NSF Sponsored Workshop on Structured Design Methods for MEMS

Executive Summary

October 15, 1996

Over the last three decades, the microelectronics industry has undergone unprecedented growth and has had an enormous impact on the nation, in what is often called the "VLSI revolution". Research activities in both industry and academia have led to the rapid introduction of advanced semiconductor process technology, hierarchically structured design methodologies, automated design tools, simulation models and rapid prototyping techniques. One key to the rapid success of the VLSI development effort was the early definition (about 1970) of a clean digital interface that separated design efforts at increasingly high levels of abstraction from the growing complexities of the fabrication processes. This allowed the designer to focus on process-independent design tools and methodologies that are available to research and academic community for rapid prototyping of VLSI chips.

Though there have been many remarkable and revolutionary advances made in Microelectromechanical Systems (MEMS) design and fabrication during the past decade, the need for structured design methods remains. At present each new MEMS development is expensive ( 1M or more), and time consuming. One contributing factor is that there is not yet an equivalent to the Caltech Intermediate Form (CIF) or the other descriptive languages which are commonly used in VLSI design. MEMS fabrication processes have matured rapidly but they are still many and varied. The time is now ripe to define and develop structured design methods, and to take advantage of the still formative nature of much of the field. The experience of the VLSI research community of 25 years ago in evolving design methodologies and fabrication processes should provide useful guidance.

Several elements appear to contribute to the successes in developing structured design methods for VLSI:

The design of mechanical systems, more generally, appears to have none of these virtues, which largely explains the lack of structured design methods for mechanical design. However, MEMS is a hybrid of VLSI and mechanical systems, employing the materials and fabrication processes of the former, while utilizing many of the energy storage and transfer domains of the latter. This report concludes that while significant differences between digital VLSI design and MEMS design clearly exist, sufficient parallels also exist to strongly encourage research on structured design methods for MEMS.

Three areas currently in use in VLSI design have also been identified as common elements for structured MEMS design:

Each of these has played a crucial role in the development of design methodologies for VLSI, and building on those prior VLSI developments will form the basis for a "clean separation" between design and fabrication of MEMS, and will provide the greatest leverage from an investment in research resources. Advances in these three areas will provide the foundation for (semi-)automatic synthesis of MEMS, perhaps by compiling a schematic or language describing desired function into a set of masks and processing information to fabricate a device or system that will robustly exhibit the desired function.

While the challenges in developing structured design methods for MEMS that preserve the "clean separation" are significant, the benefits of such methods will greatly enhance advances in MEMS. Developing structured design methods for MEMS holds the promise to significantly reduce the costs and time to create new devices and systems, and increase the complexity and robustness of devices and systems that can be designed.

Previous: Attendees
Next: Preface
Return to NSF MEMS Design Workshop Index