High Performance Multilayer Substrate Integrated Waveguide (SIW) Technics for Low-Cost Millimeter-Wave Circuits Frederic Parment

Home , Wave impedance

High performance multilayer Substrate Integrated Waveguide (SIW) technics for low-cost millimeter-wave circuits Frederic Parment

To cite this version:

Frederic Parment. High performance multilayer Substrate Integrated Waveguide (SIW) technics for low-cost millimeter-wave circuits. Optics / Photonic. Université Grenoble Alpes, 2016. English. NNT : 2016GREAT077. tel-01505416

HAL Id: tel-01505416 https://tel.archives-ouvertes.fr/tel-01505416 Submitted on 11 Apr 2017

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés.

THÈSE Pour obtenir le grade de DOCTEUR DE LA COMMUNAUTE UNIVERSITE GRENOBLE ALPES Spécialité : OPTIQUE ET RADIOFREQUENCES

Arrêté ministériel : 25 Mai 2016

Présentée par Frédéric PARMENT

Thèse dirigée par Pr Tan Phu VUONG, Codirigée par Pr Ke WU Co-encadrée par M Anthony GHIOTTO Co-encadrée par M Jean-Marc DUCHAMP préparée au sein du Laboratoire IMEP-LaHC de l’Institut National Polytechnique de Grenoble, dans l'École Doctorale Electronique, Electrotechnique, Automatique, et Traitement du Signal.

Guides d’onde Intégrés au Substrat (SIW) multicouches à haute performance pour des circuits millimétriques à faible coût

Thèse soutenue publiquement le 04 Novembre 2016, devant le jury composé de : Monsieur Eric RIUS Professeur à l’Université de Bretagne Occidentale, Président Monsieur Hervé AUBERT Professeur à l’Institut National Polytechnique de Toulouse, Rapporteur Monsieur Jean-Louis CAZAUX Responsable Innovation à Thales Alenia Space de Toulouse, Rapporteur Monsieur Maurizio BOZZI Professeur à l’Université de Pavie, Membre Monsieur Ke WU Professeur à l’Ecole Polytechnique de Montréal, Membre Monsieur Anthony GHIOTTO Maître de conférences à l’Institut Polytechnique de Bordeaux, Membre Monsieur Jean-Marc DUCHAMP Maître de conférences à l’Université Joseph Fourier, Membre Monsieur Tan Phu VUONG Professeur à l’Institut Polytechnique de Grenoble, Membre Monsieur Ludovic CARPENTIER Ingénieur de recherche au CNES de Toulouse, Membre Invité Monsieur Pierre MONTEIL Responsable Business Unit à COBHAM de Bordeaux, Membre Invité Monsieur Laurent PETIT Responsable Projet Recherche et Technologie à Radiall Voreppe, Membre Invité

La théorie, c'est quand on sait tout et que rien ne fonctionne. La pratique, c'est quand tout fonctionne et que personne ne sait pourquoi. Ici, nous avons réuni théorie et pratique : Rien ne fonctionne... et personne ne sait pourquoi !

Albert Einstein

A ma famille et mes proches…

ACKNOWLEDGEMENT

I would like to express my sincere gratitude to my supervisor, Pr Tan Phu VUONG, and my co-supervisors Pr Ke WU, Mr Anthony GHIOTTO and Mr Jean-Marc DUCHAMP for their help, guidelines and support, that contributed to the excellent research work presented in this thesis manuscript. I would like to particularly thank Anthony for his unwavering support, valuable advices, and unfailing commitment in my research project and beyond.

I would like to also express my thanks to the engineers and technicians that offered a strong support, together with valuable assistance and advices, during my experimental validations, including Mr Nicolas CORRAO from the IMEP-LaHC research center, Ms

Magali DE MATOS and Ms Isabelle FAVRE from the IMS research center, and Mr Traian

ANTONESCU, Mr Jules GAUTHIER, Mr Steeve DUBE, Mr Jean-Sebastien DECARIE and

Mr Maxime THIBAULT from the POLY-GRAMES research center.

I would like to express my recognition to the administrative staff, inluding Isabelle,

Joëlle, Valérie, Annaïck, Brigitte, Dahlila, Fabienne from Grenoble, and Rachel from the

POLY-GRAMES for their support. They have helped me a lot during my PhD on the different admisnitrative aspects.

I would like to thank my colleges and friends, including Zyad, Matthieu, Aline, Hamza,

Ayssar, Isaac, Vipin, Furat, An-Thu, Vladimir, Vlad, and others from the IMEP-LaHC, Jaber and Kuangda from the POLY-GRAMES, and Tifenn, Alexandre, and others from the IMS. I would also like to thank my supportive friends, Alexis and Alexandre.

Finally, I would like to express a deep sense of gratitude to my parents, Michèle and

Jean-Michel, and my brother, Nicolas. You helped me all along my studies. You have supported me so that I could give my best. The last but not least, I would like to take the opportunity to express my love to my girlfriend, Lucile, who supported me the most, during the good and the bad days.

TABLE OF CONTENTS

ACKNOWLEDGEMENT 5 TABLE OF CONTENTS 7 LIST OF ACRONYMS AND ABBREVIATIONS 9 SYNTHESE EN FRANÇAIS 11 CHAPTER I: INTRODUCTION 19 CHAPTER II: STATE-OF-THE-ART ON ELECTROMAGNETIC TRANSMISSION LINES 25 CHAPTER III: THE AIR-FILLED SIW TECHNOLOGY 65 CHAPTER IV: AIR-FILLED SIW FEEDING NETWORK ELEMENTS 101 CHAPTER V: AIR-FILLED SIW FILTERS AND ANTENNAS 147 CHAPTER VI: CONCLUSION AND PERSPECTIVES 183 ANNEX 1: COMPLETE ELLIPTIC INTEGRAL OF THE FIRST KIND 187 ANNEX 2: CALCULATION TO OBTAIN EQUATION (1.2) OF THE CHAPTER III 189 LIST OF PUBLICATIONS, AWARDS AND PUBLICATIONS 191 ABSTRACT 195 RESUME 195

LIST OF ACRONYMS AND ABBREVIATIONS

AFSIW Air-Filled Substrate Integrated Waveguide ALTSA Antipodal Linear Tapered Slot Antenna APHC Average Power Handling Capability (W or dBm) 퐵⃗ Magnetic induction (T) BW Bandwidth (GHz) c Light velocity (m/s) CBCPW Conductor-Backed Coplanar Waveguide CMOS Complementary Metal Oxide Semi-conductor CPW Coplanar Waveguide 퐷⃗⃗ Electric induction (C/m²) DFSIW Dielectric-Filled Substrate Integrated Waveguide DUT Device Under Test 퐸⃗ Electric field (V/m) E0 Dielectric strength (MV/m) EBG Electromagnetic BandGap EM Electromagnetic f Frequency (Hz) fc Cut-off frequency (Hz) FR-4 Flame Resistant 4 h Height of the structure (m) 퐻⃗⃗ Magnetic field (A/m) HPBW Half Power Beamwidth (°) IoT Internet Of Thing 푗 Current density (A/m2) k Wave number (rad/m) k0 Free-space Wave number (rad/m) Kc Cut-off wave number (rad/m) LTCC Low Temperature Co-fired Ceramics MHMIC Miniature Hybrid Microwave Integrated Circuit PCB Printed Circuit Board PIM Passive InterModulation PPHC Peak Power Handling Capability (W or dBm) q Electronic charge (C) RADAR RAdio Detection And Ranging RMS Root Mean Square RWG Rectangular Waveguide SIC Substrate Integrated Circuit SINRD Substrate Integrated Non-Radiating Dielectric Waveguide SIS Substrate Integrated System SISW Substrate Integrated Slab Waveguide SIW Substrate Integrated Waveguide tanδ Loss tangent TE Transverse Electric TEM Tranverse Electric Magnetic

TM Transverse Magnetic TRL Thru-Reflect-Line VNA Vector Network Analyzer W Width of the structure (m) WR Waveguide, Rectangular αg Losses constant (Np/m) βg Phase constant (rad/m) 훤 Propagation constant (m-1) δ Skin depth (m) ε Permittivity (F/m) ε0 Vacuum permittivity (F/m) εr Relative permittivity η Wave impedance (Ω) λ Wavelength (m) μ Permeability (H/m) μ0 Vacuum permeability (H/m) μr Relative permeability ρ Volume density charge (C/m3) σ Conductivity (S/m) ω Angular frequency (rad/s)

SYNTHESE EN FRANÇAIS

L’émergence de nouvelles applications qui requièrent le développement de nouvelles technologies opérant aux fréquences millimétriques et au-delà, entre 30 et 300 GHz, est principalement due à l’augmentation ininterrompue des données et à l’accroissement insatiable de la demande en bande-passante. Le manque de canal libre dans les bandes radiofréquences et micro-ondes y contribue et force l’industrie à développer de nouveaux systèmes opérant en plus haute fréquence. Parmi ces nouvelles applications, l’Internet-des-

Objets, les radars pour des systèmes types avioniques ou automobiles, les constellations de pico- et nanosatellites [1], l’imagerie médicale et sécuritaire, etc… sont un des nombreux marchés ciblés [2]. Les ondes millimétriques peuvent fournir beaucoup d’avantages pour ces futurs systèmes parmi lesquels de larges bandes passantes, une pénétration accrue dans la fumée ou le brouillard, un vaste spectre peu utilisé et bien d’autres.

Un autre point qui est tout aussi important pour l’industrie est la densité d’intégration.

En effet, une haute densité d’intégration permet de réduire les coûts par la production de masse, avec un emballage petit et léger facilitant l’adoption des produits. C’est dans le début des années 1950 que la première ligne de transmission planaire, appelée microruban, a été développée. Aujourd’hui, cette même ligne est devenue la ligne de transmission la plus populaire et utilisée jusqu’à 30 GHz. Elle combine une structure simple, bien décrite théoriquement avec des modèles implémentés dans la plupart des logiciels de conception, et elle peut être produite à prix bon marché par l’utilisation de procédés de fabrication issus des circuits imprimés. Cependant, au-delà de 30 GHz, les lignes de transmission en guides d’onde coplanaires sont préférées aux lignes microrubans, car à ces fréquences, ces dernières ont besoin de substrat fin pour éviter l’excitation de mode d’ordre supérieur, pour éviter aussi de subir des courants de fuite, qui augmentent les pertes par radiation, et elles deviennent aussi plus difficiles à fabriquer du fait dela réduction de la longueur d’onde qui requièrent une

tolérance de fabrication accrue, notamment sur les vias par exemple. Néanmoins, la contribution sur les performances induites par la rugosité de surfaces et celle de l’effet de peaux de la ligne conductrice augmente avec la fréquence. Par conséquent, ces circonstances, prises dans leur ensemble, génèrent une haute densité de courants s’écoulant au travers d’une fine région sous la surface métallique du conducteur, et les pertes s’en retrouvent d’autant augmentées [3]-[4]. Ce problème devient bien plus important dès lors que la forme des bords de la ligne conductrice n’est plus rectangulaire. Bien que le récent développement de substrat faibles pertes ait largement réduit les pertes, celles qui sont liées au conducteur deviennent prédominantes aux fréquences millimétriques et au-delà. Enfin, la tenue en puissance pour ce genre de lignes de transmission est un problème pour les applications à forte puissance. Pour toutes ces raisons, la technologie en guide d’onde rectangulaire est encore très utilisée en haute fréquence parce qu’elle surpasse en performance toutes les lignes standards. Très coûteux à produire, lourds et difficiles à interconnecter ou à intégrer avec d’autres technologies, les inconvénients du guide d’onde rectangulaire restent un challenge à résoudre encore aujourd’hui.

Pour combler le fossé entre le guide d’onde rectangulaire et les lignes de transmission planaires, une nouvelle génération de circuits planaires a vu le jour au début des années 2000, c’est-à-dire les Circuits Intégrés au Substrat (CIS) [5]. Cette nouvelle famille de lignes de transmission a reçu une attention particulière depuis plus de dix ans. Les CISs combinent les avantages des technologies sur circuits imprimés avec celle du guide d’onde : un poids léger, une structure compacte, blindée par nature, bon marché et facile a interconnecté avec des technologies traditionnelles. Cette technologie peut se décliner en plusieurs topologies : le

Guide d’onde Intégré au Substrat (SIW) [6] en Fig. 1(a), le Guide d’onde Intégré au Substrat à

Barreau Diélectrique (SISW) [7] en Fig. 1(b), et le Guide d’onde Diélectrique Intégré au

Substrat Non-Radiatif (SINRD) [8] en Fig. 1(c) comme exemples non exhaustifs. Beaucoup

d’entre eux ont été réalisés en utilisant différentes techniques de fabrication comme les circuits imprimés (PCB), les circuits micro-ondes intégrés hybrides miniatures (MHMIC), les céramiques cocuites à basse température (LTCC) ou bien les technologies sur silicium comme le CMOS [9].

Via métallisé

Trou d’air

Plan de masse du dessus Substrat Substrat diélectrique diélectrique (a) (b) Plaque de métal du dessus

Trou d’air

Plaque de métal du dessous

Le SIW est devenu le plus populaire et la topologie la plus utilisée de la famille des

CISs. Il hérite principalement ses propriétés du guide d’onde rectangulaire, et il est facile à fabriquer, car il n’a besoin que de deux rangés de vias métalisés. C’est pourquoi il est aussi facile à concevoir qu’à fabriquer. Du PCB au CMOS en passant par le LTCC, le SIW est compatible avec beaucoup de processus de fabrication. De quelques gigahertz à une centaine de gigahertz, son spectre d’application est large et varié : applications spatiales, radar automobile, ou application de sûreté comme les scanners et ainsi de suite. Le SIW bénéficie d’un grand potentiel. Néanmoins, le SIW est une ligne de transmission utilisant du substrat et souffre des pertes induites par le diélectrique, qui augmentent à haute fréquence, et diminue sa tenue en puissance. Cette limitation réduit grandement la performance d’une telle technologie, et donc son champ d’application.

Figure 2 : Comparaison des propriétés entre une ligne de transmission standard en circuit imprimé, un guide d’onde rectangulaire, un SIW et un SIW creux.

Après une brève introduction dans le Chapitre I, le contexte actuel de cette thèse est bien plus détaillé dans le Chapitre II. Cette partie présente de façon chronologique les différentes lignes de transmission électromagnétiques en partant des équations de Maxwell, des guides d’onde métalliques rectangulaires jusqu’aux lignes planaires et les structures dites

SIW. Une comparaison entre ces lignes en termes de performance mais aussi sur les procédés de fabrication est faite, ceci afin de mieux comprendre quelle est le positionememnt de la technologie SIW creux au milieu des technologies existantes.

Ainsi, dans le Chapitre III, cette nouvelle topologie de CIS est présentée dont le nom est

SIW creux. Elle consiste en une ligne de transmission réalisée grâce aux techniques des circuits imprimés multicouches avec une partie à l’intérieur en air faite par un micro-usinage laser. L’objectif d’une telle structure est de diminuer drastiquement la contribution des pertes diélectriques en remplaçant le substrat par de l’air. Les rangés de vias sont maintenues afin d’obtenir une structure complète SIW, avec aucune propagation du mode TM. La structure finale est présentée au Chapitre III avec une analyse théorique des pertes et de la tenue en puissance qui est comparée aux résultats de simulation et expérimentaux. Une comparaison des performances est donnée à la Fig. 2 entre le guide d’onde rectangulaire, le SIW, une ligne

planaire standard en PCB et le SIW creux. Il est à noter qu’une réduction de 75% des pertes comparé à une ligne SIW utilisant le même substrat a été observée lors des mesures des niveaux de pertes. Ceci a conduit à une augmentation de la tenue en puissance moyenne.

Enfin, pour interconnecter cette technologie avec celle qui est en SIW, une transition a été conçue et mesurée ainsi que des topologies de coudes. Cette transition a été conçue de façon large bande afin de pouvoir l’utiliser pour interconnecter n’importe quel composant en SIW creux avec le SIW. Pour des raisons liées à la mesure, un kit de calibration Thru-Reflect-Line a été créé en SIW creux pour dé-embeder toutes transitions se trouvant entre le dispositif à mesurer et l’analyseur de réseau.

Ensuite, plusieurs circuits passifs ont été conçus et implémentés en utilisant le SIW creux. Dans le Chapitre IV, des éléments basiques souvent utilisés dans des réseaux d’alimentation sont démontrés. Un diviseur de puissance utilisant une jonction en T et un coupleur de Riblet ont été réalisé. Une diminution des pertes a été observée. Deux coupleurs directionnels, un de type Moreno et un autre de type deux trous sont détaillés. Enfin, un déphaseur à double paroi diélectrique en SIW creux est présenté avec une analyse théorique, suivie par des résultats de simulations et de mesures. Basé sur ce composant, un déphaseur large bande compensé en SIW creux est décrit et mesuré. La compensation a été réalisée en utilisant la variation croissante de la phase avec la fréquence pour le déphaseur à double paroi diélectrique et la variation décroissante de la phase avec la fréquence pour un SIW creux

élargie. Ainsi, les deux variations se companse l’une l’autre. Un déphasage de 43° ± 6° est démontré sur toute la bande Ka.

Pour finir, des filtres et antennes basés sur le SIW creux sont présentés. Dans la première partie du chapitre V, deux filtres passe-bande ont été réalisés en SIW creux. D’une part, un filtre d’ordre trois utilisant des ponts inductifs a été conçu et comparé à sa version en

SIW. Une amélioration du facteur de qualité d’un facteur 4 environ dans le cas du dispositif

en SIW creux est à noter. D’autre part, un filtre d’ordre quatre multicouche à cavités verticalement empilées est présenté. C’est le premier dispositif dans cette technologie utilisant pleinement la possiblité de monter verticalement et donc de réduire l’empreinte d’un composant passif tel qu’un filtre. Dans la seconde partie de ce chapitre, deux antennes en SIW creux sont testées. D’abord, une antenne à fentes en SIW creux avec 4 ouvertures rectangulaires est démontrée et comparée avec sa version en SIW. Ensuite, une antenne de type ALTSA est créée en SIW creux en utilisant le processus multicouche des circuits imprimés. Les performances de cette antenne sont également comparées à sa version SIW.

Dans les deux cas, une augmentation du gain est observé ainsi qu’une amélioration de l’efficacité de radiation.

Finalement, le Chapitre VI conclut ce manuscrit. Le travail de recherche présenté dans ce manuscrit de thèse a fait l’objet de nombreuses publications incluant 5 revues internationales (4 publiées et 1 soumise), 6 conférences internationales, 6 conférences nationales et 4 communications.

References:

[1] P. Christopher, "Millimeter wave advantages for satellite communications," Military Communications Conference, 1983. MILCOM 1983. IEEE, Washington, DC, USA, pp. 555-562.C, 2010. [2] D. R. Vizard, “Millimeter-wave applications: from satellite communications to security systems,” Microwave Journal, Euro-Global Edition, vol. 49, no. 7, pp. 22–36, Jul. 2006. [3] K. Wu, "Towards the development of terahertz substrate integrated circuit technology," Silicon Monolithic Integrated Circuits in RF Systems (SiRF), 2010 Topical Meeting on, New Orleans, LA, pp. 116-119, 2010. [4] M. Bozzi, A. Georgiadis and K. Wu, "Review of substrate-integrated waveguide circuits and antennas," in IET Microwaves, Antennas & Propagation, vol. 5, no. 8, pp. 909-920, June 6, 2011. [5] Ke Wu, D. Deslandes and Y. Cassivi, "The substrate integrated circuits - a new concept for high-frequency electronics and optoelectronics," Telecommunications in Modern Satellite, Cable and Broadcasting Service, 2003. TELSIKS 2003. 6th International Conference on, pp. P-III-P-X vol.1, 2003. [6] M. Bozzi, F. Xu, D. Deslandes and K. Wu, "Modeling and design considerations for substrate integrated waveguide circuits and components," Telecommunications in Modern Satellite, Cable and Broadcasting Services, 2007. TELSIKS 2007. 8th International Conference on, Nis, pp. P-VII-P-XVI, 2007. [7] D. Deslandes, M. Bozzi, P. Arcioni and Ke Wu, "Substrate integrated slab waveguide (SISW) for wideband microwave applications," Microwave Symposium Digest, 2003 IEEE MTT-S International, Philadelphia, PA, USA, vol.2, pp. 1103-1106, 2003. [8] P. Mondal and K. Wu, "A leaky-wave antenna in substrate integrated non-radiative dielectric (SINRD) waveguide with controllable scanning rate," in IEEE Transactions on Antennas and Propagation, vol. 61, no. 4, pp. 2294-2297, April 2013. [9] K. Wu, "Substrate integrated circuits (SICs) for GHz and THz electronics and photonics: Current status and future outlook," German Microwave Conference Digest of Papers, Berlin, pp. 292-295, 2010.

CHAPTER I: INTRODUCTION

The emerging applications require the development of new technologies especially at millimeter-waves and beyond (30 GHz to 300 GHz). This is mainly due to the increase of data and hence the growth of bandwidth. The lack of free channel in the radiofrequency and microwave bands forces the industries to develop new systems at higher frequency. Among these new applications, Internet-Of-Things, automotive and aircraft radar systems, pico- and nano-satellite constellations [1] and communications, medical and safety imaging like scanners are one of the targeted markets as described in [2]. Millimeter-waves can provide various advantages for those future systems among which broad bandwidth, the penetration through smog or fog, the large unused frequency spectrum and so on.

Another point which is critical for industries is the density of integration. A high- density reduces cost, allows mass-producible, small and lightweight packaging for commercial products. This is during the beginning of the 1950’s that the first planar transmission line, called microstrip, was developed. Nowadays, microstrip is the most used and popular transmission line up to 30 GHz. It combines a simple structure, theoretically well-analyzed with models implemented into software and it can be produced at a low-cost owing to the Printed Circuit Board manufacturing process. However, beyond 30 GHz,

Coplanar Waveguide (CPW) transmission lines are preferred because microstrip lines need thin substrate to avoid higher order modes excitation, undergo leakage currents that raise the radiation loss and, are more difficult to fabricate due to the reduction of the wavelength that requires high precision fabrication tolerances, like vias for example. Nevertheless, the contribution of the surface roughness and the skin effect on the performance of the stripline rises with the frequency. As consequences, both circumstances generate high current density flowing through a thin region below the metallic surface of the conductor and the losses are significantly raised [3]-[4]. This issue is even worse for planar transmission lines in general when the shape of the conductor strip edges is not rectangular. Even if the recent development

of low-loss dielectric has largely reduced the losses, the conductor losses are predominant in millimeter-waves and beyond. Finally, the power handling is also an issue for high power applications. For these reasons, the Rectangular Waveguide (RWG) technology is still preferred at high frequency because it surpasses performances from all standard transmission lines. However, such structure is too expensive, heavy and difficult to integrate or interconnect.

(a) (b)

To bridge the gap between RWG and planar transmission line, a new generation of planar circuits have been introduced during the beginning 2000’s, namely the Substrate

Integrated Circuits (SIC) [5]. This new family of transmission line has received a particular attention since the past decade. SICs combine advantages from both previous technologies: a light weight, a compact and low-cost structure that is easy to interconnect with traditional technologies and is self-shielded. SICs can be declined in several topologies: Substrate

Integrated Waveguide (SIW) [6] Fig. 1(a), Substrate Integrated Slab Waveguide (SISW) [7]

Fig. 1(b), and Substrate Integrated Non-Radiating Dielectric Waveguide (SINRD) [8]

Fig. 1(c) as a list of non-exhaustive examples. Many of them have been realized using

different fabrication techniques like PCB, MHMIC, LTCC or CMOS [9]. SIW became the most popular and used topology from the SICs family. It inherits largely its properties from the RWG, and it is easy to fabricate as it requires mainly two rows of matalized via-holes. In addition, SIW is compatible with different manufacturing techniques including PCB, CMOS,

LTCC, for example. From few GHz to a few hundred of GHz, its spectrum of applications is large and various, including space and radar applications. SIW benefits from a great potential.

Nevertheless, SIW is a substrate-based transmission line and suffers from dielectric losses, which increase at high frequency and decrease the power handling capability. This limitation reduces the performance of such technology, and hence its applications.

In this context, which is more detailed in Chapter II, a new topology of SIC, namely

Air-Filled SIW (AFSIW), is presented in this thesis. It consists of a multilayer printed circuit board transmission line with a laser micromached air-cut. The aim of such structure is to drastically reduce the dielectric loss contribution by removing the dielectric and replacing it by air. The rows of vias are maintained to obtain a complete SIW structure, without TM modes propagation. The structure with a theoretical analysis of losses and power handling capability compared to the simulated and the measured results are presented in Chapter III. A comparison of performances between the standard planar transmission line, RWG, SIW, and

AFSIW is shown in Fig. 2. To interconnect this technology with existing SIW, a transition is introduced and bend topologies are studied. For measurement purposes, an AFSIW Thru-

Reflect-Line calibration kit has been designed to de-embed all necessary transitions between the device under test (DUT) and the Vector Network Analyzer (VNA).

Figure 2: Comparison of standard transmission line, RWG, SIW and AFSIW properties.

Then, several passive circuits have been designed and implemented using this AFSIW technology. In Chapter IV, basic elements often found in feeding networks are explored. A power divider using a T-junction, together with Riblet, Moreno and two-hole couplers are demonstrated. Then, a double dielectric slab-loaded AFSIW phase shifter is introduced with analytical, simulated and measured results. Based on this component, a broadband compensated AFSIW phase shifter is proposed.

Finally, passive resonating and radiating circuits have been explored by the implementation of AFSIW filters and antennas. In the first part of the Chapter V, two passband filters are experimented using AFSIW technology. On the one hand, a third order

AFSIW filter using inductive posts compared to its SIW counterpart is realized, and on the other hand, a fourth order multilayer vertically stacked cavitie AFSIW filter is presented. In the second part, two AFSIW antennas are introduced. An AFSIW slot antenna with four apertures is demonstrated and compared to its SIW counterpart. Then, an AFSIW Antipodal

Linear Tapered Slot Antenna (ALTSA) is designed based on AFSIW technology. This antenna is also compared to its SIW version. At the end, Chapter VI concludes this thesis and presents some perpestives.

The research work, introduced in this manuscript, has been published in 5 international journals (4 published and 1 submitted), 6 international conferences, 6 national conferences and 4 communications.

References:

[1] P. Christopher, "Millimeter wave advantages for satellite communications," Military Communications Conference, 1983. MILCOM 1983. IEEE, Washington, DC, USA, pp. 555-562.C, 1983. [2] D. R. Vizard, “Millimeter-wave applications: from satellite communications to security systems,” Microwave Journal, Euro-Global Edition, vol. 49, no. 7, pp.22–36, Jul. 2006. [3] K. Wu, "Towards the development of terahertz substrate integrated circuit technology," Silicon Monolithic Integrated Circuits in RF Systems (SiRF), 2010 Topical Meeting on, New Orleans, LA, pp. 116-119, 2010. [4] M. Bozzi, A. Georgiadis and K. Wu, "Review of substrate-integrated waveguide circuits and antennas," in IET Microwaves, Antennas & Propagation, vol. 5, no. 8, pp. 909-920, June 6, 2011. [5] Ke Wu, D. Deslandes and Y. Cassivi, "The substrate integrated circuits - a new concept for high-frequency electronics and optoelectronics," Telecommunications in Modern Satellite, Cable and Broadcasting Service, 2003. TELSIKS 2003. 6th International Conference on, pp. P-III-P-X, vol.1, 2003. [6] M. Bozzi, F. Xu, D. Deslandes and K. Wu, "Modeling and design considerations for substrate integrated waveguide circuits and components," Telecommunications in Modern Satellite, Cable and Broadcasting Services, 2007. TELSIKS 2007. 8th International Conference on, Nis, pp. P-VII-P-XVI, 2007. [7] D. Deslandes, M. Bozzi, P. Arcioni and Ke Wu, "Substrate integrated slab waveguide (SISW) for wideband microwave applications," Microwave Symposium Digest, 2003 IEEE MTT-S International, Philadelphia, PA, USA, pp. 1103-1106, vol.2, 2003. [8] P. Mondal and K. Wu, "A leaky-wave antenna in substrate integrated non-radiative dielectric (SINRD) waveguide with controllable scanning rate," in IEEE Transactions on Antennas and Propagation, vol. 61, no. 4, pp. 2294-2297, April 2013. [9] K. Wu, "Substrate integrated circuits (SICs) for GHz and THz electronics and photonics: Current status and future outlook," German Microwave Conference Digest of Papers, Berlin, pp. 292-295, 2010.

CHAPTER II: STATE- OF-THE-ART ON ELECTROMAGNETIC TRANSMISSION LINES

TABLE OF CONTENTS

TABLE OF CONTENTS ...... 26 LIST OF FIGURES ...... 27 LIST OF TABLES ...... 28 1. ELECTROMAGNETIC RADIATION ...... 29

1.1. FREE SPACE PROPAGATION ...... 30 1.2. GUIDED WAVES [2] ...... 33 2. TRANSMISSION LINES ...... 35

2.1. METALLIC WAVEGUIDES ...... 35 2.2. PLANAR TRANSMISSION LINES ...... 40 2.2.1. Stripline ...... 40 2.2.2. Microstrip Line ...... 43 2.2.3. Coplanar Waveguide ...... 45 2.2.4. Slow-Wave Transmission Line ...... 48 2.2.5. Discussion ...... 49 2.3. SUBSTRATE INTEGRATED WAVEGUIDE ...... 50 2.3.1. Historical Background ...... 50 2.3.2. SIW Structure and Performance ...... 52 2.3.3. SIW Topologies ...... 55 2.3.4. SIW Circuits ...... 57 3. CONCLUSION ...... 59 REFERENCES ...... 61

LIST OF FIGURES

FIGURE 1: ELECTROMAGNETIC SPECTRUM [1]...... 29 FIGURE 2: ILLUSTRATION OF THE FREE SPACE MODEL...... 31 FIGURE 3: A PLANAR SECTION FROM A UNIFORM RECTILINEAR TRANSMISSION SYSTEM...... 34 FIGURE 4: PICTURE OF KA-BAND RECTANGULAR WAVEGUIDE COMPONENTS [2]...... 36 FIGURE 5: GEOMETRY OF RECTANGULAR WAVEGUIDE...... 37 FIGURE 6: RECTANGULAR WAVEGUIDE LOSSES [7]...... 39 FIGURE 7: (A) AVERAGE POWER HANDLING CAPABILITY (APHC) AND (B) PEAK POWER HANDLING CAPABILITY (PPHC) FOR STANDARD RECTANGULAR WAVEGUIDES [8]...... 39 FIGURE 8: EVOLUTION OF THE FLAT-STRIP LINE: FROM (A) COAXIAL LINE, (B) SQUARE LINE, (C) RECTANGULAR LINE TO (D) FLAT-STRIP TRANSMISSION LINE [9]...... 40 FIGURE 9: (A) STRIPLINE PROPOSED BY AIRBORNE INSTRUMENTS LABORATORY AND (B) THE TRI-PLATE LINE PROPOSED BY SAUNDERS ASSOCIATES INC. [9]...... 41 FIGURE 10: (A) GEOMETRY OF A STRIPLINE AND (B) A DESCRIPTION OF ITS ELECTRIC AND MAGNETIC FIELD [2]...... 41 FIGURE 11: (A) GEOMETRY OF A MICROSTRIP LINE AND (B) A DESCRIPTION OF ITS ELECTRIC AND MAGNETIC FIELD FOR THE FUNDAMENTAL MODE [2]...... 43 FIGURE 12: GEOMETRY OF A CPW WITH A FINITE THICKNESS SUBSTRATE [17]...... 45 FIGURE 13: GEOMETRY OF A CBCPW WITH A FINITE THICKNESS SUBSTRATE [17]...... 47 FIGURE 14: GEOMETRY OF (A) A SLOW WAVE CPW [27] AND (B) A SLOW WAVE MICROSTRIP [28]...... 48 FIGURE 15: EVOLUTION OF COMPUTER POWER [43]...... 52 FIGURE 16: EVOLUTION OF PUBLISHED PAPERS ON SIW EXTRACTED FROM IEEE XPLORE WEBSITE...... 52 FIGURE 17: (A) SIW GEOMETRY [46] AND (B) SIW CROSS-SECTION VIEW WITH THE

FUNDAMENTAL TE10 MODE ELECTRICAL FIELD REPRESENTATION...... 53 FIGURE 18: DIFFERENT TOPOLOGIES OF SIW: (A) SIFW, (B) HF-SIW, (C) SISW, AND (D) RSIW [54]...... 56 FIGURE 19: (A) 3-D VIEW AND (B) CROSS SECTION VIEW OF THE SW-SIW [56]...... 57 FIGURE 20: SEVERAL TOPOLOGIES OF FILTER: (A) FILTER WITH ASYMMETRIC INDUCTIVE POSTS, (B) FILTER WITH SYMMETRIC INDUCTIVE POSTS, (C) FILTER WITH CIRCULAR CAVITIES, AND (D) FILTER WITH RECTANGULAR CAVITIES AND CROSS-COUPLING [54]...... 58 FIGURE 21: (A) BLOCK DIAGRAM AND (B) PICTURE OF THE 4X4 BUTLER MATRIX [59]...... 58 FIGURE 22: TOPOLOGIES OF ANTENNAS: (A) ALTSA, (B) HORN ANTENNA, (C) LEAKY WAVE ANTENNA, AND (D) ANTENNA NETWORK...... 59 FIGURE 23: COMPARISON OF INSERTION LOSS, COST AND SIZE OF DIFFERENT RESONATOR TECHNOLOGIES [70]...... 60

LIST OF TABLES

TABLE I: ELECTROMAGNETIC EQUATIONS OF THE TE AND TM MODES IN RECTANGULAR WAVEGUIDE [2]...... 37 TABLE II: TRANSMISSION LINES PERFORMANCE COMPARISON...... 60

1. ELECTROMAGNETIC RADIATION

Electromagnetic Radiation (EMR) consists mainly of electromagnetic waves which can be split up into transverse electric and magnetic fields propagating at the speed of light through vacuum. These waves are characterized by the frequency or the wavelength of their oscillations constituting the electromagnetic spectrum as depicted in Fig. 1.

Figure 1: Electromagnetic Spectrum [1].

The physical optics theory appeared during the 17th century with the research work from

Christiaan Huygens. His research is mainly focused on the law of reflection and refraction to predict wave behavior. Then, Augustin Fresnel kept up to work on the optical domain and developed the concept of wavelength and interference.

In the middle of 19th century, James Clerk Maxwell has summarized the electromagnetic laws developed by Gauss, Ampere, Faraday, and others, and introduced the

Maxwell’s equations given by (1.1) to (1.5). He succeeded to foresee the existence of electromagnetic waves and demonstrated that the light is an electromagnetic wave. These equations consist of a set of partial differential equations which describe how electric and

29 magnetic fields are generated and altered by each other, or by charges and currents at a macroscopic level.

휌 Maxwell ⃗ 푑푖푣 퐸 = (1.1) Gauss 휀0 푑푖푣 퐵⃗ = 0 (1.2) 휕퐵⃗ Maxwell 푟표푡⃗⃗⃗⃗⃗⃗ 퐸⃗ = − (1.3) Faraday 휕푡 휕퐸⃗ Maxwell 푟표푡⃗⃗⃗⃗⃗⃗ 퐵⃗ = 휇 푗 + 휇 휀 (1.4) Ampere 0 0 0 휕푡 ⃗⃗⃗ ⃗ ⃗ Lorentz 퐹퐿 = 푞(퐸 + 푣 ∗ 퐵) (1.5) Later on, Heinrich Hertz, Alexander Popov, Edouard Branly and Nikola Tesla began to develop and study theories and applications using the shortwave radio. Guglielmo Marconi was the first to manage to transfer a message from Great Britain to Canada over the Atlantic

Ocean using a wireless telegraph in 1901. From this moment, the development of applications based on electromagnetic waves increased faster and faster.

1.1. Free Space Propagation

The propagation inside a wireless channel such as air, space or sea water can be modeled as the electromagnetic waves undergo many effects including reflection, diffraction and scattering due to the presence of obstacles in the channel. These phenomena can lead to increase propagation losses and have to be taken into account for designing a system

(transmitter, receiver or transceiver) with a link budget analysis.

The reflection effect occurs when electromagnetic waves strike an object, such as the ground, buildings or walls, which have a larger size than the wavelength of the propagating waves. Furthermore, if the radio wave falls on a propagating medium having different electrical properties, part of the energy is transmitted and part of it is reflected back. The diffraction consists of a phenomena occurring when a travelling wave interacts with a surface having sharp irregularities. This happens in the case of skyscrapers and creates a shadowed

30 region where the received energy decreases as the user moves deeper. The Huygens’ principle can explain the diffraction effect and describes the field in the shadow region as a sum of multiple secondary sources producing a new wave front in direction of the receiver. Contrary to the reflection effect, the scattering manifests itself when the obstacle on the propagating channel is smaller than the wavelength of the waves. These various phenomena can lead to a multipath loss effect, increasing or decreasing the received power and must be included in the propagation equation.

In general case, antennas are used for transmitting signals and imply that the free space equations are mainly developed inside the far-field region, also called Fraunhofer region, where the propagating waves behave as plane waves and the power decays inversely with distance. Hence, this distance, df, is usually depicted as in (1.6) and depends on the largest dimension, D, of the antenna [2].

2퐷2 푑 = (1.6) 푓 휆 The most well-known equation which models the free space propagation is the Friis equation given in (1.7) for two antennas in direct sight without obstacles [2] as shown in

Fig. 2.

λ Antenna Antenna Transmitter Receiver

Gt Pt Gr Pr Figure 2: Illustration of the free space model.

2 푃푡퐺푡퐺푟휆 푃푟(푑) = (1.7) (4휋)2푑2퐿

where Pt is the transmitted power, Pr(d) is the received power, Gr and Gt are the receiver and transmitter antenna gain respectively, d is the distance between the antennas, ηrad is the

31 radiation efficiency, and L is the coefficient taking into account all the perturbing effects. It has to be notified that the gain is related to the effective aperture, Ae, which depends on the physical size of each antenna as described in (1.8) [2].

4휋퐴푒 퐺 = 휂 (1.8) 푟푎푑 휆2 Hence, it is also easy to represent the attenuation suffered by the propagating signal through a wireless channel. It is given by the difference of the transmitted and received power, called Path Loss, PL, and it is expressed in (1.9) [2].

푃푡 푃퐿(푑퐵) = 10푙표푔10 ( ) (1.9) 푃푟 To go further, some researchers had tried to find a better equation than this from Friis to match specific applications. Longley-Rice model, Durkin’s model, Okumura’s model, and

Hata’s model are the most used for outdoor propagation model. As an example, Okumura’s model [3] is mainly used inside urban area from 100 MHz to 1900 MHz for a distance between 1 km and 100 km. This model has been essentially developed for mobile telecommunication. It is based on an empirical formula obtained from numerous measurements of signal attenuation between base stations to mobile phones. The empirical path-loss equation of Okumura is given by (1.10) [3] at a distance d.

푃퐿 (푑) = 퐿 + 퐴 − 퐺(ℎ ) − 퐺(ℎ ) − 퐺 푑퐵 (푓푐,푑) 푚푢(푓푐,푑) 푡 푟 퐴푅퐸퐴 (1.10)

where L(fc,d) is the free space path loss at a distance d and a carrier frequency fc, Amu(fc,d) is the median attenuation in addition to free space path loss, G(ht) and G(hr) are the base station and mobile antenna gain, and GAREA is the gain changing with the environment. Amu(fc,d) and GAREA are determined by the experiments. G(ht) and G(hr) are defined as follow:

ℎ푡 퐺(ℎ ) = 20푙표푔 ( ) , 30 푚 < ℎ < 1000 푚 (1.11) 푡 10 200 푡

ℎ푟 퐺(ℎ ) = 10푙표푔 ( ), ℎ ≤ 3 푚 (1.12) 푟 10 3 푟

ℎ푟 퐺(ℎ ) = 20푙표푔 ( ) , 3 푚 < ℎ < 10 푚 (1.13) 푟 10 3 푟

hr and ht are the height of the base station antenna and the mobile phone antenna respectively. This model has a derivation of 10-14 dB compared to empirical path loss and the path loss extracted from measurement.

Furthermore, contrary to Okumura’s model, it also exists indoor propagation models to take into account the specific environment related to the interior arrangement of our buildings.

For example, the covered distances are smaller than outside and the composition of each wall influences the propagation channel. When the control of the propagating waves is required by the application and the shape of the channel, it is necessary to model the behavior of signals by solving the Maxwell’s equations.

1.2. Guided Waves [2]

In the case of harmonic conditions, it is assumed that the material used as propagation channel is isotropic, linear and homogeneous. Hence, the dielectric losses and the magnetic losses are defined as depicted in equations (1.14) and (1.15). Therefore, the Maxwell’s equations become as in (1.16).

퐷⃗⃗ = 휀퐸⃗ (1.14)

퐵⃗ = 휇퐸⃗ (1.15) 푑푖푣 퐷⃗⃗ = 0 푑푖푣 퐵⃗ = 0 Maxwell’s (1.16) equations 푟표푡⃗⃗⃗⃗⃗⃗ 퐸⃗ = −푗휔휇퐻⃗⃗ 푟표푡⃗⃗⃗⃗⃗⃗ 퐻⃗⃗ = 푗휔휀퐸⃗

Based on (1.16), the wave’s equations for the electric and magnetic field can be extracted from [2] and are given in (1.17) and (1.18).

∆퐸⃗ + 휔2휇휀퐸⃗ = ⃗0 (1.17) ⃗⃗ 2 ⃗⃗ ⃗ ∆퐻 + 휔 휇휀퐻 = 0 (1.18)

In the case of guided waves inside uniform rectilinear transmission system, as depicted in Fig. 3, the wave’s equations can be applied and solved. For that, it is assumed that a planar section of the propagation material is uniform and the propagation is along z-axis in a

Cartesian coordinate system. The other axis (x, y) are used for the transverse fields. Thus, by studying the wave’s equations in this material, the dispersion equation can be written as in

(1.19) [4].

y µ ε x 0 Planar section

Figure 3: A planar section from a uniform rectilinear transmission system.

2 2 2 퐾푐 = 훤 + 휔 휇휀 with 훤 = 훼푔 + 푗훽푔 (1.19)

Γ is the propagation constant and is the same inside all the planar section, even if there is more than one material because of the continuity equations. Then, the general equations of propagation can be defined as the relation between the transverse and lengthwise components of the electromagnetic field and are defined in (1.20) and (1.21) [4].

2 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ 퐾푐 푒⃗⃗⃗휏 = 푗휔휇(푛⃗ ⋀푔푟푎푑(ℎ푧)) − (±) 훤푔푟푎푑(푒푧) (1.20) 2⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗ 퐾푐 ℎ휏 = −푗휔휀(푛⃗ ⋀푔푟푎푑(푒푧)) − (±) 훤푔푟푎푑(ℎ푧) (1.21) Based on the previous conditions and on the equations (1.20) and (1.21), only three main modes can exist and propagate in such propagation channel. Transverse Magnetic (TM) and Transverse Electric (TE) mode are characterized by the fact that there is no propagation of the magnetic and electric lengthwise field respectively. They are depicted in (1.22) and

(1.23) where ZTM and ZTE are the impedance wave of each mode [2].

1 훤 ⃗⃗⃗⃗ TM mode ℎ휏 = − (푒⃗⃗⃗휏 ⋀(±푛⃗ )) with 푍푇푀 = (1.22) 푍푇푀 푗휔휀

2 ∆푒푧 + 퐾푐 푒푧 = 0

푗휔휇 ⃗⃗⃗⃗ 푒⃗⃗⃗휏 = 푍푇퐸(ℎ휏⋀푛⃗ ) with 푍푇퐸 = TE mode ±훤 (1.23) 2 ∆ℎ푧 + 퐾푐 ℎ푧 = 0 The third main mode is the Transverse Electric Magnetic (TEM) mode, which means that there is any propagation of lengthwise electromagnetic fields. Hence, the general equations of propagation (1.20) and (1.21) cannot be used. It is required to combine equations

(1.16) and the general expression of the electric (1.24) and the magnetic field (1.25) [4].

±훤푧 퐸⃗ = (푒⃗⃗⃗휏 + 푛⃗ . 푒푧)푒 (1.24)

⃗⃗ ⃗⃗⃗⃗ ±훤푧 퐻 = (ℎ휏 + 푛⃗ . ℎ푧)푒 (1.25) As a result, the dispersion equation for the TEM mode is obtained in (1.26) and the transverse electric and magnetic fields are given in (1.27) and (1.28) respectively [4].

2 2 훤 = −휔 µ푐휀푐푎 (1.26) 푗휔µ ⃗⃗⃗⃗ 푐 푒⃗⃗⃗휏 = 푍푇퐸푀(ℎ휏^푛⃗ ) with 푍푇퐸푀 = ±훤 (1.27) 1 ±훤 ⃗⃗⃗⃗ ℎ휏 = − (푒⃗⃗⃗휏 ^푛⃗ ) with 푍푇퐸푀 = (1.28) 푍푇퐸푀 푗휔휀푐푎 Now, it is easier to describe and predict the behavior of classical transmission line.

Using TE, TM and TEM mode, the main characteristics of each transmission line can be found. And finally, it allows scientists to design them and understand such structures.

2. TRANSMISSION LINES 2.1. Metallic Waveguides

The first wave guiding structure has been tested in 1894 based on the theory of Joseph

John Thomson generalized by Lord Rayleigh [5]. He solved the boundary value problem of electromagnetic waves propagating in an arbitrary shaped waveguide. He has shown that the waves can travel with specific modes known as TE or TM modes, and a combination from them called hybrid mode. The cut-off wavelength was introduced to determine the section of a

35 hollow tube and the frequency below which waves could no longer propagate. Therefore, it became possible to develop passive devices, as shown in Fig. 4, or systems using these technologies and transmit an electromagnetic signal from one point to another while controlling the direction of propagation.

Figure 4: Picture of Ka-band rectangular waveguide components [2].

There are different topologies of waveguide such as the dielectric-loaded waveguide, the circular or ridge waveguide [6] and the well-known rectangular metallic waveguide as depicted in Fig. 5. The behavior of the last one is very well-described by solving the equations

(1.22) and (1.23) for the TM and TE mode respectively while the boundary conditions have to be applied. Then, the propagation equations of the electric and magnetic fields in each mode are obtained as described in [2]. The equations of each mode inside such waveguide are summed up in Table 1.

y z b µ ε x 0 a Figure 5: Geometry of rectangular waveguide.

Value TEmn mode TMmn mode

k 휔√µ휀 휔√µ휀

푚휋 2 푛휋 2 푚휋 2 푛휋 2 k √( ) + ( ) √( ) + ( ) c 푎 푏 푎 푏

푐 푚휋 2 푛휋 2 푐 푚휋 2 푛휋 2 f √( ) + ( ) √( ) + ( ) c 2휋 휀 푎 푏 2휋√휀푐푎 푎 푏 √ 푐푎

2 2 √푘2 − 푘2 β √푘 − 푘푐 푐 2휋 2휋

λc 푘 푘푐 푐 2휋 2휋

λg 훽 훽 휔 휔

νp 훽 훽 푚휋푥 푚휋푦 0 퐸0sin ( )sin ( ) Ez 푎 푏 푚휋푥 푚휋푦 퐻 cos( )푐표푠 ( ) 0 Hz 0 푎 푏 푗훽휋푚 푚휋푥 푚휋푦 푗휔µ푐휋푛 푚휋푥 푚휋푦 −푗훽푧 퐻 cos( )푠푖푛 ( )푒−푗훽푧 − 퐸0cos( )푠푖푛 ( )푒 Ex 2 0 푘2푎 푎 푏 푘푐 푏 푎 푏 푐

푗훽휋푚 푚휋푥 푚휋푦 푗휔휀푐푎휋푛 푚휋푥 푚휋푦 −푗훽푧 −푗훽푧 퐻0sin( )푐표푠 ( )푒 Hx 2 퐻0sin( )푐표푠 ( )푒 푘2푏 푎 푏 푘푐 푎 푎 푏 푐 푗훽휋푛 푚휋푥 푚휋푦 푗휔µ푐휋푚 푚휋푥 푚휋푦 −푗훽푧 − 퐻 sin( )푐표푠 ( )푒−푗훽푧 − 퐸0sin( )푐표푠( )푒 Ey 2 0 푘2푏 푎 푏 푘푐 푎 푎 푏 푐

푗훽휋푛 푚휋푥 푚휋푦 푗휔휀푐푎휋푚 푚휋푥 푚휋푦 −푗훽푧 −푗훽푧 − 퐸0cos( )푠푖푛 ( )푒 Hy 2 퐻0cos( )푠푖푛 ( )푒 푘2푎 푎 푏 푘푐 푏 푎 푏 푐 푘휂 훽휂 푍 = 푍푇퐸 = Z 푇퐸 훽 푘 Table I: Electromagnetic equations of the TE and TM modes in rectangular waveguide [2].

This structure presents different loss mechanisms that impact the performance of such waveguide. First, the dielectric loss is mainly due to the dissipation of the electromagnetic energy inside the dielectric material. It can be explained by the fact that the molecules are naturally polarized and when an electric field is applied, it changes this polarization. As the electric field alternatively changes its polarization, it switches the polarization of the molecules constituting the dielectric material. As a result, a dissipation of the energy appears and molecules evacuate this energy by heat. It becomes obvious that this phenomenon depends also on the frequency. The metallic walls create the second loss mechanism. Each metallic material is mainly composed of atomic ions. When a current created by the magnetic field go through a metallic surface, collisions between electrons and atoms are generated. The created energy is evacuated by heating the conductor and leads to the conductor loss. These losses can be described by (1.29) and (1.30) for the dielectric and ohmic loss respectively for a rectangular waveguide as depicted in [2]. Hence, the total loss can be extracted and is illustrated in Fig. 6 [7] based on previous equations for different standard rectangular waveguide.

Another important property of the rectangular waveguide is the power handling. This results in different phenomenon that can limit the maximum transmitted power. For example, it exists the corona and multipaction effect1 which take place generally for space and airspace applications. Here, the study is mainly focused on the average and peak power handling. The first one concerns the continuous power and the second one describes the maximum power that a wave-guiding structure can reach during a short time. Fig. 7 illustrates both of these parameters for several standard waveguides [8].

It can be concluded from Fig. 6 that rectangular waveguide is a high power and low loss transmission line. In addition, it is a bulky structure. However, it is necessary to use a precise

1 The multipactor effect is a phenomenon in radio frequency waveguides, where, under certain conditions, secondary electron emission in resonance with an alternating electric field leads to exponential electron multiplication. 38 mechanical machining process to fabricate waveguide made of brass or aluminum and then, it results in a high cost and heavy device.

푘2푡푎푛훿 훼푑 = (1.29) 2훽 푅 휔휇 훼 = 푠 (2푏휋2 + 푎3푘2) 푤푖푡ℎ 푅 = √ 푐 푐 푎3푏훽푘휂 푠 2휎 (1.30)

Figure 6: Rectangular waveguide losses [7].

(a) (b) Figure 7: (a) Average Power Handling Capability (APHC) and (b) Peak Power Handling Capability (PPHC) for standard rectangular waveguides [8].

2.2. Planar Transmission Lines

According to [9], the first planar transmission line appeared during the Second World

War. It has been introduced by V. H. Rumsey and H. W. Jamieson. They have been applied to an antenna system and power feeding network. It consisted in a flat-strip coaxial transmission line. Its evolution is described in Fig. 8 and the work was reported at an IRE meeting and in subsequent articles [10]-[13]. After several years, during the 1950’s, the first planar circuit based on microwave printed-circuit technique have been realized mainly through several expanded research and development program. Several companies or laboratories decided to involve in this promising technology.

Figure 8: Evolution of the flat-strip line: from (a) Coaxial line, (b) Square line, (c) Rectangular Line to (d) flat-strip transmission line [9].

2.2.1. Stripline

Two different laboratories originally proposed the first stripline. The two proposed version are similar as illustrated in Fig. 9. The first one, shown in Fig. 9(a), consists in strip conductor inserted between two dielectric materials and enclosed by two conducting plates.

The second one, as described in Fig. 9(b), was made using a dielectric plate enclosed by two strip conductors and two conducting plates are suspended on top and bottom of the structure, also known as tri-plate line.

(a) (b) Figure 9: (a) Stripline proposed by Airborne Instruments Laboratory and (b) the tri-plate line proposed by Saunders Associates Inc. [9].

The actual version of the stripline used as a transmission line is described in Fig. 10(a) with its electromagnetic field in Fig. 10(b). A thin conductor strip of W width is buried and centered inside a wide dielectric material. Two ground planes are added at a distance b/2 from the strip conductor and cover the entire dielectric region.

(a) (b) Figure 10: (a) Geometry of a stripline and (b) a description of its electric and magnetic field [2].

The stripline can propagate TEM modes due to its structure and it is also the most used in this case. The characteristic impedance, given by (1.31) can be extracted from the TEM mode equations (1.26) to (1.28) and a computational fitting as described in [2].

30휋 푏 푍0 = (1.31) √휀푟 푊푒 + 0.441푏

0 푓표푟 푊⁄ > 0.35 푊푒 푊 푏 = − { 푊 2 (1.32) 푏 푏 (0.35 − ) 푓표푟 푊⁄ < 0.35 푏 푏

where We is the effective width of the strip conductor as depicted in (1.32). There are two cases for the calculation of the effective width due to fringing effect. For narrow strip conductor, the fringing effect impacts the electromagnetic behavior of such transmission line and is estimated for a W/b ratio lower than 0.35. For reasons related to stripline design, a version derived from the equations (1.31) and (1.32) is given in (1.33). This relation is referenced to the same 0.35 limit as in (1.32) but it is expressed in terms of impedance 120Ω.

30휋 − 0.441 푓표푟 √휀푟 푍0 < 120훺 푊 푍0√휀푟 = (1.33) 푏 30휋 0.85 − √0.6 − − 0.441 푓표푟 √휀푟푍0 > 120훺 { 푍0√휀푟

The losses from a stripline are generated mainly by the conductor and dielectric losses.

The dielectric loss inherits largely from the TE or TM mode as described in (1.29). In the case of TEM, the cut-off wave number is cancelled, and therefore, the propagation constant becomes equal to the wave number which leads to (1.34). Concerning the ohmic loss, if t represents the thickness of the central conductor strip, it can be found by the perturbation method or Wheeler’s incremental inductance rule as in [2]. This results in the equation (1.35).

푘푡푎푛훿 훼 = (1.34) 푑 2 −3 2.7 ∗ 10 푅푠휀푟푍0 퐴 푓표푟 √휀 푍 < 120훺 30휋(푏 − 푡) 푟 0 훼푐 = (1.35) 0.16푅푠 퐵 푓표푟 √휀푟푍0 > 120훺 { 푍0푏 with 2푊 1 푏 + 푡 2푏 − 푡 퐴 = 1 + + ln ( ) (1.36) 푏 − 푡 휋 푏 − 푡 푡 푏 0.414푡 1 4휋푊 퐵 = 1 + (0.5 + + 푙푛 ( )) 0.5푊 + 0.7푡 푊 2휋 푡 (1.37)

2.2.2. Microstrip Line

The topology of the microstrip line was proposed by the Federal Communication

Laboratories at about the same time as the striplines [9]. As described in Fig. 11(a), it consists of dielectric material covered by two ground planes where on one face, metallic portions are partially removed for creating a strip conductor line. This planar transmission line is the most well-known because it is often used by Printed Circuit Board (PCB) manufacturers. Easy to interconnect with passive and active lumped components, this topology has been intensively studied and many systems based on it have been developed. The electromagnetic field for the fundamental mode of a microstrip line is illustrated in Fig. 11(b).

(a) (b) Figure 11: (a) Geometry of a microstrip line and (b) a description of its electric and magnetic field for the fundamental mode [2].

The microstrip line cannot propagate TEM mode because of the phase velocity which is different inside the dielectric region and the air above the conductor strip. Hence, the real mode propagating inside this transmission line is a hybrid mode between TE and TM mode.

However, considering that the dielectric thickness is thin compared to the wavelength of the propagating signal, it is assumed that a quasi-TEM mode could establish in such microstrip line. As a consequence, the relative permittivity εr is replaced by an effective permittivity εe, given in (1.38), because part of the field is located in the dielectric and another part outside.

The relations, depending on the characteristic impedance from (1.39), for designing this topology of line are described in (1.39) [2]. Like the stripline, the fringing effect has to be taken into account for narrow strip conductor and leads to two different equations for the impedance.

휀푟 + 1 휀푟 − 1 1 휀푒 = + 2 2 √ 푏 (1.38) 1 + 12 푊 60 8푏 푊 푊 푙푛 ( + ) 푓표푟 ≤ 1 √휀푒 푊 4푏 푏 푍0 = (1.39) 120휋 푊 푊 푊 ( + 1.393 + 0.667푙푛 ( + 1.444)) 푓표푟 ≥ 1 { √휀푒 푏 푏 푏 푊 푏 8푒퐴 푊 푓표푟 < 2 푒2퐴 − 2 푏 (1.40) = 2 휀푟 − 1 0.67 푊 (퐵 − 1 − 푙푛(2퐵 − 1) + (푙푛(퐵 − 1) + 0.39 − )) 푓표푟 > 2 { 휋 2휀푟 휀푟 푏 with

푍0 휀푟 + 1 휀푟 − 1 0.11 퐴 = √ + (0.23 + ) (1.41) 60 2 휀푟 + 1 휀푟 377휋 퐵 = (1.42) 2푍0√휀푟 The loss mechanism from a microstrip line is generated mainly by the conductor, dielectric and radiation losses. The dielectric loss inherits largely from the quasi-TEM mode

(1.43). For the ohmic loss, the relation is approximatively given in [14] by (1.44). The radiation is a phenomenon that usually appears for high frequency. This loss has been studied in [15] for microstrip transmission line and some discontinuities. (1.45) corresponds to an open circuit and (1.46) is for a matched transmission line. It can be seen from these equations that the thicker is the conductor strip, the higher are the radiation losses. The same variation is generated with the rise of the frequency or for low effective dielectric permittivity. The last one is explained by the fact that the quantity of electromagnetic energy contained in the dielectric becomes lower and more energy can propagate in the air.

푘0휀푟(휀푒 − 1)푡푎푛훿 훼푑 = (1.43) 2√휀푒(휀푟 − 1)

푅푠 휔휇0 훼푐 = 푤푖푡ℎ 푅푠 = √ (1.44) 푍0푊 2휎

2 2 2휋푡 휀 + 1 (휀 − 1) √휀푒 + 1 Open 훼 = 60 ( ) ( 푒 − 푒 푙표푔 ( )) 푟 휆 휀 3⁄ (1.45) circuit 0 푒 2휀푒 2 √휀푒 − 1

2 2휋푡 휀푒 − 1 √휀푒 + 1 Matched 훼푟 = 60 ( ) (1 − 푙표푔 ( )) (1.46) Line 휆0 2√휀푒 √휀푒 − 1

2.2.3. Coplanar Waveguide

The Coplanar Waveguide (CPW) was proposed few years after the microstrip line during the 1960’s. In [16], the first topology of such transmission line is described and the introduced structure became the actual conventional CPW. It consisted in a dielectric substrate with above a centered conductor strip surrounded by two ground planes on each side. As a separator, two air gaps are inserted on both sides of the central conductor. CPW geometry is shown in Fig. 12. To study this structure, a conformal mapping technique is used as illustrated in [16]. It is assumed that a quasi-TEM mode can propagate along such transmission line. Hence, the characteristic impedance Z0 and the effective permittivity εe, useful for designing a CPW, are extracted from [17] and given in (1.47) and (1.48). These characteristics depend on the structure parameters as shown in Fig. 12. The mathematical function K in the next equations referred to a complete elliptical integral of the first kind (See

Annex 1).

Figure 12: Geometry of a CPW with a finite thickness substrate [17].

′ (휀푟1 − 1) 퐾(푘1)퐾(푘0) 휀푒 = 1 + ′ (1.47) 2 퐾(푘0)퐾(푘1)

30휋 퐾(푘′ ) 푍 = 0 0 (1.48) √휀푒 퐾(푘0)

with, 푆 푘 = (1.49) 0 푆 + 2푊 휋푆 푠푖푛ℎ ( ) 4ℎ 푘 = 1 1 휋(푆 + 2푊) (1.50) 푠푖푛ℎ ( ) 4ℎ1 ′ 2 푘푖 = √1 − (푘푖) (1.51) Various topologies of CPW have been presented in [18]. One of the most popular is the grounded CPW where a ground plane is added beneath the dielectric substrate. This structure is largely used for application requiring heat dissipation. This structure is called Conductor-

Backed Coplanar Waveguide (CBCPW). Its geometry is shown in Fig. 13. As before, a quasi-

TEM mode is assumed to propagate along such transmission line. Based on this approximation, the analysis of [16] is extended by several papers [19]-[23] so to obtain the characteristic impedance Z0 and the effective permittivity εe as a function of the geometry for designing purposes. These are given in (1.52) and (1.53), extracted from [17].

′ 퐾(푘 )퐾(푘3) 1 + 휀푟 ′ 퐾(푘)퐾(푘3) 휀푒 = ′ (1.52) 퐾(푘 )퐾(푘3) 1 + ′ 퐾(푘)퐾(푘3) 60휋 1 푍 = 0 퐾(푘) 퐾(푘 ) √휀푒 3 (1.53) ′ + ′ 퐾(푘 ) 퐾(푘3) with, 푎 푘 = 푏 (1.54) 휋푎 tanh ( ) 푘 = 2ℎ 3 휋푏 (1.55) tanh ( ) 2ℎ ′ 2 푘푖 = √1 − (푘푖) (1.56)

Figure 13: Geometry of a CBCPW with a finite thickness substrate [17].

The CPW loss mechanism is generated mainly by the conductor, dielectric and to a lesser extent of radiation losses, particularly in high frequency. The basic chosen approach is the conformal mapping method as described in [17]. The dielectric loss inherits largely from the quasi-TEM mode (1.57). The ohmic loss is generated by the central conductor and also by the ground planes on each side of the central strip. It is assumed that the expression given in

(1.59) is more suitable for thickness t of the central metallic part five times greater than the skin depth.

휋 휀푟 훼푑 = 푞푡푎푛훿 (1.57) 휆0 √휀푒 ′ 퐾(푘1)퐾(푘0) with 푞 = 0.5 ′ (1.58) 퐾(푘0)퐾(푘1)

푅푐 + 푅푔 훼푐 = (1.59) 2푍0

Rc and Rg, given in (1.60) and (1.61), are the series and the distributed series resistance of the center strip conductor and the ground planes respectively in ohms per unit length. Rδ is the skin effect surface resistance.

푅훿 4휋푆 1 + 푘0 푅푐 = 2 2 (휋 + 푙푛 ( ) − 푘0푙푛 ( )) (1.60) 4푆(1 − 푘0)퐾 (푘0) 푡 1 − 푘0 푘 푅 4휋(푆 + 2푊) 1 1 + 푘 푅 = 0 훿 (휋 + 푙푛 ( ) − 푙푛 ( 0)) 푔 2 2 (1.61) 4푆(1 − 푘0 )퐾 (푘0) 푡 푘0 1 − 푘0

The radiation losses for CPW have been studied in [24] and [25]. It has been concluded that the wider is the central strip, the higher will be the losses. The same variation has been observed concerning the gaps width. In the case of CBCPW, the model has been presented in

[26]. The calculation is based on Finite-Difference Frequency-Domain (FDFD) method applied on an infinite or finite substrate thickness CBCPW structure. As a result, it is shown that below a frequency limit, there is no radiation loss in both cases because there is no coupling between the CPW mode and parasitic mode created close to the transmission line. In any case, it is clearly shown in [24]-[26] that in both cases the radiation losses can be neglected below the radiofrequency band.

2.2.4. Slow-Wave Transmission Line

A different transmission line topology based on planar circuit technology has been recently reported in many papers in the literature. This concept is based on slow wave effect decreasing the group velocity or increasing the group delay compared to standard structures.

For that, the capacitance effect is generally raised by the insertion of floating metallic strips below the transmission line as depicted in Fig. 14 and in equation (1.62). The resulting effect is the reduction in size of standard transmission lines.

(a) (b) Figure 14: Geometry of (a) a slow wave CPW [27] and (b) a slow wave microstrip [28].

휔 퐶0 1 휈휑 = = = (1.62) 훽 √휀푟푒푓푓 √퐿퐶

The slow wave is of a particular interest for on-chip application. As a passive device, transmission lines are imperfect components with losses and could be very long for

48 synthesizing lumped components in integrated technologies such as silicon. Hence, for economic and integration purposes, slow wave lines such as CPW [27] or microstrip [28] have undergone an important development during the 2000’s, especially on silicon and membrane based technologies.

Regarding the losses, [29] compares the performance from a CBCPW to slow wave

CPW based on IBM silicon technology. It is reported that below the millimeter wave, the losses are lower in a slow wave structure than in conventional CPW considering equal impedance. The quality factor seems to be higher than the standard topologies even beyond

30 GHz. This is of high interest for the design of lumped components including inductors and capacitors.

2.2.5. Discussion

In [30], a comparison between microstrip and CPW lines is presented at millimeter wave. Using the same GaAs technology, it is concluded that the losses and the dispersion for

CPW are better or similar to performance achieved using microstrip. The minimum loss is achieved for 60 Ω and 25 Ω for CPW and microstrip respectively. Moreover, a coplanar waveguide short circuit and an open end microstrip are compared to obtain the loss of a discontinuity. It is inferred from [30] that the CPW radiates less than microstrip. If a 50 Ω transmission line is realized in both technologies, according to [30], the losses are estimated at

0.8 dB/cm and 1 dB/cm for the microstrip and CPW respectively considering GaAs substrate whereas losses in [14] are 0.15 dB/cm and 0.3 dB/cm at 30 GHz. It can be concluded that microstrip is a low loss structure compared to CPW. Nevertheless, CPW is still preferred because it radiates less which means that it is less sensitive to the environment and it offers an easier interconnection with active devices.

From [31], a complete theoretical study of Average Power Handling Capability (APHC) has been done. It is concluded that the microstrip line can reach 126 W (51 dBm) at 10 GHz

49 for the worst case to 76.8 kW (78.9 dBm) using a better substrate. The limitation of this model is due to several parameters: the change of substrate properties with temperature as well as the physical dimensions, the thermal resistance barriers between the strip and the dielectric and also the dielectric and the ground plane, and finally, all discontinuities including bends, connectors and so on. In [32], a recent study based on a new approach on APHC with measurements has been presented. The results show that microstrip filters can reach a power between 4.2 W (36.2 dBm) to 9.2 W (39.6 dBm) at 10 GHz on MEGTRON6 substrate (εr

=3.6, tanδ = 0.006, h = 0.93 mm).

2.3. Substrate Integrated Waveguide

2.3.1. Historical Background

Since the beginning of the 1990’s, several topologies have been proposed to integrate a classical rectangular waveguide using PCB process. One of the first to realize such structure using metallic via holes was presented in the patent from Shigeki in 1994 [33]. In 1995, an

American patent has demonstrated that a waveguide filter can be implemented on a layered dielectric substrate using the Low Temperature Co-fired Ceramics (LTCC) technics [34].

The development of Substrate Integrated Waveguide (SIW) structure was correlated with the development of the PCB fabrication process and also with the accuracy of electromagnetic filled computation directly related to the computer processing capability. The issue was to model the post-holes wall efficiently by finding the best numerical analysis method. For that. Marcuvitz, in his book [35], has summed up several models of inductive post, partially inserted into or penetrating through the waveguide from the broad wall. It was assumed that the insertion posts are small enough and cannot be modeled with a simple method until Leviathan [36], using the Method Of Moment (MOM) introduced in [37], analyze the case of single large post. Since that, Li [38] extended the results to the case of multiple posts array. Then, the model has undergone several improvements and allowed

Hirokawa [39] to publish and demonstrate the potential of a waveguide antenna using metalized via post array integrated in a dielectric substrate. In this device, the vias are drilling and plated to replace the thick metal wall. This new topology of waveguide inherits the advantages of the PCB process in terms of cost and compactness. In 1998, Uchimura [40] has demonstrated the metallization of the well-known structure called afterwards SIW using electromagnetic software.

In 2000, Tzuang [41] has proposed a simple structure integrating a via-post waveguide, a microstrip used as feed transmission line and a tapered transition as interconnect. All the circuit was fabricated using PCB lithography on a Rogers 4003 substrate of 0.508 mm thickness. The aim was to realize a 5th order filter integrating the feedline and the device on the same board with common substrate and ground. Based on the previous works, Ke Wu [42] has presented a new planar SIW in 2001 that became the conventional adopted structure.

Hence, the structure of a classical SIW is made by the top and bottom broad walls covered by copper sheet from the PCB, and the metallic walls are realized by two rows of metalized via.

Since last decades, the operating speed of computer has been largely increased as shown in Fig. 15. For example, the last target for the future is to surpass the brain capacity. This phenomenon created the required conditions to implement commercial electromagnetic software on computer in order to solve the equations by massive iterated simulations. This

Computer-Aided Design (CAD) software accurately solves the complex calculation of hybrid electromagnetic field, eases the research and so, fosters the development of SIW. From few published papers in 2001, the evolution of the number of publications increases exponentially as shown in Fig. 16. This great interest for this technology is mainly due to its advantages but also its potential for targeted applications.

Figure 15: Evolution of computer power [43].

400 300 200

papers 100

Total published Total 0 2001 2002 2003 2005 2007 2009 2011 2013 2015 Years Figure 16: Evolution of published papers on SIW extracted from IEEE Xplore website.

2.3.2. SIW Structure and Performance

In a complete RF system, both planar transmission line and waveguide have the same problem of integration and interconnect with a different technology. On the one hand, the waveguides are low loss structure but are expensive to fabricate. On the other hand, the planar transmission lines have low Q-factor, but are light weight and have cheap fabrication process.

A novel guided structure called Substrate Integrated Waveguide, as shown in Fig. 17(a), combines advantages from both technologies and hence it is put forward by the researchers as described in [44] and [45].

(a) (b) Figure 17: (a) SIW geometry [46] and (b) SIW cross-section view with the fundamental TE10 mode electrical field representation.

Basically, the geometry of SIW is close to rectangular dielectric waveguide. However, by adopting the PCB fabrication process, used for microstrip, the height is reduced drastically to the thickness of PCB substrate. The two metallic broad walls are substituted by two copper sheets and the narrow walls are replaced by two rows of plated vias which are drilled through the substrate. If the vias are close enough, the two rows become equivalent to two electrical walls for the electromagnetic waves and hence, a pair of via in the cross section can generate a current loop.

As SIW inherits largely properties from metallic waveguide, a cut-off frequency exists below which there is no propagation. It can be calculated for the fundamental mode as described in equation (1.63) and (1.64) extracted from [47] that deals with the dispersion characteristics of SIW structure.

푐 푓 = 0 푎 푐10 푒푓푓 (1.63) 2√휀푟 푑2 푎푒푓푓 = 푎푓 − (1.64) 0.95푝

In [48], an accurate formula for the effective width of an SIW is given in (1.65). This new equation takes into account the dispersion effect generated by the d/af ratio. Thus, the

SIW propagation constant can match this from a rectangular waveguide as depicted in [48].

푑2 푑2 푎푒푓푓 = 푎푓 − 1.08 + 0.1 (1.65) 푝 푎푓

Therefore, by adjusting correctly the space between vias, the diameter and the width, the leakage loss from SIW structure are negligible and a design procedure is proposed in [48]. It can be noticed that one of the major change in behavior of a SIW compared to a rectangular waveguide is the surface current. In conventional waveguide, the current surface circulates freely in any direction but in the case of SIW, due to via holes, the current flow is limited to vertical direction. As the structure of via holes is periodically spaced, the side wall current cannot circulate alongside the SIW across each regular interval. That is why there is only the

Transverse Electric TEm0 mode that can propagate, and as a result, the first mode called fundamental mode is the TE10 mode as illustrated in Fig. 17(b).

The main issue is to control and minimize the losses from transmission line, especially for SIW structure. There are three main loss mechanisms. First, the dielectric loss is mainly generated by the dielectric from a chosen PCB board like in rectangular dielectric waveguide.

It depends on the substrate height and the width of the structure as demonstrated in [49]. It can be explained by the dielectric area contained between two rows of vias and, the top and bottom copper sheet. If this area increases, which means that the height or the width is raised, then the dielectric stocks more energy and a greatest part of this energy is transformed into heat. The second losses are created by the conductor used in the SIW. The ohmic loss is subject to the height of the structure and the conductivity of the metal. The higher the SIW is, the lower the conductor loss will be [49]. The last but not least, the radiation loss can occur if conditions are not fulfilled on the diameter and the space between two vias as expressed in

(1.66) pulled out [44]. Hence, in [49], the total losses are estimated to 17.4 dB/m at 20 GHz

(or 0.174 dB/cm to compare with results presented in discussion (2.2.5)) in the case of SIW dimensioned with af = 3.97 mm, d = 0.635 mm, s = 1.016 mm using a substrate with εr = 2,

σd = 0.001 S/m and h = 0.254 mm.

휆푔 푑 < 푎푛푑 푝 ≤ 2푑 (1.66) 5

To go further, as SIW can pretend to replace conventional waveguide for some applications, it could be interesting also to evaluate its power handling capability. This characteristic is well described in [50]. In this paper, the Average and the Peak Power

Handling Capabilities (APHC and PPHC respectively) are explored for SIW. Concerning the

APHC, which represents the behavior for the maximum continuous power, the SIW achieves

450 W, 85 W and 75 W at 10 GHz, 20 GHz and 30 GHz respectively for a Rogers 5880 substrate with with εr = 2.2, tanδ = 0.0009 and h = 0.508 mm. The APHC depends on the total losses, the properties of the material, temperature of the environment and the SIW, and the effective width as shown in (1.67). It is inferred in [50] that the higher is the SIW, the better is the APHC and also, the greater is the relative permittivity, the lower is the APHC. Moreover, the APHC is insensitive to the d/p ratio. Regarding the PPHC, which describes the maximum power handled during a short time, the SIW reaches 1614 kW, 825 kW and 538 kW at

10 GHz, 20 GHz and 30 GHz respectively considering Rogers 5880 substrate with εr = 2.2, tanδ = 0.0009 and h = 0.508 mm. The PPHC is related to the relative permittivity, the cut-off frequency and the frequency, the effective width, the height and the dielectric strength linked to the substrate as expressed in (1.68). It is concluded in [50] that the greater is the space between via, the higher is the PPHC, and also, the smaller is the via diameter, the larger is the

PPHC.

푢 푑 푎푒푓푓(ℎ푐 + ℎ푐 + 2ℎ푟)(푇푆퐼푊 − 푇∞) 푃 = (1.67) 퐴푃퐻퐶 훼 휔 2 √(1 − ( 푐) ) 1 휔 (1.68) 푃 = √휀 푎 푏퐸2 푃푃퐻퐶 4 푟 377 푒푓푓 0

2.3.3. SIW Topologies

Even if the SIW technology reduces the height of the hollow waveguide, the compactness is mainly determined by the geometry of the structure and thus, the cut-off frequency of the TE10 fundamental mode. The bandwidth is also a critical point and is related

55 by the cut-off frequencies of the TE10 and TE20 modes which correspond to the monomode bandwidth [54]. Different waveguide topologies have been proposed to enhance these two characteristics as illustrated in Fig. 18 and explained in [54]. The Substrate Integrated Folded

Waveguide (SIFW) reduces the size by more than 50% but increases the losses. The Half-

Mode SIW (HF-SIW) is based on the cut of one of the broad metallic walls resulting in a virtual magnetic wall. This leads a reduction of the width of 50% in theory. The two previous topologies can be used together to form the Folded Half-Mode SIW (FHMSIW) [55] which decrease drastically the size by 75% compared to the same structure in conventional SIW.

Substrate Integrated Slab Waveguide (SISW) and Ridge SIW (RSIW) improve the bandwidth.

According to [54], RSIW achieves a 73% bandwidth enhancement and the SISW reaches a

232% bandwidth improvement and a size reduction of 40% compared to the conventional

SIW topology. Recently, a new topology close to RSIW has been presented in [56]. This structure is called Slow Wave SIW (SW-SIW) and it is based on a double-layer substrate stack with a bottom layer including internal metallized via-holes connected to the bottom conductive plane as depicted in Fig. 19. This topology achieves a 40% reduction in size compared to SIW.

Figure 18: Different topologies of SIW: (a) SIFW, (b) HF-SIW, (c) SISW, and (d) RSIW [54].

Figure 19: (a) 3-D view and (b) cross section view of the SW-SIW [56].

2.3.4. SIW Circuits

As the SIW largely inherits its properties from rectangular waveguide, many active and passive components and systems made in conventional waveguide technology can be implemented using the PCB process in SIW technology.

Concerning the passive structures, most of the conventional waveguide components have been redesigned based on SIW technologies. Among the passive components, the filters have been widely explored [51]-[53]. In [54], a general review of filters is discussed and some topologies are presented as illustrated in Fig. 20. In [57], an overview of bandpass filters using SIW is made with a comparison between several results from different research works.

An invited paper [58] has recently published a general overview on filtering structure. On the one hand, the authors focused on the filters synthesis with finite dissipation loss and the synthesis of selective combiners and multiport multiplexers. On the other hand, they describe the design of different topologies of filter such as bandstop filters, filters using acoustic wave resonators, pseudo-elliptic filter and at the end, they deal with SIW filters and multiplexers. In

[59], a complete SIW Butler matrix of 4x4 ports is demonstrated at 60 GHz. It is composed of several 90° hybrid couplers, cross-overs and phase shifters as depicted in Fig. 21. This circuit

57 shows the potential of SIW technology for front-end modules at millimeter-wave frequencies and beyond.

Figure 20: Several topologies of filter: (a) Filter with asymmetric inductive posts, (b) Filter with symmetric inductive posts, (c) Filter with circular cavities, and (d) Filter with rectangular cavities and cross-coupling [54].

(a) (b) Figure 21: (a) Block diagram and (b) picture of the 4x4 Butler matrix [59].

There are few active circuits based on SIW. This is essentially due to the fact that SIW is a buried transmission line compared to microstrip or CPW. Nevertheless, more and more researchers focus on a complete integration called System-on-Substrate (SoS) encouraged by new manufacturing and design possibilities. The realization of active SIW systems consists mainly of the integration of active devices such as transistors or diodes. The top or bottom copper sheet of the SIW is often processed to host microstrip or CBCPW on which active devices are implemented. In [60] and [61], oscillators have been proposed in 2003 and 2014.

In [62] and [63], mixers have been proposed in 2006 and 2012. In [64] and [65], amplifiers have been proposed in 2008 and 2012.

Finally, antennas are one of the key components for transceiver systems. Examples of

SIW antennas are given in Fig. 22. Using a substrate with a high permittivity, antenna size

58 can be decreased compared to the same structure based on hollow waveguide. Hence, SIW antennas are appropriated for systems that require compactness and efficiency. As shown in

[66] for an ALTSA antenna, in [67] for a horn antenna and in [68] for a leaky-wave antenna, there are several existing topologies of antennas that have been demonstrated based on SIW.

They can also form an array to enhance performance or create the required conditions for beam forming or beam steering as demonstrated for example in [69].

(a) (b)

3. CONCLUSION

In this chapter, a chronological overview of transmission line has been presented. The performances of each transmission line are compared in Table II. Also, the different technologies are compared in Fig. 23. It can be inferred from Table II and Fig. 23 that SIW is the most promising technology as its performance combines the advantages from waveguides and planar transmission lines. Nonetheless, the losses still remain high and the power average capability is still too low for some critical applications. An alternative SIW based technology must be introduced to respond to market needs in terms of loss and power handling capability.

Power handling Self- Self- Loss Compactness Cost capability shielded packaged

Waveguide

Planar Transmission line

SIW

Table II: Transmission lines performance comparison.

Figure 23: Comparison of insertion loss, cost and size of different resonator technologies [70].

REFERENCES [1] http://www.scientificamerican.com/article/graphic-science-electromagnetic- bombardment/ [2] Pozar, D. M., Microwave Engineering, John Wiley and Sons, (New York, 1998). [3] Y. Okumura and al., "Field strength and its variability in UHF and VHF land-mobile radio service," Rev. Elec. Commun. Lab.. vol. 16, 1968. [4] C. A. Balanis, Advanced Engineering Electromagnetics, Second Edition, John Wiley and Sons, 2012. [5] Packard K.S., "The origin of waveguides: A case of multiple rediscovery," IEEE Trans. Microwave Theory Tech., 1984. - vol. MTT-32, N 9.- P. 961-969. [6] T.N. Anderson, "Rectangular and ridge waveguide," IRE Trans. Microwave Theory Tech., vol. MTT-4, pp. 201-209, 1956. [7] http://www.microwaves101.com/encyclopedias/waveguide-loss [8] T. Koryu Ishii, Handbook of Microwave Technology, Volume I: Components and Devices, Academic Press, (San Diego, 1995). [9] R. M. Barrett, "Microwave printed circuits—The early years," IEEE Trans. Microwave Theory Tech., vol. MTT-32, pp. 983-990, 1984. [10] R. M. Barrett and M. H. Barnes, "Microwave printed circuits," IRE National Conf. on Airborne Electrics, Dayton, OH, May 23-25, 1951. [11] R. M. Barrett and M. H. Barnes, "Microwave printed circuits," Radio and T.V. News, vol. 46, p. 16, Sept. 1951. [12] R. M. Barrett, "Etched sheets serve as microwave components," Electronics, vol. 25, p. 114-118, June 1952. [13] R. M. Barrett, "Microwave printed circuits-A historical survey," IRE Trans. Microwave Theory Tech., vol. MTT-3, p. 1-9, Mar. 1955. [14] Denlinger E. J., and al., "Losses of microstrip lines," IEEE Transactions on Microwave Theory and Techniques, vol. 28, no. 6 (1980). [15] M. D. Abouzahra and L. Lewin, "Radiation from microstrip discontinuities," IEEE Trans. Microwave Theory Tech., vol. MTT-27, pp. 722-723, Aug. 1979. [16] C. P. Wen, "Coplanar waveguide: A surface strip transmission line suitable for nonreciprocal gyromagnetic device applications," IEEE Trans. Microwave Theory Tech.,vol. 17, no. 12, pp. 1087—1090,Dec. 1969. [17] Simons R. N., Coplanar Waveguide Circuits, Components, and Systems, John Wiley and Sons, Avril 2001. [18] J. L. B. Walker, "A survey of european activity on coplanar waveguide," in 1993 IEEE MTT-S Int. Microwave Symp. Dig., vol. 2, Atlanta, GA, pp. 693—696, June14—18, 1993. [19] C. Veyres and V. Fouad Hanna, "Extension of the application of conformal mapping techniques to coplanar lines with finite dimensions," Int. J. Electron., vol. 48, no. 1, pp. 47—56, Jan. 1980. [20] G. Ghione and C. U. Naldi, "Parameters of coplanar waveguides with lower ground plane," Electron. Lett., vol. 19, no. 18, pp. 734—735, Sept. 1983. [21] G. Ghione and C. U. Naldi, "Coplanar waveguides for MMIC applications: effect of upper shielding, conductor backing, finite-extent ground planes, and line-to-line coupling," IEEE Trans. Microwave Theory Tech., vol. 35, no. 3, pp. 260—267, March 1987. [22] S. S. Bedair and I. Wolff, "Fast and accurate analytic formulas for calculating the parameters of a general broadside — coupled coplanar waveguide for (M)MIC

applications," IEEE Trans. Microwave Theory Tech., vol. 37, no. 5, pp. 843—850, May 1989. [23] Y.-C. Wang and J. A. Okoro, "Impedance calculations for modiﬁed coplanar waveguides," Int. J. Electron., vol. 68, no. 5, pp. 861—875, May 1990. [24] N. I. Dib, W. P. Harokopus Jr., L. P. B. Katehi, C. C. Ling, and G. M. Rebeiz, "Study of a novel planar transmission line," 1991 IEEE MTT-S Int. Microwave Symp. Dig., vol. 2, pp. 623—626, Boston, MA, June 10—14, 1991. [25] N. I. Dib, W. P. Harokopus Jr., G. E. Ponchak, and L. P. B. Katehi, "A comparative study between shielded and open coplanar waveguide discontinuities," Int. J. Microwave Millimeter-Wave Computer-Aided Eng., vol. 2, no. 4, pp. 331—341, Oct. 1992. [26] F. Schnieder, T. Tischler, and W. Heinrich, "Modeling dispersion and radiation characteristics of conductor-backed CPW with finite ground width," IEEE Transactions on Microwave Theory and Techniques, vol. 51, 2003. [27] A.-L. Franc, E. Pistono, N. Corrao, D. Gloria, and P. Ferrari, "Compact high-Q, low- loss mmW transmission lines and power splitters in RF CMOS technology," in IEEE MTT-S Int. Microw. Symp. Dig., Jun. 2011, pp. 1–4. [28] A. L. C. Serrano, A. L. Franc, D. P. Assis, F. Podevin, G. P. Rehder, N. Corrao, and P. Ferrari, "Modeling and characterization of slow-wave microstrip lines on metallic- nanowire-filled-membrane substrate," IEEE Transactions on Microwave Theory and Techniques, vol. 62, 2014. [29] A. Sayag, D. Ritter and D. Goren, "Compact modeling and comparative analysis of silicon chip slow wave transmission lines with slotted bottom metal ground planes," IEEE Trans. Microw. Theory Techn., vol. 57, no. 4, pp. 840-847, 2009. [30] R. W. Jackson, "Coplanar waveguide vs. microstrip for millimeter wave integrated circuits," IEEE MTT-S Symp. Dig., vol. 86, no. 1, pp. 699-702, 1986. [31] I. Bahl and K. Gupta, "Average power-handling capability of microstrip lines," IEE J. Microw. Opt. Acoust., vol. 3, no. 1, pp. 1-4, 1979. [32] M.A. Sánchez-Soriano, Y. Queré, V. Le Saux, C. Quendo, and S. Cadiou, "Average power handling capability of microstrip passive circuits considering metal housing and environment conditions“, Components, Packaging and Manufacturing Technology, IEEE Transactions on, pp. 1624 – 1633, vol. 4, Issue: 10, Oct. 2014. [33] F. Shigeki,“Waveguide line,“ Japanes patent: JP 06 053711, 1994. [34] K. Leahy, B. Flanick, K. A. Zak, and A. Piloto, "Waveguide filters having a layered dielectric structures," Jan. 1995. [35] N. Marcuvitz, Ed., Waveguide Handbook, M.I.T. Rad. Lab. Ser., vol 10. New York: McGraw-Hill 1951. [36] Y. Leviatan, P. G. Li, A. T. Adams, and J. Perini, "Single-post inductive obstacle in rectangular waveguide," IEEE Trans. Microwave Theory Tech., vol. MTT-31, pp. 806-8012, Oct. 1983. [37] R. F. Harrington, Field Computation by Moment Methods, ser. IEEE Press/OUP Electromagnetic Wave Theory, Piscataway, NJ: IEEE Press, 1993. [38] P. Li, A. Adams, Y. Leviatan, and J. Perini, "Multiple-post inductive obstacles in rectangular waveguide," IEEE Trans. Microwave Theory Tech., vol MTT-32, no. 4, pp. 365-373, Apr. 1984. [39] J. Hirokawa and M. Ando, "Single-layer feed waveguide consisting of posts for plane TEM wave excitation in parallel plates," IEEE Transactions on Antennas and Propagation, vol. 46, no. 5, pp. 625-630, May 1998.

[40] H. Uchimura, T. Takenoshita and M. Fujii, "Development of the laminated waveguide," IEEE MTT-S Int. Microw. Symp. Digest, pp. 1811-1814, 1998. [41] C.-K. C. Tzuang, K.-C. Chen, C.-J. Lee, C.-C. Ho and H.-S. Wu, "H-plane mode conversion and application in printed microwave integrated circuit," Proc. 30th Europ. Microw. Conf., pp. 1-4, 2000. [42] D. Deslandes and Ke Wu, "Integrated microstrip and rectangular waveguide in planar form," IEEE Microwave Compon. Lett., vol. 11, pp. 68-70, Feb 2001. [43] H. Moravec, "When will computer hardware match the human brain?," Journal of Evolution and Technology, vol. 1, 1998. [44] K. Wu, D. Deslandes, and Y. Cassivi, "The substrate integrated circuits - a new concept for high-frequency electronics and optoelectronics," TELSIKS'03, pp. P-III - P-X, Nis, Yugoslavia, Oct. 1-3, 2003. [45] M. Bozzi, L. Perrigrini, K. Wu, and P. Arcioni, "Current and future research trends in substrate integrated waveguide technology," Radio Engineering, vol. 18, no. 2, June 2009. [46] D. Deslandes and K. Wu, "Accurate modeling, wave mechanism, and design consideration of a substrate integrated waveguide," IEEE Trans. Microw. Theory Tech., vol. 54, no. 6, pp. 2516-2526, Jun. 2006. [47] Y. Cassivi, L. Perregrini, P. Arcioni, M. Bressan, K. Wu, and G. Conciauro, "Dispersion characteristics of substrate integrated rectangular waveguide," IEEE Microwave Wireless Compon. Lett., vol. 12, no. 9, pp. 333-335, 2002. [48] F. Xu and K. Wu, "Guided-wave and leakage characteristics of substrate integrated waveguide," IEEE Trans. Microw. Theory Tech., vol. 53, pp. 66-73, 2005. [49] M. Bozzi, L. Perregrini, and K. Wu, "Modeling of conductor, dielectric and radiation losses in substrate integrated waveguide by the boundary integral-resonant mode expansion method," IEEE Transactions on Microwave Theory and Techniques, vol. MTT-56, no. 12, pp. 3153-3161, Dec. 2008. [50] Y. J. Cheng, K. Wu and W. Hong, "Power handling capability of substrate integrated waveguide interconnects and related transmission line systems," IEEE Trans. Adv. Packag., vol. 31, pp. 900-909, 2008. [51] A. E. Mostrah and al., "C-Band inductive post SIW alumina filter for a space application. Experimental analysis of the thermal behavior," 2010 Asia-Pacific Microwave Conference, Yokohama, 2010, pp. 103-106. [52] B. Potelon, J. F. Favennec, C. Quendo, E. Rius, C. Person and J. C. Bohorquez, "Design of a substrate integrated waveguide (SIW) filter using a novel topology of coupling," in IEEE Microwave and Wireless Components Letters, vol. 18, no. 9, pp. 596-598, Sept. 2008. [53] J. Garreau and al., "Ultra-compact X-band SIW filter in LTCC technology using high permittivity substrate for a space application," Microwave Symposium Digest (MTT), 2012 IEEE MTT-S International, Montreal, QC, Canada, 2012, pp. 1-3. [54] M. Bozzi, A. Georgiadis, and K. Wu, "Review of substrate-integrated waveguide circuits and antennas," IET Microw. Antennas Propag., 2011, vol. 5, Iss. 8, pp. 909- 920. [55] G. H. Zhai , W. Hong , K. Wu , J. X. Chen , P. Chen , J. Wei and H. J. Tang, "Folded half mode substrate integrated waveguide 3 dB coupler," IEEE Microw. Wireless Compon. Lett., vol. 18, no. 8, pp. 512-514, 2008. [56] A. Niembro-Martin, V. Nasserddine, E. Pistono, H. Issa, A.-L. Franc, T.-P. Vuong and P. Ferrari, "Slow-wave substrate integrated waveguide," IEEE Trans. Microw. Theory Techn., vol. 62, no. 8, pp. 1625-1633, 2014.

[57] S. S. Sabri, B. H. Ahmad and A. R. Bin Othman, "A review of substrate integrated waveguide (SIW) bandpass filter based on different method and design," 2012 IEEE Asia-Pacific Conf. on Appl. Electromagn., pp. 210-215. [58] R. V. Snyder, A. Mortazawi, I. Hunter, S. Bastioli, G. Macchiarella, and K. Wu, "Present and future trends in filters and multiplexers," IEEE Transactions on Microwave Theory and Techniques, vol. 63, no. 10, pp. 3324-3360, Oct. 2015. [59] C.-J. Chen and T.-H. Chu, "Design of a 60-GHz substrate integrated waveguide Butler matrix—A systematic approach," IEEE Trans. Microw. Theory Tech., vol. 58, no. 7, pp. 1724, 2010. [60] Y. Cassivi and K. Wu, "Low cost microwave oscillator using substrate integrated waveguide cavity," IEEE Microwave and Wireless Component Letters, vol. 13, no. 2, Feb. 2003, pp. 48-50. [61] J. Long and Q. Kewei, "A novel x-band oscillator using substrate integrated waveguide resonators," Electronic Packaging Technology (ICEPT), Aug. 2014, pp. 1359 – 1361. [62] J.-X. Chen, W. Hong, Z.-C. Hao, H. Li and K. Wu, "Development of a low cost microwave mixer using a broadband substrate integrated waveguide (SIW) coupler," IEEE Microw. Wireless Compon. Lett., vol. 16, no. 2, pp. 84-86, 2006. [63] Z.-Y. Zhang, Y.-R. Wei and K. Wu, "Broadband millimeter-wave single balanced mixer and its applications to substrate integrated wireless systems," IEEE Trans. Microw. Theory Techn., vol. 60, no. 3, pp. 660-669, 2012. [64] M. Abdolhamidi and M. Shahabadi, "X-band substrate integrated waveguide amplifier," IEEE Microw. Wireless Compon. Lett., vol. 18, no. 12, pp. 815-817, 2008. [65] Z. Wang, S. Adhikari, D. Dousset, C.-W. Park and K. Wu, "Substrate integrated waveguide (SIW) power amplifier using CBCPW-to-SIW transition for matching network," IEEE MTT-S Int. Microw. Symp. Dig., 2012. [66] N. Ghassemi and K. Wu, "Planar high-gain dielectric-loaded antipodal linearly tapered slot antenna for E- and W-Band gigabyte point-to-point wireless services," IEEE Transactions on Antennas and Propagation, vol. 61, no. 4, pp. 1747-1755, April 2013. [67] Y. Cai, "Compact wideband SIW horn antenna fed by elevated-CPW structure," IEEE Trans. Antennas Propag., vol. 63, no. 10, pp. 4551-4557, 2015. [68] J. Liu, D. R. Jackson, and Y. Long, "Substrate integrated waveguide (SIW) leaky- wave antenna with transverse slots," IEEE Trans. Antennas Propag., 2012, vol. 60, no. 1, pp. 20-29. [69] K. Wu, "Multi-dimensional and multi-functional substrate integrated waveguide antennas and arrays for GHz and THz applications: An emerging disruptive technology," European Conference on Antennas and Propagation (EuCAP), pp.11-15, Apr. 2013. [70] X. P. Chen and K. Wu, "Substrate integrated waveguide (SIW) filter: Basic design rules and structure features," IEEE Microw. Mag., vol. 15, no. 5, pp. 108-116, 2014.

CHAPTER III: THE AIR-FILLED SIW TECHNOLOGY

TABLE OF CONTENTS TABLE OF CONTENTS...... 66 LIST OF FIGURES ...... 67 LIST OF TABLES ...... 69 1. AIR-FILLED SIW TECHNOLOGY ...... 70

1.1. AIR-FILLED SIW GEOMETRY ...... 71 1.2. BASIC ELEMENTS ...... 74 1.2.1. Transition from SIW to AFSIW ...... 75 1.2.1.1. Geometry ...... 75 1.2.1.2. Fabrication Tolerances ...... 77 1.2.1.3. Measurements ...... 79 1.2.2. Bends ...... 81 1.2.3. Thru-Reflect-Line Calibration Kit...... 83 1.2.4. Conclusion...... 85 2. AIR-FILLED SIW PERFORMANCE...... 85

2.1. LOSSES ...... 85 2.1.1. Dielectric Losses ...... 86 2.1.2. Conduction Losses...... 86 2.1.3. Surface Roughness Effect ...... 87 2.1.4. Transmission Loss Comparison ...... 88 2.1.5. Experimental Validation ...... 89 2.1.6. Further Improvements ...... 90 2.2. POWER HANDLING CAPABILITY ...... 92 2.2.1. Average Power Handling Capability ...... 93 2.2.2. Peak Power Handling Capability ...... 94 2.3. CONCLUSION ...... 96 3. CONCLUSION ...... 97 REFERENCES ...... 99

LIST OF FIGURES

FIGURE 1: AIR-FILLED (A) CROSS SECTIONAL VIEW AND (B) STRUCTURE...... 72 FIGURE 2: AFSIW E-FIELD FOR (A) W = 0 MM AND (B) W = 0.508 MM...... 73

FIGURE 3: THEORETICAL AFSIW FC10 CUT-OFF FREQUENCY VERSUS W FOR W1 = 7.04MM...... 73 FIGURE 4: COMPARISON OF THE PHASE CONSTANT FOR AFSIW AND RWG USING STANDARD WR-28...... 74 FIGURE 5: GEOMETRY OF THE PROPOSED TRANSITION WITH CROSS SECTIONAL VIEW...... 75

FIGURE 6: SIMULATED |S11| AND |S21| FOR DIFFERENT LENGTH AT KA-BAND...... 76 FIGURE 7: SIMULATED |S11| AND |S21| FOR DIFFERENT LENGTH AT U-BAND...... 76 FIGURE 8: GEOMETRY OF SYMMETRIC (A) AND NON-SYMMETRIC (B) AFSIW...... 78 FIGURE 9: EFFECT OF MISALIGNMENT ON THE S-PARAMETERS IN (A) AND A ZOOM ON |S21| (B) OF THE TRANSITION IN KA-BAND...... 78 FIGURE 10: EFFECT OF THE MISALIGNMENT ON THE BANDWIDTH OF THE TRANSITION AT KA- BAND (A) AND AT U-BAND (B) WITH E-FIELD MAGNITUDE DISTRIBUTION (C) SIMULATED AT 38.5 GHZ FOR A MISALIGNMENT OF Δ = 0.254 MM...... 79 FIGURE 11: PICTURE OF THE FABRICATED INNER SUBSTRATE (A) OF THE BACK-TO-BACK TRANSITION BEFORE MULTILAYER PROCESSING WITH L1 = 5 CM AND L2 = 1 CM, AND (B) ITS E-FIELD MAGNITUDE DISTRIBUTION SIMULATED AT 33 GHZ...... 79 FIGURE 12: MEASURED AND SIMULATED |S11| AND |S21| OF THE BACK-TO-BACK TRANSITION AT KA-BAND...... 80

FIGURE 13: MEASURED AND SIMULATED |S11| AND |S21| OF THE BACK-TO-BACK TRANSITION AT U-BAND WITH L1 = 1.84 MM AND L2 = 0.84 MM...... 81 FIGURE 14: GEOMETRY OF THE RIGHT-ANGLE (A) AND CHAMFERED (B) ANGLE BENDS EXTRACTED FROM [20]...... 82 FIGURE 15: GEOMETRY AND E-FIELD OF EACH BEND TOPOLOGY IN AFSIW TECHNOLOGY: (A) RIGHT-ANGLE, (B) CHAMFERED ANGLE, AND (C) ROUNDED ANGLE...... 83 FIGURE 16: SIMULATED S-PARAMETER FOR EACH BEND TOPOLOGY IN AFSIW TECHNOLOGY: (A) RIGHT-ANGLE, (B) CHAMFERED ANGLE, (C) ROUNDED ANGLE, AND (D) TRANSMISSION LOSS COMPARISON...... 83 FIGURE 17: BLOCK DIAGRAMS FOR A TYPICAL TRL CALIBRATION KIT DE-EMBEDDING PROCEDURE: (A) THE MEASUREMENT PROCEDURE, (B) A THRU STANDARD CONNECTION, (C) A LINE STANDARD CONNECTION, AND (D) A REFLECT STANDARD CONNECTION...... 84 FIGURE 18: PICTURE OF REFLECT (A), THRU (B) AND LINE (C) STANDARDS OF THE AFSIW CALIBRATION KIT [15]...... 85 FIGURE 19: HAMMERSTAD-JENSEN COEFFICIENTS VERSUS FREQUENCY FOR DIFFERENT SURFACE ROUGHNESS...... 87 FIGURE 20: THEORETICAL AND SIMULATED TOTAL TRANSMISSION LOSSES FOR AIR-, ROGERS 5880-, AND 6002- FILLED SIWS AT KA- (A) AND U-BAND (B)...... 89 FIGURE 21: MEASURED AND SIMULATED INSERTION LOSSES OF AIR- AND 6002 FILLED SIW AT KA- (A) AND U-BAND (B)...... 90 FIGURE 22: PICTURE OF (A) AFSIW AND (B) ROGERS 6002 SIW STRUCTURES...... 90

FIGURE 23: UNPOLISHED (A) AND POLISHED (B) FR-4 SUBSTRATE COPPER SURFACE...... 91 FIGURE 24: MEASURED AND SIMULATED |S11| AND |S21| USING UNPOLISHED (A) AND POLISHED (B) SURFACE...... 91 FIGURE 25: THEORETICAL, SIMULATED, AND MEASURED TRANSMISSION LOSSES DIFFERENCE BETWEEN THE AFSIW STRUCTURES BASED ON UNPOLISHED AND POLISHED TOP AND BOTTOM COPPER SURFACES...... 92 FIGURE 26: THEORETICAL APHC FOR AFSIW COMPOSED OF DIFFERENT SUBSTRATE MATERIALS AND ROGERS 5880 AND 6002 SIW IN KA-BAND (A) AND U-BAND (B)...... 94 FIGURE 27: THEORETICAL PPHC FOR AFSIW AND ROGERS 5880 AND 6002 SIW OVER KA- BAND...... 95

LIST OF TABLES

TABLE I: DIMENSIONS OF AFSIW STRUCTURE FOR DIFFERENT ADDRESSED FREQUENCY BAND.72 TABLE II: MATERIAL PROPERTIES...... 88 TABLE III: THEORETICAL LOSS CONTRIBUTION FOR 0.508 MM THICK AFSIW...... 90 TABLE IV: THEORETICAL PROPERTY COMPARISON OF SIW OF SAME HEIGHT OPERATING AT KA- BAND (VALUES GIVEN AT 33 GHZ)...... 96 TABLE V: THEORETICAL PROPERTY COMPARISON OF SIW OF SAME HEIGHT OPERATING AT U- BAND (VALUES GIVEN AT 50 GHZ)...... 97 TABLE VI: PERFORMANCE COMPARISON OF DIFFERENT TRANSMISSION LINES...... 98

1. AIR-FILLED SIW TECHNOLOGY

Since several decades, there is a huge interest for technological breakthroughs in microwave engineering [1]. Today’s emerging millimeter-wave applications including high- resolution imaging, high-speed wireless data links and short-range radar applications are in demand for low-cost and high performance integrated circuits. Moreover, the great demand in compact and light weight technology to replace the bulky metallic Rectangular Waveguide

(RWG) for nano- and pico-satellite is growing [2].

One of the most popular and promising technology that received a particular attention from the scientific community during the last decade is the Substrate Integrated Waveguide

(SIW) [3]. SIW is a branch of the Substrate Integrated Circuits (SICs) family which includes

Substrate Integrated Slab Waveguide (SISW) [4], Substrate Integrated Non-Radiating

Dielectric (SINRD) guide [5], Substrate Integrated Image Dielectric Guide (SIIDG) [6],

Substrate Integrated Inset Dielectric Guide (SIINDG), Substrate Integrated Insular Guide

(SIIG), Ridge Substrate Integrated Waveguide (RSIW) [7] and Slow-Wave Substrate

Integrated Waveguide (SW-SIW) [8]. Hence, SIW provides a light weight, compact, low-cost, self-shielded and integrated alternative to conventional bulky waveguide and planar Printed

Circuit Board (PCB) transmission lines including microstrip, Coplanar Waveguide (CPW) and Conductor-Backed CPW (CBCPW) [9]. It has been adapted on many substrates such as textile [10], paper [11], and flexible materials [12]. The main limitation of SIW is its losses that still remain relatively high compared to RWG, mainly due to its dielectric loss.

Recently, in [13], we proposed a new topology of SIW. As the Air-Filled SIW (AFSIW) inherits largely its properties and its structure from SIW and rectangular waveguide, all the manufacturing processes and theory from SIW can be used and extended to air-filled topology as realized in [14]. In this chapter, the geometry of the AFSIW is described and equations related to its structure are given. Then, some basic elements such as bends and transitions

70 based on AFSIW are studied and presented. Finally, AFSIW performance in terms of loss and power handling capability are described. Material presented in this chapter has been published in two international conference papers [13] and [15], and one journal [14].

1.1. Air-Filled SIW Geometry

The AFSIW structure is based on multilayer PCB process as illustrated in Fig. 1. The air part is bounded inside the substrate 2 by two metalized via rows and enclosed by substrate 1 and 3 to insure the conducting boundaries. A rectangular shape for the vias has been chosen to avoid calculation of effective width and thus, the broadside dimension of SIW is approximately equal to the spacing between the inner walls of vias as explained in [16]. This assumption is of critical importance for enhancing simulation speed using electromagnetic simulators and, hence, reduces the necessary time for designing a complete structure.

Therefore, two continuous metallic walls can be implemented on electromagnetic software instead of metallic via-slot arrays.

The cut-off frequencies for TEmn mode for the classical SIW can be mainly calculated using the equation (1.1) extracted from the RWG theory.

푐 푚휋 2 푛휋 2 푓 = √( ) + ( ) (1.1) 푐푚,푛 2휋√휀푟 푊 ℎ

If the two dielectric slabs are small enough (as explained later), equation (1.1) provides a good approximation. If not, these slabs have to take into account to determine the cut-off frequency of the fundamental TE10 mode. Hence, equation (1.2) extends (1.1) for any dimension of slabs in the case of the fundamental mode (detailed calculations are given in

Annex 1). For a given frequency band, the cut-off frequency is generally chosen 1.25 times the useful minimum frequency to avoid the dispersion region that occurs in waveguide structures. To solve equation (1.2) and obtain the dimensions of the AFSIW structure, the

Newton-Raphson method of iteration is used.

√휀푟 ∗ (푊1 − 푊2 ) ∗ 휋푓푐 푊2 ∗ 휋푓푐 푡푎푛 ( 10 ) = √휀 푐표푡 ( 10 ) (1.2) 푐 푟 푐

The air-cut and the via holes are made using a laser micro-machining process with a resolution of 50.8 µm. Due to manufacturing limitation issues and mechanical integrity, dielectric slabs must remain on each side of the air section with a minimum width of

0.254 mm for the Poly-Grames Research Center PCB process. The dimensions of an AFSIW structure operating in different frequency-band are calculated using (1.2) and the results are summed up in Table I. The cut-off frequency of the fundamental mode fc10 is chosen such that the lowest band frequency corresponds to 1.25* fc10 in order to avoid the high dispersive region occurring near the cut-off frequency [17].

Substrate 3 W1 wvia

Substrate 2 h Substrate 1

w w Copper Substrate Air

(a) (b) Figure 1: Air-filled (a) cross sectional view and (b) structure.

Frequency band (GHz) Ka (26-40) U (40-60)

Ratio w/W1 (%) 7.22 11.04

Dimensions (mm) Space between two 0.4 0.381 vias

Via width (wvia) 0.381 0.381 Table I: Dimensions of AFSIW structure for different addressed frequency band.

The influence of both dielectric slabs for small value of w is negligible on the behavior of the fundamental mode TE10 as its E-filed is mainly confined in the center of the structure.

As shown in Fig. 2, the E-field of the fundamental mode is weak next to the via rows and it is not altered by the presence of the dielectric slab. This is confirmed by Fig. 3 illustrating the

72 theoretical fc10 of AFSIW versus the width w for a constant W1 obtained from (1.2). It can be noticed from Fig. 3 that the cut-off frequency only shifts of 4.7% for a slab width of w = 2 mm which corresponds to a quantity of more than 56% of dielectric material inside the

AFSIW.

(a) (b) Figure 2: AFSIW E-field for (a) w = 0 mm and (b) w = 0.508 mm.

(GHz)

10 c

f 15

10 0 1 2 3 4 5 6 7 2w (mm)

Figure 3: Theoretical AFSIW fc10 cut-off frequency versus w for W1 = 7.04mm.

Another important parameter is the phase constant β which characterizes the phase for a transmission line and the mode, evanescent or propagating. An arrangement consisting of ferrite loaded waveguide has been presented in [18]. In this paper, the characteristic equation for the propagation constant Γ=(α+jβ) is calculated taking into account the ferrite slabs. As

SIW structure inherits its properties from RWG, the characteristic equations depending on frequency can be carried out on AFSIW loaded with two dielectric materials. Hence, the work from [18] and [19] can be extended to calculate the phase constant in a lossless case for simplicity (α = 0). The characteristic equation for β is given in equation (1.3) to (1.7) that can be solved by Newton-Raphson method. In Fig. 4, the phase constant of TE10 mode for AFSIW and RWG designed for the operating Ka-band are theoretically calculated and compared to the simulation. It can be seen that there is a good agreement as shown in Fig. 4.

푝 푡푎푛[푘 (푊 − 2푤)] = 푠 1 푞 (1.3) 퐾 푚 ( ) ( ) where 푝 = 2 ( ) 푐표푠ℎ 퐾푚푤 푠𝑖푛ℎ 퐾푚푤 (1.4) 푘푠 퐾 2 2 ( ) 푚 2( ) 푞 = 푠𝑖푛ℎ 퐾푚푤 − ( ) 푐표푠ℎ 퐾푚푤 (1.5) 푘푠 2 ( )2 2 with 퐾푚 = − 2휋푓 휀푟휀0 µ0 + 훽 (1.6) 2 ( )2 2 푘푠 = 2휋푓 휀0µ0 − 훽 (1.7)

200 Theoritical RWG Simulated RWG 150 Theoritical AFSIW Simulated AFSIW 100

50 Phase constant (rad/m) constant Phase

0 20.5 21 21.5 22 22.5 Frequency (GHz) Figure 4: Comparison of the phase constant for AFSIW and RWG using standard WR-28.

The observed discrepancy between the two cut-off frequencies is mainly due to the difference in width from both waveguide where AFSIW width is set to 7.04 mm and for the

RWG is 7.11 mm as defined for the WR-28 standard.

Therefore, the AFSIW can achieve similar behavior to small-height dielectric loaded

RWG with the advantages of PCB manufacturing process and integration with other planar technologies. The top and bottom layer can use low-cost substrate such as FR-4 to implement base-band digital circuits or interconnect with active circuits so to obtain a very compact and cost effective system.

1.2. Basic Elements

After developing the AFSIW theory, it is required to interconnect this technology for experimental demonstration and integration purposes. This is the first issue to figure out in order to perform measurements with conventional measurement set up. Moreover, it is also

74 interesting to focus on bends. This kind of discontinuities is useful to create compact and dense systems. However, it can generate some leakage, and significantly increase transmission loss.

1.2.1. Transition from SIW to AFSIW

To interconnect AFSIW to SIW circuits, it is necessary to develop a specific transition for measurement purposes and co-integration with other planar technologies. The proposed transition is inspired by a linearly tapered transition of SIW on high-to-low dielectric substrate introduced in [20]. Nevertheless, it has been chosen to optimize the transition by maintaining the cut-off frequency of the fundamental mode nearly constant along the transition to avoid dispersion and achieve wideband operation. In [20], the high dielectric substrate penetrates into the low dielectric material. In the case of this transition, the opposite is adopted because the E-field is mainly concentrated in the center of the structure and hence, to minimize the dielectric losses, the high dielectric substrates are removed to be replaced by low permittivity material (Air).

1.2.1.1. Geometry

Figure 5: Geometry of the proposed transition with cross sectional view.

The designed transition is illustrated in Fig. 5. As shown in Fig. 5, the dimension W1 and W2 are raised along the transition. For designing such transition, there are several steps to follow. First, the design procedure involves a discretization of the transition in several

75 sections. For each section, W1 and W2 parameters are determined by using (1.2) to keep the same cut-off frequency. Then, the distance between each section is optimized in simulation to obtain the desired return loss.

A study on the effect of the transition length on S-parameters has been achieved at Ka- band and U-band based on Rogers RT/Duroid 6002 substrate for several lengths: 5, 10 20 and

30 mm. Fig. 6 shows the |S11| and |S21| in Ka-band and Fig. 7 illustrates the |S11| and |S21| in U- band. It can be seen from Fig. 6 and 7 that a better matching is obtained for longer transitions.

Moreover, the shorter the transition is, the lower the transmission losses are. This leads to a trade-off between the entire length of the transition and the return loss. It can be observed that there is a discontinuity between the Ka- and U-band results. This is mainly due to the cut-off frequency which moves from 21.2 GHz at Ka-band to 32 GHz at U-band and the associated reduction in width.

0 0 L=5 mm -10 L=10 mm L=20 mm -0.1 -20 L=30 mm -0.2 -30

-0.3 |S11|(dB) -40 |S21|(dB) L=5 mm L=10 mm -50 -0.4 L=20 mm L=30 mm -60 -0.5 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) Figure 6: Simulated |S11| and |S21| for different length at Ka-band.

0 0 L=5 mm -0.1 -10 L=10 mm L=20 mm -0.2 -20 L=30 mm -0.3

-30 -0.4 |S11|(dB) |S21|(dB) -0.5 -40 L=5 mm -0.6 L=10 mm -50 L=20 mm -0.7 L=30 mm -60 -0.8 40 42 44 46 48 50 52 54 56 58 60 40 42 44 46 48 50 52 54 56 58 60 Frequency (GHz) Frequency (GHz) Figure 7: Simulated |S11| and |S21| for different length at U-band.

1.2.1.2. Fabrication Tolerances

As the fabrication tolerances is of a great interest for millimeter and sub-millimeter wave since the wavelength approaches the dimensions of the SIW structure, the effect of fabrication tolerances on such transition has been studied. During the manufacturing process, a laser micromachining is used to remove the dielectric part. During this processing step, some tolerances can conduct to a misalignment of the air portion as illustrated in Fig. 8.

Considering misalignments of value Δ = 0, 0.254 and 0.508 mm (referred to Fig. 8), Fig. 9 shows simulation results of transition S-parameters operating at Ka-band with 20 mm length.

It can be observed from Fig. 9(b) that a misalignment increases the losses at low frequency and some parasitic peaks appear at high frequency beyond 36 GHz that limit the operating bandwidth of the transition.

To physically explain the presence of those peaks, characteristic equation (1.2) has been generalized for non-symmetrical structure with two dielectric slabs and is given in (1.8) (all parameters are shown in Fig. 8) with εr the relative permittivity, k0 the wave number, and a, t and d are geometric parameters shown in Fig. 8.

As previously, equation (1.8) has to be solved using the Newton-Raphson method. At each discretized cross section of transition, the cut-off frequency of the TE10 and TE20 modes can be calculated solving (1.8). As shown in Fig. 10(a) and (b), fc10 and fc20 are plotted in function of the position along the aligned and misaligned transition at both Ka- and U-band. It can be inferred from Fig. 10(a) and (b) that the cut-off frequency of the fundamental mode stays nearly constant and hence, the misalignment weakly impacts this mode. Nevertheless, the cut-off frequency of the second mode presents an inflection point for both frequency bands. In the higher part the operating frequency band, both modes TE10 and TE20 coexist. In the case of perfectly symmetrical structure, the two modes do not interfere with each other.

When an unsymmetrical structure is considered, for example in the case of a misalignment,

77 both modes couple and parasitic peaks are generated within the transition. As a result, mismatch is increased and thereby, the transmission losses rise. It can be seen from Fig. 10(a) and (b) that the inflection point is around 34 GHz and 52 GHz for Ka-band and U-band, respectively. These frequencies represent the lower limits beyond which the phenomenon can be created because of manufacturing process tolerances. A picture of a misaligned transition in Fig. 10(c) extracted from an electromagnetic simulator shows the phenomenon at 38.5 GHz where the TE20 mode can be observed.

푡푎푛⁡(√휀푟푘0(푎 − 푡))푡푎푛(푘0(푡 − 푑)) − 1 √휀푟 (1.8) √휀푟푡푎푛(푘0(푡 − 푑)) + 푡푎푛⁡(√휀푟푘0(푎 − 푡)) = 푡푎푛(√휀푟푘0푑)

(a) (b) Figure 8: Geometry of symmetric (a) and non-symmetric (b) AFSIW.

0 -0.2 |S11| for =0 mm -10 |S11| for =0.254 mm -0.225 |S11| for =0.508 mm -20 |S21| for =0 mm -0.25 |S21| for =0.254 mm -30 |S21| for =0.508 mm -0.275

-40 |S21|(dB) -0.3

|S11| and |S21| (dB) |S11|and =0 mm -50 -0.325 =0.254 mm =0.508 mm -60 -0.35 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) (a) (b) Figure 9: Effect of misalignment on the S-parameters in (a) and a zoom on |S21| (b) of the transition in Ka-band.

44 65

40 60 55 36 f for =0.254 mm

c10 (GHz) (GHz) f for =0 mm 50

c10

c20 c20

f 32 f for =0.254 mm f for =0.254 mm c20 45 c10

f for =0 mm f and

and and f for =0 mm 28 c20 c10

c10 40 c10

f f for =0.254 mm f c20 f for =0 mm 24 35 c20

20 30 0 5 10 15 20 0 5 10 15 20 Position from the air-filled SIW to Rogers 6002 filled SIW (mm) Position from the air-filled SIW to Rogers 6002 filled SIW (mm) (a) (b)

(c) Figure 10: Effect of the misalignment on the bandwidth of the transition at Ka-band (a) and at U-band (b) with E-field magnitude distribution (c) simulated at 38.5 GHz for a misalignment of Δ = 0.254 mm.

1.2.1.3. Measurements

Using the previously described procedure, a Ka-band back-to-back transition has been fabricated using 0.508 mm thick Rogers RT/Duroid 6002 substrate as the middle layer, and standard FR-4 substrate as substrate 1 and 3 (see Fig. 1) before being mounted using multilayer PCB process. The inner substrate is shown in Fig. 11(a) and the E-field magnitude distribution simulated at 33 GHz is given in Fig. 11(b). The air-cut is realized using a laser micromachining with a precision alignment of 25 µm to avoid parasitic resonances. In order to characterize this prototype, Conductor Backed Coplanar Waveguides (CBCPW) are added at input and output with CBCPW to SIW transitions.

(a)

(b) Figure 11: Picture of the fabricated inner substrate (a) of the back-to-back transition before multilayer processing with L1 = 5 cm and L2 = 1 cm, and (b) its E-field magnitude distribution simulated at 33 GHz.

Measured and Simulated S-parameters are given in Fig. 12. Test fixture using TRL calibration is used to perform measurement on a Vector Network Analyzer (VNA) in order to de-embed effects of connectors and CBCPW to SIW transitions. The simulations are taking into account the dielectric and conductor losses. There is good agreement between simulated and measured results. It can be seen that |S11| is lower than -15 dB from 27 GHz to 40 GHz and the transmission loss is estimated to 0.6 ± 0.2 dB over the Ka-band. It can be deduced from Fig. 12 by neglected the effect of the AFSIW losses that one transition achieves

0.3 ± 0.1 dB transmission loss.

0  0 -10 -1

-20 -2 

-30 -3

|S11|(dB) |S21||S21|(dB) (dB) -40 -4 |S11| meas. |S11| simu. -50 |S21| meas. -5 |S21| simu. -60 -6 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 12: Measured and simulated |S11| and |S21| of the back-to-back transition at Ka-band.

The previous transition has been re-designed using the same procedure to obtain an optimal dispersion and a return loss of less than -20 dB over the U-band for a total length of

5 mm. The simulated and measured results of a back-to-back transition are shown in Fig. 13.

Measured |S11| is lower than -13.5 dB from 40 to 60 GHz. Measured insertion loss (|S21|) is

0.42 ± 0.11 dB over the U-band. It can be deduced, neglecting losses in the AFSIW, that the transition achieves 0.21 ± 0.055 dB transmission losses.

0 0

-0.5 -5  -1 -10 Simulated Measured -1.5 -15 -2

-20

|S21||S21|(dB) (dB) |S11|(dB)  -2.5 -25 -3

-30 -3.5

-35 -4 40 42 44 46 48 50 52 54 56 58 60 Fréquence (GHz) Figure 13: Measured and simulated |S11| and |S21| of the back-to-back transition at U-band with L1 = 1.84 mm and L2 = 0.84 mm.

1.2.2. Bends

Bends are key components used in feeding networks. In [21], different topologies of bends are studied using Electromagnetic Bandgap (EBG) waveguide technology. As EBG waveguides and SIW partially inherit their properties from RWG, a similar comparison of different topologies of bends can be realized using AFSIW technology.

As it is done in [21], several topologies of bends are proposed: right-angle bend, chamfered bend, and rounded bend. The right-angle bend is shown in Fig. 14(a). The critical size is the diagonal of the angle region. This dimension is generally chosen to be equal to the width of the wave guiding structure. Otherwise, the right-angle region can excite the TE20 mode because the cut-off frequency of this mode is decreased, as illustrated in Fig. 14(a).

Hence, it can be coupled with the fundamental TE10 mode as explained in [21]. This results in an important attenuation of the signal. To fix the problem, one of the proposed solutions is to chamfer the angle as depicted in Fig 14(b). Nevertheless, if the angle is too chamfered, the diagonal will be too small and, as a consequence, the cut-off frequency of the fundamental mode will rise. Therefore, the desired operating frequency band is no longer addressed and the transmission losses are increased. That is why, the dimension of the angle is critical and the width at the corner has to be the same that of the uniform wave guiding structure. The same conclusion can be applied on the rounded bend.

(a) (b) Figure 14: Geometry of the right-angle (a) and chamfered (b) angle bends extracted from [21].

Based on [21], the three previous topologies of bend are explored in AFSIW technology and are shown in Fig. 15 with their E-field. The right-angle bend is illustrated in Fig. 15(a).

An inductive post (metalized via) is added in the corner to reduce the width of this region and hence, maintain the same cut-off frequencies for the first two propagating modes (TE10 and

TE20). The chamfered and rounded bends are depicted in Fig. 15(b) and 15(c) respectively.

The previous design step has been used for these two discontinuities. The simulated results are given in Fig. 16 for each topology. It can be seen from Fig. 16 that the performance of each topology of bend is similar. In Fig. 16(d), the transmission loss generated by this discontinuity for the different bends is compared. Similar transmission loss of

0.011 ± 0.004 dB is achieved for all the topologies at Ka-band. It can be noticed from

Fig. 15(a) that the inductive post is in air section and some dielectric substrate has to be used to attach it from the dielectric wall. This dielectric has not been taken into account in the simulated results presented in Fig. 16. As a result, the transmission loss will significantly increase and also, mismatch can be generated due to the variation of the impedance in this region. For all these reasons, this topology of bend will not be selected for interconnect with passive devices as presented in the Chapter IV. It only remains the rounded and chamfered bends. Both solutions are equivalent even if it is easier to match the rounded bend than the chamfered one as shown in Fig. 16(b) and (c). Also, for practical reason, the design of rounded bend is facilitate on software compared to the chamfered one where the diagonal Ldiag

82 has to be calculated or adjusted by optimization. Finally, the rounded bend is preferred to all the other topologies.

(a) (b)

(c) Figure 15: Geometry and E-field of each bend topology in AFSIW technology: (a) right- angle, (b) chamfered angle, and (c) rounded angle.

0 0 S S -5 11 -5 11 -10 S -10 S -15 21 -15 21 -20 -20 -25 -25 -30 -30 -35 -35 -40 -40

-45 -45 S-parameters (dB) S-parameters -50 (dB) S-parameters -50 -55 -55 -60 -60 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) (a) (b) 0 S 0.025 -5 11 Bend with via -10 S Chamfered bend 21 0.02 -15 Rounded bend -20 0.015 -25 -30 0.01 -35

S-parameters (dB) S-parameters -40 0.005 -45 Transmissionloss (dB)

-50 0 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) (c) (d) Figure 16: Simulated S-parameter for each bend topology in AFSIW technology: (a) right- angle, (b) chamfered angle, (c) rounded angle, and (d) transmission loss comparison.

1.2.3. Thru-Reflect-Line Calibration Kit

The TRL calibration was first presented in [22] in 1979 and consists of three standards:

Thru, Reflect and Line connections.

The correction of deviations induced by the network analyzer during measurement process can be corrected using this calibration technique. The deviation is mainly caused by non-idealities inside the network analyzer and the cables used to interconnect the DUT to the analyzer. The TRL calibration procedure is illustrated in Fig. 17 [23].

Reference Planes Error Box B Port 1 Error Box A Port 2 Transition A Transition B

DUT

P P ’ P ’ P 1 1 2 2 (a) Reference Planes Error Box B Port 1 Error Box A Port 2 Transition A Transition B

P P ’ P ’ P 1 1 2 2 (b) Reference Planes Error Box B Port 1 Error Box A Port 2 Transition A Transition B

γl Z0, e P P ’ P ’ P 1 1 2 2 (c) Reference Planes Error Box B Port 1 Error Box A Port 2 Transition A Transition B

P1 P1’ P2’ P2

Reflector (d) Figure 17: Block diagrams for a typical TRL calibration kit de-embedding procedure: (a) The measurement procedure, (b) a Thru standard connection, (c) a Line standard connection, and (d) a Reflect standard connection.

Recently, a calibration kit using SIW technology has been presented [24]. To improve measurement accuracy, an AFSIW TRL calibration kit has been proposed. It allows placing the reference plane inside the AFSIW to de-embed transitions and all elements affecting the

84 response of the DUT. The line is generally designed for a phase shift of 90° compared to Thru at the center frequency of the desired frequency band. Pictures of the calibration kit reflect, thru and line elements are shown in Fig. 18.

(a) (b) (c) Figure 18: Picture of reflect (a), thru (b) and line (c) standards of the AFSIW calibration kit [15].

1.2.4. Conclusion

In this section, several topologies of bends have been presented. The rounded bend has clearly demonstrates is potential to be used in feeding network or passive circuits. It can be also used as interconnect for measurement purposes. A back-to-back transition from SIW to

AFSIW has been demonstrated. Measurements are performed at Ka- and U-band and compared to simulation with good agreement. Such transitions are useful and critical for interconnecting other planar technologies like CBCPW. Finally, a TRL calibration kit using

AFSIW technology is proposed to perform measurements. In the following section, the

AFSIW experimental characterization in terms of losses and power handling capability is introduced.

2. AIR-FILLED SIW PERFORMANCE 2.1. Losses

AFSIW, as the conventional waveguide, has dielectric and conductor losses including the influence of surface roughness effect. Each loss mechanisms will be detailed in the next subsections. It is assumed that the leakage loss is negligible as the width of each via and the spaces between two vias respect the SIW design rules introduced in [26]. Theoretical and simulated SIW transmission loss of are compared to AFSIW transmission loss at Ka- and U-

85 band. Finally, measurements are performed in both frequency bands to validate theoretical and simulation results.

2.1.1. Dielectric Losses

The involved loss mechanism for dielectric inside AFSIW is clearly similar to RWG and can be physically explained as described in Chapter II Section 1.2.1. In SIW topologies,

E- and H-fields propagate in the dielectric medium filling the space between conducting via- holes. The attenuation related to the dielectric loss is obtained using (1.9) [25]. As the air fills the AFSIW structure, the loss tangent is set to zero and this material can significantly improve the transmission loss.

푘2푡푎푛훿 훼푑 = (1.9) 2훽

2.1.2. Conduction Losses

As explained in Chapter II Section 1.2.1., the loss mechanism resulting from the conductor losses in AFSIW is analogous to RWG. The Ohmic loss is caused by surface resistivity Rs of the conductor walls or vias usually plated with copper. It can be obtained from (1.10) [25].

휋푓µ 푅 = √ 0 (1.10) 푠 휎

Then, considering the induced surface current on the inner walls of the structure, for

TE10 mode, the attenuation related to the conductor loss is given by (1.11) [25]. It can be observed from equation (1.11) that the thicker is the AFSIW (h), the lower the conductor loss is. Furthermore, the larger the structure is (W), the lower the ohmic loss is. This phenomenon is explained by the fact that for big structure, the energy is less confined, the density of energy decreases and thus, it generates lower surface current on metallic walls and vias.

2 푅푠 2 3 2 푅푠 2휋 푘 훼푐 = (2ℎ휋 + 푊 푘 ) = ( + ) (1.11) 푊3훽ℎ휂푘 훽휂 푊3푘 ℎ

2.1.3. Surface Roughness Effect

The conductor losses can significantly increase at high frequencies as the skin depth approaches the conductor surface roughness value. From [27] and [28], an empirical formula was developed to take into account the surface roughness of a relatively smooth surface. This is verified in Fig. 19, in which coefficient KHammerstad is calculated for different root mean square (RMS) conductor surface roughness Hrms. However, for rough surface, the

Hammerstad-Jensen model saturates. Hence, a total attenuation due to conductor losses is obtained considering coefficient KHammerstad, multiplying the conductor attenuation αc (1.12).

The coefficient KHammerstad is calculated using the empirical formula (1.13).

훼푐+푟 = 훼푐 ∗ 퐾퐻푎푚푚푒푟푠푡푎푑 (1.12) 2 2 퐻푟푚푠 퐾퐻푎푚푚푒푟푠푡푎푑 = 1 + 푎푟푐푡푎푛 (1.4 ∗ ( ) ) (1.13) 휋 훿

1.8

1.6

Hammerstad H = 0.1 µm

1.4 rms K H = 0.3 µm rms H = 0.6 µm 1.2 rms H = 0.9 µm rms

1 0 5 10 15 20 25 30 35 40 Frequency (GHz) Figure 19: Hammerstad-Jensen coefficients versus frequency for different surface roughness.

For conventional SIW, the inner surface roughness of the copper foil has to be consider

(Hrms_inner = 0.4 µm for Rogers 6002 and 5008). However, for AFSIW structure, the outer surface roughness is taken into account, for substrate 1 and 3 as depicted in Fig. 1

(Hrms_inner = 0.3 µm for Rogers 6002 and 5008). As the outer roughness is better than the inner one, the total conductor losses are reduced.

2.1.4. Transmission Loss Comparison

For comparison purposes at Ka- and U-band, Rogers RT/Duroid 5880 and 6002 SIW, commonly used at millimeter-wave frequencies, with a thickness of h = 0.508 mm, are chosen as references. Properties of these substrates are given in Table II.

Rogers 6002 Rogers 5008 Air

Dielectric constant, εr, @ 2.94 2.2 1 10 GHz Dissipation factor, tanδ, @ 0.0012 0.0009 0 10 GHz

Transition glass temperature, 350 260 / Tg (°C)

Dielectric strength, E0, 239 360 36 (kV/cm) Table II: Material properties.

Considering equation (1.1), two SIW are also designed for addressing Ka-band which leads to widths of 4.77 and 4.11 mm for Rogers 5880 and 6002 respectively. The obtained widths at U-band for SIW based on Rogers 5880 and 6002 are 3.16 and 2.72 mm respectively.

The AFSIW dimensions are given in Table I at Ka- and U-band considering Rogers 6002 as the inner layer. Fig. 20 shows the theoretical and simulated total loss including dielectric, ohmic and surface roughness losses at Ka- and U-band respectively. It can be observed that theoretical and simulated results are in very good agreement. Transmission losses using

AFSIW are substantially reduced by a mean value of 0.068 dB/cm and 0.104 dB/cm compared to 5880- and 6002-filled SIWs, respectively at Ka-band. At U-band, transmission losses using AFSIW are reduced by a mean value of 0.098 dB/cm and 0.152 dB/cm compared to 5880- and 6002-filled SIWs, respectively.

0.2 0.3

Rogers 6002-filled SIW 0.25 Rogers 6002-filled SIW 0.15 0.2

Rogers 5880-filled SIW 0.1 0.15 Rogers 5880-filled SIW

0.1 Air-filled SIW based on Rogers 6002 Air-filled SIW based on Rogers 6002

0.05 TransmissionLoss (dB/cm) TransmissionLoss (dB/cm) 0.05 Theory Theory Simulation Simulation 0 0 26 28 30 32 34 36 38 40 40 45 50 55 60 Frequency (GHz) Frequency (GHz) (a) (b) Figure 20: Theoretical and simulated total transmission losses for air-, Rogers 5880-, and 6002- filled SIWs at Ka- (a) and U-band (b).

2.1.5. Experimental Validation

To validate the theoretical and simulated results, AFSIW and 6002-filled SIW transmission lines have been fabricated and measured at Ka- and U-band using Through-

Reflect-Line (TRL) calibrations to remove cables, connectors and transitions effects.

Theoretical and measured transmission losses are compared in Fig. 21 in both frequency bands. To feed the AFSIW line, the transition from air- to dielectric-filled SIW introduced in a previous section has been used. It can be observed in Fig. 21(a) that theoretical and the measured results are in good agreement. The AFSIW has a total loss of about 0.045 dB/cm, which is less than a third of the total loss of 6002-filled SIW in Ka-band. A picture for both

SIW topologies is given in Fig. 22. Only the inner layer of the AFSIW structure can be seen on

Fig. 22(a). In Fig. 21(b), the measured transmission losses for the 6002- and AFSIW are

0.4 ± 0.13 dB/cm and 0.122 ± 0.122 dB/cm, respectively over the U-band. A fair agreement between simulated and measured results is obtained. Measurements at millimeter-wave frequencies are particularly challenging.

0.2 Rogers 6002-filled SIW 0.8 Simulated Measured 0.15 0.6 Rogers 6002-filled SIW

0.1 0.4

Air-filled SIW based on Rogers 6002 Air-filled SIW

0.05 0.2 TransmissionLoss (dB/cm) Theory Transmissionloss (dB/cm) Measured 0 0 26 28 30 32 34 36 38 40 40 42 44 46 48 50 52 54 56 58 60 Frequency (GHz) Frequency (GHz) (a) (b) Figure 21: Measured and simulated insertion losses of air- and 6002 filled SIW at Ka- (a) and U-band (b).

(a) (b) Figure 22: Picture of (a) AFSIW and (b) Rogers 6002 SIW structures.

2.1.6. Further Improvements

Table III summarizes the dielectric, ohmic and surface roughness losses of an AFSIW of 0.508 mm thickness operating at Ka-band and U-band considering a 0.3 µm surface roughness (typical value for Rogers 6002 substrate). It can be seen that the surface roughness loss provides a non-negligible contribution of about 31% and 39% in the Ka- and U-band, respectively. Employing standard FR-4 substrate for the top and bottom substrates with higher surface roughness can certainly result in a higher conductor loss. To achieve a higher performance, conductor losses can be reduced by polishing the top and bottom substrate

(substrates 1 and 3 on Fig. 1) copper surfaces that forms the AFSIW.

Frequency Ka-band (@ 33GHz) U-band (@ 50 GHz) dB/cm % dB/cm % Dielectric loss 0 0 0 0 Ohmic loss 0.0297 68 0.0375 61 Surface roughness loss 0.0144 32 0.024 39 Total loss 0.0441 100 0.0615 100 Table III: Theoretical loss contribution for 0.508 mm thick AFSIW.

In Fig. 23(a), a standard unpolished FR-4 substrate copper surface profile obtained with an optical profilometer is shown. A maximum peak and RMS roughness of 15.6 µm and

0.898 µm are obtained, respectively. For demonstration purposes, the copper surfaces has been polished to reduce the surface roughness using a lapping and polishing machine with a

1000 grade silicon carbide (SIC) disk. The polished FR-4 substrate offers a maximum peak and RMS roughness of 6.53 µm and 0.389 µm, respectively Fig. 23(b).

Alpha = 45° Beta = 30° µm Alpha = 45° Beta = 30° µm 6.5 15 14 6 13 5.5 15.6 µm 12 5 11 4.5 6.53 µm 10 4 9 3.5 8 7 3 6 2.5

5 2 4 1.5 3 1 2 1.2 mm 3.42 mm 1 0.5 2.68 mm 1 mm 0 0 (a) (b) Figure 23: Unpolished (a) and polished (b) FR-4 substrate copper surface.

For an experimental validation, the AFSIW based on unpolished and polished FR-4 substrates have been fabricated. Both AFSIW S-parameters have been measured by using

AFSIW Through-Reflect-Line (TRL) calibrations. Measured S-parameters are given in

Fig. 24 for the transmission lines based on unpolished and polished FR-4, respectively. Both circuits achieve a reflection coefficient (|S11|) below -15 dB over the entire Ka-band.

0 0 0 0   -5 -0.5 -5 -0.5 S11 simu. S11 simu. -1 -1 -10 -10 S11 meas. S11 meas. S21 simu. S21 simu. -1.5 -1.5 -15 S21 meas. S21 meas. -15 -2 -2

-20 -20 

|S11|(dB) |S21||S21|(dB) (dB)

|S21||S21|(dB) (dB)  -2.5 |S11|(dB) -2.5 -25 -25 -3 -3

-30 -3.5 -30 -3.5

-35 -4 -35 -4 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) (a) (b) Figure 24: Measured and simulated |S11| and |S21| using unpolished (a) and polished (b) surface.

Fig. 25 shows the measured, theoretical and simulated transmission loss difference between the two circuits. Over the Ka-band, theoretical, simulated and measured mean

(averaged) transmission loss improvements are 1.13 dB/m, 0.89 dB/m and 1.56 dB/m, respectively. It can be seen that the theoretical, simulated and measured results are in good agreement. The difference observed for the measured results can be explained by the empirical nature of the Hammerstad-Jensen model. This demonstrates also that it is possible to improve the performance of the AFSIW structure by adding a polishing step during the manufacturing process.

4 Theoritical Simulated Measured

0 26 28 30 32 34 36 38 40 TransmissionLoss (dB/m) Difference Frequency (GHz) Figure 25: Theoretical, simulated, and measured transmission losses difference between the AFSIW structures based on unpolished and polished top and bottom copper surfaces.

2.2. Power Handling Capability

For high power applications, it is important to evaluate the power handling capability of transmission lines. Depending on applications, the power limitation in relation with continuous wave signal and also the peak power limitation that can be induced by pulsed signal or complex communication systems must be known. For space applications, Corona effect and multipaction have to be taken into account during the design step for a component.

The multipaction effect is a vacuum discharge phenomenon produced by a RF field between two conductor surfaces. When this discharge happens, a secondary electron emission is produced, called electron multipaction. In radar or communication systems, Passive

Intermodulation (PIM) must be considered. In this section, the Average Power Handling

Capability (APHC) and Peak Power Handling Capability (PPHC) of AFSIW are theoretically

92 determined in Ka- and U-band. APHC and PPHC are the basic metrics to estimate the power handling capability of a transmission line.

2.2.1. Average Power Handling Capability

As losses introduced in the previous section are transformed into heat, the integrity of the structure may be altered and therefore a power limit exists. To characterize this phenomenon, the APHC is a commonly used metric. It is obtained by a heat transfer analysis.

The APHC is determined by the lowest transition glass temperature of the materials constituting the multilayer PCB structure. As the heat distortion temperature of copper

(1093°C) is higher than the transition glass temperature (Tg) of substrate materials (about several hundreds of °C), the substrate is the limiting material. The APHC can be determined using [29] and is given in equation (1.14):

(ℎ푢 + ℎ푑 + 2ℎ ) ∗ 푊 ∗ (푇 − 푇 ) 푃 = 푐 푐 푟 푔 ∞ (1.14) 퐴푃퐻퐶 훼

푢 푑 where ℎ푐 and ℎ푐 are the convection heat-transfer coefficient for the top and bottom surfaces of the SIW, respectively, ℎ푟 is the radiation heat-transfer coefficient, W is the SIW total width, Tg is the lowest transition glass temperature of the materials constituting the SIW,

T∞ is the room temperature, and α (Np/m) is the total transmission loss, including the influence of surface roughness. Using (1.14), the APHC of both reference SIW and the

AFSIW considering a multilayer structure composed of Rogers 6002 and low-cost standard and Hi-Tg FR-4 as substrates 1 and 3 (Fig. 1) are compared in Fig. 26(a) at Ka-band and in

Fig. 26(b) at U-band.

55 52 Rogers 6002-filled SIW (Tg = 350°C) 53 Rogers 5880-filled SIW (Tg = 260°C) Rogers 6002-filled SIW (Tg = 350°C) Air-filled SIW using Hi-Tg FR-4 (Tg = 180°C) 50 Rogers 5880-filled SIW (Tg = 260°C) Air-filled SIW using standard FR-4 (Tg = 115°C) Air-filled SIW using Hi-Tg FR-4 (Tg = 180°C) 51 Air-filled SIW using Rogers 6002 (Tg = 350°C) Air-filled SIW using standard FR-4 (Tg = 115°C) 48 Air-filled SIW using Rogers 6002 (Tg = 350°C)

APHC (dBm) APHC (dBm)

47 44

45 42 26 28 30 32 34 36 38 40 40 42 44 46 48 50 52 54 56 58 60 Frequency (GHz) Frequency (GHz) (a) (b) Figure 26: Theoretical APHC for AFSIW composed of different substrate materials and Rogers 5880 and 6002 SIW in Ka-band (a) and U-band (b).

It can be observed that the AFSIW structure using standard FR-4 has an APHC comparable to both references. If Hi-Tg FR-4 is used, then the APHC of the AFSIW surpasses that of the two conventional SIWs. Finally, if the structure is made of only Rogers 6002, the

APHC is increased by more than 5 dB compared to the 5880 and 6002 SIWs in both of the frequency bands. In that case, the SIW can handle an APHC of about 54 dBm (251 W) at Ka- band, and 52 dBm (158 W) at U-band. This property of the AFSIW makes it of interest for the design of power combining and feeding circuits for high-power applications.

2.2.2. Peak Power Handling Capability

For transmitters involving pulsed or modulated signals, the determination of the maximum peak power before breakdown is of primordial importance. Breakdown phenomenons include multipaction and ionization [30]. In this section, only the ionization breakdown concerning ground based systems will be considered to determine the PPHC (Peak

Power Handling Capability). The dielectric slabs of width w (see Fig. 1) are neglected as the energy is concentrated in the center of the waveguide, and also as the dielectric strength, E0, is higher for solid than gaseous dielectrics. As references, the dielectric strength of air is set to

E0 = 36 kV/cm. To determine the PPHC, only the dominant TE10 mode is considered. For this mode, at ionization breakdown, the electric field inside an SIW is given by (1.15):

휋푥 퐸 (푥) = 퐸 푠𝑖푛 ( ) 푦 0 푊 (1.15) Then the PPHC is obtained by the following integral (1.16): W h 1 2 푃푃퐻퐶 = E x 푑푥푑푦 2휂   y (1.16) 푇퐸 0 0 377 휂푇퐸 = 푓2 with 휀 √(1 − 푐 ) (1.17) √ 푟 푓2

Finally, by solving (1.16), the PPHC is given by (1.18):

2 √(1 − 푓푐 ⁄ ) 푓2 (1.18) 1 푃푃퐻퐶 = √휀 푊 ∗ ℎ ∗ 퐸2 4 푟 377 0 Theoretical and simulated PPHC of air- and reference dielectric-filled SIWs are plotted in Fig. 27 over Ka-band. Considering an identical height, the PPHC of the AFSIW is drastically reduced compared to its dielectric-filled counterpart. The AFSIW can handle peak power of up to 18.5 kW at 26.5 GHz.

4 10 Theory Simulation Rogers 5880-filled SIW

3 10 Rogers 6002-filled SIW

2 PPHC (kW) PPHC 10

Air-filled SIW based on Rogers 6002

1 10 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 27: Theoretical PPHC for AFSIW and Rogers 5880 and 6002 SIW over Ka-band.

The ratio given in (1.19) allows us to compare the PPHC of air-filled and dielectric filled SIWs of respective total width Wair and Wdiel and same height h. It is obtained using

(1.18) and considering air and dielectric of permittivity εrair and εrdiel and dielectric strength

E0air and E0diel, respectively.

휀 ∗ 푊 ∗ 퐸 2 푃푃퐻퐶 √ 푟푑푖푒푙 푑푖푒푙 0푑푖푒푙 훤 = = (1.19) 푃푃퐻퐶 휀 ∗ 푊 ∗ 퐸 2 푎푖푟 √ 푟푎푖푟 푎푖푟 0푎푖푟

For comparison, it is considered that the same cut-off frequency for both SIW transmission lines, combining (1.1) and (1.19) the following expression is obtained:

2 퐸0 훤 = ( 푑푖푒푙 ) (1.20) 퐸0푎푖푟 From (1.20), it can be deduced that AFSIW handles lower peak power than dielectric- filled SIW and it is only due to the dielectric strength which is lower for gaseous than solid material.

2.3. Conclusion

Table IV and Table V summarizes performances achieved by 6002-, 5880- and air filled- SIWs of same height h = 0.508 mm, and cut off-frequency fc10 at the center frequency of the Ka- and U-band, respectively. The AFSIW surpasses the two other SIWs in terms of loss, Q-factor [25] and APHC. However, the AFSIW does not achieve a very high PPHC compared to its dielectric-filled counterpart. Also, the AFSIW has a larger footprint.

However, this larger footprint can be seen as an advantage at millimeter-wave frequencies and beyond as it relaxes fabrication tolerances.

6002 SIW 5880 SIW AFSIW

Dielectric loss (dB/m) 0.081 0.0523 0 0.0297 Ohmic loss (dB/m) 0.053 0.0452 Surface roughness loss (dB/cm) 0.0258 0.022 0.01437 Total loss (dB/cm) 0.16 0.12 0.044 Q-factor 187 251 680 54.78 APHC (dBm) 47.23 46.82 PPHC (kW) 1046 (90.2 dBm) 2375 (93.8 dBm) 23.6 (73.7 dBm) Width (mm) 4.11 4.77 7.04 Table IV: Theoretical property comparison of SIW of same height operating at Ka-band (values given at 33 GHz).

6002 SIW 5880 SIW AFSIW Dielectric loss (dB/m) 0.123 0.079 0 0.038 Ohmic loss (dB/m) 0.069 0.058 Surface roughness loss (dB/cm) 0.042 0.036 0.024 Total loss (dB/cm) 0.234 0.173 0.062 Q-factor 195 264 740 49.84 APHC (dBm) 44.12 43.8 PPHC (kW) 695 (88.4 dBm) 1577 (92 dBm) 15.4 (71.9 dBm) Width (mm) 2.715 3.16 4.604 Table V: Theoretical property comparison of SIW of same height operating at U-band (values given at 50 GHz).

3. CONCLUSION

In this chapter, a new SIW topology called AFSIW is presented. This structure is based on a multilayer PCB process. The dielectric is removed in the inner substrate (see Fig. 1) and air replaces it. Two more layers are added in order to enclose the inner layer and maintain current continuity. Some dielectric slabs still remain on each side of the propagating part for mechanical purposes. However, it has been shown that these slabs have negligible effect on the AFSIW performance as the E-field is confined in the center.

It is demonstrated that AFSIW surpasses its dielectric-filled counterpart in terms of transmission loss, Q-factor and average power handling capability. However, its peak power handling capability is reduced. For interconnection with SIW circuits, an air- to dielectric- filled SIW transition is theoretically and experimentally studied. Its bandwidth limitation, revealed from experiments, is closely related to fabrication tolerances. The theory of small reflection [31] to maintain the impedance along the transition and optimize the reflection coefficient could be developed.

To compare the performance of the AFSIW to the existing wave guiding technologies,

AFSIW is included in Table VI. It can be seen that AFSIW is an alternative technology that

97 bridges the gap between conventional waveguide and standard SIW in terms of loss and power handling capability.

Another group has proposed simultaneously an alternative structure to the AFSIW called Empty-SIW (ESIW) [32]. This topology uses long metallic walls instead of plated vias.

The ESIW is therefore much closer to a small height rectangular waveguide fabricated using

PCB material. This structure is also interesting and presents a particular point of interest as it employs air as dielectric. As it will be shown in the next paragraph, the AFSIW as the advantage of offering higher flexibility for the design of passive devices as the dielectric slabs can be used to control the propagation constant and as conductive vias achieving inductive effects can be inserted within the AFSIW.

Power Self- Self- Loss handling Compactness Cost shielded packaged capability

Waveguide

Planar Transmission

line

SIW

AFSIW

Table VI: Performance comparison of different transmission lines.

REFERENCES [1] D. Deslandes, and K. Wu, "Single-substrate integration technique of planar circuits and waveguide filters," IEEE Transactions on Microwave Theory and Techniques, vol. 51, no. 2, pp. 593-596, Feb. 2003. [2] H. Heidt, J. Puig-Suari, A. S. Moore, S. Nakasuka, and J. Twiggs: "Cubesat: A new generation of picosatellite for education and industry low-cost space experimentation," in 14TH Annual/USU Conference on Small Satellites, SSC00-V-5, Session Technical Session V: Lessons Learned - In Success and Failure, 2000. [3] K. Wu, D. Deslandes, and Y. Cassivi, "The substrate integrated circuits - a new concept for high-frequency electronics and optoelectronics," in Proc. 6th Int. Conf. on Telecom. in Modern Satellite, Cable and Broadcasting Service TELSIKS, vol. 1, 1–3 Oct. 2003, pp. P–III–P–X. [4] D. Deslandes, M. Bozzi, P. Arcioni and K. Wu, "Substrate integrated slab waveguide (SISW) for .wideband microwave applications," IEEE Int. Microwave Symp. 2003, pp. 1975-1978, Philadelphia, PA, USA. [5] F. Xu and K. Wu, "Substrate integrated non-radiative dielectric waveguide structures directly fabricated on printed circuit boards and metallized dielectric layers," IEEE Trans. Microwave Theory Tech., vol. 59, no. 12, pp. 3076-3086, 2011. [6] A. Patrovski and Ke Wu, "Substrate integrated image guide (SIIG)-A planar dielectric waveguide technology for millimeter-wave applications," IEEE Trans. Microwave Theory Tech., vol. 54, no. 6, pp. 2872-2879, 2006. [7] Djerafi, T., Aubert, H. and Wu, K. "Ridge substrate integrated waveguide (RSIW) dual-band hybrid ring coupler," IEEE Microwave and Wireless Components Letters, vol. 22, no. 2, pp. 70-72, Feb. 2012. [8] A. Niembro-Martin , V. Nasserddine , E. Pistono , H. Issa , A.-L. Franc , T.-P. Vuong and P. Ferrari, "Slow-wave substrate integrated waveguide," IEEE Trans. Microwave Theory Techn., vol. 62, no. 8, pp. 1625-1633, 2014. [9] D. Deslandes, and K. Wu, "Single-substrate integration technique of planar circuits and waveguide filters," IEEE Trans. Microwave Theory Techn., vol. 51, no. 2, pp. 593-596, Feb. 2003. [10] M. Bozzi and R. Moro, "SIW components and antennas based on eco-friendly materials and technologies: State-of-the-art and future applications," Wireless Sensors and Sensor Networks (WiSNet), 2014 IEEE Topical Conference on, Newport Beach, CA, 2014, pp. 58-60. [11] M. Bozzi, S. Moscato, L. Silvestri, N. Delmonte, M. Pasian and L. Perregrini, "Innovative SIW components on paper, textile, and 3D-printed substrates for the Internet of Things," 2015 Asia-Pacific Microwave Conference (APMC), Nanjing, 2015, pp. 1-3. [12] S. Kim, M. M. Tentzeris and H. Aubert, "An inkjet-printed flexible broadband multilayer SIW coupler for antenna array systems," 2014 IEEE Antennas and Propagation Society International Symposium (APSURSI), Memphis, TN, 2014, pp. 324-325. [13] F. Parment, A. Ghiotto, T.P. Vuong, J.M. Duchamp, and K. Wu, "Broadband transition from dielectric-filled to air-filled substrate integrated waveguide for low loss and high power handling millimeter-wave substrate integrated circuits," IEEE MTT-S International Microwave Symposium Digest, pp. 1-4, 2014. [14] F. Parment, A. Ghiotto, T.P. Vuong, J.M. Duchamp, and K. Wu, "Air-filled substrate integrated waveguide for low loss and high power handling millimeter-wave substrate

integrated circuits," IEEE Trans. Microwave Theory Techn., vol. 63, no. 4, pp. 1-11, Apr. 2015. [15] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp and K. Wu, "Low-loss air-filled substrate integrated waveguide (SIW) band-pass filter with inductive posts," Microwave Conference (EuMC), 2015 European, Paris, 2015, pp. 761-764. [16] F. Xu , X. Jiang and K. Wu, "Efficient and accurate design of substrate integrated waveguide circuits synthesized with metallic via-slot arrays," IET Microwave Antennas Propagation, vol. 2, no. 2, pp. 188-193, 2008. [17] http://www.microwaves101.com/encyclopedias/waveguide-mathematics#cutoff [18] K. J. Button and B. Lax, "Theory of ferrites in rectangular waveguides," IRE Trans. on Antennas and Propagation, vol. 4, pp. 531-537, Jul. 1956. [19] A. Ghiotto, S. Adhikari and K. Wu, "Ferrite-Loaded Substrate Integrated Waveguide Switch," IEEE Microwave and Wireless Components Letters, vol. 22, no. 3, pp. 120- 122, Mar. 2012. [20] N. Ghassemi, I. Boudreau, D. Deslandes, and K. Wu, "Millimeter-wave broadband transition of substrate integrated waveguide on high-to-low dielectric constant substrates," IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 3, no. 10, pp. 1764-1770, Oct. 2013. [21] A. Suntives and R. Abhari, "Design and characterization of the EBG waveguide-based interconnects," IEEE Trans. Advance Packaging, vol. 30, no. 2, pp. 163-170, 2007. [22] Engen, G.F., and Hoer, C.A.: "Thru-reflect-line: an improved technique for calibrating the dual six-port automatic network analyzer," IEEE Trans. Microwave Theory Tech., 1979, 27, (12), pp. 987–993. [23] L. Li, K. Wu and L. Zhu, "Numerical TRL calibration technique for parameter extraction of planar integrated discontinuities in a deterministic MoM algorithm," IEEE Microwave and Wireless Components Letters, vol.12, no.12, pp.485-487, Dec. 2002. [24] E. D. Caballero, A. Belenguer, H. Esteban and V. E. Boria, "Thru-reflect-line calibration for substrate integrated waveguide devices with tapered microstrip transitions," IET Electron. Lett., vol. 49, no. 2, pp. 132-133, 2013. [25] Pozar, D. M., Microwave Engineering, John Wiley and Sons, (New York, 1998). [26] F. Xu and K. Wu, "Guided-wave and leakage characteristics of substrate integrated waveguide," IEEE Trans. Microwave Theory Tech., vol. 53, pp. 66-73, 2005. [27] E. O. Hammerstad and O. Jensen, "Accurate models for microstrip computer aided design," IEEE MTT-S International Microwave Symposium Digest, pp. 407-409. [28] T. Liang, S. Hall , H. Heck and G. Brist, "A practical method for modeling PCB transmission lines with conductor surface roughness and wideband dielectric properties," Proc. IEEE Int. Microwave Symp. Dig., pp. 1780-1783, 2006. [29] Y. J. Cheng, K. Wu, and W. Hong, "Power handling capability of substrate integrated waveguide interconnects and related transmission line systems," IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 31, no. 4, pp. 900- 909, Nov. 2008. [30] Ming Yu, "Power-handling capability for RF filters," IEEE Microwave Magazine, vol. 8, no. 5, pp. 88-97, Oct. 2007. [31] B. Khalichi, S. Nikmehr, and M. Ebrahimpouri, "Designing wideband tapered-slot antennas: A novel method using SIW technology," IEEE Antennas and Propagation Magazine, vol. 87,no. pp. 60-70, Aug. 2015. [32] A. Belenguer, H. Esteban and V. Boria, "Novel empty substrate integrated waveguide for high-performance microwave integrated circuits," IEEE Trans. Microwave Theory Tech., vol. 62, no. 4, pp. 832-839, 2014.

100

CHAPTER IV: AIR- FILLED SIW FEEDING NETWORK ELEMENTS

101

TABLE OF CONTENTS TABLE OF CONTENTS...... 102 LIST OF FIGURES ...... 103 LIST OF TABLES ...... 106 1. INTRODUCTION...... 107 2. POWER DIVIDER AND COMBINER ...... 107

2.1. TRANSMISSION LINE MODEL ...... 108 2.2. T-JUNCTION GEOMETRY ...... 109 2.3. RESULTS AND COMPARISON ...... 110 3. COUPLERS ...... 112

3.1. RIBLET SHORT-SLOT COUPLER ...... 113 3.1.1. Theory and Design Approach ...... 113 3.1.2. AFSIW Riblet’s Coupler Geometry ...... 115 3.1.3. Results and Comparison...... 116 3.2. DIRECTIONAL COUPLER ...... 118 3.2.1. Broadband Directional Moreno Coupler...... 119 3.2.1.1. Moreno Coupler Structure ...... 119 3.2.1.2. Experimental Results ...... 121 3.2.2. Two-Hole Coupler...... 124 3.2.2.1. Directional Two-Hole Coupler Structure ...... 124 3.2.2.2. Experimental Results ...... 127 4. PHASE SHIFTERS ...... 130

4.1. DOUBLE DIELECTRIC SLAB-LOADED AIR-FILLED SIW PHASE SHIFTER...... 131 4.1.1. Equal-Length Unequal-Dielectric-Slab-Width Phase Shifter ...... 131 4.1.2. Equal-Length Unequal-Dielectric-Slab-Width-and-Total-Width Phase Shifter 135 4.1.3. Broadband Compensated Phase Shifter...... 136 4.2. SINGLE DIELECTRIC SLAB-LOADED AIR-FILLED SIW PHASE SHIFTER ...... 140 5. CONCLUSION ...... 141 REFERENCES ...... 143

102

LIST OF FIGURES

FIGURE 1: SCHEME OF (A) A POWER DIVIDER AND (B) A POWER COMBINER...... 107 FIGURE 2: TRANSMISSION LINE MODEL OF A T-JUNCTION POWER DIVIDER [12]...... 108 FIGURE 3: SIW T-JUNCTION POWER DIVIDER [6]...... 109 FIGURE 4: (A) GEOMETRY AND (B) E-FIELD MAGNITUDE DISTRIBUTION SIMULATED AT F = 33 GHZ OF THE T-JUNCTION WITH W = 7.064 MM, W = 0.508 MM, D = 0.45 MM, T = 0.808 MM, AND I = 2.064 MM...... 109 FIGURE 5: PHOTOGRAPH OF THE FABRICATED T-JUNCTION INNER SUBSTRATE...... 110 FIGURE 6: MEASURED AND SIMULATED S-PARAMETERS OF THE KA-BAND T-JUNCTION...... 110 FIGURE 7: MEASURED AND SIMULATED PHASE IMBALANCE OF THE KA-BAND T-JUNCTION. ... 111 FIGURE 8: MEASURED AND SIMULATED AMPLITUDE IMBALANCE OF THE KA-BAND T-JUNCTION...... 111 FIGURE 9: PICTURE OF THE (A) BETHE-HOLE, (B) SCHWINGER AND(C) RIBLET COUPLERS [12]...... 112 FIGURE 10: PICTURE OF (A) THE QUADRATURE AND (B) THE RAT-RACE HYBRID COUPLER [12]...... 113 FIGURE 11: SCHEME OF A 90° HYBRID COUPLER...... 114 FIGURE 12: SIW 90° HYBRID COUPLER [13]...... 116 FIGURE 13: (A) GEOMETRY AND (B) E-FIELD MAGNITUDE DISTRIBUTION SIMULATED AT F = 35 GHZ OF THE 90° HYBRID COUPLER WITH W = 7.068 MM, W = 0.508 MM,

C1 = 0.869 MM, AND C2 = 8.292 MM...... 116 FIGURE 14: PHOTOGRAPH OF THE FABRICATED 90° HYBRID COUPLER INNER SUBSTRATE...... 117 FIGURE 15: SIMULATED |SII| AND |SIJ| OF THE 90° HYBRID COUPLER IN KA-BAND...... 117

FIGURE 16: MEASURED |SII| AND |SIJ| OF THE 90° HYBRID COUPLER IN KA-BAND...... 117 FIGURE 17: MEASURED AND SIMULATED 90° HYBRID COUPLER PHASE IMBALANCE...... 117 FIGURE 18: MEASURED AND SIMULATED 90° HYBRID COUPLER AMPLITUDE IMBALANCE...... 118 FIGURE 19: GEOMETRY OF THE PROPOSED MORENO COUPLER WHERE INNER AND TOP SUBSTRATES CONTAIN THE AFSIW AND DFSIW RESPECTIVELY...... 120 FIGURE 20: THEORETICAL PHASE OF COUPLED SIGNALS AT THE ISOLATED PORT 2...... 121 FIGURE 21: PICTURES OF THE FABRICATED AFSIW MORENO COUPLER: (A) THE INNER SUBSTRATE, (B) THE TOP SUBSTRATE, AND (C) THE ASSEMBLED PROTOTYPE...... 122 FIGURE 22: MEASURED AND SIMULATED |S11| AND |S31|...... 123 FIGURE 23: MEASURED AND SIMULATED |S21| AND |S41|...... 123 FIGURE 24: E-FIELD MAGNITUDE AT 30 GHZ: (A) TOP AND (B) BOTTOM VIEW...... 123 FIGURE 25: EFFECT OF 0.1 MM SLOT MISALIGNMENT ON |S21| AND |S41|...... 124 FIGURE 26: GEOMETRY OF THE PROPOSED DIRECTIONAL TWO-HOLE COUPLER: TOP VIEWS OF THE (A) INNER AND (B) TOP SUBSTRATES...... 125 FIGURE 27: COMPARISON BETWEEN THEORETICAL AND SIMULATED |S31| AND |S41| FOR (A) FORWARD COUPLING WITH D1 = 2 MM AND D2 = 2.27 MM AND (B) BACKWARD COUPLING

WITH D1 = 10 MM AND D2 = 3.52 MM...... 126

103

FIGURE 28: DFSIW E-FIELD MAGNITUDE AT 30 GHZ FOR (A) FORWARD COUPLING AND (B) BACKWARD COUPLING...... 127 FIGURE 29: FABRICATED TWO-HOLE COUPLER: (A) THE INNER AND THE TOP SUBSTRATE BEFORE ASSEMBLING AND (B) THE MOUNTED PROTOTYPE...... 127 FIGURE 30: MEASURED AND SIMULATED |S11| AND |S21|...... 128 FIGURE 31: MEASURED AND SIMULATED |S31| AND |S41|...... 129 FIGURE 32: (A) AFSIW AND (B) DFISW E-FIELD MAGNITUDE AT 28 GHZ...... 129 FIGURE 33: SIMULATED SLOT X-DIRECTION MISALIGNMENT EFFECT ON |S31| AND |S41| OF THE PROTOTYPED COUPLER...... 129 FIGURE 34: AFSIW STRUCTURE...... 130 FIGURE 35: ILLUSTRATION OF THE EQUAL-LENGTH UNEQUAL-DIELECTRIC-SLAB-WIDTH AIR- FILLED SIW PHASE SHIFTER...... 131 FIGURE 36: THEORETICAL AND SIMULATED PHASE SHIFTS FOR DIFFERENT ELEMENTS CONSTITUTING THE DOUBLE DIELECTRIC SLAB LOADED AIR-FILLED SIW PHASE SHIFTER.132 FIGURE 37: SIMULATED PHASE SHIFTER TRANSITION LENGTH L PARAMETRIC SWEEP...... 132 FIGURE 38: FABRICATED EQUAL-LENGTH UNEQUAL-DIELECTRIC-SLAB-WIDTH AIR-FILLED SIW PHASE SHIFTER INNER SUBSTRATE...... 133 FIGURE 39: SIMULATED AND MEASURED S-PARAMETERS OF THE DOUBLE DIELECTRIC SLAB LOADED PHASE SHIFTER...... 133 FIGURE 40: SIMULATED AND MEASURED RELATIVE PHASE SHIFTS WITH E-FIELD MAGNITUDE DISTRIBUTION SIMULATED AT 28 GHZ...... 134 FIGURE 41: SIMULATED AND MEASURED S-PARAMETERS OF THE PHASE SHIFTER IN U-BAND. 134 FIGURE 42: SIMULATED AND MEASURED RELATIVE PHASE SHIFT ...... 135 FIGURE 43: THEORETICAL PROPAGATION OF A REFERENCE AIR-FILLED SIW AND AN UNEQUAL- DIELECTRIC-SLAB-WIDTH-AND-TOTAL-WIDTH AIR-FILLED SIW...... 136 FIGURE 44: ILLUSTRATION OF THE COMPENSATED PHASE SHIFTER...... 137 FIGURE 45: THEORETICAL PROPAGATION CONSTANT DIFFERENCE BETWEEN A REFERENCE AIR- FILLED SIW AND AN UNEQUAL-DIELECTRIC-SLAB-WIDTH AIR-FILLED SIW (TYPE 1) AND BETWEEN A REFERENCE AIR-FILLED SIW AND AN UNEQUAL-DIELECTRIC-SLAB-WIDTH-AND- TOTAL-WIDTH AIR-FILLED SIW (TYPE 2)...... 137 FIGURE 46: PICTURE OF THE FABRICATED COMPENSATED AIR-FILLED SIW PHASE SHIFTER INNER SUBSTRATE...... 137 FIGURE 47: SIMULATED AND MEASURED S-PARAMETERS OF THE PROPOSED COMPENSATED PHASE SHIFTER...... 138 FIGURE 48: SIMULATED AND MEASURED PHASE SHIFTS OF THE PROPOSED COMPENSATED PHASE SHIFTER WITH E-FIELD MAGNITUDE DISTRIBUTION SIMULATED AT 33 GHZ...... 138 FIGURE 49: SIMULATED AND MEASURED RELATIVE PHASE SHIFTS USING THE METHOD DESCRIBED IN [31], FOR A TOTAL LENGTH OF 10 MM FOR THE DOUBLE DIELECTRIC SLAB

LOADED PHASE SHIFTER AND FOR A TOTAL LENGTH LREF = 11.6 MM FOR THE REFERENCE...... 139 FIGURE 50: STRUCTURE OF THE SINGLE DIELECTRIC SLAB LOADED AIR-FILLED SIW PHASE SHIFTER...... 140 FIGURE 51: SIMULATED S-PARAMETERS OF THE SINGLE DIELECTRIC SLAB-LOADED PHASE SHIFTER...... 140

104

FIGURE 52: SIMULATED RELATIVE PHASE SHIFT OF THE SINGLE DIELECTRIC SLAB-LOADED PHASE SHIFTER...... 141 FIGURE 53: SIMULATED E-FIELD MAGNITUDE OF THE SINGLE DIELECTRIC SLAB LOADED PHASE SHIFTER (A) AT 32 GHZ AND (B) AT THE RESONANT FREQUENCY OF 36.6 GHZ...... 141

105

LIST OF TABLES

TABLE I: COMPARISON OF AIR- AND DIELECTRIC-FILLED SIW T-JUNCTIONS PERFORMANCES OVER THE KA-BAND...... 111 TABLE II: COMPARISON OF AIR- AND DIELECTRIC-FILLED SIW 90° HYBRIDS PERFORMANCES IN KA-BAND...... 118 TABLE III: DIMENSIONS OF THE PROTOTYPED KA-BAND MORENO COUPLER...... 121 TABLE IV: DIMENSIONS OF THE PROTOTYPED KA-BAND TWO-HOLE COUPLER...... 126 TABLE V: SIMULATED FIRST THREE MODE CUT-OFF FREQUENCIES OF THE AIR-FILLED SIW AND DOUBLE DIELECTRIC SLAB LOADED AIR-FILLED SIW SECTIONS...... 135 TABLE VI: COMPARISON BETWEEN MEASURED AIR-FILLED SIW EQUAL-LENGTH AND UNEQUAL- LENGTH COMPENSATED PHASE SHIFTER PERFORMANCES...... 139

106

1. INTRODUCTION

Previous chapter demonstrated the interest of the AFSIW, this chapter presents passive components commonly found in millimeter-wave systems including power dividers, couplers and phase shifters based on this new technology and compares their performances to their dielectric-filled SIW counterpart. In the first section, the T-junction power divider, proposed in [1], is explored. Then, different topologies of couplers are introduced including the Riblet coupler, the Moreno coupler and the two-hole coupler, respectively published in [1], [2] and

[3]. Finally, phase shifters, published within a journal [4] and a conference paper [5], are presented with a new topology of compensated phase shifter.

2. POWER DIVIDER AND COMBINER

The power divider and power combiner are elementary components found in feeding networks. They are both illustrated in Fig. 1 where α is the coefficient of division. Such components can usually combine or distribute signals in two or more ports. Various topologies of power dividers exist including the T-junction [6], the magic-T [7] and the

Wilkinson [8]. Those topologies have already been adapted in SIW [9]-[10] and LTCC [11] technologies. In this section, the theory of a T-junction based on scattering parameter and impedance is presented. The performance of a T-junction demonstrator based on AFSIW is compared to its SIW counterpart.

P =α P P Power 2 1 2 Power P Divider P = P + P 1 1 2 3 Combiner (α) P =(1-α) P P 3 1 3 (a) (b) Figure 1: Scheme of (a) a power divider and (b) a power combiner.

107

2.1. Transmission Line Model

A lossless T-junction power divider can be modeled as three converging transmission lines. The junction can be seen as a discontinuity and it can generate fringing fields and higher order modes, depending on the used technology (planar, waveguide…) [12]. This leads to stored energy in this discontinuity region which can be assimilated as a lumped susceptance,

B [12]. The transmission line model is shown in Fig. 2.

Figure 2: Transmission line model of a T-junction power divider [12].

For a matched input with characteristic impedance Z0, the input admittance Yin is given by equation (1.1) [12] where Z1 and Z2 are output impedances.

1 1 1

푌푖푛 = 푗퐵 + + = (1.1) 푍1 푍2 푍0

For lossless case and if B = 0, then equation (1.1) can be rewritten as (1.2) and all characteristic impedances become real.

1 1 1 + = (1.2) 푍1 푍2 푍0

As a result, Z1 and Z2 can be chosen to achieve different division coefficients. If B is not negligible, some compensations or reactive elements can be implemented to compensate as explained in [12].

108

2.2. T-junction Geometry

Using the advantages from SIW technology, the proposed T-junction geometry is illustrated in Fig. 3. This topology has been presented in [6]. This H-plane structure implements an inductive post to achieve a good matching.

Figure 3: SIW T-junction power divider [6].

An AFSIW T-junction power divider based on [6] has been designed. It is shown in

Fig. 4(a). This topology is similar to [6]. The proposed AFSIW structure implements an inductive post and two vias at the input port to improve the input matching. The AFSIW T- junction has been design to operate at Ka-band. Its electric field magnitude is shown in

Fig. 4(b) to illustrate the proper division in phase and amplitude of the input signal. The inductive post also compensates the susceptance B as explained at the end of Section 2.1.

Moreover, the diameter and the position of the inductive post and the input vias can be adjusted to optimize the return loss. For mechanical purposes, this inductive post is attached with dielectric to the side walls.

(a) (b) Figure 4: (a) Geometry and (b) E-field magnitude distribution simulated at f = 33 GHz of the T-junction with W = 7.064 mm, w = 0.508 mm, d = 0.45 mm, t = 0.808 mm, and I = 2.064 mm.

109

2.3. Results and Comparison

The fabricated prototype has been made using Rogers 6002 substrate of 0.508 mm thickness. The measurement has been performed using a VNA and a TRL calibration kits to de-embed the transitions. The prototyped AFSIW T-junction is illustrated in Fig. 5 with reference planes.

Reference plane Reference plane

Figure 5: Photograph of the fabricated T-junction inner substrate.

Simulated and measured results are given in Fig. 6. Measured |S11| is lower than -15 dB from 26 to 39.5 GHz. Measured transmissions, |S21| and |S31|, are -3.1 ± 0.5 dB over the all

Ka-band. It can be noticed that the transmission losses are improved by the use of the AFSIW when comparing with [6] in which results are between -3.3 and -4 dB for a dielectric-filled

SIW T-junction. A phase imbalance of ± 2.5° is obtained from 26 to 38 GHz in measurement as illustrated in Fig. 7. An amplitude imbalance of ± 0.4 dB is achieved in the same frequency range as shown in Fig. 8. The variation in phase and amplitude imbalance is mostly related to the measurement accuracy.

0 0  -3 -10

-6

-20

|Sij| (dB) (dB) |Sij| |Sij|  (dB) |Sij|

|S11|(dB) -9 S11 Simulated -30 S11 Measured Sij Simulated -12 Sij Measured

-40 -15 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 6: Measured and simulated S-parameters of the Ka-band T-junction.

110

-10 Measured Simulated

Phase Imbalance(°) Phase -20 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 7: Measured and simulated phase imbalance of the Ka-band T-junction.

-5 Measured Simulated -10

Amplitude imbalance (dB) imbalance Amplitude 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 8: Measured and simulated amplitude imbalance of the Ka-band T-junction.

Table I compares the performances of 6002- and air-filled SIW T-junctions having

10 mm long access port and based on the same substrate thickness (0.508 mm). The AFSIW allows achieving, over the all Ka-band, a better matching and offers 0.2 dB lower losses than its 6002-filled counterpart for comparable total dimensions. This is a substantial improvement for the design of high performances front-end feeding networks. Considering for example the design of an 8 x 1 antenna array feeding network composed of four dividing stages, a 0.8 dB of loss improvement would be achieved using air-filled SIW instead of dielectric-filled SIW.

Performance 6002-filled SIW Air-filled SIW

|S11| (dB) < -13.95 < -17

|S21| et |S31| (dB) < -3.32 < -3.12

Dimensions (mm) 13.34 x 23.59 16.79 x 26.56 Table I: Comparison of air- and dielectric-filled SIW T-junctions performances over the Ka-band.

111

3. COUPLERS

Directional couplers are four port networks. They are basic components, which are found in numerous microwave and millimeter-wave systems [13]. They are used for example in antenna array and feeding network designs. Coupling mechanisms have been studied for long time [14] and resulted in several topologies including the well-known Bethe hole,

Schwinger [15] and Riblet couplers which are illustrated Fig. 9.

(a) (b)

To characterize such components, some parameters have been established. They are given in equation (1.3) to (1.5) [12]. In (1.3), C is the coupling factor that indicates the fraction of the input power that is coupled to the output port. In (1.4), D is the directivity that evaluates the coupler’s ability to isolate forward and backward coupled waves (or the coupled and uncoupled ports). In (1.5), I is the isolation that measures the power delivered to the uncoupled port.

Coupling 푃1 factor 퐶 = 10푙표푔 ( ) (푑퐵) (1.3) 푃3

112

푃 Directivity 퐷 = 10푙표푔 ( 3) (푑퐵) (1.4) 푃4

푃 Isolation 퐼 = 10푙표푔 ( 1) = 퐷 + 퐶 (푑퐵) (1.5) 푃4 The hybrid coupler is a special case of directional coupler for which the coupling factor is 3 dB. Two examples are the quadrature hybrid, shown in Fig. 10(a) and that offer a 90° phase shift between the coupled and the through port, and the Rat-Race, depicted in

Fig. 10(b), which provides a 180° phase shift.

(a) (b) Figure 10: Picture of (a) the quadrature and (b) the Rat-Race hybrid coupler [12].

3.1. Riblet Short-Slot Coupler

The Riblet short slot coupler has been studied for a long time [16]. The first prototype has been demonstrated by Riblet himself based on the waveguide technology in 1952 [16].

Recently, this topology of coupler has been designed based on SIW technology. As an example, SIW 90° hybrid couplers based on Riblet’s structure are reported in [13], [17] and

[18].

3.1.1. Theory and Design Approach

The Riblet short-slot coupler consists of a 90° hybrid coupler, with a symmetrical structure using a cavity to generate interferences between TE10 and TE20 modes. A scheme is given in Fig. 11. To explain the behavior of such component, [13] has developed equations

113 performing an even–odd mode analysis. ai are the incident waves, and bi are the outgoing waves.

Port 1 Port 2

Port 4 Port 3

Figure 11: Scheme of a 90° Hybrid coupler.

For an even-odd mode analysis, the incident waves at port 1-4 are as in (1.6) and the outgoing waves are given by (1.7).

푎1 푎1 ⁄2 ⁄2 푎1 0 0 0 푎 = 푎 + 푎 = + = ( ) (1.6) 푒푣푒푛 표푑푑 0 0 0 푎1 − 푎1 0 ( ⁄2) ( ⁄2) 푏 = 푏푒푣푒푛 + 푏표푑푑 0 0 −푗훽 퐿 −푗훽 퐿 푎1 푒 10 푎1 푒 20 = ( ) + ( ) 2 푒−푗훽10 퐿 2 −푒−푗훽20 퐿 0 0 (1.7) 0 −푗훽 퐿 −푗훽 퐿 푎1 푒 10 +푒 20 = ( ) 2 푒−푗훽10 퐿−푒−푗훽20퐿 0 As the Riblet short slot coupler has a symmetric structure and considering that port 1 is matched, then the S-parameters are given by (1.8)-(1.11).

푆11 = 0 (1.8)

(훽 − 훽 )퐿 푏2 1 −푗훽 퐿 −푗훽 퐿 10 20 −푗(훽 +훽 )퐿/2 푆21 = = (푒 10 +푒 20 ) = 푐표푠( )푒 10 20 (1.9) 푎1 2 2

푏3 1 −푗훽 퐿 −푗훽 퐿 푆 = = (푒 10 −푒 20 ) 31 푎 2 1 (1.10) (훽 − 훽 )퐿 10 20 −푗(훽 +훽 )퐿/2 = −푗푠푖푛( )푒 10 20 2

114

푆41 = 0 (1.11)

As the Riblet structure is a hybrid junction, |S | = |S | = 1 , and hence, (1.12) can 21 31 √2 be deduced from (1.10).

−1 퐿 휋 sin |푆31| = (훽10 − 훽20) ↔ 퐿 = (1.12) 2 2(훽10 − 훽20)

2 2휋 2 휋 with 훽10 = = √푘 − ( ) (1.13) 휆푔1 푊

2 2휋 2 2휋 and 훽20 = = √푘 − ( ) (1.14) 휆푔2 푊

Now, (1.20) gives a relation between the length and the width of the cavity. In fact, it is required to introduce another equation to obtain both dimensions. For that, in [13], the design equations incorporating the wave cancellation effect have been derived from [19]. As a result, the slot length to minimize the reflect waves is obtained from [13] and is given in (1.15), where n is a non-negative integer.

(2푛 + 1)휋 퐿 = (1.15) 2훽10

Finally, by combining equations (1.15) and (1.12), it is possible to obtain the design equations (1.16) and (1.17) for the cavity width W and length L.

(푛 + 1).(3푛 + 1) 푊 = 휆 √ (1.16) 0 (4푛 + 1)

(푛 + 1). (3푛 + 1) 퐿 = 휆 √ (1.17) 0 12

3.1.2. AFSIW Riblet’s Coupler Geometry

The SIW Riblet coupler geometry is illustrated in Fig. 12. This topology has been presented in [16] and consists in an H-plane structure.

115

Figure 12: SIW 90° hybrid coupler [13].

Based on [13], an AFSIW Riblet short slot coupler has been designed. It is depicted in

Fig. 13(a). This topology is clearly similar to those used in [20] where matched H-plane impedance steps are introduced. These steps provide more flexibility to match and optimize the cavity of such coupler. In Fig. 13(b), the simulated E-field magnitude of this structure is shown when exciting port 1. It can be observed that the signal is equally divided in amplitude at port 2 and port 3 with a 90° phase imbalance, and that port 4 is isolated.

Copper Substrate Air

Port 1 Port 4

w C2 C1

Port 2 Port 3

(a) (b) Figure 13: (a) Geometry and (b) E-field magnitude distribution simulated at f = 35 GHz of the 90° hybrid coupler with W = 7.068 mm, w = 0.508 mm, C1 = 10.87 mm, and C2 = 8.29 mm.

3.1.3. Results and Comparison

The fabricated prototype is shown in Fig. 14. Simulated and measured S-parameters are respectively depicted in Fig. 15 and 16. Measured and simulated |S11| are equal to or below

-15 dB between 32 GHz and 36 GHz. On the same frequency band, the measured |S21| and

|S31| are equal to -4.6 ± 0.3 dB. They are lower than the simulated results which are

-3.2 ± 0.2 dB. This difference can be attributed to the fact that in simulation the 90° bends are not considered and also the multilayer process provides a poor electrical contact between the two adjacent ports. A phase imbalance of 90° ± 2.9° is obtained from 32 to 36 GHz in

116 measurement (Fig. 17). An amplitude imbalance of ± 0.3 dB is achieved in the same frequency range as shown in Fig. 18.

Reference Reference plane plane

Figure 14: Photograph of the fabricated 90° hybrid coupler inner substrate.

0 0 -5  -5

-10  -10

| | | (dB) (dB)

| | (dB) -15 -15

41 31 31

-20 -20

| | | and and |S |S

| | and |S -25 -25

21 21

11 S11

|S |S |S -30 S21 -30 S31 -35 -35 S41

-40 -40 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 15: Simulated |Sii| and |Sij| of the 90° hybrid coupler in Ka-band.

0 0  -5 -5

-10 -10

||(dB) (dB)

|(dB) -15  -15

41 31 31

-20 -20

| and |S|S ||and and

| and |S |and -25 -25

11 21 21

|S |S |S -30 S11 -30 S21 -35 S31 -35 S41 -40 -40 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 16: Measured |Sii| and |Sij| of the 90° hybrid coupler in Ka-band.

180 Measured 150 Simulated 120 90 60

30 Phase imbalance (°) imbalance Phase 0 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 17: Measured and simulated 90° hybrid coupler phase imbalance.

117

-5 Measured Simulated -10 Amplitude imbalance (dB) imbalance Amplitude 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 18: Measured and simulated 90° hybrid coupler amplitude imbalance.

Table V compares the performances of 6002 and air-filled 90° hybrids having 5 mm long port access, same substrate thickness (0.508 mm) and designed to achieve ± 0.3 dB of amplitude imbalance in the 32 to 36 GHz bandwidth. The air- and 6002-filled SIW 90° hybrids achieve comparable input matching (S11) and isolation (S41). However, the AFSIW

90° hybrid improves by 0.15 dB the losses compared with its 6002-filled counterpart for a slight increase in dimensions. This is again a substantial improvement for the design of high performances front-end feeding networks. The benefit of using air-filled instead of dielectric- filled SIW for the design of 90° hybrids would be even more interesting for higher frequencies.

Performance 6002-filled SIW Air-filled SIW ± 0.3 dB amplitude imbalance bandwidth (GHz) 32 - 36 32 - 36 |S11| (dB) < -15.3 < -15 |S21| et |S31| (dB) -3.35 ± 0.3 -3.2 ± 0.3 |S41| (dB) < -15.5 < -15

C2 (mm) 6.28 8.29 Table II: Comparison of air- and dielectric-filled SIW 90° hybrids performances in Ka-band.

3.2. Directional Coupler

In most millimeter-wave systems, signal monitoring must be performed using a directional coupler. For example, it can be also used in antenna array and feeding network designs. Coupling mechanisms have been studied for long time [21] and resulted in several topologies including the well-known Moreno, Schwinger, Bethe hole, and multi-hole

118 couplers. The last one offers a parallel broad-wall structure that can be designed to achieve backward or forward coupling.

The Moreno structure is a crossed broad-wall waveguide directional coupler. It is intrinsically able to provide flat coupling and good directivity over a wide bandwidth. It consists of one or two cross slot apertures placed diagonally for magnetic and electrical field couplings. This allows a magnetic circular polarization, which ensures no reflection in the primary part where the excitation is provided [22]. This device can be realized in rectangular waveguide (RWG) technology, providing low insertion loss and high power handling desired for space or radar applications, and monitoring power source, for example. However, RWG is expensive, bulky and requires complex interconnects with base-band digital circuits.

3.2.1. Broadband Directional Moreno Coupler

3.2.1.1. Moreno Coupler Structure

The Moreno coupler is a well-known 4-port passive device consisting in two crossed broad-wall waveguides. It has been theoretically studied in [23], taking only into account magnetic coupling contributions and neglecting electrical coupling.

The proposed multilayer PCB Moreno coupler geometry is shown in Fig. 19. It consists of crossed AFSIW and DFSIW coupled through two crossed slots (A1 and A2), located near the AFSIW and DFSIW sidewalls where the magnetic fields of the fundamental TE10 mode of both structures are strong. It will be considered throughout this section that port 1 is the input port, port 2 is the isolated port, port 3 is the through port and port 4 is the coupled port.

Using a symmetrical pair of crossed slots, high isolation is obtained by canceling the two signals coupled trough each slot. It can be observed in Fig. 19 that the two slots are coupled at the isolated port to the magnetic dipole with a 180° phase difference while at the coupled port they are coupled to the magnetic dipole in phase. The Φ1 and Φ2 phase shifts of

119 the signals coupled trough A1 and A2 respectively at the isolated port are given in (1.18) and

(1.19).

훷1 = 훽퐷퐹푆퐼푊 ∗ (푎퐴퐹푆퐼푊 − 2푑1) (1.18)

훷2 = 훽퐴퐹푆퐼푊 ∗ (푎퐷퐹푆퐼푊 − 2푑2) + 180 (1.19)

where βAFSIW and βDFSIW are the propagation constants in the AFSIW and DFSIW respectively, aAFSIW and aDFSIW are the width of the AFSIW and DFSIW respectively, and d1 and d2 are the distance between the sidewalls and the A1 and A2 crossed slots respectively (see

Fig. 19).

It has been shown in [22] that the directivity is improved for relatively narrow slots, a relative thin common wall and a distance from the nearest sidewall of a quarter of the waveguide width. Furthermore, the orientation of the crossed slots can increase the directivity by a few dB. In [24], 45° crossed slots has been demonstrated as the optimal configuration.

Based on equations (1.18) and (1.19), and dimensions given in Table III, the determination of the phases at the isolated port 2 is achieved. In Fig. 20, Φ1 and Φ2 are plotted with their difference ΔΦ. It can be seen that ΔΦ is 179° ± 1.2° which means that both coupled signals cancel each other at port 2 to provide a high isolation.

AFSIW Through y w Port 3

aAFSIW DFSIW x

A2 d2 (βDFSIW) w Isolated s Coupled a d1 Port 2 DFSIW Port 4 A1 l

(βAFSIW) Copper Substrate Air Invisible elements Input Port 1

Figure 19: Geometry of the proposed Moreno coupler where inner and top substrates contain the AFSIW and DFSIW respectively.

120

Dimensions aAFSIW aDFSIW w d1 d2 ws l

Values 7.04 4.11 0.508 2.085 1.235 0.35 1.7 (mm)

Table III: Dimensions of the prototyped Ka-band Moreno coupler.

300

240

180  2 120  Phase (°) Phase 1  60

0 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 20: Theoretical phase of coupled signals at the isolated port 2.

3.2.1.2. Experimental Results

Using the previously described design procedure, a Ka-band AFSIW Moreno coupler

has been designed and fabricated with 0.508 mm thick Rogers RT/Duroid 6002 material of

permittivity εr = 2.94 for the middle and top substrates, and 0.8 mm thick FR-4 material for

the bottom substrate. A picture of the assembled demonstrator is shown in Fig. 21(c). The

AFSIW was made in the inner substrate using a laser micromachining process as shown in

Fig. 21(a). Then, the top substrate, implementing the coupled DFSIW, shown in Fig. 21(b),

and the bottom FR-4 substrate are mounted using a multilayer PCB process. To characterize

this prototype, conductor-backed coplanar waveguides (CBCPW) feed each port through

CBCPW to DFSIW transitions. All dimensions, illustrated in Fig. 19, are given in Table III.

121

DFSIW to AFSIW transition Port 1

Port 3 CB-CPW to DFSIW transition AFSIW (a) SIW Port 3

Reference planes Port 2 Port 4 Port 2 Port 4

Port 1

Crossed slots (c)

(b)

Figure 21: Pictures of the fabricated AFSIW Moreno coupler: (a) the inner substrate, (b) the top substrate, and (c) the assembled prototype.

Measured and simulated S-parameters are shown in Fig. 22 and 23. Measurement was performed using a Vector Network Analyzer (VNA) and a test fixture using a TRL calibration to de-embed connectors and CBCPW to SIW transitions. Measurement reference planes are shown in Fig. 21(c). Simulations are taking into account conductor and dielectric losses. As shown in the Fig. 22 and 23, measured and simulated results show good agreement. The observed differences are mainly due to the fact that in simulation, the AFSIW to DFSIW transitions are not considered. The measured return loss (|S11|) and isolation (|S21|) are lower than -15.5 dB and -31 dB over all the Ka-band. It can be inferred from Fig. 20 and 23 that the higher is the isolation, the closer is the phase difference to 180°. The measured insertion loss

(|S31|) and coupling (|S41|) are 0.6 ± 0.2 dB and -18.7 ± 1.3 dB on overall the frequency range. The simulated E-field magnitude is shown at 30 GHz in Fig. 24. It can be observed that the signal feed at port 1, is properly coupled at port 4, and that port 2 is isolated.

122

-10

-20

-30

-40

|S11| simulated |S11| and |S31| (dB) |S11|and -50 |S11| measured |S31| simulated |S31| measured -60 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 22: Measured and simulated |S11| and |S31|.

-10

-20

-30

-40

|S21| and |S41| (dB) |S21|and |S21| simulated -50 |S21| measured |S41| simulated |S41| measured -60 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 23: Measured and simulated |S21| and |S41|.

3 3 AFSIW

2 4 4 2

DFSIW

1 1 (a) (b) Figure 24: E-field magnitude at 30 GHz: (a) top and (b) bottom view.

Finally, the fabrication tolerance of the proposed Moreno coupler is studied. During the fabrication process, a precision laser micromachining is used to remove the central dielectric portion so to obtain the AFSIW whereas the crossed slots are wet etched. During these processing steps, some fabrication tolerances can lead to misalignment of the air part or the crossed slots. To illustrate the effect of this misalignment, simulated |S21| and |S41| with 0

123 and 0.1 mm misaligned slots in x and y directions are shown in Fig. 25. It can be observed from Fig. 25 that this design is quite robust to fabrication tolerances. The coupling (|S41|) is not affected whereas the isolation (|S21|) is increased by a few dB at low frequencies.

-15

x=0 mm and y=0 mm -25 |S41| x=0.1 mm and y=0 mm x=0 mm and y=0.1 mm -35 x=0.1 mm and y=0.1 mm

-45 |S21| |S21| and |S41| (dB) |S21|and

-55 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 25: Effect of 0.1 mm slot misalignment on |S21| and |S41|.

3.2.2. Two-Hole Coupler

3.2.2.1. Directional Two-Hole Coupler Structure

The proposed multilayer PCB directional two-hole coupler geometry is shown in

Fig. 26. It consists of parallelized AFSIW and DFSIW coupled through two crossed-slots (A1 and A2), located near the DFSIW sidewalls where the magnetic field for the fundamental TE10 mode is the strongest. It will be considered throughout this section that port 1 is the input port, port 2 is the through port, port 3 is the coupled port and port 4 is the isolated port.

The proposed multilayer PCB directional two-hole coupler geometry, based on different dielectric-loaded SIW, is shown in Fig. 26. It consists of parallelized AFSIW and DFSIW coupled through two crossed-slots (A1 and A2). The DFSIW is made on the top substrate of the AFSIW three-layer PCB structure. It will be considered throughout the paper that port 1 is the input port, and port 2 is the through port.

To theoretically obtain the coupling and isolation values, the first step is to determine the electric αe and magnetic αm polarizabilities of the coupling slots. In [26] and [27],

124 electrolytic-tank measurements have been performed to obtain both polarizabilities for different slot shapes, including cross-shaped slots of width ws and length l.

Then, the theoretical analysis, reported in [25] to determine the forward cF and reverse cR coupling coefficients of a single slot by taking into account unequal-width equal-cutoff- frequency coupled AFSIW and DFSIW, can be used considering the permittivity of the

DFSIW, which phase constant and impedance are given by (1.20) and (1.21), respectively.

훽퐷퐹푆퐼푊 = √휀푟훽퐴퐹푆퐼푊 (1.20)

and 푍퐷퐹푆퐼푊 = √휀푟푍퐴퐹푆퐼푊 (1.21)

where βAFSIW and βDFSIW are the AFSIW and DFSIW propagation constants, respectively, and ZAFSIW and ZDFSIW are the AFSIW and DFSIW impedances, respectively.

F R Finally, the forward C and reverse C coupling of two slots, spaced by a distance d1, are obtained from (1.22) and (1.23).

퐶푅 = 푐푅(1+ 푒 −푗(훽퐴퐹푆퐼푊+훽퐷퐹푆퐼푊)푑1) (1.22)

퐶퐹 = 푐퐹(푒−푗훽퐴퐹푆퐼푊푑1 + 푒−푗훽퐷퐹푆퐼푊푑1) (1.23)

Through Coupled Port 2 Port 3 w AFSIW Copper Substrate Air DFSIW Invisible elements A2 A2

y w d1 s d2 x A1 A1 l

(βAFSIW) (βDFSIW)

aAFSIW aDFSIW Input Isolated Port 1 Port 4

(a) (b) Figure 26: Geometry of the proposed directional two-hole coupler: top views of the (a) inner and (b) top substrates.

125

As an example, 0.508 mm thick Rogers RT/Duroid 6002 material of permittivity

εr = 2.94 for the middle and top substrates and AFSIW and DFSIW of respective widths

aAFSIW and aDFSIW (given in Table IV) determined for broadband Ka-band operation are

considered. Also, cross-shaped slots of ws = 0.31 mm, l = 1.6 mm and at distance d2 from the

AFSIW sidewalls are used. The comparison between theoretical and simulated coupling and

isolation are shown in Fig. 27(a) for forward (d1 = 2 mm and d2 = 2.27 mm), and in Fig. 27(b)

for backward (d1 = 10 mm and d2 = 3.52 mm) coupling. The corresponding simulated DFSIW

E-field magnitudes at 30 GHz are shown in both Fig. 28(a) and Fig. 28(b). It can be observed

in Fig. 27 that a good agreement between theory and simulation is obtained. The proposed

theoretical analysis is useful to rapidly generate a preliminary design procedure of the

proposed coupler.

Dimensions aAFSIW aDFSIW w d1 d2 ws l

Values (mm) 7.04 4.11 0.508 2.5 2.27 0.31 1.6

Table IV: Dimensions of the prototyped Ka-band two-hole coupler.

0 0

-20 -20 |S31| |S31|

|S41| -40 -40 |S41|

-60 -60

|S31| and |S41| |S31|(dB) and |S31| and |S41| |S31|(dB) and Simulated Simulated Theory Theory -80 -80 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) (a) (b) Figure 27: Comparison between theoretical and simulated |S31| and |S41| for (a) forward coupling with d1 = 2 mm and d2 = 2.27 mm and (b) backward coupling with d1 = 10 mm and d2 = 3.52 mm.

126

Port 3

Port 4 (a) (b) Figure 28: DFSIW E-field magnitude at 30 GHz for (a) forward coupling and (b) backward coupling.

3.2.2.2. Experimental Results

For demonstration purposes, using the previously described theoretical design equations, a 20 dB forward coupler achieving more than 20 dB of directivity in the Satcom uplink Ka-band going from 27 to 31 GHz has been designed based on 0.508 mm thick Rogers

RT/Duroid 6002 material of permittivity εr = 2.94 for the middle and top substrates. The

AFSIW, as shown in Fig. 29(a), was made in the inner substrate using a laser micromachining process. Then, the top substrate, implementing the coupled DFSIW, also shown in Fig. 29(a), and a bottom 0.806 mm low-cost FR-4 substrate are mounted using a multilayer PCB process.

Port 2 Port 2 Top Substrate

CBCPW to DFSIW DFSIW transition Port 3 Port 3

Reference Crossed AFSIW planes y slots

x Port 4 Port 4 DFSIW to AFSIW transition Inner Substrate Port 1 Port 1

(a) (b)

Figure 29: Fabricated two-hole coupler: (a) the inner and the top substrate before assembling and (b) the mounted prototype.

127

A picture of the assembled demonstrator is shown in Fig. 29(b). All the dimensions illustrated in Fig. 26 are given in Table IV. To simplify the measurements, the experimental prototype deployed with conductor-backed coplanar waveguides (CBCPW)-to-DFSIW and

DFSIW-to-AFSIW transitions.

The measurements have been achieved using a Vector Network Analyzer (VNA) and a test fixture. A through-reflect-line (TRL) calibration has been used to de-embed connectors and CBCPW to SIW transitions. Reference planes are shown in Fig. 29(b). The S-parameters are depicted in Fig. 30 and Fig. 31. Conductor and dielectric losses are taken into account in the simulations. A good agreement between the theoretical, measured and simulated results is obtained. In the theory, as well as in the simulation, the AFSIW to DFSIW transitions are not involved. This explains the observed discrepancy with respect to the measured results. In its frequency range of operation (27 to 31 GHz), the coupler achieves the measured insertion loss, coupling and isolation of 0.28 ± 0.15 dB, 20.8 ± 0.24 dB and 46.76 ± 3.92 dB, respectively. A measured return loss of better than 20 dB is obtained over the entire Ka-band.

The simulated E-field magnitude is shown at 28 GHz in Fig. 32. It can be observed that the signal, fed at port 1, is properly coupled at port 3, and that port 4 is isolated.

0 |S21| -10

-20

-30

|S11| simulated -40 |S11| |S11| measured |S11| and |S21| (dB) |S11|and |S21| simulated -50 |S21| measured BW -60 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 30: Measured and simulated |S11| and |S21|.

128

-10

-20 |S31| -30 |S41| theoritical -40 |S41| simulated |S41| |S41| measured

-50 |S31| theoritical |S31| and |S41| (dB) |S31|and |S31| simulated -60 |S31| measured BW -70 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 31: Measured and simulated |S31| and |S41|.

Port 2 Port 3

AFSIW

DFSIW Port 1 Port 4 (a) (b) Figure 32: (a) AFSIW and (b) DFISW E-field magnitude at 28 GHz.

Finally, misalignments of the wet etched cross-shaped slots in connection with the laser micro-machined AFSIW section may occur in fabrication. To illustrate the effect of this misalignment, simulated |S31| and |S41| with -0.1, 0 and 0.1 mm misaligned slots in x- direction are compared in Fig. 33. As it can be observed this design is very robust to misalignment.

-10

-20 |S31| -30 |S41|

-40

-50 x=0.1 mm

|S21| and |S41| (dB) |S21|and x=0 mm -60 x=-0.1 mm BW -70 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 33: Simulated slot x-direction misalignment effect on |S31| and |S41| of the prototyped coupler.

129

4. PHASE SHIFTERS

Phase shifter is one of the principal components required in the development of millimeter-wave circuits and systems. Dielectric-filled SIW phase shifters have been reported in the literature. In [27], a broadband phase shifter based on the synthesis of a low dielectric slab in the middle of the SIW using an array of air holes was proposed. In [30], a tunable non reciprocal ferrite loaded SIW phase shifter based on a non-reciprocal ferrite loaded SIW was reported. In [31], a phase shifter based on phase channels made of SIW resonators loaded with extra metallic posts was presented. And in [32], delay line and equal-length unequal-width

SIW guided-wave structures were studied for the design of phase shifters. Both structures are used to design an unequal-length compensated phase shifter for achieving broadband characteristics.

Copper Substrate Air Substrate 3

Substrate 2

Substrate 1

W w 2

h W1

Figure 34: AFSIW structure.

Based on the double dielectric slab-loaded AFSIW shown in Fig. 34, the propagation properties can be controlled by adjusting the slab width w and total width W1 to vary the propagation constant β to design phase shifters.

130

4.1. Double Dielectric Slab-Loaded Air-Filled SIW Phase Shifter

In this section, equal-length unequal-dielectric-slab-width and equal-length unequal- dielectric-slab-width-and-total-width phase shifters are introduced. Then, based on the two previous phase shifters offering reverse varying behaviors against frequency, a broadband compensated equal-length phase shifter is proposed. For demonstration purposes, Rogers

RT/Duroid 6002 with thickness h = 0.508 mm and relative dielectric constant εr = 2.94 is used for the inner substrate (substrate 2 in Fig. 34). For Ka-band operation, AFSIW widths

W1 = 7.04 mm and w = 0.508 mm are obtained. Standard FR-4 substrates are used for the top and bottom substrates (substrate 1 and 3 in Fig. 34).

4.1.1. Equal-Length Unequal-Dielectric-Slab-Width Phase Shifter

The geometry of the proposed equal-length unequal-dielectric-slab-width phase shifter is illustrated in Fig. 35. It is based on the propagation constant difference offered by varying the two slab widths w (Fig. 34). As the obtained transmission line section has an unmatched guided-wave impedance, two linearly tapered transitions of length l have been implemented to improve return loss. The slabs width wp and transmission line section length L can be used to achieve the desired phase shift.

Copper Substrate Air

w wp Dielectric Linear W 1 slabs transitions

w L l

Figure 35: Illustration of the equal-length unequal-dielectric-slab-width air-filled SIW phase shifter.

131

For demonstration purposes, a 90° phase shifter is designed. Considering L = 4 mm and wp = 1.5 mm, the phase difference between the reference air-filled SIW and the double dielectric-slab width SIW obtained in theory and simulation are plotted in Fig. 36. Then, linearly tapered transitions are designed. Their effect on return loss is illustrated in Fig. 37 for different taper length l. It can be seen that a trade-off between the total length and return loss is obtained for l = 3 mm, thus allowing a matching of better than -10 dB over Ka-band.

The linear tapered transition phase contribution is theoretically determined by a discretization in segments of ∆l length and the determination of propagation constant for each segment. Their contribution considering eight segments is shown in Fig. 36. Both contributions of the double slab loaded SIW and the transitions, are added up to obtain the overall phase shift. It can be seen in Fig. 36 that good agreement between theoretical and simulated results is achieved.

150 Theoritical 120 Simulated Over All 90 Double Slabs Loaded SIW 60

Phase Difference (°) Difference Phase 30 Linear Tapers 0 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 36: Theoretical and simulated phase shifts for different elements constituting the double dielectric slab loaded air-filled SIW phase shifter.

-10

l=0mm -20 l=1.5mm l=3mm |S11|(dB) l=6mm -30

-40 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 37: Simulated phase shifter transition length l parametric sweep.

132

A picture of the fabricated prototype is shown in Fig. 38. Simulated and measured S- parameters are plotted in Fig. 39. Measurements were performed with a Vector Network

Analyzer (VNA) and a test fixture using TRL calibration to de-embed effects of connectors,

CBCPW to SIW transitions and SIW to air-filled SIW transitions. Simulations are taking into account conductor and dielectric losses. Measured and simulated results are in good agreement. The measured |S11| is lower than -10 dB and measured |S21| is 0.32 ± 0.2 dB over the entire Ka-band. It can be noticed that the insertion loss is improved by the use of air-filled

SIW compared to the SIW phase shifters proposed in [31] with an insertion loss of 1.5 dB at

25 GHz and in [32] for which insertion losses are 0.7 dB at 32 GHz.

Fig. 40 shows the achieved simulated and measured phase shifts relatively to an air- filled SIW transmission line of equal length, with the simulated E-field magnitude at 28 GHz.

The phase shifter achieves a relative phase shift of 90° at 34.4 GHz with a ± 5° frequency band of 31.5 to 35.6 GHz. The amplitude imbalance between the two circuits is

0.19 dB ± 0.14 dB within this frequency range. As it can be observed in Fig. 40, this phase shifter achieves an increasing tendency versus frequency.

l SIW to air-filled SIW CBCPW to SIW L

Equal-length unequal-dielectric-slab-width phase shifter SIW Figure 38: Fabricated equal-length unequal-dielectric-slab-width air-filled SIW phase shifter inner substrate.

0 0

-5  -0.5 -1 -10  -1.5 -15 -2

-20

|S21||S21|(dB) (dB) |S11|(dB) -2.5 S11 simulated -25 S11 measured -3 S21 simulated -30 S21 measured -3.5

-35 -4 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 39: Simulated and measured S-parameters of the double dielectric slab loaded phase shifter.

133

130 2

120 1.5 1 110 0.5 100  0 90  -0.5 80

Phase Difference (°) Difference Phase -1 Amplitude Imbalance (dB) Imbalance Amplitude 70 Simulated -1.5 (dB) Imbalance Amplitude Measured 60 -2 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 40: Simulated and measured relative phase shifts with E-field magnitude distribution simulated at 28 GHz.

The same phase shifter has been realized in U-band (40-60 GHz). The dimensions are

W1 = 4.6 mm, w = 0.508 mm, wp = 0.51 mm, and refer to Fig. 35. Simulated and measured results are given in Fig. 41. The measured |S11| is lower than -13.9 dB and the measured |S21| is 0.15 ± 0.14 dB over the U-band. It can be noticed that the insertion losses are improved by the use of the air-filled SIW when comparing with [31] and [32] in which insertion losses are

0.7 dB at 32 GHz and 1.5 dB at 25 GHz, respectively. Fig. 42 shows the achieved simulated and measured relative phase shift in U-band of the proposed phase shifter relatively to an air- filled SIW transmission line of equal length. The phase shifter achieves a phase difference of

± 2.5° over the entire band.

0 0  -0.5 -10 -1

-1.5

-20 -2

|S21||S21|(dB) (dB) |S11|(dB) -2.5  -30 -3

Simulated -3.5 Measured -40 -4 40 42 44 46 48 50 52 54 56 58 60 Frequency (GHz) Figure 41: Simulated and measured S-parameters of the phase shifter in U-band.

134

50 40 30 20 10 Simulated Measured 0 Relative phase shift (°) shift phase Relative 40 42 44 46 48 50 52 54 56 58 60 Frequency (GHz) Figure 42: Simulated and measured relative phase shift

Table V summarizes the cut-off frequencies of the first three modes of the air-filled

SIW and the double dielectric slab loaded air-filled SIW sections of the demonstrated Ka- band phase shifter. The third quasi-TE30 mode appears well above the frequency of interest for both sections. However, the second quasi-TE20 mode cut-off frequency of the double dielectric slab loaded air-filled SIW section falls within the Ka-band. Nevertheless, the proposed symmetrical structure and feeding does not excite this mode. Though, as studied in

Chapter III, Section 1.2.1.2, the quasi-TE20 mode can be coupled and creates resonances in the case of misalignment.

Double dielectric slab Section Air-filled SIW loaded air-filled SIW 18 Quasi-TE10 cut-off frequency (GHz) 22 28.3 Quasi-TE20 cut-off frequency (GHz) 42 46.55 Quasi-TE30 cut-off frequency (GHz) 61.1 Table V: Simulated first three mode cut-off frequencies of the air-filled SIW and double dielectric slab loaded air-filled SIW sections.

4.1.2. Equal-Length Unequal-Dielectric-Slab-Width-and-Total-Width Phase Shifter

The phase shifter described in the previous sub-section was based only on the slab width variation. However, a broadband response cannot be achieved because of the increasing propagation constant difference slope over frequency. To obtain a broadband phase shifter, compensation with an unequal-length phase shifter was proposed in [32]. However, unequal

135 length phase shifter is not very practical from a design point of view. To achieve compensation, a decreasing propagation constant difference versus frequency can rather be obtained adjusting the total width W1 (Fig. 35). Fig. 43 plots the propagation constant of the reference air-filled SIW (W1 = 7.04 mm and w = 0.508 mm) and an enlarged total width air- filled SIW (W1 = 7.64 mm and w = 0.508 mm). It can be seen in this figure that the difference between the two propagation constants offers a decreasing tendency versus frequency.

800 200 Air-filled SIW 700 Enlarged total width Propagation constant difference 160 600

500  120 400

300 80

200  40

Propagation constant (rad/m) Propagation 100

Propagation constant difference (rad/m) constant difference Propagation 0 0 (rad/m) constant difference Propagation 18 20 22 24 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 43: Theoretical propagation of a reference air-filled SIW and an unequal- dielectric-slab-width-and-total-width air-filled SIW.

4.1.3. Broadband Compensated Phase Shifter

In [32], an unequal-length broadband compensated phase shifter taking advantages of the opposite slops of a dispersive phase shift versus frequency of the unequal-length delay line SIW structure and the equal-length unequal-width SIW structure is proposed. However, for practical case, it is important to achieve an equal-length broadband compensated phase shifter to design complex feeding networks. Such a phase shifter is possible based on the opposite tendency of the equal-length unequal-slab-width and the equal-length unequal-width- slab-width-and-total-width air-filled SIW structures presented in the previous sub-sections.

The proposed compensated structure is illustrated in Fig. 44. It consists of the two cascaded phase shifters.

136

As illustrated in Fig. 45, for W1 = 7.04 mm, w = 0.508 mm, wp = 0.8 mm and wp2 = 0.3 mm, the first phase shifter offers an increasing propagation constant difference versus frequency while the second provides a decreasing propagation constant difference against frequency. Cascading the two structures, therefore it is possible to compensate the phase shift to obtain a broadband response. Linearly tapered transition of length l1 and l2 are used for matching purposes.

Copper Substrate Air

wp2 w wp Dielectric Linear l2 W1 slabs transitions

l L2 w L1

Figure 44: Illustration of the compensated phase shifter.

200 Type1 Type2 160

120

0 Propagation constant difference (rad/m) constant difference Propagation 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 45: Theoretical propagation constant difference between a reference air-filled SIW and an unequal-dielectric-slab-width air-filled SIW (type 1) and between a reference air-filled SIW and an unequal-dielectric-slab-width-and-total-width air-filled SIW (type 2).

CBCPW to SIW Equal-length unequal-dielectric-slab-width SIW to air-filled SIW phase shifter

SIW Equal-Length Unequal-Dielectric-Slab-Width-and-Total-Width Phase Shifter Figure 46: Picture of the fabricated compensated air-filled SIW phase shifter inner substrate.

137

For demonstration purposes, a prototype, illustrated in Fig. 46, has been designed and fabricated. Its dimensions are W1 = 7.04 mm, w = 0.508 mm, L1 = 4 mm, l = 3 mm,

L2 = 4 mm, l2 = 3 mm, wp = 0.8 mm and wp2 = 0.3 mm. The total length of the compensated phase shifter and its air-filled SIW transmission line reference are 20 mm.

Simulated and measured S-parameters of the proposed compensated phase shifter are shown in Fig. 47. The measured |S11| is lower than -12.5 dB and the measured |S21| is

-0.35 ± 0.17 dB over the Ka-band. Fig. 48 shows the achieved simulated and measured phase shifts over the Ka-band, with the simulated E-field magnitude at 33 GHz. The phase shifter achieves a relative phase shift of 43° ± 6° over the entire Ka-band with an amplitude imbalance of 0.23 ± 0.2 dB between measured and simulated |S21|.

0 0

-0.5 -5 S11 simulated  S11 measured -1 -10 S21 simulated S21 measured -1.5 -15 -2

-20

|S21||S21|(dB) (dB) |S11|(dB)  -2.5 -25 -3

-30 -3.5

-35 -4 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 47: Simulated and measured S-parameters of the proposed compensated phase shifter.

60 2

1.5 50  1 40 0.5 30  0 -0.5 20

Phase Difference (°) Difference Phase -1 Amplitude Imbalance (dB) Imbalance Amplitude 10 (dB) Imbalance Amplitude Simulated -1.5 Measured 0 -2 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 48: Simulated and measured phase shifts of the proposed compensated phase shifter with E-field magnitude distribution simulated at 33 GHz.

138

60 2

1.5 50  1 40 0.5 30  0 -0.5 20

Phase Difference (°) Difference Phase -1 Amplitude Imbalance (dB) Imbalance Amplitude 10 (dB) Imbalance Amplitude Simulated -1.5 Measured 0 -2 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 49: Simulated and measured relative phase shifts using the method described in [31], for a total length of 10 mm for the double dielectric slab loaded phase shifter and for a total length Lref = 11.6 mm for the reference.

For comparison purposes, a double dielectric slab-loaded phase shifter, introduced in the previous section, is compensated with a longest reference as proposed in [32]. Its dimensions are W1 = 7.04 mm, w = 0.508 mm, L = 4 mm, l = 3 mm and wp = 1.5 mm. Using the EM solver, the reference length for the compensation effect is found for Lref = 11.6 mm whereas the phase shifter is 10 mm long. This results in a length difference of ∆L = 1.6 mm.

Fig. 49 shows the achieved simulated and measured phase shifts over the Ka-band. This phase shifter achieves a relative phase shift of 46.3° ± 6° over the entire Ka-band with an amplitude imbalance between the two circuits of 0.13 ± 0.13 dB.

The equal-length and unequal-length compensated phase shifters achieve a comparable performances as summarized in Table VI. For design simplicity of complex networks, the equal-length compensated phase shifter is preferred.

Bandwidth ΔL Amplitude imbalance Phase shift

(GHz) (mm) (dB) (°) Equal-length (this Ka-band 0 0.23 ± 0.2 43 ± 6 work) Unequal-length [32] Ka-band 1.6 0.13 ± 0.13 46 ± 6 Table VI: Comparison between measured Air-filled SIW equal-length and unequal- length compensated phase shifter performances.

139

4.2. Single Dielectric Slab-Loaded Air-Filled SIW Phase Shifter

The single dielectric slab loaded air-filled SIW is another geometry that could be considered to allow adjusting the propagation properties of an air-filled SIW and achieve a phase shifter. Such phase shifter geometry is illustrated in Fig. 50. Linear tapered transitions are also used for matching purposes. A 90° phase shifter, centered at 33 GHz and based on this structure, is designed for W1 = 7.04 mm, w = 0.508 mm, L = 4 mm, wp = 1.5 mm and l = 3 mm. Simulated S-parameters are plotted in Fig. 51. It can be seen that the bandwidth of such a structure is limited.

Copper Substrate Air L

w wp Linear W 1 Dielectric transitions slabs w

Figure 50: Structure of the single dielectric slab loaded air-filled SIW phase shifter.

0 0

-5 -5

-10 -10

-15 -15

-20 -20

|S11|(dB) |S21||S21|(dB) (dB) -25 -25

-30 S11 -30 S21

-35 -35 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 51: Simulated S-parameters of the single dielectric slab-loaded phase shifter.

The simulated relative phase shift is shown in Fig. 52. It can be observed that a strong variation over frequency is generated. Therefore, this structure is not desired for a broadband operation. To understand the reason of this limitation, electric field (E-field) magnitude before

140

(at 32 GHz) and at the resonance frequency of 36.6 GHz are shown in Fig. 53. It can be seen in Fig. 53(b) that there is a higher-order mode. Due to un-symmetricity, this higher-order mode is coupled to the fundamental mode and creates a resonance, limiting the bandwidth.

200

160

120

Relative Phase Shift (°) Shift Phase Relative 40

0 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 52: Simulated relative phase shift of the single dielectric slab-loaded phase shifter.

(a) (b) Figure 53: Simulated E-field magnitude of the single dielectric slab loaded phase shifter (a) at 32 GHz and (b) at the resonant frequency of 36.6 GHz.

5. CONCLUSION

In this chapter, two elementary circuits, T-junction and 90° hybrid coupler have been demonstrated over Ka-band. All presented simulated and experimental results are showing good agreement. Then, a directional Moreno coupler and a broadband air-to-dielectric-filled

SIW directional two-hole coupler intended for high-performance substrate integrated systems based on air-filled substrate integrated waveguide have been presented. They are showing significant interest for signal monitoring applications in high-performance microwave and millimeter-wave substrate integrated systems based on AFSIW. Finally, double dielectric slab-loaded air-filled SIW phase shifters has been introduced and studied. During the first step, the double dielectric slab propagation properties have been theoretically derived to

141 obtain the propagation constant over geometrical properties of the air-filled SIW. Then, based on the phase propagation control, equal-length unequal-dielectric-slab-width and equal-length unequal-dielectric-slab-width-and-total-width phase shifters have been proposed and a broadband equal-length compensated phase shifter is demonstrated. At the end, a single dielectric slab-loaded air-filled SIW phase shifter is investigated for comparison and its bandwidth limitation is highlighted. All the proposed devices, based on a low-cost transmission line, are envisioned to be deployed in applications where low noise, efficient or high power handling integrated front-ends are desired, for example, for emerging millimeter- wave sensing, telecommunication and energy transfer systems.

142

REFERENCES

[1] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp, and K. Wu, "Air-filled substrate integrated waveguide for low-loss and high power-handling millimeter-wave substrate integrated circuits," in IEEE Transactions on Microwave Theory and Techniques, vol. 63, no. 4, pp. 1228-1238, April 2015. [2] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp and K. Wu, "Broadband directional moreno coupler for high-performance air-filled SIW-based substrate integrated systems," 2016 IEEE MTT-S International Microwave Symposium (IMS2016), San Francisco, CA, 2016, pp. 1-3. [3] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp, and K. Wu, "Air-to-dielectric- filled two-hole substrate integrated waveguide directional coupler," in IEEE Microwave and Wireless Components Letters, submitted in August 2016. [4] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp, and K. Wu, "Double dielectric slab-loaded air-filled SIW phase shifters for high-performance millimeter-wave integration," in IEEE Transactions on Microwave Theory and Techniques, vol. 64, no. 9, Sept. 2016. [5] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp, and K. Wu, "Air-filled SIW transmission line and phase shifter for high-performance and low-cost U-band integrated circuits and systems," Millimeter Waves (GSMM), 2015 Global Symposium On, Montreal, QC, 2015, pp. 1-3. [6] S. Germain, D. Deslandes, K. Wu, "Development of substrate integrated waveguide power dividers," CCECE 2003 proceedings, pp. 1921-1924, May 2003. [7] He Fanfan, Wu Ke, Chen Xiao-Ping, Han Liang, and Hong Wei, "A planar magic-T structure using substrate integrated circuits concept," IEEE Microw. Sympo. Digest, pp.720-723, 2010. [8] N. A. Smith and R. Abhari, "Compact substrate integrated waveguide Wilkinson power dividers," Proc. IEEE Antennas Propag. Soc. Int. Symp., pp. 1-4. [9] M. Pasian, L. Silvestri, M. Bozzi, L. Perregrini, and K. K. Samanta, "E-plane 3-dB power divider/combiner in substrate integrated waveguide technology," Microwave Conference (EuMC), 2015 European, Paris, 2015, pp. 478-481. [10] Z. Xie, B. Liu, Y. Zhao, B. Tian, and S. Jia, "A novel Ka band multi-layer SIW power divider," Cross Strait Quad-Regional Radio Science and Wireless Technology Conference (CSQRWC), 2011, Harbin, 2011, pp. 634-636. [11] H. Abuzaid, A. Doghri, K. Wu, and A. Shamim, "SIW based multilayer transition and power divider in LTCC technology," Microwave Symposium Digest (IMS), 2013 IEEE MTT-S International, Seattle, WA, 2013, pp. 1-3. [12] D. M. Pozar, Microwave Engineering, John Wiley and Sons, (New York, 1998). [13] C.-J. Chen and T.-H. Chu, "Design of a 60-GHz substrate integrated waveguide Butler matrix—A systematic approach," IEEE Trans. Microw. Theory Tech., vol. 58, no. 7, pp. 1724, 2010.

143

[14] S. B. Cohn and R. Levy, "History of microwave passive components with particular attention to directional couplers," IEEE Trans. Microw. Theory Tech., vol. 32, pp. 1046–1054, 1984. [15] A. Doghri, T. Djerafi, A. Ghiotto, and K. Wu, "Substrate integrated waveguide directional couplers for compact three-dimensional integrated circuits," in IEEE Transactions on Microwave Theory and Techniques, vol. 63, no. 1, pp. 209-221, Jan. 2015. [16] H. J. Rlblet, "The Short-Slot Hybrid Junction", Proc. IRE, vol. 40, pp. 180-184. [17] W. M. Abdel-Wahab and S. Safavi-Naeini, "Low loss H-shape SIW hybrid coupler for millimeter-wave phased arrays antenna systems," Proc. IEEE Antennas Propag. Soc. Int. Symp., pp. 1-2. [18] A. A. M. Ali, H. B. El-Shaarawy and H. Aubert, "Compact wideband double-layer half-mode substrate integrated waveguide 90°deg coupler," in Electronics Letters, vol. 47, no. 10, pp. 598-599, May 12 2011. [19] S. Yamamoto, J. Hirokawa, and M. Ando, "A half-sized post-wall short-slot directional coupler with hollow rectangular holes in a dielectric substrate," IEICE Trans. Electron., vol. 88, no. 7, pp. 1387–1394, Jul. 2005. [20] A. Doghri, T. Djerafi, A. Ghiotto, and K. Wu, "Broadband substrate-integrated- waveguide six-port applied to the development of polarimetric imaging radiometer," Proc. Eur. Microw. Conf., pp. 393-396, 2011. [21] S. B. Cohn and R. Levy, "History of microwave passive components with particular attention to directional couplers," IEEE Trans. Microw. Theory Tech., vol. 32, pp. 1046–1054, 1984. [22] J. A. Ball and T. M. Sulda, "Small aperture crossed waveguide broad-wall directional couplers," in Proc. Inst. Elect. Eng., 2000, pp. 249–254. [23] M. Surdin, "Directive couplers in wave guides," in Journal of the Institution of Electrical Engineers - Part IIIA: Radiolocation, vol. 93, no. 4, pp. 725-736, 1946. [24] T. N. Anderson, "Directional coupler design nomograms," Microwave Journal, 2, pp. 34-38, 1959. [25] K. Rambabu, H.A. Thiart, and J. Bornemann, "Simplified analysis for the initial design of rectangular waveguide cross-slot couplers," IEE Proc.- Microwave Antennas Propag., vol. 152, no. 4, August 2005, pp. 226- 230. [26] S. B. Cohn, "Determination of aperture parameters by electrolytic-tank measurements," in Proceedings of the IRE, vol. 39, no. 11, pp. 1416-1421, Nov. 1951. [27] S. B. Cohn, "The electric polarizability of apertures of arbitrary shape," in Proceedings of the IRE, vol. 40, no. 9, pp. 1069-1071, Sept. 1952. [28] A. Ali, N. J. G. Fonseca, F. Coccetti, and H. Aubert, "Analysis and design of a compact SIW-based multi-layer wideband phase shifter for Ku-band applications," 2010 IEEE Antennas and Propagation Society International Symposium, Toronto, ON, 2010, pp. 1-4. [29] I. Boudreau, K. Wu, and D. Deslandes, "Broadband phase shifter using air holes in substrate integrated waveguide," Proc. IEEE MTT-S Int. Microw. Symp., pp.1-4, Jun. 2011.

144

[30] S. Adhikari, A. Ghiotto, S. Hemour, and K. Wu, "Tunable non-reciprocal ferrite loaded SIW phase shifter," in Proc. IEEE MTT-S Int. Microw. Symp., pp. 1-3, Jun. 2013. [31] T. Yang, M. Ettorre, and R. Sauleau, "Novel phase shifter design based on substrate integrated waveguide technology," IEEE Microw. Wireless Compon. Lett., vol. 22, no. 10, pp. 518–520, Oct. 2012. [32] Y. J. Cheng, K. Wu, and W. Hong, "Substrate integrated waveguide (SIW) broadband compensating phase shifter," in IEEE MTT-S Int. Microw. Symp. Dig., 2009, pp. 845– 848.

145

146

CHAPTER V: AIR- FILLED SIW FILTERS AND ANTENNAS

147

TABLE OF CONTENTS

TABLE OF CONTENTS ...... 148 LIST OF FIGURE ...... 149 LIST OF TABLE ...... 151 1. INTRODUCTION ...... 152 2. AFSIW FILTERS ...... 152

2.1. AFSIW BAND-PASS FILTER WITH INDUCTIVE POSTS: ...... 153 2.1.1. Design approach: ...... 154 2.1.1.1. Asymmetrical inductive posts: ...... 154 2.1.1.2. Symmetrical inductive posts: ...... 157 2.1.2. Experimental results: ...... 158 2.1.2.1. SIW bandpass filter: ...... 158 2.1.2.2. AFSIW bandpass filter: ...... 159 2.1.2.3. Comparison:...... 160 2.2. AFSIW MULTILAYER BAND-PASS FILTER: ...... 161 2.2.1. Geometry: ...... 161 2.2.2. Experimental results: ...... 163 3. ANTENNAS ...... 165

3.1. AFSIW SLOT ANTENNA: ...... 165 3.1.1. SIW and AFSIW slot antenna design: ...... 166 3.1.1.1. AFSIW slot antenna: ...... 166 3.1.1.2. SIW slot antenna:...... 168 3.1.2. Experimental results: ...... 169 3.1.2.1. Fabrication: ...... 169 3.1.2.2. Results and discussion: ...... 170 3.2. AFSIW ANTIPODAL LINEARLY TAPERED SLOT ANTENNA: ...... 172 3.2.1. SIW and AFSIW ALTSA design: ...... 173 3.2.1.1. SIW ALTSA: ...... 173 3.2.1.2. AFSIW ALTSA: ...... 174 3.2.2. Experimental results: ...... 175 3.2.2.1. Fabrication: ...... 175 3.2.2.2. Results and discussion: ...... 176 4. CONCLUSION: ...... 179 REFERENCES ...... 180

148

LIST OF FIGURES

FIGURE 1: TOPOLOGY OF SEVERAL SIW FILTERS: (A) IN CERAMIC USING INDUCTIVE POSTS, (B) IN PCB USING THREE DIMENSIONAL STUBS AND (C) IN PCB USING FERRITES AND VARACTORS FOR TENABILITY PURPOSES...... 153 FIGURE 2: SIW FILTERS USING IRIS [9] (A) AND INDUCTIVE POSTS [3] (B) AS COUPLING ELEMENTS...... 154 FIGURE 3: PICTURE OF A BANDPASS FILTER USING 2 CAVITIES AND 3 INDUCTIVE POSTS, WHERE

W IS THE SIW WIDTH, D IS THE CONDUCTIVE VIA DIAMETER, XI IS THE OFFSET AND LI IS THE PHYSICAL LENGTH OF A CAVITY...... 155 FIGURE 4: CROSS SECTIONAL VIEW IN (A), TOP VIEW IN (B) AND EQUIVALENT CIRCUIT IN (C) OF AN INDUCTIVE POST...... 155 FIGURE 5: (A) EQUIVALENT INDUCTIVE POST AND (B) K-INVERTER MODEL...... 156 FIGURE 6: PICTURE OF A BAND-PASS FILTER USING 2 CAVITIES AND 3 PAIRS OF INDUCTIVE

POSTS, WHERE W IS THE SIW WIDTH , D IS THE INDUCTIVE POST DIAMETER, XI IS THE OFFSET

AND LI IS THE CAVITY PHYSICAL LENGTH...... 157 FIGURE 7: PICTURE OF THE DIELECTRIC-FILLED SIW BAND-PASS FILTER...... 158 FIGURE 8: MEASURED AND SIMULATED RESULTS OF THE DIELECTRIC-FILLED SIW BAND-PASS FILTER WITH SIMULATED E-FIELD MAGNITUDE AT 31 GHZ...... 159 FIGURE 9: PICTURE OF THE AIR-FILLED SIW BAND-PASS FILTER...... 159 FIGURE 10: MEASURED AND SIMULATED RESULTS OF THE AIR-FILLED SIW BAND-PASS FILTER, WITH SIMULATED E-FIELD MAGNITUDE AT 31 GHZ...... 160 FIGURE 11: SEVERAL TECHNOLOGIES USED FOR MULTILAYER FILTER: (A) LTCC [13], (B) WAVEGUIDE [14], AND (C) PCB [15]...... 161 FIGURE 12: CROSS SIDE VIEW OF THE MULTILAYER AFSIW FILTER...... 162 FIGURE 13: TOP VIEW OF THE LAYER 4...... 162 FIGURE 14: TOP VIEW OF THE LAYER3...... 162 FIGURE 15: TOP VIEW OF THE LAYER 2...... 163 FIGURE 16: PICTURE OF THE AIR-FILLED MULTILAYER SIW BAND-PASS FILTER: (A) THE LAYER 2 CONTAINING THE AFSIW FEED AND CAVITIES C1 AND C4, (B) THE LAYER 3 WITH THE TWO VERTICAL IRIS, (C) THE LAYER 4 INCLUDING CAVITIES C3 AND C4, AND (D) TOP OR BOTTOM LAYER ENCLOSING THE STRUCTURE...... 163 FIGURE 17: MEASURED AND SIMULATED RESULTS OF THE MULTILAYER AIR-FILLED SIW BAND- PASS FILTER...... 164 FIGURE 18: MEASURED AND SIMULATED RESULTS OF THE MULTILAYER AIR-FILLED SIW BAND- PASS FILTER AFTER INVESTIGATION ON THE VARIABILITY PROCESS...... 164 FIGURE 19: TOPOLOGY OF SEVERAL SIW ANTENNAS: (A) HORN ANTENNA, (B) TEXTILE CAVITY- BACKED PATCH ANTENNA (WITH TOP AND BOTTOM VIEW) AND (C) YAGI-UDA ANTENNA...... 165 FIGURE 20: SLOT ANTENNA: (A) WAVEGUIDE H-PLANE [22] AND (B) SIW 4X4 ARRAY [23]...... 166 FIGURE 21: AFSIW SLOT ANTENNA GEOMETRY...... 167 FIGURE 22: SIW SLOT ANTENNA GEOMETRY...... 168

149

FIGURE 23: FABRICATED (A) TOP AND (B) INNER AFSIW SLOT ANTENNA, AND (C) SIW SLOT ANTENNA...... 169 FIGURE 24: MEASURED AND SIMULATED SIW (A) AND AFSIW (B) SLOT ANTENNA |S11|...... 170 FIGURE 25: SIW SLOT ANTENNA MEASURED AND SIMULATED HORIZONTAL POLARIZATION XY- PLANE (A) AND YZ-PLANE (B) RADIATION PATTERNS AT 30.5 GHZ...... 171 FIGURE 26: AFSIW SLOT ANTENNA MEASURED AND SIMULATED HORIZONTAL POLARIZATION (A) XY-PLANE AND (B) YZ-PLANE RADIATION PATTERNS AT 30.5 GHZ...... 171 FIGURE 27: TOPOLOGIES OF ATSA: (A) LINEAR [33], (B) PARABOLIC [34] AND (C) FERMI SHAPED [35]...... 172 FIGURE 28: SIW ALTSA GEOMETRY...... 173 FIGURE 29: AFSIW ATSA GEOMETRY...... 174 FIGURE 30: FABRICATED AFSIW ALTSA 6002 INNER SUBSTRATE...... 175 FIGURE 31: FABRICATED FR4 TOP AND BOTTOM SUBSTRATES OF THE AFSIW ALTSA...... 176 FIGURE 32: FABRICATED (A) SIW AND (B) AFSIW ALTSA...... 176 FIGURE 33: MEASURED AND SIMULATED SIW (A) AND AFSIW (B) ALTSA |S11|...... 177 FIGURE 34: SIW ALTSA MEASURED AND SIMULATED HORIZONTAL- (SOLID LINES) AND VERTICAL-POLARIZATION (DASHED LINES) XZ-PLANE (A) AND XY-PLANE (B) RADIATION PATTERNS AT 35.5 GHZ...... 177 FIGURE 35: AFSIW ALTSA MEASURED AND SIMULATED HORIZONTAL- (SOLID LINES) AND VERTICAL-POLARIZATION (DASHED LINES) XZ-PLANE (A) AND XY-PLANE (B) RADIATION PATTERNS AT 35.5 GHZ...... 177 FIGURE 36: SIW AND AFSIW ALTSA SIMULATED DIRECTIVITY AND RADIATION EFFICIENCY VERSUS FREQUENCY...... 178

150

LIST OF TABLES

TABLE I: DIMENSIONS OF DIELECTRIC-FILLED SIW BAND-PASS FILTER...... 159 TABLE II: DIMENSIONS OF AIR-FILLED SIW BAND-PASS FILTER...... 160 TABLE III: COMPARISON OF THE TWO FILTERS IN BOTH TECHNOLOGIES...... 160 TABLE IV: DIMENSIONS OF MULTILAYER AIR-FILLED SIW BAND-PASS FILTER...... 164 TABLE V: AFSIW SLOT ANTENNA DIMENSIONS (MM)...... 168 TABLE VI: SIW AND AFSIW SLOT ANTENNA DIMENSIONS (MM)...... 169 TABLE VII: COMPARISON BETWEEN SIW AND AFSIW SLOT ANTENNA...... 172 TABLE VIII: SIW AND AFSIW ALTSA DIMENSIONS (MM)...... 175 TABLE IX: COMPARISON BETWEEN SIW AND AFSIW ALTSA...... 178

151

1. INTRODUCTION

Filters and antennas are essential components found in millimeter-wave system frontend. Their losses directly impact receiver noise figure and transmitter efficiency.

Therefore low-loss technologies are highly desired. In this chapter, filters and antennas based on the proposed low-loss AFSIW technology are presented. The purpose of this chapter is not to reinvent filters and antennas, but to demonstrate some implementation based on AFSIW and compare performance to their SIW counterpart.

The first section of this chapter focuses on filters. A filter is a two-port network used to control the frequency response in a certain frequency band and provide out of band attenuation and in band transmission. Typical frequency responses include low-pass, high- pass, bandpass, and stop-band functions. Two different topologies of filters are explored: a filter based on inductive posts and a filter based on stacked cavities, which has not been proposed for publication yet. The AFSIW filters based on inductive post has been published in [1].

Finally, AFSIW antennas are introduced in the second section. An antenna can be defined as a transitional component which is able to convert guided waves into radiated free space waves. Two topologies of antennas, a slot array antenna, which has not been proposed for publication yet, and an Antipodal Linear Tapered Slot Antenna (ALTSA) are presented.

The AFSIW ALTSA antenna has been published in [2].

2. AFSIW FILTERS

Numerous SIW filtering structures have been reported based on PCB [3] technology including for example inductive posts filters [5], compact three dimensional filters [6] and simultaneously electrically and magnetically tuned filters [7], all shown in Fig. 1. Also, some

152

SIW filters have been demonstrated based on multilayer PCB [8] or LTCC [4] manufacturing process.

(a) (b)

(c) Figure 1: Example of some reported SIW filters: (a) based on inductive posts [4], (b) based on PCB using three dimensional stubs [6] and (c) based on PCB using ferrites and varactors for tunability purposes [7].

The purpose of this section is to demonstrate the interest of the AFSIW technique compared to its dielectric-filled counterpart. For that, an AFSIW band-pass filter using inductive posts is proposed, studied, and demonstrated as well as an AFSIW multilayer bandpass filter.

2.1. AFSIW Band-Pass Filter Based on Inductive Posts

The most well-known filtering structure is composed of cavities coupled through iris as depicted in Fig. 2(a) [9]. It exists also filters based on inductive posts that have been reported in the literature for both low-pass and band-pass applications as shown in Fig. 2(b) [3]. This section is focused on the last structure which has been chosen for demonstration based on

AFSIW.

153

(a)

(b) Figure 2: SIW filters using (a) iris [9] and (b) inductive posts [3] as coupling elements.

The most common way to make an inductive post is to insert a conductive via inside the waveguide, through the substrate. It is relatively easy to fabricate such via using the PCB technics owing to the electrodeposition and drilling (or laser cutting) process whereas it is more difficult to realize vias in a RWG.

2.1.1. Design Approach

The equations developed hereafter are valid for SIW only in connection with its TE10 mode.

2.1.1.1. Asymmetrical Inductive Posts

The first case is concerned with asymmetrical inductive posts used for designing band- pass filters. An example of such an SIW filter is illustrated in Fig. 3.

154

Copper Substrate

W li

d xi

Figure 3: Picture of a bandpass filter using 2 cavities and 3 inductive posts, where W is the SIW width, d is the conductive via diameter, xi is the offset and li is the physical length of a cavity.

The equivalent circuit of a single inductive post embedded in a waveguide was studied in [10]. Fig. 4 shows the cross sectional view (a), the top view (b) and the equivalent circuit

(c) of such a topology. The inductive post is made of a circular post of diameter d placed at offset x (0 ≤ x ≤ W/2) from the wall in a waveguide of width W and height h.

W jXb jXb

x h d W Z0 jX Z0 x a T T d (a) (b) (c) Figure 4: Cross sectional view in (a), top view in (b) and equivalent circuit in (c) of an inductive post.

Normalized inductance Xa and capacitance Xb equations are given in [10]. The biggest is the post diameter, the highest is the normalized capacitance and the lowest is the normalized inductance.

The inverter theory is introduced in [11]. It is based on the electrical characteristic equivalence model illustrated in Fig. 5 and equation (1.1) and (1.2) using the K-inverter model of the equivalent reactance and electrical length Φ of the cavity obtained between two inductive posts. X represents the inductive post equivalent reactance.

155

Φ/2 Φ/2

jX K

(a) (b) Figure 5: (a) Equivalent inductive post and (b) K-inverter model.

퐾 훷 = tan | | (1.1) 푍0 2

2 ∗ 푋 훷 = −푡푎푛−1 ( ) (1.2) 푍0

The normalized inverter design values for Chebyshev band-pass response are defined by equations (1.3) to (1.9) [11].

휆푔1 − 휆푔2 ω휆 = (1.3) 휆푔0

휆푔 + 휆푔 휆 = 1 2 (1.4) 푔0 2

퐾0,1 휋 ω휆 = √ (1.5) 푍0 2 푔0푔1ω1

퐾푖,푖+1 휋ω휆 1 | = (1.6) 푍0 푖=1:푛−1 2ω1 √푔푖푔푖+1

퐾푛,푛+1 휋 ω휆 | = √ (1.7) 푍0 푖=1:푛−1 2 푔푛푔푛+1ω1 퐾 푖,푖+1⁄ 푋푖,푖+1 푍 = 0 (1.8) 푍 퐾 2 0 1 − ( 푖,푖+1⁄ ) 푍0

휆푔 1 l = 0 (휋 + (훷 + 훷 )) (1.9) 푖 2휋 2 푖 푖+1 where, gi are the low-pass elements for Chebyshev response tabulated in [11], ωλ is the guided-wavelength fractional bandwidth. λg0, λg1 and λg2 are the guided wavelengths at the

156 center, the upper and lower frequencies respectively. li is the physical length of the cavity between two pairs of inductive posts and n is the number of cavities.

The first step for designing a band-pass filter with inductive posts is to compare the normalized inductance obtained by (1.8) and in [10]. If there is a difference between them, diameter d and offset x of the inductive post must be adjusted in order to equalize the two normalized inductances. Hence, the length of each cavity can be calculated based on (1.9) and the waveguide filter can be implemented on 3D electromagnetic software for adjusting x, d and l parameters by optimization.

2.1.1.2. Symmetrical Inductive Posts

The other case deals with the design of symmetrical inductive posts as illustrated in

Fig. 6. There are two advantages compared to the case of asymmetrical inductive posts as explained in [12]. The first is the suppression of TEi0 even modes with i = 2,4,6… [3]. The other is related to the fact that the first and last pair of inductive posts can be closer to the load impedance without affecting its susceptance as explained in [12]. This is why this solution is used to compare the air-filled SIW technology with its dielectric-filled counterpart in the next section. Copper Substrate

W li

d xi

Figure 6: Picture of a band-pass filter using 2 cavities and 3 pairs of inductive posts, where W is the SIW width, d is the inductive post diameter, xi is the offset and li is the cavity physical length.

157

The design procedure is similar to the case of an asymmetrical inductive post except that instead of using equations derived in [10], the normalized inductance for a pair of inductive posts are given in [12].

2.1.2. Experimental Results

For demonstration purpose and comparison, third-order dielectric-filled and air-filled

SIW filters using symmetrical inductive posts are designed using the above described procedure and optimized using CST Microwave Studio, a 3D electromagnetic simulation software. Rogers RT/Duroid 6002 having a relative permittivity of εr = 2.98 and a thickness of

0.508 mm is used. The air-cut, shown in Fig. 9, is made using a laser micro-machining.

Diameter d of all the vias is 0.508 mm for both filters. For measurement, an E8391A 67 GHz

VNA from Agilent with a SC5226 text fixture from Anritsu is used. Dielectric-filled and air- filled SIW TRL calibration kits, simultaneously fabricated with the filters are used to de- embed all the transitions.

2.1.2.1. SIW Bandpass Filter

A picture of the fabricated SIW band-pass filter is shown in Fig. 7. It has a width of

W = 4.099 mm for Ka-band operation and a length of l = 14 mm after de-embedding. All the dimensions, referring to Fig. 6, are given in Table I.

Figure 7: Picture of the dielectric-filled SIW band-pass filter.

158

i 1 2 3 4 2.5 xi (mm) 0.486 0.765 0.765 / li (mm) 2.974 3.27 2.974 Table I: Dimensions of dielectric-filled SIW band-pass filter.

Simulated and measured results are given in Fig. 8. The measured center frequency is

31.83 GHz. A -3 dB bandwidth of 15.8% is obtained (29.31 GHz to 34.35 GHz). Over this band, the insertion loss is 0.56 ± 0.24 dB. The frequency shift between simulated and measured results is due to the precision of the manufacturing process, and also, the variation of the permittivity of the substrate.

0 -5 -10 -15 -20 -25 -30 -35

|S11| and |S21| (dB) |S11| and -40 Measured -45 Simulated -50 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 8: Measured and simulated results of the dielectric-filled SIW band-pass filter with simulated E-field magnitude at 31 GHz.

2.1.2.2. AFSIW Bandpass Filter

A picture of the fabricated air-filled SIW band-pass filter is shown in Fig. 10. The SIW has a width of W = 7.04 mm for Ka-band operation and a length of l = 21 mm after de- embedding. All the dimensions, referring to Fig. 7, for the filtering part are given in Table II.

Figure 9: Picture of the air-filled SIW band-pass filter.

159

i 1 2 3 4 1.054 xi (mm) 1.054 1.586 1.586 / li (mm) 5.056 7.745 5.056 Table II: Air-filled SIW band-pass filter dimensions.

Simulated and measured results are given in Fig. 11. The measured center frequency is

31.65 GHz. A -3 dB bandwidth of 15.04% is obtained from 29.27 GHz to 34 GHz. Over this band, the insertion loss is only 0.18 ± 0.17 dB.

0 -5 -10 -15 -20 -25 -30 -35

|S11| and |S21| (dB) |S11| and -40 Measured -45 Simulated -50 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 10: Measured and simulated results of the air-filled SIW band-pass filter, with simulated E-field magnitude at 31 GHz.

2.1.2.3. Comparison

For comparison, measured results have been summarized in Table III. The insertion loss is reduced by a factor of 3 using the air-filled SIW compared to its dielectric-filled counterpart. Therefore, the Q-factor of this filter is improved by the use of the air-filled SIW technology for comparable bandwidth at roughly the same center frequency.

Center Q-factor Dimensions Bandwidth IL (dB) frequency (Sim/Meas) W*L (mm) (%) (GHz) @f0 Air-filled 7.04*16.64 0.18 ± 0.17 15.04 31.65 908/706 Dielectric- 4.099*9.73 0.56 ± 0.24 15.8 31.83 269/157 filled Table III: Comparison of the AFSIW and SIW filters.

160

2.2. AFSIW Multilayer Band-Pass Filter

As AFSIW is essentially a multilayer structure, it could be interesting to implement a multilayer filter to reduce the overall footprint. In the literature, different technologies have been explored for multilayer filters. In [13], LTCC has been used to realize a dual-band filtering structure in Ka-band. In [14], a waveguide filter with vertically stacked cavities has been presented in UHF band. In [15], using a multilayer PCB process, the authors have reported a 6th order filters in Q-band. All the previous filters are illustrated in Fig. 11.

(a) (b)

(c)

Figure 11: Several technologies used for multilayer filter: (a) LTCC [13], (b) waveguide [14], and (c) PCB [15].

In this context, it is proposed to implement a multilayer 4th order filter based on AFSIW technology.

2.2.1. Geometry

In this section, the design of a multilayer AFSIW filter is reported. The objective is to demonstrate the potential of this technology and to dispose of a reference design for comparison purposes. After a preliminary dimensioning, the CST Microwave Studio 3D EM simulator has been used for simulation and fine tuning.

The AFSIW multilayer filter geometry is depicted in Fig. 12-15. It is fed by a CBCPW with a first transition to SIW and a second to the AFSIW. The horizontal coupling windows

161 are realized with standard iris leading to magnetic coupling whereas the vertical coupling windows are rectangular holes creating electric coupling. This filter is composed of five layers. The layer 1 and 5 are the top and bottom substrates used to enclose the entire structure.

Layer 4 contains cavities C2 and C3 as described in Fig. 13. Layer 3 includes vertical iris with lv1 = lv2 and wv1 = wv2 as illustrated in Fig. 14. Layer 2 contains cavities C1 and C4, with wiris1 = wiris3 and also the feeding line as depicted in Fig. 15. The thickness of the irises liris1, liris2, and liris3 are set to 1.397 mm.

Layer 5

Layer 4 C2 C3 Layer 3 wv1 w wv2 Layer 2 Port 1 C1 C4 Port 2 Layer 1 Copper Substrate Air

Figure 12: Cross side view of the multilayer AFSIW filter.

liris2

C C w 2 3 a AFSIW wiris2

l C2 lC3

Figure 13: Top view of the layer 4.

l v1 lv2

w wv1 v2

Figure 14: Top view of the layer 3.

162

l liris1 iris3

C1 C4

aAFSIW wiris1 w wiris3

lC1 lC4 Copper Substrate Air

Figure 15: Top view of the layer 2.

2.2.2. Experimental Results

For demonstration purpose, forth-order multilayer AFSIW filter using electric and magnetic coupling is designed and optimized using CST Microwave Studio. Rogers

RT/Duroid 6002 having a relative permittivity of εr = 2.98 and a thickness of 0.508 mm is used for each layer. The air-cut, shown in Fig. 16(a) to (c), is made using a laser micromachining. For measurement, an E8391A 67 GHz VNA from Agilent with a SC5226 text fixture from Anritsu is used. An air-filled SIW TRL calibration kit is used to de-embed all the transitions.

A picture of the fabricated AFSIW multilayer band-pass filter is shown in Fig. 16. It has a width of aAFSIW = 7.04 mm for Ka-band operation, a total height of 2.54 mm and a length of l = 17 mm. All the dimensions, referring to Fig. 12-15, are given in Table IV.

CBCPW to SIW SIW to air-filled SIW l

C1 C4

Reference planes (a) (b)

aAFSIW C2 C3

(c) (d) Figure 16: Picture of the air-filled multilayer SIW band-pass filter: (a) the layer 2 containing the AFSIW feed and C1 and C4 cavities, (b) the layer 3 with the two vertical iris, (c) the layer 4 including C2 and C3 cavities, and (d) top or bottom layer enclosing the structure.

163

Dimensions aAFSIW w wv1 lv1 wiris1 wiris2 liris lC1 lC2 lC3 lC4 Value (mm) 7.04 0.508 3.665 2.647 2.695 1.747 1.397 6.239 6.768 6.763 6.27 Table IV: Dimensions of the multilayer air-filled SIW band-pass filter.

Simulated and measured results are compared in Fig. 17. The measured center frequency is 33.03 GHz. A -3 dB bandwidth of 10.1% is obtained (31.35 GHz to 34.7 GHz).

Over this band, the insertion loss is 1 ± 0.36 dB. The frequency shift, of about to 1%, between simulated and measured results is due to the precision of the manufacturing process and the thermal dilatation of the substrate during the electrodeposition and the assembly process. It has been evaluated by electromagnetic software retro-simulation shown in Fig. 18, that the dimensions w, lv1 and wiris1 have changed from about 25 to 100 µm compared to the original values given in Table IV. As it can be seen, there is a fair agreement between the obtained experimental and post-simulated results in Fig. 18.

0 -5 -10 -15 -20 -25 -30 -35 -40 Measured

|S11||S21| and (dB) -45 Simulated -50 26 28 30 32 34 36 Frequency (GHz) Figure 17: Measured and simulated results of the multilayer air-filled SIW band-pass filter.

0 -5 -10 -15 -20 -25 -30 -35

|S11| and |S21| (dB) |S11| and -40 Measured -45 Simulated -50 28 30 32 34 36 Frequency (GHz) Figure 18: Measured and simulated results of the multilayer air-filled SIW band-pass filter after investigation on the variability process.

164

3. AFSIW ANTENNAS

Numerous SIW antenna structures have been reported based on different technologies including PCB [16], LTCC [17] or CMOS [18]. A few examples are a SIW horn antenna [19], a SIW textile cavity-backed patch antenna [20] and a SIW Yagi-Uda antenna [21], all shown in Fig. 19.

(a) (b)

(c) Figure 19: Example of reported SIW antennas: (a) horn antenna [19], (b) textile cavity-backed patch antenna [20] (with top and bottom view) and (c) Yagi-Uda antenna [21].

The objective of this section is to demonstrate the interest of the AFSIW technique for the design of antennas. For demonstration purposes, an AFSIW slot array antenna and

AFSIW ALTSA are proposed, demonstrated and compared to their SIW counterpart.

3.1. AFSIW Slot Array Antenna

One of the simplest ways to implement an antenna based on waveguide topology is to open the structure by rectangular apertures (or slots) that are periodically spaced. This structure is known as the slot array antenna. Numerous published papers refer to such structures, including for example a H-plane waveguide slot antenna or a SIW slot antenna respectively reported in [22] and [23], and illustrated in Fig. 20. Slot array antenna has been studied for a

165 long time and the gain of this antenna can be improved by increasing the number of slots of the array. As the losses generated by the feeding structure influence the performance of such antenna, the interest of the AFSIW for the design of such antenna compared to SIW is obvious.

In the objective of offering a high performance substrate integrated alternative to existing solutions, a slot array antenna based on AFSIW is proposed. The design of a SIW and an AFSIW slot array antenna operating at Ka-band are introduced. Then, measured and simulated results are reported with a comparison between the two antennas.

(a) (b) Figure 20: Example of slot array antennas: (a) waveguide H-plane [22] and (b) SIW 4X4 array [23].

3.1.1. SIW and AFSIW Slot Array Antenna Design

In this section, an AFSIW slot array antenna is introduced. It is designed based on previous work reported on RWG technology. A conventional SIW slot array antenna is also reported in the objective of having a reference design for comparison purposes. For both designs, after a preliminary dimensioning, the CST Microwave Studio 3D EM simulator has been used for simulation and fine tuning.

3.1.1.1. AFSIW Slot Array Antenna

As the AFSIW technology is closer to a rectangular waveguide, it will be introduced first. The implementation of such antenna based on RWG has been largely studied by R. S.

Elliot in [24]-[27]. Nevertheless, recently, [29] and [30] proposed simplified design equations based on Elliot’s equations and A. F. Stevenson [28] formulation. Hence, using those

166 equations, and by neglecting the influence of the AFSIW dielectric slabs, the design of such antenna based on AFSIW is quite straight forward.

The AFSIW slot array antenna geometry is depicted in Fig. 21. The radiating part consists of four rectangular apertures (or slots), fabricated by cutting and metalizing four rectangular holes on the top substrate. In general case, the last slot is placed at a λg/4 distance from the shunt termination, and the distance, from center to center between two slots, is roughly λg/2.

The AFSIW feeding line is composed of three Rogers 6002 material with thickness t = 0.508 mm. The two metalized surfaces of both the top and bottom substrates are kept. The ground plane is H-shaped to avoid radiation pattern tilt due to the unsymmetrical geometry of the feeding line and the two diagonally positioned holes used to align the three layers during the assembling as it can be seen from Fig. 23(a).

From the design equation provided in [29] and [30], and by taking into account the

AFSIW fabrication process specificities, the AFSIW slot array antenna can be designed.

Then, fine tuning using a 3D EM software allows optimizing the design. The theoretical and optimized dimensions are reported in Table V.

Copper Substrate Air Slots

lslot l1

l 3 w l6 l4 aAFSIW wslot

l 7 Figure 21: AFSIW slot array antenna geometry.

167

Parameters From theory Optimized aASIW 7.04 7.04 W 0.508 0.508 wslot 0.486 0.54 lslot 4.946 4.91 l1 3.572 2.52 l2 2.61 2.73 l3 2.1 1.03 l4 0.848 0.51 l5 31.58 31.58 l6 17 17 l7 12.1 12.1 Table V: AFSIW slot array antenna dimensions (mm) obtained from theory and after optimization.

3.1.1.2. SIW Slot Array Antenna

The referenced slot array antenna geometry is depicted in Fig. 22. It is fed by a SIW.

The design of the four slots is substantially identical to its air-filled counterpart except that they are not cut by laser machining but wet etched. Based on previous theoretical approach, and using equations reported in [23], the geometry of the SIW slot array antenna can be obtained. Its optimized dimensions are reported in Table VI and compared to its AFSIW counterpart dimensions. As for the AFSIW design, the ground plane of this antenna is also H- shaped for comparison purposes.

Copper Substrate Air Slots

lslot l1

l 3 w l6 l4 aSIW wslot

l 7 Figure 22: SIW slot array antenna geometry.

168

Parameters SIW slot array antenna AFSIW slot array antenna aASIW/SIW 4.11 7.04 w 0.508 0.508 wslot 0.7 0.54 lslot 3.2 4.91 l1 0.42 2.52 l2 1.3 2.73 l3 0.16 1.03 l4 0.16 0.51 l5 20.383 31.58 l6 17 17 l7 12.1 12.1 Table VI: Optimized SIW and AFSIW slot array antenna dimensions (mm).

3.1.2. Experimental Results

3.1.2.1. Fabrication

Slots Reference plane CBCPW to SIW

Alignment holes SIW to air-filled SIW (a) (b) Slots CBCPW to SIW y z ϴ x Reference plane (c) Figure 23: Fabricated (a) top and (b) inner substrate of the AFSIW slot array antenna, and (c) SIW slot array antenna.

169

3.1.2.2. Results and Discussion

The return loss (|S11|) measurements have been performed using a Vector Network

Analyzer. Dielectric- and air-filled SIW Through-Reflect-Line (TRL) calibrations are used to de-embed effects of the connectors and the transitions for the dielectric- and air-filled SIW slot array antenna, respectively. The reference planes are shown in Fig. 23. It can be seen from Fig. 24(a) and (b), that the measured -10 dB bandwidths (and center frequency f0) are

1.12 GHz (30.63 GHz) and 2.64 GHz (29.86 GHz) for the dielectric- and air-filled SIW slot array antennas, respectively. The observed discrepancy between the simulated and measured resonant frequency is less than 1%. The radiation patterns and gain have been measured in a compact range anechoic chamber from the Poly-Grames Research Center. The xy-plane and the yz-plane (shown in Fig. 23) co-polarized radiation patterns are reported in Fig. 25(a) and

(b), respectively for the SIW slot array antenna, and Fig. 26(a) and (b), respectively for the

AFSIW slot array antenna. A good agreement between simulated and measured results is observed. Measured gains of 7.98 dBi and 11.5 dBi have been obtained at 30.5 GHz for the dielectric- and air-filled SIW slot antenna, respectively.

0 0 -5 -5 -10 -10 -15 -15 -20 -20 -25 -25

S11 (dB) S11 S11 (dB) S11 -30 -30 -35 -35 -40 -40 Simulated Simulated -45 -45 Measured Measured -50 -50 25 27 29 31 33 35 25 27 29 31 33 35 Frequency (GHz) Frequency (GHz) (a) (b) Figure 24: Measured and simulated (a) SIW and (b) AFSIW slot array antennas |S11|.

170

15 15 10 10 Simulated Measured 5 5 0 0 -5 -5 -10 -10 -15 -15

Gain (dBi) Gain -20 (dBi) Gain -20 -25 -25 -30 -30 Simulated -35 -35 Measured -40 -40 0 30 60 90 120 150 180 -90 -60 -30 0 30 60 90  (°)  (°) (a) (b) Figure 25: SIW slot array antenna measured and simulated co-polarized (a) xy-plane and (b) yz-plane radiation patterns at 30.5 GHz.

15 15 10 10 Simulated Measured 5 5 0 0 -5 -5 -10 -10 -15 -15

Gain (dBi) Gain Gain (dBi) Gain -20 -20 -25 -25 -30 -30 Simulated -35 -35 Measured -40 -40 0 30 60 90 120 150 180 -90 -60 -30 0 30 60 90  (°)  (°) (a) (b) Figure 26: AFSIW slot array antenna measured and simulated co-polarized (a) xy-plane and (b) yz-plane radiation patterns at 30.5 GHz.

The simulated efficiency at 30.5 GHz of both AFSIW and SIW antennas are 97.4% and

95.8%, respectively, taking into account dielectric, conductor and surface roughness losses from the reference plane. At 30.5 GHz, the AFSIW slot array antenna provides a gain and an efficiency improved by 3.5 dB and 1.6%, respectively. Performances of both antennas are compared in Table VII. The AFSIW slot array antenna achieves a bandwidth twice larger than its SIW counterpart. Compared to its dielectric-filled counterpart, the AFSIW ALTSA achieves a narrower xy-plane half power bandwidth (HPBW) resulting in a higher gain.

Thanks to a lower loss feeding, the AFSIW slot array antenna achieves a higher efficiency.

171

AFSIW slot aray SIW slot array antenna Unit antenna 1.12 2.64 GHz Measured -10 dB BW 3.7 8.7 % Measured xy-plane HPBW* 114 93 ° Measured yz-plane HPBW* 36 20 ° Simulated radiation efficiency* 95.8 97.4 % Measured gain* 7.98 11.5 dBi

Table VII: Comparison between SIW and AFSIW slot array antennas (*at f0 = 30.5 GHz).

3.2. AFSIW Antipodal Linearly Tapered Slot Antenna

As SIW offers an excellent compromise between the metallic waveguide and the planar transmission lines, numerous SIW-based antennas have been reported in literature, including

SIW linear [33], parabolic [34] or Fermi [35] shaped Antipodal Tapered Slot Antennas

(ATSA), all illustrated in Fig. 27. ATSA provides a wide bandwidth and also a narrow beam high gain end-fire radiation. However, losses are an order of magnitude higher for the SIW- based antennas compared to the metallic waveguide-based antennas, mainly caused by substrate losses.

In the objective of offering a high performance substrate integrated alternative to existing solutions, a broadband Antipodal Linear Tapered Slot Antennas (ALTSA) based on

AFSIW is reported. The design of a SIW and an AFSIW ALTSA operating at Ka-band are introduced. Then, measured and simulated results are reported with a comparison between the two antennas.

(a) (b)

172

3.2.1. SIW and AFSIW ALTSA Design

In this section, first of all, the design of a conventional SIW ALTSA is reported. The objective is not to reinvent ALTSA but to dispose of a reference design for comparison purposes. Then, based on the first design, an AFSIW ALTSA, occupying the same area, is detailed. For both designs, after a preliminary dimensioning, the CST Microwave Studio 3D

EM simulator has been used for simulation and fine tuning.

3.2.1.1. SIW ALTSA

The referenced (dielectric-filled) SIW ALTSA geometry is depicted in Fig. 28. It is fed by a SIW. The radiating part consists of two metallic arms flared on opposite sides of the substrate. To obtain a broadband matching: the two flaring metallic parts are overlapped [33], three linearly tapered sections, between P1 and P2, P2 and P3, and P3 and P4, are implemented, and the arm-ends are rounded to obtain a broadband matching. As recommended in [36], the length Lant of the arms is chosen to respect 3λ0 ≤ Lant ≤ 8λ0, with λ0 the free-space wavelength at the center frequency of the operating band. A Rogers 6002 substrate with dielectric constant εr = 2.94, and thickness t = 0.508 mm is used. Final dimensions, including the coordinates of points P1, P2, P3 and P4 forming the three linear sections are reported in Table

VIII.

Figure 28: SIW ALTSA geometry.

173

3.2.1.2. AFSIW ALTSA

The AFSIW ALTSA geometry is depicted in Fig. 29. The design of the two flared antenna arms is substantially identical to its dielectric-filled counterpart except that they are not wet etched this time but made by using a laser cutting machining process. As shown in

Fig. 29, the dielectric between the two arms of the antenna is removed and replaced by air.

The AFSIW feeding line is composed of three substrates. However, the radiating part consists of only the top and bottom substrates. For the inner substrate, Rogers 6002 material with thickness t = 0.508 mm is used. The top and bottom substrates are FR4 with dielectric constant εr = 4.4, and thickness t2 = 0.8 mm.

Figure 29: AFSIW ATSA geometry.

The two surfaces metallization of both the top and bottom substrates are kept. To avoid the excitation of parasitic substrate modes [37], metalized vias are placed with a quarter- wavelength spacing on the two flared arms and their edge is metalized as shown in Fig. 29.

The edge of the inner substrate at the end of the AFSIW is also metalized to avoid substrate mode excitation and to improve impedance matching. To optimize simulation time, instead of substrate materials with conductive vias, the top and bottom substrate can be replaced by

Perfect Electrical Conductor (PEC) for preliminary design. The AFSIW ALTSA dimensions are reported in Table VIII.

174

Parameters SIW ALTSA AFSIW ALTSA W 4.11 7.04 Lfeed 16.4 16.4 Want 20.5 20.5 Lant 34.3 34.3 P1 ( 16.4 ; 9.15 ) ( 16.4 ; 8.78 ) P2 ( 19.4 ; 10.45 ) ( 19.4 ; 10.08 ) P3 ( 23.4 ; 11.05 ) ( 23.4 ; 10.68 ) P4 ( 45.88 ; 11.87 ) ( 45.95 ; 11.81 ) P5 ( 47.04 ; 16.34 ) ( 47.08 ; 16.16 ) r1 5 5 r2 2.5 2.5 Table VIII: SIW and AFSIW ALTSA dimensions (mm).

3.2.2. Experimental Results

3.2.2.1. Fabrication

To interconnect both antennas with laboratory instruments, K end-launcher connectors from South-West Microwave are used with a broadband transition from CBCPW to dielectric- filled SIW [31]. For the AFSIW ALTSA another broadband transition from conventional dielectric- to air-filled SIW providing -15 dB return loss and 0.3 ± 0.1 dB transmission loss over the 26.5 to 40 GHz frequency range is used [32]. Instead of conductive vias, a continuous metalized wall is used for the AFSIW feeding line as shown in Fig. 30. This provides a negligible dielectric loss improvement as shown in Chapter III, Section 1.1. The fabricated FR4 top and bottom substrates of the AFSIW ALTSA are shown in Fig. 31. A precision laser micro-machining is used to remove the AFSIW central dielectric portion and make the vias. Metallization is achieved in two steps. A sputtering process is used to deposit a very thin copper layer. It is followed by an electro-deposition method to obtain a thicker metallization. A picture of the two fabricated antennas is shown in Fig. 32.

CBCPW SIW Metalized Via AFSIW

Simulation reference plane Metalized Walls

CBCPW to SIW SIW to AFSIW transistion transistion Figure 30: Fabricated AFSIW ALTSA 6002 inner substrate.

175

Metalized Walls

Metalized Via

Figure 31: Fabricated FR4 top and bottom substrates of the AFSIW ALTSA.

y ϕ (a) Reference planes z x

(b)

Figure 32: Fabricated (a) SIW and (b) AFSIW ALTSA.

3.2.2.2. Results and Discussion

The return loss (|S11|) measurements have been performed using a Vector Network

Analyzer. Dielectric- and air-filled SIW Through-Reflect-Line (TRL) calibrations are used in order to de-embed effects of the connectors and the transitions for the dielectric- and air-filled

SIW ALTSA, respectively. The reference planes are shown in Fig. 32. It can be seen from

Fig. 33(a) and (b), that the measured -10 dB bandwidth are 13.4 GHz and 14 GHz for the dielectric- and air-filled SIW ALTSA, respectively. The radiation patterns and gain have been measured in a compact range anechoic chamber in the Poly-Grames Research Center. The xz- plane and the xy-plane (shown in Fig. 32) horizontal- (collinear with the PCB) and vertical-

(orthogonal with the PCB) polarization radiation patterns are reported in Fig. 34(a) and (b), respectively for the SIW ALTSA, and Fig. 35(a) and (b), respectively for the AFSIW

ALTSA. A good agreement between simulated and measured results is observed. Measured gains of 7.7 dBi and 11.7 dBi have been obtained at 35.5 GHz for the dielectric- and air-filled

SIW ALTSA, respectively. Isolation between the horizontal- and vertical- polarization of

8.5 dB is obtained for the AFSIW ALTSA compared to 5.9 dB for its dielectric-filled counterpart. This poor isolation is due to the fact that the two radiating flares are not on the

176 same plane. In simulation, the E-fileds are tilted by 31° and 15° for the SIW and AFSIW antennas respectively at 35.5 GHz. This isolation can be improved using thinner substrates.

0 0 Simulated Simulated Measured Measured -10 -10

| (dB) | -20 (dB) | -20

|S -30 -30

-40 -40 26 28 30 32 34 36 38 40 26 28 30 32 34 36 38 40 Frequency (GHz) Frequency (GHz) (a) (b) Figure 33: Measured and simulated (a) SIW and (b) AFSIW ALTSA |S11|.

15 15 Simulated Simulated 10 Measured 10 Measured 5 5 0 0 -5 -5

Gain (dBi) Gain -10 (dBi) Gain -10 -15 -15

-20 -20 0 30 60 90 120 150 180 -90 -60 -30 0 30 60 90  (°)  (°) (a) (b) Figure 34: SIW ALTSA measured and simulated horizontal- (solid lines) and vertical- polarization (dashed lines) (a) xz-plane and (b) xy-plane radiation patterns at 35.5 GHz.

15 15 Simulated Simulated 10 10 Measured Measured 5 5 0 0 -5 -5

Gain (dBi) Gain Gain (dBi) Gain -10 -10 -15 -15

-20 -20 0 30 60 90 120 150 180 -90 -60 -30 0 30 60 90  (°)  (°) (a) (b) Figure 35: AFSIW ALTSA measured and simulated horizontal- (solid lines) and vertical- polarization (dashed lines) (a) xz-plane and (b) xy-plane radiation patterns at 35.5 GHz.

The simulated broadband directivity and efficiency of both antennas are shown in

Fig. 36. Dielectric, conductor and surface roughness losses are taken into account. Over the

Ka-band, the AFSIW ALTSA provides a directivity and an efficiency improved by

177

3.5 ± 0.5 dB (3.9 dB at 35.5 GHz) and 2.85 ± 0.36% (2.9% at 35.5 GHz), respectively.

Performances of both identical radiating part dimension antennas are compared in Table IX.

The two antennas achieve a comparable -10 dB bandwidth. Compared to its dielectric-filled counterpart, the AFSIW ALTSA achieves a narrower xy-plane half power bandwidth

(HPBW) resulting in a higher gain. This improvement is due to the dielectric material exemption that allows providing a narrower equivalent aperture. Thanks to a lower loss feeding and a radiating part exempted of dielectric material, the AFSIW ALTSA achieves a higher efficiency. Over the Ka-band, the 3.5 ± 0.5 dB gain improvement offered by the

AFSIW ALTSA compared to its dielectric-filled counterpart avoids the use of a 2x1 array with its associated feeding network loss, which decreases the overall efficiency and increases footprint.

14 100

12  AFSIW ALTSA 10  98 8 

6 96 Directivity (dBi) Directivity  SIW ALTSA

4 efficiencyRadiation (%) Directivity Radiation efficiency 2 94 26 28 30 32 34 36 38 40 Frequency (GHz) Figure 36: SIW and AFSIW ALTSA simulated directivity and radiation efficiency versus frequency.

SIW AFSIW *at 35.5 GHz Unit ALTSA ALTSA Radiating part dimensions 34.3 x 20.5 34.3 x 20.5 mm Measured -10dB BW 26.6 – 40 26 – 40 GHz Simulated polarization tilt angle* 31 15 ° Measured xz-plane HPBW* 56 56 ° Measured xy-plane HPBW* 56 39 ° Simulated radiation efficiency* 96.8 99.4 % Measured gain* 7.7 11.7 dBi Measured isolation between the horizontal- and vertical- 5.9 8.5 dB polarization* Table IX: Comparison between SIW and AFSIW ALTSA.

178

4. CONCLUSION

In this chapter, the demonstration of AFSIW filters and antennas has been introduced.

In the first section, an AFSIW filter using inductive posts and an AFSIW multilayer filter have been explored. All simulated and measured results are in good agreement. It has been shown from full-wave electromagnetic simulations and the measurements that the insertion loss of a third-order air-filled band-pass filter has been significantly reduced by a factor of 3 compared to that of the dielectric-filled SIW counterpart. It has also been demonstrated that it is possible to vertically stack cavities to reduce high order filter footprints.

Then, in a second section, the implementations of an AFSIW slot array antenna and an

ALTSA have been reported. Compared to their SIW counterpart, they offer a significantly narrower xy-plane half-power beamwidth and radiation efficiency improvement, resulting in higher gain.

These low-cost and low-loss structures are of high interest for millimeter-wave applications that require low noise and efficient front-ends such as for emerging millimeter- wave imaging and radar systems. Other types of filter topologies and antennas reported based on RWG or SIW could be redesigned based on AFSIW.

179

REFERENCES

[1] F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp and K. Wu, "Low-loss air-filled Substrate Integrated Waveguide (SIW) band-pass filter with inductive posts", European Microwave Conference (EuMC), 2015 European, Paris, 2015, pp. 761-764. [2] A. Ghiotto, F. Parment, T. P. Vuong, and K. Wu, "Millimeter-wave air-filled SIW antipodal linearly tapered slot antenna", in IEEE Antennas and Wireless Propagation Letters, vol.PP, no.99, pp.1-1. [3] T. B. Cao, P. Lorenz, M. Saglam, W. Kraemer, and R. H. Jansen, "Investigation of symmetry influence in substrate integrated waveguide (SIW) band-pass filters using symmetric inductive posts", in European Microwave Conference (EuMC), 2008, 27– 31 Oct. 2008, pp. 492– 495. [4] J. Garreau and al., "Ultra-compact X-band SIW filter in LTCC technology using high permittivity substrate for a space application", IEEE MTT-S International Microwave Symposium Digest (MTT), 2012, Montreal, QC, Canada, 2012, pp. 1-3. [5] M. Le Coq and al., "Miniaturized C-band SIW filters using high-permittivity ceramic substrates," in IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 5, no. 5, pp. 620-626, May 2015. [6] A. Doghri, A. Ghiotto, T. Djerafi, and K. Wu, "Compact and low cost substrate integrated waveguide cavity and bandpass filter using surface mount shorted stubs ", IEEE MTT-S International Microwave Symposium, Montreal, 17-22 June 2012. [7] S. Adhikari, A. Ghiotto, and K. Wu, "Simultaneous electric and magnetic two- dimensionally tuned parameter-agile SIW devices", IEEE Transactions on Microwave Theory and Techniques, vol.61, no.1, pp. 423-435, Jan. 2013. [8] M. Bozzi, F. Mira and A. Georgiadis, "A novel multilayered SIW filter with two mono-modal cavities and three poles," European Microwave Conference (EuMC), Rome, 2014, pp. 520-523. [9] Tae-Soon Yun, Hee Nam, Ki-Byoung Kim and Jong-Chul Lee, "Iris waveguide bandpass filter using substrate integrated waveguide (SIW) for satellite communication," 2005 Asia-Pacific Microwave Conference Proceedings, 2005, pp. 4. [10] N. Marcuvitz, Waveguide Handbook, Peter Peregninus Ltd., 1985. [11] G. L. Matthaie, L. Young, and E. M. T. Jones, "Microwave filters, impedance matching networks, and coupling structures," Book Mart press , North Bergen ,NJ, USA, Nov. 1985. [12] H. Gruenberg, “Symmetrical placed inductive posts in rectangular wave guide,” Radio and Electrical Engineering Division, National Research of Canada, Ottawa, Canada, December 1951. [13] B. J. Chen, T. M. Shen and R. B. Wu, "Dual-band vertically stacked laminated waveguide filter design in LTCC technology," in IEEE Transactions on Microwave Theory and Techniques, vol. 57, no. 6, pp. 1554-1562, June 2009.

180

[14] Y. Jang, J. Kim, S. Kim and K. Lee, "Design and fabrication of a compact 3- dimensional stacked type dielectric ceramic waveguide bandpass filter," in IEEE Microwave and Wireless Components Letters, vol. 24, no. 10, pp. 665-667, Oct. 2014. [15] S. Cadiou and al., "SIW Q-band filters using advanced multilayer PCB technology," European Microwave Conference (EuMC), Rome, 2014, pp. 1052-1055. [16] N. Ghassemi and K. Wu, "Planar high-gain dielectric-loaded antipodal linearly tapered slot antenna for E- and W-Band gigabyte point-to-point wireless services," in IEEE Transactions on Antennas and Propagation, vol. 61, no. 4, pp. 1747-1755, April 2013. [17] S. B. Yeap, X. Qing, M. Sun and Z. N. Chen, "140-GHz 2×2 SIW horn array on LTCC," IEEE Asia-Pacific Conference on Antennas and Propagation (APCAP), Singapore, 2012, pp. 279-280. [18] Y. Shang, H. Yu, C. Yang, Y. Liang and W. M. Lim, "A 239–281 GHz Sub-THz imager with 100MHz resolution by CMOS direct-conversion receiver with on-chip circular-polarized SIW antenna," Proceedings of the IEEE 2014 Custom Integrated Circuits Conference, San Jose, CA, 2014, pp. 1-4. [19] M. Esquius-Morote, B. Fuchs, J. F. Zürcher and J. R. Mosig, "Novel thin and compact H-plane SIW horn antenna," in IEEE Transactions on Antennas and Propagation, vol. 61, no. 6, pp. 2911-2920, June 2013. [20] R. Moro, M. Bozzi, S. Agneessens and H. Rogier, "Compact cavity-backed antenna on textile in substrate integrated waveguide (SIW) technology," European Microwave Conference (EuMC), Nuremberg, 2013, pp. 1007-1010. [21] X. Zou, C. M. Tong, J. S. Bao and W. J. Pang, "SIW-fed yagi antenna and its application on monopulse antenna," in IEEE Antennas and Wireless Propagation Letters, vol. 13, no. , pp. 1035-1038, 2014. [22] D. J. Kim and J. H. Lee, "Compact resonant slot array antenna using partial H-plane waveguide," in IEEE Antennas and Wireless Propagation Letters, vol. 9, no. , pp. 530- 533, 2010. [23] Li Yan, Wei Hong, Guang Hua, Jixin Chen, Ke Wu and Tie Jun Cui, "Simulation and experiment on SIW slot array antennas," in IEEE Microwave and Wireless Components Letters, vol. 14, no. 9, pp. 446-448, Sept. 2004. [24] R. Elliott and W. O'Loughlin, "The design of slot arrays including internal mutual coupling," Antennas and Propagation Society International Symposium, 1986, 1986, pp. 319-322. [25] R. Elliott, "An improved design procedure for small arrays of shunt slots," in IEEE Transactions on Antennas and Propagation, vol. 31, no. 1, pp. 48-53, Jan 1983. [26] R. Elliott and L. Kurtz, "The design of small slot arrays," in IEEE Transactions on Antennas and Propagation, vol. 26, no. 2, pp. 214-219, Mar 1978. [27] G. Stern and R. Elliott, "Resonant length of longitudinal slots and validity of circuit representation: Theory and experiment," in IEEE Transactions on Antennas and Propagation, vol. 33, no. 11, pp. 1264-1271, Nov 1985. [28] A. F. Stevenson. "Theory of slots in rectangular waveguide," Journal of Applied Physics, vol. 19, no.2, 1948, pp. 24~38.

181

[29] F. Qingyuan, S. Lizhong and J. Ming, "Design and simulation of a waveguide slot antenna," Millimeter Waves (GSMM), 2012 5th Global Symposium on, Harbin, 2012, pp. 131-134. [30] H. M. El Misilmani, M. Al-Husseini, and K. Y. Kabalan, "Design of slotted waveguide antennas with low sidelobes for high power microwave applications," Progress In Electromagnetics Research C, vol. 56, 15-28, 2015. [31] S. Hemour, S. Adhikari, A. Ghiotto, and K. Wu, "Low magnetic biased SIW-based isolator: Effect of the rising temperature on the performance of the isolator", in Proc. IEEE Wireless and Microwave Technology Conference (WAMICON), pp. 1-3, 6 June 2014. [32] F. Parment, A. Ghiotto, T.P. Vuong, J.M. Duchamp, and K. Wu, "Broadband transition from dielectric-filled to air-filled substrate integrated waveguide for low loss and high power handling millimeter-wave substrate integrated circuits", in Proc. IEEE MTT-S Int. Microw. Symp., pp. 1-3, Jun. 1-6, 2014. [33] Z. C. Hao, W. Hong, J. X. Chen, X. P. Chen, and K. Wu, "A novel feeding technique for antipodal linearly tapered slot antenna array", in Proc. IEEE MTT-S Int. Microw. Symp., pp. 1641-1644. Jun. 2005. [34] A. Doghri, A. Ghiotto, T. Djerafi, and K. Wu, "Corrugated SIW L-folded antipodal parabolic tapered slot antenna", in Proc. IEEE Asia-Pacific Microwave Conference, Kaohsiung, pp. 893-895, Dec. 2012. [35] K. Phalak, Z. Briquech and A. Sebak, "Surface integrated waveguide fed Antipodal Fermi-Linear Tapered Slot Antenna at 28 GHz," Antenna Technology and Applied Electromagnetics (ANTEM), 2014 16th International Symposium on, Victoria, BC, 2014, pp. 1-2. [36] J. Zucker, "Surface- and leaky-wave antennas", in Antenna Engineering Handbook, H. Jasik, Ed. New York, NY, USA: Mc- Graw-Hill, 1961. [37] C. Schuster and W. Fichtner, "Parasitic modes on printed circuit boards and their effects on EMC and signal integrity", IEEE Transactions on Electromagnetic Compatibility, vol .43, no. 4, pp. 416-425, Nov. 2001.

182

CHAPTER VI: CONCLUSION AND PERSPECTIVES

183

This work introduced the Air-Filled SIW (AFSIW) technology and its application on the design of passive components.

In the second chapter, a general historical review of electromagnetic propagation transmission line has been presented. From Rectangular Waveguides to Substrate Integrated

Circuits, an overview of performance reached by each introduced technology is presented. A comparison has been done in this chapter. It shows a performance gap between RWG and

SIW in terms of losses and power handling.

In the third chapter, the AFSIW is introduced. A theoretical, computational and experimental analysis of losses and a theoretical analysis of power handling capability have been reported. A comparison of performance has been realized between AFSIW and SIW transmission lines. It has been demonstrated that AFSIW surpasses its dielectric-filled counterpart in terms of transmission loss, Q-factor and average power handling capability. For interconnection with SIW circuits, an AFSIW to SIW transition is theoretically and experimentally studied. Its bandwidth limitation, closely related to fabrication tolerances, is revealed from experiments. For measurement purposes, an AFSIW Thru-Reflect-Line calibration kit has been realized to de-embed transitions and connectors between the VNA and the device under test. Finally, a study on AFSIW bends has been performed to characterize the behavior of such transmission line. It revealed that the rounded bend is the most appropriate.

In the fourth chapter, passive devices used in feeding network such as power dividers, couplers and phase shifters have been explored in Ka-band. First, a power divider based on a

T-junction has been realized and compared to its SIW counterpart with simulated and measured results. Then, two directional couplers have been demonstrated. A Moreno coupler and a two-hole coupler have been theoretically studied and a comparison between simulated and measured results has been performed. At the end of this chapter, an equal-length unequal-

184

dielectric-slab-width phase shifter is theoretically study and measured. For this phase shifter, the results shown that the phase constant increase with the frequency. Whereas a wider structure leads to a decreasing tendency of the phase constant with the frequency. Based on this observation, a broadband compensated phase shifter has been designed, fabricated and measured. All these components have obtained a substantial reduction of losses.

In the fifth chapter, filters and antennas have been explored using AFSIW advantages.

First, a third order bandpass filter with inductive posts has been realized to test the behavior of this technology with resonating elements. For the same characteristics, the AFSIW version surpasses the SIW one in terms of Q-factor and losses. For demonstration purposes, a vertically stacked cavitie filter has been made and measured. This filter consists of five layers and this was the first components developed in AFSIW with such a high number of layers.

The reached performances are good for a first prototype and demonstrated the potential of the

AFSIW technology to achieve compact filters with high quality factor. Then, AFSIW slot antennas consisting of 4 rectangular apertures and AFSIW antipodal linear tapered slot antennas are presented and compared to their SIW versions. It has been demonstrated that the use of the AFSIW improved the gain and increased the radiation efficiency.

The AFSIW topology has largely reduced the losses, improving the performance compared to SIW, and showing a great potential for deployement in applications that are highly interersting for the industry. However, as a new multilayer topology of SIW, AFSIW, recently developed in this thesis, has several points to be improved among which:

 The assembling process has to be reconsidered to be more efficient and mass-

producible.

 A complete study on the material used to assemble the different layers has to be

rigorously carried out. Its influence on the perfromance has to be characterized.

185

For example, the top and bottom can be replaced by Alumina substrate that has a

higher thermal dissipation than Rogers 6002 substrate.

 A thermal study on a transmission line and cavity has to be performed to

quantify the mechanical dilatation on performances.

 Performing power handling tests on AFSIW transmission lines and cavities as

well as PIM or multipaction.

Some other passive components as well as active componants can be redesigned based on AFSIW. As an example, adaptive loads, Ortho-Mode Transducer (OMT) and tunable or reconfigurable passive components based on this promising technology are to be explored.

Concerning active device, local oscillator or implementation of a power amplifier has to be investigated. Finally, it would be of interest to demonstrate a complte system such as a transmitter or receiver, implementing some of the components introduced in this thesis, and showing the advantages of the AFSIW in term of integration, compactness, cost and performances.

186

ANNEX 1: COMPLETE ELLIPTIC INTEGRAL OF THE FIRST KIND

A complete elliptical integral of the first kind depends on angular eccentricity, denoted

α, which is one of parameters arising in the study of an ellipse as illustrated in Fig. 1. It is defined in equation (A1.1).

a b α

0 F

Figure 1: angular eccentricity of an ellipse

푏 훼 = cos−1 ( ) (A1.1) 푎 The complete elliptical integral of the first kind is given in equation (A1.2) and is described in function of the coefficient k = sin α.

휋 ⁄2 푑휃 퐾(푘) = ∫ (A1.2) 2 2 0 √1 − 푘 푠푖푛 휃

187

188

ANNEX 2: CALCULATION TO OBTAIN EQUATION (1.2) OF THE CHAPTER III

2 st 휕 2 1 Step : ( + 푘푑 ) ℎ푧 = 0, for 0 ≤ 푥 ≤ 푑, 휕푥2

2 휕 2 ( + 푘푡 ) ℎ푧 = 0, for 푑 ≤ 푥 ≤ 푡, 휕푥2

2 휕 2 ( + 푘푎 ) ℎ푧 = 0, for 푡 ≤ 푥 ≤ 푎, 휕푥2

nd 2 2 2 Step : 훽 = √휀푟 ∗ 푘0 − 푘푑 , for 0 ≤ 푥 ≤ 푑,

2 2 훽 = √푘0 − 푘푡 , for 푑 ≤ 푥 ≤ 푡,

2 2 훽 = √휀푟 ∗ 푘0 − 푘푎 , for 푡 ≤ 푥 ≤ 푎,

퐴 cos(푘푑푥) + 퐵 sin(푘푑푥), for 0 ≤ 푥 ≤ 푑, rd 3 Step : ℎ푧 = { 퐶 cos(푘푡(푡 − 푥)) + 퐷 sin(푘푡(푡 − 푥)), for 푑 ≤ 푥 ≤ 푡, 퐸 cos(푘푎(푎 − 푥)) + 퐹 sin(푘푎(푎 − 푥)), for 푐 ≤ 푥 ≤ 푎,

4th Step :

푗휔휇0 (−퐴 sin(푘푑푥) + 퐵 cos(푘푑푥)), 푘푑 for 0 ≤ 푥 ≤ 푑, 푗휔휇 푗휔휇 퐸 = 0 ∗ 휕퐻푧 → 푒 = 0 (퐶 sin(푘 (푡 − 푥)) − 퐷 cos(푘 (푡 − 푥))), for 푑 ≤ 푥 ≤ 푡, 푦 2 휕푥 푦 푘 푡 푡 푘 푡 for 푡 ≤ 푥 ≤ 푎, 푗휔휇0 (퐸 sin(푘푎(푎 − 푥)) − 퐹 cos(푘푎(푎 − 푥))), { 푘푎

th 5 Step : Boundary condition : 퐸푦 = 0 푎푡 푥 = 0 푎푛푑 푥 = 푎 → 퐹 = 0 푎푛푑 퐵 = 0

Continuity of tangential fields :(퐸푦, 퐻푧) 푎푡 푥 = 푑 푎푛푑 푥 = 푡

−퐴 sin(푘 푑) 퐶 sin(푘 (푡−푑))−퐷 cos(푘 (푡−푑)) 푑 = 푡 푡 6th Step : 푎푡 푥 = 푑, { 푘푑 푘푡 퐴 cos(푘푑푑) = 퐶 cos(푘푡(푡 − 푑)) + 퐷 sin(푘푡(푡 − 푑))

189

−퐷 퐸 sin(푘 (푎−푡)) = 푎 푎푡 푥 = 푡, { 푘푡 푘푎 퐶 = 퐸 cos(푘푎(푎 − 푡))

7th Step : Equation of propagation with 훽 = 0

tan(√휀푟 ∗ 푘0 ∗ (푎 − 푡)) ∗ tan(푘0 ∗ (푡 − 푑)) − tan(√휀푟 ∗ 푘0 ∗ 푑) ∗ (1 − ) √휀푟

= √휀푟 tan(푘0 ∗ (푡 − 푑)) + tan (√휀푟 ∗ 푘0 ∗ (푎 − 푡))

190

LIST OF AWARDS, DISTINCTIONS AND PUBLICATIONS

4 Awards and Distinctions: 1. 2016 IEEE MTT-French Chapter Best Student Paper Award.

2. 2016 International Summer School on Microwave Systems for the IoT EuMA PhD grant.

3. 2014 IEEE MTT-French Chapter Best Student Paper Award.

4. 2014 IEEE International Microwave Symposium PhD Student Sponsorship Initiative.

7 Peer Reviewed Journals (4 Published and 3 Submitted): 1. F. Parment, A. Ghiotto, T.-P. Vuong, J.-M. Duchamp, and K. Wu, "Air-filled substrate integrated waveguide for low-loss and high power-handling millimeter-wave substrate integrated circuits," IEEE Trans. on Microw. Theory and Tech., vol. 63, pp. 1228-1238, April 2015.

2. F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp, and K. Wu, "Double dielectric slab-loaded air-filled SIW phase shifters for high-performance millimeter-wave integration," IEEE Trans. on Microw. Theory and Tech., vol. 64, no. 9, pp. 2833-2842, Sept. 2016.

3. F. Parment, A. Ghiotto, T. P. Vuong, J. M. Duchamp, and K. Wu, "Air-to- dielectric-filled two-hole substrate integrated waveguide directional coupler," in IEEE Microwave and Wireless Components Letters, (Submitted in August 2016).

4. F. Parment, A. Ghiotto, T.-P. Vuong, J.-M. Duchamp, and K. Wu, "Millimeter- wave air-filled substrate integrated waveguide slot array antenna," in Electronics Letters, (Submitted in November 2016).

5. F. Parment, A. Ghiotto, T.-P. Vuong, J.-M. Duchamp, and K. Wu, "Ka-band compact and high performance bandpass filter based on multilayer air-filled SIW," in Electronics Letters, (Submitted in November 2016).

191

6. A. Ghiotto, F. Parment, T.-P.Vuong, and K. Wu, "Air-filled SIW antipodal linear tapered slot antenna for high-performance millimeter-wave substrate integrated circuits and systems," IEEE Antennas and Wireless Propagation Letters (Published in August 2016).

7. I. S. S. Lima, F. Parment, A. Ghiotto, T. P. Vuong and K. Wu, "Broadband dielectric-to-half-mode air-filled substrate integrated waveguide transition," in IEEE Microwave and Wireless Components Letters, vol. 26, no. 6, pp. 383-385, June 2016. 6 International Conference Papers (Including 2 Invited Papers) (*Speaker): 1. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, and K. Wu, "Broadband transition from dielectric-filled to air-filled substrate integrated waveguide for low loss and high power handling millimeter-wave substrate integrated circuits," IEEE MTT-S International Microwave Symposium Digest, June 2014. (IEEE MTT-French Chapter Best Student Paper Award)

2. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, and K. Wu, "Air-filled SIW transmission line and phase shifter for high-performance and low-cost U- band integrated circuits and systems," IEEE Global Symposium on Millimeter Waves (GSMM), May 2015.

3. A. Ghiotto*, A. Doghri, F. Parment, T. Djerafi, T.-P.Vuong, and K. Wu, "Three-dimensional SIW and high-performance air-filled SIW for millimeter- wave substrate integrated circuits and systems," IEEE Global Symposium on Millimeter Waves (GSMM), May 2015. (Invited Paper)

4. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, and K. Wu, "Low-loss air-filled substrate integrated waveguide (SIW) band-pass filter with inductive posts," IEEE European Microwave Conference (EuMC), September 2015.

5. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, and K. Wu, "Broadband directional moreno coupler for high-performance air-filled SIW-based substrate integrated systems," IEEE MTT-S International Microwave Symposium Digest, May 2016. (IEEE MTT-French Chapter Best Student Paper Award)

6. A. Ghiotto*, F. Parment, T.-P.Vuong, and K. Wu, "Multilayer-substrate integration technique of air-filled waveguide circuits," IEEE MTT-S Numerical Electromagnetic and Multiphysics Modeling and Optimization for RF, Microwave, and Terahertz Applications, July 2016. (Invited Paper)

192

6 National Conference Papers (*Speaker): 1. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, et K. Wu, "Antenne ATSA en bande Ka basée sur la technologie guide d’onde intégré au substrat et réalisée en circuit imprimé multicouches," Assemblé Générale du GDR-Ondes, Octobre 2013.

2. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, et K. Wu, "Antenne ATSA en bande Ka basée sur la technologie GIS et réalisée en PCB multicouches," Journées Nationales Microondes (JNM), Juin 2015.

3. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, et K. Wu, "Nouvelle technologie GISA pour la conception de dispositifs dans le domaine des applications millimétriques à hautes performances," Journées Nationales Microondes (JNM), Juin 2015.

4. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, et K. Wu, "Nouvelle technologie GISA à hautes performances pour la conception de déphaseur dans le domaine millimétrique," Assemblée Générale du GDR-Ondes, Octobre 2015.

5. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, et K. Wu, "Nouvelle technologie GISA pour la conception de filtre passe-bande à pont inductif à hautes performances dans le domaine millimétrique," Assemblée Générale du GDR-Ondes, Octobre 2015.

6. I. S. S. Lima*, F. Parment, A. Ghiotto, T.-P. Vuong, et K. Wu, "Transition large bande entre un SIW et un demi-SIW (HMSIW) utilisant de l’air comme diélectrique," Assemblée Générale du GDR-Ondes, Octobre 2015.

5 Workshops and Communications (*Speaker): 1. F. Parment*, A. Ghiotto, T.-P. Vuong, J.-M. Duchamp, et K. Wu, "Broadband transition from dielectric-filled to air-filled substrate integrated waveguide for low loss and high power handling millimeter-wave substrate integrated circuits," Assemblée Générale du Chapitre IEEE MTT France, Bordeaux, 17 Octobre 2014.

2. F. Parment*, A. Ghiotto, T.-P.Vuong, J.-M. Duchamp, et K. Wu, "Nouvelle technologie SIW-air pour les applications millimétriques," 2015 CST Workshop Series, Grenoble, June 2015.

193

3. F. Parment*, A. Ghiotto, T.-P. Vuong, J.-M. Duchamp, et K. Wu, "Nouvelle technologie SIW pour les applications spatiales," 2016 CST Workshop Series, Grenoble, France, June 2016.

4. A. Ghiotto*, F. Parment, T.-P. Vuong, J.-M. Duchamp, et K. Wu, "Emerging air-filled SIW technology for high performance integrated ciruits and systems at millimeter-wave and beyond," EuMA International summer school Microwave Systems for the IoT, European Courses on Microwvaes (EuCoM), Pavia, Italy, 7- 9 Sept. 2016.

5. A. Ghiotto*, F. Parment, T.P. Vuong, K. Wu, “Emerging Air-Filled SIW technology For High Performance and Low-Cost Integrated Circuits and Systems at Millimeter-Wave and Beyond,” IEEE MTT-S International Microwave Symposium, Workshop, Honolulu, Hawai’i, Accepted, 4-9 June 2017.

194

ABSTRACT

The substrate integrated waveguide (SIW) technology, introduced in the early 2000s, has presently trigged a huge interest from academia to industry with the focus on the design and development of low-loss, compact, integrated, self-packaged and low-cost microwave and millimeter-wave circuits, antennas and systems. However, the classical metallic waveguide technology, which offers better performances such as lower insertion loss and higher power handling, has still been used in the design of microwave and millimeter-wave systems, despite its higher cost and bulky structure. To offer a highly integrated, further loss-reduced, low-cost alternative to the conventional waveguide and also to allow a wide-spread use of the millimeter-wave spectrum, this thesis research introduces a new SIW structure called Air- Filled SIW (AFSIW). This new structure has been theoretically and experimentally studied in details with a substantial amount of results. At millimeter wave frequencies, compared to the SIW topologies, the proposed AFSIW scheme exhibits a substantially lower insertion loss (three times at Ka-band, for example) and a much higher average power handling capability (four times, at Ka-band for example). Numerous AFSIW passive components have been investigated designed and demonstrated, which take advantages of the well-established multilayer printed circuit board (PCB) fabrication process. Couplers, phase shifters, power dividers, antennas and filters have been modeled, designed, prototyped and measured based on the introduced technology. Their performances have theoretically and experimentally been compared with their SIW counterparts to demonstrate and validate the benefits of the proposed technology. RESUME

La technologie SIW, introduite dans les années 2000, suscite aujourd’hui un très vif intérêt pour la conception de circuits micro-ondes et millimétriques compacts, intégrés, faible coût et blindés par nature. Cependant, les guides d’onde métalliques, qui offrent de bien meilleures performances en termes de pertes d’insertion et de tenue en puissance, malgré un coût bien plus important, sont encore incontournables pour de nombreuses applications millimétriques. Afin de proposer une alternative intégrée et faible coût au guide d’onde rectangulaire, et de permettre une large exploitation du spectre millimétrique, cette thèse propose une nouvelle structure SIW appelée SIW creux. Cette nouvelle structure a été étudiée théoriquement et expérimentalement. Aux fréquences millimétriques, comparativement au SIW, le SIW creux offre des pertes d’insertion trois fois plus faible ainsi qu’une tenue en puissance moyenne quatre fois plus importante. De nombreux dispositifs passifs SIW creux ont été conçus en prenant avantage du procédé de circuit imprimé multicouche mis en œuvre. Des coupleurs, déphaseurs, diviseurs de puissance, antennes et filtres ont été réalisés basés sur la technologie introduite. Leurs performances sont théoriquement et expérimentalement comparées avec leur contrepartie SIW afin de démontrer les avantages de la nouvelle technologie proposée.

195