Click for next page ( 96


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 95
Chapter 6 ORGANIZATIONAL, FUNDING, AND POLICY ISSUES The foregoing chapters presented technical aspects of the subject of materials and high-density electronic packaging. The topic was presented from the points of view of systems considerations, existing approaches, and the materials themselves. In this chapter, the business, organizational, policy, and political aspects of the subject are examined. The committee recognizes that some of the discussions below go beyond technical matters, which are the focus of the report. Nevertheless, these discussions are pertinent to giving readers an understanding of the overall industry situation and provide a basis for some of the recommendations developed by the committee. This inevitably leads to comparisons of the U.S. versus the Japanese approach, involving issues of competitiveness, research funding, vertical integration, and the range of business vision in the United States. The 1980s could well be labeled the decade of "electronics consortia," and this mode of organization must also be considered. One must acknowledge that the loss of the U.S. technological edge in electronics can be related to the U.S. electronic industry's concentration on IC chip design and overall systems aspects. By contrast, the Japanese industry moved into the volume commercial marketplace, with particular focus on the packaging area, that gave them a major advantage in the technological advancement of their industry. The U.S. industry placed emphasis on the high- end, low-volume technologies, hoping for a ~ trickledown" effect. This was contrary to the Japanese approach of getting into the low-end, high-volume products, that were then upgraded to the high-end technologies where their new dominance is appearing today . In addition, large Japanese electronics companies are fully integrated from raw materials through the consumer product, giving them full control of all fabrication steps. U.S. companies have found it difficult to assimilate this type of integration into their manufacturing schemes and philosophy. INTEGRATION The U.S. electronics industry is fragmented laterally. Manufacturers do not share component, system, and process design information beyond those necessary to establish needed industry standards. This leads to nationally expensive redundance, at least some of which could be eliminated without exposing the public to the negative aspects that can result from monopoly control. U.S. corporations are, however, reluctant to engage in activities that could bring antitrust enforcement actions. There is some progress in cooperation, but caution is evident. 95

OCR for page 95
96 The U.S. electronics busir~ess is also fragmented vertically, arid it is the lower levels of the supply and manufacture ng chain that are most closely involved with the materials aspects of packaging and interconnection. The following sequence illustrates the problem for organic maters als employed in chip encapsulation. A compounder (i . e ., an independent company) purchases epoxy resins, accelerators, curing agents, silica fillers, antimony oxide (a flame retardant), and other components from various chemical suppliers, quite possibly ~ different supplier for each compound. The chemicals are supplied according to specifications, but they are normally not "pure" by electronics standards. Epoxy resins are usually somewhat ambiguous in chemical structure, and the silica powder varies in particle size distribution and possibly in surface characteristics. The compounder normally is not sophisticated in the characterization of incoming supplies. The materials are mixed (compounded) by a proprietary process that is controlled to some degree, but not precisely. Solid preforms, which are about 2 in. in diameter and 16 in. thick, are made and sent to the encapsulation facility, a different company, commonly in the Far East. The preform is preheated, placed in a plunger cylinder , heated further, and forced through runners of a transfer mold and into cavities containing individual chips (perhaps as many of 300 at a shot). In this process, the epoxy precursors react to form a highly crosslinked polymer, release heat, and undergo first a decrease in viscosity, then an increase to form a rigid solid case surrounding the chip. This encapsulated device is tested (under considerable stress) and shipped to an assembly factory, where it is put on a circuit board and soldered in place. (It should be noted that the circuit board, organic or ceramic, also has arrived at the assembly factory by a similarly complex chain of suppliers and processors.) Following more testing, the board is plugged into the system by the f inal sys tem manufac surer . Clearly, the system works remarkably well, and the existence of complex electronic apparatus attests to the dedication and expertise of the various contributors along the way. There is, however, very little information flow along ache chain because the materials and processes employed are secret and proprietary at each fabrication stage. This places great importance on the consistency for each process . Men problems ari se, it can be difficult to identify the source. A molder can be expected to find fault with the epoxy compound, even when it is known that the molding process is not quite right. And so it goes at each stage . As packaging and interconnection become more and more demanding (and fine lines, large lead counts, and higher power dissipation will test the technology), process maintenance, quality assurance, and production schedules will be increasingly difficult to guarantee. The chain of transfer from one company to another creates even greater challenges for research and development. The U.S. companies involved are traditionally secret-process-based, and this mentality runs deep. An organization does not transmit technical details to its customers, and, although there are signs of movement to a more cooperative posture, reform requires changes in this ingrained behavior.

OCR for page 95
97 The alternative is ache entry of a vertically integrated U.S. company that could carry through several or all of the steps in material production, compounding, and molding. No serious U.S. effort is known in this direction for chip encapsulation, printed circuit board manufacture, or ceramic detail production. SYSTEMS VERSUS MATERIALS APPROACHES A central technical issue is the balance of package and interconnect materials considerations, building-block approaches, system-architectural approaches, and system-organization approaches ~ Past packaging approaches have not been materials-focused but have stressed configuration of known materials. Past system-performance approaches have focused on parallelism and enhanced building blocks having tighter physical configurations. These approaches might have come to a point of diminishing returns, but progress continues to be made, and rate of improvement does not appear to be tapering off. Even so, it is clear that materials contributions to the enhancement of system capability will be centrally important in the next decade and should be addressed. l . The political issues being addressed in this report are probably the more important ones because these issues appear to limit progress in resolving today's technical problems. A few of these political issues are scrutinized here. First, the relationship between the materials supplier community and the user community is universally unsatisfactory. The suppliers are perceived as secretive, concerned with establishing a proprietary position for their financial advantage, unwilling to invest for a long-term goal, and unwilling to speculate beyond their current status. The users, in turn, are perceived as indifferent to physical limitations, unwilling to share their evaluation results with their suppliers, playing off suppliers against each other, and always blaming the materials when things do not work out as they would like. The problem is that the suppliers and the users are not of a common technical culture and are separated by marketing and purchasing communities that are not concerned with technical progress. There is need for a technical community in which the users and suppliers can engage in technical interchange without being cornered by the interplay of financial, accounting, and business ~ activities. This is difficult to accomplish in this corporate-based culture. A second political issue that will have a major impact on the packaging and interconnect approach is that of funding and control. Funding is one of the levers with which it is possible to move the suppliers and the users of packaging and interconnect materials to work with each other. It can be noted that government funding seems to perform this unifying role especially well, s ince all the participants involved sense that they are benefiting from funding that does not come from their own pockets. Government funding seems to flow naturally from the Department of Defense, but for packaging, the most serious problem appears to be the divergence of military packaging from the

OCR for page 95
98 packaging approaches of mainstream developers of high-performance systems. is easy to cite some specifics: Military systems insist on the use of hermetic packages, but most high-performance systems in the commercial world use plastic-packaged (or unpackaged) parts. Military system technology tends to lag a generation behind that used in the commercial world because military systems development and deployment have at least a 10-year time frame, whereas commercial systems must get to market 3 years after the start of development, or they will have missed the market and another supplier will have gotten there first. Military systems are tied to conservative approaches, such as hermeticity, no organics inside the package, and the value of old technologies in the face of rapid progress in the commercial world. Nevertheless, the military services have funding, and their willingness to fund (if not to use) advanced technology makes them a partner to be sought. The most serious issue is the choice of vehicle for the materials development activities. To develop materials in isolation from a system application is a futile exercise even under the best of circumstances' but, when the materials are required to be structurally and electrically stable and to be subj ect to a variety of environments and applications, it is impossible deco imagine a successful outcome. The materials user should choose a challenging system application and pick a system supplier (or suppliers ~ that will support them In applying their materials expertise to the system in a real-time (scheduled) program. PROGRAMS AND CONSORTIA During the 1980s, several U.S. cooperative technical ventures have been implemented. These programs came in response to a number of Japanese efforts that began in the mid-1970s that are perceived to be extremely successful. Competing efforts began to appear in Europe in 1983. Some are governmental in origin, and others were organized privately. In addition, some are solely funding operations, whereas others have established facilities for technology development and demonstration. Some of the programs have a healthy component of packaging technology, but none is focused on packaging and interconnection. In the following paragraphs, several programs are described briefly in the context of this report. Very-High-Speed Integrated Circuits (VHSIC) The VHSIC program was intiated by the Department of Defense in 1980 as the program, Advanced Technology Weapons Systems. Funding of more than a billion dollars has been committed to cover 10 years' work. An aggressive processor design program was developed, and contract work was undertaken by Honeywell, Hughes, IBM, TI, TRW, and Westinghouse. Packaging and PWBs are a

OCR for page 95
99 part of the program, and significant advances have been made in chip carrier, pin grid array, and TAB technologies. Microwave and Millimeter-Wave Monolithic Integrated Circuits (MIMIC) The MIMIC program was initiated by DOD in 1987 as a spin-off of VHSIC. The purpose is to develop high-speed transmitters and receivers for phased- array radar. Initial funding of $11 million was made available, and work was undertaken by a large number of firms. Semiconductor Research Corporation (SRC) The SRC was formed in 1982 to support generic research in universities on behalf of industrial members (initially AT&T, DEC, DuPont, HP, and IBM). Research results and trained engineers are said to be the main products. The funding level was about $20 million in 1986. Government agencies also have begun to provide funds under the SRC program. Packaging research has received considerable emphasis, giving rise to broadly-based thermal management and reliability results. Microelectronics and Computer Technology Corporation (MCC) The MCC began operations in 1983 with funding by share owners AND, Control Data, DEC, Harris, Honeywell, Motorola, National, NCR, RCA, and Sperry (and later Allied, Mostek, Martin Marietta, Rockwell, and BMC). Over 300 employees conduct R&D at the MCC facility in Austin, Texas, and to date over $200 million has been supplied by the member firms. The purpose of MCC is to provide techniques of manufacture and design, and packaging is one of the specifically identified areas of focus. Significant programs in TAB technology are already well established. S EMATECH Sematech is a consortium of electronics companies established in 1987 to help U.S. chip makers by developing advanced microelectronic manufacturing technology and equipment. It receives DOD funding in the amount of 6100 million, and industrial members will supply funding at the same level. The physical facility is located in Austin, Texas; corporate members include AMD, AT&T, DEC, Harris, HP, Intel, IBM, LSI Logic, Micron, Motorola, National, NCR, Rockwell, and TI. Other Considerations Thus, it seems that packaging is an important consideration in the various consortia formed recently to support advanced electronics system

OCR for page 95
100 manufacture. None of the consortia or programs are specifically directed at materials problems or packaging as the primary issue. Therefore, it may be reasonable to work wi thin the structure of the existing consortia and programs to enhance an integrated approach to package and interconnect design in which materials properties and processing are given emphasis at every step. As circuit performance limitations become more and more package-based, the j oint programs will gravitate to des ign factors at the package level, and the materials and combinations of materials involved will be central issues. An alternative approach could lie in the establishment of a U.S. consortium for pre-competitive research and development for packaging and interconnection materials and processes. Both industry and government funding would be desirable, and the organization could contribute positively to cooperation between materials users and materials suppliers. The materials focus would permit emphasis on synthesis, compounding, processing, reliability, and characterization. Many of the facilities needed are expensive and must be operated by knowledgeable people, and the existence of a materials-oriented consortium would provide economies of scale through pooling of such a capability. Coupling to systems-oriented organizations would be essential. The Table in the Executive Summary illustrates features of the consortium, Sepatech as composed with Sematech (The Air Force is now attempting to implement such a cooperative.) It may be perceived that consortia that are concerned with research and applied research aspects of package design can involve cooperation among industrial companies yet have little effect on their ability to compete at the completed system level. It will be important for U.S. companies to learn this lesson in the near future, and programs along these lines deserve encouragement. Somehow, corporate members should receive.incentives in some form to place first-class people into the consortia activity. THE UNIVERSITY ROLE U.S. universities are outstanding in all areas of engineering and science and, as such, should offer a solid foundation for the electronics industry, including materials issues in packaging and interconnection. At present, gratifying efforts are being made to improve the coupling of academia with industry. However, cultural mismatches impede progress. There is too little understanding of needs and directions by the university community, and industry has been ineffective in establishing lines of (two-way) communication. The consortia can help, but they should realize, as part of their founding concepts, that university interactions require both money and the time commitment of bright, well-connected people. Engineering departments are the sensible academic points of contact, and benefits accrue to both parties. The high proportion of foreign students in U.S. engineering departments is a concern, particularly for DOD-funded work. There is often a degree of insularity among university departments and colleges, a factor that can make it difficult for science departments to interact effectively with the electronics industry. Competition for funding and for good students is a source of difficulty. Industrial firms, consortia,

OCR for page 95
101 and funding agencies will be well-advised to require reasonable levels of interdisciplinary activity. Again, materials and systems factors cannot be separated. Materials compatibility also is critical, and isolated studies and development of individual materials is unlikely to be as valuable as work in context. The Engineering Research Centers (ERCs) program of the National Science Foundation could provide an innovative focus for materials aspects of packaging. Industrial participation and interdisciplinary organization are strongly encouraged in the ERC format. Substantial ongoing funding and program accountability are necessary features of the established ERC format. The ERCs bring together many elements that are perceived as essential in a successful pursuit of materials for high-density electronics packaging. Not to be neglected are the contributions made by the private "think- tank" types of laboratories and institutions such as, Battelle, SRI, A.D. Little, David Sarnoff Research Center, SWRI, and MITRE Corporation. These organizations are funded by government, industry, and foundations to examine various technical and policy questions. They are capable of addressing specific questions that are often too complex or too sensitive for others to undertake without apparent major conflicts of interest. Universities also are uniquely qualified to integrate the knowledge base for areas such as packaging. Packaging and interconnection involve complex engineering judgments involving diverse properties and structures. Modeling, scaling theory, and physical design fundamentals require a breadth of knowledge and facility for codification that are well - adapted to academic engineering programs . Again, work in isolation is less effective . Universities also have the culture and freedom to pursue breakthrough technologies, as opposed to incrementally improving the materials that are normally employed. Current materials have real limitations, and ultimate limitations of materials properties in systems terms are now within sight. The payback is becoming increasingly important for adoption of materials with lower dielectric constant, higher electrical conductivity, higher thermal conductivity, and controlled coefficient of thermal expansion. University work can probe and bring understanding to the physical limitations in these characteristics, beginning with long- range programs that address materials structures that are far removed from materials of commerce. The U.S. electronics industry has many examples of new corporations started by academic staff members that are based on research from university laboratories. It is, perhaps, surprising that more of these firms are not based on materials developments. As materials limits are approached, the opportunities for commercial applications of exotic materials will proliferate. Smaller companies can more easily commercialize materials measured in grams rather than tons, but success will depend on the technical depth and patent protection. University laboratories should be highly successful in this general area, and the near future is ripe for materials development in the electronics packaging business.

OCR for page 95
102 To fully enjoy the fruits of the U.S. academic community, it is important for U.S. industry to improve the industrial-academic interface. Industrial cultures tend to be secretive to an excessive degree, which is harmful to the corporations that the secrecy is designed to protect. Universities are perceived as receptive to industrial collaboration only as a means to funding. It is a fact that engineers and scientists in industry have difficulty in being received as equals in technical discussions with academics. These cultural characteristics are antagonistic to cooperative working relationships, and they can be overcome by participation in projects in which goals and rewards are shared. Thus, universities represent a U.S. resource that can provide considerable leverage in electronic packaging. Generally, the universities will be most effective in connection with the knowledge base and knowledge integration, but, where revolutionary advances occur, academic research can be translated quickly into commercial products and processes. The materials area today is attractive in this connection. [Note: Recently, the Air Force (Wright R&D Centre) initiated action on a cooperative program to determine the areas in which industry would be willing to participate. Progress has been made in stimulating interest by all parties to work together on packaging problems for mutual benefit, e.g., the MCC-Lehigh University cooperative under the guidance of R.J. Jaccodine of Lehigh University.] EMIGRATION OF TECHNOLOGY In the past three decades, the United States has moved from a world position of technical dominance to a position of insecurity and reliance on foreign supply of high-technology goods. The emergence of a world market has eliminated any protection based on geography. Manufacture of many items has moved abroad for reasons of wages, cost of supplies, and entry into foreign markets. Engineering and design are following manufacturing. A major cause is that the U.S. workforce has declined both in education and motivation. All of these factors bear strongly on the future of the U.S. standard of living. In electronics, the area of consumer products has been taken over mainly by Japan. The Japanese have also become pre-eminent in memory production and in building machines for integrated circuit manufacture. Japan is now mounting a strong effort in the computer area. On the materials side, the Japanese are self-sufficient, but the United States is not. In both ceramics and organic polymers, Japanese firms lead the United States in many important segments. Specifically, electronic ceramic materials come principally from Japan, and now the suppliers have integrated forward into the manufacture of encapsulation modules (e.g., chip carriers and pin grid arrays). They also are dominant in the supply of transfer molding epoxy compounds. In printed wiring boards, multichip modules, chip-board assembly technology, and other interconnection media, both countries appear to be equally capable. The United States will do well to study and understand the causes of its loss in relative position in these critical technologies. Studies have been conducted for other technology sectors, but the materials chain may have unique features because of its high degree of vertical and horizontal

OCR for page 95
103 fragmentation. Loss of technical capability must be arrested, and domestic production in critical areas must be restored. The present trends leave the United States vulnerable both militarily and financially. For materials, U.S. industries, universities, and technologies are fundamentally strong, but there are signs of weakness. For example, too few young people are attracted to ceramic and polymer science and engineering. Without wishing for government management of business, it would be desirable that encouragement be offered in critical technologies. PRIVATE FUNDING AND PUBLI C FUNDING The programs and consortia summarized in this chapter provide interesting examples of funding options, but many forms can be made to work. There are, however, forces at work in the United States today that work against effective long-range approaches to maintaining a high-technology industry. Funding of research and development must be kept reasonably stable. Good programs require a hierarchy of knowledge that takes years to acquire. A field can attract excellent people only if satisfying careers can be predicted with confidence. An infrastructure of communication links supports funding, development efforts, and the manufacturing and marketing that feed back the ongoing supporting funds. Once lost, these links are difficult to rebuild. Private funding in engineering and research is less robust. Many factors characteristic of U.S. business patterns today inhibit the long-range view necessary for good engineering and research. The mergers, takeovers, and resulting restructurings of major corporations are harmful to the creation of wealth and the maintenance of long-term satisfying employment within the United States. As corporations become more thinly capitalized, they become more susceptible to inevitable fluctuations in business. This leads to episodic retrenchment in long-range, high-technology segments of the business, an effect that is deadly. Government agency funding may become unstable as a consequence of the deficit, negative balance of payments, and other factors that lead to calls for cuts in the budget. The Department of Defense, however, has been remarkably consistent in its support of engineering and science, and optimism in this regard extends into the Bush administration. So, too, the Department of Energy's National Laboratory system has been well supported, and the National Science Foundation budget permits funding of a number of important technical initiatives. Thus, federal funding appears to be in reasonably good shape in spite of the negative factors mentioned. The Department of Defense has special reasons to support work on materials for high-density electronic packaging and interconnection. Military circuitry traditionally has been based on IC chips that are contained in hermetic packages. As intrachip communication becomes ever more important as a limitation on circuit performance, the single-chip hermetic packaging requirement becomes increasingly burdensome. Therefore, it is important that new approaches be found that will achieve the most advanced systems requirements without losing the long-term reliability presumed to accrue from hermetic packaging. New materials will play an important role in resolving

OCR for page 95
104 this problem, and enhanced DOD funding can be expected for materials applications in packaging and interconnection. recommended Mixed government and private funding is a healthy approach and is _ _ _. Stability can be improved through funding diversity, and national policy needs can be injected into programs at a point of leverage. should be riven to programs directed to the nation's Somehow, preference should be given to programs directed to the nation's fiscal welfare. For all the right reasons, programs in materials engineering and science, in the context of advanced electronic packaging and interconnection, deserve continuing substantial support and attention. CONCLUS ION The United States has lost control of major segments of the supply chain of materials that are essential to electronic packaging and interconnection. Furthermore, the organizational culture and legal framework of U. S . industry is such that it is likely that further erosion of the U.S. world position will occur . Thus, the supremacy ~ or even the viability) of the United States in computers, telecommunications equipment, aircraft, defense hardware, automobiles, space technology, and other key areas could well go the way of the consumer electronics business. New mechanisms and policies must be developed to support and stimulate domestic materials suppliers to work closely with electronics systems houses in long-range programs aimed at recapturing world leadership.