MEMS Technology Emerges

(This article was originally published in IEEE Computer, vol. 32, no. 6, June 1999, pp. 61-66.)

Kirk L. Kroeker

In the research labs since the 1980s, microelectromechanical systems (MEMS) are now materializing as commercial products. MEMS are semiconductor chips that integrate mechanical elements, sensors, actuators, and electronics on a silicon substrate. They are mainly used as pressure and chemical sensors, light reflectors, and switches. But MEMS technology is also being used as microactuators -- small mechanisms that trigger a mechanical (as opposed to electrical) system.

The electronics involved in MEMs are fabricated using IC process sequences (like CMOS or BiCMOS). But the micromechanical components are fabricated using compatible micromachining processes that either selectively etch away parts of the silicon wafer or add new structural layers to form the mechanical and electromechanical devices.

MEMS promise to revolutionize many consumer technologies by making system-on-chip technology more realizable. In fact, MEMS technology is a growing market presence. A recent research report by Semiconductor Equipment and Materials International, one of the largest trade organizations for the semiconductor industry, projected that the MEMS device industry will grow from this year's current level of $1.93 billion to nearly $10 billion annually in the year 2000. Several other market research studies have projected even higher growth.

Despite such optimistic statistics, investment in MEMS design and production is insufficient. Most MEMS devices are modeled using analytical tools that result in a relatively inaccurate prediction of performance behavior. As a result, MEMS design is usually trial and error, requiring several iterations before a device satisfies its performance requirements.

This design methodology -- combined with the length of time and high cost associated with MEMS prototyping -- makes it difficult to develop commercial products. The availability of suitable design tools combined with computer networks to provide access to high-performance workstations and supercomputers could alter this situation. Such tools will be coming to market in the near future.

Meanwhile, here is a sampling of a few exciting MEMS technologies that are already emerging.

MEMS for display applications

Texas Instruments' Digital Micromirror Device is a MEMS technology that incorporates an array of fast digital micromirrors onto a memory chip. TI designed digital light-processing systems based on the DMD to present images that seem less pixellated than conventional LCD displays.

Each digital light switch of a DMD is an aluminum micromirror, 16 nm2, that can reflect light in one of two directions, depending on the state of the underlying memory cell. The mirror is rotated by electrostatic attraction, which is produced by voltage differences developed across an air gap between the mirror and the memory cell. The mirror rotation is limited by mechanical stops to plus or minus 10 degrees. With the memory cell in the "on" state, the mirror rotates 10 degrees. With the memory cell in the "off" state, the mirror rotates in the other direction 10 degrees.

A comparison between an LCD-projected image (left) and a Digital-Micromirror-Device-projected image (right). Both the LCD and DMD photos were taken under the same conditions, with each projector optimized for focus, brightness, and color. Note the high level of pixelation in the LCD image in contrast to the nearly seamless DMD image. DMD offers superior picture quality because the DMD mirror pixels are separated by only 1 nm, which eliminates most pixelation.

When the DMD is combined with a suitable light source and projection optics, the mirror reflects incident light either into or out of the pupil of the projection lens, using a simple beam-steering action. The DMD accepts electrical impulses that represent levels of brightness at its input. It then outputs digital light in the form of optical impulses to the eye. Because of the short pulse duration, the optical impulses are interpreted by the eye of the observer as analog light containing up to one billion or more color and gray-scale combinations per pixel. The fast switching time results in a lag-free image, which means that screen refresh rates can be much faster than LCD technology.

The light pulse durations are determined by the precise division of time, so digital light can be very accurate. The resulting projected image faithfully reproduces the original source material and the image is more stable than other types of displays, independent of temperature or age of the projector, and is free from photo degradation effects, even up to brightness levels necessary for electronic cinema.

MEMS research

Henry Guckel at the University of Wisconsin, Madison, has been performing silicon research in MEMS since 1977.

Guckel's work has resulted in the creation of two main categories of devices: the polysilicon planar pressure transducer (which is currently being used for automotive applications by SSI Technologies to give drivers feedback on oil, transmission, and even tire pressure) and the polysilicon resonant transducer (which is being developed with Honeywell for use as a high-resolution transducer for various applications, including those that measure pressure, acceleration, strain, temperature, and vibration). Transducers convert physical or mechanical energy (motion, temperature) into electrical energy (current).

In 1988, Guckel started a research project in the area of deep X-ray lithography and electroplating. He extended and improved the original process to apply to sensors and actuators and most recently to precision engineering. This process is currently used in applications from MCNC, Sandia National Laboratory, MEMStek Products, Brookhaven National Laboratory, and Semaphore TRI.

The fabrication of micron-scale mechanical components using X-ray lithographic techniques has recently received attention as an emerging new technology that could have far-reaching consequences in MEMS and other manufacturing. Structures that are difficult to fabricate with precision, such as long square holes with small internal dimensions and arbitrary orientations, become straightforward using this "photon machining" technique.

MEMS for microrelay applications

A relay is an electrical switch that relies on the movement of a mechanical piece to open or close a circuit. A small current energizes the relay, which closes a gate, allowing a larger current to flow through. MCNC recently introduced a MEMS microrelay designed for telecommunications, automated test equipment, and automotive applications. MCNC researchers consider the new microrelay to be the world's smallest commercially available electromechanical relay, measuring just 1.5 mm × 1 mm with a height of 600 microns.

The thermally actuated microrelay is designed with nickel-surface micromachined components that have gold contacts for high conductivity. The metal-to-metal contacts offer low on-resistance and high off-resistance. Such characteristics are said to make the microrelay suitable for a range of telecommunications applications, including RF applications, modems, subscriber-line interfaces, and cross-connect switching.

According to the company, the microrelay may ultimately be used to replace larger diodes in telecommunications switching applications, which would mean reduced size and the ability, therefore, to build in more capabilities. Other possible applications include automated test equipment, in which the microrelay can be used in automotive electronics, including accessory bus connects.

The advantages of MCNC's MEMS microrelay technology stem from the fact that the MEMS relay is manufactured using standard, semiconductor-compatible processing. Therefore, thousands of devices can be produced on a single wafer, leading to higher device-to-device reproducibility and potentially lower costs. In addition, the MEMS relay is silicon-based, which means it can be integrated with other silicon-based devices.

CONCLUSION

What are the advantages of MEMS devices over current devices that perform the same functions? MEMS devices can be so small that hundreds of them can fit in the same space as one single device that performs the same function. And using MEMS means that you don't need to use many cumbersome electrical components, since the electronics can be placed directly on the MEMS device. This integration also has the advantage of picking up less electrical noise, which can improve the precision and sensitivity of sensors.

Also, by using IC processes, hundreds to thousands of these devices can be fabricated on a single wafer. This kind of mass production can greatly reduce the price of individual devices. Thus, MEMS devices will likely be much less expensive than their macroworld counterparts once they become more prevalent. For a great many applications, MEMS is sure to be the technology of the future.

Kirk L. Kroeker is a freelance editor and writer. Contact him at http://kroeker.net.