PAGE 1
Commercializing High-Temperature Superconductivity June 1988 NTIS order #PB88-246442
PAGE 2
Recommended Citation: U.S. Congress, Office of Technology Assessment, Commercializing High-Temperature Superconductivity, OTA-ITE-388 (Washington, DC: U.S. Government Printing Office, June 1988). Library of Congress Catalog Card Number 88-600535 For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, DC 20402-9325 (order form can be found in the back of this report)
PAGE 3
Foreword Less than two years ago, superconductivitytotal loss of resistance to electricitycould be achieved only at temperatures near absolute zero. Since the discovery of high-temperature superconductivity (HTS), research laboratories around the world have pushed the temperature limits steadily upward, opening the way to commercial applications with potentially revolutionary impacts. The scientific race is becoming a commercial race, one featuring U.S. and Japanese companies, and one that the United States could lose. Indeed, American firms may already be falling behind in commercializing the technology of superconductivity. Japanese companies have been more aggressive in examining possible applications of HTS, and what it might mean for competitive strategy. While payoffs on R&D may lie a decade or more in the future, managers in Japan have been willing to take the risks, Although a number of U.S. companies have also begun major efforts in HTS, most American managers, under pressure to show short-term profits, have been more inclined to wait and see. So far, the U.S. Government has supported the development of HTS in its traditional wayby putting money into R&D, mostly through the mission agencies. Federal agencies moved quickly to channel money to HTS when news of the discoveries broke. The breadth and depth of the response in government agencies and Federal laboratories, and in the university system, shows the continuing vitality of the scientific enterprise in the United States. Although Federal dollars will help support a technology base that the private sector can build upon, the U.S. Government is not providing direct support for commercialization. Nor have we any policy or tradition for this kind of supportunlike countries such as Japan. Postwar U.S. technology policy coupled R&D funding with indirect measures, such as tax policy, to stimulate commercial innovation. So long as American companies remained well ahead of the rest of the world in technical skills and management ability, this approach proved successful. With the continuing decline in competitiveness across many sectors of the U.S. economy, it no longer seems good enough, The Senate Committees on Governmental Affairs, Energy and Natural Resources, and Commerce, Science, and Transportation, together with the House Committee on Science, Space, and Technology requested the assessment of which this report is part. OTAs Energy and Materials Program is also conducting a more comprehensive examination of the science and technology of high-temperature superconductivity, and the future research agenda, as the second part of this assessment. Their report will appear in 1989. OTA is grateful for the assistance provided by many people inside and outside of government during the preparation of this report. Full responsibility for the contents rests with OTA. ///
PAGE 4
High-Temperature Superconductivity Review Panel Claude Barfield American Enterprise Institute Arden Bement TRW, Inc. Harvey Brooks Harvard University Kent E. Calder Princeton University Jeffrey Frey University of Maryland Jacques S. Gansler The Analytical Science Corp. Donald A. Hicks University of Texas at Dallas David A. Hounshell Harvard Business School and University of Delaware Clifford K. Jones Westinghouse Electric Corp. Henry Kressel E.M. Warburg, Pincus & Co., Inc. Harry Kroger Microelectronics &Computer Technology Corp. Ira Magaziner Telesis William D. Manly Consultant Edward Mead E. I. du Pont de Nemours & Co. Egils Milbergs Institute for Illinois Thomas A. Pugel New York University Maxine Savitz The Garrett Corp. David J. Teece University of California, Berkeley James Wong Supercon, Inc. Donald L. Wood General Electric Co. Many other people reviewed various drafts of this report. Special thanks are due to H. Kent Bowen, Duane Crum, Simon Foner, Donald N. Frey, Henry H. Kolm, and Alexis P. Malozemoff. OTA appreciates and is grateful for the assistance and thoughtful critiques provided by all the reviewers, none of whom, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for the report and the accuracy of its contents. iv
PAGE 5
OTA Project Staff: Commercializing High-Temperature Superconductivity Lionel S. Johns, Assistant Director OTA Energy, Materials, and International Security Division Audrey Buyrn Industry, Technology, and Employment Program Manager John A. Alic, Project Director W. Wendell Fletcher, Deputy Project Director Dorothy Robyn Martha Caldwell Harris Clare Delmar Jane A. Alexander Carollyne Hutter Contributors Gregory Eyring Laurie Evans Gavrin Joan Lisa Bromberg The Laser History Project Kenneth Flamm Glenn R. Fong Robert R. Miller David C. Mowery David Sheridan Granville J. Smith, II W. Henry Lambright, Maureen ONeill Fellows, and Leslie A. Jones Syracuse University Administrative Staff Edna M. Thompson, Administrative Assistant Diane D. White, Administrative Secretary
PAGE 6
Workshop on High-Temperature R&D in Japan and the United States, January 1988 Ronald H. Baney Dow Corning Fernand D. Bedard U.S. Department of Defense Justin L. Bloom Technology International, Inc. Roger W. Boom University of Wisconsin H. Kent Bowen Massachusetts Institute of Technology Richard C. Bradt University of Washington Donald W. Murphy AT&T Bell Laboratories Jeffrey Frey University of Maryland William Gallagher IBM Edward Mead E.I. du Pont de Nemours & Co. Charles T. Owens National Science Foundation Robert Pohanka Office of Naval Research Richard J. Samuels Massachusetts Institute of Technology Roland W. Schmitt Rensselaer Polytechnic Institute Joseph L. Smith, Jr. Massachusetts Institute of Technology Larry Smith Microelectronics & Computer Technology Corp. Nancy J. Tighe Microphage Systems Corp. Theodore Van Duzer Richard Harris University of California, Berkeley National Bureau of Standards Harold Weinstock Young B. Kim Air Force Office of University of Southern California Scientific Research M. Brian Maple University of California, San Diego
PAGE 7
CONTENTS Chapter 1: Summary . . . . . . . . . . . . Chapter 2: Commercialization: Government and Industry . . . . Chapter 3: Superconductivity in Japan and the United States . . . Chapter 4: U.S. Technology Policy: Issues for High-Temperature Conductivity. . . . . . . . . . . . . . Chapter 5: Strategies for Commercial Technology Development: 3 17 51 83 High-Temperature Superconductivity and Beyond .................,,123 Appendix A: Glossary. . . . ..., ..., . ..., . ., ..151 Appendix B: The Technology of Superconductivity . . ........,.,154 Index ....... . . . . . . . . ................,,...167 vii
PAGE 8
Chapter 1 Summary
PAGE 9
CONTENTS Commercialization: Public and Private Dimensions . . . . . High-Temperature Superconductivity: U.S. and Japanese Responses. . Principal Findings. . . . . .. .. . . .. ... ... ... .... Federal Funding for HTS R&D . . l R&D and Commercialization . . . . . . . . . HTS in the United States and Japan. . . 0 . . . . . HTS and U.S. Technology Policy . . . . . . . . .. Federal Government Strategies . . . . . . .. ... ...... Box 3 6 7 7 8 9 11 12 Box Page A. The U.S. Position in Technology . . . . . . . . 4
PAGE 10
Chapter 1 Summary During 1987, high-temperature superconductivity (HTS) became a symbolof the new and unexpected, of what was right and wrong in U.S. science, technology, and industry, of U. S.Japan competition in high technology. In December of the preceding year, two scientists at IBMs Zurich research laboratory caught the worlds attention with their discovery of superconductivity in the range of 35 to 40 K (degrees Kelvin, i.e., degrees above absolute zero) nearly double the record temperature for total loss of resistance to electricity. Within 2 months, transition temperatures had doubled once moreto over 90 Kwith near-simultaneous discoveries of a second family of ceramic superconductors in the United States, China, and Japan. In March 1987, thousands of scientists jammed a hotel ballroom in New York to hear the latest findingsa meeting dubbed the Woodstock of physics. The race to higher temperatures was on. With it came warnings that the United States could lose out to foreign competitors in commercializing a technology with potentially revolutionary impacts. Indeed, one of the principal findings of this assessment is that American companies may already have begun to fall behind. Japanese firms have been much more aggressive in studying possible applications of HTS, and have more people at work, many of them applications-oriented engineers and business planners charged with thinking about ways to get HTS into the marketplace. In the midst of the excitement, four congressional committeesthe House Committee on Science, Space, and Technology, and the Senate Committees on Governmental Affairs, Energy and Natural Resources, and Commerce, Science, and Transportation-asked OTA to examine a series of questions that ranged from public and private sector responses to HTS (here and abroad) to the advantages and disadvantages of a new Federal agency for supporting the development of commercial technologies. This special report begins with a look at U.S. strengths and weaknesses in technology development and commercialization (ch. 2), both in general and for HTS. The analysis then goes on (in ch. 3) to the strategies of U.S. and Japanese companies, as managers in each country look ahead to the new opportunities. The fourth chapter presents 20 policy options for congressional consideration; the context is U.S. technology policy as a whole, with HTS as a special case. Most of the policy options deal, in one way or another, with the management of the Federal R&D budget. Chapter 5, the last, considers three broad alternatives for speeding commercialization. 1 IApp. B, at the end of this report, summarizes the technology of superconductivity, including prospective applications, with estimates of time horizons for commercialization. (The glossary in app. A includes many of the specialized terms that apply to superconductivity.) OTA will follow this special report with a more detailed examination of the science and technology of superconductivity, the research agenda, and potential applications, to be published in 1989. COMMERCIALIZATION: PUBLIC AND PRIVATE DIMENSIONS U.S. competitiveness in both smokestack and high-technology industries has been slipping for years. Loss of technological advantage has been one of the reasons (box A). On the face of it, this seems paradoxical. The U.S. Government spends more on R&D than government and industry together in Japan. Federal R&D dollars help create a vast pool of technical knowledge that the private sector (including foreign firms) can draw upon. Beyond this, U.S. technology policies have relied heavily on indirect incentives for innovation and commercialization by industry. This approach leaving R&D priorities largely to the mission agencies, trusting to in3
PAGE 11
4 direct policies to stimulate commercialization declining competitive ability across much of worked well in the earlier postwar period, when the U.S. economy, it no longer works well American corporations were unchallenged inenough. In recent years, many U.S. companies ternationally. On the evidence of steadily have had trouble turning existing technical
PAGE 12
5 knowledge into successful products and processes, and getting new technology out of the laboratory and into the marketplace (ch. 2). Of course there is more to commercialization than R&D and technology development. Government policies affect business decisions and competitiveness, not only through technology and science policy, but also through sectorspecific measures (e.g., Government funding for the microelectronics consortium Sematech), and regulatory and macroeconomic policies. U.S. financial markets, for example, have been steadily deregulated. Among the results: greater pressures on industry for short-term investment decisions. OTA has examined the broad range of policy influences on U.S. competitiveness in many other assessments. Here, the analysis focuses on those linked more or less closely to technology itself. They fall into two groups: l policies that affect innovation and commercialization directly, notably the Federal R&D budget; l those with indirect impacts. Federal R&D helps create a technology base that private firms draw on during commercialization. Sometimes companies start development projects because of new research results; other times, they find they need critical pieces of knowledge, perhaps from earlier R&D, to complete a project, or to solve a manufacturing problem. Federally funded projects in lowtemperature superconductivity (LTS), for example, laid the foundation for applications of superconducting magnets in medical imaging equipment. The second group of policies works indirectly-through incentives (or disincentives) for private firms. Some of these policies reduce financial or technical risks, or increase rewards for successful innovators. Tax treatment of capital gains, for instance, affects decisions by prospective entrepreneurs; R&D tax credits make a difference for companies with profits that can be offset. Other such policies work through their influence on demand. Governments purchase military systems and computers, cars and trucks, consulting and construction services. Sometimes, they regulate prices or allocate production among suppliers (as the U.S. Government has done for years in agriculture). With the knowledge base ever larger and more specialized, the great majority of American firms, large and small, can no longer expect to be self-sufficient in technology. The pace and complexity have simply outstripped their ability to keep up. Industry depends more heavily than ever before on the huge Federal R&D budget$60 billion, about half of all U.S. R&D spending. Nonetheless, the U.S. Government has left most questions of R&D funding to the mission agencies, with their focused interests and immediate needs. While other countries have crafted policies for direct support of commercial technologies, the United States has not. policy makers here have argued that direct measures lead to harmful economic distortions. Instead, many say, deregulationremoving the roadblocks to innovationwill tap reservoirs of American ingenuity and entrepreneurial vigor that would otherwise be stifled, But most of the roadblocks have come down over the past 15 years, while U.S. competitiveness has continued to slip. To be sure, Federal agencies are paying more attention to the impacts of day-to-day decisions on competitiveness than during the 1970s. Antitrust enforcement reflects global, rather than simply domestic, competition. The national laboratoriesparticularly those overseen by the Department of Energy (DOE)have been seeking ways to work more effectively with industry. With recognition spreading that military R&D spending may not offer the spinoffs and synergies of earlier years, Congress has been debating the merits of a change in direction for technology policies. But it is fair to say that international competitiveness still plays a minor role in U.S. policies compared with those of countries that have learned to export as effectively as Japan, West Germany, or South Korea. The United States is still searching for workable approaches to competing in a relatively open international economy, one in which American companies no longer have big advantages in technology or management skills.
PAGE 13
. 6 HIGH-TEMPERATURE SUPERCONDUCTIVITY: U.S. AND JAPANESE RESPONSES Why all the excitement over HTS? The media have held out the promise of more efficient generation and transmission of electric power, magnetically levitated trains, electromagnetic launchers for space weaponry. Perhaps more important, HTS-based electronics could eventually become building blocks for more sensitive medical diagnostic systems, and faster, more powerful computers. The most important impacts will probably be those that cannot yet be anticipatedthe point maybe facile, but it is true. Even at liquid nitrogen temperaturesfar below room temperature but far above the operating temperatures of older LTS superconductorsthe prospects have attracted as much attention as any scientific development since the laser or gene splicing. Although no one had made a practical conductor or electronic device from the new materials, the Nobel Prize committee gave its 1987 physics award for the Zurich discoveriesthe quickest in history. Early 1988 saw the discovery of several more families of HTS ceramics. Yet the ultimate prizesuperconductivity at room temperaturelies ahead, and no one knows whether it can be achieved, even in theory. Activity has been feverish on the policy front as well as in the research laboratory. Within a few months of the initial discoveries, Federal agencies redirected $45 million in fiscal 1987 funds from other R&D to HTS (ch. 4, table 8). The scientific breakthroughs prompted a dozen bills during the first session of the 100th Congress, proposals ranging from study commissions to a national program on superconductivity. All reflected, in one way or another, concern over commercialization. The policy drama reached a peak in July 1987, when President Reagan brought three ranking cabinet officers to the Federal Conference on Commercial Applications of Superconductivity; in an unprecedented appearance, he announced an n-point initiative for the support of HTS (box B, ch. 2). In a similarly unprecedented move, the Administration closed the meeting to all foreigners except representatives of the press. Although the Presidents message focused on executive branch actions, he stated that the Administration would also be proposing new legislation. The following months brought a sense of anticlimax, with no sign of the promised legislative package. Questions of R&D funding then came to the fore, as the end of the fiscal year passed with no resolution of the budget impasse between the President and Congress. Only at the end of the calendar yearseveral months into fiscal 1988did Federal agencies know for certain how much money they would have for HTS R&D. Taken together, Federal agencies will spend nearly $160 million for superconductivity R&D in fiscal 1988, over half ($95 million) on the new materials (and the rest for LTS). The Department of Defense and the Energy Department together account for three-quarters of the HTS budget, and received most of the increase. DOE, for instance, will have nearly twice as much HTS money as the National Science Foundation (NSF). With NSF a primary patron of university research, the governments priorities seemed rather haphazard, given the great strength of the Nations universities in basic research. Most of the Federal HTS money will go to government laboratories, contractors, and universities that are well removed from the commercial marketplace. The Presidents legislative package, which reached Congress in February 1988, did not address R&D funding. Consistent with the Administrations emphasis on indirect incentives for commercialization, the package included provisions that would further liberalize U.S. antitrust policies, and extend the reach of U.S. patent protection. On the industry side, most American firms viewing payoffs from HTS R&D as uncertain and distanthave declined to invest heavily (ch. 3). A few major corporationse.g., Du Pont,
PAGE 14
7 IBM, AT&Tare mounting substantial efforts. A number of small firms and venture startups have also been pursuing the new technology. By and large, however, American companies have taken a wait-and-see attitude. They plan to take advantage of developments as they emerge from the laboratorysomeone elses laboratoryor buy into emerging markets when the time is right. Unfortunately, reactive strategies such as these have seldom worked in industries like electronics over the past 10 to 15 years, while many American firms seem to have forgotten how to adapt technologies originating elsewhere. Corporate executives in Japan, in contrast, see HTS as a major new opportunityone that could set the pattern of international competition for the 21st century. Japanese companies have made substantial commitments of people and funds, pursuing research and applicationsrelated work in parallel. Firms in more lines of business are at work than in the United States. Steel companies and glassmakers, as well as chemical producers and electronics manufacturers, are seeking new businesses, ways to diversify. Japanese managers see in HTS a road to continued expansion and exporting, and are willing to take the risks that follow from such a view. For years, the claim was common that Japanese firms got a free ride from U.S. R&D. More recently, Americans have realized that Japanese corporations have no need to imitate or to be followers; they have highly competent and creative technical staffs, fully capable of keeping up or taking the lead in fields ranging from automobile design to gallium arsenide semiconductors, opto-electronics, and ceramics. Giving the Japanese the credit they deserve has intensified U.S. anxieties over commercialization. Only in sciencein basic researchdo Japans capabilities remain in question. For the Japanese, HTS presents an opportunity to show the worldand themselvesthat they can be leaders there too. Companies like IBM and Du Pontor Hitachi and NEChave R&D budgets exceeding a billion dollars. They have skilled engineers and scientists to put to work on the technical problems of HTS, money to bet on new opportunities. But these firms are a small minority in both countries, and the competition on them alone. will not depend Photo credit: IBM Conducting strips of HTS material deposited on substrate. PRINCIPAL FINDINGS Federal Funding for HTS R&D It would be hard to criticize the magnitude of U.S. Government spending on HTS. Federal agencies have about $95 million for HTS R&D in fiscal 1988, more than twice the 1987 total. Although little of this represents new budget authority, the U.S. Government will spend more this year on HTS than Japans Government has budgeted for HTS and LTS together. With the Administration seeking $135 million for HTS in fiscal 1989, current and proposed spending might seem more than enough to support rapid commercialization. But totals can be misleading. After all, the United States spends far more on R&D than competing nations, yet U.S. industry has been unable to keep a useful lead in technology. There are many reasons, some of them having to do with the allocation of R&D funds. Nearly 70 percent of
PAGE 15
8 Federal R&D spending goes for national defense; some of this money helps build the technology base for commercial industries, some does not. The story will be the same in HTS. 1. Of $95 million that the U.S. Government has budgeted for HTS R&D during fiscal 1988, the Department of Defense (DoD) will spend $46 million and DOE $27 million. NSF is next at $14.5 million. No other agency has more than about $4 million. R&D funded by DoD and DOE will help support commercialization, but a dollar spent by one of these agencies will probably buy substantially less in terms of the Nations technology base than a dollar spent by NSF. A good deal of DoDs R&D will go for specialized applications in defense systemsincluding the Strategic Defense Initiativewith limited potential for commercial spinoffs; defense missions shape even the basic research supported by the Pentagon. DOE will distribute most of its money to the national laboratories; relationships between DOE laboratories and the private sector have begun to changea trend to be applaudedbut the laboratory system has yet to demonstrate the ability to transfer technologies rapidly and effectively to the private sector. (See Policy Options 1,2,4 in ch. 4, and discussion in chs. 4 and 5.) 2. While the Federal R&D total for HTS may seem impressive, little of it represents new money. This was necessarily the case in fiscal 1987, when agencies had no choice but to redirect existing funds. For fiscal 1988, given the pressures on the Federal budget, agencies have continued to take money from other R&D categories to pay for HTS. Congress may wish to examine the trade-offs necessary at the agency level to finance HTS R&D, and consider appropriating new money for fiscal 1989. (Options 1, 2, 3, 9.) 3. R&D priorities and funding decisions often made at relatively low levels in the agencieshave major and lasting impacts on commercialization. So do mechanisms for inter-agency coordination. The pressures on the Federal budget make good management of agency resources even more important. (Options 1, 2, 3, 6, 8.) But getting the most out of the Federal investment in HTS R&D will take more than inter-agency coordination and effective technology transfer. Successful commercialization will require continuity in R&D funding so that people and organizations can plan ahead. The United States will need graduate-level scientists and engineers educated in fields ranging from materials processing to the physics of electron devices. Most of these people get their training in university programs that depend heavily on Federal support. Likewise, the national laboratories and Federal mission agencies must know where they are going, and how much money they can expect along the way. Industry needs to know whether and when it can look for new research results from Federal R&D. Multiyear R&D planning and budgeting for HTS, on a trial basis, could help set patterns for the future. (Options 2, 3, 4, 5.) R&D and Commercialization No one can say whether superconductivity at room temperature will be possible in the near future or in the distant future. Regardless of progress in finding materials with higher superconducting transition temperatures, 5 to 10 years of R&D probably lie ahead before the technology base will be able to support substantial commercial development. Successful commercialization, in any case, takes more than R&D. It depends on market conditionson a companys ability to anticipate or create demand, and to exploit it. Linking engineering development, marketing, and manufacturingsomethin g Japanese companies excel atis crucial. So is management commitment to the long term. 1. Processing and fabrication methods will be critical for applications of HTS. American companies have fallen down in manufacturing skills across the board; the more heavily process-dependent HTS turns out
PAGE 16
9 2. 3. to be, the more difficult it will be for U.S. firms to keep up with the Japanese. A strong processing emphasis in Federal R&D could help compensate for low priorities in American corporations, a major source of U.S. competitive difficulty. (Options 2, 6, 15, 16 in ch. 4.) The Defense Advanced Researc h Projects Agency (DARPA) solicited zoo proposals on HTS during the summer of 1987, hoping to have $50 million to spend on processing-related R&D. When the final 1988 budget figures came down (in December 1987), DARPA found itself with only $15 million. Nonetheless, even at this lower level the program should be able to make a substantial contribution to commercialization, if well managed and sustained over a number of years. (As this report went to press, the Defense Department had just imposed a freeze on new outside R&D, including this program.) HTS R&D funded by defense agencies will help American companies, but the potential for commercial spinoffs will diminish as military requirements become more specialized and diverge from commercial needs. The list of new technologies and new industries that has emerged from DoD-sponsored R&D is an impressive one: computers; semiconductors; lasers; much automated manufacturing know-how. Why should things be any different with HTS? Because both the United States and the rest of the world have changed. The defense sector has grown apart from the rest of the U.S. economy; DoD money has less impact as other countries focus more of their resources, both public and private, on commercial technologies. At the least, continuing attention to technology transfer from defense contractors and Federal laboratories will be necessary to take commercial advantage of DoD (and DOE) spending. (Options 11, 12, 13, 14, 15.) Just as for technologies like microelectronics, commercializing HTS will require contributions from many disciplines physicists, chemists, materials scientists, electrical, electronic, and chemical engi4. neers. Multidisciplinary research works in industry because it must, but does not come easily in universities (here or in other countries). Federal policies that help establish multidisciplinary R&D within the university system will contribute to strong foundations for HTS and other technologies. (Options 9, 10.) NSF has embarked on a renewed attempt to stimulate multidisciplinary R&D through its program for Engineering Research Centers, and its proposed Science and Technology Centers. Consistent support will be required for these centers to take hold and become a permanent feature of the R&D landscape. HTS will demand a good deal of trial-anderror development (as was true in LTS). With U.S. difficulties in commercialization much more a matter of technology than science, Federal policies that increase support for engineering researcheven more, that seek to redirect research and education in engineering toward practical industrial problemscould have substantial longterm significance. (Options 4, 5, 9, 19.) HTS in the United States and Japan Japans Government took the better part of a year to shape its policy response to HTSa response that, when it emerged, looked not at all like the highly centralized program some Americans had expected. Much of the effort has been directed at getting the three parts of the R&D systemindustry, the universities, Japans national laboratoriesto work effectively together. The Japanese see HTS as a test case for their turn toward basic research, and are giving it high priority. Moreover, lacking energy reserves, they have strong incentives for R&D (in LTS as well as HTS) promising savings in electric power consumption. Japanese firms compete aggressively at home and abroad; they get consistent government supportfor instance, from national laboratories that work effectively with the private sectorbut succeed in international markets on their own merits. In some if not all indus-
PAGE 17
. 10 tries, Japanese companies turn R&D into new products and processes faster than American firms. They target markets effectively, linking R&D to market needs better than many U.S. companies, and manage their factories at least as well as they manage their R&D laboratories. These strengths will pay off in HTS. 1. 2. A few large American companies are putting substantial resources into HTS. But the list is short: AT&T, IBM, Du Pont, a few others. The financial criteria that drive decision-making in American corporations work against a technology like HTS one with uncertain prospects, and profits that lie well in the future; the short-term view fostered by U.S. financial markets could put American companies behind the Japanese within 2 or 3 years, if they are not behind already. A handful of small U.S. companies and startups with venture funding have also been moving into HTS. Although smaller U.S. firms may well develop creative solutions to some of the practical problems of the new technology, these companies do not have the production and marketing capabilities necessary for a major role. They will have a difficult time growing and competing with integrated Japanese multinationals. American managers, by and large, believe HTS should remain in the laboratory until more scientific knowledge is in hand. They emphasize the uncertaintiesadmittedly greatand the lack of evidence promising quick returns from R&D investments. To them, uncertainty urges caution rather than signifying opportunity. American firms have not made commitments to HTS that compare to those in Japan in terms of scale (as indicated by people at work) or scope (as indicated by people assigned to applications-related projects). Many American companies with the technical skills and the money to pursue HTS have taken a wait-and-see attitude. Typically, they have a few people tracking progress in the field. Some of these companies may be able to catch up and 3. 4. compete when applications begin to appear. Others will be left behind. Most Japanese managers believe HTS to be closer to the marketplace than do their American counterparts. Seeking growth and diversification, they have assigned more people to HTS than U.S. firms, and may also be spending more money. The Japanese have committed funds, not only to research, but to evaluating prospective applications. Executives there see HTS as a vehicle for creating new businesses, while Americans are more likely to view it in terms of existing lines of business. And if American managers have been reluctant to commit resources to HTS, the Japanese seem confident that investments now will pay offsome time and in some way. The Japanese could be wrong. In spending money on feasibility studies and engineering analyses, they may miss other opportunities. But given the scale of current investmentsin the range of $200 million dollars in each country (including both government and industry R&D), small compared to overall corporate R&D spendingthere is much to be said for taking the risks. OTAs analysis suggests that commercialization of HTS will proceed somewhat faster than many American managers anticipate, though not so fast as many in Japan expect. If this proves the case, Japanese companies could well come out ahead in the race to commercialize HTS. Japans Government will spend about $70 million for superconductivity R&D (high temperature and low) in 1988. 2 Although ministries and agencies spent much of 1987 jockeying for position, Japan now has in place a set of policies intended to compensate for the bottlenecks and weaknesses in the countrys R&D system: universities with only a few islands of excellence; national laboratories which, although some Comparisons with U.S. Government spending must be treated with caution: fiscal years in the two countries are 6 months out of phase; Japanese budget figures leave out salaries for research workers in universities; national defense has little influence in shaping Japans HTS R&D.
PAGE 18
5. have enviable reputations, cannot claim the breadth or depth of their U.S. counterparts. If their system as a whole still shows weaknesses, in superconductivity, Japans R&D is broadly based and high in quality. With R&D centered in major corporations, government policies aim to strengthen the infrastructure for developing HTS, and stimulate greater cooperation and interaction among industry, universities, and the national laboratories. Japanese officials view international cooperation in HTS research as a potential complement to their countrys own efforts. Much more than a matter of image, they see in internationalization a means of stimulating creativity in Japans universities and government laboratories. In turn, U.S. industry stands to gain by testing Japans willingness to open up its research system. (Options 18, 19, 20 in ch. 4.) HTS and U.S. Technology Policy Japanese companies place high priorities on technology as a competitive weapon; it is not only in HTS that U.S. companies risk falling behind. Business-funded R&D in Japan totals 2.1 percent of gross national product, compared with 1.4 percent here. Fewer high-level managers in American firms have technical backgrounds; they may not fully appreciate the role of R&D in business strategy and international competition. To executives fighting a takeover, research may look like a luxury; after a merger, it may seem expendable. Gaps in the U.S. technology base open where neither Government nor industry has immediate requirements for R&D results. The very unexpectedness of the discoveries in HTS points to the need for ongoing Government support of long-term research. Failure by the private sector to invest in generic R&D, much of it incremental, or in risky projects with potentially big payoffs, throws more of a burden on the Federal Government. 1. Many areas of science and technology, although vital for U.S. competitiveness, get adequate financial support from neither public nor private sources. American corporations have been turning away from long-term, high-risk R&Dthe kind of work called for in commercializing HTS. Knowledge that could help American firms compete is not available when needed. Underinvestment has been most serious in fields that lack glamoure.g., manufacturing. (Options 6, 7, 8, 15, 16 in ch. 4.) Like industry, the Federal Government spends most of its R&D dollars on development. Government money goes primarily for mission-oriented projects. When Federal agencies pay for R&D on civilian, commercial technologies, they have often made poor choicesparticularly when the R&D goals are well removed from agency missions. Without substantial changes in U.S. technology policies, industry can expect only limited help from the Federal Government in the commercialization of HTS and other new technologies. Recently, some State Governments have been more active than the Federal Government in stimulating industrial technology development. (Options 6, 8, 15, 16, 17.) Federal funding aimed at filling gaps between fundamental research and product/process development could help speed
PAGE 19
12 utilization of new technologies, including HTS. Funding for long-term, high-risk projects with potential commercial applications could be even more important. But policies during the 1980s have been moving in the opposite direction; with the Administration cutting budgets for civilian applied research, the overall thrust of U.S. technology policy has turned away from support for commercial R&D. Instead, the Government has relied on indirect measures for stimulating industrial innovation and technology development. OTAs analysis suggests that the indirect approach, emphasizing measures such as looser antitrust enforcement and stronger patent protection, does not, by itself, go far enough. Direct support for commercial technologies has never had a fair trial in the United States. Indirect measures certainly have a place: for example, incentives for corporate basic research could help U.S. competitiveness. So could a supportive climate for cooperative R&D ventures (and, perhaps, Federal cost sharing). Even so, given that policies such as the R&D tax credit (in place since 1981) have had little apparent effect in filling the holes in the Nations technology base, it seems at least as important for the Federal Government to reconsider direct funding of applied industrial R&D. When it comes to commercial technology development, the needs are two-fold: l support for generic, pre-competitive technologiesthose that can help a wide range of companies compete more effectively, without giving any one of them a big advantage; and l support for long-term, high-risk projects. Much of the generic R&D would be relatively straightforwardincremental research with a strong engineering focus. The long-term, highrisk thrust could be modeled to some extent on the work DARPA undertakes for the military. FEDERAL GOVERNMENT STRATEGIES The last chapter of this report discusses three strategies for commercialization. These strategies imply choices going well beyond the individual policy options referred to above and discussed in detail in chapter 4most of which are discrete and relatively narrow. The first of the three strategiesflexible responsethe current, de facto approach, builds on the proven strengths of the U.S. system. These strengths include diversity in funding and conducting R&D: NSF, with its mandate for financing high-quality university research regardless of field; defense agencies, with their unmatched budgets; national laboratories, reservoirs of skilled professionals. Despite its acknowledged strengths, the flexible response strategy seems unlikely to provide adequate support for HTS. The funding picture summarized above for HTS and discussed in detail in chapter 4 shows the drawbacks of the flexible response approach. Most of the Federal dollars for HTS will go to mission agencies with little experience in commercializationto DoD and DOE. NSFprimary sponsor of untargeted university research in science and engineeringhas not had the money to fund many of the highly rated proposals it has received. No one in Government has an overview of Federal support for HTS. Few mechanisms exist for debating and determining priorities. Congress could, of course, choose a more aggressive response to HTSthe second of the three strategies analyzed in chapter 5. Three elements set this strategy off from the current approach: l l more money to NSF for basic research on HTS (and perhaps for one or more interdisciplinary university centers), an insurance policy against missed opportunities; Federal Government cost-sharing in collaborative R&D programs organized and guided by industry (with the Federal money extending the R&D time horizons, ensur-
PAGE 20
13 ing more support for generic work and high-risk research); and l a working group of experts drawn from universities, industry, and Government to help shape consensus on HTS R&D priorities, and make decisions on Federal costsharing. This second strategy would direct Federal funds into HTS R&D that might otherwise be underfunded, and particularly into industry. The added cost would be modest$20 million or $30 million per year, well spent, should make a big difference. The last of the strategies goes beyond HTS, taking up the question of direct Federal support for commercial technology development. As part of such a strategy, OTA considers the merits of increased funding for engineering research, along with the advantages and disadvantages of a Federal technology agency. The analysis emphasizes the problems of defining an acceptable mission for such an agencyone charged with supporting industrial technologies and of avoiding specialinterest hand-outs. Without a mission statement that can impose discipline over the agencys decisions, both day-to-day management and the establishment of broad priorities pose real difficulties. Nonetheless, a Civilian Technology Agency might be able to provide useful support for commercialization if its activities were centered on generic R&D, intended to fill holes in the Nations technology base, and on a menu of long-term, high-risk projects. The three strategies in chapter 5 are by no means exclusive of one another. As Federal policies shift in response to the new competitive circumstances of American industry, and as the science and technology of superconductivity continue to evolve, Congress and the Administrationalong with private industry will need to remain flexible and open to new ideas. Technological innovation may demand policy innovation. Uncertainty makes planning difficult for both public and private sectors one of the reasons for a strategic framework to aid in the many decisions that lie ahead.
PAGE 21
Chapter Commercialization Government and Industry
PAGE 22
CONTENTS Page Summary .. The Government Role . ., Support for Industry: Direct and Indirect . R&D Funding and Objectives. . ., . Commercialization l Four Examples l Inputs to Commercialization: Technology and the Strengths, Weaknesses, and Strategy . . . Product/Process Strategies +. . . . ...0....,. *, ******.. l .,.,*****, l .*.**,. l ***,**, l *,****. .*, **.*. . . .., ,s .,. .* ...*.* l ******* l ,* .*,** 17 20 20 23 27 27 28 32 33 34 35 35 40 42 44 .**..***,,* *.*...*.,*, *********** Marketplace l 4********* ,0 .99.,... Research, Development, and Engineering: Parallel or Sequential? . Commercializing HTS . ...........0.,,...,.0,.0,,,,,,..,,,.,, Microelectronics, and Other Precedents *******,,**..*..***..*.****, HTS Technologies . ,,*, ***o*,***,**,*,., Concluding Remarks .. Appendix 2A: R&D and Commercialization: Four Examples. . . . Boxes Box Page B. The Reagan Administrations Superconductivity Initiative . . . 24 C. The Discovery of High-Temperature Superconductivity ,.. D. Commercialization in Europe l ,.,, l 26 l 43 *,****** l .**.** Figures Figure 1. U.S. Government R&D by Mission, 1988 . . . . Page 25 28 l ** 30 .. .,.00 0 .* ****. Systems 2. The Process of Commercialization. . . . . . 3. Development Stages for Magnetic Resonance Imaging (MRI) Tables Table 1. R&D Budget by U.S. Government Agency, 1988 . . 2. Distribution of Costs for Development and Introduction of New Products and Processes . . . . ... . Page 25 32 l 36 38 l ****** .. .,.*. l **,.,* 3. U.S. 4. U.S. Strengths and Weaknesses in Commercialization . Advantages and Disadvantages in Commercializing HTS .
PAGE 23
Chapter 2 Commercialization: Government and Industry SUMMARY The United States invents and Japan commercializes. So say some. Is it true? If so, this would suggest not only that American companies fail to capitalize on technologies developed here, but that Japanese firms get a free ride on American R&D. Furthermore, if this has happened in other industries and with other technologies, it could happen with high-temperature superconductivity (HTS). Has American industry really had that much difficulty in commercializationin designing, developing, manufacturing, and marketing products based on new technologies? Yesin some industries and with some kinds of technologies. In other casesfor example, biotechnology or computer softwareAmerican firms continue to do better at commercialization than their overseas rivals. Nonetheless, the competitive difficulties of American semiconductor firms have long since shown that continuing U.S. advantages in high technology cannot be assumed. And sectors like consumer electronics demonstrate that, when it comes to engineering, if not science, Japan has been a formidable presence since the 1960s. Commercialization is the job of the private sector. Government plays a critical role in two respects: 1. R&D funding. Federal agencies will spend some $60 billion on R&D in 1988. Government dollars create much of the technology base that companies throughout the economy draw on. In 1988, the U.S. Government will spend some $95 million on HTS R&D. This is about as much as the American firms surveyed by OTA say they will spend on superconductivity R&D (LTS as well as HTS) in 1988. (See ch. 3, box F). 2. The environment for innovation and technology development. A host of policies ranging from regulation of financial markets, to protection for intellectual property, and education and trainingaffect commercialization by companies large and small. Private firms use scientific and technical resultsmore or less freely available, including knowledge originating overseasin their efforts to establish proprietary advantages. Universities and national laboratories create much of the science base. Some industrial research contributes to the storehouse of scientific knowledge. All three groupsuniversities, government laboratories, industrycontribute to the larger technology base (which includes science but goes well beyond it). Much technical knowledge remains closely heldprotected by patents, by secrecy (classification for reasons of national security, trade secrets), or simply as proprietary expertise. Much proprietary information resides in peoples heads, in organizational routines, management styles, as tacit know-how. Companies also write down some of their organizational knowledge: in product drawings and specifications; in process sheets, manuals, and computer programs for running production lines and entire factories. The manufacturing skills that helped Japanese semiconductor manufacturers outstrip their American competitors depend heavily on proprietary know-how, much of it embodied in the skills of their employeesskills that people often cannot fully articulate or explain. Commercialization of HTS will depend on scientific knowledge, much of this available to anyone who can understand it. It will also depend on know-how, hard-won learning and experiencemaking good thin films, orienting the grains in superconducting ceramics to increase current-carrying capacity. Knowledge of markets will count too. Government contributes directly through support for the technology and science base. 17
PAGE 24
18 Federal agencies may spend their HTS R&D budgets wisely, or not. National laboratories may transfer technologies to the private sectors quickly, or only after long bureaucratic delays. Government policies also affect commercialization indirectly. Patents and legal protection of trade secrets help firms stake out proprietary technical positions. Education and training policies (and immigration policies) affect the labor pool from which companies hire people who can understand the science of HTS, envision new computer architectures based on superconducting electronics, grasp the market opportunities created by the new materials. No one anticipated superconductivity at 90 or 125 0 K. No one can predict what will come next. More likely than not, 5 to 10 years of R&Dmuch of it supported by Federal agencieslie ahead before HTS markets will have much size or begin to grow rapidly. A few niche products could come sooner. So could some military applications. New discoveries could change the picture radically. The ways in which the Federal Government spends its R&D dollars matter right now. Policy makers may have a bit more leisure to review the other channels of policy influence on commercialization of HTS. The stakes are highfor the private sector, and for government decisionmakers. Potential for dramatic breakthroughs, coupled with great uncertainty, makes for difficult decisions. OTA sees no reason to rule out the possibility of room-temperature superconductivity (next month, next year). Room-temperature superconductivityin a cheap material, easy to work withhas almost unimaginable implications. Companies with proprietary technical positions could reap huge rewards. The risks of inaction are high; on the other hand, progress could stall. High expectations and media hype could be followed by disillusionment, difficulty in raising capital, inaction on the policy front. Biotechnology has already lived through several such waves. HTS probably will too. Early applications of new technologies tend to be relatively specialized, of modest economic significance. The public may lose interest, financial markets downgrade the prospects. No one can know, at this point, whether HTS could turn out to be a solution in search of a problem. The laserinvented in 1960never seemed to live up to expectations. And yet solid-state lasers eventually made fiber-optic communications possiblean innovation with vast impacts on a worldwide scale (including, for example, a new source of competition for satellite communications systems). It was not that the possibilities went unrecognized. 1 Prospective applications of the laser to eye surgery and optical communications got immediate attention; but while ophthalmologists quickly began using lasers, little progress was made in communications for 15 years. It took, not only solid-state lasers, but low-loss glass fibers to make optical communications a reality. In the early years of laser technology, no one fully anticipated the possibilities for fiber-optic communications networks; they snuck in through the back door. The same could happen with HTS. One of the tasks for public policy is to bring stability to the early years of new technologies, building a base for later commercialization. Industry will not do this alone, absent the potential for near-term profits. OTAs analysis suggests that commercialization of HTS will proceed somewhat faster than many American companies expect, though not as fast as the Japanese companies that have been making heavy investments seem to anticipate. (Ch. 3 outlines U.S. and Japanese business strategies toward HTS.) As American companies move down the learning curves that mark out accumulated knowledge and experience in HTS, federally funded R&D will provide critical support for the technology base that all firmsbut particularly smaller companies draw from. IThe Maturation of Laser Technology: Social and Technical Factors, prepared for OTA by J.L. Bromberg, The Laser History Project, under contract No. H3-521O, January 1988, pp. 7-9. Theodore Maiman, who built the first laser in 1960, stressed the communications possibilitiesmulti-channel capability, low cost per channelat the press briefing announcing his invention.
PAGE 25
19 The analysis in this chapter leads to the following conclusions: l Small, entrepreneurial firms will be well placed to develop commercial applications of HTS. The conditions are right: a new science-based technology; synergistic links with existing industries, including lowtemperature superconductivity (LTS) and electronics; venture capital for good ideas. But while small companies have been a major source of U.S. strength in high technology, few can assemble the financing, the technological breadth, or the production and marketing capabilities to grow as fast as their markets. l Larger American corporations may find that they are starting out behind some of their Japanese rivals. The new HTS materials are ceramics, Japanese firms have a useful lead in both structural and electronic ceramics. Some of this expertise will transfer to HTS. So will a good deal of knowhow developed for fabricating microelectronic devicesanother field where Japanese firms have demonstrated themselves to be at least as good and sometimes better than American companies. l Processing and fabrication techniques will be critical for commercialization. American companies have fallen down in manufacturing skills across the board; the more heavily process-dependent HTS applications turn out to be, the more difficult it will be for U.S. firms to keep up with the Japanese. l Product as well as process technologies will demand much trial-and-error development. Japanese engineers and Japanese corporations are good at this. American companies are not. To the extent that commercialization of HTS depends on step-bystep, incremental improvementsbruteforce engineeringU.S. companies will be in relatively poor positions to compete. l R&D funded by the U.S. Government will help American companies in commercializing HTS, but the spinoffs from defenserelated R&D may not be large or longlasting if military requirements become l highly specialized and diverge from commercial needs. Indirect policy measuresintended to remove the roadblocks to commercialization and increase the rewards for innovators and entrepreneurscan also help. But the indirect approach alone will not be an adequate response to the coming international competition in HTS. What about U.S. commercialization in generalthe backdrop for the statements above? l l l l Mobility among scientists, engineers, and managers has spurred rapid growth and technological innovation in postwar U.S. high-technology industries ranging from computers and semiconductors (starting in the 1940s and 1950s) to biotechnology (beginning in the late 1970s). Venture capital for small, high-technology firms, likewise, has been a consistent source of competitive strength, one that will continue to work to U.S. advantage in HTS. Many larger American companies have pulled back from basic research and riskier technology development projects. Ease in establishing new small firms compensates in part for these relatively conservative investment decisions; indeed, negative decisions on proposed R&D projects sometimes spawn startups that go on to commercialize new technologies. Some of this will probably happen in HTS. With few American firms self-sufficient in technology, a lack of long-term R&Din the private sector, and managements that look for home-run opportunities rather than building technologies and markets step-bystep, the Federal Government has, by default, become a primary source of support for technology development. As yet, agency missions do not reflect this new role. Despite the onslaughts of foreign firms since the late 1960s, many American companies have not yet made the changes in their own organizations necessary to compete more effectively. Paying little more than lip service to well-known engineering methods such as simultaneous prod-
PAGE 26
-.20 uct and process design, they fail to give manufacturing high priority. Neither managers nor engineers in the United States have learned to take advantage of technologies originating overseas. l Industry cannot justifiably blame the U.S. Government for its failures. Compared with most other industrial economies, U.S. policies create a favorable environment for innovation and commercialization. The indirect policy approach the U.S. Government has traditionally relied on to stimulate innovation and commercialization worked well for many years. Today, with foreign competition stronger than ever before, it seems time to explore new directions. The Federal R&D budget has grown rapidly over the postwar period. Management practices in government agencies, mechanisms for setting priorities, for ensuring an adequate technology base, have not kept pace. The climate for innovation can always be improved, the barriers reduced. But the barriers are low already, and limited scope remains for policies intended simply to unleash American industry to compete more effectively. Indeed, the short-term perspectives of U.S. corporations, many of which have been unwilling to keep pace with foreign investments in new technologies, stem in part from the removal of another set of barriersderegulation in U.S. financial markets. Unless the United States learns to match the kinds of supports for commercialization that have proven effective elsewheretopics treated in more detail in later chaptersonly small improvements can be expected. U.S. industry could fall behind in HTS, and in the uses this new technology will find. THE GOVERNMENT ROLE HTS is fresh from scientific laboratories, but many commercial innovations begin with existing knowledge, gleaned from textbooks, design manuals, the schoolhouse of experience. The work of commercialization centers on engineering: development of new products and new manufacturing processes. Companies support their development groups with marketing people, and in some cases with research. Sometimes new science is part of commercialization, but not always. The process may begin with an idea that is old, but has never been reduced to practice because of gaps in the technology base. The automobile, the airplane, and the liquid-fueled rocket all had to await needed pieces of technical knowledge. The Wright brothers learned to steer and stabilize their flying machine. Despite years of trial and error (and centuries of speculation), they were the first to find a way around these technical barriers. Superconductivity itself, discovered in 1911, has a long history as a specialized field of physics, and a shorter historybeginning about 1960as a technology that private firms sought to exploit. Appendix B, at the end of this report, summarizes the science and technology of superconductivity at both low temperatures (e.g., where liquid helium commonly provides cooling) and high (above the boiling point of liquid nitrogen). Support for Industry: Direct and Indirect What does this have to do with government? Today, governments finance much of the R&D that provides the starting point for commercialization. Companies everywhere start with this publicly available pool of technical knowledge in their search for proprietary know-how and competitive advantage. Second, public policies influence the choices companies make in financing their own R&D, and in using the knowledge available to them. Tax and regulatory policies encourage or discourage investments in commercial technology development. Patents create incentives, high capital gains taxes disincentives.
PAGE 27
21 Smaller companies depend heavily on externally generated knowledge; many manufacturing firms with hundreds of employees have few if any engineers on their payrolls. But if smaller companies have the greatest needs, science and technology move so fast today that big companies also rely heavily on government R&D. Moreover, pressures for near-term profits have forced many larger U.S. corporations away from basic research. In the United States, a few hundred large companies account for the lions share of industry-funded R&Dthree firms (IBM, AT&T, General Motors) for more than 15 percent. Half of all U.S. R&D dollars come from the Federal treasury. The fraction is smaller in most other countries, but in all industrial economies public funds pay for a substantial share of national R&D. The reasons begin with health and with national defense, but competitiveness has been one of the rationales: the first government research laboratories, established in the early years of this century in Britain, Germany, and the United States, were intended to help domestic industries meet foreign competition. Foreign firms have access to many of the results of federally funded R&D, just as U.S. firms can tap some of the technical knowledge generated with foreign government support. Governments seek to use technology policy to help domestic firms compete, while commercial enterprises seek to take advantage of the world store of technical knowledge. Technology policy begins with R&D spendingsetting broad priorities, making funding decisions at the project level, agency management. Other tools include intellectual property protection, which can help domestic firms establish and protect a technological edge. Of course, many countries also provide direct funding for commercially oriented R&D. The U.S. Position in Technology Past OTA assessments have examined U.S. competitiveness in a number of industries, and linked technological position with competitiveness; the most recent found signs of slowdown in U.S. R&D productivity, as well as evidence that newly industrializing countries have made surprising gains in technology. 2 Principal findings from these earlier assessments include: l l l l Technology is vital for competitive success in some industries (including services like banking). In others, it may be secondary. But in all or nearly all sectors, the technological advantages of American firms have been shrinking for years. The United States may be able to retain narrow margins in some technologies. Parity will be the goal in others. Regaining the advantages of the 1960s will, in the ordinary course of events, be impossible. In newer technologies, those that have developed since the 1960s, the Japanese have been able to enter on a par with American firms, and to keep up or move ahead. Examples include optical communications, and both structural and electronic ceramics. European firms, in contrast with the Japanese, have had trouble turning technical knowledge into competitive advantage. Today, U.S. military and space expenditures yield fewer and less dramatic spinoffs than two decades ago. The U.S. economy is vast and diverse. Defense R&Dincreasingly specialized when not truly exotic cannot provide the breadth and depth of support needed for a competitive set of industries. Japan and several European countries place higher priorities on commercial technology development than does the United States. R&D spending by Japanese industry reached 2.1 percent of gross domestic product in 1986, compared with 1.4 percent here. Productivity, Innovation, Competitiveness, Commercialization Import penetration in steel and consumer electronics, going back two decades, marks the beginnings of the wave of concern over lagging zzn~erna~jona~ competition h Services (Washington, DC: Office of Technology Assessment, July 1987), ch. 6. Also see Development and Diffusion of Commercial Technologies: Should the Federal Government Redefine Its Role? staff memorandum, Office of Technology Assessment, Washington, DC, March 1984.
PAGE 28
22 U.S. productivity growth and competitiveness. Commercialization is simply the latest catch phrase for problems that are all of a piece. The ongoing policy debate has centered on the proper mix of policies in the United States, where government has been reluctant to intervene as directly or as deeply in the affairs of industry as, say, in Japan or France. During the Carter Administration, an interagency task force, supported by a panoply of private-sector advisory committees, labored for 18 months to produce a Domestic Policy Review of Industrial Innovation (DPR). The recommendations included: 3 l l l l l l easier licensing of federally owned patents; stronger ties between universities and industry; help for small, entrepreneurial firms through small business innovation research grants; removal of unnecessary regulatory barriers; signals to industry that antitrust policy did not bar cooperative R&D; tax incentives for R&D and innovation. Plainly, the focus was on indirect policies. In one form or another, most of these steps have been taken. Other recommendations of the Carter DPR, dealing with direct support for technology development, were not implemented. After Congress passed the Stevenson-Wydler Technology Innovation Act in 1980, the Reagan Administration declined to act on the central provisions of the legislation, which called for a network of Centers for Industrial Technology charged with supporting commercial technology development.4 The Reagan White House began its own study of the problems in mid-1983, creating a Commission on Industrial Competitiveness headed by John Young, president of Hewlett-Packard. When the Commission delivered its findings a year and a half later, many of the recommendations were familiar: balance in regulations; better labor-management cooperation; stronger protection for intellectual property. 5 Although its leadoff recommendation called for a new Department of Science and Technology (which got a frosty reception from an Administration committed to scaling back the Federal bureaucracy), the Young Commission, like the Carter DPR, stressed the indirect influences of Federal policies on technology development. The Commission helped turn the spotlight on technology transfer from the national laboratories, and urged use of the tax system to encourage private-sector R&D. During the 1980s, then, the environment for technology development continued to evolve along the lines mapped out by the Carter DPR. Congress included an R&D tax credit in the 1981 tax bill, and extended italthough at a lower level-in 1986. In 1982, Congress passed the Small Business Innovation Development Act, requiring Federal agencies to set aside 1.25 percent of extramural R&D budgets exceeding $100 million for awards to smaller companies. With the executive branch adopting a muchrelaxed enforcement policy for antitrust, the National Cooperative Research Act of 1984 explicitly permitted certain forms of joint privatesector R&D, while limiting private antitrust suits to actual (rather than treble) damages. The Administration also began negotiations with the governments of several foreign countries For a brief summary, see J. Walsh, What Can Government Do for Innovation? Science, July 27,1979, p. 378, together with N. Wade, Carter Plan to Spur Industrial Innovation, Science, NOV. 16, 1979, p. 800. In addition to agency participants, several hundred people from outside government took part in the Carter DPR; for the reports of the private sector committees and subcommittees, see Advisory Committee on Zndustriid Innovation: Final Report (Washington, DC: Department of Commerce, September 1979). 4 Section 6 of the Stevenson-Wydler Technology Innovation Act of 1980 (Public Law 96-480) directed the Secretary of Commerce to provide assistance for the establishment of Centers for Industrial Technology. Section 8 extended this authority to the National Science Foundation, The centers were envisioned as supporting generic technologies at the pre-competitive stage those that could benefit many companies and industries. Commonly cited examples included R&D on welding processes, or on steelmaking. See Implementation of P.L. 96-480, StevensonWydler Technology Innovation Act of 1980, hearings, Subcommittee on Science, Research, and Technology, Committee on Science and Technology, U.S. House of Representatives, July 14, 15, 16, 1981 (Washington, DC: U.S. Government Printing Office); also International] Competition in Services, op. cit., pp. 364-365. 5Global Competition: The New Reality, vols. I and 11 (Washington, DC: U.S. Government Printing Office, January 1985). Most of the 30 members of the Young Commission were businessmen.
PAGE 29
23 where pirating of U.S. intellectual property has been at its worst. At the same time, Federal laboratories particularly those funded by the Department of Energy (DOE)were seeking tighter linkages and better working relationships with private industry. During the early 1980s, the Federal laboratory system had come in for some rather harsh scrutiny. An outside review panel (headed by David Packard, one of Hewlett-Packards founders and a former Pentagon official) called for closer interactions with the private sector, setting the stage for efforts still underway to open up the laboratories and place their relationships with industry on a new footing (ch. 4). Meanwhile, State Governments began taking more active roles in technology policy. President Reagans proposed Superconductivity Competitiveness Act (box B) continues the stress on indirect policies. The draft legislationwhich would further relax U.S. antitrust policy, while extending the reach of patent protection would apply quite generally to U.S. industry: there is little that is specific to HTS. R&D Funding and Objectives If the weight of explicit shifts in U.S. technology policy during the 1980s has been on the indirect side, the direct role of the Federal Government has also changedthough not in the direction of support for commercial technology development. Government R&D has grown under the Reagan Administration, but much of the expansion has been for defense. Support for commercially oriented R&D has lagged, and in many cases been cut back. %ee, for example, P.M. Boffey, National Labs Reel Under Criticism and Investigation, New York Times, Aug. 24, 1982, p. cl. The Packard report, below, appeared as Report of the White House Science Council Federal Laboratory Review Panel (Washington, DC: Office of Science and Technology Policy, May 1983). For more recent perspectives, see F.V. Guterl, Technology Transfer Isnt Working, Business Month, September 1987, p. 44; and E. Lachica, Federal Labs Give Out Fruit of More Research For Commercial Uses, WaJ] Street ]ourna], Feb. 1, 1988, p. 1. Department of Defense (DoD) R&D went from $20.1 billion in fiscal 1982 to $37.9 billion in 1988 (table 1). DoD R&D, plus the defenserelated portion of DOE spending (about half the Departments R&D), account for nearly 70 percent of all Federal R&D (figure 1); the great majority consists of applied research and the engineering of weapons systems. As figure 1 suggests, the U.S. Government has not paid much attention, relatively speaking, to R&D of interest to companies outside the defense, aerospace, and health sectors. And in the 1980s, Federal agencies have backed away even further (e.g., from energy R&D). The Reagan Administration has held that government has no business supporting commercial technology development. Fundamental research, yes, but anything more would be a subsidy unjustified and likely to create harmful economic distortions. The basic research portion of the DoD budget does contribute quite directly to the Nations store of commercially relevant technical knowledge. The Pentagon, for example, provides nearly 40 percent of Federal support for university research in engineering. 7 In constant dollars, however, DoD basic research (budgeted at $892 million for fiscal 1988) remains at roughly the same level as in 1967, while the total DoD R&D budget has been steadily expanding in real terms. Based on 1987 obligations, the Federal R&D budget breaks down as follows into the three broad categories of basic research, applied research, and development: Basic research . . . . . ..$ 8.8 billion Applied research. . . . . 9.0 billion Development . . . . . 38.7 billion $56.5 billion The National Science Foundation follows, at about 30 percent. Universities carry out half of all DoD-sponsored basic research. See Directions in Engineering Research: An Assessment of Opportunities and Needs (Washington, DC: National Academy Press, 1987), pp. 46 and 63. Recently, the military has spent a little more than 2 percent of its R&D budget on fundamental research; 5 percent of private industrys R&D total goes for basic work.
PAGE 30
24 Many agencies subdivide these categories Basic research itself covers a wide range of further. 8 activities. Some of this really is untargeted sciencework that could be called pure re8N0 figures for 1988 were available as this report was being search, Nobody expects that astrophysics or the completed. Distinctions between these categories are necessarily arbitrary; for the Federal Government definitions, see Science Indicators: The 1985 Report (Washington, DC: National Science 6.3 Advanced Development Board, 1985), p. 221. DoD subdivides its R&D budget into six 6.4 Engineering Development subcategories, designated as follows: 6.5 Management Support 6.1 Research 6.6 Operational Systems Development 6.2 Exploratory Development Several of these are further subdivided,
PAGE 31
25 Table 1 .R&D Budget by U.S. Government Agency, 1988 Obligational Percentage of authority total (billions of Federal R&D dollars) budget Department of Defense (military functions only) Department of Health and Human Services . Department of Energy National Aeronautics and Space Administration National Science Foundation . . . $37.9 b 63.20/o 7.2 12.0 5.1 8.5 4.8 8.0 1.5 2.5 All others. . . . . 3.5 5.8 $60.0 100 /0 aExclud~~ $2 billion in obligations for R&D facilities. %he three services expect to commit a total of $29.4 billion in fiscal 1988-$15.2 billion for the Air Force, $9.5 billion for the Navy, and $4.7 billion for the Army. Adding in the rest of the DoD R&D budget (e.g., spending by agencies such as the Defense Advanced Research Projects Agency) brings the total to $37.9 billion. SOURCE: Special Analyses: Budget of the United States Government, Fiscal Year 1989 (Washington, DC: U.S. Government Printing Office, 1988), pp. J-3, J-5. Figure 1. U.S. Government R&D by Mission, 1988 Environment, natural resources, Energy, 2 % National defense, 68% SOURCE: Federal R&D Funding by Budget Function, Fiscal Years 1987-1989, NSF-813-315 (Washington, DC: National Science Foundation, 1988), p 4 Superconducting Super Collider will lead to results of much practical use in the foreseeable future. Understanding is the motive. Other projects, likewise defined as basic research for budgetary purposes, nonetheless bear quite directly on agency missions. Almost all the R&D funded by the National Institutes of Health (NIH, part of the Department of Health and Human Services) could be termed directed research. NIH supports much fundamental sciencee.g., in molecular biologybut it does so with a view toward eventual improvements in health care; many NIH-sponsored projects have quite specific objectives such as a cure for AIDS, or better understanding of the growth of cancerous cells. Likewise, DoD and DOE R&D serve agency missions. Research in physics laid the foundations for nuclear weapons, with DOE inheriting much of the ongoing support for physics from the Atomic Energy Commission. When the armed services or the Defense Advanced Research Projects Agency (DARPA) sponsor work in the behavioral sciences, they seek insights into the responses of fighter pilots to sensory overloads, or knowledge that will help make artificial intelligence a practical tool for battle management. Research carried on in industrial laboratories, almost by definition, has a practical orientation. So does engineering research in universities and nonprofit laboratories. Plainly, distinctions such as that between untargeted and directed research will always be arbitrary. Nonetheless, such distinctions help in thinking about R&D and how it supports commercialization. Within directed research, further distinctions can be made. Incremental work, for example, takes a step-by-step approach toward reasonably well-defined goals. The problems may be technically difficult, but the territory has been at least partially explored. Much of the work on synthesis of new materials that laid the groundwork for the discovery of HTS (box C) falls in this category, as does the many years of R&D aimed at improving the properties of LTS materials. 84.754 () 88 2 : QL 3
PAGE 32
[Page Omitted] This page was originally printed on a gray background. The scanned version of the page is almost entirely black and is unusable. It has been intentionally omitted. If a replacement page image of higher quality becomes available, it will be posted within the copy of this report found on one of the OTA websites .
PAGE 33
27 Most research serves the needs of governpetition have driven the scientific enterprise ment or industry. Military needs, social objecat least since the end of the 19th century. tives such as health care, and industrial comCOMMERCIALIZATION Both industry and government support directed research. Promising results lead naturally into development. Research and development then go on in parallel, with research outcomes suggesting new avenues for development, and problems encountered in development defining new research problems. While the U.S. Government has a long tradition of support for basic research and missionoriented R&D, it usually leaves pursuit of commercial technologies to the private sector. This policy worked well for many years. For instance, continued development of fiber-reinforced composite materialslighter and with greater stiffness, strength, and toughness than many metalsbuilds on a technology base that has been expanding at a rapid rate since the 1950s. 9 The primary stimulus came from the military, where composites found their first applications in missiles, later in manned aircraft. penetration into commercial aircraft followed. When it comes to technologies where Federal agencies have been less active, U.S. firms have often fallen behind. Although the U.S. Government has spent several hundred million dollars for R&D on structural ceramics since the early 1970s (app. 2A, at the end of the chapter), the effort has been a small one compared with fiber composites. Japan, meanwhile, has established a useful lead in structural (as well as electronic) ceramics. In semiconductors, American firms established a commanding lead during the 1950s and 1960s, when military procurement provided much of the demand (ch. 4). In later years, as production swung towards civilian markets, Japanese firms closed the gap. BAdvanced Materials by Design: New Structural Materials Technologies (Washington, DC: Office of Technology Assessment, June 1988). Four Examples In addition to summarizing Federal programs on ceramics, appendix 2A outlines the evolution of the video-cassette recorder (VCR)a quite different case from any of those mentioned so far. The appendix also reviews the development of magnetic resonance imaging (MRI) systems, a relatively new product of the medical equipment industry, and LTS magnets. Magnets wound with niobium-titanium alloy the most widely used LTS conductorfind uses not only in MRI, but in scientific research. The examples in appendix 2A illustrate something of the range and complexity of commercialization. Sometimes government R&D support is critical (LTS magnets), sometimes nearly irrelevant (the VCRalthough much of the underlying technology of magnetic recording did benefit from ongoing government-sponsored R&D). Sometimes governments try to push a technology, to little avail (ceramics for gas turbine engines). For MRI, the major policy impacts had little to do with R&D: commercialization depended on regulatory approvals, as well as Medicare and Medicaid payment policies. The starting point may be new science, creating new opportunities (MRI, LTS magnets), or it may be the prospect of a huge market if development problems can be solved (VCRs). Inter-firm competition may be intense and international (MRI, VCRs), or it may be largely irrelevant (LTS magnets, where much of the work was undertaken within Federal laboratories). Government agencies supported R&D on LTS magnets as part of larger, ongoing programs: high-energy physics research, nuclear fusion. Development of niobium-titanium wire for these magnets has been mostly a matter of painstak-
PAGE 34
28 Photo credit: University of Kansas Medical Center Magnetic resonance image of human face. ing engineering. Federal funds paid for much of the work. 10 The U.S. Government pushed structural ceramics technologies for different reasons. Most recently, DOE has supported work on ceramics for gas turbine engines, hoping to overcome their efficiency limitations; with greater fuel economy (and low enough manufacturing costs), the hope was that turbines could compete with gasoline and diesel engines for automotive applications. While these objectives are consistent with DOEs mission, there has been little pull from the marketplace. @Superconductive Energy Storage, vol. IV, DOE/ET/2660235, Final Technical Report, January 1976 to October 1981, prepared by the Applied Superconductivity Center, University of Wisconsin-Madison, for the U.S. Department of Energy, July 1983, ch. III. Inputs to Commercialization: Technology and the Marketplace Product or process developmentwhether adapting LTS magnet technology for medical imaging systems, or generic techniques for computerized process control to steelmaking depends on at least two inputs from outside the development group itself. The first of these is knowledge drawn from the technology base, including science, engineering, and shopfloor know-how (figure 2). The second input is knowledge of marketswhat potential customers want and need. Steelmaker may improve their process control systems because their customers want better formability, which requires more precise control of melt chemistry. The purchasers of steel maybe seeking to provide Figure 2 .The Process of Commercialization SOURCE: Office of Technology Assessment, 19SS.
PAGE 35
29 their own customers with products (automobile fenders, dishwashers) that have betterlooking painted surfaces. R&D and Marketing Innovations follow their own paths. Figure 3 summarizes the later stages for MRIthose after initial research and feasibility demonstration. Science came first, the complete chronology beginning in 1936 with theoretical predictions of the underlying phenomenon of nuclear magnetic resonance. Experimental demonstrations followed a decade later, with the first twodimensional images (e.g., of a wrist) in 1973. Heavy continuing involvement by physicians and scientists made MRI something of an exception. Normally, commercialization is a job for engineers, supported on the one side by knowledge flowing from the technology and science base, and on the other by information on customer, wants, needs, and perceptions. Much of the early work in HTS will be undertaken by multidisciplinary groups including physicists, chemists, materials scientists, and ceramists, along with electrical, chemical, and mechanical engineers. The known HTS materials are oxide ceramicsbrittle and difficult to work with. Learning to use them means drawing where possible on past R&Dwork undertaken earlier and for other purposes on structural and electronic ceramics, as well as processing, fabrication, and design techniques from microelectronics. As applications come into view, companies will call on marketing tools ranging from feasibility studies (which may include detailed projections of manufacturing costs) to consumer surveys. Technical objectives shift as prospective markets emerge; some firms use technology gatekeepers to help match research results and market needs. This is an area where U.S. and Japanese strategies in HTS contrast markedly, with Japanese companies much quicker to begin thinking about applications and the marketplace (see ch. 3). Judging market prospects can be harder than judging prospects for technical progress. Furthermore, market prospects often depend on technological capabilities. Early efforts by Matsushita and Toshiba to design VCRs for household use failed: production costs were high; recording times were short. Not many people would pay upwards of $1000 for a machine limited to 30 minutes per cassette. But improvement was steady. RCAs VideoDisc died in the marketplace in part because the company miscalculated the speed with which VCR manufacturers could reduce their costs to match RCAs target price (initially, $500 at retail). RCA also underestimated the weight consumers would place on off-the-air recording capability, and, failing to grasp the implications of rapidly growing rentals of videotapes, prohibited rentals of its discs. HTS Markets It is too early to reach many conclusions about markets for HTS. The more obvious highcurrent, high-field applicationsmagnets, electric generators, coil and rail guns for the militaryhave all been analyzed for feasibilPhoto credit: Argonne National Laboratory HTS wire, flexible before firing,
PAGE 36
30 Figure 3.Development Stages for Magnetic Resonance Imaging (MRI) Systems I 1 Technical analyses Preliminary design, Commercialization l Primarily aimed at developing the technology of MRI and, learning to use it (establishing effkacy, subject safety, etc ) *pdm~~ aimed at tearing prototype MRI syetem SOURCE: Baaed on Haalti Tactuwbgy Caaa Sfudy 27: Nuclear Magnetk Rtbsomrrcu knagkrg Tachrmbgy (kfhahington, DC (Mce of Technology Assessment, September 19S4), ch 4 ity, but no one knows much about making practical wire, cables, or current-carrying tapes from the new materials (app. B). These will need higher current densities than yet in view. Good thin film fabrication methods, the precondition for applications to sensors and electronics, will probably be easier to achieve. Even so, as of mid-1988, there had been no public announcements of reproducible HTS Josephson junctions (JJs). Many of the technical questions on which practical applications depend will not be answered until R&D groups learn to fabricate JJs easily. Later sections of the report discuss these technical matters in more detail. Here the point is simply that, until the technological prospects come into sharper focus, it will be impossible to do more than speculate about markets. And even then, uncertainty will remain high. No onescientists, engineers, marketing special-
PAGE 37
31 ists, science fiction writerscan predict with much accuracy how a new technology like this will eventually be applied. Nor can potential customers say what they might want, or be willing to pay, if they cannot imagine the possibilities. It is the unexpected that will probably have the greatest impact. Success and Failure What makes for success or failure in the marketplace? Product and process engineering, marketing skills, luck, sometimes research results. No one has a recipe, any more than a recipe for room-temperature superconductivity. Costs are central for some products, but for others MRI is one competition revolves around non-price features. Many hospitals will readily pay a premium of several hundred thousand dollars for an MRI system with superior imaging performance. At the same time, small private clinics or rural hospitals make up a niche market for which a number of manufacturers have designed low-cost systems. Products may come out too late or too early. A company may fall behind its competitors and never get much market penetration. Early innovators in the semiconductor industry have sometimes failed and sometimes succeeded. 11 The pioneer minicomputer manufacturer, Digital Equipment Corp., whose PDP-8 established this part of the market, went onto become the second largest computer firm in the world, On the other hand, the microcomputer pioneers Altair, Imsai, polymorphic Systemsdisappeared. Toshiba invented helical scan recording but the company ended up licensing Sonys Betamax technology (which itself has lost much ground to VHS). Cost and Risk As firms move further along the development path, mistakes become more costly. Only one often projects launched at the R&D stage ever l~see, for example, lrlter~~tiOrA Competitiveness in Electronics (Washington, DC: Office of Technology Assessment, November 1983), Appendix C: Case Studies in the Development and Marketing of Electronics Products, Semiconductors: The 4K Dynamic MOS RAM, pp. 524-531. brings in profits. Before reaching the marketplace, half of all R&D projects fail for technical reasons; poor management or financial stringencies kill two or three more. Of those that do enter production, some never earn enough to cover development costs. The vast majority of project budgets go for product engineering, process design and development, tooling and production start-up, and test marketing. Introducing an MRI system means investments of $15 million and up for R&D alone; pilot production and field trials require much larger financial commitments. Seldom does research account for more than 10 percent of total project outlays, although the distribution of costs varies a good deal from project to project and industry to industry. The distribution also varies between the United States and Japan. As table 2 shows, Japanese companies (for the industries and time period examined) spent a bit less on R&D than the average American firm, and much less on manufacturing startup and product introduction. They budgeted more in gearing up for productionon facilities, tooling, and special manufacturing equipment (a difference that may also reflect higher projected volumes). Japanese firms no doubt have lower startup costs because they invest more in frontend process development. Yet a substantial difference remains. Adding the percentages for tooling and equipment to those for manufacturing startup gives a total of 40 percent for the U.S. companies, 54 percent for the Japanese. The greater proportion of total project expenses for tooling and equipment reflects the higher priorities Japanese managers place on manufacturing as an element in competitive strategy. Such priorities will make a difference in commercialization of HTS, which will depend critically on process know-how. U.S. firms have underinvested in process technology for years one reason for competitive slippage in industries ranging from steel to automobiles to electronics.
PAGE 38
32 Table 2.Distribution of Costs for Development and Introduction of New Products and Processes a Percentage of total project cost Research, development Prototypeor Tooling and Manufacturing Marketing and design pilot plant b equipment startup startup U.S. companies . . . 26940 17% 23% 17% 17% Japanese companies . . 21 16 44 10 8 asuw~y figureS from 1 g~ for 50 matched pairs of u,S, and Japanese firms, The total of 100 included ~ chemical companies, 30 machinery, 20 eleCtriCd and electronics, and 14 from the rubber and metals industries. bFor cases of ~roduCt develo~rnent, the costs are for prototyping; for process development, they include investments in pilot plWltS. NOTE: Totals may not add because of rounding. SOURCE: E. Mansfield, The Process of Industrial Innovation in the United States and Japan: An Empirical Study, unpublished seminar paper presented Mar. 1, 1988 in Washington, DC. Competitive Advantage What does it take to use technology effectively? The examples mentioned above, and others, point to the following common factors: l Appropriate use of technology and science, new and oldwhether a company generates the knowledge internally, or gets it elsewhere. Much of the science base for HTS will be available to everyone. To establish a competitive advantage, companies will have to develop proprietary knowhow, and do it ahead of their competitors. l Effective linkages between engineering and marketing. Customers for many of the early applications of HTSin military systems, electronics, or perhaps energy storagewill be technically astute. Marketing will count, but not so heavily as for consumer products. l Effective linkages between product development groups and manufacturinga point already stressed for HTS. Managerial commitment to risky and uncertain R&D projects. The next chapter explores this dimension more fully for HTS. What are the conditions under which American firms have trouble in commercialization in the effective utilization of technical knowledge, new or old? Under what circumstances do American firms perform best? Effective policies depend on the answers to such questions. Generally speaking, OTA assessments have found the problems to be most acute when it comes to applications of existing technology by firms in older industriesand particularly when it comes to shopfloor manufacturing technologies. In the earlier years of high technology, the United States had potent competitive advantages: entrepreneurship and venture capital; a decentralized science infrastructure with many centers of excellence both inside and outside the Nations universities; flexible labor markets, with high mobility among engineers, scientists, and managers. These strengths have begun to wane. In many industries, Japanese companies are out-engineering American firms. Even in high technology, the Japanese have been able to move quickly from the laboratory to the marketplace. The days when U.S. companies could take their time in commercializing R&D are past. STRENGTHS, WEAKNESSES, AND STRATEGY What does the discussion above (and in app. problems they failed to anticipate. Powertrains 2A) imply for U.S. abilities in commercializa, wore out quickly in long-distance driving. Comtion? The first point is simply that taking a new panics like Honda found themselves trying to product into the marketplace is always diffisell cars with fenders that would rust through cult. In their efforts to penetrate the U.S. marafter one or two northern winters. Federally ket, Japanese automakers suffered from many mandated recalls were frequent.
PAGE 39
33 Product/Process Strategies Japans automakers overcame these difficulties. They redesigned their products to suit U.S. needs and tastes, establishing deserved reputations for quality and reliability. They built strong dealer organizations that helped them understand American consumers. In contrast to European manufacturers, the Japanese began developing vehicles specially tailored to the U.S. marketsmall pickup trucks, four-wheel drive vehicles, sports and luxury models. Most were variants on products sold in Japan and other foreign markets, but a fewsuch as Nissans Pulsarwere designed primarily in the United States to appeal to Americans. 12 Japans automakers learned many lessons from their American rivals, and learned them well. The credit goes to the industry, which benefited from government policies, but not nearly so directly as, say, Japanese computer manufacturers. In the past several years, with their upmarket moves and new brand names, Japans automakers have taken another leaf from Alfred Sloan: in turning their automobiles into highfashion products, they have introduced new models much more quickly than American or European firmsa necessary capability for implementing such a strategy. Design/development/tooling cycles for Japanese automakers have shrunk to little more than half those in the United States; Hondas model cycle is down to 40 months, while American firms take 5 or 6 years. 13 U.S. automakers have J, Yamaguchi, Quick-change open-top car matches closed coupe in body rigidity, Automotive Engineering, February 1987, p, 167. When Toyota models got poor ratings on U.S. crash tests, the company quickly made design changes that upped their scores L. McGinley, Car Crash Rankings: Safety Guide Or Numbers That Dont Add Up? Wa]] Street ~ournal, Dec. 1, 1987, p. 39. la Hondas R&D Mastermind, Automotive Industries, November 1987, p. 52, More generally, see H. Takeuchi and 1, Nonaka, The new new product development game, Harvard Business Review, January-February 1986, P 137; J. Bussey and D.R. Sease, Manufacturers Strive To Slice Time Needed To Develop Products, Wall Street ~ournal, Feb. 23, 1988, p. 1; R. Poe American Automobile Makers Bet On CIM To Defend Against Japanese Inroads, Datamafion, Mar. 1, 1988, p, 43; K.B. Clark and T. Fujimoto, Overlapping Problem Solving in Product Development, working paper, Harvard Business School, April 1988, Part of the Japanese advantage may come simply from putting looked to computer-aided engineering to narrow the gap. The Japanese, however, appear to succeed through quite conventional approaches to engineering development, carefully managed. Certainly they do not have the lead in such computer-intensive techniques as numerical analysis of vehicle structures, aerodynamic modeling and simulation, or analytical predictions of vehicle ride, vibration, and handling. The high-fashion, product differentiation strategy is new for Japanese companies only in the automobile industry. It is one the Japanese have used in the past in cases like consumer electronics and motorcycles. Successful targeting of marketswhether for consumer goods, for capital equipment (machine tools), or for intermediate products (semiconductor chips)has been a hallmark of Japans competitive success. 14 As discussed in the next chapter, Japanese companies have already put a good deal of effort into thinking about new applications of superconductivity; they may well locate some of the possible market niches before American firms. The Japanese have often carved out substantial markets by starting from small niches; large, integrated Japanese firms have been more more engineers to work: GM, Ford, and Chrysler employ a total of 30,000 engineers, Toyota, Honda, and Nissan more than 40,000J, McElroy, Outsourcing: The Double-Edged Sword, Automotive Industries, March 1988, p. 46. While it takes much longer for American firms to introduce new products in some industries, according to a recent survey, the U.S.-Japan difference in design and development times does not hold across the boardE. Mansfield, The Process of Industrial Innovation in the United States and Japan: An Empirical Study, unpublished seminar paper presented Mar. I 1988 in Washin@on, DC. Professor Mansfields survey does show that Japanese companies were generally much quicker than American firms when product development efforts began with licensed technologies. Moreover, Japanese firms willingly absorb substantially higher costs to shorten their development cycles. W% the Japanese approach to product planning and marketing, see J.K. Johansson and I. Nonaka, Market research the Japanese way, Harvard Business Review, May-June 1987, p. 16; P. Marsh, The ideas engine which drives Japan, Financial Times, May 29, 1987, p. 14; P. Marsh, why research is i n th e driving seat, Financial Times, June 2, 1987, p. 12; C. Lorenz, Serum and Scramblethe Japanese Style, Financial Times, June 19, 1987, p. 19; P.S. Leven, Repatriate product Design, Across the Board, December 1987, p. 39; C. Rapoport, How Honda research runs free and easy, Financial Times, Feb. 16, 1988, p. 10.
PAGE 40
34 aggressive than their American counterparts in pursuing specialized products, including advanced materials. Japanese companies are willing to start with small-volume production and grow with their marketsa strategy likely to prove successful in HTS, indeed one that may prove necessary. How do companies based in Japan do so well at defining and attacking market segments, particularly in countries foreign to them? Most Japanese companies do use market research techniques, although table 2 showed they spend less on this than American companies. As some U.S. firms also realize, the best marketing research often remains as informal today as it was 50 years agoa matter of good judgment from within the company more than consulting firms, focus groups, and consumer surveys. Japanese firms in many industries have also capitalized on the quality of their goods. Lagging quality not only leaves customers unhappy, it raises manufacturing costs. Quality and reliability problems have plagued American industries ranging from automobiles to semiconductors. Careful control of the production process will be necessary for fabricating the new HTS materials, as it is for high-technology electronic and structural ceramics, or for integrated circuits. The primary point is this: by the 1960s, American firms had come to think of their skills in engineering and marketing as far and away the best in the world. If this was true then, it is true no longer. Many U.S. companies have not yet faced up to the need to do better. Others periodically rediscover such well-known management and engineering practices as simultaneous engineering, design for production, or quality engineering, but fail to follow through with actions that institutionalize them. Some still look to techniques like quality circles for miracle cures. Research, Development, and Engineering: Parallel or Sequential? Simultaneous engineering means nothing more than tackling product and process development in parallel, with overlapping responsibilities in design and manufacturing groups, if not a fully integrated approach. Simultaneous engineering may be hard to achieve in a modern American corporation, but in principle is nothing but common sense. A hundred years ago, technology was simpler and no one had discovered any need to separate design and manufacturing. The chain can be extended back to research. But given the uncertainties that accompany the search for new knowledge, and the high costs of downstream development, many U.S. executives have come to view research, development, and product planning as sequential processes. Only when consistent, verifiable, and potentially useful results begin to emerge from the laboratory do American companies think about incorporating engineers into the effort. Even at this point, research may remain separated from development: the scientists pass along their findings, but the two groups continue to work independently. Under these circumstances, the entire process can become almost purely sequentialrunning from applied research to product planning and development to manufacturing engineering, with little overlap. Technology-based Japanese companies, in contrast, have developed simultaneous or parallel processes to a high level. Many are now busy integrating backward into researcha task they see as necessary for commercializing high technologies like HTS. Already, they do a better job of responding to design and marketing requirements through incremental, applied research. Japanese managers, moreover, tend to be optimistic about research in general and about HTS specifically (ch. 3). Perhaps because they mix development and engineering personnel into project groups at an early stage, the belief seems pervasive that useful results of one sort or another will inevitably emerge from HTS R&D. Japanese managers have strong convictions on these matters. They believe it wrong to think about technical developments as proceeding more-or-less linearly from basic research to applied research, then to development and product design, and finally to process engi-
PAGE 41
35 neering. More to the point, they are acting on these beliefs in HTS. American managers know just as well that many of the steps should take place in parallel. But for reasons ranging from trouble in learning to manage parallel processes effectively (one reason for longer product development cycles), to the characteristics of U.S. financial markets, they do not always act on this knowledge, When it comes to HTS, American managers have been relatively cautious; they want to see results from the laboratory before taking the next step. COMMERCIALIZING HTS There is a bright side. The United States retains major sources of strength in commercializing new technologies. Table 3which draws heavily on past OTA assessments of competitiveness summarizes advantages and disadvantages of U.S. firms. Table 4 outlines the implications for HTS. Later chapters expand on many of the points in these two tables, particularly where the Federal Government has policy leverage. Table 3 has a simple message: the United States has a number of areas of advantage, coupled with several serious handicaps. Those handicapsemphasis on short-term financial paybacks, low priorities for commercial technology development and for manufacturing have put U.S. firms at a severe disadvantage in competing with Japan. Some of the consequences can already be seen in HTS. On the other hand, American firms have often been successfulat least in the pastwhen new science has led to new products and new industries, especially where fast-growing and volatile markets promise rich rewards (table 3, factor 1). American companies perform less well, and often poorly, at incremental innovationmore-or-less routine improvements to existing products and processes. These kinds of problems have been much more prevalent in steel than in chemicals, in machine tools than in computer software, in automobiles than in commercial aircraft. Most of the success stories came in the years before U.S. industry had much to worry about from international competition. Table 4 summarizes the lessons that past performance and events thus far hold for HTS, and compares the strengths and weaknesses of American companies with those in Japan. Some of the U.S. entrants will be new companies, started specifically to exploit HTS and staffed by people with strong credentials in related fields of science and technology. Other firms will move in from a base in LTS. Both kinds of companies should be able to respond effectively to the problems and opportunities that emerge in the early years of HTSwith good ideas and a strong science base, together with venture capital and entrepreneurial drive, leading to success in specialized products and niche markets. The picture could change as the technology stabilizes and financial strength becomes more important. When production volumes increase, manufacturing capabilities will grow more important. Companies will have to carefully tailor products to emerging markets, and find capital for expansion. U.S. industries that flourished as infants have run into difficulty as competitorsprimarily the Japanesecaught up and pulled ahead in the race to capitalize on new approaches to factory production or new knowledge concerning electron devices; in the years ahead, the biotechnology industry could stumble, just like the semiconductor industry. 15 Microelectronics, and Other Precedents A decade ago the semiconductor industry still seemed a bastion of U.S. strength. The Japanese were nibbling at the margins, no more. Today, the Federal Government finds itself putting money into the new consortium Sematech, trying to help American firms regain a technological lead lost seemingly overnight. 15s0 far, however, there has been little sign of such slippage. See New Developments in Biotechnology: U.S. Investment in Biotechnology (Washington, DC: Office of Technology Assessment, July 1988).
PAGE 42
36 Table 3.U.S. Strengths and Weaknesses in Commercialization U.S. strengths U.S. weaknesses Comments Factor 1. Industry and market structure: market dynamics. In the past, U.S. firms performed well in rapidly growing industries and markets, especially during the early stages in R&D-intensive industries. American companies have had trouble coping with slow growth or contraction. Although new technologies promising greater productivity might improve competitive Performance in inOther countries frequently look to public policies to help companies and their employees adjust to decline. dustries like steel, corporate executives frequently choose to invest in unrelated businesses. Where foreign firms might take a more active approach to managing contraction, American companies sometimes let troubled divisions struggle along, without new invest. ment, until profits disappear. Then they shut the doors. Factor 2. Blue and gray collar labor force. High labor mobility helps American companies Many development projects depend on craftsIn the past, U.S. wage rates worked to the disattract the people they need. men who can fabricate prototypes and modify advantage of American firms, while creating inthem quickly based on test results and field excentives for investments in R&D and new manperience. In some U.S. industries, shortages of ufacturing technologies that could raise skilled Iabore.g., technicians, modelmakers productivity. Today, international differences in have begun to slow commercialization. labor costs are less of a factor than in the U.S. apprenticeship programs have been in de1970s. cline. Vocational training reaches greater fractions of the labor force in nations like West Germany; large Japanese companies invest more heavily in job-related training for blueand gray. collar employees than do American firms. Factor 3. Professional and managerial work force. Mobility among managers and technical proAmerican companies underinvest in process fessionals has stimulated early commercial(as opposed to product) technologies. This is ization in high-technology industries. New part of a bigger problem: too many managers products have reached the marketplace more and engineers in the United States avoid the quickly because people have left one company factory floor: and started another to pursue their own ideas. l for managers, marketing or finance has been Deep and well-integrated financial markets the road to the top. e.g., for venture capitalhave helped. s engineersschooled according to an applied science modelhave been insensitive, not only to role of manufacturing, but to the significance of design and marketing. Put simply, the engineering profession has divorced itself from the marketplace, and the needs and desires of potential customers (particularly when it comes to consumer products). Compounding these problems, many American companies underutilize their engineers. Finally, many U.S. firms provide little support for continuing education of their technical employees. Managers and professionals in the United States sometimes place individual ambition over company goals. Competition among individuals may make cooperation within the organization more difficult (e.g., between product engineering and manufacturing). More upper level managers in Japanese and West German firms have technical backgrounds than in the United States; they appear more sensitive to the strategic significance of manufacturing, and in at least some cases to new technological opportunities. Factor 4. Industrial Infrastructure (also see Factor 6 below). American companies can call on a vast array of U.S. competitiveness in capital goods like maAt present, the independent computer software vendors, suppliers, subcontractors, and service chine tools has slipped, compounding the proband services industry is perhaps the preemifirms for needs ranging from fabrication of proIems in manufacturing technology. nent illustration of U.S. infrastructural strength. totypes to financing, legal services, and mar Arms-length relationships between American keting research. Few other countries have a firms and their vendors and suppliers may not comparable range Of capabilities so easily be as conducive to commercialization as the available. a relationships found in Japan (relations which might be classified as close and cooperative, or perhaps with equal accuracy as coercive and dependent). % the importance of specialty firms for the U.S. microelectronics industry, particularly those supplying semiconductor manufacturing equipment, see hrterrrat/orra/ CornpetWveness In Electrons (Washington, DC: November 19S3), PP 144-145. On sewice inPuts, see Irrternatlonal Conwet/t/on In Services (Washington, DC: July I@~, pp. 32-34 and 55-57.
PAGE 43
37 Table 3.U.S. Strengths and Weaknesses in CommercializationContinued U.S. strengths U.S. weaknesses Comments Factor 5. Technology and science base (also U.S. strength in basic researchboth science and engineeringhas been a cornerstone of commercialization. The national laboratory system is a major resource, although one that has not been turned to the needs of industry. Multidisciplinary R&Dessential in industrial (and government) laboratorieshas been the exception rather than the rule in American universities. Foreign university systems, however, have probably been even worse at multidisciplinary research. see Factor 7 below). U.S. strength in basic research has not always been matched by strength in applied research, nor in the application of technical knowledge. The Nation depends heavily on a relatively small number of large corporations for industrial R&D and the development of new commercial technologies. When R&D is not close enough to anyones interests, gaps open in the technology base. Moreover, U.S. firms seem to be falling behind in their ability to move swiftly from the R&D laboratory to the marketplace. Diffusion of technology within the U.S. economy has been a persistent and serious problem. American engineers and their employers have often remained unfamiliar with technologies developed elsewhere, reluctant to adopt them. This reluctance is evident, for instance, when it comes to rules of thumb and informal proceduressanctioned by experience if not by scientific knowledge. Examples include shopfloor practices for job scheduling and quality control. The science base and technology base are not identical. The latter spreads much more broadly, encompassing, for instance, the intuitive rules and methodsmany of them tacit rather than formally codifiedthat lie at the heart of technological practice. The semiconductor and biotechnology industries have both sprung from scientific advances. But the theoretical foundations for each remain relatively weak. As a result, progress depends heavily on experience and empirical know-howagain, part of the technology base but not the science base. Japanese and German firms give commercial technology development higher priorities. Governments in these countries also give more consistent support to generic, pre-competitive R&D. Factor 6. The business environment for innovation and technology diffusion (also see Factor 7 below). ---Clusters-of-knowledge and skills such as found in the Boston area, or Silicon Valley, help speed commercialization. While some of this entrepreneurial vitality can be linked to major research universities, other regions have become centers of high-technology development even though lacking well-known schools like MIT or Stanford. The size and wealth of the U.S. market, and the sophistication of customersespecially business customerswork to the advantage of innovators; indeed, foreign companies sometimes come to the United States simply to try out new ideas. Many American firms seem preoccupied with home runsmajor breakthroughs in the marketplaceunwilling to begin with niche products and grow gradually. Poor labor relations sometimes slow adoption of new technology. Reluctance among American engineers and managers to learn from shop-floor employees hurts productivity and competitiveness. Companies in other parts of the world may be somewhat more willing to cooperate in R&D. Linkages between universities and industry could be stronger, but nonetheless probably function better in the United States than elsewhere. Business and consumer confidence encourage innovation and rapid commercialization. Over the past few years, business confidence appears to have ebbed somewhata casualty of Federal budget deficits, trade imbalances, rapid exchange rate swings, and the evident inability of the Government to address these issues. At the same time, the political stability of the United States remains a major strength. Factor 7. The policy environment for innovation and technology development. The United States h-as a deeply rooted commitment to open markets and vigorous competition. (So does Japan, when it comes to domestic markets and domestic competition.) With widespread economic deregulation since the early 1970spIus a tax system and financial markets that reward entrepreneursstartups and smaller companies have often been leaders in commercializing new technologies. Purchases by the Federal Government have stimulated some industries, particularly in their early years. Examples range from aircraft and computers to lasers and semiconductors. A broad range of other U.S. policiese.g., strong legal protections for intellectual property-helps companies stake out and exploit proprietary technological positions. Deregulated U.S. financial markets bear some Many government policies act on commercial iof the blame for the risk aversion and short-term zation indirectly. Industries have evolved in decisions common in American business. different ways in different countries, in part beSometimes U.S. regulatory policies delay comcause of these influences: mercialization. Examples include approvals for Along with antitrust, financial market drugs and pharmaceutical products. regulationse.g., rules covering holdings of stock in one company by othersaffect the extent of vertical and horizontal integration. l Tax policiestreatment of capital gains, R&D and investment tax creditsinfluence corporate decisions on investments in new products and processes. l Antitrust enforcement helps set the environment for inter-firm cooperation in R&D. Trade protection can reduce the risks of new investment, thereby stimulating commercialization. On the other hand, protected firms may grow complacent and decline to invest in new technologies. l Technical standards sometimes act to speed the adoption of new technologies. If premature or poorly conceived, however, they can impede commercialization. l Education and training have enormous long run impacts on commercialization and competitiveness. SOURCE: Office of Technology Assessment, 1988.
PAGE 44
38 As yet, no one knows very much about the technical problems that will have to be overcome in commercializing the new superconductors. Still, parallels have begun to emerge. In microelectronics, product and process knowhow are closely tied. 16 This will also be the case in HTS, where the companies that move down learning curves the fastest will reap competitive advantages. Semiconductor firms must grapple with difficult technical problems in the heat of fierce competitive struggles: understanding the effects of purity and defect population in the silicon IeInternatjona] competitiveness in Electronics, OP. cit., ch. e. As the example of Trilogy Systems illustrates, firms must be able, not only to design, but to build new types of devices; Trilogy had to abandon its planned line of computers after finding it could not fabricate the wafer-scale integrated circuits required. crystals with which production begins; process variables for the steps in diffusion or formation of oxide layers. Costs depend on yield the fraction of functional chips produced. Both yield and quality depend on the design of the chip as well as control of the manufacturing process. With the technology of semiconductor devices ahead of the underlying science, chip designers and process engineers must proceed on a largely empirical basis as they work toward ever denser and more powerful circuits. New applications of HTS will likewise require tailoring of material properties on a microscopic scale, probably without much theoretical guidance. Companies in the semiconductor industry must solve problems today so they can compete in the marketplace tomorrow. HTS is not Table 4.U.S. Advantages and Disadvantages in Commercializing HTS U.S. advantages U.S. disadvantages Comments Factor 1. Industry and market structure; market dynamics. New science and technology make for condiAt some point, financing constraints may make Past U.S. successes in high technology came tions under which American firms should be it difficult for startups and smaller U.S. compawhen international competition was a minor able to commercialize quickly and compete efnies to continue in HTS on an independent bafactor. Foreign firms have now proven they can fectively. sis. Mergers may be necessary for growth. move quickly from the R&D stage to the market place. Mergers or other arrangements driven by financing needs sometimes help, sometimes hurt. Ties with larger companies may stifle innovation. In biotechnology, linkages between small firms and larger companies have helped with regulatory approvals and process scaleUPS American semiconductor firms, however, have seldom been willing to sacrifice their independence for new capitalone reason they have fallen behind large, integrated Japanese competitors. Factor 2. Blue and gray collar labor force. Some American companies start with a core of Japanese companies with ceramics businesses So far, few American ceramics firms have been employees having experience in low-temperacan draw on larger numbers of people with releprominent in HTS R&D. ture superconductivity (LTS). A portion of these vant skills. These employees will help give Jaskills will translate to HTS. At the same time, pan a head start in certain kinds of HTS R&D given that the new HTS superconductors are e.g., mechanical behavior, processing and fabrifundamentally different materialsceramics cation. Japanese firms also have many workers rather than metalsa wide array of quite differwith extensive and transferable experience in ent skills will be needed. Some of the skilled microelectronics. employees may come from related industries, including electronics. Factor 3. Professional and managerial work force. Managers, engineers, and scientists moving Decisionmakers in American companies, large At least initially, HTS startups will have into U.S. HTS companies from industries like and small, may not be willing or able to make managerial staffs with strong technical backmicroelectronics will bring new insights and long-term commitments to HTS-related R&D, grounds. Some larger U.S. firms with the renew ideas. particularly more basic work. sources to compete in HTS-related markets
PAGE 45
39 Table 4.U.S. Advantages and Disadvantages in Commercializing HTSContinued U.S. advantages U.S. disadvantages Comments Processing and fabrication will pose difficult technical problems, of a sort that American companies have not been very good at solving. Much of the R&D needed to develop HTS will be empirically-based engineering, with heavy doses of trial and error. Japanese companies do very well at this kind of development, often better than their American counterparts. may chose other investments because managers fail to understand the technology or recognize the opportunities. Managers with previous experience in LTS may tend to err on the side of conservatism. On the other hand, HTS has had more than its share of exaggerated publicity already. A cautious view of HTS, born of past experience in LTS, could prove realistic. Factor 4. Industrial infrastructure (also see Factor 6 below). The generally strong U.S. infrastructure for high When it comes to the science and technology Japans HTS infrastructure exists mostly inside technology should be an advantage in HTS. of ceramics, specifically, the U.S. infrastructure large, integrated companies. In the United is weak. American HTS companies with States, startups will have to rely heavily on help ceramics-related technical problems may have from outside. The US. approach has advantrouble finding help. tages in flexibility and creative problem-solving, while Japans reliance on internal resources creates reservoirs of skills and expertise that will be very effective over the longer term. Factor 5. Technology and science base (also see Factor 7 below). Despite lack of attention to ceramics compared Military and civilian applications of HTS will diIn 1966, U.S. engineering schools granted 3700 with Japan, the United States has a relatively verge rapidly, limiting the spillover effects from PhDsbut only 14 in ceramics. strong base in materials R&D. DoD R&D spending. Without major policy shifts, Federal agencies In the early years of HTS development, the dewill fund little R&D that directIy supports comfense emphasis of federally supported R&D will mercialization. Nonetheless, the United States work in some ways to the U.S. advantage. Fundis beginning to address the problems of transing from the Department of Defense (DoD) will ferring federally funded R&D to industry. help train engineers and scientists, and may American companies will probably be at a dissupport the development of some dual-use HTS technologies (e.g., powerful magnets). DoD advantage for years to come in solving the support for processing R&D could be especially manufacturing-related problems of HTS. To important. make progress here, American scientists and engineersincluding those engaged in univerA number of national laboratories have the resity researchmust be willing to spend more sources, including specialized equipment, to of their time working on industrial problems help with the technical problems of HTS. (even if the scientific and university communities continue to view practical work as less than fully respectable). Without substantial efforts in manufacturing R&D, some American companies could be forced into partnerships with Japanese firms simply to get access to processing know-how. Factor 6. The business environment for innovation and technology diffusion (also see Factor 7 below). U.S. markets should prove receptive to new products based on HTS. Some foreign companies could find they need an R&D presence here simply to keep up. With Japanese firms starting on a par with American companies, know-how from abroad may prove essential for keeping pace. Many American companies have been unable or unwilling to reach useful technology transfer agreements with Japanese firms. Lack of experience in doing business with the Japanese could become a significant handicap in HTS. University-industry relations in the United States seem to be following patterns similar to those in biotechnology, with strong and productive linkages developing. Small U.S. firms have begun devising strategies for commercializing HTS. Many larger American firms with the resources to compete in HTS, however, seem to be adopting a wait-andsee attitude. Factor 7. The policy environment for innovation and technology development. So far, the U.S. policy approach seems conAfter the initial announcement of the AdminisWhile Federal procurements helped the U.S. ducive to entrepreneurial startups in HTS. trations 1 l-point superconductivity initiative, semiconductor industry get off the ground in There is little indication that the 1966 changes little was heard for 7 monthsa long time in the 1960s, poor experience with demonstration in U.S. tax lawwhich increased rates on capsuch a fast-moving field. Budgetary uncertainprojects in energy and transportation has ital gainshave choked off funds for HTS ties, moreover, delayed decisions on Federal soured prospects for some kinds of policy startups. R&D funding well into the 1966 fiscal year, hamoptions that otherwise might provide stability pering progress in universities, industry, and and support for HTS during a long period of the national laboratories. gestation. Some companies continue to express concern that U.S. antitrust policies will limit opportunities for consortia and other forms of joint R&D. However, OTA has not learned of any case in which U.S. antitrust enforcement has in fact stopped firms from cooperating in R&D. SOURCE: Office of Technology Assessment, 1988.
PAGE 46
40 yet at this stage. There is no market. The race is still a scientific race. But if HTS lives up to expectations, some of the history of microelectronics may be replayed. Commercialization, indeed, may begin with specialized electronic devicesperhaps very sensitive detectors of electromagnetic signals, or high-speed digital circuits (app. B). HTSbased devices maybe used in conjunction with semiconductors. Other parallels are nontechnicalmatters of industrial structure, corporate decisionmaking, and public policy. Relatively small U.S.-based semiconductor firms find themselves competing with vertically integrated Japanese multinationals, enterprises with far more money and manpower. These same Japanese firms have made heavy commitments to HTS R&D. Government policies for HTS in Japan, while far removed from the (false) stereotype of industrial targeting, show many familiar features: notably, pragmatic attention to bottlenecks that might slow commercialization by Japans very aggressive private sector. The Japanese firms that have made so much progress with electronic and structural ceramics will be well placed when it comes to fabricating wires, cables, tapes, and other forms of conductors made from the new HTS materials. Learning to make practical conductors from the new materialsfor the circuitry inside computers, or for electrical windings in generators or energy storage systemswill require a great deal of trial-and-error development. Japanese companies do well at this kind of engineering. Some of the specific skills in ceramics processing they have developed will transfer, just as will some of their skills in semiconductor processing. American firms, in contrast, have fallen down badly in processing and manufacturing skills over the past two decades. HTS Technologies Appendix B outlines prospective applications of HTS (table B-l), including estimated time frames for commercialization (table B-3). Early applications of HTS will be highly specialized military equipment, niche markets on the commercial side (perhaps in scientific apparatus, or for nondestructive inspection). Japanese firms will provide strong competition from the beginning. High-Current, High-Field Applications; Electrical Machinery and Equipment Japans lack of energy resources means strong motives for commercializing HTS in order to conserve electrical power. Even though superconductivity offers relatively small efficiency increments (because large-scale conventional equipment is highly efficient already), Japanese companies may make more rapid progress than American firms in superconducting motors and generators, as well as transformers and energy storage systems. Similar forces are at work for magnetically levitated trains, where the motivation comes from a heavy existing commitment to fixed-rail passenger transportation natural in a small and crowded nation like Japan. Summarizing: l l l l Both the United States and Japan start from a roughly equivalent experience base in LTS motors and generators, but the Japanese have more work underway at present, and will probably pull ahead when and if suitable HTS conductors become available. Each country has one or more active LTS energy storage projects (large superconducting rings in which electrical current circulates indefinitely, to be withdrawn when needed). With SDI funding, two U.S. firms are designing a prototype ring that would quickly dump the stored energy into powerful lasers. Japans R&D has been directed at storage for electric utilities, where discharge rates will be much lower. While the U.S. effort will yield some lessons for utilities, design trade-offs will bias the prototypeand the knowledge gained from ittowards the quick-discharge military application. DoD R&D aimed at coil and rail guns and other high-power, high-field applications (e.g., ship propulsion) could strengthen the generic technology base in HTS, helping commercial industries indirectly. When it comes to possible applications such as magnetically levitated trains, the
PAGE 47
41 United States starts out behind, having halted R&D in 1975 (see box K, ch. 3). However, it is not yet clear that HTS would offer much advantage here. l In medical electronicse.g., MRIthe United States has a substantial lead in know-how and experience, one that should persist (although LTS might again continue to be the technology of choice for some time). Pursuing most of these applications will demand technical resources and experience, as well as financing, on a scale beyond that of the small, entrepreneurial firms that emerged in the early years of LTS, and those being started today to exploit specialized HTS applications. If big U.S. companies prove reluctant to move into markets for electric power equipmentand smaller entrants cannotintegrated foreign producers will probably take the lead, internationally and perhaps in the U.S. market. Superconducting Electronics Progress in thin films for electronics should be more rapid than for the conductors needed in high-power applications. When it comes to electronics, the Japanese will probably benefit to some extent from R&D on Josephson-based computers; government and industry in Japan continued work on JJs for computing after U.S. companies dropped most of their own activity (see box J in the next chapter). Josephson junctions, however, function only as two-terminal devices, weak amplifiers at best. No one knows how to make useful threeterminal devices like transistors from superconducting materials. A practical three-terminal superconducting device, even one restricted to liquid helium temperatures, could open up a broad range of opportunities. Whether this will be possible is an open question. JJs also make for highly sensitive detectors of infrared and other electromagnetic radiation. LTS sensors and in the future perhaps HTS sensors (e.g., for satellites, where passive cooling should keep operating temperatures below the transition temperatures of the new materials)have potentially important military applications. As a result, DoD has funded a good deal of R&D over the years on these devices, as well on superconducting components for very powerful computers. A S DoD R&D increasingly focuses on HTS, some of its workperhaps in sensors will contribute to commercial spin-offs. Japanese companies will prove able competitors over the long run in both devices and systems applications of HTS. In their efforts to catch up with IBM and other U.S. computer firms, Japans integrated manufacturersseveral of which make chips, computers, and telecommunications hardwarehave been spending heavily on R&D for years. They are seeking a technological window that would help them overtake the United States in hightechnology electronics, and particularly in computersa field where American firms remain broadly superior. The Japanese see HTS as a possible window. Smaller American firms will probably find electronics markets attractive. Hypres, for example, founded by an ex-IBM physicist after the computer manufacturer scaled back its LTS JJ computer project in 1983, introduced a very high-speed LTS-based data-sampler in 1987. The company hopes its experience base will give it advantages in HTS. Other small LTS specialists also plan to move into HTS by building on their past work with the older materials.
PAGE 48
42 CONCLUDING The next chapter looks in some detail at corporate strategies toward HTS in the United States and Japan. European countries, too, have excellent science and engineering capabilities in both private and public sectors. But longstanding problems in capitalizing on these strengths suggest that European firms will not be able to keep up in the race to commercialize HTS. Box D summarizes the reasons. Companies everywhere look for proprietary advantages from R&Dpatentable inventions, expertise they can protect through trade secrets. Semiconductor companies, for example, each have their own process technology. Much of the information is closely held; some of it is embodied in the skills of their employees. In LTS as well, proprietary know-how helped small companies stake out positions in the manufacture of wire and in specialized electronics. The Japanese developed a great deal of proprietary technology in commercializing the VCR. The story is one of Japanese success in innovationengineering design and development, market research, mass production manufacturing. But in related markets like personal computers there is little evidence of slippage, despite many past predictions of a Japanese takeover. Nor have the Japanese been able, for instance, to move from success in high-density memory chips to microprocessors. U.S. leads in computer software, or biotechnology, may have narrowed in the last 5 years, but not by much. Japans bet on structural ceramics may not pay off; the technical problems of achieving reliability in very brittle materials could prove too difficult. Scientific knowledge and technological understandingnot the sameinteract throughout such development efforts. Sometimes new science leads to new technology. This has been the case in superconductivity, beginning with its discovery in 1911, but especially since the 1960s. In other cases, demand for new technology spurs scientific advance. Much military R&D works this way. Corporations are more likely to invest their own money in R&D, and take the risks of commercialization, if they expect rapid market growth. Government policies can reduce these risks. Trade protection does so, along with financial subsidies, and government purchases of a companys products. Strong legal protections for proprietary technology make R&D more attractive. Some governments go so far as to give financial help to customers for new technologies (computers and industrial robots in Japan). But with new knowledge eventually becoming available everywhere and to everyone, those who use it fastest and most effectively will come out ahead in international competition. U.S. industry has been falling behind in the use of new technical knowledge, in part because of slow product development cycles. In many fields, the Japanese are not only doing a better job of engineering than their American rivals, they are doing it faster. Speed in moving from research to production and the marketplace will be a major factor in competitive success in HTS, just as in industries like automobiles or semiconductors. American firms have also had trouble as production volumes rise, and been poor at incremental product/process improvements. Their production capabilities enabled Japans semiconductor manufacturers to establish themselves in world markets and compete successfully with American firms that had the lead in many of the functional aspects of circuit design. Many of the manufacturing techniques needed for HTS electronics will be similar to those for semiconductor devices (and ceramics). Still, at the level of R&D and product development teams, Japanese firms do not seem to operate in greatly different fashion from successful American companies. The differences that do exist are important, however: 17 ITK. Imai, I. Nonaka, and H. Takeuchi~ Managing the New Product Development Process: How Japanese Companies Learn and Unlearn, The Uneasy AIJiance: Managing the ProductivityTechnology Dilemma, K.B. Clark, R.H. Hayes, and C. Lorenz [eds.) (Boston, MA: Harvard Business School Press, 1985), pp. 372-373. Also see the Commentary by J.L. Doyle, p. 377.
PAGE 49
43 l l American firms tend to proceed through narrow technical expertise; Japanese coma more analytical and sequential approach, panics staff their development groups with one of narrowing down the alternatives. greater numbers of generalists, including Japanese firms operate in a looser style, people from sales and marketing. They may with more room for trial-and-error. also involve the firms suppliers. product development groups in the United Japanese companies use product developStates rely more heavily on engineers with ment groups as a device to break down
PAGE 50
44 some of the rigidities in their corporate culturese.g., the seniority systemand to create a place where creativity can flourish. Many American firms would like to think they dont suffer from such problems, but probably do. At the same time, all of the attributes of Japanese product development efforts can be found in some American firms. It is the more successful Japanese firms that are visible in the United States: we seldom hear about the failures. HTS poses difficult technical challenges. Japanese companies will, no doubt, solve some of the purely technical problems before American firms. Japanese companies will also do well at scaling up HTS manufacturing processes. Some will succeed in defining profitable markets. In short, they will prove highly capable and competitive in HTS. And while many large U.S. corporations have been turning away from longterm, high-risk R&Dthe kind of work that will be called for in commercializing HTS-the Japanese are making a major effort to show the world they can be as creative and innovative in science as they are in technology. It would be a grave mistake to assume that American firms will have a head start in HTS because of U.S. skills in research. The suddenness of the turnaround in microelectronics should have pounded home the message that both industry and Government will need to do things differently in the future. APPENDIX 2A: R&D AND COMMERCIALIZATION: FOUR EXAMPLES Ceramics for Heat Engines 1 The U.S. Government has spent perhaps $300 million since the early 1970s pursuing ceramic engines. Much of the money has gone for applied research and development on components, and for demonstrations. Success has been elusive. Over the past two decades, advanced ceramics have come into widespread use in electronics, as well as for specialty applications such as wear parts and cutting tools. Ceramics hold their strength at high temperatures much better than metals, but are brittle. If reliable ceramic combustors and rotor blades could be made for gas turbines, operating temperatures could be raised, making possible smaller, lighter, and more efficient powerplants. Increased Automobile Fuel Efficiency and Synthetic Fuels (Washington, DC: Office of Technology Assessment, September 1982), pp. 144145; T. Whalen, Development ProgrammesUSA, Proceedings of the First European Symposium on Engineering Ceramics, Feb. 25-26, 1985 (London: Oyez Scientific and Technical Services Ltd., 1985), p. 177; K.H. Jack, Silicon Nitride, Sialons, and Related Ceramics, High-Technology Ceramics: Past, Present, and Future, W.D. Kingery (ad.) (Westerville, OH: American Ceramic Society, 1986], p. 259; Ceramic Technology for Advanced Heat Engines, Publication NMAB-431 (Washington, DC: National Academy Press, 1987); J. Zweig, Deja vuyet again, Forbes, N OV 16, 1987, p. 282; Case Studies of Flagship Technology, prepared for OTA by W.H. Lambright and M. FeUows, Syracuse Research Corp., under contract No. H3-5565, Dec. 31,1987, ch. IV; R.P. Larsen and A.D. Vyas, The Outlook for Ceramics in Heat Engines, 1990-2010: Results of a VVorldwide Delphi Survey, Paper No. 880514, prepared for the 1988 International Congress, Society of Automotive Engineers, Detroit, Feb. 29-Mar. 4, 1988; Advanced Materials by Design: New Structural MateriaJs Technologies, op. cit., ch. 2. possible defense applications include stationary power units and engines for tanks, trucks, and cruise missiles (ceramic components may never be reliable enough for manned aircraft). In 1971, DoDs (Defense) Advanced Research Projects Agency embarked on a ceramics R&D program, funding mission-oriented work of interest to the Army and the Navy on ceramic gas turbines, as well as research into design methodologies for brittle materials. The DARPA program continued into 1977, with funding that averaged slightly over $10 million annually. The Army continued some ceramic engine work thereafter, but DOE (then the Energy Research and Development Administration, ERDA) soon emerged as the primary source of support for applications-oriented ceramics R&D. The ERDA program, in which NASA also participated, aimed at a gas turbine engine for trucks, seeking better fuel economy. Gas turbines make more sense for trucks than for passenger cars, which operate most of the time at light loads, where turbines give poor fuel economy. However, the focus on truck engines did not last. In 1980, responding to a high-level political call for the reinvention of the automobile, DOE created a new program one that would demonstrate small gas turbines for passenger vehicles. Initially funded at $20 million annually, the incoming Reagan Administration sought to scale the effort back (along with other energy R&D); lobbying by industry contractors helped keep things going.
PAGE 51
45 Recent Federal spending (for all structural ceramics R&D) has averaged about $50 million per year (figure 2A-1), but the turbine programs appear to have moved prematurely into development and demonstration, before establishing an adequate technology base, Industry cost sharing has been relatively low; companies that saw more value in the work presumably would be willing to kick in money at a higher level. Rather than turbines, Japanese firms have put much of their effort into piston engines, both gasoline and diesel. While brittleness is a serious problem in ceramics for piston engines, it is easier to deal with than in highly stressed rotating blades. Moreover, ceramics can be introduced incrementally, substituted for a few parts in an otherwise conventional design. Some of the technical problems of structural ceramics overlap those that will be encountered in commercializing HTS ceramics. A stronger basic research effort in ceramics, rather than the demonstration projects of recent years, might have put the United States in a better position to commercialize the new superconductors. Video-Cassette Recorders* Beginning in the 1950s, and through the following decade, half a dozen and more Japanese companies raced to develop low-cost VCRs. Commercialization meant solving a long chain of tough engineering problems, so that VCRs could be produced cheaply with the features consumers wanted. Helical scan video-tape recording technology first patented by Toshiba (table 2A-1)became a critical feature in VCRs, although Toshiba itself never capitalized on its early lead in helical scanning. Matsushita entered pilot production first, in 1973, but shortly withdrew, deciding its technology was not good enough. Two years later, nearly 20 years after the U.S. firm Ampex built the first 21nternational Competitiveness in Electronics, op. cit., pp. 70, 119-123, and 186-187; R.S. Rosenbloom, Managing Technology for the Longer Term: A Managerial Perspective, The Uneasy Alliance: Managing the Productivity-Technology Dilemma, op. cit., p. 297; R.S. Rosenbloom and M.A. Cusumano, Technological Pioneering and Competitive Advantage: The Birth of the VCR Industry, California Management Review, vol. XXIX, summer 1987, p. 51. Figure 2A-1.-Federal Funds for Structural Ceramics R&D
PAGE 52
46 Table 2A-1.Chronology of Video-Tape Recorder Developments 1951 1953 1954 1956 1958 1962 1969 1970-71 1971 1971 1972 1973 1975 1976 1988 R&D begins at RCA. RCA demonstrates fixed head scanner. Toshiba files patent applications for helical scanning; prototype follows in 1959. (Earlier U.S. and European patents were never reduced to practice.) Ampex introduces broadcast model videotape recorder (VTR) with rotating scanning heads. (VTRs use reel-to-reel tape, rather than cassettes.) Several Japanese firms, including Sony and Matsushita, embark on R&D directed at VTRs for consumer markets; RCA drops its work on consumer model VTRs. Sony introduces its first helical-scanning VTR, intended for institutional markets (business and industry, schools); JVC follows in 1983. Sony announces first video-cassette recorder (VCR), replacing reel-to-reel tape with a cartridge. Ampex Instavideo camera/recorder system shown in prototype formnever marketed because of production problems. Sony U-Matic marketed for institutional use at $1000. RCA resumes VTR R&D, drops out again in 1974. JVC develops prototype of its VHS system. Matsushita enters pilot production with a consumer VCR, but withdraws after a few months. Sony introduces Betamax for home use. JVC brings VHS recorders to market. Sony to begin selling VHS machines alongside its lagging Betamax system. SOURCES: W.J. Abernathy and R.S. Rosenbloom, The Institutional Climate for Innovation in Industry: The Role of Management Attitudea and Practices, The 5-Year Outlook for Science and Technology 1981: Source Mater/a/s, Volume 2, NSF 81-42 (Washington, DC: National Science Foundation, 1981), p. 407; R.S. Rosenbloom, Managing Technology for the Longer Term: A Managerial Perspective, The LJneasy A//iance: Managing the Productivity-Technology Dllenrma, K.B. Clark, R.H. Hayes, and C. Lorenz (eds.) (Boston, MA: Harvard Business School Press, 1985), pp. 317-327; R.S. Rosenbloom and M.A. Cusumano, Technological Pioneering and Competitive Advantage: The Birth of the VCR Industry, California Management Review, vol. XXIX, summer 1987, p. 51. broadcast recorders (the size of a closet and selling for $50,000), Sonys Betamax opened the consumer market. That an American firm produced the first videotape recorders for broadcast applications was close to irrelevant. The Betamax represented the fourth generation of Sonys engineering development the seventh generation if the companys earlier industrial and institutional models (e.g., the U-Matic, which appeared in 1971) are included. Japanese companies, competing fiercely with one another, persisted with the VCR for years, in the face of many disappointments. It may be true in a narrow sense to say that the United States invented the videotape recorder and the Japanese commercialized it. But in fact, some 15 companies American, Japanese, European demonstrated 9 different technical approaches to home video in the early 1970s. It took many years of money and manpower commitments by Japanese companies to win the race, and a great deal of highly creative engineeringfocusing on manufacturing, as well as product design. Once the VCR became a commercial reality, competition centered on cost reduction, better image quality (where manufacturers of magnetic tape made major contributions), and longer recording and playing times. Firms like Zenith and RCAwhich now put their labels on foreign-made machinesnever pursued consumer VCRs with the doggedness of the Japanese. After about 1980, no American company could have entered without some sort of breakthrougha product that would have opened a new round of competition. The Japanese were simply too far down the learning curve. South Korean firms were in a different position: with wage rates well below those in Japan, they had potential cost advantages. When the Japanese refused them licenses, Korean firms developed their own VCRs. The essential ingredients in Japanese success? First, willingness to make long-term investments in risky and expensive product development efforts. Second, the manufacturing capability to massproduce precision electro-mechanical components such as the helical read-write heads that proved a key in turning the video-tape recorder into a household product. Commercialization of the VCR exemplifies the kind of incremental improvement and market-oriented engineering that the Japanese have been so good at. Is the VCR story exceptional? Not really, and certainly not in the context of consumer electronics, an industry that had stagnated in the United States by the mid-1970s. Price competition in traditional products like color TVs was fierce, imports were flooding the marketplace, and the stronger U.S. firms like RCA and GE were diversifying into other lines of business. Still, the risks did not stop RCA from investing in the VideoDisc. 3 Indeed, the VideoDisc was a bold 3M. B.W. Graham, RCA and the VideoDisc: the business of research (Cambridge, UK: Cambridge University Press, 1986]. The company ultimately lost more than half a billion dollars.
PAGE 53
47 choice. If successful, it would have given RCA a unique productsomething that none of its Japanese rivals had. In contrast, pursuit of VCR technology would have meant competing in a class of products that the Japanese plainly would be able to build cheaply and well. RCA managers knew from experience in color TV production how difficult this would be, particularly given the Japanese strategy of attacking consumer electronics markets worldwide (whereas RCAs consumer sales had been confined to the U.S. market). MRI Systems 4 Magnetic resonance imaging has been the biggest market for conventional superconducting technologies over the past few years. In 1987, two dozen companies worldwide sold a total of 500-plus MRI systems to hospitals and clinics. At roughly $2 million each, industry sales came to perhaps $1 billion. Both production and sales are concentrated in the United States. Commercialization took many years, following research showing that nuclear magnetic resonance (NMR)a discovery made by physicists could be a powerful tool for medical diagnosis. To construct an MRI image, the patient must be placed within a strong magnetic fieldcommonly produced by an LTS magnet. A computer processes the resulting NMR signals, creating an image the physician can examine (like an X-ray). MRI provides better contrast and resolution, particularly for the brain and spinal cord, than competing diagnostic imaging techniques, including ultrasound and CT scanning. During the middle 1970s, more than a dozen companies in the United States, Europe, and Japan began working to commercialize MRI. Some dropped out along the way. Others were bought by stronger firms, or merged with competitors. Japanese companies entered late, and have not been very active outside their home market. European firms led the way in engineering development. The British company EMI built the first prototype in 1978, and Bruker, a West German manufacturer, followed the next year. Both these companies had prototype systems operating in clinical settings by 1981. Shortly thereafter, EMI decided to leave the medical equipment business, and sold its technology to a competitor. By the end 4Health Technology Case Study 27: Nuclear Magnetic Resonance Imaging Technology (Washington, DC: Office of Technology Assessment, September 1984]; Superconductive Materials and Devices, Business Technology Research, Wellesley Hills, MA, September 1987, pp. 38-50. of 1983, eight firms had commercial prototypes availablethree American companies, four European, and one based in Israel, Early in design and developmente.g., during the stage labeled alternative conceptual design in figure 3 (earlier in the chapter)each firm faced decisions on its magnet system. The alternatives permanent magnet, resistive (non-superconducting), superconductingcarried advantages and disadvantages in terms of factors such as initial and operating costs, as well as field characteristics like strength and stability. Most companies chose LTS magnets, with several pursuing conventional magnet designs in parallel. Because the design of the magnet affects image qualitya central concern in purchasing decisions by hospitalsfeedback from the clinical studies and clinical testing stages (figure 3 played a vital role in refining prototype designs.) This brief description illustrates, first, the ways in which research may enter the commercialization process. In this case, the R&D ranged from nuclear physics (the NMR phenomenon itself), to the medical studies demonstrating that MRI could be a valuable diagnostic tool, to computerized signal processing and superconducting magnet design, MRI systems emerged as viable commercial products in 1984. It was only then that designs stabilized and production became relatively routine, at least in the leading companies. It took 38 years to go from scientific discovery (experimental verification of NMR) to marketplace success. Commercialization in the sense of engineering development spanned the years 1977 to 1984. Regulatory approvals were an early hurdle. The U.S. Food and Drug Administration spent several years evaluating the new technology. Manufacturers had to estimate the effects of third-party payment policies on market growth: Would Government agencies responsible for Medicare and Medicaid give a quick okay to the new technology? Or would they delay? How about the big insurance plans like Blue Cross/Blue Shield? In fact, hospitals were not generally reimbursed for MRI services until late in 1985. Furthermore, with MRI systems costing several million dollars, State government certificate-of-need approvals became a precondition for many sales. U.S. firms did not have the initial lead. Nonetheless, they quickly emerged in the forefront as the technology moved out of the laboratory and became a practical tool for medical diagnosis. A major reason for U.S. success was simply that this country is the biggest market in the world by far for medical equipment. That the U.S. economy is the worlds
PAGE 54
48 largest and most diverse is both an advantage and a disadvantage for American firms. They are at home here, but their domestic markets are a magnet for foreign firmswho may be willing to lose money in the United States for the sake of learning and experience. Designing and developing a new product from scratch, as in the case of the first MRI systems, represents a major corporate commitment. An all-new product takes much more time and money than the incremental redesigns, improvements, and new models that come later and constitute most of the routine work of product/process development. The all-new product (or manufacturing process) will also, in the ordinary course of events, depend more heavily on new knowledgee.g., research results. Feedback from the R&D laboratory and the marketplace remain important even for routine development work, however. Once the medical community accepted MRI, competing firms quickly began differentiating their products through stress on image quality, good service, and reliability. LTS Magnets 5 Federal R&D, much of it for high-energy physics experiments and research into nuclear fusion, underlies development of the LTS magnets found in MRI systems. Wound with cable made from niobiumtitanium alloy filaments embedded in a copper matrix, and cooled with liquid helium, the magnet accounts for up to a quarter of the cost of an MRI system. D. Larbalestier, et al., High-field Superconductivity, Physics Today, March 1986, p, 24; L. Hoddeson, The first large-scale application of superconductivity: The Fermilab energy doubler, 1972 -1983, Historical Studies in the Physical and Biological Sciences, vol. 18, 1987, p. 25; Superconductive Materials and Devices, op. cit., pp. 33-61; Technology of High Temperature Superconductivity, prepared for OTA by G.J. Smith 11 under contract No. J3-21OO, January 1988; Governments Role in Computers and Superconductors, prepared for OTA by K. Flamm under contract No, H3-6470, March 1988, pp. 56ff. Also see app. B. Until MRI markets began to grow, most superconducting magnets were custom-designed for scientific equipment. The late 1960s saw the first major application of niobium-titanium, a bubble chamber built at Argonne National Laboratory. A much larger federally supported projectthe Tevatron particle accelerator, completed in 1983 at the Fermi National Accelerator Laboratoryconsumed more than 30,000 miles of niobium-titanium wire for its nearly 1000 magnets. Most of the wire came from Intermagnetics General Corp. (IGC), established by several former General Electric employees in 1971. IGC and other small, specialized firms had begun moving into LTS as major corporationsWestinghouse as well as GEwithdrew, finding that market growth did not live up to their expectations. Development of the processing techniques for LTS magnet wire was a lengthy and complex task, one that would have taken much more time without the demand provided by the Tevatron. Private firms drew on the publicly supported technology base, and also helped to extend it, as they developed the know-how needed for manufacturing LTS wire and cable (the Fermi Laboratory designed and built the magnets internally). It took many years to raise the critical current densities of niobium-titanium wire to the levels needed for the Tevatron and for MRI. The task hinged on the relationship between fluxoids (each of which contains a magnetic flux quantum)a matter of physicsand the microstructure of the wire. Through careful microstructural controlspecially tailored sequences of wire drawing and heat treatmentmetallurgists and materials specialists were able to create fine dispersions of second-phase particles. These particles pin the fluxoids, keeping them from moving, The pinning can raise the critical current densityhence current carrying capacityby 10 times or more. R&D aimed at optimizing the processing technology began in the late 1960s, and still continues, with engineering development guided by theoretical understanding.
PAGE 55
Chapter 3 Superconductivity in Japan and the United States
PAGE 56
CONTENTS Summary l .,, ,, ,...... Corporate Strategies l R&D and Business Planning in the United States . . U.S. Strategies in High-Temperature Superconductivity . Japanese R&D l Japanese Strategies in High-Temperature Superconductivity U.S. and Japanese Strategies Compared. . . . . Japans HTS Initiatives l Government Resources for Superconductivity . . . Who Has the Lead Role, Government or Industry? . . Rivalry or Cooperation? The Internationalization of Japanese Superconductivity R&D . . . . . . . Concluding Remarks. . . . . . . . . . Boxes Box E. F. G. H. I. J. K. L. R&D in U.S. Corporations . . . . . . . Superconductivity R&D in U.S. Companies, . . . Venture Startups in HTS .*.**.* l .,*.*.* l **...** *,***. Superconductivity R&D in Japanese Companies . . HTS R&D at Nippon Steel . . . . . . . Josephson Junction Computer R&D: From the United States Japans Magnetically Levitated Train Program . . . .* **,**, l .***** ,. 0..,, .. ...0, l *****. .,, ...0 l .****. . . l ****.* ,** l *. .. ..,.. 6 .* *..*, l *..*.. .. **,*,, l .**,.* .0, ,, l *.**., to Japan ,. ..,,, Prospects for U.S.-Japan Cooperation in Superconductivity R&D . Tables Table 5. R&D Funded by Business and Industry. 6. Major Japanese Government Programs and in Superconductivity. . . . . . .....,.,..,.....,.,. Activities l **.....*.,*,.*,,*..,. Page l l l l ** l .. l l l 51 52 54 57 60 63 66 67 68 75 77 79 . l l l l l l 55 58 60 65 67 71 74 78 Page l 61 68
PAGE 57
Chapter 3 Superconductivity in Japan and the United States SUMMARY The first 10 weeks of 1988 saw the discovery of two more copper-oxide based superconducting materialsone with bismuth as a critical ingredient, the other thallium. These two compositionsboth with critical temperatures in the range of 100 degrees Kelvinjoined those containing rare earth elements (e.g., lanthanum, yttrium) that scientists around the world had been studying for a year. Laboratory resources had been heavily committed to the yttriumbarium-copper-oxide familythe so-called 1-2-3 superconductors and the scientists had been making good progress in improving current densities and learning to make thin films. Then, all of a sudden, two entirely new compositionsequally complex, five elements in each, partially understood structures and phase diagrams. Two new worlds to explore. Heaven for the scientist (though more sleepless nights). Hell for the businessman. Business planners and government strategistsat General Electric and Sumitomo, MITI and the Pentagonnow faced still more choices. Superconductors came in at least three varieties: 1. The old, low-temperature superconducting (LTS) materialsmetal alloys like niobiumtitanium, well understood but calling for cooling to near liquid helium temperatures might still remain the material of choice for some applications. Very sensitive detectors of enemy submarines or brain waves might have to be operated at liquid helium temperatures in any event, to get noise levels down. 2. The 1-2-3 ceramicsbrittle, not very stable, but with properties that people had begun to understand. 3. The latest high-temperature superconducting (HTS) compoundsthose containing bismuth or thalliumstill a mystery, but potentially easier to work with and perhaps having better combinations of properties than the 1-2-3s. Then there is the fourth categoryeverything as yet undiscovered. With no theory, only enlightened empiricism to guide the search, not even the biggest laboratories can explore all the possibilities. Choices must be made, priorities set, resources allocated. For a company, 50 people working on HTS means 50 people who cannot work on other projects that might, in the long run, be equally important. This chapter is about those choices, and how they are made, in U.S. and Japanese companies, and in Japans Government. Chapters 4 and 5 deal with the choices facing the U.S. Government. Corporate managers in the United States and Japan look at the world differently. In seeking strategies for profits and growth, they make different kinds of choices, set different priorities, because they operate in contrasting economic, political, and social environments. Companies that do business on a global scaleIBM, Du Pont, Nippon Steel, Hitachimay have much in common in their view of the world, but there are important differences between them as well. It may be a clich to say that Japanese firms put more weight on growth and market share than on short-term profits, but it is true, and it makes a difference in R&D strategies, business plansthe entire array of competitive choices. The U.S. startups, financed with venture capital, that sprang up during 1987 have no counterparts in Japan. Nor do the small LTS specialists mentioned in the preceding chapter. Japans joint government-industry R&D projectsa fixture of that countrys industrial and technology policieshave no counterparts here. Business planners must decide how many people and how much money to put toward superconductivity. They must decide how to spend that money, and what kind of people to assign. Is it too early to think about applica51
PAGE 58
52 tions? Does it make sense to continue exploring LTS technologies? Managers in the United States and Japan have made diverging choices: l l l l A few large American companies are pumping substantial resources into HTS. But many other U.S. firmsorganizations with the resources to pursue HTS if they wishedhave taken a wait-and-see attitude. They may have a few people working on HTS R&D, but mostly just to keep track of the technology. Most of the effort in the United States is going toward research. American managers believe HTS should remain in the laboratory until more scientific knowledge is in hand. Perhaps a dozen large, integrated Japanese multinationalsmanufacturers not only of electrical equipment and electronic systems, but of ceramics, glass, and steel are pursuing multi-pronged R&D strategies in superconductivity. As in the semiconductor industry, these resource-rich companies could prove potent rivals for smaller American firms hoping to stake out a position. Japanese companies are conducting research but also thinking about applications. They are putting more effort than U.S. firms into thinking through what HTS might mean for the companys strategy. In general, managers in Japan believe that HTS is closer to the marketplace than do American managers. They also see HTS as a means of creating new businesses, while American managers are more likely to view it in the context of their existing business. The breadth of the Japanese effort substantially exceeds that of the United States. Managers in the larger American companies believe that if HTS takes off, they will be able to catchup or buy in. Japanese managers want CORPORATE What place does R&D have in the strategies of American firms? How do managers think to move down the HTS learning curve in real time. They believe that advantages established now will last. Scientists, managers, and venture capitalists involved in the HTS startups in the United States believe the same thing, but they are few, small, and weak compared with the Japanese companies. Taken as a whole, the U.S. approachdriven by the need to show financial paybacks in the short termcould leave American industry behind Japan within a few years. Such an outcome is not assured. HTS could languish in the research laboratories. Or HTS could evolve like the laser industrynever quite matching the expectations of the enthusiasts, driven heavily by military needs, lacking the revolutionary impacts of the computer or the semiconductor chip. On the other hand, HTS could grow and spread and expand like the digital computer. Computersespecially the microprocessors and single-chip microcomputers found in microwave ovens and TV sets, banking machines and machine tools, Chevrolets and 767shave penetrated innumerable products and manufacturing processes. The same could eventually happen with HTS technologies. American companies, by and large, have taken the conservative view; Japanese companies have taken the optimistic view. If technical developments in HTS proceed as swiftly over the next 2 or 3 years as they did during 1987, then Japanese companies that have been laying the groundwork for commercialization will be in a stronger position. Superconductor fever has swept through Japans Government too, with ministries vying with one another for the lead in policy. The picture has now stabilized, but 1987 saw many actors seeking center stageand few signs of the coordinated, monolithic policy machine that some Americans still think of as Japan, Inc. STRATEGIES about HTS? How does the business culture in the United States differ from that in Japan? Ef-
PAGE 59
53 fective government policies depend on an understanding of the attitudes and practices of managers, the forces that condition their decisions. As it happens, there is substantial truth to the commonplace observation that Japanese managers take a longer view than Americans. This difference shows up in R&D decisions on HTS. For years, American firms have been criticized for short-sightedness. 1 Managers are under pressure from Wall Street and institutional investors (pension and mutual funds, insurance companies) to show high and increasing quarterly earnings. Failure to do so can lead to a loss in stock values, and vulnerability to hostile takeovers. Jobs and egos are on the line, so the argument goes; few chief executive officers or division heads can survive many mediocre quarterly reports. Techniques used by American managers for evaluating investment alternativesdiscussed in the next sectionreinforce the pressures to sacrifice long-term opportunities for short-term profits. Instead of investing in R&D that will increase their firms storehouse of proprietary know-how, managers cut R&D to reduce costs. Instead of investing in new plant and equipment to increase productivity and flexibility, I More than a dozen years ago, an experienced U.S. R&D manager wrote that the root cause of the present and future decline of U.S. technological prominence is a temporal mismatch between the natural pace of innovation and the time horizon of most U.S. industrial corporations . this root cause is overlooked by the managers of major U.S. industries because they have a warped set of values R.D. Dean, Jr. The Temporal MismatchInnovations Pace vs Managements Time Horizon, Research Management, May 1974, p. 12. A recent survey of nearly 140 U.S. companies found greater emphasis on near-term lower-risk results-oriented work in their R& DTrends in the Chemical Industry, Results of the MarchMay, 1987 Survey of ACS Corporate Associates. (ASC is the American Chemical Society. The survey covered corporate members from other industries as well.) For 1988, the National Science Foundation has forecast the lowest rate of real, inflation-adjusted growth in R&D since 1977. Even the Electric Power Research Institute, financed by regulated utilities, evidently feels many of the same pressures as publicly owned corporations. According to the Institutes president, We now must clearly demonstrate that there is value in what we are doing and that it falls in an acceptable business time frame, This is a remarkable difference from when we started [1973] EPRIs New President Looks to the Future, New Technology Week, Feb. 1, 1988, p. 8. they slash payrolls, keep the old equipment running while spending no more than absolutely necessary on maintenance, and move labor-intensive production offshore or to the Sunbelt. Rather than putting money into core businesses, managers diversify (from steel to real estate, from manufacturing to services), buy up other companies rather than build their own, and seek paper profits. The picture may be a caricature, but it has a good deal of truth in it. 2 How have these pressures affected corporate decisions on HTS? What other factors enter into R&D decisions? How, specifically, do U.S. managers view HTS compared with their Japanese counterparts? The next section of this chapter examines the R&D strategies of American firms. Later sections turn to Japan. 3 The findings in brief: l l American managers have been notably more reluctant to commit resources to HTS a technology with highly uncertain prospects. They view profits as lying well in the future. Japanese executives, in contrast, seem confident that investments now will pay off some time and in some way. Their view of the future is quite a different one from that of American managers. These contrasting views reflect the business environments and investment climates in the two countriesindeed, the entire complex of factors that affects management decisions. 2A typical example: Tektronix, a leading manufacturer of instrumentation and computer work stations, will fire 1,000 whitecollar employees in a bid to boost earnings. When the company announced that it would close down some R&D projects, and scale back its marketing and sales staff, a stock market analyst said, Theyre addressing the right issues. See J.P. Miller, Tektronix Plans To Dismiss 6 percent Of Its Workers, Wall Street Journal, Mar. 7, 1988, p. 12. sMost of the information on company views of HTS comes from interviews in the United States and Japan during late 1987 and early 1988, and from surveys of U.S. and Japanese firms. The U.S. National Science Foundation, through its Tokyo office, conducted the survey of Japanese companies for OTA.
PAGE 60
54 R&D and Business Planning in the United States Funding Decisions American firms approach R&D much like any other investment. With some exceptions, a decision on individual R&D projects or divisional R&D budgets will be viewed in the same light as a decision to invest in new production equipment, acquire another company, or sell the firms Manhattan headquarters and move to New Jersey. Box E describes the process. R&D carries higher risks than many other corporate investments, in the sense that outcomes are less certain. Moreover, the projects with the greatest uncertainty tend to be those with longer payback periods. As explained in box E, such projects must promise exceptionally large rewards, or the investment money will go elsewhere. Research that loses out in private corporations might nonetheless benefit the country as a whole. If no one company can reap the rewards, none may invest. That is why the Reagan Administration has continued relatively liberal funding for basic research, even though cutting back on more applied work. Companies do little basic research because, from their perspectives, it does not pay. But the social returns from a portfolio of such investments can be great. R&D Management American companies normally engage in R&D to support existing business activities, or those that have emerged from reasonably careful planning exercises. Even the two remaining giants of U.S. corporate researchIBM and AT&Tseek, in their own quite different ways, to guide and manage R&D in support of overall corporate goals. Oriented toward results, American executives see corporate R&D as an activity to be guided by the firms overall objectives. Only rarely do they look to R&D as a means of uncovering wholly new business opportunities. When they do, they tend to seek home runs (like the Xerox copier or Polaroid photography) rather than the incremental advances that have a central place in Japanese corporate strategies. Du Pont would dearly love another product like Nylon, and Intel another invention with the impact of the microprocessor, but who knows where these might come from? Inventions cannot be planned, and no company will spend much money on an unguided search. Furthermore, in big organizations with ample R&D budgets, projects that might be exciting technically can get lost in the corporations grand strategy. Even though they might promise high rates of return, if the overall market looks relatively small, a big company may not be tempted. Low-temperature superconductivity provides a number of examples, and HTS will probably bring more. Most firms give their R&D managers latitude in initiating work on their own, hoping for results that will eventually contribute to the bottom line. 4 Individual managers, moreover, do not always follow corporate policy. Workinglevel people bootleg research that might not be approved higher up. Top management normally lets project leaders and departmental managers follow their own judgment, so long as not too much money is involved. Star researchers, likewise, may be left alone to pursue their hunches and intuitions (which is how HTS was discovered in IBMs Zurich laboratory). A few companies let people spend some fraction of their timeusually smallfollowing personal research interests. Such policies tend to be pursued for reasons of morale. They help create a more comfortable environment for industrial scientists, a more academic setting. If the results bring in money for the company, this will normally be viewed as a lucky accident; in most U.S. firms, most R&D scientists and engineers work within carefully managed groups, on projects that corporate management first approves and then monitors. qThis latitude seems increasingly circumscribed. For instance, many U.S. research managers must take such decisions as whether to spend, say, $250,000 to join an R&D consortium, all the way to the top of the management hierarchy. See Roundtable: Physics Research In Industry, Physics Today, February 1988, p. 54.
PAGE 61
55
PAGE 62
[Page Omitted] This page was originally printed on a gray background. The scanned version of the page is almost entirely black and is unusable. It has been intentionally omitted. If a replacement page image of higher quality becomes available, it will be posted within the copy of this report found on one of the OTA websites .
PAGE 63
5 7 These practices follow quite naturally from accepted business practices in the United States. They are part of the received wisdom wisdom that says the rare inventive genius will in any case strike out on his or her own, founding a new company to exploit whatever is really new. U.S. Strategies in High-Temperature Superconductivity As the survey results in box F indicate, at least 28 American companies are spending at the level of $1 million or more on superconductivity R&D. On the average, they have nearly 20 professionals at work. Many of the companies working in HTS have LTS experience. During 1987, most of HTS R&D went toward understanding the 1-2-3 ceramics, and toward processing-related work. The high proportion of scientists compared to engineers reflects the basic character of the research. A number of American companies, big and small, have been conducting R&D on superconductivityand perhaps producing LTS wire or magnetsfor two decades or more. Some attempted to commercialize LTS products, later to scale back or abandon their work. In the 1970s, LTS-based Josephson junctions (JJs) excited considerable interest in U.S. electronics and computer companies. The enthusiasm faded, and much of the U.S. work was eventually dropped. (The Japanese persevered with JJs, as discussed in box J, later in chapter.) In other cases, companies like GEwith its line of medical imaging equipment incorporating LTS magnets (ch. 2)have gone onto become major forces in the marketplace. Still, some of the Americans with experience in LTS view the new ceramic superconductors with considerable skepticism. Earlier disillusionment may have affected the current strategic posture of some American firms. Indeed, many U.S. R&D managers feel it is too early even to think about applications of HTS. They think much more research will be needed to characterize the new materials. Moreover, they believe that commercial payoffs are likely to be in the distant future. Media hype has had little influence on them, These views affect funding for HTS R&D. But if research in HTS is called for, who, on the American side, will do it? With few exceptions, U.S. industrial R&D laboratories avoid science. Their job is to support the operating units. U.S. industrial research grew rapidly during the early 20th centuryled by companies like GE (which established a corporate R&D facility in 1900), Du Pont (which followed in 1902), AT&T (1911), and Eastman Kodak (1912), But decline has set in, for reasons that range from corporate decentralization to the shortening of time horizons. Today, few American corporations pursue much basic research with their own funds. Thus, when American managers state that HTS belongs in the laboratory, they often mean someone elses laboratory. Two strategic scenarios, then, encompass most American firms: 1. The first includes the companies that have taken a careful look at how HTS might affect their businesses, assuming continuing advances in the technology. Such assessments often entail a complete review of the firms product linesa process some firms have begun by revisiting earlier evaluations of LTS. At this stage, such an assessment is no easy task, given the uncertainties. No one can predict which of the new families of Photo credit Westinghouse Rotor for prototype LTS generator ~[4_r/54 f ?IR j : i
PAGE 64
58
PAGE 65
59 materials might prove most useful, what kinds of problems the design of practical magnets or Josephson junctions might bring, whether three-terminal devices will emerge (ch. 2)much less the costs. There is a second problem, one creating even more uncertainty. Will somebody discover superconductivity at room temperature this year? Next year? In 2050? This makes all the difference for any economic evaluation. In fact, most U.S. companies have based their assessments on liquid nitrogen operating temperaturesan assumption leading to relatively pessimistic evaluations except for quite specialized applications. The typical view goes something like this: l l l l The primary need is for materials characterization, work that can be carried out (and is) at literally hundreds of academic, government, and corporate laboratories around the world. Our company could spend a lot of money on HTS without much chance of a breakthrough. Even then, the research would probably not result in proprietary advantage. In any case, the first applications are likely to be in defense systems, where cost constraints are less severe. Under these circumstances, the best strategy is to hedge the companys bets by tracking the science and technology worldwide, without investing heavily. Such a strategy implies willingness to alter course if someone else makes a major breakthrough (not necessarily in operating temperaturesa big increase in critical current densities might be enough for at least some U.S. firms). These companies many of them currently spending at the $1 million to $5 million level (box F), and with perhaps a dozen people assigned to HTSwill keep a core group at work. But they are not ready to jump into the HTS R&D race. 2. The second strategic scenario includes those companies, most of them large, with strength in research and the ability to pursue HTS R&D on a significant scale. The list is short: AT&T, IBM, Du Pont. Bellcore, Westinghouse, GE, and a few others might be added, along with several major defense contractors. Here, the presumption that HTS should remain in the laboratory is not a bar. Of course, not even IBM or AT&T can do everything; these companies too face the choice of investing money and manpower in HTS or in alternative R&D projects. But HTS exerts a powerful attraction, not only on working scientists, but on those who manage research. Finally, for some of these companies, success in HTS R&D could have pervasive impacts on their businesses. In a company like IBM, which already maintains a portfolio of equally uncertain R&Dmost with far less potential impact HTS quite naturally gets a high priority. A few American firms, then, have 50 or 60 people assigned to HTS, and some work underway that verges on development. There is also a third group, not large, consisting of startups with venture financing (see box G), plus other small firms. Government money for R&D could pull a few more American firms into HTS R&D. But much of this money will go for defense projects. Even in companies that include military or space divisions along with other operationsas IBM, AT&T, and many other large corporations do the two sides of the company normally operate largely separated from one another (ch. 4). In summary, most U.S. companies have adopted a wait-and-see attitude toward HTS. They may have assigned a group of people to monitor developments. Perhaps they conduct research on a small scale. But few major U.S. firms have placed superconductivity among their top R&D priorities. The others see good reasons for their decisions, of course. Risks and uncertainties are high; judgments differ. But if HTS develops more rapidly than they anticipate, few U.S. companies will be able to respond as quickly as the aggressive Japanese firms that have already begun laying groundwork for commercialization.
PAGE 66
60 Japanese R&D As in the United States, the R&D strategies of Japans corporations flow from more general managerial attitudes. In important respects, Japanese executives exhibit decisionmaking behavior that differs from that here. As noted in the preceding chapter, U.S. and Japanese management styles also show many similarities, particularly in high-performing companies, but some strategic choices that make sense in an American context may be incompatible with Japanese views. Corporate Research in Japan Patterns of industrial R&D have been changing rapidly in Japan. American firms, accustomed to advantages in technology, must also adaptperhaps to being first among equals. Japanese firms have a tougher job. They are trying to catch up and take the leadand trying to do so with people and organizations that, until recently, started by licensing and adapting foreign technologies. s This takes money, and Japanese industry has been willing to spend it. Table 5showing the rapid rise in businessfunded R&D in Japandemonstrates the strength of that commitment. Japanese firms see technology development as a key ingredient in com5 See International Competition in Services (Washington, DC: Office of Technology Assessment, July 1987), ch. 6.
PAGE 67
61 Table 5.R&D Funded by Business and Industry Business-funded R&D expenditures 1981 1983 1986 a United States: Billions of dollars . . $35.9 $43.2 $58.2 As percentage of all U.S. R&D . . . . 50.0 % 50.0% 49.8% Japan: Billions of yen . . . Y 4364 Y 5451 X 7000 Billions of dollars . . $19.8 $22.9 $41.6 As percentage of all Japanese R&D. . . 72.9% 75.9% 77.8% a Estimated. SOURCE: International Competition in Services (Washington, DC: Office of Technology Assessment, July 1987), p 205 petitive strategies. While business-funded R&D in the United States has been going up almost as fast in real terms, the overall lead of the United States in private sector R&D stems simply from the greater size of the American economy; on the average, Japanese firms spend substantially more on R&D as a percentage of sales. In earlier years, major Japanese corporations began by scanning the world for technology, often with the aid of Japans large trading companies, as well as the government. When possible, they set one potential source for technology against another to minimize licensing costs. Japanese companies followed with engineering excellence, highly developed manufacturing systems, and carefully targeted marketing strategiesoften competing aggressively at home before launching their export drives. But the world has changed for Japanese companies. In many technical fields, they have reached parity with the West. American and European firms, in any case, are much more wary of licensing than even 10 years ago. There is little more for Japan to assimilate. Japanese firms must either wait for new ideas to appear elsewhere or step up their own research. Even for companies not pressed by increasing competition from newly industrializing countries like South Korea, the first choice is a recipe for disaster at home. Thus Japans major corporations are working hard to generate new technical knowledge. This search for proprietary technologies means more basic research. 6 As American companies turn away from relatively fundamental work, Japanese firms are turning toward it. Many American R&D managers give the Japanese little chance of accomplishing much, at least over the next 5 or 10 years. They view Japans cultureand the organizational environment in Japanese firmsas hostile to creative research. Many Japanese would agree. Their engineers may be superb at painstaking product development efforts, but, at least according to the stereotype, research demands individuality and creativityqualities discouraged in Japan. This stereotype is greatly exaggerated: a closer look suggests that creativity in engineeringsomething the Japanese have amply demonstrateddiffers little from creativity in research and in science. In fact, U.S. scientists and R&D managers directly involved in HTS research give their Japanese counterparts high marks for their work. Moreover, in related fields like ceramics, the Japanese already have the lead in commercialization. 7 While Americans still see Japan as lagging generally in science, Wee, for instance, S.K. Yoder, Japanese Launch Bid to Lead the World in Pure Science, Wall Street )ournal, June 3, 1987, p. 26, Also P. Marsh, The search for some home-grown heroes, Financial Times, July 6, 1987, p. 15, which quotes Tokyo Universitys Professor Shoji Tanaka, Japans best-known superconductivity expert, as follows: For a long time the Japanese people had the feeling they were behind in science. But now the inferiority complex is starting to vanish. We do have a relatively inflexible university system. But young people are changing and will force their professors to adopt different ideas. In searching for creative scientists and engineers to staff their research laboratories, Japanese companies are hiring more women and foreignersM. Kanabayashi, An Acute Shortage of Engineers Threatens Japans Research Goals, Wall Street )ourna], Oct. IS, 1985, p. 32; Poor lab facilities hamper plan to attract foreign researchers, Japan Economic Journal, Apr. 16, 1988, p. 5. Industry and government in Japan have put considerable emphasis on the life sciences in their overall drive for research excellence. See Commercial Biotechnology: ArI International Ana)ysis (Washington, DC: Office of Technology Assessment, January 1984), pp. 505-510. Later sections of this chapter discuss government policies in support of basic research in Japan. High-Technology Ceramics in Japan, Nh4AB-418 (Washington, DC: National Academy Press, 1984); Ceramic and Semiconductor Sciences in Japan, 1987, PB 88-122478 (Washington, DC: Department of Commerce, 1987); Advanced A4ateria]s by Design: New Structural Materials Technologies (Washington, DC: Office of Technology Assessment, June 1988).
PAGE 68
62 there is no basis for complacencyand certainly not when it comes to superconductivity. Time Horizons How about the longer term view that Japanese managers reputedly take? This stereotype holds upas can already be seen in HTS. The reasons begin with notions of success and failure that differ substantially between the business cultures of the United States and Japan. In the United States, management perceptions of the factors that determine the value of a firms stock heavily influence decisions. When surveyed, U.S. managers rank profits (as measured by return on investment), and increase in share price, as their primary objectives. Japanese executives also view return on investment as important, but put it below another goalmarket sharewhich appears no better than third in rankings by American managers. 8 Furthermore, Japanese companies need not worry too much about the price their stock commands, given the way Japans financial markets work. Equity remains less important than debt in corporate financing, and new stock issues are the exception in raising funds. The now-standardand often oversimplifiedarguments concerning costs of capital also come into play here; plainly, on a present value basis, or indeed almost any reasonable criterion, lower costs of capital in Japan make long-term projects more attractive. 9 Japanese companies, then, typically use different decision rules in evaluating investment alternatives. Managers in Japan see R&D as a means for maintaining or increasing market share, both at home and abroad, with market share a necessity for holding on to a com8J.C. Abegglen and G. Stalk, Jr., Kaisha: The Japanese Corporation (Tokyo: Charles E. Tuttle, 1987), pp. 176ff. While decision criteria in Japan certainly differ from those here, there is little consensus in the Weston the extent to which Japanese firms rely on financial measuresor, more precisely, on what kind of measures they use, and for what purposes. See, for example, Part Two Discussion Summary, The Llneasy Alliance: Managing the Productivity-Technology Dilemma, op. cit., p. 283. ozn~erna~jon~ competitiveness in Electronics (Washington, DC: Office of Technology Assessment, November 1983), ch. 7. petitive position in dynamic markets. Japans rapid postwar economic growth, and success in exporting, has made market share the top priority; when sales are expanding rapidly, grabbing as big a share as possible, and holding on to it, become the key to profits. The emphasis on growth reflects a belief among Japanese executives that only large companies can remain financially viable in international competition. Japanese industry has spawned few of the entrepreneurial startups so much a part of the scene in U.S. high-technology industries although many policymakers in Japan would like to create a place for them. Other factors and practices reinforce the view that growth is all-important. Larger Japanese companies historically have attempted to provide lifetime employment for a portion of their work force. Managers continue to view this as an obligation, and growth makes it easier to sustain employment. Layoffs tend to be seen as evidence of management failure, rather thanas in the United Statesa consequence or symptom of economic downturn. This sense of obligation helps shape corporate goals and managerial behavior. Where American executives would slash payrolls, Japanese companies will often accept lower profits. What does this mean for R&D? A continuous search for new products and new markets, including those that might not fit very comfortably into ongoing operations. Where American companies look to R&D to support existing businesses, Japanese companies are just as likely to see it as a means of creating new businesses. Where American firms look for home-run opportunities, their Japanese counterparts have been more willing to start small and grow new businesses gradually. Government-Industry Relations The antagonism with which so many U.S. corporate managers view government also contrasts with typical Japanese attitudes. American managers feel, by and large, that the Federal Governments role should be tightly circum-
PAGE 69
63 scribed. Wherever possible, economic matters should be left to the private sector. Later sections of the chapter describe how government and business in Japan have worked together to promote HTS. Here, the point is simply that Japanese executives have relatively comfortable working relationships with government. Japanese managers tend to feel that employees of their government are competent and deserving of respect, even when they disagree vehemently on matters of policy. Managers in Japan pay attention to goals and objectives announced by the Ministry of International Trade and Industry (MITI), or the Ministry of Finance. This does not mean they will necessarily follow the paths that MITI or other ministries attempt to lay down. Contrary behavior is common. On the whole, Japanese executives would prefer to go along with government, while working to mold policy in ways they regard as desirable. R&D Management Traditionally, Japanese R&D has focused on engineeringproduct/process development, rapid transfers to manufacturing. Despite the engineering perspective, managers put less weight on short-term outcomes, and show more willingness to invest in projects that will not yield positive cash flows until well into the future. Japans national goal of technological independence pushes companies in the same direction. Personal opinions by managers carry great weightespecially when the advocates of particular projects enjoy high standing as researchers or managers. One mans recommendation can lead to a major new R&D project in Japan something that would be highly unusual in an American company. Competition among Japanese firms combines with cultural characteristics to yield another contrast with the United States. Japanese companies tend to emulate one another; when one begins research in a field like HTS, others follow. Few executives will risk letting a direct competitor engage in R&D without investigating the subject themselves. Similarly, companies are uncomfortable at the thought of closing down a research program that others are continuing. For such reasons, Japanese firms spend a good deal of effort tracking their competitors day-to-day R&D efforts. American companies, which tend to look to R&D for means of differentiating themselves, show more interest in products soon to hit the market. Finally, given the way Japanese firms make decisions, it should be no surprise to learn that they stick with R&D efforts once begun. One executive commented in an OTA interview, In Japan, we continue research projects unless persuasive reasons are mustered against them. In the United States, I get the feeling that projects are cancelled in the absence of good arguments supporting continuation. The difference is subtle, but important. We tend to be optimistic on research results; you tend to be pessimistic. Such attitudes may be remnants of an earlier time, when success came easier. Still, they contribute to the persistence that has been so important to Japans accomplishments in commercialization. The typical Japanese approach to R&D carries disadvantages as well as advantages. With little systematic guidance for comparing one project to others, and subject to the influence of strong personalities, Japanese firms risk bad decisions. This weakness could become more important in the future; given their lack of experience in managing fundamental research particularly if companies follow one another down blind alleys. But for now, the freer hand that Japanese managers have in allocating resources to long-term, high-risk projects is a notable strength. Japanese Japanese mously, see Strategies in High-Temperature Superconductivity R&D managers, almost unaniHTS as a revolutionary technology, one that promises radical change. Skepticism, common in the United States, has been rare in Japan. Implicit in some Japanese views, explicit elsewhere, has been the assumption that room temperature superconductivity is not
PAGE 70
. 6 4 far away. (Otherwise, even the more optimistic Japanese scientists and engineers see the potential as relatively limited.) A corollary follows: Japanese executives believe that HTS will be a major battleground for international competition over the next two or three decades. 10 All those interviewed by OTA believed that Japan would have to depend on home-grown technologies in the future. It follows that early exploitation of HTS holds a rare opportunity. Japanese managersin sharp contrast to their U.S. counterpartshave little doubt that HTS will be a central element in competitive strategies. Not surprisingly, then, commitments in Japan as indicated by industrial employees assigned to superconductivity R&Dsubstantially exceed those here. Box H, based on a survey of Japanese firms conducted for OTA by the U.S. National Science Foundation (NSF), reveals the following contrasts with the United States: l Although reported budgetary outlays by U.S. industry exceed those in Japan (at $97 million compared with $90 million), Japanese firms reported 900 people working on superconductivity (versus 625 here). l Total Japanese R&D spending on superconductivity for 1988industry plus governmentshould exceed $160 million, U.S. spending $250 million. Such comparisons must be treated with caution, however. The company surveys are incomplete, the fiscal years for the two governments are 6 months out of phase, and the exclusion of some salaries from the Japanese Government budget figures makes the estimate for Iosixty percent of 167 Japanese companies responding to a mid1987 survey expected a $2o billion world superconductivity market by 2000Superconductor Industry Survey Conducted, JPRS ReportScience & Technology, Japan, JPRS-JST-87-068L, Foreign Broadcast Information Service, Oct. 29, 1987, p. 1 [translated from Nikkei Sangyo Shindmn, July 28, 1987]. Electronics applications were ranked most promising, followed by energy storage, and then by a variety of other electric power applications. Nearly 85 percent of those responding to a different survey foresaw applications of room temperature superconductors in industrial equipment by 2010Waga kuni ni okeru Gijutsu Kaihatsu no Hoko ni kansuru Chosa [Survey of Trends in Technology Development in Japan], no. 4, Kagaku Gijutsucho [Science and Technology Agency], 1987, p, 12. l Japan low by some unknown amount. Finally, spending levels say nothing about the outputs of R&D. Given all these uncertainties, the contrast in numbers of industrial employees takes on the greatest weight. Japanese firms are emphasizing prospective applications more heavily. Many companies in Japan are continuing to invest in LTS projects, most of them heavily developmental. They have more engineers assigned to superconductivity R&D than the American firms. The strength of the Japanese commitment is visible not only in numbers of people, but in the range of businesses represented among the companies that have begun to invest. HTS R&D spans glass and ceramics, shipbuilding and steel, in addition to microelectronics, computers, consumer electronics, and electrical equipment. The Japanese firms can nonetheless be grouped into two classes: 1. 2. Some have relatively extensive experience in LTS. This group includes manufacturers of superconductin g magnets (e.g., for medical imaging systems). A number have been involved in Japans magnetically levitated train project (see box K later in this chapter). Toshiba, Mitsubishi Electric, and Hitachi have all built and tested prototype LTS generators. Sumitomo Electric, Japans leading producer of wire and cable, supplied superconducting wire for many of these projects. Sumitomo is also Japans (and the worlds) leader in small synchrotronswhich may emerge as a critical technology for production of next-generation integrated circuits. Finally, a number of Japanese firms have continued to pursue R&D on LTS Josephson devices, with highperformance computers in mind (see box J, later in the chapter). Others, new to superconductivity, began their research programs only after the discoveries in Zurich, Houston, and Tokyo. Some view HTS as important for existing businesses; others seek diversification. The first group includes electrical equipment manufacturers and other suppliers of cap-
PAGE 71
65 l l l IBy May 19SS, 4!# j~ M had j@ned f$l%C @e Ml members+epwt Memwtmdum 41S5, Tokyo Office of the U.S. National *~~ pom#atiun# $3,1-. m$yayreac%ed dy 25 of thwe, and an even mwller fraotion (5 of 45) of aaaociate members. For soma of thtwe compasdee, ofwmree, _*@~@ MY b @ new WCM of R&D, with Mtle internal a~tity. AR currency aonvemione in tbfr hex, aa dttawhere in the chapter, have been made at 130 yen to the dollar.
PAGE 72
66 ital goods. It also includes steelmaker, who have begun speculating, for instance, that magnetic levitation of strand-cast products could lead to better surface quality and higher yields (box I). The steelmaker are also trying to diversify, along with glass companies and shipbuilders. All these industries are in decline; opportunities for diversification and continued growth hold great attractions in Japan. Regardless of industry, many Japanese firms are pursuing research and applications development in parallel fashion. ll They have basic work underway, mostly in characterization, and people searching for materials with better properties. Development projects include work on thin films and efforts to fabricate wires. In these activities, the Japanese are proceeding much like their counterparts at the leading U.S. industrial laboratories. Japanese efforts differ in one major respect from those in the United States. Many firms in Japan have groups at work on feasibility studies and exploratory what if exercises. (Government programs, treated later in the chapter, show the same thrust.) These groups have a specific task: to think about possible commercial applications. In some cases, the efforts have already been carried to the stage of IIAs Sumitomo Electric Vice President Nakahara explains it, Japanese companies and other research organizations should pursue basic research and applications on parallel tracks to ensure cross-fertilization of effortsChoc/endo to wa IVanka what is Superconductivity], Nihon Keizai Shimbunsha, 1987, p. 91. preliminary designs and marketing analyses. The work is highly speculative, of course, but the Japanese believe it will help prepare for commercialization. Only a few U.S. companies have begun similar efforts. U.S. and Japanese Strategies Compared The Japanese see applications coming relatively quickly. When queried in the spring of 1987, scientists and research managers in Japan called for more basic research in their country, and efforts to develop applications based on patents filed in the United States. 12 Like Americans, they viewed superconductivity as largely a research enterprise for nowwith the research laying groundwork for commercial competition that would come soon, perhaps as soon as one to three years. In essence, Japanese companies are pursuing a three-pronged R&D strategy: 1) basic research; 2) development, aimed mostly at processing; and 3) product planning and market evaluation. The last of these carries the gravest potential consequences for U.S.-Japan competition. If technical developments in HTS proceed as rapidly over the next two or three years as during 1987, Japanese firms will be in better positions to move toward commercial applications than American companies. If U.S. firms wait to think about product and process developments until the research results IZChodendo Busshitsu: Nichibei Gokaku no Kaihatsu Kyoso [Superconducting Materials: Japan and the U.S. on a Par in Competition for Development], Nikkei Sangyo Shimbun, May 12,1987.
PAGE 73
6 7 Box I.HTS R&D at Nippon Steal Why did the largest steel producer in the world put 40 people to work on HTS in February 1987? Nippon Steel may be the leader in market share, but its sales have been decliningpartly the result of structural changes in the Japanese economy, leading to declining demand for steel, and partly a consequence of greater competition in international markets. Company strategists have two major tasks: 1) finding ways to make steel more cheaply, thereby helping the company compete in its primary if shrinking businesses; and 2) identifying new opportunities. Superconductivity fits both objectives. Planners see Nippon Steel as bringing three primary technical strengths to HTS. First, the company has always designed much of its own production equipment. It has process engineering skills, not only for making steel, but for titanium and other metals as well. Second, the firm has expertise in wire manufacture-a technology that could turnout to be important as HTS matures. Finallyand most important-Nippon Steel has worked hard over the years to develop technical capabilities in ceramics. Originally, most of this work was in refractories for furnaces. More recently, the company has sought to diversify into high-technology ceramics, and also into silicon production for the semiconductor industry. To the extent that the new superconducting materials will demand expertise in ceramics and other advanced materials, Nippon Steel believes it will be well-placed. None of these perceived strengths may turn out to be sufficient to place the firm in the forefront of HTS. Nippon Steels executives might be grasping at straws. Nonetheless, the company has looked with some care at 50 or more potential applications of HTS. Some of these analyses have been taken to the point of comprehensive feasibility studies. For example, company engineers have evaluated the prospects for continuous strand casting using superconducting magnets to confine and float molten steel, followed by in-plant materials handling are in, they may lose out competitively. Japanese companies will already have thought through those steps, weighed the potential problems, considered alternativesperhaps anticipated some of the follow-on technical work and even begun to pursue it. The Japanese approach probably costs more some R&D groups will pursue false trails, companies may be paying the salaries of too many people working on also based on magnetic levitation. overlapping projectsbut the eventual rewards could more than make up for this. Japanese managers find it strange that American companies believe they can track a technologys development, waiting for the right time to begin product development, without actively and aggressively pursuing that technology in their own laboratories. (Of course, many Japanese firms did just this not so many years ago.) JAPANS HTS INITIATIVES Many countries are pursuing HTS research, but talk of a superconductivity race has focused on the United States and Japan. 13 Sumitomos many hundreds of patent applications, for example, have drawn widespread attention. The 13See, for example, Two Different Cadences in the Superconductor Race, Washington Post, May 20,1987, p. Al; U.S. Leading Slightly in Superconductor Race, Japan Economic Journal, June 13, 1987, p. 15. This section is based in part on interviews with government officials in Japan during the fall of 1987. For background on Japrace is certainly a real one in terms of science. Laboratories around the world confirmed superconducting behavior in the thalium-based materials in a matter of hours. anese industrial policy, see International Competitiveness in Hectronics (Washington, DC: Office of Technology Assessment, November 1983), pp. 413-422. On patenting, below, see S.K. Yoder, Rush to Exploit New Superconductors Makes Japan Even More Patent-Crazy, WaZl Street Journal, Aug. 27, 1987, p. 18.
PAGE 74
68 Preoccupation with Japanese efforts is hardly a surprise, given Japans huge trade surplus with the United States, and the longstanding view by some that Japanese companies have been getting a free ride from American research. 14 Press reports have suggested that Japan is taking off in superconductivity, with government agencies in the lead. 15 MITI asked for 3.5 billion yen (about $27 million at 130 yen to the dollar) for superconductivity in fiscal 1988, well above the previous years level (550 million yen, $4.2 million). (Japanese budget figures do not include breakdowns for highand low-temperature superconductivity.) But the image of a government-coordinated, crash program in HTS is false. The policy environment for superconductivity remained in flux in Japan during 1987. The difference is this. After the middle of the year, the outlines of Japanese Government policies began to solidify. By early 1988, they had taken shape. U.S. policies, in contrast, remain in a state of considerable disarray. Government Resources for Superconductivity In Japan, four principal agencies have competed with one another for resources to support HTS (table 6): the Science and Technology Agency (STA); MITI; the Ministry of Education (Monbusho); and the Ministry of Transport. Key players in government, industry, and universities have been seeking to get into the superconductivity game, looking for money and the authority to expand their programs. Much of the substance of Japanese industrial and technology policies emerges from behind the scenes, the product of long-standing ties among government officials, corporate executives, and, in a case like HTS, senior professors in the leading universities. A host of advilqThiS perception is m exaggeration. See ]nternatjona] Competition in Services, op. cit., pp. 202-203. ]5For example, ton one hand, companies vie to beat each other to the patent office and marketplace. But at the same time, archrivals join forces on certain tasks when speed is essential and the research is risky. And the MITI orchestrates it all. See S.K. Yoder, Superconductivity Race Shows How Japan Inc. Works, Wall Street Journal, Aug. 12, 1987, p. 6. The same article calls superconductivity a Japanese obsession. Table 6.Major Japanese Government Programs and Activities in Superconductivity Science and Technology Agency (STA): l l l l Multicore Project Nine laboratories and other government organizations participating in work on theory and database development, materials characterization, processing, and technology transfer. The lead laboratories are the National Research Institute for Metals (NRIM, with work on theory, databases, thin films, and a superconducting generator) and the National Institute for Research on Inorganic Materials (materials synthesis and new material development, crystal structure determination, microstructure control). Japan Atomic Energy Research Institute R&D on superconducting magnets and applications. Japan Research Development Corporation Primarily measurement work. New Superconductivity Materials Research Association (Forum) Primarily information exchange. Ministry of International Trade and Industry (MITI): l Electrotechnical Laboratory (ETL), plus other MITI facilities R&D on superconducting electronics (e.g., Josephson devices), as well as new materials (including superconducting polymers). l Moonlight Project Superconducting generator. l Support for Technologies Needed for Research and Processing Thin film fabrication techniques; low temperature processing of bulk materials. Research Associations (See text.) International Superconductivity Technology Center (ISTEC) Ministry of Education (Monbusho): l University research support Examples: Professor Tanakas group at Tokyo University, working particularly on new materials; Professor Mutos work on theory at Tohoku University; support for the Ceramics Center at Tokyo Kogyo University. Ministry of Transportation: l Support for the Magnetically Levitated Train project at the Railway Technical Institute SOURCE: Office of Technology Assessment, 1988. sory committees, reports, and budget proposals also contributed to the rise of superconductivity on Japans policy agenda. The statements of the Council for Science and Technologyan advisory group chaired by the Prime Minister and including the directors of MITI, STA, and other major ministries are typical. In August 1987 the Council issued a report on superconductivity calling for hybrid basic research, involving experts from many fields. Japans national laboratories should promote cooperation with industry and universities, and reward in-
PAGE 75
6 9 dividual excellence, the Council urged, envisioning HTS R&D as a major step in improving Japans creative research capabilities. STA and the Multicore Project Each agency is independently pursuing these broad goals. In the summer of 1987, when HTS became the focus of attention worldwide, STA set up a Committee for the Promotion of Research and Development of Superconducting Materials. The nine members represent the major STA laboratories, the Japan Research Development Corporation, and the Interministerial R&D Division of the STA Research and Development Bureau. The Committees report stressed the many unknowns concerning superconductivity, and recommended a high priority for basic researchnot surprising, given the STA mission. 16 Highlighting the role of government, particularly STA, the Committee urged that Japans national laboratories, already credited with significant contributions in HTS, accelerate their efforts even more. While emphasizing the need for research, the members also advocated preparations for commercialization of new products. Their report touches on opportunities for international cooperation, recommending that Japans joint government-industry R&D programs be opened to foreign participation, and that Japan seek to make a global contribution to HTS. At the same time, the Committee underscored the importance of making research results from STA laboratoriesof which there are fivewidely available. The STA Committee sees these laboratories, and the agencys new Multicore Project, as the bridge between university science and corporate applications. Companies and universities participate through the New Superconductivity Materials Research Association, which had about 130 members at the end of 1987. The Multicore Project, with a budget of more than 2 billion yen (about $16 million) for fiscal 1988, aims to strengthen the capabilities of STA labINew Developments in Superconducting Materials R& D, Science and Technology Agency, Tokyo, Sept. 21, 1987. oratories in HTS, and to speed transfer of research results to industry. 17 A number of STA laboratories will get more money for HTSnotably the National Research Institute for Metals (NRIM), which plans a major thrust in materials characterization. NRIM scientists discovered the bismuth oxide HTS composition in January 1988. The National Institute for Research on Inorganic Materials and the Atomic Energy Research Institute will get most of the rest of the STA money. Nine laboratories in total will participate in the Multicore Project, with core research work going on in each. Given this decentralized approach (no physical relocations are planned), STA will rely on a steering committee with representatives from industry and universities, as well as government, for coordination. The steering committees job will be a difficult one, the more so ifas STA officials hopeMITI laboratories can also be pulled into the Multicore Project. At present, however, MITI has its own quite independent plans. MITI Programs A recent report by the Advisory Committee on Superconductivity Industrial Technology Development, made up of representatives from industry and universities, reflects the perspective of MITIto Western eyes, the most visible agent of Japanese industrial policy 18 Like STA, the MITI committee sees the government role as one of helping industry make use of research results from the national laboratories and universities. But MITI goes further in adITChodendo Zairyo Kenkyu Muruchikoa Projekuto 63 nendo Kisan Yokyu Sokatsuhyo [Budget Request for Multicore Project for FY1988]; also Superconductor R&D to Industrial Application, JPRS Report-Science& Technology, Japan, JPRS-JST-88007-L, Foreign Broadcast Information Service, Mar. 11, 1988, p. 100 [translated from Nikkan Kogyo Shimbun, Jan. 1, 1988]. The Multicore name signifies that multiple organizations form the core of the project, emphasizing the thrust toward coordination and reorientation rather than an all-new initiative. The project accounts for about two-thirds of STAS fiscal 1988 budget request for superconductivity, which totals 3.1 billion yen. (Japans fiscal year begins in April.) STA has also sought funding for a superconducting generator project. Iechodendo sangyo (lijutsu Kaihatsu Kondankai (1988).
PAGE 76
70 vocating national projects, not only for R&D on the new materials, but for applications in electronics and electrical machinery. Box J notes the Ministrys support for Josephson computing technologiesa field where Japan began by following the path laid down by IBM and other U.S. companies, then persisted after American firms cut back their efforts. A good portion of MITIs 1988 superconductivity budget will go toward applications. Examples include a new project on thin films and Josephson devices, part of the Technologies for Next Generation Industries program of the Agency for Industrial Science and Technology (AIST is part of MITI). The Ministrys 70 megawatt (MW) generator project, based on LTS technology and also scheduled for more than $10 milliona hefty slice of the 1988 MITI superconductivity budgetfollows several years of feasibility studies. Motivated in part by the search for energy savings, goals for the 8-year project range from improvements in methods for processing superconducting wire to construction of a complete prototype. 19 Officials say that HTS technologies will be utilized if available. Late 1987 saw a major step for the 70 MW generator project, the formation of a research association (kenkyu kumiai). As is typical of Iochodendo Hatsuden Kanren Kiki-Zairyo Gijutsu no Fizabirite Chosa Kenkyu, March 1987. The original proposal, advanced in 1985, was much more ambitious, calling for a 200 MW generator to be built in 5 years. Photo credit: IBM Corp. Memory cells in experimental Josephson junction integrated circuit chip these research associationscentral mechanisms of Japanese technology policyMITI not only helped with the planning, but will assist with administration and furnish ongoing financial support. Likewise typical of MITI projects, the research association brings together participants with a range of technical strengths, and companies from different industries: the members include two cryogenic engineering firms, the Fine Ceramics Center, the Central Electric Power Research Institute, and a number of major electric power and electronics firms. MITI sees its role as supporting industry not only by creating incentives for applications-related R&D, but by spurring productive interactions among firms and industries that might not otherwise collaborate. MITI, like STA, also runs its own laboratories. The Electrotechnical Laboratorywhich has won worldwide respect for its research- has been involved in superconductivity since the middle 1960s, when the laboratory began R&D on LTS magnets for MHD (magnetohydrodynamic) power generation under the Moonlight Project. More recently, ETL has attracted particular notice for its work on niobium-based Josephson devices (box J). ETLs overall 1987 budget came to $57 million; like many organizations in the United States, the laboratory was able to reprogram funds internally for HTS during 1987; MITI will get $2.5 million for ETL research on the new superconductors in 1988. The laboratory has several groups, and about 40 people in total, working on superconductivity (the ETL research staff numbers 560). The Ministry seeks to involve private corporations in its efforts through mechanisms ranging from research associations to advisory boards and symposia. Industry is MITIs major constituent, and the Ministrys HTS programs will follow patterns laid down over the years for supporting other industries and other technologies-e. g., semiconductors, computers, biotechnology. The Ministry of Education The Monbusho, which supports university research, has a larger R&D budget than any other arm of the Japanese Government. Sup-
PAGE 77
71 Box J.-Josephson Junction Computer R&D: From the United States to Japan 1 The pursuit of a Josephson-based computer has taken quite different paths in the United States and Japan since the early 1970s. Josephson devices provide the fastest electronic switches known, hencein principle-the fastest digital computers. Because they are Superconducting devices, with very little power dissipation, JJs can be packed tightly together. Theoretically, therefore, a computer built with JJs could be very compact, as well as extraordinarily fast and powerful. U.S. Efforts Three U.S. corporations pursued JJ R&D for computer applications: AT&T, IBM, and Sperry Univac (which later merged with Burroughs to form Unisys). Each made significant contributions to the JJ technology base. Beginning in the 1960s, more than 10 years of research at AT&Ts Bell Laboratories produced a much better understanding of the physics of JJs. IBM went much farther, building a prototype of the circuitry for a complete computer, as well as exploring fabrication methods for JJ logic and memory chips. Sperry concentrated on JJs made from refractory materials such as niobium and niobium nitride (instead of the lead alloy used by IBM), and developed processing methods for highperformance, all-niobium circuits. All three companies had scaled back or abandoned their JJ projects by the early 1980seach for its own reasons. AT&T terminated the Bell Laboratories program in 1979 after deciding that the technical hurdles to practical applications were formidable. Sperry abandoned its effort to develop a JJ computer in 1983, after closing its Sudbury, Massachusetts, research center, the focus of the work. (JJ research by Sperrys Defense Systems Division, aimed at sensors, continued.) IBM, with the most ambitious program, was spending about $20 million annually by the early 1980s, with the National Security Agency (NSA) providing about $5 million of this. Although NSA urged continuation, IBM drastically scaled back its effort in 1983, ending pursuit of a working computer, after its Yorktown Heights Laboratory was reorganized and the JJ work came under new management. Logic chips based on IBMs experimental production technology performed adequately, but the memory did not; the new management team estimated that improving the memory chips would add another 2 years to the schedule. By that time, management reasoned, continuing progress with more conventional silicon and/or gallium arsenide chips would make it hard for a JJ-based machine to offer compelling advantages in speed or processing power. Before ending its JJ program, IBM came close to an agreement with Sperry for joint development of Josephson technologies. IBM had the most advanced. designs but was struggling to fabricate them, while Sperry had proven processing technologies. The agreement was almost 18 months in the making, and had apparently cleared the antitrust hurdle after the NSA proposed taking the project under its wing. But the agreement was never consummated because Sperrys management decided to decentralize its R&D among its operating divisions, and reassigned its JJ computer group to the Defense Systems Division--a reassignment that key technical employees declined. *Much of fhe information in thia box is based on interviews. Also see A.L. Robinstm, New Superconductors for a Supercomputer, Science, Jan. 1, 1S82, p. 40; A.L. Robinson, IBM Drope Superconducting Computer Project, Scfence, Nov. 4, 1983, p. 492; JTech Panel Report on Opto-& Microelectronics, Japaneee Technology Evaluation Program, Science Applications International Corp. under contract No. TA-83-SAC02294 from tbe U.S. Department of Commerce, May 1985, $action 11; Govemrnents Role in Computers and Superconductors, prepared for OTA by K. Flamn under contract No. H3-6470, March 1988, pp. 47-52.
PAGE 78
72 port for superconductivity is relatively new, however, going back only to 1984. In 1987, Monbusho funded 41 mostly small projects on superconductivity at Japanese universities, with spending totaling $4.3 million. 20 For fiscal 1988, Monbusho spending on superconductivity will ZODaigaku Kankei ni okeru Chodendo Kenkyu no Tsuishin ni tsuite [Concerning Support for Superconductivity Research in Universities], Monbusho. The Ministry of Educations figures for support of university R&D normally exclude salaries, which are paid out of other accounts. (Other Japanese Government agencies typically include salaries in their published R&D budget figures, just as in the United States.) reach about $14 million. Although the Education Ministry has placed superconductivity at the top of its list for greater support, it ranks third behind MITI and STA in its 1988 budget for direct support of superconductivity. In addition to the many small projects it funds, Monbusho provides much of the support for a few large programs headed by internationally known scientists. For instance, Professor Shoji Tanakas group at Tokyo University will receive more than $700,000 during 1988. Professor Tanaka (who recently retired
PAGE 79
73 from the university to direct research at MITIs International Superconductivity Technology Center, ISTEC), has also been awarded a 3-year grant of $1.6 million for Specially promoted Distinguished Research. Professor Yoshio Muto, at Tohoku University, whose group has been designated one of 30 priority research projects, will get more than $2 million over the 3-year period 1988-1990. These large universitybased efforts generally include participants from a number of universities, selected by the lead professor. Taking university research as a whole, the scope in Japan is narrower than in the United States, and the quality substantially lower. Japan has fewer centers of excellence, and a more rapid drop-off in quality as one moves down the scale. The best institutions in Japan are very good. There simply are not that many of them. Superconductivity, however, has been an exception. Before the discoveries in HTS, the field had been something of a backwater. Interest had been declining in the United States, more so than in Japanese universities. Recently, American scientists have given the Japanese high marks for research in superconductivity. Tokyo, Tohoku, and Kyoto Universities have been getting about three-quarters of Monbusho superconductivity funding. Some of the research groups at these schoolse.g., Tokyo Universitys in superconductivityare on a par with the best in the world. And even in their less known schools, the Japanese excel at some kinds of worknotably painstaking empirical research. Most important, R&D in Japans universities is improving rapidly, in part because of the efforts of the younger faculty members trained in the United States. The Ministry of Transport and the Maglev Train Japans magnetically levitated (maglev) train projectbox Kwhich has been underway for two decades, is scheduled to get more than $4 million during 1988 from the Ministry of Transport, which oversees the effort. While current prototypes use LTS magnet systems for both suspension and propulsionand a relatively small fraction of the programs funds go toward superconductivity R&Dthe engineers leading the project hope that HTS materials can eventually be incorporated. The maglev program typifies the kind of longterm, continuing effortin this case beginning in the 1960sthat Japanese decisionmakers expect will pay off in eventual commercialization. Although maglev R&D supported by the U.S. Government ended in 1975, the Japanese have persevered. To Japan, the linear motor car has become a symbol of indigenous technology development. Summarizing the Government Role As the many different programs mentioned above suggest, the scene has changed rapidly in Japan. Major ministries involved in superconductivity R&D steered more money to HTS during 1987, and have substantially higher budgets for 1988. A superconductivity city has even been proposed, where research would be centralized and applications tested. 21 Increases in government spending send unmistakable signals to industry, as well as to the universities and national laboratories. Japanese industrial policy works primarily through incentives. Ministries seek advice from zINo agency has linked itself with the proposal, which seems to be a trial balloonSuperconductivity City Project, Science and Technology in Japan, November-December, 1987, p. 43. Photo credit: Japan External Trade Organization Japans prototype linear motor car, a magnetically levitated train.
PAGE 80
7 4
PAGE 81
75 business leaders in the early stages of policy development. The processes through which officials in government and the private sector interact, and informal encouragement of industry efforts by government, arguably play a role at least as important as direct financial support. Government funding for R&D projects tends to be modest; consistently, private industry has paid for three-quarters or more of all Japanese R&D, compared with about half in the United States (table 5). At the same time, public funds for the maglev train provided a stimulus for companies like Sumitomo Heavy Industries to build their experience base in superconductivity and related problems in cryogenic engineering. Japanese companies participate as members of consortia formed by MITI to undertake projects such as the 70 MW generator. In the Japanese view, such projects leverage public investment, helping break technological bottlenecks and diffuse results to industry. Companies participating in Japans government-sponsored R&D efforts normally contribute about half of the project funding. When it comes to basic research, the government share is about 50 percent in Japan, versus two-thirds here. As pointed out earlier in the chapter, many leaders in business, government, and universities in Japan are pushing for improvements in basic research, seeking greater creativity and originality. Because of budget pressures, Government support for basic science has been growing at an annual rate of only 3 percent, slower than the overall rate of R&D growth in Japan. Thus the government share of basic research funding has been declining. As the yen rises relative to the dollar, Japanese spending, when translated into dollars, appears more impressive. Calculated at exchange rates current at the end of 1987, direct Japanese Government support for superconductivityexclusive of salariesseems likely to be about $70 million during 1988. Budget figures in Japan do not break out LTS and HTS, but a good portion of the total will no doubt support ongoing work with low-temperature materials. Funding increases have been sharp, coming after a period of relatively low spending on superconductivity (leaving aside such projects as the linear motor car, where superconductivity is a means to an end). In fiscal 1986, for example, MITI spent about $2 million on LTS technologies. And set against overall Japanese Government R&D supportitself relatively small compared to corporate R&D superconductivity remains a minor item. 22 MITIs 1988 HTS budget exceeds that of the other agencies, but it would be a mistake to conclude that MITI is tightly coordinating Japans superconductivity policies. In OTAs interviews, MITI officials argued that, at this stage in the development of HTS, competition among ministries and research groups should be seen as healthy. STA staff, meanwhile, hopes that MITI laboratories will eventually join the Multicore Projectwhile conceding that this is unlikely in the near term. Who Has the Lead Role, Government or Industry? Westerners often misconstrue relationships between government and industry in Japan. MITI and other ministries may try to influence corporate decisions, but Japans Government does not issue directives to industry. A more accurate picture of Japanese policymaking sees government-industry interactions based on processes of reciprocal consentcontinuing discussion and negotiation. 23 Corporate leaders are heavily involved in building consensus and helping shape government programs. HTS will be no exception. In superconductivity, industry has influenced government policies through frequent meetings with ministry officials. At least a third of the members of MITIs Advisory Committee on Superconductivity Industrial Technology Development come from the private sector. More than a hundred Japanese corporations belong to the STAs newly formed Shin Chodendo 22The Japanese Government budget for all science and technology activities totals 1,700 billion yen for fiscal 1988about $13 billion. See Report Memorandum #147, Tokyo Office of the U.S. National Science Foundation, Feb. 5, 1988. z3R,J, Samuels, The Business of the Jzqmnese State (Ithaca, NY: Cornell University Press, 1987).
PAGE 82
76 Zairyo Kenkyukai [New Superconductivity Materials Research Association], best known as the superconductivity forum. The forum, chaired by Dr. Shinroku Saito, serves as a window between corporate members (who pay an annual fee of about $1,000) and the universities and national laboratories involved in the Multicore Project. According to the director of STAs Research and Development Bureau, the forum will hold workshops and symposia, undertake brainstorming in support of the Multicore Project, and encourage cooperation in research, both domestically and internationally. Many participants, including Dr. Saito, also advise other ministries; thus the forum helps build linkages within the Japanese Government. The Fine Ceramics Center (FCC) illustrates a different mechanism. Government and industry have both provided money for an extraordinarily well-equipped laboratory in Nagoya, with participating companies sending scientists and engineers. In contrast to some other MITIsponsored R&D efforts, many of which have had staffs viewed as second rate, the FCC appears to have attracted highly qualified people. The companies continue to pay their salaries, and they help transfer technology from the Nagoya laboratory back to their employer. When it comes to Japans national laboratories, opinions differ as to whether corporations give more than they receive. For instance, at the National Research Institute for Metals, which normally hosts a half-dozen people from industry who work alongside NRIM scientists, laboratory officials contend that they have been ahead of industry in at least some areas of superconductivity. Organizations like NRIM also let contracts to companies, including large corporations (New Japan Steel, Toshiba). While laboratory managers view contract research as a mechanism for helping industry, the companies which make little profit on such worktend to see it as part of their contribution to the larger national effort. In the future, government scientists may have a chance to spend time working in corporate laboratoriessome of them much better equipped than government facilitiesbut so far this has been rare. Legal provisions, only recently relaxed through new legislation, have limited such arrangements. Direct cooperation in research between companies and universities has likewise been limited. This is also changing, however. Professor Tanaka recently had 10 scientists from a group of private companies working in his Tokyo University laboratories. The Monbusho reports a total of 300 cooperative projects linking universities and companies during 1987, 11 of them (all with Monbusho sponsorship) in superconductivity. 24 In some contrast to efforts in the United States, many of which seek to push universities into doing industrially relevant work (see the next chapter), rhetoric in Japan stresses cooperation in projects of interest to both sides. Japanese leaders, like those in many countries, view ties among universities, industry, and government as weak. Statements on science and technology policy continually highlight the need for more effective working relationships. Industry tries to help by donating equipment to the universities, but professors worry aloud that industry will steal their best research workers. At the same time, senior professors typically help steer their graduates to particular companies, helping build long-lasting informal communications networks. Professional societies and study groups also bring people together, providing opportunities for working-level scientists and engineers from industry, government, and the universities to share information. In this respect, they replicate the function of high-level advisory committees involving senior professors, corporate executives, and ministry officials. Recent changes in the lawfor instance, making it easier for faculty members to consult encourage interactions with industry, but many Japanese officials think further steps will be needed. Broad success in basic research would seem to demand such cooperation. Given the ZSangaku Ittai e Hirogaru Koryu [Expanding Exchange Between Industry and Universities], Nihon Keizai Shirnbun, Jan. 20, 1988, p. 1,
PAGE 83
77 slow growth in the government R&D budget, the industry role is a critical one; superconductivity promises to be a prime test case. As noted earlier in the chapter, many Japanese companies Sumitomo Electric is a good examplehave been expanding their basic research efforts, while also pursuing parallel programs of applied R&Din HTS. The new opportunities have pushed many firms toward more basic workwhich they see as the necessary preliminary to commercializationand sensitized them to the importance of university science. Even so, a major reorientation of Japanese R&D toward fundamental research will require institutional, cultural, and political shifts. The university system is widely viewed as hierarchical and stifling, offering inadequate incentives to bright young researchers. Change has begun, but it is not clear how far it will go or how deeply it will penetrate. Rivalry or Cooperation? The Internationalization of Japanese Superconductivity R&D International cooperation in HTS has been a central theme in pronouncements by government officials in Japan. MITI has opened its HTS programs to foreign companies. The STA states that it will promote international collaboration under the Multicore Project. In addition, the Key Technology Centersponsored by MITI and the Ministry of Posts and Telecommunicationshas provided financial support for foreign engineers and scientists who wish to work in Japan. Finally, the Ministries of Foreign Affairs and Education have announced a postdoctoral fellowship program that will bring recent graduates of overseas universities to Japan for research. Possibilities for U.S.-Japan cooperation in HTS also exist on an agency-to-agency basis. The U.S. National Bureau of Standards and ETL have been exchanging scientists for a number of years. Both have informally expressed interest in cooperation on HTS-related standards. Why the stress on internationalization? There are two major reasons: l The United States, along with other nations, has been pressing Japan to make a l greater contribution to global welfareone commensurate with the size of the Japanese economy and Japans technological capabilities (box L). Among other things, this implies a greater commitment to sciencethe fruits of which should benefit alland to the transfer of technologies to other parts of the world. Opening Japanese research institutions to greater foreign participation would be a first step. In many official policy statementsincluding those on HTSJapan has pledged to take such actions. More than just altruism, internationalization would-serve Japans interests as well. Foreign scientists and engineers will help invigorate Japanese laboratories, encouraging new approaches to research, and breaking down some of the traditions which particularly in the universitiesseem roadblocks to creativity. Japanese leaders also realize that they may have to open their own doors to retain access to R&D from other countries. With science and technology holding the keys to continuing economic growth in the 21st centurya firm belief in Japaninternationalization can be viewed as a strategic and economic imperative. The stress on international cooperation does not signify any slackening in Japans efforts to develop indigenous technologies. The Japanese view it as a complement to these effortsfar more than a matter of image, it is an intrinsic element in Japans strategy for competing in a world of intensifying global rivalries. So far, as box L indicates, rhetoric has overshadowed results. MITIs pitch for international collaboration focuses on the International Superconductivity Technology Center, established in January 1988. ISTEC, which gets financial support from MITI, will be located near Tokyo on a site formerly owned by Tokyo Gas. More than 85 Japanese companies have signed on as founding members. Although the initial fee for full members is about $800,000, with annual charges of about $100,000, the costs are considered donations and earn the companies tax benefits. Associate members, who pay
PAGE 84
78
PAGE 85
7 9 much less, will not be able to participate in research or have immediate access to R&D results, but simply to ISTEC publications and symposia. ISTEC plans not only to conduct research in its own facilities, but to support R&D in other institutions, review and evaluate research for its members, and carry out feasibility studies on applications of HTS. To benefit from full membership, foreign companies would need Japanese-speaking employees, skilled in relevant technologies, on site in the ISTEC laboratory. As of May 1988, no foreign companies had joined as full members, although several had signed on with associate status. Are American companies missing a bet by not joining ISTEC? For smaller firms, the costs pose a major barrier. But a number of U.S. companies with R&D operations in Japan could certainly afford them. Some form of jointly sponsored membershipe.g., through an industry association, or a joint venture such as Microelectronics & Computer Technology Corp. might also be possible. If ISTEC yields impacts comparable to past MITI-promoted R&D efforts e.g., the very large-scale integrated circuit project of the late 1970sthen participation could pay off. Even if the results in terms of research outcomes prove meager, active participation helps keep tabs on the competition. This, after all, has been a primary motive for Japanese firms to join in such group efforts. CONCLUDING REMARKS The U.S. business culture differs from that in Japan, R&D strategies in American companies tend to be driven by hard-headed calculation of risks and rewardswhich does not encourage aggressive commitments to HTS. Most U.S. firms hope that someone else will do the fundamental research. Many American executives feel uncomfortable with this short-term approach, sometimes defensive. But they see little choice, given the way U.S. financial markets operate. Japanese executives work in a society and an economy with a different set of traditions and rules. They too must worry about profit levels, but these are not the most important influence on their behavior, Japanese managers think first about growth and market position. Furthermore, they are acutely aware that they can no longer depend on technologies from the United States and Europe. Managers in Japan are attempting, often with some fumbling, to increase their firms research capabilities, seeing this as one road to continued expansion. Given the differences in attitudes and in approach to R&D, HTS has stimulated contrasting responses. As a generalization, large Japanese companies have more people at work on HTS, doing a greater variety of things. Japanese managers see HTS as a technology of paramount importance for global competition in the 1990s and beyond. U.S. executives might agree, but they also see the risks and uncertainties more starkly. They believe commercial products are farther offthat HTS will remain in the laboratory for some years to come. Thus, most U.S. R&D efforts could be described as selective and probing. In contrast, most of the Japanese efforts are relatively broad, with people already assigned to think about applications. In pursuing their strategies, Japanese companies are studying superconductivity now as a potential commercial technology. American companies are not. The Japanese companies could be wasting their time and money. At this point, no one knows. But if the pace of discovery in the future matches that of the past year, Japanese companies will be better positioned. In the United States, some of the first HTS applications may well come from small, startup firmsfinancially weak, and likely to face difficulties in growing. The pattern is clear in biotechnology, where startups have had to link with larger companies to proceed with com-
PAGE 86
80 mercialization. Thus far, of course, the startups are outnumbered by big American companies like Du Pont and IBM. In Japan, the large, diversified, and financially strong companies have the field largely to themselves. These Japanese firms are poised to mov e quickly into production and marketing, on a worldwide scale if they choose. In the past, this asymmetry in industrial structure has ha d powerful impactse.g., on competition in microelectronics, where American firms have fallen behind for reasons that include lack of financial muscle. It remains to be seen how the story will unfold in biotechnologyor in HTS. In the United States, cooperation between Government, industry, and the universitie s tends to be ad hoc, motivated by particular circumstances. There is no indication that HTS will be an exception. Japanese companies compete intensely with one another, but are nonetheless quite capable of cooperating on projects judged to be in their interests, especially when MITI or government agencies seek to foster these projects. Japanese Government funding for R&D in superconductivity will not match spending by the U.S. Government (although the exclusion of salaries from some of the Japanese budget figures makes comparisons difficult). Including both LTS and HTS, the U.S. Government will spend more than twice as much in fiscal 1988. More important, however, Japanese firms have many more people at work on HTS than American firms. Companies commercialize, and companies in Japan have stronger commitments to superconductivity. Neither country has a coordinated national initiative. Both seek to promote cooperation among universities, industry, and national laboratories. While business and government in Japan do not always find it easy to cooperate, they do exchange views and work toward consensus. And, if Japans policy cannot be described as a coordinated plan, policy directions have been debated much more thoroughly than here. By the beginning of 1988, policy objectives in Japan had been reasonably clearly defined. They show a clear recognition of specific needs and specific problems impeding commercialization, and the Japanese Government aims to help solve them.
PAGE 87
Chapter 4 U.S. Technology Policy: Issues for High-Temperature Superconductivity
PAGE 88
CONTENTS Page Summary . . . . . . . . . . . . . . 83 Government Support for HTS R&D..... . . . . . . . 86 Funding Levels. . . . . . . . . . . . . 86 Defense-Related R&D. . . . . . . . . . . . 93 HTS R&D in the Energy Department Laboratories. . . . . . 99 Other Mission Agencies: NBS and NASA... ........................10 1 NSF and the University Role. . . . ............................102 Disciplinary Boundaries . . . . . . . . ..........102 NSF Centers . . . . . . . . ....................103 Technology Transfer: The Federal Laboratories .......................11 0 New Rules for the Laboratories. . . . . . . . . ...110 Transferring HTS R&D . . . ..............................111 Laboratory Personnel 112 Cooperative R&D . . . . . . . . ................113 State Programs and Approaches . . . . . . . . ....114 Technology Interchange With Japan . . . . . . . . ..115 Participation and Monitoring, . ...............................11 5 Language Training; Fellowships in Japanese Laboratories .............116 Technical Information. . . . . . . . .............117 Foreign Access to U.S. Technology ................................11 7 Concluding Remarks. . . . . . . . . ..............118 Boxes Bo x Page M. Coordinating the Federal Effort . . . . . . . . 90 N. DoD and Postwar High-Technology Development . . . . . 96 0. Multidisciplinary Research in American Universities. ...............105 P. Reciprocity in Flows of Technology and Information. . ............118 Figure Figure Page 4. National Science Foundation Research Support ..................,..104 Tables Table Page 7. Summary Guide to Policy Options . . . . . . . . 84 8. Federal Funding for HTS R&D... . . . . . . . . 87 9. Issue Area I: Funding Levels and Priorities for Federal R&D . . 88 10. Department of Defense Funding for HTS R&D . . . . . 94 11. Energy Department Funding for HTS R&D . . . . ., ...,...100 12. National Science Foundation Funding for HTS R&D ................103 13. Issue Area II: Strengthening Interactions Among Universities, Industry, and Government .****** l ******* l ******. l m.***** .*.***** l .***.*. l 108 14. Issue Area HI: Technology Interchange with Japan ...............,.116
PAGE 89
Chapter 4 U.S. Technology Policy: Issues for High-Temperature Superconductivity SUMMARY The preceding chapter discussed company strategies toward high-temperature superconductivity (HTS) in the United States and Japan, as well as the policies of the Japanese Government. The question now becomes: How can U.S. Government initiatives help American companies with commercialization? Both this chapter and the next deal with Federal policies and what they mean for HTS. Both also go beyond superconductivity, taking up broader issues that affect commercialization and competitiveness, Many of these policy issues are matters of ongoing concern to Congress and the executive branch: the Federal R&D budget and its management; the health of university research; technology transfer from national laboratories to industry. Table 7 provides a guide to some 20 policy issues and options discussed in this chapter; tables 9, 13, and 14, which follow later, give more detail. As a glance at table 7 makes clear, many of the issues and options have relevance that goes far beyond HTS. By the same token, many of the policy questions important for HTS can only be understood in the broader context of U.S. technology policy. Federal agencies will spend some $60 billion on R&D this year (ch. 2). Industry will spend about as much, with private firms also conducting more than half the Government-funded total under contract. All companies that use technology live to some extent off the publicly financed storehouse of technical knowledge. The path to commercialization begins with this technology base. The overall size of U.S. R&D expenditures more than twice as much as Japan, and far more than any of the Western European nations presents something of a paradox. How is it possible, given spending on science and technology exceeding $125 billion, that the United States has a problem in technology? Why doesnt American industry have what it needs to compete? The question has two kinds of answers, both partially true. The first is that technology is not, in fact, the problemthat difficulties in commercialization and competitiveness lie elsewhere. The analysis in chapters 2 and 3 indicated that technology is part of the problemthough far from the whole problem. The second answer is that not enough of the R&D money goes toward commercially relevant technology development. Any analysis of the Federal role in commercialization must begin with a look at how the Government spends its $60 billion: l l l Nearly 70 percent goes for defense, up from 57 percent at the beginning of the Reagan Administration. The United States devotes a much larger share of total R&D outlays for military projects than most other countries. Defense gets less than 5 percent of the Government R&D budget in Japan. Much of the Federal moneythis year, about $20 billiongoes to the 700-plus national laboratories. For the most part, these laboratories do not have a good track record in transferring technology to civilian industry. While recent initiatives by Congress and the Administration have sought to strengthen interactions between the laboratories and industry, the process of change is just beginning. Outside of defense, aerospace, and health, Federal agencies spend little on applied research and development. Given the shortterm orientation of most of the R&D paid for by private industry, a wide gap often separates basic research and commercial technologies a gap that neither Government nor industry has been filling. 83
PAGE 90
. 84 Table 7.Summary Guide to Policy Options Issue area Option Relevance 1. Funding Levels and Priorities for Federal R&D (see Table 9 for details) A. Funding Levels for HTS l New money, agency priorities . . . . B. Continuity of Funding l Multi-year benchmark plan. . . . . . l Two-year funding trial . . . . . . C. National Science Foundation Budget Overall NSF budget increase. . . . . l Funding for university laboratory equipment . D. Weaknesses in the Industrial Technology Base l Review of U.S. technology base . . . . l Basic research tax credit . . . . . E. Setting Priorities for Federal R&D l Strengthen the Office of Science and Technology Policy. . . . . . . . . . Il. Strengthening Interactions Among Universities, Industry, and Government (see Table 13 for details) A. University-Industry Interactions; Multidisciplinary Research l Funding for NSF centers . . . . . Postdoctoral fellowships . . . . . B. Government-Industry Interactions: Technology Transfer and Joint R&D l l l l l l l Oversight on technology transfer from the national laboratories . . . . . . . Pilot program for transfers of HTS technology resulting from DoD-sponsored R&D . . . Technology transfer demonstration projects . Personnel exchanges . . . . . . Cooperative R&D with industry. . . . . Sharing costs with private R&D consortia . Support for State Government initiatives . . Ill. Technology Interchange with Japan (see Table 14 for details) Seed grant for office in Japan to monitor developments in HTS . . . . . . . . l Research participation and language training . Japanese technical literature. . . . . 1 . 2 . 3 . 4 . 5 . 6 . 7 . HTS HTS, but potentially broader HTS, could be broader general general ail commercial technologies general 8 . general 9 general 10 : : : : : general 11 . general 12 . HTS, but potentially broader 13 . general 14 . HTS could get special attention 15 . HTS, but potentially broader 16 . HTS, but potentially broader 17 . general, but HTS could get special attention 18 . HTS 19 . general 20. . HTS, but potentially SOURCE: Office of Technology Assessment, 1988. At present, the Federal Government maybe spending as much on HTS as the private sector. The agencies expect to spend $95 million on HTS in fiscal 1988. OTAs industry survey (ch. 3) found that 55 U.S. firms plan to spend about $97 million on superconductivity R&D (LTS as well as HTS) in 1988. While $95 million sounds like a lot, nearly half will go for military projects. Department of Defense (DoD) objectives shape R&D goals even at the level of basic research. Nonetheless, much of the fundamental understanding of HTS that results from DoD-sponsored research will support the overall technology base
PAGE 91
85 for HTS. Moreover, the Defense Advanced Research Projects Agency (DARPA) has emphasized processing in its HTS R&D; this work should yield commercial spinoffs. In general, however, civilian and military technologies have been diverging, as DoDs needs grow ever more specialized. This pattern is already evident in HTS, where prospective applications include passive shielding for protection from nuclear radiation, or sensors for the Strategic Defense Initiative (app. B). Moreover, in a period of tight budgets, DoD decisionmakersfrom project and program managers to laboratory directors and Under Secretaries scrutinize the R&D budget to make sure that immediate military needs get the highest priority. Basic research suffers in such periods, along with other work that might be of use on the civilian side of the economy. The Department of Energy (DOE) and its laboratories will get the lions share of the nonmilitary fundingnearly 30 percent of the Federal total. Ten DOE laboratories may have more to spend on HTS in 1988 than NSF will distribute to the Nations universities. DOEs basic research, like that of DoD, will help support the technology base. As for commercial technology, the laboratories are trying to develop new cooperative ties with U.S. industry. However, it could take years for effective working relationships to develop; in the absence of such relationships, DOE R&D may not make a major contribution to commercial technology development. The National Science Foundation (NSF) share of the HTS R&D budget, going almost entirely for university research, declined from 25 percent of the Federal total in 1987 to 15 percent in 1988. The universities do get some funding from DOE and DoD (especially through the basic research programs of the Air Force and the Navy). But the allocation of Federal R&D funds seems out of balance, given the great strength of American universities in basic research. Continuity of funding over the next 5 to 10 years will be just as important as the level and allocation in any one year (Options 1, 2, 3). The Federal budget for HTS is really nothing more than the cumulation of agency decisions and appropriations. Both Congress and the Administration could benefit from a better sense of the overall dimensions of the Federal effort, so that priorities could be weighed rather than simply emerging at the end of the yearly budget process. A benchmark, multi-year funding plan for HTS, which could be adjusted periodically (not at alla rigid blueprint), would help in making good decisions. Congress might also choose to experiment with multi-year authorizations and 2-year budgeting. These steps could help avoid too much duplication in agency R&D (some overlap can be desirable), as well as cuts in other needed R&D to provide money for HTS (little of the Federal total represents new money specifically appropriated for superconductivity). Many fields of science and technology vital for competitiveness do not get adequate research support; technical knowledge that could help American firms compete is not available when they need it. Often, the underinvestment is most severe in fields that lack glamour and the promise of immediate payback (examples range from materials synthesis to corrosion and wear)just those likely to suffer when more money must be found for an exciting new opportunity like HTS. Given the constraints on the Federal budget, any decision to begin filling some of these gaps by spending more on civilian R&D must begin with good information and a government-wide perspective, matters addressed in Options 6 and 8. Commercializing HTS will require multidisciplinary R&Dphysicists, chemists, materials scientists, and engineers. NSF can play a vital role in supporting multidisciplinary research in universities, where such work has seldom caught on (Options 9 and 10). While the Reagan Administration proposed doubling the Foundations budget over a 5-year period, Congress gave NSF very little increase for fiscal 1988. Sustained growth in the NSF budget will be needed if the agency is to increase its support of traditionally underfunded areas, including engineering researcha critical priority for competitiveness.
PAGE 92
86 If the Federal laboratories, in their turn, are to provide much help in commercialization, they will need to make sustained commitments to working with the private sector (Options 11 through 17). Congress, in several recent laws, has stressed the need for closer linkages between the laboratory system and industry. Agency responses have been mixed. With experience limited, it might be prudent for the DOE laboratories to adopt an experimental approach, beginning with pilot projects, rather than plunging into a full-fledged program of cooperative endeavors. Personnel exchange programs could also help shift the culture of the laboratories; scientists and engineers working in the laboratory system need to understand how industry functions and how the marketplace works if they are to help in commercialization. Chapter 3 outlined Japans proposals for international cooperation in HTS research. So far, American firms have not responded with much enthusiasm. Options 18, 19, and 20 suggest steps the Federal Government could take to help industry and professional groups test Japans openness to foreign R&D participation, and to monitor Japanese technical developments. Given the importance of person-to-person contact in technology transfer, early steps should include language training for U.S. engineers and scientists, so they can work inside the Japanese research system. Although the analysis that follows covers a broad range of issues related to HTS, it does not pretend to be a comprehensive discussion of U.S. technology policy. Nor do the 20 policy options address all the problems identified in earlier chaptersshort-term decisionmaking in U.S. industry, for example. This chapter has a more modest aim: examining alternatives for managing the Federal R&D budget to more effectively support the Nations commercial technology base without detracting from agency missions. Most of these are incremental policy adjustments; chapter 5 looks at more comprehensive alternatives. GOVERNMENT SUPPORT FOR HTS R&D Funding Levels Funding for HTS R&D has grown dramatically since the end of 1986; table 8 gives the best available estimates. 1 It is hard to criticize the totals; indeed, the increases shown in table 8 seem generous in a time of budgetary pain. Although little of the money represents new budget authority, in 1988 the U.S. Government will probably spend more on HTS alone than Japans Government will spend on HTS and LTS together. The 1988 total approaches the ILow-temperature superconductivity (LTS) has shared in the expansion. For years, DOE and DoD have funded LTS projects such as energy storage and superconducting machinery (e.g., for ship propulsionAppendix B). Federal spending for LTS increased from $40 million in fiscal 1987 to $84 million in 1988. Agency requests for LTS in the 1989 budget come to about $83 million. (Both the 1988 and 1989 figures include Strategic Defense Initiative (SDI) contract work on superconducting magnetic energy storage.) recommendation$100 millionof a National Academy of Sciences (NAS) panel. 2 But the totals do not tell the whole story. HTS could remain in the laboratory for many years. During much of this time, the Federal Government will remain a primary source of R&D funds. Effective support for commercialization will require stability in Federal funding, attention to priorities, and good management of agency budgets. In their fiscal 1988 budgets for HTS, some agencies fared much better than others. DOE and NSF spent roughly equal amounts on HTS in 1987; the Energy Department will have more than twice as much this year, while NSFs in~Research Briefing on High-Temperature Superconductivity, Committee on Science, Engineering, and Public Policy, National Academy of Sciences, Washington, DC, 1987, p. 19. The panel, noting that corporate funding might add a comparable amount, termed this a good beginning in addressing the challenges and opportunities offered by the new materials.
PAGE 93
87 Table 8.Federal Funding for HTS R&D Fiscal year budget (millions of dollars) 1987 1988 1989 (actual) (estimated) (requested) Department of Defense a . . . . . . . . . $19. 0 $46. 0 $63. 0 Department of Energy . . . . . . . . . . 12.5 27.2 38.7 National Science Foundation . . . . . . . . 11.7 14.5 17.2 National Bureau of Standards . . . . . . . . 2.8 b 9.3 National Aeronautics and Space Administration . . . . 4.2 6.7 Bureau of Mines . . . . . . . . . . . 0.1 0.1 0.1 $44. 9 $94. 8 $135. 0 awOrking figures, subject to change. bExcludes$lsO,Ooo correlated work. SOURCE: Preliminary agency data and budget estimates provided to the Subcommittee on Superconductivity of the Committee on Materials, May 1988 crease is only 25 percent. NSF officials have said they have received many more highly-rated proposals on HTS than they can support in fiscal 1988. Meanwhile, the DOE laboratories which typically get nearly two-thirds of the Departments basic research fundsmay have more money for HTS than NSF will provide the Nations universities. The NAS panel emphasized the need for new money for HTS to avoid cuts in other, perhaps comparably important, R&D. When the excitement over HTS reached a peak early in 1987, the fiscal year was well underway. Thus almost all the 1987 funding came through redirecting of dollars originally allocated to other research. In many cases, scientists and engineers with Federal contracts and grants took the lead in this process, seeking approval from agency contract monitors to move into HTS. Faced with little growth in R&D budgets, most agencies have had little choice but to continue pulling money from other fields to pay for HTS. The National Bureau of Standards (NBS)with a budget that grew 16 percent from 1987 to 1988 is probably alone in being able to fund its HTS work without sacrifices elsewhere. In a period of tight budgets, when there may be no way to avoid sacrificing one kind of research to pay for another, good decisions on priorities within and across agencies become more important than ever. Doing a better job of formulating R&D budgets could help identify conflicts earlier, and perhaps ease their resolution. For HTS, stability over time will be as important as next years R&D totals. a At present, most of the funds for HTS come from the general R&D authorizations of the agencies. Rather than this piecemeal approach, Congress could take a broader look at the Federal effort in HTS, and provide overall guidance, through such mechanisms as a single piece of legislation that would provide multiyear authorizations of appropriations, defining the responsibilities in HTS for each agency. This approach is discussed in more detail in table 9 (Option 1). It carries dangers: for example, possible micromanagement by Congress. On the ether hand, if implemented in too weak a form, the effort could end up as little more than a paper exercise, with little or no influence on the actual allocation of HTS R&D support across the agencies. As a further step, Congress could direct the Administration to prepare a multi-year estimate of funding expectations for HTS R&D (see Option 2 in table 9). Some of the proposals on HTS before the 100th Congresse.g., H.R. 3217, as sIn a well publicized episode, a recent NSF effort to reduce uncertainty in university research programs backfired. Managers in the Foundations Materials Research Division, expecting a substantial funding increase in fiscal 1988, made too many longterm commitments during 1987, When the Federal budget was finally approved, and the money was not there, NSF cut back on ongoing multi-year grants (which are conditional on availability of funds) in order to support some new starts. See Statement on Funding Levels for the Division of Materials Research, National Science Foundation, Mar. 3, 1988.
PAGE 94
88 Table 9.issue Area 1: Funding Levels and Priorities for Federal R&D Issue Options for Congress Advantages Disadvantages A. Funding Levels for HTS On the surface, Federal funding for HTS R&D seems generous$95 million for fiscal 1988. The difficulties lie beneath the surface: l Little of the total is new money. Few agencies got the increases in their R&D budgets they had planned on for fiscal 1988. They have taken money for HTS from other research. Universities have had difficulty in lining up funds. Ten DOE laboratories may well get more for HTS during fiscal 1988 than NSF will have for all the Nations universities. The Administration is requesting a hefty increase for HTSto $135 million in 1989and is calling for a substantial rise in non-defense R&D. If Congress pares back the R&D budget to accommodate other needs, the new money issue, along with allocations of R&D funds among the agencies, could be central issues, not just for HTS, but for R&D generally. B. Continuity of Funding HTS could easily require a decade or more of steady R&D support before a technology base adequate to support commercialization emerges, with a continuing need for Congress and the executive branch to assess funding levels, as well as allocations across agenciese.g., support for processing R&D, and whether it is adequate to support commercialization. Stop-and-go funding has been a common problem for U.S. science and technology policyand a serious onein part because of year-by-year budgeting for Federal R&D. A period without newsworthy research results could lead to a dry spell in HTS R&D budgets. C. National Science Foundation Budget Despite the Administrations announced objective of a doubling in the NSF budget between 1988 and 1992, the Foundations fiscal 1988 appropriation grew by only 6 percent (compared with a request of 17 percent). NSF has had to postpone increases in funding for multidisciplinary R&D centers and for research in engineering, traditionally underfunded. Laboratory equipment in many American universities is inadequate for either research or teaching. OPTION 1. Provide a legislative framework defining the overall Federal commitment to HTSfor example, a single bill providing specific multi-year authorizations for HTS R&D by agency. The authorizations would signal the congressional appropriations and budgeting committees, as well as the agencies, concerning the relative shares of funds for HTS R&D to be given to each agency. OPTION 2. Direct the Administration to prepare a multi-year estimate of Federal funding expectations for HTS R&D. This might be a rolling 5to 10-year plan, directed at commercial (rather than military) applications, and intended to be revised periodically (not a rigid, inflexible set of research targets). Private sector input could be built into the process. OPTION 3. Direct the Administration to experiment with a 2-year funding cycle for HTSpossibly beginning with a pilot program at NSF. (Section 201 of Public Law 100-119 encourages congressional committees to experiment with multi-year authorizations and 2-year appropriations. ) OPTION 4. Consider substantial increases in the NSF budget over the next few years. Budget increases along the lines of President Reagans proposal for a doubling of the Foundations budget over 5 years would permit NSF to double or triple its funding for engineering research to the $400 million to $500 million levelwithout sacrifices elsewhere. OPTION 5. Appropriate substantially more money to NSFan added $100 million or more per yearfor equipment grants to the Nations universities for both research and teaching. A single framework for funding decisions could help keep Congress aware of potential imbalances among the R&D agencies. Multi-year authorizations, along with the multi-year planning exercise discussed in Option 2, and the experiment in multi-year funding discussed in Option 3, could help make the point to universities, the laboratories, and to industry that Congress intends to sustain the Governments commitment to HTS over time. As a mechanism for helping policymakers gain perspective on annual budget proposals, multi-year estimates should be useful both to Congress and the agencies. The effort could improve agency coordination, limit overlap in R&D funding, improve the quality of scientific and technical advice to Federal agencies, and reduce the likelihood that money for superconductivity will come at the expense of other needed R&D. If successful for HTS, the approach might become a model for other fields. Uncertainty over funding for HTS during 1987 and early 1988, and particularly over the prospects for new money, made it hard for research groups in government, universities, and industry to plan, and delayed some projects. Such problems cannot be totally avoided in a fastmoving field like HTS. But a 2-year budget cycle would help keep R&Don a steady course. More money for engineering would be a major step, not only in commercializing HTS, but in supporting U.S. industrial competitiveness across the board. NSF will spend $171 million on engineering research in fiscal 1988, only 10 percent of the agencys research budget. Gifts from the private sector can help, but the problem is far too big to be solved in this way alone. Government action would help improve the Nations technological capabilities, Congressional guidance could turn into micromanagement of Federal R&D, or pork-barreling. Without proper oversight from upper levels in the Administration, such an exercise could turn into an agency wish list, with little utility for making tough budget decisions. Moreover, any effort to develop a government-wide perspective would probably be seen by some as top-down planningthreatening agency autonomy and flexibility. Multi-year budget estimates, finally, would probably have limited utility unless the agencies supported the conceptwhich few do now. In the absence of improvements in mechanisms for establishing R&D priorities, a 2-year budget cycle would do little to overcome the fundamental budgeting problems posed by competition for limited funds, To some extent, a 2-year cycle might reduce the flexibility of the system, with potentially serious consequences in periods of rapid technological advance. Given the size of the Federal budget deficit, a significant increase for one agency could well come at the expense of others. The increases in civilian R&D included in the Presidents fiscal 1989 budget request$300 million for NSF, $400 million for DOE, $2.5 billion for NASAcannot be accommodated within the framework agreement worked out between Congress and the Administration in late 1987 unless Congress adjusts other budget items downward. Unless accompanied by an overall increase in NSFs budget (see Option 4), more funds for equipment could cut into the Foundations research budget.
PAGE 95
89 Table 9.Issue Area 1: Funding Levels and Priorities for Federal R&DContinued Issue Options for Congress Advantages Disadvantages D. Weaknesses in the Industrial Technology Base Despite the size of the U.S. R&D budget, gaps open in the technology base where neither industry nor government provide support. Prior OTA assessments have pointed to some of the problems; many more certainly exist. The first step toward a solution is to characterize the weaknesses more fully American companies conduct relatively little basic research. Under the Tax Reform Act of 1986 (Public Law 99-514) companies get a more favorable tax credit for basic research they fund in universities than for work performed internally. Both the general R&D tax credit and the basic tax credit for work sponsored at universities expire at the end of 1988. E. Setting Priorities for Federal R&D Competition for Federal R&D dollars seems bound to grow more intense, with conflicting demands between big science and small, defense and civilian R&D, and basic research and more applied work. Establishing priorities and sticking to theme g weighing the pros and cons of expenditures such as required for a Superconducting Super Collider, or the National Aerospace Planerequires a government-wide perspective. This is the job, in principle, of the Office of Science and Technology Policy (in the Executive Office of the President). OPTION 6. Request a detailed review of Given the budget deficit, it is more imStudying the problem without taking the U.S. technology base by the National portant than ever that R&D decisions be steps to solve it would accomplish little, Academies of Science and Engineering. based on sound analysis, Less glamSuch a review might encompass: orous. less visible fields tend to suffer l l funding levels for both basic and apmost in such periods, with harmful implied research across the broad range pacts that show up only in later years, of scientific and technical disciplines when the damage has been done. important for industrial competitiveness, with particular attention to actual and potential bottlenecks and to technical fields (like manufacturing) that historically have been underfunded; processes for setting research priorities and determining funding levels within and across Federal agencies. OPTION 7. Permit a separate tax credit for basic research conducted within the firm. To have much impact, an in-house research credit would have to be as favorable as current rules applying to basic research paid for by industry but conducted at universities, and more favorable than tax credits for internal R&D under the 1986 tax act, A basic research credit could supplement the overall R&D tax credit if Congress decides to make it permanent for 1989 and beyond. If Congress lets the existing credit expire, a special provision might be crafted-perhaps on a trial basisfor basic research within industry. OPTION 8. Give the Office of Science and Technology Policy access to the staff resources and advisory processes needed, not only to monitor science and technology issues in the agencies, but to assume an effective decisionmaking role within the executive branch. A basic research credit for work within the firm would create stronger incentives for attacking technical problems that fail to excite much interest in universities. A strengthened OSTP would permit the Executive Office of the President to develop and articulate priorities for science and technologybacked up with analytical depth and detail that have not been possible, given the Offices current staff (about 30) and budget (about $1.9 million). Creating new tax credits runs counter to the spirit of tax reform, while enforceable guidelines for basic research could be difficult to define, OSTP will have little influence unless the President wants it to, Lacking this, congressional action to strengthen the Office would make little difference, SOURCE: Office of Technology Assessment, 1988. introducedwould direct the executive branch to provide, on a one-time basis, a Federal program plan for superconductivity, including estimated funding levels by agency for a five-year period. H.R. 3217 would assign the overall responsibility to the Executive Office of the President, with roles for the Office of Science and Technology Policy and the National Critical Materials Council. It provides for consultation with the mission agencies, as well as universities and industry. The proposal would also create a more formal structure for coordination among agencies, (Box M discusses interagency coordination of HTS R& D.) Any effort to develop Government-wide estimates risks being seen as top-down planningthreatenin g agency autonomy, professionalism, and flexibility. Nonetheless, viewed as a mechanism for helping policymakers gain perspective on annual R&D budget proposals, multi-year estimates could be useful both to Congress and to the agencies. 84-754 0 88 4 : QL 3
PAGE 97
91 With Congress appropriating money annually for research programs that may go on for years, the ups and downs in R&D funding have also stimulated frequent proposals for multiyear authorizations and/or appropriations. 4 Al though Congress has been reluctant to move in this direction, growing concern over the budget process as a whole has led to discussion of a two-year budget cycle. As a more modest step toward a longer-term perspective on R&D decisions, Congress could experiment with multiyear funding in a single agencyperhaps NSF (Option 3). The experiment might be undertaken by programs in, say, the engineering directorate or the materials research division both of which support HTS. Neglect by Government and industry of commercial R&D has slowed the passage of technology from laboratory to marketplace, harming U.S. productivity and competitiveness. Less glamorous fields, particularly in engineering, seldom attract funding commensurate with their potential economic significance. Chapter 2 stressed U.S. underinvestment in processing R&D; other examples include materials synthe4 For discussion of some of the possible mechanisms, see U.S. Science and Engineering Base: A Synthesis of Concerns about Budget and Policy Development, GAO/Reed-87-65 (Washington, DC: U.S. General Accounting Office, March 1987), pp. 22-34. sis (box C, ch. 2). 5 For such reasons, and despite the huge U.S. investment in R&D, the technology base no longer seems adequate to support a competitive set of industries. In government, lack of mechanisms for setting priorities, coupled with stop-and-go funding for some kinds of R&D, have contributed to the problems. Gaps and holes in the technology base emerge particularly in fields that Federal agencies DoD, DOE, NASA (the Na*On the lack of R&D in construction technologies, see International Competition in Services (Office of Technology Assessment, July 1987), pp. 138-144. Other examples include: l l l l l l Direct reduction of iron to steel. Railway technology. (Given the importances of rail transportation for the Nations economy, support has been woefully inadequate compared to, say, aeronautical engineering.) Process control models for the fabrication of microelectronic devices. Theoretical foundations for software engineering. (Better understanding could lead to greater productivity in programming, helping break a major bottleneck in U.S. industry.) Fundamental understanding of combustion processes. (Environmental pollution from stationary powerplants, burning of solid wastes, and automotive engines costs the United States billions of dollars each year. Lack of a research base in combustionin terms of thermodynamics, chemical kinetics, fluid mechanics, heat transfermakes it difficult to develop inherently clean combustion processes.) Corrosion and wear. (These processes, so familiar and pervasive as to seem inevitable, have economic costs measured in billions of dollars annually; wear, in particular, has never attracted much scientific attention or research support.) Also see Directions in Engineering Research: An Assessment of opportunities and Needs (Washington, DC: National Academy Press, 1987].
PAGE 98
. 92 Photo credit: Argonne National Laboratory HTS superconducting wire, ready for testing. tional Aeronautics and Space Administration)view as too far from their missions, and that, in the view of corporate managers, will not yield financial returns in the short or medium term. Among the other causes: relatively low levels of support for engineering research, and Federal R&D programs that have often gone astray when not tightly linked to agency missions Congress could begin to enlarge the pool of commercially-relevant technology by appropriating additional funds to NSF, allowing the Foundation to expand its support for engineering research without taking money from other areas (Option 4). NSFs mission embraces the strengthening of the Nations science and engineering base; yet its current spending on engineering research ($171 million) does not amount to three-tenths of a percent of the overall Federal R&D budget. Congress might also provide additional money to NSF specifically for laboratory equipment. Equipment in the Nations engineering schools averages 20-30 years old; a quarter of it cannot even be used. 6 An additional $100 million annually, to supplement NSFs current spending of $250 million a yearwould help (Option 5). NSF ranks no better than fifth in R&D spending among Federal agencies. Any search for a broad solution to the problems in commercial technology will have to look beyond NSF and the university research it sponsors. Given the pressures on the Federal budget, a realistic first step might be to identify the weaknesses in the existing technology base, and begin establishing priorities for allocating the limited funds available. Congress could ask the National Academies of Sciences and Engineering to begin this task (Option 6). As a complementary measure, aimed at encouraging American firms to undertake more fundamental research, Congress might consider changes to the Research and Experimentation Tax Credit. 7 At present, industry finances only a fifth of all U.S. basic research. Federal agencieswhich pay for two-thirds (universities fund the remainder)do not set priorities based on commercial relevance. Giving companies greater incentives to conduct work inhouse would help focus basic research on industrial needs. Congress could institute a special basic research tax credit for work conducted within 8P. Doigan and M, Gilkeson, Engineering Faculty Demographics: ASEE Faculty & Graduate Student Survey, Part II, Engineering Education, January 1987, p, 212. The National Research Council suggests that an increase of $3o million or more for engineering equipment alone would be appropriate Directions in Engineering Research: An Assessment of Opportunities and Needs, op. cit., pp. 50-51. Also see Scientific Equipment for Undergraduates: Is It Adequate? staff paper, Science, Education, and Transportation Program, Office of Technology Assessment, Washington, DC, September 1986. Introduced in 1981, the credit was reduced from 25 percent of qualifying R&D expenditures to 20 percent in the 1986 Tax Reform Act. On its effectiveness, see International Competition in Services, op. cit., p. 364. Current law allows companies more favorable tax treatment for support of basic research at universities or other qualified R&D organizations than for work carried out at their own facilities,
PAGE 99
93 industry (Option 7). Assuming that Congress extends the existing R&D tax credit, now set to expire at the end of 1988, or makes it permanent, basic research conducted internally could be given more favorable treatment than other qualifying R&D. Finally, Congress could ask the Academies for recommendations on an R&D strategy aimed specifically at strengthening the Nations commercial technology base (as noted in Option 6). Such an exercise might help OSTP carry out its policy and planning functionsincluding legislative mandates that the office has had limited success in fulfilling. As discussed under Option 8 in table 9, OSTP may need strengthening if it is to be an effective arbitrator among agencies and interest groups seeking Federal R&D funds. In a period of intense competition for scarce dollars, a Government-wide perspective is needed more than ever in setting and enforcing priorities Defense-Related R&D Funding Patterns DoD has been supporting superconductivity R&D for more than three decades because of the potential applications in military systems. In this light, the dominance of DoD in Federal support for I-ITS (shown earlier in table 8 ) should be no surprise; much of the work is a natural follow-on to earlier sponsorship of LTS R&D. The three services, together with DARPA and the Strategic Defense Initiative Organization (SDIO), maintain their own programswith the DARPA and SDIO efforts the biggest by far (table 10). Three-fourths of DARPA funds, and a high proportion from SDIO, go to industry. DARPA states that as much as 60 percent of the processing R&D contracts currently in negotiation could go to firms that are not traditionally part of the defense industry. As for the services, about two-thirds of their HTS R&D funding is currently going to universities; if HTS follows the typical pattern for basic research in the services, this fraction may eventually decline somewhat (universities perform about half the 6.1 (basic) research paid for by the services, with government laboratories and industry sharing the remainder). DARPAs widely publicized processing initiative accounts for nearly all that agencys 1988 total of $18 million. With no R&D facilities of its own, DARPA will support processing-related work in industry, universities, and laboratories overseen by other agencies. The primary objective: speeding development of fabrication techniques for HTS coatings, thin and thick films, wires and other conductors. DARPA officials view the effort as a natural extension of the agencys ongoing program in manufacturing technology for advanced ceramics. After receiving about zoo proposals during the summer of 1987responses to a solicitation that assumed funding of up to $50 million for 1988 the agency announced in January that some 16 companies and 4 universities had been selected to enter into contract negotiations. When DoD placed a temporary freeze on some of its outside R&D (including DARPAs) in May 1988, nearly all of the contracts remained to be awarded. The freeze was in effect when this report went to press in June 1988. SDIOs HTS R&Dsecond to DARPAs in fundingfocuses on relatively near-term applications. The organization works closely with the services and other agencies, looking to technology insertion working groups for advice on where to direct its R&D dollars. Like other parts of DoD, SDIO contracts extensively with industry. In addition to HTS, the organization funds considerable work on LTSfor instance, a design competition on magnetic energy storage for powering large lasers, budgeted at $11 million currently and $13 million for fiscal 1989. R&D sponsored by the services reflects their missions. Much of the Air Force effort goes toward possible applications in electronics, funded (principally through the Air Force Office of Scientific Research) in universities and the Air Forces own laboratories. The Office of Naval Research is likewise putting most of its current HTS money into basic research (6.1). While the Army also has a program underway, the level is low (as expected, given that the
PAGE 100
94 Table 10.-Department of Defense Funding for HTS R&D a Fiscal year budget (millions of dollars) 1987 1988 1989 (actual) (estimated) (requested) Army . . . . . . . . . . . . . . . $1. 0 $2. 0 $3. 0 Navy . . . . . . . . . . . . . . . 5.0 7.0 9.0 Air Force . . . . . . . . . . . . . . 4.0 7.0 8.0 Defense Advanced Research Projects Agency (DARPA) . . . . 4.0 b 18.0 20.0 Strategic Defense Initiative Organization (SDIO) . . . . . 5.0 12.0 23.0 $19.0 $46.0 $63.0 aworking figures, subject to change. DoD also spends substantial Sumson low-temperature superconductivity. Dlncludes$2 minion from the Balanced Technology lnitiativO. SOURCE: U.S. Department of Defense, April 1988. Army traditionally funds relatively little R&D compared to the other two services). Each of the services has formed internal working groups to coordinate its effort. lmplications for HTS With DoD paying for nearly half the governments HTS R&D, the obvious question follows: What does this mean for commercial development, and for the civilian side of the economy? In the past, Federal dollars for both R&D and procurement provided much of the impetus for vibrant commercial industriesaircraft, computers, microelectronics. At the same time, as summarized in box M, DoDs very success in driving technology forward has led to a split between military and civilian applications, with defense systems growing steadily more specialized. Some would claim that military spending has undermined U.S. industrydistorting the technological enterprise by diverting the best and brightest engineers and scientists from civilian industries, skewing university research (and, through the research interests of faculty, university curricula), and turning companies aside from the costdriven discipline of the marketplace. In this view, rather than providing fertile ground for spinoffs, DoD support for HTS might divert resources from commercialization Indeed, there seems little reason to expect that spinoffs from DoD funding for HTS will have impacts as significant as those that spurred earlier high-technology industries. Since the 1950s and 1960s, technology transfer from the military to the civilian side of the economy has slowed, for reasons that include the expanding curtain of secrecy surrounding DoD and its contractors. With military systems growing steadily more esoteric, it would be unwise to rely on DoD support for HTS as a substitut e for civilian R&D. This does not mean that DoD R&D cannot be a valuable complement. Two broad questions will determine the effects of DoD spending on the commercia l prospects for HTS: 1) What are DoDs objectives with respect to HTS, and how do they compare with commercial needs? and 2) How much money will go to generic R&D, and thus offer potential for commercial spillover regardless of ultimate system requirements? In mid-1987, a DoD working group examined the R&D that would be needed to exploit HTS in military systems. The working group, in an options paper described as a map of the territory rather than a predetermined itinerary, concluded that an aggressive program to bring HTS to the point of military-specific applications would cost about $500 million over a 5year period. 8 The working groups options paper, which assumes that technology, not money, 8Superconductivity Research and Development Options: A Study of Possible Directions for Exploitation of Superconductivity in Military Applications, U.S. Department of Defense, July 1987. Summary figures for the 5-year program plan, totalling $506 million, appear on pp. 122 and 123. In the first 3 years (fiscal 1988 to 1990], the working group called for $293 million twice the $150 million DoD expenditure mentioned in the Presidents July 1987 superconductivit y initiative, and far more than defense agencies are likely to spend over this period, judging from preliminar y budget figures.
PAGE 101
95 would be the limiting factor, discusses R&D in several broad categories: l materials characterization, including efforts to find HTS compositions with higher transition temperatures ; l processing R&D ; l small and large scale applications and demonstrations While there are no signs that the 5-year spending plan will go forward as outlined in the working groups report, the budget estimates provide a baseline for considering DoDs view of prospects and priorities in HTS. Sixty percent of the 5-year total would go for applications $306 million. processingwhich holds mor e potential for commercially relevant R&D results would get $129 million, or 25 percent; the options paper allocates $71 million for materials characterization, equally generic. The breakdown by budget category paints a similar picture: basic research (6.1) accounts for 29 percent of the total, compared to 38 percent for exploratory development (6.2), and 33 percent for advanced development (6.3A). Viewed either way, basic research and generic R&D would get a substantial share of the resources, as befits a new technology. Most but not all of the applications work would be of interest primarily to the military. Examples include infrared sensors, detectors for submarines, and electromagnetic coil/rail guns. Some applications projects might generate commercial spinoffs: electronic devices for digital systems; motors, generators, and other electrical power equipment. (As discussed in app. B, these applications could, in principle, be implemented with LTS technology; indeed, even were HTS reduced to practice, LTS might provide superior performance.) Still, superconducting motors and generators for military applications, to take one example, will differ fundamentally from those for civilian applications. DoDs interest stems largely from the advantages that superconductin g motor-generator sets could have for ship propulsion and on board aircraft. Such propulsion systems would offer new freedom in packaging the major systems within a ships hull; for submarines, in particular, there would be more room for weapons. Compact design becomes a primary design criteria. For civilian power generation, in contrast, greater efficiency is the objective, with size (and weight) of little import. From a design standpoint, superconducting generators for the military and for electric utilities would have relatively little in common. Only in the most general sense would knowhow from one transfer to the other. Processing technology will be particularly important for HTS. Wire manufacture and fabrication received little emphasis in LTS R&D until becoming a bottleneck to applications. Years were then spent learning to produce niobiumtitanium wires and windings with the needed properties. A similar experience in HTS could put U.S. firms behind, given that processing is an area in which Japanese firms will undoubtedly excel. Here, DARPAs processing program should help. Many of the processing and fabrication methods ultimately developed will be similar regardless of end-application, and DoD officials have frequently stated that results will remain unclassified to the extent possible. (In part for such reasons, H.R. 3024, the proposed National Superconductor Manufacturing and processing Technology initiative, would give DARPA a lead role in the Federal Government for processing-related work. The 100th Congress had taken no action on this bill, which assigns subsidiary roles to DOE, NSF, and NBS, as OTAs report went to press.) DoD work aimed at high-performance computers, where applications will depend in part on thin-film fabrication capabilitiese.g., for Josephson junctionscould likewise have positive impacts on the civilian economy. Not only DARPA, but the National Security Agency has traditionally supported work aimed at highperformance computing (box N). If DoD were to follow a spending plan something like that outlined by the working group i.e., roughly half a billion dollars over five or six yearscivilian industry would surely benefit from some of the technology developed. Despite the stress on applicationsnoteworthy, given the relative pessimism of U.S. industry
PAGE 103
97 Finally, DoD also funded a considerable amount of visionary research-one of (D)ARPAs jobs. Here, the military mission did not always dictate R&D objectives, or even provide much guidance: (D)ARPA supported work in artificial intelligence and the behaviorial sciences inthe absence of nearterm military applications. Military and Civilian Technologies: Diverging Objectives During the Vietnam years, defense R&D growth slowed; DoD has never built its support for generic technology development back to pre-Mansfield Amendment levels. (The Mansfield Amendment, part of the military authorization bill for fiscal years 1970 and 1971, sought to tie DoD R&D more closely to defense needs.) Meanwhile, military high technology moved steadily away from civilian high technology. In the face of pressures from the Pentagon, DARPA too has turned toward projects for which it can more easily demonstrate military relevance, and steered a greater fraction of its funding to traditional defense contractors. 4 As computers, for example, proliferated on the civilian side of the economy, prices dropped and the government role as primary customer declined. Computer firms took more of the R&D burden on themselves, adapting their products to the needs of banks, insurance companies, and manufacturing firms. Even so, defense agencies have continued to support both basic research and high-risk, high-cost development projectswork that could have major impacts in the future; as noted in chapter 3 (box J), the National Security Agency provided partial support for IBMs research on superconducting computer components. Military demand also continues to provide substantial support for supercomputer manufacturers. 5 In semiconductors, the story is similar. Military procurements accounted for about half of all U.S. production in 1960. By the middle 1970s, the military had become no more than a minor customer for all except the most highly specialized chips; today, military sales run at less than 10 percent of the U.S. market, In 1979, the Pentagon found itself forced to create the VHSIC (Very High-Speed Integrated Circuit) program, an effort to take advantage of advances on the commercial side of the industry, where applications had long since outrun those in military systems. The Pentagon likewise provided much of the early R&D support for lasersin the early 1960s, twice the industrys own spendingand today continues to pay for most of the work on high-power lasers. 6 Military R&D, including fundamental research, has been conducted primarily in DoDs own laboratories, or those of its contractors, not at universities. As customers, the services have sought laser rangefinders for tanksthe first significant application on the defense side-and beam weapons. Civilian applications, meanwhile, began with eye surgery. Today, the growing divergence between military and commercial technologies is visible in at least three ways: CDARPA has weathmod a number of thaee cy cka-tolerarwe for vtiionary research followed by a turn bwktoward applications, engineering, and hardware. See Targeting the Cbnpter, op. cit., p. 190; &dao The Advanced Reseamh Pr@acts AS$IW% 1S88-1974, Richard J. Barber Associataa, Washington, IX, Deoemk 1975. W. Koameteky, %percomputere and IWional E%@: M@@nins U.S. preemi-ce in ~ Eme@%I 1-DS Su@rcomP~~ers: A KeY to U.S. Scientific, Tedmdogkxd, and &duut?&J Pmm&w@@, @ K&kknd and J,H. poo~ (ads.] (P&w Y@rk Praeger, 1987), P 10. @The Maturation of Laser Technology; Social and Technical Factors, prepared for OTA by J.L. Bromberfg, The Laser History Project, under contract No. H2-521O, January 1988; R.W. Seidel, From glow to flow: A history of military laser research and development,* op. cit. 7Advanced Materials by Design: JVew Structund MateriaZs Twhnologies (Washington, DC: Office of Technology Assessment, June 1988). l%ose on the commercial side of the aircraft industry envision an airplane that could fly halfivay around the globe in z hours, reaching speeds of Mach 5 (i.e., 5 times the speed of sound). DoD aces the NASP as a possible launch vehicle for SDL among other things, and the military version would have to reach Mach 25.
PAGE 104
98 the options paper calls for a lot of money in the gap will grow: DoD has requested $63 miltotal, and a hefty infusion of funds for the more lion for HTS in fiscal 1989, much less than the generic work. working groups recommendation. With funds tight, defense agencies normally preserve their But DoD will almost certainly not have this applications programs as best they can; they much money for HTS, as table 10 indicates. The will have to continue with materials characterifiscal 1988 total--$46 million--is well under the zation and processing to support downstream $68 million called for in the options paper, and development in HTS developments, but the
PAGE 105
99 temptation will be to go no further into basic work than absolutely necessary 9 The final point is this. R&D management in any mission agency entails a continuous series of large and small decisions. These deal with such matters as funding levels and priorities, research targets, intramural versus extramural projects, contract and program managers constantly weigh alternatives for expenditures ranging from a few thousand dollars to many millions. Broad objectives are set at upper levels; people lower down make their choices guided by these objectives (though often with considerable autonomy). But at all levels, Do D decisionmakers from program managers to laboratory directors and the Under Secretary for Acquisition (who has overall responsibility for DoD R&D)have their eyes on military needs, not those of the civilian economy. This is their job. Directives from outside the Pentagon may influence these day-to-day decisions, but not by much. HTS R&D in the Energy Department Laboratories DOE laboratories have actively sought major roles in HTS, typically for reasons including diversification beyond their primary missions. As table 8 indicated, DOEs budget for HTS R&D exceeds that of NASA, NSF, and NBS combined; table 11 gives the allocation within the Department. If usual patterns prevail, twothirds or more of DOEs basic research dollars will be shared among DOEs nine multiprogram laboratories (the National laboratories) and a number of more specialized research facilities, The Energy Department and its predecessors have been the patron of big science in the Federal Government since the days of the Manhattan Project. While the Federal Governmen t owns the DOE laboratories, most are operated under contractsome by universities, some by private corporations 10 The laboratories hav e a collective budget well into the billions, and employ about 15,000 scientists and engineers, Several have strong foundations in superconductivity, stemming from years of work on LTS magnets for high-energy physics and fusion research, along with projects such as Brookhavens 10-year effort on superconducting power transmission. By one estimate, DOE has spent $100 million on LTS R&D over the last two decades, in addition to $200 million for purchases of materials and equipment. Given this history, and the Departments responsibilities for energy R&D, it is no surprise that the laboratories have garnered the majority of non-DoD Federal dollars for HTS. A number of the laboratories have excellent equipment for synthesizing and characterizing the new HTS materials. They have physicists, chemists, and engineers with the skills and experience to contribute to the science and technology base for HTS. But while many of these laboratories produce excellent science (as well as mission-oriented weapons development) they have little experience in helping industry The technical objectives of DoD 6.1 basic research are commonly shaped to considerable extent by military needs. DoDs own options paper notes: while DOD will surely benefit significantly from efforts of other organizations (DOE, NSF, DoC, NASA) in areas of materials characterization, theory, and search for high-transition-temperature materials, it is essential that DSRD [Defense Superconductivity Research and Development] itself include substantive activity in these areas. Much of the remainder of DSRD activity is so highly applications driven that DSRD characterization, theory, and search activities are essential as a means to provide focus in directions of greatest perceived impact on DoD applications. weight considerations are paramount in many DoD applications (as in those of NASA), and DoD has other stressing requirements related to mechanical and thermal shock, as well as to radiation hardness, all of which dictate that DoD-specific characterization investigations be pursued. Superconductivity Research and Development Options: A Study of Possible Directions for Exploitation of Superconductivity in Military Applications, op. cit., pp. 19-20. IOEight multiprogram national laboratories have gotten most of the DOE funds for HTS. These laboratories. and contractors as of 1988, are: Laboratory Argonne Brookhaven Lawrence Berkeley Oak Ridge Pacific Northwest Lawrence Livermore Los Alamos Sandia Contractor University of Chicago Associated Universities, Inc. University of California Martin Marietta Energy Systems, Inc. Battelle Memorial Institute University of California University of California Sandia Corp. (a subsidiary of AT&T Technologies) Livermore, Los Alamos, and Sandia are weapons laboratories. Single-program DOE laboratories active in HTS include Ames Laboratory [operated by Iowa State University) and the Solar Energy Research Institute (operated by the Midwest Research Institute).
PAGE 106
100 Table 11. Energy Department Funding for HTS R&D a Fiscal year budget (millions of dollars) 1987 1988 1989 (actual) (estimated) (requested) Office of Energy Research Basic Energy Sciences. . . . . . . . . . . High Energy&Nuclear Physics . . . . . . . . Defense Programs. . . . . . . . . . . . Office of Conservation & Renewable Energy Energy Storage & Distribution . . . . . . . . Energy Utilization Research . . . . . . . . . Office of Fossil Energy Advanced Research & Technology Development. . . . . . . . . . $10.2 $15.1 $16.7 0.2 0.2 0.3 1.6 6.7 6.7 0.2 4.4 12.9 0.1 0.4 2.0 0.2 0.3 0.2 $12.5 $27,2 $38.7 aExCluding the De~artmentls Small Business Innovation Research Program. DOEcurrently spends more on LTSR&Dthanon HTS$28.5miIIionon LTSinfiscal 1987, S39.5million in 1988. lnfiscal 1989, the Department isseeking $52 million for LTSR&D (afigure that excludes $34 million forprocurement of materlalsandcomponents) NOTE: Totals may not add because of rounding. SOURCE: U.S. Department of Energy, 1988.
PAGE 107
101 commercialize new technologies. DOE plans to require the laboratories to involve industry and the universities in their HTS work to a greater extent than usual; for DOEs R&D to have impacts on commercialization commensurate with the Departments budget allocations, these efforts will have to succeed. In 1988, more than half of DOEs HTS budget$15 million of $27 million totalwill be channeled through the Basic Energy Sciences program (B ES, table 11). While some BES funds go to universities and to industry, most of the programs HTS work during 1987 was undertaken within the laboratory systema pattern that will probably continue. 11 BES has established two joint programs in HTS, each involving three laboratories. Under an arrangement worked out in 1987, Argonne, Ames, and Brookhaven will concentrate on processing R&D for bulk materials, while Oak Ridge, Los Alamos, and Lawrence Berkeley will work primarily on materials synthesis, thin films, and electronic devices. The Administrations 1989 budget request would give BES a 10 percent increase for HTS. Another DOE officeconservation and renewable energywill spend nearly $5 million in fiscal 1988 for R&D related to possible electric power applications. Initial activities included a number of feasibility studies, including a jointly funded effort with the Electric Power Research Institute examining possible end uses. If the presidents 1989 budget is adopted, conservation and renewable energy could find its HTS budget tripling. Most of this would go to the offices energy storage and distribution group. In April 1988, DOE announced that it would provide relatively small sums to 10 DOE laboratories (eight of the multiprogram facilities, Ames Laboratory and the Solar Energy Research Institute) for work related to electric IiThe Division of Materials Sciences, which controls most of the money for HTS within BES, spent 63 percent of research funds totaling $155 million within DOEs own laboratories during fiscal 1987. About 35 percent went to universities (including support for graduate student research at national laboratories), and 1.8 percent to industry. See Materials Sciences Programs: Fiscal Year 1987, DOE/ER-0348 (Springfield, VA: National Technical Information Service, September 1987), p. F-3. These figures do not include $15.5 million in equipment funds. energy storage and distribution. Future funding under this program will depend in part on the ability of the laboratories to involve industry and universities. As table 11 shows, the only other DOE program with significant funding for HTS engages in defense R&D. Most of this workbudgeted at $6.7 million for 1988, with next years request at about the same leveltakes place at the three weapons facilities. The sections of this chapter dealing with technology transfer consider DOEs prospective contributions to commercialization of HTS for instance, the likelihood of productive collaborative efforts between the Departments laboratories and private industry. If cooperative arrangements and rapid technology transfers to industry are to flourish, the laboratories will have to change in style and culture. Table 13, later in the chapter, includes a number of specific policy options for accelerating this shift. Other Mission Agencies: NBS and NASA For more than three decades, the National Bureau of Standards, part of the Commerce Department, has been engaged in research on LTS materials. President Reagans superconductivity initiative gave NBS the responsibility for establishing a superconductivity center focusing on electronic applications. While NBSs technical achievements have been impressive e.g., a precision voltage standard incorporating 19,000 Josephson junctionsthe Bureau is small compared to many other Federal laboratories, and superconductivity a minor part of its work. The NBS appropriation for 1988 included $2.8 million for HTS projects (table 8) on measurement methods, standard reference materials, and devices for measuring weak magnetic fields. The Administration seeks a major increase for NBSto $9.3 millionfor fiscal 1989. The National Aeronautics and Space Administrations HTS R&D will aim at eventual applications such as remote sensing, power and
PAGE 108
102 propulsion, and space communications. 12 In space, simple passive cooling systems could keep the new materials below their transition temperatures. As a result, HTS holds considerable interest for NASA. At the same time, space missions demand very high reliability, thus painstaking development and testing; deployment on an actual mission is probably many years in the future. In some contrast to the other major R&D agencies, NASA has not rushed into HTS; the agencys R&D is still in the planning stages. NASA reprogrammed some $4.2 million for HTS during fiscal 1988mostly for feasibility studies (table 8), and is seeking twice as much for 1989. The preliminary program plan cited above calls for spending $48 million on HTS over the period 1988-94. Even at this level, however, it seems unlikely that NASA R&D would have much impact on commercialization of IZNASA Technology Program Plan: High Temperature f%perHTS: mission requirements are apt to be too conductivity Technology, Preliminary Program Plan, Vol. Z, Naspecialized. tional Aeronautics and Space Administration, Feb. 3, 1988. NSF AND THE UNIVERSITY ROLE The National Science Foundation is a mission agency too, but its responsibilities differ greatly from those of DoD, NASA, or DOE. The NSF mission: to support research because this is in the public interest (for reasons including economic growth and competitiveness). Almost all NSFs research dollars go to the university system, which the United States depends on far more than other industrialized economies. NSF expects to spend $14.5 million in fiscal 1988 on HTStable 12. With only a few U.S. companies putting much effort into basic research, many of the preliminaries to commercialization of HTS will take place on the Nations campuses. Are the universities up to the job? In the short run, the answer is plainly yes. But the work of commercialization will go on for years, and as it shifts from research toward applications, a set of perennial problems in engineering research, and in university/industry relationships, could hinder the process. These problems stem from the inhospitality of universities to multidisciplinary research, and the differing goals of university and industry R&D. Disciplinary Boundaries Many of the Nations universities have strong if often small HTS research efforts. As HTS technology moves ahead, multidisciplinary R&D will be essential. Progress will depend on the physics communitye.g., for theoretical guidance and an understanding of the ways in which structure, particularly at the atomic level (crystallography, flux pinning sites) determines properties (critical current densities). Chemists will add their skills, particularly in materials synthesis and characterization, as well as in processing. Materials scientists will have the job of understanding microstructural and substructural effects (grain boundaries, twins, dislocation structures), and of linking these with processing (e.g., thermal-mechanical sequences). Materials engineers will develop processing techniques that yield the needed structures (hence properties) at reasonable costs. Design of electronic devices will fall mostly to electrical engineers and physicists. Electrical and mechanical engineers will develop high-power/high-field applicationse.g., for energy storage systems. Each group has its own language, its own assumptions and preconceptions, its own world view. To the lay person, science and technology may seem all of a piece. They are not. In private firms, multidisciplinary groups function effectively because they mustotherwise the company would not be able to compete. Over the past decade, American companies have worked hard at this, as they have faced up to the loss of technological advantages in world competition. Firms like IBM and AT&Tleaders in HTS R&Dhave been seeking better ways
PAGE 109
103 Table 12.National Science Foundation Funding for HTS R&D Fiscal year budget (millions of dollars) 1987 1988 1989 (actual) (estimated) (requested) Directorate for Mathematical and Physical Sciences: Materials Research $8. 0 $10.0 $12.0 Chemistry . . . . . . . . . . . . . 0.3 0.4 0.6 Physics. . . . . . . . . . . . . . . 2.2 2.3 2.4 Engineering Directorate: . . . . . . . . . . . 1.3 1.9 2.2 $11,7 $14.5 $17.2 NOTE: Totals may not add because of rounding SOURCE: National Science Foundation, 1988. of moving new technology from the research laboratory, through development, and into production. The steady advance of technical knowledgewhich inevitably entails greater specialization and fragmentationonly makes this more difficult. The job is one for management, and a continual struggle. Universities find it even more difficult to accommodate such work, lacking the imperatives of the corporation. Specialization and fragmentation begin on campus. Indeed, disciplinary boundaries account for some of the technology gaps noted earlier in this chapter. No one undertakes needed R&D because no group of engineers or scientists looks on the problems as part of its territory (welding, wear, ceramic processing). HTS will probably face some of these kinds of problems. NSF Centers Federal agencies have tried to encourage interdisciplinary research in the universities, using the carrot of R&D money, but funds for programs like NSFs Engineering Research Centers (ERCs) remain small compared to those for single-investigator projects. Figure 4 shows the trends over three decades at NSF. Individual project support remains at about 70 percent of the NSF totalwell above the level of the mid-1960s. 13 Still, NSF-sponsored research IsAbout 13 percent of NSFs fiscal 1987 budget went for multidisciplinary research centersDepartment of Housing and Urban DevelopmentIndependent Agencies Appropriations for 1988, Part 4, hearings, Subcommittee on HUD-Independent Agencies, Committee on Appropriations, U.S. House of Representatives (Washington, DC: U.S. Government Printing Office, 1987), p. 74. centers could number 80 or more by the middle 1990s, if the Foundation gets the budget increases it has been seeking. Currently, about one-fifth of the NSF engineering budget goes for the ERCs, the first of which were established in 1985. In the Foundations 1988 spending plan, the ERCs account for $33 million ($15 million less than NSF originally sought) of the $171 million allocated to engineering. 14 The ERCs are relatively small and focused e.g., on Optoelectronic Computing (University of Colorado). Annual funding levels have ranged from $1.5 million to $3.5 million. While NSF expects many proposals for HTS centers in the future, superconductivity does not fall within the purview of any of the 14 ERCs approved through the end of 1987. Indeed, this group of 14 includes only one center in the area of materials (and it is scheduled to lose its NSF support)--perhaps because the Foundation also funds about a dozen interdisciplinary MaterilqIn th e first z years of the program, the Foundation approved 11 ERCs (expending $27.7 million, with industry, States, and localities more than doubling the NSF contribution). Current plans call for up to 18 ERCs by the end of 1989. Under the program, NSF agrees to support centers for up to 11 years, with evaluations after 3 and 6 years. The Foundation recently announced it will discontinue support for two of the initial centers, following their 3-year reviews. For further background, see The New Engineering Research Centers: Purposes, Goals, and Expectations (Washington, DC: National Academy Press, 1986), and Educating Scientists and Engineers: Grade School to Grad School (Washington, DC: Office of Technology Assessment, June 1988), ch. 3. On NSFs proposed S&T centers, below, see Science and Technology Centers: Principles and Guidelines (Washington, DC: National Academy of Sciences, 1987); also C. Norman, NSF Centers: Yes, But Science, July 3, 1987, p. 21.
PAGE 110
104 Figure 4.National Science Foundation Research Support 100 90 80 70 30 20 10 0 1960 1965 1970 1975 SOURCE: National Science Foundation, 1988 als Research Laboratories (MRLs) under a separate program (see box O). As discussed in box O, ARPA (later DARPA) which, over the years, has financed a good deal of work in superconductivityoriginally sponsored the MRLs. Five of the MRLs have moved into HTS research, with $3.5 million of NSFs 1988 support for HTS going toward these activities. The MRLs represent an early attempt by the Federal Government to change the ground rules for university research; the ERCs, along with NSFs proposed Science and Technology (S&T) centers represent the latest. Announced by President Reagan in his 1987 State of the Union Message, the S&T centers could eventually become the largest NSF program for interdisciplinary research support. Universities submitted more than 300 proposals after this program was announced (plus a comparable number of planning proposals), a third of them in the general area of materials (and some of these on superconductivity). Given the slow growth in its budget, discussed above (table 9, Option 4), the 1980 1985 1990 Foundation has not yet found money for the S&T centers. In February, the Administration announced that none would be funded during the 1988 fiscal year. Instead, the Administration will seek a one-time appropriation of $150 million in fiscal 1989 to fund 10 to 15 S&T centers for 5-year periods. If Congress provides the money for these centers, it is possible that one or two of those approved by NSF might have a focus on superconductivity. Funding for the Industry/University Cooperative Research Center programwell on the way to proving its worthhas been flat in recent years. Nor has the ERC budget grown as NSF had hoped. As discussed under Option 9 in table 13, additional funds will be needed to expand the center programs. Growth in these programs will not have much impact on HTS unless one or more of the proposals that would focus on superconductivity wins the competition for funds. While Congress could direct NSF to launch a center specifically for HTS, this would bean unfortunate precedent, given that the Foundation has traditionally avoided tar-
PAGE 111
105
PAGE 112
[Page Omitted] This page was originally printed on a gray background. The scanned version of the page is almost entirely black and is unusable. It has been intentionally omitted. If a replacement page image of higher quality becomes available, it will be posted within the copy of this report found on one of the OTA websites .
PAGE 113
107
PAGE 114
108 geted R&D. (Chapter 5 discusses a number of alternative approaches.) Some academics have feared that increases in funding for centers and other multidisciplinary programs would come at the expense of single-investigator and small-group research. While a legitimate concern, figure 4 shows that the relative shift has been small. Without growth in the NSF budget, competition for limited funds will intensify. Independent research must be preserved. Even so, it would seem prudent to risk erring on the side of support for the new multidisciplinary centers, rather than on the side of a continuation of traditional funding patterns. There are other ways as well to foster a multidisciplinary environment in the university system: for example, federally funded postdoctoral fellowships could be designed to encourage scientists and engineers planning academic careers to to move laterally into related fields e.g., from chemistry to materials, from electrical engineering to solid state physics (Option 10). Table 13.lssue Area II: Strengthening Interactions Among Universities, Industry, and Government Issue Options for Congress Advantages Disadvantages A. University-industry interactions; Multidisciplinary Research OPTION 9. Congress could: Commercialization of HTS requires multidisciplinary R&D, To do a better job of training people who can help American firms compete, universities will need to encourage multidisciplinary research and teaching. Federal agencies, notably NSF, have been increasing support for multidisciplinary research, but have had limited funds to accomplish this, Most of the incentives in American universities reward those who pursue conventional research careers; few encourage faculty members to cross disciplinary boundaries. Provide full funding for NSF to launch its proposed interdisciplinary Science and Technology centers. The Foundation seeks a one-time appropriation for fiscal 1989 of $150 million to support 10 to 15 centers for 5 years. l Appropriate funds at the $5 million or above Ievel for NSFs Industry/University Cooperative Research centers over each of the next several years, ensuring that the newer centers do not overshadow this program. Congress might also consider renewed support for the industry/University Cooperative Projects Program. Ample continuing support for NSFs Engineering Research Centers, provided evacuations indicate they are effective, also seems appropriate. OPTION 10. Direct NSF, along with other agencies that fund postdoctoral fellowships, to establish programs specifically for scientists and engineers who chose to move to a related field for a year or more of research. B. Government-industry interactions: Technology Transfer and Joint R&D Over the past few years, Congress has OPTION 11. Conduct early oversight on the enacted several pieces of legislation intendresponses of major R&D agencies ed to encourage transfer of technology from particularly the Departments of Defense and Federal laboratories to industry. These proEnergyto recent laws and executive vide a framework for reform, with decenbranch actions aimed at speeding technoltralized decision-making at the laboratory ogy transfer and commercialization of fedIevel. While some of the laboratories have erally funded R&D. responded enthusiastically to the new laws, it is not clear that the agenciesespecially at higher levelshave embraced this mandate, More support for multidisciplinary research and teaching could help train engineers and scientists to do a better job of bridging the gaps between research and design, development and production, R&D and marketing. Not only will this be vital for competitiveness in HTS, it is vital throughout the U.S. economy. According to the National Research Council, such fellowships would facilitate communication among disciplines and seed the faculty with individuals who are experienced in the cross-disciplinary approach. d The oversight process could help Congress determine whether further changes in the legislative framework are needed. Matters that might be examined include: l Whether to require that Federal agencies issue regulations for implementing the provisions of the Federal Technology Transfer Act of 1986, The law does not require agencies to issue implementing regulations; indeed, it specifies that they shall not delay implementation until rules are issued. But the situation is a new one for industry too, and lack of guidelines may discourage them from approaching the laboratories, Without a corresponding increase in NSFs overall budget (see Option 4 in Table 9), money for centers could come at the expense of individual and small group researchone of the outstanding strengths of the American university system. Without complementary changes in the university environment, such moves might hurt the career prospects of those accepting fellowships. Reforms take time to implement, It may be too early to get an accurate reading of agency responses to the new rules for technology transfer, The oversight process itself could mean that responsible officials spend time answering inquiries that otherwise would go into improving transfer processes. aD/reCt/on3 in Enginwring ReSearch: A n ASSeS~ment of @poflun/t/es and Needs (Washington, Dc: National Academy press, 1987), p. 67.
PAGE 115
109 Table 13.lssue Area II: Strengthening Interactions Among Universities, industry, and Government-Continued Issue Options for Congress Advantages Disadvantages Technology transfer may get few resources and little attention when it is not viewed as part of the agencys own mission, For HTS, effective transfer mechanisms could be especially important, DoD, with more money to spend on this technology than other agencies, has fewer reasons for working hard to transfer R&D results to commercial (non-defense) industry Demonstration projects could help Identify better methods for transferring technologies to industry, but little funding has been available. The same is true of demonstration projects involving R&D cooperation between the national laboratories and industry. If the national laboratories are to transfer technologies to industry effectively, many more laboratory employees will need to understand industrial needs and marketplace realities While industrial (or university) scientists can arrange to work in a Federal laboratory with little difficulty, the primary need is for movement in the other directionfrom the laboratories to industry. DOEs national laboratories are seeking a major role in helping U.S. industry commercialize HTS, but as yet have limited experience in cooperative R&D with the private sector, Working out R&D arrangements that suit industrys needs without detracting OPTION 12. Direct DoD, working with DOE and the Federal Laboratory Consortium for Technology Transfer, to use on a trial basis an intermediary or adjunct organization for transfer of HTS technology to nondefense firms. The intermediary would need to have well-established working relationships with the private sector, and strong motives for making the transfer process function effectively. OPTION 13. Appropriate or allow more money to be set aside for the Federal Laboratory Consortium for Technology Transfer to undertake three or more demonstration projects on technology transfer and/or R&D cooperation over the next year or two. Projects with outcomes relevant to several agencies would be most useful. Possibilities include: pilot programs at the State level (see Option 17 below); l development of guidelines, and trials, involving intermediary organisations (see Option 12 above), l preparation and testing of a technology transfer training program for laboratory (and industry) employees. OPTION 14. Authorize and encourage temporary exchanges of technical personnel (and sharing of personnel), as well as cooperative R&D projects between industry and the national laboratories. HTS could get special attention. Alternatively, Congress could create a broader exchange program to send engineers and scientists from national laboratories to private corporations for periods of 6 months to 2 years. One hundred fellowships per year would begin reaching enough laboratory employees to make a difference. Laboratory engineers and scientists could be required to work on problems of mutual interest, with the Government paying half their salaries and maintaining pension eligibility benefits. OPTION 15. Direct DOE to encourage an experimental approach to cooperation with industry, As the Departments laboratories establish pilot centers for HTS R&D, and engage in other collaborative efforts with industry and universities, each center could l Actions taken by the laboratories to improve institutional support for technology transfer through personnel policies and provisions for royalty sharing with inventors, Effects of agency mission on the course of technology transfer. Congress might also ask DoD and DOE how, specifically, their procedures will apply to HTS. l The success of the Federal Laboratory Consortium in living up to its mandate under the 1986 Act. Given DoDs funding Ievels for HTS R&D, transfers to the civilian side of the economy could have substantial impacts on commercialization. Once R&D results were approved for transfer by DoD, the intermediary could take on the job of working with industry, minimizing interference with the primary missions of DARPA, SDIO, and the services. Regardless of the mechanism chosen, an HTS technology transfer program could be viewed as a demonstrationwith high visibility and potential relevance for other technologies. The FLC received about $700,000 during 1987 under a set-aside specified in Public Law 99-502, with only 5 percent available for demonstration projects, Additional funds for demonstrationsperhaps $300,000 per yearwould begin to address the need. Such a program would serve a need largely unmetgiving laboratory employees handson industrial experience, thereby speeding commercialization. Fellowships could be made available to laboratory personnel on a competitive basis. Temporary assignments in universities would not serve the same purpose, nor would programs that focus only on bringing industry people into the laboratories. Cost sharing by companies would help ensure that the laboratory fellows worked on commercially relevant problems. An experimental approach would help the laboratories learn to work with industry without consuming a disproportionate share of HTS research dollars, Trying a number of different approaches implies learning from the results, hence provision for evaluation; Transfers from DoD might come to be viewed as substitutes for R&D funding by civilian agencies, to the possible detriment of commercial technology development, Each technology transfer situation is unique, putting limits on the lessons to be learned, Nor can a cookbook approach to technology transfer function effectively. Such a program carries risks of conflict of interest, as well as the appearance of subsidy, Moreover, the laboratories might find industry hiring away some of their more valuable people. Some firms might fear they could lose control over proprietary technology. Relying too heavily on cooperation between the laboratories and industry, particularly to the exclusion of other policies for speeding commercialization, would be a mistake. There is a second dan-
PAGE 116
110 Table 13.lssue Area II: Strengthening Interactions Among Universities, Industry, and GovernmentContinued Issue Options for Congress Advantages Disadvantages from broader laboratory missions could require considerable experimentation Under the right circumstances, collaborative R&Dinvolving several private firms in pre-competitive projectscould be an efficient mechanism for building the HTS technology base. Yet the time horizons of industry consortia are unlikely to be that much longer than those of individual firms State Governments have a broad range of economic development tools at their disposal. In addition to the direct funding for R&D that some have provided, States could help commercialize HTS through programs that accelerate the diffusion of research results to industry. At present, however, linkages between State Governments and national laboratories within their borders tend to bead hoc and not very well established. be designed somewhat differently, even though all were charged with aiding in commercialization OPTION 16. The Federal Government could make funds for HTS R&D available on a cost-sharing basis to industry consortia, provided the funding agency determines that public money will serve to extend the R&D time horizons. OPTION 17. Congress could: l Provide small planning grants to the States for strengthening R&D-based economic development initiatives, including grants for the evaluation of existing programs, H may take 5 years or more for States to put new programs in place; planning grants available now could mean better capabilities at the State level when HTS technologies begin moving out of the laboratory, l Fund several State Government pilot projects embodying different approaches to the transfer and commercialization of federally-funded HTS R&D (conducted in universities as well as national laboratories). Direct Federal agencies to give greater weight to support from State Governments in evaluating proposals for university-based R&D centers, and other proposals where commercialization is a major objective. to succeed, the laboratories will have to be self-critical Approaches that worked for HTS could be adopted elsewhere. Cost-sharing of longer-term R&D would address a critical problem for U.S. competitiveness. The Federal contribution could involve provision of facilities (e.g., at a national laboratory) and/or temporary assignments of personnel to a consortium, in addition to financing, Strengthened capacities in the States to assist smaller businesses in commercializing innovative technologies would complement Federal SBIR (Small Business Innovation Research) programs, particularly Phase Ill efforts, Planning grants could also help the States find ways of bridging the gap between Phase I and Phase II awards. ger as well: DOE and the laboratories might find it difficult to shut down cooperative projects that proved ineffective, or were no longer needed Any project involving Federal funding would be subject to the vagaries of the budget process. Unless Government, as well as industry, lengthened its time horizons, money could be wasted. On the other hand, costsharing, once started, might be difficult to stopeven if, in time, the justification vanished. Few State programs have been evaluated by independent parties; little is known about the approaches that work best Federal assistance could end up favoring States that might need help the leaste g those that already have well-developed programs SOURCE: Office of Technology Assessment, 1988 TECHNOLOGY TRANSFER: THE FEDERAL LABORATORIES Much of the Federal funding for HTS R&D for commercialization until they have proven is going to government laboratoriesmostly fathemselves. cilities run by DoD and DOE, but also to NBS and NASA research centers. These laboratories differ in missions, in their historical ties with New Rules for the Laboratories industry, and in operating arrangements. While NBS has long had good relations with indusFor many yearsas congressional hearings try, and DoD laboratories often work closely and an accumulation of studies pointed to the with military contractors, few Federal laboralarge fraction (said to be 90 percent) of federtories have accomplished much in commercialially owned patents never licensed or otherwise zation. This has not, after all, been one of their commercializedthe U.S. Government has tasks. Whether the laboratory system will be sought to stimulate commercial use of publicly able to contribute much beyond a general funded R&D. Since 1980, Congress has enacted strengthening of the HTS technology base a series of laws intended to give industry greater remains an open question. Certainly it would access to the laboratory system, and to speed be a mistake to rely heavily on the laboratories transfers of technology to the private sector.
PAGE 117
111 As a result of patent law changes in 1980, small businesses, non-profit organizations, and universities can gain title with relative ease to inventions they make in the course of R&D paid for by Government. In 1984, Congress extended this statutory policy to contractor-operated laboratories, including several DOE facilities. (The statutory policy does not extend to weapons laboratories, or DOE laboratories operated by large, for-profit businesses, although th e Administration has initiated changes here as well.) The most recent step, the 1986 Federal Technology Transfer Act, seeks tighter links between government-operated laboratories and industry. This law: l provides clear authorization for government-owned and -operated laboratories to enter into cooperative R&D with private firms l Gives the Federal Laboratory Consortium on Technology Transfer (FLC) a statutory charter. About 400 laboratories, representing 11 agencies, belong to the FLC, which was organized to facilitate use of federally developed technologies 15 l provides for agencies to return licensing income to the originating laboratory, and requires that at least 15 percent of royalties or other income go to the employees responsible l Directs laboratory directors to consider technology transfer activities in performance evaluations and promotions, and to include it in job descriptions. The 1986 Act decentralizes many administrative responsibilities, giving substantial discretion to the laboratory directors. Beyond these statutory changes, President Reagans April 1987 Executive Order 12591, on facilitating access to science and technology, establishes guidelines for all the laboratories. While many of the laboratories have expanded their technology transfer activities over the past several years, the pace of change at the agency level has often been slow. Moreover, s Strategic Plan: 1988-1992, Federal Laboratory Consortium for Technology Transfer Administrator, Fresno, CA, October 1987. the discretionary authority given to laboratory directors in the 1986 Act applies only to government-operated laboratories, not to contractor-operated facilities like DOEs. To help determine whether further policy modifications might be needed, Congress could conduct oversight on the responses of the mission agencies to the 1986 Technology Transfer Act, other recent changes in the law, and to Executive Order 12591 (Option 11, table 13). Transferring HTS R&D While a new framework for technology transfer exists, it is far from clear that industry and the laboratories will be able to forge effective partnerships for commercializing technologies like HTS. Many of the formal barriers have come partway down, but the culture of these 700-plus institutions insulates them from industry and marketplace. The laboratories also differ greatly in style and tradition. Some stress engineering, others research for the sake of research. Policies with much to recommend for a DOE facility maybe irrelevant for NIH, while conflicting with DoD security requirements. Technology Transfer from DoD Much of the Federal funding for superconductivity passes through DoD, which operates more than 70 laboratories, a pattern that will probably continue. While defense agencies work hard at transferring technologies to military contractors, diffusion to the civilian side of the economy poses special problems. These begin with the frequent requirements for secrecy, and end with the likely reluctance of the Pentagon to accept such a burden as a major ongoing responsibility. In authorizing an HTS program for fiscal years 1988 and 1989, Congress instructed DoD (and DOE, when its laboratories receive DoD funds) to give special attention to transfers of technology to the private sector. 16 Apparently, ~esection 218 of the National Defense Authorization Act fOr 1988 and 1989 (Public Law 100-180) earmarked $60.56 million annually for 2 years for a DoD program on HTS. Congress appropriated only $15 million for fiscal 1988 under this provision, which went to DARPA for initial funding of its processing R&D effort, described earlier in the chapter.
PAGE 118
. 112 DoD intends to use existing mechanisms to implement the requirements of the Defense Authorization Act, rather than establish procedures specifically for HTS. While current practices may suffice for transferring HTS technologies to defense industries, they will probably be less effective for transfers to firms on the civilian side of the economy. Instead, it might make sense to assign the task of working with non-defense firms to an intermediary organization (Option 12). A number of arrangements seem feasible. Several DOE laboratoriesincludin g Argonne, Oak Ridge, and Ameshave set up adjunct organizations to handle technology transfer. DoD could contract directly with an existing organizatione.g., a not-for-profit R&D laboratory like Battelle. An intermediary charged exclusively with transferring technical knowledge to commercial enterprises could play a useful role during the stages of technology development and commercialization processes that are not germane to DoDs mission. Demonstration and Evaluation Technology transfer has significance going beyond HTS. So does cooperation in R&D between industry and the national laboratories. But both in the laboratories and at middle and upper ranks in the agencies, commitment to meaningful change has not always been visible. Information about what works would help; successful demonstration projects could have considerable impact (Option 13). The FLC has the authority to conduct demonstrations, but its set-aside funds from the agencies paid for only one such project during 1987. Making technology transfer function effectively and efficiently will demand systematic, empirically-based analysis of transfer processes (including cooperative R&D), and of subsequent impacts on innovation and commercialization. Demonstration projects without critical evaluation of results may not accomplish much. Laboratory Personnel People transfer technology much more effectively in person than through reports, and they do so best when they work together (rather than in meetings). Transferring HTS technologies from the national laboratories means: 1) bringing people from industry into the laboratories to work on HTS, perhaps through cooperative R&D projects; and 2) sending people from the laboratories to industry so they can learn what commercialization is all about. In the short run, the first of these steps has much to offer for HTS. The second step is necessary for lasting changes in the culture of the laboratories, and for long-run success in better integrating the laboratories into the Nations R&D infrastructure. Because of possible conflicts with DoD missions, personnel exchanges have greater potential attraction at DOE (non-weapons) laboratories. The laboratory system attracts many highly competent people with more interest in research than in the practical problems of industryno surprise, given that commercialization has not been a mission of the laboratories. Some people join a laboratory precisely because they have no wish to work on industrial problems. They may be highly capable professionals, dedicated to research, but even if motivated to work with industry, laboratory employees may not know howthrough lack of exposure to corporate life and the realities of the marketplace. Agency policies have been broadened in recent years, so that many Federal employees can do consulting (on their own time), or take leaves of absence to work in industry. Both these steps will make a difference. So could a program of temporary appointments sending laboratory personnel to the private sector (Option 14). Although industry employees can come to the laboratories quite easily, flow in the other direction will have more impact in changing the laboratory culture. Congress could explicitly authorize and encourage fellowships and/or exchanges. Several of the HTS bills introduced in the 100th Congress authorize industrial fellowships at Federal laboratories, but not fellowships that would send laboratory employees to industry. Others have tied personnel exchanges to cooperative R&D programs. The need is a broad one: there seems no necessary
PAGE 119
113 reason to tie personnel exchanges either to cooperative R&D or to HTS. 17 Cooperative R&D Cooperative research bringing together laboratory employees with those from industry and perhaps from universitiesoffers another way to integrate the laboratories more effectively into the Nations technological infrastructure. The possible arrangements serve needs ranging from efficiencyavoiding too much duplication to lengthening industrys time horizons, as discussed in the next chapter (see box R on collaborative R&D). The discussion below focuses on HTSe.g., approaches such as the proposed National Laboratory Cooperative Research Initiative Act, S. 1480 in the 100th Congress. 18 DOE has itself moved toward closer cooperation with industry. In April 1988, the Secretary of Energy designated three national laboratoriesLos Alamos, Argonne, and Oak Ridge as superconductivity pilot centers. The pilot center approach, including expedited procedures for contracting and project approval, and transfer of intellectual property rights, had been initially proposed by Los Alamos, which had been asked by the Secretary to explore mechanisms for cooperative ventures. DOE user facilities have begun attracting the attention of private firms: the Departments figures show 1600 industrial visits in 1987, compared with 260 in 1981. 19 But collaborative ~TThe presidents Commission on Executive Exchange, in response to the April 1987 Executive Order, has been working on a small-scale plan for exchanges of technical personnel between industry and Federal laboratories. Industrial participants will probably be limited to relatively large companies that can afford to share the administrative expenses, as well as picking up part of the costs of the exchange. Contractor-operated laboratories may not be coveredthus excluding most DOE facilities. InIntroduced in 1987, S. 1480 also includes provisions for cooperative R&D on mapping the human genome and semiconductor manufacturing. The bill, in its original and modified versions (Senate Amendment 1627, introduced in March 1988) would direct the Secretary of Energy to establish cooperative centers at DOE national laboratories for HTS R&D, and give the laboratories greater autonomy in negotiating agreements with private companies and universities. leThe Ig87 estimate comes from DOEs Laboratory Management Division, that for 1981 from the statement of Dr. James projects with industry, though on the rise, involved just 57 companies and R&D valued at about $110 million during 1987 (for the multiprogram laboratories). Given this so-far modest showing, Congress might direct DOE to take an explicitly experimental approach to cooperation with industry (Option 15). Rather than a full-scale effort, structured trials could help industry and the laboratories find ways of working together while avoiding unrealistic expectations and the danger of steering too many HTS R&D dollars to untested programs. Pilot projects could take a variety of forms: firms might work with the laboratories singly or in groups; potential rivals could choose to pursue pre-competitive projects jointly; firms with similar R&D objectives, though in different businesses, could cooperate, along with those having supplier-customer relationships. Industry will have to take much of the initiative if the DOE laboratories are to aid in commercialization of HTS. Companies must be willing to search out areas of expertise in the laboratory system, and seek to take advantage of themcontributing a substantial share of project funds. If Federal dollars cover too high a fraction, the company may no longer feel it has a stake in outcomes; projects can stray from the needs of commercial technology development. The laboratories might also find that the only companies working with them were those with few prospects for commercial success, and little choice but to take whatever help DOE might offer. At the same time, while few firms are likely to make substantial financial commitments without guarantees of influence over research goals, industry cannot have too much control, else planning horizons will shorten: unless cooperative projects have riskier and/or more generic R&D objectives than companies would pursue on their own, there is little justification for Government participation. These considerDecker at the Joint Hearing on Technology Transfer before the House Committee on Science and Technology and the Subcommittee on Energy Research and Development of the Senate Energy and Natural Resources Committee, Sept. 4, 1986 (Washington, DC: U.S. Government Printing Office, 1987), p. 22.
PAGE 120
114 ations suggest projects that last 3 to 5 years or more, with industry cost sharing in the range of 40 to 60 percent. It might also be appropriate for Federal agencies to share costs with industry-based collaborative R&D ventures (Option 16also see ch. 5). The justification? Longer time horizons. While national laboratory participation might sometimes be desirable, there seems no reason to make this a precondition. Regardless of final policy decisions on cooperative R&D, mechanisms for evaluating differing approaches, and disseminating the results and not just the success storieswill be needed. The approaches emerging could have relevance going well beyond superconductivity. State Programs and Approaches Over the past decade, many States, in the name of economic development, have established programs for supporting high-technology businesses. Some already support HTS. Among the more visible initiatives: l l l l l advanced technology centers intended to attract and work with high-technology industry; centers of excellence at state-supported universities (severale.g., at the University of Houston and the State University of New York at Buffalohave been established in superconductivity); small business innovation research programs, patterned after those at the Federal level (adopted quite recently by half a dozen States); technology extension services, intended to help companies attack technical problems and diffuse know-how to industry; financial assistance for start-up firms and small businesses. Many State governments have also established advisory commissions and councils on science and technology. The variety and innovative nature of State programs also mean that some of the undertakings have been fragmentary. Few States have comprehensive efforts. The New York State Science and Technology Foundation, and Pennsylvanias Ben Franklin Partnership, have been among the more extensive. Still, State governments have some tools to call upon for aiding HTS technology transfer and commercialization that are not available to the Federal Government. They also have compelling motivationsjobs and income. OTA has previously suggested that Federal matching funds for State programs such as technology extension services could be appropriate. 20 Additional possibilities (Option 17) include small planning grants to the States for strengthening R&D-based economic development initiatives. Pilot projects might include a demonstration program for State technology extension services, as provided for in S. 907 (incorporated in the omnibus trade bill passed by Congress but vetoed by President Reagan in May 1988). This option could complement existing Federal Small Business Innovation Research (SBIR) programs. Under the Small Business Innovation Development Act of 1982, Federal agencies must allocate 1.25 percent of extramural R&D budgets exceeding $100 million for SBIR awards: l l l Phase 1 contracts provide up to $50,000 for demonstrating the merit of an idea. Under Phase II, agencies may award up to $500,000 for taking Phase I concepts to the pre-prototype stage. In Phase III, companies can proceed with development using non-federal funds, or seek non-SBIR money from Federal agencies. Development and Diffusion of Commercial Technologies: Should the Federal Government Redefine Its Role? staff memorandum, Office of Technology Assessment, Washington, DC, March 1984, pp. 10 and 48. Precedents for assistance to the States include Federal funding during the 1960s and early 1970s for strengthening State and local capacities in dealing with issues of technology and science. At the time, these efforts were primarily focused on public technology e.g., the direct needs of State and local governments, rather than economic development and business needs.
PAGE 121
115 SBIR money has gone in the past for LTS reness advisory services, some of which offer search; Federal agencies, including DoD and assistance in applying for Federal SBIR grants. DOE, are currently evacuating well over 100 planning grants could help the States go furSBIR proposals for HTS awards. ther in complementing the Federal effort. Con. So far, half a dozen States have established gress might also consider raising the Phase I small business innovation research programs ceiling from $50,000which buys relatively little todayto, say, $100,000. of their own. Many States operate small busiTECHNOLOGY INTERCHANGE WITH JAPAN President Reagans superconductivity initiative speaks of reciprocal opportunities for the United States and Japan to cooperate in R&D. The Japanese Governments pronouncement s on HTS also stress international cooperation e.g., foreign participation in government-sponsored superconductivity projects such as the New Superconductivity Materials Research Association, and the International superconductivity Technology Center (ISTEC, ch. 3). If nothing else, Japan has sought to respond to criticism that its research system has been closed to foreigners. But agencies of the Japanese Government also have quite concrete benefits in viewnotably, new perspectives and new ideas that could strengthen Japans capabilities in basic science. More than symbolism, international cooperation could help the Japanese reach their own objectivea more creative R&D system. The Japanese also realize that they risk loss of access to research from the United States and other countries if they do not open up their own laboratories. New developments in HTS will continue to come from Japanese laboratories. American companie s as well as individual scientists and engineersstand to gain from participating in cooperative projects, but only in full partnership with Japanese companies and Japanese scientists. More than direct benefits are at stake. Hands-on involvement in Japanese R&D will help Americansas organizations and individualsunderstand how the Japanese compet e so effectively. Participation and Monitoring As this report was being completed, no U.S. firm had agreed to join one of Japans cooperative projects as a full member, although a few had become affiliates. American companies, at this point, feel that the costs are too highand not only the fees (full membership in ISTEC runs about $800,000). To benefit from full membership, a company would have to assign one or more highly competent professionalsfluent in Japaneseto the cooperative project. Scientists with relevant skills and experience are rare in both countries. American firms would be reluctant to send one of their best people to Japan, even if they had someone who spoke the language. (Evidently, few U.S. subsidiaries in Japan have not had much success in hiring toprank engineers and scientists.) For smaller U.S. firms, especially those without Japanese affiliates, any form of participation may be difficult to justify. If U.S.-based professional societies or trade associations were permitted to join Japans governmentsponsored projects, spreading the costs, American industry could gain better access to Japanese HTS R&D. Alternatively, a number of American firms might form a joint venture for such purposes as monitoring HTS R&D in Japan, keeping members aware of opportunities for individuals as well as companies, and helping transfer technology to the United States. The U.S. Government could support such an effort, perhaps by helping finance an office in Japan, or as part of a larger program such as NSFs Japan Initiative (see Option 18 in table 14). Federal support for such an office would build on precedents including aid provided by the Commerce Department in 1984 to the American Electronics Association for a trade office in Tokyo. Some American companies already operate Japanese affiliates primarily as listening posts,
PAGE 122
116 Table 14.lssue Area Ill: Technology Interchange with Japan Issue Options for Congress Advantages Disadvantages Smaller U.S. companies may not have the resources to keep up with fast-breaking developments in Japan. While the Japanese have offered foreign firms opportunities to participate in governmentsponsored cooperative projects, the response to date has been tepid. A joint venture or an organization such as a professional society or trade association might be able to spread the costs and help American industry gain access to Japanese HTS R&D. Few technical professionals in the United States have the language skills or inclination to take temporary appointments in Japanese laboratoriesthe most direct means for transferring technology and know-how from Japan to the United States, and a necessary step in improving American understanding of Japans research system. No more than a tiny fraction of U.S. scientists and engineers will learn Japanese in the near future, leaving an ongoing need for prompt translations of Japanese scientific, technical, and business publicationsincluding informally circulated gray literature. OPTION 18. Provide a seed grant to a professional society or trade association for an office in Japan to monitor developments in HTS, with funds sufficient to operate the office for perhaps 5 years. A non-profit organization such as the American Institute of Physics, the American Chemical Society, or the Federation of Materials Societies should bean acceptable vehicle in Japanese eyes. OPTION 19. Monitor progress in implementing NSFs Japan Initiative, appropriating additional money if U.S. funds together with contributions from Japan cannot sustain all components of the program. OPTION 20. Direct the Department of Commerce, as part of its responsibilities under the Japanese Technical Literature Act of 1986 (Public Law 99-382), to establish a program specifically for gathering, evaluating, and disseminating information on Japanese science/technology and business activities as they relate to HTS. The effort might be viewed as a model for improving the effectiveness of Commerces programs on Japanese information. For insightful evaluations of technical efforts in Japan, Commerce will probably need help from Federal agencies with greater expertise in engineering and science. SOURCE: Office of Technology Assessment, 1988. just as the Japanese maintain technology centers in the United States. The Japanese should be willing to allow group participation by American firms in information-exchange activities such as the New Superconductivity Materials Research Association. Moreover, there is no reason why a corporation setup as a joint venture should not qualify to join ISTEC. Language Training; Fellowships Laboratories in Japanese If the United States is to make better use of R&D conducted in other countries, more Americans will have to seek and take temporary assignments in foreign laboratories. At present, One way or another, the United States should take the Japanese up on their offers to cooperate in HTS research. In addition to serving as liaison to forums like the New Superconductivity Materials Research Association, a professional society or trade organization could help screen the latest scientific and technical information in Japanese, identifying HTS research reports for translation and distribution. This function could complement the current effort, quite small, by the Commerce Department under the Japanese Technical Literature Act (see Option 20 below). To take advantage of R&D opportunities, and Japanese technical know-how, more Americans need, not only language training, but experience working In Japan. Access to foreign scientific and technical information has become increasingly important as U.S. technological advantages have diminished. In the past, technical translations from German and Russian have been more common than from Japanese. Japan might gain more than the United States from cooperation in HTS. Much of the U.S. work will take place in universities, where it will be relatively open; much of Japans work will take place in industrial laboratories. Sending more engineers and scientists to Japan, and funding language training for professionals, will be only small steps forward, Longer-term needs begin with language training in U.S. primary and secondary schools. Translations and technical evaluations will need substantially higher funding levels to accomplish much. During fiscal 1987, Commerce reprogrammed $300,000 to implement the Japanese Technical Literature Act; in 1988, the Department plans to reprogram $500,000 for this purpose. An aggressive effort on HTS alone could well consume most or all of this. Screening and evaluation of technical information is particularly important, but expensive. To this point, U.S. companies have shown little interest in Japanese technical and scientific literature. few U.S. engineers or scientists have the right combination of technical qualifications, language skills, and motivation to work in Japan (in part because they may feel that their employerand the U.S. labor market as a whole will not reward them for learning Japanese and spending time there). As many as 7,000 Japanese engineers and scientists are currently at work in the United Statesin government facilities, as well as universities; perhaps 500 Americans work in Japanese laboratories. 21 Japan offers research fellowships for foreigners under the sponsorship of its Key Technol21E. LaChiCa, U. S., Japanese Negotiators Deadlocked on Tapping Each Others Technology, Wall Street Journal, Jan. 22,1988.
PAGE 123
717 ogy Center, as well as the Ministries of Foreign Affairs and Education. Given the paucity of language skills within the U.S. technical communityand such less tangible but no less substantial barriers as difficulty in finding employment for husbands or wivesnot many Americans have sought out these opportunities. Cultural differences and high living costs create obstacles particularly for more senior American engineers and scientists, including those in R&D management positions. NSFs Japan Initiative, launched in 1988 with an allocation of $800,000 from the Foundations budget for bilateral programs, is designed to encourage more Americans to take advantage of research opportunities in Japan. The program offers fellowships for language training, as well as financial support while in Japan. Japans Prime Minister Takeshita, moreover, has announced that that his government will give NSF $4.8 million to finance work in Japan by U.S. researchers, With this offer, near-term funding for the Japan Initiative seems adequate. NSF is seeking $1.6 million for the program in fiscal year 1989; Congress might monitor progress in implementing the program, and appropriate additional money if U.S. funds (together with contributions from Japan) cannot sustain all its elements (Option 19). Congress might also wish to consider greater Federal assistance to American schools and universities for language training. Technical Information Most U.S. scientists and engineers necessarily will continue to rely upon translations for technical information from Japan. Congress could direct the Commerce Department to expand its small program for translations of Japanese technical literature under the Japanese Technical Literature Act, perhaps appropriating funds specifically for information on HTS (Option 20). A professional society or trade organization (as discussed under Option 18), in addition to serving as liaison to organizations like the New Superconducting Materials Forum, could help screen the latest scientific and technical information on HTS. While major scientific findings from Japanese laboratories normally see publication in English or another Western language, less of the Japanese engineering literature is translated. Moreover, Japan produces a large volume of gray literature company, university, and government reports, as well as other informal documents not widely circulated. The gray literature, hard to acquire outside Japan, often includes important technical and business information. Foreign Access to U.S. Technology The U.S. Government has signaled Japan and other nations that it may restrict outflows of information from Federal HTS research. The July 1987 Federal conference, at which President Reagan announced his superconductivit y initiative, was itself off-limits to representatives of foreign governments. In the private sector, the Council on Superconductivity for American Competitiveness limits its membership to U.S. corporations and citizens. One title of the Administrations proposed Superconductivity Competitiveness Act, sent to Congress in February 1988, would permit agencies to withold scientific and technical information requested under the Freedom of Information Act under some circumstances. The scientific community has long argued that restrictions on information exchange harm its enterprise, and can only be justified on strict grounds of national security. But the question for HTS is rather different. Proposals such as the Administrations seem to assume that the United States is far enough ahead in HTS to have something to protect. OTA has found no evidence supporting such an assumption. Lacking a decisive lead in the R&D race, measures seeking an equitable two-way flow with countries such as Japan have much more to recommend them. Of course, the threat of embargoes on scientific and technical information helps keep the pressure on other nations to provide access to their own research systems (box P).
PAGE 124
118 CONCLUDING In the months following the initial breakthroughs in superconductivity, the U.S. Government moved quickly to redirect R&D support. The growth in fundingfrom virtually nothing in fiscal 1986, to $45 million in 1987, and $95 million in fiscal 1988demonstrates the responsiveness of the system. Yet there are real reasons for concern: lack of new money for HTS; the possibility of a reaction against continuing high levels of R&D spending unless exciting new results keep coming in; heavy de facto reliance on DoD and DOE to generate technology that industry can commercialize. Budget uncertainties in the R&D agencies, which lasted well into the current fiscal year, put many federally funded projects on hold, slowing U.S. progress. Defense-related spendingcenterpiece of the Federal R&D budget since World War IIleads to major new commercial products or processes less frequently than in earlier years. The reasons include a drop in support for both generic and high-risk, long-term R&D relative to the overall DoD R&D budget, as well as growing isolation of the defense sector of the economy. Meanwhile, funds for applied research that would fill the gap between basic science and the short-term projects conducted by industry have been cut back: the U.S. Government spends little money on work that would strengthen the foundations for commercial industries. At present, DoD has roughly half the Federal money for HTS. The field is new, still in the research stages. Much DoD-sponsored R&D over the next few years should yield broadly useful results. Thus DoDs ample resources could become a major asset in commercializing HTS. But the Pentagon will begin steering dollars to support mission-specific applications as soon as these are in viewindeed, may already be doing so. When it comes to the Department of Energy, which is getting 30 percent of the Federal funds, the primary questions concern the ability of the national laboratories to forge new cooperative relations with industry. The laboratories are changing. But the system is a big one, burdened with inertia; commercialization has not been a significant mission. It will take a major departure from business as usual for the laboratories to have much impact on commercialization of
PAGE 125
11 9 HTS (which is not to say that when the next set of opportunities comes along, the DOE laboratory system may not be in a better position to respond). The Federal Government pays for about twothirds of the R&D carried out on the Nations campuses, chiefly through awards to individual faculty members, Given the short-term orientation of most business-funded R&D, the NSF budget for HTS is particularly critical. Policies aimed at breaking down some of the disciplinary barriers in American universities, and creating environments where truly interdisciplinary research could flourish, would help broadly in the commercialization of this and other technologies. The R&D budget is not the whole of technology policy. Nor is technology the only ingredient in successful commercialization. All of the policy options covered in this chapter, taken together, would be no more than a first step in addressing the competitive difficulties of American industry. Still, the money Federal agencies spend on R&D, the ways they spend it, their efforts to transfer technologies to industryactions taken every day by more than a dozen agencieshave enormous long-run impacts. When companies search for competitive advantages through proprietary technology they draw continually on this publicly funded, publicly available technology base. More effective interactions among the major players in the R&D systemindustry, the national laboratories, universitieswould speed the generation of technical knowledge, and, perhaps more importantly, its use. In a time of budgetary stringencies, mechanisms for establishing R&D priorities across the agencies become more critical than ever. This is perhaps the single most important poin t raised in this chapter. To maintain its competitiveness, the United States must generate today the technical knowledge that industry will depend on tomorrow, For HTS, this means, not only effective mechanisms for setting R&D priorities, but stability and continuity in funding. These needs imply another: a strategic view of the ways in which federally funded R&D can spur economic growth and competitiveness the subject of the next chapter.
PAGE 126
. Chapter 5 Strategies for Commercial Technology Development: High-Temperature Superconductivity and Beyond
PAGE 127
CONTENTS Page Summary . . . . . . . . ........................123 Strategy l: Flexible Response . . . . . . . .........128 The Current Approach . . . . . . . . . ........128 Strengths and Weaknesses . . . . . . . . .......129 Strategy 2: Unaggressive Response to HTS ..........................13 2 A Larger Role for NSF . . . . . . . . ...........132 Industrial Consortia with Federal Cost Sharing ......................133 A Working Group on Commercialization. ................,..,.......13 7 Strategy 3: A Federal Technology AgencyThree Alternatives .........,138 An Umbrella Agency for Science and Technology ...................140 Higher Priorities for Technology and Engineering ...................141 An Agency for Commercial Technology Development ................142 Concluding Remarks.. . . . . . . . . .............147 Boxes Box Page Q. Strategy: Key Ingredients . . . . . . . . . . .124 R. National Laboratories or Universities?. ............................13 1 S. Collaborative R&D . . . . . . . . ..............135 T. Aggressive Support for Commercialization: Other Possibilities ........139 U. An Advanced Technology Program . . . . . .........143 V. An Advanced Civilian Technology Agency, as Proposed in S.1233 .. ...144 W. Project Review . . . . . . . . ..................146 Table Table Page 15. Desirable Features in a Federal Agency for the Support of Commercial Technology Development. . .........................14 7
PAGE 128
Chapter 5 Strategies for Commercial Technology Development: High-Temperature Superconductivity and Beyond SUMMARY Together, a collection of government actions constitutes a strategy, just as the actions of a corporations upper management constitute a strategy. De facto strategies, though hardly unheard of in business, are more common in government. indeed, to some, the very notion of strategy implies a measure of loss in one of the primary strengths of U.S. technology policies the flexibility of Federal agencies, their ability to respond quickly to new circumstances, Regardless of approach for promoting commercial development of high-temperature superconductivity (HTS)and regardless of whether the approach is called a strategysuccess will require diversity in sources of funding and in the R&D programs that Government money supports, Earlier chapters stressed the uncertain in prospects for HTS. With several agencies involved, good ideas will get a hearingat the Defense Advanced Research Projects Agency (DARPA), if not the Department of Energy (DOE) or the National Science Foundation (NSF). Duplication, in any case unavoidable, can spur competition. Continuity, likewise, will be important. To encourage U.S. industry to take a longer view, the Federal Government must do so itself (a need addressed by several of the policy options in ch. 4). In keeping with continuity and stability in funding, any strategy for HTS should avoid high visibility. If policy makers or the public look to Federal programs for near-term breakthroughs, disillusionment will follow. Technological advance cannot keep up with expectations fed by the media The Federal Government will have to let mar ket forces drive HTS technology as much as possible. Historically, governments have done a poor job of trying to anticipate what markets will demand, if not the course of technological evolution; picking R&D fields ripe for major technical advance is one thing, picking winning commercial applications quite another Policies that pull technologies into the marketplace in more subtle wayse.g., through government procurementcan work, especially in conjunction with R&D funding designed to push the technology along. Market pull coupled with technology pushin a policy environment that encourages collaboration among industry, universities, and Governmentwill help speed commercial technology development. Box Q discusses these and other operating principles in more detail. This chapter considers three approaches through which the Federal Government might foster commercialization of HTS: l l Flexible response, Strategy 1the current, de facto U.S. policygrows naturally out of postwar U.S. technology policy. Characterized by strong support for basic science and for mission-oriented technology development, direct measures for supporting commercial technology development find little place. An aggressive response, Strategy 2, would differ in three major ways from current policies. First, NSF would have more money for HTSin essence, an insurance policy to make certain that good ideas for basic research have a shot at funding. Second, the Federal Government would share in the costs of private sector collaborative R&D ventures. The rationale: more work within industry on long-term, high-risk HTS R&D. Third, a working group of experts from industry, universities, and government would be assembled to decide which collaborative R&D proposals were worthy of support, and to otherwise advise on policy measures, 123
PAGE 129
124 Box Q.--Strategy: Key Ingredients Disagreements over strategies for technology development reflect differences of opinion over technical questions (How long will it take to develop flexible conductors made from the new superconductors?) and over market conditions (What applications will be most attractive at liquid nitrogen temperatures?), as well as over matters of political preference (Is it proper for government to support commercial technology development directly?). Some policiese.g., support for basic researchfind nearly universal support. So do some of the other ingredients that would find a place in almost any strategy for supporting HTS: Diversity in sources of R&D support. No one knows what new discoveries may emerge in superconductivity, where they will come from, or when. A portfolio of research makes more sense than one or a few centers of excellence; so do multiple sources for contracts and grants. Continuity in support, over a period that could easily be a decade (as emphasized in ch. 4). A judicious balance of technology push and market pull. Technology push via R&D support works best when accompanied by policies such as government purchases and demonstration projects that help pull high-risk, high-cost technologies into the marketplace. Overall, however, government should let industry drive technology as much as possible. Measures that encourage collaboration among universities, industry, and Government. Diversity Many sources of support, and many centers for R&D, may mean duplication of effort, but that is not necessarily bad. Overlap breeds competition and helps ensure that no path goes unexplored. There is another side to diversity. Accountability can suffer: if everyone is responsible for HTS, then no one is fully accountable. Still, lacking an overriding goalsending a spacecraft to Mars, or building a magnetically-levitated (maglev) rail system along the Eastern seaboardwhere is the need for centralized responsibility? Development of HTS is a broad objective, and also a fuzzy one: the technology cuts across the missions of a number of agencies. NSF funds research proposals rated highly on grounds of their promise in advancing science and technology; the subject of the research carries less weight. Not without controversy, the process nonetheless has found wide acceptance. But when it comes to projects in the mission agencies, decisionmakers do not always see eye-to-eye on what should be supported. In the early days of computer technology, visionary projects such as Whirlwind and ILLIAC were as fiercely opposed by one set of agencies as they were favored by those paying the bills; U.S. computer technology would not have advanced so quickly had any one agency been solely responsible. 1 Continuity In reality, diversity is seldom a problem in the decentralized U.S. systemit comes naturally. The more common problem is continuity. Stop-and-go decisions have bedeviled U.S. technology policies, as chapter 4 makes clear. Congress passes the Stevenson-Wydler Act, but a new Administration does not fully implement it. The executive branch seeks to double the NSF budget over five years, but Congress does not appropriate the funds. If there is a secret to Japanese technology policy, it lies in continuity-stability in commitment and financing, without rigidity. Publicly-funded R&D programs in Japan, many of which have enhanced the competitiveness of Japanese industry although budget levels have been modest, often begin with an 8or 10-year planning horizon. R&D priorities change over the course of the maglev train program (ch. 3) or the fifth-generation computer project as results come in and new directions open. Budgets may change too. But sharp reversals are rare. A decade-long time horizon stands for all to see as a demonstration of commitment by both government and industry. R&D sponsored by the U.S. Government often lives from one budget cycle to the next; the consistency seen in Japanese policies Governments Role in Computers and Superconductors, prepared for OTA by K, Flamm under contract No, H36470, March 1988, pp. 9-27.
PAGE 130
125 has few precedents, even in defense (just as few American firms have shown the persistence in product development that led to export successes like video-cassette recorders for the Japanese). In the United States, the spectacular success of a few flagship efforts like the Manhattan Project and Apollo left a trail of unrealistic expectations. NSFs RANN program (box O, ch. 4) sprang from the notion that the technical expertise needed to put men on the moon could be turned, almost as directly, to social problems. Operation Breakthrough, the Department of Housing and Urban Developments effort to revolutionize the technology of residential construction, grew from the same soil. RANN came and went in half a dozen years, Operation Breakthrough even more quickly. HTS will be equally vulnerable. The publics expectations have been raised by a year and a half of scientific discoveries, a Federal conference featuring the President and three cabinet officers, and a Nobel Prizeall accompanied by ample media coverage. In the absence of steady progress, public support may wane; the painstaking and laborious work needed to turn science into useful technology spawns few headlines. Low Visibility In the United States, publicly-supported R&D programs have generally been more successful, and more durable, when they avoid high visibility. 2 Apollo was the exception, not the rule. If policy makers or the public look to a Federal HTS program for immediate technological or commercial triumphsin the extreme, as a flagship in the international competitive strugglethey will be disappointed. A highvisibility, crisis-driven program such as the synthetic fuels initiative of the 1970s may collapse on itself, and in doing so harm the cause of future support for related efforts. Technology Push and Market Pull Visibility, by itself, did not do in the synthetic fuels program. The failure lay in attempts by government bodies to anticipate the course of technological evolution and the needs of the marketplace. (Japan has not been immune: the Ministry of International Trade and Industry tried to force Honda out of the auto business in the 1960s.) The lesson, repeated many times over: in the absence of convincing reasons for doing otherwise, let market forces drive technology. Although picking winners is something that Federal agencies have never done well, policies that serve to pull the market in more subtle ways have proved beneficial. They are particularly effective in conjunction with R&D funding designed to push the technology along. To elaborate, the Federal government has confined its role in (non-defense) technology development largely to funding research, on the assumption that the commercial market could and would create the necessary demand for resulting products. But this is an area where the market does not function perfectly. Because product development efforts in high-risk technologies are extremely expensiveaccounting for nearly 80 percent of R&D costs firms must have some confidence that there are customers at the end of the tunnel; potential customers often do not have enough information or certainty to provide that assurance, however. Thus, there is a role for government in helping assure an early market for such products. In computers and semiconductors, Federal Government procurement provided that assurance. The defense-space share of the total computer hardware market was 100 percent in 1954, and it exceeded 50 percent until 1962. Similarly, during the early years of integrated circuit production, defense and space procurement accounted for almost 100 percent of sales. Given assurances of stable indeed growingdemand, companies raised their own R&D spending. Technological advances coupled with learning and scale economies led to dramatic price reductions for computers. Even more important was the demonstration effect: successful use of computers by military and space agencies proved their value to a skeptical business community. Defense procurement was effective at pulling computer and semiconductor technologies along for two reasons. The governments mission-based needs meant that agencies evaluated technological Collaborative Research: An Assessment of Its Potential Role in the Development of High Temperature Superconductivity, prepared for OTA by D.C. Mowery under contract No. H36730, January 1988, pp. 13-14 and 67-88.
PAGE 131
126 alternatives carefully and provided valuable user feedback to suppliers. Moreover, agency needs, and the technologies they spawned, meshed closely with business needs. (As one computer executive observed, Space and defense computer applications .,. served as a crystal ball for predicting the future direction of computer use in industry.] Federal demonstration projects have been similarly effective under the same conditions, i.e., when mission-based government support of a developing technology steers it in a direction which converges with commercial interests. Lacking this, though, the synergy can quickly vanish. Because of the growing divergence between military and civilian technologies and markets, defense procurement no longer has the positive impact it once did on commercial technology development. That issue aside, procurement and demonstration projects have been less effective when they have been done without the guidance of an agency mission. Governments can strengthen market forces in other ways: in Japan, government-financed enterprises buy computers and robots and lease them to end users. The result? Guaranteed markets for the manufacturers, and reduced risk for the customers, who can turn back the equipment if they find it unsatisfactory. Finally, government regulations can also pull technologies into the market, sometimes with good results. Federal fuel economy standards created incentives for American automakers to improve their capabilities in engineering and producing small cam. Technical standards (e.g., for computer languages) can be an important spur to technology diffusion. 3 As this discussion suggests, market pull policies create vexing dilemmas for governments. Policies to support the adoption of publicly-funded R&D results are an essential component of a government effort to develop technology. But insofar as these incentives for technology adaption target specific applications, policymakers are placed in the position of trying to forecast the course of technological evolution and anticipate the commercial market. This is a task they have done poorly in the past. Interactions Among Industry, Universities, and Government A final principle, again emerging from the postwar history of high technology: collaborative interactions among universities and industry speed technological advance, particularly when supplemented by government R&D support and procurement. Coupling between industry and universities played a major role in the development of computing during the 1940s. The first practical electronic machine was built at the University of Pennsylvania. After the war, many of the key scientists and engineers left university and government laboratories to staff fledgling computer manufacturers like Univac. In the late 1970a, genetic engineering and biotechnology--supported in universities and the laboratories of the National Institutes of Health with Federal dollars--moved rapidly into the private sector, aided by abundant infusions of venture capital. Today, the United States remains well ahead of Japan and Europe in biotechnology. At this stage, coupling among universities, industry, and government makes sense for HTS: much of the research remains well-suited to academic settings; firms in many industries want a window on the technology; Federal agencies are already putting money in. %ee lntefmtio.ual Competft&m k 8arv&Iu (#WW@On, N: OffiW Of Technology Aaweamen& ~ 19s7). pp. 315-317, Airline -tiomih by MrMing p~ff competiticm, fOr@ carriers tO other insane d Mer@fa@ their uervicea. Thy vied to offer truvelem the lateet equipment in order to compete cm apeed uncl comfort, buying planes from Boeing, LocldwL and McDonnelWouglaa. These manufacturers, in turn, developed new mxlds. l A Federal technology agency. Strategy 3 that would substantially raise priorities for considers proposals for altering governengineering research; and direct support ment responsibilities for science and techfor civilian, commercial technologies. The nologya subject that goes far beyond the second two hold more promise than the particular needs of HTS. OTA analyzes first. Of course, such changes would take three variants: a cabinet-level science and time, meaning that they have little to offer technology agency: institutional changes in terms of the immediate needs in HTS.
PAGE 132
- 127 Each of the strategies has advantages and disadvantages. Strategy 1 is in place and working; the initial Federal response to HTS illustrates the considerable strength of the traditional approach. Mission agencies moved quickly, funneling millions of dollars into HTS R&Din a matter of months. The scientific community moved even faster, with large numbers of skilled professionals shifting into HTS from related fields. The magnitude of the response reflects the sheer size of the pool of scientific expertise in the United Statesitself a major source of advantage. Several weaknesses are apparent in the U.S. response. First, the universities are having a hard time competing for funds; DOE laboratories are getting roughly as much money for HTS research as NSF has for all the Nations universities. The second weakness: almost complete reliance on mission agencies to support HTS R&D. As a result, not enough R&D money flows to non-defense industrieswhich might not be a problem, were American firms pursuing HTS as aggressively as Japanese firms. Neither DOE nor the Department of Defense (DoD) can be expected to provide broad support for industrial R&D in HTS. HTS technologies stemming from DoD R&D may eventually find their way into the marketplace, but time lags that made little difference in the 1950s and 1960s, when the United States dominated the technological frontier, can be fatal in the 1980s. OTAs analysis suggests that continuing along the lines of Strategy 1the current approach will more than likely leave the United States behind in superconductivity. In mid-1988, the U.S. position in HTS looked like a strong one. But this is because HTS remains largely a matter for scientific inquiry. With progress toward applications, the picture will change. At best, a lead in science creates small advantages, often fleeting. The real contest will be over applications engineering and manufacturingwhere Japan excels, and where proprietary technology, much of it developed in industry, will make the difference. Strategy Z an aggressive Federal response, would, first of all, assign NSF a greater role in sponsoring R&D. Although basic research in HTS does not seem underfunded, more money for the Foundationperhaps $20 million over a 5-year period, specifically for HTSwould guard against missed opportunities in relatively fundamental work. With enough funding available for its research center programs, NSF would also be in a position to support one or more proposals for centers dedicated to HTS. As the second step in a more aggressive strategy, the Federal Government could share in some of the costs of R&D conducted by industry consortia. This is the simplest and quickest way to steer more resources into applied research and generic technology development. Government could direct public funds to R&D industry views with favor, but where the benefits would be difficult for individual firms to capture. A working group on HTS would be assembled to help carry out this second strategy, and in particular to make decisions on cost-sharing. The aggressive response approach requires agreement on an R&D agendaa consensus that should not leave out the universities or the national laboratories, even though industrys view of commercial needs would have to come first. The primary task for the working group: deciding which R&D consortia receive Federal dollarsa sticky issue, one involving decisions going beyond the scientific merits of alternative projects. This strategy skirts the picking winners problem raised by Federal subsidies for private R&D by, in effect, allocating government resources to the highest bidders, That is, among proposed R&D consortiaall of which were technically qualifiedfunds would go to those whose members were willing to make the longest term financial commitment and self-finance the highest fraction of total costs. These criteria would allocate government financing to joint R&D ventures with the greatest expected payoffs over the medium to long term. The aggressive strategy for commercializing HTS could substantially improve prospects for rapid commercialization of HTS in the United States at relatively modest cost.
PAGE 133
128 Under Strategy 3, OTA addresses prospects for a Federal agency charged with supporting commercial technologies, A perennial issue in U.S. science and technology policy, many alternatives have been proposed over the years. The possibilities range from a small, independent agency with a budget of less than $100 million, to a cabinet-level Department of Science and Technology pulling together some (though hardly all) of the R&D activities of existing agencies, The cabinet-level Department of Science and Technology came forward once more in 1985 as the lead recommendation of the Young Commission on competitiveness, but found no more support than in the past. Alternatively, is it possible to envision a smaller Federal agency with industrial technology as its mission? Once again, proposals have been commone.g., for a national technology foundation, paralleling NSFs role in support of science, or a civilian version of DARPA. The latter has attracted particular attention, given DARPAs enviable reputationa small group of creative people, with the judgment and experience to seek out and support the best ideas. But DARPA has a mission, and a critical onesupport of long-term R&D with potentially big payoffs in military systems. Lack of a comparable mission is the potential Achilles heel for a civilian technology agency. All such proposals face a common problem: providing money for industrial R&D, in the name of commercialization and competitiveness, without a well-understood and widelyaccepted mission, (Competitiveness is a notoriously slippery conceptmore so than national defense or health.) Lacking such a mission to lend discipline to the process of setting priorities and making funding decisions, a Federal technology agency could easily end up subsidizing marginal projects, The more of its funds such an agency channeled to industry, the deeper the possible pitfalls. Direct funding of industrial R&D raises the specter of subsidies won by lobbying rather than merit. Yet if the technology is to be useful to industry, then much of it should be developed by industry, Dealing with the many and contentious issues posed by a Federal agency for commercial technology development would be difficult, although not necessarily impossible. If such an agency is to support R&Din the public interest, it will need to find ways of identifying what that public interest is, convincing potential critics that it has done so fairly, and that the results justify continuing support. Plainly, the three strategies analyzed in this chapter are not exclusive. They do represent differing views of the strengths and weaknesses of the U.S. approach to technology development. Those who believe that the fundamental strengths of the U.S. system remain intact feel that industry will be able to commercialize HTS when the time is right. Those advocating a more aggressive policy stress the dangers of a business-as-usual mentality, given the surprising speed with which U.S. industry has lost its earlier advantages in high technology. The underlying worries over loss of competitive advantage lead those who would favor the third strategy, or something like it, to argue that the United States needs to thoroughly overhaul its approach to technology policy. STRATEGY 1: FLEXIBLE RESPONSE The Current Approach This strategy presumes that the existing policy framework is appropriate and sufficient for supporting HTS. 1 To those who advocate this For general background, see A.H. Teich and J.H. Pace, Science and Technology in the USA (Essex, UK: Longman, 1986); also H. Ergas, Does Technology Policy Matter? Technology approach, a major departure from the current course would be prematureat least during the early stages of HTS. With a good deal more basic research required to overcome the technical obstacles posed by the new materials, the and Global Industry: Companies in the World Economy, B.R, Guile and H. Brooks (eds.) (Washington, DC: National Academy Press, 1987], p. 191.
PAGE 134
129 Presidents initiative (box B, ch. 2), along with other executive branch actions, will provide a sound basis for industry to commercialize HTS when the time comes, And, so the argument goes, if the pace quickens, or foreign competition intensifies, there will be ample opportunity to agree on a stepped-up response. This is the de facto U.S. strategy. The Federal Government is following traditional channels, relying on existing institutional arrangements, and avoiding direct support for commercial technology development. A continuation of this approach (indeed, almost any approach) will mean: l l l Heavy ongoing funding for defense applications of HTS. The Strategic Defense Initiative (SDI) will continue providing a good deal of the money, and DARPA will probably continue to have a prominent place as well. Both industry and universities would get research moneysome of the the latter through DoDs University Research Initiativebut within half a dozen years, aerospace firms and military systems houses would probably be conducting the bulk of DoD-sponsored superconductivity R&D. Although DARPA has stated that the results of its processing contracts will remain unclassified, such a policy will be subject to change, depending on outcomes. As R&D moves on to defense-specific applications, classified programs may become common. DOE and the National Aeronautics and Space Administration (NASA) would pursue their own mission-oriented projects, with most of the Energy Departments money going to the national laboratories. Much of DOEs support will probably go for R&D and demonstration projects directed at electric power applications. DOE and NASA might pick up a few projects of interest to DoD. NSF would continue to fund HTS in the universities, with some of the Foundations support going to individual investigators and some to research centers. Other ongoing shifts in U.S. technology policy would proceed along lines suggested in chapter 4: l l l l The executive branch will continue its efforts to open up the national laboratories, as well as to strengthen university-industry relationships and stimulate technology development and transfer through such initiatives as NSFs Engineering Research Centers (ERCs) and agency Small Business Innovation Research (SBIR) programs. The Administration would also continue to press for stronger intellectual property protection, both at home (process patents) and overseas (negotiations with foreign governments aimed at stronger laws and tougher enforcement]. Some State governments would channel support, direct and indirect, to HTS as part of technology-based economic development programs, Venture-financed companies dedicated to HTS would continue to emerge. Private firms, both new and established, would negotiate collaborative R&D arrangements, nationally and perhaps internationally. Over the next several years, the United States will continue to take the course outlined above. Will such a response, by itself, be adequate? The following analysis indicates that it will not. Strengths and Weaknesses In many respects, the current approach to HTS illustrates the great strength of the U.S. system of technology development. Although superconductivity had become something of a scientific backwater by the mid-1970s, NSF for years supported people like Paul Chu at the University of Houston (an institution much like a hundred others below the top ranks in terms of research funding or prestige), and at least a few large U.S. corporations maintained small superconductivity research programs. Moreover, when HTS broke, American scientists could quickly take advantage of facilities ranging from neutron scattering equipment to the National Magnet Laboratory at MITfacilities already in place, the result of years of Federal
PAGE 135
130 funding. Agencies with their own laboratories DoD, DOE, NASAbegan new R&D internally, while contracting out other work. U.S. scientists not only responded quickly, but in large numbers, as measured by the flood of proposals to NSF, and papers published in professional journals and delivered at scientific meetings. This response reflects the sheer size of the pool of scientific and technical expertise in the United States-a notable strength. It also reflects the flexibility of the U.S. R&D system. NSF was perhaps the most agile of the Federal agencies, moving quickly to provide funds largely redirectedto individual research groups and to the Materials Research Laboratories (MRLs). NSF-funded investigators working in related areas were able to shift their attention immediately to HTS, because of the flexibility of Foundation grants. Scientists with DoD and DOE contracts or grants were also able to move quickly. In addition, DoD redirected millions of dollars in a few months, as various defense agencies exercised their much-valued fiscal autonomy. Each went its own way, with a resulting diversity of technical approaches that is probably healthy overall. Likewise in DOE, laboratories competed to stake their claim in the newly discovered territory of HTS, resulting in an aggressive, if somewhat fragmented, effort, Interagency coordination, though largely informal, has been relatively effective: program managers and contract monitors working in superconductivity know one another and feel a shared sense of responsibility (box M, ch. 4). Like government, venture capitalists reacted quickly to the new opportunities, lining up technical experts, many of them university faculty members, and quickly investing nearly $20 million in entrepreneurial startups (box G, ch. 3). These startups are just one illustration of close industry-university links in HTSanother significant asset of the U.S. system. The October stock market plunge led at least one HTS startup to cancel plans for a public offering but, overall, availability of capital has not been a major constraint. In sum, U.S. R&D in HTS will continue to benefit from the unparalleled breadth, depth, flexibility, and diversity of the Nations research system. It is easy to see why many people feel the current U.S. response to HTS is sufficient at least for now. But weaknesses have also begun to surface, and others will probably appear over the next year or two. Funding for basic research in HTS could be a problem, though probably not a serious one. As the ongoing flood of technical papers indicates, the scientific effort remains broad and intense. (At the March 1988 meeting of the American Physical Society, more than 600 of 3,500 papers presented dealt with superconductivity. Most were written by scientists based in the United States.) There are no obvious gaps in fundamental science: people somewhere are pursuing almost every possibility imaginable. More than likely, the ongoing university efforts will suffice to train enough people for industrys eventual manpower needs. On the other hand, basic research in HTS may already have reached its peak, More sophisticated laboratory equipment will be needed to keep up in the future (e.g., for work on thin films), and costs will rise, Some of the investigators who used existing grants and contracts to move into HTS will have trouble getting new money to continue; they will have to show real promise, rather than routine results, to qualify for ongoing support. And even if there is enough money in total for basic research in HTS, the money might be better spent (box R). As noted earlier, the national laboratories have an HTS budget in 1988 roughly equal to that of NSF, an allocation that seems out of proportion, given the trouble universities have had getting funds. Even if mission-oriented R&D funded by DOE and DoD were to transfer to the commercial marketplace, it would not do so immediately, The time lags made little difference in the 1950s and 1960s, when American companies were far ahead in technology. Today, the United States cannot afford to wait while know-how diffuses at its own pace from Federal laboratories to the private sector.
PAGE 136
131 Box R.National Laboratories or Universities? Postwar U.S. science and technology policy has relied heavily on the university system, with the national laboratories concentrating on mission-oriented R&D. Since the budget cutbacks of the early 1970s, many laboratories have sought to broaden their R&Da trend that, arguably, has already cut into the share of Federal resources flowing to the universities. With DOE efforts to move beyond the big science role inherited from the Atomic Energy Commission (i.e., big physics) into fields such as mapping the human genome, it may not be much of an exaggeration to say that the Energy Department seems to be trying to become a general-purpose science agency. This expansion raises issues of balance. Compared to the DOE laboratories, the university system is more open and supports a more diverse set of R&D projects. The universities operate with proven systems of self-governance, intellectual autonomy, and quality control through peer review. Bad science cannot hide for too long. Perhaps most important, industrys need for trained people gives the universities a special claim on Federal R&D funds. No other set of research institutions trains scientists and engineers for industry in large numbers. When these people move to the private sector, they take the latest knowledge with them. While the laboratories performance in technology transfer has certainly improved over the five years since the Packard Commission report (ch. 2), some of the remaining problems concern the amount of autonomy that should be given to mission-oriented facilities operated under contract to the Government. In addition, the laboratories are poorly positioned to deal with problems related to manufacturing. Laboratory personnel, unfamiliar with industry and the marketplace, often ignore the need to address processing early enough in development. University engineering departments have also fallen into this trap, but seem to be doing more to dig themselves out. Moreover, universities have been willing and able to work with industry; the DOE laboratories, until recently, have shown few signs of the flexibility needed to adapt their ways to industrys needs. At present, some people in industry view the laboratories with suspicionand also as competitors for Federal R&D dollars. 1 The laboratories can claim advantages over the university systemincluding a capacity for interdisciplinary research, sophisticated facilities, and experience with large-scale projects. But, as emphasized in chapter 4, even if policies put in place to change the laboratory system prove successful, the process will take time. Neither the universities nor the national laboratories should or could become centerpieces of a strategy for commercializing HTS. Their strengths lie elsewhere. Y%, for example, National Labs Struggle With Technology Transfer, New Technology Week, Sept. 28, 1987; Allan S. Gelb (Director, Marlow Industries, Inc.), testimony to the House Subcommittee on Energy Research and Development, Committee on Science, Space, and Technology, Oct. 20, 1987. Granted, the mission agencies are trying to address national concerns over competitiveness in their pursuit of HTSe.g., through DARPAs processing R&D initiative. However, the near-term focus of the DARPA processing program in HTS may not take advantage of the agencys own strengthsfunding of visionary research. Finally, DARPA must ultimately serve DoD missions, which means that when civilian and military needs diverge, commercialization will recede as an objective. Only if the agency can link its HTS R&D with military objectives will it find continued support within the Pentagon. Other problems aside, the program is relatively smallonly about $18 million in 1988and it was not fully underway as this report went to press, In short, the lack of more direct mechanisms for Federal support of commercially oriented R&D has become a weakness in the U.S. approach to technology development. The issue is not overall funding levels for HTS. The issue is the allocation of those funds: missionoriented R&D does not provide enough support for commercialization to ensure that American firms will be able to keep up in HTS. The problem is particularly acute because of the wait-and-see attitude in much of American industry. As described in chapter 3, the Japa-
PAGE 137
132 nese are putting more effort into exploring applications. Furthermore, Japanese companies, with their strengths in engineering and manufacturing, would probably be able to catch up even if U.S. firms were first to reach the market with innovative products. In sum, the current U.S. response to HTS displays the strengths and weaknesses that have characterized the performance of American companies in high technology. The U.S. effort looks formidable in the middle of 1988, but that is to be expected: the challenges so far have been largely matters for the research laboratory. If there is a surprise, it is that the Japanesenot known for innovation in sciencehave already posted such a strong showing. As HTS moves toward applications, science will recede in importance. Basic research results, by their nature, will diffuse rapidly, providing little in the way of national advantage in commercialization. (Patent coverage sufficiently broad and strong to lock up a critical class of HTS materials seems unlikely.) Rather, the critical technological advantages are likely to reside in proprietary know-how associated with processing and fabrication techniques, and design-manufacturing relationships precisely where Japanese companies have demonstrated an advantage over many of their American counterparts. A continued response along the lines of Strategy 1 is quite likely to fall short: the widely expressed fear that Americans will win in science, while the Japanese take the commercial markets could come true. Support for science and for military technologythe essence of this strategyserved the United States well from 1950 to the middle 1970s. But the lesson of the past 15 years is clear: in a world of increasingly effective national competitors, these two levers no longer suffice. STRATEGY 2: AN AGGRESSIVE RESPONSE TO HTS Three primary features distinguish this second 1. 2. 3. strategy from the current approach: A larger role for NSF, both in funding individual research at universities, and through the establishment of one or more university centers in superconductivity. Federal cost-sharing of long-term, high-risk R&D planned and conducted by industry consortia. A working group on commercialization of HTS charged with helping shape consensus on an R&D agenda, and making decisions on Federal cost-sharing. This strategy preserves the strengths of the traditional U.S. approach to technology development, while compensating for the weaknesses brought out by stronger international competition i.e., lack of breadth in industrial R&D, and heavy reliance on mission agencies for Federal support. Step One: A Larger Role for NSF The initial step toward a more aggressive response to HTS should be straightforward: give the National Science Foundation more money for university research. For reasons outlined earlier in this report, a dollar spent by NSF should contribute more to commercial development, on the average, than a dollar spent by DoD or DOE. In view of this, NSFs existing 15 percent share of the total Federal R&D budget for HTS is too small. There are two complementary ways for Congress to expand NSFs role, An otherwise unrestricted appropriation, earmarked for HTS money that the Foundation could spend on superconductivity as it sees fitwould permit NSF to fund some of the highly rated HTS proposals that it is currently forced to turn away. As noted above, OTA has found no evidence of serious underfunding in basic research on
PAGE 138
133 superconductivity. Nonetheless, with an NSF research budget that has been flat in real terms for several years, funds for condensed matter physics were cut back during fiscal 1988. 2 (Ch. 4, which discussed this and other symptoms of the pressure on the NSF budget at some length, included a number of policy options addressing the general problem.) As part of Strategy 2, Congress might consider appropriating, say, $20 million (in additional new money) for NSF for the 5-year period beginning in fiscal 1989, specifying that the funds go for HTS research. Such a step would help ensure that the basic science underlying superconductivity gets adequate support, without cutting into budgets for NSF-sponsored R&D in other fields. In addition (or as an alternative), Congress could authorize and appropriate funds to NSF specifically for one or more university centers dedicated to HTS research, To give NSF maximum flexibility, the centers could be established under one of several existing programs, or through a new program altogether. While none of the Foundations existing or proposed center programs (discussed inch. 4 see especially box N) seems ideally suited to the needs in HTS, an ERC comes closest. Although a number of the MRLs have good experimental facilities, and active research in superconductivity, industry involvement has rarely been a major goal. Nor are the proposed S&T centers, although emphasizing multidisciplinary research, likely to focus as strongly on industry interactions as the ERCs. While the program is relatively new, and as yet few of the ERCs have themselves demonstrated close working relationships with the private sector, their focus means the ERC program fits the needs in HTS more closely than other candi20vercom mit ments by the Division of Materials Research during 1987, in the expectation of a substantial budget increase, forced the cuts. Not only NSF, but DoD and DOE maintain that a considerable number of highly rated research proposals are going unfunded. The problem is not a new one, particularly for NSF. But the problem has gotten worse, and program managers understandably feel uncomfortable trying to draw lines between proposals that are almost indistinguishable in quality. dates. (None of the existing ERCs has a research agenda embracing HTS.) Several bills introduced in the 100th Congress e.g., H.R. 3048 and H.R. 3217would instruct NSF to establish a program for interdisciplinary National Superconductivit y Research Centers. Would a new center program for HTS one with the explicit mission of building a strong technology base for commercialization, and one with teeth in the requirements for multidisciplinary work and industry involvement do more for HTS than funding for one or more ERCs or S&T centers? The answer has to be yes, if Congress appropriates the money and if NSF moves relatively quickly. (The Foundation would have to solicit new proposals, while it already has proposals in hand for S&T centers on superconductivity,) In sum, congressional funding for several (say, one to three) new multidisciplinar y centers in HTS could represent a modest but important step. It would not be realistic to expect such a measure to expedite commercialization dramatically. University centers, even at major schools, would no doubt remain relatively small in scale and scope. Sums of $10 million to $20 million annually (perhaps $5 million, at most, per center) are about the maximum that make sense for an NSF center programthe Foundation does not do business on a scale much above this, Such centers would serve primarily as a source of new ideas and trained peoplean important contribution to commercialization, not to be undervalued, Step Two: Industrial Consortia with Federal Cost-Sharing As another element in a more aggressive strategy, Congress could direct the Administration to partially offset the costs of joint industrial investment in long-term, high-risk R&D. Such a policydesigned to address the gaps in HTS R&D in U.S. industrywould be based on two premises discussed at length in this report, First, just as the Federal Government supports basic research, it must bear part of the burden
PAGE 139
134 Photo credit: Superconductive Components, Inc. Pellets of HTS ceramic material of exploring risky, radical industrial technologies, which often provide large public benefits but only small private returns. Second, although DoD has borne much of this burden in the past, the growing specialization of defense technologies, and continuing pressure to meet the immediate needs of the services, mean that the United States now lacks a consistent champion for major new technologies with potential impacts on the civilian side of the economy. Partial Federal support for one or more industry consortiaestablishethe, as discussed in box S, to share R&D costsis a simple and workable policy to address this problem. Financing some fraction of joint industry efforts with Federal funds (or tax expenditures) would pull more resources into applied research and generic technology development, and raise the overall level of R&D investment. The approach would direct public funds into areas that industry itself thinks will have the highest payoffs, but where the benefits would be difficult for individual firms to capture. In some respects, the approach envisioned for HTS resembles that of Sematech, the microelectronics industrys new R&D consortium (to which more than a dozen firms have pledged 1 percent of their revenues). While providing substantial funding, the Federal Government would be a largely silent partner, with an R&D agenda put together and managed by member firms. And as with Sematech, there would be explicit linkages with universities and DOE national laboratories, so as to tie publicly-funded basic research to the joint R&D. The differences between Sematech and the HTS R&D consortia envisioned for Strategy 2 are perhaps more important. First, the semiconductor industry itself proposed and fought for a Federally-supported venture. The companies likely to be involved in HTS consortia have made no such effort, and are not likely to; thus, the job of initiating and organizing such programs would fall in part on Government. Second, DoD, through DARPA, serves as the financial channel to Sematech. A report by a Defense Science Board Task Force argued that the industrys troubles imperiled national securityone reason for DoDs oversight role. Moreover, DoD was perhaps alone among Federal agencies in having the technical expertise to monitor microelectronics R&D. Even if the DoD arrangement proves satisfactory in the case of Sematech, it holds small promise as a model for HTS. DoDas well as DOE and NSF could be involved with HTS consortia, but none of these agencies should oversee them, lest their purposes be subordinated to ongoing agency missions. This point is discussed further in the following section, which deals with institutional mechanisms. Finally, Sematech has a focused R&D agenda, stemming from a consensus within the industry that manufacturing technology has been a major source of competitive difficulty. Rather than a single, well-defined focus, HTS lends itself to multiple agendas, different sets of participants, Three among many possible candidates: l Electric Utility Applications.Utilities normally make highly conservative investment decisions. They are unlikely to adopt a new technology until sure it will work reliably for many years. Thus, there may be a useful role for Government in accelerating the development of the engineering database and field service experience in HTS through support for cooperative projects (which could involve DOE laboratories),
PAGE 140
135 Box S.Collaborative R&D 1 The breakthroughs in superconductivity brought forth many proposals echoing the theme of strength through collaboration. The Presidents initiative urged Federal agencies to cooperate with universities and the private sector. Legislation has been introduced in Congress with similar objectives. In part, these calls for collaboration represent a response to rising R&D costs and the loss of U.S. technological advantage. To some extent, they stem from a misapprehension of the sources of Japans success (the myth of cooperation discussed in ch. 3). Although joint research is nothing new, the past decade has seen a steady growth in U.S. R&D consortia, and a marked change in research focus. Firms that once cooperated only on matters such as technical standards have increasingly banded together to undertake pre-competitive R&Dprojects on new technologies with direct commercial relevance. The best-knownthe joint venture Microelectronics and Computer Technology Corp. (MCC)has begun exploratory work on HTS. The economics of joint R&D hold considerable appeal. By pooling resources and avoiding duplication, cooperative research increases the potential leverage of each firms R&D budget, while limiting the costs to any one firm of a failed project. Joint efforts also enable participants to monitor developments in a technical field without developing a full capability in-house. While the benefits of cooperative research are sizable, so are the limitations. Most important, joint endeavors cannot substitute for in-house R&D efforts, only complement them. The participating firms must absorb the results, and transform them into commercially relevant products or processessomething that requires a sophisticated and independent internal effort. For such reasons, government efforts to encourage cooperative research programs in industries where firms pursue little or no R&D on their own have seldom succeeded. 2 A second limitation may prove more serious. Ideally, cooperation in research should help lengthen project time horizons; but many of the participants in cooperative ventures shift R&D strategies in two ways: 1) by focusing their internal research on still shorter-term work; and 2) by seeking to move the cooperatives agenda away from basic research and toward more applied projects. The trend has been evident in MCC, which recently restructured its largest programin advanced computer architecturesto emphasize more immediate paybacks. Collaborative efforts involving universities (or government laboratories) have an easier time focusing on longterm research, often with little interference from participating firms, because the firms are not seeking specific R&D results so much as access to skilled graduates and faculty expertise. Of course, the financial commitments are generally much smaller than those for participation in a joint venture such as MCC (and may be viewed in part by the firm as good corporate citizenship). While cooperative R&D programs housed in universities help maintain a strong technological infrastructure, they should not be viewed as engines of commercialization. Finally, the sheer difficulty of organizing and managing a collaborative research venture creates its own set of limits. Fundamental issuesreaching agreement on an R&D agenda, finding ways to share technologies and business information, controlling costs and determining intellectual property rightscan pose enormous obstacles, particularly when the collaborators are also competitors. These are among the reasons that cooperative research accounted for only $1.6 billion of the more than $50 billion spent by American industry on R&D in 1985. Of this total, 85 percent ($1.4 billion) went to support R&D cooperatives in the communications, gas, and electric utility industries, whose members do not compete directly. It is no surprise to find that industry leaders rarely put cooperative R&D very high on the list of steps needed for rebuilding U.S. competitiveness. 3 Joint R&D has a role to playin HTS and in solving the more general problems visible in the U.S. technology basebut that role will inevitably be circumscribed by tensions between competition and cooperation among the participating companies. Firms seek proprietary technologies in order to compete with one another. Cooperation between firms in the same business can only go so far; cooperation between firms having supplier-customer relationships, or those in different industries with common R&D goals, holds more promise. ISee Collaborative Research: An Assessment of Its Potential Role in the Development of High Temperature Superconductivity, prepared for OTA by D.C. Mowery under contract No. H36730, January 1988. zNSFS Industry/University Cooperative Research Center program (IUCR, box O, ch. 4) was initially designed to substitute for in-house R&D in technologically moribund industries. An NSF evaluation found that Companies with little research background, such as the utilities and furniture companies, are traditionally conservative with respect to new technology and depend on their suppliers for whatever changes they adopt. An Analysis of the National Science Foundations University-Zndustry Cooperative Research Centers Experiment (Washington, DC: National Science Foundation, 1979). IUCR programs have gone on to more success in other industries. sFor the results of a recent survey of corporate executives, see C The Role of Science and Technology in Economic Competitiveness, final report prepared by the National Governors Association and The Conference Board for the National Science Foundation, September 1987, pp. 20-30. The R&D spending figures for cooperatives come from P.F. Smidt, US Industrial Cooperation in R& D, remarks at the Annual ESPRIT Conference, Brussels, Belgium, Sept. 25, 1985.
PAGE 141
136 l l HTS Magnets. Strong magnetic fields, by dampening thermal fluctuations, can give purer, more uniform crystals of silicon and other semiconductor materials, making this a natural candidate for superconducting magnets. R&D directed at HTS-based magnets could, at the same time, help move HTS out of the research laboratory and lessen U.S. dependence on foreign sources of semiconductor wafers. X-ray lithography using compact synchrotronsa promising candidate for making next-generation integrated circuits likewise could serve as a spur for HTS R&D while filling a chink in U.S. microelectronics capabilities. Superconducting Computer Components. HTS interconnects may help improve the performance of computers. Hybrid semiconducting/superconducting electronics may also prove viable. Collaborative projects on prototype circuitry, with subsequent Federal purchases of very high-performance machines if the technology panned out, might help speed commercial development of HTS, while at the same time preserving the U.S. lead in high-end machines. Such a program would have a substantial basic research component, but involve manufacturability and applications issues as the technology began to mature. Whatever the substantive agenda, the private sector should take the lead, insofar as practical, with Government participation limited to that necessary to achieve the Government objectivesleveraging critical R&D, filling gaps in the technology base, lengthening time horizons. Federal cost-sharing justifies such procedural rules as these: l Government participation should be conditional on significant investment by industrysay 50 percent or more. (The Sematech formula percent from both industry and DoD, and 20 percent from State and local Governmentprovides an alternative.) The less of their own money companies contribute, the greater the risk of R&D that strays from marketplace needs. So long as Government funding decisions can be limited to choosing among alternative apl l l l l preaches, all of which industry is prepared to back, some of the problems of picking winners can be skirted. By also requiring, as a minimum, a 3-year financial commitment on the part of consortium members, the Government will have added assurance that public funds will help support mediumto long-term R&D. (MCC requires only a l-year commitment, down from 3 years initially; this change is both cause and consequence of pressures for tangible early R&D results.) Companies should be either in or out. MCC, which permits members to join projects selectively, has found itself trying to wall off some of its work to prevent R&D results from leaking to companies that have not joined a particular project. Entrance requirements should be transparent, lest the consortium become a smokescreen for anti-competitive behavior; any eligible U.S. firm should be able to join (those without majority U.S. ownership might reasonably be barred). people transfer technologies most effectively. To see that knowledge flows out of the consortium, employees of the member companies should be heavily represented among the R&D staff. It seems reasonable to insist that half or more of a consortiums staff come from member companies (rather than new hiring). Assignments might be temporary, but should be long enough perhaps 6 months as a minimumfor meaningful contributions to R&D, and for learning purposes. Furthermore, member firms should be encouraged to send their best people. MCC dealt with this by retainingand exercising liberallythe right to reject employees sent by shareholders (initially, it turned down 95 percent). 3 University involvement can also help attract the best industry scientists. No less important if the consortium is to affect commercialization, member firms must conduct ongoing complementary R&D sB.R. Inman, Collaborative Research and Development, Commercializing SD] Technologies, S. Nozette and R.L, Kuhn (eds.) (New York: Praeger, 1987), p. 65.
PAGE 142
of their own. MCC encourages shadow research by members, paralleling the joint ventures work. This ups the ante for members; Digital Equipment Corp. spends half again its investment in MCC seeking ways to use the consortiums resultsa level of commitment that has, however, been rare. Parallel efforts will be particularly important for firms with little or no experience in superconductivity. They will have more learning to do than companies with backgrounds in, say, LTS. l If one purpose of Federal funding is to stimulate visionary R&D, it may make sense to discourage too much publicity, By reducing the pressurespolitical and otherfor short-run success stories, the chance of success stories over the longer term should go up. A long-term orientation also suggests that a consortiums work stop well short of fullscale commercial development efforts i.e., at the prototype stage, leaving further development to the members own efforts. Step 3: A Working Group on Commercialization The final step in the aggressive strategy would be to establish a working group of experts drawn from industry, universities, and governmentwith a limited mandate to promote the new technology. Such a groupwith a lifetime fixed at, say, 10 yearscould serve a number of important functions. The first is fact-finding and analysis. Currently, no public or private body has a continuing responsibility to provide authoritative policy guidance concerning such questions as: What problems do we need to start on now to assure rapid commercialization? How much money will it take? Are some HTS R&D areas getting too much money? Which areas are not getting enough? While such questions never have definitive answers, the first step toward good decisionsgiven that the Federal Government will have to make decisions in any case is to understand what is going on in both government and industry, here and in other countries. The problem is not inadequate coordination. Rather, the problem is that no one has the task of drawing even a crude map of the road to commercialization, and setting the necessary priorities along the way. To do so will require solid and timely analysis, on a continuing basis. (The Presidents Wise Mens advisory committee on HTS will evidently prepare a one-time report, rather than provide ongoing policy guidance,) While the working groups responsibilities would involve decisions on Federal cost-sharing in response to proposals from private sector consortia, it might first have to engage in consortium-building and facilitation. In contrast to Sematech, HTS consortia are not likely to organize themselves spontaneously, at least until guidelines for Federal cost-sharing have been set down. The working group might be able to play a match-making role, helping bring together companies, universities, and government laboratories, and aiding them in reaching consensus on research needsa function that could continue even after the joint R&D effort was underway; as the experience of Sematech and many other joint ventures demonstrates, conflict will be inherent in any consortium of independent firms, The working groups ability to get things done will flow in part from its power to allocate Federal resources. Decisions on who is to get Government money will hinge not only on technical questions, but on economic judgments. Technical evaluations of competing projects can rely on the tried and tested approach of external review by recognized experts without a stake in the outcome, Evaluating the economic meritsi.e., likely impacts on commercializationwould be more difficult, but the problem can be sidestepped to considerable extent by using procedural rules to allocate public funds. First, a consortiums proposed R&D agenda would have to meet minimum criteria, which the working group would set, based not only on technical merit, but, as discussed above, on rules for participation, and provisions for member funding and R&D time horizons, Then, among the qualifying proposals, funds would go to the highest bidders as measured by length of time commitment par-
PAGE 143
138 ticipants in the consortium were willing to make, and the fraction of total costs they were willing to self-finance. (A third criterionthe investments of members in complementary internal R&Dis also relevant, but it would be difficult to determine what was complementary and what was not.) These measures should help to allocate government resources to joint R&D projects with the longest term payoffs, and the greatest value to industry. If the working group makes decisions on funding, the issue of administering such a program remains. As an ad hoc body outside the ordinary apparatus of Government, the working group would need to look to an existing agency for help with staffing and managing its responsibilities. This poses a dilemma. As discussed above, the working group is not simply an advisory body, but a center for decisionmaking on commercialization policies. Yet only a minority of its members would be Federal employees, it would go out of existence after some period of years, and the intent is not only to complement the activities of existing agencies but, to considerable extent, to substitute for the agenciesto undertake tasks that they do not (and perhaps cannot). Attaching the working group, even for administrative convenience, to an existing agency could undermine its impact. Each of the three agencies heavily involved in funding HTS R&DDoD, DOE, and NSF has noteworthy strengths: experience in funding LTS research, technical competence, and the administrative tools needed for monitoring the expenditure of Government funds. But each has flaws as well: DoDs military mission will always come first; later if not sooner, commercial technology development will probably devolve into a secondary objective of consortia with Pentagon involvement. DOE has less experience with the private sector than DoD e.g., in managing extramural R&Dand a narrower base of technical expertise. For NSF, the assignment would be a substantial departure from the norm; the Foundation has limited ties to the private sector, and few employees with industrial experience. Its past attempts to foster applied research have not met with great success. Are there other possibilities? The Commerce Department is seldom seen as a technology agency. The Office of Science and Technology Policy (OSTP), as pointed out in the preceding chapter, has a small staff and is not set up to handle the kind of tasks the working group would need to pass along, (A Federal technology agency, as described in the final section of this chapter, might be well-suited; but even if Congress were to pass legislation creating such an agency, it would not be ready in time to serve the working group.) In the end, the best solution is probably to set up the working group as an ad hoc independent body, with a small staff of its own, and attach it to OSTP. As such, it should have the necessary qualities for promoting commercialization of HTS: the flexibility and substantive depth to learn by doing, tailoring its procedures to the special needs of HTS as these became apparent, The mission agencies will take care of their own needs in superconductivity. What is lacking is an organization to look after the broader national interest in commercializing this new technology. Federal support of joint privatesector R&D investments addresses the need. This is not the only alternativebox T summarizes othersbut it seems a promising one. STRATEGY 3: A FEDERAL TECHNOLOGY AGENCYTHREE ALTERNATIVES The Federal structure for science and techR&D. In 1950, after prolonged debate, Congress nology policy has changed little since the late passed the authorizing act for the National Sci1950s. Within DoD, the Office of Naval Reence Foundation. The same year saw major search set the post-1945 pattern for support of new legislation setting the National Institutes
PAGE 144
139 Box T.Aggressive Support for Commercialization: Other Possibilities Flagship Projects The symbolic and potential economic importance of HTS have led to calls for a flagship approach. With HTS already a symbol, among other things, of U.S.-Japan economic competition, advocates of the flagship alternative see virtue in Government initiatives that likewise have symbolic value. A flagship should rally industry, universities, and the public sector, building on the enthusiasm created in the media and galvanizing the Nations creative and entrepreneurial vigor in a race, not to the moon, but over the hurdles of some earthbound Olympicsan Olympics of science, technology, and competitiveness. This countrys strength is in meeting crisesthe Manhattan project, Sputniknot in incrementalism, say advocates of this approach. Political consensus, and a bold national effort, could pull superconductivity out of the laboratory and into the global marketplace. Visibility can be a strength as well as a potential source of weakness. As appealing as such images might seem, HTS is not the kind of technology that lends itself to a massive, concentrated effort. Although bills have been introduced dealing with magnetically levitated trains, public and political attitudes toward rail transportation would probably have to change a good deal before maglev would have broad appeal. Talk of an energy crisis evokes little response today. Superconducting computers will just be black boxes; no one much cares what is inside. Defense systems do not fill the bill either. There is another dilemma. A flagship has visibility. It must succeed, if only for such reasons. This forces technological conservatism on decisionmakers. Apollos achievements were in systems engineering and largescale project management, not in revolutionary technologies; space is no place for trying out unproven technologies. Pressures for successor pressures to avoid the appearance of failuremean safe choices by the managers of such projects. If the goal is Government support for long-term, risky technology development, the flagship approach has little to offer. DoD Processing R&D Processing will be vital in commercializing HTS, and several legislative proposals have made it a central element. For example, H.R. 3024 would authorize $400 million for a 5-year, DARPA-centered effort also involving DOE, NSF, and NBS. DARPA has left a deep imprint on U.S. high technology, most of all in computers. Chapter 4 discussed DARPAs current HTS processing initiativean effort that could fill an important gap in superconductivity R&D. Giving DARPA the lead in a more ambitious effort, as in H.R. 3024, might seem appropriate, On the other hand, programs aimed explicitly at (non-defense) commercialization fall well outside DARPAs historical mission and experience; in part because it seems to critics in the Pentagon too far removed from military needs, DARPAs HTS program has not found widespread support inside DoD. Congressional enthusiasmreflected in a direct appropriationmade the program possible, but if Congress loses interest, the program could fade away. Only if DARPA could link the R&D with military objectives would it get internal support in the face of budget pressures and competing DoD demands. In the past, such pressures have periodically led DARPA to abandon longer-term R&D and embrace more immediate military objectives. DARPAs transfer of the MRLs to NSF marked the first of these periods (box O, ch. 4). After the end of the Vietnam War, fatter R&D budgets enabled DARPA to move back toward long-term research. But in the early 1980s, the pendulum swung once again, with the Strategic Computing Program a prime case in point; DARPA has channeled much of the programs funds to military contractors, rather than the university laboratories responsible for most of its earlier successes in computing technology, and set program objectives that will appeal to the services. Ironically, over the last few years, DARPA has behaved much like American corporationsstressing projects with quick payoffs. Finally, there seems little question that major breakthroughs in HTS resulting from DoD-sponsored R&D would be classified, should the Pentagon feel that, as a practical matter, they could thereby be kept from the Soviet Union. The DOE Laboratories Although no one has formally proposed the designation of a lead laboratory for HTS, the suggestion has been in the air, There are at least 10 DOE laboratories with work of one sort or another underway in HTS. The chief argument for greater concentration and centralization is one of efficiency: with all the laboratories on the HTS bandwagon, duplication of effort will be hard to avoid. By giving a clear mandate to one, DOE should be better able to manage the division of labor. On the other hand, with HTS remaining primarily a matter of researchresearch with fuzzy objectivescentralization for its own sake has little to offer. The conventional management wisdom that basic research is cheap, the benefits of competition among scientists great, makes central control unnecessary and undesirable. Nor is there an obvious candidate for a lead laboratory in HTS.
PAGE 145
140 of Health on a course it has continued to follow. At the end of the 1950s, the Soviet Sputniks spurred another set of changes: the establishment of NASA and DARPA. NASA grew out of the National Advisory Committee for Aeronautics (NACA). Founded in 1915 to conduct research and testing, NACA remained small until World War II, when its staff grew to nearly 7,000 people. NASAs staff eventually reached five times that level, while its budget grew even faster (NASA contracted out much more of its work). DARPA, setup in 1958and originally given the mission of developing a U.S. space program, later passed to NASAquickly established itself as the home of long-range R&D within the Defense Department. Since this period, the Federal R&D budget has grown steadily, but organizational changes have been minor. In 1950, the Federal Government spent about $1 billion on R&D. Today, half a dozen Federal agencies each spend over $1 billion annually, and more than a dozen others spend lesser amounts. Given this growth in R&D spending, and the increasing concern over the Nations ability to utilize its technology effectively, many proposals to reorganize Federal science and technology functions have come before Congress during the 1980s. (This is nothing new: in 1913, during the debate preceding the formation of NACA, some of the opponents of a new organization for aeronautics research saw it as a stalking horse for a cabinet-level science department.) An Umbrella Agency for Science and Technology The more ambitious sounding proposals often call for a science and technology (S&T) agency %ee, for example, J.A. Shannon, The National Institutes of Health: Some Critical Years, 1955-1957, Science, Aug. 21, 1987, P. 865. On NACA and NASA, below, see F.W. Anderson, Jr., Orders of Magnitude:A History of NACA and NASA, 1915-1980, NASA SP-4403 (Washington, DC: National Aeronautics and Space Administration, 1981); and A. Roland, Model Research: The Nationzd Advisory Committee for Aeronautics, 1915-1958, vol. 1, NASA SP-4103 (Washington, DC: National Aeronautics and Space Administration, 1985); also Collaborative Research: An Assessment of Its Potential Role in the Development of High Temperature Superconductivity, prepared for OTA by D.C. Mowery under contract No. H36730, January 1988, pp. 29-34. to consolidate Federal R&D functions. Advocates of consolidation argue that an umbrella organization would lead to clearer priorities, less duplication, and greater efficiencyin a word, to better management. They point, for instance, to the more than 700 national laboratories, managed, often quite loosely, by many different agencies, and note the frequent criticism that the laboratory system has come under. In fact, calls for a Department of Science and Technology tend to be a bit misleading. Because the mission agencies control most of the Federal R&D budget, the resulting changes would necessarily be modest. When the Young Commission called for a cabinet-level S&T agency in 1985, it offered no guidance on how such a proposal might be implemented. 5 The problem is clear enough. Some 70 percent of Federal R&D goes for defense and space. Much of the rest pays for health-related research. It is hard to envision moving more than a few bits and pieces of DoDs current R&Dsay a billion dollars or sointo another part of Government. Moreover, since creating the Atomic Energy Commission in 1946, Congress has kept nuclear weapons research isolated; currently, nuclear weapons account for about half of DOEs R&D budget. The second largest R&D agency, Health and Human Services (DOE is third), operates a research armNIHwith a hundred-year tradition of excellence. Why risk disrupting organizations like NIH in the name of management efficiency? Even the strongest advocates of an S&T department acknowledge that consolidation could not go too far without harming the ability of agencies to manage R&D in support of their own missions. But without pulling much of the R&D that is currently the responsibility of these agencies under the new S&T umbrella, there Globa~ Competition: The New Reality, vol. I (Washington, DC: U.S. Government Printing Office, January 1985), p. 51. The Commission simply said that the department should include major civilian research and development agencies. For extensive discussion of proposals for an S&T agency, see the special issue of Technologyln Society, vol. 8, Nos. 1/2, 1986, entitled A Department of Science and Technology: In the National Interest?
PAGE 146
.. would be little left, It would be hard to take seriously a cabinet-level S&T agency that would oversee perhaps 10 percent of the Federal R&D budget. Given this dominance of R&D by the mission agencies, most of the legislative proposals for reorganization have had quite modest objectives. H.R. 2164, for example, is fairly typical. This billintroduced in the 100th Congress to create a Department of Science and Technologywould pull together NSF and the National Bureau of Standards (NBS), together with several smaller Commerce Department programs, while also also creating a National Bureau of Technology Transfer and an Advanced Research Projects Foundation. The lattercharged with supporting generic, industrial R&Drepresents one variant of a recurring proposala proposal that, according to the analysis below, has more in its favor than an umbrella agency for science and technology. Higher Priorities for Technology and Engineering Proposals for a technology agency that would stand alongside NSFperhaps called a National Technology Foundation (NTF)start with the premise that technology does not always depend on science. Indeed, development of new technologya goal-directed, problem-solving activitydiffers fundamentally from scientific research. Even where the interrelationships are close, as they often are in high technologies, the two activities depend on different kinds of people, with different skills and expertise. Science seeks understanding. Technology seeks satisfactory solutions to practical problems. Science looks to technology for toolscomputers to unravel the structure of DNA, or to guide powerful telescopes as they scan the heavens. By the same token, technology looks to science for tools: knowledge of DNA leads to new pharmaceutical products; theoretical insights into computer software now guide the design of hardware. U.S. problems in commercialization lie in technology, not science. And while scientists make their contributions to innovation and 141 Photo credjf Biomagnet/c Technologies /nc Brain scan using equipment incorporating superconducting quantum interference devices (SQUIDS) competitiveness, the engineering profession carries much of the burden (ch. 2). Raising priorities within Government for engineering researchwork directed at technology rather than sciencewould be a straightforward and positive step toward renewed competitiveness. NSF has made considerable progress at this in recent years, most notably through its ERC program, while amendments to the NSF charter have also given engineering more prominence. Moreover, the Foundations current director, Erich Bloch, who came to NSF from industry, has provided strong leadership for initiatives such as the ERCs. But Bloch will not be there forever. And NSFs fundamental job is the support of sciencescience for its own sake. The Foundation cannot tilt too far toward engineering without provoking a strong reaction from its primary, and well-organized constituencyuniversity scientists. Indeed, the ERCs have already provoked such a reaction. NSF
PAGE 147
142 is unlikely to shift its priorities much further unless pushed from the outside. Currently, about 10 percent of NSFs budget goes for engineering. It is hard to envision the Foundation, as presently constituted, increasing the proportion for engineering to more than 15 or 20 percent. The Engineering Directorates budget has been spread among several thousand departments in the Nations nearly 300 engineering schools. With relatively low levels of support from NSF, faculty have turned to DoD for money, skewing research toward specialized military problems. (This trend affects curricula and course contents as well, although less directly.) An NTF, independent of NSF and DoD, could be a powerful lever for moving university research back toward the civilian side of the economy, and for steering engineering education back towards practical industrial problems. Nonetheless, OTAs past analyses have found restructuring NSFmaking it, say, into a National Science and Technology Foundation to be a more attractive option than creating a separate National Technology Foundation. 6 Science and engineering do depend on one another. Thus, an integrated agency, charged with supporting both engineering and science, makes more sense than two parallel agenciesprovided sufficient resources can be guaranteed for engineering. In this variant of Strategy 3, with the focus on engineering research in the universities, impacts on commercialization would be long-term and indirectboth a strength and a weakness. Government money would not go directly for commercial technology development in industry, avoiding the problems such a step would raise. But it would take time before the resources flowing to new research in engineering could make a difference for competitiveness and commercialization. In particular, creating an NTF would not do much for HTS. eDevelopment and Diffusion of Commercial Technologies: Should the Federal Government Redefine Its Role? staff memorandum, Office of Technology Assessment, Washington, DC, March 1984. Other suggested reorganizations would target industrial technologies more directly. Box U, for example, discusses one recent proposal, in this case a reorientation of NBS. An Agency for Commercial Technology Development For years, DARPA has enjoyed an enviable reputation: an elite band of non-bureaucrats, able to pick technological winners and drive them forward. Why not, many have asked, do the same on the civilian side of the economy? The response follows just as quickly. DARPA can pick winning technologies because it has a reasonably clearcut mission, whereas a civilian DARPA would have a much fuzzier charge. Nonetheless, at a time when U.S. industry has lost ground competitively, the notion of a civilian DARPA holds considerable appealan agency devoted to championing high-risk, longterm projects, technologies that could make a real difference in international competition, The huge U.S. trade deficit, and especially the imbalance with Japan, is todays Sputnik. Somebody has to do something. DARPA has, in fact, been able to anticipate technologies important to civilian industry. Although commercialization per se has not been DARPAs goal, for most of the agencys history the defense mission has not tightly constrained its decisions. Rather, DARPA has invested in what it regarded as high-payoff technologies, on the rationale that DoD would ultimately benefit as a purchaser. If DARPA can make make technically sound decisions, a civilian agency should also be able to do so. But the DARPA analogy can be taken only so far. A civilian DARPA, by its nature, would be much more difficult to run efficiently: nurturing new technologies intended to succeed in the marketplace is a more complex and exacting undertaking than supporting a technology for which the Government is the end user. In addition, a civilian DARPA would have high visibility politically. Technology development is now seen as the sine qua non of economic prosperity. This means that a civilian DARPA would be under strong pressure from
PAGE 148
143 Box U.An Advanced Technology Program Among reorganization proposals, the proposed Technology Competitiveness Act has come closest to implementation. The billincorporated in the omnibus trade package passed by Congress in the spring of 1988 (and vetoed by the President)would: rename the Commerce Departments National Bureau of Standards (NBS) the National Institute of Standards and Technology; l authorize regional centers for the transfer of manufacturing technology; provide for technical assistance to State technology programs; l establish a clearinghouse on State and local initiatives on productivity, technology, and innovation; create an Advanced Technology Program as part of the revamped NBS. The technology transfer and State assistance provisions in the bill could be useful. But it is the Advanced Technology Program (ATP) that would most directly address U.S. needs for industrial technology. The ATP would assist businesses in applying generic technologies, and in research needed for refining manufacturing technologies and for rapid commercialization of new scientific discoveries. This would be accomplished through, among other things, aid to joint R&D ventures (including ATP participation in such ventures under some circumstances). The bill also authorizes cooperative agreements and contracts with small business, and involvement of the Federal laboratories in the program. Although the trade bill itself does not appear to specifically authorize appropriations for the ATP, a predecessor bill in the Senate (S. 907) would have authorized $15 million for the ATP in its first year. special interests, States, and Congress itself to steer resources to particular projects, In other words, a civilian DARPA could easily become a pork barrel. Political interests could override economic sense. Could a Civilian Technology Agency (CTA) avoid these pitfalls? What might it look like, and what it would do? In many versions it would be small and lean (DARPAs staff numbers about 125), emphasizing flexibility and making use of experienced professionals on temporary assignment from the established mission agencies, In more expansive alternatives, a CTA might pull in relevant functions from elsewhere in Government, such as support for universitybased engineering research (from NSF), a technology extension effort, and perhaps aerodynamics programs (from NASA). Box V outlines one recently proposed agency. Beyond questions of size and scope, a CTAs effectiveness would depend heavily on four questions: 1) its mission; 2) project selection and monitoring; 3) the quality of its staff; and 4) intramural research. Mission The CTAs central mission would be to extend the time horizons of U.S. industrial R&D, and help fill some of the gaps in the Nations technology base. More specifically, it would be responsible for supporting two rather different kinds of work. The first is long-term, highrisk 17&D at pre-commercial stages, with the goal being relatively dramatic advances in technology. For example, candidate projects in the manufacturing area might include direct reduction steelmaking, or expert systems for shopfloor production scheduling. HTS examples could begin with three-terminal electronic devices, or integration of semiconductor and superconducting electronics, The second area is generic technology development, which would typically be incremental.
PAGE 149
144 Box V.An Advanced Civilian Technology Agency, as Proposed in S. 1233 S. 1233the Economic Competitiveness, International Trade, and Technology Development Act of 1987reported by the Senate Governmental Affairs Committee in 1987. It was then incorporated in that houses omnibus trade bill, later to be dropped. S. 1233 is of interest here because of its provisions for an Advanced Civilian Technology Agency (ACTA), which would have been part of a new Department of Industry and Technology-the latter created through a major reorganization of Federal Government responsibilities. The ACTA provisions in Title I, Part III of S. 1233 represent the closest. that Congress has yet come to implementing some form of civilian DARPA. l Intended to support technology development and commercialization through contracts and grants, cost-shared with industry, the agency would give particular attention to risky, long-term projects. Technology-related functions transferred from the Commerce Department, including NBS, would stand alongside the ACTA (rather than becoming part of it). The bill authorized an ACTA budget big enough to make a difference$80 million in the first year, rising to $240 million in the third year. Financial support to industry would be permitted through the stage of prototype development. S. 1233 would provide for a high-level outside advisory board, but in most other respects leave agency operations up to the Secretary of the new department and his or her deputies. Report language calls for a small professional staff (35, initially), coming largely from industry, with considerable use of scientists and engineers on loan from the private sector. IAS explicitly stated in Economic Competitiveness, International Trade, and Twhnologyl?ewdopment Act of 19S7: Report of the Committea on Governmental Affairs, United Stfites SmMte, To Accompany S. 1233, Report No. 100-82 (Washington, DC: U.S. Government Printing Office, lune 23,19871, P 10. In *tion 122 of the bill, the a8ency fS directed to coordinate i~ activities with those Of ~th DARpA and NSF~ the dir@ctors of which are to be members of its advisory board. Many projects here would aim to reduce research to practice. The work would help many companies, but would rarely lead directly to proprietary advantage. Examples (again from manufacturing): nondestructive evaluation techniques, especially those suited to real-time operation as part of feedback control systems; small hand tools for mass production that are ergonomically designed for ease and speed of use (something that gets little attention in the United States compared to Japan and Europe). HTS examples include processing of the new ceramic materials, and magnet design and development for applications such as separation of steel scrap, or refining of ores. Why is mission so important? Because it is a precondition for accountability. DARPAs mission creates discipline over the decisions of its staff and managers: only so long as DARPA can show that the work it funds will support future military requirements can the agency expect support from the Office of the Secretary of Defense and the relevant committees in Congress. For a civilian agency, vague statements concerning commercialization or competitiveness will not do. Lack of agreement on mission is one of the reasons why none of the many bills introduced over the years to create a new technology agency has become law. Consensus on mission is critical for any agency with substantial budget authorityand given the size of the U.S. economy, and the needs for industrial technology, a CTA would have to have an annual R&D budget of $100 million or more for meaningful impact (DARPAs current budget is about $800 million). Whatever form a CTA might take, it would never have a mission as clearly defined as that of DARPA or NASA. The overall goalsupporting commercial technologies in order to support the international competitiveness of U.S. industrydoes not lend itself to neat and clean decisionmaking. Competitiveness is difficult to measure, harder to predict, and depends only partially on technology. Many people and many groups may view competitiveness as a legitimate goal. But as a practical matter, it would
PAGE 150
145 not be possible to judge the merits of a CTAs workor evaluate the outcomes of completed R&D projectsby linking that work to competitiveness. The linkages are too loose, the causal connections often spanning many years. Even so, it should be possible to define the technological objectives of a CTA tangibly enough to provide a handle on mission. Project Selection and Monitoring With the charge of supporting commercial technology development, much of a CTAs budget would have to go for contracts with the private sector (including consortia), on a costshared basis. Some money might also flow to non-profit laboratories, and to universities through grants and contracts, But the point, after all, is to channel direct support to industrial technology, supplementing the many indirect measures the Federal Government already calls on. A CTA would not be able to rely exclusively on review panels or outside experts to develop an overall strategya broad view of where resources should goor for help in setting priorities. Outside experts, by definition, have a narrow view. The further science and technology advance, the greater the specialization among experts. The CTA would have to depend on the collective judgment of its own staff for strategy and priorities. How about project-specific decisions? Money for private firms raises questions. The agency would have to choose projects on grounds that would be accepted as fair. Again, the answer begins with a competent staff, combined with merit review processes (box W). The CTA Staff Federal support for commercial technologies will always run the risk of devolving into little more than a program of subsidies for industry, with much of the money going to marginal projects, The primary guarantee against that danger is to staff the CTA with professionals who have the independence of judgment and the technical knowledge to make good decisions and stick to them, To gain the respect of their industrial counterparts, CTA employeestechnical specialists, program managers, administratorswould need a good grasp of market realities, as well as of industrys technical requirements. They would need to function as part of a peer group that includes industrial scientists, engineers, and R&D managers. If the agencys managers were to provide exciting work, give employees substantial responsibilities, and maintain a selective and competitive personnel policymore like DARPA or the Office of the U.S. Trade Representative than the Commerce Departmenta CTA should have little trouble in assembling a capable staff, Finding people with strong technical credentials is relatively easy (American universities excel at deep but narrow training of engineers and scientists). Breadth and experienceindustrial experience, in particularare harder to find, Bringing in people from industry might require exceptions to normal civil service requirements, Given that the Federal Government already employs many highly competent engineers and scientists, the CTA could begin by assembling a core staff borrowed from other agencies. Continued use of detailees would ensure a steady flow of fresh perspectives, while also helping with inter-agency coordination and technology transfer. Industry sabbaticals that sent CTA employees to the private sector for periods of 6 months to 2 years could help serve the same purpose (as suggested inch. 4 for national laboratory employees), Intramural R&D Although most of its projects would be contracted to industry, it would also seem desirable for a CTA to carry out in-house R&D in its own facilities. This need not be a large-scale undertaking (say, 5 percent of the agencys budget). But it would allow staff members to keep their hands in. Some technical employees might rotate through the CTAs laboratories. Others could spend part of their time engaged in R&D more or less continuously. ,, ~,, I 1 t I :
PAGE 151
146 Box W.-Project Review When it comes to selecting projects for extramural funding, Federal mission agencies rely primarily on expert technical reviews. DoD conducts many of its technical reviews internally, but also kinks to external bodies, permanent and ad hoc, on occasion. l NSF and NIH use outside review panels extensively, aiming at merit-based rankings reflecting the collective judgment of a group of recognized experts-the peer review model. Not perfect, these processes can be criticized if the reviewers do not have appropriate qualifications (a problem particularly for internal agency reviews), or represent the conventional wisdom when the need may be to break the mold of ongoing research. Sometimes reviewers may favor their friends (true anonymity may be impossible in a specialized field). But the general approach has been widely accepted. It works, and-if applied appropriately-should work for sponsorship of industrial R&D by a CTA. Once a broad agenda of R&D priorities had been set, the CTA could look to review panels for merit-based judgments, mixing outside engineers and scientists with the agencys own staff. The outside people would have to come from organizations without a direct stake in outcomes. There is no reason why a materials scientist working for an electronics firm could not give a fair review to the materials-related portions of HTS proposals on electrical machinery. Alone, such an individual would not have the expertise. In a group, he or she would contribute a useful perspective. Contract monitoring, necessarily, would be the responsibility of the CTA staff. Agency employees would need to keep a critical distance from sponsored work, and be willing to cut off funding to companies that failed to perform (something DoD has difficulty doing on occasion). ll%e Advisory Group on Electron Devices, ~ longstanding committee of specialists fiorn the three military services, industry, and tmiversities, exerted considerable influence in shaping the Very HiglMpeed Integrated Circuit program, See Federal Support for Industrial Technology: Lessons From VHSIC and VLSI, prepared for OTA by G.R. Fong under contract No, H3651o, December 1S87. NSF uaee criteria for rating proposals that includw technical competence of the proposed reeearch, baaed on past achievements, as well as the details of the proposal; intrin$ic merit of the research, meaning the likely impacts on scientific advance; and relevance. See Guide to Frogramw FiaceJ Yesr 1988, NSF 87-57 (Washington, DC: National Science Foundation 1987), p. ix. For a recent examination of scientific peer review, see University Funding: Information on the Role of Peer Review at NSFwrd NIH, GAOIRced-87-87FS (Washington, DC: U.S. General Accounting Office, March 1987). This activity would also force staff members to demonstrate that they can produce what they want others to produceR&D that is relevant. The test is simple. If the intramural R&D is picked up and used by industry, the CTA staff has passed. Table 15 summarizes this and other features of a CTA. Pitfalls The primary difficulties for any agency charged with supporting commercial technologies are likely to be political rather than technical. The problems of defining an R&D agenda, and recruiting a staff with the right mix of skills and experience would be straightforward compared with the problems of establishing credibility within the broader system of U.S. policymaking. Like any Government institution that seeks to endure, a CTA would have to respect notions of democratic virtue. That requires in addition to accountabilityprudence i n spending public funds, fairness in dealings with the private sector, and some degree of balance with respect to regional interests. 7 For a CTA to become a reality, any proposal would have to satisfy constituencies having very different interestssome conflicting. The Frost Belt, seeking to rebuild its technological base and infrastructure, would no doubt want 7E. Bardach, Implementing Industrial Policy, The Industrial Policy Debate, C. Johnson (cd.) (San Francisco: Institute for Contemporary Studies Press, 1984), p. 103. The discussion following draws heavily on Bardach, Also see H. Heclo, Industrial Policy and the Executive Capacities of Government, The Politics of Industrial Policy, C.E. Barfield and W.A. Schembra (eds.) (Washington, DC: American Enterprise Institute, 1986), p. 292; and International Competitiveness in Electronics (Washington, DC: Office of Technology Assessment, November 1983), ch. 12, especially pp. 475-482,
PAGE 152
147 Table 15.Desirable Features in a Federal Agency for the Support of Commercial Technology Development Budget.$100 million to $500 million annually in early years, exclusive of industry cost sharing, with 90 percent or more going for the support of R&D projects. These projects might be split roughly as follows: l Industry, both single companies and consortia-80 percent l Universities and non-profit research institutes percent l Internal agency projects percent Cost-sharing on industry projects at 40 to 60 percent seems appropriate to ensure that companies view the work as important. Staff.At the $500 million level, the agency would probably need about 250 professional employees. At any one time, about half the professional staff time would be devoted to intramural R&Dwith technical employees expected to spend some fraction of their time, over a period of years, actively engaged in R&D. Substantial use of detailees from other Federal mission agencies, as well as people on leave from universities and industry would be desirable. The agencys permanent staff members could also be expected to spend periodic tours in industry. Intramural R&D.At the $500 million level for the agency, 5 percent for intramural R&D means $25 million annually (and perhaps 100 full-time equivalent professionals), It would seem preferable to maintain a number of relatively small efforts, spread quite widely across the spectrum of industrial technologies; given that the primary function of intramural R&D is to maintain staff expertise, breadth would be essential. Even with half a dozen R&D areas, many staff members with other specialities would have to spend time in industrial laboratories to maintain hands-on R&D skills. SOURCE Office of Technology Assessment, 1988 parity with the Sun Belt; small business interests would probably begin lobbying for setasides. Long before it opened its doors, an agency with the mandate to restore U.S. technological competitiveness would face pressures from companies in financial straits, seeking, for instance, relaxation in the CTAs requirements for cost-sharing. Other threats to neutral allocation of CTA resources would be almost as quick to materialize. Perceived inequities between competing firms and industries would be all but impossible to avoid even with long-term R&D. CTAfunded R&D on magnetic separation of steel scrap would help minimills at the expense of integrated steel producers. Advances in magnetic levitation rail technology would threaten aircraft manufacturers and airline companies. Such pressures could easily jeopardize the CTAs intended focus. Not only would distressed industries be pushing for a quick fix, a national emergencyan energy crisis, saywould bring calls for technological solutions. With high turnover likely in its political leadershipon average, assistant secretaries remain in Government for only 18 monthsa CTA would be under constant pressure to show results. Pressures for immediate results would coexist with pressures to maintain funding for major projects, even if they proved flawed. Managers would be reluctant to admit mistakes, andcompared to their private sector counterpartshave less incentive to do so. Even flawed projects develop constituencies, moreover, ready to argue for continued funding. In sum, a CTAlike any public institution would have to win favor from enough well-situated constituents to continue its work. That is as it should be. The dangerand a very real oneis that, as an institution charged with spending money, a CTA would become just another forum for the distributive clashes that already consume Congress and much of the executive branch. Said one close observer of science policy, We have the most highly-developed system of interest groups in the world, and theyve discovered R& D. CONCLUDING REMARKS The Federal Governments responsibility for nology policy is not only to find and support promoting technology is plainest in two cases: such R&D, but to stimulate industry to use the support for basic research, and investment in results in timely fashion. risky and speculative technologies, In both, substantial public benefits may coincide with meaBasic research continues to flourish under ger private returns. The problem for U.S. techthe present system of U.S. support for science
PAGE 153
148 and technology: when the breakthroughs in HTS occurred, the Nation had the resources and flexibility to mount a considerable effort in short order. To support advanced technologies, however, the United States has traditionally relied on the mission agenciesDoD, in particular. The approach reflects a philosophical distaste for government involvement in the economy, and also the belief that government cannot anticipate the needs of the marketplace; spinoffs, rather than direct financing, have supported many of the new technologies that American industry commercialized. The approach worked well for several decades. But the world has changed. Military technologies have grown steadily more specialized, the defense sector more isolated from the rest of the economy, If DoD R&D funding was ever a cornucopia for U.S. industry, it is no longer. Second, other countries have caught up in technology. Today, both Japanese and West German firms spend higher proportions of their revenues on R&D than American firms. That spending has been one of the critical elements in their competitive success, The emerging pattern in HTS seems much like that in microelectronics. Japanese firms are investing their own funds heavily. Government policies support their efforts. In the United States, only a small fraction of the Federal money for HTS finds its way into industry, and most of this will pass through DoD. These and other indicators lead to the conclusion that a continuation of current policies for supporting commercialization of HTS will leave U.S. industry behind its strongest international competitors, The United States may continue to dominate the science of superconductivity, and might pioneer in commercial innovations. But the contest will eventually come down to engineering and manufacturing, where American industry has fallen down in recent years, and where the Japanese continue to improve OTA has analyzed two alternatives to the business-as-usual approach. One of the choices creation of a Federal technology agency, with HTS as a piece of its territoryholds promise for the future. Such an agency might support industrial technology directly; many proposals have envisioned a kind of civilian DARPA, established to focus on R&D relevant on the civilian side of the economy. The pitfalls are not so much technicalmaintaining a sound portfolio of projectsas political. A CTA would have to deal with the demands of distressed industries, depressed regions, and companie s simply attracted by a pot of R&D money, Whatever their merits, the alternatives under Strategy 3 cannot offer near-term support for HTS. The ad hoc measures outlined under Strategy 2 could. This approach Federal costsharing of joint R&Dwould be explicitly designed to promote an industry-centered agenda of long-term, high-risk R&D in superconductivity. Governments rolecarried out through a working group on commercializationwould be as facilitator, as well as financier, helping to establish consensus on a research agenda, and securing cooperation from universities Federal laboratories, and mission agencies, The three elements in Strategy 2 meet the needs summarized at the beginning of the chapter: diversity and continuity of Federal support; market-driven decisions; technology push complemented by market pull; low visibility; collaboration among industry, universities, and Government. In conjunction with ongoing activities in the mission agencies, they would substantially improve the odds on U.S. industr y in the race to commercialize HTS.
PAGE 154
Appendixes
PAGE 155
Appendix A Glossary Coil/rail gun: Uses a rapidly changing magnetic field in a spiral coil (coil gun) or a linear conductor (rail gun) to accelerate a projectile via magnetic forces. Much greater velocities can be reached than are possible with gas expansion (as in a conventional gun). COMAT, Committee on Materials: An interagency group under the Federal Coordinating Council for Science, Engineering and Technology chaired by the White House science advisor. COMAT's Superconducting Materials Subcommittee, chartered in June 1987, is comprised of program directors and other representatives of Federal agencies involved in superconductivity R&D, Critical current density: The maximum value of the electrical current per unit of cross-sectional area that a superconductor can carry without reverting to the normal (non-superconducting) state. The critical current density drops as the temperature rises toward the transition temperature, and as the magnetic field increases, CTA, Civilian Technology Agency: Several legislative proposals over the years would establish a Federal CTA focused on commercial technology development. DARPA, Defense Advanced Research Projects Agency: A Defense Department R&D funding agency that gives most of its support to long-term, high-risk projects, Examples include artificial intelligence, and, currently, processing of hightemperature superconductors. DPR, Domestic Policy Review of Industrial Innovation: A Carter Administration study of alternative Federal policies for stimulating technological innovation. Electromagnetic launcher: See coil/rail gun. ERCs, Engineering Research Centers: Cross-disciplinary research centers funded by the National Science Foundation at universities. ETL, Electrotechnical Laboratory: This Japanese laboratory, administered by the Ministry of International Trade and Industry, has been involved in superconductivity research since the mid-1960s FCC, Fine Ceramics Center: A laboratory in Nagoya, Japan jointly funded by industry and the Ministry of International Trade and Industry. participating firms send researchers to work at the facility, which opened in the spring of 1987, Fiber-optics: Use of glass fibers to transmit light (produced by lasers) for telecommunications and computer networking. Optical fibers can carry much more information than electrical wires. FLC, Federal Laboratory Consortium on Technology Transfer: A network of technology transfer officers from 400 Federal laboratories and eleven agencies for facilitating transfers of technology from the laboratories to industry. First setup informally in 1974, the FLC was given a statutory basis in the Federal Technology Transfer Act of 1986 (Public law 99-502). HTS, high-temperature superconductor: Refers to materialsfour classes of which have been discovered since 1986with much higher transition temperatures than previously known superconductors. (See LTS.) Intermetallic compound: Chemical compounds of nominally fixed composition, one or more elements of which are metals. Most intermetallic compoundse.g., the superconductor niobiumtin (Nb 3 Sn)are brittle and therefore hard to work with, ISTEC, International Superconductivity Technology Center: An organization for superconductivity R&D set up by Japans Ministry of International Trade and Industry. IUCRs, Industry-University Cooperative Research Centers: National Science Foundation program that provides seed grants for cooperative R&D at universities, JJ, Josephson junction: Superconducting electronic devices that can be used to sense electromagnetic radiation and also as digital switches (hence as logic devices in computers). o K, degrees Kelvin: The absolute scale of temperatures, with O 0 K (-4940 F) equal to absolute zero (a temperature that can be approached but never reached, Logic chips: Integrated circuits consisting of arrays of gates each of which implements a Boolean function such as AND, OR, NOR, NAND. Computer processors are built from logic chips, as are many specialized digital systems. LTS, low-temperature superconductor: Materials that become superconducting only when cooled to a few degrees above absolute zero. All superconductors discovered before 1986 were lowtemperature materials, with 230 K (-418 0 F) the highest known transition temperature, (See HTS). Magnetically-levitated train (maglev): Trains suspended and propelled by magnetic forces offer the prospects of much higher speeds than can be achieved by conventional wheel-on-rail technologies. A prototype superconducting maglev train in Japan (also called a linear motor car) has achieved speeds of over 300 miles per hour. 151
PAGE 156
152 Magnetometers: Sensors which measure magnetic field strength. Because magnetic fields accompany so many physical phenomena, magnetometersincluding ultrasensitive versions made from superconducting deviceshave many uses. (See JJ, SQUID.) Magnetic resonance imaging (MRI): Refers to equipment and techniques used in medical diagnosis for imaging the soft tissues of the body. MRI systems often use superconducting magnets. MCC, Microelectronics and Computer Technology Corp: A joint venture that conducts R&D. MCCs program to develop and evaluate electronic applications of HTS had 13 participants as of the spring of 1988. MITI: Japans Ministry of International Trade and Industry. Monbusho: Japans Ministry of Education, which supports university research. MRLs, Materials Research Laboratories: Now supported by the National Science Foundation, several of the MRLs are conducting research on HTS. Multicore Project: Established by Japans Science and Technology Agency to link nine laboratories and government organizations working on HTS with one another and with industry. NBS: National Bureau of Standards of the U.S. Department of Commerce. New Superconductivity Materials Research Association: Generally called the superconductivity forum, this association was set up by Japans Science and Technology Agency. It provides workshops, symposiums, and other opportunities for interaction among corporations, universities, and national laboratories. NSA: U.S. National Security Agency. NSF: U.S. National Science Foundation. 1-2-3 superconductor: One of a new class of hightemperature superconductors, typified by yttriumbarium-copper-oxide and called 1-2-3 because of their generic chemical formula: RBazCOs07.X, with R almost any one of the rare-earth elements.. Much of the research on the new superconductors has focused on the 1-2-3 materials, which typically have transition temperatures above 90 0 K OSTP, Office of Science and Technology Policy: Headed by the White House science advisor, OSTP is part of the Executive Office of the President. Perovskite: Refers to the crystal structure shared by the 1-2-3 and other high-temperature superconductors. Rail gun: See coil/rail gun. RAM chips: Integrated circuits that provide random access memory for computers and other digital systems. SBIR, Small Business Innovation Research: A Federal program, in operation since fiscal 1983, which requires Federal agencies to set aside a small percentage of extramural R&D budgets for contracts with small businesses. SDI, SDIO: The Strategic Defense Initiative, and the Defense Department organization that runs it. Sematech: An R&D consortium, financed by 14 member companies (as of the spring of 1988) and the U.S. Government, established to pursue improvements in semiconductor manufacturing technologies. Signal-to-noise ratio: An important parameter for sensors, the signal-to-noise ratio compares the signal the sensor is intended to measure with background noise (one source of which is thermal, rising with temperature). SMES, superconducting magnetic energy storage system: A coil or solenoid of superconducting wire in which an electric current can circulate, storing energy until needed for purposes such as feeding an electric utility grid or powering a free-electron laser. SQUID, superconducting quantum interference device: A very sensitive instrument, built with Josephson junctions, used to detect magnetic signals. STA, Science and Technology Agency: Under the Prime Ministers Office in Japan. S&T centers, Science and Technology centers: Multidisciplinary centers proposed for funding by the National Science Foundation. Stevenson-Wydler Act: The Stevenson-Wydler Technology Innovation Act of 1980 (Public law 96480, as amended), placed increased emphasis on technology transfer from the Federal laboratories. The 1986 Federal Technology Transfer Act (Public law 99-502) amended the 1980 act to provide (among other things) more emphasis on cooperative research between federally operated laboratories and industry. Superconductivity: Total loss of resistance to direct electrical currents. Superconducting magnet: An electromagnet wound with superconducting wire. Essentially all the power consumed goes for refrigeration to keep the coil windings below their superconducting transition temperatures. Three-terminal electronic device: One which, like a transistor, can amplify a signal substantially. (See two-terminal device.) Transition temperature: The highest temperature
PAGE 157
. at which a material becomes a superconductor, also known as the critical temperature. The transition temperature drops as the magnetic field and current density increase. Two-terminal electronic device: one which, like a Josephson junction, can serve only as a weak amplifier. (See three-terminal device.) URI, University Research Initiative: A Department of Defense program, started in 1986, intended to support university capabilities in research, and training of scientists and engineers, in disciplines important for national defense. 153 VHSIC program: An R&D program begun by the Department of Defense in 1979 to develop advanced integrated circuits for military systems. VLSI program: Joint government-industry R&D effort in Japan for developing very large-scale integrated circuits (VLSI), in existence from 1976 to 1980. X-ray lithography: Creation of patterns for fabricating integrated circuits using X-rays, Because Xrays have shorter wave lengths than visible light, they can produce finer patterns, hence denser circuits.
PAGE 158
Appendix B The Technology of Superconductivity Electric currents travel through superconductors with no resistance, hence no losses (provided the current is steadyalternating currents meet resistance even in superconductors). When current flows in an ordinary conductor, say a copper wire, some power is lost. In a light bulb or an electric stove, resistance creates light and heat, but in other cases the energy is simply wasted. With no resistance, magnets wound with superconductors can create very high fields without heating up and dissipating energy. Motors and generators with superconducting windings could be smaller, lighter, and more efficient than those built with copper. Very high magnetic fields might be used to fire projectiles, to float molten metal in a steel mill, to levitate trains. Table B-1 lists some of the possible applications. During the 1960s, low-temperature superconductors (LTS)specially developed metal alloys like niobium-titaniumcame into use in specialized applications such as magnets for scientific research. Some current LTS applicationse.g., ultrasensitive magnetic field detectorswill not be superseded by high-temperature superconductivity (HTS) (because of higher thermal noise at higher temperatures). Other applications, possible but not practical with LTS, could become much more attractive with HTS. As table B-1 indicates, superconductors not only banish electrical losses, and provide the basis for very sensitive detectors of magnetic fields and other radiation, but can also be used to produce the fastest possible electronic switching devices. In addition, superconductors exclude magnetic fields, which means they can be used as radiation shields. (The exclusion of external magnetic fields is termed the Meissner effect, after the physicist who discovered the phenomenon in 1933.) Many of the applications listed in table B-1 have been goals for engineers and scientists since superconductivity was discovered early in the century. I Much of the material in this appendix is drawn from Superconductive Materials and Devices, Business Technology Research, Wellesley Hills, MA, September 1987; and Technology of High Temperature Superconductivity, prepared for OTA by G.J. Smith 11 under contract No. J3-21OO, January 1988. Also see Physics Today, Specia] Issue: Superconductivity, March 1986; A.P. Malozemoff, W.J. Gallagher, and R.E. Schwall, Applications of High-Temperature Superconductivity, Chemistry of High-Temperature Superconductors, ACS Symposium Series 3s1, D.L. Nelson, M.S. Whittingham, and T.F. George (eds.] (Washington, DC: American Chemical Society, 1987), p. 280; Research Briefing on HighTemperature Superconductivity, Committee on Science, Engineering, and Public Policy, National Academy of Sciences, Washington, DC, 1987. Along with powerful magnets for a variety of purposes, prototype generators, electrical transmission lines, and computer chips have all been made, operating in most cases at liquid helium temperatures (about 4 K, or 4 degrees above absolute zero, figure B-l). But the very low temperatures have been a barrier for many of the applications possible in principle. High-temperature superconductors and liquid nitrogen cooling (77 0 K, figure B-1) would bring relatively modest improvements in costs, system complexity, and practicality for most of these applications. Liquid nitrogen temperatures, after all, are only high compared with the near-absolute zero of liquid helium. Where system designs already incorporate LTSe.g., magnetic resonance imaging (MRI, ch. 2)it might not pay to change over simply to take advantage of HTS. In applications where unattended operation is desirablee.g., military surveillance or geophysical explorationthe much reduced boil-off rate of liquid nitrogen would be a major advantage. HTS also holds obvious attractions for applications in space; beyond low-Earth orbit, passive cooling may suffice to maintain superconductivity in the new materials. Nonetheless, it is quite possible that continued R&D will bring new applications of HTS that cannot yet be anticipated. superconductivity at room temperature, moreover, would be truly revolutionary. Compact and efficient small motors and actuators, for example, could find uses ranging from household products and automobiles to machine tool drives and power-packs for replacing aircraft hydraulic systems. Superconductivity Above its transition (or critical) temperature, a superconductor exhibits electrical resistance like any other material. Below the transition temperature, the material has zero resistance to direct current (DC): a steady electric current will circulate in a superconducting coil forever, so far as anyone knows. Variation in the flow of current does lead to electrical losses; thus a superconductor dissipates energy when turned on or off, or when carrying an ordinary alternating current (AC losses). However, these losses are much less than those in a good normal conductor (e. g., copper) at the same temperature. 154
PAGE 159
155 Table B-1 .Representative Applications of Superconductivity -. Large-scale passive: Shields, waveguides Superconductors screen or reflect electromagnetic radiation; possible applications range from coating of microwave cavity walls to protection from the electromagnetic pulses of nuclear explosions. Bearings Repulsive forces created by exclusion of magnetic flux make non-contact bearings possible. High-current, high-field: Magnets Medical imaging LTS magnets widely used in commercial systems. Scientific equipment LTS magnets used in fusion experiments and particle accelerators. Magnetic separation Possible uses include separating steel scrap, purifying ore streams, desulfurizing coal, and cleaning up stack gases, At least one LTS magnet is in current use for purifying Kaolin clay. Magnetic levitation Levitated trains have been extensively studied, with prototypes in Japan and Germany. Launchers, coil/rail guns Electromagnetic launching systems can accelerate objects to much higher velocities than gas expansion; possible applications range from small guns for military purposes to aircraft catapults and rapidly repeatable Earth satellite launching. Other Powerful magnets could eventually find a very wide range of uses. Examples: compact synchrotrons for lithographic processing of integrated circuits; growth of the crystals for integrated circuits (a strong magnetic field yields more nearly perfect wafers of silicon and other semiconductor materials); MHD (magnetohydrodynamic) systems for energy conversion. MHD thrusters might also be used in place of propellers to drive ships and torpedos. Other static applications Electric power transmission Prototypes of LTS underground lines have demonstrated feasibility, but such installations are not costeffective (compared with overhead high-tension lines) at present. Energy storage Solenoids wound with superconducting cable could store electrical energy indefinitely as a circuIating current; i n addition to utility applications (e. g., load leveli rig), superconducting storage could find uses in miIitary systems (e.g., pulsed power for large lasers). Cheap and reliable superconducting energy storage would eventually find many other applications. Rotating machinery Generators A number of LTS prototypes have been buiIt to investigate possible electric utility applications. Motors, motor-generator sets Used in conjunction with a superconducting generator, a superconducting motor could be an efficient alternative to mechanical power transmission for applications such as ship and submarine drives, railway locomotives, and perhaps even for helicopters. Sufficiently low costs would open up many industrial applications. Electronics: Passive Superconducting wiring (interconnects) for computers, on-chip or between chip, could help increase processing speed. Sensors SQUIDS (superconducting quantum interference devices) made from Josephson junctions (JJs) are the most sensitive detectors of electromagnetic signals known; applications range from detecting neural impulses in the human brain to geophysical exploration, detection of submarines in the deep ocean from airplanes or, potentially, from space, and nondestructive inspection. Digital devices JJs can also be used for digital switches, opening up such applications as computer logic and memory; competitive three-terminal devices with substantial gain may eventually be developed; combined semiconductorsuperconductor devices or systems also hold many attractions. Other devices Analog/digital converters, voltage standards, many types of signal processors, and microwave mixers can all be designed, in principle, with superconductors; some of these applications (e. g., voltage standards) have been reduced to practice with LTS JJs. NOTE This Iist IS based on known properties of known materials. It IS not exhaustive, even for existing, well. understood Iow-temperature superconductors SOURCE Off Ice of Technology Assessment, 1988 Until 1986, the highest known transition temperature was 23 K (in niobium-germanium). Then two IBM scientists in Zurich discovered a new class of materials that showed superconductivity at 35-40 K (see box C, ch. 2). Shortly thereafter, the compositions now termed the 1-2-3 superconductors were found, with transition temperatures in the vicinity of 95 K. The first of the 1-2-3 materials announced contained yttrium, barium, and copper oxide. More recently, copper-oxide ceramics containing bismuth or thallium have been found. These have superconducting critical temperatures in the range, respectively, of 110 K and 125 K. In April 1988, the first HTS compositions containing no copper were announced, with transition temperatures up to 30 K.
PAGE 160
156 Figure B-1 .Temperature Scales University of Leiden had developed cryogenic refrigeration equipment capable of reaching these very low temperatures, In 1913, Kamerlingh-Onnes found that lead became superconducting at 7.20 K. At this point, the Leiden group built a superconducting magnet with lead windings, only to find the lead reverting back to its normal state when the magnetic field reached a few hundred gaussa severe limitation on practical use, given that a common kitchen magnet creates a field of about 1,000 gauss. (The average magnetic field of the Earth is one-half gauss; the windings in electric motors create fields of about 10,000 gauss,) Critical Properties. Later it was learned that maintaining the superconducting state requires that both the magnetic field and the electrical current density, as well as the temperature, remain below critical values that depend on the material. Figure B-2 shows this schematically, while table B-2 gives the critical values of temperature and field for a number of superconducting materials, Practical applications, in general, require that both the transition temperature and the critical current density be high; in some cases, relatively high magnetic fields are SOURCE: Office of Technology Assessment, 1988. necessary as well. A major advantage of the new materials, of course, is the potential for simpler, less expensive cooling, For technical reasons, superconductors must be operated well below their transition temperatures as a rule of thumb, at half to three-quarters the transition value. In fact, then, liquid nitrogen temperatures will be marginal for the 1-2-3 compositions (although the practical advantages of operating in the range of 40 0 K rather than 4 K can be great), If the more recently discovered ceramics, with critical temperatures of 110 K and up prove to have otherwise useful properties, liquid nitrogen cooling will almost certainly prove adequate. All the HTS compositions so far discovered are ceramics, rather than metals. They are new materials, poorly understood. All ceramics are brittle; they require processing and fabrication methods very different from metals and alloys. Superconducting Properties and Behavior In 1911, Heike Kamerlingh-Onnes, a Dutch physicist, made the astounding discovery that mercury lost all electrical resistance at 4 K. Earlier, Kamerlingh-Onnes and his research group at the Figure B-2. Dependence of the Superconducting State on Temperature, Magnetic Field, and Current Density Temperature, T A Critical surface if outside) SOURCE: Off Ice of Technology Assessment, 1988
PAGE 161
157 Table B-2.Critical Values for Superconducting Materials Temperature Magnetic field Current density a (degrees Kelvin) (gauss) (amps per square centimeter) Aluminum . . ... . . . . . . Mercury . . . . . . . . . Lead . . . . . . . . . . . Niobium. ., . . . . . . . . . Niobium (75%) titanium (25%) ., . . . . . Niobium tin . . . . . . . . . 1-2-3 ceramic (YBa 2 Cu 3 0 69 ) . . . . . . aAt Zero magnetic f(eld buppercnhcal field (Type H superconductor cAt42 K dAt77 K Thehlghestvaluesare reached wlthor!ented Slngie-crystat films SOURCE Offrceof Techr?ology Assessmen! 1988 By the late 1930s, scientists had distinguished Type I and Type II superconductors. Type I materials, in which the phenomenon had first been studied, shift abruptly to their normal state above the critical magnetic field. Type II superconductors exhibit a mixed state between two values of magnetic field, the lower and upper critical fields. The new HTS materials show Type II behavior, with extremely high critical fields (table B-2)indeed, so high in the 1-2-3 compositions that simply measuring them has proven very difficult. As Kamerlingh-Onnes found with his lead-wound magnet, Type I superconductors have critical fields too low to make useful magnets. While Type II materials have much higher critical fields, the new HTS materials have proved to have disappointingly low values of the third critical parameterthe current density (table B-2). Raising allowable current densities in the 1-2-3 ceramics from the values found initially in polycrystalline samples (consisting of many grains, randomly oriented)below 10 3 amps per square centimeterquickly became a major research target. For most applications, improvements of 100 times or moreto the range of 10 5 or 10 6 amps per square centimeterwill be needed, This is important not only for high-power applications; electronic devices carry small currents, but current densities are high because cross-sectional areas are microscopic. Not a fundamental limitation, the low current densities are materials processing problems (critical current density depends on the microstructure of the material, hence on its processing). Many years of R&D were needed to raise the critical values for Type II superconductors like niobium-titanium to the values shown in table B-2, Similar effort lies ahead for the new HTS materials; so far, progress has been most rapid in thin films, From the 1930s to the 1980s.The 1930s and 1940s saw a good deal of progress in the cooling and refrigeration systems needed to reach very low temperatures, spurred in part by wartime needs for liquid oxygen. 2 After the end of the Second World War, the newly formed Office of Naval Research established a major research program in low-temperature physics. One consequencerapid improvements in the technology for producing liquid heliummade experimental research in superconductivity much easier. Federal support during the 1960s included development of very powerful superconducting magnets, principally for conducting experiments in highenergy physics and nuclear fusion, as well as exploratory studies of possible electronic applications of superconductivity. The first of the current generation of conductors, niobium-tin, was developed during the early 1960s. A brittle intermetallic, very difficult to work with, niobium-tin found little use. Within a few years, almost all LTS magnets were being wound with niobium-titaniummore ductile, though with a lower critical field. With steady progress in processing techniques, niobium-titanium conductors could be fabricated as braided cables each containing thousands of very fine filaments. Small filamentsless than the diameter of a human hairreduce the AC losses stemming from variations in current, and have other highly desirable properties for magnet applications. These filaments are embedded in a copper matrix, capable of carrying the full current in the event of an accidental loss of superconductivity. The conductors must be flexible enough for winding, and jGovernments Role in Computers and Supercondu{:tors, prepared for OTA by K. Flamm under rontra( t No H3f?J470, March 1988, pp 43ff.
PAGE 162
158 mechanically strong because the magnetic field creates very high forces. Many years of R&D have led to steady improvements in LTS magnets. The discovery of the Josephson effect in the early 1960s opened up a new class of possible applications in electronics. Josephson junctions (JJs), made with a thin insulating layer (a matter of a few atomic diameters) separating two superconductors, can act as very fast electronic switchescomfortably exceeding the fastest semiconductor devices. Because they dissipate so little powerabout 1,000 times less than semiconductorsJJ electronics can be more tightly packed, which also contributes to speed. Theory Theoretical work had likewise moved ahead, culminating in 1957 in the Bardeen-Cooper-Schrieffer (BCS) model, for which a Nobel Prize was later awarded. The BCS theory explains superconductivity in terms of interactions between electrons (which carry current) and phonons (atomic vibrations). Under normal conditions, the electrons collide with atoms, leading to electrical resistance and energy loss. In the superconducting state, electrons move in coordinated fashion without any collisions. While the BCS model explains superconductivity in the old materials, so far HTS has baffled the theorists. Until 1986, most of those working in the field had agreed that BCS superconductivity at temperatures above about 30 0 K was impossible. Given the recent experimental results, the theoretical community has been scrambling to extend the BCS model or find some new explanation. Although the Josephson effect was first predicted theoretically, and later confirmed by experiment, theory gave little or no guidance in the search for LTS materials with higher transition temperatures. Thus the situation is really no worse for HTS. Looking for materials with higher transition temperatures remains largely a matter of trial-and-error, guided by intuitionlaboratory research that is time-consuming and expensive. Processing and Fabrication Tailoring HTS materials for applications either in electronics or where high currents and/or fields are needed (e.g., electric power) will entail design and processing at size scales from the atomic level on up. Engineers and scientists engaged in applications developments, as well as materials processing, will have to concern themselves with electronic structure (energy gaps), crystal structure (the arrangement of atoms in the material), microstructure (grain boundaries), and the fabrication of films, filaments, tapes, and cables. The HTS ceramics are not only brittle, and chemically reactive, but highly anisotropicmeaning that properties vary with direction within a grain of the material. The 1-2-3 ceramics, for instance, show differences of as much as 30:1 in critical current density depending on grain orientation. Some of the current density limitations can be traced to anisotropy, but grain boundaries seem to be the primary culprit. Some of the processing and fabrication techniques familiar from work with electronic and structural ceramics hold promise for the new superconductors. Bulk samples of HTS material can be made by hot pressing, extrusion, and tape casting, among other methods. The anisotropy in the 1-2-3 materials has led many research groups to seek processes for aligning the grainse.g., extruding a slurry of single crystals in a high magnetic field to create a tape. Semiconductor fabrication techniques, likewise, can in some cases be adapted for making thin films. Past work on niobium-tin, a brittle intermetallic compound, may also hold lessons for HTS materials, which, like other ceramics, cannot deform plastically. Because they break easily and without warninglike glass, being very sensitive to small imperfections (hence the scribed lines used to cut glass)practical applications may require specialized in-situ processing, as well as careful design to minimize strain. Magnets wound with niobium-tin are made starting with strands of niobium in a copper-tin alloy matrix flexible and ductile. Heat treatment after the wires have been drawn and wound into coils causes the tin to combine with the niobium, forming the superconducting compound, with its vastly different properties. Some R&D groups have pursued similar processes for ceramic superconductors. Progress has been faster with thin films, which can be created via a wide range of well-known techniquese.g., sputtering, and evaporation by molecular or electron beams, Finding good substrates on which to deposit the HTS layer has been the primary problem. The HTS compounds react chemically with many otherwise desirable substrate materials, including those used for integrated circuits (silicon, sapphire). Strontium titanate gives high current densities compared to other choices, but is expensive and has otherwise undesirable properties. Silicon would be ideal as a step toward combining semiconductor and superconducting electronics. While the temperatures so far required for creating the proper HTS composition have posed
PAGE 163
159 difficulties, many research groups have been working on the problems, with encouraging results. Applications Much of the excitement over HTS has been stirred by speculation concerning such possible applications as low-loss electric power transmission or magnetically levitated trains. In some of these cases, commercialization will depend more on system costs and progress in competing technologies than on the specifics of HTS. Both transmission lines and levitated trains have been demonstrated with LTS materials. Superconducting transmission lines, which must be run underground because of the cooling requirements, may eventually prove cost-effective relative to conventional underground transmission; thus far, however, these applications have not moved out of the test stage, Maglev trains could be built by the end of the century in Japan and West Germany. Competing technologies sometimes present a moving target for superconductivity: after more than 10 years of R&D aimed at a Josephson computer, IBM concluded that competing semiconductor technologies were improving rapidly enough that its approach to JJ computer elements would probably not bear fruit. More than likely, then, 5 to 10 years of R&D lie ahead before many applications of HTS emerge. Those that come earlier are likely to be highly specializedperhaps in military systems, perhaps targeted on very demanding civilian needs (for example, Hypres very high-speed data sampler, which incorporates LTS electronics]. The ongoing R&D will involve: 1. 2. 3 Basic research, both theoretical and experimental, aimed at explaining HTS, at finding new materials and exploring their properties, and at understanding structure-property relationships. Applied research, focused particularly on development of processing methods and optimization of material properties through manipulation of processing variables, A great deal of R&D will be needed before routine production of tapes and multifilamentary conductors could begin, with substantial improvements in critical currents an early step. Josephson junctions for electronics will also be difficult to reduce to practice. Applications engineering (for HTS)e.g., development of prototype chips containing many JJsincluding extensive testing under realistic operating conditions (environmental exposure, 4. 5. thermal cycling, mechanical vibrations, electrical surges, loss of temperature control), Joining techniques for conductors will be needed; so will repair methods. Process engineeringmanufacturing methods for routine (rather than laboratory) production. Problems here will include yields and reliability in superconducting circuits, and methods for producing long continuous lengths of superconducting cable. Inspection, testing, and quality control procedures will need a good deal of attention. Systems engineeringdesign, development, and demonstration of applications in which superconducting components are integrated into such end products as computers, electrical generators, and coil or rail guns. For instance, without further progress in transition temperatures, HTS interconnects in computers will require cooling to liquid nitrogen temperatures. Fortunately, these temperatures also offer performance advantages for semiconductor chips. Many of these activities can go forward in parallel. In some cases it makes sense to proceed sequentially. For instance, applied research aimed at increasing current density can and should proceed in conjunction with process R&D, because processing affects microstructure, and microstructure affects current density. But work on production scaleup must wait until the effects of processing variables can be reasonably well understood. On the other hand, research intended to discover whether a particular processing techniquee.g., laser annealingcompromises some properties will be needed early. High-Current, High-Field Applications Magnets.Most past applications of superconductivity have involved the design and construction of powerful magnets wound with LTS materials and cooled with liquid helium. Such magnets have been used in scientific experiments (e. g., the Tevatron, ch. 2), and in MRI. Learning to design and build magnets helps with more demanding applications, such as rotating machinery. Almost all the power consumed by a superconducting magnet goes to operate the cooling system. For a big magnet wound with copper, resistive losses far outweigh the refrigeration costs for an equally powerful LTS magnet. Indeed, large copperwound magnets need their own cooling systems just to carry off the heat generated through resistance. The cost comparison below, for a bubble chamber
PAGE 164
160 magnet at Argonne National Laboratorya typical early scientific applicationshows that a conventional magnet would cost five times more to operate: 3 Annual operating costs (thousands of dollars) Superconducting Conventional magnet magnet (actual) (estimated) Electrical power. ... . $17.5 $550 Cooling . . 81.3 4 Maintenance . . . 5.2 ? $104 $554 + Superconducting magnets have other advantages compared with conventional magnets. Stability is easier to achieve, for instance. In a conventional magnet, the field strength varies as the windings heat up and expand. The stability characteristics of LTS magnets give them advantages both in scientific apparatus and in MRI. With LTS magnet technology well in hand, HTS designs will have to perform at least as well (in terms of characteristics such as stability) before their simpler cooling systems and lower operating costs will make them competitive. Fabricating the conductors will be difficult, Stable operation and protection against overheating in the event of refrigeration failures require multifilamentary cables, just as for LTS, with filament diameters of a few microns. 4 Given the brittleness of the new ceramics, methods for producing filaments and for fabricating cables are not yet in sight. Once HTS wire and cable become available, applications-specific requirements will come to the fore. MRI, for example, while requiring highly stable fields for good image quality, does not otherwise make heavy demands on the magnet system. Still, joining methods that eliminate resistive imperfections will be needed for image quality comparable with that already achieved using LTS. MRI systems are expensive, and savings from simpler cooling will not make that much difference for commercial competition. Magnetic separation is 3 P. J. Reardon, High Energy Physics and Applied Superconductivity, IEEE Transactions on Magnetics, vol. MAG-13, 1977, p, 705, This magnet, for Argonnes 12-foot bubble chamber, draws 1800 amperes, producing a field of 1.8 X 10 3 gauss. The cost figures assume 140 operating days per year. As another example, a magnetic separator for purifying Kaolin clay (table B-1 ) consuming 270 kilowatts [kW) if built wltb a conventional magnet, plus another 30 kk% for cooling the magnet, could today be replaced with an LTS magnet that needed no more than 60 kW, all for refrigerating the windings, 4Report of the Basic Energ~ Sciences Advisorj Committee Panel on High-Tc Superconducting Afagnet Applications in Particle Phjsics, D~E/Ef7-0358 (Washington, DC: Department of Energy, December 1987), pp. 9-12. another story. Here, for instance, cheap but powerful magnets could be used to sort scrap metal for recycling, in refining ores, purifying chemicals, removing sulfur from pulverized coal, and cleaning up waste water, In all these applications, cost, reliability, and ease of use (including maintenance) by a largely blue-collar labor force become significant design considerations. Design considerations for maglev trains likewise include cost, reliability and longevity, and safety. But the political and economic questions loom even larger than for, say, desulfurizing coal, In the United States, investments in fixed-rail transportation would have to clear obstacles ranging from opposition by airlines to high costs for rights-of-way. In Japan, where the needs and constraints differ, R&D on HTS-based maglev is much more likely to go forward. Electric Power and Utility Applications.Magnets have no moving parts. Technical complexities grow in electrical machinery, and in the entire range of electric utility applications. Transformers, for example, would demand more attention to AC losses than magnets, while superconducting transmission lines will almost certainly have to go underground, so long as refrigeration is required. Underground lines are costly, although already in use in many urban areas. Still, the over-riding design requirement is reliability, Utilities are quite willing to trade off higher operating costs against lower probability of failures and down-time, A disabling failure, after all, can lead, not only to a blackout, but to an ongoing need to purchase power from other suppliers until repairs have been completed. In general, HTS-based generators will need conductors similar to those for magnets. Dynamic forces, however, will add to static forces, while cooling also becomes more difficult. Large conventional generators already have efficiencies greater than 98 percent. Superconducting field windings can increase this to more than 99 percent, In a large machine, an improvement of 0.5 percent to 1 percent in efficiency can be significantreducin g th e losses by halfwhile superconducting generators have the additional advantage (for utility applications) of increasing network stability (they are less sensitive to shifts in electrical load). Worldwide, at least two dozen LTS generator R&D projects have been undertaken since the middle 1960s, but none has gone beyond construction and testing of a prototype. Utilities will have to be convinced that such machines offer reliable service over periods of many years before investing; HTS will not affect the economics much compared to LTS, and, lacking even the experience base of
PAGE 165
161 LTS systems, the new materials have an added hurdle to overcome. Energy storage ringswith no moving parts, and tolerable failure modeswill almost certainly come first. Other Electrical Machinery.For non-utility applications, characteristics other than efficiency and reliability come to the fore: superconducting machines promise to be smaller and lighter than conventional motors and generators by half and more, These are the attractions for ship propulsion, where a superconducting generator driving a superconducting motor could eliminate the gearing and shafting between turbine (or other prime mover) and propeller. With much more freedom in packaging, nuclear submarines could carry more weapons (or be smaller), So could surface ships. Submarines might also prove quieter, perhaps even faster. Moreover, with the motor/generator set(s) providing speed control (and reversing), efficiency during part load operation would rise (the turbine can run at its optimum speed). As table B-1 indicated, other, more cost-sensitive, applications for motor/generator sets might also open up at some point. And of course, given high enough operating temperatures, the many large electric motors used throughout industry (ranging from pump, fan, and blower drives to machine tools and rolling mills) would be candidates for replacement. Electronics From the beginning, Josephson junctions have been the basis for many superconducting electronic devices, SQUIDssuperconducting quantum interference devices, simple circuits incorporating JJs have extremely high sensitivity levels, which have led to a considerable range of practical uses for LTS SQUIDS. The Josephson effect can also be exploited for computer logic and memory; although a number of practical problems stand in the way, JJs could in principle replace semiconductor chips in powerful digital processors (box J, ch. 3). Sensors.SQUIDs can detect the very faint signals produced by the human heart (10 -6 gauss) and brain (1 0 -9 gauss). These simple circuits can also measure a wide variety of other electromagnetic signals (anything with an associated magnetic signature from DC up to microwave frequencies). SQUIDS are about 1,000 times more sensitive than the next best magnetic field detectors, They can sense the disturbances in the Earths magnetic field caused by a submarine deep in the ocean, or the field distributions caused by geologic formations holding oil or mineral deposits, Requiring, in simplest form, only one or two JJs (rather than the large numbers required in computer applications), LTS SQUIDs typically fabricated from niobiumare now made routinely. To minimize thermal noise, SQUIDS should be operated at the lowest possible temperature, and in any case at less than half to two-thirds of the superconducting transition temperature, At liquid nitrogen temperatures, for instance, sensitivity will be 20 times poorer than at liquid helium temperature, Even so, an HTS SQUID would still be a more sensitive magnetic field detector than any of the alternatives except an LTS SQUID, If they can be built successfully, HTS SQUIDS will quickl y find a considerable range of applications (though none of these are likely to be high-production-volume applications). Computers and Other Digital Systems.JJ-based electronic devices promise switching speeds 10 times faster than the very best compound semiconductors. Because the energy losses are several orders of magnitude smaller, JJ-based integrated circuits could be packed much more densely, However, the practical problems of making JJ-based chips far exceed those of SQUIDS. Even if the practical problems were solved, Josephson computers might not be commercialized. The competing technologies extend well beyond silicon and gallium arsenide chips: a good deal of R&D has been going into alternative computer architectures such as massively parallel processors. Much of this work seeks increases in processing power without major advances in components. Still, faster chips will always promise faster machines, But, in a further contrast with SQUIDswhich are the most sensitive magnetic field detectors known the theoretical limits of JJ-based logic devices fall well short of what might eventually be possible, for example, using optical switching. Thus the window of opportunity for JJ-based computing may never open. (It may never open for optical computing, either.) On the other hand, advances in device designand, in particular, a practical three-terminal device that would erase the primary drawback of JJ chips, low gaincould open a broad new frontier. 5 It is simply too early to say, R&D in the United States and Japan on LTS-based JJ computing illustrates some of the problems that designers of HTS logic and memory would face. IBM was able to build logic chips with 5,000 junctions reliably, but had trouble with cache memory. 5S. G, Davis, The Superconducti\e Computer In Your Future, Datamation, Aug. 15, 1987, p 74,
PAGE 166
162 (Fast logic does no good without fast cache memory for support.) IBMs prototype memory chips, with over 20,000 JJs, proved susceptible to errors caused by slight variations in control currenta good example of the kind of problem that a 3terminal device would help solve. More recently, Japanese companies have built several kinds of LTS chips incorporating niobium JJs. Fujitsus 4-bit microprocessor, 25 times faster than a similar silicon chip, and 10 times faster than a gallium-arsenide microprocessor, consumes only 0.5 percent as much power as either. NEC has produced a 1,000 bit dynamic memory, containing 10,000 JJs; access time is a factor of 200 better than for silicon. The first applications of HTS in computers may be interconnectselectrical pathways joining otherwise conventional chips. Signal dispersion and other problems associated with transmitting electrical pulses within the processor limit performance; practical means for incorporating HTS interconnects should find ready application in large and powerful machines. Moreover, at liquid nitrogen temperatures, superconductors and semiconductors could operate compatibly in hybrid designs. Ordinary semiconductors cannot be used at liquid helium temperatures; even if they could be made to operate in otherwise satisfactory fashion, semiconductors would dissipate too much heat, overwhelming the cooling system. Given that hybrid LTS-semiconductor systems are not feasible, past work on Josephson computing has involved either all-superconducting chips, or unique designs with controlled temperature gradients. The Hypres data sampler, for example, uses an integrated circuit cooled to liquid helium temperature on one end onlythat end holding about 100 LTS JJs. Three-terminal devices could be a big step forward in superconducting electronics, making possible logic designs at the chip level much like those now used with semiconductors. It could well be, however, that major advances in HTS electronics would come only with devices that departed in a major way from currently known electronic devices, The first requirement, in any case, is mastery of thin-film fabrication technology. Military Systems As table B-1 indicated, possible defense applications of superconductivity range from shielding against nuclear blasts to high-speed computers and motor-generators for ships. Conceptually, there may be little difference between military and commercial applications. But in practice, differences will be pervasive at levels all the way from devices and components (e.g., radiation hardening) to the system configuration itself (cost-performance tradeoffs much different than for commercial markets). Computing requirements for smart weaponsfor example, real-time signal processingtend to be quite different from those important in the civilian economy. Thus, as development proceeds, military uses of superconductivity will diverge in many respects from civilian applications. Some of the military applications could be compelling. Submarine detection with SQUID-based sensors, for instance, offers at least a factor of 1 0 improvement over current methods. Conventional electric generators for shipboard or vehicle use, or for producing electric power under battlefield conditions, produce about 2 horsepower per pound; prototype LTS generators have already reached 25 horsepower per pound. Superconducting coil or rail guns promise increases in projectile velocities of 5 to 10 times. The U.S. Department of Defense (DoD) has funded superconductivity R&D since the early 1950s, contributing to the development of large, high-field magnets, electrical machinery, LTS sensors, and superconducting computers. DoD (and the Department of Energy) also supported much of the materials processing R&D that proved necessary to achieve high current densities in LTS wire and cable. Since 1983, the R&D objectives of DoD programs in LTS have been redirected, and the programs have grown, as a result of the Strategic Defense Initiative (SDI). For SDI, HTS shielding, waveguides, and sensors (for use in space) hold obvious attractions, while LTS work also continues; early in 1988, Bechtel and Ebasco began an SDI-funded design competition on LTS magnetic energy storage for powering groundbased free-electron lasers. SDI has also targeted very high-frequency communication systems, where LTS could offer substantial improvements in performance and extended frequency range. Here, the 1-2-3 ceramics seem to offer theoretically promising electronic characteristics (i. e., larger energy gaps). They would also avoid the many practical problems that liquid helium cooling poses in a military environment. DoD has also renewed its attention to two of the prospective high-field, high-power applications ship propulsion, and coil/rail guns, Military funding of R&D on LTS machinery began in the middle 1960s, with a 300()-horsepower prototype completed several years ago. Magnetohydrodynamic (MHD) thrusters offer a wholly different alternative, doing away with propellers, as well as shafts and gear-
PAGE 167
163 ing. In 1978, the Defense Advanced Research Projects Agency began funding R&D on electromagnetic launchers, or coil/rail guns. The initial goal, apparently, was a cannon for the Navy. With the advent of SDI, much of the DoD work has been redirected toward higher velocity systems, capable of launching a projectile into space. Like the commercial applications, the requirements, whether for machines or for coil/rail guns, start with good conductors. Developing the Superconductivity Technology Base Table B-3 gives a sampling of expert opinion on timing for a number of the applications discussed above, Without too much oversimplification, the R&D needed for supporting these and other applications can be pictured as in figure B-3. Leaving aside military applications, particularly those in which the superconductor serves as a passive shielding medium, sensors and other relatively straightforward electronics applications will probably come first. As noted earlier, without new and much more tractable families of HTS materials, learning to make practical wire and cable will be a long and tedious process. As a result, the highcurrent, high-field applications will be slower in reaching the marketplace than thinand thick-film electronics. The R&D tasks outlined in figure B-3 will take a wide range of skills. Materials synthesis and characterization demands well-equipped laboratories and sophisticated experimental techniquese.g., X-ray Table B-3.Estimated Development Times for Prospective HTS Applications Application Time a SQUIDS . . . . . . . Less than 5 years Sensors. . . . . . . 5 years Computer interconnects . . . Less than 5 years Superconducting computer. . . Long-term Multifilamentary composite cable . 5 to 10 years Magnet system . . . . . over 10 years Magnetic energy storage . ... . Long-term Transmission lines . . . . Long-term Electrical generators . . . . Long-term a untll ~mall.scale commercial Production SOURCE Technology of High Temperature Superconductivity, prepared for OTA by G J Sm!th II under contract No J3-21OO, January 1988, p. V.97, based on interviews and neutron diffraction, electron microscopy, molecular beam epitaxy. Making thin and thick films of the 1-2-3 materials with adequate currentcarrying capacity will probably mean oriented grain structuresa good deal more difficult in production than in the laboratory. Fabricating useful Josephson junctions will mean controlling the deposition of very thin layers. The processing techniques are likely to be more demanding than related semiconductor processing technologies. Still, there is much that can be learned from related technologies, not only in microelectronics, but in ceramics. Applications of both structural and electronic ceramics demand very pure starting materials, careful control of processing (and thereby structure), and sensitive nondestructive inspection techniques, Some of this experience base will translate to HTS, especially to fabrication processes for filaments and wires. As figure B-3 suggests, cryogenics technologies will be needed for most applications of HTS (in the absence of room-temperature superconductivity], Space is the exception, Even if much higher transition temperatures emerge, good performance may still require coolinge.g., to minimize electrical noise, or increase current-carrying capacity, Although much of the speculation concerning HTS has assumed liquid nitrogen cooling, closed-cycle refrigeration systems can reach temperatures as low as 100 K, and would probably be the technology of choice in many systems. Much of the R&D needed for commercialization of HTS will have to go on more-or-less simultaneously. For simplicity, one-way arrows join the boxes in figure B-3: a more realistic picture would be full of feedback loops representing the flows of knowledge accompanying development of a complex new technology (ch. 2), As conductor fabrication technology evolves, the design constraints for magnets and machines will take shape. System level studies of digital processors will feed back to the device level. Developing the technology base in HTS means multidisciplinary research, and productive interactions among universities, national laboratories, and industry. Developing a technology base quickly, so that U.S. industry can keep up with Japanese industry, will mean taking risks, and managing overlapping R&D projects. The examples of industries ranging from automobiles to microelectronics (ch. 2) demonstrate that competing in HTS will require an R&D system that effectively supports parallel development on many different but inter-related problems.
PAGE 168
164 Figure B-3. HTS R&D Making Making wires, tapes, films and other high-current conductors Thick Thin I v Shielding, interconnects, Devices, sensors (many types) Systems Magnets Machinery Electric utility applications SOURCE: Office of Technology Assessment, 1988
PAGE 169
Index
PAGE 170
Index Advanced Civilian Technology Agency (ACTA), proposal for, 144 Advanced Technology Program, proposal for, 143 Altair, 31 American Superconductor Corp., 60 Ampex, 45 Antitrust, enforcement and U.S. industrial competitiveness, 5, 6, 22 Applications high-current, high field, 40-41, 65, 136, 159-160, 162-163 LTS, 154, 157 military, 162-163 prospective HTS, 40-41, 44-48, 134-136, 159-163 superconducting electronics, 41, 136, 161-162 Appropriations. See Budgets; Funding Atomic Energy Commission U.S. (AEC), 25, 131, 140 AT&T, R&D funding by, 7, 21, 55, 56, 57, 58, 59, 71, 72, 102 Bardeen-Cooper-Schieffer (BCS) model, 158 Bednorz, J. C., 26 Bellcore, R&D by, 56, 59 Bloch, Eric, 141 Boeing, 98 BRITE (Basic Research on Industrial Technologies in Europe), 43 Budgets CTA R&D, 144, 147 Federal Government R&D, 5, 9, 17, 23, 83, 87 Japanese superconductivit y R&D, 10, 61, 65, 68, 69, 70, 71-72, 73, 75 see also Funding California, University ofBerkeley, 58, 60 Carter, Jimmy, 22 Central Electric Power Research InstituteJapan, 70 Ceramics commercializing R&D results pertaining to structural for use in heat engines, 27, 28, 44-45 Japanese expertise and leadership in, 19, 27, 61 superconductivity materials made of, 3, 6, 27, 4445, 155, 156, 157, 158, 162 China, 3 Chu, Paul, 26, 129 Civilian Technology Agency (CTA) desirable features of, 144, 145-146, 147 mission of, 143-145 potential problems with, 146-147 project selection and review for, 145, 146 proposals for, 13, 126, 141-142, 143 Collaboration industrial R&D, 133-137 industry/Federal Government R&D, 5, 12-13, 23, 86, 113-114 industry /university/Government, 123, 124, 126 industry/university HTS, 59, 60, 130, 131 international HTS research, 11, 77-79, 86, 115-118 in Japanese R&D system, 9-10, 11, 62-63, 73-77 policy issues and options concerning strengthening university /industry/Government, 8, 9, 85-86, 104-115 Commercialization, 3-5, 17-20, 42-44 aggressive support strategy for, 132-138, 139 European HTS, 43 factors influencing U.S. HTS, 35-40 Federal Governments role in high-technology, 3-5, 9, 20-27 flagship approach to technology, 125, 139 overview of technology, 3-5, 17-20 strategies for HTS technology, 12-13, 123-142 summaries of specific examples of technology} R&D, 27-28, 44-48 working group on, 137-138 see also Management; Marketing; Research and development [R&D) Committee on Materials (COMAT), 90 Conductus Inc., 60 Costs. See Budgets; Funding Current density, critical, 156, 157 Defense Advanced Research Projects Agency (DARPA)DoD, 139 as precedent for CTA, 142-143 R&D support by, 9, 25, 44, 85, 93, 104, 139, 163 Defense. See Department of Defense, U.S. (DoD); Military Department of Commerce, U.S. (DOC), 138, 143, 145 Department of Defense, U.S. (DoD), 90, 91 high-technology development [postwar history), 96-98 HTS R&D funding by, 6, 12, 24, 41, 85, 93-99, 112, 162-16 3 R&D budget of, 23, 87 see also Defense Advanced Research Projects Agency (DARPA); Military Department of Energy, U.S. (DOE), 90, 91 collaboration with industry by, 5, 23 HTS R&D funding by, 6, 12, 24, 28, 44, 85, 99-101 see also National laboratories Development, See Commercialization; Research and development (R&D) Digital Equipment Corp., 31, 137 Domestic Policy Review of Industrial Innovation (DPR), 22 Du Pont, R&D by, 6, 7, 51, 57, 58 Eastman Kodak, 57 Economic Competitiveness, International Trade, and Technology Development Act, 144 Economic Policy Council (EPC), 90 Education HTS personnel, 130, 133 U.S. engineering, 4, 142 167
PAGE 171
168 Electric power HTS applications involving, 160-161 shortage in Japan, 9, 40 Electric Power Research Institute, 90, 101 Electronics, superconducting, 41, 136, 161-162 Energy Research and Development Administration (ERDA], See Department of Energy (DOE) Engineering education, 4, 142 parallel, process for HTS, 34-35, 159, 163 raising priority of research on, 13, 141-142 Engineering Research Centers (ERCs)NSF, 9, 103, 104, 106, 107, 108 ESPRIT (European Strategic Program of Research in Information Technology), 43 Eureka (program), 43 European Community (EC), 43, 78 Expenditures. See Budgets; Funding Exxon, 55 Fabrication. See Processing Federal Government. See Government, Federal Federal laboratories. See National laboratories Fine Ceramics Center (FCC)Japan, 70, 76 Food and Drug Administration, U.S. (FDA), 47 France basic research support by, 26 government involvement in industry affairs by, 22 R&D support by, 43 Funding continuity in R&D, 8, 85, 91, 123, 124-125 Federal HTS R&D, 6, 9, 24, 84, 86-93 Federal objectives of R&D, 9, 23-27, 84 industrial R&D, 6-7, 21, 54, 57, 58, 60, 61, 84 R&D at U.S. universities, 6, 12, 85, 102, 103, 104, 132-133 see also Budgets Fujitsu, 71, 72 General Electric (GE), R&D support by, 46, 55, 57, 59 General Motors Corp. (GM), 21, 98 Germany, Federal Republic of. See West Germany Government, Federal (U.S.) HTS funding by, 6, 9, 24, 84, 86-93 industry support by, 3-5, 12, 20-23, 128, 131, 133-137 procedural rules for R&D cost sharing by, 136-137 R&D budget, 5, 9, 17, 23, 83, 87 R&D funding objectives of, 9, 23-27, 84 technology commercialization strategy options for, 12-13, 123-148 see also Military; individual agencies in Government, State, HTS support by, 11, 23, 114-115, 129 Heat engines, 44-45 History, of superconductivity development, 26, 154-158 Hitachi Co. business strategy of, 51 R&D support by, 7, 64, 71 Hoechst AG, 43 Honda, 32, 33 Houston, University of, 58, 129 Hypres Co., 41, 159 IBM, 41, 43, 51, 98 R&D funding by, 7, 21, 55, 56, 58, 59, 71, 72, 102 transition temperature breakthrough by, 3, 26, 155 Imsai, 31 Incentives indirect, for commercialization, 5, 6, 12, 20, 22 Japanese energy shortage as HTS R&D, 9, 40 Industry competitiveness of U.S. high-technology, 3-5, 8-9, 32 direct support from Federal Government to, 5, 12, 20-23, 128, 131, 133-137 HTS R&D consortia, 127, 133-137 HTS R&D strategies of U. S., 6-7, 57-59 R&D funding by, 6-7, 21, 54, 57, 58, 60, 61, 84 R&D strategies of Japanese, 9-11, 29, 33-35, 51-52, 60-67 R&D strategies of U. S., 4, 6-7, 10, 34, 35, 51-52, 53-59 wait and see attitude of U. S., 4, 6-7, 10, 52, 53, 55-56, 131-132 Industry/University Cooperative Research Centers (IUCRs)NSF, 107 Intermagnetics General Corp. (IGC), 48 International Superconductivity Technology Center (ISTEC)Japan, 65, 72, 77-79 Investment. See Budgets; Funding Japan basic research support by, 7, 9, 26, 61, 68, 75 ceramics technologies expertise of, 19, 27 corporate R&D strategies in, 9-11, 29, 33-35, 51-52, 60-67 HTS policy of, 9, 10-11, 67-79, 86 HTS R&D initiatives in, 3, 7, 9, 10-11, 67-80 industry support by government in, 9-10, 22, 62-63, 73-77 national laboratories of, 9, 69, 70, 76, 78 policy issues and options concerning technology interchange with, 9, 86, 116-117 U.S. technology interchange with, 11, 77-79, 86, 115-118 Josephson junctions (JJs) computer applications using, 41, 64, 71-72, 161-162 reproducible HTS, 30, 41 R&D for, 57, 70, 163 Keyworth, George, 91 Knowledge base Federal Governments contribution to, 3, 5, 11-12, 21, 96 for superconductivity, 163 gaps in U.S. technology, 11, 83, 85, 91, 136 private sectors access to, 3, 5, 21 Korea, Republic of. See South Korea Legislation. See individual statutes
PAGE 172
. Low-temperature superconductivity (LTS), 5, 6, 25, 65 historical development of, 154, 156, 157-158 magnet technology, 29, 40, 41, 48, 64, 154 properties and behavior of, 156, 157 R&D funding for, 93, 95, 99 see also Josephson junctions (JJs); Magnets Maglev (magnetically levitated] trains, 73, 74, 159, 160 Magnetic resonance imaging (MRI), commercializing R&D results involving, 27, 29, 31, 41, 47-48, 154, 160 Magnets HTS, 29, 160 LTS, 48, 157, 159-160 Maintenance, See Management Management Japanese industrial R&D, 63 U.S. industrial R&D, 54-57, 135 Marketing policy, 143, 144 R&D applications, 8-9, 29-31 research, 43 Market pull, policies encouraging, 123, 124, 125-126 Massachusetts Institute of Technology (MIT), 60, 96 Materials Research Laboratories (MRLs), 103-104, 106-107 Matsushita, 29, 45 Microelectronics consortium, 5, 35, 134, 136 as precedent for HTS commercialization, 35-40 Microelectronics & Computer Technology Corp. (MCC), 79, 135, 136, 137 Military commercial spin-offs from R&D spending by, 4, 5, 9, 84-85, 94-95 R&D directed toward, 25, 83, 84, 85 superconductivity applications, 162-163 see also Department of Defense, U.S. (DoD) Ministry of Education (Monbusho)Japan, HTS support by, 68, 70-73 Ministry of International Trade and Industry (MITI)-Japan, 63, 71, 72 HTS support by, 68, 69-70, 75, 76 International Superconductivity Technology Center (ISTEC), 65, 73, 77-79 Ministry of TransportJapan, HTS support by, 68, 73, 74 Mitsubishi Electric, R&D support by, 64, 74 Muller, K. A., 26 Multicore Project (STA)Japan, 69, 76, 77 Muto, Yoshio, 73 National Academy of Sciences (NAS), 86, 87 National Aeronautics and Space Administration, U.S. (NASA], 91-92, 101-102 National Bureau of Standards, U.S. (NBS), 24, 77, 87, 90, 101 National Cooperative Research Act, 22, 24 National Critical Materials Council (NCMC), 89, 90-91 169 National defense. See Department of Defense, U.S. (DoD); Military National Institutes of Health, U.S. (NIH), 25, 138-140 National laboratories HTS R&D in, 85, 110-114, 131, 139 LTS magnet research at, 48 technology transfer to private sector from, 19, 22, 23, 83, 86, 110-114, 130 see also Department of Energy, U.S. (DOE) National Security Agency, U.S. (NSA), 71, 97 National Science Foundation, U.S. (NSF), 90, 91 Engineering Research Centers, 9, 103, 104, 106, 107, 108 HTS R&D funding by, 6, 12, 85, 102, 103, 104, 132-133 Industry/University Cooperative Research Centers (IUCRs), 107 Japanese industry superconductivity survey by, 64, 65-66 Japan Initiative of, 116 multidisciplinary R&D encouragement by, 9, 85, 102-103, 105, 106, 107 Research Applied to National Needs (RANN) program of, 107, 125 science and technology (S&T) centers (proposed), 104 National Technology Foundation (NTF), proposal for, 141-142 NEC Co., R&D support of, 7, 71, 72 Netherlands, The, 43 Nippon Steel, 51, 67 Nippon Telephone & Telegraph (NTT), 71 Nissan. 33 Office of Science and Technology Policy (OSTP), 89, 90, 93, 138 1-2-3 ceramics. See Ceramics Onnes, Heike, 156, 157 Oxide ceramics. See Ceramics Packard, David, 23 Packard Commission, 131 Patents, as indirect incentive to industrial investment, 6, 20, 22, 111 Personnel CTA, 145, 147 exchange program (industry/national laboratories), 86, 112-113, 116-117 Japanese superconductivity R&D, 64, 65 U.S. superconductivity R&D, 11, 64, 130, 133 Philips, 43 Planning. See Management Policy commercialization and U.S. technology, 3-5, 9, I 1 12, 20-27, 83-86, 128-13 2 issues and options for Congress, 83-119 Japanese HTS, 9, 10-11, 67-79, 124-125 see also Strategies Polymorphic Systems, 31 Private sector. See Industry
PAGE 173
170 Processing as central element in HTS commercialization, 139 HTS applications marketing and methods of, 8-9, 30 R&D funding of, 95, 98 techniques for HTS materials, 157-159 RACE (R&D in Advanced Communication-technology for Europe), 43 RCA R&D support by, 55 VCR commercialization and, 29, 46-47 Reagan, Ronald HTS initiative of, 6, 34, 128-129 R&D support by, 12, 23, 83, 85, 90, 91 Young Commission and, 22 Regulation, as indirect incentive for industrial investment, 5, 20, 22 Research Federal support for directed, 25-27, 83 Japanese support for basic, 7, 9, 26, 61 U.S. support for basic, 12, 23, 24, 56, 83, 130, 133, 139 marketing, 34 multidisciplinary, 9, 85, 102-103, 105-107, 133 Research and development (R&D), 3-5, 7-11, 17-20 collaborative HTS, 133-137 continuity in funding, 123, 124-125 coordinating Federal, 89, 90, 128, 140-141 corporate strategies for, 4, 6-7, 9-11, 29, 32-35, 51-59 CTA intramural, 145-146, 147 diversity in support for, 123, 124 Federal budget for, 5, 9, 17, 23, 83, 87 Federal funding of HTS, 6, 9, 23-27, 84, 86-93, 128, 133-137, 140-141 industrial funding of, 6-7, 21, 54, 57, 58, 60, 61, 84, 163, 164 marketing of, 8-9, 29-31 parallel, 34-35, 66, 67, 163 policy issues and options concerning funding levels and priorities for Federal, 8, 9, 85-86, 87-93 science and technology (S&T) agency for coordinating, 128, 140-141 U.S. industry view of long-term, 4, 6-7, 10, 52, 53, 55-56, 131-132 Research Applied to National Needs (RANN) programNSF, 107, 125 Saito, Shinroku, 76 Science. See Research Science and Technology Agency (STA)Japan, HTS support by, 68, 69, 76 Science base. See Knowledge base Sematech, 5, 35, 134, 136 Siemens, 43 Sloan, Alfred, 33 Small Business Innovation 22, 115 Small Business Innovation 115 Development Act, Research, Federal (SBIR), Sony, VCR development by, 31, 46 South Korea export policy effectiveness of, 5, 61 VCR development by, 46 Sperry Univac, 71, 72 SQUIDS, 161, 162 Stanford University, 58, 60 Stevenson-Wydler Technology Innovation Act, 22, 143 Strategic Defense Initiative (SDI) HTS use in, 24, 85 LTS research for, 40, 162 Strategic Defense Initiative Organization (SDIO), 93 Strategies Japanese R&D, 9, 10-11, 29, 33-35, 51-52, 60-67 key ingredients of Federal HTS development, 123, 124 options for Federal Government HTS commercialization, 12-13 product/process, 33-34 U.S. industrial R&D, 4, 6-7, 10, 34, 35, 51-52, 53-59 see also Policy Tanaka, Shoji, 72-73, 76 Taxes, as indirect incentive to industrial investment, 5, 20, 22, 92-93 Technologies commercialization of (overview), 3-5, 17-20 DoDs postwar development of, 96-98 industrial competitiveness and development level of, 4, 21 strategies for commercializing HTS, 12-13, 123-142 Technology base. See Knowledge base Technology push, 123, 124, 125-126 Technology transfer from Federal laboratories, 9, 22, 23, 83, 86, 110-114, 130 policy issues and options concerning U.S./Japanese, 9, 86, 115-117 Technology Transfer Act, 111, 118 Toshiba R&D support by, 64, 74, 76 VCR development by, 29, 31, 45 Training. See Education Transition temperatures current flow and, 154 discovery of ceramics higher, 26, 155, United States. See Government, Federal Universities 163 Us. funding for HTS R&D at U. S., 6, 12, 85, 102, 103, 104, 132-133 HTS R&D in Japanese, 72-73 multidisciplinary R&D within, 9, 85, 102-103, 105107, 13 1 University Research Initiative (URI)DoD, 107 U.S. Steel, 55
PAGE 174
171 Video-cassette recorders (VCRs), commercialization of Westinghouse, 59 R&D results involving, 29, 31, 45-47 Visibility, level of R&D programs, 123, 125, 137, 139 Young, John, 22 Young Commission, 22, 128, 140 West Germany export policy effectiveness of, 5 Zenith, 46 R&D support by, 43
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EBOIKVH24_T1JKCK INGEST_TIME 2017-05-24T21:03:38Z PACKAGE AA00055484_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
|