SAN FRANCISCO – At recent Semicon West events, most fab tool vendors dismissed the need to migrate to the next-generation, 450-mm wafer size, saying the technology is too costly.
During this year’s event, there were no major 450-mm announcements—at least on the surface. To be sure, though, there were some vibes in the air that the technology is picking up steam. Many of vibes are mere rumors. Some is pure speculation. Others are educated guesses.
Besides a nebulous feeling in the air, there is suddenly a possible breakthrough in 450-mm. There is a new ''cost sharing’’ program between the industry and fab tool makers—a move that could accelerate 450-mm in the market. And on top of that, SEMI has recently unveiled some new 450-mm standards.
The ''momentum’’ is building in the 450-mm supply chain, said Scott Kramer, vice president of manufacturing at chip-making consortium International Sematech, which is leading the charge in 450-mm.
''There is sincere interest in 450-mm’’ in the industry as a whole, said Bob MacKnight, president and chief executive of Crossing Automation Inc., a fab tool automation vendor.
''There are no orders for 450-mm tools right now,’’ said Brian Trafas, chief marketing officer at KLA-Tencor Corp., but ''my sense is that there will be a big announcement in the next one or two quarters.’’
Trafas was unsure what exact announcement is pending. Some suspect that a company—namely Intel Corp.—could be ready to announce a 450-mm prototype fab. Others think that the announcement calls for another bombshell: The 450-mm roadmap could get pushed out.
For some time, 450-mm technology has been stalled due to cost and resistance by the fab tool makers to develop the necessary systems. Now, most believe that the three 450-mm proponents--Intel, Samsung and TSMC--will push the tool makers towards the next-generation wafer size. Tool makers may have no choice: Intel, Samsung and TSMC represent a major percentage of the capital equipment buyers in the market.
The three companies have talked about 450-mm prototype fabs by 2012. Some see 450-mm production in 2015. Others see 2018. Some believe 450-mm fabs will never happen, saying the R&D costs are too expensive.
No one is still quite sure who will pay for the tools or the R&D, that is, until now. With little or no fanfare, International Sematech is pushing a new ''cost sharing program’’ with fab tool vendors to accelerate the development of the next-generation wafer size.
International Sematech (ISMI) ''is accepting submissions from process and metrology equipment suppliers for 450-mm tool transition projects to be considered for ISMI cost-sharing and other assistance. We want to understand the technical and commercial challenges your company faces in building tools to accommodate 450 mm wafers,’’ according to the chip-making consortium.
For some time, Sematech has been developing and defining the fab and tool technologies for 450-mm. The pressing problem is to get fab tool vendors on board. During an interview at Semicon West, Kramer said the chip-making consortium is ''engaged’’ with undisclosed fab tool and materials vendors, some of which were reluctant to jump on the 450-mm bandwagon in the past.
In fact, there were ''zero’’ 450-mm tools that were in development in the past, said Tom Jefferson, 450mm program manager for International Sematech. Now, ''there are a lot more than that,’’ he said.
At present, there are a smattering of announcements and standards for 450-mm. The real tipping point for 450-mm will occur when a chip maker announces a 450-mm pilot line and the associated node for the technology, said KLA-Tencor’s Trafas. That way, fab tool vendors can make a more solid commitment, he said.
In any case, the debate rages on for the technology. The probability that the semiconductor industry will one day use 450-mm silicon wafers in manufacturing has increased significantly, though the chance of achieving volume production by the original target date of 2012 is remote, according to a recent report by analysts at Semico Research Corp.
In another viewpoint, the shift towards 450-mm fabs is still unlikely for the foreseeable future due to costs, according to VLSI Research Inc. At a minimum, it is going to take a commitment of at least $1 billion to a consortium to take 450-mm technology to the next level, which is to develop early production prototype tool sets, according to the firm.
Hoping to accelerate the possible migration to 450-mm prototype fabs by 2012, International Sematech has aggressively moved on several fronts. In recent times, the chip-making consortium has moved into the ''test wafer generation'' stage and installed the early tools in its 450-mm prototype clean room in Austin, Texas, including metrology and wet-clean systems.
At one time, Sematech said the facility was using the first 450-mm handlers from Brooks Automation Inc. and carriers from Entegris Inc. Japan's Nippon Mining & Metals Co. Ltd claimed to have developed the first 450-mm polycrystalline silicon wafers.
During a presentation to fab tool vendors at this year’s event, Sematech claimed it has achieved the following milestones in the arena:
If Sematech is serious they better bring the wafer mfgs. into the mix. One maker producing some poly wafers isn't going to cut it. Intel and Samsung will probably fix a target price of 1.5 to 2 times 300 mm for the 450s ignoring the fact that they will have more than 3 times the silicon in them including kerf and grinding loses but, not crystal pulling. If they demand first through (no remelt) crystal the costs will be staggering. I hope PV can absorb the reallocated material!
The litho problems will be harrowing as the line widths are approaching physical limits of resolution while the local and global flatness demands on the wafer and equip makers are just as difficult to achieve. This will be the hardest migration yet.
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.