STRATEGY FOR AMERICAN LEADERSHIP IN ADVANCED MANUFACTURING
– 15 –
have been the driving force behind the exponential increase in transistor density and, concurrently,
reduction in the power per transistor. The industry is now facing fundamental performance limitations
of the CMOS technology, diversification of the market beyond processors and memory, and intense
global competition. Therefore, there is a critical need to develop fundamental materials, devices, and
interconnect solutions to enable future computing and storage paradigms beyond conventional CMOS
semiconductors and the ubiquitous von Neumann computer architecture and classical information
processing/storage methods.
The technical priorities for this objective are semiconductor design tools and fabrication; and new
materials, devices, and architectures.
Semiconductor Design Tools and Fabrication. A significant barrier to innovation in the semi-
conductor and microelectronics industry is a lack of affordable access to design tools and fabrication
foundries for integrated circuits that use advanced semiconductor materials and processes. The
specialized equipment needed to design and produce multilayer structures with novel materials is
extremely costly to purchase and maintain and requires highly-trained operators. For these reasons,
such facilities do not represent prudent investment choices for individual design firms, as the machines
must be engaged in continuous production to justify their high cost.
The continuous production requirement limits the exploration of new fabrication materials and
designs, since the time required for a changeover to new processes interrupts high-value production.
In addition, introducing new materials into a conventional semiconductor foundry can contaminate an
entire facility, making it unsuitable for further production. There is a need to establish semiconductor
foundries and the associated design tools to give designers throughout the United States access to the
fabrication services needed to experiment with and commercialize circuit designs in advanced
semiconductor materials. Broad access to facilities for fabricating computing hardware from exotic
materials, insulators, and biological cells is needed to research, develop, and efficiently implement the
new computer architectures that will be used in future neural computers.
Prioritize investment in capabilities to ensure that new microelectronics technologies are retained
and manufactured domestically. Beginning with the prototyping stage, investigate ways to
provide agile manufacturing capabilities that allow the creation of new devices and testing of new
materials. Create models that give more access to design tools and domestic microelectronics
foundries.
New Materials, Devices, and Architectures. For over 50 years, much of the increased computing
capabilities, advances in communication, and improved standard of living around the globe has been
enabled by the exponential growth in electronics performance, captured in Moore’s law, which states
that computing performance will double every two years by doubling the number of transistors on a
chip. Since 2012, the ability to continue this growth has been challenged by theoretical limitations on
the smallest manufacturable feature size of transistors. The quest for continued performance gains
requires further development of new technologies, including 3D systems-on-chip integration, tunneling
field-effect transistors, spintronics, integrated photonics, integration of III-V compound
semiconductors with silicon-based devices, and quantum information systems. Board-level
technologies are also a priority, including 3D printing with integrated electronics, die-bonded flexible-
hybrid circuits, and roll-to-roll manufacturing. Research should also include precision sensing
(including time, space, gravity, and electromagnetism), health and asset monitoring sensors,
photovoltaics, and medical devices. Finally, investment in quantum computing must continue to be a
priority to maintain global leadership in future complex electronics design and computing capabilities.