Waveguide-Integrated MEMS Concepts for Tunable Millimeter ...705226/FULLTEXT01.pdf · Waveguide-Integrated MEMS Concepts for Tunable Millimeter-Wave Systems ZARGHAM BAGHCHEHSARAEI

Waveguide-Integrated MEMS Concepts forTunable Millimeter-Wave SystemsZARGHAM BAGHCHEHSARAEI

Waveguide-Integrated MEMS Concepts forTunable Millimeter-Wave Systems

ZARGHAM BAGHCHEHSARAEI

Doctoral ThesisStockholm, Sweden 2014

Front cover picture:Scanning electron microscope (SEM) images of a fabricated device, consisting of aMEMS-reconfigurable surface and on-chip electrostatic combdrive actuators, to beintegrated in a WR-12 rectangular waveguide as the first millimeter-wave MEMSwaveguide switch (left) and close-up view of the MEMS-reconfigurable surface show-ing gold-covered 5-µm wide vertical cantilevers and 25-µm wide horizontal suspen-sion bars (right).

TRITA-EE 2014:012ISSN 1653-5146ISBN 978-91-7595-062-4

KTH Royal Institute of TechnologySchool of Electrical Engineering

Department of Micro and Nanosystems

Akademisk avhandling som med tillstånd av Kungliga Tekniska högskolan framläggestill offentlig granskning för avläggande av teknologie doktorsexamen i elektrisk mät-teknik fredagen den 4 april 2014 klockan 10.00 i F3, Lindstedtsvägen 26, Stockholm.

Thesis for the degree of Doctor of Philosophy at KTH Royal Institute of Technology,Stockholm, Sweden.

© Zargham Baghchehsaraei, April 2014

Tryck: Universitetsservice US AB, Stockholm, 2014

iii

Abstract

This thesis presents two families of novel waveguide-integrated compo-nents based on millimeter-wave microelectromechanical systems (MEMS) forreconfigurable systems. The first group comprises V-band (50–75 GHz) andW-band (75–110 GHz) waveguide switches and switchable irises, and theirapplication as switchable cavity resonators, and tunable bandpass filters im-plemented by integration of novel MEMS-reconfigurable surfaces into a rect-angular waveguide. The second category comprises MEMS-based reconfig-urable finlines integrated as phase shifters into a rectangular waveguide arrayto demonstrate beams steering with a phased array antenna.

The first group of the presented reconfigurable waveguide components isbased on a novel MEMS-reconfigurable surface structured in the device layerof a silicon-on-insulator (SOI) wafer using metallized mono-crystalline siliconas structural and functional material. The chip containing the reconfigurablesurface is integrated in the cross-section of a WR-12 rectangular waveguideperpendicular to the wave propagation. The reconfigurable surface is modifiedfor different states by on-chip push-pull electrostatic comb-drive MEMS actu-ators. The switch is ON when the reconfigurable surface is in its transmissivestate and OFF when the reconfigurable surface is in its blocking state for thepropagating wave. This millimeter-wave waveguide switch shows an insertionloss and isolation very similar to high-performance but bulky mechanical ro-tary waveguide switches, despite being extremely compact (30 µm thick), andthus combines the high electrical performance of mechanical switches with thesize of (high power consuming and inferior performance) PIN-diode waveg-uide switches. This thesis also investigates the optimization to decrease thenumber of contact points for the OFF state and presents a device yield anal-ysis. The same concept is developed further to MEMS-switchable inductiveand capacitive irises, with the performance similar to ideal irises. With suchMEMS-reconfigurable irises a switchable cavity resonator was implementedand the potential of tunable bandpass filters are demonstrated. Since thesedevices feature all-metal design as no dielectric layers are utilized, no dielectriccharging effect is observed. Furthermore, this thesis investigates the low-lossintegration of millimeter-wave MEMS-reconfigurable devices into rectangularwaveguide with conductive polymer interposers.

The second group of components comprises finlines which are fabricatedout of two bonded silicon wafers with bilateral gold structures integrated intoa WR-12 rectangular waveguide. A 2-bit waveguide phase shifter is designedfor 77-GHz automotive radar. Such phase shifters are used as individualbuilding blocks of a two-dimensional antenna array for beam steering front-ends.

Zargham Baghchehsaraei, zargham@ kth. seDepartment of Micro and Nanosystems, School of Electrical EngineeringKTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden

v

To Zhaleh and Rahim, my parents, whom I always adore,and to Farzam, my lovely brother, whom I always admire

Contents

List of Publications ix

List of abbreviations xiii

List of radar-frequency letter bands(IEEE standard 521-1984) xv

1 Introduction and outline 1

2 Background to RF and microwave MEMS 32.1 Microelectromechanical systems . . . . . . . . . . . . . . . . . . . . . 32.2 RF and microwave MEMS . . . . . . . . . . . . . . . . . . . . . . . . 3

2.2.1 RF MEMS switches for millimeter-wave applications . . . . . 52.2.2 Tunable MEMS planar filters and phase shifters for millimeter-

wave applications . . . . . . . . . . . . . . . . . . . . . . . . . 62.3 Challenges of RF MEMS technology . . . . . . . . . . . . . . . . . . 72.4 SOI-based RF MEMS devices . . . . . . . . . . . . . . . . . . . . . . 8

3 Background to conventional tunable waveguide components 93.1 Waveguide switch technologies . . . . . . . . . . . . . . . . . . . . . 93.2 Tunable cavity resonators and waveguide filters . . . . . . . . . . . . 103.3 Variable phase shifters . . . . . . . . . . . . . . . . . . . . . . . . . . 12

4 Previous work on tunable MEMS waveguide components 154.1 MEMS waveguide switches . . . . . . . . . . . . . . . . . . . . . . . 154.2 Tunable MEMS cavity resonators and waveguide filters . . . . . . . . 174.3 Tunable MEMS waveguide phase shifters . . . . . . . . . . . . . . . 18

5 MEMS-reconfigurable surfaces for millimeter-wave waveguidecomponents 215.1 Concept of MEMS-reconfigurable surfaces . . . . . . . . . . . . . . . 215.2 SOI RF MEMS fabrication platform . . . . . . . . . . . . . . . . . . 245.3 Waveguide integration and RF measurement setup . . . . . . . . . . 25

vi

Contents vii

5.4 Waveguide switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265.4.1 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265.4.2 RF characterization . . . . . . . . . . . . . . . . . . . . . . . 30

5.5 Switchable waveguide irises . . . . . . . . . . . . . . . . . . . . . . . 335.5.1 Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335.5.2 RF simulation and characterization . . . . . . . . . . . . . . . 34

5.6 Application examples of reconfigurable waveguide irises . . . . . . . 355.6.1 Switchable cavity resonator . . . . . . . . . . . . . . . . . . . 365.6.2 Tunable/switchable waveguide filters . . . . . . . . . . . . . 37

5.7 MEMS actuator characterization and reliability features . . . . . . . 385.8 Discussion and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . 39

6 MEMS-based reconfigurable antenna array for beam steeringfront-ends 416.1 Beam steering front-ends . . . . . . . . . . . . . . . . . . . . . . . . . 416.2 Novel MEMS-based waveguide-integrated phase shifters . . . . . . . 42

6.2.1 Concept and design . . . . . . . . . . . . . . . . . . . . . . . 426.2.2 Prototype fabrication . . . . . . . . . . . . . . . . . . . . . . 466.2.3 RF characterization . . . . . . . . . . . . . . . . . . . . . . . 47

6.3 Transmit-array antenna for beam steering front-ends . . . . . . . . . 486.4 Discussion and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . 50

7 Integration challenges of microwave MEMS devices into waveg-uides 517.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 517.2 Prototype integration concept . . . . . . . . . . . . . . . . . . . . . . 527.3 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557.4 Discussion and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . 58

8 Conclusions 59

Summary of Appended Papers 61

Acknowledgments 65

References 69

Paper Reprints 85

List of Publications

The thesis is based on the following peer-reviewed international journalpapers:

1. “MEMS reconfigurable millimeter-wave surface for V-band rectangular-waveguide switch,”Z. Baghchehsaraei, U. Shah, J. Åberg, G. Stemme and J. Oberhammer,International Journal of Microwave and Wireless Technologies, vol. 5, no. 3,pp. 341–349, June 2013.

2. “Integration of microwave MEMS devices into rectangular waveguide withconductive polymer interposers,”Z. Baghchehsaraei, M. Sterner, J. Åberg and J. Oberhammer, IOP Jour-nal of Micromechanics and Microengineering, vol. 23, no. 12, pp. 125020,November 2013.

3. “Waveguide-integrated MEMS-based phase shifter for phased array antenna,”Z. Baghchehsaraei, A. Vorobyov, J. Åberg, E. Fourn, R. Sauleau and J.Oberhammer, IET Microwaves, Antennas and Propagation, in print, acceptedon October 2013 (doi: 10.1049/iet-map.2013.0256).

4. “Parameter analysis of millimeter-wave waveguide switch based on a MEMS-reconfigurable surface,”Z. Baghchehsaraei and J. Oberhammer, IEEE Transactions on MicrowaveTheory and Techniques, vol. 61, no. 12, pp. 4396–4406, December 2013.

5. “MEMS-reconfigurable irises for millimeter-wave waveguide components,”Z. Baghchehsaraei and J. Oberhammer, submitted to IEEE Microwave andWireless Components Letters, 2014.

6. “MEMS-reconfigurable transmit-array antenna for 77-GHz automotive radarapplications,”A. Vorobyov, Z. Baghchehsaraei, E. Fourn, J. Oberhammer and R. Sauleau,submitted to IEEE Transactions on Antennas and Propagation, 2014.

ix

x List of Publications

The contribution of Zargham Baghchehsaraei to the different publica-tions:

1. major part of design, all fabrication, all experiments, and major part of writing

2. part of design, part of fabrication, part of experiments, and major part ofwriting

3. part of design, major part of fabrication, and major part of writing



6. part of fabrication and major part of writing

The work has also been presented at the following reviewed internationalconferences:

1. (*) “MEMS-reconfigurable waveguide iris for switchable V-band cavity res-onators,” Z. Baghchehsaraei and J. Oberhammer, IEEE International Con-ference on Micro Electro Mechanical Systems (MEMS), San Francisco, USA,January 2014.

2. “Monocrystalline-silicon microwave MEMS,” J. Oberhammer, U. Shah, Z.Baghchehsaraei, F. Töpfer, M. Sterner, N. Somjit and N. Chekurov, Procs.Progress in Electromagnetics Research Symposium (PIERS), Stockholm, Swe-den, August 2013.

3. (**) “Millimeter-wave SPST waveguide switch based on reconfigurable MEMSsurface,” Z. Baghchehsaraei, U. Shah, J. Åberg, G. Stemme and J. Ober-hammer, IEEE International Microwave Symposium (IMS), Seattle, USA,June 2013.

4. (***) “MEMS 30 µm-thick W-band waveguide switch,” Z. Baghchehsaraei,U. Shah, S. Dudorov, G. Stemme, J. Oberhammer and J. Åberg, IEEE Eu-ropean Microwave Conference (EuMC), Amsterdam, Netherlands, October2012.

5. “MEMS 30 µm-thick W-band waveguide switch,” Z. Baghchehsaraei, U.Shah, S. Dudorov, G. Stemme, J. Oberhammer and J. Åberg, IEEE EuropeanMicrowave Integrated Circuit Conference (EuMIC), Amsterdam, Netherlands,October 2012.

6. “TUMESA - MEMS tuneable metamaterials for smart wireless applications,”J. Ala-Laurinaho, D. Chicherin, Z. Du, C. Simovski, T. Zvolensky, A. V.Räisänen, M. Sterner, Z. Baghchehsaraei, U. Shah, S. Dudorov, J. Ober-hammer, A. V. Boriskin, L. Le Coq, E. Fourn, S. A. Muhammad, R. Sauleau,

xi

A. Vorobyov, F. Bodereau, G. El Haj Shhade, T. Labia, P. Mallejac, J. Aberg,M. Gustafsson and T. Scheir, IEEE European Microwave Conference (EuMC),Amsterdam, Netherlands, October 2012.

7. “Iris-based 2-bit waveguide phase shifters and transmit-array for automotiveradar applications,” A. Vorobyov, E. Fourn, R. Sauleau, Z. Baghchehsaraei,J. Oberhammer, D. Chicherin and A. Räisänen, IEEE European Conferenceon Antenna and Propagation (EUCAP), Prague, Czech Republic, March 2012.

8. “MEMS based waveguide phase shifter for phased array in automotive radarapplications,” A. Vorobyov, R. Sauleau, E. Fourn, J. Oberhammer and Z.Baghchehsaraei, IEEE European Conference on Antenna and Propagation(EUCAP), Rome, Italy, April 2011.

The work has also been presented at the following international work-shops:

9. “MEMS-reconfigurable millimeter-wave surfaces for waveguide switches, irisesand resonators,” Z. Baghchehsaraei and J. Oberhammer, GigaHertz Sym-posium, Gothenburg, Sweden, March 2014.

10. (****) “V-band single-pole-single-throw MEMS rectangular waveguide switch,”Z. Baghchehsaraei and J. Oberhammer, Micromechanics and MicrosystemsEurope (MME), Espoo, Finland, September 2013.

11. “Phased-array antenna based on MEMS fin-line phase shifters for W-bandbeam-steering applications,” Z. Baghchehsaraei, A. Vorobyov, R. Sauleau,E. Fourn and J. Oberhammer, GigaHertz Symposium, Stockholm, Sweden,March 2012.

12. “Microwave MEMS activities at KTH- Royal Institute of Technology,” J.Oberhammer, N. Somjit, M. Sterner, Z. Baghchehsaraei, U. Shah, F. Sa-haril, S. Braun and G. Stemme, Micronano System Workshop (MSW), Stock-holm, Sweden, May 2010.

13. “MEMS tunable metamaterials for beam steering millimeter wave applica-tions,” D. Chicherin, M. Sterner, Z. Baghchehsaraei, J. Oberhammer, S.Dudorov, Z. Du, T. Zvolensky, A. Vorobyov, M. de Miguel Gago, E. Fourn,R. Sauleau, T. Labia, G. El Haj Shhade, F. Bodereau, P. Mallejac, J. Åberg,C. Simovski and A. V. Räisänen, NATO Research Workshop on AdvancedMaterials and Technologies for Micro/Nano-devices, Sensors, and Actuators:From Fundamentals to Applications, St. Petersburg, Russia, June 2009.

xii List of Publications

Comments:

• (*) This conference paper received one out of four Outstanding Student PaperAwards at 2014 IEEE International Conference on Micro Electro MechanicalSystems, San Francisco, CA, USA.

• (**) This conference paper was finalist for Best Student Paper Award at 2013IEEE International Microwave Symposium, Seattle, WA, USA.

• (***) This conference paper was finalist for Best Paper Award at 2012 IEEEEuropean Microwave Conference, Amsterdam, The Netherlands.

• (****) This workshop presentation received a Best Poster Award at 2013Micromechanics and Microsystems Europe Workshop, Espoo, Finland.

List of abbreviations

3D Three dimensionalBOX Buried oxideCPW Coplanar waveguideDC Direct currentDLP Digital light processingDMTL Distributed MEMS transmission linesDRIE Deep reactive ion etchingFCR Fire-control radarHPBW Half power beamwidthHRSS High-resistivity silicon substrateMEMS Microelectromechanical systemsRF Radio frequencyRIE Reactive ion etchingRMS Root mean squareSEM Scanning electron microscopySIW Substrate integrated waveguideSOI Silicon on InsulatorSPDT Single pole double throwSPST Single pole single throwTRL Thru-reflect-lineTTD True-time delayVNA Vector network analyzerYIG Yttrium iron garnet

xiii

List of radar-frequency letter bands(IEEE standard 521-1984)

Letter Designation Frequency Range

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

xv

Chapter 1

Introduction and outline

This thesis presents novel works in the field of radio-frequency microelectrome-chanical systems (RF MEMS) for V-band (50–75 GHz) and W-band (75–110 GHz)reconfigurable waveguide components. The main motivation of doing research onMEMS for microwave applications emerges from excellent signal handling perfor-mance of microwave MEMS devices in terms of insertion loss, isolation, linearity, aswell as, large tuning range over a large bandwidth, besides, low power consumptionof MEMS actuators, e.g. for electrostatic actuation mechanism.

The millimeter-wave spectrum, historically mainly exploited for defense andspace applications, is of increasing interest for civilian applications, including auto-motive radar at 76–81 GHz [1], high speed wireless communications at 60 GHz [2, 3]and E-band (71–76 GHz, 81–86 GHz) [4], remote sensing for security and surveil-lance [5, 6], and medical diagnosis [7]. Advantages of going to higher frequenciesare larger available bandwidths, higher spatial resolution, better discrimination be-tween materials, and smaller antenna and sensor interfaces [8–10].

Over the past several decades, the need for versatile devices with the abilityto process and transfer high amount of data, on one hand, and the trend in ex-tending the application of the millimeter-wave spectrum, on the other hand, drivethe development of new technologies. Most of these technologies for microwavefrequencies are based on planar transmission lines, e.g. microstrip and coplanarwaveguides (CPW). Nevertheless, waveguide technology is still necessary at mi-crowave and millimeter-wave frequencies due to low signal loss, but also at radiofrequencies where high Q-factor or high power handling is required. In order tomake full advantage of waveguide technology, components inserted in the waveg-uide for signal manipulation should be of equally low loss. Although waveguidesystems can benefit significantly from the characteristics of RF MEMS, there hasbeen very limited research work on waveguide-integrated RF MEMS systems, wherenumerous specific challenges arise such as complexity of the reliable low-loss inte-gration of the MEMS chip into the waveguide and consideration requirement forthe interference of MEMS actuators, including the bias routing, with the RF signal.

1

2 Chapter 1. Introduction and outline

Novel concepts and fabricated components are presented in this thesis to addressthe limitations of the state-of-the-art waveguide switches, cavity resonators, filtersand phase shifters. In addition, a novel method for integration of MEMS chipsinto rectangular waveguides based on commercial conductive polymer interposer ispresented which features very low insertion loss.

The remainder of this thesis is divided into seven chapters and is organized asfollows.

Chapter 2 gives a brief introduction to microelectromechanical systems ingeneral and focuses, in particular, on RF and microwave MEMS sub-field. Thelast section of the chapter gives a short overview of RF MEMS devices based onsilicon-on-insulator (SOI) technology which is the main fabrication platform usedin Chapter 5.

Chapter 3 gives a background on conventional methods for tunable waveguidecomponents such as waveguide switches, tunable cavity resonators, and tunablewaveguide filters.

Chapter 4 covers the previous works of various research groups on tunablewaveguide components based on integrated MEMS devices.

Chapter 5 presents millimeter-wave waveguide switches and switchable cav-ity resonators based on novel MEMS-reconfigurable surfaces, and demonstrates aconcept for tunable waveguide bandpass filter.

Chapter 6 introduces a novel concept of tunable waveguide phase shifterswith integrated MEMS elements and application of such phase shifters for a two-dimensional antenna array for beam steering front-ends at 77 GHz.

Chapter 7 investigates a novel integration method based on commercial con-ductive polymer sheets to address the challenges for integration of MEMS devicesinto waveguides, such as increased insertion loss, spurious resonances, reliability ofmechanical mounting, and routing of the actuation voltage.

Chapter 8 summarizes the conclusions of this thesis.

Chapter 2

Background to RF and microwave MEMS

This chapter starts with a brief introduction to microelectromechanical systems,with focus on RF/microwave MEMS, and ends with RF MEMS devices based onsilicon-on-insulator (SOI) technology, which is the main fabrication platform usedin this thesis.

2.1 Microelectromechanical systems

Microelectromechanical systems (MEMS) are integrated micro-scaled devices ex-tending microelectronics with miniaturized mechanical structures and transducerswhich provide a link to the physical surroundings. Transducers, which are definedas components converting energy from one form to another, are categorized intosensors and actuators. Sensors are mostly used to measure a physical parameter bytaking energy from the surrounding and report in form of a transduced energy tothe device, whereas, actuators, which work the other way, are used to manipulatethe surroundings by converting an input energy from the device into a mechani-cal work output. Among many physical principles which can be used as actuationmechanism for MEMS actuators, electrostatic, thermal, piezoelectric, and magneticmechanism are employed most commonly. MEMS devices are fabricated usinglithography-based microfabrication techniques, which are heavily borrowed fromthe semiconductor industry and further enhanced with other micromachining tech-niques. Typical examples for MEMS-based sensors are accelerometers, gyroscopes,microphones, pressure sensors, and gas sensors; and examples for commerciallysuccessful MEMS actuators are inkjet printhead nozzles and micromirror arrays fordigital light processing (DLP) video projectors [11–16].

2.2 RF and microwave MEMS

MEMS devices are extremely diverse and they are divided into a number of cate-gories based on the application area, with devices for radio frequency (RF) applica-

3

4 Chapter 2. Background to RF and microwave MEMS

tions defined as RF MEMS. RF MEMS devices include, basically, micromachinedparts which are interacting with electrical signals of the RF range to provide RFfunctionalities such as switching, filtering, tuning, phase shifting, matching, and po-larization changing. The mechanical nature of their interaction with high frequencyelectrical signals offers numerous advantages over semiconductor counterparts suchas lower insertion loss, higher isolation, better linearity, higher Q-factors, widertuning range over a wider bandwidth, and extremely low power consumption [17–20].

The most basic RF MEMS devices are RF MEMS switches [17, 21–24] andtunable capacitors [22, 25–27], which have been developed to very mature RF com-ponents [28–31] since their introduction. The majority of more advanced recon-figurable RF MEMS systems typically feature conventional RF circuits which areenhanced by integrated MEMS actuators, mostly switches and varactors, to pro-vide new reconfigurable functionality [32]. Besides RF MEMS switches and tunablecapacitors, examples of RF MEMS devices include tunable inductors [33, 34], tun-able phase shifters [35–38], tunable filters [39–41], MEMS resonators for filters andreference oscillators [42], tunable impedance matching circuits [43–45], and recon-figurable antennas [46–48]. Furthermore, micromachined inductors [49–51], micro-machined transmission lines [52] and waveguides [53, 54], and microwave acousticdevices [55] are typically considered RF MEMS components since they involve mi-cromachining processes in their fabrication.

Among different possibilities for actuation of RF MEMS devices, electrostaticdrive is the most widely used actuation mechanism, due to its simplicity, com-pactness, good reliability, extremely low power consumption characteristics, andsuitability for high-volume wafer-scale manufacturing using standard semiconduc-tor processes and materials [56].

To clarify the RF MEMS device interacting with signals in the frequency rangeabove 30 GHz, the term microwave MEMS is used. The MEMS switch demon-strated in 1991 by Larson [57] for DC-to-45 GHz is considered the first microwaveMEMS device, and despite its immaturity, the outstanding performance in compar-ison to GaAs-based solid-state switches of the time, has been the main motivator forRF/microwave MEMS research. With the application frequency moving towardsmillimeter-wave, the advantage of RF MEMS in terms of signal performance, e.g.insertion loss, are becoming more prominent. Furthermore, increase in the applica-tion frequency leads to size reduction of the RF/microwave devices, thus allowingintegration of more compact systems.

Since the novel reconfigurable millimeter-wave waveguide components presentedin this thesis are based on MEMS switching elements, a general background on RFMEMS switches is given in the next part. Moreover, two relevant RF MEMS devicesto this thesis, i.e. tunable filters and phase shifters, are briefly discussed below forplanar circuits in the millimeter-wave frequency to cover a broader background tothe field, considering that the research on RF MEMS has been extensively focusedon transmission-line devices. An overview on waveguide counterparts is given inChapter 4.

2.2. RF and microwave MEMS 5

2.2.1 RF MEMS switches for millimeter-wave applicationsRF MEMS switches are, essentially, microfabricated miniaturized relays containingintegrated MEMS actuators which conduct the electrical signal in one state andbreaks the signal propagation in another state, typically to achieve open- or short-circuit. Most designs are only mechanically monostable in one state and are kept inthe other state either passively by a locking mechanism or actively by an externalforce, and they can transit between the two states either by restoring mechanicalforce or by imposing an external effort. Examples of microwave MEMS switchesdemonstrated for V-band and W-band are [58–64]. Moreover, in [65] a DMTLcomponent consisting of four capacitive swithces is presented which is consideredthe first demonstration of a MEMS component at frequencies above 110 GHz (aboveW-band).

RF MEMS switches can, mainly, be classified according to their actuation mech-anism, contact type, circuit configuration, and direction of displacement [17].

Actuation mechanism

Among widely used actuation mechanisms, electrostatic principle is usually pre-ferred due to the advantages such as fast switching time, large actuation force, lowpower consumption, and simple design, in expense of high voltage requirement fora reliable switch featuring high isolation.

Contact type

The contact between the movable part of the switch and the fixed part, can beeither ohmic metal-to-metal or capacitive with a dielectric layer between the twoparts. The metal-contact switch can potentially conduct signals all the way fromDC to 100 GHz, in contrast to capacitive switches where the low-frequency portionof the signal is blocked and only high-frequency application is feasible. To decreasethe capacitance coupling of both switch types in the open state, a large switchdisplacement is required.

Circuit configuration

The switch can be placed either in series with the transmission line, to allow thesignal to propagate in the transmission line through the switch contact in the closedstate, or as a shunt to ground, to short-circuit the signal to the ground and blockthe propagation in the transmission line in the closed state.

Direction of displacement

The direction of displacement of the conventional MEMS switch elements with re-spect to the substrate can either be out of plane of the substrate for the verticalswitching or in the plane of the substrate for lateral switching; much less frequently


switches employ torsional movements. Vertical switches are usually fabricated inthe shape of thin metal membranes or cantilevers by surface micromachining tech-niques, whereas, lateral switches are typically patterned in the wafer by bulk mi-cromachining techniques. In design of vertical switches there is excellent controlover the contact surface and isolation geometry defined mainly by photolithogra-phy. On the other hand, the movable structure of vertical MEMS switches usuallyinvolves deposition of multiple layers of different materials, for achieving the rightmechanical properties, resulting in residual stress and bending of the structure. Tocompensate for the bending and to achieve a robust switch with predictable actua-tion voltage, complex stress-compensation techniques during the fabrication [66] orlow-stress process steps [67] are required. Furthermore, thin moving structures ofvertical switches make them prone to plastic deformation and creep, especially atelevated temperatures, limiting their power handling capabilities. In comparison,lateral switches benefit from excellent mechanical properties of the bulk substratematerial, e.g. monocrystalline silicon [68], and involve simple photolithographysteps to define actuator geometry, but they suffer from a less controllable switchcontact surface, e.g. due to difficulty in metal deposition on the side walls or scal-loping of the structure from the Bosch [69] deep etching process step [70].

It is desirable to have a compact reliable MEMS switch featuring low RF inser-tion loss, high RF isolation, and low actuation voltage. However, there is a compro-mise between these features, for example, to achieve high isolation for electrostaticseries MEMS switches, the contact separation has to be increased, which leads toincrease in actuation voltage and decrease in contact force resulting in higher resis-tivity. To avoid the increase in actuation voltage, either a weaker switch membraneor a larger electrode area is necessary, with the former causing reliability issues byincreasing the chance of stiction as a result of weak restoring force and the latterresulting in larger switch size. To address this issues some ideas are proposed inthe literature such as curved electrodes [71], combination of vertical and lateralactuators [72], and push-pull concepts allowing active opening and active closingwith external force [73–75].

2.2.2 Tunable MEMS planar filters and phase shifters formillimeter-wave applications

Most planar tunable MEMS filters working at millimeter-wave frequency band, arebased on coplanar waveguide (CPW) technology and can be classified into twomajor groups. In the first group MEMS elements are used as localized varactorsbetween the transmission line and the ground plane, for example as shunt capacitorto transmission line [76], as variable LC in lumped-element filters [77, 78], as tuningelement of stubs to change their electrical length [77, 79], or as shunt bridges tochange inter-resonator coupling [80]. In the second group, the MEMS switches areintegrated only in the signal line for example to reconfigure the inter-connection ofperiodic structures [81]. Moreover, filter banks [82] are implemented by replacingsemiconductor switching elements with MEMS switches to improve RF loss.

2.3. Challenges of RF MEMS technology 7

There are many methods in use for tunable MEMS phase shifters for microwavefrequencies [17], but two widely-used concepts for the millimeter-wave portion of theelectromagnetic spectrum are MEMS-switched true-time delay-line (TTD) phaseshifter networks [59, 61, 83] and distributed MEMS transmission-line (DMTL)phase shifters [84, 85]. MEMS-switched TTD phase shifters consist of various phase-shifting sections in a cascade arrangement, where MEMS switches are employed indifferent positions to change the total length of the signal line resulting in differ-ent phase shifts. Although TTD phase shifters inherit all advantages of MEMSswitches, the total insertion loss degrades significantly at millimeter-wave frequen-cies due to the application of multiple switches and the necessary length of thetransmission line. An alternative method, DMTL phase shifting, that is typicallyused for broadband applications, is based on periodically loading the transmissionline with capacitive MEMS bridges to change the propagation coefficient affectingthe phase shift. Another novel method, presented recently, is to load the trans-mission line with multiple λ/2-long silicon blocks that can be moved vertically toachieve binary-coded phase shifter [86].

2.3 Challenges of RF MEMS technology

The compromise in using RF MEMS devices in comparison to their solid-statecounterparts is the reduced response speed [87, 88], e.g. switching time, highermanufacturing cost (including packaging), and usually design-specific process de-velopment, but the main challenges towards all-purpose mass-market application ofRF MEMS devices, over the last two decades, are, besides cost, long-term reliabilityissues [17, 89–92], which are mainly related to high-current handling, mechanicaldegradation of movable electrodes, and dielectric charging phenomena.

The major failure mechanism in metal-contact RF MEMS devices is the contactdegradation causing stiction in the contact area. Soft metals, such as gold, are moresusceptible to stiction [93], and different approaches are suggested in the literatureto overcome this problem, for example increasing the passive restoring force of thedevice at the cost of higher actuation voltages [94] and active opening concepts [95,96]. Additionally, there has been intensive research effort to replace gold with hardmaterial stacks, e.g. silver/tungsten/rhodium and copper/tungsten/gold, in orderto increase the reliability of contact performance even though hard metals requirehigher contact forces to compensate for the higher resistivity [91].

In electrostatic MEMS devices with dielectric isolation layer between the elec-trodes, dielectric charging due to charge injection or charge trapping, is the majorreliability issue affecting actuation behavior of the devices due to actuation voltagedrift [97]. There are numerous suggestions investigated in the literature to mitigatedielectric charging issues, e.g. using bipolar actuation voltage [98, 99], reducingactuation voltage to decrease the field strength in the dielectric [98, 99], employingdielectric materials with lower trap density and fast recovery time [100], and usingauxiliary (side) pull-down electrodes with stoppers [17].


(a)

(c)

(b)

(d)

Figure 2.1. Electrostatically actuated SOI-based lateral RF MEMS devices from the literature.(a) Switch [103]. (b) Tunable capacitor [104]. (c) Tunable directional coupler [105]. (d) Tunablefilter [106].

Besides, the power handling capability of conventional RF MEMS devices islimited due to reliability issues. For instance, thin all-metal RF MEMS devicesare prone to plastic deformation at high cycle numbers and prolonged down-stateposition due to creep, fatigue, and microwelding [91]. In the up-state position highRF power leads to self-actuation of the MEMS device [101, 102].

2.4 SOI-based RF MEMS devices

Monocrystalline silicon is an attractive material for MEMS, due to its excellent me-chanical properties, good thermal conductivity, compatibility with highly developedsemiconductor technology and micromachining tools. It is also available in largevariety of substrates in terms of size, crystal orientation, and doping levels [68, 91].

Silicon-on-insulator (SOI) wafers have long been used for fabrication of inte-grated circuits to suppress parasitic effects, such as leakage currents and latch-upeffects [107]. They facilitate fabrication of monocrystalline-silicon based MEMSdevices to be fabricated by bulk micromachining techniques, in addition to, bene-fiting RF MEMS device performance, for example by blocking DC leakage currentin 3D CPW structures [108]. The advantages of SOI micromachining over tradi-tional bulk micromachining include flexibility in choosing the initial thickness ofthe structure based on the thickness of the device layer, as well as, availability ofthe insulating and etch stop layer. Examples of lateral SOI RF MEMS devices,shown in Fig. 2.1, are switches [103, 109, 110], tunable capacitors [104], tunabledirectional couplers [105], and tunable filters [106].

Chapter 3

Background to conventional tunablewaveguide components

Waveguide technology is superior to transmission lines in terms of loss and powerhandling, specifically at extremely high frequencies.

The novel work presented in this thesis is primarily concerned with hollow metalwaveguides and the term waveguide is reserved for this type, however, examples ofdielectric waveguide and substrate integrated waveguide (SIW) components aregiven in this chapter and the next chapter, to cover a broader overview of tunablewaveguide components. This chapter gives a background to some conventionaltechnologies for tunable waveguide components. RF MEMS tunable waveguidecomponents are reviewed separately in Chapter 4.

3.1 Waveguide switch technologies

Switches are utilized in various microwave systems to control signal flow or to selectamong different signal sources. For example, in fire-control radars (FCR) switchesallow to share one power source for multiple antennas with different functionalityincluding identification and tracking, whereas in secure communication systemsswitches can be employed to switch among several power sources to one output.

Conventional electrical waveguide switches comprise mainly two concepts. First,in rotary motor-based mechanical switches a metallic block is utilized to break thewaveguide path [111], as shown in Fig. 3.1(a). This method has very low inser-tion loss and high isolation, but is heavy, bulky, very slow and requires high powerfor switching. Second, PIN-diode switched elements are integrated into waveg-uides [112], as shown in Fig. 3.1(b). This type switches in nanoseconds but haspoor insertion loss/isolation performance, needs relatively high biasing currents,and introduces distortion products.

Moreover, ferrite materials are used to fabricate waveguide switches, featur-ing faster switching speed, lower power consumption, and longer life time com-

9

10 Chapter 3. Background to conventional tunable waveguide components

P2

P1 P3

P4

P2

P1 P3

P4

(a)

Mechanically switched:

rotary motor switch

P2P1

(b)

Electronically switched:

PIN-diode switch

(c)

Magnetically switched:

ferrite switch

P2

P1

P3

Figure 3.1. Three main concepts for conventional waveguide switches. (a) Rotary waveguideswitch [111]. (b) PIN-diode waveguide switch [112]. (c) Ferrite waveguide switch [113].

pared to mechanical switches, in addition to lower distortion level and better signalperformance compared to PIN-diodes. Ferrite waveguide switches for single-poledouble-throw (SPDT) switch configuration are available in two main types, latch-ing junction circulator switch and differential phase shift switch, up to Ka band.In the junction circulator switch, shown in Fig. 3.1(c), the integrated ferrite res-onators are magnetized by applying a current pulse through a wire loop to controlthe direction of circulation allowing the signal flow from one port to either of theother two [113, 114]. In the differential phase shift switch, the input power is splitinto two identical signals by a waveguide tee and a relative phase shift of +90 ◦ or−90 ◦ is introduced between these two signals before they are recombined througha coupler guiding the signal to one or the other of two output ports.

3.2 Tunable cavity resonators and waveguide filters

Tunable high frequency components, including tunable filters, are some of the mostessential parts of modern multi-frequency systems as they may result in reducedcost, miniaturized system size, and decreased complexity in comparison to tradi-tional implementations, for instance switched filter banks. The same as waveguideswitches, most tunable waveguide filters can be categorized in three basic typesaccording to tuning mechanism, i.e. mechanically tunable, electronically tunable,and magnetically tunable filters [115].

Waveguide cavity filters are necessary at microwave and millimeter-wave fre-

3.2. Tunable cavities and waveguide filters 11

quencies, but also at radio frequencies where high Q-factor or high power handlingis required [116]. The resonant frequency of cavities can be varied by interferingwith the field within the cavity, typically by mechanical means, e.g. with ad-justable tuning screws or posts [117], shown in Fig. 3.2(a), as the earliest and themost popular method, generally used for post-production manual tuning, however,electromechanical tuning is possible by using motors to drive the tuning screws.This method features very good power handling capability and low insertion loss atlow cost, but it increases the size and offers low tuning speed. Tuning by insertinga lumped susceptance in the cavity without modifying the coupling between thecavities imposes a limitation over the possible tuning range because the pass-bandresponse becomes more and more asymmetric as the cavities are tuned more andmore away from the reference resonance frequency, thus for wide-band tuning range,inter-resonator couplings should also be tuned.

For mechanical tuning of cavity filters piezoelectric actuators are also used,instead of stepper motors. For example, in [118, 119], shown in Fig. 3.2(b), atunable evanescent-mode cavity filter over 3–5.6 GHz range is presented basedon a substrate integrated waveguide (SIW) technique. The filter consists of twocavity resonators formed by conductive vias and connected by coupling irises. Eachresonator has a capacitive post in the center formed by four vias and a cylindricalcopper plate. A round metallized piezoelectric actuator is mounted on top of thesubstrate with the integrated cavities for tuning functionality. Unloaded qualityfactors (Qu) between 250 and 650 have been demonstrated over an octave tuningrange.

For electronically tuning mechanism, varactor diodes are employed to realizetunable waveguide filters offering very fast tuning speed over a wide tuning range,for instance a two-pole finline filter is presented in [120] for the Ka-band. Thesefilters are tuned by adjusting the DC voltage supplied to reverse-biased varactordiodes to change their capacitance, and thus affecting the resonance frequency of theresonator connected to the varactor diode. The main drawbacks of varactor-tunedwaveguide filters, similar to PIN-diode waveguide switches, are their relatively largeinsertion loss and low power handling capability, which are mainly attributed tothe low quality factor and nonlinearity characteristics of varactor diodes.

Furthermore, magnetically tunable concepts have been used for tunable waveg-uide filters for many years by employing ferrimagnetic and ferromagnetic materialsfor applications requiring ultra-wideband tuning range [121, 122]. The most popu-lar tunable microwave filters, introduced in 1958 [123], use ferrimagnetic resonators,typically yttrium iron garnet (YIG), and gyromagnetic coupling. The YIG filter, of-ten used up to Ku-band, features multi-octave tuning range, spurious free response,and very high selectivity. The drawbacks of YIG filters is their moderate tuningspeed in the order of gigahertz per milliseconds and low power handling capabilitylimited by excitation of spurious resonances at high powers, and also relatively largepower consumption [115, 124]. The principles to design multi-stage filters with YIG-tuned spheres from 0.5 to 40 GHz are described in [124, 125]. A tunable waveguidefilter using YIG films with insertion loss of 1 dB in the range 33–50 GHz is demon-


(a)

Manual tuning

of a cavity filter

(b)

Tuning of a SIW filter

with piezoelectric actuators

Figure 3.2. Examples of mechanical tuning of waveguide filters. (a) Tunable cavity filter withindividual tuning knobs from Ducommun Technologies. (b) Two-coupled substrate integratedevanescent-mode cavity filter tuned by piezoelectric actuators [119].

strated in [126], however, it is preferred to employ highly anisotropic hexagonalferrites instead of YIG spheres to reduce the value of the external magnetic field,and therefore, to be able to tune to higher frequencies in millimeter-wave spectrum.For example, a four-pole filter is reported in [127] with an insertion loss of about6 dB over 50–70 GHz tuning range. Moreover, ferrite-loaded evanescent waveguidefilters and various configurations of printed E-plane filters are reported to improvethe limitations of YIG filters such as low power handling capability [115]. In [128],another configuration for magnetically tunable waveguide filters is proposed by re-placing the sidewalls of the iris cavity with electromagnetic crystals and tunabilityin 31.6–33.2 GHz range with insertion loss better than 4 dB is demonstrated.

3.3 Variable phase shifters

The most important application of tunable phase shifters, driving their develop-ment, is in phased-array/beam-steering antennas for example for radar. Majorconventional concepts for phase shifting waveguide devices, the same as previouslydiscussed waveguide components, include mechanical tuning, integration of PIN-diodes into waveguides, and ferrite-based magnetic tuning.

A variable phase change is obtained in dielectric-slab phase shifters by mechan-ically displacing the dielectric structure across the waveguide either manually, byplastic screws, or electronically, by step motors, with an example of such a conceptshown in [129] for the Ku-band. Another method for mechanical phase shiftingin waveguides is based on a rotating half-wavelength section between two fixedquarter-wave sections made of circular waveguide iris polarizers. This concept wasdescribed many years ago in [130], was discussed theoretically and experimentally

3.3. Variable phase shifters 13

in [131], and has been improved for operation over 75–100 GHz range in a morerecent publication [132], shown in Fig. 3.3(a).

Semiconductor active devices, e.g. PIN-diodes, can be integrated into waveg-uides to obtain electronically controlled phase shifters. For example, it is shownin [133] that by loading the sub-resonant slots of a waveguide wall with PIN-diodes,an electronically controllable phase shifter at Ka-band can be obtained. The slotsare designed sub-resonant to avoid any significant radiation or high reflections. Thestructure of this phase shifter is shown in Fig. 3.3(b).

Historically, ferrite phase shifters have been the dominant technology for waveg-uide phase shifting. They are divided generally into four types, i.e. variable per-meability, toroidal, dual mode, and rotary field. Besides, ferrite phase shifters canbe divided into reciprocal or nonreciprocal depending on dependency of the vari-able differential phase shifts to the direction of the propagation, or they can becategorized as latching or non-latching depending on the requirement for supplyingconstant current to sustain the magnetic bias field. Comprehensive and chronolog-ical reviews on ferrite materials and their application for phase shifters, in additionto different concepts for implementation of ferrite phase shifters, can be found inthe literature [134–138]. Some examples of millimeter-wave ferrite waveguide phaseshifters from the literature and some available products should be given here.

The application frequency range of latching nonreciprocal toroid/twin-toroidand latching reciprocal dual-mode phase shifters was demonstrated up to millimeter-wave range [134]. In these phase shifters the phase change is affected by modifyingthe propagation constant of the transmission line, which contains the ferrite, bycontrolling the relative permeability of the ferrite through the flux level existing inthe closed magnetic circuit. An example of ferrite-based dual-mode phase shifter ispresented in [139] with insertion loss of 1.6 dB at 60 GHz for the latching phase shiftbelow 363 ◦ consuming switching power of 4.5 mW. Moreover, commercial and mil-itary millimeter-wave ferrite waveguide phase shifters are available from numerousproducers. For example dual-mode phase shifters from MAG (Microwave Appli-cations Group), have insertion loss of maximum 1.25 dB, maximum return loss of15.56 dB, RMS phase error of 4 degrees and switching time of 60 µs at Ka-band.Ferrite Domen Co. supplies different toroidal and dual-mode phase shifters up to40 GHz with nominal insertion loss in 1–2.2 dB range, phase shifts from 0–337.5degrees up to 0–500 degrees, and switching time below 100 µs depending on theproduct model. A high power toroidal ferrite phase shifter for Ka-band is avail-able from DEV COM, shown in Fig. 3.3(c), with insertion loss of less than 1 dB,phse shift range of −360 ◦, switching speed of faster than 15 µs and average powerhandling of 120 W. Besides, mechanically-tunable rotary vane phase changer fromFLANN MICROWAVE has insertion loss of 2.5 dB and handles maximum powerof 0.2 W at W-band.

Other technologies have also been utilized to improve the characteristics ofmillimeter-wave phase shifters in the recent decade. In [140], a waveguide phaseshifter using tunable electromagnetic crystal sidewalls was reported with a phaseshift of 360 ◦ and an insertion loss of <2 dB at 38 GHz. A magnetic field tun-


(a)

Mechanically tunable

phase shifter

(b)

Electronically tunable

phase shifter

(c)

Magnetically tunable

phase shifter

Figure 3.3. Examples of conventional tunable waveguide phase shifters. (a) Mechanically tunablerotary phase shifter for 94 GHz [132]. (b) PIN-diode tunable waveguide phase shifter for Ka-band [133]. (c) Magnetically tunable toroidal phase shifter from DEV COM for Ka-band.

able dielectric phase shifter with maximum phase shift of 60 ◦ at a magnetic fieldstrength of 255 000 A/m and an insertion loss of better than 4 dB at 80 GHz wasdemonstrated in [141].

Chapter 4

Previous work on tunable MEMS waveguidecomponents

RF MEMS devices are promising candidates for integration into waveguides dueto excellent signal handling performance, thus allowing to make full advantage ofwaveguide technology.

This chapter gives an overview of some tunable microwave and millimeter-wavewaveguide components based on waveguide-integrated or waveguide-mounted MEMSactuators.

4.1 MEMS waveguide switches

Potentially, RF MEMS technology for waveguide switches could combine the ad-vantages of two main conventional methods discussed in Chapter 3, i.e. an elec-tromechanical rotary switch and a PIN-diode switch, as summarized in Table 4.1.

To the best knowledge of the author, there have only been two attempts byanother research group to build a MEMS waveguide switch; one was implementedwith electrothermal actuators, with insertion loss of 1–2.8 dB and return loss of15 dB in the Ku and K-bands [142], and another one with electrostatic actuators,with an insertion loss of 0.9–1.2 dB and return loss of 20 dB over 7% of the Ku [143].

15

16 Chapter 4. Tunable MEMS waveguide components

Table

4.1.C

omparison

ofdifferenttechnologies

form

illimeter-w

avew

aveguidesw

itches.

Rotary

motor

PIN

diodeM

EM

S

Insertionloss

<0.3–0.8

dB<

2.0–6.5dB

As

goodas

rotaryswitches

dueto

mechanicalswitching

natureIsolation

>40

dB>

20dB

As

goodas

rotaryswitches

Speed≈

100m

s(Slow

)ns

(Fast)µs

(Medium

speed)P

ower

consumption

High

Medium

Verylow

(e.g.electrostatic

actuation)V

olume

Bulky

Com

pactC

ompact

Weight

Heavy

LightLight

Reliability

Medium

High

Requiresa

reliabledesign

concept[90–92]

4.2. Tunable MEMS cavity resonators and waveguide filters 17

4.2 Tunable MEMS cavity resonators and waveguide filters

Although MEMS technology is mainly utilized in planar circuits, there are numerousexamples of MEMS cavity resonators and waveguide filters presented in the litera-ture, however, they are demonstrated mainly for applications below millimeter-waverange. To show the advantages and drawbacks of MEMS technology, Table 4.2 sum-marizes the performance of mechanical, YIG, and RF MEMS tunable waveguidefilters.

Fig. 4.1 provides examples of state-of-the-art tunable MEMS waveguide cavityresonators from the literature [88, 144–147].

To realize tunable MEMS waveguide or cavity filters, cavity resonators of thetop two rows of Fig. 4.1 were demonstrated, in the same papers where they wereintroduced, to implement two-pole filters, with 1.3–2.4 dB insertion loss and tuningrange of 2.71–4.03 GHz in [144], and with 3.55–2.38 dB insertion loss and tuningrange of 3.04–4.71 GHz in [146]. Furthermore, a two-pole evanescent-mode cavityfilter, with electrostatic cantilever tuning elements, is presented in [145] having 4.91–3.18 dB insertion loss for tuning range of 4.07–5.58 GHz with actuation voltage of0 to 65 V.

To the knowledge of the author, to date, there is no millimeter-wave tunableMEMS waveguide filter or cavity resonator demonstrated, even though, some at-tempts are reported in the literature up to K-band. For example, the feasibility ofconstructing a two-pole micromachined tunable ridge waveguide filter with embed-ded thermal MEMS actuators is demonstrated up to 24 GHz with 200 MHz tuningrange and insertion loss of 3 dB in [148]. In another attempt, an all-silicon cavitytunable resonator with capacitive post, demonstrated in [149], is tuned from 6.1to 24.4 GHz with measured unloaded Q from 300 to 1000, but with the requiredactuation voltage as high as 600 V.

Table 4.2. Comparison of different technologies for tunable waveguide filters, adapted form [115].

Parameter Mech. YIG MEMS

Insertion loss (dB) 0.4–2.5 3–8 0.8–6Power handling (W) High (<500) Low (< 2) Low (< 2)Tuning speed (GHz/ms) Very slow 0.5–2 102

Tuning range High High LowMillimeter-wave capability up to Ka-band Yes Yes

18 Chapter 4. Tunable MEMS waveguide components

A Tunable Miniaturized

RF MEMS Resonator With

Simultaneous High Q

(500–735) and Fast

Response Speed (< 10 − 60 μs),

J. Small et al.,

2013

Performance

Highly Loaded Evanescent

Cavities for Widely Tunable

High-Q Filters,

H. Joshi et al,

2007

High-Q Tunable Microwave

Cavity Resonators and

Filters Using SOI-Based

RF MEMS Tuners,

X. Liu et al.,

2010

Ku Band High-Q

Tunable Surface-Mounted

Cavity Resonator Using

RF MEMS Varactors,

R. Stefanni et al.,

2011

Tuning range: 2.3–4.6 GHz

Unloaded Q: 360–702

Actuation voltage: -10 to 90 V

Piezoelectricdisk

Actuator

Tuning range: 1.9–5 GHz


Actuation voltage: 0 to 122 V







Electrostaticdiaphragm

Electrostaticfixed-fixed

beam

Electrostaticcurled

cantilvers

Figure 4.1. Examples of reconfigurable MEMS cavity resonators reported in the literature [88,144–147].

4.3 Tunable MEMS waveguide phase shifters

A comparative summary is presented in Table 4.3 to highlight the advantages ofMEMS technology over ferrite for tunable waveguide phase shifter.

Similar to MEMS waveguide switches, there has not been much work on tun-able MEMS waveguide phase shifters. A single ridge transmission type phase shifterwith waveguide-integrated MEMS actuators presented in [150] has achieved a phaseshift of 70 ◦ and an insertion loss of 3.6 dB at 98.4 GHz, and a triple ridge trans-mission type phase shifter has achieved a phase shift of 134 ◦ and an insertion lossof 2.4 dB at 92.8 GHz. In [151] an analog-type phase shifter for W-band has been

4.3. Tunable MEMS waveguide phase shifters 19

Table 4.3. Comparison between ferrite and MEMS tunable waveguide filters, adapted from [138].

Parameter Ferrite MEMS

Cost Very expensive LowPower consumption ≈ 10 W (≈ 1 W latching) NegligiblePower handling Very high (1 kW) Typically < 50 mWSwitching speed µs (medium, inductance) µs (medium, mechanical)Microwave loss < 1 dB/360 ◦ (X band) ≈ 2.3 dB/337.5 ◦ (K band)Size Large SmallReliability Excellent Good after billion cycles

demonstrated consisting of a tunable MEMS high-impedance surface placed as aback-short or as a sidewall of a rectangular waveguide with a phase shift span of240 ◦ and an insertion loss of maximum 3.5 dB at 83.4 GHz.

Chapter 5

MEMS-reconfigurable surfaces formillimeter-wave waveguide components

This chapter introduces the concept of novel MEMS-reconfigurable surfaces forwaveguide-integrated switch and switchable waveguide irises. Then, as applica-tion examples of switchable MEMS waveguide irises, an implementation of the firstswitchable waveguide cavity resonator and a demonstration of a switchable waveg-uide filter concept are presented. Further details and discussion of the materialreviewed in this chapter are provided in Paper 1, Paper 4, and Paper 5.

5.1 Concept of MEMS-reconfigurable surfaces

The MEMS waveguide components presented in this chapter are based on MEMS-reconfigurable surfaces that are inserted perpendicular to the wave propagation intoa WR-12 rectangular waveguide (with inner dimensions 3.099 mm × 1.549 mm andfrequency range 60–90 GHz). These surfaces consist of distributed metallized ele-ments, which can be reconfigured to total transmissive state or the blocking statefor the wave propagation, by on-chip MEMS actuators outside the waveguide cross-section. A simplified 3D schematic of the proposed concept is shown in Fig. 5.1,designed as a waveguide switch component. The distributed metallized elements ofthe reconfigurable surface, designed in shape of vertical cantilevers, are divided intoa set of fixed and a set of movable elements via horizontal suspension bars. Horizon-tal suspension bars of the fixed set are anchored at the narrow waveguide sidewallsin the waveguide cross-section, whereas, of the movable set are connected to electro-static comb-drive push-pull MEMS actuators, outside the waveguide narrow walls,providing synchronous lateral movement. In the surface-deactivated state, shownin Fig. 5.1(b), there is a gap between the vertical cantilevers of the fixed and themoving elements and, thus, electromagnetic wave can freely propagate throughthe transmissive surface. In the surface-activated state, shown in Fig. 5.1(c), themovable vertical cantilevers are laterally moved into contact with the non-movable

21

22 Chapter 5. MEMS-reconfigurable mm-wave waveguide components

rectangularwaveguide on-chip MEMS

actuators

on-chip MEMSactuators

fixedhorizontal bars

movinghorizontal bars

TE10E

(a) Schematic view

(b) Transmissive state (b) Blocking state

Figure 5.1. Concept of MEMS-reconfigurable surface, depicted for a structure with uniformly dis-tributed elements, with 3 moving and 2 fixed horizontal suspension bars and 7 cantilever columns,applicable as a waveguide switch component. (a) 3D schematic view. (b) Surface-deactivated(transmissive) state, and (c) Surface-activated (blocking) state.

vertical cantilevers, to form vertical columns which are blocking the wave prop-agation as they are short-circuiting the electric field lines of the dominant TE10mode.

The MEMS-reconfigurable surfaces utilized in this thesis have the followingdimensions. The width (horizontal) and thickness (in direction of the wave prop-agation) of the vertical cantilevers are 5 µm and 30 µm, respectively, and thusprovide minimum insertion loss in the transmissive state and very high isolation inthe blocking state. In the surface-activated state, the moving and the fixed verticalcantilevers are in contact with 5 µm overlap, whereas, in the surface-deactivated

5.1. Concept of MEMS-reconfigurable surfaces 23

(a) overview of a sample chip for MEMS waveguide switch application

(b) Close-up view of a chip

verticalcontactcantilevers

anchorof fixedhorizontalbar

reconfigurable

surface

folded beam

springs

comb-drive

actuator

Figure 5.2. SEM images of a MEMS-reconfigurable surface for waveguide switch application. (a)Overview of a sample chip consisting of a reconfigurable surface, one stage of comb-drive actuatorwith folded beam springs on each side. (b) Close-up view of different elements of the reconfigurablesurface including vertical cantilevers, as well as, two sets of movable and fixed horizontal bars.


(d) Underetching the BOX (e) Metallization

TiW/Au SiO2Si

30 µm

2 µm

450 µm

(a) SOI wafer (b) DRIE of the handle wafer (c) DRIE of the device layer

Figure 5.3. The SOI RF MEMS process flow used to fabricate the chips containing MEMS-reconfigurable surfaces. (a) A standard silicon-on-insulator (SOI) wafer. (b) Anisotropic deepreactive ion etching (DRIE) of the handle wafer. (c) DRIE of the device layer. (d) Wet etchingof the buried oxide (BOX) layer in hydrofluoric acid to release the moving parts. (e) Sputteringof gold on the top and bottom sides of the chip.

state, they are 24 µm apart.Fig. 5.2 shows SEM images of an overview and close-up view of a MEMS-

reconfigurable surface designed for switch application.

5.2 SOI RF MEMS fabrication platform

The two-mask SOI MEMS process flow outline in Fig. 5.3 are used to fabricate thechips for MEMS-reconfigurable waveguide components presented in this chapter. Asilicon-on-insulator (SOI) wafer with resistivity of 1–20 W-cm is used as substrate[see Fig. 5.3(a)], which consists of a 30-µm thick device layer, 2-µm thick buried-oxide (BOX) layer, and 450-µm handle wafer. The fabrication process starts withthermal oxidation of the wafer and patterning of the oxide for the deep-siliconetching hard mask, followed by anisotropic deep reactive ion etching (DRIE) of thehandle wafer [see Fig. 5.3(b)] and subsequently of the device layer [see Fig. 5.3(c)]using Bosch process [152]. Then wet etching in hydrofluoric acid (HF) is doneto remove the sacrificial BOX layer under the moving structures [see Fig. 5.3(d)]followed by critical point drying to avoid any stiction of the released structures.A 1-µm thick layer of gold is sputtered on both sides of the wafer using 50 nm oftitanium tungsten as adhesion layer [see Fig. 5.3(e)] to completely metallize boththe switching elements and the walls of the handle wafer. Since the entire chip,including the top and the bottom faces and the sidewalls, is completely coveredin this last step, SOI wafer of any resistivity can be used for the fabrication ofMEMS-reconfigurable devices and there is no need for expensive high-resistivitySOI wafers which are common in RF MEMS devices. Fig. 5.4 shows a cross-sectional 3D illustration of the chip with a MEMS-reconfigurable surface and an

5.3. Waveguide integration and RF measurement setup 25

(a) Schematic cross-section view of a chip (b) An ion-milled vertical cantilver

Si

Au

Figure 5.4. Cross-sectional views of the chip containing a MEMS-reconfigurable surface anda vertical cantilever. (a) 3D schematic cross-sectional view of the final chip. (b) Wafer cut of avertical cantilever, by ion-milling, showing the metal coverage on the sidewalls of the reconfigurablesurface elements.

TiW/Au SiO2Si Brass Conductive polymer

conductive

polymer

sheets

chip

tailor-made flanges

recesses for chip

self-alignment

Figure 5.5. 3D exploded view of different components used to securely integrate the MEMSchip containing the reconfigurable surface into a WR-12 waveguide. The main components aretwo tailor-made flanges with recesses for chip self-alignment and two pieces of conductive polymerinterposers, besides, the MEMS chip.

SEM image of the cross section of an ion-milled vertical cantilever element showingthe metal coverage on the sidewalls.

5.3 Waveguide integration and RF measurement setup

Fig. 5.5 shows an exploded 3D rendered schematic of the chip-to-waveguide assem-bly, including the details of the tailor-made WR-12 flanges with recesses for chip


self-alignment to the waveguide. The figure also shows conductive polymer inter-posers applied only between the wide walls of the waveguide and the device layer ofthe chip, leaving a gap between the surface of the narrower wall of the waveguideand the chip, which allows for the mechanical connection between the reconfig-urable elements in the waveguide cross-section and the on-chip actuators outsidethe waveguide walls. The details of this integration method and a discussion areprovided in Chapter 7 and Paper 2.

The RF measurements of the fabricated chips integrated into the WR-12 waveg-uide were performed using an Agilent E8361A vector network analyzer (VNA) with110 GHz millimeter-wave test heads, calibrated using thru-reflect-line (TRL) cali-bration and a V-band waveguide calibration kit.

5.4 Waveguide switches

This section discusses the parameters involved in the design of a waveguide switch,which can be switched ON or OFF by a MEMS-reconfigurable surface, and showsmeasurement results of such a device.

5.4.1 DesignAs mentioned in the previous section, in practice, conductive polymer interposersare applied only between the wide walls of the waveguide and the top surface ofthe fabricated chips. Therefore, it is of interest to compare two different assemblycases, with the interposer layer covering all walls of the waveguides or only the widewalls, before discussing the waveguide switch design. Since the wide waveguide wallsare primarily carrying the surface currents, it is expected to observe insignificantdifference on the performance between these two cases for thin conductive polymerinterposers. Fig. 5.6 shows the simulation results of an assembly consisting ofconductive polymer layer between two WR-12 waveguide sections. The simulationis done for two above cases with the polymer thickness as a parameter. For polymerlayer thickness of 120 µm, which is equal to the nominal thickness of the polymersheet used in this chapter, the simulated insertion loss for both cases is below 0.02dB, i.e. negligible.

The waveguide-switch performance in the blocking and in the non-blocking stateis determined by the number of vertical columns and the number of sections of thecolumns, i.e. the number of horizontal suspension bars with associated verticalcantilevers, and on the distribution of these elements in the cross-section of thewaveguide. First, we study the influence of vertical cantilever columns and horizon-tal suspension bars on the performance for a reconfigurable surface with uniformdistribution of vertical cantilever elements along the wide wall of the waveguide.Then, we study the optimization for decreasing the number of contact points basedon non-uniform distribution.

The switch design study, presented here, is based on simulations of an ANSYSHFSS model shown in Fig. 5.7(a). The reconfigurable surface elements are modeled

5.4. Waveguide switches 27

Frequency (GHz)

0.1 mm

0.5 mm

1.0 mm

1.5 mm

solid line:

polymer covering

all walls

dashed line:

polymer covering

only wide walls

S (d

B)

21

0

−0.1

−0.2

−0.3

−0.4

60 65 70 75 80 85 90

Figure 5.6. Simulated comparison of interface performance between two WR-12 waveguideswith conductive polymer interposer with polymer thickness as parameter for two cases: interposerlayer covering all walls of the waveguides or only the wide walls.

as 30-µm thick gold blocks with a resistivity of 2.049 µW-cm taking into accountall metal losses, as compared to the elements of the fabricated devices consistingof silicon blocks with 1-µm gold coverage. The simulation model includes all ver-tical cantilever elements and all horizontal suspension bars of the same geometryas in the final mask layout. For high-accuracy simulation results, the waveguidefeedthroughs are also considered in the model, even though the simulation resultsof Fig. 5.6 confirmed that an opening in the narrow walls of the waveguide in theorder of polymer thickness used here, has negligible effect on the assembly loss.Fig. 5.7(b) and (c) show the HFSS-simulated RF performance of the waveguideswitch for the uniform distribution of vertical cantilever elements, with the num-ber of horizontal suspension bars and vertical cantilever columns as parameters, at75 GHz (E-band center frequency). Vxx marks on the figure indicates the numberof vertical cantilever columns.

Vertical cantilever columns are parallel to the electric field lines, therefore, in-creasing their number improves the blocking of the wave propagation in the OFFstate resulting in better isolation, however, it impacts negatively the ON-state in-sertion loss. On the other hand, being perpendicular to the electric field linesinside the waveguide, horizontal suspension bars have a very weak effect on the RFperformance of the switch in the OFF state. However, increasing the number ofhorizontal bars results in shorter vertical cantilevers and thereby a higher degree ofsplitting of the sections which are intended to short-circuit the electric filed linesin the OFF (blocking) state. Consequently, less fraction of the wave is reflected in

28 Chapter 5. MEMS-reconfigurable mm-wave waveguide componentsS

(d

B)

11

9 13 17 21 25−50

−40

−30

−20

−10

0V20 V30,35

V15V10

V10V15

V20

V30V35

number of horizontal bars

−0.25

−0.20

−0.15

−0.10

−0.05

0

S (d

B)

21

−30

−25

−20

−15

−10

−5

0 0

−0.5

−0.4

−0.3

−0.2

−0.1

−0.30

−0.35

Port 1 (Input)

Waveguide walls (copper)

Switch structure (gold)

2.03 mm

mechanical

feedthrough

(a) HFSS model

(b) Switch OFF state (blocking)

9 13 17 21 25

number of horizontal bars

(c) Switch NO state (non-blocking)

V20

V10

V30

V15

V35

V20

V10

V30

V15

V35

Figure 5.7. Simulated study of different MEMS-reconfigurable waveguide switches changing thenumber of horizontal suspension bars and vertical cantilever columns. (a) HFSS model shownfor the switch with 21 horizontal bars and 35 vertical columns in the OFF state (blocking). (b)OFF-state (blocking) simulation results. (c) ON-state (non-blocking) simulation results.

the ON (non-blocking) state resulting in lower insertion loss due to longer path ofthe electric field lines. As a design guide, if a larger number of vertical columnsis desired to increase the isolation of the switch, the degradation in insertion losscan be compensated by also increasing the number of horizontal suspension bars,however, beyond a certain number, increasing the number of horizontal suspen-sion bars does not significantly improve the insertion loss (saturation effect), forinstance beyond 17 for 20 vertical columns, thus a compromise has to be made. Acombination of 20 vertical columns with 21 horizontal suspension bars was foundto achieve an optimum compromise in performance between the ON and the OFF


S (d

B)

21

−25

−20

−15

−10

−5

0

−0.05

−0.04

−0.03

−0.02

−0.01

0

S (d

B)

11

60 70 80 90

Frequency (GHz)

−0.06

−0.05

−0.04

−0.03

−0.02

−0.01

0

−35

−30

−25

−20

−15

−10

−5

0

(b) Switch OFF state (blocking)

60 70 80 90

Frequency (GHz)

(c) Switch ON state (non-blocking)

12

354

35

4 2 1

35

24

1

12

354

Structure 2:

16 columns at density of

20 columns, 4 edge columns

removed

Structure 3:

16 columns distributed

uniformly

Structure 4:

16 columns, weak

sinusoidal distribution

Structure 5:

14 columns, strong

sinusoidal distribution

Structure 1:

20 columns distributed

uniformly

(a) Illustration of the 5 studied distributions

Figure 5.8. Investigation of influence of the distribution of cantilever columns conducted for5 structures in HFSS. (a) OFF-state simulation results. (b) ON-state simulation results. (c)Illustration of the 5 studied distributions with the red line indicating the distribution of thecolumn density.

state, resulting in an ON-state insertion loss and return loss better than 0.035 dBand 22.55 dB, respectively, and OFF-state isolation and reflected power better than28.49 dB and 0.023 dB, respectively, for a total of 840 vertical cantilevers and 440contact points. In conclusion, the blocking quality of the switch is determined bythe number of vertical columns and weakly depending on the number of horizontalsuspension bars, whereas the transmissive properties of the switch are dependenton the ratio of the horizontal bars to the vertical columns.

The density of the electric field lines of the dominant TE10 mode has a sinusoidaldistribution along the axis parallel to the wider walls of the rectangular waveguide


Figure 5.9. Measurement setup with the chip integrated into the tailor-made flanges.

with a maximum at the center [153]. Thus, it is expected to gain an improvedOFF-state isolation by sinusoidally distributing the same number of columns asuniformly distributed design, or, to have similar OFF-state isolation at a reducednumber of vertical columns and thus improved ON-state performance.

A comparative investigation is performed for five structures, by simulation inHFSS, while varying number and distribution of vertical columns to confirm theabove expectation. The other parameters of these five structures are identical, inparticular having 17 horizontal bars. Drawings of the distributions of the cantilevercolumns of these five structures are shown in Fig. 5.8(a) and the simulation resultsare presented in Fig. 5.8(b) and 5.8(c). Comparing Structure 1 with Structure 2,shows that removing a total four cantilever columns and thus decreasing the totalnumber of columns by 20%, has negligible effect on the performance. Structure 5with 14 sinusoidally distributed cantilevers (30% decrease in the number of contactcantilevers as compared to the structure with 20 cantilever columns) and highestcolumn density in the center shows a better performance than Structure 4 withuniform distribution of 16 cantilevers. As a conclusion, as expected, the importantfactor is the column density in the center where the intensity of the electric field isvery strong, and a non-uniform distribution can substantially decrease the numberof contact points with just slight decrease in the performance.

Further elaboration on the design steps and parameters of the waveguide switch,consisting of a MEMS-reconfigurable surface, is given in Paper 1 and Paper 4.

5.4.2 RF characterizationA photograph of the RF measurement setup of the fabricated chips with the re-configurable surface integrated into a WR-12 waveguide is shown in Fig. 5.9. Thecalibration is performed at the coaxial-to-waveguide adapters. Therefore, the mea-surements show the performance of the integrated chips combined with the tailor-made assembly setup, including the modified flanges and the conductive polymerinterposers.


E

(a)

S (d

B)

21

S (d

B)

11

S (d

B)

11

S (d

B)

21

Frequency (GHz)

60 62 64 66 68 70−50

−40

−30

−20

−10

0

−40

−30

−20

−10

0

−2.4

−2.1

−1.8

−1.5

−1.2

−0.9

−0.6

−0.3

0

60 62 64 66 68 70

Frequency (GHz)

V20H09/A V20H17/B V20H21/C

V30H17/D V30H21/E V35H21/F

V0H0 (reference)/R

Switch OFF state (blocking)

Switch ON state (non-blocking)

−5.0

−4.5

−4.0

−3.5

−3.0

−2.5

−2.0

−1.5

−1.0

−0.5

0

(b)

C

D

R

R

E

AA

B

AB

F

CD

D

E

F

F

EB

D

A

C

A

C

Figure 5.10. RF characterization results from five prototypes of the MEMS waveguide switchesand a reference measurement on a dummy chip showing the losses of the setup (common legendto both sub-figures). (a) OFF-state (blocking). (b) ON-state (non-blocking).

Fig. 5.10 shows measured performance of six test waveguide switches, in additionto the results of a reference chip. The data of six test chips includes losses by thechip-to-waveguide assembly and, therefore, to show the loss of the assembly setup


RF MEMS

Waveguide Switch,

M. Daneshmand et al.,

2004

Actuation mechanism

Frequency (GHz)

Insertion loss (dB)

Return loss (dB)

Isolation (dB)

Multiport MEMS-Based

Waveguide and Coaxial

Switches,

M. Daneshmand et al., 2005

Paper 4

Electrothermal

Ku and K-band

1–2.8

15

15

Electrostatic

7% of Ku-band

0.9–1.2

20

36–40

Electrostatic

V-band

0.65

>15

>30

Figure 5.11. RF MEMS waveguide switches reported in the literature [142, 143] and Paper4 [154].

separately, the measurement is done on a reference chip which does not have anyMEMS elements and consists of only a micromachined waveguide frame (V00H00).The test chips contain either 20, 30, or 35 vertical columns (chip nomenclatureVxx), and either 9, 17, or 21 horizontal suspension bars (chip nomenclature Hxx).These different implemented designs contain from 360 to 1470 vertical cantileversand between 200 and 770 parallel-actuated contact points.

In the OFF state (blocking), in the investigated frequency range of 60–70 GHz,the three chips with 30 or 35 vertical columns with an isolation of 35–45 dB havebetter RF perform than the three chips with 20 vertical columns, which have anisolation of 30–35 dB, thus the behavior is consistent with the expectations ofsimulation study. The best ON-state performance is measured for the V20H21 andV35H21 devices at 60 GHz with insertion loss of 0.4 dB while all devices haveON-state insertion loss between 0.4 and 1.1 dB over the whole frequency range.Device V20H9 has an overall best insertion loss, which is below 0.6 dB from 60to 69 GHz. The measurement of the reference chip shows that the setup itselfcontributes about 0.3 dB to these losses, which is the major part of the losses of thebest-performance measurements. The reference loss includes the losses due to chipmounting imperfections, polymer interposers, tailor-made flanges, and standardflanges for connection to the waveguide adapters.

The ON-state return loss is between 15 and 25 dB for all devices over the 60–70 GHz frequency range while the reference chip has a return loss between 25 and

5.5. Switchable waveguide irises 33

30 dB.As a conclusion, the ON-state insertion loss of the MEMS waveguide switches

presented here is significantly better than PIN-diode waveguide switches with inser-tion loss in the range 2–6.5 dB, and it performs as good as rotary motor waveguideswitches with insertion loss in the range 0.3–0.8 dB (Table 4.1), even including thelosses of the chip-to-waveguide integration. The measured isolation is 10–20 dBbetter than PIN-diode waveguide switches and the designs with 30 and 35 verticalcolumns have an isolation equal to rotary motor waveguide switches.

The comparison among the performance of the MEMS-reconfigurable waveguideswitch presented in this chapter and the MEMS waveguide switches reported in theliterature [142, 143] is summarized in Fig 5.11.

Further discussion on the RF characterization of the novel MEMS waveguideswitch is available in Paper 4.

5.5 Switchable waveguide irises

If the vertical cantilevers are only placed on specific areas of the waveguide cross-section, other microwave structures such as MEMS-reconfigurable waveguide irisesare possible. This section gives an overview of the design concept and RF char-acterization of inductive and capacitive irises. Some application examples of suchreconfigurable-irises are given in Section 5.6.

5.5.1 ConceptA simplified schematic view of the chips with a MEMS-reconfigurable inductiveand capacitive iris is shown in Fig. 5.12. The vertical cantilevers of the MEMS-reconfigurable surface are placed in the areas mimicking the standard inductive(Fig. 5.12(a)) and capacitive (Fig. 5.12(b)) irises while the switching from thedeactivated state to the activated state and vice versa, are provided by on-chipMEMS actuators similar to the MEMS waveguide switch discussed in the previoussection. In the deactivated-iris state the transmissive surface is transparent, andin the activated-iris state the vertical cantilevers are short-circuiting the electricfield lines of the TE10 mode and thus, blocking the wave propagation in the irisareas. The electric field lines in the iris-activated and iris-deactivated states ofan inductive iris and a capacitive iris are shown in Fig. 5.12(c) and Fig. 5.12(d),respectively.


(a) MEMS-reconfigurable inductive iris (b) MEMS-reconfigurable capacitive iris

(b)

TE10E

TE10E

reconfigurable iris areareconfigurable iris area

(c) Electric field lines of switchable inductive iris (d) Electric field lines of switchable capacitive iris

Figure 5.12. Simplified illustration of MEMS-reconfigurable iris concept. (a) 3D renderedschematic of a chip including the MEMS-reconfigurable inductive iris. (b) 3D rendered schematicof a chip including the MEMS-reconfigurable capacitive iris. (c) Configuration and electric fieldlines for an inductive iris in the deactivated (top) and activated (bottom) states. (d) Configurationand electric field lines for a capacitive iris in the deactivated (top) and activated (bottom) states.

5.5.2 RF simulation and characterizationThe measured performance of a reconfigurable inductive iris, compared to HFSS-simulation results, is shown in Fig. 5.13. The inductive iris has 21 horizontal barsand seven cantilever columns on each side, with an opening of 1.075 mm in thewaveguide H-plane. For the activated-iris state the simulation result is also com-pared to an ideal reference iris with filled metal areas, and for the deactivated-iriscase it is compared to a reference chip, with no MEMS elements, in the same setup.The measurements fit the simulations very well. The insertion loss of 0.4–1.0 dB inthe deactivated-iris state is mainly attributed to the waveguide assembly as shownby the reference-chip measurements.

Fig. 5.14 shows the HFSS-simulated and experimental performance of a capac-itive iris with 20 vertical cantilever columns and an E-plane opening of 0.777 mm.Also here, the measurements fit the simulations very well (insertion loss increasedby 0.2–0.4 dB due to assembly).

5.6. Application examples of reconfigurable waveguide irises 35

(a) Iris-activated (ON) state (b) Iris-deactivated (OFF) state

Activated (ON)

Frequency (GHz)

60 70 80

Deactivated (OFF)

simulated block iris

simulated reconfigurable iris

measured reconfigurable iris



measured reference chip

S (

dB

) 1

1

0

−1.0

−2.0

0

−4

−8

−12

−16−3.0

S (d

B)

21

−30

−20

−10

0

−1.0

−0.8

−0.6

−0.4

−0.2

0

−1.2

S (

dB

) 1

1

S (d

B)

21

Frequency (GHz)

60 70 80

Figure 5.13. RF characterization and HFSS simulation results of a MEMS-reconfigurable in-ductive iris. (a) Iris-activated/ON state. (b) Iris-deactivated/OFF state.

(a) Iris-activated (ON) state (b) Iris-deactivated (OFF) state

Activated (ON)

Frequency (GHz)

60 70 80

Deactivated (OFF)

simulated block iris





S (

dB

) 1

1

S (d

B)

21

−30

−20

−10

0

−1.0

−0.8

−0.6

−0.4

−0.2

0

−1.2

S (

dB

) 1

1

S (d

B)

21

Frequency (GHz)

60 70 80

0

−15

−20

−25

−10

−1.0

−0.8

−0.6

−0.4

−0.2

0

-5

Figure 5.14. RF characterization and HFSS simulation results of a MEMS-reconfigurable ca-pacitive iris. (a) Iris-activated/ON state. (b) Iris-deactivated/OFF state.

5.6 Application examples of reconfigurable waveguide irises

As application examples of switchable waveguide irises, we demonstrate the firstswitchable waveguide cavity resonator and present the concept of a tunable / switch-able waveguide filter. Further details of this application examples are provided in


(a) (b)

TiW/Au

SiO2

Si

Brass

Cu

Conductive

Polymer

WR-12waveguide

inputport

shortedport

fixe

d s

ho

rt

reconfigurablesurface

λ/2 = 3.057 mm

resonant cavity

Figure 5.15. Measurement setup for a switchable cavity resonator where the chip with a MEMS-reconfigurable inductive iris is integrated odd multiple of half-wavelengths from the back-shortedend of a WR-12 waveguide. (a) Camera photo. (b) Schematic drawing.

(a) (b)

−3

−2

−1

0

−4

S (d

B)

11

Frequency (GHz)

60 70 8065 75

deactivated

activated

−12S (d

B)

11 −8

−4

0

Frequency (GHz)

60 70 8065 75

Q=186.13activated

deactivated

−16

Figure 5.16. Iris-activated and iris-deactivated configurations of the MEMS-switchable cavityresonator. (a) Measurement results. (b) HFSS simulation.

Paper 5.

5.6.1 Switchable cavity resonator

By placing a switchable inductive iris at an odd-number multiple of a half wave-length from the back-shorted end of a rectangular waveguide, as shown in Fig. 5.15,a switchable cavity resonator was implemented. The measurement and simulationresults of the cavity resonator in the ON and the OFF states are shown in Fig 5.16.The simulations were done with a gold thickness of only 0.5 µm for appropriatemodelling of the metal losses. The resonator can clearly be switched ON and OFF,shown by the sharp ON -state resonance with a Q-factor of 186.13 at 68.87 GHz.In the OFF state, the insertion loss is less than 2 dB, including the measurementsetup, without any resonance in the frequency range of interest.

5.6. Application examples of reconfigurable waveguide irises 37

(b)

−30

−20

−10

0

S (d

B)

11

Frequency (GHz)

60 9070 80

deactivated

activated

(a)

−30

S (d

B)

21 −20

−10

0

Frequency (GHz)

60 70 9080−40

deactivatedactivated

BW=0.95%

Figure 5.17. HFSS-simulated results of the switchable first-order bandpass filter shown foractivated and deactivated states. (a) Insertion loss. (b) Reflection coefficient.

5.6.2 Tunable/switchable waveguide filtersBy using multiple MEMS-switchable irises in a waveguide, more complex compo-nents such as switchable/tunable millimeter-wave waveguide filters can be imple-mented.

Fig. 5.17 shows the HFSS simulation results for a first-order bandpass waveguidefilter which can be switched ON and OFF with MEMS-reconfigurable inductiveirises. The simulation results indicate that the filter has an insertion loss of 0.42 dBat the resonance frequency of 73.6 GHz with 3-dB bandwidth of 0.7 GHz in the ONstate and reflection coefficient of better than 0.15 dB in the OFF state.

Since comb-drive actuators provide the possibility of controlled displacementalong the whole movement range, it is also possible to utilize the reconfigurablesurface to tune the center frequency of the bandpass filter. Fig. 5.18(a) shows theHFSS model of the first-order bandpass filter where cavity resonators are made bysymmetric inductive irises and a MEMS-reconfigurable capacitive iris is placed be-tween the fixed inductive irises to change the effective length of the cavity resonator.Fig. 5.18(b) shows that the resonance frequency can be tuned from 73.5 GHz to78.5 GHz through a 20 µm displacement of the reconfigurable surface elements.The metal losses are modelled using a 1 µm gold cladding of the movable elements.


(a) (b)

−30

21 −20

−10

0

−40

Frequency (GHz)

60 70 9080

0 μm2 μm

4 μm

20 μm

S (d

B)

port 1

port 2

2.0 mm 3.1

mm

1.5

5

Figure 5.18. Simulated concept of a tunable first-order bandpass filter by integration of MEMS-reconfigurable capacitive iris within a cavity resonator: (a) HFSS model. (b) Simulated tuningperformance for 20 µm MEMS displacement range.

5.7 MEMS actuator characterization and reliability features

The on-chip MEMS actuators of the MEMS-reconfigurable designs discussed above,comprise laterally-moving electrostatic comb-drives, arranged in a push-pull con-figuration and placed symmetrically to the left and the right of the narrow wallsof the waveguide. Two designs are implemented, design I and design II, consistingof eight and sixteen total comb-drive actuators, respectively. The devices featurean all-metal design, i.e. no dielectric layers are exposed to neither the RF field northe DC (actuation) field, which virtually eliminates the reliability limitations bydielectric charging. Mechanical stoppers are implemented for limiting the actuatormovements. Furthermore, due to the SOI RF MEMS device concept, all mov-ing elements are utilizing monocrystalline silicon as the structural material, whichprovides high mechanical reliability. This silicon core is symmetrically metallized,which provides temperature compensation. As the device is bulk micromachined,there are no stress control/stress uniformity problems typical for surface microma-chined MEMS devices. The mechanical spring has a spring constant of 1.97 N/m(actuator design I) or 3.85 N/m (actuator design II), which is relatively soft, but incontrast to conventional MEMS switches the spring is not used to provide a restor-ing force but only to guide the movement of the actuator, as the device is activelypushed and pulled in both directions, i.e. featuring active opening and active clos-ing. The total movement is 24 µm or ± 12 µm around the rest position. Fig. 5.19shows the measured actuation voltage repeatability test for a chip with actuatordesign I, operated by a single-stage comb-drive with 106 fingers. After an initial 40cycles of stabilization, the mean actuation voltage is 40.0 V with a standard devi-ation (1σ) of 0.0605 V for the 20 subsequent cycles. The simulated passive releasetime for a critically-damped case is 91 µs (10%–90% fall time), which correspondsto the simulated mechanical resonance frequency of 5.85 kHz for the actuator witha spring constant of 1.97 N/m. However, the fact that the device is actively openedand actively closed allows for faster operation, in contrast to conventional MEMS

5.8. Discussion and outlook 39

40

30

20

10

0

41.5

41.0

40.5

40.0

39.540200 60 20100 30 605040

cyclecycle

Actu

atio

n v

olta

ge

(V

)

Figure 5.19. Measured actuation voltage repeatability for a chip with spring constant of 1.97N/m and one stage of comb-drive actuators on each side of the reconfigurable surface.

transmission-line switches whose switching time is limited by the passive openingtime and by design for critical damping.

A chip with actuator design II and a two-stage comb-drive actuator has a mea-sured average actuation voltage of 44.13 V. This device was tested in continuousactuation in contact mode during 14 hours, achieving 4.3 million cycles with a saw-tooth input signal of 80 Hz, without noticing any degradation in performance oractuation voltage. The comb-drives are designed for high lateral robustness of 92V (measured), giving a margin of at least a factor of two to the nominal actuationvoltage. Actuation voltage overdrive may be done up to this voltage which sets thelimit for the switching speed.

5.8 Discussion and outlook

The millimeter-wave waveguide components based on MEMS-reconfigurable sur-faces shown in Paper 1, Paper 4, and Paper 5 demonstrate a promising con-cept for switchable and tunable millimeter-wave waveguide systems. The proposedMEMS waveguide switch performs significantly better than PIN-diode waveguideswitches, shows performance comparable to rotary motor waveguide switches, evenincluding the assembly loss, and consumes negligible power in comparison to bothwaveguide switch technologies. The simulation and measurement for device-levelyield analysis of MEMS waveguide switches, discussed in detail in Paper 4, showsthat a device with 20 vertical cantilever columns and 21 horizontal suspension barsstill achieves an OFF-state isolation of better than 20 dB with 86% yield of verticalcantilevers and this device with 95% yield performs very similar to a 100%-yielddevice with measured OFF-state isolation of 30–37 dB.

The switchable iris concept presented here allows for completely activating anddeactivating the irises and thus enables new possibilities for MEMS-reconfigurablewaveguide devices, such as resonators and filters which can be switched ON andOFF, whereas, previous attempts of tunable cavity resonators and waveguide filtersonly provide limited tuning functionality.


Practical challenges of the concept presented here are the difficulty in gettinga thick metallization multiple of skin depth on the sidewalls in the fabrication andmaintaining a very low-loss assembly during the measurement.

Initial reliability and repeatability tests were conducted for the proposed devices,and to improve the reliability of these waveguide-integrated MEMS-based conceptsseveral suggestions are open to investigation such as replacing the gold with a hardermetal stacks. Simulations of MEMS waveguide switches indicate that electricalcontact is actually not necessary between the cantilevers, as capacitive contact withgaps in the order of up to few tens of nanometers is sufficient. I.e the performancein the blocking state was found not to be dependent on the contact force, contactresistance, and the number of cantilevers having physical contact. However, it issuggested to use parallel-plate actuators in combination with comb-drive actuatorsto increase the contact force for switchable waveguide iris concepts.

During this study, RF high-power handling performance of the MEMS waveg-uide switch could not be determined, as the power levels available to the authors inthe V-band are limited to -19 dBm, however, it would be interesting to determinepower handling capability of presented waveguide components.

Chapter 6

MEMS-based reconfigurable antenna arrayfor beam steering front-ends

This chapter gives a short introduction to beam steering front-ends, then presentsthe novel work done on MEMS-based waveguide-integrated phase shifters and uti-lization of such waveguide-integrated phase shifters in a phased-antenna array at77 GHz. Further details of these works are covered in Paper 3 and Paper 6.

6.1 Beam steering front-ends

Electronic beam steering, i.e. changing the direction of the main lobe of a radi-ation pattern of an antenna or of multiple stationary antenna elements, is widelyused in military and commercial applications such as radar systems and wirelesscommunications. The most popular method is to use phase shifters to create avariable phase gradient in the signal fed to antenna elements of an array, as shownin Fig. 6.1 [155]. Phase shifters are the critical components of such systems becausetheir characteristics strongly influence the signal performance and they affect thecost of the system.

The phase shifting devices in modern phased array antenna systems are typi-cally digital phase shifters, that is, the device can switch between predefined fixedphase states. Passive beam-forming is superior to active methods in terms of noise,linearity, and power handling, thus RF MEMS switches by far are the best can-didates for implementing high-performance phase shifters due to the advantagesdiscussed in Chapter 2.

There has been numerous efforts in realization of low-cost high performancebeam-steering systems for the millimeter-wave range based on semiconductor de-vices [1, 156–160] and MEMS technology [161–164].

41

42 Chapter 6. MEMS-based reconfigurable antenna array

RF

SourcePower

Divider

Variable

Phase

Shifters

Antenna

Elements

Equiphase Front

Beam Dire

ction

θ Radiation Angle

φ

φ

φ

φ

Figure 6.1. Schematic drawing of a basic concept of an electronic beam-steering front-end withphased-array antenna.

6.2 Novel MEMS-based waveguide-integrated phase shifters

Most MEMS-based beam-steering systems employ planar phase shifters due to tech-nological advances and ease of fabrication, however, integration of variable phaseshifting elements into the feed of an individual aperture antenna can potentiallyimprove the performance of the phased-antenna array. In this section a MEMS-based waveguide-integrated phase shifter is presented. In the next section, 6.3, atwo-dimensional antenna array is demonstrated which consists of unit cells similarto such waveguide-integrated phase shifters.

6.2.1 Concept and designThe electronically tunable waveguide phase shifter presented here consists of afinline filter integrated into a precision-machined waveguide, benefiting from lowinsertion-loss and high Q-factor characteristics of finline filters [165]. Finline filtersallow implementation of compact and leight-weight millimeter-wave components,as they are easy to integrate in millimeter-wave systems.

The 3D illustration of the basic concept is shown in Fig. 6.2, consisting of themicromachined chip with bilateral finline structure, i.e. the metallic finline elementsplaced on both sides of the dielectric substrate, integrated along the longitudinaldirection of a WR-12 rectangular waveguide (with inner dimensions 3.099 mm ×1.549 mm, and frequency range 60–90 GHz). Slots at the bottom and the top of thewaveguide allow bias feedthroughs for MEMS actuators of the microchip. Takingthe biasing lines into account is important for any MEMS array device, as therouting of the biasing lines often interferes with RF performance. The phase shifterchips are designed in an X-shape form, as shown in Fig. 6.2(a), in order to reducethe slot length in the top and the bottom faces of the waveguide, thus keeping

6.2. Novel MEMS-based waveguide-integrated phase shifters 43

Non-conductive

glue or film

MEMS

cantilevers

Silicon

substrate

Finline

Slot for

biasing lines

Waveguide

MEMS contact

cantilevers

0.0

2 m

m

1.205 mm

0.1 mm0.19 mm 0.18 m

m

0.02 mm

Gold (MEMS cantilevers)

Silicon (ε =11.9, tanδ=0.003)Gold (Fins)

2 µm

gap

Ohmic contact

(a)

(b)

(c)(d)

Figure 6.2. Concept of digital finline phase shifter based on MEMS switching elements. (a) 3Drendered view of the finline chip integrated into a WR-12 waveguide. (b) Side view of one finstructure with MEMS switching elements and comb-like fixed fin slot pattern of the ground layer.(c) Open-state configuration of the MEMS cantilever switch. (d) Closed-state configuration of theMEMS cantilever switch.


State 1: only switch 6 is closed

State 2: switches 4 and 11 are closed

State 3: switches 3 and 9 are closed

State 4: switches 1 and 11 are open

Figure 6.3. Illustration of phase shifter slot configuration for four distinct phase shifter states.

the radiation loss and the disturbance of the currents in these waveguide faces ata minimum. A longitudinal slot (20 µm wide) is patterned in the middle of eachfin, as shown in Fig. 6.2(b). Eleven MEMS cantilever switches are placed acrosseach fin slot, i.e. there are 22 total MEMS switches for the bilateral structure.The cantilvers can influence the wave propagation by short-circuiting the slot. Thephase of the wave transmitted from the input port to the output is controlled bya suitable combination of open and closed states of the individual cantilevers. Thedistance between the cantilever plate and the electrode is 2.0 and 0.8 µm in theopen and closed states respectively, as shown in Fig. 6.2(c) and (d).

For the proof of concept, the prototype chips of the MEMS-based finline phaseshifters are designed in fixed MEMS states, i.e. the actuators are not functional. Inthe fixed states, all actuation and biasing elements, e.g. all gold fin layers with allslots for the MEMS biasing lines, are included both in the simulation models and inthe fabricated designs, to include their influence on the RF performance. However,all cantilevers in the closed state are realized with gold short-circuiting connectionlines in the slot on the substrate surface. Moreover, the simulation model includesmetal and dielectric losses.

The phase of the transmitted wave is controlled by short-circuiting the slotat specific points, which imitates the closed cantilever state of the non-moveableMEMS. Four different embodiments are implemented, shown in Fig. 6.3, which cor-respond to four different phase states, shown in HFSS-simulated results of Fig. 6.4.From the simulation data of Fig. 6.4(a), the phase responses of the four chips atthe nominal design frequency of 77 GHz, normalized to State 4 are 0 ◦, 95 ◦, 158 ◦,and 246 ◦. The simulated reflection coefficient is lower than 10 dB for all states atthe design frequency, and the simulated insertion loss, S21, is better than -1 dB forall states at 77 GHz. State 1 and State 2 of this design are very narrowband.


Ph

ase

of S

(

de

g)

21

−180

−150

−120

−90

−60

−30

0

30

60

90

120

150

180

State 1State 2State 3State 4

133°

45°

-18°

-113°

S (d

B)

11

−30

−25

−20

−15

−10

−5

0


S (d

B)

21

State 1State 2State 3State 4−3

−2

−1

0

Frequency (GHz)76 77 7875.5 76.5 77.5

−0.5

−1.5

−2.5

(b)

Frequency (GHz)76 77 7875.5 76.5 77.5

Frequency (GHz)76 77 7875.5 76.5 77.5

(a)

(c)

−35

−3.5

Figure 6.4. HFSS-simulated performance of the finline filter designed at 77 GHz. (a) Phaseresponse. (b) Reflection coefficient. (c) Insertion loss.


(a) Adhesive bonding

(d) Lift-off on the other side

Ti/Au PhotoresistSi BCB

(c) Lift-off on one side

(f) Free-etching of the chips(e) DRIE of one wafer

(b) Lithography and Metallization

Figure 6.5. Process flow of the MEMS-based phase shifters. (a) Adhesive bonding of two 300-µmthick high-resistivity silicon wafers with a BCB layer. (b) Lithography and deposition of titaniumadhesion layer and gold on the front-side by evaporation technique. (c) Lift-off process. (d)Evaporate deposition of Ti/Au on the backside followed by lift-off. (e) Anisotropic deep reactiveion etching of the front side and the backside by Bosch process.

6.2.2 Prototype fabricationThe fabrication process of micromachined phase shifters is outlined in Fig. 6.5.First, two 300-µm thick high-resistivity silicon substrates (HRSS) are bonded usingbenzocyclobutene (BCB) as an intermediate adhesion layer (Fig. 6.5(a)). HRSSwith nominal resistivity of higher than 13 kW.cm is chosen because of low-losstangent of these substrates at room temperature in the entire W-band [166]. Alift-off process is used to pattern the metal structures by evaporating of 1 µm layerof gold on the front-side (Fig. 6.5(b) and (c)). The front-side is then protected byphotoresist followed by repeating the metallization lift-off process on the backside,to achieve a bilateral finline structure (Fig. 6.5(d)). Then, deep reactive ion etchingof the front-side wafer is done by Bosch process (Fig. 6.5(e)). Finally, deep reactiveion etching is also done on the backside to completely free-etch the chips while theyare kept in place on a temporary carrier wafer using a temporary thermal-releasetape (Fig. 6.5(f)).

Camera pictures of some prototype microchips are shown in Fig. 6.6, after beingreleased from the thermal-release adhesive tape.


Figure 6.6. Photographs of finline-based phase shifter chip prototypes.

(a) (b)

clamps

lidX-shaped chip

clamps

lid

X-shaped chip

Figure 6.7. Integration of X-shape phase shifter prototype chips into tailor-made WR-12 waveg-uide for RF measurement. (a) 3D rendered model. (b) camera picture.

6.2.3 RF characterizationA modified 120-mm long WR-12 waveguide section, shown in Fig. 6.7, is usedto securely mount the X-shape micromachined phase shifter prototypes into themeasurement setup. The modified section has recesses in the wide walls of thewaveguide to guarantee the accurate positioning of the chip in the center, and, itincludes a 10-mm long copper lid and two sliding clamps to keep the chip fixed.

The measured phase shift and S-parameters of the prototype chips mountedinto the modified waveguide are shown in Fig. 6.8. The measured phase states,shown in Fig. 6.8(a), relative to State 4, are 0 ◦, 56 ◦, 189 ◦ and 256 ◦, for the designfrequency of 77 GHz, which corresponds to an effective 1.78 bit. The measuredreflection coefficient and insertion loss of the four states are shown in Fig. 6.8(b)and (c). The measured reflection coefficients show that the reflected power is muchmore than the amount expected from the simulations. The insertion loss datashows the same behavior as it is measured better than 5.7 dB for the state withthe lowest loss. To investigate the reason for this discrepancy, the reproducibilityof the assembly and the influence of fabrication tolerances were tested, with thedetails provided in Paper 3. It is concluded that a major reason for the deviationof the measured S-parameters from the simulation data is the narrow-band designof the devices, which makes the performance very sensitive to design parametervariations.


(b)

(a)

Frequency (GHz)

Ph

ase

of S

(

de

g)

21

75 76 77 78 79 80−180

−120

−60

0

180

120

60

(c)

S (d

B)

21

−30

−25

−20

−15

−10

−5

0

Frequency (GHz)

75 76 77 78 79 80

Frequency (GHz)

S (d

B)

11

75 76 77 78 79 80−30

−25

−20

−15

−10

−5

0



State 1

State 2

State 3State 4

State 1State 2

State 3

State 4

State 1

State 2

State 3

State 4

Figure 6.8. RF characterization results of the finline filter designed for 77 GHz. (a) Phaseresponse. (b) Reflection coefficient. (c) Insertion loss.

6.3 Transmit-array antenna for beam steering front-ends

The MEMS-based phase shifter concept discussed above can be used as individ-ual building blocks of a waveguide-based transmit-array antenna for beam-steeringfront-ends, in a configuration shown schematically in Fig. 6.9. The transmit-arrayantenna is fed by a pyramidal horn antenna placed at a certain distance, and thephase shifter states (four in this design) are used to compensate for the differentelectrical path lengths between the horn and each cell of the radiating aperture. Thephase shifter elements discussed in the previous section was designed for proof-of-concept phase shifting principle, and not for antenna array application. Therefore,four rectangular-shape chips with simulated phase responses of −90 ◦, 15 ◦, 110 ◦

and 178 ◦ at 77 GHz were designed to be used as unit-cell of the transmit-arrayantenna for automotive radar beam-steering front-end.

The assembled transmit-array antenna, shown in Fig. 6.10, fabricated in bronze,have 10 × 21 elements and a total dimension of 54.55 mm × 41.9 mm × 10.205mm. There is a recess cut into the waveguide element to fix the rectangular-shapephase shifter chip.

6.3. Transmit-array antenna for beam steering front-ends 49

External feed

(horn antenna)

Array of proposed

phase shifters

MEMS-based

waveguide-integrated

phase shifter cell

Figure 6.9. Transmit-array based on the proposed finline MEMS-based phase shifter elements.

(b) (c)

(a)

Figure 6.10. Fabricated transmit-array antenna: (a) Integration of phase shifter chips into thearray. (b) Assembly of the antenna. (c) Complete transmit-array antenna.

Since the phase shifter chips are fabricated in the fixed MEMS states, the abilityof the antenna array for beam-steering is demonstrated by shifting the horn antennafeed in the E-plane parallel to the antenna array, with 4-mm step size equal to thefeed aperture dimension, as shown in Fig. 6.11(a). In this case, all phase shiftersare tuned to have maximum radiation at broadside. From the measurement resultsof Fig. 6.11(b) it can be seen that the antenna array exhibits good performance upto 15 ◦ at the designed frequency of 77 GHz, but the gain decreases rapidly and theside lobe level increases beyond this value. The half power beamwidth (HPBW) ofthe antenna array is 4.5 ◦–5.5 ◦ up to 12 mm displacement of the horn feed.

Further details on the antenna array design and performance characterization


0 4 mm 8 mm 12 mm 16 mm 20 mm

No

rm. G

ain

(d

B)

Angle (degree)

0

-5

-10

-15

-20

-25

-30-90 -60 -30 0 30 60 90

Antenna array

Shift along E-plane

(a)

Schematic illustration of beam-steering

measurement scenario

(b)

Antenna-array radiation patterns

with horn feed displacement as parameter

Figure 6.11. Antenna-array beam-steering measurement. (a) Schematic illustration of horn feeddisplacement to measure beam-steering ability of the antenna array. (b) Radiation patterns of theantenna array at 77 GHz for horn feed displacement as parameter.

are provided in Paper 6.


The proof-of-concept simulation results of MEMS-based waveguide-integrated phaseshifters presented in this chapter has shown promising performance, however, thecharacterization results demonstrated drastically higher insertion loss than ex-pected. The discrepancy is attributed mainly to the narrow-band design, and notto fabrication process and assembly setup, confirmed by additional measurements.

It is interesting to investigate and extend this concept further to achieve a fully-functional MEMS-reconfigurable waveguide phase shifters, but it is recommendedto optimize the design for wide-band performance. The fabrication process of fully-functional device is anticipated to require process-optimization iterations due tonecessary surface machining techniques.

Chapter 7

Integration challenges of microwave MEMSdevices into waveguides

This chapter starts with a background on challenges for integration of microwaveprototype chips into waveguides and gives some application examples of adhesivetapes and sheets in microsystems. Afterwards, it presents a novel low-loss methodof integrating microwave MEMS chips into rectangular waveguides using a double-sided adhesive conductive polymer interposers for reflective and transmissive sur-faces. Further details and investigation of this method is provided in Paper 2.

7.1 Background

A major practical obstacle of millimeter-wave systems based on waveguides withintegrated MEMS chips is the reliable, reproducible and low-loss assembly of themicromachined MEMS chips into the conventionally machined waveguides. Theassembly solution needs to provide a stable chip fixture for mechanical reliabilityand reproducibility of the measurement without leaving any air-gap between thechip and the waveguide wall, which contributes to loss and often introduces spuri-ous resonances. A few methods have been reported to integrate MEMS chips intowaveguides including assembling the MEMS devices directly into the waveguidewithout any interfacing layer [142, 143] and with copper foil pads as interposerlayer [150], using screws to connect the hard-wired MEMS devices and the waveg-uide [167, 168], and using conductive epoxy to attach the MEMS membrane to thewaveguide wall [146].

In microsystems, adhesive tapes and sheets have been employed for differentapplications, for examples as an interface material for assembly [169] and bond-ing [170, 171]. Moreover, conductive tapes and sheets have been used in applica-tions requiring electrical connection, for example, as vertical interconnects [172], asan electrical contact of sensor electrodes [173], and as an interface to mount MEMSchips onto a waveguide [174].

51

52 Chapter 7. Integration challenges of MEMS devices into waveguides

This chapter discusses the integration aspects of reflective and transmissiveMEMS chips which are electrically connected and mechanically fixed to rectangularwaveguides at 60–110 GHz frequency range using conductive polymer sheets.

7.2 Prototype integration concept

Fig. 7.1 shows an illustration of the proposed chip-to-waveguide integration conceptfor reflective and transmissive MEMS surface applications, based on conductivepolymer interfacing layer. Furthermore, a cross-sectional view comparing the air-gap interfacing and the new interfacing, with conductive polymer interposers isshown in Fig. 7.2. For the air-gap proximity integration method, the waveguideis not electrically connected to the MEMS chip, and the resulting air gap causesradiation losses and often spurious resonances, especially in the millimeter-wavefrequencies. Direct touching of the waveguide on the chip must often be avoidedsince the rough surface topography of conventionally machined components resultsin local stress maxima which break the brittle microchips. Furthermore, the lack ofmechanical connections introduces complexities in the mounting and alignment. Incomparison, these issues are addressed by the proposed interfacing concept, wherethe air gap is filled by the conductive polymer sheet. This concept combines thefollowing features.

• Good and uniform RF and DC electrical connection between the MEMS chipand the waveguide frame is provided by a conductive polymer layer of lowresistivity (8 mW.cm), avoiding any unintentional discontinuities in the high-frequency current distribution which increase the radiation loss and introducespurious resonances.

• Secure and reliable mechanical connections of the MEMS chip into the waveg-uide by double-sided adhesive interposer as the interposer addresses the mis-match in the surface roughness and fabrication tolerances between the micro-machined MEMS chip and the waveguide with macro-scale topography. Thesoftness of the compliant conductive polymer layer with a Young’s modulusin the lower GPa range significantly reduces the risk of breaking the brittleMEMS chip at local high stress contact points.

• Waveguide feedthroughs for DC biasing of MEMS actuators through the poly-mer frame without decreasing the RF performance or short-circuiting to thewaveguide frame is easily possible by structuring the polymer interposer andpassing the biasing lines along the narrow walls of the waveguide to avoid anyinterruption of the major waveguide surface currents in the dominant mode,thereby avoiding unnecessary radiation loss.

• Mechanical interconnection of MEMS actuator through the polymer openingis possible by perforating the interposer to accommodate mechanical connec-tions in applications where the MEMS actuators are required to be placed

7.2. Prototype integration concept 53

rectangular

waveguide

conductivepolymerspacer

MEMSchip

MEMSreconfigurablereflective

surface

feedthroughfor DC biasing

rectangular

waveguide

groundplane

MEMSreconfigurabletransmissive

surface

shuttleconnectionof actuators

groundplane

(a) (b)

conductivepolymer

interposer

Figure 7.1. 3D view of the proposed integration concept for integrating MEMS chips into arectangular waveguide by structured double-sided adhesive conductive polymer interposers. Theconcept is displayed for integration of (a) a reflective surface, and (b) a transmissive surface.


(a)

Chip-to-waveguide integration of

a chip containing a reflective surface

(b)

Chip-to-waveguide integration of

a chip containing a transmissive surface

groundlayer

dielectriclayer

biasingline

MEMSreconfigurable

surface

waveguide

conductingspacer

Air-gap

interfacing

New

interfacingwaveguide

dielectriclayer

MEMSreconfigurable

surface

conductingspacer

device layer

waveguide

conductingspacer

Air-gap

interfacing

New

interfacing

Au SiO2 Cu Conductive polymerSi

Figure 7.2. Schematic illustration to compare an air-gap and the proposed concepts for two typesof MEMS chips. (a) MEMS chip with a reflective surface; (b) MEMS chip with a transmissivesurface. In the air-gap concept substantial radiation loss could happen between the waveguideswhich is addressed in the new concept by replacing the gap with a compliant conductive polymerinterposer. Moreover, the new concept allows connecting the waveguide to the ground plane ofthe MEMS chip and provides the possibility to pass DC bias lines or mechanical connection ofthe MEMS actuators between the two waveguide.

outside of the waveguide frame in order not to interfere with the propagatingwave.

• Reworkable bonding interface is provided by this concept as the bonding inter-face between the conductive adhesive sheet and the waveguide is not perma-nent. Reworkable interfaces are essential in testing, calibration, and prototypedevelopment where MEMS chips are frequently replaced.

As compared to these features, the disadvantages involved with the proposedinterfacing concept are as follows.

• Lack of hermeticity of polymer interposer. Even though it is generally adust and a water barrier, it is not gas-tight, and if feedthroughs need to beimplemented, the particle and water protection are lost.

• Patterning of the adhesive polymer sheet is required for shaping the openingsfor DC biasing or mechanical interconnection of the actuator, therefore, acutting plotter or similar equipment is recommended. Sub-millimeter sizedpieces of soft polymer sheets can be challenging to integrate on small chips into

7.3. Characterization 55

waveguides working in the high millimeter-wave frequency range. The manualassembly demonstrated in this prototype work needs further automation tobe suitable for volume production.

• Alignment issues and safety distance requirements for MEMS designs consid-ering that the softness of polymer sheet results in significant extension in lat-eral direction under high assembly pressure between the chip and the waveg-uide, even though it is beneficial for secure and reliable mounting. Therefore,some alignment precaution and consideration of some design safety marginsbetween the MEMS movable parts and the applied polymer sheet is necessary.

7.3 Characterization

The double-sided adhesive conductive polymer interposer utilized here, to mountthe prototype MEMS chips, is commercially available Ted Pella Silver ConductiveSheet 16086-1 with a nominal thickness of 125 µm. They are patterned by a Graph-tec CE5000 cutting plotter, with a nominal positioning accuracy of 5 µm, and aremanually applied to the chips.

A comparison between the RF performance of two assembly prototypes, onewith a conductive polymer interface layer between the waveguide and the chipand the other with a direct metal-to-metal interface, is shown in Fig. 7.3, with thefigure including also the performance of a waveguide short as a reference. A 250- µmopening is perforated the polymer ring for electrical feedthroughs on both narrowsides. The results illustrate the advantages of the compliant polymer and confirmsthat the conductivity of the polymer interface is high enough in this frequencyrange, as the prototype with conductive polymer interface performs as good asthe one with metal contact interface below 95 GHz, and even better than thedirect metal contact above 95 GHz. This can be explained by local air gaps inthe direct waveguide-to-chip interfacing due to mismatching surface topography ofthe waveguide frame, in particular on the wider sidewalls of the waveguide. Theintroduced loss of the conductive polymer interface is less than 0.4 dB as comparedto the reference standard short.

The direct metal interface is not suitable for the integration of a MEMS chip,and thus, for comparison to the above discussion, it is interesting to investigatethe return loss of an assembly setup consisting of an air-gap interface between ametal waveguide and a reflective surface, imitated by a standard fixed-short flange.Fig. 7.4 shows the measured performance of such an assembly for different air-gapwidths. For small air gaps below 30 µm, the return loss can be as low as 1.5 dB.However, such an air gap interface results in many spurious resonances and verysmall usable bandwidths. For air gaps above 50 µm, the performance drasticallydegrades, and the gap between the waveguide flanges results in sharp resonancepeaks. This comparison emphasizes the performance improvement of the proposedpolymer interface, in particular for gaps wider than 50 µm, and even if the polymeris only used on the wider sides of the waveguide.


75 80 85 90 95 100 105 110–6

–5

–4

–3

–2

–1

0

Frequency (GHz)

Standardshort

Metalcontact

Conductivepolymer

S (d

B)

11

Figure 7.3. RF characterization results comparing two different chip-to-waveguide connectionmethods for the reflective surface applications, conductive polymer interface and a direct metalinterfacing, including the measurement results of a waveguide short as reference. The proposedinterface performs as good as or better than direct metal contact, due to the compliancy of thepolymer surface.

20 µm

30 µm

40 µm

50 µm

100 µmgap (µm)

75 80 85 90 95 100 105 110

Frequency (GHz)

–30

–25

–20

–15

–10

–5

0

no

rma

lize

d to

a s

tan

da

rd s

ho

rt

S (d

B)

11

Figure 7.4. RF characterization of an air-gap interface to a reflective surface normalized to astandard waveguide short for different gap distances, showing the poor interface performance ofsuch interfaces.

To investigate the integration performance for a transmissive surface applica-tion, Fig. 7.5 shows the measured performance of a complete assembly setup, wherethe transmissive chip is integrated into modified flanges and the WR-12 waveguide,comparing three integration cases: (1) Conductive polymer interposer applied be-

7.3. Characterization 57

S (d

B)

11

S (d

B)

21

−50

−40

−30

−20

−10

0

Frequency (GHz)

60 65 70 75

−2.0

−1.5

−1.0

−0.5

0

Frequency (GHz)

60 65 70 75

Metal

contact

Conductive

polymer on the top side

Conductive

polymer on both

sides

Metal

contact

Conductive

polymer on

the top side

Conductive

polymer on

both sides

Figure 7.5. Measurement results of a complete assembly with an integrated transmissive chip,including the two modified waveguide flanges, comparing three cases: direct contact between themetal layer on the chip and the waveguides, polymer interposer between the top metal layer ofthe chip and the waveguide, and polymer interposer applied on both the top and the bottom facesof the chip. The assembly using the polymer interposer shows much more stable results than thedirect metal contact, as the latter case might lead to partial air gaps between the chip and thewaveguide. Applying the polymer interposer on both sides of the chip improves the result slightlycompared to applying the interposer only on one side of the chip.

tween the top layer of the chip and the wide walls of the waveguide; (2) conductivepolymer interposer covering the interface of the chip and the wide walls of thewaveguide, both on the top and the bottom layer of the chip; (3) direct metal con-tact between the waveguide and the metal layer on the chip. The transmissive chip,which has a through-silicon opening as large as the waveguide inner dimensions, iscompletely covered with 1 µm layer of gold. The results confirm that the compliantconductive polymer interposer removes any partial air gap, improves the perfor-mance, and results in stable measurements over the whole frequency range. Theinsertion loss of the whole assembly setup is between 0.27 dB and 0.59 dB, and thereturn loss is better than 20.38 dB where the conductive polymer is only appliedon one side of the chip, compared to the improved insertion loss between 0.21 dBand 0.53 dB, and return loss of better than 21.01 dB where the polymer interposeris applied on both sides of the chip. Therefore, applying the conductive polymer


interposer on both sides of the chip improves the performance insignificantly incomparison to the case where the interposer is only applied on one side of the chip.

Further data on performance analysis of the conductive polymer interfacingconcept is provided in Paper 2, including the effect of resistivity and thickness ofpolymer intersposers, as well as, the width of the opening for the feedthroughs.


The simulation and measurement results for the integration of RF MEMS chips intoa waveguide with a commercial conductive polymer interposer were investigatedup to 110 GHz in this chapter, showing the advantages of this method over theintegration with an air-gap interfacing, with the latter method causing up to 3dB reflected power for gaps as small as 20 µm, introducing spurious resonances,and having very small usable bandwidth. The proposed method performs as goodas or better than direct-metal contact investigated for both reflective surface andtransmissive surface applications.

The new method can be utilized in significantly higher frequencies, however,some additional considerations are required.

Since the cutoff frequency of the waveguide is related to the inner dimensions ofthe waveguide, waveguides with smaller dimensions are used for higher frequencyapplications, and therefore the performance becomes more sensitive to fabricationand assembly tolerances, making the manual assembly methods challenging.

The radiation loss from the opening of the waveguide is proportional to thefrequency, thus to scale up in frequency without performance degradation, thinnerpolymer sheets should be used.

Since the silver particles introduced in the polymer sheet to provide electricalconductivity are much smaller than the wavelength even at 1 THz (free-space wave-length of 300 µm), it is believed that the conductivity of the conductive polymersheet used in this work is suitable for even terahertz frequency range.

Other interesting aspects for industrial adaption of the method would be theinvestigation of other commercial conductive polymer interposers provided by dif-ferent manufactures and the development of automated integrating systems.

Chapter 8

Conclusions

In this thesis novel microelectromechanical (MEMS) devices for millimeter-wavewaveguide applications in two categories are presented, and, also, a novel methodfor integration of MEMS chips into waveguides, by using conductive polymer inter-posers, is investigated.

In the first part of the thesis a 30-µm thick MEMS-reconfigurable surface isintroduced and it is utilized to present the first waveguide switch above 40 GHz,which has similar excellent signal performance as rotary motor-based waveguideswitch, but has a drastically reduced size. Furthermore, the concept of MEMS-reconfigurable surface is extended to the first MEMS-switchable waveguide iris todemonstrate a V-band switchable waveguide cavity resonator. Besides, the po-tential of MEMS-reconfigurable irises for tunable/switchable waveguide bandpassfilters has been shown.

Moreover, a new configuration of a digital waveguide phase shifter is proposed,based on distributing MEMS switches along the finline structure of a waveguidebandpass filter, allowing the phase shift control by changing the MEMS switchcombinations. The proof-of-concept prototypes for WR-12 waveguide were fabri-cated and employed in a phased-antenna array for 77 GHz beam steering front-end.

In addition, to address the integration difficulties and complexities of waveguide-integrated MEMS-based high frequency systems, a mechanically-robust electricallylow-loss chip-to-waveguide integration concept was investigated based on commer-cial conductive polymer interposers for applications of up to 100 GHz.

59

Summary of Appended Papers

Paper 1: MEMS reconfigurable millimeter-wave surface for V-band rectangular-wave-guide switchThe paper presents a concept of a microelectromechanical systems (MEMS)waveguide switch based on a reconfigurable surface, whose working principleis to block the wave propagation by short-circuiting the electrical field linesof the TE10 mode of a WR-12 rectangular waveguide. The reconfigurablesurface consists of up to 1260 micromachined cantilevers, which are movedsimultaneously by integrated MEMS comb-drive actuators. Measurements ofproof-of-concept prototypes, which were not fully functional and were fixatedmechanically, show very with promising results where the devices blockingwave propagation in the OFF-state with over 30 dB isolation for all designs,and allow for transmission of less than 0.65 dB insertion loss for the bestdesign in the ON-state for 60–70 GHz.

Paper 2: Integration of microwave MEMS devices into rectangular waveguidewith conductive polymer interposersThe paper investigates a novel method of integrating microwave MEMS chipsinto millimeter-wave rectangular waveguides. The fundamental difficulties ofmerging micromachined with macromachined microwave components, in par-ticular surface topography, roughness, mechanical stress points, and air gapsinterrupting the surface currents, are overcome by a double-side adhesiveconductive polymer interposer. This interposer provides a uniform electricalcontact, stable mechanical connection and a compliant stress distribution in-terlayer between the MEMS chip and a waveguide frame. Moreover, both DCbiasing lines and mechanical feedthroughs to actuators outside the waveguideare demonstrated which is achieved by structuring the polymer sheet xuro-graphically. FEM simulations were carried out for analyzing the influence ofdifferent parameters on the RF performance.

Paper 3: Waveguide-integrated MEMS-based phase shifter for phased array an-tennaThe paper investigates a new concept of waveguide-based W-band phaseshifters for applications in phased array antennas. The phase shifters are

61

62 Summary of Appended Papers

based on a tunable bilateral finline bandpass filter with 22 microelectrome-chanical system (MEMS) switching elements, integrated into a custom-madeWR-12 waveguide with a replaceable section, whose performance is also in-vestigated in this study. The individual phase states are selected by changingthe configuration of the switches bridging the finline slot in specific positions.MEMS chips have been fabricated in fixed positions, to prove the principle,that is, they are not fully functional, but contain all actuation and biasing-line elements. The measured phase states are 0, 56, 189 and 256°, resultingin an effective bit resolution of 1.78 bits at 77 GHz.

Paper 4: Parameter analysis of millimeter-wave waveguide switch based on aMEMS-reconfigurable surface

The paper presents a novel concept of a millimeter-wave waveguide switchbased on a MEMS-reconfigurable surface with insertion loss and isolation verysimilar to high performance but bulky rotary waveguide switches, despite itsthickness of only 30 µm. A set of up to 1470 micromachined cantilevers ar-ranged in vertical columns are actuated laterally by on-chip integrated MEMScomb-drive actuators, to switch between the transmissive state and the block-ing state. In the blocking state, the surface is reconfigured so that the wavepropagation is blocked by the cantilever columns short-circuiting the electricalfield lines of the TE10 mode. A design study has been carried out identify-ing the performance impact of different design parameters. A device-levelyield analysis was carried out, both by simulations and by creating artificialdefects in the fabricated devices, revealing that a cantilever yield of 95% issufficient for close-to-best performance. Life-time measurements of the all-metal, monocrystalline-silicon core devices were carried out for 14 hours afterwhich 4.3 million cycles were achieved without any indication on degradation.Furthermore, a MEMS-switchable waveguide iris based on the reconfigurablesurface is presented.

Paper 5: MEMS-reconfigurable irises for millimeter-wave waveguide components

The paper presents MEMS-reconfigurable waveguide irises based on reconfig-urable surfaces integrated into a WR-12 rectangular waveguide (60–90 GHz).The reconfigurable surface is only 30 µm thick and incorporates 252 simul-taneously switched contact points short-circuiting the electric field lines ofthe TE10 mode. This novel concept allows for completely switching ON andOFF inductive or capacitive irises for building switchable frequency-selectivewaveguide devices. The measurements in the ON and OFF states matchsimulation results very well. Furthermore, a switchable cavity resonator wasimplemented with a MEMS-reconfigurable inductive iris, resulting in a mea-sured ON -state reflection coefficient of 14.5 dB with a Q-factor of 186.13 atthe resonance frequency of 68.87 GHz. Moreover, it is shown that by usingmultiple MEMS-reconfigurable irises in a waveguide, more complex compo-

63

nents such as switchable/tunable millimeter-wave waveguide filters can beimplemented.

Paper 6: MEMS-reconfigurable transmit antenna arrays for 77-GHz automotiveradar applicationsThe paper presents an alternative solution consisting of a waveguide-basedtransmit-array antenna with phase shifters used as individual building blocks.A new antenna array configuration is based on WR-12 waveguide. The pro-posed configuration of the integrated phase shifters is based on a bilateralfinline bandpass filter associated to MEMS elements as switches and oper-ates at 77 GHz. The phase states are controlled by changing MEMS statesintegrated in the finline slot. The proposed antenna array model has beenprototyped and characterized.

Acknowledgments

This thesis represents the outcome of my doctoral research at the department ofMicro and Nanosystems (MST), KTH Royal Institute of Technology, however, tome, it is considered a milestone for more than one decade of university educationand research work, which would have been impossible to achieve without the helpand support of many people, to only some of whom it is possible to give particularmention here.

First of all, I would like to express my sincere gratitude to my supervisor Prof.Joachim Oberhammer for giving me the opportunity to continue as a PhD studentin his team, even before I officially hand-in my master thesis, for giving endlessresearch advice, for scrutinizing my scientific writings and helping with the im-provement, and, finally, for guiding me whenever I was lost into insignificant de-tails. I am also thankful to Prof. Göran Stemme for officially welcoming me to thedepartment, leading the MST group, and giving research insights.

Financial support for this work has been provided partly through grants from theEuropean Commission via the FP7 TUMESA project and European Community’sSeventh Framework Programme via xMEMS.

Furthermore, I would like to thank the following people who have contributedto my professional and personal time at the MST group, in no particular order:Niclas Roxhed, for his guidance about cleanroom-related tasks and charming loud“Hi Zargham” in the mornings | Frank Niklaus, for proof-reading this thesis andgiving inputs for improvement | Mikael Sterner, for supervising my master the-sis, for helping me getting started with the lab work in Kista, for giving valuableideas during processing discussions, and for other assistance along the way | UmerShah, for always being nice and positive, for sharing the office with me, and for nothesitating to help with anything varying from technical discussions to simple ad-ministrative tasks, and for always volunteering for after-work activities | NutapongSomjit, for updating particularly me and Umer about scientific and general inter-esting news on facebook and having an on-going discussion about them, and forgiving the group Thai food for his birthdays | Gaspard Pardon, for being a goodofficemate and roommate during the MEMS conference, and for initiating deep in-teresting discussions about general topics | Nikolai Chekurov, for helping with theion-milling, for sharing the evening working shifts with me, and for being open todiscuss about whatever topic in a humorous tone | Fritzi Töpfer, for taking the

65

66 Acknowledgments

initiative in bringing the group closer by off-hour activities, and for visiting ouroffice sometimes just to say hi | Erika Appel and Ulrika Pettersson for managingthe administrative and economic issues | Kristinn B. Gylfasson (a.k.a KBG), for hisbroad and profound knowledge which he is always ready to share | Wouter van derWijngaart, for his valuable insights about any topic often supported by out-of-the-box ideas, and for injecting the charming spirit into the group during fika | HansSohlström, for refreshing my skills on measurement techniques and their pitfalls,and for helping with some education-related administrative tasks | Kjell Norén, forbeing “the” lab engineer of the group during my PhD time – with skills rangingfrom circuit design to manually milling of complex aluminum structures – and forfabricating holders for critical point dryer and sputter machine | Mikael Karlsson,for his nice attitude towards everything and everyone, and for occasional stand-upchats in front of the lab | Simon Bleiker, for his objectiveness, and for describingSwiss political system, and for speaking his views on societal matters | ValentinDubois, for his French charm, and for bringing interesting processing discussions |Mikael Antelius, for his calm and nice attitude, and for discussions about Sweden| Andreas Fischer, for help with wire bonding, for giving me the modified templateof this thesis (the idea of vertical separators between names), and for being a goodmotivator for me to take a scuba diving course | Hithesh Gatty, for teaching me themilling machine, and for sending us interesting Nature and Science papers | StaffanJohansson, for his interesting discussions about tech. stuff, and for his incrediblesilltårta | Martin Lapisa, for help with the wafer bonder and optical profilometer,and other clean-room issues | Stephan Schröder, for the help with the wire bonderand metallization, and for sharing the teaching assistantship of a course with me |Sergey Dudorov, for his extensive advice and guidance on RF/microwave problems| Fredrik Forsberg, for being a very knowledgable person and speaking his thoughtshonestly | Henrik Gradin, for the talks which happened mostly out of the workenvironment | Laila Ladhani, for skimming through some parts of this thesis eventhough it was only a-couple-of-day notice | Carlos Errando Herranz, for taking careof MST website and sending detailed reminders | Farizah Saharil, for the happy andpositive attitude | Jonas Hansson, for always giving an unexpected cool comment| Floria Ottonello Briano, for explaining the Italian way of life and expectations |Stefan Braun for making sure that everyone was aware of own’s kitchen responsi-bility and visiting us at MST every now and then | Tommy Haraldsson, for lettingme absorb some polymer/lab-on-a-chip knowledge when he was talking to Gaspardand Laila in our previous office, and for humorous comments during fika | Chianty(Xiamo) Zhou for shifting the average age of the group significantly | Adit Decharat,for taking the night shifts at the MST lab | Niklas Sandsträm, for always being inthe right mood, and for taking care of the MST web stuff | Mikael Bergqvist, forfitting very quickly into the group from the first day, and for sending us nice hikingpictures when he is on vacation | Alexander Vastesson, for taking the initiative forstuff related to the lab at MST | Mina Rajabi and Reza Zandi Shafagh, for joiningour group to increase the Persian population, and for providing me the possibilityto speak also Persian during the working hours | Gabriel Lenk, Maoxiang Guo, and

67

Xuge Fan, for joining as PhD students when I am about to finish | Stavros Yika,for being our officemate, and involving us in some electronic/electromagnetic ques-tions | Björn Samel, for occasional helping in the cleanroom of Kista even after hisposition as lab manager for Acreo | Cecilia Aronsson, for always unselfishly helpingwhen I was stuck with a cleanroom tool | Jan Åberg, for his advice with anythingRF related and for fabricating a holder for my MEMS waveguide switches, as wellas, tailor-made waveguide sections for TUMESA project | Gunnar Malm, for guid-ance with the vector network analyzer (VNA) | Alexander Vorobyov, for a fruitfulcollaboration during the TUMESA project leading to numerous publications | Ro-nan Sauleau, for giving inputs for the manuscript of papers related to TUMESAproject | Dmitry Chicherin and Antti Räisänen of Aalto University, and ErwanFourn of University of Rennes for the good collaboration in the TUMESA project| Alan Cheshire and Gabriel Roupillard, for helping with the Centura etching tool.

I would like to acknowledge the personnel from our cleanroom facility, in partic-ular, Reza Nikpars, Magnus Lindberg, Roger Wiklund, Aleksandar Radojcic, andPer-Erik Hellström for their effort in keeping the equipment in good shape andassisting me in the cleanroom.

Living abroad very far from home and for a long time has its difficulties, and Iam grateful to all my friends, because of whom this has become a lot easier. I amalso thankful to my girlfriend, Ramona, who has supported me, specifically, duringthe long nights of thesis writing for the past months.

Most importantly, my deepest gratitude belongs to the ones who have supportedme unconditionally, by all means, during my lifetime and taught me to believe inmyself, although they had to live a continent apart, to the west and to the east,for years; thus, my special thanks belong to my parents and my brother, who havecherished with me every exciting moment and supported me when I needed.

Zargham BaghchehsaraeiStockholm, SwedenMarch 11, 2014

References

[1] J. Wenger, “Automotive radar - status and perspectives,” in Proc. IEEECompound Semiconductor Integrated Circuit Symp., Oct. 2005.

[2] P. Smulders, “Exploiting the 60 GHz band for local wireless multimedia ac-cess: prospects and future directions,” IEEE Commun. Mag., vol. 40, no. 1,pp. 140–147, 2002.

[3] R. Daniels and R. Heath, “60 GHz wireless communications: emerging re-quirements and design recommendations,” IEEE Vehicular Technology Mag-azine, vol. 2, no. 2, pp. 41–50, Sept. 2007.

[4] V. Dyadyuk, J. Bunton, J. Pathikulangara, R. Kendall, O. Sevimli, L. Stokes,and D. Abbott, “A multigigabit millimeter-wave communication system withimproved spectral efficiency,” IEEE Trans. Microwave Theory Tech., vol. 55,no. 12, pp. 2813–2821, Dec. 2007.

[5] R. Appleby and R. Anderton, “Millimeter-wave and submillimeter-wave imag-ing for security and surveillance,” Proc. IEEE, vol. 95, no. 8, pp. 1683–1690,Aug. 2007.

[6] J. A. Nanzer, Microwave and Millimeter-Wave Remote Sensing for SecurityApplications. Norwood, MA: Artech House, 2012.

[7] F. Topfer, S. Dudorov, and J. Oberhammer, “Micromachined 100GHz near-field measurement probe for high-resolution microwave skin-cancer diagnosis,”in IEEE MTT-S Int. Microw. Symp. Dig., Montreal, QC, Canada, June 2012,pp. 1–3.

[8] H. H. Meinel, “Millimeter wave applications and technology trends,” annalsof telecommunications - annales des télécommunications, vol. 47, no. 11-12,pp. 456–468, 1992.

[9] M. Marcus and B. Pattan, “Millimeter wave propagation; spectrum manage-ment implications,” IEEE Microwave, vol. 6, no. 2, pp. 54–62, 2005.

69

70 References

[10] V. Lubecke, K. Mizuno, and G. Rebeiz, “Micromachining for terahertz ap-plications,” IEEE Trans. Microwave Theory Tech., vol. 46, no. 11, pp. 1821–1831, Nov. 1998.

[11] G. T. A. Kovacs, Micromachined Transducers Sourcebook, 1st ed. New York:McGraw-Hill, 1998.

[12] S. D. Senturia, Microsystem Design. Boston, MA: Kluwer Academic, 2000.

[13] S. Franssila, Introduction to Microfabrication. New York: Wiley, 2004.

[14] C. Liu, Foundations of MEMS, 2nd ed. Prentice Hall, 2011.

[15] M. J. Madou, Fundamentals of Microfabrication and Nanotechnology, 3rd ed.New York: CRC Press, 2011.

[16] R. Bogue, “Recent developments in MEMS sensors: a review of applications,markets and technologies,” Sensor Review, vol. 33, no. 4, pp. 300–304, 2013.

[17] G. M. Rebeiz, RF MEMS Theory, Design and Technology, 1st ed. Hoboken,New Jersey: Wiley, 2003.

[18] V. K. Varadan, K. J. Vinoy, and K. A. Jose, RF MEMS and Their Applica-tions. Chichester, UK: Wiley, 2003.

[19] H. J. D. L. Santos, RF MEMS Circuit Design for Wireless Communications.Norwood, MA: Artech House, 2002.

[20] H. A. C. Tilmans, “MEMS for wireless communications: ’from RF-MEMScomponents to RF-MEMS-SiP’,” IOP J. Micromech. Microeng., vol. 13, no. 4,pp. S139–S169, Jul. 2003.

[21] E. R. Brown, “RF-MEMS switches for reconfigurable integrated circuits,”IEEE Trans. Microwave Theory Tech., vol. 46, no. 11, pp. 1868–1880, Nov.1998.

[22] J. Yao, “RF MEMS from a device perspective,” IOP J. Micromech. Microeng.,vol. 10, pp. R9–R38, Dec. 2000.

[23] R. Mahameed and G. M. Rebeiz, “RF MEMS capacitive switches for widetemperature range applications using a standard thin-film process,” IEEETrans. Microw. Theory Tech., vol. 59, no. 7, pp. 1746–1752, Jul. 2011.

[24] R. Stefanini, M. Chatras, P. Blondy, and G. M. Rebeiz, “Miniature MEMSswitches for RF applications,” J. Microelectromech. Sys., vol. 20, no. 6, pp.1324–1335, Dec. 2011.

[25] T. Kawakubo, T. Nagano, M. Nishigaki, K. Abe, and K. Itaya, “RF-MEMStunable capacitor with 3 v operation using folded beam piezoelectric bimorphactuator,” J. Microelectromech. Sys., vol. 15, no. 6, pp. 1759–1765, Dec. 2006.

References 71

[26] T. G. S. M. Rijks, P. G. Steeneken, J. T. M. van Beek, M. J. E. Ulenaers,A. Jourdain, H. A. C. Tilmans, J. D. Coster, and R. Puers, “Microelec-tromechanical tunable capacitors for reconfigurable RF architectures,” IOPJ. Micromech. Microeng., vol. 16, no. 3, pp. 601–611, Mar. 2006.

[27] H. Yamazaki, T. Ikehashi, T. Saito, E. Ogawa, T. Masunaga, T. Ohguro, andH. S. Y. Sugizaki, “A high power-handling RF MEMS tunable capacitor usingquadruple series capacitor structure,” in Proc. IEEE MTT-S Int. Microw.Symp. Dig., Anaheim, CA, USA, May 2010, pp. 1138–1141.

[28] J. Muldavin and G. Rebeiz, “Inline capacitive and DC-contact MEMS shuntswitches,” IEEE Microwave Wireless Compon. Lett., vol. 11, no. 8, pp. 334–336, Aug. 2001.

[29] S. Majumder, J. Lampen, R. Morrison, and J. Maciel, “A packaged, high-lifetime ohmic MEMS RF switch,” in IEEE MTT-S Int. Microw. Symp. Dig.,vol. 3, Philadelphia, PA, USA, June 2003, pp. 1935–1938.

[30] T. Seki, S. Sato, T. Masuda, I. Kimura, and K. Imanaka, “Low-loss RF MEMSmetal-to-metal contact switch with CSP structure,” in Proc. Int. Solid-StateSensors, Actuators and Microsystems Conference (TRANSDUCERS), vol. 1,Boston, MA, USA, June 2003, pp. 340–341.

[31] D. Czaplewski, C. Nordquist, G. Patrizi, G. Kraus, and W. Cowan, “RFMEMS switches with RuO2 − Au contacts cycled to 10 billion cycles,” J.Microelectromech. Syst., vol. 22, no. 3, pp. 655–661, June 2013.

[32] G. Rebeiz, K. Entesari, I. Reines, S.-J. Park, M. El-Tanani, A. Grichener,and A. Brown, “Tuning in to RF MEMS,” IEEE Microwave, vol. 10, no. 6,pp. 55–72, Oct. 2009.

[33] S. Chang and S. Sivoththaman, “A tunable RF MEMS inductor on siliconincorporating an amorphous silicon bimorph in a low-temperature process,”IEEE Electron Device Lett., vol. 27, no. 11, pp. 905–907, Nov. 2006.

[34] J.-I. Kim and D. Peroulis, “Tunable MEMS spiral inductors with optimizedRF performance and integrated large-displacement electrothermal actuators,”IEEE Trans. Microwave Theory Tech., vol. 57, no. 9, pp. 2276–2283, Sept.2009.

[35] G. Rebeiz, G.-L. Tan, and J. S. Hayden, “RF MEMS phase shifters: designand applications,” IEEE Microwave, vol. 3, no. 2, pp. 72–81, Jun. 2002.

[36] G. McFeetors and M. Okoniewski, “Distributed MEMS analog phase shifterwith enhanced tuning,” IEEE Microwave Wireless Compon. Lett., vol. 16,no. 1, pp. 34–36, Jan. 2006.

72 References

[37] K. Van Caekenberghe and T. Vaha-Heikkila, “An analog RF MEMS slot-line True-Time-Delay phase shifter,” IEEE Trans. Microwave Theory Tech.,vol. 56, no. 9, pp. 2151–2159, Sept. 2008.

[38] R. Malmqvist, C. Samuelsson, B. Carlegrim, P. Rantakari, T. Vähä-Heikkilä,A. Rydberg, and J. Varis, “Ka-band RF MEMS phase shifters for energystarved millimetre-wave radar sensors,” in Int. Semiconductor Conference(CAS), vol. 1, Oct. 2010, pp. 261–264.

[39] K. E. V. Sekar, M. Armendariz, “A 1.2–1.6-GHz substrate-integrated-waveguide RF MEMS tunable filter,” IEEE Trans. Microw. Theory Tech.,vol. 59, no. 4, pp. 866–876, Apr. 2011.

[40] G. R. K. Entesari, “A 12–18-GHz three-pole RF MEMS tunable filter,” IEEETrans. Microw. Theory Tech., vol. 53, no. 8, pp. 2566–2571, Aug. 2005.

[41] A. Abbaspour-Tamijani, L. Dussopt, and G. Rebeiz, “Miniature and tunablefilters using MEMS capacitors,” IEEE Trans. Microw. Theory Tech., vol. 51,no. 7, pp. 1878–1885, July 2003.

[42] C.-C. Nguyen, “MEMS technology for timing and frequency control,” IEEETrans. Ultrason., Ferroelect., Freq. Contr., vol. 54, no. 2, pp. 251–270, Feb.2007.

[43] T. Vähä-Heikkilä, J. Varis, J. Tuovinen, and G. M. Rebeiz, “A 20-50 GHzRF MEMS single-stub impedance tuner,” IEEE Microwave Wireless Compon.Lett., vol. 15, no. 4, pp. 205–207, Apr. 2005.

[44] Q. Shen and N. Barker, “Distributed MEMS tunable matching network us-ing minimal-contact RF-MEMS varactors,” IEEE Trans. Microwave TheoryTech., vol. 54, no. 6, pp. 2646–2658, June 2006.

[45] J. Papapolymerou, K. Lange, C. Goldsmith, A. Malczewski, and J. Kleber,“Reconfigurable double-stub tuners using MEMS switches for intelligent RFfront-ends,” IEEE Trans. Microwave Theory Tech., vol. 51, no. 1, pp. 271–278, Jan. 2003.

[46] J.-C. Chiao, Y. Fu, I. M. Chio, M. DeLisio, and L.-Y. Lin, “MEMS reconfig-urable Vee antenna,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., vol. 2,Anaheim, CA, USA, 1999, pp. 1515–1518.

[47] C. won Jung, M.-J. Lee, G. P. Li, and F. De Flaviis, “Reconfigurable scan-beam single-arm spiral antenna integrated with RF-MEMS switches,” IEEETrans. Antennas Propagat., vol. 54, no. 2, pp. 455–463, Feb. 2006.

[48] S. Cheng, P. Rantakari, R. Malmqvist, C. Samuelsson, T. Vaha-Heikkila,A. Rydberg, and J. Varis, “Switched beam antenna based on RF MEMSSPDT switch on quartz substrate,” IEEE Antennas Wireless Propagat. Lett.,vol. 8, pp. 383–386, March 2009.

References 73

[49] H. Jiang, Y. Wang, J.-L. Yeh, and N. Tien, “On-chip spiral inductors sus-pended over deep copper-lined cavities,” IEEE Trans. Microwave TheoryTech., vol. 48, no. 12, pp. 2415–2423, Dec. 2000.

[50] Y. Jun-Bo, C. Yun-Seok, K. Byeong-Il, E. Yunseong, and Y. Euisik, “CMOS-compatible surface-micromachined suspended-spiral inductors for multi-GHzsilicon RF ICs,” IEEE Elect. Dev. Lett., vol. 23, no. 10, pp. 591–593, Oct.2002.

[51] J. Wu and M. Zaghloul, “Robust CMOS micromachined inductors with struc-ture supports for Gilbert mixer matching circuits,” IEEE Trans. Circuits Syst.II: Express Briefs, vol. 56, no. 6, pp. 429–433, June 2009.

[52] I. Llamas-Garro and A. Corona-Chavez, “Micromachined transmission linesfor millimeter-wave applications,” in in Proc. Int. Conf. on Electronics, Com-munications and Computers, Feb. 2006, pp. 15–15.

[53] C. Collins, R. Miles, R. Pollard, D. P. Steenson, J. Digby, G. Parkhurst,J. Chamberlain, N. Cronin, S. Davies, and J. Bowen, “Technique for micro-machining millimetre-wave rectangular waveguide,” IEEE Electronics Letters,vol. 34, no. 10, pp. 996–997, May 1998.

[54] N. Vahabisani and M. Daneshmand, “Monolithic wafer-level rectangularwaveguide and its transition to coplanar waveguide line using a simplified3-D fabrication process,” IEEE Trans. Compon. Packag. Manuf. Technol.,vol. 4, no. 1, pp. 168–176, Jan. 2014.

[55] R. Aigner, “MEMS in RF filter applications: Thin-film bulk acoustic wavetechnology,” Sensors Update, vol. 12, no. 1, pp. 175–210, Feb. 2003.

[56] G. M. Rebeiz and J. B. Muldavin, “RF MEMS switches and switch circuits,”IEEE Microwave, vol. 2, no. 4, pp. 59–71, Dec. 2001.

[57] L. Larson, R. Hackett, M. Melendes, and R. Lohr, “Micromachined mi-crowave actuator (MIMAC) technology-a new tuning approach for microwaveintegrated circuits,” in IEEE Microw. Millimeter-Wave Monolithic CircuitsSymp., Boston, MA, USA, June 1991, pp. 27–30.

[58] J. Rizk, G.-L. Tan, J. B. Muldavin, and G. Rebeiz, “High-isolation W-bandMEMS switches,” IEEE Microwave Wireless Compon. Lett., vol. 11, no. 1,pp. 10–12, 2001.

[59] J. Rizk and G. Rebeiz, “W-band microstrip RF-MEMS switches and phaseshifters,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., vol. 3, Philadelphia,PA, USA, June 2003, pp. 1485–1488.

74 References

[60] D. Mercier, P. Charvet, P. Berruyer, C. Zanchi, L. Lapierre, O. Vendier,J. Cazaux, and P. Blondy, “A DC to 100 GHz high performance ohmic shuntswitch,” in IEEE MTT-S Int. Microw. Symp. Dig., vol. 3, June 2004, pp.1931–1934 Vol.3.

[61] A. Stehle, G. Georgiev, V. Ziegler, B. Schoenlinner, U. Prechtel, H. Seidel, andU. Schmid, “RF-MEMS switch and phase shifter optimized for W-band,” inProc. IEEE European Microw. Conf. (EuMC), Amsterdam, The Netherlands,Oct. 2008, pp. 104–107.

[62] B. Ghodsian, P. Bogdanoff, and D. Hyman, “Wideband DC-contact MEMSseries switch,” IET Micro Nano Letters, vol. 3, no. 3, pp. 66–69, 2008.

[63] T. M. Vu, G. Prigent, J. Ruan, A. Rumeau, P. Pons, and R. Plana, “Fabri-cation and characterization of RF-MEMS switch in V-band,” in Proc. Asia-Pacific Microw. Conf., Singapore, Dec. 2009, pp. 202–205.

[64] P. Rantakari, R. Malmqvist, C. Samuelsson, R. LeBlanc, D. Smith, R. Jons-son, W. Simon, J. Saijets, R. Baggen, and T. Vähä-Heikkilä, “Wide-bandradio frequency micro electro-mechanical systems switches and switchingnetworks using a gallium arsenide monolithic microwave-integrated circuitsfoundry process technology,” IET Microwaves, Antennas Propagation, vol. 5,no. 8, pp. 948–955, June 2011.

[65] T. Vähä-Heikkilä and M. Ylonen, “G-band distributed microelectromechan-ical components based on CMOS compatible fabrication,” IEEE Trans. Mi-crowave Theory Tech., vol. 56, no. 3, pp. 720–728, Mar. 2008.

[66] K. Cao, W. Liu, and J. Talghader, “Curvature compensation in micromirrorswith high-reflectivity optical coatings,” J. Microelectromech. Syst., vol. 10,no. 3, pp. 409–417, Sept. 2001.

[67] A. Tarraf, J. Daleiden, S. Irmer, D. Prasai, and H. Hillmer, “Stress inves-tigation of PECVD dielectric layers for advanced optical MEMS,” IOP J.Micromech. Microeng., vol. 14, no. 3, pp. 317–323, 2004.

[68] K. Petersen, “Silicon as a mechanical material,” Proceedings of the IEEE,vol. 70, no. 5, pp. 420–457, May 1982.

[69] H. V. Jansen, M. J. de Boer, S. Unnikrishnan, M. C. Louwerse, and M. C.Elwenspoek, “Black silicon method X: a review on high speed and selectiveplasma etching of silicon with profile control: an in-depth comparison be-tween Bosch and cryostat DRIE processes as a roadmap to next generationequipment,” IOP J. Micromech. Microeng., vol. 19, no. 3, pp. 033 001–1–033001–41, Feb. 2009.

References 75

[70] I. Schiele and B. Hillerich, “Comparison of lateral and vertical switches forapplication as microrelays,” IOP J. Micromech. Microeng., vol. 9, no. 2, pp.146–150, Jun. 1999.

[71] R. Legtenberg, J. Gilbert, S. Senturia, and M. Elwenspoek, “Electrostaticcurved electrode actuators,” J. Microelectromech. Syst., vol. 6, no. 3, pp.257–265, Sept. 1997.

[72] Y.-H. Jang, Y.-S. Lee, Y.-K. Kim, and J.-M. Kim, “High isolation RF MEMScontact switch in V and W-bands using two directional motions,” ElectronicsLetters, vol. 46, no. 2, pp. 153–155, Jan. 2010.

[73] S. Pacheco, C. Nguyen, and L. Katehi, “Micromechanical electrostatic K-bandswitches,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., vol. 3, Baltimore,MD,USA, June 1998, pp. 1569–1572.

[74] W. C. Tang, T.-C. H. Nguyen, M. W. Judy, and R. T. Howe, “Electrostatic-comb drive of lateral polysilicon resonators,” Sens. Actuators A, Phys.,vol. 21, no. 1–3, pp. 328–331, Feb. 1990.

[75] E. Y. D. Hah and S. Hong, “A low-voltage actuated micromachined mi-crowave switch using torsion springs and leverage,” IEEE Trans. Microw.Theory Tech., vol. 48, no. 12, pp. 2540–2545, Aug. 2000.

[76] D. Peroulis, S. Pacheco, K. Sarabandi, and L. Katehi, “Tunable lumped com-ponents with applications to reconfigurable MEMS filters,” in Proc. IEEEMTT-S Int. Microw. Symp. Dig., vol. 1, Phoenix, AZ, USA, May 2001, pp.341–344.

[77] J.-H. Park, H.-T. Kim, Y. Kim, and Y.-K. Kim, “Tunable millimeter-wavefilters using a coplanar waveguide and micromachined variable capacitors,”IOP J. Micromech. Microeng., vol. 11, no. 6, pp. 706–712, Oct. 2001.

[78] H.-T. Kim, J.-H. Park, Y.-K. Kim, and Y. Kwon, “Low-loss and compact V-band MEMS-based analog tunable bandpass filters,” IEEE Microwave Wire-less Compon. Lett., vol. 12, no. 11, pp. 432–434, Nov. 2002.

[79] E. Fourn, C. Quendo, E. Rius, A. Pothier, P. Blondy, C. Champeaux, J. C.Orlianges, A. Catherinot, G. Tanne, C. Person, and F. Huret, “Bandwidthand central frequency control on tunable bandpass filter by using MEMS can-tilevers,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., vol. 1, Philadelphia,PA, USA, June 2003, pp. 523–526.

[80] C. Lugo, G. Wang, J. Papapolymerou, Z. Zhao, X. Wang, and A. Hunt, “Fre-quency and bandwidth agile millimeter-wave filter using ferroelectric capaci-tors and MEMS cantilevers,” IEEE Trans. Microwave Theory Tech., vol. 55,no. 2, pp. 376–382, Feb. 2007.

76 References

[81] J.-H. Park, S. Lee, J.-M. Kim, H.-T. Kim, Y. Kwon, and Y.-K. Kim, “Recon-figurable millimeter-wave filters using CPW-based periodic structures withnovel multiple-contact MEMS switches,” J. Microelectromech. Syst., vol. 14,no. 3, pp. 456–463, June 2005.

[82] K. Chan, R. Ramer, and Y. Guo, “RF MEMS millimeter-wave switchablebandpass filter,” in IEEE International Wireless Symposium (IWS), Beijing,China, Apr. 2013, pp. 1–4.

[83] J. B. Rizk and G. Rebeiz, “W-band CPW RF MEMS circuits on quartzsubstrates,” IEEE Trans. Microwave Theory Tech., vol. 51, no. 7, pp. 1857–1862, July 2003.

[84] N. Barker and G. Rebeiz, “Optimization of distributed MEMS transmission-line phase shifters-U-band and W-band designs,” IEEE Trans. MicrowaveTheory Tech., vol. 48, no. 11, pp. 1957–1966, Nov. 2000.

[85] B. Lakshminarayanan and T. Weller, “Optimization and implementationof impedance-matched true-time-delay phase shifters on quartz substrate,”IEEE Trans. Microwave Theory Tech., vol. 55, no. 2, pp. 335–342, Feb. 2007.

[86] N. Somjit, G. Stemme, and J. Oberhammer, “Binary-coded 4.25-bit W-bandmonocrystalline-silicon MEMS multistage dielectric-block phase shifters,”IEEE Trans. Microwave Theory Tech., vol. 57, no. 11, pp. 2834–2840, Nov.2009.

[87] B. Lacroix, A. Pothier, A. Crunteanu, C. Cibert, F. Dumas-Bouchiat,C. Champeaux, A. Catherinot, and P. Blondy, “Sub-microsecond RF MEMSswitched capacitors,” IEEE Trans. Microwave Theory Tech., vol. 55, no. 6,pp. 1314–1321, June 2007.

[88] J. Small, M. Arif, A. Fruehling, and D. Peroulis, “A tunable miniaturizedRF MEMS resonator with simultaneous high Q (500-735) and fast responsespeed (<10-60µs),” J. Microelectromech. Syst., vol. 22, no. 2, pp. 395–405,April 2013.

[89] J. DeNatale and R. Mihailovich, “RF MEMS reliability,” in Proc. Int. Solid-State Sensors, Actuators and Microsystems Conference (TRANSDUCERS),vol. 2, Boston, MA, USA, Jun. 2003, pp. 943–946.

[90] C. Goldsmith, J. Maciel, and J. McKillop, “Demonstrating reliability,” IEEEMicrowave, vol. 8, no. 6, pp. 56–60, 2007.

[91] M. Sterner, N. Somjit, U. Shah, S. Dudorov, D. Chicherin, A. Räisänen, andJ. Oberhammer, “Microwave MEMS devices designed for process robustnessand operational reliability,” Int. Journal of Microwave and Wireless Tech-nologies, vol. 3, no. 5, pp. 547–563, Oct. 2011.

References 77

[92] G. Rebeiz, C. Patel, S. Han, C.-H. Ko, and K. Ho, “The search for a reliableMEMS switch,” IEEE Microwave, vol. 14, no. 1, pp. 57–67, 2013.

[93] R. A. Coutu, Jr, P. E. Kladitis, K. D. Leedy, and R. L. Crane, “Selecting metalalloy electric contact materials for MEMS switches,” IOP J. Micromech. Mi-croeng., vol. 14, no. 8, pp. 1157–1164, June 2004.

[94] L. L. Mercado, S.-M. Koo, T.-Y. T. Lee, and L. Liu, “A mechanical approachto overcome RF MEMS switch stiction problem,” in in Proc. IEEE ElectronicComponents and Technology Conf., New Orleans, LA, USA, May 2003, pp.377–384.

[95] I.-J. Cho, T. Song, S.-H. Baek, and E. Yoon, “A low-voltage push-pull SPDTRF MEMS switch operated by combination of electromagnetic actuation andelectrostatic hold,” in Proc. IEEE Int. Conf. Micro Electro Mechanical Sys-tems (MEMS), Miami, Florida, Jan. 2005, pp. 32–35.

[96] J. Oberhammer and G. Stemme, “Active opening force and passive contactforce electrostatic switches for soft metal contact materials,” J. Microelec-tromech. Syst., vol. 15, no. 5, pp. 1235–1242, Oct. 2006.

[97] W. M. van Spengen, “Capacitive RF MEMS switch dielectric charging andreliability: a critical review with recommendations,” IOP J. Micromech. Mi-croeng., vol. 22, no. 7, p. 074001, June 2012.

[98] J. Reid and R. Webster, “Measurements of charging in capacitive microelec-tromechanical switches,” Electronics Letters, vol. 38, no. 24, pp. 1544–1545,Nov. 2002.

[99] R. Herfst, P. Steeneken, and J. Schmitz, “Time and voltage dependence ofdielectric charging in RF MEMS capacitive switches,” in Proc. IEEE Int.Reliability Physics Symposium, Phoenix, AZ, USA, Apr. 2007, pp. 417–421.

[100] J. Webster, C. Dyck, J. Sullivan, T. Friedmann, and A. Carton, “Performanceof amorphous diamond RF MEMS capacitive switch,” Electronics Letters,vol. 40, no. 1, pp. 43–44, Jan. 2004.

[101] X. Rottenberg, S. Brebels, B. Nauwelaers, R. Mertens, W. De Raedt,and H. Tilmans, “Modelling of the RF self-actuation of electrostatic RF-MEMS devices,” in Proc. IEEE Int. Conf. Micro Electro Mechanical Systems(MEMS), Maastricht, The Netherlands, Jan. 2004, pp. 245–248.

[102] R. Malmqvist, R. Jonsson, C. Samuelsson, S. Reyaz, P. Rantakari, A. Ouacha,T. Vaha-Heikkila, J. Varis, and A. Rydberg, “Self-actuation tests of ohmiccontact and capacitive RFMEMS switches for wideband RF power limitercircuits,” in Int. Semiconductor Conference (CAS), vol. 1, Sinaia, Romania,Oct. 2012, pp. 217–220.

78 References

[103] M. Sterner, N. Roxhed, G. Stemme, and J. Oberhammer, “Static zero-power-consumption coplanar waveguide embedded DC-to-RF metal-contact MEMSswitches in two-port and three-port configuration,” IEEE Trans. ElectronDevices, vol. 57, no. 7, pp. 1659–1669, 2010.

[104] U. Shah, M. Sterner, and J. Oberhammer, “Multi-position RF MEMS tunablecapacitors using laterally moving sidewalls of 3-D micromachined transmis-sion lines,” IEEE Trans. Microwave Theory Tech., vol. 61, no. 6, pp. 2340–2352, June 2013.

[105] ——, “High-directivity MEMS-tunable directional couplers for 10–18-GHzbroadband applications,” IEEE Trans. Microwave Theory Tech., vol. 61,no. 9, pp. 3236–3246, Sept. 2013.

[106] ——, “Basic concepts of moving-sidewall tuneable capacitors for RF MEMSreconfigurable filters,” in Proc. IEEE European Microw. Conf. (EuMC),Manchester, UK, Oct. 2011, pp. 1087–1090.

[107] G. K. Celler and S. Cristoloveanu, “Frontiers of silicon-on-insulator,” Journalof Applied Physics, vol. 93, no. 9, pp. 4955–4978, April 2003.

[108] H. Gamble, B. Armstrong, S. J. N. Mitchell, Y. Wu, V. Fusco, and J. Stew-art, “Low-loss CPW lines on surface stabilized high-resistivity silicon,” IEEEMicrowave Guided Wave Lett., vol. 9, no. 10, pp. 395–397, Oct. 1999.

[109] M. Tang, A.-Q. Liu, A. Agarwal, Z.-S. Liu, and C. Lu, “A single-pole double-throw (SPDT) circuit using lateral metal-contact micromachined switches,”Sens. Actuators A: Phys., vol. 121, no. 1, pp. 187–196, Feb. 2005.

[110] D. Yamane, W. Sun, H. Seita, S. Kawasaki, H. Fujita, and H. Toshiyoshi,“A Ku-band dual-SPDT RF-MEMS switch by double-side SOI bulk micro-machining,” J. Microelectromech. Syst., vol. 20, no. 5, pp. 1211–1221, Oct.2011.

[111] E. Hettlage and G. Ruff, “Rotary waveguide switch having arcuate waveguidesrealized by planar faces,” US Patent 4 967 170 A, Oct., 1990.

[112] R. B. Greed, “Waveguide switch,” EP Patent 0 399 739 A2, Nov., 1990.

[113] W. Passaro and J. McManus, “A 35 GHz latching switch,” in Proc. IEEEMTT-S Int. Microw. Symp. Dig., vol. 66, no. 1, Palo Alto, CA, USA, May1966, pp. 270–274.

[114] J. Helszajn and J. Sharp, “Cut-off space of a gyromagnetic planar disk res-onator with a triplet of stubs with up and down magnetisation,” IET Mi-crowaves, Antennas Propagation, vol. 6, no. 5, pp. 569–576, Apr. 2012.

References 79

[115] J. Uher and W. J. R. Hoefer, “Tunable microwave and millimeter-wave band-pass filters,” IEEE Trans. Microwave Theory Tech., vol. 39, no. 4, pp. 643–653, Apr. 1991.

[116] R. Sorrentino, L. Pelliccia, and S. Bastioli, “Recent progress in miniaturizedand reconfigurable filters for advanced communications and space applica-tions,” in Microwaves, Radar and Remote Sensing Symposium (MRRS), Kiev,Ukraine, Aug. 2011, pp. 20–23.

[117] G. Craven and C. K. Mok, “The design of evanescent mode waveguide band-pass filters for a prescribed insertion loss characteristic,” IEEE Trans. Mi-crowave Theory Tech., vol. 19, no. 3, pp. 295–308, Mar. 1971.

[118] H. Joshi, H. H. Sigmarsson, S. Moon, D. Peroulis, and W. Chappell, “High-Qfully reconfigurable tunable bandpass filters,” IEEE Trans. Microwave TheoryTech., vol. 57, no. 12, pp. 3525–3533, Dec. 2009.

[119] S. Moon, H. H. Sigmarsson, H. Joshi, and W. Chappell, “Substrate integratedevanescent-mode cavity filter with a 3.5 to 1 tuning ratio,” IEEE Trans.Microwave Theory Tech., vol. 20, no. 8, pp. 450–452, Aug. 2010.

[120] J. Xu, X.-P. Liang, and K. Shamsaifar, “Full wave analysis and design ofRF tunable filters,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., vol. 3,Phoenix, AZ, USA, May 2001, pp. 1449–1452.

[121] P. S. Carter, “Magnetically-tunable microwave filters using single-crystalyttrium-iron-garnet resonators,” IRE Trans. Microw. Theory Tech., vol. 9,no. 3, pp. 252–260, May 1961.

[122] H. Brand and N. Krause, “Tunable millimetre-wave ferrite bandpass filterwith reduced biasing field,” Electronics Letters, vol. 13, no. 19, pp. 576–578,Sept. 1977.

[123] R. W. DeGrasse, “Low-loss gyromagnetic coupling through single crystal gar-nets,” J. Appl. Physics, vol. 30, pp. 1555–1559, Apr. 1958.

[124] G. L. Matthaei, E. Young, and E. M. T. Jones, Microwave Filters, Impedance-Matching Networks, and Coupling Structures. Norwood, MA: Artech House,1980.

[125] R. Fjerstad, “Some design considerations and realizations of iris-coupled YIG-tuned filters in the 12-40 GHz region,” IEEE Trans. Microwave Theory Tech.,vol. 18, no. 4, pp. 205–212, Apr. 1970.

[126] M. Tsutsumi, H. Hasegawa, and H. Shimasaki, “Waveguide bandpass filter ofmillimetre waves using polycrystalline yttrium iron garnet,” in IEE Proceed-ings Microwaves, Antennas and Propagation, vol. 142, no. 3, Jun. 1995, pp.251–256.

80 References

[127] H. Tanbakuchi, D. Nicholson, B. Kunz, and W. Ishak, “Magnetically tunableoscillators and filters,” IEEE Trans. Magn., vol. 25, no. 5, pp. 3248–3253,Sept. 1989.

[128] J. Higgins, H. Xin, and A. Sailer, “Tunable millimeter wave band pass filterusing electromagnetic crystal sidewalls,” in Proc. IEEE MTT-S Int. Microw.Symp. Dig., vol. 3, June 2004, pp. 1295–1298.

[129] F. Arndt, A. Frye, M. Wellnitz, and R. Wirsing, “Double dielectric-slab-filled waveguide phase shifter,” IEEE Trans. Microwave Theory Tech., vol. 33,no. 5, pp. 373–381, May 1985.

[130] A. Fox, “An adjustable wave-guide phase changer,” Proceedings of the IRE,vol. 35, no. 12, pp. 1489–1498, Dec. 1947.

[131] N. Sultan, “Generalized theory of waveguide differential phase sections andapplication to novel ferrite devices,” IEEE Trans. Microwave Theory Tech.,vol. 19, no. 4, pp. 348–357, Apr. 1971.

[132] G. Pisano, G. Savini, L. Pietranera, K. Isaak, B. Johnson, M. Gervasi, L. Pic-cirillo, B. Maffei, and S. Melhuish, “A 90-GHz waveguide variable phaseshifter,” IEEE Microwave Wireless Compon. Lett., vol. 17, no. 3, pp. 208–210,Mar. 2007.

[133] M. Abou-Khousa, M. Baumgartner, S. Kharkovsky, and R. Zoughi, “Ka-band vector reflectometer based on simple phase-shifter design,” IEEE Trans.Instrum. Meas., vol. 60, no. 2, pp. 618–624, Feb. 2011.

[134] W. E. Hord, “Microwave and millimeter-wave ferrite phase shifters,” Mi-crowave J., vol. 32, pp. 81–94, 1989.

[135] D. Parker and D. Zimmermann, “Phased arrays-part II: implementations, ap-plications, and future trends,” IEEE Trans. Microwave Theory Tech., vol. 50,no. 3, pp. 688–698, Mar. 2002.

[136] J. Adam, L. Davis, G. F. Dionne, E. Schloemann, and S. Stitzer, “Ferritedevices and materials,” IEEE Trans. Microwave Theory Tech., vol. 50, no. 3,pp. 721–737, Mar. 2002.

[137] L. Whicker, “The evolution of ferrite and other passive control components: Apersonal account,” in Proc. IEEE MTT-S Int. Microw. Symp. Dig., Montreal,QC, Canada, June 2012, pp. 1–3.

[138] V. G. Harris, “Modern microwave ferrites,” IEEE Trans. Magn., vol. 48, no. 3,pp. 1075–1104, Mar. 2012.

[139] C. Boyd, “A 60 GHz dual-mode ferrite phase shifter,” in Proc. IEEE MTT-SInt. Microw. Symp. Dig., Dallas, TX, USA, 1982, pp. 257–259.

References 81

[140] J. Higgins, H. Xin, A. Sailer, and M. Rosker, “Ka-band waveguide phaseshifter using tunable electromagnetic crystal sidewalls,” IEEE Trans. Mi-crowave Theory Tech., vol. 51, no. 4, pp. 1281–1288, Apr. 2003.

[141] M. Popov, I. Zavislyak, and G. Srinivasan, “Magnetic field tunable 75-110GHz dielectric phase shifter,” Electronics Letters, vol. 46, no. 8, pp. 569–570,Apr. 2010.

[142] M. Daneshmand, R. Mansour, and N. Sarkar, “RF MEMS waveguide switch,”IEEE Trans. Microwave Theory Tech., vol. 52, no. 12, pp. 2651–2657, Dec.2004.

[143] M. Daneshmand and R. Mansour, “Multiport MEMS-based waveguide andcoaxial switches,” IEEE Trans. Microwave Theory Tech., vol. 53, no. 11, pp.3531–3537, Nov. 2005.

[144] H. Joshi, H. H. Sigmarsson, D. Peroulis, and W. Chappell, “Highly loadedevanescent cavities for widely tunable high-Q filters,” in Proc. IEEE MTT-SInt. Microw. Symp. Dig., Honolulu, HI, USA, June 2007, pp. 2133–2136.

[145] S.-J. Park, I. Reines, C. Patel, and G. Rebeiz, “High-Q RF-MEMS 4-6-GHz tunable evanescent-mode cavity filter,” IEEE Trans. Microwave TheoryTech., vol. 58, no. 2, pp. 381–389, Feb. 2010.

[146] X. Liu, L. Katehi, W. Chappell, and D. Peroulis, “High-Q tunable microwavecavity resonators and filters using SOI-based RF MEMS tuners,” J. Micro-electromech. Syst., vol. 19, no. 4, pp. 774–784, Aug. 2010.

[147] R. Stefanini, J. Martinez, M. Chatras, A. Pothier, V. Boria, and P. Blondy,“Ku band high-Q tunable surface-mounted cavity resonator using RF MEMSvaractors,” IEEE Microwave Wireless Compon. Lett., vol. 21, no. 5, pp. 237–239, May 2011.

[148] W. Yan and R. Mansour, “Micromachined millimeter-wave ridge waveguidefilter with embedded MEMS tuning elements,” in Proc. IEEE MTT-S Int.Microw. Symp. Dig., San Francisco, CA, USA, June 2006, pp. 1290–1293.

[149] M. S. Arif and D. Peroulis, “A 6 to 24 GHz continuously tunable, microfabri-cated, high-Q cavity resonator with electrostatic MEMS actuation,” in Proc.IEEE MTT-S Int. Microw. Symp. Dig., June 2012, pp. 1–3.

[150] D. Psychogiou, Y. Li, J. Hesselbarth, D. Peroulis, C. Hierold, and C. Hafner,“Continuously variable W-band phase shifters based on MEMS-actuated con-ductive fingers,” Int. J. Microw. Wirel. Technol., vol. 5, no. 4, pp. 477–489,Apr. 2013.

82 References

[151] D. Chicherin, M. Sterner, J. Oberhammer, S. Dudorov, J. Åberg, andA. Raisanen, “Analog type millimeter wave phase shifters based on MEMStunable high-impedance surface in rectangular metal waveguide,” in Proc.IEEE MTT-S Int. Microw. Symp. Dig., Anaheim, CA, USA, May 2010, pp.61–64.

[152] F. Laermer and A. Schilp, “Method of anisotropic etching of silicon,” USPatent 6 531 068, Mar., 2003.

[153] R. E. Collin, Foundation for Microwave Engineering, 2nd ed. New York:McGraw-Hill, 1992.

[154] Z. Baghchehsaraei and J. Oberhammer, “Parameter analysis of millimeter-wave waveguide switch based on a MEMS-reconfigurable surface,” IEEETrans. Microw. Theory and Techniques, vol. 61, no. 12, pp. 4396–4406, Dec.2013.

[155] A. J. Fenn, D. H. Temme, W. P. Delaney, and W. E. Courtney, “The devel-opment of phased-array radar technology,” Lincoln Lab. J., vol. 12, no. 2, pp.321–340, 2000.

[156] I. Gresham, N. Jain, T. Budka, A. Alexanian, N. Kinayman, B. Ziegner,S. Brown, and P. Staecker, “A compact manufacturable 76-77-GHz radarmodule for commercial ACC applications,” IEEE Trans. Microwave TheoryTech., vol. 49, no. 1, pp. 44–58, Jan. 2001.

[157] M. Schneider, “Automotive radar – status and trends,” in Proc. GermanMicrowave Conf., Apr. 2005, pp. 144–147.

[158] A. Natarajan, A. Komijani, X. Guan, A. Babakhani, and A. Hajimiri, “A 77-GHz phased-array transceiver with on-chip antennas in silicon: Transmitterand local LO-path phase shifting,” IEEE J. Solid-State Circuits, vol. 41,no. 12, pp. 2807–2819, Dec. 2006.

[159] A. Artemenko, A. Mozharovskiy, A. Maltsev, R. Maslennikov, A. Sev-astyanov, and V. Ssorin, “Experimental characterization of E-band two-dimensional electronically beam-steerable integrated lens antennas,” IEEEAntennas Wireless Propagat. Lett., vol. 12, pp. 1188–1191, Sept. 2013.

[160] E. Topak, J. Hasch, C. Wagner, and T. Zwick, “A novel millimeter-wave dual-fed phased array for beam steering,” IEEE Trans. Microwave Theory Tech.,vol. 61, no. 8, pp. 3140–3147, Aug. 2013.

[161] J. Schoebel, T. Buck, M. Reimann, M. Ulm, M. Schneider, A. Jourdain,G. Carchon, and H. Tilmans, “Design considerations and technology assess-ment of phased-array antenna systems with RF MEMS for automotive radarapplications,” IEEE Trans. Microwave Theory Tech., vol. 53, no. 6, pp. 1968–1975, June 2005.

References 83

[162] K. Van Caekenberghe, “RF MEMS on the radar,” IEEE Microwave, vol. 10,no. 6, pp. 99–116, Oct. 2009.

[163] M. Sterner, D. Chicherin, A. Räisänen, G. Stemme, and J. Oberham-mer, “RF MEMS high-impedance tuneable metamaterials for millimeter-wavebeam steering,” in Proc. IEEE Int. Conf. Micro Electro Mechanical Systems(MEMS), Sorrento, Italy, Jan. 2009, pp. 896–899.

[164] C.-C. Cheng, B. Lakshminarayanan, and A. Abbaspour-Tamijani, “Aprogrammable lens-array antenna with monolithically integrated MEMSswitches,” IEEE Trans. Microwave Theory Tech., vol. 57, no. 8, pp. 1874–1884, Aug. 2009.

[165] F. Arndt, J. Bornemann, D. Grauerholz, and R. Vahldieck, “Theory anddesign of low-insertion loss fin-line filters,” IEEE Trans. Microwave TheoryTech., vol. 30, no. 2, pp. 155–163, Feb. 1982.

[166] M. N. Afsar and H. Chi, “Millimeter wave complex refractive index, complexdielectric permittivity and loss tangent of extra high purity and compensatedsilicon,” Int. J. Infrared Millim. Waves, vol. 15, no. 7, pp. 1181–1188, Jul.1994.

[167] L. Pelliccia, S. Bastioli, F. Casini, and R. Sorrentino, “High-Q MEMS-reconfigurable waveguide filters,” in Proc. IEEE European Microw. Conf.(EuMC), Paris, France, Sept. 2010, pp. 1126–1129.

[168] L. Pelliccia, P. Farinelli, and R. Sorrentino, “MEMS-based high-Q recon-figurable E-plane filters,” in Proc. IEEE European Microw. Conf. (EuMC),Yokohama, Japan, Oct. 2011, pp. 369–372.

[169] A. W. Martinez, S. T. Phillips, and G. M. Whitesides, “Three-dimensionalmicrofluidic devices fabricated in layered paper and tape,” Proc. Natl. Acad.Sci. USA, vol. 105, no. 50, pp. 19 606–19 611, Dec. 2008.

[170] T.-Q. Truong and N.-T. Nguyen, “A polymeric piezoelectric micropump basedon lamination technology,” Journal of Micromechanics and Microengineering,vol. 14, no. 4, pp. 632–638, Feb. 2004.

[171] L. Xie, S. C. Chong, C. S. Premachandran, M. Chew, and U. Raghavan,“Development of an integrated bio-microfluidic package with micro-valvesand reservoirs for a DNA lab on a chip (LOC) application,” in in Proc. 56thElectron. Compon.Technol. Conf., San Diego, CA, USA, May 2006, pp. 693–698.

[172] T. Frisk, N. Sandström, L. Eng, W. van der Wijngaart, P. Månsson, andG. Stemme, “An integrated QCM-based narcotics sensing microsystem,” LabChip, vol. 8, pp. 1648–1657, 2008.

84 References

[173] Y. A. Chee, A. Bakir, and D. Wicaksono, “Proprioceptive sensing system fortherapy assessment using textile-based biomedical micro electro mechanicalsystem (MEMS),” in IEEE Sensors, Taipei, Taiwan, Oct. 2012, pp. 1–4.

[174] H. Rajagopalan, Y. Rahmat-Samii, and W. Imbriale, “RF MEMS actuatedreconfigurable reflectarray patch-slot element,” IEEE Trans. Antennas Prop-agat., vol. 56, no. 12, pp. 3689–3699, Dec. 2008.

Paper Reprints

85

Documents

Waveguide-Integrated MEMS Concepts for Tunable Millimeter ...705226/FULLTEXT01.pdf · Waveguide-Integrated MEMS Concepts for Tunable Millimeter-Wave Systems ZARGHAM BAGHCHEHSARAEI