MEMS for mobile communications: microelectromechanical (MEM) components and systems can enhance future wireless systems.

New functionality and extreme miniaturization of electronic systems are required for future wireless systems. The third and fourth generation of multi-mode/ multi-band wireless terminals will provide connectivity to new services, Internet and Mcommerce data, and still/video images. In addition, end users will be interested in the new functionality and features of the terminals.

Specifically, microelectromechanical (MEM) components and systems can reduce the size and cost of the initial third-generation devices by increasing the level of integration and by creating possibilities for tuning and directing the components. (1,2,3) They will also provide new architectures for radio systems and power management. Furthermore, MEMS will enable integration of new functionality, sensors and actuators into wireless terminals.

New technical solutions are needed for the user interfaces. Micromechanical sensors and microphones are available as components, and they can be seen as building blocks for more complex smart systems consisting of several integrated sensor elements. Entertainment and games are important drivers for new functionality, as sensors and actuators in mobile terminals. Micromirrors and microlenses create new possibilities for camera and display implementations, such as projection displays.

The role of MEMS in wireless communications will be described in a two-part series. In this article, the potential microsystem applications will be discussed, with the key focus on radio frequency (RF) MEMS applications. The benefits and limitations of the different MEM components will also be presented. The second article will review packaging, which is a key factor in determining the total cost and the overall size of the microsystems.

General MEMS Applications

The applications of microsystem technologies in mobile communications can be categorized into four main areas:

* enhanced and miniaturized RF implementations

* power/energy generation and management

* novel mass memory

* sensors and actuators for adding new functionality, including microphones, microphone arrays and optical MEMS.

The technical drivers for microsystems are: miniaturization; integration of multifunctional devices into compact systems; system-on-package and system-on-chip solutions for RF and sensor applications; and embedded intelligence in microsystems to control the communication and power consumption.

The key technologies that can create new added value in mobile communication applications are:

* components and microsystems for GSM/ WCDMA RF front-end, including low insertion loss and high isolation RF switches, tunable capacitors, high quality (Q-value) planar and three-dimensional (3-D) inductors and transformers, and VHF resonators

* components and microsystems for millimeter wave applications, including smart antennas and antenna arrays, cavity resonator-based filters and tunable phase shifters. This frequency range is possibly related to wireless local area network (WLAN) applications in the time frame of the research.

* components and microsystems for adaptive power management, including integrated 3-D power coils and transformers, microrelays, micromechanical resonators with high Q-value, and active cooling elements

* power generation, including microturbines, microgenerators, microengines, hydrogen- or methanol-based micro fuel cells, and polymer batteries

* mass memory technologies, including micromachined read/write of atomic resolution memories

* micromechanical sensors and actuators for user interfaces, including accelerometers, tilt sensors, gyroscopes, pressure sensors and magnetometers

* biometric identification, including fingerprint sensors

* cameras and displays, including micromirrors, microlenses, scanners and projection displays

* components for speech and voice, including micromechanical silicon microphones, microphone arrays, microspeakers and vibration suppression systems.

RF MEMS Applications

RF switches

Micromechanical components can be used in RF applications. The micromechanics enable electromechanical relays in the chip scale. MEM switches can be either galvanic or capacitive. The key drivers for the micromechanical RF switches are:

* good isolation, > 30 dB @ 2 GHz

* low insertion loss, < 0.3 dB @ 2 GHz

* possibility to isolate the switch control signal from the RF signal.

Several devices are under development. (2) However, problems remain to be solved: reliability of the electrical contact; long switching time (~0.1 ms); and high voltage level needed for the switch control.

The reliability of the switch contacts has been improved, but the miniaturization of devices has some fundamental limitations. Small contacts with small amounts of conductive material are vulnerable to wear. Furthermore, the switching time and the control voltage depend on each other: the faster the switch, the higher the necessary control voltage.

The problem with the large control voltage in some structures can be solved by using short-duration high-voltage spikes to trigger the switch but lower control voltages to keep it closed.

Tunable capacitors

Micromechanical tunable capacitors typically resonate mechanically at frequencies below 50 kHz. The mechanical resonance frequency determines the switching time. The mechanical inertia can be used to isolate the mechanical dynamics from the RF electrical signals. This approach creates the key benefit of micromachined tunable capacitors: The RF signals do not modulate the capacitance values. However, the capacitance value is influenced by the root mean square (rms) signal level of the RF signal.

The main design problem of tunable capacitors is the elimination of the series resistance. Figure 1 shows an electroplated metal film bridge capacitor combined with a planar inductor. The use of metal films solves the problem of series resistance; 0.05 to 0.5 [OMEGA] resistance levels can be achieved. However, the metal films create a new problem due to the large mismatch between the thermal expansion of the metal film and the substrate. Small enough temperature dependence can be achieved by mechanical structures that compensate the strain due to thermal expansion. (4)

High Q-value inductors

In general, the quality (Q-value) of the inductor depends on three different factors:

Q [approximately equal to] [omega]L/[R.sub.s]([omega]) x (substrate loss factor) x (self resonance factor)

where the first term is the geometrical quality factor of the coil depending on the operating frequency [omega], the inductance L, and the series resistance [R.sub.s]([omega]). The self-resonance factor of the coil depends on the parasitic capacitance of the coil. The optimization of the first term of this equation for a planar inductor on an ideal substrate leads to linear dependence between the coil area and the Q-value; for example, a coil with 1 [mm.sup.2] area optimally has a Q-value of 50.

High Q-value suspended coils on a silicon wafer have been successfully fabricated (Figure 1). However, the integration of the planar inductors on a silicon wafer has several drawbacks: a large area is required for a sufficiently high Q-value; the inductors interfere with surrounding circuits; and the substrate generates noise and causes parasitic capacitance.

[FIGURE 1 OMITTED]

Three-dimensional (3-D) structures have a better inductance-to-series resistance ratio than planar geometries. Most commercial discrete RF coils are wire-wound 3-D solenoids. For example, a 1-mm-long solenoid with three turns and a diameter of 1 mm has an inductance of 6 nH and a Q-value over 100 at 900 MHz. The overall size (volume) of the solenoid is the critical factor to improve the quality; miniaturization has its limits. Integration of 3-D inductors on a chip or in a package is the key challenge. Micromachining has been applied for creating 3-D solenoid inductors on chip. (5,6)

High-frequency micromechanical resonators

Several research groups have demonstrated micromechanical resonators up to 100 MHz. (2,3,7,8) The intrinsic losses of the silicon resonators are very small because of the favorable crystalline structure. Recently, a micromechanical resonator utilizing an 11.8 MHz longitudinal bulk acoustic wave mode has shown a mechanical Q-value over 200,000. (8)

The size of the micromechanical resonators at 10- to 100-MHz frequencies is in the order of 10 to 100 lam. Several resonators can be fabricated on one silicon die.

Micromechanical resonators can be used as building blocks of integrated MEM systems. The integration of these elements with ICs will also be possible in the long run. Integrated reference oscillators based on micromechanical resonators can be feasible when the stability and low enough temperature dependence of the MEM resonators can be established.

References

(1.) Petersen, K. 1982. Silicon as a mechanical material. Proceedings of the IEEE, May, Vol. 70, pp. 420-457.

(2.) Yao, J. 2001. RF MEMS from a device perspective. J. Micromech. Microeng., Vol. 10, pp. R9-R38.

(3.) Nguyen, C., et al. 1998. Micromachined devices for wireless communications. Proceedings of the IEEE, August, Vol. 86, pp. 1756-1768.

(4.) Nieminen, H., et al. 2001. European Patent Application 01660182.5-2203, filed 2.10.2001.

(5.) Yoon, J.-B., et al. 1999. High-performance three-dimensional on-chip inductors fabricated by novel micromachining technology for RF MMIC. Tech. Digest of 1999 IEEE MTT-S Int. Microwave Syrup., pp. 1523-1526.

(6.) Young, D., et al. 1998. A low-noise RF voltage controlled oscillator using on-chip high-Q three-dimensional coil inductor and micromachined variable capacitor. Tech. Digest of Solid State Sensor and Actuator Workshop, pp. 128-131.

(7.) Wang, K., et al. 1999. VHF free-free beam high-Q micromechanical resonators. Proceedings of the 12th IEEE International Conference on Micro Electro Mechanical Systems. Orlando, Florida, Jan. 17-21, pp. 453-458.

(8.) Mattila, T., et al. 2001. 14 MHz micromechanical oscillator. Proceedings of the 11th International Conference on Solid-State Sensors and Actuators, Munich, Germany, June 10-14.

Vladimir Ermolov, Heikki Nieminen, Kjell Nybergh, Tapani Ryhanen and Samuli Silanto are all with Nokia Group, Finland; e-mail: [e-mail].

Part 2 will be featured next month.

A version of this article was originally presented at 2001 SMTA International