Doubling what can be put on a piece of silicon every two years is Moore’s Law, and what the Semiconductor Industry Association seeks to achieve.
By James Buchanan
Semiconductors are likely something that most people give little thought to. As long as our electronic devises work, why consider what’s in them?
But for George M. Scalise, president of the Semiconductor Industry Association, semiconductors are a $110 billion industry that employs approximately 225,000 people in the U.S. alone. And as the president of the industry’s largest member organization, it is his job to guide the SIA as it maps out the future of this one very important industry.
“We fund a lot of research,” says Scalise, “about $80 million per year in research to drive the technology that will allow us to continue to implement Moore’s Law — which says that we can double the number of transistors we can put on a piece of silicon real-estate every two years.”
In 1965, Gordon E. Moore, a co-founder of Intel Corp., observed that the number of transistors on an integrated circuit doubles approximately every 24 months – ergo, his eponymous law.
“We have been able to live up to Moore’s Law consistently since 1965,” says Scalise. “Back then it was one transistor or diode on a chip of silicon, but today we are now approaching 2 billion.
“We can continue to shrink the dimensions of the device, both lines and widths. As a consequence, we are now in the 65 nanometer range. Pretty soon we will go down to 10, which is as far as we can go, because then you’re down to atomic layers and you just can’t go down any further. Physics just doesn’t allow it.”
Staying on the cutting edge of technological innovation is perhaps the most important endeavor of the SIA. Its 95 corporate members are responsible for 85 percent of the semiconductor production in the US. To give that figure some context, according to the SIA, in 2003 the semiconductor industry manufactured approximately 90 million transistors for every man, woman and child on the earth. By 2010, the SIA expects that number to increase to 1 billion transistors.
Guiding that level of growth requires not only a dedication to research, but organizing that research and its future path for the SIA’s members.
“We put together a technology road map about every two years that we originated about 15 years ago, and it looks out 15 years to define the issues we need to address to maintain the implementation of Moore’s Law,” says Scalise. “About 10 years ago we decided to internationalize it to not only bring the best and brightest minds in the U.S. to deal with the issue, but to bring them together from all around the world.
“To do that, we have about 400 of the best technologists dealing with the various disciplines, whether it’s lithography, or etch, or cleans, or on and on — and they look at the road map and these technology working groups will come up with their piece of it and we’ll consolidate that and publish it.”
The net result, Scalise says, is an incredible document.
“It is unrivaled in industry for defining a pathway for everyone involved – our suppliers, equipment manufacturers, applications users,” says Scalise. “You can go on down the list and everyone looks at this to see how they can continue to implement their business model. It is a very unique and important and well done document.”
Scalise goes on to say that the bulk of the research is handled by research universities through a well defined process the SIA has been using for 25 of its 30 year history. It is simply based on identifying research priorities and then requesting proposals from these institutions for research funding.
The SIA has also established certain research facilities as centers of excellence within its Focus Center Research Program. These are UC Berkeley, Georgia Tech, Carnegie Melon, MIT, and UCLA. These each tie in with another five to 10 universities, which Scalise says makes a total of 30 to 40 institutions handling some aspect of research for the semiconductor industry.
“It’s a very well organized, and very well functioning system that deals with the various issues very effectively,” Scalise says.
One particular focus, Scalise notes, is the drive toward developing workable nanotechnology concepts.
“Nanotechnology will get implemented in two time frames,” says Scalise. “As we go down through the final stages of scaling, the first implementation of nanotechnology will unfold – deposition methodologies and things of that nature. Later, we’ll get into device structures and so on. That will be the second phase and will be the real nanotechnology going forward.”
In the end, Moore’s Law is hopefully preserved and the semiconductor industry remains as vital and strong as it ever has, Scalise concludes.