Will Tiny Robots Build Diamonds One Atom At A Time?

By MICHAEL D. LEMONICK

On its face, the notion seems utterly preposterous: a single technology so incredibly versatile that it can fight disease, stave off aging, clean up toxic waste, boost the world's food supply and build roads, automobiles and skyscrapers--and that's only to start with. Yet that's just what the proponents of nanotechnology claim is going to be possible, maybe even before the century is half over.

Crazy though it sounds, the idea of nanotechnology is very much in the scientific mainstream, with research labs all over the world trying to make it work. Last January President Clinton even declared a National Nanotechnology Initiative, promising $500 million for the effort.

In fact, nanotechnology has an impeccable and longstanding scientific pedigree. It was back in 1959 that Richard Feynman, arguably the most brilliant theoretical physicist since Einstein, gave a talk titled "There's Plenty of Room at the Bottom," in which he suggested that it would one day be possible to build machines so tiny they would consist of just a few thousand atoms. (The term nanotechnology comes from nanometer, or a billionth of a meter; a typical virus is about 100 nanometers across.)

What would such a machine be good for? Construction projects, on the tiniest scale, using molecules and even individual atoms as building blocks. And that in turn means you can make literally anything at all, from scratch--for the altering and rearrangement of molecules is ultimately what chemistry and biology come down to, and manufacturing is simply the process of taking huge collections of molecules and forming them into useful objects.

Indeed, every cell is a living example of nanotechnology: not only does it convert fuel into energy, but it also fabricates and pumps out proteins and enzymes according to the software encoded in its DNA. By recombining DNA from different species, genetic engineers have already learned to build new nanodevices--bacterial cells, for example, that pump out medically useful human hormones.

But biotechnology is limited by the tasks cells already know how to carry out. Nanotech visionaries have much more ambitious notions. Imagine a nanomachine that could take raw carbon and arrange it, atom by atom, into a perfect diamond. Imagine a machine that dismembers dioxin molecules, one by one, into their component parts. Or a device that cruises the human bloodstream, seeks out cholesterol deposits on vessel walls and disassembles them. Or one that takes grass clippings and remanufactures them into bread. Literally every physical object in the world, from computers to cheese, is made of molecules, and in principle a nanomachine could construct all of them.

Going from the principle to the practical will be a tall order, of course, but nanomechanics have already shown that it's possible, using tools like the scanning tunneling electron microscope, to move individual atoms into arrangements they'd never assume in nature: the IBM logo, for example, or a map of the world at one ten-billionth scale, or even a functioning submicroscopic guitar whose strings are a mere 50 nanometers across. They've also designed, though not yet built, minuscule gears and motors made of a few score molecules. (These should not be confused with the "tiny" gears and motors, built with millions of molecules, that have already been constructed with conventional chip-etching technique. Those devices are gargantuan compared with what will be built in the future.)

Within 25 years, nanotechnologists expect to move beyond these scientific parlor tricks and create real, working nanomachines, complete with tiny "fingers" that can manipulate molecules and with minuscule electronic brains that tell them how to do it, as well as how to search out the necessary raw materials. The fingers may well be made from carbon nanotubes--hairlike carbon molecules, discovered in 1991, that are 100 times as strong as steel and 50,000 times as thin as a human hair.

Their electronic brains could themselves be made from nanotubes, which can serve both as transistors and as the wires that connect them. Or they may be made out of DNA, which can be altered to carry instructions that nature never intended. Armed with the proper software and sufficient dexterity, a nanorobot, or nanobot, could construct anything at all.

Including copies of itself. To accomplish any sort of useful work, you'd have to unleash huge numbers of nanomachines to do every task--billions in every bloodstream, trillions at every toxic-waste site, quadrillions to put a car together. No assembly line could crank out nanobots in such numbers.

But nanomachines could do it. Nanotechnologists want to design nanobots that can do two things: carry out their primary tasks, and build perfect replicas of themselves. If the first nanobot makes two copies of itself, and those two make two copies each, you've got a trillion nanobots in no time, each one operating independently to carry out a trillionth of the job.

But as any child who's seen Mickey Mouse wrestle with those multiplying broomsticks in The Sorcerer's Apprentice can tell you, there's a dystopian shadow that hangs over this rosy picture: What if the nanobots forget to stop replicating? Without some sort of built-in stop signal, the potential for disaster would be incalculable. A fast-replicating nanobot circulating inside the human body could spread faster than a cancer, crowding out normal tissues; an out-of-control paper-recycling nanobot could convert the world's libraries to corrugated cardboard; a rogue food-fabricating nanobot could turn the planet's entire biosphere into one huge slab of Gorgonzola cheese.

Nanotechnologists don't dismiss the danger, but they believe they can handle it. One idea is to program a nanobot's software to self-destruct after a set number of generations. Another is to design nanobots that can operate only under certain conditions--in the presence of a high concentration of toxic chemicals, for example, or within a very narrow range of temperature and humidity. You might even program nanobots to stop reproducing when too many of their fellows are nearby. It's a strategy nature uses to keep bacteria in check.

None of that will help if someone decides to unleash a nanotech weapon of some sort--a prospect that would make computer viruses seem utterly benign by comparison. Indeed, some critics contend that the potential dangers of nanotechnology outweigh any potential benefits. Yet those benefits are so potentially enormous that nanotech, even more than computers or genetic medicine, could be the defining technology of the coming century. It may be that the world will end up needing a nanotech immune system, with police nanobots constantly at microscopic war with destructive bots.

One way or another, nanotechnology is coming.