For much of 2014, there has been little to report on the progress of 450mm. No surprise that it appears most suppliers have pulled back on 450mm investments and focus. This is evidenced by how many times the phrase 450mm is mentioned in some key player’s annual 10K filings with the Securities and Exchange Commission (SEC), as shown.
Though it was tough for me to find consistent annual data, in English, from Japanese companies, it is notable that SUMCO, reported to be in a key relationship with G450C, goes from twice mentioning 450mm wafers in their 2012 annual report, to no mention at all by the time their 2013 report is released.
Likewise, chipmakers and G450C have pushed out the expected 450mm transition. The industry 450mm leader, Intel, mentions 450mm nine times in their 2012 10K SEC filing, but just five in their 2013 filing. Conventional wisdom notes 2020 as a more likely transition year for 450mm wafers. This makes Dan Hutcheson, look like a techno-prophet™, based upon his prediction back in 2008.
It’s All in the Timing
What key was turned to slow the development of 450mm? Certainly ASML’s late 2013 announcement that they were putting 450mm on hold provided some influence. It clearly signified that, even with the substantial chipmaker-provided investments, 450mm EUVL was not going to be the technology used to pattern the much-needed test wafers. In hindsight, the July 2013 announcement of G450C’s large investment in Nikon for a 450mm-capable immersion scanner was an early indicator that an alternative lithography path to EUVL (and nano imprint, for that matter) was being pursued.
Full-blown semiconductor process equipment development requires numerous patterned test wafers along with an inspection system to look at the results. KLA has inspection equipment available for 450mm wafers, however, without a supply of test wafers, OEMs are stuck. Without this supply, it seems some slowdown in 2014 was inevitable. But this may all change in 2015.
Will Nikon Save the Day?
This past summer, G450C showed-off fully patterned 450mm wafers, made possible with Nikon’s immersion lithography tool. A 450mm capable system, based on Nikon’s 300mm NSR-S630D, is scheduled for delivery to G450C in April, 2015. This provides a credible tool for patterning the thousands of wafers needed for 450mm equipment development. When this tool is installed and operational, it will be a milestone for 450mm development.
A Missing Piece
Lithography, inspection, and prime, bare wafers are critical for producing patterned test wafers. The availability of the actual 450mm bare wafers to pattern is an area where there may be valid reason for concern. Granted, high quality (prime) 450mm wafers have been available, but will they be available in quantity and when? Particularly, will there be the quantity required for equipment development and testing at a reasonable cost?
The path to high-volume, reasonably priced 450mm wafer production is unclear. The technical and financial challenges for prospective 450mm wafer manufacturers have been well documented by SUMCO, SEMATECH, and the ITRS. The research and development costs, along with the current 450mm market uncertainty, appear to have removed a number of wafer manufacturers from the competition. As the dynamics of the silicon market have made this segment of the industry difficult to navigate, it is no surprise that even reference to 450mm wafers is now nearly absent in almost all wafer manufacturers’ annual reports available for review.
I am currently researching a technology that may provide a path to a lower-cost supply of 450mm test and production wafers. This advancement appears to provide an upgrade path for existing 300mm crystal pullers, at reduced technical risk, and with improved safety (EHS), as compared to 450mm plans I’ve seen. I’ll share more when I have some analysis to present.