SEMI Equipment Productivity Working Group (EPWG)
Findings Concerning a Wafer Size Transition to 450 mm
November 2005 – April 2008
An open, rigorous and defensible analysis of the economic benefits of a wafer-size transition to 450 mm wafers shows that the best way to enable continued technical progress and profitable growth for the entire semiconductor industry is to reject 450 mm engagement until a need for it is substantiated, and to focus work and resources on targeted 300 mm Prime opportunities.
The move to a 450 mm wafer-size transition being proposed is based on faulty assumptions and a flawed economic model, and should be rejected based on fact, not accepted based on intuition or emotion.
Beginning in 2004, SEMATECH executives were forecasting a significant relative increase in the manufacturing cost of IC’s starting in approximately 2006. They proposed that a near-term transition from 300 mm silicon wafers to 450 mm wafers was an essential element of an effort to combat this forecasted cost increase. Given the recent painful transition from 200 mm to 300 mm wafers, these assertions attracted great attention and concern throughout the industry. In previous years, both SEMATECH and SEMI have issued reports indicating that the combination of slower industry growth and increasingly challenging technology goals is creating a significant gap in RD&E resources available to fund future innovation. In an effort to address these concerns collaboratively, Board Members from SEMI and SEMATECH held joint meetings which resulted in the formation of a Joint Productivity Working Group (JPWG). The group was tasked with replacing “intuitive consensus” with “analytical rigor” in guiding investment priorities for the industry.
The JPWG has met for over two years, yet has not been successful in reaching agreement—between manufacturers and suppliers, or even among chip manufacturers—on the relative merits and drawbacks of a possible 450 mm transition. With few representatives of the chip-making community participating in the JPWG effort, the SEMI EPWG1 was formed and has completed a series of economic studies that are presented here.
Economic Findings and Conclusions
The EPWG has conducted simulation, survey, and modeling to support research, analysis, and discussion of the role of wafer scale-up in enabling continuous productivity gains. Initial work has contradicted ISMI’s initial assertion that wafer processing cost curves were shifting unfavorably. Rather, no statistically significant shift in these costs is evident.
Subsequent EPWG work over the past two years has shown 450 mm should be an extremely low-priority area for industry investment. Given the industry’s limited R&D resources and the significant technical challenges ahead, careful analysis shows that 450 mm wafer scale-up represents a low-return, high-risk investment opportunity for the entire semiconductor ecosystem. Based on bottom-up analyses of fab economics and operations, EPWG has shown that the economic benefits of a wafer scale-up are intrinsically limited, and in fact are shrinking (due to a shifting cost structure). Furthermore, a 450 mm transition would not be justified by market demand until well into the 2020’s (if at all), and its economic impact on chip costs would be neutral to negative—definitely insufficient to recoup the significant investment in the transition.
Unfortunately, the JPWG activity has suffered from minimal participation by chip manufacturers. In general, suppliers have typically outnumbered customers (excluding consortia) by at least 4:1 in most JPWG meetings; and discussion materials may not have been properly shared with all sponsoring chipmaker executives. As a result, ISMI has emerged as the key contact and counterpart for suppliers.
The EPWG cautioned ISMI against further 450 mm work based on their strong analytical findings. ISMI has not tried to refute these bottom-up analyses, and it has not provided any credible analysis to substantiate the need or the economic merits for a scale-up to 450 mm in the foreseeable future. Based on analysis, the EPWG believes that ISMI has, unfortunately, taken a fundamentally flawed approach in launching a 450 mm program. In particular, in moving forward with their effort, ISMI:
1. Based their 450 mm decision on a top-down, unsubstantiated, 30% cost savings "expectation" survey, rather then rigorous analysis.
2. Bundled high-return and lower-cost automation / architecture shifts, tool innovation, and continuous improvement benefits with the high cost of a wafer-size increase. This overstates the value of a 450 mm transition, and these approaches can be decoupled to improve return on R&D investment.
3. Created a premature and unnecessary so-called “blue-diamond” deadline for a “go/no-go” decision on 450 mm, forcing a flawed decision to be made with limited data and minimal analysis.
Since the “blue-diamond” decision date, the attempt to engage ISMI in objective 450 mm discussion has reached an impasse. The EWPG has leveraged solid economic rationale, technical knowledge, and operational insights in a transparent mode to evaluate and prioritize potential initiatives. The lack of reasonable discussion based on fact, rigorous analysis, and thoughtful consideration has led the EPWG to request discontinuing the JPWG engagement on 450 mm work while publishing our work to date. The EPWG will continue its efforts on 300 mm Prime (300 mm NGF) work.
Risk of Proceeding with a Flawed Decision
The debate over the merits and disadvantages of a wafer scale-up is not, as some have portrayed it, an academic battle of dueling models. Instead, it is an attempt to tackle what is one of largest threats to industry prosperity in our time.
While most semiconductor manufacturers are highly skeptical of 450 mm, the current dynamics are clearly pushing suppliers for engagement in this low-return, high-risk endeavor. This presents the entire industry – not just the supplier community – with a clear and present danger of wasting precious resources.
SEMI and EPWG members are concerned that industry-level joint activity (which should have been an opportunity for advancing the broad interests of the industry) might circumvent the real need of the majority of the industry and ultimately result in a slowdown in pace of technology.
Summary, Recommendations and Next Steps
The EPWG has analytically shown that a transition to 450 mm should not be a high priority issue for the semiconductor industry. Work of any type (including standard-setting) on 450 mm is liable to be premature and suboptimal, thereby distracting and diverting resources from activities with proven cost-saving or cycle-time improvement value.
Given this, the EPWG will be focusing their analysis efforts on the economics of optimizing 300 mm for the markets and business models which exist today and will predominate in the future. Initial EPWG analysis and surveys show that highly agile fabs which can handle rapidly changing consumer trends and markets could significantly enhance the value of the 300 mm wafer size generation. If properly planned and executed, this effort could and should push out or even eliminate the need for larger wafers while benefiting the entire semiconductor ecosystem.
Therefore, our next steps will be independent analysis as well as collaboration with ISMI and interested chipmakers on key topics intended to optimize the return-on-investment for development activities in the 300 mm NGF. These findings and strategies will be documented elsewhere.
Our analysis concluded that the best way to enable continued technical progress and profitable growth for the entire industry would be to reject 450 mm engagement until a need for it is substantiated, and to focus work and resources on targeted 300 mm Prime opportunities. Ultimately it is, of course, up to each chipmaker and equipment supplier to individually determine the best means to meet their customers’ and investors’ requirements. We intend for the analysis presented here to provide insight and data for industry participants to make the optimal roadmap, timing, and investment decisions best suited for their companies.
1 The SEMI Equipment Productivity Working Group consists of senior executives from ASML, Applied Materials, Axcelis, Brooks Automation, KLA-Tencor, Lam Research and TEL.