The 450 mm Debate
by Stanley Myers, president and CEO of SEMI
Many in the industry agree that economics, as well as technical challenges, will be a future barrier to technology advancement. The challenges we face in funding device scaling and materials advancements have sparked another debate, namely a proposed transition to 450 mm wafers.
Semiconductors are powerful technology enablers, spawning countless numbers of new electronics products that have improved the lives of just about everyone on the planet. The Moore’s Law correlation of reduced cost and increased performance has delivered ICs with price points and capabilities not imagined even a few decades ago. Innovation, the lifeblood of the semiconductor industry, has been the driving force of this technology advancement. While there has always been a cost to innovation, the technology demands we are facing today require a whole new level of industry research and development funding.
Many in the industry agree that economics, as well as technical challenges, will be a future barrier to technology advancement. The challenges we face in funding device scaling and materials advancements have sparked another debate, namely a proposed transition to 450 mm wafers. It has been estimated that the cost of the 300 mm wafer transition was $12 billion, with the investment spread over the 1996 to 2003 timeframe. As we look at a transition to 450 mm, it appears that such a transition could take as long as eight years to bring to market and, according to some industry estimates, could cost well over $20 billion.
The difficulty with this debate is that most suppliers feel they have yet to recoup their investment in the 300 mm transition. Additionally, industry experts agree that existing and planned 300 mm fabs are currently operating below their optimum performance. A move to 450 mm at this time would not only diffuse efforts to improve the performance of 300 mm fabs, but also create competition for limited R&D funding.
The motivation for moving to larger wafers is improved silicon productivity. However, this can be achieved through means other than traditional wafer diameter scaling. Many in the industry, including SEMI, support an approach that is designed to maximize the efficiency and lifetime of existing platforms. The “300 mm Prime” concept, first coined by ISMI and still being fine-tuned in terms of its definition, is aimed at maximizing the return on the industry’s existing investment in 300 mm wafer fabs. We believe this approach is also more suited to the requirements of the market today. The shift to consumer-driven electronics has resulted in a short run, high product mix for fabs. These requirements are better served by more productive use of existing 300 mm fabs.
Clearly, there will be fewer device makers willing to make the transition to a 450 mm wafer substrate. The optimum output for a 300 mm fab is 80 thousand wafers per month. For a 450 mm fab, the optimum could be in the range of 120 thousand to 150 thousand wafers per month, resulting in an investment of $12 billion to $15 billion per fab, according to some industry experts.
Before a successful transition to 450 mm can be conceived, the industry will need to fund the investment in a way that provides a reasonable return to the supply infrastructure whose participation will be required. At this time, it is not clear that there is a benefit to the industry that outweighs the increased cost of the equipment required.
In the past, leading U.S. chip manufacturers were willing to bear the brunt of the research and development costs necessary for a transition to larger wafer sizes. However, when the industry migrated from 200 mm to 300 mm wafers, the equipment and materials companies carried a significant portion of the research and development burden with the promise of financial returns as the device industry moved to these larger wafers for production.
It doesn’t make sense for the equipment and materials community, which we represent, to invest in a transition to 450 mm wafers in the near future. That’s because industry revenues are not growing in line with rising research and development costs. The best course forward is to allow the industry to maximize its investment in 300 mm before considering a new wafer size transition.