Features

SAN JOSE — Scientists at Hewlett-Packard Co. and UCLA said Wednesday they have patented a means of getting around a significant hurdle in the race to build computer chips at the molecular level. 

The breakthrough could give researchers an efficient way to control the flow of information on such minuscule circuits, a requirement if tiny but enormously powerful molecular computers are ever to become a reality. 

Researchers in the burgeoning field of nanotechnology hope to someday create computers small enough to be sprinkled like dust, embedded in materials or perhaps even injected into the bloodstream to serve as diagnostic sensors. 

With molecular circuitry, a supercomputer could easily fit in a person’s hand. 

Today’s computers are run by silicon chips crammed with millions of transistors that turn on or off extremely quickly. 

But many researchers believe that by the middle of the next decade, because of the physical limitations of silicon, computer makers will cease being able to fit more and more transistors on a chip. 

In hopes of getting around that barrier, scientists are seeking ever-smaller materials for computer parts. Some have shown that individual sections of molecules can made into switches that turn on or off. 

The Hewlett-Packard-UCLA team, which is partly funded by the Defense Department, already has patented a way to connect molecular-scale switches with chemical “wires” that are just six to 10 atoms wide and two atoms tall — about 100 times smaller than the tiniest wires on chips today. 

They also have described a way to make the molecular systems run despite imperfections found everywhere in nature. 

Last year, R. Stanley Williams of HP and Philip Kuekes and James Heath of UCLA developed a chemical process and computer program that would allow the circuitry to be mapped like perpendicular city streets — so a computer’s central processing unit could know exactly where on the molecular grid certain information is being stored. 

But simply providing routes for electrical impulses to travel is not enough. For a molecular circuit to really work, researchers need to have a way of managing the way the signals travel. 

So to govern their small city, the HP-UCLA team propose creating a rough equivalent of traffic lights — an electric-chemical process that can sever the conductivity of certain points on the grid. Essentially, that creates a series of wires of varying lengths, mimicking the circuitry in existing computers. 

That process won a patent in November. 

“I believe that in 10 years we definitely will have hybrid molecular-silicon circuitry,” Williams said. “Molecules will take over more of the computational tasks of the system and the silicon will become just the input-output device and the power supply.” 

The federal government finds so much promise in molecular computing it is spending $604 million to support research in the field, up 43 percent from last year, said Mihail Roco, who heads the White House’s National Nanotechnology Initiative. 

He said the HP-UCLA process announced Wednesday “is a significant step in creating real circuits” that can connect to larger systems. 

“This breakthrough is a manifestation of the fact that we are moving from long-term, exploratory research to applied research in this field,” Roco said. “The developments are much faster than expected.” 

James M. Tour of Rice University, a nanotechnology researcher who has launched a private company to capitalize on developments in the field, said the HP-UCLA researchers still need to find a molecule that would hold up in circuitry for long-term use. 

But he said they appeared to have made an important step, “and they are to be praised.” 

——— 

On the Net: 

Researchers: http://www.hp.com, http://www.cnsi.ucla.edu 

Government nanotechnology site: http://www.nano.gov 

Tour’s company: http://www.molecularelectronics.com