• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Cramming More Onto a Computer Chip

Posted on: Tuesday, 20 May 2008, 06:00 CDT

Few people outside the world of semiconductor research know about the work of Veena Misra or Ki Wook Kim at N.C. State University.

But the two engineers are certain everyone will know if they and their colleagues fail: The march of technological progress as we know it will cease.

They are sure of this. Moore's Law -- named after Intel co-founder Gordon Moore -- has predicted it.

"Our whole standard of living will stop evolving," said Dinesh Mehta, vice president of Semiconductor Research Corp. in Durham. "Our quality of life will stop evolving."

The Semiconductor Research Corp., or SRC as its known in the industry, is the world's leading university-research consortium for semiconductors. It has invested $1.1 billion in research since it was founded in 1982.

The work of Misra and Kim is part of the newest round of grants totaling $11.2 million spread among three dozen universities.

The goal is to develop a radically new chip when today's semiconductors reach their technological limits. That is expected to happen in about 12 years.

"We have reached a point where people take it for granted, that technology will always improve," said Larry Sumney, president of SRC. "That should never happen."

In its simplest form, semiconductors are a series of transistors placed on a silicon-based chip; think of companies in Silicon Valley.

But that simplicity vanishes at today's standards, where 2.1 billion transistors are placed on a square wafer that is about 1/2 an inch across.

Shrinking transistors to a size measured in nanometers is the result of Moore's Law -- an industry standard that dictates that the number of transistors on a chip should double every two years.

Meeting the standard has allowed remarkable progress in the technology that propels everything from medical equipment to airline travel.

But it might be easier for most consumers to think in terms of televisions, cameras, computers and appliances, said Henry Becker, president of Qimonda North America, a global supplier of digital memory based in Cary.

Why chips matter

"If you walked in Circuit City or Best Buy and could make everything disappear that uses a chip, the shelves would be practically empty," Becker said. "It is that much a part of our lives."

But the expectation that your next computer will be more powerful than the last one, or today's technology will be cheaper tomorrow, only holds true if the industry meets Moore's Law.

And the current structure has its limits.

"The transistors are already so small you can't see them," Sumney said. "Eventually we will reach the point where they won't work because there is only room to move one electron, which is not reliable. You will need a different way to move information."

This is where researchers such as Misra and Kim come in.

To move information across a space at faster speeds, Misra and graduate student Steven Novak are experimenting with a recent discovery known as domain wall logic that relies on magnetic fields.

It faces significant challenges, Misra said. To begin with, the approach has only worked on large instruments at slow speeds. But it holds significant promise.

Kim also hopes that his area of research will push the industry toward an answer. He is working with substances such as graphene -- a thin layer of carbon atoms -- to greatly reduce the amount of heat today's chips generate.

The two engineers are part of large teams of researchers whose work is sponsored by SRC.

Alarmed by the decline in U.S. research dollars for semiconductor chips, industry leaders created SRC more than two decades ago as a way to focus the industry.

To meet that goal, SRC solicits proposals from chip makers about basic research necessary to continue advancements.

Kudos for SRC

Once a proposal is selected, the fees that each company pays to join the consortium are used to help pay for university research.

SRC coordinates research among more than 110 universities divided roughly into short-term, midterm and long-term needs of the industry.

In recognition of its work, the organization in 2007 was awarded the National Medal of Technology, the country's highest honor for groups or individuals that improve quality of life.

Research money for the next generation of chips strengthened those federal ties with money coming from the National Institute of Standards and Technology. The National Science Foundation has also paid for some projects.

The national stature of the group comes despite the semiconductor industry's relatively low profile in North Carolina.

SRC settled in Durham, Sumney said, largely because former Gov. Jim Hunt was a strong supporter of the electronics industry, and the Microelectronics Center of North Carolina had just been formed.

But Sumney said states such as California, Texas and New York have since shown more interest in nurturing the industry, which means the SRC staff spends a fair amount of time outside the state.

That matters little in the scheme of things, he said. The race for the next generation of chips is an international competition.

"China, Taiwan, Japan -- we are all searching for that breakthrough," Sumney said. "There is no reason it can't happen here."

Source: The News & Observer

More News in this Category