Miniaturization has allowed a growing number of electronic functions to be packed into pocket-sized designs and, in the case of microelectromechanical-systems (MEMS) devices, even mechanical functions can be included in those designs. MEMS devices are perhaps best known as miniature switches, although they are also gaining popularity when used as frequency sources, such as oscillators. They can be packaged in similarly small enclosures as purely electronic devices, but how do MEMS devices differ from similar-sized electronic circuits, such as monolithic-microwave integrated circuits (MMICs)?
The most essential difference between MEMS and MMICs is that MEMS are designed to function with moving parts, as electromechanical devices, while MMICs are meant to operate fully electronically. Because of the mechanical capabilities, MEMS devices are capable of producing high-frequency resonances (serving as resonators and oscillators), as well as the reverse function of detecting vibration (as in audio microphones).
At RF and microwave frequencies, companies such as SiTime have built comprehensive portfolios of oscillators and timing products based on MEMS technology which challenge or exceed the performance of traditional quartz crystal temperature-compensated crystal oscillators (TCXOs).
With the capability to fabricate microminiature moving parts, RF MEMS technology is a natural candidate for certain component functions, such as switches. Companies such as Analog Devices, MEMtronics Corp., Radant MEMS, and WiSpry have established strong track records in reliable, high-performance RF/microwave switches based on MEMS technology, with open and closed switching states often determined by the position of a mechanical part, such as a cantiliever.
For example, MEMS switches from Analog Devices (Fig. 1) are fabricated on silicon (Si) substrates, building on the company’s vast experience in designing and constructing silicon integrated circuits. The microminiature switches are based on the use of electrostatically actuated cantilever beam switching elements. When applied, the electrostatic force is sufficient to overcome the spring force of the cantilever beam and form a metal-to-metal contact to close the conductive path of the switching elements.
MEMS switches have been developed on Si wafers using a variety of different approaches, including Si CMOS semiconductor processes to produce MEMS switches capable of operation at millimeter-wave frequencies for possible use in 5G wireless communications networks. As with transistors and other semiconductor devices, MEMS components have been fabricated on substrate materials other than Si, including gallium arsenide (GaAs) and gallium nitride (GaN) substrates.
In contrast, MMICs are essentially miniaturized RF and microwave circuits, designed and fabricated using many different semiconductor processes and with many different levels of integration. MMICs may have thousands of monolithic active and passive circuit elements and components, with a substrate measuring just a few square millimeters.
Analog circuit functions may be active (such as amplification), or passive (e.g., attenuation), although functional control is performed by means of applied energy. MMIC technology allows highly integrated portions of systems, such as receiver front ends, to be mass produced using proven high-frequency semiconductor processes—for example, those based on silicon germanium (SiGe), Si, GaAs, and GaN substrate materials.
The MMIC circuits can be impedance matched to 50 Ω at input and output ports for ease of cascading in multiple MMIC designs. Although RF/microwave MMICs are available from many suppliers in die form, they are probably more often used in packaged form, added to printed-circuit-board (PCB) designs in the firm of packaged surface-mount-technology (SMT) components.
A single chip or packaged MMIC can replace what was once a receiver front end constructed with a long list of discrete components and matching elements. These include frequency mixers, filters, oscillators, amplifiers, and attenuators and the transmission lines required to interconnect them.
Although the majority of MMIC devices are active in nature (such as amplifiers, receivers, and transceivers), the technology also lends itself to the production of low-cost, highly repeatable passive components (filters and frequency multipliers, to name two). For example, Custom MMIC has established itself as a reliable supplier of both active and passive MMICs, including a line of passive MMIC frequency doublers. The most recent model, the CMD226N3 frequency doubler (Fig. 2), turns input signals from 7 to 11 GHz to output signals from 14 to 22 GHz.
Because it is passive, signal power is lost from input to output, with typical conversion loss of 9 dB. But because the component is fabricated as a MMIC, it can also be housed in a miniature QFN-type surface-mount package to save circuit-board space, and its input and output ports are matched to 50 Ω, eliminating the need for matching circuit elements or transmission lines on the circuit board.
The passive nature of the doubler also minimizes its contributions to system phase noise, compared to an active frequency doubler in which amplification can also increase the amount of phase noise. For its small size, the MMIC frequency doubler can also handle fairly large input signals, to about 0.5 W (+27 dBm), enabling it to be used in a variety of different applications, including commercial satellite communications (satcom) systems and military radar systems.
Both MEMS and MMICs rely on semiconductor processing methods to form small device features on various types of semiconductor wafers. The motivation for both types of components is the same—to save size, weight, and cost for the circuit functions they are implementing—with investments in MMICs tracing back to defense system requirements established by ARPA (now DARPA) for highly integrated circuits capable of operating at low power levels for portable applications, including manpack radios.
Defense and aerospace needs for both MEMS and MMICs are still strong, but a number of emerging commercial applications are creating needs for both types of devices in areas of the spectrum that have been lightly used (compared to lower frequencies) to this point: the millimeter-wave region. The increasing use of millimeter-wave radar devices in automotive collision-avoidance systems and the expected need of mass-produced millimeter-wave radio MMICs and MEMS switches for short-haul links in 5G wireless communications networks has encouraged a number of MEMS and semiconductor companies to pursue higher-frequency devices for the expected large-volume demands for millimeter-wave components.
For example, Plextek RFI is a long-time innovator in GaAs MMIC technology currently involved in exploring the possible opportunities for active components in 5G systems. The company notes that no single millimeter-wave frequency band will serve the needs of short-haul links in all 5G systems around the world, due to the different spectrum allocations made by different regulatory groups, such as the FCC in the U.S. The FCC has designated spectrum at 37 GHz (37.0 to 38.6 GHz) and 39 GHz (38.6 to 40.0 GHz) on a flexible trial use basis for 5G systems.
Of course, practical implementation of communications links for these frequency bands assumes the availability of affordable components at these frequencies, which represents an opportunity for a mass-producible device technology such as GaAs MMIC technology. Plextek RFI’s design teams have already examined possible design solutions for the two frequency bands, including separate amplifiers or a single broadband amplifier to cover both bands, but created a GaAs MMIC that incorporates two switchable amplifiers to cover the two frequency bands with the best combination of small size, performance, and low cost.
The switchable GaAs MMIC amplifier chip is fabricated on a high-volume 0.15-µm GaAs pHEMT process and housed in an SMT-compatible air-cavity QFN plastic package. The innovative design is just one of the GaAs MMIC devices that the company is developing for 5G applications which, along with automotive electronic safety systems, should drive a growing need for both RF MEMS and RF/microwave MMICs for years to come.