The global financial crisis of 2007-2008 has forced a sharper focus on the manufacturing sector in the United States and other advanced economies.1 The bursting of financial bubbles and the spectacular collapse of major financial institutions has dispelled the once widely held view that the decline of U.S.-based manufacturing can be shrugged off in light of the advent of a dynamic and innovative financial services sector.2 The recession that followed this crisis hit U.S.based manufacturing hard—between December 2007 and December 2009, it lost 2 million jobs, or 17 percent of the total workforce, and manufacturing employment fell to its lowest level since March 1941.3 By mid-2013, roughly 12 million Americans worked in manufacturing, down from a peak of about 20 million in the
1 A Korean industry analyst observed in 2012 that “the global financial crisis made even advanced economies emphasize the importance of manufacturing. Korea must foster high-value-added manufacturing industries.” Chae Seung-byung, senior analyst, Samsung Research Institute, “Manufacturing Sector at Its Worst in Three Years,” The Korea Herald Online, December 9, 2012.
2 “Not so long ago, it was fashionable in Whitehall and Westminster to claim that [Britain’s] relative industrial decline didn’t matter. Then came the banking crisis, which showed what happened when you bet the lot on the City and the bet didn’t come off.” Aditya Chakrabortty, “David Cameron’s Talk About Reviving British Industry is Nonsense,” The Guardian, November 1, 2010, see James Dyson, Ingenious Britain: Making the UK the Leading High Tech Exporter in Europe, March 2010.
3 Megan M. Barker, “Manufacturing Employment Hard Hit During the 2007-09 Recession,” Monthly Labor Review, April 2011, p. 33. U.S. manufacturing output peaked in December 2007, bottoming out 19 percent below the 2007 average in December 2009. By early 2013, 5 years later, U.S. manufacturing output was still 4 percent below the 2007 average. Federal Reserve data reproduced in SCDigest Editorial Staff, “Supply Chain News: How Is US Manufacturing Doing Five Years After the Great Recession?” Supply Chain Digest, April 10, 2013.
late 1970s.4 Surveying these developments, the President’s Council of Advisors on Science and Technology (PCAST) warned in 2011 that the U.S. competitive position in manufacturing was declining relative to other high-wage countries such as Japan and Germany, and the United States was losing leadership in technology-intensive industries employing high-skilled workers.5
At both the federal and state levels, a significant aspect of the policy response to the “Great Recession” has been initiatives to bolster the U.S. manufacturing sector to reverse its long-term decline, foster onshore manufacturing operations, create new domestic jobs, and exploit the U.S. science and research base. In 2011, the Administration established the Advanced Manufacturing Partnership (AMP) to coordinate investments in manufacturing by governments, universities, and industry.6 In 2012, the Administration announced that it would create 15 manufacturing research institutes around the country to strengthen U.S. manufacturing infrastructure, the National Network for Manufacturing Innovation (NNMI).7 A number of new state and regional initiatives feature significant investments in local manufacturing innovation, infrastructure, and workforce development.8
U.S. Advantages in Manufacturing
Notwithstanding the erosion of many industries, the United States may be able once again to position itself to compete internationally in manufacturing.
4 See Figure 2 in Martin Neil Baily and Barry P. Bosworth, “US Manufacturing: Understanding Its Past and Its Potential Future,” The Journal of Economic Perspectives 28, no. 1 (2014): 3–25. See also, Pamela M. Prah, “Has U.S. Manufacturing’s Comeback Stalled?” USA Today, July 30, 2013.
5 PCAST, Report to the President on Ensuring American Leadership in Advanced Manufacturing (June 2011), i.
6 AMP is being coordinated through the National Institute of Standards and Technology (NIST). The President has defined the object of AMP as leveraging existing programs in manufacturing, augmented by $500 million in investments. “President Obama Launches Advanced Manufacturing Partnership,” White House Press Release, June 24, 2011.
7 The NNMI features collaboration between the private sector, universities, and government organizations at the federal, state, and local levels. The first research institute, the National Additive Manufacturing Institute, will be located in Youngstown, Ohio, and will specialize in additive manufacturing, a potentially paradigm-shifting manufacturing process based on 3-D printing. The process uses 3-D software to control machines that project various materials including resins, metal powders, and polymers to “print” 3-D in layers, with little if any waste of materials. “Obama Names Youngstown as Model for the New Tech,” Youngstown Vindicator, February 13, 2013; “Obama Push for 3-D Hub to Turn ‘Rust Belt’ City Into ‘Tech City,’” Columbus Examiner, August 17, 2012.
8 See National Research Council, Best Practices in State and Regional Innovation Initiatives (Washington, DC: The National Academies Press, 2013), 78–83, 115–116, 131–140, 149–161, 174–182. For a detailed case study of initiatives in New York State to promote innovation and manufacturing in semiconductors, see National Research Council, New York’s Nanotechnology Model, rapporteur C. Wessner (Washington, DC: The National Academies Press, 2013).
The United States has the world’s foremost system of research universities, which have a long tradition of working with industry to promote innovation in manufacturing.9 For many generations the United States has been a leader in the introduction and improvement of mechanized production methods, and U.S. manufacturers have responded to the recession, in part, through intensified investments in automated production.10 The advent of new techniques for domestic extraction of natural gas and oil is lowering the energy costs of U.S.-based manufacturers relative to their foreign competitors.11
A 2012 survey by Massachusetts Institute of Technology found that a number of factors were combining to favor “re-shoring” of manufacturing in the United States, citing U.S. advances in automation and manufacturing techniques, declines in domestic natural gas prices, and the advent of additive manufacturing and nanotechnology.12 A widely cited 2013 survey by the Boston Consulting Group (BCG) of 200 U.S. executives of large manufacturers found that 21 percent were either already relocating production to the United States or planning to do so within the next 2 years, and another 33 percent were considering or would do so in the near future, which represents sharp increases from 2012’s results of 10 and 27 percent, respectively.13 Based on these trends, the BCG forecasted that U.S. manufacturing would return $70 billion to $115 billion of export business to the United States by 2020.14 An international survey of global executives by
9 National Research Council, State and Regional Innovation Initiatives, 49–59; N. Rosenberg and R.R. Nelson, “American Universities and Technical Advance in Industry,” Research Policy 23 (1994): 323–348.
10 “Making a Comeback—Bluffs Manufacturer’s Revival in the Wake of the Recession Echoes Nationally,” Omaha World-Herald, June 21, 2013; “More Businesses are Returning to Made-in-USA Mentality,” Spokane Spokesman-Review, May 26, 2013. In 1854, an English machine tool maker, Joseph Whitworth, who had been appointed as a British commissioner to the New York Industrial Exhibition, presented a report to Parliament based on site visits to American manufacturers. Profoundly impressed with American methods and the workforce, he stated: “It is this [comparative labor scarcity] and this eager resort to machinery whenever it can be applied, to which, under the guidance of superior education and intelligence, the remarkable prosperity of the United States is mainly due. . . . The workmen hail with satisfaction all mechanical improvements, the importance and value of which, as releasing them from the drudgery of unskilled labor, they are enabled by education to understand and appreciate.” “English Views of American Manufacturers: Special Report of Mr. Joseph Whitworth,” Greenough’s American Polytechnic Journal of Science, Mechanic Arts and Engineering 3–4 (1854): 344–345.
11 Price Waterhouse Cooper, “Shale Gas: A Renaissance in U.S. Manufacturing? (2011), Accessed on April 14, 2014, <http.www.pwc.com/en_US/us/industrial-products/assets/pwc-shale-gas-as-us-manufacturing-renaissance.pdf>.
12 MIT, “U.S. Re-Shoring: A Turning Point,” MIT Forum for Supply Chain Innovation and Supply Chain Digest (Cambridge, MA: MIT, 2012); TD Economics “Onshoring, and the Rebirth of American Manufacturing,” October 15, 2012.
13 “U.S. Manufacturers ‘Relocating’ from China,” Financial Times, September 23, 2013.
14 “Overseas Jobs are Coming Home—S.C. Business,” Columbia, SC, The State, September 8, 2013.
the consultancy Stanton Chase cited concern about product and process quality as a key motivating factor:
[A] significant geographical separation of research & development and production is not always ideal . . . [I]t is increasingly seen as beneficial to have part of production close to R&D.15
Emblematic of this trend, in 2013 Motorola, then owned by Google, indicated that its new Moto-X smartphone would be manufactured at a Flextronics facility in Fort Worth, Texas—the only smartphone to date to be made in the United States.16 The same trends are observable in Europe. A 2012 survey of German manufacturers found that the share of German companies that offshored a part of production declined between 2003 and 2009, and a November 2013 survey found that in the preceding 12 months, 11 percent of UK companies brought production back home while only 4 percent moved production offshore.17
During the past half century, U.S. industry has performed particularly well in sectors where U.S. firms were the first movers in entirely new technologies, such as microelectronics, software and computer systems, and biotechnology. In a number of other technology fields in which U.S. researchers made cutting-edge discoveries, however, most if not all of the manufacturing activity and manufacturing jobs have come to be located outside of the United States.18 “Flexible electronics” is an emerging technology field with revolutionary potential that represents an opportunity to establish another new high-technology industry that would be a major source of domestic manufacturing income and jobs and that would address national needs in areas such as national security, energy, and sustainable growth.
15 Dieter Hagmann of Stanton Chase in “Reshoring Is an Issue for Europe Too,” Finanz & Wirtschaft, October 23, 2013.
16 “More Companies Making Commitments to Made in the USA,” Bucks County Courier Times, June 30, 2013. A 2013 study by the MIT Forum for Supply Chain Innovation cited a survey of companies that found that one-third of them were considering “reshoring” offshore manufacturing into the United States and that 15 percent were “definitively” planning to do so (ibid). On January 2014, Google announced that it would sell its Motorola business to Lenovo of China for about $2.9 billion. At this writing, the impact of this transaction, if any, on Motorola’s manufacturing arrangement with Flextronics is unclear. Lenovo issued a statement in January 2014 that “there are now no plans to change Motorola’s approach to manufacturing.” “What Lenovo’s deal could mean for American manufacturing,” Engadget, January 2014.
17 “Offshoring and Reshoring Trends: European Data,” The Operations Room, September 27, 2013; Bernhard Dachs et al., The Offshoring of Production Activities in European Manufacturing (Austrian Institute of Technology and Fraunhofer Institute for Systems and Innovation Research ISI, December 2012); “Reshoring Trend in the UK,” FDI—Foreign Direct Investment (UK), February 1, 2014; “The Mark of the Makers Seems to Have Got Going,” Coventry Telegraph, February 10, 2014.
18 U.S.-invented “lost technologies” produced mainly abroad include lithium-ion batteries, oxide ceramics, wafer steppers, solar cells, interactive electronic games, laptop computers and many types of industrial robots. PCAST, Ensuring American Leadership in Advanced Manufacturing, 4–5.
The Opportunity in Flexible Electronics
“Flexible electronics” refers to electronic devices that can be bent, folded, stretched, or conformed regardless of their material composition without losing functionality.19 Although conventional electronics such as integrated circuits or solar cells are typically built on thick inflexible substrates, “flexible electronics—built on substrates like plastic or metallic foil—can be folded, wrapped, rolled, and twisted with negligible effect on its electronic function.”20 (See Box 1-1.)
Flexible electronics have multiple potential applications that include bendable and reliable displays for smartphones that can be rolled up or folded; stretchable electronics “skin” and other medical devices that can monitor health; “smart fabrics” that can modify their characteristics in response to external stimuli; and flexible photovoltaic and lighting systems that can be adapted to irregular or curved surfaces. With respect to national defense, flexible electronics can support a variety of applications for equipment that is rugged, lightweight, waterproof, and versatile, including displays, batteries, communications, and physiological monitoring systems that may be embedded in soldiers’ uniforms. To the extent such devices can be produced through additive processes such as printing, utilizing roll-to-roll methods, another benefit will be a dramatic reduction in the production cost of many kinds of electronic devices, a significant development given the escalation in cost of semiconductor fabs to multiple billions of dollars. The energy savings and biodegradability likely to be associated with most flexible electronics technologies will make major contributions to sustainability.21 Although estimates vary as to the eventual market for flexible electronics products, the range for potential annual revenues by the early 2020s is pegged by informed observers at $75 billion to $190 billion.22
The United States is in a strong position to compete successfully in the global flexible electronics industry. The U.S. system of academic research universities remains the world’s best and enjoys a strong track record with respect to commercializing research.23 U.S. universities and government organizations are engaged in a broad range of research topics directly relevant to flexible electronics products and associated production processes. The U.S. microelectronics
19 Closely related fields are defined by materials composition (“organic electronics,” or electronic devices composed of organic as opposed to inorganic matter) and production process (“printed electronics,” or electronic devices manufactured through printing technologies). NorTech FlexMatters, “Developing a Roadmap for Northeast Ohio’s Flexible Electronics Sector.”
20 Muhammad A. Alam and Satish Kumar, “Definition of Flexible Electronics,” ed. Bharat Bhushan, Encyclopedia of Nanotechnology (Dordrecht, Springer Reference, 2014).
21 Chemical Sciences and Society Summit, Organic Electronics for a Better Tomorrow: Innovation, Accessibility, Sustainability, September 2012.
22 Adam Page and Smithers Pira, “Market for Organic and Printed Electronics,” in OE-A, Organic and Printed Electronics (2013), 34; IDTechEx, “Printed, Flexible and Organic Electronics Sees 15.3% CAGR,” Printed Electronics World, May 15, 2013.
23 Richard C. Atkinson and William A. Blanpied, “Research Universities: Core of the U.S. Science and Technology Systems,” Technology and Society 30 (2008): 30–48.
Flexible Electronics at the Frontier
According to a recent report by the National Academy of Engineering, “One of the frontier goals in electronics research is to transform conventional fabrication processes to meet the demands of soft, pliant, and often easily damaged surfaces. Research in new materials and patterning technologies has enabled flexible electronics that push the boundaries of how electronics are made and used toward the possibility of incorporating electronic control and power sources into any object.
Unlike conventional silicon electronics that are limited to rigid wafers, flexible electronic devices have been demonstrated on plastics, paper, fibers, and even biological tissues. These flexible devices enable a wide range of applications, in fields ranging from energy sustainability to smart sensor networks to bioelectronics. Some specific examples are energyefficient, stretchable lighting, lightweight photovoltaics, smartsensing wallpaper, and dissolvable electronic implants.
To make flexible electronics that are compatible with delicate surfaces, lowtemperature processing is required. This need has led to the development of materials such as organic conductors and semiconductors as well as advanced solutionbased techniques that enable lowtemperature processing. Thus flexible electronics not only enable novel applications but also promote the use of alternative manufacturing technologies, such as rolltoroll printing for electronics.”a
aYuehLin (Lynn) Loo and Tse Nga (Tina) Ng, “Flexible Electronics,” in National Academy of Engineering, Frontiers of Engineering: Reports on Leading-Edge Engineering from the 2013 Symposium (Washington, DC: The National Academies Press, 2014).
industry is the world leader not only in technology but also in manufacturing of the most advanced devices. Furthermore, the printing industry, which has relevant technologies and know-how with respect to manufacturing processes, is well established in the United States, although buffeted by the decline of book and newspaper readership.24 U.S.-based companies are global leaders in the manufacturing of a number of key materials with applications in flexible electronics including Corning (bendable glass), Universal Display Corporation (phosphorescent emitter materials), and DuPont Teijin (films with electronics applications). The U.S. Army and Air Force and other elements of the defense establishment have been supporting the development of flexible displays for military and dual-use applications for more than a decade, and other federal
24 IBIS World, Out of Print: The Industry Struggles as Printed Media Lose Consumers to Web, August 2012; “Gravure Makes Inroads in Printed Electronics,” Printed Electronics Now, February 2011; “Printed Electronics: Fexo and Screen,” Printed Electronics Now, February 2009; “A New Industry Shapes the Future of Printing,” Printed Electronics Now, December 2008.
agencies have provided R&D support. 25 Promising startups spun off from U.S. research universities are commercializing academic discoveries such as Cambrios (silver nanowires with touchscreen applications at MIT); MC10 (stretchable electronics at the University of Illinois), and Imprint Energy (bendable batteries at the University of California at Berkeley).26 U.S. companies have also pioneered the development of organic photovoltaics.
Despite these strengths, the United States faces major challenges in establishing a strong manufacturing base in this emerging field. Most R&D in flexible electronics conducted by major U.S. firms has been precompetitive. The technological hurdles to commercial applications in the form of new products and processes have proven sufficiently daunting worldwide that they have confounded numerous optimistic forecasts of industry growth.27 Repeated delays in commercialization of flexible electronics technologies attributable to materials and manufacturing challenges have affected large established electronics firms as well as startups and have occurred in all of the geographic regions examined in this study. These difficulties underscore the reality that widespread commercialization will require further advances in efficient manufacturing technology, materials, equipment, and processes.
Asian manufacturers have developed such a complete dominance in the area of conventional electronic displays that they are poised to dominate flexible displays with mainstream commercial applications in products such as smartphones, TVs, and e-readers.28 To date, a number of innovative U.S. companies developing
25 The 2010 World Technology Evaluation Center, Inc. (WTEC) study group that reported on European research programs in flexible electronics reported that “principal strengths of U.S. efforts, as perceived by European scientists [include] strong support from Federal agencies such as NSF, ONR, Department of Defense, Department of Energy, etc.” WTEC, European Research and Development in Hybrid Flexible Electronics (July 2010), xvi. Mariana Mazzucato, a professor of economics at the University of Sussex, observes in a 2013 book that “despite the perception of the U.S. as the epitome of private sector–led wealth creation, in reality it is the State that has been engaged on a massive scale in entrepreneurial risk taking to spur innovation.” She details four highly successful U.S. institutions including the Defense Advanced Research Projects Agency and the Small Business Innovation Research program. Mariana Mazzucato, The Entrepreneurial State: Debunking Public vs. Private Sector Myths (London, New York, and Delhi: Anthem Press, 2013), 72–86.
26 “A new battery that could revolutionize wearables,” Gigaom, January 8, 2013; “MC10 Reshaping Electronics,” Flexible Substrates, September 2012; “Honey I shrunk the technology—NU making a big difference in the world of nanotech,” Chicago Sun-Times, August 13, 2007.
27 “Printed Electronics—Many New Directions,” Printed Electronics World, February 21, 2011; Lawrence Gasorian, “Notable Developments in Flexible Glass,” Flexible Substrate, January 2013; Nick Colaneri, “Manufacturing Flexible Displays: The Challenges of Handling Plastic,” Solid State Technology, May 1, 2013.
28 HSBC Global Research, Flexible Display: Fantastic Plastic—A Shape Shifting Game Changer, April 2013.
new flexible electronics technologies have chosen to manufacture their designs in Asia or have been acquired outright by Asian firms.
Although the U.S. government is investing in R&D relevant to flexible electronics through a number of institutional channels, and several states are making significant commitments to R&D, cumulative government financial support in the United States falls short of government funding in this field in Europe and East Asia.
Perhaps most significantly, while the United States has developed an onshore supply chain to support flexible electronics manufacturing, no large U.S.-based champion has yet emerged that is prepared to engage in large-scale commercial manufacturing of products that integrate the various flexible and printed electronics technologies being developed by the U.S. research base. Some industry experts reckon that if only market forces act, then a champion never will emerge, particularly in the displays field, reflecting the demonstrated risks associated with investments in this sector.29
The Role of Public Investments
A 2011 study by PCAST identified flexible electronics as one of four “early-stage technologies that have transformative potential” that may likely not develop if left to market forces (“market failures”).30 Flexible electronics is a “General Purpose Technology” (GPT), that is, a technology capable of having a major impact in a broad number of sectors in manners similar to those that occurred with
29 DuPont’s Steven C. Freilich offers a perspective from the history of the display industry that underscores the market deterrents that inhibit investment in promising but uncertain new technology areas. “In 2005, four major technologies were jostling for market share as the size of display panels grew ever larger: the traditional cathode ray tube (CRT); plasma displays, that was pushing the CRT for dominance in the mid-sized displays; liquid crystal displays (LCDs), which dominated the high definition hand-held market and had suddenly solved problems in larger dimensions; and rear projection, which had dominated the larger sizes. Plasma was quickly pushed by LCDs out of the mid-sized market, which then pressured the rear projection displays, leaving the market to just two technologies instead of four. The investment of every one of these technologies amounts to billions of dollars per fabrication unit, and yet companies must be prepared to shift quickly to keep up with evolving technologies, consumer tastes and price changes. You may think that [the display competition] is over now, but up in the corner [of this chart] you find organic light emitting diodes, and in a couple of years we’re going to see this shift happening all over.” Steven C. Freilich, “DuPont Reflections on Photovoltaics,” in National Research Council, The Future of Photovoltaic Manufacturing in the United States (Washington, DC: The National Academies Press, 2011), 68.
30 PCAST observed with respect to flexible electronics that the R&D “needed to advance this technology is costly and complex, requiring facilities for prototyping and pilot-scale manufacture. Yet, the benefits are broadly applicable across industries and cannot be fully captured.” The other three sectors identified by PCAST were nanoscale carbon materials, next-generation optoelectronics, and nanotechnology-enabled medical diagnostic devices and therapeutics. PCAST, Report to the President, 19–20.
respect to the steam engine, microelectronics, and the Internet.31 However, as the academic analyses of GPTs have concluded, the private sector will not necessarily make the investments necessary to develop such technologies. Professor Vernon W. Ruttan, a development economist at the University of Minnesota, observed in 2006:
A second issue is whether the private sector can be relied upon as a source of major new general purpose technology development. The quick answer is that it cannot! Each of the general purpose technologies that I have reviewed has required at least several decades of public support to reach the threshold of military and commercial viability. Decision makers in the private sector seldom have access to the patient capital implied by a time horizon measured in decades rather than years.32
A recent study by Mariana Mazzucato of the University of Sussex cites considerable academic research and empirical evidence to the effect that large-scale and long-term government investment “has been the engine behind almost every GPT in the last century,” including the U.S. system of mass production, aviation technology, nuclear power, space technology, information technology, and the Internet.33
The question currently confronting the United States is whether industry, universities, and government can work together to realize the opportunity presented by flexible electronics in the face of significant technological hurdles and intense international competition. In particular, the question is whether U.S. research discoveries and technological advances in flexible electronics can be translated into significant onshore manufacturing capability and domestic manufacturing employment. The United States arguably has no peer in basic research in this field. However, the President’s National Science and Technology Council has
31 The term was coined by Timothy F. Bresnahan and Manuel Trajtenberg in 1995 and is widely used in analyses of innovation and its effects on economic development. Bresnahan and Trajtenberg, “General Purpose Technologies: Engines of Growth?” (NBER Working Paper No. w4148, 1995). GPTs are identified by three characteristics: (1) they are extremely pervasive and are used as inputs across wide areas of the economy; (2) they continually improve over time to the benefit of their users; and (3) they facilitate innovation. Elhanan Helpman and Manual Trajtenberg, “A Time to Sow and a Time to Reap: Growth Based on General Purpose Technologies” (NBER Working Paper No. 4854, September 1994).
32 Vernon W. Ruttan, “Is War Necessary for Economic Growth?” Clemons Lecture, Saint Johns University, Collegeville, Minnesota (October 9, 2006), 30.
33 Mariana Mazzucato, The Entrepreneurial State. Among others Mazzucato cites the work of Dr. Vernon W. Ruttan, who demonstrated that military and defense research, development, and procurement “have been major sources of technology development across a broad spectrum of industries that account for an important share of United States industrial production.” Ruttan, “Is War Necessary for Economic Growth?”
highlighted shortcomings in the U.S. system of applied research, which is the practical application of science—an area where many U.S. companies have excelled individually, but which remains an area of relative weakness in the overall landscape of U.S. manufacturing innovation.34
The Innovation Infrastructure
The challenge facing the United States in innovation is to create an infrastructure that recognizes and capitalizes on the changing nature of research where the boundaries between basic and applied research have become more blurred.35 Technological advances are now enabling a degree of creativity in applied research of the kind traditionally associated with basic research—a phenomenon that is being exploited systematically in intermediate research organizations such as Europe’s Fraunhofer institutes and IMEC and in leading companies such as DuPont, General Electric, Proctor and Gamble, and Intel. Dr. Kristina Johnson, then the U.S. Undersecretary of Energy, underscored this fundamental shift in a 2011 presentation at a National Academies symposium.36 The traditional view of scientific research, articulated by Vannevar Bush, she observed, was two-dimensional, with knowledge flowing in one direction from basic research to application, development, and commercialization.
But this model, to the extent it was ever realistic, was changed by the advent of transistors, computers and the Internet—the features of the information age. In the old model, scientists would think about how things work, and engineers
34 One factor underlying the U.S. position on applied research has been the downsizing and even disappearance in the country of large industrial laboratories, which have historically been centers of applied research and innovation aimed at commercial objectives. See Erich Bloch, “Seizing U.S. Research Strength,” Issues in Science and Technology, Summer 2003. Corporate laboratories have been affected, among other things, by financial restructuring. “General Electric donated the David Sarnoff Research Laboratory to SRI International, which General Electric acquired in its purchase of RCA. Regional telephone companies formed from the split of AT&T created Bell-core as a separate laboratory, and soon it was sold. Kodak acquired Sterling Drug and then sold. General Motors took over Hughes Aircraft, and DuPont acquired Conoco. General Electric aerospace merged into Martin Marietta, which in turn merged into Lockheed. Kodak transferred research from its central laboratory into operating divisions. Allied Signal, Armstrong World Industries, and W. R. Grace completely eliminated corporate support for central research.” Roli Varma, “Changing Research Cultures in U.S. Industry,” Science, Technology, and Human Values, Autumn 2000, 400.
35 It is beyond the scope of this study to survey the vast, evolving, and occasionally discordant theoretical literature on innovation and its management, exploitation, and diffusion. However, see generally John E. Ettlie, Managing Innovation: New Technology, New Products, and New Services in a Global Economy (Burlington, MA: Elsevier-Butterworth-Heinemann, 2006); Bengt-Ake Lundvall (ed.), National Systems of Innovation: Toward a Theory of Innovation and Interactive Learning (London: Anthem Press, 2010); Ruud E. Smits, Stefan Kuhlmann and Philip Shapira (eds.), The Theory and Practice of Innovation Policy: An International Research Handbook (Cheltenham and Northampton, MA: Edward Elgar, 2012).
36 Kristina Johnson, “The U.S. Department of Energy’s Perspective,” in National Research Council, Future of Photovoltaics Manufacturing, 150.
would make them work. The new availability of knowledge allowed the engineers to invent new things as well. They could design them more intelligently, using mathematical models. Computers brought everyone the same platform and tools, allowing not only engineers but social scientists to become more quantitative and take more analytic approaches to the deployment of technology. This is a fundamental change that has not been clearly recognized.37
Although some individual U.S. firms have been able to break down traditional walls between basic and applied research in a highly successful manner (see Box 1-2), as the MIT Task Force on Production and Innovation recently concluded, these firms are somewhat anomalous in the U.S. industrial landscape.38
The Missing Middle
But while individual U.S. companies have proven adept at translating scientific knowledge into useful products and processes, in contrast to countries such as Germany, the United States does not have a nationwide infrastructure supporting translational innovation by companies—particularly small companies—broadly and systematically. In Germany, small companies can access sophisticated computer design and simulation tools and platforms, prototyping facilities, precision measuring and test equipment, and a deep bench of expertise at a dense network of Fraunhofer institutes and research centers bridging the space between the research base and industry. Equipment makers can test their machines on pilot lines in a factory environment.39 Although some of these capabilities are present
37 Ibid. A complementary perspective is provided by Michael Idelchik, GE’s Vice President for Advanced Technologies at GE Global Research, who told an MIT task force on innovation that today’s advanced industrial R&D involves “the management of concurrent, non-sequential interactions with multiple exchanges between scientists and engineers and manufacturing specialists and with the product passing back and forth between the hands of experts with fundamentally different competencies.” He cited the example of GE’s intermetallic turbine blades for the new GEnx engine. “The new alloy was patented in 1989, but the complexities of working with this material meant that it took until 1992 to be cast. The casting involved continuous interactions between [specialists in] manufacturing and R&D. There are patents in the blades, but most of the casting and materials depend on trade secrets. . . . The first engine test was in 1993—and this took collaboration between designers, R&D, and manufacturing.” Suzanne Berger, Making in America: From Innovation to Market (Cambridge, MA, and London: MIT Press, 2013), 58.
38 The MIT Task Force on Production and Innovation documented a number of highly innovative U.S. companies and industry groups that have scaled up and manufactured new technologies successfully. But it observed that these examples, while “promising,” were also “fortuitous.” “When we looked in detail at their origins, we found special individuals and companies who were willing to go far beyond their ordinary job descriptions and ways of doing business. The United States abounds in entrepreneurial people but without institutional bases, their efforts often remain isolated and hard to sustain. These positive but chancy interactions are unlikely to produce complementary capabilities on the scale needed to transform regional ecosystems across the United States.” Berger, Making in America, 211.
39 See Appendix A2, “Fraunhofer Gesellschaft: The German Model of Applied Research,” in National Research Council, 21st Century Manufacturing: The Role of the Manufacturing Extension Partnership (Washington, DC: The National Academies Press, 2013), 224–284.
Intel Corporation—Integrating Basic and Applied Research
One of the most successful microelectronics enterprises of all time, Intel Corporation not only reduced barriers between basic and applied research, but also was created as a spinoff of Fairchild with no central research laboratory whatsoever: [B]ecause of problems Gordon Moore and Bob Noyce experienced in transferring technology from Fairchild’s central research laboratory to their manufacturing lines, they established Intel without a separate R&D laboratory. They did this because competition is driven by time-to-market and technology transfer from a central laboratory is too slow to synchronize with the product cycle. In addition, the exactness of understanding that researchers wish to pursue is too rarely met in semiconductor development where process designers make empirical choices and try them to see if they work.a
Intel continued to operate on the Noyce principle of “minimum information,” that is, rather than trying to understand completely a given phenomenon, perhaps resulting in a published scientific article, Intel “tries to get by with as little information as possible. . . . Locating development and manufacturing together allows Intel to explore variations on its existing technologies very efficiently.” Semiconductor companies that “made great investments in longterm research at the expense of focus on the next product cycle have lost market share to Intel, a company that has kept the next product cycle in clear focus while annually investing $2 billion in R&D.”b This approach, which was an important factor in Intel’s success, was carried over into Sematech, an industrygovernment applied research consortium, with the result that the developmental pace of the U.S. semiconductor industry accelerated dramatically.c
aElias C. Carayannis and James Gover, “The SematechSandia National Laboratories Partnership: A Case Study,” Technovation, 2002, 586. The MIT Task Force on Production and Innovation, which interviewed numerous innovative companies, reported that “there was general consensus across the senior executives that bringing new ideas through early stages of prototyping, testing, demonstration, and pilot production works most efficiently when it is kept close to key scientists and engineers within the company. . . . Proximity is important not only for control and to avert disaster, but to accelerate them to market and to explore and develop multiple variants (and price points) of a new product.” Berger, Making in America, 57–58.
bCarayannis and Gover, “The SematechSandia National Laboratories Partnership,” 586. At the same time, there are limits to the extent to which applied research can be pushed forward on its own without parallel advances in basic science. For example, in a 2011 interview, Dr. Bernard Kippelen, Director of Georgia Tech’s Center for Organic Photonics and Electronics, cited significant recent strides achieved in improving the performance of organic thinfilm transistors (OTFTs) and the potential for further improvements. He said, “One of the biggest hurdles [to further progress] is the mismatch between the performance level reached by OTFTs and the lack of fundamental understanding of their properties. There is a need for further study as to develop predictive modelling capabilities, especially predictive control of morphology of organic semiconductors.” “A Look at Printed Electronics: Printed Electronics Now Interview with Dr. Bernard Kippelen,” Printed Electronics Now, July 2011.
c“Moore’s Law” is a rule of thumb articulated by Intel’s Gordon Moore that the number of transistors on integrated circuits will double approximately every 2 years, reflecting advances
in microelectronics technology. The existence of Moore’s Law and its acceptance by the semiconductor industry gave individual companies a method for benchmarking its own development. With the advent of Sematech, the industry as a whole was able to accelerate the pace of development through collective understandings on specific needs and the timing of their introduction. “What Moore’s Law has done is to give the research process a cadence. You try to get from one node, or minimum feature size, to the next as fast as possible. That has served to excite the industry to beat the roadmap, and they have done that. It wouldn’t have happened without that expectation or cadence that Moore’s Law provides.” Larry Sumney, “The Semiconductor Research Corporation (SRC): A Proven Means to Fund Relevant Research,” in National Research Council, Future of Photovoltaic Manufacturing, 186.
at individual sites in the United States, nothing comparable to the German system as a whole exists.40
The Administration’s National Network for Manufacturing Innovation is intended to address this gap, closely linking the U.S. research base to industry and the workforce. A recent government “Frequently Asked Questions” relating to NNMI framed the issue as follows:
Q: Don’t we already have many Federal programs that achieve the goals of the NNMI?
A: The simple answer is no. There are no current federal programs that have the required attributes to significantly influence the nation’s competitiveness and successfully bridge “the missing middle” in the manner and magnitude proposed for the NNMI. A whole-of-government approach is needed.41
40 The MIT Task Force on Production and Innovation observed in 2013 that “The American company stands alone. The German firm has access to a rich and diverse set of external resources . . . to which it contributes, but which it does not have to generate on its own. It hires employees who have been educated in technical schools and universities linked to hands-on experience and credentialing within companies. . . . For a price that seems acceptable even to small and medium businesses, the German firm can get expert advice and use expensive equipment at para-public institutions like the Fraunhofer Institutes.” Berger, Making in America, 87.
41 “National Network for Manufacturing Innovation Frequently Asked Questions,” accessed June 7, 2013, <http://manufacturing.gov/docs/nnmi_faa.pdf>. National Science and Technology Council, National Network for Manufacturing Innovation: A Preliminary Design (Washington, DC: Executive Office of the President, January 2013), 2. Under Secretary of Commerce for Standards and Technology Patrick Gallagher stated in 2012 congressional testimony that an interagency group comprised of the Departments of Commerce, Defense, and Energy under the National Science and Technology Council’s Committee on Technology had concluded, after study, that “the acceleration of innovation for advanced manufacturing required bridging a number of gaps in the present U.S. innovation system, particularly the gap between R&D activities and, the deployment of technological innovations in domestic manufacturing production.” Patrick D. Gallagher, “Assembling the Facts: Examining the Proposed National Network for Manufacturing Innovation,” May 31, 2012.
Although it is beyond the scope of this study to access the NNMI, it should be noted that the initiative draws on the example of Germany’s Fraunhofer-Gesellschaft, a public network of large-scale institutes of applied industrial research.42 As this study indicates, at least 10 Fraunhofer institutes are currently deploying major resources, including equipment, prototyping and simulation tools, demonstration lines, and their own deep competencies, to the development of processes and products across a broad spectrum of potential commercial applications of flexible electronics. There is nothing comparable in the United States, notwithstanding the existence of several noteworthy research centers at U.S. universities addressed in Chapter 7. The Fraunhofer’s methods have a track record of commercial success, reflected in the competitive prowess and reputation of German engineering and manufactured goods around the world and Germany’s relative success in retaining advanced manufacturing jobs at home.43 Whether or not NNMI replicates the Fraunhofer’s level of achievement, and depending on its ultimate areas of technology focus, the NNMI has the potential to help bridge the gap between basic research and commercialization in flexible electronics.
Modeling the Innovation Process
The National Institute of Standards and Technology (NIST) has developed a sophisticated schematic, breaking down the innovation process into basic components and identifying the points in that process where public sector investment can have a positive impact. Analysis by Gregory Tassey, formerly NIST’s chief economist, identifies the development of generic technology platforms (enabling technologies that can be shared by everyone) and “infratechnologies” as key areas in which public participation (augmented by private efforts) can make major contributions to innovation. (See Figure 1-1.)
Infratechnologies are tools which include “measurement methods for R&D and production control, technical support for interfaces between components of systems technologies, scientific and engineering databases, techniques such as quality assurance procedures, and test methods for facilitating marketplace transactions of complex technology-based products. They are ubiquitous in technology-based industries and often exert their impacts in the form of industry standards.”44 Infratechnologies support R&D, but they also play a critical role in the later developmental stages of production, system integration, and market development.
42 Congressional Research Service, The Obama Administration’s Proposal to Establish a National Network for Manufacturing Innovation (January 29, 2014), 4.
43 See, generally, National Research Council, 21st Century Manufacturing, 224–284.
44 Gregory Tassey, “Modeling and Measuring the Economic Roles of Technology Infrastructure” Economics of Innovation and New Technology (January 2008), p. 624. See also Gregory Tassey, “Technological Infrastructure and the Role of Government,” Working Paper Series, December 19, 1995; Gregory Tassey, The Technology Imperative (Cheltenham and Northhampton, MA: Edward Elgar, 2007).
FIGURE 1-1 Targets for science, technology, innovation and diffusion policy.
SOURCE: Figure 6, “Managing the entire technology life cycle: Policy roles in response to market failures” in Gregory Tassey, “Beyond the business cycle: The need for a technology-based growth strategy,” Science and Public Policy (2013) 40(3):293-315, by permission of Oxford University Press.
Tassey concludes that infratechnologies mitigate risks associated with commercialization of new products and thus facilitate entrepreneurial investment.45 NIST Director Patrick Gallagher observed in 2010 that Tassey’s model is a useful taxonomy of stages in the R&D process which enables a more efficient application of government support at the stages in the process where it can have the greatest effect.
[The model] goes from planning and production and goes to value add[ed] and then what you see is the development of technology in these lower boxes, which include entrepreneurial activity, risk reduction technology methodologies, which include standards development and regulators and other things that
45 Infratechnologies reduce risk, among other ways, by establishing industry norms, standards and benchmarks which ensure that a new product will “fix” or “interface” with other products in a given technology system. The development of such standards by Sematech—and industry-government consortium—with respect to semiconductor manufacturing equipment and materials was critically important to the vendor base which supplied semiconductor manufacturers. See Tassey, “Technological Infrastructure” (1995) op. cit. pp. 11–12; Tassey, “Modeling the Economic Roles of Technology Infrastructure” (2008) op. cit. pp. 620–622.
lower the risk of introducing a technology into the market. This is where the real technology innovation is occurring. . . . [Y]ou can divide the technologies into several pieces. We tend to think of the technology as the intellectually protected part of it, the IP part of technology. But in fact it’s also equally important to have technology infrastructure, manufacturing process technology is vital to productivity increases and, in fact, are essential for this system to work. And very often proprietary technologies are built on underlying generic technology platforms. . . . [F]rom a government side it means that we should really start focusing on these generic technology issues.46
The global flexible electronics industry is in its infancy. Scaled-up production for commercial applications exists in only a few areas (e-paper, RFID tags, OLED screens). Most of the government programs surveyed in this study are found in the early stages of development—basic research and the development of generic technology platforms (such as roll-to-roll manufacturing processes) and “infratechnologies.” While some government-supported R&D in flexible electronics may result in proprietary technology owned by individual firms (such as the German Fraunhofer’s contract research for industry), it is overwhelmingly concentrated on industry-government consortia developing generic technologies that can be used by multiple participating (and in some cases nonparticipating) industry partners. Scale-up incentives for production are far less common than in mature sectors like semiconductors and in some countries may be inhibited by rules (EU state aids requirements) or domestic opposition (United States) in the future. In China, where massive state financial support is being deployed to support scaling up of production of AMOLED displays,47 the principal state actors are municipal governments such as those of Shanghai, Ordos, Huizhou, and Beijing.48 Given the local boosterism that is intrinsic to many economies, including our own, the local character of government support for production that is evident in China may prove the norm in many of the countries surveyed in this report.
The international competitive prospects confronting any U.S. effort to establish onshore manufacturing capability in flexible electronics are sobering. East
46 Patrick Gallagher, “Strengthening the Connections: Research Innovation and Economic Growth,” University of Pittsburgh, October 7, 2010. Tassey’s criticism of neoclassical economics, and the alleged indifference of conventional economic thinking to the offshoring of manufacturing and the importance of industrial technology, has drawn sharp criticism, not so much because his model itself is flawed but because of the importance he places on the role of government in the innovation process. See Claude Barfield, “Commentary on Gregory Tassey’s ‘Rationales and Mechanisms for Revitalizing U.S. Manufacturing R&D Strategies,’” Journal of Technology Transfer (2010), vol. 35, issue 3.
47 AMOLED (active-matrix organic light-emitting diode) is a display technology for use in mobile devices and televisions.
48 See Chapter 6.
Asian firms dominate the manufacture of conventional displays and are using their installed manufacturing base in that field to leverage their entry into flexible displays with consumer applications. They are backed by government programs that emphasize applied research in industry and government research institutes, some of which are among the best of their kind in the world. Large Asian industrial groups enjoy not only ample financial resources but also extensive industrial and technological competencies in relevant fields such as microelectronics, optoelectronics, materials science, and printing. Particularly in Korea and Taiwan, company management has proven adept at making bold, rapid moves into risky but promising new fields.
East Asian efforts in flexible electronics are heavily concentrated in the area of displays for consumer use, holding out the prospect that a far broader U.S. approach, based on applications in areas such as medical devices, photovoltaics, lighting, and smart textiles, could be successful. However, a European effort based on just such an approach is already well under way, with a growing emphasis on establishing indigenous manufacturing capability in what in European parlance is termed “OLAE” (organic and large area electronics). Europe enjoys not only a strong fundamental research base but also a formidable infrastructure for applied research in relevant technology areas, which includes Germany’s renowned Fraunhofer institutes, a new group of research centers in the United Kingdom, and world-class institutes in the Low Countries (IMEC and its affiliate, the Holst Centre) and Finland (VTT). The European developmental effort is broad in both a geographic and technological sense, is supported by successive layers of government at the community, national, regional, and local levels, and engages companies with a long tradition of collaboration to achieve technological objectives.
Collaborative applied research efforts involving companies, government, and universities may present advantages in fields where technological and commercial risks are substantial and where no single company, no matter how large and well endowed, can command the resources and full range of technologies needed to successfully commercialize new technologies. There is a growing consensus that research partnerships, whether formal or informal, represent an organizational form that can encourage collaboration by reducing technological risks to participants, reducing time to market, fostering cost savings, and improving appropriability of R&D results. (See Box 1-3.)
Most current U.S. development initiatives in flexible electronics are not connected or mutually supportive. In contrast, the European Union and its Member States have overwhelmingly opted to pursue the development of flexible electronics through a variety of collaborative organizational structures—despite differences in language, nationality, and innovation culture. This is indicative
Economic Foundations for Applied Research Partnerships
Recent reviews of the academic and business literature as well as a growing body of empirical evidence affirm that research joint ventures that include firms, universities, and government organizations are effective.a In their review of this literature, Combs and Link (2003) argue that “a rich theoretical foundation has emerged upon which one can, after the fact, justify the historic and current policy interest in research partnerships.” They find that wellstructured research partnerships and the collaborations they create improve cost efficiencies, increase competition in the marketplace, and improve consumer surplus through improved products.b In their survey, Caloghirou, Ioannides, and Vonortas (2003) note further that “interfirm cooperative agreements to create and disseminate technological knowledge are now viewed as veritable competitive mechanisms, right at the strategic core of most companies in high technology industries.”
The analytical literature documents a variety of incentives for firms to join a research joint venture that includes R&D costsharing, pooling of risks and reducing uncertainties, internalizing knowledge spillovers, ensuring the continuity of R&D efforts through access to finance, complementary resources and skills, and advancing R&D through capturing research synergies and the effective deployment of extant resources. Joint venture participants also find that research joint ventures can improve their strategic flexibility, promote shared technical standards, and increase the market power of the industry.c Indeed, Cologhirou and colleagues make the case that “it would not be unreasonable to argue that there is some evidence in support of each of these arguments.”
aSee Yannis Caloghirou, Stavros Ioannides, and Nicholas S. Vonortas, “Research Joint Ventures,” Journal of Economic Surveys 17, no. 4 (2003): 541–570. Kathryn L. Combs and Albert N. Link, “Innovation Policy in Search of an Economic Foundation: The Case of Research Partnerships in the United States,” Technology Analysis and Strategic Management 15 (2003): 177–187. Albert N. Link and John T. Scott, “Public/Private Partnerships: Stimulating Competition in a Dynamic Market,” International Journal of Industrial Organization, 19, no. 5 (2001): 763–794. See also Albert N. Link and Donald S. Siegel, Innovation, Entrepreneurship, and Technological Change (New York: Oxford University Press, 2007). Stephen Martin and John T. Scott, “The Nature of Innovation Market Failure and the Design of Public Support for Private Innovation,” Research Policy 29, no. 4–5 (2000): 437–448.
bEffective management and organization contribute to the success of research consortia. See Lee G. Branstetter and Mariko Sakakibara, “When Do Research Consortia Work Well and Why? Evidence from Japanese Panel Data” (NBER Working Paper No. 7972, October 2000). See also Sarah M. Greene, Gene Hart, and Edward H. Wagner, “Measuring and Improving Performance in Multicenter Research Consortia,” JNCI Monographs 2005, no. 35: 26–32.
cIn spite of the fact that much is known about how to effectively manage consortia in the United States, there remains much unevenness in putting this knowledge into practice. Early studies evaluating the National Science Foundation cooperative R&D centers concluded that R&D consortia are a good idea with payoffs in the range of 3 to 1, similar to the National Labs payback. See Barry Bozeman, “Technology Transfer and Public Policy: A Review of Research and Theory,” Research Policy 29, no. 4–5 (April 2000): 627–655. Studies that followed, taking a more systematic, often quasiexperimental design approach to evaluation found a greater variation in outcomes. For example, some later studies found that the most important benefits
of university consortia and cooperative R&D centers were indirect. See Eliezer Geisler, “Technology Transfer: Toward Mapping the Field, a Review, and Research Directions,” The Journal of Technology Transfer 18, no. 3–4 (Summer–Fall 1993): 88–93. Firm benefits were often paid back in human capital: participating firms got to work and recruit the best university technical students as part of project work. There continue to be good examples. See Jonathan A. Morell, “Why Are There Unintended Consequences of Program Action, and What Are the Implications for Doing Evaluation?” American Journal of Evaluation 26, no. 4 (December 2005): 444–463.
that they see that the advantages associated with such structures are real and applicable in this sector.49
Can a collaborative effort work in the United States? Dr. Malcolm Thompson, former Chairman and CEO of RPO, Inc., a developer of touch screens, outlined the advantages for the United States of a consortium approach in flexible electronics in a 2010 National Academies conference.50 He pointed out that in light of the fact that addressing the relevant process, materials, and equipment challenges was beyond the ability of even a very large company, a collaboration would reach across the relevant topical areas, generate precompetitive research results capable of broad adoption, and minimize duplication of cost and effort. He summarized the alignment of interests and challenges confronting industry, government, and academia that favored a collective effort. (See Table 1-1.)
Many industry and government leaders closely associated with the development of flexible electronics technologies advocate the establishment of a “Sematech-style” industry-government manufacturing consortium to facilitate the development of flexible electronics manufacturing competencies and the necessary tools and materials.51 That fact, and the reality that few other comparable
49 “After ten years of sustained efforts, Europe has now managed to create a leading position in R&D on OLAE (organic and large area electronics), mainly thanks to R&D collaboration efforts between research institutes, SMEs and large companies established through the support of public funding.” The FP7-ICT Coordination Action OPERA and EM Commission DG INFSO Unit G5 “Photonics,” An Overview of OLAE Innovation Clusters and Competence Centres (September 2011), 4.
50 Malcolm J. Thompson, “A Consortium in Flexible Electronics for Security, Manufacturing and Growth in the US,” in National Research Council, Flexible Electronics for Security, Manufacturing, and Growth in the United States, rapporteur S. Shivakumar (Washington, DC: The National Academies Press, 2013).
51 In 2009, a blue-ribbon panel of experts sponsored by the National Science Foundation and the Office of Naval Research, which conducted a thorough study of European R&D in flexible electronics, recommended that the United States establish a Sematech-like organization for hybrid
TABLE 1-1 Factors Favoring a U.S. Consortium in Flexible Electronics
U.S. analogues are readily apparent, suggests that a closer assessment of the relevance of the Sematech experience to the challenges facing the United States in flexible electronics is warranted.
Sematech is widely acknowledged to have played a major role in the U.S. semiconductor industry’s market resurgence and recapture of global leadership in the 1990s.52 Numerous subsequent U.S. and foreign R&D initiatives have used Sematech as a model and point of reference; indeed, the very name “Sematech” has taken on an iconic cast not necessarily coextensive with the actual experience of the Sematech consortium.53
flexible electronics for precompetitive research, involving multiple companies and universities. Ananth Dodabalapur, “Panel on Flexible Hybrid Electronics,” WTEC Workshop: International Assessment of R&D in Flexible Hybrid Electronics (June 30, 2009). The U.S. Army’s launch of the Flexible Electronics and Display Center at Arizona State University in 2004 was “based on a scaled-down version of the Sematech model.” “Army to Create a ‘Sematech’-Like Consortium to Develop U.S. Flexible Display Industry,” Manufacturing & Technology News, May 2, 2003.
52 Kenneth Flamm and Qifei Wang, “Sematech Revisited; Assessing Consortium Impacts on Semiconductor R&D,” in National Research Council, Securing the Future: Regional and National Programs to Support the Semiconductor Industry (Washington, DC: The National Academies Press, 2003).
53 “Sematech has become a model for how industry and government can work together to restore manufacturing industries—or help jump-start new ones. The National Alliance for Advanced Transportation Battery Cell Manufacture, formed in 2008, was designed on the model of Sematech, for instance. So is the Department of Energy’s new SunShot Initiative, which aims to reduce the cost of solar energy by 75 percent by 2020.” “Lessons from Sematech,” MIT Technology Review, July 25, 2011. The National Academies’ Steering Committee for Government-Industry Partnerships made a comparable observation in 2003 in assessing the applicability of the Sematech model to an initiative to develop U.S. industrial capability in solid-state lighting. “[F]irms in the semiconductor industry
In fact the “Sematech model” per se does not appear to represent a ready-made solution for the challenges confronting the nascent U.S. flexible electronics industry. Sematech was a consortium embracing large players in an established industry (i.e., IBM, Intel, AT&T, Texas Instruments) pursuing research themes in an agreed technology area (i.e., complementary metal oxide semiconductor or CMOS) pursuant to certain commonly accepted principles (such as Moore’s Law). Moreover, Sematech was a national effort arising out of a perceived national emergency affecting U.S. defense readiness in the Cold War.54 None of these elements is present in flexible electronics.
Moreover, although it is often overlooked, the initial implementation of Sematech itself—despite the factors cited above—proved so difficult that the application of the Sematech model remains a challenge for even well-established industries, to say nothing of newer industries such as flexible electronics.55 Cultural differences between companies exerted powerful centrifugal force at Sematech, and it is difficult to envision how they would not do so in future flexible electronics consortia of even more diverse members.56 A number of original Sematech members pulled out when the Sematech research agenda did not match their own objectives.57 Similar problems, writ large, are likely to confront consortia formed to implement a Sematech-style mission in flexible electronics, particularly given the diversity of potential technology paths.
At the same time, advocates of collaboration in flexible electronics point to the role that intermediary organizations can play in bridging the cultural factors that divide relevant industry segments. They point out that this new field will require industries that have seldom, if ever, worked together, such as printing and
appear to have had the advantage of a much clearer research path than that being confronted by the solid-state lighting industry today.” National Research Council, Government-Industry Partnerships for the Development of New Technologies (Washington, DC: The National Academies Press, 2003), 86n.
54 Defense Science Board, Task Force on Semiconductor Dependency, November 30, 1986.
55 “The Sematech semiconductor equipment consortium represents another model for successful public–private cooperation in technology development. But it has not been replicated in other industries.” Vernon W. Ruttan, “Is War Necessary for Economic Growth?” September 26, 2006, 24.
56 At Sematech “IBM employees, the ‘white shirts,’ were put off by the more informal Californians from Advanced Micro Devices. AT&T people were conspicuous for their secretiveness and reluctance to speak up. People from Intel were too blunt.” Although these differences were gradually overcome, similar issues will pose a hurdle for prospective consortia. “Sematech Sets Model for Industry Consortium,” Fort Worth Star Telegram, November 10, 1993. A clash between Sematech’s first CEO, Bob Noyce, and COO, Paul Castrucci, arising out of different managing styles, led to the abrupt departure of the latter in 1989. Larry D. Browning and Judy C. Shetler, Sematech: Saving the U.S. Semiconductor Industry (College Station, TX: Texas A&M University Press, 2000), 86–99.
57 Withdrawals included Micron Technology, LSI Logic, and Harris Corporation.
electronics, to cooperate closely in the development of new industrial processes. They believe that R&D collaboration can make this happen, albeit not without difficulty, to facilitate relationships between companies and technological disciplines that “don’t necessarily talk to each other.”58
Despite the many differences that exist between the situation of the U.S. semiconductor industry in the 1980s and the flexible electronics industry of today, a number of aspects of the Sematech experience are relevant to the latter.59 The federal contribution to Sematech was matched by industry, whose top executives made a strong commitment to the consortium in terms of their own companies’ resources and talent. Sematech included not only semiconductor manufacturers, but also equipment and materials suppliers, “who help[ed] set and drive an agenda, [convening] a large network around individual problems and bring[ing] solutions more quickly.” Technology roadmaps and development of common manufacturing standards facilitated the development of necessary equipment and materials and reduced time to markets. Sematech also facilitated industry partnerships with universities, NIST, and the Department of Energy’s national laboratories.60
Although the Sematech model is arguably inapplicable in toto, many lessons can be derived from specific operational aspects of the Sematech experience—as well as the broader experience of the U.S. semiconductor industry—which are potentially instructive with respect to the value of collaboration in this new field.
Industry leadership. Most retrospective accounts of Sematech concur that a critical factor in its success was the readiness of the major players in the semiconductor industry to make very substantial, even painful, contributions to the success of the project. Company CEOs and other senior executives served on Sematech’s Board of Directors. “Technical advisory groups,” comprised of company representatives with specialized expertise, advised on specific thematic projects. Many semiconductor device companies sent their most talented engineers to the consortium, a burden that weighed heavily on smaller firms but an initiative that ensured that the assignees’ companies received relevant research results.61 IBM executives such as Sandy Kane, Obi Oberoi, and Paul Castrucci
58 Conference call, FlexTech Alliance, November 15, 2013.
59 One difference is that much of the semiconductor manufacturing was initially located in the United States, whereas many of the companies that are interested in flexible electronics, especially for consumer applications, are not necessarily interested in locating manufacturing in the United States.
60 Michael Polcari, President and CEO of Sematech, “The Sematech Model: Potential Applications for PV,” in National Research Council, Future of Photovoltaics Manufacturing, 182.
61 Sam Harrell, who served as an assignee to Sematech, recalls that “so if you were assigned in a lithography program, you worked in the lithography program. But you were also responsible to make sure that the lithography companies who were members got the information they needed. They got access, whether it’s technical meetings or other kinds of opportunities to get together with customer and so on. . . . Most of the [Sematech] employees were not in the technical mainstream. Most of the technical mainstream were assignees.” “Oral History of Sam Harrell,” November 16, 2004, Austin, Texas (Computer History Museum, 2011). Ted Malanczuk, Vice President of Wafer Fabrication and Technology at National Semiconductor, said in 1988 that the assignment of secondees to Sematech was “not an easy task. National will have to donate up to 25 of its best.” LSI Logic, one of two
played key roles in the launch of the consortium, and as Castrucci later recalled, IBM executive (later President) Jack Keuhler “was behind [Sematech] 100 percent. They were putting $50 million a year into it.”62 This industry-led feature of Sematech stands in contrast to the character of most European research consortia, which are government-led projects in which participating companies pursue themes of interest to public authorities.63
To date no comparable major corporate U.S. “champions” have emerged in flexible electronics, a fact that represents a significant stumbling block to the realization of a U.S. manufacturing presence. However, a number of world-class U.S. companies are active in precompetitive and upstream research, including Hewlett-Packard, DuPont, Eastman Kodak, Corning, 3M, Universal Display Corporation, and GE. The Palo Alto Research Corporation (PARC), a research arm of Xerox—to which are attributable many of the innovations underpinning the Internet era—is pursuing flexible electronics research themes.64 Although the critical mass of industry commitment, resources, and personnel to a joint developmental effort has not yet occurred, industry leadership in this field remains possible, particularly with supportive public policies.
Roadmapping. Technology roadmaps are evolving industry plans in areas of emerging technology that identify short-, medium- and longer-term technological objectives for the purpose of forming a rough consensus about technological requirements, the timing of needed solutions, and the future evolution of the technology. By their nature, roadmaps rely on forecasts about the market or consumer preferences, and such predictions can be difficult, especially for new markets. Participants in the symposia convened for this study have emphasized the importance of technology roadmapping in flexible electronics to stimulate the
Sematech members with less than $1 billion in annual sales, was reportedly “still wrestling with how to pull at least five key people from its operations to send to Sematech.” “Sematech Beckons Brightest Engineers Going to Promised Land,” San Jose Mercury News, March 14, 1988.
62 “Oral History of Paul Castrucci,” July 18, 2008, Mountain View, California (Computer History Museum, 2008), 24; “Join Hands in the Semiconductor Race,” Los Angeles Times, June 22, 1989. Charlie Sporck, then head of National Semiconductor and one of the founders of Sematech, observed in an oral history interview that he “went around and found tremendous support including IBM. IBM is not really a big joiner. But that led to the foundation of Sematech.” “Interview with Charlie Sporck,” February 21, 2000, <http://silicongenesis.stanford.edu/transcripts/sporck.htm>. IBM also contributed its extensive knowledge and experience gained from 3 years of operations in processing 200 mm (8 inch) diameter wafers at a time (late 1980s) when most U.S. semiconductor makers had not yet made the transition to 200 mm. AT&T contributed proprietary technology for its static random access memory (SRAM) to serve as a process vehicle as well as key process technology. Robert R. Schaller, Technological Innovation in the Semiconductor Industry: A Case Study of the International Technology Roadmap for Semiconductors (ITRS). (Ph.D. dissertation, George Mason University, 2004), 458; Browning and Shetler, Sematech, 57–59, 72–74.
63 Browning and Shetler, Sematech; National Research Council, Government-Industry Partnerships, 92.
64 Ross Bringans, Palo Alto Research Center, Inc., “Challenges and Opportunities for the Flexible Electronics Industry,” in National Research Council, Flexible Electronics for Security, Manufacturing, and Growth.
development of common industry standards, identify technological gaps, and enable equipment and materials suppliers to produce products for which real needs and applications do exist.65 The relevance of recent experience with roadmapping in semiconductors was acknowledged.
In the semiconductor industry, roadmaps have become essential because semiconductor manufacturing is virtually completely automated, utilizing complex, highly specialized machines and processes developed by multiple vendors that cannot readily be integrated without coordination between players and a clear understanding of what technologies will be required within a given timeframe. Although flexible electronics may ultimately involve manufacturing processes that are simpler relative to those involved in semiconductor fabrication, they will nevertheless also be highly automated, involve multiple technological actors and disciplines, and require close and sophisticated coordination (e.g., roadmapping) in order to succeed.
Technology roadmapping began within individual semiconductor companies in the late 1970s.66 Collective benchmarking by the industry first occurred in the context of the Defense Department’s Very High Speed Integrated Circuit (VHSIC) program, launched in 1980, which was an Army/Navy/Air Force effort to develop semiconductors with military applications exclusively based on substrates made of silicon, the principal technology then in use in the commercial semiconductor industry.67 The Semiconductor Research Corporation (SRC), a research consortium formed in 1983 by the Semiconductor Industry Association
65 Dan Gamota, “iNEMI Flexible Electronics Roadmap: From Concept to Product,” in National Research Council, Flexible Electronics for Security, Manufacturing, and Growth.
66 Bob Galvin, then the Chairman of Motorola, recalled that beginning in the mid-1970s he spent many hours in the company’s semiconductor laboratories, “dropping in on people, finding out what they’re working on.” He concluded that this front-line knowledge needed to be shared throughout the company in a systematic way. He told his executive team that “I want a roadmap of what we’re doing and where we are going,” giving them a simple chart (x and y axis) showing what Motorola was developing and resources needed, with a review in 6 months’ time. The managers incorporated other factors into the process, such as product life cycles and learning-curve economies, eventually producing a forecasting tool—the “roadmap”—which came to play a central role in Motorola’s strategic planning. Interviews and email exchanges with Bob Galvin and Bill Howard, Motorola, cited by Schaller, Technological Innovation in the Semiconductor Industry.
67 Previously defense-related microelectronics R&D efforts had focused on technologies based on substrate materials such as gallium arsenide and germanium, which had important defense-related advantages but limited commercial application. VHSIC established a significant government connection with industry. Larry Sumney, who served as program director of VHSIC, went on to serve as Executive Director of the Semiconductor Research Corporation (SRC), the first R&D consortium in the U.S. semiconductor industry, and later served as acting head of the nascent Sematech consortium. He observed that VHSIC program reviews by participating companies “were a public way for individual firms to gauge or benchmark themselves against each other in a form of open competition. Driven by definitive technology targets (i.e., ‘near micron’ 1.25 microns for Phase I, and ‘sub-micron’ 0.5 micron for Phase II, the VHSIC program contained some of the key factors that could be interpreted as a technology roadmap exercise, if not explicitly intended as such.” Schaller, Technological Innovation in the Semiconductor Industry, 427–428.
to support university-based R&D, originated the first industry-wide goal-setting technology “roadmap” in 1984-1985, a set of 10-year research targets to make possible next-generation dynamic random access memory (DRAM) devices, with the idea “to organize university research through a ‘roadmap’ versus the traditional ‘white space’ approach to academic inquiry.68
Sematech embraced and facilitated roadmapping from its inception, with planning actually predating the formal startup of the consortium in 1988 in industry-government-academic workshops convened in 1987 and 1988.69 Paul Castrucci, an IBM executive who served as Sematech’s first chief operating officer, took part in the workshops and described in a subsequent oral history interview how they quickly broke down boundaries between people representing different organizations.70 He took the Sematech job based on his experience in the workshops:
68 Ibid., 442–443. Larry Sumney observed in 2011 that SRC brings relevant sectors together with government in a manufacturing research ecosystem. The sectors lead by jointly identifying the most urgent R&D needs at the precompetitive level, and government provides incentives by co-funding research: “Given the diversity of participants, this ecosystem can be distributed but very coordinated. We see a flow of related ideas and technologies moving in both directions between industry and academia, with government playing a major role. Larry Sumney, “Semiconductor Research Corporation,” in National Research Council, Future of Photovoltaic Manufacturing, 188. An industry participant who took part in the SRC exercise later recalled that “from the very beginning there was general enthusiasm on talking about forecasting the future,” and the discussion was so productive that “consensus on the future was easily reached.” Schaller, Technological Innovation in the Semiconductor Industry, 444, citing telephone interview with Jim Daughton, Honeywell, August 4, 2000.
69 Sematech’s startup team conducted an intense series of industry-government planning workshops in 1987 and 1988, which produced the “Black Book,” a roadmap for achieving technological goals together with resources and business initiatives needed to achieve those goals. The Black Book set forth a 5-year plan for pushing technology levels ahead (Phase I, 0.8 micron design rules; Phase II, 0.5 micron; Phase III, 0.35 micron). The planning workshops primarily involved engineers and scientists from companies, the government, and SRC organized into subsets of 10-15 people according to specific technology topics with some overlap of related workshops. The workshops included “many of the industry’s best people” and among other things produced the “first real database of key technologists that ‘fanned and penetrated the armor’ of the semiconductor community.” Schaller, Technological Innovation in the Semiconductor Industry, 451–458.
70 At a 1988 Sematech workshop, “They had 60 people there, different backgrounds, national laboratories, device, people (tooling) people, and universities. And the idea of this two and a half day workshop was to define the 0.25 micron process tools and ground rules. Give the ground rules and the process to Motorola guys, have them design an SRAM in two and a half days. I say ‘this I got to see.” I couldn’t get the Army of Engineers to do that in two and a half months. So they broke them into groups of 20 each, that mix I talked about. It was run by a TI guy, I can’t remember his name, but they were worrying about the materials. Lou Parillo from Motorola had the job of the processing tools, I was in that group. And [Obi] Oberoi from IBM had the back end of the line [a mixture of] 20 people in each of those groups, mixture. So I’m in the process group, 20 of us. Lou’s up front with a flipchart. Nobody’s talking. We’ve all been told, ‘don’t talk when you go out.’ . . . I said, ‘Come on, this is not proprietary, let’s get going.’ So people started opening up. And Lou got down to maybe step five and then he said, ‘I really don’t know how to do this next step.’ The guy over in the corner says ‘Well I know how to do it. Da, da, da, da.” After 8 hours, we had a complete process description and all the tools. It was a combined intelligence of those 20 people, collaboration.” “Oral History of Paul Castrucci,” 23–24.
The secret of success in the future is going to be collaboration. But there are a lot of people who don’t understand that. But Sematech was based on collaboration. So I says ‘I’ll take that job, John. You know, I think I can do something with it.’ I saw what happened in that workshop.71
The Sematech workshops were instrumental in giving direction and focus to a vast and comprehensive investment program by the U.S. semiconductor industry and the U.S. government. The roadmapping process made it possible to identify certain potential “show stopping” technological problems in advance and to enable redeployment of resources to address those problems in a timely manner.72 Sam Harrell, one of the founders of Sematech, later recalled,
Those were working sessions which drove to some conclusion about the needs and requirement of the industry and what [were the] most likely alternatives to meet those needs and requirements. Those were very powerful interactions that had never been able to happen before. . . . Sematech’s [proposed] $100 million from the government and $100 million from industry was peanuts compared to what the industry spends on its own balance sheets. Suppliers alone spend $1.4 billion a year on RD&T [research, development, and testing]. The member companies spend $6 to $7 billion a year on RDT&E [evaluation] in a comparable basis. What the strategic workshop road maps did was to set in motion a bunch of focusing activities of $8 or $9 billion worth of effort, not just $200 million worth of effort.73
The emerging field of flexible electronics is already characterized by numerous roadmaps generated by industry groups, consulting organizations, and regional development agencies. However, no universally accepted roadmap yet exists for flexible electronics, and it is not clear that a single comprehensive roadmap is feasible or appropriate over the near term, given the multiplicity of potential manufacturing processes, substrates and conductive materials, encapsulation techniques, and end products. The semiconductor roadmaps of the 1980s and 1990s involved certain technological uncertainties, such as the future of lithography, but all reflected silicon-based CMOS technology and many generally accepted
71 Ibid., 24.
72 Intel’s Paulo Gargini, who participated in semiconductor industry roadmapping in the 1990s, recalls that “in the 1997 roadmap for the first time, for instance, I had the idea that by the middle of the next decade, 2004/5/6, we had to introduce high-k metal gate. And at the end of the meeting, especially the university people were terrorized, because they thought they were not going to think about it until 2010. At that point, this was in ’97 . . . in the next six years you have to be ready, so and it was really a terror for all of them all of a sudden, something that was really a low key project, all of a sudden was becoming very important. The benefit of it was that indeed, fortunately for all of us, many of the universities began working on it. . . .” “Oral History of Paolo Gargini,” July 27, 2011, Mountain View, California (Computer History Museum, 2011).
73 Cited in Browning and Shetler, Sematech, 42. Beginning in 1998, the U.S. semiconductor industry expanded the roadmap process to include foreign participants, and the evolution of global semiconductor technology is now assessed pursuant to the International Technology Roadmap for Semiconductors (ITRS).
International Electronics Manufacturing Initiative (iNEMI) Roadmaps
The International Electronics Manufacturing Initiative has been roadmapping the needs of the electronics industry since 1994. The iNEMI roadmap is modeled on the International Technology Roadmap for Semiconductors (ITRS) and includes 21 individual roadmap chapters in its most recent biannual edition. Five external organizations including the ITRS contribute to the chapters. Dan Gamota, a member of the WTEC task force that reported on European developmental efforts in flexible electronics in 2010, chairs the Printed and Organic Electronics Technical Working Group, which has been developing, publishing, and updating the iNEMI Roadmap for Printed and Organic Electronics in consultation with numerous relevant public and private entities since 2006. Dr. Gamota indicates that the iNEMI roadmap is intended to stimulate the development of industry standards and to identify “gaps and needs” in the industry supply chains, including potential “showstoppers.”
manufacturing processes. Despite that fact, the establishment of generally accepted roadmaps was a challenging exercise.74
In a recent discussion of the applicability of semiconductor-style roadmapping to the photovoltaic manufacturing industry, it was pointed out that the diversity of photovoltaics (PV) manufacturing themes in areas such as deposition methods, cell architectures, and materials (polycrystalline versus monocrystalline silicon) meant that a single semiconductor-like equipment technology roadmap “would not be appropriate for PV.”75 Full pursuit of the opportunities in this emerging field is likely to require multiple roadmaps, some of them quite divergent in technological terms with respect to materials and processes.76 The role
74 Intel’s Paolo Gargini, a participant in early semiconductor industry roadmapping exercises, recalls that the process was “dominated by many people who were highly theoretical, and I understood, after a while, that the motivation was to submit proposals to the government for funding, and one of the problems that came back from the government, the national labs, university and industry, would go and propose programs to the government, with completely different roadmaps. . . . “Oral History of Paolo Gargini.”
75 Doug Rose, Sunpower, “Observations on Building a PV Roadmap Model,” in National Research Council, Future of Photovoltaic Manufacturing, 198–201.
76 Dr. Steven Freilich of DuPont, speaking in 2011 of the opportunities presented by thin-film photovoltaic devices on flexible substrates, cited the experience of semiconductor roadmapping in the area of lithography as a precedent for “controlling a fast-moving technology and fast-moving markets.” The semiconductor roadmap explored five “technologies of note” between 2000 and 2003, one of which was dropped. The roadmap “was able to lay out the objectives for each technology, performance goals, milestones and timing. This is important to a materials supplier . . . because it clearly shows when a research program is not performing well.” Steven C. Freilich, “DuPont Reflections on Photovoltaics,” in National Research Council, Future of Photovoltaic Manufacturing, 69.
played by consortia in roadmapping in the semiconductor context, although not perfectly analogous, remains instructive.77
Supply chain integration. A rationale for one or more flexible electronics consortia is to “align the interests of end product manufacturers to tools and materials suppliers.”78 A critically important conclusion of Sematech’s early planning workshops was a unanimous view that “the chipmakers’ biggest headaches stemmed not only from technological problems but more from difficulties with suppliers” with a need to synchronize the “cadence” of development of tools, materials, and process technology by suppliers needed to support the device makers’ advances in design.79
Semiconductor device makers complained that U.S.-based equipment and materials vendors lacked ability to recognize and address quality problems associated with their products, while the suppliers faulted the device firms for nonstandardized requirements, unwillingness to share relevant data, and an over-emphasis on squeezing the lowest price, rather than the best quality, out of suppliers. Intel’s Gordon Moore, who participated in Sematech during its early years, recalls that Sematech “worked with these [supplier] companies to develop reliable tools, to teach them total quality control, and to help them understand the needs of the industry and the increasing sophistication of the manufacturing process.”80 Suppliers’ equipment was put through rigorous endurance testing, a process dubbed
77 Mark Pinto of Applied Materials cited the problems associated with establishing an industry chain in an emerging sector at the National Academies 2011 symposium on photovoltaics. He observed that in the solar industry it was hard to form collaborations “because manufacturing and technical development were still competitive.” There were “some three dozen solar start-ups in Silicon Valley developing nearly the same number of [copper indium gallium arsenide] processes, all of them competing. Believe me, I know, because they all want equipment, and it’s hard to find.” However, he pointed out, even in such situations, “there are areas in which collaboration is useful,” such as the modeling necessary to explore optics and electronics. “You can plug in whatever band structure you want for whatever elements and do the model from first principles. This helps to figure out what’s going on and how to design these structures. Every company should benefit from that.” National Research Council, Future of Photovoltaic Manufacturing, 117.
78 Sam Harrell is former President of SEMI-Sematech, a consortium formed by U.S. tool and materials suppliers to support collaboration with Sematech. “Oral History with Sam Harrell.” In 1989, Kristopher Lee, Program Manager for Sematech’s tool development program, estimated that U.S. semiconductor production lines could only be used for 25 to 40 percent of a given work week. “Most of the time, they are shut down, because one or more machines requires repair or maintenance.” “Sematech Racing the Clock in Effort to Regain US Lead in Chipmaking,” Dallas Morning News, July 22, 1989.
80 Moore observes that “much of the important work required to improve manufacturing equipment did not have to be done individually but could be done by the consortium centrally. The consortium developed a cost-of-ownership model for manufacturing tools that described the problems in detail.” Gordon Moore, “The Sematech Contribution,” in National Research Council, Securing the Future, 99–100. Sematech developed a program called “Partnering for Total Quality” (PFTQ) that benchmarked U.S. suppliers’ performance against global trends (which showed that “U.S. equipment makers were almost looking at going out of business.”) and, jointly with suppliers, developed protocols for PFTQ partner training. Browning and Shetler, Sematech, 135.
“Iron Man.”81 Conversely, suppliers successfully demanded that device makers develop and share nonproprietary standards.82 Sematech roadmaps were modified to include “the intricate coordination between [Sematech’s quality program] with equipment and material suppliers.”83 One supplier company executive, Scott Kulicke, who was instrumental in setting up SEMI-Sematech, a parallel organization of suppliers that interfaced with Sematech, observed in 2004 that
[w]hen all was said and done, the part of Sematech that really worked was in the roadmap, and in the ability of using the roadmap and the Sematech infrastructure to focus the equipment companies on a few stated objectives so we did a better job of bringing it out on time and to spec. And that allowed the semiconductor guys to regain product parity in a rapidly evolving, technologically evolving environment.84
In flexible electronics, contrasted with the semiconductor industry of the 1980s, no established group of major device makers exists to demand cost savings and improved quality from upstream suppliers. Rather, a flexible electronics food chain has emerged in the United States that—for the moment at least—primarily serves offshore device makers, but which constitutes a viable infrastructure to support potential device manufacturing in the United States. The toxic character of relations between U.S. semiconductor device makers and their vendors in the 1980s likely would have foreclosed the revival of semiconductor manufacturing in the United States had it not been addressed in a foursquare manner, primarily through the Sematech initiative. In flexible electronics, the U.S. competitive position could probably not survive a comparable protracted period of culture clash between whatever indigenous device producers may emerge ultimately, on the one hand, and suppliers, on the other. Consortia, and other forms of collaboration, may play a role comparable to that of Sematech in mitigating conflict and forging
81 Intel’s Paolo Gargini recalls that prior to Sematech’s formation, Intel “[B]uilt two adjacent chambers in which we would put a Japanese equipment in one chamber, and we will put a US equivalent in the next chamber, and then we will compel the CEO of the company to watch his equipment fail after two hours, while the Japanese equipment had been going for a week.” After Sematech was formed, Intel “essentially … gave Sematech the whole procedure (of testing the endurance of the equipment) that we had developed here in [Intel] and Sematech created the name Iron Man (i.e. endurance test) that was essentially, the equipment was going to be put through the Iron Man until it broke down and then there would be the list of defects (failures) and the supplier was supposed to fix it (the equipment failures), and so forth, so the major benefit of Sematech was really to help not the IC companies, but really, to help the equipment company come back to speed.” “Oral History of Paolo Gargini,” 27.
82 Dan Hutcheson, President of the consultancy VLSI Research, observed in 1993 that prior to Sematech “the semiconductor industry looked at the equipment industry like used-car salesmen. The way it usually worked was that a chip maker would take two equipment vendors and pit them against each other. So they’d get the prices down, but there was not incentive to produce reliable equipment.” “Sematech Sets Model for Industry Consortium,” Fort Worth Star-Telegram, November 10, 1993.
83 Browning and Shetler, Sematech, 135.
84 “Oral History of C. Scott Kulicke,” September 22, 2004 (Computer History Museum, 2004).
strong and lasting ties within the industry supply chain. One of the most urgent challenges will be the development of standards.
Development of common standards. The development of widely accepted industry standards is a necessary step in the evolution of a new technology-intensive sector. Standards enable industry participants and customers to understand, measure, and compare various materials, devices, and processes. The existence of common standards is also necessary to enable the interoperability of equipment and systems developed by different companies in complex manufacturing operations. In most cases the development of workable and generally acknowledged industry standards requires extensive collaboration. NIST plays a key role in the development of industry standards, but in the United States such standards “are not handed down by the government, but produced through a collaborative process.”85 Cooperative research organizations may not be essential in all cases to such collaboration, but the historical record illustrates the positive role they can play.
For example, at the time Sematech was formed the semiconductor fabrication process utilized company-specific computer-integrated manufacturing (CIM) systems that “evolved over many years to support the particular needs of an individual company or factory,” a reality that resulted in “high support costs and long delays to add new functionality . . . low customer satisfaction and low development/support staff morale.”86 In 1994, Sematech and Texas Instruments eleased a jointly developed protocol, CIM Framework 1.0, which allowed software applications to interoperate with each other in the same manner that Microsoft Windows–compatible applications interface in computers using the Windows operating system. This innovation enabled manufacturers to avoid installation of multiple proprietary CIM systems that were not necessarily compatible with other suppliers’ tools. “The use of cooperative equipment standards in an otherwise competitive environment reduced duplication and enhanced manufacturing productivity.”87 A retrospective assessment by NIST, which participated in the effort pursuant to a cooperative research and development agreement (CRADA), observed that Sematech played a key role in developing an industry consensus supporting the CIM framework.88
85 Comments of Keat Rochford, Acting Director, Electronics and Electrical Engineering Laboratory, National Institute of Standards, “Measurement and Standards: The Role of NIST,” in National Research Council, Future of Photovoltaic Manufacturing, 163.
86 Sematech (CIM) Framework Specification Version 2.0.
87 Browning and Shetler, Sematech, 176.
88 S.L. Stewart and James A. St. Pierre, NIST, “Experiences with a Manufacturing Framework,” Business Object Design and Implementation (1997), 138. “To achieve Sematech’s objectives, it is not sufficient to produce a specification, even one of technical excellence. There must also be widespread agreement among both the suppliers and users of manufacturing software that applications should be based on the specification. This consensus is a necessary step in the road to adoption and success. Sematech has already involved the users’ groups from its member companies in the process of developing the specification. It is also contacting independent suppliers and providing orientation
Like semiconductors, flexible electronics production will be highly automated and require the development of industry standards for equipment, software, and materials. Printed electronics manufacturers have complained, for example, that materials received from suppliers are of inconsistent quality, while suppliers counter that users need to specify “a few broad categories of standard material quality, as the semiconductor industry did early on and [that] would allow suppliers to focus on their limited development funds on controlling the key specifications to the necessary degree for multiple customers.”89
Given the diversity of flexible electronics technologies, it is unlikely that a single public or private organization can establish industry-wide standards and, in fact, standards setting has moved forward in some fields already.90 However, standards setting may evolve in this industry as NIST indicates. As in other sectors, the progress will be industry-driven, and consortia are likely to play a constructive leadership role.91
It is far from inevitable that the formation of industry standards for flexible electronics will await consensus among U.S.-based firms. In 2009, Germany’s research ministry BMBF sponsored a collaboration, MaDriX, by local firms to develop standards for the printed electronics industry. This effort involved €15 million, €8 million of which was provided by BMBF. The effort included a printed electronics manufacturer, PolyIC, materials suppliers BASF, Evonik Industries, and ELANTAS Beck and Siemens, which provided the necessary inspection systems. “Through MaDriX, material parameters will be determined and an unvarying test environment for new materials introduced. By doing so, companies can develop new materials quickly and more efficiently. Standardizing test conditions facilitates use of a statistical measurement system developed by Jacobs University Bremen and Siemens to enable companies to compare results.92
Leveraging the federal laboratories. A number of federal laboratories are currently engaged in research within or relevant to the field of flexible electronics,
and training about the CIM Framework in scheduled classes and public conferences. We believe that this process of awareness, involvement, and training is absolutely essential to the success of a CIM Framework, and we recommend that it be continued and expanded to the limits of the resource available” (ibid).
89 “Printed Electronics Suppliers Look to Learnings from Electronics World to Help Scale Volume Production,” Printed Electronics Now, January 2010. The consultancy IDTechEx observed in 2011 that in flexible electronics, “for most devices, there are no globally approved test standards, so while one company may claim exceptionally high mobility for an organic semiconductor, the trade-off may be lifetime and using a very high voltage. Similarly, for lighting and displays, there needs to be test standards to fairly compare different materials and devices,” IDTechEx, Printed, Organic & Flexible Electronics: Forecasts, Players & Opportunities 2011-2021 (2011), 232.
90 The National Renewable Energy Laboratory in Denver has developed standards for evaluating and comparing materials and devices in the area of organic photovotaics and has been globally acknowledged as the independent test center for these technologies. IDTechEx, Printed, Organic, and Flexible Electronics, 232–233.
91 NIST, Global Standards Information, <http://gsi.nist.gov/global/index.cfm/L1-5/L2-44/A-165>.
92 “MaDriX Sets New Standards for the PE Industry,” Printed Electronics Now, September 2009.
including NIST and the Ames, Lawrence Berkeley, and Oak Ridge National Laboratories.93 NIST has been working to address “basic unknowns” associated with flexible electronics materials and nanoscale structures “through sophisticated measuring instruments that range from x-rays and synchrotrons to acoustic surface measurement and scanning tunneling microscopes . . . to help the technology developers better understand the materials science that underlie manufacturing behaviors.”94 The federal laboratories have a broad statutory mandate to transfer technology to industry, including the conclusion of CRADAs.95 However, as Carol Battershell of the Department of Energy (DOE) noted in 2011, a “perceived weakness [exists] in the commercialization activities of the national laboratories.”96
The federal laboratories are engaged in a number of initiatives designed to accelerate transfer of technology to industry, with implications for flexible electronics.97 The Sematech experience offers a case study in how a consortium can lay the groundwork for productive future collaborations between individual firms and the labs. A variety of cultural and institutional factors had historically inhibited working relationships between the semiconductor industry and the federal laboratories, notwithstanding the extraordinary resources and relevant competencies available at some federal facilities.98 Sematech played a significant
93 See Chapter 6, “Lawrence Berkeley National Laboratory Develops Micro-Activator that Flexes Under Laser Light,” Flexible Substrate, January 2013; “ORNL Develops Carbon Nanotube Conductive Coatings for Flexible Electronics,” Flexible Substrate, September 2011.
94 Eric K. Lin, NIST, “Advancing Technology Through Measurement Science at NIST,” in National Research Council, Future of Photovoltaic Manufacturing, 112.
95 Stevenson-Wydler Technology Innovation Act of 1980 (P.L. 96-480); Bayh-Dole Act of 1980 (P.L. 96-517); Federal Technology Transfer Act of 1986 (P.L. 99-502); National Competitiveness Technology Transfer Act of 1989 (P.L. 101-189); National Technology Transfer and Advancement Act of 1995 (P.L. 104-113); Technology Transfer Commercialization Act of 2000 (P.L. 106-404); America COMPETES Act (P.L. 110-69).
96 Caroll Battershell, Senior Advisor for Commercialization and Development, Energy Efficiency and Renewable Energy, U.S. Department of Energy, “Bringing Department of Energy Innovations to Market,” in National Research Council, Future of Photovoltaic Manufacturing, 155. See also NIST, Federal Laboratory Transfer Fiscal Year 2010, August 2012; General Accountability Office, Technology Transfer: Clearer Priorities and Greater Use of Innovative Approaches Could Increase the Effectiveness of Technology Transfer at Department of Energy Laboratories, June 2009.
97 DOE operates a Process Development and Integration Laboratory that supports manufacturing-scale R&D and enables DOE National Laboratories to partner with companies to develop commercially relevant manufacturing processes. DOE’s Technology Commercialization Fund facilitates the transfer of technologies from the laboratories to companies in a manner designed to bridge the “valley of death” between research and later stage funding. The Entrepreneur in Residence program enables selected entrepreneurs to spend a year in National Laboratories “mining any knowledge that is available,” with a commitment to look preferentially at the labs’ technologies for commercialization. John Lushetsky, “DOE Solar Energy Technologies Program: Accelerating the U.S. Solar Industry,” and Carol Battershell, “Bringing Department of Energy Innovations to Market,” in National Research Council, Future of Photovoltaic Manufacturing, 153–157.
98 Sandia National Laboratory in New Mexico, which was DOE’s leading research laboratory for semiconductors, held responsibility for all electronics and microelectronics systems in the U.S. nuclear arsenal. Sandia had an on-site semiconductor design, manufacture, and failure analysis capa-
role as “matchmaker” between the industry and federal laboratories, establishing operations at Sandia and Oak Ridge National Laboratories, and NIST.99 Collaboration with federal laboratories proved particularly valuable for industry-leading device makers and smaller equipment makers.100
Improving performance. Sematech’s collaborative approach to manufacturing challenges has enhanced the members’ productivity and efficiency in numerous ways. A system of “blind benchmarking” enables companies to compare their performance metrics with those of other companies.101 “Equipment productivity teams are joint efforts by members to identify common problems with a tool, work together with the tool supplier, and share information about how to make a given tool perform more efficiently.”102 There is no reason why closely comparable, if not identical, practices could not be applied in industry collaborations in flexible electronics.
bility. In collaboration with Sematech, “Sandia’s technical capabilities easily met and often exceeded Sematech’s needs.” But the industry regarded the federal laboratories as insensitive to commercial time scales and other imperatives, and the latter saw the industry as inexperienced in pursuing public return from research. Companies viewed federal laboratories as variants on university research laboratories without tenure pressure, whose researchers “are more concerned with pushing the frontiers of science than meeting the demands of making timely analysis and recommendations for the engineering of manufacturing equipment.” Elias Carayannis and James Glover, “The Sematech-Sandia National Laboratories Partnership—A Case Study,” Technovation, 22 (2002): 588.
99 In 1989, Sematech and Sandia concluded an agreement to create the Semiconductor Equipment Technology Center (SETEC) as a vehicle for concentrating Sandia’s resources on the development of tool designs and methodologies and equipment performance and reliability. This collaboration led to a $110.7 million CRADA that afforded the U.S. industry access to “the use of first-rate facilities and a complete range of science-based expertise upon which to call when faced with real industry problems.” In 1992, Sematech and Sandia concluded a CRADA to establish the Contamination-Free Manufacturing Research Center to solve defect problems associated with larger die sizes and smaller feature sizes required by next-generation semiconductor devices. In 1993, Sematech entered into a CRADA with NIST, with DOE contributing $53.6 million and Sematech $49.4 million, to support semiconductor manufacturing R&D in areas of core competence including materials analysis, equipment modeling and design, and equipment and software reliability. Browning and Shetler, Sematech, 173–174; John McGrayer, SETEC/Semiconductor Manufacturing Technologies Program: 1999 Annual and Final Report (Sandia Report SAND2000-2845).
100 McBayer, SETEC, 38; Carayannis and Gover, “Sematech-Sandia,” 588–589.
101 Blind benchmarking involves members’ sharing 50 performance metrics on a nondisclosed basis. Each company recognizes its own data but not that of other companies. One member, seeing the lower cost of electricity paid by other members, used the data to demonstrate to its power company that its rates were not competitive and secured a reduction. Blind benchmarking can also serve as an early performance wake-up call. “When you see that your 10 percent improvement still leaves you behind by 50 or 60 percent, you realize you have to do something different.” Michael Polcari, “The Sematech Model: Potential Applications for PV,” in National Research Council, Future of Photovoltaic Manufacturing, 183.
102 In one case, a Sematech member company found that most of the power used by the tool was consumed by the pumps. The research team identified which pumps could be idled at which times, saving substantial energy consumption. The member company then worked with the tool suppliers to adjust the idle modes for maximum efficiency. The information was made available to all Sematech members. Ibid., 183.
Building human capital. Byron Clayton of NorTech, a technology-based economic development intermediary that has established a flexible electronics cluster in northeast Ohio, emphasized at the symposia and meetings convened for this study that training a workforce to support flexible electronics manufacturing is “hugely important.” He indicated that local universities and junior colleges need to know in advance the types of skill sets that will be required so that they can implement appropriate curricula. Failure to address workforce concerns will result in local manufacturers “looking elsewhere.”103
Although the renaissance of the U.S. semiconductor industry is most commonly associated with Sematech, the education and training role of SRC, which was formed to support and coordinate relevant university-based research, was also important. Dr. John E. Kelly, who heads IBM’s global research efforts, observes that SRC was “instrumental in sustaining the essential pipeline of skills. If you decide you’re going to compete on innovation, versus lowest cost only, you have to have the skills to innovate.” At the time SRC was formed in 1982, fewer than 100 students and faculty were conducting silicon-based semiconductor R&D in the United States; by 2008, that figure had grown to 2,226, with a research output “larger in some dimensions than that of some of the largest corporations in the industry.” SRC “put thousands of highly qualified students into the industry.”104 SRC’s research agenda, which directs government and industry funds to university-based research teams, focuses on the space between “blue sky” basic research and early product development in all parts of the semiconductor value chain. All projects are governed by research contracts with milestones worked out with the principal investigator; intellectual property (IP) is shared, with the universities owning the IP and SRC members allowed royalty-free, non-exclusive access. SRC “make[s] sure there’s no blocking IP.”105
Sematech-style industry collaborations and public–private partnerships are sometimes characterized as “industrial policy,” with a variety of negative and counterproductive connotations. This study indicates that such industrial policies are being implemented across Europe and East Asia in flexible electronics, including even in countries with robust Manchestrian traditions, like Britain, and ferocious intra-industry rivalries that have historically inhibited collaboration, like Korea. At least in the European case, the merits of public–private
103 National Research Council, Flexible Electronics for Security, Manufacturing, and Growth in the United States: Summary of a Symposium (Washington, DC: The National Academies Press, 2013).
104 John E. Kelly, “Collaboration for Success in Semiconductors,” in National Research Council, Future of Photovoltaic Manufacturing, 73. As of 2011 SRC had invested more than $1.3 billion contributed by members and government, supported 7,500 graduate students through 3,000 research contracts, 1,700 faculty, and 241 universities. SRC support resulted in more than 43,000 technical documents, 326 patents, 579 software tools, and work on 2,315 research tasks or projects. Larry Sumney, “Semiconductor Research Corporation,” in National Research Council, Future of Photovoltaic Manufacturing, 184–185.
105 Sumney, “Semiconductor Research Corporation,” in National Research Council, Future of Photovoltaics Manufacturing, 185.
The FlexTech Alliance
The FlexTech Alliance, based in San Jose, California, is a consortium that promotes the supply chain for electronic displays and printed, flexible electronics. Like its organizational predecessor, the U.S. Display Consortium, it encourages collaboration among companies, academia, and research organizations, and it funds R&D projects in the display supply chain. The scale of FlexTech is extremely modest by current international standards, with combined funding support from the Army and Air Force Research Laboratories of $6 million over three years, with matching industry costsharing. This is sufficient to fund one or two tool development projects, several project demonstrators, and some materials and device R&D annually.
FlexTech representatives (including company executives) have emphasized the importance of fostering increased collaboration among hitherto disparate but welldeveloped industry sectors in the United States: electronics, printing, and packaging, among others. They believe that collaboration could be manifested in a variety of ways, as it is in Europe—the formation of industry clusters, joint universityindustry R&D efforts, supply chain partnerships, and specialized consortia. “What you create as a whole is greater than the sum of the parts.”a
aConference call with FlexTech on November 15, 2013. See Chapter 7 for an extended description of the FlexTech Alliance.
collaborations have been validated by the eminent U.S. scientists and engineers who comprised the 2010 World Technology Evaluation Center, Inc. (WTEC) mission. Under such circumstances, and viewed against the concededly one-off experience of Sematech, the question presented does not appear to be why the United States should adopt a public–private collaborative approach in this new field, buy why it should not do so. The WTEC panel observed that in Europe,
There is close industry-university-innovation center cooperation in precompetitive research. . . . In addition to substantial technical challenges, the cost of research and development in the flexible electronics area is perceived as a significant risk to individual companies, many of which have core expertise in only a subset of required competencies (e.g., circuit/systems design, coating, printing, and materials). European innovation centers mitigate the financial risk by sharing costs among multiple commercial enterprises and by leveraging substantial government funding at both the national and European Union (EU) levels. On the technical front, the centers foster a highly synergistic and interdisciplinary environment in which the complementary expertise of industrial, government, and academic scientists is combined to achieve new systems design goals (e.g., ultra-low-power
systems in foil), enhanced device performance, broader materials choices, and practical, low-cost manufacturing approaches.106
Examining and comparing selected innovation programs both foreign and domestic, and their potential to advance the production of flexible electronics technology in the United States, an ad hoc committee of the National Academies reviewed the goals, concept, structure, operation, funding levels, and evaluation of foreign programs similar to major U.S. programs.107 Specifically, the committee examined the role of research consortia around the world to advance flexible electronics technologies, comparing their structure, focus, funding, and likely impact, in order to determine what appropriate steps the United States might consider to the develop the industry in the United States.108
This volume surveys flexible electronics developmental efforts under way in Europe, East Asia, and North America with a focus on government support. Chapter 2 summarizes the enormous promise of flexible electronics as well as the technological obstacles to realizing that promise. Chapter 3 presents an overview of government promotional efforts in the three regions surveyed. Chapter 4 examines the beginnings of a competitive landscape in an industry where most of the technology is not yet mature in a commercial sense, with a highlight on expert SWOT (strengths, weaknesses, opportunities, threats) analyses of relevant regions and countries. Chapter 5 surveys European promotional efforts in flexible electronics, with a focus on programs being implemented by the European Union and by national and regional governments in the United Kingdom, the Netherlands, Belgium, Germany, and Finland. Chapter 6 presents an overview of East Asian promotional efforts featuring South Korea, Taiwan, Japan, and China. Chapter 7 summarizes federal and state initiatives in North America. This volume concludes with the committee’s consensus findings and recommendations on the potential of the U.S. flexible electronics industry, the possible contributions of intermediate institutions to facilitate applied research and manufacturing, and other measures to support the development of the flexible electronics industry in the United States.
106 WTEC Panel Report on European Research and Development in Hybrid Flexible Electronics (July 2010), xvi, 7. The WTEC panel members were Dr. Ananth Dodabalapur (Chair), Ashley H. Priddy Centennial Professor in Engineering at the University of Texas at Austin; Dr. Ana Claudia Aria, Manager of the Printed Electronic Devices at PARC Inc. (formerly XEROX-PARC); Dr. C. Daniel Frisbie, Professor of Chemical Engineering and Material Science at the University of Minnesota; Mr. Daniel Gamota, co-founder and President of Printovate Inc., a developer of large area electronics in lighting, renewable energy, and point-of-care diagnostics; Dr. Tobin J. Marks, Professor of Chemistry, Material Science, and Engineering at Northwestern University; Dr. Colin E. C. Wood, Research Professor at Texas State University at San Marcos, and Dr. Grant Lewison, an expert on bibliometrics and a Senior Research Fellow at University College, London.
107 For a list of the members of this committee, see the Front Matter of this report.
108 See the Preface for the complete text of the Committee’s Statement of Task.
A Review of Sources and Their Limitations
The committee’s review of programs and policies to support national flexible electronics industries has drawn from a variety of sources, including
- Presentations by U.S., Korean, Taiwanese, and German participants at a National Academies symposium convened by the committee in 2010 and presentations by U.S. and Dutch industry representatives in a workshop convened by the committee in 2011.
- Site visits by the committee to the Flexible Display Center at Arizona State University and to flexible electronics firms research centers in the FlexMatters cluster developing in northeast Ohio.
- A 2010 assessment by the WTEC study of European R&D programs in flexible electronics.
- Reports by government and parapublic organizations such as the European Commission, VTT (Finland), the European Union’s (EU’s) OPERA project, the UK House of Commons Innovation, Universities, Science, and Skills Committee, the German National Academy of Science and Engineering (acatech), NorTech (Ohio), the UK Plastic Electronics Leadership Group, PolyMap (EU), and Taiwan’s ITRI.
- Reports of U.S., EU, and UK industry associations.
- Company annual reports and regulatory filings.
- Consultant studies, for example, HSBC’s Fantastic Plastic and IDTechEx’s Printed, Organic & Flexible Electronics Forecasts, Players & Opportunities (2011).
- Websites and annual reports of each of the organizations known to be involved in flexible electronics R&D, such as individual Fraunhofer institutes, Finland’s VTT, the Holst Centre in the Netherlands, ITRI in Taiwan, and various research organizations in Korea and Japan.
- European databases such as CORDIS and the UK’s EPSRC database.
- National Academies’ studies, for example, 21st Century Manufacturing: The Role of the MEP Program (2013), which includes an extended appendix on leading manufacturing support programs in Asia and Europe.
- Informed trade press, including publications such as Flexible Substrate, Plastic Electronics, The Emitter, Printed Electronics Now, Nanowerk, and Printed Electronics World.
- Media reports from East Asia and Europe including economic, technical, and mainstream publications such as Chosun Ilbo, JoongAng Daily, Taiwan Economic News, the South China Morning Post, and China Daily and government press agencies such as Yonhap (S. Korea), Xinhua (China), Central News Agency (Taiwan), and Jiji (Japan).
- Relevant academic articles and books.
These sources are cited in full when referred to in the report.
The committee has, within the scope of its budget, made every effort to draw on publicly available information about programs to support the flexible electronics industry. However, information available about these national programs is limited.
- Existing academic work with respect to government support for flexible electronics is very limited. Although there are numerous technical scientific articles about research findings, these typically address scientific arcana and rarely contain information relevant to the committee’s task.
- Published consultants’ studies mostly describe what individual companies are doing and what applications could emerge in flexible electronics. Although useful in this regard, they shed little light on government policy measures.
- Commercial data are limited and consist largely of forecasts—inherently subjective—because this is a relatively new industry.
- Articles, studies, or government white papers about relevant programs in East Asian countries are often not made public and are rarely published in English.