For a decade or longer, scientists have been fascinated by the potential of micromachines -- devices so small that some of the components are invisible to the naked eye. Although potential applications have only begun to be conceived, the range of possibilities includes: * Tiny micromedical pumps implanted in the human body to drive internal drug-delivery systems. * Low-cost, high-performance gyroscopes for use in automobile or aircraft navigation systems. Miniature devices capable of manipulating individual cells in microsurgical procedures or DNA research -- or, perhaps, removing micron-size particles from the surface of a computer chip. Applying fabrication methods similar to those used in the production of integrated circuits, various research teams have pursued the development of micromachines and "microengines" to power the tiny devices. Until recently, most of the work has been limited to the development of single-level structures. However, a team at Sandia National Laboratories, Albuquerque, has found a way to produce more complex devices -- including interconnected motors and gears -- using microelectronic fabrication methods. The Sandia micromotor and external gears are etched on the same silicon-wafer substrate, using a repetitive photolithography process, with multiple layers of silicon dioxide sandwiched between layers of polysilicon. The research, undertaken as part of a defense program initiative to improve locking mechanisms on nuclear weapons, has considerable commercial potential. "We are now in the process of negotiating on several different [commercial] applications for the technologies," says Paul J. McWhorter, an electrical engineer and manager of Sandia's Intelligent Micromachine, Microsensor & CMOS Technology Dept. McWhorter headed the 20-member team that achieved the breakthrough. Other key team leaders included Jeff Sniegowski, senior member of the technical staff who led the effort to develop the fabrication methods; and Ernest Garcia, a mechanical engineer who oversaw precision design work on the micromechanical. components. "We view the work as having the potential to continue to enhance the safety and security of nuclear weapons," asserts Robert J. Eagan, Sandia's vice president for electronics, materials research, and components. "That's the driving force. But the potential goes well beyond that. Just think of a machine that has intelligence, but that is very small and can do things which would be very difficult with other kinds of instruments. In your wildest dreams, you can conceive of some exciting medical applications." A videotape produced by Sandia offers a greatly magnified view of the micromachine in operation: A motor consisting of two tiny silicon combs with a shuttle between them generates 0.5 microwatts of power. Energized by on-off electric voltages, the stationary combs take turns pulling the shuttle, using electrostatic attraction. An attached shaft turns a minuscule drive gear -- which is 30 microns across, less than the diameter of a human hair -- one quarter of a rotation. A second comb-drive engine, at right angles to the first, causes the next quarter turn. By alternating forces, the two engines convert reciprocating motion into rotary motion. The small gear has teeth that mesh with those of a second gear that is 30 times larger, causing it to turn. In one application, the larger gear -- which has two holes -- serves as an optical shutter because a beam of light can pass through the openings in some positions but not in others. Development of the optical shutter is expected to have military application as well as significant potential for telecommunications switching devices. McWhorter believes his team's work represents "a very big step forward," though he quickly notes that it is really a "logical next step" in the ongoing micromachining work that began at the University of California-Berkeley about 10 years ago. Others have built micromotors, he notes. "But what we've demonstrated here is the ability to build motors and tools -- and then to connect the motors and tools together, all on the chip, to get these little systems that can really work. With simpler technologies, you could build a motor, or maybe you could build a tool, but it really takes a more sophisticated technology to be able to tie the whole thing together." A number of obstacles had to be overcome. "It's not easy meshing gears that are that tiny, "McWhorter points out. "Imagine trying to make gears out of something a hundred times thinner than a sheet of paper and aligning them to the right height to turn each other." But that wasn't the only hurdle. "At some point," McWhorter notes, "you need to be able to integrate electronics with the micromachines, so that you are not only building a small mechanical device -- you are building a small mechanical device with intelligence. And that is one of the things we've been able to demonstrate here- -- an ability to build intelligent micromachines." A major challenge for the team was to devise a way to fabricate components capable of interacting in a three-dimensional framework. "We can build mechanical structures that have three levels of polycrystalline silicon -- or polysilicon," says McWhorter. "And having three levels of polysilicon enables the fabrication of more complex systems" than were possible in the past. The first level contains the engine, the second the gears, and the third the linkages connecting the engine to the gears. Moreover, the Sandia devices also accommodate a fourth level for electrical interconnections. "As far as we know," McWhorter says, "this is the very first complete demonstration of four-level technology, which enables a huge variety of systems." Alternative micromachine fabrication methods require steps such as plating and assembly, whereas the Sandia approach accomplishes the entire process using photolithography and silicon etching, much the way microelectronic circuits are manufactured. The significance is that the devices can be made relatively inexpensively with large-volume batchproduction methods -- perhaps in facilities that once produced microprocessor chips such as Intel's 286 or 386. The Sandia development has broad implications for a variety of scientific endeavors where there is "a need to be able to work in the micro domain . . . to manipulate and do things in very small dimensions," McWhorter observes. He imagines, for example, that intelligent micromachines might one day be deployed to remove defects from computer chips. "In integrated- circuit manufacturing, you are making very high-density microelectronic devices," he says. "And a particle one micron in size can be a killer defect on a chip. So you might like to have a small machine that could go in and gingerly pull that particle off the surface of the wafer without scratching the chip. It would involve manipulation at a very small scale." One of the first real-world applications for the technology, the Sandia researchers expect, will be the manufacture of micromachined switches for locking devices on nuclear weapons systems. Since the tiny motor and gearing has much less mass than its macro-world counterparts, it stands a better chance of surviving an impact, notes Garcia, the primary components designer. The switches would serve as a mechanical interlock "to keep electrical signals from flowing in the event of a crash or a fire," explains Eagan. "And that prevents detonation of the weapon when you don't want it to happen." In the commercial arena, one potential application is in optical switches for telecommunications systems based on fiber optics. Micromachines able to reposition tiny mirrors could function as optical switching devices. Longer-term, the technology might be used to fabricate minuscule gyroscopes for automobile navigation Systems. Coupled with tiny accelerometers -- which have already been made with micromachining methods -- the gyros would help track the position of a car and relay the information to a dashboard display. (Accelerometers measure linear distance, while the gyros detect a change in direction -- such as turning a corner.) It is likely, McWhorter suggests, that any such application would be designed to work in conjunction with satellite-based global positioning systems (GPS), filling in the gaps when GPS signals are blocked out by mountainous terrain or other obstacles. Further development work is needed to extend the lifespan of the gyros for automotive applications, observes Sniegowski, the fabrication engineer. "We're still learning on that aspect of it," he says. "We have a way to go. But, hopefully, we can extend the wear properties to get us well into multiple billions of revolutions." Perhaps the real importance of the work of the Sandia team is that it may lead to new discoveries in the future. "From a scientist's point of view," says Sniegowski, "the microengine itself is leading us to understand much better how mechanisms work at these dimensions. There is the whole issue of friction and wear and how one element interacts with another element. As you shrink things down in size, you have effects coming into play that normally don't play a role in a macroscopic environment." At very small scales, McWhorter adds, "Things like inertia and gravity are not as important as new effects like 'stiction,' which is a surface effect that tends to draw objects together, causing them to stick. So your normal scaling rules and normal mechanical-design rules do not really work on these small devices." "We are learning a lot about how these things interact in terms of friction," says Sniegowski. "As we learn more, it could lead to new ways of thinking, new technologies, new ways of designing machinery at these dimensions. In the end, I suspect that the work we're doing will give us a better idea of whether we can shrink down conventional-type machinery to the micro dimension -- or if we have to totally rethink things and go some other route. "So far," he adds, "it looks very promising that we can shrink down conventional-type machinery designs. It is working for us. In fact, it has worked better than we ever expected."