ISS SPEAKERS SAY SEMICONDUCTOR INDUSTRY WILL OVERCOME NEAR TERM TECHNICAL HURDLES
Challenges Will be in Finding Cost Efficient New Materials
Half Moon Bay, January 11, 2005 – The semiconductor industry will overcome the hurdle of the next two technology nodes, but the "long laundry list" of required new materials will need to be shortened to make it economically viable, according to speakers on the second day of SEMI ISS 2005.
Based on the number of design starts, the 90-nanometer node will be a fairly smooth transition, and the industry can "take it for granted that 45 and 32 will happen", according to Sass Somekh, president of Novellus Systems.
"The number of designs in 90 nm is staggering -- 350 designs that are ongoing or already taped out," he said. For the 65-nanometer node, there are 40 designs, and a "handful" for 45-nm. "That gives us several more years of intense technology competition. Intel will lead the way and everybody else will need to catch up," he said.
Somekh said that despite slowing growth in the semiconductor industry, "there is no end in sight to the digital revolution." The downside: more price pressure driven by cost conscious consumer spending. "That cost pressure translates through the supply chain, which means efficiency and productivity are very important," said Somekh.
In order to compete at the leading edge, device makers are forming more technology alliances, which are a form of consolidation because they result in fewer equipment buying decisions. And the device makers that remain in the market are focusing on technology as a differentiator, argued Somekh.
"The major customers are building mega 300 mm fabs, and these [IDMs] would like to see very strong competition among the [equipment and materials] suppliers," he said.
To make the transition to the 45-nanometer node possible, the industry will have to shorten its "long laundry list" of new materials required, according to Bill Siegle, senior vice president and chief technologist, Advanced Micro Devices Inc.
Introducing a new advanced material into the semiconductor stream requires close cooperation across a broad range of disciplines and typically takes longer and costs more than originally anticipated, he explained. "We eventually get there, but it’s always pretty messy," said Siegle.
He cited silicon on insulator (SOI) and low-k dielectrics as examples of new materials that had a difficult transition from the lab to the fab. On the positive side, the industry is often able to extend the life of older materials technologies.
Looking further ahead, Siegle said that CMOS technology will need to be replaced by new nanostructures within the next decade, and this will pose many new materials challenges. "Maintaining Moore’s Law is not only going to require materials advances…but is also going to require innovation in system architecture and design," he said.
SEMI ISS 2005 continues at the Ritz Carlton, Half Moon Bay, until January 12. From January 13-14, the venue will host the SEMI Strategic Materials Conference (SMC).
SEMI is a global industry association serving companies that develop and provide manufacturing technology, materials and services to make semiconductors, flat panel displays (FPDs), micro-electromechanical systems (MEMS) and related microelectronics. SEMI maintains offices in Austin, Beijing, Brussels, Hsinchu, Moscow, San Jose (Calif.), Seoul, Singapore, Tokyo, Shanghai and Washington, D.C. For more information, visit SEMI at www.semi.org.