Molecular Structure Used in Circuit System
March 23, 2000


A New Era in Technology for Computers

By JOHN MARKOFF, New York Times

Racing ahead of the most optimistic projections, a team of American computer researchers and chemists have quietly formed a company trying to open a new era of digital electronics by creating immensely powerful computing circuits based on trillions of individual building blocks, each no larger than a single molecule.

The potential commercialization of the technology -- thought to have vast potential, but generally believed to be at least a half a decade or decade in the future -- has surprised computer industry executives.

The company, Molecular Electronics Corp., was incorporated in December and opened its first office in Chicago in January. Its founding scientists are some of the leading researchers in a still immature field that might someday replace today's multibillion-dollar chip manufacturing factories with pure chemical processes growing tiny electronic circuits into vast arrays that make huge memory systems and perhaps powerful parallel computers.

"People in Silicon Valley view this as off the fringe," said James Ellen, a molecular electronics researcher at Mitre Corp. While he did not have detailed knowledge of the progress made by the Molecular Electronics team, he said, researchers in the field have been making significant progress in recent years, and he would not be shocked by early commercialization.

The founding scientists of the new company include three chemists, a theoretical physicist and an electrical engineer from some of the nation's most prestigious universities and research laboratories.

The company has obtained substantial initial financing from private investors and has hired a chief executive, Harvey Plotnick. In interviews this week the company's executives said that while they had still not built a working prototype, they believed they had solved all of the difficult research obstacles and would be able to create working models in 18 to 24 months.

"We have already achieved initial results, and we'll move from there and show people these are really doable things," said Mark Reed, co-founder and chief technology officer of the new company. Reed is chairman of the electrical engineering department of Yale University and said he planned to continue in his academic position.

He was also an author of a November 1999 article in Science magazine with a Rice University chemistry professor, James Tour, that described the creation of a single molecule switch, a basic component of any future system that might be created using chemical processes known as self-assembly.

Tour is a co-founder of the new company. The group also includes chemists at Pennsylvania State University, who have done pioneering work in self-assembly and the creation of so-called nanowires, and Brosl Hasslacher, a theoretical physicist.

Although the group has refused to say what its first product or prototype might be, the inclusion of Hasslacher, who has done pioneering work in parallel computing, suggests that it might be aiming higher than simply creating ultrahigh-capacity computer memories.

Several people close to the company said that the first prototypes would be memory devices and that Molecular Electronics researchers had already succeeded in the laboratory in creating arrays of molecule-sized switches from which information can be stored and retrieved.

This achievement, if it has indeed occurred, has not been publicly announced.

The molecular computing technology, which has been given the name moletronics by a Pentagon official, is based on a radically different manufacturing technique from the one used to build today's semiconductor chips.

A principal obstacle to Moore's Law, which states that the the number of transistors that can be etched on a sliver of silicon doubles every 18 months, is the exponential increase in the cost of new factories. Currently Intel Corp., the world's largest chipmaker, is spending more than $2 billion for each new factory.

"We're working with laboratory beakers," Hasslacher said. Molecular Electronics researchers are making the molecular structures relatively inexpensively inside a laboratory. Such a huge cost saving could give the products of the fledgling industry a chance against the powerful semiconductor.

Despite the researchers' optimism, all those involved in the start-up acknowledge major challenges. One is that molecular memories and other computing systems are likely to have far more faulty components than today's semiconductors. Thus, ways must be found to map out bad circuits and at the same time take advantage of billions of circuits that function correctly.

Molecular Electronics is not the first company to try to commercialize the new technology. In 1997, California Molecular Electronics Corp. was created to pursue similar efforts in the field. Last year the company received a National Science Foundation grant to pursue the development of components of a molecular-scale switching device. In February, California Molecular issued a $6 million direct public offering via the Internet.