Miniaturization Technologies (November 1991)

Miniaturization Technologies

November 1991

OTA-TCT-514NTIS order #PB92-150325

Recommended Citation:

U.S. Congress, Office of Technology Assessment, Miniaturization Technologies, OTA-TCT-514 (Washington, DC: U.S. Government Printing Office, November 1991).

For sale by the U.S. Government Printing OfficeSuperintendent of Documents, Mail Stop: SSOP, Washington, DC 20402-9328

I S B N O - 1 6 - 0 3 5 9 8 3 – X

Advances in miniaturization technologies have had dramatic impacts on our lives. Radios, com-puters, and telephones that once occupied large volumes now fit in the palm of a hand. Dozens of sen-sors are sent on spacecraft to the planets and on instruments into the human body. Electronic brainsare in everything from bombs to washing machines.

This report analyzes various technologies that may be important for future advances in miniatur-ization. Current research in the United States and other nations is pushing the limits of miniaturizationto the point that structures only hundreds of atoms thick will be commonly manufactured. Researchersstudying atomic and molecular interactions are continuing to push the frontiers, creating knowledgeneeded to continue progress in miniaturization. Scientists and engineers are creating microscopic me-chanical structures and biological sensors that will have novel and diverse applications.

OTA characterizes U.S. research and development in miniaturization technologies as the best inthe world. Despite the growing prowess of foreign research, American researchers continue to innovateand push the frontiers of miniaturization. The more elusive challenge is to translate success in the labo-ratory to success in the global marketplace.

OTA gratefully acknowledges the contributions of the workshop participants, contractors, review-ers, and contributors who provided information, advice, and assistance. OTA, of course, bears soleresponsibility for the contents of this report.

(JJOHN H. GI B BO NSDirector

. . .111

Miniaturization of Electronic and Mechanical Devices Workshop, Feb. 19, 1991

Karl Hess, ChairmanProfessor of Electrical Engineering

Fernand BedardResearch PhysicistNational Security Agency

T.H. Phillip ChangIBM T.J. Watson Research Center

Harold CraigheadDirectorNational Nanofabrication FacilityCornell University

Eric DrexlerVisiting Scholar

B e c k m a n I n s t i t u t e -

University of Illinois

Wen Hsiung KoProfessor of Electrical EngineeringProfessor of Biomedical EngineeringMember of the Electronics Design CenterCase Western Reserve University

Carl KukkonenDirectorCenter for Space Microelectronics TechnologyJet Propulsion Laboratory

Richard S. MullerDepartment of Electrical Engineering/

Computer ScienceUniversity of California at Berkeley

Stanford Computer Science DepartmentStanford University

.

David K. FerryRegents ProfessorDepartment of Electrical EngineeringArizona State University

Kaigham GabrielAT&T Bell Labs

Stephen JacobsenDirectorCenter of Engineering DesignUniversity of Utah

James S. MurdaySuperintendent of ChemistryNaval Research Laboratory

Richard PotemberJohns Hopkins Applied Physics Lab

Arati PrabhakarDirectorMicroelectronics Technology OfficeDARPA

George Sai-HalaszResearch Staff MemberIBM T.J. Watson Research Center

NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the workshop participants. Theworkshop participants do not, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibilityfor the report and the accuracy of its contents.

Miniaturization Technologies Project Staff

John Andelin, Assistant Director, OTAScience, Information, and Natural Resources Division

James W. Curlin, Program ManagerTelecommunication and Computing Technologies Program

Project Staff

Sunil Paul, Project Director

Contractor Staff

Harold Craighead, Cornell UniversityMichael Skvarla, Cornell University

Clayton Teague, National Institute of Standards and Technology

Administrative Staff

Liz Emanuel, Office AdministratorJo Anne Young, Secretary

Karolyn St. Clair, PC Specialist

Miniaturization Technologies

Reviewers

John AlicIndustry, Technology and

Employment ProgramOffice of Technology Assessment

Samuel BaldwinEnergy and Materials ProgramOffice of Technology Assessment

Robert BateTexas Instruments

Fernand BedardNational Security Agency

Robert BirgeCenter for Molecular ElectronicsSyracuse University

Frederico CapassoAT&T Bell Labs

Kenneth DavisElectronics DivisionOffice of Naval Research

Bob DennardIBM

Eric DrexlerStanford Computer Science

DepartmentStanford University

Larry DworskyMotorola, Inc.

David K. FerryArizona State University

Kaigham GabrielAT&T Bell Labs

Joseph GiachinoFord Motor Co.

L Val GiddingsU.S. Department of Agriculture

Lance GlasserDARPA

Henry GuckelUniversity of Wisconsin-

Madison

George HazelriggElectrical and Communications

Systems DivisionNational Science Foundation

Karl HessUniversity of Illinois

Stephen JacobsenCenter of Engineering DesignUniversity of Utah

Wen Hsiung KoElectrical Engineering and

Applied PhysicsCase Western Reserve University

Carl KukkonenCenter for Space Microelectronics

TechnologyJet Propulsion Laboratory

Rolf LandauerIBM

Ralph MerkleXerox Palo Alto Research Center

Lester W. MilbrathState University of New York

at Buffalo

Richard S. MullerDepartment of Electrical

Engineering/Computer ScienceUniversity of California at Berkeley

James S. MurdayNaval Research Laboratory

Robyn NishimiBiological Applications ProgramOffice of Technology Assessment

Wally ParceMolecular Devices

Carl PilcherOffice of Space Science

and ApplicationsNational Aeronautics and

Space Administration

Richard PotemberJohns Hopkins Applied Physics Lab

Arati PrabhakarDARPA

George Sai-HalaszIBM T.J. Watson Research Center

Charles ShanleyMotorola, Inc.

Clayton TeagueNIST

William TrimmerPrinceton University

William TroutmanAT&T

Matthew WeinbergEnergy and Materials ProgramOffice of Technology Assessment

Sheryl WinstonBiological Applications ProgramOffice of Technology Assessment

Ken D. WiseThe University of Michigan

NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the reviewers. Thereviewers do not, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for thereport and the accuracy of its contents.

vi

ContentsPage

Chapter 1. Introduction and Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1FINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . “ 1WHY IS MINIATURIZATION IMPORTANT? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ... ... ......4COMPETITIVENESS OF U.S. MINIATURIZATION TECHNOLOGIES . . . . . . . . . . . . . . . . . . . . . . . . 5

Semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Quantum Effect Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Molecular and Biological Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ... ... .......7Packaging and Interconnection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Biosensors and Chemical Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Micro-Mechanical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Fabrication Technology Research and Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Chapter 2. Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13INTEGRATED CIRCUITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 13

Physical Limits to Transistor Miniaturization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Continuing the Trend in Miniaturization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Fabrication Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

COMPUTING SYSTEMS TECHNOLOGIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Display Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 23Data Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Interconnection and Packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

SENSORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ......27Charge Coupled Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Chemical and Biological Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Micromachined Silicon Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

ACTUATORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Micro-Electro-Mechanical Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Integrating Actuators, Sensors, and Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Appendix A. Fabrication Technology for Miniaturization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Appendix B. Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

BoxesBox Page

2-A Molecular Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202-B Bedside Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292-C A New Way To Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

FiguresFigure Page

1-1 How Small Is Small? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-2 Cost of a New Memory Fabrication Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31-3 Price History of Electronic Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41-4 Use of Surface Mount Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61-5 Semiconductor R&D Spending-Top Four Merchant Companies in the U.S. and Japan 62-1 TransistorTrends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2-2 Microprocessor Clock Speeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152-3 The Photolithographic Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182-4 Three Approaches to Electronics Miniaturization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222-5 Relative Cost of a Chip at Different Levels of Interconnection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252-6 Relationship of Die Size and Yield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262-7 Impact of Surface Mount Technologies on Calculator Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272-8 Cut-away View ofa Silicon Pressure Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302-B-1 Bedside Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292-C-1 Surface Micromachining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32A-1 Fabrication Sequence for a Metal-Oxide-Semiconductor (MOS) Circuit . . . . . . . . . . . . . . . . . . . . . . . . . 36A-2 Ion Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

vii

Chapter 1

Introduction and Summary

“Small is Beautiful.” The truth of that state-ment is debated in economic and sociologicalcircles, but when it comes to technology, there isno debate; small is beautiful because small is fast,small is cheap, and small is profitable. The revo-lution begun by electronics miniaturization dur-ing World War II is continuing to change theworld and has spawned a revolution in miniatur-ized sensors and micromechanical devices.

Miniaturization plays a major role in the tech-nical and economic rivalry between the UnitedStates and its competitors. It translates to marketshare and competitive advantage for many com-mercial and scientific products. Those compa-nies and nations that can successfully developand capitalize on miniaturization developmentswill reap handsome rewards. Personal comput-ers, portable radios, and camcorders are exam-ples of products that created massive new mar-kets through miniaturization: they added billionsof dollars to the GNP of countries where theywere designed and built.

FINDINGS

The United States remains strong in miniatur-ization technologies research and development(R&D), although the lead over other nations isless substantial than it has been in the past.

U.S. researchers continue to innovate and pro-duce world-leading research despite strong re-search programs in Japan and Europe. There aresome areas where Japanese or European re-search surpasses the United States in quantityand in a few cases in quality as well. But on thewhole, U.S. researchers lead in miniaturizationtechnology R&D. The danger is that U.S. com-panies will lag other nations in implementingadvanced technologies, especially when new

technology is driven by a product or market dom-inated by another nation’s industry.

The trends in silicon electronics miniaturiza-tion show no signs of slowing in the near future.

The current pace of miniaturization will pro-duce memory chips (dynamic random accessmemory, DRAM) with a billion transistors andthe capacity to store 1 billion bits (1 gigabit) of in-formation around the year 2000.1 Transistors willcontinue to shrink until the smallest feature isaround 0.1 micron (1 micrometer or one millionthof a meter). By comparison, today’s most ad-vanced mass-produced integrated circuits havefeatures as small as 0.8 microns. A human hair is50 to 100 microns in width (see figure l-l).Achieving such tiny features will require a hugeengineering and research effort. New fabricationequipment and processing techniques must bedeveloped, pushing the cost of chip fabricationplants to over $1 billion, compared to hundreds ofmillions for a current state-of-the-art plant (seefigure 1-2). It is likely that despite the high costs,chips having features around 0.1 micron will bemanufactured; progress beyond the era of O.l-mi-cron transistors is uncertain.

The technology of semiconductor manufactur-ing is being applied to other fields to create newcapabilities. Sensors created with semiconduc-tor manufacturing technology hold the promiseof widespread applications over the next 10years.

Micromechanical sensors for pressure and ac-celeration have used semiconductor manufactur-ing technology for several decades now, but re-cent innovations allow further miniaturization,greater flexibility, and compatibility with micro-electronics. A wider range of sensors can now befabricated using micromechanical structures.

l~i~ ~omPre~ t. t~ay~s ~mt dense memoy chip, which have about 4 million transistors and hold 4 million bits (4 megabits) ofinformation.

-1-

2 ● Miniaturization Technologies

Figure 1-1 -How Small Is Small?

Human hand m wide

Grain of sand mm

Human hair50-100 microns

wide

Smoke particle microns

DNA nm wide

Atoms 1-4

100m

1 foot0.3m

1 0-1m

1 i n c h25.4 mm

10-2m.

1 0- 3m .

l o- 4m .

-

1 0-lOrn-

– 1 meter (m)

—

0.1 m Printed circuit board100 mm 0.05-1 m wide

– 0.01 m10 mm1 centimeter

- 1 millimeter(mm)

- 0.1 mm100 microns

- 0.01 mm10 microns

- 1 micrometer1 micron

- 0.1 micron100 nm

- 0.01 micron10 nm

- 1 nanometer (rim)10A

. 0.1 mm

Integrated circuit chip (die) mm wide

Micromechanicalcomponents

10-1000 microns wide

Transistor on integrated circuit2-20 microns wide

(smallest feature -O.8 micron)

Quantum electronics structures

wide

Atomic lettering usingscanning tunneling microscope

65A high1 angstrom

Each increment on the vertical scale indicates change in scale bye factor of 10. On the right of the scale are examples of miniaturizationtechnologies. For comparison, several objects of similar scale are shown on the left.SOURCES: of Technology Assessment, 1991. Data from Philip Morrison, (San Francisco, CA: American

Books, 1982); David Michael Valley, CA: Jones and 1987), p. 183;Douglas M. cd., Van Scientific Encyclopedia (New NY: Van Nostrand Reinhold Co., 1983).

Chapter l–Introduction and Summary ● 3

$1,800

$1,600

$1,400

$1,200

Figure 1-2–Coat of a New Memory Fabrication Facility (DRAM Fab)

0’ $1,000

//,

–

$600 –

$400 –

$200 –

I I I I I6 4 K 256K 1Mb 4Mb 16Mb 64Mb 256Mb1979 1982 1985 1988 1991 1994 1997

SOURCE: Dataquest (August 1991) and Graydon Texas Instruments, personal Sept. 18, 1991.

Biosensors and chemical sensors that can sensegases and chemicals are being perfected using in-tegrated circuit manufacturing techniques andwill be used in medical, food processing, andchemical processing applications. The economicsof microelectronics fabrication will result in thesenew sensors becoming cheap and ubiquitoussince hundreds or thousands can be created on asingle wafer. Integrating sensors with electronicspromises to increase the versatility of sensors forconsumer, medicine, automotive, aerospace,and robotics markets.

Materials and surface science research is criti-cal to further advancement of all miniaturiza-tion technologies.

In every miniaturization technology-from sil-icon microelectronics to quantum electronics, tomicromechanics and biosensors—better under-

standing of materials and surface interactionswill be a critical part of further technology ad-vances. Better characterization of manufacturingprocesses will be necessary to make future gener-ations of miniaturized semiconductors. Makingpractical miniaturized biosensors and chemicalsensors will require better understanding of howto bond molecules to surfaces. Progress in micro-mechanics will depend on how well the mechani-cal and surface properties of materials like siliconare understood. Resolving problems in quantumelectronics and molecular computing—the fron-tier of electronics miniaturization-are highlydependent on improved understanding and con-trol of materials and surfaces. Basic research onmaterial properties and surface interactions—es-pecially in semiconductor processing and man-ufacturing-will be necessary for further minia-turization in many technology areas.


Packaging is playing an increasingly largerrole in miniaturization of electronics.

Trends in miniaturization are creating pressurein the electronics industry to improve packaging:

1.

2.

3.

The proliferation of electronics into porta-ble devices, consumer electronics, automo-biles, and industrial applications is forcingmore compact and ruggedized packaging;

The miniaturization of transistors causesthem to operate faster and is forcing atten-tion to better packaging because fast elec-tronics must be packed close together toavoid delay in sending signals from chip tochip; and

As the costs of integrating more and moretransistors onto the same piece of siliconincrease, alternative ways to integrate tran-sistors into a single package, such as multi-chip modules and surface mount technolo-gies, are becoming more attractive.

10

0.001

n

WHY IS MINIATURIZATIONIMPORTANT?

Miniaturization has inherent advantages,among them higher speed, lower cost, and greaterdensity. Smaller electronics devices are generallyfaster because the signals do not have to travel asfar within the device. Packing more functionalityinto a smaller or same-sized device reduces thecost of electronics. For example, a l-megabitDRAM chip has four times the memory capacityof a 256-kilobit DRAM, but costs only abouttwice as much and occupies the same space as thelower capacity chip. Since the number of compo-nents (e.g., chips) on a circuit board largely deter-mines the cost of the system, every l-megabitDRAM used instead of four 256-kilobit DRAMreduces the number of memory chips on a boardand reduces cost. Similarly, decreasing transistorsize and greater integration has caused the priceof logic devices to decline (see figure 1-3).

Miniaturization is important because it cancreate new markets by enabling new applications.Development of the microprocessor-a tiny com-

Figure 1-3-Price History of Electronic Logic

I Discrete d \ \transistors

r scale \ \I

Large-scale

i n t e g r a t i o n —

1 I I I I 1955 1960 1965 1970 1975 1980 1985 1990 1995

The price per gate-a circuit that performs a simple Iogic function -has continued on a steep downwardtrend since the introduction of the integrated circuit.SOURCE: John S. Mayo, “The Microelectronics in Communications,

(San Francisco, CA: Freeman & Co., 1977), p. 106. Copyright(c) 1977 ScientificAmerican, Inc. All rights reserved. Additional data from Graydon TexasInstruments, personal communication, Sept. 30, 1991.


puter on an integrated circuit—in 1970 led to astill expanding personal computer market cur-rently valued at $70 billion per year. Flat paneldisplays and improved chip packaging have led tobattery-powered portable personal computersthe size of notebooks (and some even smaller—the size of a checkbook), one of the newest mar-kets created by miniaturization. Personal com-munications is a developing major consumermarket created by miniaturization. With thesimultaneous reduction in the size of personalcomputers and cellular telephones, the two aremerging into a cordless appliance that can com-municate with the rest of the world through a net-work. Nippon Electric Co. (NEC) already offers alaptop computer with a built in cellular phone forsale in Japan; similar products are under devel-opment by American companies.

COMPETITIVENESS OF U.S.MINIATURIZATION

TECHNOLOGIES

The competitive position of U.S. R&Din min-iaturization technology remains strong, althoughcompetition from Japanese and European indus-try and governments has increased. As a result,the U.S. lead is less substantial than it was in thepast. The rejuvenation of postwar Japanese andEuropean economies has resulted in greaterquality and quantity of research in those nations.There are now many more sources of competitionin research worldwide.

Although U.S. R&D strength is still sound,U.S. industry has a mixed record in the imple-mentation of miniaturization technology. Manyminiaturization technologies that are crucial tothe success of consumer electronics, for example,were embraced by Japanese industry more quick-ly than by U.S. industry. Surface mount technolo-gy, which allows more electronic chips to be

placed on a circuit board, has had greater pene-tration into Japanese industry than U.S. industry(see figure 1-4). Concerned with packing moreelectronics into portable consumer products, e.g.,cameras, calculators, and stereo receivers, Japa-nese consumer electronics companies were eagerto reduce the size of their products. U.S. compa-nies that produce computers, industrial controls,and other large systems serve markets that arenot as concerned with size and portability.

Semiconductor

Shrinking the size of transistors and their inter-connections is being pursued vigorously by U.S.industry, government, and universities. Themainstream approaches to transistor miniatur-ization use silicon with designs similar to those ofpast devices. In many respects, the United Statesleads world research on smaller transistors. De-sign of miniaturized transistors draws heavily onbasic sciences and computer modeling, areas inwhich the United States remains strong. Imple-menting smaller transistors in products, however,is a strength of Japanese industry. DRAM chipshave historically been the first commercial prod-ucts produced with each new generation of semi-conductor manufacturing equipment. SinceDRAM chips are made primarily by Japanesecompanies, they are the leaders in implementingsmall transistors in products.

R&D in semiconductor technology is done pri-marily by industry in both the United States andJapan. The Federal Government spends about$0.5 billion per year on semiconductor R&D.2

Japanese merchant semiconductor firms, com-panies that sell chips to other companies, havebeen outspending U.S. merchant firms on R&D(see figure 1-5). Total U.S. and Japanese industryR&D spending is roughly equal; $3.7 billion isspent in the United States and $4 billion inJapan.3

@ffice of Management and Budget estimates for fiscal year 1990.3Dataquest estimates for 1990”


Figure 1-4-Use of Surface Mount Technologies

100%

o%Japan W. U.S. Japan W. U.S. Japan W. U.S.

Europe Europe Europe1985 1991 1995

Use of surface mount technologies (SMT), which can reduce the thickness and volume of electronic products, haslagged in the United States compared to other nations. This graph shows past, current, and projected estimates ofSMT use on printed circuit boards for Japan, Western Europe and the United States. - -

SOURCE: Office of Assess 1991. Data from VLSI R Inc. (Fall

Figure 1-5-Semiconductor R&D Spending-Top Four Merchant Companies in theUnited States and Japan

$3

Top fourJapanese merchant

companies$2 — (then-yr. conversion)

~ @ @- 0 0 - - - -

# & ~

- - - - - - - - -- ‘Top four Japanese

merchant companieso (avg. conversion)m

$1● U.S. merchant

companies

1985 1986 1987 1988 1989 1990

U.S. semiconductor merchant companies are outspent in R&D by Japanese merchant firms. There are twocurves for Japanese spending. The more dramatic increase results from converting yen to dollars at annualconversion rates. The other curve accounts for changes in the exchange rate by using a single conversion rateaveraged over 5 years.SOURCE: Office of Technology Assess ment, 1991. Data from Dataquest.


Quantum Effect Devices

Significant momentum is left in the currenttrend in silicon miniaturization. But what hap-pens when today’s silicon designs hit physicallimits in further miniaturization, as many expertsexpect will happen in about 10 years? One optionwill be to change the way switching devices (usedto make computer logic) are made. Several alter-native approaches are being pursued by theUnited States, Japan, and Europe. One ap-proach—quantum effect electronics—is still are-search topic, but is receiving worldwide support.In the United States, about $8 million are spentannually by the Federal Government in this area.Most of the research is being sponsored by theDefense Advanced Research Project Agency(DARPA), with significant efforts at other DoDresearch agencies and the National Science Foun-dation (NSF). Industry support for this basic re-search is sparse; AT&T Bell Laboratories andTexas Instruments both have research activitiesin the area. IBM has a small research effort aswell. Most of the research in the United States isbeing conducted at universities supported bygovernment agencies.

The United States leads in research related toquantum effect devices, but this lead competeswith other nations’ efforts. Japanese efforts aresignificant and are beginning to make an impact;the Japanese involvement at recent conferenceshas grown. The United States can expect to main-tain its lead in research, but, as the researchmoves toward development and implementation,the competition with Japanese researchers andcorporations grows more fierce. Japanese indus-tries have the advantage of being the world lead-ers in semiconductor processing technology andoptoelectronics technology.4 U.S. companieshave the advantage of proximity to the world’sleading researchers in the field. The JapaneseGovernment will spend less than $3 million in fis-cal year 1991 on quantum effect device research.With two new projects initiated in 1991 and 1992,

funding will increase to about $6 million per yearover the next 5 years. Japanese industry is activelypursuing quantum effect device research. Fujitsuand NTT have the two largest research efforts.Estimates of Japanese industrial investment inthe field are difficult to confirm, but the sum maybe more than double the government investment.The Ministry of International Trade and Industry(MITI) 1992 Large Scale Program, “AngstromTechnologies,” will also fund research in quantumelectronics. The program will spend 25 billion yen($183 million) over ten years to conduct researchon technologies at the scale of angstroms. It is notknown, however, how much funding will go toquantum electronics.

Molecular and Biological Computing

One proposed way to continue the miniaturiza-tion of computers is to use individual moleculesas switching devices in place of today’s semicon-ductor transistors. U.S. investment in researchrelated to molecule-based computing has dimin-ished substantially since a period of intense inter-est in the early 1980s. NSF was the principal fund-ing source for much of the original research.NSF’s current funding for molecule-based com-puting is a few hundred thousand dollars andmay end next year. Funding from U.S. industry islimited to a few venture capital firms and DigitalEquipment Corp. Much of the effort in theUnited States has turned to development of mo-lecular or biologically derived materials for appli-cations in computing-related areas. For example,films of molecules rather than individual mole-cules 5 are being considered for use in opticaldisks and other data storage technologies.

Japanese and European governments continueto fund molecular computing research. How-ever, differences in terminology complicatecomparisons of programs between countries.Portions of one of the Japanese Exploratory Re-search for Advanced Technology (ERATO) proj-ects are designed to pursue molecule-based com-

qMany ~ument app]imtions for suFr]attius—a structure used to create quantum effect detius-invoke oPto electronics.!$Films of bactenorh~owin, a light ~nsitfie biologi~] molecule, are being prepared by re~archers at syracu~ university, Mitsubishi, and a

few American companies.

8. Miniaturization Technologies

puting 6 and the Ministry of International Tradeand Industry (MITI) Basic Technologies forFuture Industries Program spends about 300 mil-lion yen ($4 million) on research for biologicalmolecular computing.

Packaging and Interconnection

As electronics become more ubiquitous, faster,and denser, the pressure on improving electronicpackaging technology will increase. Federal fund-ing has not addressed the issue of packaging as awhole, but many different programs fund re-search on packaging. Technologies such as sur-face mount technologies (SMT) and multichipmodule (MCM) technology can place more semi-conductor chips into a system than traditionalpackaging. SMT is maturing and is seeing wide-spread application. Over 50 companies and re-search institutions worldwide are now pursuingMCM or related technology.7 One focus of theMicroelectronics and Computing TechnologyConsortium (MCC)–a consortium of U.S. com-panies—is packaging and interconnection tech-nology, including MCM technology. A NationalResearch Council report in 199@ found that theAmerican and Japanese industries are aboutequally matched in printed wiring boards, multi-chip modules, and other interconnection technol-ogy. The report also found that U.S. industry wasdependent on Japanese suppliers for materialsrequired for several packaging technologies.

Biosensors and Chemical Sensors

There is intense worldwide interest in develop-ing biosensors and chemical sensors, particularly

in Japan. Biosensors can detect the presence ofchemicals or molecules such as glucose, urea, andoxygen. Diverse applications are seen in medicaldiagnosis, industrial process monitoring and con-trol, fermentation process control, and food qual-ity monitoring. Most of the R&D is being done byindustry. Within the Federal Government, theDepartment of Defense (DOD) is a primary sup-porter of biosensor R&D with the National Insti-tutes of Health supporting some development.

According to a 1989 Japanese Technology andEvaluation Center (JTEC) Panel report, the U.S.efforts in chemical and biological sensors trailthose of Japan.9 The panel report rated Japaneseefforts more advanced in commercialization,product development and quantity of basic re-search. In quality of research, the U.S. and Japa-nese efforts were considered equal. OTA inter-views in the biosensor industry indicate thesituation has not changed much since the JTECreport. According to the report, in 1985 therewere more Japanese publications and patents forbiosensors than U.S. publications and patents.10

A consortia has been established by 35 Japanesecompanies to conduct R&Din biological sensors.The annual budget for the consortia is about $2million from industry matched by funds from theMinistry of Agriculture, Forestry and Fisheries.ll

In addition, a large number of Japanese compan-ies are funding independent development of bio-sensors. 12

Micro-Mechanical Systems

Development of new manufacturing tech-niques by researchers at American universitiesduring the 1980s has led to an expansion of inter-est in micromechanics around the world. Build-

6Kunitake molecuiar architecture project (1987-92).TDenni~ Hemel and HasSan Hashemi, “Hybrid wafer Scale Integration, “ MCC Ikchnical Report P/I-329-89, 1989.8National Re=arch Council, commission on Engineering and Ikchnical Systems, National Materials Advisov Board, Matefia~~or Hi@-

Density Electronic Packuging and Interconnection (Wshington, DC: National Materials AdvisogI Board, March 1990).9u.s. Depa~ment of Commerw, C4JTECH (Japanese lkchnology Evaluation Program) Panel Report on Advanced Sensom in Japan,” JanUaV

1989, p. 184.IofibliMtions: 59 Japan, 35 United States; patents: 74 Japan, 9 United States.llHi& Tec~olo~ Business, September-October 1989, p. 29.12National science Foundation, JaPnew ~ChnoloW and Evaluation Center (~c), Viewgraphs for “~c Workshop on Bioprocess Engi-

neering in Japan” (Washington, DC: National Science Foundation, May 21, 1991); and U.S. Congress, Office of Technology Assessment, Bio-technology in a Global Economy, OTA-BA-494 (Washington, DC: U.S. Government Printing OffIce, October 1991).


ing upon established semiconductor processingtechniques, researchers have been able to fabri-cate elaborate mechanical structures: motors andturbines as wide as a hair (see photograph), canti-levers capable of measuring acceleration, andgear assemblies smaller than a fleck of dust.These new techniques are closely related to tech-niques traditionally used to create mechanicalstructures for use as pressure sensors. The firstapplications with sizeable markets for the newtechnology are in sensors and instrumentation.Future applications may involve micromechani-cal actuators or systems of actuators and sensors.Some niche applications are already using micro-mechanical actuators made with conventionalmilling and extrusion techniques (non-microelec-tronics techniques).

Researchers at U.S. universities are the ac-knowledged leaders in micromechanical sensorresearch. The European nations, especially Ger-many, have a strong technology base in sensortechnology and are supported with extensive gov-ernment funding. Germany’s Karlsruhe NuclearResearch Center and Fraunhofer Institute co-developed a new lithographic process for makingrelatively thick (hundreds of microns) microme-chanical structures, called LIGA. Although theirindustrial research in pressure and accelerationsensors is impressive, the Japanese trail theUnited States and Europe in R&D.

Although still relatively modest in scale, in thelast few years funding for micromechanics hasbeen increasing rapidly. In the United States, re-search in the field is sponsored by several differ-ent agencies, including DOD, NSF, the NationalInstitutes of Health (NIH), the National Aero-nautics and Space Administration (NASA), andthe Department of Energy (DOE). Total Federalspending in the field is planned to be over $15 mil-lion in 1991-from under $6 million in 1990 and1991. DOD and NSF were early supporters of thetechnology. DARPA has been spending about $2million per year, and plans to spend about $3 mil-lion per year in 1992. NSF has been funding muchof the research at universities since 1983 at about$2 million per year. Efforts at the DOE have beenmostly at Sandia and Lawrence Livermore Na-tional Laboratories. Starting in 1991, DOE willspend about $16 million over 3 years at LouisianaTechnical University (LTU) as directed by Con-gress. Other State governments including Cali-fornia and Louisiana have shown interest in pro-moting micromechanics research.

On the whole, U.S. industry is playing "waitand see” while trying to sort out what commercialapplications the technology might have. Sensorand instrumentation companies, however, seeclearer potential for applications and are pushingforward more aggressively. Analog Devices, Inc.,for example, has developed the first commercialproduct that uses the new manufacturing tech-nique—an acceleration sensor targeted for auto-mobile airbag deployment applications. Othercompanies are abandoning micromechanics. Apioneer in the field, AT&T Bell Laboratories, ter-minated its efforts in micromechanics. U.S. in-dustry spent over $20 million in fiscal year 1991on micromechanics R&D.

In Europe the largest and most advanced re-search efforts are being pursued by Germany. In1990, the Germany Federal Ministry for Researchand Technology initiated a 4-year program thatwill spend 400 million marks ($230 million) on re-search. The German Government recently an-nounced it will extend the program for at leastanother year. The Karlsruhe Nuclear ResearchCenter and the Fraunhofer Institute are conduct-


ing leading research in micromechanics. TheFraunhofer Institute is designed to encourageparticipation from industry and receives about 50percent of its funding for micromechanics fromindustry. A consortium of German companieshas licensed technology from Karlsruhe and isdeveloping industrial applications based on thetechnology. Germany companies are interested inapplications of the technology. Bosch has thelargest effort, and Siemens and Messerschmidtalso have significant R&D efforts in the field.

Other European nations and industries arealso pursuing micromechanics research. Themost notable research efforts besides Germanyare in Switzerland and the Netherlands. The In-stitute of Microelectronics in Neuchatel, Switzer-land, is developing solid-state sensors and is ac-tive in the field of micromechanics research. TheUniversity of Twente and the University of Delftin the Netherlands are active in the field, with anew institute formed at the University ofTwente—the Micro-Electromechanical Sensorsand Actuators (MESA) Institute. Almost everynation in Europe has research activity in micro-mechanics—much of which is at universities.

Perceiving themselves as behind in microme-chanics research, Japanese researchers are vigor-ously pursuing the technology with a new MITIprogram that will spend 25 billion yen ($183 mil-lion) over the next 10 years. The program, initi-ated in April of 1991, aims to develop microro-bots for health and industrial applications.Micromechanical systems will be a major part ofthe program’s research, although portions of theprogram will bean extension of a previous MITIprogram to develop miniature robots. Fiscal year1991 funding for the program is only about $3 mil-lion, but that will increase to over $20 million an-nually as the program accelerates to full speed.Japanese industry typically adds its own invest-ment of labor and equipment to work sponsoredby MITI, so the total research effort associatedwith this program is substantially larger than thegovernment investment. In addition to the MITIprogram, the Japanese Science and Technology

Agency (STA) funds research at several universi-ties in Japan, including the University of Tokyo,13

at a total of about $1 million per year. Japaneseindustry is pursuing research in the field, with1991 expenditures estimated at over $20 million.One of the largest industrial research efforts is atToyota’s central R&D facility in Nagoya. Otherindustrial research is underway at NTT, NECCorp., Ricoh Corp., IBM-Tokyo, and MatsushitaCentral Research. The MITI project has in-creased the interest of industry in micromechan-ics, and R&D can be expected to increase overthe next several years.

Fabrication Technology Researchand Development

Making miniaturized electronics, sensors, andmicromechanics requires increasingly sophisti-cated manufacturing equipment as each genera-tion of miniaturized components demands great-er precision in fabrication.

Lithography—the technique used to etch fea-tures into integrated circuits—is one of the mostchallenging hurdles for future miniaturization ofintegrated circuits. The size of a transistor, thelines connecting transistors, and other devices ina circuit can only be made as small as the resolu-tion of the tools used to make them. As the size oftransistors become smaller, making tools of ade-quate resolution becomes increasingly difficultand more expensive. The current approach to in-creasing the resolution of lithography has been toreduce the wavelength of the light source, pro-gressing from visible wavelengths, to ultraviolet,to deep ultraviolet. Now many experts are pre-dicting that the trend will continue to x-ray lithog-raphy, but the outcome is far from certain. Someexperts predict ultra-violet lithography will beuseful in creating features as small as those possi-ble with x-ray lithography. Using a techniquecalled phase–shifting, this approach would re-quire very sophisticated masks and computersoftware to be successful. Other lithographic ap-

lsothem include ~hoku University, University of Osaka Prefecture, and K..shu university.

Chapter 1-–Introduction and Summary ● 11

preaches such as electron-beam and ion-beamare considered viable options.

The debate surrounding x-ray lithography hasbeen well documented.14 The governments andindustries of the United States, Japan, and Euro-pean nations are all investing heavily in R&D forthe next generation of lithography tools. Congressallocated $60 million to DARPA in fiscal year1991 to develop x-ray lithography technology, in-cluding research on mask development and laserx-ray sources. There are a total of five synchro-

trons currently in the United States for lithogra-phy research. Two more will come on line by 1993in Baton Rouge, Louisiana and Upton, New York.Industrial support for x-ray lithography usingsynchrotrons consists primarily of IBM and Mo-torola. AT&T is focusing on x-ray lithography us-ing laser sources instead of synchrotrons. In Ja-pan and Germany, government and industry aretaking an aggressive approach to x-ray lithogra-phy development. There are nine synchrotrons inJapan and two synchrotrons in Berlin for lithog-raphy research.

laFor enmp]e, we Mark Crawford, “The Silicon Chip Race Advances Into X-rays,” Science, VO1. 246, Dec. 15, 1989, pp. 1382-1383; U.S.Congress, Congressional Budget Office, UsingR&l) Consom”a for Cornmmiallnnovahon: SEM4TECH, X-ray h”thography, and High-ResolutionSystems (Washington, DC: Congressional Budget Office, July 1990), pp. 69-87; U.S. Congress, Senate Committee on Commerce, Science, and‘llansportation, Subcommittee on Science, lkchnology, and Space, Sem”conductozs and the Electronics Zndustry, Serial No. 101-771, May 17,1990; and John Markoff, “Etching the Chips of the Future,” New Mrk Times, June 20, 1990, p. D1.

Chapter 2

Technology

INTEGRATED CIRCUITS

Since World War II, system designers havetried to create smaller electronics. Reduction inthe size of vacuum tubes progressed slowly aspractical limits were encountered. Replacementof vacuum tubes with transistors in the 1960s wasa breakthrough in miniaturization. Transistorsbecame smaller, but each transistor was pack-aged individually, thus mounting large numbersof transistors on a circuit board still resulted inlarge assembly units. Invention of the integratedcircuit overcame that problem by embeddingmany transistors on a single silicon chip thatcould be then connected to other components ona circuit board. Integrated circuits acceleratedthe drive toward transistor miniaturization be-cause, unlike discrete transistors, the connectionsbetween transistors could shrink with the transis-tor. The smaller the transistor and its intercon-nections, the greater the transistor density (thenumber of transistors in a given chip area). Aschip designers were able to pack more and moretransistors onto a single silicon chip, the numberof transistors per chip increased nearly ahundred fold each decade (see figure 2-l). As thesize of a transistor shrinks it operates faster, lead-ing to faster computation (see figure 2-2).

If this trend continues, sometime after 2000 thesmallest feature on an integrated circuit will beabout 0.1 micron (1 micron is 1 micrometer or onemillionth of a meter). For comparison, today’sdynamic random access memory (DRAM) tran-sistor has a smallest feature of about 0.8 micron.The equivalent capacity of a DRAM1 chip will beover 1 billion bits (1 gigabit). With O.1-microntransistors, microprocessors will contain over 400million transistors. Industry experts believe that

such densities are possible and that there are nofundamental physical limits to prevent achievingthem.

The time needed to overcoming engineeringand manufacturing problems to reach these highdensities is uncertain. Several problems must beovercome to develop high-density chips, includ-ing: 1) higher resistance of the minute connec-tions between transistors and within transistors;2) the tendency of very small transistors (about0.1 micron) to “leak,” rendering them useless; and3) lack of high-volume manufacturing equipmentcapable of creating very small features.

Fabrication technologies, especially lithogra-phy, have paced the miniaturization of transis-tors in the past and may ultimately determine thepractical economic limits of transistor miniatur-ization. Fabrication equipment used to makecircuits at 0.1 microns must be affordable andcapable of sustaining reliable, high-volume pro-duction. Electron-beam lithography can now pro-duce features much smaller than 0.1 micron, butis unable to manufacture a sufficient volume ofchips to be economical.2 X-ray lithography andphase-shifted masks (another lithography tech-nique using visible light) are the two most likelycontenders for achieving manufacturable O.l-mi-cron transistors. Other prospects include projec-tion electron beam, and projection focused ionbeam.

Physical Limits to Transistor Miniaturization

What are the prospects for increasing densitiesof transistor based integrated circuits beyond 1gigabit DRAM and 400 million transistor micro-processors? On this question, experts’ opinions

ID~ is the ~nmay means ofstonng information in a computer system temporanlywhile the computer is working on the info~ation. Massstorage (e.g., hard disk) is used for permanent storage.

zElectron.beam lithography is Currentb used in production of masks and may be useful in some production applimtion% e.g., manufacturingchips for high-performance computers.

-13-


Figure 2-1 –Transistor Trends

Decreasing transistor size. . .100

10

1

0.1

0.01

I

1970 1980 1990

and increasing die (chip) size. . .1

6 4 M A / n’

❑

I I I I n101970

of Technology Assessment,Texas Instruments.

1980 1990

is resulting in rapid growth in the number of transistors per chip.

10,OOO

100 I10

II I 1 I

1970 1980 1990

Data from Intel David Ferry, Arizona State and Graydon


Figure 2-3-The Photolithographic Process

Silicon dioxide

Photoresist

Ultraviolet radiation

Hardenedphotoresist

Photolithography is the process by which a microscopic pattern is transferred from a photomask to a layer of material in a circuit. In thisillustration a pattern is shown being etched into a silicon dioxide layer (shaded) on the surface of a silicon wafer. The oxidized wafer (1) is firstcoated with a layer of a light-sensitive material called photoresist (2) and then exposed to ultraviolet light through the photomask (3). Theexposure renders the photoresist insoluble in a developer solution; hence a pattern of the photoresist is Ieft wherever the mask is opaque (4).The wafer is next immersed in a solution of hydrofluoric acid, which selectively attacks the silicon dioxide, leaving the photoresist pattern andthe silicon substrate unaffected (5). In the final step the photoresist pattern is removed by means of another chemical treatment (6). Thereare variations on this process such as use of photoresists that become soluble instead of insoluble (4), and use of reactive gases instead ofliquid acid solutions for etching (5).

SOURCE: Adapted from William G. “The Fabrication of Microelectronic Circuits,” (San Francisco, CA: Freeman & Co., 1977), p. 47. Copyright (c) 1977 Scientific American Inc.–George Kelvin.

most common current technology uses visible or The current state-of-the-art photolithographyultraviolet wavelengths of light for the photolitho- uses ultraviolet light, typically with a wavelengthgraphic process. of less than 400 nanometers. Use of excimer laser

Chapter 2-–Technology .19

sources could achieve wavelengths less than 200nanometers. By improving mask technology, op-tics, and resists, ultraviolet systems may be us-able to 0.25- or 0.2-micron minimum featuresize—dimensions that are expected to be neededin the mid-1990s. There are several lithographytechnologies that can create features smaller than0.25 or 0.2 microns. The likely candidates for thenext generation of lithography tools are x-ray li-thographies, optical lithographies using phaseshift masks, electron-beam lithographies, andion-beam lithographies. Combinations of thetechniques are also possible. (For a more detaileddescription of the photolithographic process, seeapp. A.)

Materials and the principles of surfacesciences are fundamental to advancing miniatur-ization of transistors and electronics. Improve-ments in semiconductor manufacturing rely onunderstanding how materials react to processingand fabrication techniques. Physical phenomenaunderlying the critical fabrication steps are notwell understood; knowledge has developed most-ly from experience, not derivation from physicallaws. Better understanding of the physical lawsaffecting small-scale structures would accelerateadvances in miniaturization.

There are more speculative ideas for futurefabrication technologies that would require evenmore rigorous understanding of surface interac-tions. Some exploratory work is currently under-way to use proximal probes to fabricate inte-grated circuits. Using a scanning tunnelingmicroscope (STM) or some variation of an STM,scientists are beginning to fashion crude struc-tures from individual atoms and clusters ofatoms. Researchers have been able to move ordeposit atoms on or below surfaces to “draw”figures, maps, and company logos (see photo-graph). Even if these techniques do not result inusable manufacturing tools, they will increase ourunderstanding of surface interactions of atoms,leading to better understanding of traditionalsemiconductor manufacturing. Even more spec-ulative is the prospect of creating molecular-sizedrobots and machines that could be programmed

Photo credit: IBM

Scientists at IBM’s Almaden Research Center used a scanningtunneling microscope (STM) to move xenon atoms around on anickel surface and spell out the company name. The distancebetween each atom of xenon is about 13 angstroms.

to manufacture virtually any molecular structure(see box 2-A).

COMPUTING SYSTEMSTECHNOLOGIES

Computer systems have shrunk dramaticallyduring the last 30 years. Mainframe computersonce filled large rooms and required air condi-tioning to dissipate the heat generated by vacuumtubes and early transistors. Computers have be-come so small that laptops and notebook com-puters are more powerful than mainframe com-puters of 10 years ago. Equivalent power will soonbe available in a checkbook-size package. Pastminiaturization of electronics has played the ma-jor role in bringing this capability about and fu-ture miniaturization advances will be driven pri-marily by electronics (see figure 2-4). Othercomponents in addition to integrated circuitsalso had to shrink in size to accommodate today’snotebook-sized computers.Increased massmemory density, greater circuit board density,and thinner flat panel displays have improvedcomputer system performance and alloweddownsizing. Trends toward smallness are ex-pected to continue as customers demand higherquality, higher performance, and portability fromsmaller boxes.

Chapter 2–Technology .17

The technical problems associated with mak-ing practical memory or logic systems from mo-lecular devices are substantial and are not likelyto be solved in this decade. The most vexing prob-lem is the interface (connection) between the mol-ecule and the circuit itself, a problem common tomany forms of miniature transistors. At 0.1 mi-cron, transistors begin to experience problemswith high resistance, but a molecular transistorwould be smaller still—0.001 microns or less. Inorder to avoid the electrical interface problem,some researchers are exploring optical interfaces.The diameter of an optical interface, however, islimited by the wavelength of light. The stability ofthe molecules is also a problem. To be useful forcomputation, a molecule must remain in a specif-ic configuration until changed by an external sig-nal. Individual molecules sometimes change con-figurations unexpectedly. Using clusters ofmolecules makes them easier to connect to cir-cuits and reduces the errors in the stored data,but makes the computing device larger.

Related research is underway in the use ofbiologically derived or organic material for stor-ing information or computing. These approachesminimize the interface and stability problems ofindividual molecule computation. A few U.S.companies are working on organic materials foroptical disks and at least one Japanese firm, Mit-subishi, is working on an optical disk that will usebacteriorhodopsin as the storage material.5 Fewmaterials have suitable properties for computing;this has been a major stumbling block for molec-ular computing. U.S. research is directed at bet-ter understanding materials that might serve asthe basis of future molecule-based computers.The more immediate results may be faster datastorage, but proponents of molecule-based com-puting hope that the experience gained will leadto true molecule-based computing and process-ing systems in the future.

Fabrication Technologies

There are several manufacturing technologiesthat are important in making semiconductor de-vices. The most critical manufacturing technolo-gy by many accounts, is lithography–the tech-niques used to pattern and etch transistors andtheir interconnections on a substrate like siliconor GaAs.

There are several different lithographic tech-nologies including photolithography, electron-beam lithography, ion-beam lithography, and x-ray lithography. In addition, there are severalnovel technologies still in early stages of develop-ment—e.g., scanned photolithography, com-bined patterning and growth, and proximal probefabrication.

One of the most critical steps in lithography isexposing the resist—usually an organic com-pound layered on top of the semiconductor wa-fer–to an energy source (see figure 2-3). Theenergy source can be optical, ultraviolet, or x-ray.It can also be a beam of charged electrons or abeam of charged atoms (ions). Some of the resistis exposed to the energy source and some is not,depending on whether the resist lies under atransparent or opaque portion of the mask. In theplaces exposed to the energy source, the resist ismodified so that certain chemicals can dissolveit.6 Exposing the wafer with a series of etching

chemicals removes the resist and a layer of mate-rial (usually silicon dioxide) underneath. The re-maining resist is then removed by another chemi-cal that does not affect the semiconductormaterial. A layer of material is typically depos-ited again and the process is repeated—as manyas 12 or more times—to make an integrated cir-cuit. The minimum feature size that can beformed is determined primarily by the precisionthat the energy source can be focused to discrimi-nate between the areas of the resist that are ex-posed to the energy and those that are not.7 The

5Robefi Birg~, Syacuw University, personal communication, Ju& 29) 1991-

%ere are different types of resists. Exposure to radiation causes some types to become susceptible to the subsequent etching step, and causesothers to become resistant to etching.

TOther key factom are the ability to align subsequent layers of masks to create the proper vertical geometxy and the ability to control the rateand direction of etching.


Figure 2-3-The Photolithographic Process

Silicon dioxide

/ Photoresist/ /

Ultraviolet radiation

Hardenedphotoresist

Photolithography is the process by which a microscopic pattern is transferred from a photomask to a layer of material in a circuit. In thisillustration a pattern is shown being etched into a silicon dioxide layer (shaded) on the surface of a silicon wafer. The oxidized wafer (1) is firstcoated with a layer of a light-sensitive material called photoresist (2) and then exposed to ultraviolet light through the photomask (3). Theexposure renders the photoresist insoluble in a developer solution; hence a pattern of the photoresist is Ieft wherever the mask is opaque (4).The wafer is next immersed in a solution of hydrofluoric acid, which selectively attacks the silicon dioxide, leaving the photoresist pattern andthe silicon substrate unaffected (5). In the final step the photoresist pattern is removed by means of another chemical treatment (6). Thereare variations on this process such as use of photoresists that become soluble instead of insoluble (4), and use of reactive gases instead ofliquid acid solutions for etching (5).

SOURCE: Adapted from William G. Oldham, “The Fabrication of Microelectronic Circuits,” Microekctronics (San Francisco, CA: W.H.Freeman & Co., 1977), p. 47. Copyright (c) 1977 Scientific American Inc.–George V. Kelvin.

most common current technology uses visible or The current state-of-the-art photolithographyultraviolet wavelengths of light for the photolitho- uses ultraviolet light, typically with a wavelengthgraphic process. of less than 400 nanometers. Use of excimer laser


sources could achieve wavelengths less than 200nanometers. By improving mask technology, op-tics, and resists, ultraviolet systems may be us-able to 0.25- or 0.2-micron minimum featuresize—dimensions that are expected to be neededin the mid-1990s. There are several lithographytechnologies that can create features smaller than0.25 or 0.2 microns. The likely candidates for thenext generation of lithography tools are x-ray li-thographies, optical lithographies using phaseshift masks, electron-beam lithographies, andion-beam lithographies. Combinations of thetechniques are also possible. (For a more detaileddescription of the photolithographic process, seeapp. A.)

Materials and the principles of surfacesciences are fundamental to advancing miniatur-ization of transistors and electronics. Improve-ments in semiconductor manufacturing rely onunderstanding how materials react to processingand fabrication techniques. Physical phenomenaunderlying the critical fabrication steps are notwell understood; knowledge has developed most-ly from experience, not derivation from physicallaws. Better understanding of the physical lawsaffecting small-scale structures would accelerateadvances in miniaturization.

There are more speculative ideas for futurefabrication technologies that would require evenmore rigorous understanding of surface interac-tions. Some exploratory work is currently under-way to use proximal probes to fabricate inte-grated circuits. Using a scanning tunnelingmicroscope (STM) or some variation of an STM,scientists are beginning to fashion crude struc-tures from individual atoms and clusters ofatoms. Researchers have been able to move ordeposit atoms on or below surfaces to “draw”figures, maps, and company logos (see photo-graph). Even if these techniques do not result inusable manufacturing tools, they will increase ourunderstanding of surface interactions of atoms,leading to better understanding of traditionalsemiconductor manufacturing. Even more spec-ulative is the prospect of creating molecular-sizedrobots and machines that could be programmed

Photo credit: IBM

Scientists at IBM’s Almaden Research Center used a scanningtunneling microscope (STM) to move xenon atoms around on anickel surface and spell out the company name. The distancebetween each atom of xenon is about 13 angstroms.

to manufacture virtually any molecular structure(see box 2-A).

COMPUTING SYSTEMSTECHNOLOGIES

Computer systems have shrunk dramaticallyduring the last 30 years. Mainframe computersonce filled large rooms and required air condi-tioning to dissipate the heat generated by vacuumtubes and early transistors. Computers have be-come so small that laptops and notebook com-puters are more powerful than mainframe com-puters of 10 years ago. Equivalent power will soonbe available in a checkbook-size package. Pastminiaturization of electronics has played the ma-jor role in bringing this capability about and fu-ture miniaturization advances will be driven pri-marily by electronics (see figure 2-4). Othercomponents in addition to integrated circuitsalso had to shrink in size to accommodate today’snotebook-sized computers.Increased massmemory density, greater circuit board density,and thinner flat panel displays have improvedcomputer system performance and alloweddownsizing. Trends toward smallness are ex-pected to continue as customers demand higherquality, higher performance, and portability fromsmaller boxes.


Box 2-A—Molecular Machines

With electronics continuing to shrink and advances in the ability to make tiny mechanical devices,one wonders if there is a limit to the ability to manipulate small structures. Some theorists speculate thathumans could control individual molecules precisely enough to build molecule-sized machines and ro-bots. These machines would be so small that millions would fit in one of the micromotors described else-where in this report. Various terms have been used to describe the concepts;l here they will be referred toas “molecular machines.”

The process of DNA replication and protein generation demonstrates that molecules can store in-formation and use it to fabricate complex molecular structures. Molecular machine theorists claim hu-mans will be able to create molecular machines to perform similar functions.2 These molecular robotscould be programmed to manufacture virtually any molecule-based structure Possible--everything fromhamburgers to spaceships. According to proponents, such technology could control pollutants, createflawless materials, and provide almost limitless computing power. But they warn that there are potential-ly dangerous applications such as weapons more powerful than nuclear bombs, or machines that repli-cate uncontrollably, reducing the earth to a “gray goo.”

Significant barriers prevent the immediate implementation of molecular machine concepts. Theonly tools that can directly manipulate molecules are proximal probes, e.g. the scanning tunneling micro-scope (STM). Proximal probes allow imaging and manipulation of atoms on a surface. Although STMtechnology is advancing rapidly, manipulating molecules and atoms is awkward and time consuming-asimple molecule has yet to be fabricated.3 Problems with reliability of molecular-sized systems mightmake them impractical or delay their development. Because of their extremely small size, molecularmachines are especially susceptible to influences such as thermal noise and radiation damage. Molecularmachine system designers must compensate for these damaging effects. In many molecular machineapplication concepts, e.g., random access memory (RAM) or gene sequencing, molecular scale systemsmust interface with larger scale systems. In addition, molecular machines must interface with the outsideworld in order to be “programmed.” The problems of interfacing molecules to larger scale systems ham-per the ability to miniaturize many devices,4 including many of those proposed by molecular machinedesigners.

Should Policymakers Be Concerned?

Because of the tremendous impact molecular machine technology might have, there are calls forgovernment to monitor progress in the field and fund research toward its realization.5 OTA attempted todetermine the importance of molecular machine ideas to policymakers by assessing the basis of the mo-lecular machine concepts and prospects for their development.

lot~er le~ ~ cment~ and in the past include “nanotechnolo~,” “Feynman machines,” “eutaxic control,” “=mblem,”and “narmbots.” The term “nanotechnol@’ is particularly confusing because it is also used to refer to sub-micron electronics,micro-sensors, quantum electronics, and micromechanics. As a result, molecular machine concepts are often portrayed in the samelight as those other technologies that are very different in size, technology base, and time to realization.

ZFor more ~e~iled d~riptions of molemdar machine concepts, see ~~ Drexler, En@s of c~~n (N~wyork ~: ‘ChorBooks, 1986) or references 7 and 8.

sDonald Eider, IBM ~~den R_ch Center, personal communication, May 1991. Fabrication of a simPle molecule areasonable near-term expectation.

Xsec t~ dc~ptions Of transistor miniaturization in this chapter for details.Sb Sumn G. Hadden et a#.,A.r$&g j##o~ec~~ and Atomic Scale Technologies (MAST) (Austin, ~: Universiq of ~ms at

Austin Board of Regents, 1989); and Chris Peterson, “Molecular Manufacturing for Space,” Forsight Updlzte, No. 12, p. 8.

Conflmmd on next page

Chapter 2–Technology ● 21

Despite the skepticism of many researchers contacted by OTAand the evident controversial natureof their ideas, there has been little written criticism of molecular machine Concepts.6 The two seminalarticles on the topic of molecular machines were written by Richard Feynman7 and Eric Drexler.8 Thearguments for and against molecular machines tend to be conceptual and not refined to the point ofdiscussing architectures or systems the debate so far seems to center on whether such devices are possi-ble within bounds of natural law. Written criticism tends to focus on more specific suggestions and archi-tectures in the related field of molecule-based electronics rather than molecular machine concepts. Thescarcity of criticism maybe due to the reluctance of scientists to denounce new concepts in publications.9

There are many concepts of what impossible within the bounds of natural law--e.g., steam-poweredcomputers and interstellar travel-but they do not exist as technologies. While science can determinewhether a concept is feasible, technology development is influenced by unpredictable economic and so-cial factors. As a technology, molecular machines are non-existent; the only work to date has been con-ceptual and computational modeling.

When will molecular machine technology be developed? Estimates from the proponents of the con-cepts are 10 to 30 years, while others predict from centuries to “never.” One of the basic components of amolecular machine technology base would be a protoassembler, a molecular machine capable of fabri-cating other molecular machines, The earliest prediction for development of protoassemblers is 5 to 10years. l0

Is There a Government Role?

To date, no proposal for research on molecular machine development has been received by a Feder-al agency. Basic scientific and engineering research in the fields of materials science, chemistry, molecu-lar biology, advanced electronics, molecular modeling, and surface science are being funded by manyFederal agencies and would be necessary precursors to the realizationof molecular machines. It is impos-sible to estimate the level of funding, however, since there is no exact definition of precursor technolo-gies. There are a few small research efforts explicitly addressing molecular machines concepts in U.S.academia and industry and Japanese government.ll

Development of a framework for government regulation and oversight of molecular machine tech-nologies has been suggested by several analysts, driven by fears of abuse or accidents associated with thedevelopment of these technologies. The communities of researchers working on these precursor tech-nologies is rather small and the concern over accidents or misuse of the technology is well known amongthem. Government regulation at this stage would be premature, might hamper emerging research ef-forts, and have uncertain advantages. The question of regulation and oversight should be revisited andanalyzed in greater depth if developments in the field bring the technology closer to reality. The develop-ment of the first protoassembler might bean appropriate milestone to reconsider government regulatoryinvolvement.

%YEAccmducted alittxature seamhof articles that reference the two articles and analyzed them for critical and supporting argu-xnents.

TR, FWnn, ~~erc~~ p~enw of Room at the ~ttom,}> Mi.niazurizution, A. Gilbert (cd.) (New York, NY: Reinhold, 196~), PP.282-2%.

8KR ~re~er, “Molecular Engineer@: ArI Approach to the Development of General Capabilities for Molecular Manipula-tion: Proceedings of the NationulAcademy of Science USA, vol. 78, No. 9, September 1981, pp. 5275-5278.

gRolf Lan&uer, ‘*Poor Signal to Noise Ratio in Science,” f?Y~ “c Patterns in Compkx Systems, J.A.S. KeIso, A.J. Mandell, M.F.$$hlesinger (eds.) (Singapore: World Scientific, 1988), pp. 388-394.

IQwC Dr~yder, For~i@t Institute, personal communication, March 1991.~l~n tie Unitd Statm, one researcher is Performing mohcular modeling at Xerox Palo ~fo Research center. A ~@wrofit Owa-

nization, Institute for Molecular Manufacturing, in Palo AJto, Gliforniawas formed this year and plans to qxmwr research in thefi.Iture. No projects in .lapan or Europe are explicitly directed at xnolecular machine development, although the ~otani MolecuhDynamics project sponsored under the Japanese Iilx#oratory Research for Advanwd %chnology (E-) program addressedmolecular machine concepts in addition to its regular line of scientific investigation.

297-927 0 - 91 -- 3 : (IL 3


Smaller transistor sizeallows more transistorsper area

Figure 24–Three Approaches to Electronics Miniaturization

Make Transistors Smaller

❑

❑ Make Die LargerLarger area allows moretransistors per die (chip)

Multi-chip modulesand

multi-chip packages

Three basic strategies are used to miniaturize electronics systems. By making each transistor smaller (l), more can be placed on an inte-grated circuit, incorporating the functions of other chips. Increasing the size of the chip (2) has a similar effect by increasing the availablearea for transistors and their interconnections. These first two approaches have been the driving force behind electronics miniaturization forat Ieast the past two decades. Improved packaging (3) is a way to improve use of space on a printed circuit board. Atypical printed circuitboard has only a small fraction of its space covered with integrated circuits; the rest is primarily packaging and interconnection.

SOURCE: Office of Technology Assessment, 1991.


Display Technology

The size of a typical cathode ray tube (CRT)desktop computer display is about a cubic foot.New technologies, e.g., as liquid crystal display(LCD) and electroluminescence displays, haveallowed the development of flat panel displaysthat occupy a tenth of the volume. Flat paneldisplay technologies are discussed in detail in theOTA report The Big Picture: HDTV & High Reso-lution Systems. New technologies required to ad-vance high-resolution video technology are manyof the same that are required for advanced elec-tronics and more powerful computer systems.8

Japanese producers of flat panel displays favorLCD technology and are increasingly adapting itto larger screens, but current fabrication tech-niques are limiting screen size to about 15 inches.Further increases in screen size await improvedlithography.

Data Storage

“Mass Storage” refers to storage technologieswith large capacities, including optical and mag-netic media that retain data when the power is off.These devices encode data on a surface that canchange its magnetic or optical characteristics. Bymoving the surface under a device that can senseand change the characteristics of the surface—aread/write head—data is stored and retrieved.

Semiconductor memory storage increases ca-pacity as constraints on transistor miniaturiza-tion can be overcome, but mass storage operateson different physical principles and therefore hasdifferent limitations. There are two major typesof mass storage technologies—magnetic and op-tical. The trends in magnetic storage show no signof slowing as they approach densities that will put

1 billion bits onto 1 square inch. Researchers atIBM’s Almaden Research Center have demon-strated such densities.9 Today’s storage densitiesare about 50 to 100 million bits per square inch inhigh-end storage systems. Optical storage tech-nology uses lasers to write and read data from anoptical disk and is limited by the size of the laserbeam spot on the disk. The spot size, in turn, islimited by the wavelength of the laser source.Most optical disks today use a red (wavelength of800 nanometers) laser to write and read data.Efforts are underway now to develop blue lasers(wavelength of 400 to 500 nanometers) that wouldhave smaller spot sizes and be able to increase thecapacity of optical disks by four times.

Another technique that promises much higherdensities uses STM1O to write and read data. TheSTM and related instruments use sharp tips inclose proximity to a surface to create an atom-by-atom image of the surface. These tools also havethe capability to physically modify surfaces byeither etching away or building up a few hundredatoms at a time. Such techniques could result inmassive data density-more than 1,000 billionbits (1 terabit) per square centimeter. The majorproblem with STM and related approaches todata storage is that the data access speed is pain-fully slOW.ll There is research currently underwayat a few labs to combine many tips in parallel,increasing the speed of writing and reading.These approaches remain speculative, but mightyield useful technology in the long term.

Interconnection and Packaging

The basic packaging component of most com-puter systems is one or more circuit boards. Vari-ous components are placed on the circuit boardand interconnected with strips of metal. Due tomanufacturing costs, the number of components

8~e techniques ~wd t. make ~D displays are many of the same that are used to make integrated Circuits. For more information see OTA’sBackground Paper, The Big Picture: HDTV& High-Resolution Systems, OTA-BP-CIT-64 (Washington, DC: U.S. Government Printing Office,June 1990}

gRobefi M. ~ite, “Peripherals,” IEEE Spectrum, February 1990, pp. 28-30.IOSee app. A for details on the STM.none ~stimate is that 32 ~entunes would be required t. ~te one ~uare centimeter with one profirna] probe. see James S. Murday and

Richard J. Colton, “Proximal Probes: lkchniques for Measuring at the Nanometer Scale,” in ChemiszryandPhysics of Solid Surfaces, R. Vanelowand R. Howe (eds.) (New York, NY: Spnnger-Verlang, 1990), p. 347.


is the greatest contributor to the overall cost of anelectronics system (see figure 2-5). System design-ers seek to reduce the number of components on aboard to improve reliability as well. The intercon-necting strips of metal on a board and connec-tions between boards are more likely to breakthan connections elsewhere in the system. Eversince the invention of the integrated circuit, thebest way to reduce costs and increase reliability isby putting more transistors on integrated cir-cuits, hence reducing the number of circuit boardcomponents. By decreasing the size of each tran-sistor and their interconnections, more transis-tors have been crammed onto integrated circuits.

While the density of circuits on a chip havebeen increasing, the size of the chip has beenincreasing as well. The first microprocessor in1971 was on a 19,000 square mil (1 mil = one-thousandth of an inch) chip. For the next 20 years,chip size increased by about 14 percent per year.Current microprocessors are on chips of about260,000 square roils. Chip size may rise even fast-er over the next few years, yielding a chip size of625,000 square roils in production by 1994.12

Since the probability that a chip is defective in-creases with its size, chip size is limited by thenumber of defects on a wafer (see figure 2-6).

If the increase in chip size is taken to its limit,an entire wafer can become one large circuit.Since a typical wafer produces about a hundredchips, a huge number of transistors could be puton one wafer. But there are significant problemswith such an approach. There are defects in everywafer (primarily caused by particulate from themanufacturing equipment), so a wafer-scale cir-cuit is virtually guaranteed to have defects. Withthe high defect rate, strategies for routing connec-tions around defective components must be con-sidered in circuit design. Another problem is thatthe complexity of interconnecting wafer-scale cir-cuits increases the number of processing steps.

Finally, the fine lines of metal used to link theimbedded transistors together are not suited forthe relatively long interconnections needed forwafer-scale integrated systems. There are somespecialized applications that could use wafer-scale integration such as image processing andlarge memory storage,

13 but unless fabricationproblems are solved, it will likely not be an impor-tant technology. Applications that have regulardesigns —e.g., memory is essentially the same cir-cuit pattern repeated over and over—haveachieved marginal success with wafer-scale inte-gration because errors in the circuit are relativelyeasy to detect and correct or bypass.

Hybrid wafer-scale integration—known asmultichip module (MCM)—is another approachto increasing circuit density. MCMs take severalchips, place them on a substrate (usually siliconor a ceramic), and connect them with thin films ofconducting metal using techniques like thoseused in integrated circuits. Multichip modulesplace chips closer together than is possible usingsingle chip packages —an important advantage aschip speeds increase. The distance a signal cantravel in one clock cycle—the “heartbeat” of acomputer system—shrinks as cycle times short-en. Machines with clock cycles of 2 nanosecondswill soon be available,14 which corresponds to 500megahertz. (The fastest of today’s personal com-puters operate at about 33 to 50 megahertz.) Reli-ability is also enhanced because the number ofprinted circuit board connections and board-to-board connections are reduced.

MCMs have been successfully used in the IBM3090 and the NEC SX-AP; high manufacturingcosts, however, keep them from being used inmass production items like personal computersand workstations.15 The Microelectronics &Computer Technology Corp. (MCC)–-a privateconsortium of electronics companies—has devel-oped a MCM technology capable of producing

lzJe~ Sulllvan, c’~e Nem Generation of Electronics Design Automation Technology for Systems in Silicon,” MCC l’kchnical Repofl No.CAD-043-90, Jan. 24, 1990.

13’(A Dream Remembered,” The EconomisI, Nov. 17, 1990, p. 12.IdDennis He~ell and HasSan Hashemi, “Hybrid Wafer Scale Integration,” MCC llchnieal Report P/I-329-89, 1989.IsIbid.


Figure 2-5–Relative Cost of a Chip at Different Levels of Interconnection

Chip in wafer, untested

Testing and yieldper good chip

Package, packagingand testing

Space on printed-circuit board

Share of back panel and wiringShare of cabinet and

power supply

Total

Averagecost

$0.55

$0.15

$1.00

$0.35

$2.63

Space on the printed circuit board is one of the most expensive commodities in a system design. By integrating more functions onto a chip,system designers can reduce the number of chips on board, lowering overall system cost. The illustration shows the steps that a chip goesthrough on its way to becoming part of a computer system and the relative cost per chip for each stage.

SOURCE: Adapted from Ivan E. Sutherland and A. Mead, “Microelectronics and Computer Science,” (SanFrancisco, CA: Freeman &Co., 1977), p. 113. Copyright(c) 1977 by Scientific American, Inc. All rights reserved. Data fromGraydon Texas Instruments, personal communications, Sept. 20, 1991.


Full wafer

Figure 2-6- Relationship of Die Size and Yield

■ Defects

Yield = O

Large die

Yield = 25%

Small die

Yield = 71%

Smaller die (chip) size means higher yield (percentage of good chips) for a given density of defects.

SOURCE: Adapted from Sand, Systems, Howard K. (cd.) (Scottsdale, AZ: DM Data Inc.,1986), p. 7-12.

computers that are fast (3 nanosecond cycle time,or 333 megahertz), use 10 million gates, and cost$5,000. Costs must be reduced further for thetechnology to be affordable for workstations andpersonal computers.

A major obstacle to widespread use of MCMsis the lack of standardized interfaces for the hard-ware and software needed to design the modules.For the last 30 years the electronics industry hasbeen building an infrastructure that is focused onthe same objective: improving the performance ofthe integrated circuit. All the manufacturingequipment and computer software is geared fordesigning and optimizing a single chip. It is diffi-cult to obtain “naked” chips (dice) outside theirpackage from semiconductor vendors. Com-puter-aided design and testing are all difficultwith MCMs.

Dissipating the heat generated by dense cir-cuitry is a major problem facing circuit designers.This becomes more difficult as the number oftransistors in a chip increases. The simplest ap-proach to accomplish this in a computer system isto flow air over the circuits. In high-performancesystems (supercomputers and some main-frames), however, where speed is critical, othermeans for cooling are often used. Some super-computers flow inert fluids, e.g., freon, directlyover the circuitry. Sometimes fluid distributionsystems attached directly to the chip package are

used to dissipate heat, a scheme used in severalmainframe computers.

Many options for increasing the density of elec-tronic circuits now being investigated operatebest at very low temperatures and require exter-nal refrigeration units. Advancements in refriger-ation and heat dissipation technology are impor-tant to pushing the miniaturization of electronicsdevices. This is because as the cross-section of aconductor decreases, the resistance to currentflow rises. Greater resistance means greater heatgeneration because the same current flowingthrough a high resistance material creates moreheat than in a low resistance material.

The need for better interconnection is greatestin two areas: 1) consumer electronics and 2) high-performance computers. Consumer electronicsmust combine low costs with high dependability.High-performance computers demand ultra-fastcomputing speeds and extensive interconnection;while the field is highly competitive, cost is less afactor than performance. In high-performancecomputing, the need for better interconnectionresults from the short distance that an electronicsignal can travel before another clock cycle starts.Supercomputers push clock cycles to the limit, sodesigners must devise ways to minimize intercon-nection distance. The Cray supercomputers havean unconventional circular shape in order to keepcomponents close together and reduce the time


needed to communicate among components.Massively parallel computers face an even moreserious interconnection problem because manymore components (processors) must be con-nected. A conventional serial (von Neumann ar-chitecture) mainframe has only one central pro-cessing unit (CPU) that is centrally located sothat other components can be closely connectedto the CPU. A massively parallel computer mayhave over 1,000 CPUs that must be intercon-nected.

In consumer electronics, a major objective is tomake the product smaller, lighter, cheaper, andmore reliable. Each of these objectives demandbetter interconnection and packaging. Japaneseindustry, with its emphasis on consumer electron-ics, tends to adopt packaging innovations to meetthe demands of the market. The impact can beseen in the many consumer products around us(see figure 2-7).

SENSORS

Sensors are devices that can monitor andtranslate observed conditions—light level, accel-eration, pressure, or temperature—into a signal.The signal can then be transmitted, processedthrough a system or stored as data. A mercurythermometer, for example, responds to an envi-ronmental condition—temperature—and trans-lates to a visual readout indicated by the level ofmercury gauged against a calibrated scale. Anelectrical temperature sensor might sense thechange in temperature by measuring the changein voltage across a material. The resulting electri-cal signal can be manipulated and displayed forread-out.

There are a wide variety of sensors; each can beclassified as imaging or non-imaging.16 Imagingsensors take many measurements of radiated en-ergy from the imaged target. Camcorders, forexample, use a charge coupled device (CCD) asan imaging sensor to convert a visible light imageto an electrical signal that can be recorded on a

Figure 2-7–impact of Surface Mount Technologieson Calculator Thickness

1969

1973

1975

1976

1977

1978

1979

26mm

9mm

7mm

5mm

1.6mm

Implementing surface mount technologies in calculators had adramatic impact on thickness. Surface mount technologies in-clude surface mounted packages as well as surface mounted key-boards and displays.SOURCE: VLSI Research.

VCR tape. Non-imaging sensors act at a singlepoint, typically in contact with the object beingsensed. Sensors can be made to sense differentsubstances and energies: electromagnetic radi-ation, temperature, pressure, acceleration, chem-icals, and biological materials. Of the many sen-sor technologies, a few are key to miniaturizingsystems: CCDs, chemical and biological sensors,and micromechanical sensors.

Charge Coupled Devices

CCDs are solid-state image sensors that detectlight. They are used in cameras—from camcord-ers to professional cameras to astronomical cam-

t. this include some tactile sensors that detect sensations of touch an of


eras on satellites. The resolution of a CCD isdetermined by the number of picture elements(pixels) on a CCD. The devices are a special kindof integrated circuit with problems similar to oth-er integrated circuit manufacturing. Reducingthe size of each pixel improves the pixel densityand the image resolution. Increasing the size ofthe CCD can increase the number of pixels, butdefects in the circuit cause yield to plummet asthe CCD size increases. Manufacture of CCDs isdominated by Japanese industry. More than 20Japanese companies manufacture CCD chipsand they are the leaders in small pixel size, andinnovative design.

17 In the United States, Tek-tronix, Texas Instruments, and Kodak produceCCDs designed for military, industrial and spaceapplications. The majority of the image-sensingmarket is in consumer electronics.18

Imaging sensors are particularly important tothe Department of Defense (DOD). DOD relieson imaging data for things like aiming missiles,detecting rocket launches, and enhancing night-vision. Most DOD applications sense areas of theelectromagnetic spectrum that are unique to mili-tary needs.

Chemical and Biological Sensors

By combining information technologies withbiotechnologies, researchers are developing newsensors that are cheaper, faster, more versatile,and more efficient than previous generations.These sensors, known as biosensors, can detectgases, chemicals, and biological molecules. Thefirst biosensor was invented in 1%2 by Leland C.Clark Jr., but it was not until the 1970s that bio-sensors came into practical use. The early biolog-ical and chemical sensors were relatively large;Clark’s first glucose sensor was about a centime-

ter in diameter.19 There are many ways to makebiological and chemical sensors; some can bemade using techniques borrowed from integratedcircuit manufacturing. By applying this technolo-gy, it is now possible to make sensors that are onlya few thousandths of a centimeter wide.

Smaller biosensors have two advantages: 1)they can be placed in areas that were previouslyinaccessible, and 2) they are fast since the mea-surement can be done in the field instead of at acentral lab. Because they offer similar economiesof production as microelectronics, biosensorswill become cheap and widely available. Biosen-sors will be useful in portable systems. A marketanalysis by Arthur D. Little, Inc. in 1991 pre-dicted that portable biosensors sales would reach$1 billion by 2000.20 With sensors for substanceslike glucose, urea, and carbon monoxide availablein an inexpensive and small package, portablediagnostic kits could be made available (see box2-B). The industrial food processing industry isexpected to make use of portable biological sen-sors to determine the freshness of food. Becausebiosensors are small enough to be placed on thetip of a hypodermic needle, blood chemistrycould be monitored continuously by placing sev-eral biosensors onto a chip inserted on a cathetertip into the patient during surgery. The capabilityto measure chemicals concentrations will be use-ful for process industries such as biotechnologyand chemical production.

Miniaturization of biosensors and chemicalsensors relies on a device called a chemical fieldeffect transistor (ChemFET).21 ChemFETs aresimilar to the field effect transistors (FET) thatare used in normal microelectronics. A micro-electronics FET conducts electricity (turns on)when a voltage is applied to one of it inputs (gate),creating an electric field in the FET. A ChemFET

ITu.s. Depafiment of Commerce, ‘TI’E(X-I (Japanese Technology Evaluation Program) Panel Report on Advanced Sensors in Jwm” JanU-ary 1989, p. 40.

181bid., p. 44.lgJerome S. Schulz, “Biosenmrs,” Scientific 4UriC~, August 1991, pp. 64-69.z@Strong Gro~h for Biosensors,” New Technolo~ Week, Feb. 19, 1991, P. 8.21~w knom as ion ~nsitjve fje]d effect transistor (ISF~. “CHEMF~” js usually used to refer to sensors that use whole mOleCUleS, while

“lSFET” refers to sensors designed with ions (charged atoms or molecules).

Chapter 2–Technology 29

Box 2-B--Bedside Analysis1

A hand-held analyzer currently un-der evaluation at the Hospital of theUniversity of Pennsylvania demon-strates how biosensors might findtheir way into clinical use. It simulta-neously makes six commonly re-quested chemical measurements on apatient’s blood-sodium, potassium,chloride, urea nitrogen, glucose, andhematocrit-producing results in lessthan 2 minutes. The bedside tests costmore than ones performed in a cen-tral laboratory, but their immediacymay make them more effective.

The device achieves accuracy com-parable to that of laboratory equip-ment by using a disposable cartridgecontaining six biosensors and a cali-bration sample. A medical workerplaces 60 microliters of blood in thecartridge; the analyzer then measuresboth the calibration sample and thepatient sample. It displays test resultsand also stores them, keyed to timeand the patient’s identification num-ber, for later analysis: The cartridge-based design adopted by manufactur-ers will make it possible to perform adifferent set of tests once the appro-priate sensors have been developed.

from Schultz, 1991, pp. 64-69.

Figure 2-B-1 -Bedside Analysis

Display

Sample

Calibrationsolution

Cartridge

Sample inlet

Sensors

SOURCE: Adapted from Jerome S. Schultz, Scientific American, 1991, pp. 64-69. Copyright (c) 1977 American, Inc. All

rights reserved.

uses a FET that is specially coated with a chemi-testosterone, the male hormone, to produce acal or molecules that will create an electric fieldwhen the sensor is exposed to a specific chemical,gas, or molecule.

Fiber-optics technology aids biosensor andchemical sensor development and miniaturiza-tion. Certain chemical and biological moleculeswill change their optical properties (either give offor absorb light differently) when exposed to otherchemicals or molecules. For example, the en-zymes dehydrogenase and luciferase react with

secondary chemical that gives off light (fluo-resces). A fiber-optic-based glucose sensor mightone day serve as the basis of an implantable artifi-

22 In the meantime, glucose sensorscial pancreas.are being incorporated into the latest generationblood sugar monitors for diabetics. These devicesare portable, take measurements in seconds com-pared to hours with earlier monitors, and somefunction without drawing blood by measuringabsorption of light through skin.23

footnote 18” 6, 1991, p. 4.


New generations of biosensors and chemicalsensors will take advantage of theircommon heri-tage within formation technology to integrate log-ic functions into the same package with biosen-sors. Integrating logic with sensors can result inmore useful devices. For example, a glucose sen-sor with integrated logic might respond different-ly to blood sugar levels at different times of day.Integrated circuit fabrication holds promise forhigh integration by placing biosensors and elec-tronics on the same chip. It also promises reduc-tion in prices because large numbers of sensorscan be fabricated at one time. Major challengesremain, however, such as isolating the microelec-tronics from the environment while allowing thesensor access to the environment.

Researchers are actively pursuing refinementsto the fabrication processes that will make wide-spread use of biosensors possible. The yield anduniformity of the manufacturing process needs tobe improved before large scale production ispractical. Shelf life is a problem with biosensors;compared to microelectronics and fiber optics,biosensors are perishable. Biomolecules thatmust be affixed to a chip or optical fiber are notstable outside their cellular environments. As aresult, biosensors can only be stored for weeks ormonths—sometimes years. This limitation pro-hibits some applications, but for some industriesshort lifetime products are normal and expected.In medicine, for example, instruments are oftendiscarded after use to avoid risk of contamina-tion. The price and capabilities of biosensors inother applications (e.g., detecting explosive gasesor critical chemicals in a process) are valuableenough that users are willing to accommodatelimited lifetimes.

Micromachined Silicon Sensors

A significant field for future sensor develop-ments will be mechanical sensors—devices thatrely on the mechanical properties of a material tosense energy of their surroundings. Since 1958,silicon has been used in pressure sensors. By

fabricating a covered cavity, changes in ambientconditions can be detected by monitoring theresistance across resistors on a membrane cover-ing the cavity (figure 2-8). Improvements in sili-con processing technology have reduced the sizeand cost of silicon pressure sensors, with majorgains made during the 1980s. Cost has gone fromabout $1,000 per sensor in the 1960s to a fewdollars per sensor today. The 1958 silicon pres-sure sensor that was half an inch wide is now onehundredth of an inch wide. Growth in the 1980swas very strong; silicon pressure sensors are 60percent of the pressure sensor market—up from40 percent in 1985 and 16 percent in 1980.24

Silicon mechanical sensors are also used todetect acceleration. Accelerometers are madewith a mechanical silicon structure that places amass of material—the “proof mass’’—on the endof a thin “arm” of silicon. Acceleration makes the

Figure 2-8-Cut-away View of a SiliconPressure Sensor

Top entrypressure

Implantedstrain gages

Silicon frame

Electrostaticglass-siliconbond

Silicon diaphragm

Glass

Bottomentry

pressure

SOURCE: Adapted from et al., Sensorsand Microstructure (Fremont, CA: NovaSensor),1990, p. 7.3.

CA: NovaSensor, June p.


silicon arm bend; the degree of bending deter-mines the acceleration. Accelerometers are find-ing applications in automobiles for airbag de-ployment and automatic suspension systems.Analog Devices, Inc., for example, is using a sur-face-micromachined cantilever as the basis for aaccelerometer. The separation between the canti-lever and the chip surface changes with accelera-tion, changing the voltage. The voltage changecan be translated into an acceleration measure-ment.

The miniaturization of silicon sensors resultedfrom developments in micromachining–a tech-nique to precisely shape the surface of a materialsuch as silicon. Silicon in its crystalline form hasdifferent orientations in the crystal that can bespecified for a particular silicon wafer. Chemicalsused to etch silicon “eat away” crystalline siliconat different rates depending on orientation of thecrystal to the surface. This process is called pref-erential etching. By combining preferential etch-ing with special chemicals that stop short theetching process,complex structures can becreated in silicon. This technique is known as“bulk micromachining,” and has been used sincethe 1960s.

Use of “sacrificial layers’’—technique devel-oped during the 1980s--can create more intricatestructures. Sacrificial layers are thin films (usual-ly less than 10 microns) that are removed by etch-ing chemicals to release movable parts from thesubstrate (see box 2-C). The process of usingmultiple layers of sacrificial layers to create com-plex structures is called surface micromachining.Using surface micromachining, researchers atthe University of California at Berkeley and atAT&T Bell Labs fabricated a working motorabout the width of a hair (100 microns) in 1988(see photograph). That achievement set off aflurry of research activities. Laboratories aroundthe world were soon duplicating and improvingupon the original work. Work at the Massachu-setts Institute of Technology (MIT) improved thesacrificial layer technique, making it more com-patible with traditional silicon electronics pro-cessing. The Karlsruhe Nuclear Research Centerin Germany developed a variation on the surface-

micromachining technique known as “LIGA.” Ituses x-ray lithography technologies to fabricatemicrocomponents from other materials includingplastic and metal.

The new complex structures have created aneed for better understanding and characteriza-tion of the materials used in the structures. Sili-con is a well known material, but not enough isunderstood of its mechanical properties. Struc-tures that are released from the substrate, forexample, can warp due to strains in the material.In microelectronics these strains are not impor-tant because structures remain connected to thesubstrate, but for surface micromachining, thesestrains can ruin a component. Wear and frictionare not well understood at such small scales, andresearchers need to understand these propertiesto improve the reliability of micromotors andother movable parts. Characteristics of surfaceinteractions, flow of fluids, and air flow are notwell understood at such small scales. Under-

Photo credit: Analog Devices, Inc.

An acceleration sensor (accelerometer) made by Analog De-vices, Inc. will be the first commercial product that uses surfacemicromachining, a new processing technique that uses sacrificialmaterial Iayers to free structures from a substrate. The microstruc-ture, similar to ones fabricated at the Berkeley Sensor and Actua-tor Center, can be seen in the center of this microphotograph ofthe sensor chip. The sensor is surrounded by electronics thatprovide calibration and signal conditioning.


Box 2-C--A New Way To Machine

l. The pattern of the gear is transferred to thesubstrate by shining ultra-violet or x-ray lightthrough a stencil-like mask.

2. This sequence is repeated several times toachieve a structure that has alternating layers ofsilicon and “sacrificial layers." (See the descrip-tion of fabrication technologies in this chapterand app. A for details.)

3. The sacrificial layers can be dissolved in achemical that doesn’t disturb the silicon. Afterthe sacrificial layer is removed, the gear is freeto rotate. A restraining hub prevents it from fly-ing off the surface of the chip.

Figure 2-C-1 -Surface Micromachining

SOURCE: Sharon , “Welcome to Newsweek, vol.117, No. r. 15, 1991, pp. 60-61. Jared

pyright (c) 1991, k, Inc. Allrights

standing fluid flow at such small scales will beimportant if the technology is used for chemicaland biological processing applications.

Future progress in micromachined sensors arein three areas of improvement:

1.

2.

3.

design and fabrication processes will im-prove performance and expand applicabili-ty;manufacturing technology will becomemore important as sensors become com-modities; and

packaging of the sensor is the third chal-lenge, since reduction in packaging costshas a significant impact on cost of the finalproduct.

Manufacturing technologies and procedureshave received insufficient attention in the past.25

This could hold particular peril for U.S. industryif it is not remedied as it competes with Japanesecompanies attuned to the value of robust man-ufacturing methods.

ACTUATORS

Actuators translate a signal to motion or force.A solenoid valve is a common type of actuatorthat relies on electromagnetic forces to move aplunger arm to open and close a valve. Electro-static actuators are promising small-scale actua-tors because electrostatic fields can easily becreated in micro-structures and are relativelypowerful at small scales. Recent developmentsindicate electromagnetic actuators are promisingas well. Small actuators have also been madefrom piezo-electric material26 and shape memoryalloys.27

Micro-Electro-Mechanical Actuators

Surface micromachining techniques are usedto fabricate micro-structures with moving parts.

p. 2.18. contract or expand when a voltage applied.

change shape with a change in

Chapter 2–Technology .33

Researchers at the University of California atBerkeley and AT&T Bell Labs fabricated motorsin 1988 using these new techniques. These motorsare typically about 100 microns in diameter (thewidth of a hair) and about 3 microns deep. Mi-nuscule motors have also been fabricated at Uni-versity of Utah and AT&T Bell Labs using moretraditional machining techniques that are assmall as 500 microns in diameter. These motorsproduce more power than the smaller motorsproduced by surface micromachining and mightbe useful in applications such as microsurgery ordrug delivery. 29 Small actuators could also beused in consumer items, (e.g., cameras, cassetteplayers, video recorders, and toys) robotics anddefense applications.

Tiny levers, gears, and other mechanical as-semblies can be etched from silicon with the samesurface micromachining techniques used to makemotors. Gears, levers, and the like might be usefulfor transmitting motion and force. At such smallscales, these mechanisms might be useful forma-nipulating light, low-mass objects. There aremany potential applications in optics; several ma-terials that can transmit, emit, or reflect light canalso be used in the construction of microactua-tors. Tiny mirrors on movable levers or gearscould serve as optical switches or light modula-tors by moving from one position to another.Moving chemicals, fluids, and cells is currentlythe object of much research. Electric fields andultrasonic waves are demonstrated techniquesthat are being refined through research and de-velopment at university and industrial laborato-ries. Hewlett-Packard and Canon use bulk-mi-cromachined devices to dispense ink for theirthermal ink jet printers.30

Integrating Actuator, Sensors,and Electronics

One of the most exciting aspects of making tinystructures in silicon and other materials is the

prospect of combining sensors, actuators, andelectronics into integrated systems on a singlechip. The greatest prospect for near term applica-tions is probably science and engineering instru-mentation.

Chemical and biological applications also ap-pear promising: small tweezers might be useful tohold specimens (e.g., cells) in place while they aremanipulated by other devices, such as injectionneedles. Drug delivery systems might be madesmall enough to be worn by patients, or eveninserted into the body. Chemical processing thatoften uses complex lab equipment might be per-formed on a chip or wafer in a portable system.Biotechnology procedures normally performed ina batch process could be done on a cell-by-cellbasis with greater control over the results.

At Lawrence Livermore Labs in California,researchers are working on spectrometers thatcould be used to monitor environmental hazard-ous materials. Lawrence Livermore is also devel-oping a DNA sequencer on a chip. At least onemajor U.S. company is developing a gas chroma-tography system on a chip. Hitachi announced in1989 that it had a prototype cell fusion system31

that fused as many as 60 percent of cells com-pared with about 2 percent using conventionaltechniques.

Looking even further into the future, “microro-bots” might be fashioned from micromechanicalcomponents. These tiny robots could perform avariety of functions. One of the objectives of theMITI micromechanics program is to develop mi-crorobots that would be capable of inspectinginaccessible or hazardous locations such as in jetengines, nuclear plants, and the human body.Here in the United States, similar research isbeing conducted at NASA's Jet Propulsion Labo-ratory and a few universities. The objective of theNASA program is to make microrobots and mi-crorovers useful for exploration of other planetsand the moon.

281~~chniqua include ~~mion, diamond-point machining, and EDM (electro-di=harge rnaChin@”~small moto~ of this @ arc already being used for drug delive~ for livestock.sophillip Wrth, Hewlett Packard, personal communication, May 29, 1991.31cell fusion is a technique Uwd t. create ~1~ ~Pble of producing large quantities of monoclinal antibodies.

Appendix A

Fabrication Technology for Miniaturization

Many ofadvances in

INTRODUCTION

the technologies that have enabledminiaturization were first developed

for microelectronics and allow both lateral andthickness control in the creation of structures.Although the early techniques and tools weredirected at silicon, recent years have seen in-creased attention to materials other than silicon:compound semiconductors, superconductors,metals, and insulators. Furthermore, they are be-ing applied to more diverse areas: micromechani-cal structures, biosensors, and chemical sensors.

The same processes that have allowed for de-creases in size also allow parallel production ofmany devices. Thousands or millions of transis-tors or other devices can refashioned simulta-neously on one chip the size of a thumbnail, andmany chips preprocessed at the same time on onewafer. The key steps in the process of creatingthese structures are:

● lithography,

● pattern transfer, and

● characterization.

LITHOGRAPHY

Integrated circuit fabrication techniques allrely on lithography, an ancient technique firstused for artistic endeavors. The basic processinvolves covering an object with a thin layer ofmaterial (ancient artisans sometimes used wax)that can be patterned but will resist subsequentprocessing and protect the material underneath.Industrial lithographers use a hydrocarbon poly-mer, appropriately called “resist.” A pattern isproduced in the resist—usually by exposure to

visible light or ultraviolet (UV) radiation —expos-ing selected areas of the material below. The ex-posed areas can then be modified in some way.Successive applications of this basic process pro-duce a multilevel structure with millions of indi-vidual transistors in a square centimeter as de-scribed in figure A-1.

Photolithography is the most widely used formof lithography and is likely to continue to be so forthe near future. UV light, usually with a wave-length less than 400 nanometers, is used to createpatterns resist layers on flat surfaces. Diffractionof light—the interference of light waves with oneanother—limits the resolution or minimum re-producible feature size. Diffraction is minimizedby reducing the wavelength of the exposing radi-ation. Mercury vapor lamps are routinely used asa source of 436 nanometer (G-line) and 365 nano-meter (I-line) light. Shorter wavelength excimerlaser sources are a new source of UV light withoperation possible at 248 nanometers and in thefuture 193 nanometers and other wavelengths. Bycombining shorter wavelengths with improvedmask technology (including phase shiftedmasks2), better optics, and more sophisticatedresist chemistries, it is likely that photolithogra-phy will be usable to 0.25-microns minimum fea-ture sizes and possibly below 0.2 microns.

The longevity of UV lithography is a subject ofdebate in the semiconductor industry. Many ex-perts argue the limits of diffraction and depth offocus will prevent use of UV photolithographybelow 0.2 microns and that x-ray lithography,which uses substantially shorter wavelength radi-ation, will be the successor to UV lithography.X-ray lithography is a candidate for high volumeproduction of integrated circuits with line widthsbelow 0.5 microns. The most studied and devel-oped form of x-ray use is “proximity printing,”

IH.G. Craighead and M. Skvarla, “Micro and Nanofabrication lkchnology, Applications & Impact,” contractor repofi prepared for the OffIceof lkchnology Assessment, April 1990; and Tlm Studt, “Thin Films Get Thinner as Research Heats Up,’’ R&D Magazine, March 1990, pp.70-80.

Zphase shift masks function by Careful& controlling light diffraction, using the constructive and destructive interference to help create thecircuit pattern. Phase shift masks hold greatest promise for manufacturing memory and other ICS with regular, repeated patterns.

-35-


Figure A-1 -Fabrication Sequence for a Metal-Oxide-Semiconductor (MOS) Circuit

Silicon nitride Silicon dioxide

1

First polysilicon

p-Type silicon

Insulatingoxide film

Contact “window”

Aluminum

Fabricating this MOS circuit element (a two-level n-channel-negative charge carrier- polysilicon-gate metal-oxide-semiconductor) re-quires six masking steps. The first few process steps involve the selective oxidation of silicon with the aid of a film of silicon nitride, which actsas the oxidation mask. A thin film of silicon dioxide is grown over the entire wafer, and a Iayer of silicon nitride is deposited from a chemical (l).The layer is selectively removed in a conventional photolithographic step in accordance with the pattern on the first mask (2). A p-typedopant (e.g., boron) is implanted using the silicon nitride as a mask, followed by an oxidation step, resulting in a thick layer of silicon dioxidein the unmasked-(3). The silicon nitride is then removed in selective etchant that does not attack either the silicon or silicon dioxide (4).Since silicon is consumed in the oxidation process, the thick oxide layer is partly recessed” into the silicon substrate. The first layer of poly-crystalline silicon is then deposited and patterned in the second masking step (5). A second insulating film of oxide is grown or deposited,followed by the deposition of the second polysilicon layer, which is in turn patterned in the third masking step (6). A short etch in the hydro-fluoric acid at this stage exposes certain regions to an implantation or diffusion of n-type dopant. A thin layer of silicon dioxide is depositednext, and contact “windows” are opened with the fourth mask (7). Finally a layer of aluminum is deposited and patterned in the fifth maskingoperation (8). The wafer will also receive a protective overcoating of silicon dioxide or silicon nitride (not shown); the fact that openings mustbe provided in this overcoating at the bonding pads accounts for 6 masking steps. Vertical dimensions are exaggerated for clarity.SOURCE: William G. ‘The Fabrication of Microelectronic Circuits,” (San Francisco, CA: Freeman & Co.,

1977), p. 48. Copyright (c) 1977 by Scientific American, Inc.–George Kelvin.

AppendixA-Fabrication Technology for Miniaturization .37

where the mask is placed on the substrate materi-al and its pattern is reproduced by exposure tox-rays, resulting in a substrate pattern the samesize as the mask. Practical problems that must beaddressed include complexity of mask technolo-gies, difficulty in alignment of multiple layers,economical x-ray sources, and resolution limita-tions. X-ray lithography operates similarly to UVphotolithography, but with much shorter wave-lengths–around 5 nanometers. Proximity print-ing requires a bright source of x-rays. Synchro-tron sources are the highest intensity sources ofx-rays but they are very expensive. Most expertsfeel it is unlikely that proximity printing withx-rays will be practical below 0.2 microns, whileothers hold that the technology will be usable to0.1 microns.

More recent exploratory work in x-rays in-volves “reduction projection lithography.” X-raysare focused through a mask that is kept awayfrom the substrate. The principal advantages ofthis approach are cheaper x-ray sources (a syn-chrotron is not required) and improved resolu-tion. In this country, AT&T Bell Laboratoriesand other laboratories are investigating this tech-nology. One of the greatest challenges to makingprojection x-ray lithography useful for manufac-turing is the optics used in the process; they arecomplex, involving the creation of multiple-layerfilms with precisely controlled thickness.

Another way to get beyond the diffraction limitof radiation-based lithography is to use electronsor atoms to expose the resist. Diffraction does notlimit the resolution of electron-beam lithographybecause the quantum mechanical wavelengths ofhigh energy electrons are exceedingly small. Elec-trons scatter quickly in solids, limiting practicalresolution to dimensions greater than 10 nanome-ters—significantly greater than current demandsof any practical technology. Electron-beam lith-ography has demonstrated resolution as small as2 nanometers (0.02 microns) in a few materials.Electron beam technology, however, is limited inusefulness because an electron beam must bescanned across the entire wafer. Electron-beamlithography tools are in use in universities, in

laboratories, and in industry for mask-makingand small manufacturing production runs.

There are approaches still in research that mayyield a more versatile lithography tool. A possiblealternative is to use photolithography for largefeature definition and reserve the scanned elec-tron beam for the critical dimensions. A varietyofapproaches for parallel exposure, rather thanserial scanned exposure, of electron-beams havebeen studied for years. There is currently activeresearch in electron beams reduction projectionusing masks at AT&T Bell Laboratories, and inmultiple source systems and proximity printing atIBM. These approaches would exploit the resolu-tion and alignment possibilities of electronbeams with the speed of parallel exposure tech-niques.

Ion beam lithography is in many ways similarto electron beam lithography with beams ofcharged atoms (ions) taking the place of elec-trons. Recent advances in ion sources have in-creased the utility of scanned focused ion beams.Compared to photons (x-rays and light) or elec-trons, ions chemically react with the substrate,allowing a greater variety of modifications. Somescattering effects are also reduced compared toelectrons and, as with electrons, diffraction limitsof ion beams are negligible. The fundamentalresolution limits for charged particle lithogra-phies are below 1(M) nanometers. Ion beam lithog-raphy suffers the same draw back as electronbeam systems; it relies on a serially scannedbeam. That limits its applications to productsthat do not require high throughput—e.g., masksand small batches of electronics. Research effortsare underway to develop parallel projection andmultiple source ion beam lithographic equipmentthat might overcome the limitations of seriallyscanned ion beam lithography.

Scanned probe modification of surfaces, usingrecently developed scanning tunneling microsco-py (STM) methods, is being done by researchersin many research laboratories. Since STMs canmanipulate individual atoms, such approacheshave theoretical resolution limits of single atoms.This approach is still in early phases of researchand is focusing on basic physics measurements

38• Miniaturization Technologies

and better understanding of the behavior of sur-faces. The lack of stability of STM tips and slowspeed, however, make it far from a practical tech-nology for manufacturing in the foreseeable fu-ture.

PATTERN TRANSFER

The various lithography systems produce reliefpatterns in resists, and a subsequent patterntransfer step is necessary to fabricate the struc-ture. There are a number of options available andeach involves some engineering tradeoffs. Oncethe resist pattern is in place, and selected areas ofthe underlayer are exposed, there are three possi-ble next steps: 1) add to, 2) remove, or 3) modifythe exposed areas.

Adding to the exposed area involves either dep-osition or growth of a material on the exposedareas. Growth is different from deposition be-cause it involves consumption of the surface (usu-ally combining with some introduced chemical)to create the new substance. The most commonexample is a silicon surface being consumed inthe process of forming silicon dioxide. It involvesheating the silicon wafer in the presence of oxy-gen, causing the growth of silicon dioxide fromthe exposed silicon. A number of materials aregrown in this way, since the consumption of thesurface results in excellent adhesion, better elec-trical and mechanical properties, and improve-ments in other materials properties. This specificreaction has been exhaustively investigated andperfected and now can be used to grow oxidefilms whose thickness, and even lateral dimen-sions, can be measured in atomic layers (a fewangstroms).

Deposition creates new layers of material onthe exposed area. These techniques usually in-volve evaporating or sputtering a piece of mater-ial-the target-so its atoms or molecules fly offand land on the sample. Deposition techniquesinclude:

● spin-on,

● thermal evaporators,

● sputtering,

. laser ablation deposition,

● chemical vapor deposition; and

● molecular beam epitaxy such as organo-metallic chemical vapor deposition.

Spin-on deposition is the simplest depositiontechnique. It involves spinning the sample while aliquid is poured onto it; the spinning action dis-tributes the liquid evenly. Subsequent heatingbakes the liquid into a solid thin film.

Thermal evaporators use a heated filament oran electron beam to vaporize material. The mate-rial passes through a vacuum to impinge on thesample, building up a film. Evaporators emit ma-terial from a point source, resulting in “shadow-ing” and sometimes causing problems with verysmall structures.

Sputter deposition systems erode atoms from abroad target that then travel to the sample. Thisresults in better line width control and uniformi-ty, especially with high aspect ratio (height/width)structures. Refinements of these processes, e.g.,in situ cleaning with an ion gun, or using electroncyclotron resonance to confine and densify aplasma, result in better adhesion, higher qualityfilms, and more versatility. Using lasers or par-ticle beams in conjunction with deposition pre-sents the possibility of defining patterns on thesample in a single step.

Laser ablation deposition uses intense laserradiation to erode a target and deposit the mate-rial onto a substrate. This technique is particular-ly useful in dealing with compounds of differentelements —e.g., yttrium-barium-copper oxide su-perconductor films.

Chemical vapor deposition (CVD) depositsthin films by passing reactive gases over the sam-ple. The substrate is heated to accelerate deposi-tion. CVD is used extensively in the semiconduc-tor industry and has played an important role inpast transistor miniaturization by making it pos-sible to deposit very thin films of silicon. MostCVD is performed in vacuum, but new tech-niques allow operation without vacuums. Radiofrequency (RF) or photon radiation can be used

Appendix A--Fabrication Technologyfor Miniaturization .39

to enhance the process and is known as plasmaenhanced CVD (PECVD).

Some deposition techniques are so precise thatmaterial can be built up literally atom by atomwith the crystal structure of the new materialexactly matching that of the underlying layer. Inmolecular beam epitaxy (MBE), the sample isplaced in an ultra high vacuum in the path ofstreams of atoms from heated cells that containtargets of various types. These atomic streamsimpinge on the surface, creating layers whosestructure is controlled by the crystal structure ofthe surface, the thermodynamics of the constitu-ents and the sample temperature. Organo-metal-lic chemical vapor deposition (OMVPE, or some-times MOCVD) relies on the flow of gases(hydrides like arsine and phospine or organome-tallics like tri methyl gallium and tri methyl alu-minum) past samples placed in the stream.Again, the sample surface and thermodynamicsof the processes determine the compounds de-posited. Both MBE and OMVPE provide thick-ness control within one atomic layer (a few ang-stroms) and are especially useful in creatingcompound semiconductors and exploiting quan-tum effects and band-gap engineering. Thisstructural control enables researchers to exploit

optical transitions in some materials, producinglasers, detectors, and other optical elements.

Materials modification processes are usedmainly to vary the electrical conductivity in theappropriate areas. In the past, dopants (atomsthat can either contribute or subtract electronsfrom silicon atoms) were diffused into the sub-strate thermally; now they are implanted as highenergy ions (see figure A-2). Implantation offersthe advantage of being able to place any ion at anydepth in the sample, independent of the thermo-dynamics of diffusion and problems with solidvolubility and precipitation. Ion implantation,originally developed for high energy physics, isnow an indispensable part of semiconductormanufacturing. Ion beams produce crystal dam-age in addition to the chemical or electronic effectof the dopants. Since crystal damage reduceselectrical conductivity, this effect can be ex-ploited to electrically isolate devices from oneanother. Ion beams can implant enough materialto actually form new materials--e.g., oxides andnitrides-some of which show improved wearand strength characteristics.

Subtractive processes—the removal of materi-al-are also vital to the field. Some of the stepsare still done the same way they were centuries

Figure A-2–ion Implantation

Accelerated boron ions

p-Type - silicon Thick oxide

Polysilicon

/

gate

Ion implantation is employed to place a precisely controlled amount of dopant (in this ease boron ions) below the gate oxide of a MOStransistor. By choosing a suitable acceleration voltage the ions can be made to just penetrate the gate oxide but not the thicker oxide (left).After the boron ions are implanted polycrystalline silicon is deposited and patterned to form the gate regions of the transistor. A thin layer ofthe oxide is then removed and the source and drain regions of the transistor are formed by the diffusion n- of an n-type impurity (right).

SOURCE: William G. “The Fabrication of (San CA: Freeman Co.,1977), p. 50. Copyright (c) 1977 by -George V. Kelvin.


ago. The object, with its lithographically definedareas exposed and the remainder protected bythe “resist,” is dipped into a chemical bath and areaction or dissolution is allowed to take place.Unfortunately, liquid chemicals usually have nosense of direction and will dissolve at an equalrate on all sides. As device geometries shrink, itbecomes important to control directionality.Most modern etching is done in vacuum withoutcontact with liquids.

Ion milling uses the sputtering process to bom-bard the substrate with accelerated ions. Thesputtering process relies on a physical mecha-nism whereby energy is transferred to a surfaceby impinging ions. This energy is enough thatsome atoms break free of the surface. This mech-anism erodes all materials, with small differencesin rate between any two. Selectivity is poor, andany masking material (resist) will disappear asfast as the underlayer, limiting the (relief) depth.

Reactive ion etching (RIE) can be highly selec-tive because it is a chemical reaction between ions(radicals) formed in a plasma and atoms on thesample surface, and can be highly selective. RIEhas excellent lateral control because the electricfields produced by the plasma cause ions to dif-fuse directionally. Straight parallel trenches a mi-cron or more deep can be etched without affect-ing a rather fragile resist mask; minimum featuresizes of a few nanometers are possible. RIE ishighly selective in its etching: oxygen etches hy-drocarbons (resists), fluorine compounds areuseful in working with silicon systems, and manymaterials are susceptible to chlorine etch, espe-cially the compound semiconductors like galliumarsenide. A great deal of work is directed at dis-covering the etching characteristics of other com-pounds.

The combination of physical and chemicaletching is achieved in a process called chemicallyassisted ion beam etching (CAIBE). The com-bined effects of ion sputtering and chemical reac-tion with reactive ions results in a faster etch ratethan that due to either mechanism alone. CAIBEis of particular interest to research groups inves-

tigating new materials, since it can be exploited touncover ways to etch new and exotic materials.

Unlike ancient lithographs, modern semicon-ductors must undergo many processing steps.The mask must remain intact until the substrateetch is complete, without changing feature size orshape or surface characteristics (morphology).Multiple operations are often required because ofthe processing limitations of materials. For exam-ple, to grow a thin layer of silicon dioxide at1,000°C, a resist is used to pattern a silicon ni-tride layer which is then capable of surviving thehigh temperature during the oxide growth. Multi-ple layers are often processed in succession tobetter allow a straightedge to be carried througha thick stack of material.

CHARACTERIZATION

Characterization has become an indispensabletool in the art and science of fabrication. Thedesigner needs information about device per-formance, the experimenter needs to know com-position and concentration, and the process engi-neer must have immediate confirmation of theelectrical, mechanical, optical, or chemical prop-erties of a thin film. In fact characterization, orfailure analysis, is in some cases as difficult andextensive as the original experiment. The mostcommon questions addressed by characteriza-tion concern the shape and size of the structure.Optical and electron microscopes provide thisinformation over size ranges from hundreds ofmicrons to fractions of a nanometer (near-atomicdimensions).

Instruments also exist to measure a range ofmaterials properties, including the thickness, re-fractive index, and electrical resistivity. Whilethese techniques involve the measurement of thinfilm properties over large distances, microchar-acterization can determine composition or struc-ture on a size scale comparable to the mean freepath of an electron and some techniques have anultimate resolution on the order of a few atomicdiameters.

Appendix A--Fabrication Technology for Miniaturization .41

There are a wide range of techniques usuallyidentified by acronyms, and some of the moreuseful are listed here:

SCANNED BEAM TECHNIQUES

AES –

RBS –

SEM – Scanning electron microscope:creates image of a surface bymeasuring the reflection from abeam of electrons.

TEM – Transmission electron micro-scope: determines crystal struc-ture by measuring transmissionof electrons through a thin sam-ple; has high resolution.

Auger electron microscopy: iden-tifies constituents in the surfacelayer by measuring energy of elec-trons emitted due to “Auger tran-sitions.”

Rutherford backscattering spec-troscopy: identifies constituentsby measuring the spectrum ofscattered ions off a sample.

SIMS – Secondary ion mass spectrosco-py: analyzes composition of mate-rial removed by ion bombard-ment.

XRF – X-ray fluorescence: identifiescomposition from photon emis-sion due to x-rays.

EDS (EDX)— Energy dispersive x-ray spectros-copy: identifies composition fromdispersion of x-rays.

IMMA –

EELS –

ED –

LEED –

RHEED –

Ion microprobe mass analysis:high resolution secondary ionmass spectroscopy (SIMS).

Electron energy loss spectrosco-py: identifies composition bymeasuring energy loss of elec-trons during electron interaction.

Electron diffraction: measurescrystal structure by diffraction ofelectrons.

Low energy electron diffraction:measures surface structure bydiffraction of electrons.

Reflection high energy electrondiffraction: measures surfacestructure; particularly useful forepitaxial film growth.

SCANNED PROBE TECHNIQUES

STM – Scanning tunneling microscopy:creates image of surface usingelectron tunneling between thesurface and a sharp tip near thesurface; very high resolution.

AFM – Atomic force microscopy: createsimage of surface using repulsionof electron charge distributionsbetween surface and a tip; near-atomic resolution is possible.

SxM – Other developing scanned probemicrocopies: create images ofsurface characteristics for mag-netic force, electrochemical mea-surement, etc., using the principleof the STM.

Appendix B

Glossary

Actuator: Causes mechanical force or motion inresponse to an electrical signal.

Analog: Refers to electrical signals that can varycontinuously over a range, as compared todigital signals, which are restricted to a pairof nominally discrete values.

Angstrom: One tenth of a nanometer, 10 -10

meters.

Architecture: The overall logic structure of acomputer or computer-based system.

Capacitor: Passive circuit element that storeselectrical charge, creating a voltage differen-tial. Capacitors can be fabricated within inte-grated circuits, as well as in the form of dis-crete components.

Charge-Coupled Devices (CCD): A type of solid-state electronic device used as a sensor insome types of cameras.

Circuit Board: A card or board of insulatingmaterial on which components such as semi-conductor devices, capacitors, and switchesare installed.

Clock An electronic circuit, often an integratedcircuit, that produces high-frequency timingsignals. A common application is synchroni-zation of the operations performed by a com-puter or microprocessor-based system. Typi-cal clock rates in microprocessor circuits arein the megahertz range, 1 megahertz equaling106 cycles per second.

Compact Disk (CD): An optical storage mediumused for music and for computer data, amongothers.

Deposition: An operation by which a film isplaced on a surface.

Die Bonder: Assembly equipment that bonds theback side of an integrated circuit die to vari-ous materials. A die is a small piece of siliconwafer that contains the complete circuit beingmanufactured.

Doping: Adding to a semiconducting materialsmall amounts of other elements (dopants) tochange its electronic properties.

Dynamic Random Access Memory (DRAM): Themost common type of computer memory.DRAM architecture usually uses one transis-tor and a capacitor to represent a bit, which isa memory cell in a computer.

Electron Cyclotron Resonance (ECR): A technol-ogy that uses a high-frequency microwave en-ergy source to create a plasma in a confinedregion using a magnetic field for the purposeof etching and deposition.

Etching: A process in which chemicals are usedto remove previously defined portions of thesilicon oxide layer covering the wafer to ex-pose the silicon underneath.

Flat Panel Displays: A thin display screen thatuses any of a number of technologies, such asliquid crystal display. Flat panel displays areused in laptop computers in order to keep theoverall size and weight of the machine to aminimum.

Gallium Arsenide (GaAs): A compound semi-conductor with properties necessary for veryhigh-frequency microwave (analog) devicesand optoelectronics.

Gate: A simple electronic circuit that can imple-ment a specified logical operation. In essence,gates act like switches. Computer processingunits depend on large assemblies of gates, asdo integrated circuit memory chips.

Integrated Circuit (IC): Electronic circuits, in-cluding transistors, resistors, capacitors, and ‘their interconnections, fabricated on a singlesmall piece of semiconductor material (chip).Categories of ICs such as LSI and VLSI referto the level of integration, which denotes thenumber of transistors on a chip. ICs may bedigital (logic chips, memory chips, or micro-processors) or analog.

-43-


Ion Implantation: A process in which silicon isbombarded with high-voltage ions in order toimplant them in specific locations and pro-vide the appropriate electronic characteris-tics.

Liquid Crystal Display (LCD): A liquid crystaldisplay is a technique that uses a transistorfor each monochrome or each red, green andblue dot. It provides sharp contrast andspeeds screen refresh. LCD technology iscommonly used in digital watches and laptopcomputers.

Lithography A process in which the desired cir-cuit pattern is projected onto a photoresistcoating covering a silicon wafer. When devel-oped, portions of the resist can be selectivelyremoved with a solvent, exposing parts of thewafer for etching and diffusion.

Logic Chips: ICs that manipulate digital data.

LSI: Large-scale integration. Typically involvingbetween 2,000 and 64,000 transistors.

Mainframe Computer: A system, normally in-tended for general-purpose data processing,characterized by high performance and ver-satility. Mainframes have grown steadily incapability as smaller and less expensive ma-chines have progressed.

Mask: Stencil-like grid used in creating litho-graphic patterns on semiconductor chips.

Mass Storage: Refers to peripheral equipmentfor computer memory suitable for largeamounts of data or for archival storage. Typi-cal mass storage devices are disk and tapedrives.

Memory Chips: Devices for storing informationin the form of electronic signals.

Microcomputer: Refers to integrated circuitsthat contain a microprocessing unit plusmemory, as well as to computers designedaround microprocessors or single-chip mi-crocomputers.

Micrometer: One-millionth of a meter.

Micron: One-millionth of a meter.

Microprocessor: A computer central processingunit on a single chip.

Nanometer: One-billionth of a meter.

Optical Disks: Recording media including CDsthat store information in patterns of micro-scopic pits on the surface of the disk, whichcan then be detected by a solid state laser anddetector system and reproduced as sound,images, or data.

Optoelectronic Devices: Devices that convertlight signals to electronic signals and viceversa.

RAM (Random Access Memory): Most common-ly, an integrated circuit that stores data insuch form that it can be read, erased, andrewritten under the control of a computerprocessor. Any memory location in a RAMcan be addressed directly (random access) asopposed to sequentially or serially.

Resist, Photoresist: Chemicals used in litho-graphic processing of integrated circuitswhich, much like photographic emulsions,can be exposed by light, X-rays, or other radi-ation to form patterns.

Semiconductors: Materials, e.g., silicon, thathave four electrons in their outer electronshells and have electrical conductivities thatare lower than good conductors such as met-als but higher than insulators such as glass.Other semiconductors include germanium,gallium arsenide (GaAs), iridium phosphide(InP), mercury telluride(HgTe), cadmium tel-luride (CdTe), and alloys of these compoundsemiconductors.

Sensor: Converts a pressure, temperature, or oth-er physical parameter into an electrical sig-nal, often for use in a control system. A digitalspeedometer for an automobile transformsthe output of a sensor into a miles-per-hourreading, as does an airplane’s air speed indi-cator. In the case of the automobile speedom-eter, rotary motion is converted into an elec-trical signal, while an air speed indicatordepends on the pressure created by the mo-tion of the airplane.

Appendix B-Glossary .45

Silicon: One of the most common elements foundin nature and the basic material used to makethe majority of semiconductor wafers.

Stepper: A sophisticated piece of equipmentused to transfer an integrated circuit patternfrom a mask onto a wafer.

Substrate: A piece of material, typically a semi-conductor, on which layers of materials aredeposited and etched to fabricate a device ora circuit.

Transistor: An electronic device that can be usedto switch or amplify electronic signals.

VLSI: Very large-scale integration. Typically in-volving more than 64,000 transistors.

Wafer: A thin disk, cut from silicon or other semi-conductor material. The wafer is the basematerial on which integrated circuits are fab-ricated. It is typically 4 to 8 inches in diame-ter.

Yield: In the production of microelectronic de-vices, the fraction that survive all tests andinspection, function correctly, and can besold or incorporated into the manufacturer’sown end products. Production costs dependheavily on yields, which themselves dependon circuit design, fabrication equipment, andcontrol of the manufacturing process.

Index

Accelerometer, 1,30-31Analog Devices, Inc., 9,31Angstrom technology, 7Arthur D. Little, Inc., 28Assemblers, 20,21AT&T Bell Laboratories, 7,9, 11,31,33,37Automobiles, 3,4,9

Biosensors, 3,8,28-30Blood analysis, 29Bosch, 10Bulk micromachining, 31

California, 9Canon, 33Cellular automata, 16cellular phone, 5Characterization, 3,35,40-41Charge coupled device, 27-28Chemical field effect transistor, 28Chemical sensors, 3,8,28-30Chemical vapor deposition, 38Chemically assisted ion beam etching, 40chip size, 14,22,24Clark, Leland C., 28Clock cycle, 4, 15,24Communications, 5, 16Consumer electronics, 4,5, 16,26,27,28,33

Defense Advanced Research Project Agency, 7,9, 11Department of Defense, 7,8,9,28Department of Energy, 9Digital Equipment Corp., 7Display technology, 5,23Doping, 39Drexler, Eric, 21Dynamic random access memory, 1,4,5,13,15

Electron-beam lithography, 11,13, 17,37Etching, 39-40Eutaxic control, 20

Feynman machines, 20Feynman, Richard, 21Fiber-optics, 29Flat panel display, 5,23Fluid flow, 31,33Food processing, 28Fraunhofer Institute, 9, 10Fujitsu, 7

Gallium arsenide, 16, 17German Federal Ministry for Research and Technology, 9German Karlsruhe Nuclear Research Center, 9, 10,31Glucose sensor, 28,29-30

Hewlett-Packard, 33High electron mobility transistor, 16

High-performance computers, 15,26Hitachi, 33Hybrid wafer-scale integration, 24

Interconnection, 8, 13, 15, 17,21,23-27International Business Machines, 7, 10, 11,23,24,37Ion-beam lithography, 11, 17,37Ion implantation, 39Ion milling, 40

Japanese Basic Technologies for the Future Industries Program,7

Japanese Exploratory Research for Advanced Technology, 7Japanese Large Scale Program, 7Japanese Ministry of Agriculture, Forestry and Fisheries, 8Japanese Ministry of International Trade and Industry, 7,33Japanese Science and Technology Agency, 10Japanese Technology and Evaluation Center, 8Jet Propulsion Laboratory, 33

Kodak, 28

Laptop computer, 5,19Lasers, 16, 18,23,38Lawrence Livermore National Laboratory, 9,33LIGA, 31Lithography, 10-11, 13, 17-19,23,31,32,35-38Louisiana, 9, 11Louisiana Technical University, 9

Magnetic storage, 23Mainframe computer, 19Manufacturing, 3, 13,30,32Manufacturing costs, 1,3, 13Mass storage, 17,23Massachusetts Institute of Technology, 31Materials science, 3, 16, 19,31,4041Matsushita Central Research, 10Mechanical properties, 31Medical applications, 16,28-30,33Messerschmidt, 10Micro-Electromechanical Sensors and Actuators Institute, 10Microelectronics and Computing Technology Consortium, 8,24Micromachining, 30Micromechanics, 1,3,8,9,20,30-34Micromotors, 32-33Microprocessor, 4Microrobots, 10,33Military applications, 16Mitsubishi, 17Molecular beam epitaxy, 16,38,39Molecular electronics, 3,7, 16, 17Molecular machines, 20-21Motorola, 11Multi-chip modules and packages, 4,8,24,25

Nanobots, 20Nanotechnology, 20

- 4 7 –

48 . Miniaturization Technologies

National Aeronautics and Space Administration, 9,33National Institutes of Health, 9National Research Council, 8National Science Foundation, 7,9Netherlands, 10New York, 11Nippon Electric Co., 5, 10,24Nippon Telephone & Telegraph, 7, 10Noise, 15

Optical storage, 17,23Optics, 16,33,39Optoelectronics, 7,39Organo-metallic chemical vapor deposition, 39

Packaging, 4,5,23-27,32Phase-shift lithography, 10, 19,35Pressure sensor, 30

Quantum electronics, 3,7,15,16

Reactive ion milling, 40Refrigeration, 15, 19,26Ricoh, 10

Sacrificial layer, 31-32Sandia National Laboratory, 9scaling laws, 15Scanning tunneling microscope, 19,20,23,37,41Semiconductor manufacture“ g equipment, 1,3,4,9, 10-11,

17-19,30,31-32,35-41

Siemens, 10Space applications, 16Spin on, 38Sputtering, 38Surface micro-machining, 31-32Surface mount technologies, 4,6,8Surface science, 3, 19,31,38,40-41Switzerland, 10

Tektronix, 28Texas Instruments, 7,28Thermal evaporation, 38Toyota, 10Traditional machining, 33

U.S. Congress, 9, 11Ultraviolet lithography, 10, 17,31,35,37University of California at Berkeley, 31,33University of Delft, 10University of Pennsylvania, 29University of Tokyo, 10University of Twente, 10University of Utah, 33

von Neumann architecture, 16, 27

X-ray lithography, 10, 11, 13, 17,31,32,35,37

U S G O V E R N M E N T P R I N T I N G O F F I C E : 1991 0 - 297-927 : QL 3