The Incredible Shrinking Machine

By Philip Elmer-DeWitt

/ To the naked eye, the object mounted on a postage stamp-size wafer and held aloft by a pair of tweezers is all but invisible. Even under a bright light, it looks like nothing more than a speck of dust. But magnified 160 times in an electron microscope, the speck begins to take on shape and function: a tiny gear with teeth the size of blood cells. "You have to be careful when handling these things," warns Kaigham Gabriel, an engineer at AT&T Bell Laboratories . "I've accidentally inhaled a few right into my lungs."

The miniaturization of technology, having made extraordinary progress in the 40 years since the invention of the transistor, is about to make another shrinking leap. Adapting the chipmaking equipment used to squeeze millions of electrical circuits onto slivers of silicon, researchers are creating a lilliputian tool chest of tiny moving parts: valves, gears, springs, levers, lenses and ball bearings. One team at the University of California, Berkeley, has already built a silicon motor not much wider than an eyelash that can rotate 500 times a minute.

Welcome to the world of microtechnology, where machines the size of sand grains are harnessed to do useful work. Huge numbers of microscopic sensors are already employed to measure the temperature, air pressure and acceleration of airplanes and automobiles. Delco Electronics alone sells 7 million silicon pressure sensors a year to its parent company, General Motors, for use in power-train controls and diagnostics. But scientists at Berkeley, Stanford, M.I.T., AT&T, IBM and a handful of other research centers around the world see much broader possibilities for minuscule machines. They envision armies of gnat-size robots exploring space, performing surgery inside the human body or possibly building skyscrapers one atom at a time. "Microelectronics is on the verge of a second revolution," says Jeffrey Lang, a professor of electromechanics at M.I.T. "We're still dreaming of applications."

A report to the U.S. National Science Foundation last year listed dozens of near-term uses for the new micromachines. Among them:

-- Tiny scissors or miniature electric buzz saws to assist doctors performing microsurgery.

-- Micro-optical systems to focus lasers to the precision required for fiber- optic communication.

-- Miniature machine parts that could drive a new generation of tiny tape recorders, camcorders and computers.

Engineers and industrialists are rushing to put the new technologies to use. M.I.T. has invested $20 million in a new fabrication facility for micromachining and microelectronics. Japan's Ministry of International Trade and Industry is considering allocating nearly $70 million for the development of medical microrobots. "I'm absolutely amazed at how fast this field has progressed," says George Hazelrigg, a program director at the NSF, the Government agency spearheading the U.S.'s micromechanics effort.

Human interest in tiny machines dates back to the clockwork toys of the 16th century. But it was not until this century that making things smaller became a matter of military and economic survival. Spurred by the cold war and the space race, U.S. scientists in the late 1950s began a drive to shrink the electronics necessary to guide missiles, creating lightweight devices for easy launch into space. It was the Japanese, though, who saw the value of applying miniature technology to the consumer market. In his book Made in Japan, Akio Morita tells how he proudly showed Sony's $29.95 transistor radio to U.S. retailers in 1955 and was repeatedly asked, as he made the rounds of New York City's electronics outlets, "Who needs these tiny things?"

American manufacturers eventually learned what the Japanese already knew: that new markets can be created by making things smaller and lighter. (The popular phrase in Japan is kei-haku-tan-sho -- light, thin, short and small.) Ten years ago, Black & Decker scored big when it shrank the household vacuum cleaner from a bulky 11.2 kg (30 lbs.) to a 0.75-kg (2-lb.) device dubbed the Dustbuster. Tandy and Apple Computers put the power of a room-size computer into something resembling a television-typewriter and created an industry worth $75 billion a year.

Now these breakthrough products look hopelessly oversize. Last month Compaq unveiled a 2.2-kg (6-lb.) full-powered portable computer that fits in a briefcase. Sharp and Poqet make even smaller models that slip into a suit pocket. Today there are fax machines, radar detectors, electronic dictionaries, cellular telephones, color televisions, even videotape recorders that fit comfortably in the palm of a hand.

With the advent of silicon gears, springs and cantilevers, machines will become smaller still. These miniature moving parts can be etched on silicon using a variation on the photolithographic technique used to make computer chips. To build a tiny rotating arm, for example, layers of polysilicon and a type of glass that can be removed with acid are deposited on a silicon base. A hole for the hub is lined with the glass and then filled with polysilicon. When the glass is etched away, the hub remains and the arm is free to spin around its axis.

Sensors like those made by Delco were the first to combine microelectronics and micromachines on one chip. The typical microsensor is a thin silicon diaphragm studded with resistors. Because the electrical resistance of silicon crystals changes when they are bent, the slightest stress on the diaphragm can be registered by the resistors and amplified by electronic circuits.

As prices drop, these devices will become ubiquitous. By 1995 the typical car may contain as many as 50 silicon sensors programmed to control antilock brakes, monitor engine knock and trigger the release of safety air bags. Similar sensors are already employed in the space shuttle Discovery to measure cabin and hydraulic pressures and gauge performance at more than 250 separate points in the craft's main engines.

Medical applications are also being rapidly developed. Researchers at Maryland's Johns Hopkins have made a pill slightly larger than a daily vitamin supplement that has a silicon thermometer and the electronics necessary to broadcast instant temperature readings to a recording device. By having a patient swallow the pill, doctors can pinpoint worrisome hot spots anywhere within the digestive tract. Future "smart pills" may transmit information about heart rates, stomach acidity or neural functions. Says Russell Eberhart, program manager at Johns Hopkins' Applied Physics Laboratory: "This could change the way we diagnose and monitor patients."

Researchers at Tokyo University are pursuing an even more ambitious goal. Working under Iwao Fujimasa, an artificial-heart specialist, a team of 20 scientists is building a robot less than 1 mm (0.045 in.) in diameter that could travel through veins and inside organs, locating and treating diseased tissue. The group hopes to build a prototype within three years for testing on a horse, but the researchers first must obtain gears, screws and other parts 1,000 times smaller than the tiniest available today.

The ultimate fantasy of the miniaturists is tiny robot "assemblers" that could operate at the atomic level, building finished goods one molecule at a time. This is the far-reaching goal of an embryonic discipline called nanotechnology, so named because it would require manipulating objects , measured in billionths of a meter (nanometers). In Engines of Creation, the nanotechnologist's bible, K. Eric Drexler envisions a world in which everything from locomotives to cheeseburgers is assembled from molecular raw materials, much as proteins are created from their amino-acid building blocks by the machinery of a living cell.

Working with microscopic machines presents special challenges to scientists. Not only do they risk inhaling their tools or scattering them with a sneeze, but they also have to cope with a new set of physical laws. The problem of friction, for instance, looms ever larger as parts get smaller. The tiniest dust speck can seem like a boulder. Rotating a hair-width dynamo through air molecules, says AT&T's Gabriel, "is like trying to spin gears in molasses."

But the payoff can be enormous. As electronics manufacturers have discovered, the laws of economics at the micro level are as different as the laws of physics. A manufacturer might spend a small fortune putting hundreds of moving parts and circuits onto a single silicon chip. But when that chip goes into large-scale production and millions of copies are made, the economies of scale take over, and development costs virtually disappear.

Unfortunately, there is a limit to how many transistors can be squeezed onto the surface of a chip. Thus the attraction of micromachines. They give engineers a way to shrink the moving parts of a device rather than trying to shrink its computer controls further. Some experts believe that within the next 25 years micromachinery will do for machines what microelectronics did for electronics. Given the progress over the past quarter-century, that is saying a lot.

With reporting by Scott Brown/San Francisco and Thomas McCarroll/New York