, and performs transmission and reflection measurements. A visual of the measurement setup is illustrated in Figure C.1. Figure C.1: Visual of the W band Measurement Setup. For the W band antenna measurements two type of calibrations were performed in order to remove the systematic errors fr om measured data. First, the system was calibrated using the CS-5 calibration kit. With this calibration kit a TRL calibration was

PAGE 142

126Appendix C (Continued) performed. Then the reference plane was shifte d in order to obtain a probe tip calibration. The corrected calibrated results are depicted below. -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 7580859095100105110 Freq (GHz)IL (dB) CS-5 S(2,1) Thru CS-5 S(1,2) Thru Figure C.2: IL of the CS-5 Thru Standard (A fter Reference Plane Shift to Probe Tips). -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RL (dB) CS-5 S(1,1) Thru CS-5 S(2,2) Thru Figure C.3: RL of the CS-5 Thru Standard (A fter Reference Plane Shift to Probe Tips).

PAGE 143

127Appendix C (Continued) -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 7580859095100105110 Freq (GHz)Phase (degrees) CS-5 S(2,1) Phase CS-5 S(1,2) Phase Figure C.4: Phase of the TRL Thru Standard (A fter Reference Plane Shift to Probe Tips). -70 -50 -30 -10 10 7580859095100105110 Freq (GHz)RL (dB) CS-5 S(1,1) 550 m Delay CS-5 S(2,2) 550 m Delay Figure C.5: RL of the CS-5 550 m Delay Standard (After Reference Plane Shift to Probe Tips).

PAGE 144

128Appendix C (Continued) -180 -160 -140 -120 -100 -80 -60 -40 -20 0 7580859095100105110 Freq (GHz)Phase (degrees) CS-5 (S(2,1) 550 m Delay CS-5 (S(1,2) 550 m Delay Figure C.6: Phase of the CS-5 550 m Delay Standard (After Reference Plane Shift to Probe Tips). -0.1 0.0 0.1 0.2 0.3 0.4 0.5 7580859095100105110 Freq (GHz)RL (dB) CS-5 S(1,1) Open CS-5 S(2,2) Open Figure C.7: RL of the CS-5 Open Standard (After Reference Plane Shift to Probe Tips).

PAGE 145

129Appendix C (Continued) -40 -34 -28 -22 -16 -10 -4 2 7580859095100105110 Freq (GHz)Phase (degrees) CS-5 S(1,1) Open CS-5 S(2,2) Open Figure C.8: Phase of the CS-5 Open Standa rd (After Reference Plane Shift to Probe Tips). After obtaining the antenna measurements with the probe tip calibration, an onwafer calibration was performed in order to de-embed the systematic errors that are present within the feeding line. A drawing of the calibration standards utilized for this measurements along with the calibrati on results are presented Figure C.8.

PAGE 146

130Appendix C (Continued) Figure C.9: Drawing of the On -wafer Calibration Standards. -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 7580859095100105110 Freq (GHz)IL (dB) On-Wafer S(2,1) Thru On-Wafer S(1,2) Thru Figure C.10: IL of the On-wafer Thru Standard.

PAGE 147

131Appendix C (Continued) -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 7580859095100105110 Freq (GHz)Phase (degrees) On-Wafer S(2,1) Thru On-Wafer S(1,2) Thru Figure C.11: Phase of the On-wafer Thru Standard. -60 -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)RL (dB) On-Wafer S(1,1) Thru On-Wafer S(2,2) Thru Figure C.12: RL of the On-wafer Thru Standard.

PAGE 148

132Appendix C (Continued) -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 7580859095100105110 Freq (GHz)RL (dB) On-Wafer S(1,1) Delay On-Wafer S(2,2) Delay Figure C.13: RL of the On-wafer Delay Standard. -1.0 -0.8 -0.6 -0.4 -0.2 0.0 7580859095100105110 Freq (GHz)IL (dB) On-Wafer S(2,1) Delay On-Wafer S(1,2) Delay Figure C.14: IL of the On-wafer Delay Standard.

PAGE 149

133Appendix C (Continued) -160 -140 -120 -100 -80 -60 -40 -20 0 7580859095100105110 Freq (GHz)Phase (degrees) On-Wafer S(2,1) Delay On-Wafer S(1,2) Delay Figure C.15: Phase of the On-wafer Delay Standard. -1.0 -0.8 -0.5 -0.3 0.0 0.3 0.5 7580859095100105110 Freq (GHz)RL (dB) On-Wafer S(1,1) Open On-Wafer S(2,2) Open Figure C.16: RL of the On-wafer Open Standard.

PAGE 150

134 Appendix C (Continued) -50 -40 -30 -20 -10 0 7580859095100105110 Freq (GHz)Phase (degrees) On-Wafer S(1,1) Phase On-Wafer S(1,2) Phase Figure C.17: Phase of the On-wafer Open Standard.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001966245
003 fts
005 20081020122030.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 081020s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002221
035
(OCoLC)262835149
040
FHM
c FHM
049
FHMM
066
(S
090
TK145 (ONLINE)
1 100
La Rosa, Henrry.
0 245
Investigation of a Rectenna element for infrared and millimeter wave application
h [electronic resource] /
by Henrry La Rosa.
260
[Tampa, Fla.] :
b University of South Florida,
2007.
502
Thesis (M.S.E.E.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 134 pages.
590
Co-adviser: Elias Stefanakos, Ph.D.
Co-adviser: Kenneth A. Buckle, Ph.D.
3 520
ABSTRACT: This thesis presents the rectifying antenna potential for infrared and millimeter wave energy conversion. Infrared imaging is one of the emerging technologies that have attracted considerable attention in the next generation of military, medical, and commercial applications. Moreover, with the ever-increasing congestion of the electromagnetic spectrum at RF and microwave frequencies and the establishment of firm civilian and military requirements best met by millimeter wave systems, the interest in the technology has grown and is now firmly established. During this work a 2.5GHz slot antenna, a 2.5GHz Schottky diode detector, a CPW-to-Microstrip transition, a fully integrated Rectenna element, and a 94GHz slot antenna were designed, fabricated, and tested. Results on the performance of the devices show a great deal of correlation between the simulated and measured data. To perform an initial study, the CPW-fed narrow slot antenna is designed at 2.5GHz and implemented on an FR-4 board. This investigation serves as the basis for the development of the Rectenna element at millimeter wave frequencies. In order to increase the bandwidth of the slot antenna, a 2.5GHz CPW-fed wide slot antenna with U-shaped tuning stub is realized, which provides a 60% increase in bandwidth while keeping the same radiation characteristics. In addition, a set of simulations is performed to show how a reflector plate affects the radiating properties of the slot antenna. A 2.5GHz square-law detector is also designed, fabricated, and tested in order to rectify the RF signal delivered by the antenna. The fabricated detector presents a well matched condition at the design frequency with a dynamic range found to be from --17dBm to --50dBm. The low frequency Rectenna element prototype is then integrated within a single FR-4 board.^ This is accomplished by implementing a compact via-less CPW-to-Microstrip transition. Finally, a 94GHz CPW-fed wide slot antenna is realized on a 10m high resistivity silicon membrane. This antenna shows a great deal of similarity to the 2.5GHz slot antenna. This low profile antenna presents at least a 10dB return loss over the entire W band frequency window. Simulated antenna efficiencies of up to 99% were achieved assuming a perfect conductor.
653
Microstrip slot antenna.
Schottky diode.
Detector.
Coplanar waveguide (CPW).
Wideband.
W band.
Membrane.
690
Dissertations, Academic
z USF
x Electrical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2221