Microelectronics Research and Development

March 1986

NTIS order #PB86-205200 Recommended Citation: U.S. Congress, Office of Technology Assessment, Microelectronics Research and Development–A Background Paper, OTA-B P-C IT-40 (Washington, DC: U.S. Gov- ernment Printing Office, March 1986).

Library of Congress Catalog Card Number 86-600509

For sale by the Superintendent of Documents U.S. Government Printing Office, Washington, DC 20402 Foreword

Microelectronics, the fundamental! building block of today’s pervasive information technologies, has progressed at a tremendous rate over the last few decades and has become a vital part of U.S. commerce and defense. This fast-changing field depends critically on aggressive research and development (R&D) programs. Federal participation in microelectronics R&D is twofold: the government directly funds a significant fraction of the activities, and several Federal policies indirectly affect private support of R&D. Hence, Congress has a major role to play in microelectronics R&D—a role that can be illuminated by a better understanding of today’s activities in this area. In November 1984, the House Committee on Science and Technology asked OTA to prepare a background paper on microelectronics R&D as a follow-on to the OTA assessment of Information Technology R&D: Critical Trends and Is- sues, released in February 1985. This background paper describes the current state of research and development in microelectronics by examining the range of R&D efforts and the sources of Federal and private support for R&D. It also presents potential policy concerns that stem from existing arrangements for direct Feder- al support and from changes underway in microelectronics R&D. Many members of the diverse microelectronics community, spanning the spectrum from basic research to applied development, contributed their knowledge, insight, and viewpoints to this background paper. These included scientists, engineers, managers, and observers in Federal agencies and laboratories, industry, universities, and other organizations. OTA is pleased to thank all participants for their assistance. However, OTA assumes full responsibility for the contents of this study.

Ii; OTA Microelectronics Research and Development Staff

John Andelin, Assistant Director, OTA Science, Information, and Natural Resources Division

Fred W. Weingarten, Communication and Information Technologies Program Manager

Project Staff Arati Prabhakar, OTA Fellow and Analyst

Contractor Susan P. Walton, Editor

Administrative Staff Elizabeth A. Emanuel Shirley Gayheart* Patricia M. Keville Audrey Newman

*Deceased, Dec. 11, 1985. iv Reviewers

John A. Armstrong Robert W. Keyes Vice President, Logic and Memory Research Staff Member, Semiconductor IBM Watson Research Center Science and Technology IBM Watson Research Center Donald S. Beilman President Jack S. Kilby Microelectronics Center of North Carolina Dallas, Texas Arpad A. Bergh Conilee G. Kirkpatrick Division Manager, Device Science and Director, Advanced Technology Technology Research Implementation Bell Communications Research Rockwell Microelectronics R&D Center Eliot D. Cohen Fred W. Kittler, Jr. Navy Director, Very High Speed Vice President Integrated Circuits Program J.P. Morgan Investment Space and Naval Warfare Systems Paul Losleben Command Assistant Director, VLSI Larry R. Cooper Defense Advanced Research Projects Program Manager, Electronics Division Agency Office of Naval Research Noel MacDonald Kenneth L. Davis Professor, Electrical Engineering Head, Electronics Division Cornell University Office of Naval Research Kevin J. Malloy Charles B. Duke Program Manager Senior Research Fellow Air Force Office of Scientific Research Xerox Webster Research Center Thomas C. McGill James M. Early Professor, Applied Physics Scientific Advisor to the Director California Institute of Technology Fairchild Research Center Carver A. Mead Jeffrey Frey Professor, Computer Science Professor, Electrical Engineering California Institute of Technology Cornell University Robert E. Nahory Paolo Gargini District Research Manager, Optical Coordinator, One Micron Program Properties of Semiconductors Research Intel Corp. Bell Communications Research George H. Heilmeier Martin A. Pollack Senior Vice President and Head, Photonic Devices Research Chief Technical Officer Department Texas Instruments, Inc. AT&T Bell Laboratories Evelyn Hu Richard A. Reynolds Professor, Electrical and Computer Director, Defense Sciences Office Engineering Defense Advanced Research Projects University of California at Santa Barbara Agency Anthony A. Immorlica, Jr. Director, Microelectronics Center McDonnell Douglas Corp. Sven A. Roosild Jimmie R. Suttle Assistant Director, Electronic Sciences Director, Electronics Division Division Army Research Office Defense Advanced Research Projects Paul Turner Agency Manager, Technology Assessment Joseph A. Saloom Xerox Palo Alto Research Center Senior Vice President Horst R. Wittmann M/A-COM Director, Electronic and Material Sciences George M. Scalise Air Force Office of Scientific Research Senior Vice President and Chief Edward D. Wolf Administrative Officer Professor and Director, National Research Advanced Micro Devices and Resource Facility for Submicron Donald J. Silversmith Structures Director, Solid State and Microstructures Cornell University Engineering Program National Science Foundation Michael A. Stroscio Program Manager, Electronics Division Army Research Office

Other Contributors

Robert S. Bauer Dean Eastman Manager, Integrated Optoelectronics IBM Fellow Xerox Palo Alto Research Center IBM Watson Research Center Jules A. Bellisio Lester F. Eastman Division Manager, Digital Signal Professor, Electrical Engineering Processing Research Cornell University Bell Communications Research Barry K. Gilbert Robert D. Burnham Special Purpose Processor Development Research Fellow Group Xerox Palo Alto Research Center Mayo Foundation Praveen Chaudhari Kenneth L. Guernsey Vice President, Science General Partner IBM Watson Research Center Hambrecht & Quist Melvin I. Cohen Richard Henderson Director, Interconnection Technology Hughes Research Laboratories Laboratory M.P. Lepselter AT&T Bell Laboratories Director, Advanced LSI Development Robert Dennard Laboratory IBM Fellow AT&T Bell Laboratories IBM Watson Research Center Michael A. Littlejohn James V. DiLorenzo Professor, School of Engineering Vice President, Research and Development North Carolina State University Microwave Semiconductor Corp.

vi James W. Mayer Louis Toth Professor, Materials Science and Director, Ceramics and Materials Program Engineering National Science Foundation Cornell University P. K. Vasudev James H. McAlear Senior Staff Engineer President Hughes Research Laboratories Gentronix Laboratories, Inc. William D. Warters James Meindl Assistant Vice President, Communications Professor Systems and Network Technology Stanford University Bell Communications Research C. Joseph Mogab Hugh A. Watson Microwave Semiconductor Corp. Head, VLSI Processing Technology and Fabrication Department Gordon E. Moore AT&T Bell Laboratories Chairman and Chief Executive Officer Intel Corp. Benjamin A. Wilcox Assistant Director, Materials Science Stephen Nelson Division Director of Research Defense Advanced Research Projects Cray Research, Inc. Agency Gregory H. Olsen Gerald L. Witt President Program Manager Epitaxx, Inc. Air Force Office of Scientific Research Phil Reid Irving Wladawsky-Berger TRW Vice President, Systems John B. Salazar IBM Watson Research Center Vice President, Corporate Banking Rainer Zuleeg Bank of America Microelectronics Center George E. Smith McDonnell Douglas Corp. Head, MOS Device Department AT&T Bell Laboratories

OTA Reviewers

Lauren W. Ackerman Jim Dray Eugene Frankel Research Analyst Research Analyst Senior Analyst John A1ic Greg Eyring Chuck Wilk Project Director Project Director Project Director Earl Dowdy Research Analyst

vii Contents

Chapter Page Chapter Page 1. ISSUES IN MICROELECTRONICS 4.INSTITUTIONAL SUPPORT FOR RESEARCH AND DEVELOPMENT,,. 1 MICROELECTRONICS R&D ...... 19 Main Findings of the Study: Trends Federal Support, ...... 19 in Microelectronics R&D . . . . 1 Department of Defense ., . . . . . 19 Findings About Institutional Support . 1 Research for the Navy, Air Force, Findings About Technological Trends 2 and Army...... 20 Trends in Fabrication Technology .,. 2 Defense Advanced Research Trends in Design ...... 2 Projects Agency ...... 20 Potential Policy Concerns ...... 3 DOD Special Programs . . . 21 Federal Response to Changes in National Science Foundation ., . . 22 Microelectronics R& D ...... 3 Private Sector R&D ...... 23 Federal Response to Changes Captive Manufacturers . . . . ., . . ., . 23 in the Industry ...... 3 Merchant Companies. ..., . . . ., 24 Federal Response to Technological Cooperative R&D ...... 25 Trends...... 5 Universities and R&D, ...... 25 Direct Federal Support for Micro- Foreign Activities . . . . . 26 electronics R&D: Policy Questions 5 Multiple Sources of Federal Funding: Appendix Page Pros and Cons, ...... 5 A. CURRENT MICROELECTRONICS Questions Raised by Department TECHNOLOGY ...... 27 of Defense Activities ...... 6 Electronics Technology From Vacuum Tubes to Integrated Circuits 27 2. INTRODUCTION: MICROELECTRONICS Microelectronic Devices . 28 TECHNOLOGY AND R&D ...... 9 Types of Integrated Circuits. ., ., . 28 The Impact of Microelectronics and Other Semiconductor Devices . . . 31 Evolution of the Industry ...... 9 Technologies To Produce Semiconductor The Nature of Research and Development Microelectronics, ...... 31 in Microelectronics ...... 10 Circuit Design,..,...... 31 Chip Fabrication ., . . . . 32 3.CURRENT RESEARCH AND B. GLOSSARY OF TERMS ...... 35 DEVELOPMENT ACTIVITY ...... 11 Advanced Silicon Integrated Circuits, . . . 11 Table Circuits and Packaging, ...... 11 Design and Fabrication Processes . . . . 12 Table No. Page A-1. Levels of Integration . . . . 29 Circuits for Specific Markets ...... 12 Microelectronics Based on Gallium Arsenide and Other Compound Figures Semiconductors ..,..., . . . . . 12 Figure No. Page GaAs Digital Integrated Circuits ..., 13 1. Periodic Table of the Elements . 13 Optoelectronics...... 14 A-l. Generations of Electronics Discrete Optoelectronic Devices . . . 14 Technology...... , ...... 27 Integrated Optoelectronics 15 A-2. Types of Microelectronic Products . . . 28 Superlattices andOtherQuantum- A-3. Microscopic Sizes ...... 30 Effect Structures...... 15 A-4. First of Four Mask Processes Microwave Devices ..., ...... , . 16 To Make an N-Channel MOSFET...... 32 Nonsemiconductor Integrated Circuit Technologies ...... , ..., 16 Josephson Junctions ...... 16 Biomolecular Electronics...... 16

Vlll Chapter 1 Issues in Microelectronics Research and Development

Microelectronics is the cornerstone of the in- opment. These multiple involvements make it formation technologies that pervade virtually important to understand both the structure every aspect of contemporary life. These com- of institutional support and the implications puter and communications technologies are of current technological trends in considering the basis for changes such as automation, the many Federal policies that affect micro- energy conservation, and pollution control in electronics R&D. offices, factories, automobiles, and homes; su- Today, many factors, including shifts in in- percomputers for applications from weather dustry structure and limitations posed by prediction to computational research; new ca- technological trends, raise questions concern- pabilities in financial services; new means of ing the types and levels of Federal support for storing and playing back audio and video microelectronics R&D. To address these is- recordings; advanced telephone and television sues, this OTA background paper, requested systems; and complex weapons systems for by the House Committee on Science and Tech- national defense. Each of these areas is criti- nology, describes the current state of micro- cally dependent on microelectronics technol- electronics research and development by ex- ogy. Furthermore, the microelectronics indus- amining the technologies emerging from R&D try—and the industries that depend on it— efforts and the range of institutional support are vital to the U.S. economy. for R&D. Although other relevant Federal pol- Research and development (R&D) efforts icies are discussed to some extent, the primary have fueled progress in microelectronics tech- focus of the paper is the role of direct Federal nology at an extraordinary rate. Since the in- support for microelectronics R&D. vention of the integrated circuit (IC) 27 years This chapter: 1) summarizes the OTA find- ago, the capabilities of these devices have ings, and 2) discusses potential Federal pol- more than doubled every 2 years. Currently, icy implications that they suggest.1 an IC with several hundred thousand components can be purchased for a few dollars-less than the 1950’s price for a single component. The Federal Government has historically 1 In this paper, the term “microelectronics’ is used to describe miniature electronic devices in general, Readers whose back- played vital roles, some direct and some in- ground in the technology is limited may wish to review app. direct, in microelectronics research and devel- A before proceeding.

MAIN FINDINGS OF THE STUDY: TRENDS IN MICROELECTRONICS R&D

Findings About Institutional Support universities all contribute in different ways to A broad range of organizations supports progress in the field. Among international microelectronics R&D. In the United States, activities, Japanese R&D efforts predominate; Federal agencies and laboratories, private several European nations also support micro- films, cooperative research organizations, and electronics research and development. 2

Within the Federal Government, the De- 3. Because of international competition and partment of Defense (DOD) sponsors the limited resources for equipment and per- largest share of microelectronics R&D. Fed- sonnel, cooperative research efforts are eral funds from various sources support work gaining popularity as a means to bolster in Federal laboratories, industry, universities, R&D. These activities, which typically in- and cooperative organizations. volve cooperation among different industrial, academic, and Federal organiza- Private sector R&D in the United States en- tions, represent a relatively new approach compasses a spectrum of activities in both ver- to R&D in the United States. tically integrated companies and merchant 2 4. The deregulation of the telecommunica- firms. Vertically integrated companies gen- tions industry is affecting R&D in micro- erally support a full range of activities from electronics-related areas. AT&T Bell Lab- basic research to applied development, al- oratories’ efforts are becoming more though the types and levels of R&D vary closely tied to products than were re- widely from company to company. Merchant firms tend to limit their R&D to the last stages search efforts at the predivestiture Bell of development. Some companies in each cat- Laboratories. The role of Bell Communi- egory support and participate in R&D in co- cations Research is still not completely known. operative organizations and in universities in addition to their onsite efforts. Findings About Technological Trends OTA identified four basic changes occurring in institutional support for microelectronics Microelectronics can be separated into two R&D in the United States: related parts: 1) fabrication technology, including materials, devices, and circuits; and 2) chip 1. In the last few years, many factors have architecture or design. Advances are taking converged to alter the structure of the place in both aspects of the technology. microelectronics industry. Chief among these is the Japanese challenge to U.S. Trends in Fabrication Technology competitiveness, which has led to increased shares of U.S. and international Over the next two decades, the primary markets for Japanese companies and, con- technological trend in the physical structure sequently, reduced profits for U.S. com- of microelectronics is likely to be continued panies. This may be leading to decreased miniaturization of silicon integrated circuits. R&D efforts by the industry. Because this scaling down of the dimensions 2. The exceptional capability of Japanese of ICs has been the key to better and less companies to transform research concepts expensive chips, it has been the basis of into products is well established in micro- progress in microelectronics over the last 25 electronics. Now, there is also growing years. OTA found that, according to experts, evidence that Japanese basic research ef- this trend will continue for the next 5 to 10 forts are outpacing U.S. efforts in some years and then will begin to level off in approx- areas of microelectronics, e.g., optoelec- imately 15 years, when minimum dimensions tronics. are between one-tenth and one-fifth of a micron-about one-tenth of the minimum size currently in production. (A micron is one- *“Vertically integrated” in this context refers to a company that makes microelectronics to use in the products that it sells, millionth of a meter. See figure A-3 in app. A.) e.g., a computer company that makes integrated circuits for its computers. The microelectronics division of such a company OTA identified several other technological is often termed a captive manufacturer. Some vertically in- trends related to this central finding. Devel- tegrated companies also sell their microelectronic products to others in addition to using them internally. A merchant firm, opment activities expected to influence the in- in contrast, makes microelectronics primarily to sell to other dustry soon center on advanced manufactur- end users. ing techniques required to fabricate circuits 3 with smaller and smaller dimensions. These Trends in Design advances represent incremental changes in While circuits continue to shrink and begin current silicon IC technology. to reach limitations, the power of design tools, Mid- to long-term R&D efforts, which focus and hence the flexibility of IC design, will con- on significantly different technologies with tinue to grow rapidly, OTA found. Users have promise for the next generation of microelec- been limited to building systems out of stand- tronics, are centered on: ard ICs and other components in the past. Now, complex new design systems coupled ● digital and analog (microwave) integrated with advanced manufacturing capabilities al- circuits made from gallium arsenide low users to configure chips to perform spe- (GaAs), cialized tasks. Progress in this area will hinge • optoelectronics, and on R&D activities in design software. ● quantum-effect structures. Merchant manufacturers are currently pur- OTA found that few microelectronics experts suing the expanding markets for applica- expect GaAs integrated circuits to replace sili- tion-specific ICs (ASICs), which include cus- con digital ICs. Rather, they believe that the tom and semicustom chips. As the capabilities two types of ICs will meet complementary of design systems and the networks that link needs. them to manufacturing facilities expand, an The outlook for technologies based on ma- engineer will probably be able to design an IC terials other than semiconductors appears for a specific need from a workstation, trans- limited for the next few decades. R&D activi- mit the design to a silicon foundry, and receive ties in Josephson junction technology, once a the completed special chip within a short time major contender for the next generation of ICs and for low cost. This level of flexibility could for computers, are currently receiving limited open the door to a whole new genre of elec- attention. Efforts in bimolecular electronics tronic capabilities. are only exploratory today, and their promise speculative. Most experts agree that they will not come to fruition in the next few decades, if ever.

POTENTIAL POLICY CONCERNS

The state of research and development in Federal Response to Changes in the Industry microelectronics raises several potential pol- In the early days of the merchant IC indus- icy concerns about the Federal role. One set try, that part of the microelectronics commu- of issues arises from the changes occurring in nity generally tried to minimize its interaction microelectronics R&D; a second set centers on with the Federal Government. In recent years, ongoing direct Federal support for microelec- however, changes in the industry have shifted tronics R&D. the relationship between the government and merchant firms, making it necessary for each to attend more closely to the other: This sig- Federal Response to Changes in nifies wider recognition of the impact of Fed- Microelectronics R&D eral policy on the industry and a resulting gem eral trend toward increased reliance on Federal Governmental policies need to recognize the policy by the industry. changes underway in microelectronics, both in the industry’s support of and need for R&D, Because Japanese competition is the great- and in the emerging technological trends. est challenge-that U.S. merchant microelec- 4

tronics companies face today, it is the focal ● direct Federal funding for R&D, the re- point of interactions between the industry and sults of which are available to private the Federal Government. The U.S. industry companies. has, in the last few months, asked the govern- The first two types of Federal involvement are ment for help in changing the trade imbalance 3 indirect ways to make R&D investment eas- in three separate cases. Another pressing con- ier for companies; the third approach funds cern is the high cost of capital in the United R&D directly. States relative to Japan. The lion’s share of direct Federal support for The current problems that the U.S. compa- microelectronics R&D comes from the Depart- nies face pose a paradox. Although R&D, a ment of Defense, and is therefore driven by long-term investment, cannot solve industry’s military requirements. In general, DOD-spon- immediate problems, it is a crucial ingredient sored basic research serves both military and in industrial competitiveness. Without con- commercial goals. Development activities for tinued strength in R&D, solutions to the near- military microelectronics, in contrast, do not term problems will only delay the decline of overlap completely with activities for commer- the U.S. companies. Yet microelectronics firms cial needs. For example, low-cost, high-volume that are struggling to survive are likely to ne- production capabilities are a high priority for glect R&D activity in the face of more imme- the commercial manufacture of integrated cir- diate and pressing problems. For example, cuits, but the major DOD program to advance they may find it difficult to justify funding IC technology, the Very High Speed Inte- R&D while cutting jobs at an unprecedented grated Circuit (VHSIC) program, focuses on rate. This may lead to a deterioration of the design and fabrication of a small volume of industrial R&D base. highly specialized ICs for use in military Federal policies that affect the amount of systems. R&D available to private companies can be The DOD style for funding microelectronics categorized as follows: R&D, characterized by long-term investment ● policies that generally strengthen the in R&D with a clear connection to end uses, companies by making them more com- appears to have been highly successful for petitive, and thus assuring sufficient prof- achieving military goals. Some members of the its to support R&D (e. g., international microelectronics community would, therefore, trade policies and mechanisms to lower like to see the Federal Government aim a sim- capital costs); ilar level of support at commercial needs—a ● policies to ease the financial burden or point of view that has gained momentum in lower the risks of private R&D (e.g., the the face of the current pressures on the indus- tax treatment of R&D4 and intellectual try. They cite the influential Japanese Min- property protections); and istry of International Trade and Industry (MITI) as a useful model. But opinions in the ‘Semiconductor Industry Association, “Japanese Market Bar- riers in Microelectronics: Memorandum in Support of a Peti- microelectronics community diverge sharply tion Pursuant to Section 301 of the Trade Act of 1974 As on this topic. Opponents of further direct Fed- Amended, ” June 14, 1985; “Japanese Accused on Chips, ” New eral involvement in microelectronics R&D for York Times, June 27, 1985, p. Dl; and “EpROM Makers File Claim Against Japanese ‘Dumping’, ” Electronic News, Oct. 7, the industry argue that the results of R&D will 1985, p. 1. meet commercial needs and will be available ‘Since 1981, firms have been allowed tax credits on increased to industry only if the commercial sector expenses for research and development, enhancing the R&D directs and carries out the work itself. They spending power of these companies. This R&D tax credit ex- pired at the end of 1985. view the MITI model as unacceptable in the ‘Pressure from the microelectronics industry spurred Con- American context, given the vast differences gress to legislate a new form of intellectual property protection in the Semiconductor Chip Protection Act. The act, signed in industry-government relationships between into law in November 1984, protects masks (used to fabricate Japan and the United States. ICs) from unauthorized copying, and so offers manufacturers an additional incentive to come out with new products by pro- One example of a plan to start to bridge this tecting returns on R&D investments. difference of opinion comes from the Semicon- —

ductor Research Corp. (SRC), a cooperative This has already occurred in the case of R&D organization directed by microelectron- GaAs digital integrated circuits research led ics industry leaders. SRC currently supports by the Defense Advanced Research Projects R&D at universities with funds from its mem- Agency (DARPA). The microelectronics com- ber companies. In the next few years, it may munity had conducted relatively little research ask the Federal Government to match this lev- on integrated circuits made from gallium el of support. The doubled budget would be arsenide before 1982 when DARPA decided, administered solely by SRC, which would soothe based on the results of 6 years of GaAs basic at least the concerns of SRC’s member com- research that they had sponsored, to fund a panics about the selection of research topics series of pilot production lines to demonstrate and the availability of results.6 the feasibility of using GaAs for ICs. This an- nouncement kindled the interest of a variety. Federal Response to Technological Trends of organizations, and several defense elec- Limitations to growth that stem from tech- tronics firms mobilized to build a base of activ- nological trends are less immediate than the ities in GaAs so that they could be involved economic problems that the industry faces, with the pilot lines. Three new commercial ventures have already spun off from this work. but they too pose questions about the ap- Perhaps even more significantly, there is some propriate Federal role. The shrinking of circuitry on silicon chips, on which progress has evidence that DARPA’s interest in GaAs ICs hinged thus far, has required enormous inno- may have helped to convince research organi- vation, chiefly in manufacturing technology zations such as AT&T Bell Laboratories that and engineering exploitation of the concepts the field deserves an intensive research effort. for transistors and circuit integration. But At about the same time, IBM turned its focus for alternative chip technology from progress in fabrication technology beyond the limits of silicon scaling will demand a wider Josephson junctions to GaAs. In part because range of more basic R&D activities. of DARPA’s initiative, the activities in GaAs ICs grew in a few years from a handful of iso- This technological factor may drive ex- lated efforts in individual laboratories to large panded Federal participation in R&D for po- programs sponsored by Federal agencies, in- tential alternative microelectronics technol- dustry, and universities. ogies. This could take many forms, including policies to encourage basic research in indus- Direct Federal Support for try or greater direct Federal R&D funding. Al- Microelectronics R&D: Policy Questions ternatively, Federal agencies may select specific areas in which to focus support without Ongoing direct funding of microelectronics significantly increasing their total funding R&D from Federal agencies continues to raise levels. Because the Federal Government funds policy concerns. The system of multisource a wide range of microelectronics research ef- support, with several different Federal agen- forts and interacts with a variety of R&D orga- cies funding microelectronics R&D activities, nizations, it can exert considerable leverage poses potential policy questions. And since the in key areas. In these areas, some Federal largest amount of support comes from DOD, agencies might be able to lead the way for ef- many of the concerns center on the implica- forts by participants in the commercial sector, tions of defense R&D spending. universities, and other Federal agencies by tar- geting R&D monies. Multiple Sources of Federal Funding: Pros and Cons Microelectronics, like other science and engi- ‘Interview with George M. Scalise, Advanced Micro Devices neering fields, receives R&D funding from sev- and Semiconductor Research Corp., June 1985. The doubled budget could total as much as $100 million per year, although eral different Federal agencies. Because a De- SRC’s current annual budget is only approximately$15 million. partment of Science or some other scheme for 6 centralized R&D funding is proposed and con- 2. the impact of the Strategic Defense Ini- sidered from time to time, it is important to tiative (SDI) on the structure of DOD mi- examine the pros and cons of the current croelectronics R&D. multisource arrangement. In microelectronics- Keeping information about new defense related areas, the system generally avoids po- technologies within the United States is a crit- tential pitfalls and offers advantages over a ical concern for military security; free and wide centralized system. exchange of ideas and results is an equally The potential drawbacks of the multisource crucial ingredient in scientific research. This system do not pose problems at present. One dilemma is the basis of an ongoing discussion possible difficulty is wasteful duplication of among players both within and outside DOD, effort and competition for resources if the vari- all of whom are trying to determine the appro- ous agencies do not communicate their plans priate type and level of control of defense re- to one another. The present system, too, could search results. Controls on universities, where confuse or inconvenience researchers seeking many foreign students are involved in scien- funding, particularly newcomers. However, tific research on campus, are of particular con- the researchers within an area typically com- cern. Some leading universities have banned municate with each other and are familiar with classified research on campus altogether as the full range of activities in their area. These a partial solution. Several years of debate re- informal infrastructures prevent most un- cently resulted in a policy from the White necessary duplication of effort and alert re- House (the National Security Decision Direc- searchers to the relevant funding sources in tive) which makes classification the only mech- the area. In addition, the agencies coordinate anism for control of fundamental research (i.e., R&D funding through both formal and infor- unclassified research may not be restricted).8 mal channels. The policy does not solve the problems com- The advantages of distributed funding pletely. It will, however, greatly simplify the process of deter across agencies more than compensate for the mining control by reducing the number of gray areas. potential problems. The arrangement provides researchers with multiple channels for Federal R&D funding under SDI raises two types support for promising new ideas; they can turn of concerns about DOD research: the impact to a second agency if a proposal is refused by on the structure of DOD funding in micro- the first. The present system also permits each electronics R&D, and further questions about agency to fund R&D to meet its own goals or the treatment of university research. those of its parent department. The existing The initiative’s activities in this area may situation, with several loosely coordinated in- involve a major restructuring of the funds for dependent agencies funding different aspects microelectronics R&D from the various DOD of microelectronics R&D, appears to serve its agencies, whether or not it increases the over- purpose well. all level of DOD support. The transfer of the Questions Raised by Department of Defense GaAs IC pilot lines from DARPA to the SD I Activities Organization (SD IO) is early evidence to support this possibility. Given the wide percep- Beyond the questions of the balance be- tion that the current arrangement for DOD- tween R&D for military needs and R&D for sponsored research (with several different commercial needs, DOD activities raise two agencies operating independently but commu- sets of potential Federal policy issues. These are: 71nstitute of Electrical and Electronics Engineers, “DOD’s 1. DOD control of research and develop- Perle Questions Value of Open Research on Campus,” The In- stitute, July 1985, p. 10. ment, particularly university research 8“White House Issues Secrecy Guideline, ” Science, vol. 230, activities; and Oct. 11, 1985, p. 152. 7 -. nicating with each other) works well, central- recent policy stating that SD I research at ized funding of microelectronics R&D through universities will be considered ‘‘fundamental SDIO could decrease DOD’s effectiveness in research, ” which is to be treated in accordance the field. with the new National Security Decision Directive, may have alleviated some of these The fact that SD I-funded activities are des- 0 concerns. ‘ ) On the other hand, the fact that ignated as “advanced technology develop- SDI’s Innovative Science and Technology Of- ment” rather than “research” has exacerbated fice received approximately 2,700 preliminary the concerns about DOD controls on univer- proposals from university researchers over a sity research. There has been concern that period of just 3 months]’ is strong evidence DOD would censor dissemination of all SDI of interest from that community. work, including university activities. From the perspective of a group of university scientists boycotting SDI, “the likelihood that SDI 10] nstitute of E ]ectric~ and Electronics 1? ngineers, ‘‘ SI~ I funding will restrict academic freedom . . . is Memo Bars Controls on Most ‘Star Wars’ Research in Univer- greater than for other sources of funding. ” (A sities, ” The Institute, October 1985, p. 1, ‘lDwight Duston, Innovative Sciehce and Technology office, SDI organization, personal communication, Janu~ 1986. This ‘From the ho~cott petition, as quoted in “Star Wars 130y - number includes preliminary’ proposals in areas other than cott Gains Strength, Science, vol. 230, Oct. 11, 1985, p. 152. microelectronics. Chapter 2 Introduction: Microelectronics Technology and R&D

THE IMPACT OF MICROELECTRONICS AND EVOLUTION OF THE INDUSTRY

Microelectronics technology has dramati- The dramatic growth of the technology has cally improved the capabilities of computers prompted observers to describe it as the mi- and communications systems, while also fuel- croelectronics revolution. To date, the minia- ing the growth of completely new applications, turization of circuitry, which leads to products such as personal computers. Growing ex- that perform faster and better, has been chief- tremely fast and providing increasingly power- ly responsible for this revolution. Shrinking ful and inexpensive tools to manipulate elec- the electronic devices has yielded lower cost, tronic signals, microelectronics has become the expanded performance, and higher reliability. cornerstone of information technologies. It is By any measure—cost for a given function, central to such areas as: complexity of circuits, performance—integrated circuit (IC) technology has progressed ● computers, from powerful supercom- at a tremendous rate since its inception. In puters, through business computers, to 1964, Gordon E. Moore predicted that the inexpensive personal computers; number of components on a chip would con- ● communications systems, including tinue to double annually, as it had since the switching stations and satellite communi- beginning of that decade. Twenty years later, cations; technology has not departed significantly ● consumer products, such as electronic from Moore’s law. Experts predict a slowing watches, video games, and pocket calcu- of the trend over the next 10 to 20 years. lators; ● control systems for industrial applica- At the same time, IC design will begin to tions, automobiles, and home appliances; alter the way that system engineers use ICs. and New capabilities provided by design software, ● military systems for national defense. silicon foundries, and the networks that link them allow the end user much greater flexi- Microelectronics has become a vital part of bility in designing chips for a specific appli- U.S. commerce and defense. In both sectors, cation. This trend, coupled with changes in maintaining a technological edge over the rest fabrication technology, will affect the industry. of the world is the only way to ensure security—whether military or economic. U.S. com- Many other factors, more immediate than panies claimed approximately $14 billion of the scientific and engineering changes, affect the $26-billion world market for microelec- the structure of the American microelectronics tronics in 1984. I The development of increas- industry, especially merchant firms. These in- ingly sophisticated weapons systems means clude international competition, capital re- that virtually every aspect of current military quirements, capital cost, and shifting mar- technology depends on microelectronics. kets. U.S. microelectronics companies currently face debilitating competition from Japanese sources, and other Asian countries (e.g., Ko- 1“The ,Japan(~sc, Semiconductor hlarket and S1 ~1 :10 1 I’uti- tion, ” presentat ion 1)~’ (;eorge Scalist’ for th~’ Sen]i~f)nduct{)r rea) are also preparing to enter the market. At 1 ndustr} Ass{)t’lation, J ul~ !4, 19X5. the same time, the industry is growing stead-

9 10 ily more capital-intensive, since every techno- Together, these technological, international, logical advance demands more complex equip- and economic trends and forces are altering ment and production facilities. The high cost the microelectronics industry. The changes of capital exacerbates this problem. Finally, mark the beginning of the maturing of the in- although many of the markets for uses of dustry and may signal the beginning of the microelectronic products are growing, some of end of an era of apparently limitless growth. them grow explosively and then plummet as rapidly, as did the video games market. These shifts cause instability in the industry.

THE NATURE OF RESEARCH AND DEVELOPMENT IN MICROELECTRONICS

Since microelectronics has expanded swiftly ● making ICs with smaller component de- and will continue to do so, research and devel- vices (i.e., smaller transistors, resistors, opment (R&D) activities play an extremely capacitors, interconnections) and packing crucial role in its progress. As the technology these devices closer together on the chip; and the industry mature, however, the reasons ● using larger wafers so that more chips can for supporting R&D may shift. For example, be made on a single wafer; the approaching “post-shrink” era (beyond the ● using new processes and new equipment limits of miniaturization for silicon integrated in chip fabrication; circuits) may demand renewed vigor in basic ● packaging chips in increasingly sophisti- research to find a successor to silicon ICs. cated ways; ● designing increasingly complex circuits; To suit the needs of those who use the prod- and ucts based on microelectronic devices, R&D ● designing circuits that are tailored to spe- efforts in microelectronics are aimed at mak- cific applications. ing circuits that: Scientists and engineers involved in longer ● cost less, term R&D are trying to anticipate technolog- • operate at higher speeds (higher fre- ical needs beyond the current generation of quencies), products. Their concerns center on topics such • require less power and generate less heat, ● are more reliable and last longer, and ● carry out specific functions. new materials for microelectronics (e.g., gallium arsenide); These goals are prioritized in different ways new devices to replace or augment tran- depending on the end use. For example, low sistors (e.g., quantum-effect devices); cost is typically the chief concern for chips to new equipment to fabricate these mate- be sold in high volume, but high speed may rials and devices; be the dominant requirement in making com- integrating optical and electronic cir- ponents for supercomputers, and concerns cuitry; and about power consumption may dominate for advanced design tools. ICs intended for use in satellites. For the microelectronics manufacturer, in the short term, these requirements translate into:2 2App. A: Current Microelectronics Technology provides some general background information on integrated circuit technology, and App. B: Glossary of Terms gives definitions for technical terms. Chapter 3 Current Research and Development Activity

Microelectronics research and development Design activities span all three categories. (R&D) activities can be separated into three Most of the work described here is aimed at categories: making better digital integrated circuits. Other semiconductor activities, such as opto- 1. activities to improve silicon integrated electronics and microwave devices, are also in- circuits (ICs), cluded here because they are merging to some 2. efforts for compound semiconductor mi- degree with IC technology and because all croelectronics (primarily based on gallium semiconductor R&D shares a common base of arsenide (GaAs)), and physical understanding and process tech- 3. investigations for integrated circuits nology. based on materials other than semiconductors.

ADVANCED SILICON INTEGRATED CIRCUITS

Efforts now underway to improve silicon- must lay out both the devices and the inter- based microelectronics will be the first type connections in more and more complex pat- of R&D to have practical applications. These terns. Finally, the package for the completed efforts can be grouped in three categories of chip must allow signals to enter and leave the simultaneous activities: chip at high speed, so that the package itself does not obliterate the speed advantage of the 1. the improvement of the physical circuits new circuitry. and packaging of integrated circuits, 2. the facilitation of the design and fabrica- These steps have been used to scale down tion processes, and silicon ICs over the past 25 years. Every new

3. the design of new types of ICs for specific reduction in feature size has been significantly markets. more difficult to achieve than the last, and progress has been possible only through the introduction of increasingly complex manufac- Circuits and Packaging turing technologies and device and circuit de- signs. Today’s R&D workers face the great- The process of reducing the size of devices est challenges yet. in ICs and increasing their packing densities has several parts. Scientists and engineers are Few trends, however, can continue forever. developing devices–transistors, resistors, ca- The remarkable feature of Moore’s law (the an- pacitors–that have feature sizes of less than nual doubling of the number of components 1 micron. Despite their small size, these de- on a chip) is the range over which it extends vices must be designed to operate correctly before meeting unavoidable limits. Two tech- and to control the required amount of power. nological factors limit growth. The sizes of the The interconnections required to hook the de- individual devices and the separations be- vices together are also becoming increasingly tween them eventually become so small that harder to make. Each connection must shrink the devices cannot function as desired. In in width but still conduct electrical current some instances, these dimensions are a few with virtually no resistance. The interconnec- dozen atom layers. The problems involved in tions must lie closer together but still be com- interconnecting the devices on a chip also pletely isolated from each other. Designers become virtually insurmountable. Together, 11 12 these make a relaxation in the rapid growth photolithography for better definition of of microelectronics technology inevitable. ultrasmall features. Microelectronics experts do not agree on the Manufacturing technology is a crucial part details and consequences of this slowdown, of these advances because progress in silicon but they generally do agree that these limits scaling is based on the introduction and im- will be reached during the next 10 to 20 years. provement of highly sophisticated equipment. Some examples are chemical-vapor-deposition Design and Fabrication Processes (CVD) and evaporation systems to grow thin Activities to facilitate the IC design proc- films of semiconductor crystals and metals; li- ess are focused on design tools, which simplify thography equipment and plasma etchers to circuit layout for IC designers. Particularly as define the ultrasmall features of the IC; and the design process has grown more complex furnaces and ion implanters to introduce the to accommodate the millions of devices, com- proper impurities to the wafer. puter-aided design (CAD) systems have become virtually indispensable. Currently, there Circuits for Specific Markets is no single standardized CAD system; rather, Currently, most ICs fall into a few stand- there are several different systems built by ard categories: logic chips, memory chips, and different groups of designers. As these sys- microprocessors. However, as the design and tems evolve, they will simplify the design proc- manufacturing capabilities of the IC industry ess and thereby give a wider range of users grow and become more flexible, a range of spe- great flexibility in creating new chips. cialized integrated circuits will play a more Fabrication technology includes the proc- central role. Application-specific ICs (ASICs) esses for depositing very thin layers of differ- are projected to grow from their current 12 to ent metals, insulators, and semiconductor ma- 15 percent of the total IC market to 25 to 30 terials on the silicon substrate; changing the percent of the 1990 market. ’ This category of impurity content in the semiconductor; etch- integrated circuits includes custom chips, ing the layers; and defining small features in which are designed from scratch for the par- the layers through lithography. Current R&D ticular application, and chips that can be activities are exploring better techniques to adapted by the user for the specific need. Fur- carry out each of these tasks. For example, ther enhancements of the design process will densely packed circuits may require new ma- expand users’ ability to design their own ICs. terials with special properties—high electrical conductivity, chemical stability, particular crystal structure —for interconnections. Also, ‘“A Chip Business That Is Still Growing: Innovation Spurs x-ray, electron-beam, or other lithographic Market for Application-Specific Integrated Circuits, ” Elec- techniques may be needed to replace current tronics, July 22, 1985, p. 40.

MICROELECTRONICS BASED ON GALLIUM ARSENIDE AND OTHER COMPOUND SEMICONDUCTORS From the vantage point of the chemist or umns (II-VI compounds such as cadmium tel- physicist, there is a logical progression of semi- luride). (See figure 1. In this terminology, the conductor materials in the periodic table from Roman numerals refer to the columns on the silicon (a column IV material), to compound right side of the periodic table; e.g., “III-V” semiconductors made from the columns refers to columns 111A and IVA. ) This pro- adjacent to silicon (II I-V compounds such as gression is also useful for classifying the range GaAs), to compounds made from the next col- of R&D activities in semiconductor microelec- 13

Figure 1 .—Periodic Table of the Elements

NOTE Elements n boldface are those commonly used 10 make semiconductor materials SOURCE Off Ice of Technology Assessment tronics, since the most immediate develop- The typical devices made from GaAs operate ment efforts focus on silicon, and longer term faster and consume less power than silicon de- work usually focuses on compound materials. vices. They are also less likely to malfunction in the presence of radiation. However, since GaAs and other 111-Vs are now the basis for a compound semiconductor is intrinsically a variety of discrete microelectronic devices, more complicated than a single-element semi- as described in appendix A. Current R&D efforts involving these materials focus on.: 1) conductor (silicon), GaAs is a much more difficult material to grow, to handle, and to use GaAs digital integrated circuits, 2) advanced to fabricate reliable devices. optoelectronic devices, and 3) monolithic microwave devices. R&D on design tools, which At present, many barriers impede prospects is critical in silicon technology, is also very im- for making GaAs ICs at production capacity portant for these alternative technologies. compared with silicon ICs, which are readily produced in quantity. Silicon ICs are currently fabricated on wafers with 5- or 6-inch diam- GaAs Digital Integrated Circuits eters, while 3-inch wafers are the largest current size for GaAs. Standard silicon wafers Digital integrated circuits based on mate- have far fewer defects and are less brittle than rials other than silicon continue to receive the best GaAs wafers. Some processing steps attention in the research community. GaAs- for silicon IC technology can be adapted based integrated circuits are currently the directly for GaAs ICs–portions of the litho- leading contender for next-generation technol- graphic procedure, wafer handling, clean-room ogy. Even so, it is important to note that vir- requirements. But the steps involving other tually no experts believe that GaAs ICs will materials, such as oxides and other insulators, usurp the position of silicon for most appli- metals, and polycrystalline semiconductor cations. material, must be developed specifically for Gallium arsenide has several intrinsic phys- GaAs. This requires a more complete under- ical properties that distinguish it from silicon. standing of the chemistry and physics of the interfaces between these materials. In addition sensor applications. R&D activities in opto- to all of these fundamental difficulties, GaAs electronics fall into three categories: IC technology suffers because experience with 1. advanced discrete light sources and de- it is very limited, relative to silicon. Many of tectors, the problems, however, will probably be solved 2. integrated optoelectronics, and as experience accrues. 3. superlattices and other quantum-effect The list of organizations supporting (and not structures. supporting) R&D in this area reveals quite The II I-V compound materials used for op- clearly the microelectronics community’s toelectronics include GaAs, iridium phosphide views on the applicability of GaAs ICs. DOD (InP), gallium phosphide (GaP), aluminum ar- is the leading Federal supporter of research in senide (AlAs), iridium antimonide (InSb), and this field, because military applications, par- alloys of these materials, such as aluminum ticularly in space, require the properties GaAs gallium arsenide (AIGaAs), iridium gallium ar- offers: high speed for large-scale signal proc- senide (InGaAs) and iridium gallium arsenide essing, low power to minimize bulk and energy phosphide (InGaAsP). Similarly, important II- consumption, and radiation hardness for relia- VI materials include cadmium telluride (CdTe), bility in the presence of radiation. Major com- mercury telluride (HgTe), and their alloy, mer- puter and communications companies-e. g., cury cadmium telluride (HgCdTe). These ma- AT&T, IBM, and several Japanese companies terials are designated as binary, ternary, or —are also investigating GaAs ICs, primarily quaternary depending on the number of ele- for use in the parts of their systems that re- ments found in them. The alloys actually rep- quire the highest speed, e.g., computer front resent a range of materials; for example, half ends. Supercomputer companies, most nota- the atoms in HgCdTe must be tellurium, but bly Cray, are attempting to make super- the other half may be any combination of mer- computers based on GaAs ICS. However, the cury and cadmium atoms. The properties of standard merchant chip makers (e.g., Intel, the alloy generally lie between the properties Fairchild Semiconductor, Advanced Micro De of the binary materials that compose it. The vices), which tend to concentrate almost ex- particular composition of an alloy is typically clusively on short-term development activi- chosen to have a certain desired wavelength ties, have no onsite efforts in materials other response. The standard approach to fabricat- than silicon. Some of these companies support ing optoelectronic devices is to grow thin longer term projects, including GaAs work, at layers of ternary or quaternary alloys on a sub- universities and through cooperative research strate of a binary material. organizations. This balance of support indi- cates two things: Discrete Optoelectronic Devices 1. GaAs digital ICs are beginning to find niche applications in a variety of areas, The first optoelectronic devices for fiber and optic communications were made of GaAs 2. they will probably not make a significant and AlGaAs. The best current devices, how- dent in the standard IC components mar- ever, are based on different compositions of ket for several years. InGaAsP grown on substrates of InP, structures that generate and detect light over a Optoelectronics range of wavelengths that includes those of lowest loss (1.55 microns) and lowest disper- As described in appendix A, compound sion (1.3 microns) in optical fibers. Since these semiconductors and their alloys are currently devices are relatively new, the materials and used to make devices that convert electrical processing problems have not been completely signals to light signals and vice versa. The de- resolved. R&D efforts in this area focus on vices are used for optical communications and making devices more reliable and more toler- 15 ant of extreme environments, achieving more Superlattices and Other precise control of the generated light, and de- Quantum-Effect Structures veloping production processes (with high By using sophisticated techniques to de- throughput and yield) for the devices. posit materials very precisely, crystal growers The II-VI compounds are used to fabricate can make layers as thin as a few layers of long-wavelength infrared sensors because atoms on a semiconductor substrate. Such a these materials (especially HgCdTe and CdTe) layer is approximately one-thousandth of a mi- z are sensitive to the wavelengths of inter- cron thick-one-billionth of a meter. Struc- est—from around 1 micron to the 10- to 12- tures called superlattices are formed by grow- micron range. Currently, research on these de- ing alternating ultrathin layers of two vices centers on materials properties, which different materials, e.g., GaAs and AlGaAs. are much more difficult to control in the II- Special methods such as molecular-beam VI compounds than in other semiconductors. epitaxy (MBE) and metal-organic chemical vapor deposition (MOCVD) are necessary to Integrated Optoelectronics achieve this extreme level of control during crystal growth. These advanced techniques Individual optoelectronic devices can be in- open a completely new set of options for tegrated and fabricated on a single substrate semiconductor materials. Varying the thick- much as standard electronic devices are in- ness of the layers and their composition can tegrated on a chip to make an IC. An inte- yield a superlattice material with different grated optoelectronic device, typically built on electronic and optical properties. Quantum a substrate of InP or GaAs, may be composed mechanical effects dominate the behavior of of lasers, devices to amplify and modulate the the electrons in superlattices because the light signals, and light detectors. In addition, structures have such small dimensions. Thus, the same chip may have purely electronic de- the electron transport processes can be dra- vices that process electrical signals. Such an matically different from the processes in nor- integrated chip has all the advantages of a con- mal material. ventional integrated circuit—miniaturization, high speed, low power, fabrication reliabil- Currently, the range of research efforts in ity—and also solves the alignment and vibra- this area is very wide, spanning the spectrum tional-stability requirements for the optical from work in designing better systems for elements it comprises. In addition, it brings growing these precise layers to making devices optical and electronic devices so close together based on superlattices. The devices include that signal delays between them are minimized, III-V and II-VI photodetectors, lasers, and allowing high speeds to be achieved. transistors. The greatest overall contribution of this new breed of materials will probably The basic concepts for integrated optoelec- stem from the fact that they can be tailor- tronics may be extrapolated even further in made for a particular application. Already, the future. Highly advanced crystal growth they are heralded by observers from diverse and processing techniques, currently in their vantage points as one of the most exciting infancy, could also open the door to the pos- 3 areas of research today. sibility of structures that would combine silicon, III-V, and II-VI devices on a single substrate. Such a scheme would allow the flexibility to use the optimal material for each por- tion of the complete circuit. For example, a circuit on a single chip could be composed of ‘F’igure A-3 in app. A shows how small a micron is. InGaAsP lasers, HgCdTe light detectors, and ‘For example, National .4cademy of Sciences, Committee on GaAs and silicon digital logic and memory cir- Science, Engineering, and Public Policy, The Outlook for Science cuits. Concepts of this sort are still in the spec- and Technolo~’ 198,5; and CJeorge H. Heilmeier, “Microelec- tronics: End of the Beginning or Beginning of the End?” In- ulative stage today. ternational Electron L)e\’ices Meeting, December 198L1. 16

Microwave Devices and reliable microwave circuitry for applications in radar, transmission of television and Analog microwave devices, operating at fre- telephone signals, and spectroscopy. A couple quencies from approximately 1 to 60 gigahertz of companies have MMICs on the market al- (a gigahertz is 1 billion cycles per second), are ready, and others plan to market them soon.4 commonly made from GaAs to take advantage Most recently, the Department of Defense of the high speed that the material offers. Cur- (DOD) announced that it will launch a major rently, monolithic microwave integrated cir- new initiative for MMICs for defense systems. cuits (MMICs) are being developed. These cir- The new program will be analogous to the cuits combine various microwave devices on Very High Speed Integrated Circuit (VHSIC) a single substrate, typically GaAs. The devices program, which addresses digital IC technol- are not packed as densely as silicon devices ogy for DOD. on a conventional digital IC, but the smallest dimensions are about the same as or smaller than those for silicon ICs—below 1 micron. 4“MMICS Save Space, Increase Reliability, and Improve Per- MMICs will fill the demand for more compact formance, ” Electronics Week, May 20, 1985, p. 52.

NONSEMICONDUCTOR INTEGRATED CIRCUIT TECHNOLOGIES Integrated circuits may also be based on ma- most notably at IBM. However, IBM halted terials other than semiconductors. Circuits all but its basic research effort in this area sev- made of superconducting Josephson junctions eral years ago because packaging, manu- have been heavily investigated but are not at facturing, and reliability difficulties meant present expected to find any large-scale appli- that the systems could not operate as well as cations in digital microelectronics. Bimolecul- had been originally predicted. Furthermore, ar electronics is currently only a highly spec- while the superconducting technology was ulative field. struggling to get on its feet, silicon and gallium arsenide technologies kept progressing Josephson Junctions at a tremendous rate. IBM’s cutbacks in this area symbolized a significant change in the A Josephson junction is an electronic device microelectronics community’s view of IC tech- made by sandwiching a very thin insulator be- nology beyond silicon. Currently, although tween two superconductors—materials with some work on Josephson junctions continues zero electrical resistance at very low temper- in the United States and in Japan (especially atures. Like electronic devices made from efforts aimed at making high-speed analog cir- semiconductors, Josephson junctions can cuits), these devices are considered unlikely switch or store electronic signals. Despite candidates to pickup where silicon leaves off drawbacks such as extremely low operating in digital ICs. temperatures, they offer several advantages over semiconductor microelectronics. Joseph- son junctions can switch signals at unparal- Bimolecular Electronics leled speeds, require very little power, and can In contrast to semiconductor and Joseph- be scaled down to extremely small dimensions. son junction technologies, the concept of using Computer systems based on Josephson- biological systems to process electrical signals junction technology were intensively re- is still in its infancy. Some researchers are searched and developed for over two decades, turning their attention to biological systems 17 as the extrapolation of the trend to smaller in the fabrication of extremely small struc- and smaller devices leads to molecular-scale tures on semiconductors, and in electronic sen- structures. sors for medical applications. But the potential for computers based on bimolecular At present, large-scale bimolecular elec- circuitry has not yet been demonstrated. tronics is a largely speculative area. Some experts envision-a role for biological materials .

Chapter 4 Institutional Support for Microelectronics R&D

Because microelectronics is a key commer- private, Federal, academic, and cooperative cial and military technology, support for re- groups are active in the area. Other nations search and development comes from many also support microelectronics R&D; Japan in sources. In the United States, a multitude of particular is a growing presence in the field.

FEDERAL SUPPORT The largest block of Federal funding for cally dependent for its operation on semicon- microelectronics R&D comes from the Depart- ductor integrated circuits. ’ ment of Defense (DOD). The National Science Other microelectronic devices, such as sensors, Foundation (NSF) funds basic research in are also becoming increasingly important in areas related to microelectronics. Several other military technology. Federal agencies and laboratories have programs to meet their specific needs, among The defense community, like other users them the National Aeronautics and Space of microelectronics, requires high-speed in- Administration, the National Institutes of tegrated circuits that consume little power. Health, the National Bureau of Standards, Since U.S. defense policy has long been based Lawrence Livermore National Laboratory, on technological superiority, DOD needs the and Los Alamos National Laboratory. best possible signal processing and sensor capabilities. Military end users have specific Department of Defense needs as well, including products like ICs that are immune to damage from radiation. DOD support was crucial to early successes Each of the three services—the Navy, Air in microelectronics. For example, military re- Force, and Army—supports activities in quirements for smaller circuits were the pri- microelectronics R&D in DOD laboratories, mary force behind the development of the first such as the Naval Research Laboratory, the integrated circuits (ICs). While DOD has con- Air Force Avionics Laboratory at Wright- tinued to fund microelectronics R&D exten- Patterson, and the Army Electronic Device sively and remains the largest source of Fed- and Technology Laboratory at Fort Mon- eral support for microelectronics R&D, mouth, and through research agencies. Apart support from private companies has assumed from the services, DOD’s Defense Advanced a much larger role, and includes an entire in- Research Projects Agency (DARPA) is heav- dustry that has sprung up in the last few dec- ily involved in this area. Certain aspects of ades. Today, military applications constitute microelectronics R&D are also covered under approximately 10 percent of the total sales of a few DOD special programs, such as the Very microelectronic products. Microelectronics High Speed Integrated Circuits (VHSIC) pro- technology is crucial to DOD, and its impor- tance will grow further in the foreseeable future. According to two experts, IAlton L. Gilbert and Bruce D. McCombe, “Joint Services Electronics Program: An Historical Perspective, prepared for . . . [i]t is safe to say that there is not a sin- the U.S. Army Research Office, Electronics Division, April gle western military system that is not criti- 1985.

19 20 — gram and the Strategic Defense Initiative Defense Sciences Office (DSO) supports mid- (SDI). to long-term work on materials, processes, devices, and circuits. Another branch, the Infor- Research for the Navy, Air Force, and Army mation Processing Techniques Office (IPTO) The Office of Naval Research (ONR), the Air sponsors activities in very large-scale integra- Force Office of Scientific Research (AFOSR), tion (VLSI) design and architecture and some and the Army Research Office (ARO) handle production automation work. scientific research for their respective In fiscal year 1985, DSO had a budget of ap- branches of the military. All three fund sub- proximately $34 million for microelectronics, stantial amounts of basic research in micro- of which $28 million supported basic research electronics, with large shares of these funds and the remainder funded exploratory devel- going to universities. Their areas of interest opment. About 40 percent of the total budget center on materials and devices. Because the went to universities for basic research. DSO three agencies are organized in different ways, sponsors investigations of gallium arsenide a completely accurate comparison of their lev- (GaAs) and other II I-V compounds and their els of support is not possible. However, ap- alloys, and work on II-VI compounds, such as proximate annual budget figures for microelec- mercury telluride and cadmium telluride, and tronics-related work sponsored by AFOSR, their alloys. Many of these materials have in- ONR, and ARO are $24 million, $13 million, teresting combinations of properties that can 2 and $9 million, respectively. be exploited for new applications. DARPA Part of the funding from the three services’ also supports research on the integration of research offices goes to support the Joint Serv- biological materials and semiconductor elec- ices Electronics Program (J SE P). They spon- tronics for ion sensors and other devices. sor the program jointly, with each office con- As of fiscal year 1985, DSO assumed respon- tributing approximately one-third of the sibility for the administration of the GaAs IC funding. JSEP is a 38-year-old DOD program pilot lines and related activities for SDI. These that funds electronics research, including efforts were transferred to SD I from DARPA microelectronics, at a group of universities. It at the end of fiscal year 1984. SDI will spend is designed to provide its 12 member universi- $23 million in this area in 1985, and funding ties with stable, long-term funding for basic may rise to $40 to $60 million over the next research. Total JSEP annual funding has in- few years, depending on the overall level of creased since the program’s inception to $9.6 support for SDI. Most of these funds are used million in 1984; approximately 75 percent of to support the GaAs IC pilot lines that were this is spent on research on integrated circuits originally established and funded by DARPA; and other microelectronics-related areas.3 additionally, $2 to $3 million from this source goes to universities for basic research Defense Advanced Research Projects Agency activities.4 The Defense Advanced Research Projects The R&D activities supported by IPTO Agency (DARPA), which is separate from the cover the circuit-design and manufacturing services, supports long-term research for gen- portions of microelectronics technology. Ap- eral military applications, including microelec- proximately $12 million of the funds that tronics R&D. DARPA funds efforts at univer- IPTO puts into microelectronics goes to basic sities, industrial laboratories, and not-for-profit research activities in VLSI design and ar- organizations. One component of DARPA, the chitecture, most conducted at universities. In many cases, this research has later been fun- ‘~Horst R. Wittmann, Gerald I.. Witt, and Kevin J. Malloy, AFOSR; Kenneth L. Davis and Larry R. Cooper, ONR; and Jim- mie R. Suttle and Michael A. Stroscio, ARO; interviews and discussion, April 1985 to January 1986. 4Richard A. Reynolds and Sven A. Roosild, DARPA, inter- ‘Gilbert and McCombe, op. cit. views and discussion, October 1984 to January 1986. 21 neled into commercial enterprises. For exam- IC technology; SDI has microelectronics R&D ple, IPTO originally sponsored the “Cosmic components as part of a more general mission. Cube” parallel architecture work at the Cali- DOD also recently announced the start of a fornia Institute of Technology; Intel added new VHSIC-like program to advance the tech- funding to the Federal money and is now nology of monolithic microwave ICs (MMICs) building and marketing a minisupercomputer for defense applications. based on this architecture. In addition, IPTO The 10-year VHSIC program was estab- spends about $6 million per year on explora- lished in 1979 to address specific military tory development in automation and fast-turn- needs in microelectronics. The three services around efforts for ICs. These funds cover work participate in VHSIC, but the Office of the Un- on the Metal-Oxide-Semiconductor Implemen- dersecretary for Defense Research and En- tation System (MOSIS), which gives a large gineering provides overall administration. and geographically diffuse community of IC Honeywell, TRW, IBM, Hughes, Texas In- designers, particularly at universities, access struments, and Westinghouse are the prime to a silicon fabrication facility. With MOSIS, contractors for VHSIC, which is scheduled to individuals design circuits at their home fa- finish in 1989. cilities using a computer-aided design (CAD) system. The design commands are transmitted VHSIC has several goals. The primary tech- over a communication network to a manufac- nical objective is to establish processes to de- turing site, where the IC is fabricated. At pres- sign and fabricate chips with characteristics ent, designers use MOSIS to create new sili- necessary for defense needs. The program also con ICs. Activities in progress will also make intends to ease the adoption of these ICs in possible the creation of GaAs-based ICs in a military systems. In addition, VHSIC ad- similar systems ministrators view their program as a mech- anism to encourage the commercial micro- Since DARPA is charged with longer term electronics sector to develop production R&D responsibilities, the agency is shifting capabilities suited for the military market. The its focus in microelectronics away from silicon program is intended to function as a bridge efforts and toward GaAs ICs, while also be- between designers of military systems and the ginning some activities in more esoteric fields. integrated circuit community. In some cases, work that DARPA initiated and supported is being picked up by groups The technological goals of VHSIC are more interested in near-term efforts. For ex- divided into phases. The first phase of the pro- ample, DOD’s VHSIC program and the pri- gram, now nearing completion, developed pi- vate Semiconductor Research Corp. (SRC) are lot lines for the production of ICs with 1.25 taking over the funding of some silicon VLSI micron minimum feature sizes and provided programs at Stanford University that were ini- new chips for use in military systems. The sec- tially sponsored by DARPA.6 ond phase will attempt to establish pilot lines to fabricate ICs with 0.5 micron features. At DOD Special Programs the same time, commercial R&D activities In addition to the four established agencies, have been and will be developing ICs with sim- DOD has a variety of special programs that ilar feature sizes, but the VHSIC chips are de- support microelectronics R&D. Examples in- signed for specific DOD applications. Al- clude the Very High Speed Integrated Circuits though the first phase of activities took longer program and the new Strategic Defense Ini- to complete than originally anticipated, some tiative. VHSIC is wholly dedicated to silicon of the projects have now been carried out suc- cessfully. ‘>Paul I,osleben, DAR PA, interview, August 1985; and pres- Because VHSIC’s other objectives are entation on ‘ L Silicon as a Medium for New Ideas, I E ~; E Sys- tems, Man, and Cybernetics Society, Sept. 16, 1985. longer term and less concrete, their success is ‘Paul I,osleben, DARPA, interview, August 1985. harder to assess. Several signs point to 22 -. progress; VHSIC chips are used in several mil- electronics R&D at NSF is funded primarily itary systems in each of the services. And through the Directorate for Engineering. NSF while VHSIC has gotten a mixed reaction spends approximately $23 million in areas that from industry over the years, the program has include solid-state and microstructure engi- drawn the attention of at least some of the neering, quantum electronics, electronic ma- commercial IC vendors to military appli- terials, electrical and optical communications, cations. and VLSI. These funds support individuals or small groups of researchers. As conceived, VHSIC had a budget of approximately $200 million over 10 years. Sub- NSF established the National Research and sequently, its budget has been expanded to ap- Resource Facility for Submicron Structures at proximately $1 billion to support a range of Cornell University in 1977 and has provided additional activities. These include efforts to it with about $2 million annually for the last encourage incorporating VHSIC technology few years. This facility is also supported in military systems in the three services, the directly by industrial sponsors and indirectly development of a design automation system, by other Federal agencies. In addition, NSF and work on yield enhancement.7 has recently established an Engineering Re- search Center at the University of California SDI, started in 1983, is designed to study at Santa Barbara for Robotic Systems in and perhaps deploy a space-based missile de- Microelectronics. NSF plans to give this cen- fense. Since microelectronics technology would ter up to $14 million over the next 5 years, a be central to any such system, one program sum that the university will probably augment goal is the development of advanced circuitry with support solicited from industry. The cen- for space-based military operations. In addi- ter will focus on automated systems for IC tion to the DARPA GaAs work that is now manufacturing. 8 To some extent, these centers funded through the SD I program element for are evidence of a trend at NSF to concentrate sensors, SDI Innovative Science and Tech- its limited resources in a small number of large nology (1ST) office is preparing to support facilities rather than granting small bundles multiple activities in microelectronics re- of money to a large number of investigators. search. At present, the programs are being established, and funding levels are being de- NSF’s style of supporting research differs bated. Although final amounts have not been a great deal from the DOD approach. Defense set, it is possible that 1ST will spend several agencies tend to seek out channels-at univer- million dollars in this area over the next few sities, in industry, or at DOD laboratories— years. The funds may comprise dollars from to accomplish their goals, with a focus on end other DOD research pockets (e.g., DARPA, as uses. NSF, by contrast, is not a mission- in the case of the GaAs IC work, or the serv- oriented agency, so it responds to proposals ices’ research offices), so they may or may not from the research community. NSF’s funding constitute a net increase in the overall DOD for microelectronics and related areas is quite research funding for microelectronics. Cur- small compared to the total DOD support, but rently, contract monitors from other DOD re- these monies provide some counterbalance to search agencies (primarily ONR) are adminis- the defense dollars. And while NSF has not tering these funds. played the major role in advancing the technology, it has helped to build a base of qual- National Science Foundation ified scientists and engineers, e.g., by helping new university professors get started with The National Science Foundation (NSF) is small grants. This gives NSF an important charged with supporting research in a broad role as a broad basic-research agency. range of science and engineering fields. Micro- ‘Evelyn Hu, Associate Director, Center for Robotic Systems, 7Eliot D. Cohen, Navy VHSIC Program Director, interview, interview, October 1985; and NSF literature on Engineering July 1985, and additional comments, November 1985. Research Centers. .

23 —

PRIVATE SECTOR R&D

Many kinds of private sector organizations telecommunications, and military systems. are engaged in microelectronics research and Because these uses dominate the field, most development. They may be grouped into two companies with captive microelectronics oper- broad categories: ations specialize in them. 1. captive manufacturers—the parts of Since these firms are generally very large large, vertically integrated companies and relatively stable, they tend to support a that make microelectronic components for very broad range of activities, from basic re- their own products and services (typically search to product development. For example, computers or telecommunications) or for companies such as IBM and AT&T have thou- their defense systems applications sands of scientists and engineers involved in (termed “captive” because the primary different aspects of microelectronics R&D. markets for their microelectronic prod- Telecommunications and computer technol- ucts are internal); and ogies are beginning to merge because of de- 2. merchant firms-companies that make in- velopments in information technology and also tegrated circuits and other microelectron- as a result of the recent deregulation of the ic products to sell to the full range of end telecommunications industry. The latter de- users. velopment allowed new entrants into the field These categories are not mutually exclusive. and permitted AT&T, formerly excluded, to Several companies (e.g., Texas Instruments) participate in the computer marketplace. The make microelectronic components for internal divestiture of AT&T also split the well-known use as well as outside sale. However, the divi- Bell Telephone Laboratories into two research sion helps to illuminate the nature of R&D in organizations: AT&T Bell Laboratories and many companies, since the size and goals of Bell Communications Research (Bellcore), the organization are key determinants of its which serves the seven regional Bell operat- approach to R&D. ing companies jointly. Although the divestiture of AT&T officially occurred at the begin- Today, the most prominent force changing ning of 1984, its long-term effects on research microelectronics companies (and thus their are not yet completely clear. However, there R&D efforts) is Japanese competition. Mer- are preliminary indications of trends at the chant vendors are especially vulnerable. This two companies. has major implications for industrial R&D. Companies may cut back on R&D investment Some changes are clearly underway at as part of a general belt-tightening effort. AT&T Bell Laboratories. R&D activities are, Paradoxically, R&D is increasingly important overall, becoming more closely linked with in the competitive environment. Thus, the products as a result of the new competitive companies are beginning to look to the Fed- environment that AT&T faces. Even so, there eral Government for R&D support, either is ample evidence that at least in areas related through direct funding for R&D, or through to microelectronics the organization will con- Federal policies that ease the way for private tinue to pursue a broad spectrum of basic re- support, such as the tax treatment of R&D ex- search as well as development. The new envi- penditures and intellectual property pro- ronment has both negative and positive tections. implications for R&D. Scientists and engineers in the research community are concerned that, Captive Manufacturers whether or not AT&T Bell Labs shifts from basic research, it will be less likely to share Although microelectronics technology is be- the fruits of its research with others. On the ing applied in more and more ways, its pri- other hand, the pressure of competition will mary uses are still concentrated in computers, probably drive AT&T Bell Labs to move research into new products and services faster tions systems, consumer products, control than it did in predivestiture years. equipment, and defense systems. California’s The prognosis for Bellcore’s role as an R&D Silicon Valley firms (e.g., Intel and Advanced Micro Devices) are the archetypes of this cat- organization is, if anything, even less certain. Bellcore is a unique laboratory in the United egory. Generally, merchant firms have concentrated on producing standard chips (micro- States because it serves seven separate, highly processors, logic chips, and memories), which regulated companies. It is not directly linked have been used in the larger electronic sys- to any particular manufacturing facility, and tems. Custom integrated circuits, which are so will probably not experience the same shifts designed for a user’s particular needs, have in R&D as AT&T Bell Labs. In fact, Bellcore held part of the chip market for many years. to date exhibits signs of pursuing a vigorous As IC design and manufacturing become more basic research program in microelectronics- flexible, a wider range of application-specific and optoelectronics-related areas. But since it is a completely new organization, Bellcore will integrated circuits (ASICs), including custom and semicustom chips, is drawing a larger need several years to establish an identity in share of the market. During the 1985 slump R&D. in IC sales, ASICs constituted the one healthy Other manufacturers of products for indus- segment of the market.9 trial and commercial use, such as Xerox and Merchant companies, unlike their larger Hewlett-Packard, also contribute heavily to counterparts, tend to limit their R&D to the microelectronics R&D. These companies, al- last stages of development; their central con- though significantly smaller than, for example, IBM and AT&T, are still large and diverse cern is getting the latest design and fabrication technologies into production. Two major enough to support good-sized research efforts. factors converge to make longer term R&D ef- Their specific markets tend to shift the direc- forts improbable in these firms. First, the com- tion of R&D activities, so that each such companies depend on the sale of only one type of pany pursues a somewhat different research product—semiconductor devices—for their agenda. For example, Xerox’s interest in print- survival. They tend to be focused, lean opera- ing technology helped the company to achieve tions, with few discretionary dollars for basic prominence in optoelectronics research. research in an area where even a simple exper- Of the several captive microelectronics oper- imental facility costs several million dollars. ations that serve the military markets for Second, two hallmarks of the Silicon Valley microelectronics, many carry out DOD-funded culture are the ease with which workers move and internal R&D. Much of the internal R&D from one company to another, and the fre- is funded by Independent Research and De- quency with which new companies spring up. velopment (IR&D) funds which are derived Managers in merchant chip firms seldom find from overhead on defense R&D contracts. The that a heavy investment in long-term research largest players in this category are such com- pays off in this fluid environment. panies as Hughes Aircraft, Honeywell, Rock- However, in the last several years, this com- well, TRW, and McDonnell Douglas. Many munity grew concerned that its needs were not other companies that are well known for their commercial efforts are also defense contrac- being met by universities and other basic research organizations. Manufacturing research, tors. Typical examples include IBM, AT&T, for example, is increasingly crucial to the con- and Texas Instruments. tinued growth of the industry, but it had scant support among basic research organizations. Merchant Companies Merchant companies sell microelectronic ‘“A Chip Business That Is Still Growing: Innovation Spurs products to users who incorporate them in a Market for Application-Specific Integrated Circuits, ” Electron- variety of systems—computers, communica- ics, July 22, 1985, p. 40, 25

To correct this, many merchant companies ate in other R&D ventures. Several merchant today help to support external R&D activities companies also independently support re- through different channels. For example, sev- search projects at universities. All of these eral merchant semiconductor firms fund uni- activities center almost exclusively on aspects versity research through the Semiconductor of silicon technology. Research Corp. These companies also cooper-

COOPERATIVE R&D Cooperative research and development ogy Corp. (MCC).10 Each of these channels activities take different forms: organizations funds from a variety of commercial firms to that are jointly funded from several different universities and other basic-research organi- sources or research facilities that are shared zations. MCNC and MCC also carry out in- by different groups of workers. These joint ef- house R&D activities. forts represent a relatively new approach to In addition, a plethora of initiatives for R&D in the United States. In microelectronics, shared facilities are emerging from Federal several factors are driving the growing trend funding agencies, ranging from NSF, which re- toward centralized funding for R&D: cently reorganized to focus resources on a research in microelectronics requires in- group of Engineering Research Centers, to the creasingly expensive facilities, which few Innovative Science and Technology part of the participants can afford alone; Strategic Defense Initiative Organization, advances in microelectronics depend in- which is actively promoting the establishment creasingly on multiple technical disci- of research consortia. While many researchers plines, requiring a number of persons support this trend, others point out that co- trained in different areas; and operative research is hardly a panacea for cooperative research can link academia microelectronics R&D. They argue that cen- and-industry, thereby bringing necessary tralization of resources threatens innovation, funding to universities and facilitating and that research done by a large number of the process of transferring technology individual investigators, who come together from research to development to pro- in small groups to communicate and col- duction. laborate, is the most productive approach. Examples of cooperative microelectronics research organizations include the Semicon- 10The structure and Operation of these and other cooperati~re ductor Research Corp. (SRC), the Microelec- research organizations are described fully in U.S. Congress, Of- fice of Technology Assessment, Information Technology R&D: tronics Center of North Carolina (MCNC), and Critical ‘Ii-ends and Issues, ch. 6, OTA-C IT-268 (Washington, the Microelectronics and Computer Technol- DC: U.S. Government Printing Office, Februar~ 1985).

UNIVERSITIES AND R&D

Universities serve two main functions in Support for research at universities comes R&D: they perform basic research suited to from many sources, including military and an academic environment; and they educate civilian agencies of the Federal Government, and train students who subsequently perform industrial organizations, and combinations of research and development in industrial, gov- these. Microelectronics research takes many ernmental, and academic organizations. forms in this setting. Universities across the 26

Nation have individual research programs, of which have prompted disagreement in the typically in such departments as electrical field of microelectronics. Although many ob- engineering, physics, and chemistry. A hand- servers view the current trend toward large ful of universities have large research centers. campus engineering facilities as a useful way These include the National Research and Re- to train students for the activities they will source Facility for Submicron Structures at undertake in industry, some experts are con- Cornell University, Stanford University’s Cen- cerned that this trend undermines the well- ter for Integrated Systems, the Microelec- rounded education that universities ought to tronic and Information Sciences Center at the provide for their students. Thus, there is an University of Minnesota, and the Center for ongoing debate about the best way for univer- Robotic Systems in Microelectronics at the sities to fulfill this part of their mission. University of California at Santa Barbara. The university role in preparing students for the R&D community has many facets, some

FOREIGN ACTIVITIES Foreign activities in microelectronics R&D Japanese companies continue to transfer re- are so diverse that a full treatment of the topic search concepts to production with great is beyond the scope of this paper.11 However, speed. In this activity, they draw extensively several prominent features of the Japanese ef- from U.S. as well as Japanese basic research forts have important implications- for U.S. results. Another JTECH panelist states, R&D in microelectronics. They do not seem to have difficulties with Japan is the largest foreign supporter of the “not-invented-here” syndrome that slows microelectronics R&D. Although observers technology advances into the marketplace in 13 have long viewed Japanese development and the U.S. manufacturing activities as competitive with The United States has a harder time taking or superior to ‘U.S. efforts in microelectronics, the same advantage of Japanese work. One of they ‘had generally believed that the United the biggest barriers to access to Japanese re- States excelled at innovation in basic research. search by U.S. workers is the language dif- Now, however, Japanese basic research efforts ference. are drawing world-tide attention. In the words of one panelist of the Department of Com- The structure of the electronics industry in merce’s Japanese Technology Evaluation Japan strongly affects R&D. In contrast to (JTECH) Program: the United States, Japan has almost no merchant microelectronics firms. Its large, stable, It is often said that the U.S. invents and vertically integrated companies can and do in- Japan copies. . . . [S]uch generalizations are vest more heavily in long-term R&D than U.S. grossly inaccurate and certainly do not favor merchant firms. This suggests that the chal- a genuine understanding of our best competi- 12 lenge they pose to U.S. competitiveness will tor. remain and quite possibly increase. “For an extensive discussion of foreign R&D efforts in information technology, see U.S. Congress, Office of Technology Assessment, Information Technology R&D: Critical Trends and Issues, ch. 7, OTA-CIT-268 (Washington, DC: U.S. Government Printing Office, February 1985). l~Federico Capasso, AT&T Bell Laboratories, quoted in “Ja- “Robert S. Bauer, Xerox Palo Alto Research Center, quoted pan Reaches Beyond Silicon, ” ZllIIE Spectrum, October 1985, in “Japan Reaches Beyond Silicon, IEEE Spectrum, october p. 52. 1985, p. 51. .

Appendix A Current Microelectronics Technology

This appendix provides general background in- Figure A-1 .—Generations of Electronics Technology formation about microelectronics technology, the roots of which extend back to the early part of this century. Today, a vast assortment of integrated Beyond circuits (ICs) and other miniature electronic de- very large-scale Fifth generation vices are on the market. All of these components integration pass through complex design and fabrication processes before reaching consumers. t Electronics Technology From Vacuum Tubes to Integrated Circuits Very Electronics technology over the last century has Early 1980s large-scale Fourth generation advanced in a series of dramatic breakthroughs: integrated circuits the vacuum tube, the transistor, and the integrated circuit. These are often used to designate different generations of technology, as shown in figure A-1. The history of electronic devices began in the early 1900s with the invention of vacuum tubes, 1958 Integrated Third generation the first devices to reach major use in manipulat- circuits ing and amplifying electrical currents. This technology is limited by several related features. b Vacuum tubes require high voltages and, by microelectronics standards, a great deal of power. t The invention of the transistor in 1947 marked the beginning of a new generation in electronics.

The inventors fabricated the device from a crys- 1947 Transistors Second generation tal of semiconductor material to which they added controlled quantities of different impurities. The transistor had three points of electrical contact, or terminals. Like a vacuum tube, it could amplify an electrical signal, but at room temperature, and t it required only a small fraction of the power, voltage, and space that tubes demand. The transistor rapidly became the central component in circuits Early 1900s Vacuum First generation for a variety of applications. tubes By the late 1950s, transistor-based circuits dom- inated early computers and military systems. But SOURCE Off Ice of Technology Assessment the drive for increasingly complex circuitry en- countered some difficult barriers. Millions of con- demanded smaller, faster circuits that would connections were required to turn the hundreds of sume less power and could be fabricated reliably. thousands of components into a functioning The next major breakthrough in electronics was circuit—a labor-intensive process with unreliable the invention of the integrated circuit in 1958. Un- results. Further complicating things, the total cir- til this time, all electrical circuits consisted of sep- cuit size was growing unmanageable, although the arate components-transistors, diodes, resistors, individual components were quite small. The capacitors—connected by wires. The inventors of equipment in which electronic devices were used the IC recognized that the wires and other com-

27 28 ponents could also be fabricated from or on the vices in general. Figure A-2 shows the types of same semiconductor material that was used to microelectronic devices on the market today. The make transistors.1 In this way, they made an en- most numerous and widely used of these are stand- tire circuit on a single piece, or chip, of the mate- ard ICs made of silicon. Other semiconductor rial. Using this technique, engineers were able to devices, such as optoelectronic and microwave make connections and components by etching the products and discrete devices, are also part of semiconductor and depositing metals and insula- microelectronics technology. Additionally, a few tors in patterns on the chip. The new process elim- examples exist of miniature devices made from inated many of the problems associated with pro- materials other than semiconductors (e.g., mag- ducing highly complex circuits. Because the huge netic bubble storage devices), but these fall out- number of connections no longer had to be made side the scope of this paper. by soldering wires together, the IC technique was more reliable and less labor-intensive. The individ- Types of Integrated Circuits ual components could now be smaller since they would not need to be handled by people. This Integrated circuits are commonly classified in shrinkage yielded circuits that operated at higher a variety of ways and referred to with an impres- frequencies and consumed less power. sive number of acronyms. They may be catego- Progress in microelectronics over the last 25 rized by function, by the level of integration (num- years has been based on further shrinkage of the ber of components), by the type of transistors used circuitry etched onto chips. The number of com- in the circuitry, or by the underlying material (or ponents per chip has, on average, almost doubled substrate) on which they are fabricated. every year, Today, up to 1 million transistors and Types of Functions.–Integrated circuits may be other components are fabricated on a single chip either digital or analog (also called linear). In gen- that may have an area of less than 1 square centi- eral, a digital circuit switches or stores voltages meter. that represent discrete values. For example, O volts and 5 volts may represent the values O and 1, respectively. An analog circuit, in contrast, am- Microelectronic Devices plifies or otherwise modifies voltages of any value within a given range (e.g., any voltage between O In this background paper, the term “microelec- and 5 volts). Some integrated circuits are designed tronics” is used to refer to miniature electronic de- to convert signals from analog form to digital form ‘For an account of the invention of the integrated circuit, see T. Il. (A-to-D converters) or vice versa (D-to-A con- Reid, ‘The Chip,” Science 85, February 1985, p. 32, excerpted from T.R. verters). Custom chips can combine the various Reid, The Chip: 7’he Microelectronics Revolution and the Men W’ho Made It (New York: Simon & Schuster, 1985). functions. Currently, the majority of ICs are digi-

Figure A-2.—Types of Microelectronic Products

Microelectronic products . 1 1 I I Integrated Discrete Optoelectronic Microwave circuits transistors devices - devices and diodes . b I 1 t ? * 1 Custom and other Monolithic Digital Analog Lasers Discrete Ics (linear) ICs ICs and light- microwave I 4 4 < devices emitting Detectors ICs diodes (LEDs) t -1 f Microprocessors a Memories - Logic ICs 4 A Read-only Random-access memories memories (ROMs) (RAMs)

SOURCE Office of Technology Assessment 29 tal; computers, the major use of ICs, are based on Designers can program information into read- digital integrated circuits. only memories in different ways. They may put the The major product categories in digital inte- information in the physical chip design; this type grated circuits are logic circuits, memories, and is known as a mask-programmable ROM. Alter- microprocessors. All three handle information in natively, chip manufacturers fabricate ROMs that the form of electrical signals: logic circuits proc- users can program themselves (programmable ess the signals, memories store them, and micro- ROMs, or PROMs)–a much more versatile prod- processors combine the two functions. uct. Some varieties of these PROMS can be erased Logic circuits carry out the operations required and reprogrammed using light or electrical signals to manipulate data in binary digital form. They (erasable PROMS, or EPROMs). EPROMs differ perform mathematical functions. Logic ICs are from read/write memories in two ways: they typi- typically classified according to the type of tran- cally cannot be reprogrammed by the computer it- sistors used to make them. self, but they need no power source to retain in- Semiconductor memories are microelectronic cir- formation. cuits designed to store information. They fall into The capacity of a semiconductor memory is de- two groups: read/write memories and read-only termined by the number of bits that it can store. memories (ROMs). Although both groups consist The maximum capacity available in the early primarily of memories that can be accessed ran- 1970s was 1,024 bits, commonly called a kilobit or domly (i.e., any storage location can be accessed kbit. Today, 262,144-bit (256 -kbit) RAMs are on in the same amount of time), the term “random- the market, and l,048,676-bit (l-megabit) RAMs access memory” (RAM) usually refers only to are just around the corner. Prices for memories read/write memories. A computer can store a da- have dropped as dramatically as their capacity has tum in a location within a read/write memory and risen: memory costs approximately one-thousandth later retrieve it or store a new datum there. How- of a cent per bit today. ever, the computer can only retrieve information Because a microprocessor is a complete com- that is already stored (by some external means) in puter processing unit on a single IC, it can proc- a ROM. Thus, a ROM may be used to store un- ess data expressed in a series of binary digits or changing information or instructions that the com- bits (ones or zeroes). The number of bits used in puter needs regularly, while read/write memories the series determines the precision of the datum’s are used to store data as they come into the com- digital representation; precision increases with the puter and as they are processed. number of bits. The original microprocessor, made Read/write memories may be labeled “static” or in 1970 at Intel Corp., was designed to handle four “dynamic” depending on the design of the circuits bits. The industry introduced 8- and 16-bit micro- that make up the RAM. Dynamic RAMs (DRAMs) processors by the end of that decade and 32-bit store data in the form of charge on capacitors, microprocessors within the last few years. and so require that the capacitors be recharged Levels of Integration.—The level of integration regularly-every few milliseconds. Static RAMs of an integrated circuit refers to the number of (SRAMs), on the other hand, store data by chang- components packed on the chip. The progress in ing the state of transistors in the storage element, integration since the invention of the IC is shown a technique that requires a constant flow of power in table A-1. Figure A-3 shows how a transistor rather than intermittent recharging to maintain on an IC compares in size with some other micro- accurate storage. Both static and dynamic read/ scopic items. write memories need a constant supply of power Types of Transistors. -An integrated circuit may to the circuit as a whole to operate. also be classified according to the kind of transis-

Table A-1 .—Levels of Integration — ——— Design rule Number of Approximate Level of integration (microns) transistors years Scale-scale integration (SSI) ...... 30-20 2 to 64 1960-65 Medium-scale integration (MSI) ., ...... 20-10 64 to 2,000 1965-70 Large-scale integration (LSI) ...... 10-3½ 2,000 to 64,000 1970-78 Very large-scale integration (VLSI) . . ... 3½ -1 1/4 64,000 to 2 million 1978-86 Ultra large-scale integration (ULSI) ...... <1 l/4 >2 million after 1986

NOTE The numbers (n Ihls table a;e approximate NO standard de flnlt;on extsts for the dlfferenl I ntegratton levels SOURCE Adapted from James M Early and Bruce E Deal Falrchlld Research Center and C Gordon Bell Encore Computer Co.[) 30

Figure A-3. —Microscopic Sizes

I&l ++ 1 micron Transistor on an integrated circuit approx. 1-2 microns

Bacterium approx. 3 microns long / \ m Cross section of 1 Small human hair particle approx. 50 microns 15 microns

o 5 10 15 20 25 30 35 40 45 50 55 60 65 M icrons

SOURCE Off Ice of Technology Assessment Adapted fr)m .] flqure by Jlmmle R Suttle Army Research Of f!ce tor used in its design. For silicon ICs, two types chines because they are the only chips capable of of transistors are used: bipolar transistors and the high speeds that are required. On the other metal-oxide-semiconductor field-effect transistors hand, personal computer vendors tend to base (MOSFETs). These are further subdivided to de- their products on MOS chips, because cost is gen- scribe the specific structure of the device and the erally a bigger concern than speed. design of the surrounding circuit. The most com- Types of Materials. –Virtually all integrated cir- mon bipolar designations are ECL (emitter- cuits on the market today are made on substrates coupled logic) and TTL (transistor-transistor of the same material that was the base for the first logic); the most common types of MOSFETS are IC: silicon. The longevity of silicon is due to its NMOS (n-channel MOS) and CMOS (complemen- physical properties as well as to practical economic tary MOS). (See the discussion of chip fabrication considerations. Silicon IC technology is more eas- in this appendix for some additional details.) ily executed than is the technology of compound The primary differences between bipolar and semiconductors, which have more than one ele- MOS technologies are speed (bipolar transistors ment, because the chemistry of a material made are faster); power (MOS circuits require less of a single element is innately simpler. Further- power); and ease of manufacture (MOS circuits more, once silicon technology was established, eco- have higher yield). Because the yield determines, nomic factors dictated that any replacement be far in large part, the cost of the chip, bipolar chips are superior to silicon to be worth the large costs of typically more expensive. However, the advantage converting to a new system. And since silicon tech- of higher speed is worth the extra cost and power nology has never stopped progressing-or even requirements for bipolar ICs in some applications. slowed significantly-attempts to replace it have For example, most makers of large computers use been shooting at a moving target. The recent de- bipolar ICs for the central components in their ma- cline of Josephson junction technology (which is 31 based on superconductors) for computers has wires that carry electrical signals. Semiconductor largely been attributed to this problem. lasers or LEDs generate the light signals, and semi- Despite its great usefulness, however, silicon conductor photodetectors convert the received cannot satisfy all demands of microelectronics. light signals back to electrical signals. These de- Since transistors made from gallium arsenide vices are most commonly made from layers of (GaAs) can be faster and more impervious to ra- GaAs and aluminum gallium arsenide (AIGaAs) on diation damage than silicon transistors, several a substrate of GaAs, or, more recently, from layers companies are trying to introduce GaAs inte- of iridium gallium arsenide phosphide (InGaAsP) grated circuits. These ICs are particularly attrac- and iridium gallium arsenide (InGaAs) on a sub- tive for some military applications (e.g., space- strate of InP. The quaternary materials-different based circuitry), and they may also find standard compositions of InGaAsP-can generate and de- commercial applications. However, they cannot tect the wavelengths of light that travel along the currently compete with silicon technology for glass fibers with the least distortion (wavelength standard functions. For example, a fully opera- of 1.3 microns) and fading (wavelength of 1.55 tional 1,024-bit GaAs RAM is still in the labora- microns). tory phase, while silicon RAMs with over 250 The primary purpose of long-wavelength in- times the capacity have been on the market for frared (IR) detection is to “see” living or hot ob- some time already, (See ch, 3 for a more detailed jects in the dark. The Department of Defense has description of current R&D activities in this area.) spearheaded work on IR sensors, since their potential military uses are many. Typically, the de- Other Semiconductor Devices vices are made from alloys of HgTe and CdTe, because these materials can be adjusted to detect the Semiconductor materials are also used to make appropriate wavelengths of light (3 to 5 microns a host of individual microelectronic devices, as and 8 to 14 microns). Special types of silicon di- distinguished from standard silicon integrated odes can also act as IR sensors in some of these circuits. These include discrete transistors and wavelength ranges. diodes, optoelectronic devices, and microwave Microwave Devices.-Microwave (high-frequen- devices. cy) devices take advantage of the intrinsic high Discrete Transistors and Diodes.—Electronic in- speed of gallium arsenide. They generate, amplify, struments, including radios, televisions, and con- switch, and receive microwave signals with fre- trol instruments, use not only ICs but also indi- quencies from approximately 1 to 60 gigahertz (1 vidual transistors and diodes. (Diodes are two- gigahertz is 1 billion cycles per second). Radar sys- terminal devices that allow current flow in only one tems and transmission systems for telephone, tele- direction.) These components are typically made vision, and telegraph signals, which operate at from silicon, The transistors may be bipolar or these frequencies, as well as military systems, de- field-effect transistors. pend on these microwave devices. This technology Optoelectronic Devices. -Semiconductors such as is at present progressing to monolithic microwave gallium arsenide (GaAs), iridium phosphide (InP), integrated circuits (MM ICs), in which microwave cadmium telluride (CdTe), and mercury telluride circuits are fabricated on a single substrate, typi- (HgTe) are the bases for a range of other micro- cally gallium arsenide. electronic devices that require properties that silicon lacks—the ability to interact efficiently with Technologies To Produce light and the ability to operate at extremely high speeds or frequencies. Semiconductor Microelectronics Optoelectronic devices depend on the first of Numerous steps are involved in making any these properties. Optoelectronics includes light- microelectronic device. Integrated circuits are the emitting diodes (LEDs) and lasers, which convert most complex in this respect. The process can be electrical signals to light signals, and photodetec- separated into two parts: circuit design and chip tors, which convert light signals into electrical sig- fabrication. nals. The two major applications for these devices are fiber optic communications and long-wave- Circuit Design length infrared light detection. Fiber optic or lightwave communications sys- Once the application for a new integrated circuit tems use glass fibers to transmit light signals from is established, the first step in making the chip is one point to another; the fibers can replace metal the design of the circuit. The use of a wide variety 32 of computer-aided design (CAD) tools has facili- Figure A-4.– First of Four Mask Processes To Make tated all parts of the design process for large ICs an N-channel MOSFET tremendously. For a digital IC, the overall circuit will typically be a combination of logic functions n-type n-type and storage functions. The designer determines A. B the layout of the subcircuits that carry out each siIicon I I silicon I of these functions and the connections between the subcircuits. Also in this initial stage, the chip Light designer must establish mechanisms for bringing the external signals and power to the various subcircuits. Each subcircuit may itself be composed of smaller, interconnected cells of circuitry. The fundamental units for the designer are the individual components, such as transistors, diodes, resistors, and capacitors, and the connections between them. Because the process of designing circuits has grown so complex over the years, now involving over 100,000 components on a chip, design software is a major research area which is as impor- E n-type F n-type tant to progress in microelectronics technology as silicon silicon I I I I advancements in chip fabrication. As the capabilities of design tools grow, they provide increasingly greater flexibility for different chip architec- p-,type impurities tures, and therefore open the door to using IC I I I I technology in completely new ways to fabricate specialized circuits.

Chip Fabrication Chip fabrication includes making the circuitry and packaging the completed chip. The first step of the process is transferring the Metal contacts circuit designers’ layout onto the semiconductor chip. The substrate onto which the circuits are transferred is a wafer, or thin slice, of a single crystal of the semiconductor. Currently, a typical wafer is several inches in diameter, so many chips can SOURCE Office of Technology Assessment be made on a single wafer and separated after circuit fabrication. The circuit is created by deposit- thography) leave a layer of silicon dioxide (SiO2), ing thin layers of metals, insulators, and perhaps with small holes etched in it, on top of the sub- additional semiconductor materials on the wafer; strate. This layer serves as a stencil for the p-type adding n-type impurity atoms (negative charge impurities to which the chip is then exposed: only contributors) or p-type impurity atoms (positive the portions that are not covered with SiO2 become charge contributors) to the semiconductor; and p-type. etching away precisely defined portions of the vari- This process is accomplished as follows. The ous layers with chemicals or ions. starting material is a wafer of silicon to which n- Illustrating a simplified example of one part of type impurities were added during growth. The this process, figure A-4 shows the steps in the first first step is to grow a thin layer of silicon dioxide phase of the fabrication of a single transistor-an on the surface of this substrate, either by expos- n-channel MOSFET—on a silicon substrate. The ing the surface to the proper chemicals, or by heat- objective of this first phase is to add p-type impu- ing the wafer in a furnace with water vapor (step rity atoms in two small selected regions near the A in figure A-4). Next, a layer of light-sensitive ma- surface of the chip, thereby forming “p-wells.” The terial called photoresist coats the oxide (step B). first six steps of this phase (which constitute li- The portions of the photoresist that are exposed —

when the surface is irradiated with light through Fabricating an entire integrated circuit is sig- a glass mask (step C) undergo a chemical change. nificantly more complicated because a larger num- These parts of the photoresist do not wash away ber of much more complex masks are necessary. when the chip is rinsed in the chemical developer, Chips now also require extremely sophisticated but the unexposed portions do (step D). Now, the techniques for selective etching of materials, pho- chip is placed in a bath of chemicals that etch away toresist exposure, and other parts of the process. silicon dioxide, but leave the photoresist and sili- All of these processing steps must be carried out con intact (step E). Since the portions of the ox- in an environment completely free of particulate ide that are below the photoresist are not etched, —clean-room facilities. Thus, it is not surprising this procedure opens windows in the oxide layer. that the trend in chip fabrication has consistently A different chemical bath removes the remaining been towards greater automation of the processes, photoresist without damaging the oxide regions which reduces the chances of contamination or hu- (step F). The chip is now ready for exposure to the man error. Today, most machines used in chip p-type impurities, which are driven into the chip fabrication can handle cassettes containing many either from a piece of source material in a furnace large wafers of silicon, minimizing the need for hu- or by accelerating impurity ions into the material mans to handle the delicate circuitry. (step G). When the number of p-type impurities ex- When the final lithographic procedures are com- ceeds the number of n-type impurities that were plete, the large silicon wafer is tested and diced there originally, the wells will exhibit p-type be- to separate the individual chips. The good chips havior. Finally, when the p-wells have been formed, are then assembled in packages appropriate for the remaining SiO2 can be chemically removed their particular applications. Packaging technol- (step H). ogy is a crucial part of the production of ICs. After three more mask cycles, the completed de- Proper packaging techniques can shield the chip vice is composed of regions of semiconductor with from physical damage and some forms of radiation p- and n-type impurities; silicon dioxide, which acts damage, For commodity ICs that are sold to a va- as an electrical insulator; and aluminum electrical riety of end users, standard packaging systems are contacts. In present-day chips, the physical sepa- necessary to ensure that chips are interchangeable. ration between the regions with p-type impurities On the other hand, chips designed and used for may be less than 1 micron, and the entire transis- high-speed computation require special custom tor is invisible to the naked eye. Furthermore, the packages that maximize the speed with which sig- number of impurities must be controlled to within nals enter and leave the chip, since standard pack- a few parts per billion in some regions of the aging schemes could obliterate the special speed device. advantage of the chip by itself. Appendix B Glossary of Terms

Chip: A small piece (typically less than 1 square may be bipolar transistors, or one of a variety centimeter) of a semiconductor wafer. Also used of metal-oxide-semiconductor (M OS) tran- to refer to a packaged integrated circuit. sistors. Compound semiconductor: “A semiconductor made Logic chips: ICs that manipulate digital data. See of a compound of two or more elements instead app. A. of a single element like silicon, II I-V semicon- LSI: Large-scale integration. See table A-1 in app. ductors are made from elements in group 111 of A. the periodic table (such as aluminum, gallium, Memory chips: Devices for storing information in and iridium) and group V of the periodic table the form of electronic signals. See app. A. (such as nitrogen, phosphorus, arsenic, and an- Microprocessor: A computer central processing timony). Binary compounds are made with two unit on a single chip. See app. A. elements, ternaries with three elements, and Micron: One-millionth of a meter. See fig. A-3 in quaternaries with four elements. ” app. A. Custom and semicustom integrated circuits: ICs that Microwave circuit: An analog circuit designed to can be designed to varying degrees by the end process high-frequency signals. See app. A and user for a specific application. Full custom ICs ch. 3. are designed and fabricated from scratch: semi- MMIC: Monolithic microwave integrated circuit. custom chips allow the user to modify a chip for See app. A and ch. 3. the application. Optoelectronic devices: Devices that convert light Design rule: Minimum feature size. Current chips signals to electronic signals and vice versa. See typically employ design rules of 1 micron or app. A and ch. 3. greater. See table A-1 in app. A. RAM: Random-access memory. See app. A. Diode: A two-terminal electronic device that allows ROM: Read-only memory. See app. A. current to flow freely in one direction only. Semiconductor: A crystalline material whose elec- Gallium arsenide (GaAs): A compound semiconduc- trical conductivity falls between that of a metal tor with properties necessary for very high-fre- and that of an insulator. Semiconductor mate- quency microwave (analog) devices and opto- rials are used to fabricate virtually all microelectronics. There are also several efforts to electronic devices. Silicon is the most common; make high-speed digital ICs based on gallium others include germanium, gallium arsenide arsenide. (GaAs), indium phosphide (InP), mercury tellu- Integrated circuit (IC): Electronic circuits, includ- ride (HgTe), cadmium telluride (CdTe), and al- ing transistors, resistors, capacitors, and their loys of these compound semiconductors. See fig. interconnections, fabricated on a single small 1 in ch. 3 for a periodic table of the elements piece of semiconductor material (chip). Catego- showing common elements used in semicon- ries of ICs such as LSI and VLSI refer to the ductors. level of integration, which denotes the number Substrate: A piece of material, typically a semicon- of transistors on a chip. ICs may be digital (logic ductor, on which layers of materials are depos- chips, memory chips, or microprocessors) or ited and etched to fabricate a device or a circuit. analog. The transistors that ICs are made of Transistor: A three-terminal electronic device that can be used to switch or amplify electronic signals. VLSI: Very large-scale integration. See table A-1 p in app. A.