Development of next-generation IC wafer metrology must simultaneously address two major challenges: providing measurement and detection capabilities for continuously decreasing critical dimensions, and returning the investment in a 450 mm fab.
The Enabling Role of Metrology in the 450 mm Transition
Rand Cottle,1 David Nessim,2 Frank Robertson,1 Menachem Shoval3
1G450C 2Intel 3Metro450
Development of next–generation IC wafer metrology must simultaneously address two major challenges. The first–independent of wafer size and dictated by Moore's Law–is providing measurement and detection capabilities for continuously decreasing critical dimensions (Figure 1). Detection and measurement of smaller dimensions on increasing topographical complexity drives the need for higher resolution and lower signal–to–noise (S/N) ratio.
The second major challenge is dictated by the economics of larger wafer size. Returning the investment in a 450 mm fab is associated with a die output rate proportional to the increase in wafer area (nominally 2.25x higher). To avoid either increasing the number of metrology tools (for a given die output rate) or reducing metrology–based measurements, 450 mm metrology tools must achieve wafer–per–hour throughput times (TPTs) equal or better than their 300 mm wafer predecessors. Bigger and heavier stages must move faster and accelerate/decelerate very quickly from one measurement point to another without detriment to measurement (stability, accuracy, etc.) or cleanliness. It also means faster acquisition and real–time analysis of huge amounts of data from a larger surface or number of die.
Figure 1. Providing measurement and detection capabilities for continuously decreasing CDs is key to next–gen wafer metrology.
Metrology Must Lead
Unit process development can't progress without reliable metrology. The first step in establishing an early test wafer (ETW) pilot line is to acquire and characterize metrology capabilities for bare wafer particle inspection, film thickness and stress, defect inspection and review, CD, overlay, wafer geometry, and nanotopography. Having these capabilities in place as early as possible enables rapid learning at the leading edge of technology progression. Enabling metrology is consequently a major focus for the Albany, N.Y.–based Global 450mm Consortium (G450C), and many of the first tools in the G450C demo line are metrology tools.
Challenges of Early Metrology
The most commonly used metric for determining if a metrology tool is meeting performance requirements is to gauge it in terms of the ratio of tool measurement precision to process tolerance (P/T). Historically, P/T ≤ 0.1 is the goal; i.e., metrology uncertainty is consuming no more than 10% of the process tolerance for variation. As dimensions continue to shrink to 14 nm and below, it's getting increasingly difficult to achieve a P/T of 0.1. In the case of 1X nm node technology combined with 450 mm wafers, process tolerances are to a large extent uncertain at this point.
Early metrology tools are typically best–in–class 300 mm tools scaled up to handle 450 mm wafers. While actively in development, 450 mm calibration/qualification wafers are not yet available for most early development needs. 300 mm wafers are being used for initial testing. Potential differences between 300 and 450 mm wafers and film stacks add to the measurement uncertainty. An early priority for the G450C metrology team will be getting reference standards fabricated on 450 mm wafers.
The initial G450C capability will consist of a distributed toolset with some process and metrology tools at supplier or other sites. Reference artifacts are required to correlate tools from around the globe. Traceability from organizations like the National Institute of Standards and Technology (NIST) would also be highly desirable to make sure we're not just on the same page, but on the right page.
450 mm metrology tools, being the first to arrive, are the first to face and solve wafer handling issues to pave the way for subsequent tools (Figure 2). Scaled–up versions of mature 300 mm platforms must be debugged as carriers, person–guided vehicles (PGVs), load ports, equipment front–end modules (EFEMs) and end effectors are integrated with them. G450C and suppliers are working diligently to resolve issues and optimize tools during development as essential early steps in the complex journey to production–worthy 450 mm metrology.
Figure 2. 450 mm metrology tools are the first to face and hopefully solve wafer handling problems.
Israeli Metro450 Consortium
Given the challenge of 450 mm development, some key observations can be made:
- Achieving acceptable metrology TPT is pivotal to enable 450 mm high–volume manufacturing (HVM) economies of scale. This is a key goal of the Metro450 consortium.
- Early 450 mm enabling programs are focused on establishing and debugging initial 450 mm metrology handling capabilities to enable process development and proof of capability, but are for now less concerned with metrology high–volume IC manufacturability.
- No single design element solves the complex equation for 450 mm metrology tool production–worthiness. Rather, it is the combination of and synergy between several key mechanical and operational aspects that pave the way to the required performance.
- A concerted, collaborative effort between stakeholders will drive the innovation and breakthrough solutions to support 450 mm HVM with production–worthy metrology platforms.
- Waiting on the sidelines for the 450 mm bandwagon to go by is not an option for metrology OEMs wishing to maintain/enhance their market position.
Stepping up to these challenges, four Israeli metrology equipment manufacturers are collaborating with four local universities and a major global IC manufacturer. Backed by the Israeli Office of Chief Scientist (OCS), they have established the Israel 450 mm Metrology Consortium (Metro450) and defined the following five themes as the key focus of its 3+ year program:
- Metro450 and G450C will work together to enhance 14 nm and beyond metrology tool performance specs.
- CNSE/G450C capabilities will be used to produce prototype "standard calibration wafers."
- G450C metrology module performance data (availability, defects, maintainability, costs) will be shared with Metro450 as input to annual work plan validation and optimization.
- G450C wafer damage information will be shared with Metro450, and Metro450 will share diagnostics and RCA with G450C, including use of metrology to predict/prevent.
Additional collaborations on 450 mm metrology–centric projects exist among G450C, Metro450 and organizations participating in European 450 mm programs. An example is global harmonization of on–wafer and ambient micro–contamination objectives and means to measure not only particles but molecular contamination in 450 mm tools, carriers and fabrication facilities.
Figure 3. Metro450 and G450C have established a collaboration scheme centered on several key activities.
There is a huge opportunity to accelerate 450 mm development with timely availability of measurement capability meeting medium–term equipment demonstration performance objectives. This early enablement must be complemented with longer–term focus on manufacturability and evolving technology needs. G450C, Metro450 and other global organizations will collaborate to achieve these goals in as cost–effective manner as possible as the industry transitions to 450 mm wafers. n
About the Authors
Rand Cottle is a College of Nanoscale Science and Engineering (CNSE) assignee to the Global 450mm Consortium (G450C) as manager of metrology. He received his Ph.D. in materials science and engineering in 2000 from the University of Texas at Austin, and has 12 years of R&D experience in the semiconductor industry in the fields of metrology, integrations and lithography.
David Nessim is a senior program manager with Intel's Technology Manufacturing Engineering group (TME–EMEA). For the past 18 years, he has held engineering, operations, construction, manufacturing and program management positions across Intel fabs in Israel, the United States and Ireland. He is currently engaged with various industrial, government and academic programs focused on emerging technologies. He represents Intel in the Israel 450mm Metro Consortium since it was established in early 2012.
Frank Robertson is vice president and general manager, industry interface and program strategy at the Global 450mm Consortium (G450C). He began work on the 450 mm transition in 2005 while managing external programs for Intel, assuming his current position as an assignee at the founding of G450C in 2012. Prior to joining Intel in 2000, Robertson had been chief operating officer of International SEMATECH and general manager of I300I after spending 20 years in fabs, development work and program management.
Menachem Shoval has an electronic engineering degree specializing in industrial control systems. After 10 years at various organizations in Israel, he joined Intel to start the first fab in Israel in 1984, and spent the past 28 years in various engineering, technological and managerial positions with the company. With the initiation of the Metro450 Consortium in Israel, Shoval left Intel to take leadership of the consortium's board. n