Microwave Laboratory

Meisei University

 
 

    Over the past 60 years, the dominant technology for non-reciprocal microwave components has been based on ferrite, which are passive high-resistivity ferrimagnetic materials, such as YIG (yttrium iron garnet) or  compounds of iron oxides, and other elements such as aluminium, cobalt, manganese and nickel. However, ferrite components suffer from major drawbacks, inherent to ferrite materials and the permanent magnet required to bias them: they are bulky, heavy, incompatible with integrated circuit technologies, expensive, sensitive to temperature detuning at lower microwave frequencies, and sometimes ferromagnetic resonance-based devices) inapplicable at frequencies above the X-band. Recently reported self-biased ferromagnetic nanowire materials are immune to most of these drawbacks, but still exhibit prohibitively high loss and low power handling for commercial applications at this point.


    An active approach of non-reciprocal microwave components, where non-reciprocal effects are induced by the interconnection of transistors with transmission line, couplers and lumped elements in various architectures, has been explored to avoid the drawbacks of ferrite technology . Unfortunately, this approach has been restricted to marginal applications, due to several issues:
1) Architectures involving cascaded transistors, although providing high gain, suffer of stability issues, requiring stabilization resistors which seriously affect the insertion loss;
2) the simplest -- and most symmetric -- circuital architectures suffer from poor performance, such as low isolation or high insertion loss, due to inherent limited transmission efficiency, backward leakage and all-port matching difficulty;
3) performance improvement over these architectures, incorporating distributed couplers, leads to major increase in the size of the components, which partly defeats the initial purpose of size reduction, and also often breaks the symmetry of the circuit (sometimes allowing only a ¥emph{quasi}-circulator.
4) non-reciprocal circuits are not materials, and are therefore inapplicable to the many components requiring distributed material effects, such as Faraday rotation, field displacement or birefringence.


    Recently, we invented a magnetless non-reciprocal metamaterial (MNM). The classical-picture current loops formed by spin electron precession in ferromagnetic materials is mimicked by electric current loops along conducting ring-pair particles loaded with a unilateral semiconductor element, typically a field-effect transistor (FET). The rotating magnetic dipole moments are produced in the mid-plane between the two rings of each ring-pair particle as a result of the traveling-wave propagation regime due to the FETs. In the present work, the ring particles are arranged in a substrate plane (2D structure) and each particle contains only one ring, a ground plane below the ring playing the role of an electric mirror from which the operation of the resulting unbalanced MNM is identical to that of a groundless balanced two-ring particle MNM. The molecular-level analogy of MNMs to ferrites distinguishes MNMs from other non-reciprocal metamaterials, which are designed to provide some particular characteristics of ferrite-based devices, such as isolation, without exhibiting the fundamental properties of ferrites, such as Faraday rotation and field displacement.


    The MNM approach to non-reciprocity might be classified between the ferrite approach and the active-circuit approach.  An MNM is a kind of “artificial ferrite,'' since it mimics the atomic activity of a ferrite material, and hence exhibits the same fundamental properties, while it includes active elements (the FETs) like active-circuit non-reciprocal components. In contrast to active non-reciprocal circuits, where non-reciprocity directly stems from the unilaterality of FETs connected along transmission lines and other distributed or lumped components, an MNM is really a material or a “magneto-dielectric”, characterizable by a macroscopic permeability tensor. Practically, it overcomes the aforementioned drawbacks of ferrites, requiring neither a ferrite nor a magnet. It is therefore lightweight, fully compatible with monolithic microwave integrated circuit (MMIC) technology, low-cost, insensitive to temperature detuning and systematically applicable up to the upper millimeter-wave frequency range.  A diversity of proof-of-concept MNM structures and components have already been reported, including a Faraday polarizing surface, an electrically-gyrotropic MNM, an isolator, a field-displacement isolator , a perfect electromagnetic (PEMC) meta-surface, and a magnetless implementation of the non-reciprocal ferrite leaky-wave antenna.

 

WElcome to the world of Artificial electromagnetic structures enabling non-reciprocity

An MNM is a kind of “artificial ferrite,'' since it mimics the atomic activity of a ferrite material.


Practically, it overcomes the aforementioned drawbacks of ferrites, requiring neither a ferrite nor a magnet. It is therefore lightweight, fully compatible with monolithic microwave integrated circuit (MMIC) technology, low-cost, insensitive to temperature detuning and systematically applicable up to the upper millimeter-wave frequency range. 

Copyright ©2014,
Toshiro Kodera
Microwave Laboratory, Meisei University. All Rights Reserved.