Paradigm Phase Shift Amrita Chakraborty and Bhaskar Gupta
M
icroelectromechanical systems (MEMS) have been familiar in the field of microelectronics and device technology since 1965, when Harvey C. Nathanson invented the first MEMS device (known as the resonant gate transistor). Since that time, MEMS have represented a prominent field for academic research, with simplified implementations of several devices such as sensors, transducers, actuators, and accelerometers being fabricated. According to MEMS technology, all devices could be fabricated using standard integrated circuit (IC) techniques, with the device’s lateral dimensions normally ranging from a maximum of 500 nm to submicron ranges. H o w e v e r, not until the 1990s was MEMS technology first implemented to realize devices operating in the microwave and millimeter-wave frequency ranges [by IBM Research and the Hughes Research Laboratory
(HRL)]. A novel technology, RF MEMS, paved the way for the realization of a host of devices such as switches, switched capacitors, varactors, tunable filters, and phase shifters—all operating in the microwave and millimeter-wave frequency ranges. Although RF spiral inductors and interdigited capacitors can involve fabrication strategies similar to those of MEMS device
image licensed by ingram publishing
Amrita Chakraborty (
[email protected]) and Bhaskar Gupta (
[email protected]) are with the Department of Electronics and Telecommunication Engineering, Jadavpur University, Kolkata, India. Digital Object Identifier 10.1109/MMM.2016.2616155 Date of publication: 12 December 2016
22
1527-3342/17©2017IEEE
January/February 2017
implementations, they are not classified under the category of RF MEMS because all MEMS-based designs must consist of an actuating portion capable of undergoing movement (in-plane or out-of-plane vibrations) when subjected to external stimulus. Thus, in a nutshell, RF MEMS devices can be broadly classified as those devices having 1) dimensions essentially of a few microns to submicron order and 2) a movable or actuating portion that undergoes free movement upon the influence of an externally applied stimulus. The first RF MEMS-based device was the basic ON– OFF switch and varactor developed in 1991 [1] at HRL (California), under a project funded by the U.S. Defense Advanced Research Projects Agency. The switch exhibited extremely low-loss performance up to a frequency range of 50 GHz, which easily surpassed the performances of the popular switching modules, such as PIN diodes, field-effect transistor (FET) switches, or gallium arsenide (GaAs)-based devices, used in telecommunications at the time. However, the switches suffered from reliability issues and so remained at rudimentary levels. It was only in 1995 that the RF MEMS metal-to-metal contact switch, capable of operating from dc to 60 GHz, was developed by research groups at the Rockwell Science Center [2]. That same year, Texas Instruments developed the first RF MEMS-based switched capacitors or dielectric contact switches suitable for 10–120-GHz applications [3]. The standard ON–OFF switch and switched capacitor modules were identified as the basic (or, rather, the smallest) entity and underwent further development in subsequent years to realize RF subsystems such as tunable filters, resonator modules, and phase shifters. The years between 1998 and 2010 witnessed dramatic developments in the field of microwave- and millimeter-wave-based system design and integration. This period was marked by a large number of publications in the research domain of RF MEMS-based devices. Earlier, microwave devices/subsystems were mostly implemented using semiconductor switches and switching circuits. Common examples were PIN diodes, FET switches, and ferrite-based devices, in which each component had associated inherent bottlenecks that limited performance at higher frequencies.
Phase Shifters Before we detail their technical operational principles and provide a literature review of RF MEMS-based phase shifter devices, it is first important to understand the basics of phase shifters. Phase shifters can be defined as two-port passive microwave devices that allow adjustable phase changing of the incoming RF signal at the output port. Both the input and output ports should have perfect impedance matching to result, ideally, in zero attenuation of the outgoing signal. These approximations are realized by proper design criteria to obtain the best performance. Early phase shifting of the RF signal was obtained by
January/February 2017
Distributed-line—or, more precisely, distributed MEMS transmission line (DMTL)—phase shifters are by far the most popular of all RF MEMS-based phase shifters. mechanical adjustment of motors, which exhibited robust construction, consumed immense space, and were prone to aging and mechanical wear and tear. With the advent of microtechnology, however, mechanical adjustments were gradually replaced by electronic components such as PIN diodes, FET switches, and ferrite-type components, which could address all these issues and, in addition, enhance phase shift accuracy. Phase shifters find their main applications in areas of defense, where they form an integral component in phased-array radar systems. Phased-array radars comprise an array of radiating elements (in most cases, microstrip patch antenna arrays) that allow the radiated beam to be directed through desired angles in space. This beam scanning can be electronically controlled by varying the phase shift of multiple phase shifters connected in the feed network of each antenna element. Therefore, the number of phase-shift elements required for beam-scanning operations occupy relatively less space and are easily integrable in the form of monolithic microwave ICs (MMICs). In this context, it is worthwhile to present a brief background for various electronic implementations of phase shifters.
Overview of PIN Diode-Based Phase Shifters A PIN diode is a p–n junction that has a very minimally doped or intrinsic region sandwiched between the p-type and n-type regions. The addition of the intrinsic region results in various characteristics: i.e., conductivity can be controlled in forward bias, and the capacitance can be reduced in reverse bias. PIN diodes are extensively utilized in microwave circuits for amplitude modulation and attenuation and in the fabrication of high-performance RF switches, phase shifters, and limiters. In phase shifters, the PIN diodes are employed as electronic switches, switching the bias current from the forward to the reverse bias mode. The intrinsic region controls the ON–OFF state of the diode switch such that 1) under forward bias, the diode’s impedance is significantly reduced and 2) under reverse bias, it offers very high impedance to the diode. Therefore, PIN diode phase shifters can generate phase shifts by switching the signal between two different path lengths l 0 and l 0 + l, as shown in Figure 1. The phase shift (say, {) corresponds to the additional path delay bl (i.e., { = bl), where b is the propagation constant of the medium. However, the dc bias
23
circuitry must be completely isolated from the path of RF signal propagation, leading to complicated structures.
Overview of FET-Based Phase Shifters The FET switch is a three-terminal device in which switching is controlled by varying the gate voltage. The low- and high-impedance states required for switching applications are obtained by alternating the gate voltage such that it is equal to zero during one state and greater than the pinch-off voltage in the other state. In either case, the switch consumes approximately zero static dc bias power [4] and can substitute directly with the PIN diodes shown in Figure 1 for a basic switched-line phase shifter. Thus, the analysis of FET-based RF switches is highly simplified by considering only a passive mode of operation. The FET switch provides several advantages compared to PIN diodes. It shows significantly faster switching speed (~ns), very low dc power consumption, compatibility for monolithic integration, and simpler dc bias circuits because the gate (control) is inherently decoupled from the RF path at lower frequencies. The following section considers the utility of FETs in the development of MMICs.
Overview of MMIC-Based Phase Shifters The introduction of GaAs III–V compounds has, in large part, replaced the use of PIN diodes because FETs promote easy integrability with ICs and offer lower dc power consumption and faster switching speed of operation. Silicon-based complementary–metal-oxidesemiconductor (CMOS) technology provides strong competition to GaAs-based MMIC switches, which depend to a great extent on the working principle of MOSFETs. In hybrid microwave ICs (MICs), both active and lumped components are connected with distributed networks on a planar transmission line, generally of microstrip configuration, by soldering or wire-bonding techniques. MMICs have gained popularity due to the flexibility in design and implementation, whereby both active and passive components are fabricated simultaneously on direct bandgap semiconductor substrates (e.g., GaAs in this case). This technology eliminates the need to attach discrete components and thus reduces loss incurred due to wire-bond interconnects. The
Input
PIN Diodes
Output
I0 I0+I
Figure 1. A basic schematic showing phase shifter operation using PIN diodes.
24
major advantages offered by MMICs are low cost, small size, simplified packaging, better reliability, low power consumption, faster switching speed, broader operational bandwidth, the capability of being manufactured in bulk, and multifunctional performance. However, high insertion loss remains a serious limitation of MMIC-based switching elements [4].
Comparison of Devices Discussed Thus Far Table 1 compares the varied performances exhibited by PIN diodes, FET switches, ferrite-based devices, MMIC-based switches, and RF MEMS-based switching elements. RF MEMS-based devices are reported to exhibit lossless performance up to terahertz frequency ranges, thus aiding in the removal of quite a few amplifier stages commonly employed in standard radar modules. Ferrite-based devices exhibit moderate switching speed (~a few ns), low dc power consumption, and compatibility for MMIC integration. However, higher cost, fabrication complexity, and frequency limitations have curtailed their use in wireless communications and rendered them impractical for mobile satellite communication devices. Thus, even though ferrite phase shifters have low insertion loss and can handle significantly high power, their cost and fabrication complexity issues are yet to be resolved. Unlike ferrites, semiconductor phase shifters using PIN diodes and FETs are inexpensive, smaller, and commercially available in packaged and ready-to-use forms. However, their applications are limited because of the significant insertion loss incurred at higher frequency ranges and their poor power-handling capability. PIN-diode phase shifters consume more dc power (3–10 mW per diode) than their FET-based equivalents but provide low-loss performance at X-band. While FET switches may have lower power handling, PIN diodes are known to handle multikillowatt power levels in pulsed-mode operation in radars. The advantage of FET-based phase shifters is that they consume virtually zero dc power and can promote on-chip integration with low noise, thereby reducing the expense associated with subsystem assembling in phased-array radar systems. However, FET-based designs introduce a significant amount of loss in the front end, around 4–6 dB at 12–18 GHz [5], [6] and 8–9 dB at 35 GHz for higher bit designs [7], [8].
Overview of Ferroelectric Phase Shifters To conclude this overview, we offer a brief discussion of ferroelectric phase shifters because the topic is widely explored and supported by several publications to date. However, as our focus in this article is on RF MEMS-based phase shifters, we will not attempt an in-depth study.
January/February 2017
TABLE 1. A performance comparison of various developing phase shifter components. Parameters
PIN-Diode-Based RF Switch
FET-Based RF Switch
Ferrite-Based RF Switch
MMIC-Based RF Switch
RF MEMS Switch
Weight (oz)
Light (0.5–1)
Light (40
Power-handling capability (W)
~kW in pulse mode; ~200 W in CW mode
100
