MEMS Journal -- The Largest MEMS Publication in the World: An overview of RF MEMS technologies and applications

by Jacopo Iannacci, Ph.D.
Researcher, MEMS Research Unit of FBK

RF MEMS devices are microsystems manufactured in a suitable technology platform, typically characterized by having movable micro-parts capable of reconfiguring the RF characteristics of a device. Some RF micro-components are categorized as RF MEMS devices despite not having any movable or deformable parts. This is because such devices are manufactured in a surface or bulk micromachining technology, or in a combination of both, and such techniques are considered as the de-facto standard for the realization of RF MEMS devices and components. In this article, we review all of the main categories of RF MEMS lumped components and networks, and also specify their purpose, functionality, typical applications, as well as pros and cons compared to the traditional implementation of the same devices.

Besides these definition issues, the variety of RF components manufactured with MEMS technologies is quite wide, and ranges from lumped components, such as ohmic and capacitive switches, fixed and variable capacitors (also referred to as varactors), variable inductors and resonators, to more complex networks, based on a combination of these basic components. Examples of the latter class of networks are tunable RF filters, reconfigurable impedance matching tuners, switching matrices, RF power dividers and attenuators, reconfigurable phase shifters, and so on. Below is a list of the main types of RF MEMS devices along with their corresponding descriptions.

Variable capacitor (varactor)
Functionality: implementing a capacitance to be tuned by means of a controlling signal (bias)
Typical applications: lumped components within RF circuits and transceiver platforms
Pros: good linearity, large tuning range, high Q-factor (low loss), virtually no power consumption (for controlling the device)
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS)

Capacitive switch
Functionality: implementing a switch for RF signals based on a two-state shunt (to RF ground) capacitance. The switch is closed when the capacitance is low (MEMS in the “off” position), while it is open when the capacitance is high (MEMS in the “on” state)
Typical applications: Selection and de-selection of RF signals within circuits and transceiver platform
Pros: low loss, good isolation, virtually no power consumption (for controlling the device)
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating)

Ohmic switch
Functionality: micro-relay for RF (DC and AC) signals, based on metal-to-metal contacts. An RF MEMS ohmic switch can be series or shunt type. A series switch is open when the MEMS is “off” (rest position) and is closed when the MEMS is “on” (actuated position). The behavior of a shunt switch is the opposite of the one in the series configuration
Typical applications: selection and de-selection of RF (DC and AC) signals within circuits and transceiver platform; reconfiguration of the architecture and/or functionality of an RF apparatus (e.g. antenna selection)
Pros: low loss, good isolation, virtually no power consumption (for controlling the device)
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Variable inductor
Functionality: implementing an inductance to be tuned by means of a controlling signal (bias)
Typical applications: lumped components within RF circuits and transceiver platforms
Pros: good linearity, large tuning range, high Q factor (low loss), virtually no power consumption (for controlling the device)
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS)

SPDT (Single Pole Double Throw)
Functionality: realizing a T-type splitting of an RF signal in input, towards two possible output channels; the number of possible RF output lines can be larger than two; in the latter case the device is referred to as SPMT (single pole multiple throw)
Typical applications: redistribution of signals within an RF circuit and transceiver platform
Pros: low loss, good isolation, virtually no power consumption (for controlling the device), small dimensions and reduced weight
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Switching matrix
Functionality: redistributing the RF signals between multiple input lines and multiple output lines
Typical applications: redistribution of signals within an RF circuit and transceiver platform (e.g. in satellites for telecommunications)
Pros: low loss, good isolation, high complexity, virtually no power consumption (for controlling the device), small dimensions and reduced weight
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Reconfigurable RF filter and resonator
Functionality: implementing a given (and variable) filtering function of an RF signal in a certain frequency band (e.g. band-pass, low-pass filter, etc.)
Typical applications: filtering and selection of RF signal within circuits and transceiver platform (e.g. LC tanks for voltage controlled oscillators, or VCOs; MEMS resonators in oscillators)
Pros: Low loss, high Q factor, very wide reconfigurability, virtually no power consumption (for controlling the device), small dimensions and reduced weight
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Reconfigurable impedance tuner
Functionality: transforming the characteristic impedance at the input of the tuner into another one at its output
Typical applications: impedance matching between sub-blocks of an RF circuit having a different characteristic impedance (i.e. mismatch); for example, impedance tuning between the power amplifier (PA) and the transmitting antenna in a GSM system
Pros: low loss, high Q factor, very wide reconfigurability, virtually no power consumption (for controlling the device), small dimensions and reduced weight
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Phase shifter
Functionality: modifying the phase (i.e. the delay) of an RF signal between the input and the output of the network
Typical applications: Antennae based on the electronic steering; for example, in phased arrays for radar, surveillance and navigation systems
Pros: Low loss, high Q factor, very wide reconfigurability, virtually no power consumption (for controlling the device), small dimensions and reduced weight
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Power divider / attenuator
Functionality: splitting the power of an RF signal between different channels; attenuating the power of an RF signal
Typical applications: adapting the power level at the input stage of an RF amplifier, or before a block requiring a lower power level to work properly
Pros: very wide reconfigurability, virtually no power consumption (for controlling the device), small dimensions and reduced weight
Cons: fragile (needs package), reliability (medium and long term), performance drift (e.g. charge accumulation, fatigue, and so on), large controlling voltages required (not compatible with CMOS), power handling (e.g. self-heating, micro-welding)

Overall, the applications of RF MEMS devices and networks are various and follow two main progressive trends. The first one features the replacement of lumped components, currently realized in standard semiconductor technology, with the corresponding implementations in RF MEMS technology, within RF transceiver (telecommunications) platforms and systems.

Given the high-performance and good characteristics of lumped components in RF MEMS technology (like high Q-factor, low-loss, high isolation, wide tenability, etc.), their employment and integration within an RF system would enhance its performance and, rather interestingly, also its reconfigurability.

For example, RF MEMS resonators can be employed in the LO (local oscillator) of RF transceivers, improving the purity of the signal carrier, while RF MEMS variable capacitors and inductors can realize widely-tunable LC-tanks that, integrated within a VCO (voltage controlled oscillator) architecture, can significantly extend the tuning range achieved by the functional block. Again, low-loss and high-isolation RF MEMS switches can be employed to reconfigure the architecture of an RF transceiver, therefore reducing the hardware redundancy while improving the characteristics of an antenna, and making it operable for different frequency ranges, rather than including one antenna per each band.

The second main trend aims to exploit RF MEMS technology for synthesizing entire functional RF system blocks and sub-blocks based on microsystems, rather than replacing lumped components within already existing architectures. Through these efforts, entire switching matrices for the re-routing of RF channels, in terrestrial and satellite telecommunication systems, can be implemented in RF MEMS technology, considerably reducing their area occupation and weight; the latter feature is particularly critical in space applications.

Also, phase-shifters and true time delay (TTD) lines have been already implemented in RF MEMS technology, and are demonstrated in research literature. If integrated within RF systems, such networks enable the electronic steering of an antenna lobe, i.e. without any physical moving parts of the system (apart from the MEMS reconfigurable components, of course). The advantages arising from such a solution are outstanding. For instance, short and mid-range radar systems can be realized with dramatically reduced mechanical complexity and area occupation, thus making their integration and market adoption much easier.

Applications for such short and mid-range radar systems are quite varied and their impact on improving the quality of our daily life is significant. For example, object and obstacle detection (indoor and outdoor) systems could assist people with disabilities, or improve the safety of any type of vehicle with anti-collision capabilities, and so on. Of course, these are only a few examples of systems that might benefit from employing components and entire sub-blocks realized in RF MEMS technology.

In general, RF MEMS technology has not reached yet a massive market adoption in contrast to other types of MEMS based sensors and actuators such accelerometers, gyroscopes, micro-mirrors arrays. At the current time, lumped RF MEMS components are more mature than complex architectures. Among them, micro-switches are definitely the most stable RF MEMS devices nowadays, and several companies include them in their lists of available products. For example, companies such as Radant MEMS, Raytheon, Omron and MEMtronics provide micro-relays and varactors based on RF MEMS technologies.

As outlined above, implementations of complex networks based on RF MEMS technology are still rather difficult to find on the market for multiple reasons.

First, reliability and performance stability are significant issues when developing a new solution in RF MEMS technology. Indeed, the presence of movable and deformable micro-membranes makes each RF MEMS device or network prone to failure and degradation mechanisms; this is usually not a problem when dealing with their counterparts in standard semiconductor technology. The most relevant failure and degradation mechanisms affecting RF MEMS devices are contamination, wearing, fracture, delamination, micro welding, charge entrapment, electromigration, and so on. Most of these problems can be mitigated by acting at three different levels, namely, at design level, technology level, and operation level.

Secondly, despite the fact that the manufacturing costs of RF MEMS devices and networks are, in general, lower if compared to advanced CMOS processes, bringing them from the prototype level to a market product demands high initial investments.

Also, developing the proper packaging (i.e. encapsulation and protection) of RF MEMS devices, as well as the development of the active (CMOS) control circuitry and the interface to the rest of the system (i.e. integration) are also critical challenges to be resolved before making an RF MEMS device suitable for a real applications.

In conclusion, RF MEMS components such as switches and variable capacitors as well as basic switching units, like for instance SPDTs (single pole double throws) have seen the most success in the marketplace. The more complex, integrated networks based on RF MEMS technologies are promising, but still mostly in development mode. However, these new classes of RF MEMS devices and integrated systems are likely to see a new wave of commercialization in the near future.

*********************************************

Dr. Jacopo Iannacci received the M.Sc. degree in electronic engineering from the University of Bologna, Bologna, Italy, in 2003, and the Ph.D. degree in information technology from the Advanced Research Center for Electronics Systems “E. De Castro” (ARCES) Research Center, University of Bologna, in 2007. In 2005 and 2006, he was with the HiTeC-DIMES Technology Center, Technical University of Delft in Delft, The Netherlands, developing packaging solutions for RF MEMS components. Since October 2007, he has been with the Fondazione Bruno Kessler (FBK) in Trento, Italy, as a Researcher on MEMS technology. Dr. Iannacci’s scientific interest is focused on compact modeling, design, optimization, integration, packaging, and reliability of MEMS and RF MEMS devices and networks for sensors and telecommunication systems. Dr. Iannacci can be reached at [email protected].