The era of nano-manufacturing is being born in hundreds of labs that are racing to perfect a technique called self-assembly

If you just listen casually to a description of what Sandia National Laboratories has been working on, you would think it had wasted its time reinventing the wheel: It has developed a robot that can walk and pick up and deliver loads of cargo. In an age of advanced assembly and landings on Mars, that hardly sounds impressive -- except that Sandia's robot is a molecule.

Called a motor protein, it has two little feet on one end and a tail that can grab things on the other. Once a special chemical is added to the solution in which it resides, the protein begins moving along strands of fiber that are one-fifth the width of a human hair, says Bruce Bunker, a Sandia researcher who's in charge of the project.

He's betting that his experiment could play a part in heralding the arrival of a new era in manufacturing -- a leap akin to going from the Stone Age to today's complex assembly lines. In this new world, atoms and molecules will gravitate toward one another and self-assemble into components -- and then, perhaps, into a computer or an artificial organ, says Benjamin Miller, a researcher at the University of Rochester in Rochester, N.Y.

MAKING HEADWAY.

For now, making such complicated things this way is obviously science fiction. But the idea behind self-assembly is very much here, right now. Mother Nature has used self-assembly for eons to create and foster life. Mankind already uses it to make nonwrinkle trousers, fragrances, and silver polish. And it's used in microelectronics and for making anticorrosion coatings. And even in drugmaking, it's about to take off, says Mihail Roco, chair of the Nanoscale Science, Engineering & Technology Subcommittee of the National Science & Technology Committee, which coordinates the federal government's nanotech research and development efforts.

That's because scientists recently have made so much headway in designing artificial molecules that self-assemble in a predictable pattern -- an outgrowth of steady increases in research funding for such projects worldwide.

About one-quarter of the 2,000 or so nanotechnology projects the National Science Foundation now sponsors involve self-assembly -- and funding for nanotechnology should grow about 20% year-over-year, to $305 million, in fiscal 2005, says Roco. Total federal nanotech funding through a program called the National Nanotechnology Initiative, which Roco helps coordinate, should reach nearly $1 billion next year.

NO LONGER IMPOSSIBLE.

Even that figure is likely to be overshadowed by private funding. Everyone from drug companies such as Merck (MRK ) and Pfizer (PFE ) to tech whizzes 3M (MMM ), IBM (IBM ), and Hewlett-Packard (HPQ ) is investing heavily in self-assembly.

Their interest is easy to understand: Self-assembly can make manufacturing fast and cheap. In theory, anyway, humans will be able to simply design artificial molecules that assemble in a desired pattern -- and let nature take its course. A bonus is that because self-assembly devices will be made from the bottom up, there'll be less waste, says Christine Peterson, president at the Foresight Institute, a nanotechnology-policy think tank in Palo Alto, Calif.

More significant, perhaps, self-assembly and molecular manufacturing will help companies develop products that would be impossible to make using conventional methods. Take semiconductor manufacturing. For years, transistors have become smaller and smaller in order to deliver more performance from each chip. The problem is, current manufacturing methods are reaching their physical limits.

SQUASHED BY A BUG.

So, within five years, IBM will likely employ self-assembly to do the job. Last December, Big Blue demonstrated a flash-memory chip that had been created via self-assemly. IBM used special polymer molecules, derived from those found in styrofoam and plexiglass, to make transistors and connections that are one-tenth the size of today's, says Chuck Black, who leads IBM's self-assembly project.

A key advantage of this method is that it won't require a major equipment overhaul. Polymers are already widely used in semiconductor manufacturing, so every self-respecting chipmaker owns so-called spin-coating tools, which IBM uses to control self-assembly. Spin-coating gear can be thought of as a record turntable. The polymer solution is poured on top of it. Then, the whole thing is spun around, so excess solution flies off, and the polymer layer is evened out. The rotation speed controls the resulting layer's thickness, Black explains.

The equipment changes that will be necessary probably won't be visible to the naked eye. Nadrian Seeman, professor of chemistry at New York University, dreams of nanorobots so tiny they could be squashed by the foot of a bug. He recently made a molecular device that can execute a half-turn revolution when placed in solution with certain DNA strands. Within a decade, Seeman imagines, this nano-robotic arm could become part of a molecular assembly line -- in a chemical factory so small it might sit on your tabletop.

GLOW ALERTS.

Self-assembly can also help reinvent existing products. One example is a sensor being developed at the U.S. Energy Dept.'s Oak Ridge National Lab that's designed to replace the current, cumbersome methods of checking on storage tanks that contain radioactive waste. Today, such tests require humans outfitted in protective suits to use a robotic arm to extract a sample. The sample is then taken to a lab and checked for cesium, one of the biggest contributors to radioactivity.

The process costs around $1 million a tank and takes several days, says Gilbert M. Brown, senior research scientist at the lab. But self-assembly could cut both the cost and the time.

Brown is working on a sensor that will analyze the sample in minutes for a cost of around $10,000 (even better, the sensor could be reused). The device will contain a coating of special molecules that self-assemble with molecules of cesium. Radioactivity levels will then be determined by measuring the fluorescence of the new structure or structural changes on the sensor, Brown says. The device should be ready for prime time in two years or so, he predicts.

BUILDING "BONE."

Perhaps no field will experience as much reinvention thanks to self-assembly as medicine. The technique could greatly improve orthopedic implants, for instance. The average life of an implant today is about 15 years, according to the American Academy of Orthopaedic Surgeons. And when implants give out, they often break or crack the bone they're attached to. That's because it's hard to get bone to grow on the implant, which is usually made of titanium or ceramics, says Thomas Webster, an expert at Purdue University in West Lafayette, Ind. A trick involving self-assembly could fix that, though.

Webster has developed a special coating comprising molecules that self-assemble into nanotubes, a structure similar to that of bone. His proof-of-concept experiments (animal and human trials are yet to follow) show that when such a coating is applied to an implant, bone cells recognize these structures as their own and grow onto them, he says. Indeed, 60% more bone cells stick to the implant.

If that ratio holds up in future experiments, that could translate into a 60% improvement in the lifespan of implants -- and make a huge difference for patients, especially those who, say, get a kneecap replacement in their 20s and then have to repeat the surgery throughout their lives.

DISEASE SENSORS.

Self-assembly can also lead to completely new treatments. Researchers at Northwestern University in Evanston, Ill., have designed artificial molecules that form log-like structures that the researchers think could be used to create a piece of artificial spinal cord to help paralyzed patients regain some mobility, explains Mark Ratner, professor of chemistry at Northwestern. The log structures would act as a scaffold around which spinal-cord tissue or bone tissue could regenerate.

The University of Rochester's Miller is working on molecules that can self-assemble into drugs, then sense cells with disease and target them with treatments. His molecules can bind to genes that cause certain genetic diseases and block them from functioning. Such a drug could appear on the market within five years, he says. And he predicts that self-assembly will change the nature of drugmaking.

Of course, many obstacles must be overcome. Self-assembly in drugmaking still has to be proven nontoxic to humans. There's also the lengthy federal approval process to go through. Plus, scientists have yet to learn how to design the right molecules -- and figure out how to force them to assemble in certain ways.

Still, "the field is moving very fast," says NYU's Seeman. Only a year ago, he says, he showed his students a top-10 list of major challenges nanotechnology faced. Today, three are already resolved. At this pace, he adds, self-assembly and molecular manufacturing will come into commercial use sooner rather than later.

By Olga Kharif in Portland, Ore.