Download PDF Winter Bridge on Frontiers of Engineering December 15, 2022 Volume 52 Issue 4 From novel applications of microbes to DEI in engineering to the potential for hydrogen energy, Frontiers of Engineering participants tackle today’s challenging world issues. The winter issue of The Bridge showcases research by early-career engineers as shared at the 2022 US FOE symposium. Reasserting US Leadership in Microelectronics: The Role of Universities Tuesday, December 13, 2022 Author: MIT Microelectronics Group Reasserting US semiconductor leadership must leverage the resources and talent of universities in education, research, and startup formation. “US strength in semiconductor technology and fabrication is vital to US economic and national security interests.” – Congressional Research Service (2020) The United States’ longstanding leadership in semiconductors and microelectronics is under serious challenge. The solution: a concerted and ambitious national response that emphasizes manufacturing, research, innovation, and workforce development. In the ecosystem that has enabled US preeminence in microelectronics, universities play a significant role. In this paper we describe the current challenges and identify opportunities and needs for US universities in education and workforce development, research, technology translation, startups, intellectual property, academic infrastructure, and regional networks to support US semiconductors and microelectronics. Introduction Microelectronics underpins the information society. Extraordinary progress in health, communications, computation, energy, transportation, and many other areas stems from revolutionary advances in microelectronics technologies over the past 50 years. And US leadership in microelectronics has brought enormous economic progress to the nation and deterred adversaries. That commanding role, however, has eroded over time. Other countries are vigorously contesting US leadership in microelectronics, some at odds with US interests and values. As this country’s leading-edge semiconductor manufacturing capacity has dramatically dwindled, concerns have grown about vulnerable supply chains due to natural disasters, trade disputes, or military conflict. Examination of the entire microelectronics ecosystem reveals weaknesses and gaps that the US government is committed to address through legislation passed in August by Congress and signed by the president.[1] The terms “microelectronics” and “semiconductors” are often used as shorthand to refer to a variety of technologies involving multiple material systems, processes, and devices that perform various functions. Nanoscale, silicon-based CMOS (complementary metal-oxide semiconductor) logic technology is at the core of ubiquitous hardware technology. Equally strategic are memory technologies, signal processing, power electronics, communication chips, system integration technologies, sensors, and photonics, as well as the broader context of manufacturing equipment, advanced materials, packaging, circuit and system design and verification tools, and the large system integrator industry that aggregates everything for the end user. The ever-expanding diversity of materials, processes, and functions makes microelectronics a rich and rapidly changing field of surprises and unexpected opportunities. Traditional geometrical scaling of logic CMOS will remain central to virtually all applications despite a slowdown in performance gains and increasing costs with new technology generation. Innovative application-specific architectures and algorithms will significantly enhance performance, as already evidenced by data-intensive artificial intelligence applications that solve previously intractable problems. New material systems, devices, and integration technologies are opening unprecedented capabilities in communications, memory, computation, power management, and interfaces with the human body. Opportunities abound. Seizing them, however, is not straightforward. Hardware innovation is constrained by the lack of manufacturing system proximity to those doing the innovating. To ensure long-term leadership, US manufacturing of strategic semiconductor and packaging technologies must be prioritized and university activities have to get closer to it. “Without scaling [to volume manufacturing], we don’t just lose jobs—we lose our hold on new technologies. Losing the ability to scale will ultimately damage our capacity to innovate.” – Andy Grove (2010) Universities in the Microelectronics Ecosystem US universities, colleges, and community colleges contribute nearly the entire workforce of the nation’s microelectronics ecosystem. Universities also generate most of the fundamental research that identifies early opportunities and showstoppers. It is in university labs that the application potential of a new technology is often recognized first, and universities often spawn the new companies that bring pioneering concepts to the world. Most major innovation hubs around the world are near university campuses. US semiconductor manufacturing capacity has dwindled and is hampered by vulnerable supply chains due to natural disasters, trade disputes, and military conflict. In the extraordinarily fast-moving field of microelectronics, US preeminence is challenged by aging university facilities and inadequate resources. Societal changes are also a factor as interest in “hard tech” among US students wanes[2] and, as explained below, the appeal of microelectronics eludes students. Education and Workforce Development An educated, motivated, and diverse workforce is essential for any industry to thrive. To ensure US leadership in microelectronics, a dramatic expansion of the size and diversity of the microelectronics workforce is imperative. There is no more strategic convergence of university, industry, and government interests than the education of the next generation of technicians, engineers, scientists, and technical leaders in microelectronics. US graduate and postdoctoral programs attract the best talent from all over the world. Most of this talent remains in the United States and joins the university ranks or goes to work in industry or national labs. More than in other countries, US educational programs combine hands-on learning—-involving project-based experiences, design exercises, and research projects—with a well-balanced grounding in fundamentals. They also offer industry internships at the graduate and undergraduate levels, for students to acquire practical skills, learn about career prospects, and contribute toward college costs. Still, in the words of industry insiders, “the US educational system is failing to produce a sufficient number of American workers and students with the necessary STEM expertise to meet the needs of the semiconductor industry” (SIA 2019, p. 13). A number of factors explain students’ declining interest in microelectronics-related disciplines: lack of awareness of how microelectronics can help address the world’s most pressing problems (a motivation for undergraduates), the perception that this is a mature industry with little excitement ahead, lack of “physicality” (chips are hidden and of dimensions much smaller than human scale), and lack of awareness of fulfilling careers at the end of a demanding course of study. This is a systemic failure that requires concerted collective action to correct. What’s needed are systems-oriented multidisciplinary subjects, hands-on lab courses (figure 1), research experiences, design exercises using modern computer-aided design (CAD) tools, well-crafted internship programs in industry (often not available to freshmen just as they are about to select a major course of study), and support from industry mentors to attract students. Research on pedagogy should explore new teaching methods that shorten the learning curve and facilitate the technology access needed to fulfill project requirements and internship experiences. Implementing these initiatives will require sizable investments in research and educational facilities and in staff support. Changes must not, of course, detract from teaching the fundamentals—more important than ever in these rapidly evolving disciplines. A national microelectronics workforce development initiative must seek not just to expand the pool of qualified graduates in relevant disciplines but also to dramatically enrich its diversity in every dimension. Scale-up of existing programs will not accomplish this. The involvement of underserved and other educational institutions that for too long have been on the sidelines of the microelectronics enterprise is imperative. Universities should open their facilities and share their resources and know-how with colleges and community colleges, and also support the creation of educational programs, hands-on and research experiences, and internship opportunities for their students. Outreach efforts to middle schools, high schools, and community colleges must expand and deepen their reach. Many opportunities exist for economies of scale if industry and academia coordinate activities to develop and share resources and best practices. Universities also should play a role in supporting the continuing education needs of the microelectronics industry workforce, as new materials, technologies, processes, and techniques emerge all the time. Universities originate many of these innovations and are in a privileged position to prepare the existing workforce to use them. Advances in online pedagogy make it feasible to create and share educational materials on a national scale. Research Curiosity-driven, single-investigator research is central to the modern university and the foundation upon which most innovative technologies are built. Multi-disciplinary, vertically integrated, collaborative research with industry participation brings into focus promising technologies and facilitates commercialization. Partner-ships among industry, universities, and national labs are routinely assembled by mission-oriented agencies for projects relevant to national security; for example, a collaboration with Analog Devices, under DARPA sponsorship, has enabled fabrication at MIT of a carbon nanotube microprocessor (figure 2). In microelectronics, fundamental research in advanced lithography, strain engineering, scaled -transistors, wide-bandgap semiconductors, THz devices, MEMS, 2D materials and devices, circuits and systems, and AI hardware, among many examples, has fueled technological innovations with tremendous economic significance. US universities have contributed to this expensive enterprise by pooling resources and creating and managing shared facilities that can support fabrication processes and materials. But a chasm is growing between university facilities and the state-of-the-art tools and processes used in industry. Not only is the maximum wafer diameter that university facilities can handle in multistep fabrication mismatched with industry (at best, 150 mm vs. 300 mm; MIT Microelectronics Group 2021) but the performance, productivity, and reliability of university tools is in decline. This greatly limits competitiveness, university collaborations with industry and national labs, and technology translation. Problems are compounded by unaffordable equipment service plans and inadequate technical staff support. Moreover, many research grants do not cover the true cost of research that requires large integrated facilities with multistep semiconductor fabrication processes. Faculty, in their role as facility administrators, must devote substantial efforts to raising additional resources within or outside the university to make ends meet. A culture of scarcity permeates the operation. The United States urgently needs a national plan of sustained investment in both human and capital infrastructure. Investments are required to keep facilities relevant, for example with the establishment of 200 mm wafer diameter capabilities, the “sweet spot” for collaborations with industry and national labs and for technology translation (MIT Microelectronics Group 2021, appendix A). Also needed are mechanisms that provide stable support for equipment service plans and technical staff. A national coordination body should be established to provide users—not just at research universities but also at colleges, community colleges, startups, corporations, and national labs—with access to university facilities as well as unique resources such as a national 300 mm R&D center. It’s worth noting that the Semiconductor Industry Association (SIA 2019) calls for the creation of a National Semiconductor Technology Center and this recommendation made it into the recent legislation cited above. The performance, productivity, and reliability of university tools is in decline, limiting university collaborations with industry and national labs as well as technology translation. Research programs need to be expanded and their costs fully covered. A healthy mix of single-investigator grants, multidisciplinary vertically integrated programs, and collaborative university/industry/national lab initiatives over a broad intellectual front in a competitive framework is needed to support a diverse community of researchers and students. Technology Translation, Startups, and Intellectual Property Many effective technology transfer avenues from universities to industry exist. Companies that sponsor university research programs enjoy early and privileged exposure to research results through periodic updates, progress reports, formal project reviews, and informal interactions. But often the products of university research do not initially reveal their ultimate commercial value. This makes it difficult for companies to decide to license university intellectual property (IP) soon after it is conceived. For microelectronics hardware, the typical time for an invention to reach the marketplace is 10 years, as significant technology maturation is needed for the value of a new concept to become apparent. Microelectronics technology maturation requires a toolset, a baseline of established process modules, functional block designs, and strict execution protocols that reflect the manufacturing environment. Shared university facilities generally cannot accommodate these needs. Instead, an effective path for translation of new university technologies is through partnerships with prototyping facilities, national labs, and commercial foundries. These entities embody the rigor of a manufacturing environment, with enough flexibility to embrace new disruptive technologies. Fostering prototyping facilities and subsidizing industry engagement with universities to promote technology maturation should be a high priority in a national microelectronics program. University-generated tech startups also can have considerable impact. Innovation ecosystems around US university campuses attract venture capital, research labs in well-established companies, and startup incubators. Fostering the formation and growth of startups should be among the core goals of a comprehensive national microelectronics strategy. The US innovation and commercialization record is impressive, but obstacles include the high costs associated with development of microelectronics technologies and access to fabrication facilities. Startup activities could be fostered with subsidized access to university facilities (when compatible with the university’s research and educational mission), broadening the user pool of shared facilities and thus lowering the costs and time involved for all players. The inventors of a technology are often the best entrepreneurs to transition their innovations to market. Incentives to engage in technology translation activities can be created through translational fellowship programs that support students and postdocs outside their regular research activities as they explore the commercialization of the technologies they have created. US universities can retain ownership of inventions created with federal funding under the Bayh-Dole Act. This legislation was established to foster an environment that stimulates and protects innovation and incentivizes its commercialization through commercial licensing. US universities grant licenses to their patented and copyrighted inventions to both established companies and startups if the licensee demonstrates the technical and financial capabilities to develop the early-stage technology into commercially successful products. Research contracts with industry generally include terms that create options for the sponsor to license the IP generated by the research in a nonexclusive or exclusive form in a field of use. An exclusive license in a field of use is a crucial asset for a startup, as it confers to it a higher valuation and increases the ability to attract capital. Recent research contracts with industry consortia have IP terms that severely limit the ability of universities to license technology in exclusivity, as required for starts up thrive. In effect, these terms disincentivize IP generation and prioritize existing companies at the expense of future companies. When mixing industry consortia and US government research funds, as is desirable in the launch of ambitious, multidisciplinary, multiuniversity research programs, IP terms are much more restrictive than those typical of US government contracts. The sheer size of these programs and the number of consortia players involved make IP negotiations highly unbalanced. To ensure a greater role for public-private partnerships in microelectronics research, a new compact is needed for microelectronics IP generation and protection in a university environment. An organization of representatives from government, industry, academia, and the venture capital community should be created to generate policies and provide oversight. Academic Infrastructure To support university education, research, and IP generation and translation in semiconductors and microelectronics, a robust university infrastructure is paramount—not only the facilities and tools but also the staff support structures that make everything hum. For advanced microelectronics research, universities need new 200 mm facilities that combine the performance, reliability, and reproducibility of commercial manufacturing tools with the flexibility to handle a variety of materials and sample sizes and shapes. These and smaller, more versatile research tools can be operated in an economically sound model if they are shared by investigators, educators, startups, companies, universities, colleges, community colleges, and national labs. IP terms that limit universities’ ability to license technology in exclusivity, as required for startups to thrive, disincentivize IP generation. Attention to university infrastructure should extend to facilities for metrology, CAD, system design and prototyping, testing and packaging, and access to integrated circuit shuttle runs. These capabilities are often sited in private labs or otherwise out of reach for students. Shared facilities should support these resources for the use and benefit of the community, and CAD licensing arrangements and necessary cybe-rinfrastructure should be put in place to allow flexible access. The human factor is as critical as buildings and instruments. Highly qualified, motivated technical staff are integral to successful operation. It is our experience that universities can create an attractive working milieu for hiring and retaining competent personnel even in a field rich in employment opportunities. But universities also face challenges of understaffing, scarce resources, and inadequate salaries. A toolset expansion and modernization program, as argued here, must come with a con-comitant increase in the technical staff ranks with support for service contracts by outside professional entities. Junior faculty play a singular role in university microelectronics activities. US universities hire junior faculty to rejuvenate the faculty ranks, acquire new ideas, and launch new and promising research programs that expand university offerings. A national microelectronics program should invest in the creation of faculty positions at US colleges and universities and provide flexible career-initiation grants for equipment and research support in the early years of a faculty career. In addition, a renewed partnership in microelectronics between industry and academia should recruit seasoned and experienced researchers from industry to participate in university education and research as visiting scientists, professors of practice, guest lecturers, and mentors. It is equally important to establish research sabbaticals for faculty and university research personnel at prototyping facilities and industry R&D laboratories. Regional Network Efficiencies The efficacy of the comprehensive and ambitious plan proposed here can be enhanced considerably by exploiting substantial regional network efficiencies. We see ample experience in our university community of highly effective, multidisciplinary, multiinstitution research programs that pool the capabilities and expertise of those best qualified, regardless of geography. Accomplishing the goals articulated here will involve the engagement of institutions—universities, colleges, community colleges, middle and high schools, science museums—that have not traditionally been part of the US microelectronics enterprise. Furthermore, smaller educational institutions with distinguished educational or research programs that are limited in scope and size could enlarge their involvement under the proposed initiative. Widely expanding the number of players, scaling up their activities, and engaging a highly diverse population of students is essential to accomplishing the workforce education goals of this plan. It is in this quest that regional network effects can be helpful. We envision a loose confederation of institutions that coordinate research, education, outreach, and internship activities at a regional scale. The notion of “region” will necessarily differ around the country, but might enable access within a 2- to 3-hour drive. Such proximity could support regular use of facilities and participation in programs as well as the development of daylong and multiday programs that cater to a geographically dispersed community. Programs should facilitate access to technical expertise and shared experimental and design facilities if regional institutions are to support existing and new research and educational programs. Joint research projects should be created to engage neighboring institutions. Funding—from industry and perhaps also the government (federal and state)—for visiting appointments, internships, and summer research experiences will greatly assist this mission. Educational facilities, content, and know-how can be effectively pooled at a regional scale through a mixed in-person/online approach, as demonstrated during the Covid-19 disruption. Similarly, outreach and industrial internship programs can be coordinated and expanded on a regional scale. Startup support and technology transition efforts also benefit from regional proximity by cataloguing regional resources and coordinating access protocols. Across all these dimensions, regional-level meetings, conferences (figure 3), informal get-togethers, career fairs, startup exchanges, educational competitions, and other networking events can contribute greatly to the whole. Conclusion Universities are central actors in the microelectronics enterprise. Reasserting US leadership in semi-conductors must leverage the resources and talent of universities in contributing fundamental understanding and innovative technologies, educating the next-generation workforce, and nucleating and nurturing startups that bring new, disruptive concepts to the marketplace. Achieving these goals will demand sound investments in upgrading the physical and human infrastructure of US universities and the launch of industry-academia collaborative programs. References Congressional Research Service. 2020. Semiconductors: US Industry, Global Competition, and Federal Policy, Oct 26. Washington. Grove A. 2010. How America can create jobs. Bloomberg Businessweek, Jul 1. MIT Microelectronics Group. 2021. Reasserting US Leader-ship in Microelectronics – A White Paper on the Role of Universities. Cambridge: Massachusetts Institute of Technology. Available at https://dspace.mit.edu/-handle/1721.1/139740. SIA [Semiconductor Industry Association]. 2019. Winning the Future: A Blueprint for Sustained US Leadership in Semiconductor Technology. Washington. Whalen J. 2022. Economic future of US depends on making engineering cool. Washington Post, Oct 23. [1] The CHIPS Act of 2022: Section-by-Section Summary, available at https://www.commerce.senate.gov/services/files/592E23A5- B56F-48AE-B4C1-493822686BCB. [2] This is shown in the trends captured in 1959–2020 data from the National Center for Education Statistics Digest of Education Statistics table 325.47 (https://nces.ed.gov/programs/digest/d21/tables/dt21_325. 47.asp). The point was also made in Whalen (2022). About the Author:The members of the MIT Microelectronics Group are Jesús del Alamo, Dimitri Antoniadis (NAE), Robert Atkins, Marc Baldo, Vladimir Bulovic´, Mark Gouker, Craig Keast, Hae-Seung Lee, William Oliver, Tomás Palacios, Max Shulaker, and Carl Thompson. This article is adapted from a white paper of the same title posted on the MIT Microelectronics Laboratory webpage (https://dspace.mit.edu/handle/1721.1/139740).