2004 ITRS Updated

The 2004 update of the International Technology Roadmap for Semiconductors (ITRS) is now available at http://public.itrs.net . Of the 219 tables contained in the 2003 ITRS, 128 tables were updated and four new tables were added. Additionally, 12 figures were updated. Many of the difficult challenges for the ITRS were revisited and updated to reflect the most current thinking and most critical tasks for semiconductor R&D. The 2003 DRAM technology node definition remains unchanged (half-pitch of Metal 1 layer for DRAM) in 2004.

"The importance of the roadmap is that, once you establish a common roadmap for technology needs and possible solutions, then it's possible for the consortia, researchers, suppliers to work on detailed solutions," noted Paolo Gargini in a keynote speech launching the new ITRS at SEMICON Japan. Gargini has been serving as the chairman of the ITRS since 1998, and is also director of the technology manufacturing group at Intel and Intel fellow. "In the last six, seven years, the roadmap has become really a key element to compare against the internal R&D and also to guide some of the external activity for the consortia."

One important change in the 2004 ITRS is a new technology requirement color and definition for Interim Solutions are Known. "Since things are becoming more difficult, we have added a new symbol," Gargini said. "Normally, white meant it was ready for manufacturing, yellow needed some additional improvements and red very difficult, but in some cases, people are able to do it even though you cannot fully meet the specifications." The new symbol has a hatched yellow background and a red diamond.

Also, a sub-chapter on emerging research materials was added (see "ITRS Adds Emerging Tech Materials ").

According to the 2004 ITRS, lithography is not expected to be a limiter to the 65 nm generation, nor (maybe) the 45 nm node. Exposure tools with 193 nm wavelength will dominate the next technology nodes (65, 45 and maybe 32/22 nm) with both dry and immersion solutions (see "ITRS Update Puts More Emphasis, Faith in Immersion "). However, Gargini noted that EUV will challenge 193 nm at 45/32/22 nm. "193 nm lithography has finally arrived, and it appears that 65 nm will not be a problem. It may be 45 nm as well. With the addition of immersion, there is much more confidence in this approach. There has been a lot of progress on EUV. We can expect, beginning with 45 nm, that we will see some challenge from EUV suppliers," he said. Control of critical dimensions below 4 nm was identified as a major issue.

One of the more surprising aspects of the new ITRS is the emphasis on 450 mm wafer sizes, given the industry's recent and not-forgotten anguish in going from 200 to 300 mm. Introduction of 450 mm wafers is expected to occur (according to the ITRS) in the 2011-2015 timeframe. "If you look at the transition of 200 mm wafers in 1990 and 300 mm in 2001, now is the time to think about 450 mm," Gargini said. He also noted that 675 mm will probably come after that. Using 2012 as the date when 450 mm will be required, it means that, in 2005, a lot of activity on standards needs to begin. "It would be a very important year because all of the activity and definition of the wafer and the carrier will have to begin. This was one of the key accomplishments of the 300 mm, especially the cooperation between I300I and J300 was very beneficial for the whole industry. We hope to repeat the same successes," Gargini said. "Once again, there are many options with respect to different carriers, and this will have to be agreed. And also 450 mm at this point is a proposal because it will double the productivity of 300 mm, but the final number needs to be decided by the whole industry."

Mobility enhancement techniques will allow an increase in transistor performance without the introduction of high-k gates and metal electrode materials for 65 nm and maybe for the 45 nm node. The low-k roadmap is essentially unchanged for the first time in 10 years. Materials with k~2.7 are in manufacturing and k~2.4 materials will follow. "New techniques to enhance mobility are just becoming manufacturing standards, and the low-k roadmap is finally, after 10 years, finalized, and materials with 2.7 dielectric constant are in manufacturing and 2.4 will come soon," Gargini said.

Many new package solutions needed to increase functionality and reduce cost will require additional attention in the future. "One field that has received a lot of attention is packaging, because it is probably the easiest way of bringing quickly to manufacturing commercialization multiple technologies that are heterogeneous in nature," Gargini said.