imaging the future of microcomputer chips · to continually shrink the size of the basic elements...

November 1999

U.S. Department of Energy’s

Lawrence LivermoreNational Laboratory

Imaging the Future of Microcomputer Chips

Also in this issue:

• Microfluidic Sensors to Help Save Lives• A New Polymer Can Cleanse the Environment of Mercury• Transistors on Plastic Transform Flat-Panel Display

Imaging the Future of Microcomputer Chips

Lawrence Livermore has joined LawrenceBerkeley and Sandia national laboratories and aconsortium of leading U.S. semiconductormanufacturers to develop extreme ultravioletlithography (EUVL) for making the nextgeneration of high-capacity computermicrochips. The article beginning on p. 4 is aprogress report on the partnership’s efforts.Pictured on the cover is a key component ofEUVL’s success—the Ultra Clean Ion BeamSputter Deposition System. Developed atLivermore, it creates the precise, uniform,highly reflective, low-defect masks (or masterpatterns) used to “print” semiconductor circuitson next-generation silicon microchips.

About the Cover

About the Review

• •

Lawrence Livermore National Laboratory is operated by the University of California for theDepartment of Energy. At Livermore, we focus science and technology on assuring our nation’s security.We also apply that expertise to solve other important national problems in energy, bioscience, and theenvironment. Science & Technology Review is published 10 times a year to communicate, to a broadaudience, the Laboratory’s scientific and technological accomplishments in fulfilling its primary missions.The publication’s goal is to help readers understand these accomplishments and appreciate their value tothe individual citizen, the nation, and the world.

Please address any correspondence (including name and address changes) to S&TR, Mail Stop L-664,Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, California 94551, or telephone (925) 422-8961. Our electronic mail address is [email protected]. S&TR is available on the World WideWeb at www.llnl.gov/str/.

Prepared by LLNL under contractNo. W-7405-Eng-48

© 1999. The Regents of the University of California. All rights reserved. This document has been authored by theRegents of the University of California under Contract No. W-7405-Eng-48 with the U.S. Government. To requestpermission to use any material contained in this document, please submit your request in writing to the TechnicalInformation Department, Document Approval and Report Services, Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, California 94551, or to our electronic mail address [email protected].

This document was prepared as an account of work sponsored by an agency of the United States Government. Neitherthe United States Government nor the University of California nor any of their employees makes any warranty,expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness ofany information, apparatus, product, or process disclosed, or represents that its use would not infringe privately ownedrights. Reference herein to any specific commercial product, process, or service by trade name, trademark,manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring bythe United States Government or the University of California. The views and opinions of authors expressed herein donot necessarily state or reflect those of the United States Government or the University of California and shall not beused for advertising or product endorsement purposes.

SCIENTIFIC EDITOR

David Eimerl

MANAGING EDITOR

Sam Hunter

PUBLICATION EDITOR

Dean Wheatcraft

WRITERS

Arnie Heller, Ann Parker, Katie Walter, and Dean Wheatcraft

ART DIRECTOR AND DESIGNER

Kitty Tinsley

INTERNET DESIGNER

Kitty Tinsley

COMPOSITOR

Louisa Cardoza

PROOFREADER

Carolin Middleton

S&TR, a Director’s Office publication, is produced by the Technical InformationDepartment under the direction of the Office of Policy, Planning, and Special Studies.

S&TR is available on the World Wide Webat www.llnl.gov/str/.

2 The Laboratory in the News

3 Commentary by James GlazeInfinite Riches in a Little Space

Features4 Extreme Ultraviolet Lithography: Imaging the Future

As part of DOE’s Virtual National Laboratory, Lawrence Livermore is helping develop extreme ultraviolet lithography to make the next generation of microcomputer chips.

10 Handling Fluids in MicrosensorsA complete microsystem for sampling and detecting biological and chemical pathogens is still a few years away. But several unique components for handling fluids are already demonstratingremarkable results.

Research Highlights17 A Crowning Achievement for Removing Toxic Mercury20 Flat-Panel Displays Slim Down with Plastic

23 Patents and Awards

Abstracts

S&TR Staff November 1999

LawrenceLivermoreNationalLaboratory

Printed in the United States of America

Available fromNational Technical Information ServiceU.S. Department of Commerce5285 Port Royal RoadSpringfield, Virginia 22161

UCRL-52000-99-11Distribution Category UC-0November 1999

Page 4Page 20

Page 10

2 The Laboratory in the News

Adaptive optics center fundedIn late July, the National Science Foundation’s governing body,

the National Science Board, approved a proposal to establish aCenter for Adaptive Optics at the University of California atSanta Cruz, with Lawrence Livermore as an important partner.The center will coordinate the work of researchers across thecountry involved in the rapidly developing field of adaptive optics,which has major applications in astronomy, vision science, andhigh-power laser beams. (See S&TR, July/August 1999, pp. 12–19.)

The Center for Adaptive Optics, which begins operation inNovember 1999, is one of five science and technology centersapproved by the National Science Foundation this year. NSFprogram guidelines allow for financial commitments of up to$20 million over five years for each center, with an option torenew for an additional five years.

UC Santa Cruz’s 27 partner institutions in the center includeLawrence Livermore; the University of California at Berkeley,San Diego, Los Angeles, and Irvine; the University of Chicago;the California Institute of Technology; the University ofRochester; the University of Houston; Indiana University; and17 other national laboratory, industry, and international partners.

Claire Max, director of the Laboratory’s University RelationsProgram, said that Livermore is well positioned to play a bigrole in the collaboration. According to Max, “The Center forAdaptive Optics will provide the sustained effort needed to bringadaptive optics from promise to widespread use by astronomersand vision researchers.”Contact: Claire Max (925) 422-5442 ([email protected]).

Two DOE centers study CO2 storageAs part of its global climate change research program, the

Department of Energy has formed two centers to study the captureand long-term storage (sequestration) of atmospheric carbondioxide (CO2) in terrestrial and oceanic ecosystems. The ultimategoal of these centers is to make carbon sequestration a potentialcomponent of future international efforts to reduce CO2 buildup inthe atmosphere, which is believed to contribute to global warming.

The DOE Center for Research on Ocean Carbon Sequestration(DOCS) will focus on oceanic ecosystems. It is a collaborationof numerous academic and oceanographic institutions led byLawrence Livermore and Lawrence Berkeley nationallaboratories and will receive $3 million over three years.

Livermore’s Ken Caldeira and Berkeley’s Jim Bishop are thecenter’s codirectors. According to Caldeira, “The Lab and Berkeleybring complementary activities to the center. We have experiencein modeling the oceanic carbon cycle and in simulation, andBerkeley has experience in observation and monitoring.”

DOCS will research the feasibility, effectiveness, andenvironmental acceptability of ocean carbon sequestration.

Research will assess the environmental consequence ofpossibly increasing the amount of CO2 absorbed by theocean through CO2 injection into the deep ocean and CO2fertilization of ocean organisms.

The Center for Research on Enhancing CarbonSequestration in Terrestrial Ecosystems (CSITE) is also adiverse collaboration, led by Oak Ridge, Pacific Northwest,and Argonne national laboratories. It will receive $6 millionover three years.

CSITE will research ways to use plants, microbes, and soilmanagement practices to cause more carbon to be stored belowground without major sacrifices in aboveground yields. It willalso investigate lengthening the time carbon is sequestered inthe soil as a means of limiting atmospheric concentrations. Andit will study ways to measure, monitor, and verify sequestrationso that national inventories of greenhouse gas emissions canbe appropriately accounted for.Contact: Ken Caldeira (925) 423-4191 ([email protected]).

Lab part of Next-Generation InternetIn early August, the Department of Energy announced

appropriation of $15 million to finance 19 projects relatedto the emerging government-supported network called theNext-Generation Internet. The network will have thecapability of carrying massive amounts of electronic, video,and voice signals at the speed of light—that is, about athousand times faster than a standard Internet connection.

A localized version of the network already connectsagencies in the Los Angeles and San Francisco area. Theexpanded network is expected to link a select group ofagencies across the nation and around the globe.

Lawrence Livermore, Sandia/California and LawrenceBerkeley national laboratories and the Stanford LinearAccelerator Center are among the contributors to this NextGeneration Internet project.

Bill Lennon, a program leader in Next GenerationInternet research at Livermore, says that the increase infinancial support will allow researchers to create softwarethat can manage and secure data on the new network.“We have to customize the way that the data is sent, todo applications that have never been done before, . . . toidentify the holes—the things that people haven’t thoughtof,” said Lennon.

Livermore is the lead in one Next Generation Internetproject that will study weather change and predictability byconducting high-definition simulations of global weatherpatterns. The project will receive $3.6 million in supportper year for three years.Contact: Bill Lennon (925) 422-1091 ([email protected]).

S&TR November 1999

Lawrence Livermore National Laboratory

■ James Glaze is Executive Director of the Virtual National Laboratory.

MAGINE owning a desktop computer as powerful as today’ssupercomputers. A computer in which each memory chip

can hold tens of gigabytes of information. A computer that canrecognize speech in context, hold an intelligent conversation,and translate languages. Now, imagine driving a car that canfollow verbal commands, visualize the surrounding trafficenvironment, sound an alert upon encountering hazardousconditions, and provide high-resolution “heads-up” displaysof local roadways and intersections.

Such possibilities seem as unreal today as the idea of a“personal” computer seemed 30 years ago. To a large degree,today’s technological miracles—and tomorrow’s technicaldreams—are the result of the semiconductor industry’s effortsto continually shrink the size of the basic elements (transistors,capacitors, resistors, etc.) that make up a computer chip. Whenchip features get smaller, speed, reliability, functionality, andefficiency all improve, while costs decrease.

For the past 30 years, optical lithography has been theprocess by which these features are printed on semiconductorsubstrates to produce integrated circuits. This optical process,however, is rapidly approaching its limits. The internationalsemiconductor industry must decide—and soon—which next-generation lithographic technology it will embrace formanufacturing chips in the first two decades of the 21st century.

In 1997, three Department of Energy laboratories—Lawrence Livermore, Lawrence Berkeley, and Sandia—joined together in the Virtual National Laboratory (VNL) todevelop one such promising technology: extreme ultravioletlithography (EUVL). Each national laboratory contributesunique expertise to the program: Sandia brings systemdesign and process development, Lawrence Berkeley bringsprecision metrology and patterning, and LawrenceLivermore brings expertise in complex multilayer reflectivecoatings, optics design, and precision engineering. To fullydevelop and help commercialize EUVL, the VNL ispartnering with an industry consortium that includes some of the country’s leading semiconductor manufacturers.The article beginning on p. 4 gives an overview of theVNL–EUVL program and Lawrence Livermore’scontributions to it.

I

3Commentary by James Glaze

The VNL has been extremely successful to date. Forinstance, the three laboratories have garnered numerousR&D 100 awards as a direct result of technologies developedfor EUVL. Even more to the point, a year ago, the internationalsemiconductor community voted EUVL “the most promisingtechnology” for printing future generations of computer chipswith features as small as and smaller than 70 nanometers.(Current features are 180 nanometers.) The fact that EUVLtechnology is extendable is a major attribute.

Why all the interest in making something that’s already sosmall even smaller?

For the VNL, helping to power the information age throughthe development of ever-smaller chips that make computersmore compact yet more powerful furthers our nation’s economicobjectives. The VNL’s efforts also support the nationallaboratories’ missions in national security, proliferationprevention, energy, and environmental monitoring throughadvances in micromachining, sensor technology, precisionmeasurement, and supercomputing.

To answer the question in a larger context, we need onlylook around us. Much of the modern world and its economicwell-being are made possible by the computer chip. Accordingto Department of Commerce data from 1998, the semiconductorindustry is the United States’ largest manufacturing industry interms of value added, contributing 20 percent more to theeconomy than its nearest rival. As impressive as this figure is,it doesn’t begin to explain the role semiconductors have playedin enabling the revolution in today’s electronics and informationtechnologies.

The EUVL technology being developed by the VNL and itsindustry partners could take semiconductor manufacturing to theend of the “silicon cycle,” sometime in the second decade of the21st century. Then, it is predicted, the computer industry willhave fully exploited silicon as a substrate material. It will havecrammed as many features on a silicon chip as the material cansupport, and it will be time to dream again. Lawrence Livermoreand the VNL hope to again play a significant part in makingthose new dreams come true.

Infinite Riches in a Little Space


S&TR November 19994


Livermore researchers

are part of a collaboration

between private industry

and DOE laboratories

dedicated to making extreme

ultraviolet lithography

the technology of choice

for manufacturing the

next generation of

microcomputer chips.

Extreme Extreme Ultraviolet Lithography

Imaging the Future

WENTY-FIVE years ago, the computing equivalent of today’slaptop was a room full of computer hardware and a cartload of

punch cards.Since then, computers have become much more compact and

increasingly powerful largely because of lithography, a basicallyphotographic process that allows more and more features to be crammedonto a computer chip. Light is directed onto a mask—a sort of stencilof an integrated circuit pattern—and the image of that pattern is thenprojected onto a semiconductor wafer covered with light-sensitivephotoresist. Creating circuits with smaller and smaller features hasrequired using shorter and shorter wavelengths of light.

However, current lithography techniques have been pushed justabout as far as they can go. They use light in the deep ultraviolet range—at about 248-nanometer wavelengths—to print 150- to 120-nanometer-size features on a chip. (A nanometer is a billionth of a meter.) In thenext half dozen years, manufacturers plan to make chips with featuresmeasuring from 100 to 70 nanometers, using deep ultraviolet light of193- and 157-nanometer wavelengths. Beyond that point, smallerfeatures require wavelengths in the extreme ultraviolet (EUV) range.Light at these wavelengths is absorbed instead of transmitted byconventional lenses. The result: no light, no image, no circuit.

Semiconductor manufacturers are, therefore, at a critical juncture. Soon,they must decide which lithographic horse to back in the race to the next

T

S&TR November 1999 5EUVL Progress Report

What’s at Stake

. . . [T]he first 30 years of the integrated circuit had from two to fivetimes the impact on the U.S. economy as the first 30 years of therailroad. Or, to put it another way, the transformation of the nineteenthcentury U.S. economy by the railroad took 60 years to achieve half theeffect that microelectronics had over 30 years.

—Kenneth Flamm“More for Less: The Economic Impact of Semiconductors”

December 1997

Semiconductors are ubiquitous in our lives. They are found in our cars, televisions, radios,telephones, stereos, personal computers, children’s toys—even dishwashers and ovens.According to research conducted by the Semiconductor Industry Association and the WorldBank, each person worldwide uses on the average about ten million transistors in a lifetime,with this consumption increasing 55 percent each year. By the year 2008, projections arethat each person will consume a billion transistors.

It’s no wonder then that the annual sale of semiconductors worldwide will soon reachabout one trillion dollars—the equivalent of the gross national product of the UnitedStates. Furthermore, it’s no wonder that the next generation of lithographic technology—which is seen as the “gating technology” to the next generation of integrated circuits—is of such great interest to semiconductor manufacturers throughout the world.

generation of microchip manufacturing.There are currently four possiblealternatives: EUV, x-ray, electron-beam, and ion-beam lithography.

Creating a Virtual LaboratoryTwo years ago, three Department

of Energy national laboratories—Lawrence Livermore, LawrenceBerkeley, and Sandia/California—formed the Virtual National Laboratory(VNL) to research and develop extremeultraviolet lithography (EUVL)technology. The VNL is funded by theExtreme Ultraviolet Limited LiabilityCompany—a consortium of IntelCorporation, Motorola Corporation,Advanced Micro Devices Corporation,and Micron Technology, Incorporated—in one of the largest cooperative researchand development agreements withinthe Department of Energy. The three-year, $250-million venture is dedicatedto developing the EUVL technologyfor commercial manufacturing ofcomputer chips and to move thistechnology into production facilitiesin the first decade of the 21st century.

Each national laboratory bringsunique contributions to this effort.Lawrence Livermore supplies itsexpertise in optics, precisionengineering, and multilayer coatings.Sandia provides systems engineering,the photoactive polymer thin filmexposed by the light, and the light source.Berkeley contributes its AdvancedLight Source capability to generateEUV light to characterize optics andresists at the nanometer scale.

The VNL’s lithography systemuses mirrors to project the image of areflective mask onto the photoresist-coated semiconductor wafer. Ultimately,this system will enable a microchip to bemanufactured with etched circuit linessmaller than 100 nanometers in width,extendable to below 30 nanometers.

The resulting microprocessors would bea hundred times more powerful thanthose made today. Memory chips wouldbe able to store a thousand times moreinformation than at present.

“Lithography is generally viewed asthe enabling technology for each newgeneration of semiconductor devices,”

says Don Sweeney, LawrenceLivermore’s program manager forEUVL. “To put this technology intoproduction facilities in 10 years, we needto show that the technology can workunder real manufacturing conditions.”

The VNL’s current focus is onbuilding and integrating the necessary


S&TR November 19996


EUVL Progress Report

Using a prototype system,the Virtual NationalLaboratory has successfullyprinted lines as small as 50 nanometers (billionthsof a meter) wide inphotoresist. Currentlithographic tools used inthe semiconductor industryprint patterns with 180-nanometer-sizefeatures.

The optical layout of the engineering test stand for extreme ultraviolet (EUV) lithography. The EUV radiation is produced at the plasma source, transmitted through the condenser optics to the mask, reflected from the mask onto the four mirrors of the projection optics box, and delivered to the EUV-sensitive film on the semiconductor wafer. Each mirror in the system has 81 layers of reflective coatings that must be applied with extreme precision. At the short wavelengths used in the process, the total thickness of each mirror’scoatings must deviate less than an atom if the mask pattern is to be reflected without distortion. One such mirror is shown onp. 4.

Mask stage

Plasmasource

Wafer stage

Projectionoptics

Condenseroptics

Pho

to c

ourt

esy

of S

andi

a N

atio

nal L

abor

ator

ies/

Cal

iforn

ia.

technologies into an engineering teststand (ETS). Each national laboratoryspearheads specific development areasfor the ETS and for the systems beyond.Lawrence Livermore is leading theefforts to develop the optical systemsand components, thin films, masks,and submicrometer metrology requiredfor EUVL.

Brighter Light Is KeyThe ETS (see the box on p. 7)

includes a condenser optics box anda projection optics box. Both boxeshouse complex optical trains ofprecision concave and convexaspherical mirrors.

The main role of the condenseroptics box is to bring light to thereflective pattern on the mask. “Wewant to bring as much light to the maskand, ultimately, the wafer, as possible,”explains Sweeney. “The more light wedeliver, the shorter the exposure time.It’s like taking a picture with a camera.A picture taken in bright noonday sunrequires a shorter exposure time thandoes a picture of the same scene takenat twilight.”

For the semiconductor industry,brighter EUV images mean shorterexposure times, which translate tomanufacturing more chips at a faster rate.The optics design team from LawrenceLivermore and Sandia designed acondenser optics system that collectsand transports a significant fraction ofthe EUV light from the source to thereflective mask.

Once the image is reflected from themask, it travels through the projectionoptics system. According to Sweeney, theprojection optics box is the optical heartof the lithographic exposure system. “It isto the system what an engine is to a car,”he explains. The four mirrors of the ETSprojection optics system reduce theimage and form it onto the wafer.“Again, imagine using a pocket camera.The camera lens transmits an image tothe film, which—like the wafer—has alight-sensitive surface,” says Sweeney.

7


EUVL Progress ReportS&TR November 1999

The extreme ultraviolet lithographyprojection optics system in final assembly.

The Engineering Test Stand Provides a Prototype

The Virtual National Laboratory is developing, designing, and building a prototypeextreme ultraviolet lithography (EUVL) system called the engineering test stand (ETS) atSandia National Laboratories/California. The ETS uses laser-produced plasmas to supplythe extreme ultraviolet radiation needed. The radiation travels through a complexcondenser optics system before reflecting from a lithographic mask. That image is thenprojected by the projection optics onto a semiconductor wafer.

“The basic building blocks are the same as those found in systems operating at visiblewavelengths, except their forms are different, because of the short wavelength of EUV,”says Don Sweeney, Lawrence Livermore’s program leader for EUVL.

Because all materials, including nitrogen and oxygen, absorb EUV, the machine mustoperate in a vacuum and use reflective mirrors and masks. The ETS has six essentialsubsystems: a laser-produced plasma EUV source, condenser optics, projection optics, a mask, precision scanning stages, and a vacuum enclosure.

A conceptualdrawing of theextreme ultravioletengineering teststand. The goal ofthe ETS is todemonstrate howultravioletwavelengths canbe used to printpatterns onintegrated circuitsat production levelsand sizes.

The optics teams are now workingon advanced designs for the projectionoptics. They have a six-mirror design thatpromises to extend EUVL systems sothat they can print features as small as30 nanometers— a significant jump fromthe 70-nanometer limit of the ETS.

According to Sweeney, extendabilityto smaller features is an importantrequirement for whatever lithographictechnology the semiconductor industryfinally decides to back.

Applying Uniform Thin FilmsPart of the success of the EUVL

technology is due to the immense stridesLawrence Livermore has made inproducing the highly reflectivemultilayers that are used on the ETS’soptical mirrors as well as on the mask.

The projection and condenseroptical systems require mirrors thatreflect as much EUV light as possible.Manufacturing these mirrors has beena challenge because, in addition tobeing highly reflective, they must havesurface coatings that are essentiallyperfectly uniform.

Lawrence Livermore and LawrenceBerkeley developed advanced multilayercoatings of molybdenum and silicon thatcan reflect nearly 70 percent of the EUVlight at a wavelength of 13.4 nanometers.Applying these coatings evenly is adifficult task even when a mirror is flat,but EUVL mirrors are either convex orconcave. Any small nonuniformity inthe coatings destroys the shape of theoptics and results in distorted patternsprinted on the chips.

In the past year, the developmentof a new precision deposition systemprovided a major advance in applyingthese thin films to optics. (See S&TR,October 1999, p. 12.) This system,which won a 1999 R&D 100 Award,is so precise that 81 layers ofmolybdenum and silicon, each about3.5 nanometers thick, can be depositedover a 150-millimeter area so that thetotal thickness over the surface deviatesby less than an atom. The technique

can be used to coat mirrors as large as40 centimeters in diameter.

The Mask-Making ChallengeIndustry experts generally agree

that the biggest challenges and risksfor the next generation of lithographysystems involve the mask—that is, themaster pattern used to “print” thesemiconductor circuits onto the siliconwafers or chips. The technology thatsuccessfully overcomes the hurdles ofmask production has a good chance ofbecoming the preferred choice.

In EUVL, a mask is produced byapplying multilayers of molybdenumand silicon to a flat substrate. The circuitpattern is produced by applying a final

8


EUVL Progress Report S&TR November 1999

The Ultra Clean Ion BeamSputter Deposition System,developed at LawrenceLivermore, is used to produceprecise, uniform, highlyreflective masks. A keyrequirement of the next-generation lithography systemis that it produce virtuallydefect-free masks. The system contributes fewer than 0.1 defects per squarecentimeter to each mask. The ultimate goal for extremeultraviolet lithography is to addno more than 0.001 defects persquare centimeter to a finishedwafer blank.

EUV-absorbing metal layer and thenetching away the metal to form theimage of the circuit.

One key requirement is to produce amask with essentially no defects. Anysmall defect ends up being replicated, orprinted, in the lithography process ontothe computer chips being manufactured,thus damaging the chips’ complexcircuitry. A key breakthrough in thisarea was the development of an UltraClean Ion Beam Sputter DepositionSystem about two years ago. Thissystem—also an R&D 100 Awardwinner—produces precise, uniform,highly reflective masks with fewerdefects than those produced byconventional physical depositionprocesses. (See S&TR, October 1997,p. 8.) In April 1999, the team madesignificant improvements to the system’ssputtering shield design and otheroperational parameters. The systemnow consistently produces fewer than0.1 defects per square centimeter—a factor-of-8 improvement over defectdensities produced in 1998.

In fact, under the best operatingconditions, the system adds as few as0.04 median defects per square centimeterduring a coating run of 25 wafers. The

ultimate goal for the system is to addno more than 0.001 defects per squarecentimeter to the finished wafer blank.

The system has also been upgraded toprocess 200-millimeter wafers—the sizeused in industry—up from 150-millimeterwafers. The ability to process largerwafers for mask substrates means boththat the technology is working withindustry standards and that patterns forlarger chips can be placed on the wafer.

The Lawrence Livermore team hasalso conducted groundbreakingexperiments looking at the propagationof defects during multilayer film growth.All masks have defects of some kind—some more, some less. First of all, thereare defects that arrive on the waferfrom the manufacturer. These areanalogous to the pinholes and dust onefinds on photographic negatives. Defectssmaller than a certain critical size arecovered up by the film layers and presentno problem. However, defects largerthan this critical size persist through thecoating process and must be repaired orreckoned with in some way.

There are also defects created by thecoating process itself: a few atoms toomany in any one area can create a bumpthat will affect the final circuit pattern.

The question becomes, what is thiscritical size? “We’ve been modeling fora long time to see how different sizesand kinds of defects affect the finalproduct,” says Scott Burkhart, groupleader for mask blank development.“We finally conducted experimentsthat are setting the lower bound ofcritical defect size.”

The group has also made strides inrepair strategies for mask defects.“One mask can cost tens of thousandsof dollars,” notes Burkhart. “Whenpossible, repairing the defects savesthe industry a lot of money.”

Measuring at the Atomic LevelUntil recently, it was impossible to

accurately measure a mirror surfacefor high and low spots of a few atoms.An R&D 100 Award–winninginterferometer developed at theLaboratory two years ago—called thephase-shifting diffraction interferometer(PSDI)—changed all that. (See S&TR,October 1997, p. 6.)

Like all interferometers, the PSDIuses the interference pattern of twowaves of light to measure objects orphenomena. These light waves areusually imperfect because of the

9


EUVL Progress ReportS&TR November 1999

DONALD SWEENEY received his B.S., M.S., and Ph.D inmechanical engineering from the University of Michigan at AnnArbor in 1968, 1969, and 1972, respectively. He was a professorat Purdue University from 1972 to 1983, after which he became adepartment manager at Sandia National Laboratories/California andassumed responsibility for a research program in optical diagnostics.

Sweeney joined Lawrence Livermore’s AdvancedMicrotechnology Program (AMP) in early 1993 and soon became deputy programleader for Optics Technology. In February 1999, AMP reorganized and became theInformation Science and Technology Program, where he is currently deputy programleader for Extreme Ultraviolet Lithography and Advanced Optics.

About the Scientist

Schematic of the phase-shifting diffraction interferometer which uses two single-mode optical fibers.In this example, the interferometer is testing a lens, but the setup is similar for testing mirrors.

Fibercore

Single-modeoptical fiber

Single-modeoptical fiber

Semitransparentmetallic film

Sphericalmeasurement

wavefrontfrom fiber Optical

systemunder test

Sphericalreferencewavefrontfrom fiber

Aberrant wavefront

Interferencepattern

Aberrant wavefront

reflected from fiber face

imperfect condition of the surface orlens from which they emanate. Anyimperfection introduces error into themeasurements. The PSDI produces anearly perfect spherical wavefront usingdiffraction. In diffraction, light passesaround an object or through a hole,breaking up in the process. In the PSDI,two light beams pass through twoseparate optical fibers. When light exitsthe surface of each fiber, it diffracts,forming nearly perfect sphericalwavefronts. Because the two wavefrontsare generated independently, theirrelative amplitude and phase can becontrolled, providing contrast adjustmentand phase-shifting capability for thehighest possible accuracy.

The measurement wavefront passesthrough the optical system being tested,which induces aberrations in thewavefront and causes it to focus on theendface of the other fiber. Here, thewavefront reflects off a semitransparentmetallic film of the fiber end’s surfaceand interferes with the referencewavefront to generate an interferencepattern. The pattern is then recordedby a charge-coupled-device camera.

Over the past three years, manyEUV optics have been measured usingthis interferometer, including bothconcave and convex spherical andaspherical mirrors and completedprojection systems. The PSDI is now areliable production tool for measuringthe overall surface shape of thoseaspherical optics that have a specificationof 0.50 nanometers or less and hassuccessfully measured errors in thesurface shape down to 0.35 nanometers.The Livermore metrology team isupgrading the system so that it canbe used to measure errors in theoverall surface shape as small as0.15 nanometers.

EUV Pulling Ahead in the RaceLast December, the VNL’s work

paid off with a vote of confidence fromInternational Sematech, a privately

funded organization of semiconductormanufacturers.

At its annual meeting, InternationalSematech evaluated the four next-generation lithographic technologies—EUV, x-ray, electron-beam, and ion-beam—and strongly recommendedEUV lithography. “Theirrecommendation gave our effortsimportant momentum,” notes Sweeney.“It validated what we already knew:that we have a winning combination inthe three national laboratories and ourindustrial partners and that our strengthcomes from working together.”

—Ann Parker

Key Words: Extreme Ultraviolet LimitedLiability Company, extreme ultravioletlithography (EUVL), masks, phase-shifting diffraction interferometer(PSDI), precision deposition system,reflective multilayers, submicrometermetrology, thin films, Ultra Clean IonBeam Sputter Deposition System, VirtualNational Laboratory (VNL).

For further information contact Donald Sweeney (925) 422-5877([email protected]). Or visit theLawrence Livermore EUV lithographyWeb site at http://lasers.llnl.gov/lasers/IST/euvl.html.

S&TR November 199910


HandlingFluids inMicrosensorsWanted: Small, portable devices forautomatically detecting and identifyingviruses, bacteria, and toxic chemicals.

Reward: Satisfaction in knowing thatmicrosensors may help save lives.

ORK has been under way forseveral years at sites throughout

the departments of Energy and Defenseon autonomous devices for detectingbiological and chemical agents. The goalis to install them in subways, major officecomplexes, convention centers, or othersites where the public is at high risk ofexposure to a covert release of biologicalor chemical agents. They will also be partof a network of sensors that will monitorurban areas or large events such asinaugurations or the Olympics. They willfind their way onto the battlefield toprotect soldiers in action. Used to analyzeblood or other samples, such systems maydetect and diagnose diseases in the field,far from laboratories and hospitals.

W These monitoring systems must berobust and easy to operate and maintain.They must also have low powerrequirements to be truly portable andnot rely on the large batteries thatfrequently accompany so-called field-portable devices today. And of course,the systems, like the tiny componentsthat make them up, must be small andlightweight, which would not bepossible without the ongoingrevolution in microtechnology,particularly microelectromechanicalsystems. (See “The MicrotechnologyCenter: When Smaller Is Better,”S&TR, July/August 1997, pp. 11–17,and “Countering the BioweaponsThreat,” S&TR, June 1998, pp. 4–9.)

HandlingFluids inMicrosensors

S&TR November 1999 11


Microfluidics

Handheld instruments willincorporate microchip devices to take in an air or fluid sample; filter out smoke,dust, and other contaminants; mix thedesired particles with fluid reagents asneeded; and pump the mixture to sensorsat the other end to determine whatpathogens or toxins, if any, are present.Micromachined from silicon, glass,plastics, and ceramics, the componentswill have channels 20 to 200 micrometersdeep and up to a millimeter wide throughwhich fluids will travel. (Keep inmind that a human hair is just about50 micrometers in diameter.)

Complete microfluidic systemshave been a dream for more than adecade. At Livermore’s Center forMicrotechnology, experts in biology,electronics, optics, and engineering areworking together on several uniquecomponents that will make them a reality.

Two Projects, One GoalTwo microfluidic projects, one

sponsored by the Department of Energyand the other by the Department ofDefense, are currently under way at theCenter for Microtechnology. Thesponsorship is different but the goals forboth projects are to develop systems forhandling fluids in autonomous detectorsfor biological pathogens.

Robin Miles leads the engineeringteam that is developing fluidic systemsfor DOE. This project, which includesparticipants from several DOE sites, isdeveloping an autonomous device dubbedSentry, whose functions will includecontinuous or on-demand air sampling,sample preparation, automated fluidicsample handling and transport, detectionand identification of pathogens byimmunoassay and DNA recognition, andautomated data analysis and reporting.

Engineer Peter Krulevitch and histeam are working under a subcontract

to researchers at the University of TexasM. D. Anderson Cancer Center on theDoD project, which is sponsored by theDefense Advanced Research ProjectsAgency (DARPA) under the MicrofluidicMolecular System (MicroFLUMES)Program. The purpose of this project isto develop an instrument that incorporatesnew technologies for separating particles(known as fractionation), sensing them,and identifying them based on theirdielectric properties. Its first use will beto perform differential cell analysis onblood samples.

Building sophisticated,multifunctional, automated samplepreparation systems for field use is stillprimarily a research and developmentactivity. Most systems availablecommercially either assume a laboratorysetting for testing or are designed forone use only. Also, many fractionationmethods require filters that becomeclogged over time and contribute to thecarryover of particles between tests.Furthermore, most operate usingpneumatic power, which is excellentfor microfluidic actuation because itcan provide large forces over longdistances and conform to the tinytubes. But such systems require cartingaround a bulky canister of compressedair, and pneumatic valves cannot beminiaturized sufficiently to make thesesystems easily portable.

Both Livermore projects circumventthese challenges by using new methodsof pumping, fractionation, mixing, andsample concentration and purification.Krulevitch’s team has also developeda new sensing device that uses changesin impedance to identify particles.

The teams are working towardintegrating their assorted componentsinto a single instrument. Fabricationtechniques, fluid conductivities, andfluid velocities, among other concerns,

must be compatible for each overallsystem to perform optimally. Systemintegration is key for both programs.

Inside SentryBecause Sentry is intended for

detecting and identifying biological orchemical warfare agents in the field, itis being designed to collect samples fromthe air. But samples could just as easilybe blood or tissue for diagnosing disease.Processing the samples involves mixingreagents or microscopic polystyrenebeads coated with antibodies with thepotentially pathogenic particles.Pathogens in the sample will cling tobeads coated with the appropriateantibody. Then the sample will beconcentrated to facilitate the use ofDNA-based assays. Combining a DNA-based assay with an antibody-basedassay greatly increases Sentry’s overallreliability in identifying pathogens.

Livermore engineer Amy Wang isexploring the use of acoustic energy tomanipulate particles in the sample andto enhance the mixing of sample andreagent. Microscale mixing is a challengebecause small channel dimensions makeit difficult to create turbulence. Acousticmixing brings with it the advantage ofrapid mixing, no moving parts, and noneed for nozzles or external injectionof fluids to create turbulence. The abilityto mix samples on the chip will speed upthe rate of binding for immunoassays,increasing the throughput and speedof the system.

Electronic FilteringMiles is using dielectrophoresis as a

method of filtering the sample to collectthe particles of interest in Sentry.Dielectrophoresis is an electricalphenomenon that allows particles to betrapped or manipulated by applyingnonuniform electrical fields, which

S&TR November 199912


Microfluidics

(a)

(b)

(a) Bacillus globigii,a simulant foranthrax, collects onactivated electrodesas it flows throughwater. (b) Thespores do not collecton inactivatedelectrodes.

30-micrometerelectrode

Bacillus globigiispores

30-micrometerelectrode

induce electrical polarization in theparticles. Depending on the polarizabilityof the particle with respect to the mediumin which it is suspended, it will movetoward or away from regions of highfield intensity. Motion toward the regionsof high field intensity is termed positivedielectrophoresis, while motion awayfrom them is negative.

Dielectrophoretic forces providean electrically switchable means todiscriminate between particles andto manipulate them according to theirdielectric properties. This phenomenon isideally suited to microfluidic situationsbecause large field strengths andcorrespondingly high dielectrophoreticforces are readily achievable withelectrodes spaced less than a millimeterapart. Spores, bacteria, and cells, whosesizes range from 1 to 10 micrometers,

may be captured with dielectrophoreticelectrodes using less than 2 volts.Furthermore, dielectrophoretic forcesare effective even for extremely smallparticles such as DNA.

Miles notes that this use of positivedielectrophoretic forces to electronicallyfilter a sample—which has never beendone commercially before—solves acouple of problems. First, it can be usedto remove soil, smoke, pollen, moldspores, oil, and other contaminants fromthe raw sample. These contaminantsmay inhibit the operation of assays forDNA recognition based on thepolymerase chain reaction (PCR), themost reliable method of identifyingbiowarfare organisms. Contaminantsmake antibody-based assays operateless effectively, too. Antibody assaysare less specific than PCR, but they can

be faster when the target organism ispresent in high concentrations. Theseassays also allow for simultaneousdetection and identification of multiplebiowarfare agents including toxins.Finally, using an electronic method tofilter the sample eliminates the manualhandling of samples and the largevolumes of reagents and filter mediatypically used in laboratory analyses.

The dielectric properties of cellshave been characterized by numerousexperiments. But the dielectrophoreticproperties of antibody-coated beads,spores, and DNA are less well known.Miles and her team have explored themin numerous experiments.

They studied the nontoxic Bacillusglobigii spores and the vegetativebacteria Erwinia herbicola—simulantsfor anthrax and plague, respectively—to demonstrate their capture usingdielectrophoretic electrodes. As shownat left, the force of attraction to theelectrodes is sufficient to overcome theforce on the spores due to fluid flow,allowing debris in the carrier fluid to bewashed away while the spores are heldin place. Another series of experimentsdetermined the optimal capturefrequency for several bioparticles ofinterest, including various types of DNA,Bacillus globigii, Erwinia herbicola,and beads.

On the basis of these and otherexperiments, the team believes it knowsthe best way to capture particles. Threeparameters—the magnitude anddirection of the dielectrophoretic force,the frequency of the electrodes, and theelectrode geometry—are adjusted toselectively capture cells, spores,polystyrene beads, or DNA. Once theparticles are captured, they are held inplace against the flow of fresh carriersolution for a short time. Then theelectric field is removed so the particlescan be suspended in clean solution.

13


MicrofluidicsS&TR November 1999

Flow injection ports

Leads to electrodes

Flow exit

Electrode set

Component for testing the dielectrophoretic concentration of particles suspendedin water. Acoustic subsystems and magnetohydrodynamic pumps for Sentry arebeing designed to integrate with it.

Holes for fluid inputand output

Hole for electrical contact

Hole for electrical contact

Glass

Silicon

Glass

A cross section ofthe magnetohydro-dynamic micropump.It is currentlyfabricated of glassand silicon, but theuse of othermaterials is beingexplored.

Electromagnet

The figure at top right shows theexisting prototype of the componentfor testing the dielectrophoreticconcentration of particles suspendedin water. Other microfluidic pieces ofSentry are being designed to fittogether with this component in anintegrated system.

Keeping It MovingMagnetohydrodynamic (MHD)

pumps are being developed by physicistSony Lemoff and others at Livermoreto move fluid through all phases ofSentry’s microfluidic system. Theyhave been the first to demonstrateMHD pumps for aqueous solutionsthat function on a microchip and arepioneering the use of MHD pumpsfor micromachined applications.

There are several types ofnonmechanical pumps, but thus far, theMHD pump is the most effective forproducing a continuous, nonpulsatingflow in a complex microchanneldesign. An MHD pump consistsprimarily of an electromagnet and aseries of metal electrodes. Multiplepumps on the same chip can be drivenindependently by varying theirelectrode current amplitude and phaserelative to the electromagnet, thusenabling routing in complex integratedmicrofluidic systems.

As shown in the figure at the lowerright, the channels are etched through asilicon wafer with electrodes depositedon the walls of the channels. The siliconis then sandwiched between glass plateswith holes for electrical contacts andfluid input and output. An electromagnetis positioned beneath the device.

The first-generation MHD pumphad channels 380 micrometers deepand 800 micrometers wide, huge in themicrotechnology world. Lemoff notes,“We are aiming for channels that are20 by 50 or 50 by 100 micrometers. But

14



The integrated injection–separation–detection system being developed by the DARPA projectteam resides on a 2.5- by 4-centimeter substrate.

with such small cross-sectional areas,we need higher magnetic field strengthto push the fluid through.” The earlysystem used 0.025-tesla magnets, butthey are being replaced with much morepowerful 0.2-tesla ones.

Another Approach The project with the University of

Texas M. D. Anderson Cancer Centertakes a different tack. Krulevitch andhis team are responsible formicrofabrication on the project andhave designed the microfluidic chip.

For fractionation, the DARPAproject team is using negativedielectrophoresis to keep particles“levitated” in the microfluidic channelaway from electrodes. Krulevitch’steam together with the Andersonresearchers have designed themicrochannels and dielectrophoreticelectrode arrays that make thisseparation system work.

The DARPA project team has alsodeveloped an impedance sensor that

can detect particles as they leave thesystem and characterize them bothby their time of passage through theseparation column and by their sizeand dielectric frequency spectrum.This sensor consists of two electrodeson opposite sides of the channel. Theresistance and capacitance—and hencethe impedance—between the twoelectrodes is already known. When aparticle passes between the electrodes,the impedance value changes. Direct-current impedance sensors are used incommercial blood cell and microparticlecounters, but they indicate only particlesize. This new version, using alternatingcurrent, represents a considerableimprovement.

The integrated injection—separation—detection component shown belowconsists of microscopic sorting channels150 micrometers deep, 1 millimeterwide, and 10 centimeters long arrangedin serpentine fashion on a 2.5- by 4-centimeter substrate. Channels areequipped with integrated arrays of

dielectrophoretic electrodes50 micrometers wide with 50-micrometer-wide gaps. At the exitof the channel are impedance sensorswith their independent electrodes.

In operation, a micropump sends1-microliter samples of 10-micrometer-diameter surface-coated beads (fordetecting warfare agents) or humancells (for detecting disease) into themicrochannel. The samples are thenslowly flushed through with fluidfrom another small reservoir. Beadsor cells are detected as they pass theimpedance sensors after fractionationalong the channel.

At the M. D. Anderson CancerCenter, both the fractionation processand the impedance sensor using bothbeads and cells have performed wellin initial tests. Plans call for the fullyintegrated blood-analysis fractionationsystem, which is being packaged byLYNNTECH, Inc., of College Station,Texas, to be delivered to DARPA inthe fall of 2000.

Simulations of MicroactivityThe transport and manipulation of

beads and pathogenic particles mustbe predictable for all of the componentsto operate together in the DOE andDARPA systems.

To guide work on these subsystems,engineer David Clague has developedan enhanced three-dimensional, discretesimulation model that permits the studyof stationary and mobile particles inmicrofluidic devices. The model isbeing extended to incorporateintermolecular force interactions.Because the channels in microfluidicsystems are so small, intermolecularforces, which are typically masked inlaboratory-scale instruments, affectthe behavior of particles as they movethrough the system and are manipulated.

Known as lattice Boltzmann, themodel is based on the Boltzmann

15



A lattice Boltzmann model of fluid flow and particle movement in a microchannel. In thiscutaway side view of a microchannel, the sphere representing a particle moves from left toright. Inertial forces and wall interactions tend to lift the rotating sphere, translating it towardthe center of the microchannel. The DARPA project team is attempting to exploit this liftphenomenon in the design of a fractionation device based on negative dielectrophoresis.

Translationalvelocity

Rotationalvelocity

Lift velocity

equation, which easily incorporatesexternal forces, thermal diffusion, anddevice wall interactions in studies of thedynamic behavior of a collection ofparticles (see the figure at right). Thiscontrasts with traditional microfluidicmodeling based on finite-elementanalysis and boundary-element methods,which in this area of research typicallydeal only with pure fluid effects.

In the near future, Clague willincorporate electromagnetic-inducedpressure fields to study particle behaviorduring each process in the twoinstruments. As work on the componentsprogresses and the configuration of theoverall systems is determined,simulations will help explain theinterplay of fundamental-forceinteractions in each subsystem.

Fabricating the SystemsThe goal of all this work is to have

fully integrated systems. Nowheredoes this present a greater challengethan in the area of fabrication. Currently,no single-material system has provensuperior to all others. Each has its meritsand drawbacks, depending on the specificfunction of the component. When thevarious functions are being integrated,multiple-material systems must beconsidered, including associatedpackaging and interconnectiontechnologies.

The most commonly used materialsfor microfluidic systems are polymers,silicon, and Pyrex glass. Livermore hasalso begun fabricating microfluidicdevices using ceramics and other glasses.Engineer Harold Ackler is developingfabrication processes for both projectsand exploring a number of fabricationmethods for the various subsystems.

He and other Livermore researchersare developing proprietary technologythat will integrate new ceramic and glassdevices with commercially availablemicroelectronic packaging and other

proprietary Livermore fluidic interconnecttechnology. This integrative capabilitywill make the now troublesome task ofmaking fluidic and electronic connectionsas simple as plugging in a packagedintegrated circuit. Ackler notes,“Being able to integrate these devicesso easily is extremely attractive forsystems deployed in the field in whichreplacement of a component or aconsumable material like a reagentmust be quick and simple.”

The team has several material-relatedissues to handle. Eliminating the mostrapid corrosion mechanism has solvedthe problem of electrode corrosion.Problems with spores, dirt, DNA, andother materials adhering to the varioussurfaces in the system are being dealtwith. The first-generation MHD pumpwas made of silicon and Pyrex glass,but newer ceramic and glass alternativesare also being studied.

While it may be possible to fabricateall components with the same materials,some functions may not be optimal. So Livermore is taking a dual approach,examining the use of discretecomponents as well as pursuing thedevelopment of a fully integratedfabrication method wherein allcomponents are integrated into the samepiece of material. The latter approachshould reduce system size, powerrequirements, use of consumables,packaging problems, and manufacturingcosts. But if the use of discretecomponents results in superior systemperformance, Ackler will take thatroute, making use of Livermore’sintegration and packaging technology.

Complete Systems SoonThe Center for Microtechnology is

involved in numerous microfluidicprojects. In addition to the microfluidics

16


Microfluidics S&TR November 1999

ROBIN MILES joined Lawrence Livermore in 1997 as a mechanicalengineer specializing in the development of microdevices forbiological applications. Currently, she is an engineer and groupleader in the Laboratory’s Center for Microtechnology. She iscoprincipal investigator for sample preparation on an autonomous,continuous monitoring system for counterbiological warfarepathogens and principal investigator on a project to build a biological

processor using electric fields.Miles earned an B.S. in mechanical engineering from the Massachusetts Institute

of Technology, an M.S. in mechanical engineering from Stanford University, and anM.B.A. from the University of California at Berkeley.

PETER KRULEVITCH holds a B.S., M.S., and Ph.D. from theUniversity of California—all in mechanical engineering. He joinedLawrence Livermore in 1994 as a postdoctoral fellow and is currentlya microelectromechanical systems researcher in the EngineeringDirectorate’s Center for Microtechnology. He is principal investigatorfor a project to develop microfabricated cell separation and detectionsystems in collaboration with the University of Texas M. D. Anderson

Cancer Center. He is also working on projects to develop shape-memory filmmicroactuators for medical and microfluidic applications, to create finite-elementmodels of micromechanical devices, to microfabricate a temperature–pH biosensor,and to investigate the mechanical properties of thin films for multilayer mirrors.

Krulevitch is the coholder of eight U.S. patents for a variety of microdevices andtheir fabrication methods, primarily for medical and biotechnology applications.

About the Engineers

work for DOE and DARPA, the centeris participating with the University ofMinnesota and three other universitiesin a proposal submitted in recently tothe National Science Foundation for aCenter for Biomedical Microsystems.If the proposal is funded, 22 industrialpartners have committed to assist thecollaboration in developing diagnosticand therapeutic microsystems,including micropumps and othermicrofluidic subsystems.

For work under way now, a fullyintegrated microfluidic system for theDARPA project will be delivered nextyear, while completion of the DOEdetector is about three years away.Ray Mariella, director of Livermore’s

Center for Microtechnology, is pleasedwith successes to date. “Livermore hasbeen a leader in the microfabricationworld for quite a while. Producing ausable microfluidic system as part ofa detector for biological and chemicalagents will be a real feather in our cap.”

—Katie Walter

Key Words: Center for Microtechnology,dielectrophoresis, impedance sensor,magnetohydrodynamic (MHD) pump,microdevices, microfluidics.

For further information contact Robin Miles (925) 422-5048([email protected]) or Peter Krulevitch (925) 422-9195 ([email protected]).


Research Highlights

A CrowningAchievement for Removing Toxic Mercury

17

IKE its namesake, the messenger of the gods, mercury is notoriously mobile in the environment. Water-

soluble and toxic mercury readily leach out of landfills andeven wastes solidified with cement. In recent years,environmental scientists and regulators have focused onthe development of new processes to remove mercuryions from solutions more efficiently and cheaply thanpresent methods.

The technical challenge is formidable because anymethod must be impervious to the corrosive nature of thewaste streams that contain these ions. In addition, thesewaste streams can contain a variety of other metal ions(sometimes in much higher concentrations)—some ofwhich also possess the same +2 charge as mercury ions.So an effective removal process must be selective of onlymercury ions.

In response to the need for a better method for mercuryremoval, a team of Lawrence Livermore chemists (GlennFox, John Reynolds, and Ted Baumann) has designed anorganic polymer called Mercaptoplex that demonstrates anunusually strong affinity for mercury ions in solution.Tests at Livermore show that Mercaptoplex extracts morethan 95 percent of mercury ions and does so faster andmore selectively than other techniques such asprecipitation and activated carbon absorption. Originallydeveloped for use in processing nuclear fuel rods at theDepartment of Energy’s Idaho National Engineering andEnvironmental Laboratory, the molecule can also removemercury from both industrial waste streams and publicwater supplies.

Mercaptoplex has demonstrated a remarkable capacityfor removing mercury ions under a broad range ofconditions, including those currently found in governmentand industrial waste streams. In addition, the molecule can

L

be reused indefinitely after the bound mercury is removed,making the process cost-effective. Because of its ability to berecycled, the molecule minimizes the amount of secondarywaste generated during extraction, a major challenge inwaste treatment.

Three Molecules in OneMercaptoplex is really three molecules combined into one.

The business end belongs to a class of organic compoundscalled crowns, which are molecular rings that contain metal-binding atoms incorporated into their carbon frameworks.The original crowns featured oxygen atoms linked together ina ring by carbon atoms. They earned their name because themolecule looks like a crown when viewed from the side.

Glenn Fox (left), John Reynolds, and Ted Baumann have developedMercaptoplex, a polymer for removing toxic mercury from waste streams.

18 S&TR November 1999


Crown Polymers

Two repeating units of the mercury-extraction polymer, Mercaptoplex.The molecule consists of three parts: (a) a backbone of cross-linkedpolystyrene molecules (n) makes the molecule insoluble in water; (b) a nitrogen-linking unit (N) facilitates interaction between the crownand acidic solutions; and (c) a “crown” of five linked sulfur atoms (S)binds to a single mercury ion in solution.

nn

NN

S

S

S S

S S S

S S S

(a)

(b)

(c)

The oxygen atoms can be replaced with sulfur atoms to forma crown that exhibits a high affinity for mercury ions, throughthe donation of electrons to the positively charged mercury ion.(The molecule is still called a crown in chemistry parlance,although the sulfur atoms do not confer a crown appearance.)Together, the five linked sulfur atoms of Mercaptoplex forma strong complex with a single mercury ion.

The sulfur crown is attached to the second Mercaptoplexconstituent, a nitrogen-linking unit. The researchers surmisethat this unit facilitates interaction between the crown and theacidic aqueous solution. It also links the sulfur-containingcrown to the third component, a backbone of cross-linkedpolystyrene molecules (polystyrene is the chief ingredient ofthe ubiquitous Styrofoam coffee cup).

Strong Polystyrene BackboneThe Livermore chemists chose a backbone of polystyrene

because its chemistry is well understood and its simple cross-

links of divinylbenzene transform the molecule into a highlyentangled and thereby insoluble repeating unit (or polymer)that does not dissolve in water. The Livermore teampostulates that other materials, such as polymers ofpolyethylene, may also prove effective as backbones.

In solution, because of the entangled nature of theMercaptoplex polymer, it is probable that neighboringcrowns combine to trap mercury ions. For example, twosulfur atoms from one crown may combine with threesulfur atoms from a nearby crown to bind to a mercury ion.Studies using spectroscopic techniques are under way atLawrence Livermore to gain a better understanding of thebonding mechanism.

The Livermore chemists have shown that Mercaptoplex iseffective at pH ranging from 1.5 (extremely acidic) to 7.0(neutral). In contrast, precipitation, a common technique ofmercury removal, requires constant pH adjustment. If the pHgets too low (too acidic), the precipitation process produceshydrogen sulfide, a highly toxic gas, and does not removethe mercury. The other popular mercury removal process,activated carbon, also requires continuous adjustment of pH.

Mercaptoplex is also faster and more selective in removingmercury than other techniques. When mixed with solutionscontaining mercury ions, it captures virtually all of the mercurywithin 30 minutes. This extraction rate is much faster thanthat seen in other systems, which can take up to 20 hours to do their job. Baumann says the ultimate goal is to useMercaptoplex as packing for large columns to speed up thewaste treatment process. In this design (shown on p. 19), thewaste stream would simply flow through the Mercaptoplexwithout the need for mixing.

Fox notes that because typical mixed waste streams (thosecombining both toxic and radioactive materials) contain avariety of other metal ions, such as aluminum, iron, cadmium,and lead, removal of mercury requires a highly selective process.The chemists have tested Mercaptoplex in solutions of mercuryions ranging from 4 to 200 parts per million and whenconcentrations of other ions outnumber mercury by 100 to 1.In every case, Mercaptoplex has selectively removed mercurywith an efficiency of 95 percent or greater. (Baumann saysmercury removal is probably greater than 99 percent, but theamount of mercury left in solution after treatment is too smallfor the chemists to measure accurately.)

Recycling Is a Big AdvantageBecause Mercaptoplex is insoluble in water, it can be

easily separated from solution by filtration once the extraction

19Crown PolymersS&TR November 1999


Polymer backbone

Nitrogen-linking unit

Aqueouswaste, no

mercury ions

Aqueouswaste with

mercury ions CrownMercury ion

Schematic of column treatment of mercury waste using Mercaptoplex.

is complete. The mercury can then be recovered, and theMercaptoplex regenerated by a variety of treatments. Onemethod developed by the Livermore team is to use chloroformsolutions of diphenylthiocarbazone to strip the bound mercuryfrom the polymer. Under these regeneration conditions, thediphenylthiocarbazone has an even greater affinity for mercuryions than does the sulfur crown. Once rinsed and dried,Mercaptoplex has been used to effectively treat additionalvolumes of mercury. The team is investigating other methodsof stripping the mercury ions from the polymer such as electro-chemically reducing the ions to the safer metallic mercury.

In comparison to Mercaptoplex, other techniques typicallyrequire additional treatment steps and generate large amountsof secondary waste. Precipitation generates mercury sludgesthat require further treatment. Activated carbon columnsloaded with mercury are rarely regenerated, and the spentcolumns require additional processing.

At the Idaho National Engineering and EnvironmentalLaboratory, where mercaptoplex was first used, mercury isused as a catalyst to treat spent fuel rods from U.S. Navysubmarines. The Livermore process is also applicable at otherDOE sites that need selective and cost-effective treatmentsfor mixed waste.

The process should prove useful in treating industrialwaste streams and water supplies that contain mercury. For example, the Livermore team has discussed the processwith representatives from the bleach manufacturing and oil

industries, who must meet strict federal regulations concerningmercury levels in their waste streams.

By simple substitution of the sulfur atoms, the moleculecan be tailored to target other metal ions, such as cadmium,silver, and lead, commonly found in mixed waste streamsand water supplies. “There is a lot of synthetic chemistryyou can do with crowns,” says Fox, “such as modifying the number of noncarbon atoms in the ring to better bond to the ion in the solution of interest. In this way, chemistscan target a particular metal pollutant through carefulmolecular design.”

—Arnie Heller

Key Words: activated carbon, crown polymers, Idaho NationalEngineering and Environmental Laboratory, Mercaptoplex, mercury,precipitation.

For more information contact Glenn Fox (925) 422-0455([email protected]).

20 Research Highlights S&TR November 1999


Flat-PanelDisplays Slim Downwith Plastic

Plastic substrates forflat-panel displaysare flexible,transparent, andlightweight.

OR years, manufacturers of electronics with flat-paneldisplays have dreamed of using plastic as a cheaper, more

compact, more rugged, and far more lightweight alternative toglass. The Department of Defense is particularly interested inultrathin yet flexible screens as standard equipment for thePentagon’s “information warrior” of the next century. Withplastic displays, soldiers could hang satellite navigationsystem displays on their belts or keep electronic maps rolledup in a back pocket.

The most advanced type of flat-panel displays, used inmost portable computers, is active-matrix liquid-crystaldisplays. In this display, each of the million or so tiny screenpixels is controlled by thin-film transistors (TFTs) that act astiny on/off electrical switches. By turning on and off dozensof times a second, the TFTs permit continuously changingimages of words, pictures, and video.

Currently, TFTs for active-matrix displays aremanufactured onto a rigid glass substrate in a process thatinvolves baking glass sheets at temperatures of up to 600°C.This conventional process is far hotter than any plastic canwithstand without deforming and melting. But now a team of

F Lawrence Livermore researchers is showing how TFTs can bemanufactured on top of thin, flexible plastic sheets instead ofglass by keeping manufacturing temperatures at or below 100°C.

The work was carried out by a group of electrical engineers,physicists, and materials scientists in the Device and ProcessGroup in the Information Science and Technology Program ofthe Laser Programs Directorate. The research is part of alarger effort by Livermore scientists and their Department ofEnergy colleagues to apply laser-based processing techniquesto current U.S. semiconductor production problems. Theplastic substrate project, now in its third year, is funded by theDefense Advanced Research Projects Agency’s High-DefinitionSystems Program, which sponsors development of new displayconcepts that address the issues of lighter weight, improvedruggedness, lower power, higher resolution, and easier use.

Laser Pulses Fast, PreciseThe novel Livermore transistor fabrication process

combines well-established, low-temperature depositiontechniques with excimer lasers that produce pulsed beams ofultraviolet light. These lasers are a much more powerful version

21Plastic Display SubstratesS&TR November 1999


Wavelength = 308 nanometers

Polysilicon

Silicon oxide

Plastic

An excimer laserirradiates a layerof amorphoussilicon atoms ata wavelength of308 nanometersfor about35 nanoseconds.The process isrepeated up to10 times.

of the instruments that are used in eyesight correction surgery toliterally vaporize corneal tissue without damaging surroundingtissue. In fact, the lasers are so precise they can make precisenotches in human hair that can be exactly and repeatedlyduplicated. The Livermore team takes advantage of the laser’sextreme precision and ultrafast operation to melt, crystallize,and dope (add impurities to) the silicon layers forming the TFTsat substrate temperatures lower than the melting temperatureof plastic.

The Livermore team chose one of the most commonplastics for the substrate: polyethyleneterephthalate (PET),more commonly known as polyester. Thin (175 micrometers),cheap, flexible, transparent, and rugged, PET is used for manyother purposes, including the Mylar for viewgraphs. Standard10-centimeter-diameter wafers are cut from 61-centimeter-widerolls of PET. Onto these plastic circles are applied the materialsfundamental to integrated circuits: an insulator (silicon dioxide),semiconductor (crystallized silicon or polysilicon), dopantsof selected elements, and metal connectors.

The process begins with a thin layer of silicon dioxidedeposited on the plastic wafer through a conventional processcalled plasma-enhanced chemical-vapor deposition that producesuniform films of molecules. Next, the team uses sputterdeposition to apply an amorphous layer of silicon atoms tothe substrate. Both of these layers are applied at a relativelycool temperature of about 100°C to keep the plastic intact.

The excimer laser irradiates the amorphous silicon layerfrom 3 to 10 times at an ultraviolet (UV) wavelength of308 nanometers. Each pulse lasts only 35 nanoseconds (billionthsof a second) while melting the amorphous silicon. The result(shown on p. 22) is a highly ordered, polycrystalline layer ofsilicon atoms some 40 nanometers thick. (This transformedsilicon, typically called polysilicon, permits electrons tomove more easily through its highly ordered lattices.)

Plastic Doesn’t MeltDuring the melting process, the fleeting UV laser energy

is absorbed mainly in the top 10 nanometers of the amorphoussilicon layer before it diffuses downward into the plastic.That localization of the laser energy, together with the silicondioxide layer that acts as a thermal barrier, keeps the plasticsubstrate from heating and melting.

Although the silicon layer melts at 1,400°C, the plasticbarely notices the heat from the deposited laser energy. Theteam’s understanding of the physics and chemistry of the laserprocessing steps is aided by advanced simulation work doneat Livermore.

The laser beam is adjusted to cover from 1 to 11 squaremillimeters at the wafer surface. Covering the entire wafertakes about one minute. In contrast, traditional processesrequire baking glass sheets in high-temperature furnaces formany hours.

The next steps are modified, lower-temperature versionsof traditional semiconductor processing involvingphotolithography, which uses a sequence of photomasks.These masks act as photographic negatives do, allowing lightto imprint a pattern on the wafer. The pattern defines the areasto be removed through etching, doped with impurities, anddeposited with aluminum connectors.

The doping with boron and other elements is accomplishedusing another pulsed excimer laser in a technique also developedat Livermore. First, a thin layer of doping atoms is depositedusing plasma-enhanced chemical-vapor deposition. Thenrepeated laser pulses drive the atoms deep into the polysilicon.(Doping allows the polysilicon, which is essentially an insulator,to conduct electricity by giving up or attracting electrons.)

Switches Ready for ConnectionThe result is a 10-centimeter-diameter array of several

hundred simple switches ready to be joined to its neighborsand to a liquid-crystal-display system. The Livermore team iscontinuing to refine the low-temperature manufacturingprocess. In particular, it is working to achieve TFTs thatpermit electrical current with higher “mobility,” or speed.The bigger the display, the higher the desired mobility.



Plastic Display Substrates

Electron micrographs show how the excimer laser pulses transformthe amorphous silicon layer into a 40-nanometer-thick layer ofpolycrystalline silicon.

(a) Before laser

(b) After laser

Epoxy

Amorphous silicon

Plastic substrate

Epoxy

SIlicon dioxide

Plastic substrate

Polycrystalline silicon

Silicon dioxide

100 nanometers

The research has progressed sufficiently that discussionsare taking place with U.S. flat-panel-display manufacturers tolicense the technology. It is anticipated that an industry–Livermore project to develop a complete prototype wouldcombine Livermore’s plastic “backplane” of TFT-drivenpicture elements, or pixels, with liquid crystals or organiclight-emitting materials furnished by a display manufacturer.

Display manufacturers are particularly interested in thepotential to manufacture large displays inexpensively,particularly with a roll-to-roll continuous manufacturingtechnique much like the roll-to-roll printing process. In thisscenario, the plastic would roll through processing stationssimilar to those of a printing press, and finished displayswould be cut to size.

The Livermore breakthrough may well make possiblewithin a few years a new generation of ultralight, flexible,and inexpensive displays. Applications could includenotebook and desktop computer displays, instrument panels,video game machines, videophones, mobile phones, hand-held PCs, camcorders, satellite navigation systems, smartcards, toys, and a new generation of electronic devices forwhich flat-panel displays have been too heavy or too costly.Indeed, it looks as if plastic flat-panel displays will be usedby everyone, from couch potatoes to information warriors.

—Arnie Heller

Key Words: active-matrix liquid-crystal display, amorphous silicon,Defense Advanced Research Projects Agency, enhanced chemical-vapor deposition, excimer laser, flat-panel display, liquid-crystaldisplay, polysilicon, sputtering, thin-film transistor (TFT).

For further information contact Tom Sigmon (925) 422-6753([email protected]).

23Each month in this space we report on the patents issued to and/orthe awards received by Laboratory employees. Our goal is toshowcase the distinguished scientific and technical achievements ofour employees as well as to indicate the scale and scope of thework done at the Laboratory.

Patents and Awards

Patent issued to

Paul WickboldtPaul G. CareyPatrick M. SmithAlbert R. Ellingboe

Joseph P. Fitch

G. Bryan BalazsPatricia R. Lewis

Jay A. SkidmoreBarry L. Freitas

Patent title, number, and date of issue

Deposition of Dopant Impurities andPulsed Energy Drive-In

U.S. Patent 5,918,140June 29, 1999

Sparse Aperture Endoscope

U.S. Patent 5,919,128July 6, 1999

Mediated Electrochemical Oxidationof Organic Wastes Using a Co (III)Mediator in a Neutral Electrolyte

U.S. Patent 5,919,350July 6, 1999

Microlens Frames for Laser DiodeArrays

U.S. Patent 5,923,481July 13, 1999

Summary of disclosure

A semiconductor doping process that enhances the dopant incorporationby using the gas-immersion laser-doping technique. The enhanceddoping is achieved by first depositing a thin layer of dopant atoms on asemiconductor surface, followed by exposing the semiconductor to oneor more pulses from either a laser or an ion beam to melt a portion of thesemiconductor to a desired depth. This process causes the dopant atomsto be incorporated into the molten region. After the molten regionrecrystallizes, the dopant atoms are electrically active. The dopant atomsare deposited by plasma-enhanced chemical-vapor deposition or anotherknown deposition technique.

An endoscope that has a smaller imaging component, maintainsresolution of a wide-diameter optical system while increasing toolaccess, and allows stereographic or interferometric processing for depthand perspective information and visualization. Because its imagingoptics are smaller, more of its volume can be used for nonimaging tools,thus permitting smaller incisions when it is used in surgical anddiagnostic medical applications. In turn, it produces less trauma to thepatient or allows access to smaller volumes than is possible with largerinstruments. The endoscope has fiber-optic light pipes in an outer layerfor illumination, a multipupil imaging system in an inner annulus, and anaccess channel for other tools in the center. The endoscope can be usedas a flexible scope, thus increasing its utility. Because the endoscopeuses a multiaperture pupil, it can also be used as an optical array,allowing stereographic or interferometric processing.

An electrochemical cell with a cobalt (III) mediator and neutral pHanolyte, which provides efficient destruction of organic and mixedwastes. The organic waste is concentrated in the anolyte reservoir, wherethe cobalt mediator oxidizes the organics and insoluble radioactivespecies and is regenerated at the anode until all organics are converted tocarbon dioxide and destroyed. The neutral electrolyte is noncorrosiveand thus extends the lifetime of the cell and its components.

Monolithic microlens frames to enable the fabrication of monolithiclaser diode arrays. They can be manufactured inexpensively and havehigh registration and inherent focal length compensation for any lensdiameter variation. A monolithic substrate is used to fabricate a low-cost microlens array. The substrate is wet-etched or sawed with a seriesof V grooves. The V grooves can be created using wet-etching byexploiting the large etch-rate selectivity of different crystal planes. TheV grooves provide a support frame for either cylindrical or custom-shaped microlenses. Because the microlens frames are formed byphotolithographic semiconductor batch-processing techniques, they canbe formed inexpensively over large areas with precise lateral andvertical registration. The V groove has an important advantage forpreserving the correct focus for lenses of varying diameters.

Patents


Patent issued to

Thomas C. Kuklo

Peter A. KrulevitchAbraham P. LeeM. Allen NorthrupWilliam J. Benett

Patent title, number, and date of issue

Concentric Ring Flywheel withHooked Ring Carbon FiberSeparator/Torque Coupler

U.S. Patent 5,924,335July 20, 1999

Microbiopsy/Precision CuttingDevices

U.S. Patent 5,928,161July 27, 1999

Summary of disclosure

A concentric-ring flywheel with expandable separators (which functionas torque couplers) between the rings to take up the gap formed betweenadjacent rings due to differential expansion between different-radiusrings during rotation of the flywheel. The expandable separators ortorque couplers include a hooklike section at an upper end that ispositioned over an inner ring and a shelflike or flange section at a lowerend onto which the next adjacent outer ring is positioned. As theconcentric rings are rotated, the gap formed by the differential expansionbetween them is partially taken up by the expandable separators ortorque couplers to maintain torque and centering attachment of theconcentric rings.

Devices for performing tissue biopsy on a small scale (microbiopsy).By reducing the size of the biopsy tool and removing only a smallamount of tissue or other material in a minimally invasive manner,these devices reduce the risk, cost, injury, and patient discomfortassociated with traditional biopsy procedures. By using micromachiningand precision machining capabilities, it is possible to fabricate smallbiopsy/cutting devices from silicon. These devices can be used in one of four ways: (1) intravascularly, (2) extravascularly, (3) by vesselpuncture, and (4) externally. Additionally, the devices may be used in precision surgical cutting.

Robin Newmark and Roger Aines, together withcollaborators at the University of California at Berkeleyand Southern California Edison, were recently awarded the Environmental Protection Agency’s OutstandingRemediation Technology Award for their work on dynamicunderground stripping and hydrous pyrolysis/oxidation,technologies that heat soil and groundwater to removecontaminants and destroy them in place. (See S&TR, May1998, pp. 4–11.) The award officially recognizes “technicalexcellence in the development of innovative in situ thermaltreatment technologies.”

Charles A. McDonald, Jr., associate director emeritusat-large of the Laboratory and a member the U.S. StrategicCommand’s Strategic Advisory Group, was recentlypresented the Department of Defense Distinguished PublicService Award. Given on behalf of Defense SecretaryWilliam S. Cohen, the award recognizes “exceptionallysuperior civilian public service.” McDonald was specificallycited for his tireless efforts in developing ways to monitorthe safety and reliability of the nuclear stockpile.

McDonald, who retired from the Laboratory in 1993 but still works as a participating guest and consultant, was alsorecognized for his leadership of the 1997–1998 StockpileAssessment Team, a group of civilian and retired militaryanalysts who reviewed testimony from several governmentagencies and provided the U.S. Strategic Command with an in-depth assessment of the nation’s nuclear stockpile.

Paula Trinoskey, a health physicist in the Education andTraining Division of the Hazards Control Department, wasawarded emeritus status by the Board of Directors of theNational Registry of Radiation Protection Technologists.

Trinoskey has served on the Panel Examiners and the Board of Directors of the National Registry of Radiation ProtectionTechnologists. She is currently the liaison between the AmericanAcademy of Health Physics and the Board of the NationalRegistry of Radiation Protection Technologists.

Emeritus status is awarded in recognition of outstandingcontributions to the registry and has only been awarded to 22 individuals since 1976.

Awards



Patents and Awards


Extreme Ultraviolet Lithography: Imaging the Future

Lawrence Livermore has joined forces withSandia/California and Lawrence Berkeley nationallaboratories to research and develop extreme ultravioletlithography (EUVL) for use in the semiconductor industry.Lithography is the critical technology that enables computermicrochip manufacturers to place more and more features oneach chip, thus increasing computer power while shrinkingcomputer size. With current lithography techniques pushedabout as far as they can go, semiconductor industries mustsoon decide on a standard lithography technology forproducing the next generation of computer microchips.EUVL uses extreme ultraviolet light to produce microchipcircuit lines smaller than 100 nanometers in width. The threelaboratories are integrating the needed technologies into anengineering test stand to demonstrate how EUVL can meetindustry requirements. Lawrence Livermore is leadingefforts to develop the optical systems and components, thinfilms, masks, and submicrometer metrology needed to bringthis technology into everyday use in the semiconductorindustry of the future.ContactDonald Sweeney (925) 422-5877 ([email protected]).

Handling Fluids in MicrosensorsLawrence Livermore's Center for Microtechnology has

been a leader in the design and fabrication of micro-electromechanical systems (MEMS) for several years.Current work in integrated microfluidic systems is arelatively new avenue for MEMS research. These systemswill handle the air and fluid samples that are taken toidentify biological or chemical warfare agents or to identifythe cells that cause disease. (In other work at Livermore,detectors are being developed that identify the specificpathogen or disease present in the sample.) Devices smallenough to fit on a microchip of silicon, glass, or plasticwill take in samples; mix them with reagents; separate out DNA, cells, or other agents in the sample; andsense the presence of those agents.ContactRobin Miles (925) 422-5048 ([email protected]) or Peter Krulevitch (925) 422-9195 ([email protected]).

Abstracts

U.S. Government Printing Office: 1999/583-046-80024

Coming Next

Month

Coming Next

Month

Also in December• Similarities between the

human and mouse genomes helpus understand gene functionsand inherited diseases.

• Multilayer optics reveal solaractivity in greater detail.

• The spheromak offers analternative concept for achievingmagnetic fusion energy.

Modeling the Internal Combustion

Engine

Using their chemical kinetics code to model combustion

processes, Livermore researchers are developing ways to increase fuelefficiency in engines while reducing

exhaust emissions.

Science &

Technology Review

Lawrence Liverm

ore National Laboratory

P.O. B

ox 808, L-664Liverm

ore, California 94551

Printed on recycled paper.

U. S

. Postage

PAID

Albuquerque, N

MP

ermit N

o. 853

imaging the future of microcomputer chips · to continually shrink the size of the basic elements...

Documents