rf and microwave microelectronics packaging · 2013-07-18 · rf and microwave microelectronics...

RF and Microwave Microelectronics Packaging

Ken Kuang · Franklin Kim · Sean S. CahillEditors

RF and MicrowaveMicroelectronics Packaging

123

EditorsKen KuangTorrey Hills Technologies, LLC6370 Lusk Blvd, F-111San Diego CA [email protected]

Sean S. CahillBridgeWave Communications Inc.3350 Thomas RoadSanta Clara CA [email protected]

Franklin KimKyocera America, Inc.8611 Balboa AvenueSan Diego CA [email protected]

ISBN 978-1-4419-0983-1 e-ISBN 978-1-4419-0984-8DOI 10.1007/978-1-4419-0984-8Springer New York Dordrecht Heidelberg London

Library of Congress Control Number: 2009939146

© Springer Science+Business Media, LLC 2010All rights reserved. This work may not be translated or copied in whole or in part without the writtenpermission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York,NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use inconnection with any form of information storage and retrieval, electronic adaptation, computer software,or by similar or dissimilar methodology now known or hereafter developed is forbidden.The use in this publication of trade names, trademarks, service marks, and similar terms, even if they arenot identified as such, is not to be taken as an expression of opinion as to whether or not they are subjectto proprietary rights.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Preface

This book is an outgrowth of the first IMAPS (International Microelectronics andPackaging Society) Advanced Technology Workshop on RF/Microwave Packaging,held September 16–18, 2008 in San Diego, California. Wireless technologies haveundergone tremendous growth in the last decade and the interest in packagingfor high-frequency applications has grown as well. Over 30 invited speakers gavepresentations on select advanced topics in RF, microwave, millimeter-wave andbroadband packaging.

The motivation behind this conference, however, goes beyond the obvious areasof utility. When referring to fundamental engineering limits to very high speed elec-tronics, packaging and interconnect constraints figure significantly. Ever increasingdata rates are transforming digital technologies into what are essentially RF systems.The once arcane tools of the RF discipline are becoming increasingly applicable toelectronic systems in general, motivating many new considerations. RF systems,despite their small device count, have traditionally been voracious, inefficient con-sumers of power, creating significant challenges for packaging engineers to dealwith heat dissipation. Most digital devices have been much more frugal, but speedand high levels of integration turn these devices into significant heat sources as well.In light of these evolutionary trends, a sampling of the workshop participants wereasked to submit chapters on these fascinating areas of development for the work athand.

Given the diversity of voices at this workshop, and the highly interdisciplinarynature of the topics discussed, this work ranges broadly. Authors include academics,students, large industrial concerns and small entrepreneurial ventures covering avariety of areas such as performance fundamentals, design considerations, novelstructures, manufacturing methods, and advanced materials. Although some of theinformation included here may be of particular use to those studied in a specificdiscipline, an effort has been made to convey large portions of the information in away accessible to an audience with general knowledge of electronic packaging.

Chapter 1 introduces the topic with a look at the fundamentals underlyingdesign and performance trade-offs and the additional complexities encounteredat microwave and millimeter wave frequencies. In these regimes, even simpleinterconnects like wire bonds must be considered as complex circuit elements.

v

vi Preface

Chapter 2 introduces a new interconnect approach that allows low-cost high-volume packaging philosophies to be translated into the high-frequency domain.This development will have implications for digital electronic packaging as well.

Chapter 3 shows a possible path for making millimeter wave passive compo-nents with a high-volume production approach. This innovation may enable muchmore pervasive penetration of millimeter wave systems into consumer and other costsensitive applications.

Chapter 4 gives some pointers on how low cost is achieved through chip-on-board integration and packaging for millimeter wave electronics and then discussesthe particular problems of millimeter-wave circuit performance.

Chapter 5 presents the design and development of thin-film liquid crystal poly-mer (LCP) surface mount packages for X, K, and Ka-band applications. Constructedusing multi-layer LCP films, the packages are surface mounted on a printed circuitboard (PCB). Packages include a typical low pass feedthrough design, as well as anew bandpass feedthrough design.

Chapter 6 reviews the design options and the materials available to make portableproducts and discusses ways to meet packaging density and performance needs. Thematerial discussion focuses on types of organic materials used in portable productsand techniques to make PWBs thinner, lighter and cost effective.

Chapter 7 shows how advances in ceramic materials and processing are per-mitting the creation of increasingly complex multi-layer structures. Going beyondsimple routing structures, these ceramics become device elements, and criticalpackaging for micro-electro-mechanical RF systems.

Chapter 8 discusses Laminated Waveguides, characterizing them numerically,and addressing issues in material and process tradeoffs arising when consideringinterconnects on a common substrate at mm-wave frequencies with regards to inser-tion loss and isolation between interconnects. Numerical simulations illustrate thetrade-offs using laminated waveguide or common stripline in the same material set.

Chapter 9 shows the latest developments in both simulation and fabrication ofLTCC for RF/MW packaging applications. It reviews current LTCC fabricationmethods and discusses the trend for RF/MW System in Package modules, highbandwidth design and integrated antenna.

Chapter 10 discusses advances in thermally dissipative composite materials.The discussion covers constituents such as carbon nanotubes, diamond compos-ites, and puts some new spin on well-known materials such as aluminum nitride andberyllium oxide.

Chapter 11 reviews the heat sink material fabrication, application and develop-ment for RF/MW packaging. The discussion covers the traditional, second and thirdgenerations of heat sink materials.

Chapter 12 reviews the latest development of AlN 3D MCM technology forRF/MW packaging. The discussion covers the AlN HTCC process, matching withvarious tungsten pastes, impact of firing profiles and other practical design andmanufacturing issues.

We, the editors, give profuse thanks to all of the contributing authors, not justfor making the effort to inform all of us, but for jumping numerous administrative

Preface vii

hurdles at their various institutions to bring these advances to print. We also wouldlike to thank Mr. Nick Zhou for his help in formatting the draft chapters and in con-verting all figures to grey scale photos. Lastly we would like to thank Steve Elliotand Andrew Leigh from Springer for their consistent support during the long jour-ney to edit this book. Our involvement with these knowledgeable and informativefolks has made the undertaking a great pleasure. It is our sincere hope that the infor-mation conveyed here will illuminate the efforts of subsequent investigators, andinspire technological advances that will have a positive impact for our world.

Contents

1 Fundamentals of Packaging at Microwaveand Millimeter-Wave Frequencies . . . . . . . . . . . . . . . . . . . 1Rick Sturdivant1.1 Wavelength and Frequency . . . . . . . . . . . . . . . . . . . . 31.2 Lumped Elements . . . . . . . . . . . . . . . . . . . . . . . . 31.3 Transmission Lines . . . . . . . . . . . . . . . . . . . . . . . . 5

1.3.1 Dispersion . . . . . . . . . . . . . . . . . . . . . . . . 81.3.2 Dispersion Effects in High Speed Systems . . . . . . . 101.3.3 Transmission Line Distributed Effects . . . . . . . . . 121.3.4 Transmission Line Coupling and Cross Talk . . . . . . 13

1.4 Package Fabrication Methods . . . . . . . . . . . . . . . . . . 151.4.1 Co-fired Ceramics . . . . . . . . . . . . . . . . . . . . 151.4.2 Thick Film and Thin Film Ceramics . . . . . . . . . . 181.4.3 Organic Substrates . . . . . . . . . . . . . . . . . . . 19

1.5 Interconnects . . . . . . . . . . . . . . . . . . . . . . . . . . . 201.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2 Low-Cost High-Bandwidth Millimeter Wave LeadframePackages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Eric A. Sanjuan and Sean S. Cahill2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252.2 MicroCoax Approach . . . . . . . . . . . . . . . . . . . . . . . 26

2.2.1 Packaging Approaches . . . . . . . . . . . . . . . . . 292.2.2 Limitations to the Approach . . . . . . . . . . . . . . 32

2.3 MicroCoax/Leadframe Approach . . . . . . . . . . . . . . . . 322.3.1 Package I/O Structure Considerations . . . . . . . . . 332.3.2 Modelling the Signal Path . . . . . . . . . . . . . . . . 342.3.3 Performance . . . . . . . . . . . . . . . . . . . . . . . 38

2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3 Polymeric Microelectromechanical Millimeter Wave Systems . . . 43Yiin-Kuen Fuh, Firas Sammoura, Yingqi Jiang and Liwei Lin3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

ix

x Contents

3.2 Polymeric Millimeter Wave Systems usingMicromachining Technologies . . . . . . . . . . . . . . . . . . 44

3.3 Fabrication Examples of mm-Wave Components . . . . . . . . 483.3.1 Polymeric Waveguides . . . . . . . . . . . . . . . . . 483.3.2 Waveguide-Based Iris Filters . . . . . . . . . . . . . . 493.3.3 Waveguide-Based Tunable Filters and Phase Shifters . 513.3.4 Waveguide-Fed Horn Antennas . . . . . . . . . . . . . 553.3.5 W-Band Waveguide Feeding Network of a 2×2

Horn Antenna Array . . . . . . . . . . . . . . . . . . 573.4 Fundamental Characterizations of Polymer Metallization Process 59

3.4.1 Surface Roughness . . . . . . . . . . . . . . . . . . . 593.4.2 Characterization of In-channel Electroplating Thickness 613.4.3 Geometry Effects . . . . . . . . . . . . . . . . . . . . 62

3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4 Millimeter-Wave Chip-on-Board Integration and Packaging . . . . 69Edward B. Stoneham4.1 Motivation for a Chip-on-Board Approach for

Millimeter-Wave Product Manufacturing . . . . . . . . . . . . 694.1.1 The Drive for Low Cost . . . . . . . . . . . . . . . . . 694.1.2 Low-Cost Manufacturing Processes . . . . . . . . . . 704.1.3 Problems Specific to Millimeter-Wave Electronics . . . 73

4.2 A Chip-on-Board Solution . . . . . . . . . . . . . . . . . . . . 804.2.1 The Surface-Mount Panel . . . . . . . . . . . . . . . . 814.2.2 Attaching the Bare Chips . . . . . . . . . . . . . . . . 834.2.3 Wire Bond Interconnects . . . . . . . . . . . . . . . . 834.2.4 Eliminating Wire Bonds in the RF Path . . . . . . . . . 844.2.5 Cover Lamination . . . . . . . . . . . . . . . . . . . . 854.2.6 Segregation . . . . . . . . . . . . . . . . . . . . . . . 874.2.7 Testing . . . . . . . . . . . . . . . . . . . . . . . . . . 87

4.3 Application Examples . . . . . . . . . . . . . . . . . . . . . . 874.3.1 A 60-GHz Transceiver . . . . . . . . . . . . . . . . . 884.3.2 Miniaturized 60-GHz Transmitter and Receiver Modules 894.3.3 76-GHz Automotive Radar Module Package . . . . . . 89

4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

5 Liquid Crystal Polymer for RF and Millimeter-WaveMulti-Layer Hermetic Packages and Modules . . . . . . . . . . . . 91Mark P. McGrath, Kunia Aihara, Morgan J. Chen,Cheng Chen, and Anh-Vu Pham5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 915.2 Design and Fabrication of the Thin-Film LCP Package . . . . . 935.3 Lid Construction and Lamination . . . . . . . . . . . . . . . . 975.4 Results and Model of Lowpass Feedthrough . . . . . . . . . . . 98

Contents xi

5.5 Hermeticity and Leak Rate Measurement . . . . . . . . . . . . 1015.6 Reliability of LCP Surface Mount Packages . . . . . . . . . . . 102

5.6.1 Non-operating Temperature Step Stressing . . . . . . . 1035.6.2 Non-operating Thermal Shock Testing . . . . . . . . . 1035.6.3 Operating Humidity Exposure Testing . . . . . . . . . 1055.6.4 Reliability Testing Summary . . . . . . . . . . . . . . 106

5.7 Bandpass Feedthrough . . . . . . . . . . . . . . . . . . . . . . 1065.7.1 Bandpass Feedthrough Design and Fabrication . . . . . 1065.7.2 Bandpass Feedthrough Results and Discussion . . . . . 109

5.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

6 RF/Microwave Substrate Packaging Roadmap for PortableDevices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115Mumtaz Bora6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1156.2 Substrate Materials for Portable Products . . . . . . . . . . . . 1166.3 RF Substrate Materials Thermal and Electrical Properties . . . . 116

6.3.1 Standard FR-4 . . . . . . . . . . . . . . . . . . . . . . 1166.3.2 High TG FR-4 . . . . . . . . . . . . . . . . . . . . . . 1176.3.3 Polyimide . . . . . . . . . . . . . . . . . . . . . . . . 118

6.4 Cyanate Ester Blend (BT- Bismaleamide Triazine) . . . . . . . 1186.5 PTFE Based Laminates . . . . . . . . . . . . . . . . . . . . . . 119

6.5.1 PTFE Resin Coated on Conventional Glass . . . . . . 1196.5.2 PTFE Film Impregnated with Cyanate Ester or

Epoxy Resin . . . . . . . . . . . . . . . . . . . . . . . 1196.5.3 PTFE Mixed with Low Dk Ceramic . . . . . . . . . . 119

6.6 Materials Summary . . . . . . . . . . . . . . . . . . . . . . . . 1206.7 Substrate Critical Properties . . . . . . . . . . . . . . . . . . . 120

6.7.1 Dielectric Constant (Dk) . . . . . . . . . . . . . . . . 1206.7.2 Dissipation Factor/Dielectric Loss: (tan δ) . . . . . . . 1216.7.3 Glass Transition Temperature (Tg) . . . . . . . . . . . 1216.7.4 Glass Decomposition Temperature; Td . . . . . . . . . 1216.7.5 Moisture Absorption . . . . . . . . . . . . . . . . . . 1226.7.6 Coefficient of Thermal Expansion . . . . . . . . . . . 122

6.8 Materials Summary . . . . . . . . . . . . . . . . . . . . . . . . 1226.9 Portable Products Technology Roadmap . . . . . . . . . . . . . 1226.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266.11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

7 Ceramic Systems in Package for RF and Microwave . . . . . . . . 129Thomas Bartnitzek, William Gautier, Guangwen Qu,Shi Cheng, and Afshin Ziaei7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1297.2 RF-PLATFORM . . . . . . . . . . . . . . . . . . . . . . . . . 129

xii Contents

7.2.1 LTCC for Systems in Package . . . . . . . . . . . . . 1307.2.2 Design of Ceramic Packages . . . . . . . . . . . . . . 1317.2.3 Why Multi-Project Wafers Made of LTCC? . . . . . . 1317.2.4 Hermetic Capping of MEMS with Ceramic Lids . . . . 1327.2.5 LTCC Packages for Advanced RF and

Microwave Applications . . . . . . . . . . . . . . . . 1337.3 Three Examples . . . . . . . . . . . . . . . . . . . . . . . . . 135

7.3.1 4 by 4 Patch Antenna Array for Operation at 35 GHz . 1357.3.2 LTCC for 77–81 GHz Automotive Radar

Systems-in-Package . . . . . . . . . . . . . . . . . . . 1427.3.3 24 GHz Switched Beam Steering Array

Antenna Based on RF MEMS Switch Matrix . . . . . . 1457.4 RF-MEMS for Radar and Telecom Applications . . . . . . . . 155

7.4.1 Research Activities and Trends on RF-MEMS Switches 156References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

8 Low-Temperature Cofired-Ceramic Laminate Waveguidesfor mmWave Applications . . . . . . . . . . . . . . . . . . . . . . . 165Jerry Aguirre8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1658.2 The Laminated Waveguide . . . . . . . . . . . . . . . . . . . . 1668.3 Transitions to a LWG . . . . . . . . . . . . . . . . . . . . . . . 1678.4 Rectangular Waveguide Theory . . . . . . . . . . . . . . . . . 1698.5 LTCC Process . . . . . . . . . . . . . . . . . . . . . . . . . . 1748.6 Insertion Loss in an LTCC Laminated Waveguides . . . . . . . 1748.7 U- band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1778.8 V-band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1788.9 E-band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1788.10 W-band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1788.11 F-band . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1848.12 LWG-to-LWG Coupling . . . . . . . . . . . . . . . . . . . . . 1848.13 LWG vs. Stripline . . . . . . . . . . . . . . . . . . . . . . . . 1848.14 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

9 LTCC Substrates for RF/MW Application . . . . . . . . . . . . . . 189Jian Yang and ZiliangWang9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1899.2 LTCC Fabrication Process . . . . . . . . . . . . . . . . . . . . 1929.3 Current Status and Trend . . . . . . . . . . . . . . . . . . . . . 197References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

10 High Thermal Dissipation Ceramics and CompositeMaterials for Microelectronic Packaging . . . . . . . . . . . . . . . 207Juan L. Sepulveda and Lee J. Vandermark10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

Contents xiii

10.2 Ceramics and Carbon Based Materials . . . . . . . . . . . . . . 21010.2.1 Common Packaging Ceramics . . . . . . . . . . . . . 21010.2.2 LTCC . . . . . . . . . . . . . . . . . . . . . . . . . . 21010.2.3 High Performance Packaging Ceramics (BeO AlN) . . 215

10.3 Direct Bond Copper (DBC) Packaging . . . . . . . . . . . . . . 21810.4 RF/MW Brazed Packages . . . . . . . . . . . . . . . . . . . . 22010.5 Thin-Film Packaging . . . . . . . . . . . . . . . . . . . . . . . 22110.6 Thick-Film Packaging . . . . . . . . . . . . . . . . . . . . . . 22110.7 Carbon Nanotubes (CNT) . . . . . . . . . . . . . . . . . . . . 22310.8 Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

10.8.1 Metal Matrix Composites . . . . . . . . . . . . . . . . 22310.8.2 Cu/cBN Composites . . . . . . . . . . . . . . . . . . 22710.8.3 Cu/SiC Composites . . . . . . . . . . . . . . . . . . . 22810.8.4 Al/Diamond Composites . . . . . . . . . . . . . . . . 228

10.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

11 High Performance Microelectronics Packaging Heat SinkMaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233Jiang Guosheng, Ken Kuang, and Danny Zhu11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23311.2 Refractory Metal Based Microelectronics Packaging Materials . 236

11.2.1 Development, Manufacturing and Applicationof Copper Tungsten . . . . . . . . . . . . . . . . . . . 236

11.2.2 Development, Manufacturing and Applicationof Copper Molybdenum (MoCu) . . . . . . . . . . . . 241

11.2.3 Development, Manufacturing and Applicationof Copper-Molybdenum-Copper Laminatesand Copper-Copper/Molybdenum-Copper Laminates . 244

11.3 Aluminum Based Heat Sink Materials . . . . . . . . . . . . . . 24811.3.1 AlSiC Heat Sink Materials . . . . . . . . . . . . . . . 248

11.4 New Development for Microelectronics Packaging HeatSink Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 258

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262

12 Technology Research on AlN 3D MCM . . . . . . . . . . . . . . . . 267Zhang Hao, Cui Song, and Liu Junyong12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26712.2 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

12.2.1 Co-fired Spacer Rod and 2D MCM Substrate . . . . . 26912.2.2 Vertical Interconnected by BGA Solder Ball . . . . . . 26912.2.3 AlN 3D MCM Package . . . . . . . . . . . . . . . . . 26912.2.4 Technological Method . . . . . . . . . . . . . . . . . 270

12.3 Result and Discussion . . . . . . . . . . . . . . . . . . . . . . 27012.3.1 General Technological Scheme . . . . . . . . . . . . . 27012.3.2 Layout and Interconnect Design . . . . . . . . . . . . 271

xiv Contents

12.4 Matching Optimization Research on W paste and AlN Ceramics 27212.4.1 Technological Improvement Experiment of

AlN 2D MCM Substrate . . . . . . . . . . . . . . . . 27312.4.2 The Making of Spacer Rod . . . . . . . . . . . . . . . 27412.4.3 Package Technology . . . . . . . . . . . . . . . . . . 27612.4.4 Vertical Interconnected Technology Research . . . . . 276

12.5 Result of Experiment . . . . . . . . . . . . . . . . . . . . . . . 27812.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

Contributors

Jerry Aguirre Kyocera America, Inc., San Diego, CA 92123, USA

Kunia Aihara Microwave Microsystems Laboratory, University of California,Davis, CA 95616, USA

Thomas Bartnitzek VIA electronic GmbH, Robert-Friese-Straße 3, DE-07629Hermsdorf

Mutaz Bora Peregrine Semiconductors, Inc., San Diego, CA 92121, USA

Sean S. Cahill BridgeWave Communications, Inc., Santa Clara, CA 95054, USA

Cheng Chen Microwave Microsystems Laboratory, University of California,Davis, CA 95616, USA

Morgan J. Chen Microwave Microsystems Laboratory, University of California,Davis, CA 95616, USA

Shi Cheng University of Upsala, Sweden

Yiin-Kuen Fuh Department of Mechanical Engineering, University of California,Berkeley, CA 94720, USA

William Gautier EADS Innovative works, Germany

Jiang Guosheng Changsha Saneway Electronic Materials Co., Ltd, Central SouthUniversity, Changsha, Hunan, China 410012

Zhang Hao East China Research Institute of Microelectronics, Hefei, Anhui,China 230022

Yingqi Jiang Department of Mechanical Engineering, University of California,Berkeley, CA 94720, USA

Liu Junyong East China Research Institute of Microelectronics, Hefei, Anhui,China 230022

Ken Kuang Torrey Hills Technologies, LLC, San Diego, CA 92121, USA

xv

xvi Contributors

Liwei Lin Department of Mechanical Engineering, University of California,Berkeley, CA 94720, USA

Mark P. McGrath Microwave Microsystems Laboratory, University ofCalifornia, Davis, CA 95616, USA

Anh-Vu Pham Microwave Microsystems Laboratory, University of California,Davis, CA 95616, USA

Guangwen Qu University of Uppsala, Sweden

Firas Sammoura Department of Mechanical Engineering, University ofCalifornia, Berkeley, CA 94720, USA

Eric A. Sanjuan BridgeWave Communications, Inc., Santa Clara, CA 95054,USA

Juan L. Sepulveda Materials and Electrochemical Research, Tucson, AZ 85706,USA

Cui Song East China Research Institute of Microelectronics, Hefei, Anhui, China230022

Edward B. Stoneham Stoneham Innovations, Los Altos, CA 94024, USA;Endwave Corporation, San Jose, CA 95134, USA

Rick Sturdivant Microwave Packaging Technology, Inc., Brea, CA 92821, USA

Lee J. Vandermark Brush Ceramic Products, Tucson, AZ 85706, USA

Ziliang Wang Nanjing Electronics Devices Institute, Nanjing, Jiangsu, China210016

Jian Yang Nanjing Electronics Devices Institute, Nanjing, Jiangsu, China 210016

Danny Zhu Jiangsu Dingqi Sci. & Tech. Co. Ltd., Yixing, Jiangsu, China 214200

Afshin Ziaei THALES TRT, France

Chapter 1Fundamentals of Packaging at Microwaveand Millimeter-Wave Frequencies

Rick Sturdivant

Abstract The thirst for higher data rates and greater bandwidth has resulted inincreased interest in millimeter-wave systems as a means for local and wider areainformation transport. At the same time there is a real need for lower cost andmore compact systems. These requirements have led to the development of a highlyintegrated millimeter-wave System In Package (SIP), which operates beyond 40GHz. This solution uses low cost ceramic packaging as well as optimized inter-connects and transitions to allow for wide-band electrical performance. This paperwill present details on the functionality, electrical performance and packaging usedto realize this solution.

Packaging at microwave and millimeter-wave frequencies has the same challengesas packaging at lower frequencies except that there are several additional complex-ities. An example is distributed effects. This is an issue at high frequencies becausecircuit features and components can have dimensions that are an appreciable frac-tion of a wavelength. This causes circuit elements to have electrical characteristicsthat change as frequency increases. For instance, a wire bond is a simple connectionpoint at low frequencies. However, at microwave frequencies a wire bond performsmore like an inductor and at millimeter-wave frequencies it can perform more like aresonator or antenna. Distributed effect concerns dominate the design procedure forpackaging at high frequencies.

The requirement to treat metal traces as transmission lines and interconnectingvias as signal transitions adds another layer of complexity. For instance, features asminute as a bend in a signal trace will degrade performance if not carefully designed.A via carrying a signal from one layer to the next can create a signal transition whichcan limit electrical bandwidth. Often times, extreme effort is required to accuratelymodel and predict the performance of interconnects and transitions. The use of threedimensional numerical simulation tools such as the finite element or finite differencemethod is common.

R. Sturdivant (B)Microwave Packaging Technology, Inc., Brea, CA 92821, USAe-mail: [email protected]

1K. Kuang et al. (eds.), RF and Microwave Microelectronics Packaging,DOI 10.1007/978-1-4419-0984-8_1, C© Springer Science+Business Media, LLC 2010

2 R. Sturdivant

Another layer of complexity is coupling and radiation. Adjacent metal traces,traces next to integrated circuits and traces between layers can couple energy. Thiscan be a useful effect and circuits such as directional couplers and baluns can becreated using coupling. However, coupling is often the hidden enemy of the designerof microwave and millimeter-wave packaging. Coupling and radiation effects canbe difficult to model and often only reveal themselves during electrical test. Evenduring electrical testing, it can be a challenge to precisely determine the locationand cure for a coupling or radiation problem. Such effects can lead to resonances oramplifier oscillation.

Material selection is an important part of the packaging process at high frequen-cies. One of the reasons is material selection will affect the line impedance andinsertion loss of transmission lines. In addition to the material itself, the thicknessof dielectric layers and metals layers will affect the design of the transmission line.Also, the materials must be selected to minimize interaction with integrated circuitsas in the case of a glob top or under-fill application.

Another major concern when packaging at high frequencies is the thermal powerdensity that is often associated with high frequency components. This is especiallyan issue for RF power amplifiers, which can have power densities of hundreds orthousands of watts per square centimeter.

Often these requirements work against each other. For instance, the requirementto design for distributed effects and the need to provide thermal paths for high powerdevices are usually in conflict. The designer of high power packaging often spendsa large amount of effort balancing these, often conflicting, requirements.

“Normal” Packaging Issues:

• Choose compatible materials for reliability• Die attach method and interconnect method• Metal system, CTE matching• Sealing and die encapsulation

Additional Packaging Issues for Microwave and Millimeter-wave Frequencies:

• Design of the metal pattern and dielectric thickness to maintain required lineimpedance.

• Short interconnect lengths to minimize reflections.• Careful material selection to minimize effect on electromagnetic fields in inte-

grated circuits• Coupling between traces and package resonance• Active devices often have high power dissipation

Radio frequency (RF) ranges are typically described by bands of operation.Table 1.1 shows the frequency spectrum and the name or letter designation foreach band. Use of these designations is common among professionals developingproducts and technology for high frequency applications.

1 Fundamentals of Packaging at Microwave and Millimeter-Wave Frequencies 3

Table 1.1 Frequency bands and their names and letter designations. IEEE Std 521-2002 (originalstandard adopted in 1984)

Band Frequency range

HF band 3–30 MHzVHF band 30–300 MHzUHF band 300–1000 MHzL Band 1–2 GHzS band 2–4 GHzC band 4–8 GHzX band 8–12 GHzKu band 12–18 GHzK band 18–27 GHzKa band 27–40 GHzV Band 40–75 GHzW band 75–110 GHzmm band 110–300 GHz

1.1 Wavelength and Frequency

Distributed effects become important when circuit element’s physical features arean appreciable portion of a wavelength. Consider for example a 10 mm packageoperating at 10 MHz. At that frequency, the free space wavelength is about 30 mwhich is 3000 times larger than the package. However, at 10 GHz, the free spacewavelength is about 30 mm which is only 3 times larger than the package. Therefore,the features of the package can have a significant effect on the electrical performanceof the package.

Equation 1.1 shows the relationship of wavelength, frequency and materialdielectric constant. It is important to note that for materials with larger dielectricconstant, the wavelength is smaller which means that distributed effects becomemore important. This means that higher dielectric constant materials will be moresensitive, in general, to circuit features and will have more coupling.

λ = v

f√

εr(1.1)

Where: λ = wavelength, v = free space velocity = 299,792,458 m/sec, εr = dielec-tric constant of the material (unity for free space).

1.2 Lumped Elements

The issue with lumped elements is that above a certain frequency, lumped elementsno longer perform electrically like lumped elements. The frequency at which a par-ticular inductor, capacitor or resistor begins to deviate from the ideal or desired

4 R. Sturdivant

Fig. 1.1 A comparison of measured data for a lumped inductor, an ideal inductor and electromag-netic simulations of an embedded spiral inductor. Simulations performed using Microwave Office©

from Applied Wave Research [1]

frequency response depends upon the type of lumped element, manufacturingtechnique and manufacturer.

As a result, at microwave and millimeter-wave frequencies, lumped elements areoften integrated or “embedded” as part of the packaging. The performance benefitof integration is that the lumped elements can be designed to perform to higherfrequencies. Also, integration usually leads to smaller size and, often, lower cost.

Figure 1.1 compares an ideal lumped element inductor with measured data foran inductor and electromagnetic simulation results for an embedded spiral inductor.Notice how the ideal lumped inductor insertion loss (S21) increases smoothly as afunction of frequency. This is exactly what you would expect from a series inductiveelement. However, the measured data for the lumped inductor shows that the inser-tion loss begins to deviate from ideal at about 0.5 GHz. Observe from the graphhow the behavior of the lumped inductor not only deviates from the ideal inductorslope, it also has a resonance at about 1.8 GHz. This type of response means thatthe lumped inductor is not very useful above about 0.5 GHz. The figure also showsthe electromagnetic simulation results for an embedded spiral inductor. It followsthe ideal inductor performance to about 1 GHz or about double the frequency of thelumped element. Though the example is for an inductor, most capacitors and resis-tors embedded elements perform over a wider frequency range than their lumpedequivalent value components.

Embedded lumped elements can be realized using most of the available fab-rication technologies available including low temperature co-fired ceramic, hightemperature co-fired ceramic, thick-film, thin-film and laminates.


Fig. 1.2 Spiral inductor model used in the electromagnetic simulations

Figure 1.2 shows the model used in the electromagnetic simulations of theembedded inductor. The spiral inductor is formed by the metal pattern on thedielectric substrate. The bandwidth of the spiral inductor is limited by the straycapacitance to ground plane and capacitance between the windings.

1.3 Transmission Lines

A transmission line is a metal conductor that transports electrical signals. An idealtransmission line is shown in Fig. 1.3. The transmission line is connected betweena generator and a load and transfers energy between them.

Fig. 1.3 Transmission line with line impedance Zo and phase constant γ connected between asource with impedance ZS and load with impedance ZL

A transmission line can be described electrically by its characteristic impedance,Zo, and its propagation constant. The transmission line physical dimensions,

6 R. Sturdivant

substrate material, and substrate dimensions determine these parameters. The prop-agation constant has a loss term and a phase constant term. Equation 1.2 shows therelationship:

γ = α + jβ (1.2)

Where: α = attenuation constant = αc + αd, αc = loss due conductors, αd = lossdue to dielectric, β = phase constant.

The loss due to conductors is often calculated using the incremental inductancerule [2]. However, most modern transmission line analysis programs will predictconductor loss. The loss is due to the finite conductivity of the metals used in thetransmission line. The finite conductivity causes a portion of the RF energy to beconverted into heat. Another layer of complexity is that the conductor loss is depen-dent upon the skin depth. It is a measure of the distance the RF current penetratesinto the conductor. At low frequencies, the current can be considered to be uniformlydistributed within a conductor. However, at high frequencies, the RF current is con-fined to the surface of the metal. This increases the current density, which increasesthe conversion of RF energy into heat, and therefore increases insertion loss. Therelationship for skin depth is given in Equation (1.3).

δ = 1√πμσ f

(1.3)

Where: μ = permeability of the metal = μoμr, σ = conductivity of the metal,f = frequency.

The dielectric loss depends upon the material characteristics of the dielectricmaterial surrounding or supporting the transmission line metal. The parameter ofinterest is loss tangent. Loss tangent is the ratio (or angle) in the complex planeof the lossy portion of the electric field and the lossless reactive portion. Therelationship is:

∇ × H = jωε′E + (ωε′′ + σ )E

The designer rarely has access to the fundamental parameters tan δ = ωε′′+ σ

ωε′ in

the equation above. However, many manufacturers will supply loss tangent. It can beused to calculate the insertion loss due to dissipation in the dielectric. Equation (1.4)shows the relationship for the dielectric loss. This equation assumes TEM wavepropagation.

αd = π

λtan δ (1.4)

Where: λ = wavelength, tan δ = loss tangent (often given in material datasheets).Table 1.2 shows a few of the more typical transmission lines along with their

benefits, drawbacks and typical use. One of the most common transmission lines ismicrostrip. It is particularly useful in planar circuit applications.


Table 1.2 Benefits, drawbacks and uses for typical transmission lines

Transmission Line Type Benefits Drawbacks Typical Uses

Coax

1. Good isolation due

to the external ground shield.

2. Low dispersion.

3. Wide band.

1. Difficult access to the signal line due to ground shield.

2. Physically large

1. TV and cable signal.

2. Lab testing.3. Instrumen- tation.

Microstrip

1. Physically small.

2. Low cost and ease of manufacturing.

3. Large industry base of compatible

circuits and components.

4. Easy access to signal line.

1. Low isolation.

2. Higher dispersion.

3. Higher order mode propagation is MMW frequencies.

1. Printed circuit boards.

2. Micr owave hybrids.

3. Antennas4. Passive

circuits and components.

5. MMICs

Stripline


2. No dispersion.3. Physically

small.4. Low radiation

and can be low coupling.

1. Difficult access to signal line.

2. Higher order mode propagation.

1. Buried signals in PWB and ceramic packages.

2. Signal distribution.

3. Couplers and other components.

Coplanar WaveguideWith Ground

1. Small size2. Low cost

and ease of manufacturing.

3. Large industrybase of compatible circuits and components.


5. Lower dispersion

1. Prone to higher order propagation modes and resonances. Requires careful via placement.

1. Printed circuit boards.

2. Microwave hybrids.

3. Antennas4. MMICs

Coplanar Waveguide

1. Physically small.



4. Low dispersion.

1. Requires connection of grounds at discontinuities.

1. Antennas.2. Suspended

substrate circuits and components.

3. MMICs.

8 R. Sturdivant

Coax should be familiar to most as the transmission line used to connect to atelevision or cable box. Fig. 1.4 shows an example of coax. It has a center con-ductor surrounded by a dielectric. A metal ground shield surrounds the dielectric.Plastic usually encloses the assembly though some industrial and lab grade coaxialline does not. The ground shield provides excellent isolation that is important formost systems where coaxial cabling is being used to connect between subsystemsor modules. Coax can be very wide band and has very low dispersion.

Fig. 1.4 Coaxialtransmission line

Microstrip can be fabricated using many different types of technology. Typicallyit is etched onto printed circuit boards. However, it can also be fabricated by etchingthin film or by printing thick film on ceramics. Also, microstrip is often used inMonolithic Microwave Integrated Circuits (MMICs).

Stripline is used for passive components such as couplers and power dividers.It is also used as a buried transmission line in multiplayer circuits. Stripline pro-vides excellent shielding with proper via placement and very broad bandwidths.However it is difficult to connect components to the signal line since it is buried inthe dielectric. Traditionally, stripline was fabricated by a sandwich of soft circuitboards.

Coplanar Waveguide (CPW) can be used with or without a bottom side groundplane. CPW is often used in planar circuits. In these instances, a ground plane is usu-ally required to provide isolation from other circuits in buried layers. When a groundplane is used, the circuit is said to be Conductor Backed Coplanar Waveguide(CBCPW) or Coplanar Waveguide With Ground (CPWG). Both terms are used todescribe the same transmission line. If a bottom ground is used, vias must connectthe topside ground to the bottom ground. This is to reduce the effect of the zero cutoff transverse radiation mode and the patch resonance that can be set up by the topside ground planes.

1.3.1 Dispersion

In transmission lines, dispersion occurs when the propagation constant is not lin-ear with frequency. The effect can be devastating for some signals. For instance,


Fig. 1.5 Simulation results of a microstrip line on 0.25 and 0.75 mm alumina and a coaxial lineshowing the dispersion effect of the inhomogeneous microstrip transmission line. Electromagneticsimulations conducted using EM Sight Simulator from Applied Wave Research [3]

dispersion can result in increased bit error rates in wide band, high data ratetelecommunications systems.

Figure 1.5 shows the group delay simulation results from an electromag-netic analysis of a 50 � transmission line on 0.25 and 0.75 mm thick alumina(er = 9.9) and coaxial line. Group delay is the change in phase constant as a functionof radian frequency (dφ/dω). It can be used to show dispersion. For a non-dispersivetransmission line such as coax, the group delay is flat as a function of frequency.This means that for coax, the phase is linear with frequency and the transmissionline does not show dispersion. Microstrip line, on the other hand, does show dis-persion. For the 0.75 mm thick substrate, the group delay begins to diverge fromideal flat group delay at about 3 GHz. The ripple in the group delay is due to thetransmission line being slightly off from 50 �. The absolute change in group delayfrom zero to 30 GHz is about 26 pS for microstrip on 0.75 mm thick alumina, but isonly about 9.8 pS for the 0.25 mm thick alumina substrate.

Since the propagation constant in a dispersive transmission lines is not linear, theenergy propagates at a different group velocity as frequency changes. This meansthat if a wide band signal is injected onto a dispersive transmission line, some of theenergy will arrive at a different time relative to other energy at a different frequency.This results in pulse distortion in wide band systems. However, for most narrowband systems, dispersion can be ignored.

Dispersion occurs because a transmission line is not homogeneous. An exampleof a homogeneous line is coax. The dielectric material that the wave experiences

10 R. Sturdivant

is continuous. Another example is stripline, since it has a homogenous dielectric.Microstrip uses two dielectrics (air and the dielectric under the strip) and is there-fore not homogeneous. In other words, the air above the microstrip line causes thedispersion. If the dielectric were continuous with no air above, the transmission line,ideally, would not have dispersion.

1.3.2 Dispersion Effects in High Speed Systems

Dispersion is important because it affects signal integrity, which is an importantconsideration in the design of most high-speed systems. This is especially truefor telecom and datacom systems, which may be operating at many gigabits persecond. Fig. 1.6 shows the effects of dispersion on eye quality of a 10 Gb/s signalby comparing a 12.7 mm (0.500′′) length of coaxial line and microstrip line. It canbe seen that the coaxial line has essentially no effect on eye quality. The microstripline has overshoot and the eye closes early. In addition, the rise and fall times havebeen reduced.

To be fair, most designers would not choose a 0.75 mm thick alumina substratefor a 10 Gb/s application. However, the results demonstrate that dispersion can havea significant effect on eye performance.

Consider the case of the 0.25 mm thick alumina substrate. It has significantly lessdispersion and as can be seen in Fig. 1.7, the eye performance is commensuratelyimproved. The improvement is due to use of a thinner substrate. A thinner substratehas less dispersion which is the same as saying that its phase is more linear (flattergroup delay).

It has been shown that transmission line dispersion can have a negative effect oneye quality. A few general design considerations are given for reducing dispersionin microstrip as a guide to the designer:

Substrate Thickness Effects: The substrate thickness, h, affects dispersion.A thicker substrate causes more dispersion. From a simple consideration, this seemsto follow from the fact that the inhomogeneity of the transmission line is increasedwith more dielectric. A more rigorous approach is to compare the substrate thicknessto the wavelength in the material. As the ratio of substrate thickness to wavelengthin the material is reduced, the dispersion is also reduced. Therefore as frequency isincreased, the substrate thickness should be decreased to maintain a required levelof linear phase (flat group delay response).

Substrate Dielectric Constant Effects: The substrate dielectric constant alsoaffects dispersion and eye quality. For a given substrate thickness, a lower dielec-tric constant has less effect on eye performance. Again, this makes sense fromthe consideration of the inhomogeneous transmission line. The lower the dielectricconstant, the closer it is to air, which is the homogeneous (no dispersion) case. Also,a lower dielectric constant reduces the ratio of substrate thickness to wavelength.

Transmission Line Length: The effect of a dispersive transmission line can bereduced if the lengths of dispersive lines are minimized. This may seem to be


Fig. 1.6 The effect of dispersion on eye quality for a 10 Gb/s signal can be seen by comparing12.7 mm (0.500′′) length of (a) coaxial transmission line and (b) 0.75 mm thick alumina

obvious. However, the effect of long, dispersive transmission lines is often missedby circuit and board designers.

h

λ= hf

√εr

vo≤ 0.05

12 R. Sturdivant

Fig. 1.7 The effect of dispersion on eye quality for a 10 Gb/s for a 0.25 mm thick alumina substrate

A fairly conservative design goal is to have the substrate thickness less than 5%of a wavelength in the material. The equation below gives the relationship:

h = 0.05vo

f√

εr

This equation leads to the relationship for choosing the substrate thickness:Where: λ = wavelength in the dielectric material, νo = velocity (free space),f = frequency, h = substrate height.

For example, consider a 10 Gb/s signal. As a general rule, the transmission linesshould perform well to a frequency that is at least twice the frequency for a RZ(Return to Zero) signal. For 10 Gb/s, this translates into a transmission line band-width of not less than 20 GHz. For high purity alumina (er = 9.9), the substratethickness using the above equation is 0.25 mm.

The equation is fairly conservative and it is possible to achieve successful prod-ucts with a ratio of h/λ as high as 0.15 or 0.2. However, the challenge will be toreduce line lengths.

1.3.3 Transmission Line Distributed Effects

Another level of complexity in the design and development of microwave andmillimeter-wave packaging is distributed effects. They are caused by the fact that


the circuit dimensions are an appreciable portion of a wavelength. Because of dis-tributed effects, the metal traces must be treated as transmission lines. Also, smallfeatures or discontinuities in the signal traces can have a significant effect on circuitperformance.

Consider a microstrip transmission line with a simple bend. A bend is a verycommon circuit feature. Fig. 1.8 shows the bend realized three different ways. Thefirst is square which may be first choice for making a bend in a metal trace. Thesecond is mitered which has half of the metal removed from the signal line. Lastly,there is the optimum miter which has additional metal removed from the bend. Otherrealizations are possible such as a smooth curve, but are not being considered in thisanalysis.

Fig. 1.8 Three methods to realize a bend in a microstrip line

The circuits were analyzed using electromagnetic simulation from 0 to 30 GHz.The model for the optimum mitered bend is shown in Fig. 1.9 for the case of0.25 mm thick alumina (εr = 9.9).

The results are shown in Fig. 1.10. The curve shows insertion loss which is ameasure of the amount of energy that is lost due to reflections of energy, absorptionor radiation. The results show that the square bend has more insertion loss than themitered bends. The optimum mitered bend has the best performance.

Many other discontinuities can exist in a package at microwave and millimeter-wave frequencies. Because of distributed effects significant care must be taken whendesigning so that the impact on electrical performance can be miminized.

1.3.4 Transmission Line Coupling and Cross Talk

At microwave and millimeter-wave frequencies coupling can occur between metaltraces, devices and integrated circuits. The coupling can be significant and extremesteps are often taken to mitigate its effects. Fig. 1.11 shows a pair of coupledmicrostrip lines.

Energy in one line will coupled to the other line across the gap (S) between thelines. In general the coupling between the lines is:

14 R. Sturdivant

Fig. 1.9 Model used in the electromagnetic analysis of the bend. Electromagnetic simulationsconducted using EM Sight Simulator from Applied Wave Research [3]

Fig. 1.10 Insertion loss for the square bend, mitered bend and optimum mitered bend


Fig. 1.11 Coupled microstrip lines

(1) Proportional to frequency at low frequency (below the point where the lengthof the lines is a quarter wave long).

(2) A function of frequency at high frequencies and exhibits a resonant behavior.(3) Proportional to the dielectric constant.(4) Decreased as the substrate thickness is decreased.

1.4 Package Fabrication Methods

Packages and substrates can be fabricated using several commercially availabletechnologies. The fabrication methods can be broken into described and organizedmany different ways. We have divided the methods into three groups using the basematerial:

(1) Co-Fired Ceramics(2) Thick and Thin Film Ceramics(3) Organic Materials

The focus of this section will be upon the electrical characteristics as well as thegeneral benefits and drawbacks of each technology.

1.4.1 Co-fired Ceramics

Co-fired ceramics are fabricated with “green” tape layers, which are fired at hightemperatures. The result is a homogeneous ceramic with three-dimensional metalfeatures formed by metal traces and filled vias that were part of each layer priorto firing. Fig. 1.12 shows a typical co-fired ceramic process flow. At the start ofthe process, raw materials are received such as alumina, quartz, glass and binders.The materials are prepared into a slurry, then cast into “green” tape. After the tapehas dried, it is punched with vias and the vias are filled with metal. Metallizationlayers are then added using a printing method. Next, the packages may be singulatedor may remain in an array where they are stacked and sintered (or fired) at high

16 R. Sturdivant

Fig. 1.12 Typical process flow for co-fired ceramics. (Courtesy of Adtech Ceramics)

temperatures. Finally, the parts are plated and brazed if required. There are manyvariations on this basic process depending upon the material and the package design.However, the basic processing concept is similar for co-fired ceramics.

Some typical co-fired ceramics are High Temperature Co-fired Ceramic (HTCC)alumina and aluminum nitride. HTCC alumina has been used extensively forintegrated circuit and multichip module packaging for a variety of applicationsincluding microprocessors, military radar modules and high speed telecom mod-ules. The benefits are hermeticity, good metal adhesion, good thermal conductivity(20–25 W/mK), excellent compatibility with brazed metal, high strength, relativelylow cost and good electrical performance at microwave frequencies. The main drawback to HTCC ceramic at microwave and millimeter-wave frequencies is that refrac-tory metals such as Molybdenum (Mo) or Tungsten (W) must be used for the internalmetal features. The resistive losses in Mo and W are higher than gold metallizationby a factor of about 2.5.

rf and microwave microelectronics packaging · 2013-07-18 · rf and microwave microelectronics...

Documents