<<

SPECIAL ISSUE

SEPTEMBER 2001 $4.95 WWW.SCIAM.COM

Medical Nanoprobes Atom-Moving Tools Buckytube Electronics New Laws of Physics Living Machinery Nano Science Fiction

NThe ScienceANOTECH of the Small Gets Down to Business ALSO Eric Drexler on Nanorobots and Richard Smalley on Why They Won’t Work

Copyright 2001 Scientific American, Inc. september 2001 contentsSCIENTIFIC AMERICAN Volume 285 Number 3 features

SPECIAL ISSUE Magnified tip of an atomic force microscope OVERVIEW 32 Little Big Science BY GARY STIX Nanotechnology is all the rage. Will it meet its ambitious goals? And what is it, anyway? NANOVISIONS 74 Machine-Phase Nanotechnology NANOFABRICATION BY K. ERIC DREXLER 38 The Art of Building Small The leading visionary in the field forecasts BY GEORGE M. WHITESIDES how nanorobots will transform society. AND J. CHRISTOPHER LOVE The search is on for cheap, efficient ways to make NANOFALLACIES structures only a few billionths of a meter across. 76 Of Chemistry, Love and Nanobots BY RICHARD E. SMALLEY NANOPHYSICS A Nobel Prize winner explains why self-replicating 48 Plenty of Room, Indeed nanomachines won’t work. BY MICHAEL ROUKES There is plenty of room for practical innovation at NANOINSPIRATIONS the nanoscale—once the physical rules are known. 78 The Once and Future Nanomachine BY GEORGE M. WHITESIDES NANOELECTRONICS Lessons from nature on building small. 58 The Incredible Shrinking Circuit BY CHARLES M. LIEBER NANOROBOTICS Researchers have built nanoresistors and 84 Nanobot Construction Crews nanowires. Now they have to find a way BY STEVEN ASHLEY to put them together. One company’s quest to develop nanorobots.

NANOMEDICINE NANOFICTION 66 Less Is More in Medicine 86 Shamans of Small BY A. PAUL ALIVISATOS BY GRAHAM P. COLLINS Nanotechnology’s first applications may include Nanotechnology has become a favorite topic biomedical research and disease diagnosis. of science-fiction writers.

www.sciam.com SCIENTIFIC AMERICAN 5

Copyright 2001 Scientific American, Inc. departmentsSCIENTIFIC AMERICAN Volume 285 Number 3 8 SA Perspectives The National Nanotechnology Initiative brings a welcome boost to the physical sciences and engineering. 10 How to Contact SA 18 26 10 On the Web 12 Letters 30 Profile: Elizabeth Gould This neurobiologist looks at how memory 16 50, 100 & 150 Years Ago and healing in the brain may rely on the growth 18 News Scan of new neurons. Solved: the solar neutrino problem. 92 Working Knowledge Drawbacks of the cancer-fighting drug Gleevec. Fleas flee from new “spot” treatments used on pets. Retinal displays for pilots. How snowball Earth got rolling. 94 Voyages Geological tours expose the innermost secrets No more anonymous Web surfing? of New York City and beyond. Hunting jaguars with darts. By the Numbers: Reliability of crime statistics. 98 Reviews Data Points: Believers in the paranormal. Science, Money, and Politics: Political Triumph and Ethical Erosion depicts American “Big Science” as a bloated, whiny, self-important bureaucracy. 66 columns

29 Skeptic BY MICHAEL SHERMER The new religion of cryonics offers to raise its faithful dead. 102 Puzzling Adventures BY DENNIS E. SHASHA Square dancing without collisions. 103 Anti Gravity BY STEVE MIRSKY Never take off your shoes near a Komodo dragon. 104 Endpoints

ABOUT THE PHOTOGRAPHER: The work of Felice Frankel appears throughout this issue. Collaborating with scientists, Frankel creates film and digital imagery related to diverse areas of science, including nanotechnology. Her images have appeared in major national magazines and technical journals. In January 2002 the MIT Press will publish her guide to photographing science. Recently she received a three-year grant from the Alfred P. Sloan Foundation to co-author a book on nanotechnology with Harvard University’s George M. Whitesides. She and Whitesides wrote a previous book, On the Surface of Things: Images of the Extraordinary in Science (Chronicle Books, 1997).

Cover image and preceding page: Felice Frankel, with technical help from J. Christopher Love; this page, clockwise from top left: Lawrence Berkeley National Laboratory; Robert Young Pelton/Corbis; Felice Frankel, with technical help from K. F. Jensen, M. G. Bawendi, C. Murray, C. Kagan, B. Dabbousi and J. Rodriguez-Viego of M.I.T.

6 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. SA Perspectives Megabucks for Nanotech

Biologists sometimes stand accused of physics envy: fastest transistors or the strongest and lightest mate- a yearning for irreducible, quantifiable laws sufficient rials ever made. But even if the program gives special to explain the complex workings of life. But the jeal- emphasis to the physical sciences and engineering, it ousy goes both ways. Physicists, chemists and other has something for everyone. Biologists, of course, nonbiologists have long suffered from what can only have their own claim on the molecular realm. And be called NIH envy: the longing for the hefty increas- nanotechnology could supply instrumentation to es in research funding that seem to go every year to the speed gene sequencing and chemical agents to detect National Institutes of Health. tumors that are only a few cells in size. From 1970 through 2000, Of course, a program that tries to accommodate federal backing for the life sci- everyone could end up as a bottomless money sink. In ences more than tripled in con- his new book Science, Money and Politics (reviewed stant dollars, whereas money for in this issue on page 98), journalist Daniel S. Green- the physical sciences and engi- berg warns of the dangers inherent in an indiscrimi- neering has by comparison re- nate, all-encompassing approach to research that eats mained flat. But last year the up money. Skeptics have wondered whether sizable Clinton administration delivered increases are warranted for such a nascent field. A a valentine to the physics, chem- Congressional Research Service report last year raised istry and materials science com- questions about why nanotechnology merited such munities: the National Nano- generosity, given that some of its research objectives technology Initiative provided a may not be achievable for up to 20 years. big boost in funding for the sci- But the initiative is more than mere marketing. A ence and engineering of the small. portfolio of diverse ideas—unlike a program focused NANOMACHINES The initiative, moreover, seems on, say, high-temperature superconductivity—may to have staying power. The Bush help ensure success of a long-term agenda. The vari- White House has targeted a more modest but still sub- ety of research pursuits increases the likelihood that stantial increase for nanotech. If the president’s bud- some of these projects will actually survive and flour- get request passes, federal funding for nanotechnol- ish. Industry, in contrast, is generally reluctant to in- ogy, at $519 million, will have nearly doubled in the vest in broad-based research programs that may not

past two years, more than quadrupling since 1997. bear fruit for decades. COLORIZATION BY FELICE FRANKEL The initiative may prove to be one of the most bril- Because the development of tools and techniques liant coups in the marketing of basic research since the for characterizing and building nanostructures may have announcement, in 1971, of the “War on Cancer.” far-reaching applicability across all sciences, nano- Nanotechnology—the study and manufacture of technology—the focus of this issue of Scientific Amer- structures and devices with dimensions about the size ican—could serve as a rallying point for physicists, of a molecule—offers a very broad stage on which the chemists and biologists. As such, it could become a research community can play. Nanometer-scale physics model for dousing NIH envy and the myriad other skir-

and chemistry might lead directly to the smallest and mishes that occur in the yearly grab for research dollars. M. J. MURPHY, D. A. HARRINGTON AND L. ROUKES California Institute of Technology;

THE EDITORS [email protected]

8 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. How to Contact Us On the Web WWW.SCIENTIFICAMERICAN.COM EDITORIAL ADVERTISING For Letters to the Editors: www.sciam.com has electronic THE SCIENTIFIC REASON TO GO ONLINE Letters to the Editors contact information for sales Scientific American representatives of Scientific American in all regions of 415 Madison Ave. the U.S. and in other countries. FEATURED STORY New York, NY 10017-1111 or New York [email protected] Scientific American Please include your name 415 Madison Ave. Looking for Aliens and mailing address, New York, NY 10017-1111 212-451-8893 and cite the article fax: 212-754-1138 and the issue in Los Angeles which it appeared. 310-234-2699 Letters may be edited fax: 310-234-2670 for length and clarity. San Francisco We regret that we cannot 415-403-9030 answer all correspondence. fax: 415-403-9033 Chicago For general inquiries: Rocha & Zoeller Media Sales Scientific American 312-782-8855 fax: 312-782-8857 415 Madison Ave. Dallas New York, NY 10017-1111 The Griffith Group We often assume that extraterrestrials possess the 212-754-0550 972-931-9001 intelligence and the technology to contact us. But what fax: 212-755-1976 fax: 972-931-9074 or Detroit if they aren’t that smart? How would we find them? [email protected] Karen Teegarden & Associates In fact, many scientists are now suggesting that when 248-642-1773 we discover alien life—if we do—it won’t resemble the SUBSCRIPTIONS fax: 248-642-6138 U.S. and Canada Canada cunning eight-eyed rivals of Star Trek episodes. Instead, 800-333-1199 Fenn Company, Inc. they say, it will most likely come in the form of tiny Outside North America 905-833-6200 microbes. With that in mind, scientists at NASA’s Jet 515-247-7631 fax: 905-833-2116 U.K. Propulsion Laboratory are now developing ways to or The Powers Turner Group search for cells among the stars. www.sciam.com +44-207-592-8331 For address changes fax: +44-207-630-6999 www.sciam.com/explorations/2001/071601alien/ on subscriptions: France and Switzerland Scientific American PEM-PEMA Box 3187 +33-1-4143-8300 Harlan, IA 51537 fax: +33-1-4143-8330 NEW TO THE SITE or Germany Publicitas Germany GmbH www.sciam.com +49-69-71-91-49-0 REPRINTS fax: +49-69-71-91-49-30 Sweden Scientific American Jobs To order reprints of articles: Andrew Karnig & Associates An innovative career center focused on creating Reprint Department +46-8-442-7050 Scientific American fax: +49-8-442-7059 synergies between star talent and top-notch companies 415 Madison Ave. Belgium in the science and technology industries. New York, NY 10017-1111 Publicitas Media S.A. 212-451-8877 +32-2-639-8445 www.scientificamericanjobs.com fax: 212-355-0408 fax: +32-2-639-8456 Middle East and India [email protected] Peter Smith Media & PERMISSIONS Marketing +44-140-484-1321 For permission to copy or reuse BOOKSTORE fax: +44-140-484-1320 SA material from : Japan [email protected] Pacific Business, Inc. Introducing the new and improved Scientific or +813-3661-6138 American Bookstore—now with suggested reading 212-451-8546 for procedures fax: +813-3661-6139 or Korea recommended by our editors, more selections and Permissions Department Biscom, Inc. easier navigation! Scientific American +822-739-7840 fax: +822-732-3662 415 Madison Ave. Hong Kong www.sciam.com/books/ New York, NY 10017-1111 Hutton Media Limited Please allow three to six weeks +852-2528-9135 for processing. fax: +852-2528-9281

10 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. ® Letters EDITORS@ SCIAM.COM Established 1845

EDITOR IN CHIEF: John Rennie WRITES FORMER PRESIDENT JIMMY CARTER, “I read with EXECUTIVE EDITOR: Mariette DiChristina great interest ‘The Arctic Oil & Wildlife Refuge,’ by W. Wayt Gibbs. MANAGING EDITOR: Michelle Press ASSISTANT MANAGING EDITOR: Ricki L. Rusting I had the privilege of signing the 1980 Alaska National Interest NEWS EDITOR: Philip M. Yam Lands Conservation Act, which established the Arctic National SPECIAL PROJECTS EDITOR: Gary Stix SENIOR WRITER: W. Wayt Gibbs Wildlife Refuge and specifically prohibited oil development in its EDITORS: Mark Alpert, Steven Ashley, 1.5-million-acre coastal plain. I also had the opportunity to visit Graham P. Collins, Carol Ezzell, Steve Mirsky, George Musser, Sarah Simpson the area and witness its great herds of caribou, muskoxen and CONTRIBUTING EDITORS: Mark Fischetti, other wildlife. I feel that those in Congress who soon will render Marguerite Holloway, Madhusree Mukerjee, Paul Wallich their own judgment on whether to protect or drill the Arctic refuge would benefit from reading your article.” EDITORIAL DIRECTOR, ONLINE: Kristin Leutwyler SENIOR EDITOR, ONLINE: Kate Wong For further comments on this and other articles from the May ASSOCIATE EDITOR, ONLINE: Harald Franzen issue, please read on. WEB DESIGN MANAGER: Ryan Reid

ART DIRECTOR: Edward Bell SENIOR ASSOCIATE ART DIRECTOR: Jana Brenning ASSISTANT ART DIRECTORS: Johnny Johnson, Mark Clemens YOUR OIL OR YOUR WILDLIFE? Gibbs noted the findings of several re- PHOTOGRAPHY EDITOR: Bridget Gerety Those who refer to the Arctic National searchers who speculate that oil develop- PRODUCTION EDITOR: Richard Hunt Wildlife Refuge (ANWR) as the last pris- ment and caribou cannot mix. But he COPY DIRECTOR: Maria-Christina Keller tine wilderness in America [“The Arctic failed to cite published, peer-reviewed sci- COPY CHIEF: Molly K. Frances COPY AND RESEARCH: Daniel C. Schlenoff, Oil & Wildlife Refuge,” by W. Wayt entific studies indicating that oil develop- Rina Bander, Sherri A. Liberman, Shea Dean Gibbs] either are purposely misrepre- ment in Alaska’s Arctic has not affected EDITORIAL ADMINISTRATOR: Jacob Lasky senting the facts or have never visited the calving success or herd growth. Any bill SENIOR SECRETARY: Maya Harty region. permitting oil development will require ASSOCIATE PUBLISHER, PRODUCTION: William Sherman MANUFACTURING MANAGER: Janet Cermak In truth, it is 100 miles of flat, barren, the highest degree of wildlife protection. ADVERTISING PRODUCTION MANAGER: Carl Cherebin frozen tundra, a small part of the 1,100- PREPRESS AND QUALITY MANAGER: Silvia Di Placido DON YOUNG PRINT PRODUCTION MANAGER: Georgina Franco mile coastline of Alaska on the Arctic U.S. Congressman, Alaska PRODUCTION MANAGER: Christina Hippeli Ocean. It is predicted to hold about the ASSISTANT PROJECT MANAGER: Norma Jones CUSTOM PUBLISHING MANAGER: Madelyn Keyes-Milch same quantity of oil as we have imported GIBBS REPLIES: According to the 2000 Annual

ASSOCIATE PUBLISHER/VICE PRESIDENT, CIRCULATION: from Saudi Arabia during the past 30 Report of the U.S. Energy Information Admin- Lorraine Leib Terlecki years. It will keep the Alaska pipeline full istration, Saudi Arabia sent 10.7 billion barrels CIRCULATION MANAGER: Katherine Robold CIRCULATION PROMOTION MANAGER: Joanne Guralnick for at least the next 30 years. of oil to the U.S. from 1969 to 1999. That is half FULFILLMENT AND DISTRIBUTION MANAGER: Rosa Davis TED STEVENS again as much as the best guess for the 30- PUBLISHER: Denise Anderman U.S. Senator, Alaska year production from the 1002 Area. EIA ana- ASSOCIATE PUBLISHER: Gail Delott SALES DEVELOPMENT MANAGER: David Tirpack lysts predict a maximum production rate from SALES REPRESENTATIVES: Stephen Dudley, estimates the refuge of about half a million barrels a day Wanda R. Knox, Hunter Millington, Christiaan Rizy, The U.S. Geological Survey Stan Schmidt, Debra Silver that between six billion and 16 billion 10 years after development begins, with a peak

ASSOCIATE PUBLISHER, STRATEGIC PLANNING: Laura Salant barrels of recoverable oil are in ANWR. of nearly a million barrels a day about a decade PROMOTION MANAGER: Diane Schube Even if the mean estimate were discov- after that. To “keep the Alaska pipeline full” re- RESEARCH MANAGER: Aida Dadurian PROMOTION DESIGN MANAGER: Nancy Mongelli ered, it would be the largest oil field quires 2.1 million barrels a day, but only 1.1

GENERAL MANAGER: Michael Florek found worldwide in the past 40 years. million barrels flowed through it in 1999, and BUSINESS MANAGER: Marie Maher At a time when we face a significant production is falling steadily. Alaska’s Division MANAGER, ADVERTISING ACCOUNTING AND COORDINATION: Constance Holmes energy crisis and our dependence on oth- of Oil and Gas estimates that 20 years from

DIRECTOR, SPECIAL PROJECTS: Barth David Schwartz er nations for energy is rising, we must now North Slope oil fields outside of the 1002

MANAGING DIRECTOR, SCIENTIFICAMERICAN.COM: look here at home for solutions. Al- Area will produce only 408,000 barrels a day. Mina C. Lux though ANWR alone is not the answer, There are large natural variations in the DIRECTOR, ELECTRONIC PUBLISHING: Martin O. K. Paul we cannot ignore the tremendous re- populations and breeding patterns of the ani- DIRECTOR, ANCILLARY PRODUCTS: Diane McGarvey PERMISSIONS MANAGER: Linda Hertz sources that exist there. It can be safely mals that live on the North Slope, so any effect MANAGER OF CUSTOM PUBLISHING: Jeremy A. Abbate explored and should be a part of our na- of human activities on those trends may not CHAIRMAN EMERITUS: John J. Hanley tional energy strategy. be visible until many years of data have been CHAIRMAN: Rolf Grisebach collected and many confounding factors have PRESIDENT AND CHIEF EXECUTIVE OFFICER: FRANK MURKOWSKI Gretchen G. Teichgraeber U.S. Senator, Alaska been studied and appropriately controlled for. VICE PRESIDENT AND MANAGING DIRECTOR, INTERNATIONAL: Chairman, U.S. Senate Energy Moreover, some of the studies that examined Charles McCullagh VICE PRESIDENT: Frances Newburg and Natural Resources Committee effects on caribou from North Slope oil devel-

12 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Letters

opment are not immediately applicable to the test’s reliability and validity and ignore er heuristic systems as being used within 1002 Area because of differences in geogra- those that demonstrate its merits and agents that trawl the Semantic Web. I see the phy, herd size and the distribution of vegeta- sound psychometric properties. output of such systems as being very useful tion. Nevertheless, much of the older peer- They also fail to recognize that no psy- and sometimes so valuable that it is reentered reviewed research relevant to this debate was chological measure should be used in iso- into the Semantic Web as trusted data. But it in fact cited in the article. lation when making clinical decisions. can’t be the basis for the Semantic Web. In the Well-trained Rorschachers know that in- basic Semantic Web you have to be able to fol- “The Arctic Oil & Wildlife Refuge” over- terpretations based on any given test must low links successively across the globe with- states the benefits of drilling by asking be supplemented by information obtained out getting fuzzier. how much oil might ultimately be eco- through other methods. The Rorschach nomically recovered if the cost of finding has prevailed because it captures the com- E. COLI–FREE COOKOUTS it were already sunk. But exploration plexity of human functioning in a way In “Antimicrobe Marinade” [Staking costs—economic as well as environmen- that self-report measures alone do not. Claims], Gary Stix presents three new, tal—also have to be considered when LISA MERLO high-tech methods to decrease E. coli judging whether to allow drilling. If find- DOUGLAS BARNETT 0157:H7 in meat. I, for one, plan on con- ing costs are included, then at a sustained Department of Psychology tinuing to fall back on an older (Nean- West Coast price of $24 a barrel (in 1996 Wayne State University dertals used it), low-tech but foolproof dollars), the mean expected economical- method to kill the bacteria: cooking. If all ly recoverable reserve is 5.2, not 7, billion LILIENFELD REPLIES: Despite thousands of the effort at developing high-tech means barrels; at $18, about 2.4, not 5, billion studies conducted on the Rorschach, only a of control were redirected at educating barrels; and at $15, zero, not a few hun- handful of indices have received consistent everyone about known control meth- dred million barrels. Both ways of con- empirical support. Merlo and Barnett are cor- ods—avoiding cross-contamination and sidering recoverable reserves are legiti- rect that the Rorschach should not be used cooking the inside of burgers to 160 de- mate, but I think the lower figure, count- in isolation when making clinical decisions, grees or higher (“cook the pink out”)— ing finding costs, is the right one for the but they erroneously assume that adding the E. coli problem we have now might public policy decision. the Rorschach to existing test information just go away. AMORY B. LOVINS necessarily increases validity. In fact, in sev- WINKLER G. WEINBERG Chief Executive Officer (Research) eral studies validity decreased when clini- Section Chief, Infectious Disease Service Rocky Mountain Institute cians with access to other test information Kaiser Permanente Snowmass, Colo. were provided with Rorschach data. Atlanta There is little evidence that Rorschach The question before us appears to be data contribute statistically to the assess- SUPERCAVITATION, SWIMMINGLY whether our rapacious appetite for oil will ment of personality or mental illness beyond In reading “Warp Drive Underwater,” lead to the destruction of vast expanses of questionnaire data. Although the Rorschach by Steven Ashley, I tripped over the state- untouched wilderness, an irreplaceable yields immensely detailed and complex data, ment that “swimming laps entirely under- sanctuary for polar bears, white wolves we should not make the mistake of assuming water is even more difficult” than swim- and caribou. For 20,000 years, the native that these data are necessarily valid or useful. ming on the water’s surface. Actually, it Gwich’in people have inhabited this sa- is much easier to swim several laps under- cred place, following the caribou herd and SEMANTIC WEB: NOT FUZZY water with a single breath than it is to leaving the awe-inspiring landscape just as What struck me as a curious omission in swim on the surface, because the air/wa- they found it. For the sake of future gen- “The Semantic Web,” by Tim Berners- ter boundary friction is even greater than erations, I hope the answer is no to Lee, James Hendler and Ora Lassila, was the water friction. In fact, competitive drilling, despite advances in technology. any mention of the notion of fuzzy logic. swimming rules for years stipulated that ROBERT REDFORD Even today’s relatively primitive search swimmers could not become entirely sub- Sundance, Utah engines attempt to rate the relevance of merged during the breaststroke. hits to the supplied search terms. JON TOBEY ALL ABOUT INKBLOTS ROB LEWIS Monroe, Wash. Scott O. Lilienfeld, James M. Wood and Langley, Wash. Howard N. Garb [“What’s Wrong with ERRATUM To determine how far you’ve dri- This Picture?”] do not present a balanced BERNERS-LEE REPLIES: I deliberately didn’t ven, multiply (not divide, as was incorrectly analysis of the Rorschach; they overem- mention fuzzy logic, as fuzzy logic itself does stated in “Rip Van Twinkle,” by Brian C. Chaboy- phasize studies that do not support the not work for the Web. I see fuzzy logic and oth- er) the fuel supply by the gas mileage.

www.sciam.com SCIENTIFIC AMERICAN 13 Copyright 2001 Scientific American, Inc. 50, 100 & 150 Years Ago

Salmon Sense Okapi Surprise Yankee Ingenuity

SEPTEMBER 1951 tant factor in our civilization. There are ed at the present time only by the giraffe HOW SALMON GET HOME—“An explana- 400,000 of them in the U.S., and engi- and the prong-horned antelope, so-called, tion of one of the most engaging phe- neering is now our third-largest profes- of North America. So far as it can be as- nomena of nature—the salmon’s return sion, exceeded only by teaching and certained, the okapi is a living represen- from hundreds of miles at sea to its native nursing. However, engineers are in acute- tative of the Hellatotherium genus, which creek—has been proposed by Arthur ly short supply, and the number of grad- is represented by an extinct form found Hasler and Warren Wisby of the Univer- uates in the next few years will be far fossilized in Greece and Asia Minor. The sity of Wisconsin. They believe that the short of the need for new engineers.” animal is about the size of a large ox. The fish can smell its way back home. ‘It ap- coloration is, perhaps, unique among pears that substances in the water, prob- SEPTEMBER 1901 mammals. The body is of a reddish color, ably coming from the vegetation and OXYGEN FOR AERONAUTS—“An appara- the hair is short and extremely glossy. soils in the area through which the tus has been devised by a Frenchman, Only the legs and hind quarter of the an- stream runs, give each stream an odor Louis-Paul Cailletet, for the purpose of imal appear to be striped.” which salmon can smell, remember and supplying aeronauts with pure oxygen recognize even after a long period of non- when poised at a high altitude, where the SEPTEMBER 1851 exposure.’ The damming of many Pacif- extreme rarefaction of the air renders MECHANICAL REAPER—“The Illustrated ic Coast salmon streams is resulting in a them liable to asphyxiation. When the News, London, says: ‘McCormick’s reap- progressive decline of the yearly catch, as aeronauts experience the nausea arising ing machine had a fair trial, at the annual huge numbers of salmon batter them- from rarefied air, they have recourse to gathering at Mechi’s farm. It rained in tor- selves to death trying to get over the an oxygen supply. His device consists of rents, and mud and wet straw soon dams and back home.” a double glass bottle containing liquid clogged the other instruments; but we oxygen. From the reservoir extends a have the authority of, among others, Mr. ENGINEERS—“In 1850 a little more than flexible tube communicating with a small Fisher Hobbes, the well-known agricul- 5 per cent of America’s industrial power metal mask covered externally with vel- turalist, for stating that McCormick’s ma- was supplied by machines; 79 per cent vet to protect it from the cold.” chine performed its work perfectly, and was furnished by animals and 15 per cent proved itself one of the most valuable agri- by human muscles. Today 84 per cent of THE OKAPI DISCOVERED—“Sir Harry cultural inventions of the age. It has ar- our power is supplied by machines and Johnston’s discoveries in Uganda are of rived at the fortunate period, when the only 12 per cent by animals and 4 per great importance. One of the new animals steady emigration of Irish laborers threat- cent by men. As a consequence the engi- which he found was the ‘Okapi.’ It be- ens to leave our farmers short of hands at neer has become an increasingly impor- longs to a group of ruminants represent- every harvest. The proprietor will be ready to bear witness that he found no impedi- ments from British jealousy, and that his success was hailed with as much enthusi- asm as the damp weather would allow.’” [Editors’ note: Cyrus McCormick is con- sidered to be the inventor of the first suc- cessful mechanical reaping machine.]

OUR EARLY YEARS—“From small begin- nings, six years ago, the Scientific Amer- ican has attained a very honorable posi- tion in point of circulation, and conse- quent influence and usefulness. No man can spend two dollars to better advan- tage than by subscribing for it. We may confidently expect over 20,000 patrons SCIENTIFIC AMERICAN THE ENGINEER, 1951: pen, ink, French curve, cigarette to our new volume. The more we have to

feed, the better fare we will serve you.” FROM

16 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. news

PHYSICS SCAN SNO Nus Is Good News THE LATEST ON MUTATING NEUTRINOS SOLVES THE SOLAR NEUTRINO PROBLEM BY GRAHAM P. COLLINS

elltale flashes of light within a 1,000- has been counting neutrinos in ordinary wa- ton sphere of ultrapure heavy water, ter for about five years. As Super-K member deep underground in a nickel mine near Edward Kearns of Boston University ex- TSudbury, Ontario, have resolved a 33-year- plains, “Although SNO is 10 times smaller, old puzzle. In June the Sudbury Neutrino Ob- because the deuterium reaction is pretty servatory (SNO) collaboration announced strong, it has a comparable total event rate as firm evidence that elusive ghostly particles Super-Kamiokande.” called neutrinos morph from one subspecies More important than the gross numbers, to another during their flight from the sun to however, is the variety of interactions possi- Earth. The result reassures ble at SNO, giving the detector new ways to astrophysicists that their pre- distinguish subspecies, or flavors, of solar neu- cision solar models do not trinos. In both heavy and ordinary water, a contain a lurking blunder, neutrino can hit an electron, sending it ca- and it gives particle physi- reering through the liquid fast enough to pro- cists further clues to what duce a flash of Cherenkov radiation. But such lies beyond their beloved but electron scattering can be caused by any of the incomplete Standard Model three neutrino flavors: the tau-neutrino, the of particle physics. muon-neutrino and the electron-neutrino. The mystery of solar neu- (The sun’s nuclear reactions produce electron- trinos has haunted physicists neutrinos exclusively.) SNO’s heavy water can since 1968, when the first ex- single out electron-neutrinos, because that fla- periment to count those neu- vor alone can be absorbed by a deuterium nu- trinos came up with less than cleus, transforming it into two protons and an half of the expected number. electron that fly apart at high energy. Three decades of experiments SNO’s count of just the electron-neutrinos and more refined theories is lower than Super-K’s count of all flavors. TO SEE NEUTRINOS from the sun, the underground Sudbury Neutrino have only confirmed the discrepancy. The conclusion: some of the sun’s electron- Observatory (SNO), shown here The SNO project is unique in that it uses neutrinos turn into muon- or tau-neutrinos. during construction, relies on 9,456 heavy water, containing the deuterium iso- Because muon- and tau-neutrinos scatter elec- sensors to monitor a 1,000-ton tope of hydrogen, to observe neutrinos (de- trons much less efficiently than electron-neu- sphere of water. Physicists use the noted by the Greek letter “nu”). A similar de- trinos do, the small excess of scatterings at Su- Greek letter “nu” to denote neutrinos.

tector, Super-Kamiokande in Kamioka, Japan, per-K translates into a large number of muon- LAWRENCE BERKELEY NATIONAL LABORATORY

18 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. news or tau-neutrinos present. Two thirds of the neutrino masses as something to be explained electron-neutrinos from the sun are trans- by whichever theory supersedes the Standard SCAN formed by the time they reach Earth—pre- Model. The masses are tiny: the SNO results, cisely the right number to agree with the pre- combined with other data, imply that all three TO COME: dictions made using solar models. Arthur B. neutrino flavors have masses less than 1⁄ 180,000 MORE NU NEWS McDonald, head of the SNO project, says of an electron mass, the next heavier particle. In the next few months that the result “provides a very good confir- SNO’s results put sterile neutrinos—a pe- SNO researchers will: mation that we understand how the sun is culiar hypothetical fourth flavor—on shakier generating energy with great accuracy.” ground. All the observed SNO and Super-K Determine whether the amount Changes, or oscillations, of neutrino fla- data can be explained by electron-muon-tau of solar neutrino oscillation vors can occur only between flavors having oscillations. Contributions by sterile neutri- varies according to the time of unequal masses. In essence, a subtle mass dif- nos are not yet ruled out, “but the fraction of day and the season. Such variations, not discerned at ference causes the quantum waves of two dif- sterile neutrinos is not expected to be large,” Super-Kamiokande, would result ferent flavors to oscillate in and out of sync McDonald says. from extra oscillations when with each other, like two close musical notes The SNO experiment will continue for a neutrinos pass through Earth or producing beats. If the peak of each beat cor- few more years. Since June the detector has because of Earth’s varying responds to an electron-neutrino, then the been running with ultrapure salt (sodium distance from the sun. minimum in each cycle is, say, a muon-neu- chloride) added to the heavy water. The chlo- See if the oscillation depends on trino. The first strong evidence for such os- rine atoms in the salt greatly enhance the de- the neutrinos’ energy. The data cillations came in 1998 from Super-K’s study tection of neutrons, produced when neutrinos will further constrain the of high-energy cosmic-ray neutrinos. split a deuterium nucleus without being ab- neutrino masses and the so- In the simplest version of the Standard sorbed. Accurate counts of those reactions called mixing angle (which defines how much oscillation can Model of particle physics, neutrinos are mass- should be a recipe for further tasty results occur) and might eliminate some less and cannot oscillate. Most theorists view about neutrinos. oscillation theories.

ONCOLOGY Cancer in the Crosshairs WHY SOME TUMORS WITHSTAND GLEEVEC’S TARGETED ASSAULT BY DIANE MARTINDALE

pproved for use this past May, Gleevec CML is caused when chromosomes 9 and was celebrated as the first in a wave of 22 swap genes. This produces a mutation in supremely effective cancer drugs that the Abl protein, a type of internal signaling homeA in on tumors without harming healthy enzyme known as a tyrosine kinase that is in- cells. In clinical trials, 90 percent of patients volved in normal cell growth. Once mutated, in early stages of chronic myelogenous however, the Abl protein becomes hyperac- leukemia (CML), a rare blood cancer, went tive and drives white blood cells to divide in- into remission within six months of first tak- cessantly. In either form, Abl needs to bind a ing Gleevec. Despite its promise, however, molecule of a cell’s ATP to function. Gleevec Gleevec and drugs like it may be only a pro- works by docking into the pocket ordinarily visional measure. More recent reports have occupied by ATP, thereby stopping the func- found that patients in late-stage CML relapse tion of the signaling protein. The CML cells because their tumors become drug-resistant. then pack up and die. Gleevec, made by Novartis, belongs to a But why are CML cells the only ones class of drugs called small molecules. They killed, given that there are hundreds of other

The Oregonian/Corbis/Sygma are designed to either target specific receptors tyrosine kinases and similar enzymes that rely on cancer cells or disrupt their signaling path- on ATP for their activation? Why doesn’t ways, thereby marking a major departure Gleevec block these? Years ago scientists be- REASON TO SMILE: Gleevec’s pioneer from radiation and chemotherapy, the broad lieved that all ATP binding sites were identi- Brian Druker at a May press conference announcing the cancer drug’s approval.

PATRICIA CORDELL effects of which can be toxic to healthy cells. cal. Actually, they are all slightly different,

www.sciam.com SCIENTIFIC AMERICAN 19 Copyright 2001 Scientific American, Inc. says Brian J. Druker, a cancer researcher at cells are genetically malleable,” Hunter notes, news Oregon Health Sciences University who helped “and they find ways to escape, no matter how to develop Gleevec: “This was unpredicted, clever we think we’ve been.” SCAN but good for the targeted approach.” Small-molecule therapy is being viewed as Moreover, such molecules strike a blow at a treatment, not a cure, Hunter adds. In some SMALL MOLECULES’ tumor cells’ weak spot: their almost exclusive patients, particularly those in advanced stages dependence on the mutated protein. Blocking of disease, the drugs may work only to keep BIG BROTHER the errant Abl cuts off the tumor cells’ lifeline, the tumor in check, transforming it into a Small-molecule drugs such as but normal cells are unaffected because they chronic condition, much as diabetes is. Gleevec are not the only targeted possess redundancy in protein function; their Still, Gleevec is a milestone not just for therapy. Another is monoclonal backup pathways kick in, Druker explains. what it does but for the revolutionary strate- antibodies, which are created from The targeted approach offers much hope— gy it represents, says Larry Norton, head of a single cell and are designed to respond to a specific antigen. In Gleevec also blocks protein receptors in two solid-tumor oncology at Memorial Sloan-Ket- cancer therapy, most block growth other malignancies—but it is no magic bullet, tering Cancer Center in New York City. He factors from activating cell warns Tony Hunter, a molecular biologist at predicts that because of the new understand- division. For example, the the Salk Institute for Biological Studies in San ing of tumors at the molecular level, cancers monoclonal antibody Herceptin Diego. Gleevec has been used for a short time will soon be classified by their molecular targets an epidermal growth factor on certain types of breast cancer (its approval came after clinical trials that be- makeup rather than by their location in the cells. Although early results are gan in 1998), and some patients have already body, as is done today. Researchers might encouraging, monoclonal antibodies developed resistance to it—the Abl protein’s then build molecules to attack them—a tall or- are not nearly as impressive as ATP pocket mutates so the drug can no longer der, Norton says, “but one that is possible.” Gleevec. Moreover, antibodies will bind to it. In other instances, production of bind to the same receptors on normal cells, thereby causing more the Abl protein is so great that the drug—even Diane Martindale is a science writer based severe side effects. at the highest dose—cannot keep up. “Cancer in New York City.

GEOSCIENCE Triggering a Snowball DID METHANE ADDICTION SET OFF EARTH’S GREATEST ICE AGES? BY SARAH SIMPSON

DINBURGH, SCOTLAND—Once upon a hotly debated the details of the events, which time ice as much as a kilometer thick en- occurred as many as four times between 750 gulfed the earth. Glaciers scoured the million and 580 million years ago, in a time Enearly lifeless continents, and sea ice encap- known as the Neoproterozoic. But just how sulated the oceans—even in the tropics. The the earth first plunged into a snowball-style ice planet’s only solution to the deep freeze was age has been unknown. Now the original pur- to wait for volcanoes to release enough heat- veyors of the snowball earth hypothesis have trapping dioxide proposed a trigger: methane addiction. to create a runaway green- At a scientific conference in Edinburgh house effect. A brutal epi- this past June, geochemist Daniel P. Schrag sode of warming ensued, described the addiction scenario: Just before not only melting the ice the first snowball episode, the ancient earth but also baking the planet. kept warm by relying on methane—a green- Three years ago a house gas 60 times more powerful than car- group of Harvard Univer- bon dioxide. The methane had begun leaking sity researchers proposed slowly into the atmosphere when massive, icy this revolutionary idea— methane hydrate deposits within the seafloor Corbis dubbed the snowball earth became destabilized somehow. As a result, hypothesis—about the po- carbon dioxide levels decreased, and methane tential of the planet for became the world’s dominant greenhouse gas. MASSIVE GLACIERS entombed the earth hundreds severe climate reversals. The trouble with the planet’s dependence of millions of years ago, but how?

Since then, scientists have on methane is that the gas disappears quickly WOLFGANG KAEHLER

20 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. news in an oxygen-rich atmosphere. An interrup- in the oceans can also cause carbon isotope SCAN tion in the methane leak left the earth in dire values to fall and remain low over periods of need of greenhouse gases, and the climate a million years or so. But based on estimates tumbled into a deep freeze before volcanoes of sedimentation rates of the rocks in ques- could release enough carbon dioxide to make tion, Schrag and others think the crash could up for the lost methane. “I’m not sure I total- have occurred over a shorter time frame. Still, ly buy this idea—it’s outrageous,” Schrag ad- Kaufman is skeptical: “We don’t have any THE DISTANT PAST’S mitted at the conclusion of his presentation. way to look into the rock record and see a ELUSIVE CLUES “But it’s the only idea that explains the carbon methane buildup.” Information about the chemical isotope crash just before the glaciations.” “It’s difficult to test anything that old,” makeup of the ancient atmosphere Such bizarre drops in carbon isotope val- Schrag says. But you can look for carbon iso- can sometimes be gleaned from the ues have been recorded in the rocks beneath tope crashes in places where their duration remains of bacteria trapped in the jumbled layers deposited by the glaciers of may be more certain, he adds. Several work- rocks. If the number of fossilized snowball events at several locations around ers have already correlated the carbon iso- methane-loving bacteria is unusually high, for instance, the world. But as provocative as the methane tope values among Neoproterozoic rocks in scientists can reasonably assume addiction hypothesis may be, it left the con- Namibia, Australia, California, Canada and that the atmosphere was rich in ference audience with many questions. the Arctic islands of Svalbard. Dozens of oth- that particular gas. But rocks from Alan Jay Kaufman, a University of Mary- er deposits of similar age exist but have not yet snowball earth times have land geochemist who first measured some of been analyzed. The potential triggers of a experienced the heat and pressure of metamorphism, which have long the carbon isotope crashes, pointed out that a snowball earth, it seems, may be as contro- since destroyed any biological cells. dramatic decrease in biological productivity versial as the details of the event itself.

CONSERVATION Into the Jaguar’s Den HUNTING AS A MEANS TO PRESERVE THE JUNGLE’S FOREMOST PREDATOR BY ERIC NIILER AOBAS, MEXICO—Hunting jaguars is all Mexico’s Quintana Roo state in the Yucatán about keeping your head low while Peninsula. It’s a several-million-acre patch- — TAKEN DOWN: Wildlife veterinarian running through the jungle the better work of rain forest between two protected Marcela Araiza checks a darted jaguar’s Cto duck spike-covered vines and saw-toothed national parks that is rapidly being cut down health and genetic information. palm leaves that infest this remote area of by farmers, ranchers and small-scale loggers. As the forest goes, so does habitat for this re- gion’s most charismatic animal: Panthera onca, known locally as el tigre. Catching the nocturnal jaguars means en- tering the jungle in the early hours before dawn; otherwise, heat and sunlight evaporate the cat’s scent trail. After alternately running and stumbling for two hours, the hunting par- ty stops and listens to the sound of barking dogs. “They’ve treed him,” concludes Tony Rivera, who is leading us. “That’s it!” Two trackers with machetes run ahead. Soon we see a female jaguar, weighing perhaps 70 pounds, nestled 30 feet above the forest floor. It gazes impassively at us with huge brown-yellow eyes. The jaguar’s beauti- ful golden, black-spotted fur keeps it well hid-

den in the forest canopy. Mayan kings wore ERIC NIILER

22 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. news jaguar skins as battle tunics; modern-day frighten animals and could lead them to injure poachers prize them as proof of vanquishing themselves. As for the jaguars, Alan Rabino- SCAN the jungle’s most powerful predator. witz, director of the New York City–based Joe Bojalad, a big-game hunter, rests for Wildlife Conservation Society’s Global Car- a minute and squints through the rifle scope. nivore Program, claims that science isn’t the Hunters like him have paid upward of $5,000 top priority. “The project is being driven from for this opportunity. He steadies the rifle, a hunting perspective,” says Rabinowitz, who pulls the trigger and pfffftt—the shot misses. has visited the Mexican project. “Don’t tell “It didn’t go in?” he asks. In this case, big- me it’s a scientific project.” game hunting does not mean killing. Bojalad The jaguar project’s principal investigator, has an air-powered rifle that fires tranquiliz- UNAM’s Gerardo Ceballos Gonzales, de- er darts. Once asleep, the jaguar will be ex- fends the program. He says that its early prob- amined and tagged. lems with U.S. hunters—their wish to shoot Scientists don’t know exactly how many already collared animals, for example—have jaguars survive in the tropical forests of Latin been corrected with stricter protocols. America. The cats once flourished from the According to Ceballos, southern U.S. all the way to Argentina. To- the project plans to capture CUBA day only a few significant populations re- eight to 10 jaguars inside Gulf of Mexico main, mostly in Mexico, Belize and Brazil. Dart the nearby Calakmul Bio- hunting is part of a unique program sponsored sphere Reserve and another Yucatán by Unidos para la Conservación, a Mexico 10 in the mixed forest and Peninsula MEXICO Jaguar City–based conservation group, the National farming country around it Caobas Study Site Autonomous University of Mexico (UNAM), in the next year. Close to 20 and Safari Club International, a U.S.-based animals have been collared BELIZE outfit that recruits American big-game hunt- since the project began; the GUATEMALA HONDURAS ers to fund conservation and research. A team is currently tracking Pacific Ocean Mexican cement company and a cellular- five animals. EL SALVADOR NICARAGUA phone firm donate additional funds to the Once the collaring pro- project, which has been run by Unidos para la gram is completed, Ceballos plans to set up a YUCATÁN FORESTS harbor Conservación since 1997. network of motion-detecting cameras across one of the few remaining populations of jaguars. Twice again, Bojalad fires. But no luck. trails. These cameras will record both preda- The gun is passed to one of the trackers, who tor and prey and will help determine the size shimmies up a nearby tree to get a better an- of the jaguar population, now estimated at gle, takes aim and nails the jaguar. Disorient- 400 to 500 within the reserve. Outside ed by the ketamine tranquilizer, the cat scram- Brazil’s Pantanal region, this may be the bles down the tree and stumbles through the world’s largest jaguar population. “The hope WHEN ANIMALS ARE jungle with three dogs at its heels before plop- is that we can have a sustainable population ping over. of jaguars not only inside but outside the re- EASY PICKINGS For the next 45 minutes, Cuauhtemoc serve,” Ceballos remarks. One of the controversial aspects of Chavez, a graduate student from Mexico Yet Rabinowitz and others also worry the dart hunting of jaguars is the City, and Marcela Araiza, a wildlife veteri- about the presence of guides such as the 50- use of live bait: goats are tied up at night to entice the predator, narian, take skin, muscle and blood samples. year-old Rivera, a former poacher. In the late à la Jurassic Park, says Marcelo They remove more than 30 botfly larvae that 1980s Rivera ran afoul of the U.S. Fish and Aranda, a biologist at Mexico’s have burrowed into the animal’s hide and Wildlife Service for transporting jaguar skins Ecology Institute in Veracruz. then attach a GPS radio collar. Over the next into the U.S. Rivera says he’s changed his Perhaps as a result of the practice, year and a half, the collar will record the an- spots and no longer kills jaguars, although he one jaguar that was collared and imal’s location. believes there are enough jaguars in this part released in April was caught at a nearby farm in late May after it had Not everyone is convinced that dart hunt- of Mexico to sustain limited hunting. killed four sheep, a cow and a young ing is a good idea. In Africa, where rhinos, The problem here is not illegal hunting or calf. The 150-pound male was warthogs and other game are targeted, some the possibility of its return but rather defor- relocated to the Sian Ka’an conservationists have complained that ani- estation, according to Carlos Manterola, di- Biosphere Reserve, about 100 miles mals often get darted more than once, poten- rector of Unidos para la Conservación. Man- away. Project defenders say that the problem jaguar was an tially resulting in tranquilizer poisoning, and terola recently began working with local anomaly—so far none of the other

SARA CHEN that partial injections from poor shots merely community landowners to start a mahogany- collared jags has attacked livestock.

www.sciam.com SCIENTIFIC AMERICAN 23 Copyright 2001 Scientific American, Inc. news furniture workshop that will boost the value of the region’s timber and perhaps slow its SCAN cutting. He also pays local farmers if the jaguars eat livestock, an insurance policy de- signed to cut poaching. “This is not just a sci- entific project for jaguars,” Manterola insists. “We are using the jaguar as a flagship species for conservation of the Mayan jungles.”

TAKING AIM at a treed jaguar is Joe Bojalad; Tony Eric Niiler, based in San Diego, writes Rivera (center) and Carlos Manterola (right) look on.

frequently about conservation issues. ERIC NIILER

PRIVACY Surveillance by Design WILL A NEW CYBERLAW BYPASS THE U.S. CONSTITUTION? BY WENDY M. GROSSMAN FOR ME BUT NOT FOR YOU? ONDON—A legislative move in Europe pseudonymous ID. Users had to trust Hel- that would also affect the U.S. is threat- singius. Many of today’s services and soft- As counterculturalist Stewart Brand ening the sometimes controversial abil- ware, such as the Dublin-based Hushmail and has said, anonymity can be toxic Lity of Internet users to mask their real-world the Canadian company Zero Knowledge’s to community, because it can identities. The move, which is heavily backed Freedom software, keep no logs whatsoever. foster irresponsible activity and sow mistrust. But an experiment by the U.S. Department of Justice, is the cy- But if the cybercrime treaty is ratified, will some years back on the WELL, the bercrime treaty, designed to make life easy for they still be able to? Would they have to move San Francisco–based electronic law enforcement by requiring Internet service beyond the reach of the law to, say, Anguilla? conferencing system, showed that providers (ISPs) to maintain logs of users’ ac- More than that, will the First Amendment people typically wanted anonymity tivities for up to seven years and to keep their continue to protect us if anonymity is effec- for themselves—just not for others. networks tappable. The Council of Europe, tively illegal everywhere else? Says Mike God- a treaty-building body, announced its support win, perhaps the leading legal specialist in civ- of the cybercrime effort in June. il liberties in cyberspace: “I think it becomes a Anonymity is a two-edged sword. It does lot harder for the U.S. to maintain protection enable criminals to hide their activities. But it if the cybercrime treaty passes.” Godwin calls PRECEDENTS is also critical for legitimate citizens: whistle- the attempt to pass the cybercrime treaty FOR PRIVACY blowers, political activists, those pursuing al- “policy laundering”—a way of using interna- In the U.S., the right to be ternative lifestyles, and entrepreneurs who tional agreements to bring in legislation that anonymous is protected under the want to acquire technical information with- would almost certainly be struck down by First and Fourth amendments. out tipping off their competitors. U.S. courts. (On its Web site, the U.S. De- According to Mike Godwin, author of Even without the proposed legislation, partment of Justice explains that no support- Cyber Rights, two significant anonymity is increasingly fragile on the Net. ing domestic legislation would be required.) Supreme Court cases establish the precedents. Corporations have sued for libel to force ser- In real-world terms, the equivalent of the vices to disclose the identities of those who treaty would be requiring valid return ad- NAACP v. Alabama The 1958 ruling posted disparaging comments about them dresses on all postal mail, installing cameras upheld the NAACP’s refusal to online. Individual suits of this type are rarer, in all phone booths and making all cash disclose its membership lists on but last December, Samuel D. Graham, a for- traceable. People would resist such a regime, the grounds that this type of privacy is part of the right to mer professor of urology at Emory Universi- but surveillance by design in the electronic freedom of assembly. ty, won a libel judgment against a Yahoo user world seems less unacceptable, perhaps be- whose identity was released under subpoena. cause for some people e-mail still seems op- McIntyre v. Ohio Elections Services designed to give users anonymi- tional and the Internet is a mysterious, dark Commission The 1995 ruling struck ty sprung up as early as 1993, when Julf force that is inherently untrustworthy. down a requirement, instituted to control campaign spending, that Helsingius founded Finland’s anon.penet.fi, Because ISPs must keep those logs and political pamphlets include the which stripped e-mail and Usenet postings of that data, your associations would become name and address of their issuer. identifying information and substituted a an open book. “The modern generation of

24 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. traffic-analysis software not only can link to read. Whatever privacy regime we create now news conventional police databases but can give a will almost certainly wind up controlling the comprehensive picture of a person’s lifestyle bulk of our communications. Think careful- SCAN and communications profile,” says Simon ly before you nod to the mantra commonly Davies, director of Privacy International. “It heard in Europe at the moment: “If you have can automatically generate profiles of thou- nothing to hide, you have nothing to fear.” sands of users in seconds and accurately cal- Do you really want your medical records sent culate friendship trees.” on the electronic equivalent of postcards? In the not too distant future, nearly every- thing that is on hard copy today will travel via Wendy M. Grossman, who writes about e-mail and the Web, from our medical records cyberspace issues from London, is also on to the music we listen to and the books we the board of Privacy International.

DISPLAYS Eye Spy FORGET MONITORS—NOMAD PUTS TEXT AND GRAPHICS RIGHT ONTO THE RETINA BY PHIL SCOTT att Nichols, director of communica- ditions. “It also means that in a tions for Microvision, has just ar- heavy-vibration environment— rived from his crosstown walk. He’s the thump-thump-thumping of Mhurried from New York City’s upcoming Mu- the rotors—the image remains seum of Sex, where he showed off the same stabilized in space,” Lippert adds. equipment that he wants to demonstrate now. “It doesn’t bounce around like a No, please—it’s called Nomad, a retinal scan- bad newsreel.” Other manufac- ning device that can beam words and graph- turers have built head-mounted ics directly into the viewer’s eye. “They’re try- displays in the past, but Nomad ing to figure out how to create a fully interac- superimposes images without tive museum,” Nichols sheepishly explains. blocking the view (the images are The U.S. Army is also interested in No- in outline form). mad, for a less titillating function: equipping In late June the company conducted No- PILOT’S-EYE VIEW using Nomad would show an overlaid flight path. its helicopter pilots. When coupled with the mad’s first flight tests; eventually, Nichols proper software, the headset can display alti- states, the system will incorporate a voice-op- tude, heading, speed, course and weapons sta- erated computer. Microvision intends to sell tus, all presented in a nice monochrome light a commercial version as early as this fall for beam that doesn’t hamper the pilot’s view. “It between $8,000 and $10,000. projects the image desired into the visual field Helicopter pilots aren’t the only ones who of the pilot’s eye, and that image is seen at op- might take advantage of the technology, tical infinity,” says Thomas Lippert, chief sci- Nichols adds. Air-traffic controllers could NEED TO KNOW: entist at Microvision, based in Bothell, Wash. watch airplanes while their headsets broad- INSIDE VIEW “That means the pilot can keep his gaze out cast the flight data. Surgeons might have a pa- The heart of the Nomad technology the windscreen while keeping this augmented tient’s medical images alongside real-time lies in a microelectromechanical information sharply in focus.” readouts of vital signs. Firefighters could en- system (MEMS)—specifically, a Such technology could replace head-up ter smoke-filled buildings with overlaid room MEMS light-beam scanner, which displays on windscreens and virtual-reality diagrams. The system has even allowed pa- has a tiny mirror 1.5 millimeters wide. A red laser diode bounces a helmets—a goal of the U.S. Air Force for tients with macular degeneration to read once pulsed beam off the MEMS mirror, decades. Nomad is lightweight (the produc- again. And of course, there’s the Museum of which uses an electromagnetic tion version could weigh in at about one Sex. But we’ll let you fill in your own uses for system to move in two directions, pound), and because it is worn, it will swivel the headset there. creating a scanning pattern similar with the pilot’s head. That makes it ideal for to those on television screens. An optical combiner modifies the helicopters, which are inherently unstable Phil Scott is a science and technology writer beam to create an image 800

MICROVISION, INC. and difficult to fly under instrument-only con- in New York City. pixels wide by 600 pixels long.

www.sciam.com SCIENTIFIC AMERICAN 25 Copyright 2001 Scientific American, Inc. news SCAN ASTRONOMY Moons over Saturn Saturn’s family has gotten bigger. Re- searchers using 11 different telescopes around the world have reported finding 12 new moons, ranging from six to 32 kilometers in diameter and orbiting Saturn MOST MOONS, so far. DATA POINTS: in highly eccentric paths, quite unlike the A DEMON-HAUNTED generally circular orbits of its major moons, such as Titan. These tiny moons seem to be clus- WORLD tered in groups of three or four, suggesting that they are remnants of larger bodies that were fractured, probably by collisions, early in the planet’s formation. Such irregular bodies may be common for the gas giants; in the past few years astronomers have found five irregulars around Uranus and 12 around Jupiter. The latest finding, in the July 12 Nature, brings Sat- urn’s satellite brood to 30: six major moons and 24 minor ones. —Philip Yam

COMMUNICATIONS Built for Speed PSYCHOLOGY IBM attested in June to building the world’s fastest You Forgot silicon-based transistor. Potentially able to operate at 210 gigahertz, this silicon germanium transistor to Remember performs 80 percent faster than previous technol- Recent studies illustrate how easy it is ogy, breaking the 200-GHz speed barrier thought to create false memories. Jacquie Pick- Belief in some paranormal phenomena is on the rise in the U.S. to exist for silicon-based transistors. Transistor rell and Elizabeth Loftus of the Uni- speed depends largely on the distance electricity versity of Washington reported at a re- Percent Percent must travel within the device, and IBM researchers cent American Psychological Society who change believe from were able to shrink conference in Toronto that many visi- 1990 UNFILLED FIBER this distance in so- tors to Disneyland concluded that they called heterojunction had met Bugs Bunny, a Warner Bros. Psychic or 54 +8 spiritual healing bipolar transistors, character ordinarily not seen at the in which electrons happiest place on Earth. The psychol- Extrasensory 50 +1 flow along a vertical ogists had subjects read fake Disney perception path rather than tak- ads, some of which mentioned Bugs— Haunted houses 42 +13 ing the horizontal and sometimes in the presence of a card- route in convention- board cutout of him. About one third

Ghosts or spirits 38 +13 ); ILLUSTRATION BY MATT COLLINS al transistors. IBM of those later exposed to Bugs believed

Telepathy 36 0 expects that within that they had encountered and even bottom ( two years the tran- shook hands with the rascally rabbit at Extraterrestrials 33 +6 have visited the earth sistors will drive Disneyland. Other researchers found Corbis chips used in com- that people can fabricate false images Astrology 28 +3 munications equipment to 100 GHz—five times because of “causal inference.” Viewing Communication 28 +10 faster than today’s chips. (The transistors, though, the result of an action (oranges on the with the dead are incompatible with computer processors.) The ground, say), people will recall spotting little super-silicon transistors still have a way to go the cause (someone reaching for an or- Reincarnation 25 +4 before they can switch quickly enough to keep up ange at the bottom to upset the stack) ); ROBERT YOUNG PELTON top Percent change in the Nielsen ratings with the theoretical limit of fiber-optic communica- even if they never saw such an image. of The X-Files from its peak season: –29.8 tions. In the June 28 Nature, researchers at Lucent The study, which helps to explain er-

SOURCES: Gallup Organization; Nielsen Media Technologies calculated the limit to be approxi- rors in eyewitness accounts, is available Research, 2000–01 season average compared mately 100 terabits per strand of fiber. Current data at www.apa.org/journals/xlm/press_ with 1997–98 average. Psychic-healing category includes belief in the power of the transmission rates run as high as 1.6 terabits per sec- releases/july_2001/xlm274931.html

mind to heal the body. ond over a single strand. —Mariama Orange —Philip Yam R. BEEBE (NMSU) NASA (

26 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. news MEDICINE EVOLUTION SCAN Peace in the Infectious Selection Infectious disease can be a powerful driving force Nonobese in evolution. Recently scientists led by Sarah A simple treatment might one day re- Tishkoff of the University of Maryland found that lieve type 1 diabetics of daily finger changes in the frequency of certain forms, or alleles, pricking and insulin injections. In type of the gene for glucose-6-phosphate dehydrogenase 1 diabetes, immune cells wage war on (G6PD) within human populations roughly mirror insulin-secreting islet cells in the pan- the history of malaria. Some creas, resulting in improper blood glu- mutations of the gene result cose levels. (Type 2 diabetes results in reduced activity of G6PD, from cells’ becoming insensitive to in- producing anemia and, sulin, often as a result of obesity.) more advantageously, mod- Working with what they call non- erate resistance to malaria. obese diabetic mice, researchers at Looking at the history of Massachusetts General Hospital re- the various forms of the trained the attacking immune system G6PD gene within popula- to ignore any surviving islet cells. The tions most affected by investigators first killed off the attack malaria, such as those in cells, which are abnormal, and then in- Africa and India, the re- MALARIA PARASITE jected normal immune cells from searchers discovered that attacks red blood cells. healthy donor mice. As a result, dia- the alleles that encode for betic mice began producing islet- G6PD deficiency arose during the same approxi- friendly immune cells. The approach mate time that malaria became more deadly. More- seemed to effect a permanent reversal: over, the alleles spread more rapidly within hard- WWW.SCIAM.COM/NEWS the glucose levels of some 75 percent hit populations than chance alone would suggest. BRIEF BITS of the mice returned to normal. The These results, in the July 20 Science, indicate that study appears in the July 1 Journal of selective pressure from malaria can maintain and In Ethiopia, researchers Clinical Investigation. promote otherwise deleterious alleles in the human uncovered the fossils of what appears to be humanity’s —Mariama Orange gene pool. —Alison McCook oldest ancestor, who is thought to have lived between 5.2 million and 5.8 million years ago. GEOCHEMISTRY /071201/3.html ) Scientists have created

bottom More Than Shade

( patterned glass surfaces on When trees first took over the continents about 380 million years ago, they changed the world which living nerve cells can wire

Corbis in an unexpected way. Researchers knew that trees absorbed much more carbon dioxide from themselves to electrodes; the the atmosphere than their moss and alga predecessors. But according to computer models of research may lead to better the long-term carbon cycle by Yale University geochemist Robert A. Berner and his colleagues, implants, prosthetics and biosensors. /071001/1.html as photosynthesis hit an unprecedented

); DAVID MUENCH high, the atmosphere also became twice as The friction-free superfluid top rich in oxygen as it is today—40 percent rel- helium 3 can act as a quantum ative to 21 percent. That means trees could gyroscope, producing a have been responsible for the evolution of whistling sound that gets louder or quieter depending on the gigantic insects, such as the dragonflies with orientation of the earth’s axis of 70-centimeter wingspans known to exist at rotation. /070901/3.html the time. An insect’s size depends in part on how much atmospheric oxygen is available A study has found that clones to diffuse passively into its body. Berner are not genetically identical to the donor animals; in fact, the presented the findings in June at the Earth clones sometimes exhibit System Processes conference in Edinburgh, TREES may have been harbingers of big bugs. dangerously different patterns of

MECKES/OTTAWA/PHOTO RESEARCHERS, INC. ( Scotland. —Sarah Simpson gene expression. /070601/1.html

www.sciam.com SCIENTIFIC AMERICAN 27 Copyright 2001 Scientific American, Inc. news SCAN BY THE NUMBERS Measuring Bad Behavior FBI CRIME STATISTICS: USE WITH CAUTION BY RODGER DOYLE

he Uniform Crime Reports (UCR), pro- To supplement the UCR, the Bureau of duced by the Federal Bureau of Investi- Justice Statistics in 1973 started an annual gation, have figured prominently in dis- survey of about 50,000 households designed Tcussions of crime since at least the Nixon era, to count the number of crime victims. Many but their reliability has long been suspect. A respondents did not correctly recall when a major reason is substantial underreporting. crime was committed, putting it in the wrong For a variety of reasons, including distrust of year, for example, and some even failed to re- law-enforcement officials, many crimes are call crimes in which they were known to be not reported to local police departments, the victims. Despite such limitations, however, source of the FBI data. Furthermore, the num- the National Crime Victimization Surveys, as ber of crimes that police departments report they are called, are a reasonably good guide can vary from year to year depending on bud- to overall crime trends, as shown by their gets. The FBI cannot legally enforce the coop- rough concordance with data on homicide, eration of local police departments and state the best recorded of the UCR categories. THE FBI’S agencies, and so it is not surprising that for The chart compares the UCR and victim- INDEX CRIMES several years in the 1990s, six states (the largest ization data in terms of serious violent crime. in terms of population was Illinois) supplied The UCR numbers are not only much lower Criminal homicide—excludes FBI attempted homicide no data, forcing the to estimate the num- because of incomplete reporting but are mis- ber of crimes in those states. leading as an indicator of violent crime trends Rape of women by force or Local police sometimes cook the books, because reporting improved over the past threat of force—excludes either underreporting to make crime in their several decades. The extent of the improve- statutory rape area appear to be under control or overre- ment is suggested by the growing conver- porting to support requests for more funding. gence of the UCR with the victimization sur- Robbery: taking by force or threat of force anything of value— Fabrication of this kind has presumably de- vey: In 1973 the number of violent crimes re- includes attempted robbery clined as police departments have become ported by the UCR totaled only 38 percent more publicly accountable in the past few of those reported by the survey but increased Aggravated assault: attack with decades, but it still persists, as recent reports of gradually to between 80 and 84 percent in the intent of inflicting severe bodily data manipulation in New York City, Phil- second half of the 1990s and is expected to injury, usually with a weapon— excludes simple assault (assault adelphia and Boca Raton, Fla., testify. rise further. Improved reporting, together without a weapon resulting with a newly introduced system that provides in little or no injury) 1,000 greater detail on crime incidents, has in- creased the usefulness of the UCR as an ana- Burglary: unlawful entry or attempted entry to commit lytic tool, but as an indicator of national 800 National Crime crime trends it still remains deficient. a felony or theft Victimization Survey The UCR covers eight types of violent and Larceny, theft: unlawful taking or property offenses—so-called index crimes— 600 attempted taking of property— but excludes others, such as drug violations, excludes motor vehicles simple assault, vandalism, prostitution, statu- Motor vehicle theft—includes 400 tory rape, child abuse, and white-collar of- attempted theft fenses such as embezzlement, stock fraud, forgery, counterfeiting and cybercrime. These Arson—includes 200 Uniform Crime Report types of infractions are excluded from the in-

attempted arson population) 100,000 (per Rate Crime Violent dex because they are not readily brought to The Bureau of Justice Statistics’s the attention of the police (for example, em- National Crime Victimization 0 bezzlement), are rare (kidnapping) or are not Surveys provide data on the same 1960 1970 1980 1990 2000 serious enough to warrant inclusion (disor- Year crimes except for homicide, arson, derly conduct). crimes committed against SOURCES: FBI and Bureau of Justice Statistics. Violent commercial establishments and crime statistics shown above include robbery (except of businesses), rape and aggravated assault and exclude Rodger Doyle can be reached at crimes against children crimes against children younger than 12. Victimization

younger than 12. data exclude incidents not reported to the police. [email protected] RODGER DOYLE

28 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Skeptic

Nano Nonsense and Cryonics True believers seek redemption from the sin of death By MICHAEL SHERMER

Cryonicists believe that people can be frozen immedi- Drexler and Merkle are the trinity of this scientistic sect, ately after death and reanimated later when the cure for then F. M. Esfandiary is its Saul. Esfandiary, on the road what ailed them is found. To see the flaw in this system, to his personal Damascus, changed his name to FM- thaw out a can of frozen strawberries. During freezing, 2030 (the number signifying his 100th birthday and the the water within each cell expands, crystallizes, and rup- year nanotechnology is predicted to make cryonics suc- tures the cell membranes. When defrosted, all the in- cessful) and declared, “I have no age. Am born and re- tracellular goo oozes out, turning your strawberries into born every day. I intend to live forever. Barring an ac- runny mush. This is your brain on cryonics. cident I probably will.” Cryonicists recognize this detriment and turn to Esfandiary forgot about cancer, a pancreatic form nanotechnology for a solution. Microscopic machines of which killed him on July 8, 2000. FM-2030—or more will be injected into the defrosting “patient” to repair precisely, his head—now resides in a vat the body molecule by molecule until the trillions of cells of liquid nitrogen at the Alcor Life Ex- The rub is finding are restored and the person can be resuscitated. Every tension Foundation in Scottsdale, Ariz., that balance religion needs its gods, and this scientistic vision has a but his legacy lives on among his fellow between being trinity in Robert C. W. Ettinger (The Prospect of Im- “transhumanists” (they have moved be- open-minded mortality), K. Eric Drexler (Engines of Creation) and yond human) and “extropians” (they are enough to accept Ralph C. Merkle (The Molecular Repair of the Brain), against entropy). who preach that nanocryonics will wash away the sin This is what I call “borderlands sci- radical new ideas of death. These works are built on the premise that if ence,” because it dwells in that fuzzy re- but not so open- you are cremated or buried, you have zero probability gion of claims that have yet to pass any minded that your of being resurrected—cryonics is better than everlasting tests but have some basis, however re- brains fall out. nothingness. mote, in reality. It is not impossible for Is it? That depends on how much time, effort and cryonics to succeed; it is just exceptionally unlikely. The money ($120,000 for a full-body freeze or $50,000 for rub in exploring the borderlands is finding that balance just the head) you are willing to invest for odds of suc- between being open-minded enough to accept radical cess only slightly higher than zero. It takes a blindly op- new ideas but not so open-minded that your brains fall timistic faith in the illimitable power of science to solve out. My credulity module is glad that some scientists any and all problems, including death. Look how far are devoting themselves to the problem of mortality. we’ve come in just a century, believers argue—from the My skepticism module, however, recognizes that trans- Wright brothers to Neil Armstrong in only 66 years. humanistic-extropian cryonics is uncomfortably close Extrapolate these trends out 1,000 years, or 10,000, to religion. I worry, as Matthew Arnold did in his 1852 and immortality is virtually certain. poem “Hymn of Empedocles,” that we will “feign a I want to believe the cryonicists. Really I do. I gave bliss/ Of doubtful future date, /And while we dream on up on religion in college, but I often slip back into my this/ Lose all our present state, /And relegate to worlds former evangelical fervor, now directed toward the yet distant our repose.” wonders of science and nature. But this is precisely why I’m skeptical. It is too much like religion: it promises Michael Shermer is publisher of Skeptic magazine everything, delivers nothing (but hope) and is based al- (www.skeptic.com) and author of How We Believe

BRAD HINES most entirely on faith in the future. And if Ettinger, and The Borderlands of Science.

www.sciam.com SCIENTIFIC AMERICAN 29 Copyright 2001 Scientific American, Inc. Profile

Young Cells in Old Brains The paradigm-shifting conclusion that adult brains can grow new neurons owes a lot to Elizabeth Gould’s rats and monkeys By MARGUERITE HOLLOWAY

PRINCETON, N.J.—Reunion weekend at Princeton Uni- versity, and the shady Gothic campus has been inun- dated by spring showers and men in boaters and natty orange seersucker jackets. Tents and small groups of murmuring alumni dot the courtyards. Everything proper, seemingly in its place. In Green Hall, however, the same order does not prevail. Elizabeth Gould’s lab- oratory is undergoing construction, and the neurosci- entist herself would not be mistaken for an alum: her plaid blue workman’s shirt hangs loosely and unbut- toned over a T-shirt and jeans, and she confesses she often feels out of place on the conservative campus. Against a backdrop of tidy ideas about the brain, Gould and her colleagues have been messing things up and, in the process, contributing to some of the most ex- citing findings of the past decade. Her work—and that of several other neuroscientists—has made clear that new neurons are produced in certain areas of the adult brains of mammals, including primates. Moreover, these cells can be killed off by stress and unchallenging envi- ronments but thrive in enriched settings where animals are learning, and they may play a role in memory. Until recently, dogma held that mature brains were static: no cells were born, except in the olfactory bulb. One of the cornerstones of this understanding came from studies by of Yale University, who ex- amined macaque monkeys and found no evidence of the creation of nerve cells, a process called neurogenesis. The prevailing view has since held that primates—and, indeed, mammals in general—are born with all the neu- rons they are going to have. Such neural stability was ELIZABETH GOULD: CHANGING MINDS considered necessary for long-term memory. So in the late 1980s when Gould, who was then researching the Past thinking: Memories are stored by locked-in neural connections. effect of hormones on the brain as a postdoctoral fellow Present: The brain can add neurons, perhaps to establish new memories. in the laboratory of Bruce S. McEwen at the Rockefeller Hope for dementia: New neurons seem able to migrate, suggesting that University, saw evidence of new neurons in the rat hip- therapeutic cells can be guided to areas damaged by disease or injury. pocampus, she was perplexed. Gould knew from the pi- Use it or lose it: In lab animals not kept in a stimulating cognitive oneering work of Fernando Nottebohm, also at Rock-

environment, “most new neurons will die within a few weeks.” efeller, that neurogenesis occurred in adult birds—ca- PHOTOGRAPHS BY PETER MURPHY

30 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. naries and zebra finches, for instance, And she has liked being in a quiet field grow nerve cells to learn new songs— of research, one she chose when study- but she and her lab mates knew of no ing psychology at the University of mammalian parallel. “We were really California at Los Angeles. “I have no puzzled,” she recalls. “It wasn’t until interest in doing experiments that we delved far enough back into the lit- someone else is going to do a month erature that we found evidence that later if I don’t get around to it,” she new neurons are produced in the says. “You have to pick things to do hippocampus.” that are really intriguing to you, things Those earlier studies had never that you are really curious about—not been widely noticed. Beginning in the just because you want to publish on 1960s Joseph Altman, now professor VISIBLE RAT BURROWS enable Gould to observe behavior them before anyone else does.” and experiences that might lead to new neurons. emeritus at Purdue University, and Her curiosity is taking her in sev- neurologist Michael S. Kaplan inde- eral directions these days. An out- pendently recorded neurogenesis in rats and other mammals. standing question centers on what role new hippocampal neu- They saw growth in the olfactory bulb, in the hippocampus—a rons play. Do they establish new circuits or memories? Or do region important to memory—and, most strikingly, in the neo- they replace old neurons in established circuits? This year Gould cortex, which is the part of the brain involved in higher thinking. and her colleagues reported that the neurons are involved in the “But nobody picked up on the results,” Gould says. “It is a clas- creation of trace memories—memories important to temporal sic example of something appearing before its time.” information. “We had evidence that the new cells were affect- In her work with rats, Gould verified that when she altered ed by learning, and this is evidence that the new cells are neces- the normal hormonal bath the hippocampus received, cells died sary for learning,” Gould explains. She now intends to do sim- and, apparently to compensate, more cells were born. “That ilar studies in marmosets, to see whether her discoveries about was really the beginning of my interest in neurogenesis and my rats will prove true for primates. realization that it happened,” she says. Her first papers on the The 39-year-old Gould is also repeating and extending work phenomenon, published in 1992 and 1993, did not attract of a few years ago in which she found neurogenesis in the neo- much attention. cortex of macaques, a finding that remains controversial and that Gould went on to do experiments clarifying aspects of neuro- would be highly significant because of the importance of the cor- genesis. She found that stress suppressed the creation of neurons tex. Although no one has published a replication so far, William and that lesions in the hippocampus triggered the development T. Greenough of the University of Illinois says Gould’s findings of new cells—something she considers significant because it im- “do not surprise me. We have unpublished data in rats that sup- plies that the brain can heal, or be induced to heal, after injury. port the same thing.” In 1997 Gould moved to Princeton as an assistant professor. In addition, Gould has begun investigating the role of sleep Over the next few years she and her co-workers reported that deprivation in neurogenesis, an interest triggered by the birth of new neurons survived if animals lived in complex environments her third child last year. “I never really thought about the sleep and learned tasks, findings also documented in mice by Fred H. aspect until I wasn’t getting any,” she says, laughing. And she Gage of the Salk Institute for Biological Studies in La Jolla, Calif. is intrigued by the possibility that much of what we have come Gould’s work in rats contradicting the status quo had al- to understand from laboratory settings may be skewed. ready put her out on a limb. Then she observed that new neu- “Our laboratory animals are very abnormal,” Gould notes. rons are found in the hippocampus of marmoset monkeys and “They have unlimited access to food and water, and they have macaques. News of neurogenesis in primates was met with skep- no interesting cognitive experiences at all. We know that if you ticism and stiff opposition, and some suggested her methods house an animal in that setting, most of its new neurons will die were flawed. She was soon vindicated, however, thanks to con- within a few weeks after they are produced.” Gould is design- firmatory work by Rakic in macaques and by Gage in the hu- ing environments that are closer to the ones rats and marmosets man hippocampus. The findings catalyzed widespread interest experience in the wild, hoping to get closer to the truth about because they introduced the possibility of repairing the brain and the brain. “It really raises the issue of whether a lot of the things elucidating memory formation. we are looking at are really deprivation effects.” For Gould, the sudden splash of attention has been disori- Potentially shifting another paradigm doesn’t faze Gould. enting—and she does not relish it, particularly when it takes her “There has to be some fresh perspective, something new that you away from her experiments. She says she is happiest in the lab, can bring to the work that other people wouldn’t see,” she says. working under the microscope with brain slices, which she finds “Otherwise you are not making a real contribution, and you beautiful and which recall a childhood interest in being an artist. might as well just step aside and find something else to do.”

www.sciam.com SCIENTIFIC AMERICAN 31 Copyright 2001 Scientific American, Inc. OVERVIEW LittleBig Science Nanotechnology is all the rage. But will it meet its ambitious goals? And what the heck is it?

By Gary Stix

Albert Einstein, as part of his doctoral disserta- technology. The field is a vast grab bag of stuff that has to do tion, calculated the size of a single sugar molecule from exper- with creating tiny things that sometimes just happen to be use- imental data on the diffusion of sugar in water. His work ful. It borrows liberally from condensed-matter physics, engi- showed that each molecule measures about a nanometer in di- neering, molecular biology and large swaths of chemistry. Re- ameter. At a billionth of a meter, a nanometer is the essence searchers who once called themselves materials scientists or or- of small. The width of 10 hydrogen atoms laid side by side, it ganic chemists have transmuted into nanotechnologists. is one thousandth the length of a typical bacterium, one mil- Purist academic types might prefer to describe themselves lionth the size of a pinhead, one billionth the length of Michael as mesoscale engineers. But it’s “nano” that generates the buzz. Jordan’s well-muscled legs. One nanometer is also precisely the Probably not since Du Pont coined its corporate slogan “bet- dimension of a big windfall for research. ter things for better living through chemistry” have scientists Almost 100 years after Einstein’s insight, the nanometer who engage in molecular manipulation so adeptly captured scale looms large on the research agenda. If Einstein were a and held public attention—in this case, the votes of lawmakers graduate student today probing for a career path, a doctoral in Washington who hold the research purse strings. “You need adviser would enjoin him to think small: “Nanotech, Albert, to come up with new, exciting, cutting-edge, at-the-frontier nanotech” would be the message conveyed. things in order to convince the budget- and policy-making ap- After biomedical research and defense—fighting cancer and paratus to give you more money,” remarks Duncan Moore, a building missile shields still take precedence—nanotechnology former White House official who helped to organize the Clin-

has become the most highly energized discipline in science and ton administration’s funding push for nanotechnology. IMAGE BY FELICE FRANKEL, WITH TECHNICAL HELP FROM J. CHRISTOPHER LOVE

32 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. With recognition has come lots of money—lots, that is, for TIP OF ATOMIC FORCE MICROSCOPE used to probe surfaces and something that isn’t a missile shield. The National Nanotech- manipulate molecules symbolizes the nanotechnology revolution. nology Initiative (NNI), announced early last year by President Bill Clinton, is a multiagency program intended to provide a big Interest in nano is also fueled, in an aberrant way, by the funding boost to nanoscience and engineering. The $422-mil- visions of a fringe element of futurists who muse on biblical life lion budget in the federal fiscal year that ends September 30 spans, on unlimited wealth and, conversely, on a holocaust marks a 56 percent jump in nano spending from a year earlier. brought about by legions of uncontrollable self-replicating ro- The initiative is on track to be augmented for fiscal year 2002 by bots only slightly bigger than Einstein’s sugar molecules. another 23 percent even while the Bush administration has pro- (Check out the Web site for NanoTechnology magazine— posed cuts to the funding programs of most of the federal agen- http://planet-hawaii.com/nanozine/—if you want to learn cies that support research and development (see the NNI Web about an “era of self-replicating consumer goods, super-health, site at www.nano.gov). Nano mania flourishes everywhere. super-economy and inventions impossible to fabricate with More than 30 nanotechnology research centers and interdisci- first wave industrialization.”) plinary groups have sprouted at universities; fewer than 10 ex- When Clinton introduced the nanotechnology initiative in isted two years ago. Nanoism does not, moreover, confine itself a speech last year, he was long on vision and short on specifics: to the U.S. In other countries, total funding for nanotechnology nanotech, he noted, might one day store the Library of Con- jumped from $316 million in 1997 to about $835 million this gress on a device the size of a sugar cube or produce materials year, according to the National Science Foundation (NSF). with 10 times the strength of steel at a mere fraction of its

www.sciam.com SCIENTIFIC AMERICAN 33 Copyright 2001 Scientific American, Inc. Macro, Micro, Nano How small is a nanometer? Stepping down in size by powers weight. But this wasn’t just the meanderings of a starry-eyed of 10 takes you from the back of a hand to, at one nanometer, a politician. Surprisingly, the science establishment itself is a lit- view of atoms in the building blocks of DNA. The edge of each image tle unclear about what it really means when it invokes nano. denotes a length 10 times longer than its next smallest neighbor. “It depends on whom you ask,” Stanford biophysicist Steven The black square frames the size of the next scene inward. M. Block told a National Institutes of Health symposium on nanotechnology last year in a talk that tried to define the sub- HAND ject. “Some folks apparently reserve the word to mean what- ever it is they do as opposed to whatever it is anyone else does.” 1 2 What’s in a Name? THE DEFINITION is indeed slippery. Some of nanotechnol- ogy isn’t nano, dealing instead with structures on the micron scale (millionths of a meter), 1,000 times or more larger than a nanometer. Also, nanotechnology, in many cases, isn’t tech- nology. Rather it involves basic research on structures having 10 centimeters 1 centimeter at least one dimension of about one to several hundred nano- meters. (In that sense, Einstein was more a nanoscientist than a technologist.) To add still more confusion, some nanotech- 3 4 nology has been around for a while: nano-size carbon black particles (a.k.a. high-tech soot) have gone into tires for 100 years as a reinforcing additive, long before the prefix “nano” ever created a stir. For that matter, a vaccine, which often con- sists of one or more proteins with nanoscale dimensions, might also qualify. But there is a there there in both nanoscience and nano- 1 millimeter 100 microns technology. The nanoworld is a weird borderland between the realm of individual atoms and molecules (where quantum WHITE BLOOD CELL mechanics rules) and the macroworld (where the bulk prop- erties of materials emerge from the collective behavior of tril- 5 6 lions of atoms, whether that material is a steel beam or the cream filling in an Oreo). At the bottom end, in the region of one nanometer, nanoland bumps up against the basic building blocks of matter. As such, it defines the smallest natural struc- tures and sets a hard limit to shrinkage: you just can’t build things any smaller. Nature has created nanostructures for billennia. But Mihail 10 microns 1 micron C. Roco, the NSF official who oversees the nanotechnology ini- tiative, offers a more restrictive definition. The emerging field— DNA new versus old nanotech—deals with materials and systems having these key properties: they have at least one dimension 7 8 of about one to 100 nanometers, they are designed through processes that exhibit fundamental control over the physical and chemical attributes of molecular-scale structures, and they can be combined to form larger structures. The intense inter- est in using nanostructures stems from the idea that they may boast superior electrical, chemical, mechanical or optical prop- erties—at least in theory. (See “Plenty of Room, Indeed,” by 100 nanometers 10 nanometers Michael Roukes, on page 48, for a discussion of why smaller is not always better.) Real-world nano, fitting Roco’s definition, does exist. Sand- 9 wiching several nonmagnetic layers, one of which is less than a nanometer thick, between magnetic layers can produce sen- sors for disk drives with many times the sensitivity of previ- ous devices, allowing more bits to be packed on the surface of

each disk. Since they were first introduced in 1997, these gi- PHOTOGRAPHS © 1982 BY SCIENTIFIC AMERICAN LIBRARY, USED WITH PERMISSION OF W. H. FREEMAN AND COMPANY From the classic book Powers of Ten, by Philip and Phylis Morrison and the 1 nanometer office of Charles and Ray Eames. SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Once conventional silicon electronics goes bust, new nanoelectronic devices are a good bet to replace them. A likely wager, though not a sure one.

ant magnetoresistive heads have served as an enabling tech- tors that must all work together to function in a nanocomput- nology for the multibillion-dollar storage industry. er. One company, Quantum Dot Corporation, has already New tools capable of imaging and manipulating single mol- emerged to exploit semiconductor quantum dots as labels in ecules or atoms have ushered in the new age of nano. The icons biological experiments, drug-discovery research, and diagnos- of this revolution are scanning probe microscopes—the scan- tic tests, among other applications. ning tunneling microscope and the atomic force microscope, Outside biology, the earliest wave of products involves us- among others—capable of creating pictures of individual atoms ing nanoparticles for improving basic material properties. For or moving them from place to place. The IBM Zurich Research instance, Nanophase Technologies, one of the few companies Laboratory has even mounted the sharp, nanometer-scale tips in this field that are publicly traded, produces nano-size zinc used in atomic force microscopes onto more than 1,000 mi- oxide particles for use in sunscreen, making the usually white- croscopic cantilevers on a microchip. The tips in the Millipede colored cream transparent because the tiny particles don’t scat- device can write digital bits on a polymer sheet. The technique ter visible light. could lead to a data storage device that achieves 20 times or more the density of today’s best disk drives. FUNDING FOR NANOTECHNOLOGY Varied approaches to fabricating nanostructures have 1,000 emerged in the nanoworld. Like sculptors, so-called top-down practitioners chisel out or add bulk material to a surface. Mi- 800 crochips, which now boast circuit lines of little more than 100 600 DOCUMENTED SPENDING BY nanometers, are about to become the most notable example. NNI Begins In contrast, bottom-up manufacturers use self-assembly pro- NON-U.S. GOVERNMENTS 400 cesses to put together larger structures—atoms or molecules Research Budgets Research *Proposed Spending that make ordered arrangements spontaneously, given the right 200 U.S. GOVERNMENT SPENDING

conditions. Nanotubes—graphite cylinders with unusual elec- (millions of current dollars) trical properties—are a good example of self-assembled nano- 0 structures [see “The Art of Building Small,” by George M. 1997 1998 1999 2000 2001 2002* Whitesides and J. Christopher Love, on page 38]. Government Fiscal Year SOURCES: U.S. Senate briefing on nanotechnology, May 24, 2001, and National Science Foundation Beyond Silicon THE DWINDLING SIZE of circuits in electronic chips drives TRENDS IN FEDERAL RESEARCH FOR SELECTED DISCIPLINES much of the interest in nano. Computer companies with large research laboratories, such as IBM and Hewlett-Packard, have 20 substantial nano programs. Once conventional silicon electron- ics goes bust—probably sometime in the next 10 to 25 years— 15 it’s a good bet that new nanotechnological electronic devices will replace them. A likely wager, though not a sure one. No one knows whether manufacturing electronics using nanotubes or 10 LIFE SCIENCES some other novel material will allow the relentless improvements ENGINEERING in chip performance without a corresponding increase in cost Budgets Research 5 that characterizes silicon chipmaking [see “The Incredible (billions of constant dollars) of constant (billions Shrinking Circuit,” by Charles M. Lieber, on page 58]. PHYSICAL SCIENCES Even if molecular-scale transistors don’t crunch zeroes and 0 ones in the Pentium XXV, the electronics fashioned by nano- 1970 1975 1980 1985 1990 1995 2000 technologists may make their way into devices that reveal the U.S. Government Fiscal Year secrets of the ultimate small machine: the biological cell. Bio- SOURCE: National Science Foundation nano, in fact, is finding real applications before the advent of postsilicon nanocomputers [see “Less Is More in Medicine,” UPTICK: The National Nanotechnology Initiative (NNI), begun in fiscal year by A. Paul Alivisatos, on page 66]. Relatively few nanotags 2001, helps to keep the U.S. competitive with world spending (top). It also made of a semiconductor material are needed to detect cellu- provides a monetary injection for the physical sciences and engineering,

NINA FINKEL lar activity, as opposed to the billions or trillions of transis- where funding has been flat by comparison with the life sciences (bottom).

www.sciam.com SCIENTIFIC AMERICAN 35 Copyright 2001 Scientific American, Inc. Nanotechnology’s bid for respectability is colored by the word’s association with a cabal of futurists who foresee nano as a pathway to utopia.

− 9 The government’s nanotech initiative goes beyond sun- A Few 10 Milestones screen. It envisages that nanostructured materials may help re- 3.5 billion years ago The first living cells emerge. Cells duce the size, weight and power requirements of spacecraft, house nanoscale biomachines that perform such tasks as create green manufacturing processes that minimize the gen- manipulating genetic material and supplying energy. eration of unwanted by-products, and form the basis of mole- 400 B.C. Democritus coins the word “atom,” which means cularly engineered biodegradable pesticides. The field has such “not cleavable” in ancient Greek. a broad scope—and basic research is still so new in some 1905 Albert Einstein publishes a paper that estimates the nanosubspecialties—that worries have arisen about its ability diameter of a sugar molecule as about one nanometer. to deliver on ambitious technology goals that may take 20 years 1931 Max Knoll and Ernst Ruska develop the electron to achieve. “While nanotechnology may hold great promise, microscope, which enables subnanometer imaging. some scientists contend that the field’s definition is too vague 1959 Richard Feynman gives his famed talk “There’s Plenty and that much of its ‘hype’ may not match the reality of pres- of Room at the Bottom,” on the prospects for miniaturization. ent scientific speculation,” noted a Congressional Research Ser- 1968 Alfred Y. Cho and John Arthur of Bell Laboratories and vice report last year. their colleagues invent molecular-beam epitaxy, a technique that can deposit single atomic layers on a surface. Nanodreams 1974 Norio Taniguchi conceives the word “nanotechnology” ANY ADVANCED RESEARCH carries inherent risks. But to signify machining with tolerances of less than a micron. nanotechnology bears a special burden. The field’s bid for re- 1981 and Heinrich Rohrer create the scanning spectability is colored by the association of the word with a ca- tunneling microscope, which can image individual atoms. bal of futurists who foresee nano as a pathway to a techno- 1985 Robert F. Curl, Jr., Harold W. Kroto and Richard E. utopia: unparalleled prosperity, pollution-free industry, even Smalley discover buckminsterfullerenes, also known as something resembling eternal life. buckyballs, which measure about a nanometer in diameter. In 1986—five years after IBM researchers Gerd Binnig and 1986 K. Eric Drexler publishes Engines of Creation, a Heinrich Rohrer invented the scanning tunneling microscope, futuristic book that popularizes nanotechnology. which garnered them the Nobel Prize—the book Engines of 1989 Donald M. Eigler of IBM writes the letters of his Creation, by K. Eric Drexler, created a sensation for its depic- company’s name using individual xenon atoms. tion of godlike control over matter. The book describes self- 1991 Sumio Iijima of NEC in Tsukuba, Japan, discovers replicating nanomachines that could produce virtually any ma- carbon nanotubes. terial good, while reversing global warming, curing disease and 1993 Warren Robinett of the University of North Carolina dramatically extending life spans. Scientists with tenured fac- and R. Stanley Williams of the University of California at ulty positions and NSF grants ridiculed these visions, noting Los Angeles devise a virtual-reality system connected to that their fundamental improbability made them an absurd a scanning tunneling microscope that lets the user see projection of what the future holds. and touch atoms. But the visionary scent that has surrounded nanotechnol- 1998 Cees Dekker’s group at the Delft University ogy ever since may provide some unforeseen benefits. To many of Technology in the Netherlands creates a transistor from nonscientists, Drexler’s projections for nanotechnology strad- a . dled the border between science and fiction in a compelling 1999 James M. Tour, now at Rice University, and Mark A. way. Talk of cell-repair machines that would eliminate aging Reed of Yale University demonstrate that single molecules as we know it and of home food-growing machines that could can act as molecular switches. produce victuals without killing anything helped to create a 2000 The Clinton administration announces the National fascination with the small that genuine scientists, consciously Nanotechnology Initiative, which provides a big boost in or not, would later use to draw attention to their work on more funding and gives the field greater visibility. mundane but eminently more real projects. Certainly labeling 2000 Eigler and other researchers devise a quantum mirage. a research proposal “nanotechnology” has a more alluring ring Placing a magnetic atom at one focus of an elliptical ring of than calling it “applied mesoscale materials science.” atoms creates a mirage of the same atom at another focus, a Less directly, Drexler’s work may actually draw people into possible means of transmitting information without wires. science. His imaginings have inspired a rich vein of science- fiction literature [see “Shamans of Small,” by Graham P.

36 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Collins, on page 86]. As a subgenre of science fiction—rather Nano for Sale than a literal prediction of the future—books about Drexler- Not all nanotechnology lies 20 years hence, ian nanotechnology may serve the same function as Star Trek as the following sampling of already does in stimulating a teenager’s interest in space, a passion that commercialized applications indicates. sometimes leads to a career in aeronautics or astrophysics. The danger comes when intelligent people take Drexler’s APPLICATION: CATALYSTS predictions at face value. Drexlerian nanotechnology drew re- COMPANY: EXXONMOBIL newed publicity last year when a morose Bill Joy, the chief sci- DESCRIPTION: Zeolites, minerals with pore sizes of entist of Sun Microsystems, worried in the magazine Wired less than one nanometer, serve as more efficient catalysts about the implications of nanorobots that could multiply un- to break down, or crack, large hydrocarbon molecules controllably. A spreading mass of self-replicating robots—what to form gasoline. Drexler has labeled “gray goo”—could pose enough of a threat to society, he mused, that we should consider stopping devel- APPLICATION: DATA STORAGE opment of nanotechnology. But that suggestion diverts atten- COMPANY: IBM tion from the real nano goo: chemical and biological weapons. DESCRIPTION: In the past few years, disk drives have Among real chemists and materials scientists who have added nanoscale layering—which exploits the giant magneto- now become nanotechnologists, Drexler’s predictions have as- resistive effect—to attain highly dense data storage. sumed a certain quaintness; science is nowhere near to being able to produce nanoscopic machines that can help revive APPLICATION: DRUG DELIVERY frozen brains from suspended animation. (Essays by Drexler COMPANY: GILEAD SCIENCES and his critics, including Nobel Prize winner Richard E. Smal- DESCRIPTION: Lipid spheres, called liposomes, which ley, appear in this issue.) Zyvex, a company started by a soft- measure about 100 nanometers in diameter, encase an ware magnate enticed by Drexlerian nanotechnology, has rec- anticancer drug to treat the AIDS-related Kaposi’s sarcoma. ognized how difficult it will be to create robots at the nano- meter scale; the company is now dabbling with much larger APPLICATION: MANUFACTURE OF RAW MATERIALS micromechanical elements, which Drexler has disparaged in COMPANY: CARBON his books [see “Nanobot Construction Crews,” by Steven Ash- DESCRIPTION: Co-founded by buckyball discoverer Richard E. ley, on page 84]. Smalley, the company has made carbon nanotubes more Even beyond meditations on gray goo, the nanotech field affordable by exploiting a new manufacturing process. struggles for cohesion. Some of the research would have pro- ceeded regardless of its label. Fusing “nano” and “technolo- APPLICATION: MATERIALS ENHANCEMENT gy” was an after-the-fact designation: IBM would have forged COMPANY: NANOPHASE TECHNOLOGIES ahead in building giant magnetoresistive heads whether or not DESCRIPTION: Nanocrystalline particles are incorporated the research it was doing was labeled nanotechnology. into other materials to produce tougher ceramics, transparent For the field to establish itself as a grand unifier of the ap- sunblocks to block infrared and ultraviolet radiation, and plied sciences, it must demonstrate the usefulness of grouping catalysts for environmental uses, among other applications. widely disparate endeavors. Can scientists and engineers do- ing research on nanopowders for sunscreens share a common set of interests with those working on DNA computing? In some cases, these crossover dreams may be justified. A semi- conductor quantum dot originally developed for electronics and now being deployed to detect biological activity in cells is a compelling proof of principle for these types of transdisci- plinary endeavors. If the nano concept holds together, it could, in fact, lay the groundwork for a new industrial revolution. But to succeed, it will need to discard not only fluff about nanorobots that bring cadavers back from a deep freeze but also the overheat- ed rhetoric that can derail any big new funding effort. Most important, the basic nanoscience must be forthcoming to iden- tify worthwhile nanotechnologies to pursue. Distinguishing be- tween what’s real and what’s not in nano throughout this pe- riod of extended exploration will remain no small task. NANOPARTICLES are made by Nanophase Technologies.

NANOPHASE TECHNOLOGIES Gary Stix is Scientific American’s special projects editor.

www.sciam.com SCIENTIFIC AMERICAN 37 Copyright 2001 Scientific American, Inc. Copyright 2001 Scientific American, Inc. NANOFABRICATION The art of BuildingSmall

BY GEORGE M. WHITESIDES AND J. CHRISTOPHER LOVE

RESEARCHERS ARE DISCOVERING CHEAP, EFFICIENT WAYS TO MAKE STRUCTURES ONLY A FEW BILLIONTHS OF A METER ACROSS

INTRICATE DIFFRACTION PATTERNSare created by nanoscale-width rings (too small to see) on the surface of one-centimeter-wide hemispheres made of clear polymer. Kateri E. Paul, a graduate student in George M. Whitesides’s group at Harvard University, fashioned the rings in a thin layer of gold on the hemispheres using a nanofabrication technique called soft lithography.

www.sciam.com SCIENTIFIC AMERICAN 39 Copyright 2001 Scientific American, Inc. “Make it small!” is a tech- investigate their properties and potential lion transistors per second in the U.S. nological edict that has changed the applications. The age of nanofabrication alone. Photolithography is basically an world. The development of microelec- is here, and the age of nanoscience has extension of photography. One first tronics—first the transistor and then the dawned, but the age of nanotechnol- makes the equivalent of a photographic aggregation of transistors into micro- ogy—finding practical uses for nano- negative containing the pattern required processors, memory chips and con- structures—has not really started yet. for some part of a microchip’s circuitry. trollers—has brought forth a cornucopia This negative, which is called the mask or of machines that manipulate informa- The Conventional Approach master, is then used to copy the pattern tion by streaming electrons through sil- RESEARCHERS may well develop nano- into the metals and semiconductors of a icon. Microelectronics rests on tech- structures as electronic components, but microchip. As is the case with photogra- niques that routinely fabricate structures the most important applications could phy, the negative may be hard to make, almost as small as 100 nanometers be quite different: for example, biolo- but creating multiple copies is easy, be- across (that is, 100 billionths of a meter). gists might use nanometer-scale particles cause the mask can be used many times. This size is tiny by the standards of as minuscule sensors to investigate cells. The process thus separates into two everyday experience—about one thou- Because scientists do not know what stages: the preparation of the mask (a sandth the width of a human hair—but kinds of nanostructures they will ulti- one-time event, which can be slow and it is large on the scale of atoms and mol- mately want to build, they have not yet expensive) and the use of the mask to ) ecules. The diameter of a 100-nanome- determined the best ways to construct manufacture replicas (which must be page 38 ter-wide wire would span about 500 them. Photolithography, the technology rapid and inexpensive). ( atoms of silicon. used to manufacture computer chips To make a mask for a part of a com- The idea of making “nanostruc- and virtually all other microelectronic puter chip, a manufacturer first designs tures” that comprise just one or a few systems, can be refined to make struc- the circuitry pattern on a conveniently

atoms has great appeal, both as a scien- tures smaller than 100 nanometers, but large scale and converts it into a pattern Harvard University tific challenge and for practical reasons. doing so is very difficult, expensive and of opaque metallic film (usually chromi- A structure the size of an atom repre- inconvenient. In a search to find better um) on a transparent plate (usually glass sents a fundamental limit: to make any- alternatives, nanofabrication researchers or silica). Photolithography then reduces thing smaller would require manipulat- have adopted the philosophy “Let a thou- the size of the pattern in a process anal- ing atomic nuclei—essentially, transmut- sand flowers bloom.” ogous to that used in a photographic ing one chemical element into another. In First, consider the advantages and darkroom [see illustration on opposite recent years, scientists have learned var- disadvantages of photolithography. Man- page]. A beam of light (typically ultravi- ious techniques for building nanostruc- ufacturers use this phenomenally pro- olet light from a mercury arc lamp) tures, but they have only just begun to ductive technology to churn out three bil- shines through the chromium mask, then passes through a lens that focuses the im- age onto a photosensitive coating of or- Overview/Nanofabrication ganic polymer (called the photoresist) on The development of nanotechnology will depend on the ability of researchers to the surface of a silicon wafer. The parts efficiently manufacture structures smaller than 100 nanometers (100 billionths of the photoresist struck by the light can of a meter) across. be selectively removed, exposing parts of Photolithography, the technology now used to fabricate circuits on microchips, the silicon wafer in a way that replicates can be modified to produce nanometer-scale structures, but the modifications the original pattern. would be technically difficult and hugely expensive. Why not use photolithography to Nanofabrication methods can be divided into two categories: top-down methods, make nanostructures? The technology which carve out or add aggregates of molecules to a surface, and bottom-up faces two limitations. The first is that the methods, which assemble atoms or molecules into nanostructures. shortest wavelength of ultraviolet light Two examples of promising top-down methods are soft lithography and dip-pen currently used in production processes is lithography. Researchers are using bottom-up methods to produce quantum dots about 250 nanometers. Trying to make that can serve as biological dyes. structures much smaller than half of that

spacing is like trying to read print that is IMAGE BY FELICE FRANKEL, WITH TECHNICAL HELP FROM KATERI E. PAUL; COURTESY OF GEORGE M. WHITESIDES

40 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. too tiny: diffraction causes the features to CONVENTIONAL PHOTOLITHOGRAPHY blur and meld together. Various techni- cal improvements have made it possible to push the limits of photolithography. 1 A laser beam writes the The smallest structures created in mass circuit pattern for a ULTRAVIOLET LIGHT production are somewhat larger than microchip on a layer of LASER BEAM 100 nanometers, and complex micro- light-sensitive polymer 1 electronic structures have been made that rests atop a layer MASK with features that are only 70 nanome- of chromium and a ters across. But these structures are still glass substrate. The 3 not small enough to explore some of the 2 sections of polymer LENS most interesting aspects of nanoscience. struck by the beam can The second limitation follows from be selectively removed. the first: because it is technically difficult to make such small structures using light, 2 The exposed sections of GLASS SUBSTRATE SILICON WAFER WITH it is also very expensive to do so. The pho- chromium are also removed, LAYER OF PHOTORESIST tolithographic tools that will be used to and the rest of the polymer CHROMIUM LAYER make chips with features well below 100 is dissolved. The result is a nanometers will each cost tens to hun- mask—the equivalent of a dreds of millions of dollars. This expense photographic negative. 4 may or may not be acceptable to manu- facturers, but it is prohibitive for the bi- 3 When a beam of ultraviolet light is directed at ologists, materials scientists, chemists and the mask, the light passes through the gaps physicists who wish to explore nanosci- in the chromium. A lens shrinks the pattern by ence using structures of their own design. focusing the light onto a layer of photoresist on a silicon wafer. Future Nanochips THE ELECTRONICS industry is deeply 4 The exposed parts of the photoresist are interested in developing new methods for removed, allowing the replication of the pattern SILICON CHIPS nanofabrication so that it can continue its in miniature on the silicon chips. long-term trend of building ever smaller, faster and less expensive devices. It would be a natural evolution of microelectron- manuscript by hand, one line at a time. ly damages many of the materials used in ics to become nanoelectronics. But be- If electrons are not the answer, what masks and lenses. But the microelectron- cause conventional photolithography be- is? Another contender is lithography us- ics industry clearly would prefer to make comes more difficult as the dimensions of ing x-rays with wavelengths between 0.1 advanced chips using extensions of fa- the structures become smaller, manufac- and 10 nanometers or extreme ultravio- miliar technology, so these methods are turers are exploring alternative technolo- let light with wavelengths between 10 being actively developed. Some of the gies for making future nanochips. and 70 nanometers. Because these forms techniques (for example, advanced ul- One leading contender is electron- of radiation have much shorter wave- traviolet lithography for chip produc- beam lithography. In this method, the lengths than the ultraviolet light current- tion) will probably become commercial circuitry pattern is written on a thin poly- ly used in photolithography, they mini- realities. They will not, though, make in- mer film with a beam of electrons. An mize the blurring caused by diffraction. expensive nanostructures and thus will electron beam does not diffract at atom- These technologies face their own set of do nothing to open nanotechnology to a ic scales, so it does not cause blurring of problems, however: conventional lenses broader group of scientists and engineers. the edges of features. Researchers have are not transparent to extreme ultravio- The need for simpler and less expen- used the technique to write lines with let light and do not focus x-rays. Fur- sive methods of fabricating nanostruc- widths of only a few nanometers in a lay- thermore, the energetic radiation rapid- tures has stimulated the search for un- er of photoresist on a silicon substrate. The electron-beam instruments current- GEORGE M. WHITESIDES and J. CHRISTOPHER LOVE work together on unconventional meth- ly available, however, are very expensive ods of nanofabrication in the department of chemistry at Harvard University. Whitesides, and impractical for large-scale manufac- a professor of chemistry, received his Ph.D. from the California Institute of Technology in turing. Because the beam of electrons is 1964 and joined the Harvard faculty in 1982. Love is a graduate student and a member of needed to fabricate each structure, the Whitesides’s research group. He received his bachelor’s degree in chemistry from the Uni- THE AUTHORS BRYAN CHRISTIE process is similar to the copying of a versity of Virginia in 1999 and his master’s degree from Harvard in 2001.

www.sciam.com SCIENTIFIC AMERICAN 41 Copyright 2001 Scientific American, Inc. SOFT LITHOGRAPHY

Printing, molding and other mechanical processes Such techniques can fabricate devices that carried out using an elastic stamp can produce might be used in optical communications or patterns with nanoscale features. biochemical research. MAKING AN ELASTIC STAMP

1 A liquid precursor to polydimethylsiloxane (PDMS) is poured over a bas-relief master 2 The liquid is cured into a rubbery solid that produced by photolithography or matches the original pattern. 3 The PDMS stamp is peeled off the master. electron-beam lithography. LIQUID PRECURSOR TO PDMS PDMS STAMP

MASTER PHOTORESIST MICROCONTACT PRINTING

1 The PDMS stamp is inked with a solution consisting of 2 The thiols form a self-assembled monolayer on the gold organic molecules called thiols and then pressed against surface that reproduces the stamp’s pattern; features in the a thin film of gold on a silicon plate. pattern are as small as 50 nanometers.

SELF-ASSEMBLED GOLD SURFACE MONOLAYER THIOL INK

MICROMOLDING IN CAPILLARIES

SOLIDIFIED POLYMER

LIQUID POLYMER

1 The PDMS stamp is placed on a hard surface, 2 The polymer solidifies into the and a liquid polymer flows into the recesses desired pattern, which may contain between the surface and the stamp. features smaller than 10 nanometers. BRYAN CHRISTIE

42 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. conventional approaches that have not matches the original pattern with aston- are in the field of nanofluidics, an exten- been explored by the electronics indus- ishing fidelity: the stamp reproduces fea- sion of microfluidics that would involve try. We first became interested in the top- tures from the master as small as a few producing chips for biochemical research ic in the 1990s when we were engaged in nanometers. Although the creation of a with channels only a few nanometers making the simple structures required in finely detailed bas-relief master is expen- wide. At that scale, fluid dynamics may microfluidic systems—chips with chan- sive because it requires electron-beam allow new ways to separate materials nels and chambers for holding liquids. lithography or other advanced techniques, such as fragments of DNA. This lab-on-a-chip has myriad potential copying the pattern on PDMS stamps is These methods require no special uses in biochemistry, ranging from drug cheap and easy. And once a stamp is in equipment and in fact can be carried out screening to genetic analysis. The chan- hand, it can be used in various inexpen- by hand in an ordinary laboratory. Con- nels in microfluidic chips are enormous sive ways to make nanostructures. ventional photolithography must take by the standards of microelectronics: 50 The first method—originally devel- place in a clean-room facility devoid of microns (or 50,000 nanometers) wide, oped by Amit Kumar, a postdoctoral stu- dust and dirt; if a piece of dust lands on the rather than 100 nanometers. But the dent in our group at Harvard Universi- mask, it will create an unwanted spot on techniques for producing those channels ty—is called microcontact printing. The the pattern. As a result, the device being are quite versatile. Microfluidic chips can PDMS stamp is “inked” with a reagent fabricated (and sometimes neighboring be made quickly and inexpensively, and solution consisting of organic molecules devices) may fail. Soft lithography is gen- many are composed of organic polymers called thiols [see middle illustration on erally more forgiving because the PDMS and gels—materials not found in the opposite page]. The stamp is then brought stamp is elastic. If a piece of dust gets world of electronics. We discovered that into contact with an appropriate sheet of trapped between the stamp and the sur- These methods require no special equipment and in fact can be carried out by hand in an ordinary lab. we could use similar techniques to cre- “paper”—a thin film of gold on a glass, face, the stamp will compress over the ate nanostructures. silicon or polymer plate. The thiols react top of the particle but maintain contact The methods represented, in a sense, with the gold surface, forming a highly with the rest of the surface. Thus, the pat- a step backward in technology. Instead ordered film (called a self-assembled tern will be reproduced correctly except of using the tools of physics—light and monolayer, or SAM) that replicates the for where the contaminant is trapped. electrons—we employed mechanical pro- stamp’s pattern. Because the thiol ink Moreover, soft lithography can pro- cesses that are familiar in everyday life: spreads a bit after it contacts the surface, duce nanostructures in a wide range of printing, stamping, molding and em- the resolution of the monolayer cannot materials, including the complex organ- bossing. The techniques are called soft be quite as high as that of the PDMS ic molecules needed for biological stud- lithography because the tool they have in stamp. But when used correctly, micro- ies. And the technique can print or mold common is a block of polydimethyl- contact printing can produce patterns patterns on curved as well as planar sur- siloxane (PDMS)—the rubbery polymer with features as small as 50 nanometers. faces. But the technology is not ideal for used to caulk the leaks around bathtubs. Another method of soft lithography, making the structures required for com- (Physicists often refer to such organic called micromolding in capillaries, in- plex nanoelectronics. Currently all inte- chemicals as “soft matter.”) volves using the PDMS stamp to mold grated circuits consist of stacked layers of To carry out reproduction using soft patterns. The stamp is placed on a hard different materials. Deformations and lithography, one first makes a mold or a surface, and a liquid polymer flows by distortions of the soft PDMS stamp can stamp. The most prevalent procedure is capillary action into the recesses between produce small errors in the replicated to use photolithography or electron- the surface and the stamp [see bottom il- pattern and a misalignment of the pat- beam lithography to produce a pattern in lustration on opposite page]. The poly- tern with any underlying patterns previ- a layer of photoresist on the surface of a mer then solidifies into the desired pat- ously fabricated. Even the tiniest distor- silicon wafer. This process generates a tern. This technique can replicate struc- tions or misalignments can destroy a bas-relief master in which islands of pho- tures smaller than 10 nanometers. It is multilayered nanoelectronic device. toresist stand out from the silicon [see particularly well suited for producing Therefore, soft lithography is not well top illustration on opposite page]. Then subwavelength optical devices, wave- suited for fabricating structures with a chemical precursor to PDMS—a free- guides and optical polarizers, all of multiple layers that must stack precisely flowing liquid—is poured over the bas- which could be used in optical fiber net- on top of one another. relief master and cured into the rubbery works and eventually perhaps in optical Researchers have found ways, how- solid. The result is a PDMS stamp that computers. Other possible applications ever, to correct this shortcoming—at

www.sciam.com SCIENTIFIC AMERICAN 43 Copyright 2001 Scientific American, Inc. DIP-PEN LITHOGRAPHY more commonly, in monolayer films of atoms or molecules that coat the surface). AFM CANTILEVER THIOL MOLECULES Similarly, if researchers increase the cur- rents flowing from the tip of the STM, the AFM TIP microscope becomes a very small source DROP OF WATER for an electron beam, which can be used to write nanometer-scale patterns. The GOLD SURFACE STM tip can also push individual atoms around on a surface to build rings and wires that are only one atom wide. An intriguing new scanning probe fabrication method is called dip-pen lith- PYRAMIDAL TIP ography. Developed by Chad A. Mirkin of an atomic force micro- of Northwestern University, this tech- scope (AFM) is coated with nique works much like a goose-feather a thin film of thiol molecules. pen [see illustration at left]. The tip of the A minute drop of water AFM is coated with a thin film of thiol condenses between the molecules that are insoluble in water but microscope’s tip and a gold surface. react with a gold surface (the same chem- The thiols migrate from the tip to the surface, SELF-ASSEMBLED istry used in microcontact printing). MONOLAYER where they form a self-assembled monolayer. When the device is placed in an atmo- sphere containing a high concentration of water vapor, a minute drop of water least in part—by employing a rigid stamp Research Laboratory received the Nobel condenses between the gold surface and instead of an elastic one. In a technique Prize in Physics in 1986. This remark- the microscope’s tip. Surface tension called step-and-flash imprint lithogra- able device detects small currents that pulls the tip to a fixed distance from the phy, developed by C. Grant Willson of pass between the microscope’s tip and gold, and this distance does not change the University of Texas, photolithogra- the sample being observed, allowing re- as the tip moves across the surface. The phy is used to etch a pattern into a quartz searchers to “see” substances at the scale drop of water acts as a bridge over which plate, yielding a rigid bas-relief master. of individual atoms. The success of the the thiol molecules migrate from the tip Willson eliminated the step of making a STM led to the development of other to the gold surface, where they are fixed. PDMS stamp from the master; instead scanning probe devices, including the Researchers have used this procedure to the master itself is pressed against a thin atomic force microscope (AFM). The write lines a few nanometers across. film of liquid polymer, which fills the operating principle of the AFM is simi- Although dip-pen lithography is rel- master’s recesses. Then the master is ex- lar to that of an old-fashioned phono- atively slow, it can use many different posed to ultraviolet light, which solidifies graph. A tiny probe—a fiber or a pyra- types of molecules as “inks” and thus the polymer to create the desired replica. mid-shaped tip that is typically between brings great chemical flexibility to nano- A related technique called nanoimprint two and 30 nanometers wide—is brought meter-scale writing. Researchers have lithography, developed by Stephen Y. into direct contact with the sample. The not yet determined the best applications Chou of Princeton University, also em- probe is attached to the end of a can- for the technique, but one idea is to use ploys a rigid master but uses a film of tilever, which bends as the tip moves the dip-pen method for precise modifica- polymer that has been heated to a tem- across the sample’s surface. The deflec- tions of circuit designs. Mirkin has re- perature near its melting point to facilitate tion is measured by reflecting a beam of cently demonstrated that a variant of the the embossing process. Both methods can laser light off the top of the cantilever. ink used in dip-pen lithography can write produce two-dimensional structures with The AFM can detect variations in verti- directly on silicon. good fidelity, but it remains to be seen cal surface topography that are smaller An interesting cousin to these tech- whether the techniques are suitable for than the dimensions of the probe. niques involves another kind of nano- manufacturing electronic devices. But scanning probe devices can do structure, called a break junction. If you more than simply allow scientists to ob- break a thin, ductile metal wire into two Pushing Atoms Around serve the atomic world—they can also be parts by pulling sharply, the process THE CURRENT REVOLUTION in used to create nanostructures. The tip on seems abrupt to a human observer, but it nanoscience started in 1981 with the in- the AFM can be used to physically move actually follows a complex sequence. vention of the scanning tunneling micro- nanoparticles around on surfaces and to When the force used in breaking the wire scope (STM), for which Heinrich Rohrer arrange them in patterns. It can also be is first applied, the metal begins to yield

and Gerd K. Binnig of the IBM Zurich used to make scratches in a surface (or and flow, and the diameter of the wire BRYAN CHRISTIE

44 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. decreases. As the two ends move apart, gies for using single organic molecules as ally generated as simple particles in sus- the wire gets thinner and thinner until, in electronic devices such as diodes and tran- pension or on surfaces, rather than as de- the instant just before breaking, it is a sin- sistors [see “Computing with Mole- signed, interconnected patterns. gle atom in diameter at its narrowest cules,” by Mark A. Reed and James M. Two of the most prominent bottom- point. This process of thinning a wire to Tour; Scientific American, June 2000]. up methods are those used to make a break junction can be detected easily by nanotubes and quantum dots. Scientists measuring the current that flows through Top-Down and Bottom-Up have made long, cylindrical tubes of car- the wire. When the wire is slender enough, ALL THE FORMS of lithography we bon by a catalytic growth process that current can flow only in discrete quanti- have discussed so far are called top-down employs a nanometer-scale drop of ties (that is, current flow is quantized). methods—that is, they begin with a pat- molten metal (usually iron) as a catalyst The break junction is analogous to tern generated on a larger scale and re- [see “Nanotubes for Electronics,” by two STM tips facing each other, and sim- duce its lateral dimensions (often by a Philip G. Collins and Phaedon Avouris; ilar physical rules govern the current that factor of 10) before carving out nano- Scientific American, December 2000]. flows through it. Mark A. Reed of Yale structures. This strategy is required in The most active area of research in University has pioneered a particularly in- fabricating electronic devices such as mi- quantum dots originated in the labora- ventive use of the break junction. He built crochips, whose functions depend more tory of Louis E. Brus (then at Bell Labo- a device that enabled a thin junction to be on their patterns than on their dimen- ratories) and has been developed by broken under carefully controlled condi- sions. But no top-down method is ideal; A. Paul Alivisatos of the University of tions and then allowed the broken tips to none can conveniently, cheaply and California at Berkeley, Moungi G. be brought back together or to be held quickly make nanostructures of any ma- Bawendi of the Massachusetts Institute apart at any distance with an accuracy of terial. So researchers have shown grow- of Technology, and others. Quantum a few thousandths of a nanometer. By ad- ing interest in bottom-up methods, dots are crystals containing only a few justing the distance between the tips in the which start with atoms or molecules and hundred atoms. Because the electrons in presence of an organic molecule that build up to nanostructures. These meth- a quantum dot are confined to widely bridged them, Reed was able to measure ods can easily make the smallest nano- separated energy levels, the dot emits a current flowing across the organic structures—with dimensions between only one wavelength of light when it is bridge. This experiment was an impor- two and 10 nanometers—and do so in- excited. This property makes the quan- tant step in the development of technolo- expensively. But these structures are usu- tum dot useful as a biological marker [see Nanofabrication: Comparing the Methods Researchers are developing an array of techniques for building structures smaller than 100 nanometers. Here is a summary of the advantages and disadvantages of four methods. Photolithography Soft Lithography Advantages: The electronics industry is already familiar with Advantages: This method allows researchers to inexpensively this technology because it is currently used to fabricate reproduce patterns created by electron-beam lithography or microchips. Manufacturers can modify the technique to produce other related techniques. Soft lithography requires no special nanometer-scale structures by employing electron beams, equipment and can be carried out by hand in an ordinary x-rays or extreme ultraviolet light. laboratory. Disadvantages: The necessary modifications will be expensive Disadvantages: The technique is not ideal for manufacturing the and technically difficult. Using electron beams to fashion multilayered structures of electronic devices. Researchers are structures is costly and slow. X-rays and extreme ultraviolet trying to overcome this drawback, but it remains to be seen light can damage the equipment used in the process. whether these efforts will be successful. Scanning Probe Methods Bottom-Up Methods Advantages: The scanning tunneling microscope and the atomic Advantages: By setting up carefully controlled chemical force microscope can be used to move individual nanoparticles reactions, researchers can cheaply and easily assemble atoms and arrange them in patterns. The instruments can build rings and molecules into the smallest nanostructures, with and wires that are only one atom wide. dimensions between two and 10 nanometers. Disadvantages: The methods are too slow for mass production. Disadvantages: Because these methods cannot produce Applications of the microscopes will probably be limited to the designed, interconnected patterns, they are not well suited for fabrication of specialized devices. building electronic devices such as microchips.

www.sciam.com SCIENTIFIC AMERICAN 45 Copyright 2001 Scientific American, Inc. QUANTUM DOT ASSEMBLY “Less Is More in Medicine,” on page 66]. One procedure used to make quan- Crystals called quantum dots contain only a few hundred atoms and tum dots involves a chemical reaction be- emit different wavelengths of light, depending on their size. They may tween a metal ion (for example, cadmi- um) and a molecule that is able to donate become useful as biological markers of cellular activity. a selenium ion. This reaction generates crystals of cadmium selenide. The trick is to prevent the small crystals from stick- 1 A chemical reaction brings ing together as they grow to the desired together cadmium ions size. To insulate the growing particles (purple), selenium ions from one another, researchers carry out (green) and organic the reaction in the presence of organic molecules (red spheres molecules that act as surfactants, coating with blue tails). the surface of each cadmium selenide particle as it grows. The organic mole- cules stop the crystals from clumping to- gether and regulate their rate of growth. The geometry of the particles can be con- trolled to some extent by mixing differ- ent ratios of the organic molecules. The reaction can generate particles with a va- riety of shapes, including spheres, rods 2 The organic molecules act and tetrapods (four-armed particles sim- as surfactants, binding to ilar to toy jacks). the surface of the cadmium It is important to synthesize the quan- selenide crystal as it grows. tum dots with uniform size and composi- tion, because the size of the dot deter- mines its electronic, magnetic and optical properties. Researchers can select the size of the particles by varying the length of time for the reaction. The organic coating 3 When the crystal reaches its also helps to set the size of the particles. optimum size, the organic When the nanoparticle is small (on the molecules coat its surface in scale of molecules), the organic coating is a stable packing. loose and allows further growth; as the particle enlarges, the organic molecules become crowded. There is an optimum size for the particles that allows the most stable packing of the organic molecules and thus provides the greatest stabiliza- tion for the surfaces of the crystals. These cadmium selenide nanoparti- cles promise some of the first commercial products of nanoscience: Quantum Dot Corporation has been developing the crystals for use as biological labels. Re- searchers can tag proteins and nucleic acids with quantum dots; when the sam- ple is illuminated with ultraviolet light, the crystals will fluoresce at a specific wavelength and thus show the locations of the attached proteins. Many organic molecules also fluoresce, but quantum dots have several advantages that make

them better markers. First, the color of BRYAN CHRISTIE

46 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. a quantum dot’s fluorescence can be tai- mammalian cells are larger. Cells are, The development of nanotechnology lored by changing the dot’s size: the larg- however, filled with much smaller struc- will depend on the availability of nano- er the particle, the more the emitted light tures, many of which are astonishingly structures. The invention of the STM is shifted toward the red end of the spec- sophisticated. The ribosome, for exam- and AFM has provided new tools for trum. Second, if all the dots are the same ple, carries out one of the most important viewing, characterizing and manipulat- size, their fluorescence spectrum is nar- cellular functions: the synthesis of pro- ing these structures; the issue now is how row—that is, they emit a very pure color. teins from amino acids, using messenger to build them to order and how to de- This property is important because it al- RNA as the template. The complexity of sign them to have new and useful func- lows particles of different sizes to be this molecular construction project far tions. The importance of electronics ap- used as distinguishable labels. Third, the surpasses that of man-made techniques. plications has tended to focus attention fluorescence of quantum dots does not Or consider the rotary motors of the bac- on nanodevices that might be incorpo- fade on exposure to ultraviolet light, as terial flagella, which efficiently propel the rated into future integrated circuits. And does that of organic molecules. When one-celled organisms [see “The Once and for good technological reasons, the elec- used as dyes in biological research, the Future Nanomachine,” on page 78]. tronics industry has emphasized fabri- dots can be observed for conveniently It is unclear if “nanomachines” tak- cation methods that are extensions of long periods. en from cells will be useful. They will those currently used to make micro- Scientists are also investigating the probably have very limited application in chips. But the explosion of interest in possibility of making structures from col- electronics, but they may provide valu- nanoscience has created a demand for a loids—nanoparticles in suspension. Chris- able tools for chemical synthesis and broad range of fabrication methods, topher B. Murray and a team at the IBM sensing devices. Recent work by Carlo D. with an emphasis on low-cost, conve- Thomas J. Watson Research Center are Montemagno of Cornell University has nient techniques. exploring the use of such colloids to cre- shown that it is possible to engineer a The new approaches to nanofabrica- ate a medium for ultrahigh-density data primitive nanomachine with a biological tion are unconventional only because storage. The IBM team’s colloids con- engine. Montemagno extracted a rotary they are not derived from the microtech- tain magnetic nanoparticles as small as motor protein from a bacterial cell and nology developed for electronic devices. Bottom-up methods start from atoms or molecules and build up to nanostructures. three nanometers across, each composed connected it to a metallic nanorod—a Chemists, physicists and biologists are of about 1,000 iron and platinum atoms. cylinder 750 nanometers long and 150 rapidly accepting these techniques as the When the colloid is spread on a surface nanometers wide that had been fabricat- most appropriate ways to build various and the solvent allowed to evaporate, ed by lithography. The rotary motor, kinds of nanostructures for research. the nanoparticles crystallize in two- or which was only 11 nanometers tall, was And the methods may even supplement three-dimensional arrays. Initial studies powered by adenosine triphosphate the conventional approaches—photo- indicate that these arrays can potential- (ATP), the source of chemical energy in lithography, electron-beam lithography ly store trillions of bits of data per square cells. Montemagno showed that the mo- and related techniques—for applications inch, giving them a capacity 10 to 100 tor could rotate the nanorod at eight rev- in electronics as well. The microelec- times greater than that of present mem- olutions per minute. At the very least, tronics mold is now broken. Ideas for ory devices. such research stimulates efforts to fabri- nanofabrication are coming from many cate functional nanostructures by demon- directions in a wonderful free-for-all of The Future of strating that such structures can exist. discovery. Nanofabrication THE INTEREST in nanostructures is so MORE TO EXPLORE great that every plausible fabrication More information about nanofabrication can be found at the following Web sites: technique is being examined. Although International SEMATECH: www.sematech.org/public/index.htm physicists and chemists are now doing The Whitesides group at Harvard University: gmwgroup.harvard.edu most of the work, biologists may also The Mirkin group at Northwestern University: www.chem.northwestern.edu/~mkngrp/ make important contributions. The cell The Willson group at the University of Texas at Austin: (whether mammalian or bacterial) is rel- willson.cm.utexas.edu/Research/research.htm atively large on the scale of nanostruc- The Alivisatos group at the University of California at Berkeley: www.cchem.berkeley.edu/~pagrp/ tures: the typical bacterium is approxi- The Bawendi group at M.I.T.: web.mit.edu/chemistry/nanocluster/ mately 1,000 nanometers long, and The Montemagno group at Cornell University: falcon.aben.cornell.edu/

www.sciam.com SCIENTIFIC AMERICAN 47 Copyright 2001 Scientific American, Inc. NANOPHYSICS

Plentyof RoomIndeed, There is plenty of room for practical innovation at the nanoscale. But first, scientists have to understand the unique physics that governs matter there

By Michael Roukes

Back in December 1959, future plenty of room. I will not now discuss his singular intellect toward. Indeed, it Nobel laureate Richard Feynman gave a how we are going to do it, but only what has profoundly inspired my two decades visionary and now oft-quoted talk enti- is possible in principle....We are not do- of research on physics at the nanoscale. tled “There’s Plenty of Room at the Bot- ing it simply because we haven’t yet got- Today there is a nanotechnology tom.” The occasion was an American ten around to it.” gold rush. Nearly every major funding Physical Society meeting at the Califor- The breadth of Feynman’s vision is agency for science and engineering has nia Institute of Technology, Feynman’s staggering. In that lecture 42 years ago announced its own thrust into the field. intellectual home then and mine today. he anticipated a spectrum of scientific Scores of researchers and institutions are Although he didn’t intend it, Feynman’s and technical fields that are now well es- scrambling for a piece of the action. But 7,000 words were a defining moment in tablished, among them electron-beam in all honesty, I think we have to admit nanotechnology, long before anything and ion-beam fabrication, molecular- that much of what invokes the hallowed “nano” appeared on the horizon. beam epitaxy, nanoimprint lithography, prefix “nano” falls a bit short of Feyn- “What I want to talk about,” he projection electron microscopy, atom- man’s mark. said, “is the problem of manipulating by-atom manipulation, quantum-effect We’ve only just begun to take the and controlling things on a small electronics, spin electronics (also called first steps toward his grand vision of as- scale.... What I have demonstrated is spintronics) and microelectromechanical sembling complex machines and circuits that there is room—that you can de- systems (MEMS). The lecture also pro- atom by atom. What can be done now is crease the size of things in a practical jected what has been called the “magic” extremely rudimentary. We’re certainly way. I now want to show that there is Feynman brought to everything he turned nowhere near being able to commercial-

48 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. ly mass-produce nanosystems—integrat- NOVEL NANOTECH DEVICES, such as these nanoelectromechanical resonators, are enabling scientists ed multicomponent nanodevices that to discover the laws of physics that regulate the unique properties of matter at the mesoscale. have the complexity and range of func- tions readily provided by modern mi- of one to a few hundred nanometers— transitional frontier using complemen- crochips. But there is a fundamental sci- small indeed, but much bigger than sim- tary top-down and bottom-up fabrica- ence issue here as well. It is becoming in- ple molecules. Matter at this mesoscale tion methods. Advances in top-down creasingly clear that we are only begin- is often awkward to explore. It contains nanofabrication techniques such as elec- ning to acquire the detailed knowledge too many atoms to be easily understood tron-beam lithography (used extensively that will be at the heart of future nano- by straightforward application of quan- by my own research group) yield almost technology. This new science concerns the tum mechanics (although the funda- atomic-scale precision, but achieving suc-

COLORIZATION BY FELICE FRANKEL properties and behavior of aggregates of mental laws still apply). Yet these sys- cess, not to mention reproducibility, as atoms and molecules, at a scale not yet tems are not so large as to be complete- we scale down to the single-digit-nano- large enough to be considered macro- ly free of quantum effects; thus, they do meter regime becomes problematic. Al- scopic but far beyond what can be called not simply obey the classical physics ternatively, scientists are using bottom- microscopic. It is the science of the meso- governing the macroworld. It is precise- up techniques for self-assembly of atoms. scale, and until we understand it, practical ly in this intermediate domain, the meso- But the advent of preprogrammed self- devices will be difficult to realize. world, that unforeseen properties of col- assembly of arbitrarily large systems— Today’s scientists and engineers lective systems emerge. with complexity comparable to that

M. J. MURPHY, D. A. HARRINGTON AND L. ROUKES California Institute of Technology; readily fashion nanostructures on a scale Researchers are approaching this built every day in microelectronics, in

www.sciam.com SCIENTIFIC AMERICAN 49 Copyright 2001 Scientific American, Inc. It is becoming increasingly clear that we are only beginning to acquire the detailed knowledge that will be at the heart of future nanotechnology.

MEMS and (of course) by Mother Na- another in ways we do not yet understand of Technology and Henk van Houten of ture—is nowhere on the horizon. It ap- and, hence, cannot yet control. the Philips Research Laboratories (both pears that the top-down approach will Feynman’s original vision was clear- in the Netherlands) and collaborators most likely remain the method of choice ly intended to be inspirational. Were he were studying the flow of electric current for building really complex devices for a observing now, he would surely be through what are now called quantum- good while (for more, see “The Art of alarmed when people take his projec- point contacts. These are narrow con- Building Small,” on page 38). tions as some sort of gospel. He deliv- ducting paths within a semiconductor, Our difficulty in approaching the ered his musings with characteristic along which electrons are forced to flow mesoscale from above or below bespeaks playfulness as well as deep insight. Sad- [see illustration on page 54]. Late one a basic challenge of physics. Lately, the ly for us, the field that would be called evening van Wees’s undergraduate assis- essence of Feynman’s “Plenty of Room” nanotechnology was just one of many tant, Leo Kouwenhoven, was measuring talk seems to be taken as a license for lais- that intrigued him. He never really con- the conductance through the constriction sez faire in nanotechnology. Yet Feyn- tinued with it, returning to give but one as he varied its width systematically. The man never asserted that “anything goes” redux of his original lecture, at the Jet research team was expecting to see only at the nanoscale. He warned, for in- Propulsion Laboratory in 1983. subtle conductance effects against an stance, that the very act of trying to otherwise smooth and unremarkable “arrange the atoms one by one the way New Laws Prevail background response. Instead there ap- we want them” is subject to fundamen- IN 1959, and even in 1983, the com- peared a very pronounced, and now char- tal principles: “You can’t put them so plete physical picture of the nanoscale acteristic, staircase pattern. Further analy- that they are chemically unstable, for ex- was far from clear. The good news for re- sis that night revealed that plateaus were ample.” Accordingly, today’s scanning searchers is that, by and large, it still is! occurring at regular, precise intervals. probe microscopes can move atoms from Much exotic territory awaits explo- David Wharam and Michael Pepper place to place on a prepared surface, but ration. As we delve into it, we will un- of the observed this ability does not immediately confer cover a panoply of phenomena that we similar results. The two discoveries rep- the power to build complex molecular as- must understand before practical nano- resented the first robust demonstrations semblies at will. What has been accom- technology will become possible. The of the quantization of electrical conduc- plished so far, though impressive, is still past two decades have seen the elucida- tance. This is a basic property of small quite limited. We will ultimately develop tion of entirely new, fundamental physi- conductors that occurs when the wave- operational procedures to help us coax cal principles that govern behavior at the like properties of electrons are coherent- the formation of individual atomic bonds mesoscale. Let’s consider three impor- ly maintained from the “source” to the under more general conditions. But as we tant examples. “drain”—the input to the output—of a try to assemble complex networks of In the fall of 1987 graduate student nanoelectronic device. these bonds, they certainly will affect one Bart J. van Wees of the Delft University Feynman anticipated, in part, such odd behavior: “I have thought about some of the problems of building electric Overview/Nanophysics circuits on a small scale, and the problem Smaller than macroscopic objects but larger than molecules, nanotechnological of resistance is serious....” But the ex- devices exist in a unique realm—the mesoscale—where the properties of matter perimental discoveries pointed out some- are governed by a complex and rich combination of classical physics and thing truly new and fundamental: quan- quantum mechanics. tum mechanics can completely govern Engineers will not be able to make reliable or optimal nanodevices until they the behavior of small electrical devices. comprehend the physical principles that prevail at the mesoscale. Direct manifestations of quantum Scientists are discovering mesoscale laws by fashioning unusual, complex mechanics in such devices were envi- systems of atoms and measuring their intriguing behavior. sioned back in 1957 by Rolf Landauer, Once we understand the science underlying nanotechnology, we can fully a theoretician at IBM who pioneered realize the prescient vision of Richard Feynman: that nature has left plenty of ideas in nanoscale electronics and in the room in the nanoworld to create practical devices that can help humankind. physics of computation. But only in the mid-1980s did control over materials

50 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. er, the coulomb blockade phenomenon is becoming the rule, rather than the ex- ception, in weakly coupled nanoscale devices. This is especially true in experi- ments in which electric currents are passed through individual molecules. These molecules can act like coulomb is- lands by virtue of their weak coupling to electrodes leading back to the macro- world. Using this effect to advantage and obtaining robust, reproducible cou- pling to small molecules (in ways that can actually be engineered) are among the important challenges in the new field of molecular electronics. In 1990, against this backdrop, I was at Bell Communications Research study- ing electron transport in mesoscopic semiconductors. In a side project, my colleagues Larry Schiavone and Axel NANOBRIDGE DEVICE allowed Caltech physicists to first observe the quantization of thermal Scherer and I began developing tech- conductance—a fundamental limit to heat flow in minute objects. Four holes (black) etched into a silicon niques that we hoped would elucidate nitride membrane defined an isolated thermal reservoir (central green square) suspended by four narrow the quantum nature of heat flow. The bridges. One gold transducer (yellow) electrically heated this reservoir; the second measured its work required much more sophisticated temperature. Thin superconducting films (blue) on top of the bridges electrically connected the nanostructures than the planar devices transducers to off-chip instrumentation but carried no heat. The reservoir therefore cooled only through used to investigate mesoscopic electron- the silicon nitride bridges, which were so narrow that they passed only the lowest-energy heat waves. ics. We needed freely suspended devices, structures possessing full three-dimen- and nanofabrication begin to provide Projections from the International Tech- sional relief. Ignorance was bliss; I had access to this regime in the laboratory. nology Roadmap for Semiconductors, no idea the experiments would be so in- The 1987 discoveries heralded the hey- prepared by long-range thinkers in the volved that they would take almost a day of “mesoscopia.” industry, indicate that by 2014 the min- decade to realize. A second important example of new- imum feature size for transistors in com- The first big strides were made after ly uncovered mesoscale laws that have puter chips will decrease to 20 nanome- I moved to Caltech in 1992, in a collab- led to nascent nanotechnology was first ters. At this dimension, each switching oration with John Worlock of the Uni- postulated in 1985 by Konstantin Likha- event will involve the equivalent of only versity of Utah and two successive post- rev, a young physics professor at Moscow about eight electrons. Designs that prop- docs in my group. Thomas Tighe devel- State University working with postdoc- erly account for single-electron charging oped the methods and devices that gen- toral student Alexander Zorin and un- will become crucial. erated the first direct measurements of dergraduate Dmitri Averin. They antic- By 1987 advances in nanofabrica- heat flow in nanostructures. Subsequent- ipated that scientists would be able to tion allowed Theodore A. Fulton and ly, Keith Schwab revised the design of

COLORIZATION BY M. L. ROUKES control the movement of single electrons Gerald J. Dolan of Bell Laboratories to the suspended nanostructures and put in on and off a so-called coulomb island— construct the first single-electron tran- place ultrasensitive superconducting in-

Caltech; a conductor weakly coupled to the rest sistor [see illustration on page 56]. The strumentation to interrogate them at ul- of a nanocircuit. This could form the ba- single-electron charging they observed, tralow temperatures, where the effects sis for an entirely new type of device, now called the coulomb blockade, has could be seen most clearly. called a single-electron transistor. The since been seen in a wide array of struc- In the late summer of 1999 Schwab physical effects that arise when putting tures. As experimental devices get small- finally began observing heat flow through a single electron on a coulomb island be- come increasingly robust as the island is MICHAEL ROUKES, professor of physics at the California Institute of Technology, heads a scaled downward. In very small devices, group studying nanoscale systems. Among the holy grails his team is chasing are a po- these single-electron charging effects can tential billionfold improvement in present-day calorimetry, which would allow observation completely dominate the current flow. of the individual heat quanta being exchanged as nanodevices cool, and a potential Such considerations are becoming quadrillion-fold increase in the sensitivity of magnetic resonance imaging, which would en- THE AUTHOR K. C. SCHWAB, E. A. HENRIKSEN AND M. L. ROUKES increasingly important technologically. able complex biomolecules to be visualized with three-dimensional atomic resolution.

www.sciam.com SCIENTIFIC AMERICAN 51 Copyright 2001 Scientific American, Inc. ONE STEP AT A TIME scale systems behave. The discovery of the electrical and thermal conductance THE QUANTIZATION OF ELECTRICAL CONDUCTANCE quanta and the observation of the cou- In 1987 Bart J. van Wees and his collabo- lomb blockade are true discontinuities— rators at the Delft University of Technolo- abrupt changes in our understanding. gy and Philips Research Laboratories Today we are not accustomed to calling (both in the Netherlands) built a novel our discoveries “laws.” Yet I have no structure (micrograph) that revealed a doubt that electrical and thermal con- basic law governing nanotech circuits. ductance quantization and single-elec- Gold gate electrodes (bright areas) were tron-charging phenomena are indeed placed atop a semiconductor substrate among the universal rules of nano- (dark background). Within the substrate, design. They are new laws of the nano- ELECTRON GAS world. They do not contravene but aug- a planar sheet of charge carriers, called a (BELOW SURFACE) REGION DEPLETED two-dimensional electron gas, was creat- OF ELECTRONS ment and clarify some of Feynman’s (BELOW SURFACE) ed about 100 nanometers below the sur- GOLD GATE B original vision. Indeed, he seemed to face. The gates and the gas acted like the have anticipated their emergence: “At plates of a capacitor. the atomic level, we have new kinds of When a negative voltage bias was forces and new kinds of possibilities, applied to the gates, electrons within the new kinds of effects. The problems of gas underneath the gates, and slightly manufacture and reproduction of mate- DEPLETION rials will be quite different.” beyond the gates’ periphery, were A EDGE ELECTRON FLOW We will encounter many more such pushed away. (The diagram shows this THROUGH CONSTRICTION state.) When increasing negative voltage discontinuities on the path to true nano- was applied, this “depletion edge” be- technology. These welcome windfalls came more pronounced. At a certain threshold, carriers on either side of the constric- will occur in direct synchrony with ad- tion (between points A and B) became separated, and the conductance through the vances in our ability to observe, probe device was zero. From this threshold level, conductance did not resume smoothly. In- and control nanoscale structures. It stead it increased in stepwise fashion, where the steps occurred at values deter- would seem wise, therefore, to be rather mined by twice the charge of the electron squared, divided by Planck’s constant. This modest and circumspect about forecast- ratio is now called the electrical conductance quantum, and it indicates that electric ing nanotechnology. current flows in nanocircuits at rates that are quantized. The Boon and Bane of Nano THE NANOWORLD is often portrayed silicon nitride nanobridges [see illustra- transistors and the clock rates (frequen- by novelists, futurists and the popular tion on page 51]. Even in these first data cies) of microprocessors, the problem of press as a place of infinite possibilities. the fundamental limit to heat flow in keeping microchips cool to avoid com- But as you’ve been reading, this domain mesoscopic structures emerged. The plete system failure is becoming monu- is not some ultraminiature version of the manifestation of this limit is now called mental. This will only become further Wild West. Not everything goes down the thermal conductance quantum. It de- exacerbated in nanotechnology. there; there are laws. Two concrete il- ) termines the maximum rate at which Considering even this complexity, lustrations come from the field of nano-

heat can be carried by an individual Feynman said, “Let the bearings run dry; electromechanical systems (NEMS), in diagram wavelike mechanical vibration, span- they won’t run hot because the heat es- which I am currently active. ning from the input to the output of a capes away from such a small device very, Part of my research is directed to- nanodevice. It is analogous to the elec- very rapidly.” But our experiments indi- ward harnessing small mechanical de- trical conductance quantum but governs cate that nature is a little more restrictive. vices for sensing applications. Nanoscale ; NINA FINKEL ( the transport of heat. The thermal conductance quantum can structures appear to offer revolutionary This quantum is a significant para- place limits on how effectively a very potential; the smaller a device, the more meter for nanoelectronics; it represents small device can dissipate heat. What susceptible its physical properties to al- the ultimate limit for the power-dissipa- Feynman envisioned can be correct only teration. One example is resonant de- tion problem. In brief, all “active” de- if the nanoengineer designs a structure so tectors, which are frequently used for vices require a little energy to operate, as to take these limits into account. sensing mass. The vibrations of a tiny and for them to operate stably without From the three examples above, we mechanical element, such as a small can- Delft University of Technology overheating, we must design a way to can arrive at just one conclusion: we are tilever, are intimately linked to the ele- WEES extract the heat they dissipate. As engi- only starting to unveil the complex and ment’s mass, so the addition of a minute VAN

neers try to ever increase the density of wonderfully different ways that nano- amount of foreign material (the “sam- B. J.

54 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. The difficulties in communication between the nanoworld and the macroworld represent a central issue in the development of nanotechnology.

ple” being weighed) will shift the reso- cheap, ultraminiature “smart” sensors more than $1 billion, and it would prob- nant frequency. Recent work in my lab that could continuously monitor all of ably have to be retooled to be useful. But by postdoc Kamil Ekinci shows that the important functions in hospitals, in I am certain that the revolutionary nanoscale devices can be made so sensi- manufacturing plants, on aircraft, and prospects of nanoscale devices will even- tive that “weighing” individual atoms so on. The idea of ultraminiature devices tually compel such changes. and molecules becomes feasible. that drain their batteries extremely slow- But there is a dark side. Gaseous ly, especially ones with sufficient com- Monumental Challenges atoms and molecules constantly adsorb putational power to function autono- CERTAINLY A HOST of looming is- and desorb from a device’s surfaces. If mously, has great appeal. sues will have to be addressed before we the device is macroscopic, the resulting But here, too, there is a dark side. The can realize the potential of nanoscale de- fractional change in its mass is negligi- regime of ultralow power is quite foreign vices. Although each research area has ble. But the change can be significant for to present-day electronics. Nanoscale de- its own concerns, some general themes nanoscale structures. Gases impinging vices will require entirely new system ar- emerge. Two challenges fundamental to on a resonant detector can change the chitectures that are compatible with my current work on nanomechanical resonant frequency randomly. Appar- amazingly low power thresholds. This systems, for instance, are relevant to ently, the smaller the device, the less sta- prospect is not likely to be received hap- nanotechnology in general. ble it will be. This instability may pose pily by the computer industry, with its Challenge I: Communication between a real disadvantage for various types of overwhelming investment in current de- the macroworld and the nanoworld. futuristic electromechanical signal-pro- vices and methodology. A new semicon- NEMS are incredibly small, yet their mo- cessing applications. Scientists might be ductor processing plant today costs tion can be far smaller. For example, a able to work around the problem by, for nanoscale beam clamped on both ends example, using arrays of nanomechani- vibrates with minimal harmonic distor- cal devices to average out fluctuations. tion when its vibration amplitude is kept But for individual elements, the problem below a small fraction of its thickness. seems inescapable. For a 10-nanometer-thick beam, this am- A second example of how “not plitude is only a few nanometers. Build- everything goes” in the nanoworld re- ing the requisite, highly efficient trans- lates more to economics. It arises from ducers to transfer information from such the intrinsically ultralow power levels at a device to the macroworld involves read- which nanomechanical devices operate. ing out information with even greater Physics sets a fundamental threshold for precision. the minimum operating power: the ubiq- Compounding this problem, the nat- uitous, random thermal vibrations of a ural frequency of the vibration increases mechanical device impose a “noise floor” as the size of the beam is decreased. So below which real signals become in- to usefully track the device’s vibrations, creasingly hard to discern. In practical the ideal NEMS transducer must be ca- use, nanomechanical devices are opti- pable of resolving extremely small dis- mally excited by signal levels 1,000-fold placements, in the picometer-to-fem- or a millionfold greater than this thresh- tometer (trillionth to quadrillionth of a old. But such levels are still a millionth to meter) range, across very large band- COURTESY OF CALTECH ARCHIVES a billionth the amount of power used for widths (extending into the microwave conventional transistors. range). These twin requirements pose a The advantage, in some future nano- RICHARD FEYNMAN predicted the rise of nano- truly monumental challenge, one much mechanical signal-processing system or technology in a landmark 1959 talk at Caltech. more significant than those faced so far computer, is that even a million nano- “The principles of physics,” he said, “do not in MEMS work. A further complication

Ohio Historical Society; mechanical elements would dissipate speak against the possibility of maneuvering is that most of the methodologies from only a millionth of a watt, on average. things atom by atom.” But he also anticipated MEMS are inapplicable; they simply Such ultralow power systems could lead that unique laws would prevail; they are finally do not scale down well to nanometer

JOE MUNROE to wide proliferation and distribution of being discovered today. dimensions.

www.sciam.com SCIENTIFIC AMERICAN 55 Copyright 2001 Scientific American, Inc. In each new regime, some wonderful scientific phenomenon emerges. But then a thorny host of underlying, equally unanticipated problems appear.

These difficulties in communication essence, individual macromolecules. Al- alizable in our lifetimes will entail some between the nanoworld and the macro- though the grand vision of futurists may form of reporting up to the macroworld world represent a generic issue in the de- involve self-programmed nanobots that and feedback and control back down. velopment of nanotechnology. Ulti- need direction from the macroworld The communication problem will re- mately, the technology will depend on only when they are first wound up and main central. robust, well-engineered information set in motion, it seems more likely that Orchestrating such communication transfer pathways from what are, in most nanotechnological applications re- immediately invokes the very real pos- sibility of collateral damage. Quantum theory tells us that the process of mea- TAKING CHARGE suring a quantum system nearly always perturbs it. This can hold true even SINGLE ELECTRONICS when we scale up from atoms and mol- Advances in nanofabrication allowed Theodore A. Fulton and Gerald J. Dolan to build ecules to nanosystems comprising mil- a single-electron transistor at Bell Laboratories in 1987 (micrograph). In this lions or billions of atoms. Coupling a structure the controlled movement of individual electrons through a nanodevice was nanosystem to probes that report back first achieved. At its heart was a coulomb island, a metallic electrode isolated from to the macroworld always changes the its counter-electrodes by thin insulating oxide barriers (diagram). The counter- nanosystem’s properties to some degree, electrodes led up to the macroscale laboratory instrumentation used to carry out the rendering it less than ideal. Introducing experiments. An additional gate electrode (visible in the diagram but not the the transducers required for communi- micrograph) was offset from the coulomb island by a small gap; it allowed direct cation will do more than just increase control of the charge introduced to the island. Electric current flowed through the the nanosystem’s size and complexity. device from one counter-electrode to another, as in a conventional circuit, but here it They will also necessarily extract some was limited by the stepwise hopping of electrons onto and off the coulomb island. energy to perform their measurements Fulton and Dolan’s experiments demonstrate both the fundamental physics of and can degrade the nanosystem’s per- single-electron charging and the potential of these devices as ultrasensitive formance. Measurement always has its electrometers: instruments that can easily detect individual electron charges. price. Circuits that switch one electron at a time could someday form the basis for an Challenge II: Surfaces. As we shrink entirely new class of nanoelectronics. The advent of such single electronics, MEMS to NEMS, device physics be- however, also presages problems that will have to be faced as conventional comes increasingly dominated by the sur- electronic circuits are shrunk to the nanoscale. faces. Much of the foundation of solid- state physics rests on the premise that the ) surface-to-volume ratio of objects is in- ELECTRON

finitesimal, meaning that physical prop- drawing COUNTER-ELECTRODE erties are always dominated by the physics of the bulk. Nanoscale systems INSULATING BARRIER are so small that this assumption breaks

down completely. BRYAN CHRISTIE ( COULOMB ISLAND For example, mechanical devices pat- terned from single-crystal, ultrapure ma- terials can contain very few (even zero) crystallographic defects and impurities. Bell Laboratories; My initial hope was that, as a result, there would be only very weak damping of mechanical vibrations in monocrys- talline NEMS. But as we shrink mechan- GATE ELECTRODE ical devices, we repeatedly find that acoustic energy loss seems to increase in

proportion to the increasing surface-to- T. A. FULTON AND G. J. DOLAN

56 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. true as we have pushed to decreased size, enhanced sensitivity, greater spatial res- olution, higher magnetic and electric fields, lower pressure and temperature, and so on. It is at the heart of why pro- jecting forward too many orders of mag- nitude is usually perilous. And it is what should imbue us with a sense of humili- ty and proportion at this, the beginning of our journey. Nature has already set the rules for us. We are out to understand and employ her secrets. Once we head out on the quest, na- ture will frequently hand us what initial- ly seems to be nonsensical, disappoint- ing, random gibberish. But the science in the glitches often turns out to be even more important than the grail motivat- ing the quest. And being proved the fool in this way can truly be the joy of doing NANOMECHANICAL AMPLIFIER overcomes the vexing problem of communication with the macroworld science. If we had the power to extrapo- by providing up to 1,000-fold amplification of weak forces. Two suspended bridges (left and right) of late everything correctly from the outset, monocrystalline silicon carbide support the central crossbridge, to which the signal force is applied. the pursuit of science would be utterly Thin-film electrodes (silver) atop these structures provide very sensitive readouts of nanoscale motion. dry and mechanistic. The delightful truth is that, for complex systems, we do not, volume ratio. This result clearly impli- Payoff in the Glitches and ultimately probably cannot, know cates surfaces in the devices’ vibrational FUTURISTIC THINKING is crucial to everything that is important. energy-loss processes. In a state-of-the-art making the big leaps. It gives us some Complex systems are often exquis- silicon beam measuring 10 nanometers wild and crazy goals—a holy grail to itely sensitive to a myriad of parameters wide and 100 nanometers long, more chase. And the hope of glory propels us beyond our ability to sense and record— than 10 percent of the atoms are at or onward. Yet the famous 19th-century much less control—with sufficient regu- next to the surface. It is evident that these chemist Kekulé once said, “Let us learn larity and precision. Scientists have stud- atoms will play a central role, but under- to dream, gentlemen, then perhaps we ied, and in large part already understand, standing precisely how will require a ma- shall find the truth.... But let us beware matter down to the fundamental parti- jor, sustained effort. of publishing our dreams before they cles that make up the neutrons, protons In this context, so-called nanotube have been put to the proof by the wak- and electrons that are of crucial impor- structures, which have been heralded ing understanding.” tance to chemists, physicists and engi- lately, look ideal. A nanotube is a crys- This certainly holds for nanoscience. neers. But we still cannot deterministi- talline, rodlike material perfect for build- While we keep our futuristic dreams cally predict how arbitrarily complex as- ing the miniature vibrating structures of alive, we also need to keep our expecta- semblages of these three elemental interest to us. And because it has no tions realistic. It seems that every time we components will finally behave en masse. chemical groups projecting outward gain access to a regime that is a factor of For this reason, I firmly believe that it is along its length, one might expect that in- 10 different—and presumably “better”— on the foundation of the experimental teraction with “foreign” materials at its two things happen. First, some wonder- science already under way, in intimate

COLORIZATION BY M. L. ROUKES surfaces would be minimal. Apparently ful, unanticipated scientific phenomenon collaboration with theory, that we will not. Although nanotubes exhibit ideal emerges. But then a thorny host of un- build the road to true nanotechnology.

Caltech; characteristics when shrouded within derlying, equally unanticipated new Let’s keep our eyes open for surprises pristine, ultrahigh vacuum environments, problems appear. This pattern has held along the way! samples in more ordinary conditions, where they are exposed to air or water MORE TO EXPLORE vapor, evince electronic properties that Nanoelectromechanical Systems Face the Future. Michael Roukes in Physics World, Vol. 14, No. 2; are markedly different. Mechanical prop- February 2001. Available at physicsweb.org/article/world/14/2/8 erties are likely to show similar sensitiv- The author’s group: www.its.caltech.edu/~nano ity. So surfaces definitely do matter. It Richard Feynman’s original lecture “There’s Plenty of Room at the Bottom” can be read at

D. A. HARRINGTON AND M. L. ROUKES would seem there is no panacea. www.its.caltech.edu/~feynman

www.sciam.com SCIENTIFIC AMERICAN 57 Copyright 2001 Scientific American, Inc. Copyright 2001 Scientific American, Inc. NANOELECTRONICS

The

IncredibleShrinking Circuit

RESEARCHERS HAVE BUILT NANOTRANSISTORS AND NANOWIRES. NOW THEY JUST NEED TO FIND A WAY TO PUT THEM ALL TOGETHER

BY CHARLES M. LIEBER

NANOWIRES, each about five to 10 nanometers in diameter, may represent the future of electronics. They are the brown lines, made of indium phosphide, connecting the gold electrodes in this micrograph. These wires have been put to truly diverse uses—as memory and logic and as arrays of light-emitting diodes.

www.sciam.com SCIENTIFIC AMERICAN 59 Copyright 2001 Scientific American, Inc. Do we really need to keep on making circuits els of organization. The basic building block is usually the tran- smaller? The miniaturization of silicon microelectronics seems sistor or its nanoequivalent—a switch that can turn an electric so inexorable that the question seldom comes up—except current on or off as well as amplify signals. In microelectron- maybe when we buy a new computer, only to find that it be- ics, transistors are made out of chunks of semiconductor—a comes obsolete by the time we leave the store. A state-of-the- material, such as impure silicon, that can be manipulated to art microprocessor today has more than 40 million transistors; flip between conducting and nonconducting states. In nano- by 2015 it could have nearly five billion. Yet within the next electronics, transistors might be organic molecules or nano- two decades this dramatic march forward will run up against scale inorganic structures. scientific, technical and economic limits. A first reaction might The next level of organization is the interconnection—the be, So what? Aren’t five billion transistors enough already? wires that link transistors together in order to perform arith- Yet when actually confronted with those limits, people will metic or logical operations. In microelectronics, wires are met- no doubt want to go beyond them. Those of us who work to al lines typically hundreds of nanometers to tens of microns keep computer power growing are motivated in part by the in width deposited onto the silicon; in nanoelectronics, they are sheer challenge of discovering and conquering unknown ter- nanotubes or other wires as narrow as one nanometer. ritory. But we also see the potential for a revolution in medi- At the top level is what engineers call architecture—the over- The use of molecules for electronic devices remained theoretical until a recent confluence of advances in chemistry, physics and engineering.

cine and so many other fields, as extreme miniaturization en- all way the transistors are interconnected, so that the circuit can ables people and machines to interact in ways that are not pos- plug into a computer or other system and operate independently sible with existing technology. of the lower-level details. Nanoelectronics researchers have not As the word suggests, microelectronics involves compo- quite gotten to the point of testing different architectures, but nents that measure roughly one micron on a side (although we do know what abilities they will be able to exploit and what lately the components have shrunk to a size of almost 100 weaknesses they will need to compensate for. nanometers). Going beyond microelectronics means more than In other ways, however, microelectronics and nanoelec- simply shrinking components by a factor of 10 to 1,000. It also tronics could not be more different. To go from one to the oth- involves a paradigm shift for how we think about putting er, many believe, will require a shift from top-down manufac- everything together. turing to a bottom-up approach. To build a silicon chip today, ) Microelectronics and nanoelectronics both entail three lev- fabrication plants start with a silicon crystal, lay down a pat- page 58 tern using a photographic technique known as lithography, ( Nanoelectronics and etch away the unwanted material using acid or plasma. Overview/ That procedure simply doesn’t have the precision for devices Silicon chips, circuit boards, soldering irons: these are the that are mere nanometers in width. Instead researchers use the

icons of modern electronics. But the electronics of the future methods of synthetic chemistry to produce building blocks by Harvard University may look more like a chemistry set. Conventional techniques the mole (6 × 10 23 pieces) and assemble a portion of them into can shrink circuits only so far; engineers will soon need to progressively larger structures. So far the progress has been im- shift to a whole new way of organizing and assembling pressive. But if this research is a climb up Mount Everest, we electronics. One day your computer may be built in a beaker. have barely just reached the base camp. Researchers have created nanometer-scale electronic components—transistors, diodes, relays, logic gates— Smallifying Machines from organic molecules, carbon nanotubes and THE USE OF MOLECULES for electronic devices was sug- semiconductor nanowires. Now the challenge is to wire gested more than a quarter of a century ago in a seminal pa- these tiny components together. per by Avi Aviram of IBM and Mark A. Ratner of North- Unlike conventional circuit design, which proceeds from western University. By tailoring the atomic structures of or- blueprint to photographic pattern to chip, nanocircuit design ganic molecules, they proposed, it should be possible to will probably begin with the chip—a haphazard jumble of as concoct a transistorlike device. But their ideas remained large- many as 1024 components and wires, not all of which will ly theoretical until a recent confluence of advances in chem- even work—and gradually sculpt it into a useful device. istry, physics and engineering.

Of all the groups that have turned Aviram and Ratner’s IMAGE BY FELICE FRANKEL, WITH TECHNICAL HELP FROM YU HUANG

60 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. NANOTRANSISTORS

ROTOXANE BENZENETHIOL MOLECULAR TRANSISTORS could be the building blocks of electronics on the nanometer scale. Each of the two molecules shown here conducts electricity like a tiny wire once a SWITCH OFF chemical reaction— oxidation reduction—alters SWITCH OFF its atomic configuration and switches it on. In the SWITCH ON diagram, each stick SWITCH ON ADDED represents a chemical ELECTRON bond; each intersection of two sticks represents a carbon atom; and each ball represents an atom other than carbon.

ideas into reality, two teams—one at the University of Califor- My own research group at Harvard University is one of sev- nia at Los Angeles and Hewlett-Packard, the other at Yale, Rice eral that have focused not on organic molecules but on long, and Pennsylvania State universities—stand out. Within the past thin inorganic wires. The best-known example is the carbon year, both have demonstrated that thousands of molecules clus- nanotube, which is typically about 1.4 nanometers in diame- tered together can carry electrons from one metal electrode to ter [see “Nanotubes for Electronics,” by Philip G. Collins and another. Each molecule is about 0.5 nanometer wide and one Phaedon Avouris; Scientific American, December 2000]. or more nanometers long. Both groups have shown that the Not only can these nanoscale wires carry much more current, clusters can behave as on/off switches and might thus be usable atom for atom, than ordinary metal wires, they can also act as in computer memory; once on, they will stay on for 10 min- tiny transistors. By functioning both as interconnections and as utes or so [see “Computing with Molecules,” by Mark A. Reed components, nanowires kill two birds with one stone. Anoth- and James M. Tour; Scientific American, June 2000]. That er advantage is that they can exploit the same basic physics as may not sound like a long time, but computer memory typical- standard silicon microelectronics, which makes them easier to ly loses its information instantly when the power is turned off; understand and manipulate. even when the power is on, the stored information leaks away In 1997 Cees Dekker’s group at the Delft University of and must be “refreshed” every 0.1 second or so. Technology in the Netherlands and Paul L. McEuen’s group, Although the details differ, the switching mechanism for then at the University of California at Berkeley, independent- both molecules is believed to involve a well-understood chem- ly reported highly sensitive transistors made from metallic car- ical reaction, oxidation reduction, in which electrons shuffle bon nanotubes. These devices could be turned on and off by a among atoms within the molecule. The reaction puts a twist in the molecule, blocking electrons as surely as a kink in a hose CHARLES M. LIEBER spent much of his childhood building—and blocks water [see illustration above]. In the “on” position, the breaking—stereos, cars and model airplanes. He is now the Mark clusters of molecules may conduct electricity as much as 1,000 Hyman Professor of Chemistry at Harvard University, where he di- times better than in the “off” position. That ratio is actually rects a group of 25 undergraduate, graduate and postdoctoral re- rather low compared with that of typical semiconductor tran- searchers who focus on nanoscale science and technology.

sistors, whose conductivity varies a millionfold. Researchers are THE AUTHOR Lieber recently founded NanoSys, Inc., with Larry Bock of CW Ven- now looking for other molecules with even better switching tures and Hongkun Park of Harvard. Their modest goal: to revolu- properties and are also working to understand the switching tionize chemical and biological sensing, electronics, optoelec-

SLIM FILMS process itself. tronics, and information storage.

www.sciam.com SCIENTIFIC AMERICAN 61 Copyright 2001 Scientific American, Inc. single electron but required very low temperatures to operate. a diode, which allows electric current to pass in one direction This past July Dekker’s team swept away this limitation. The only. Finally, my group has demonstrated a very different type researchers used an atomic force microscope to create a single- of switch, a nanoscale electromechanical relay. electron transistor that could function at room temperature. Dekker and his co-workers have also fashioned a more con- Hot Wire ventional field-effect transistor, the building block of most in- A MAJOR PROBLEM with nanotubes is that they are difficult tegrated circuits today, out of a carbon nanotube, and McEuen’s to make uniform. Because a slight variation in diameter can group has combined metallic and semiconductor nanotubes into spell the difference between a conductor and a semiconductor,

SINGLE- DNA Computing STRANDED DNA

WHY LIMIT OURSELVES TO ELECTRONICS? Most efforts CHIP to shrink computers assume that these machines will continue to operate much as they do today, using 1 Single DNA strands are attached to a silicon chip. electrons to carry information and transistors to They encode all possible values of the variables in an process it. Yet a nanoscale computer could operate by equation that the researchers want to solve. completely different means. One of the most exciting possibilities is to exploit the carrier of genetic COMPLEMENTARY information in living organisms, DNA. STRAND The molecule of life can store vast quantities of INVALID VALID data in its sequence of four bases (adenine, thymine, SOLUTION SOLUTION guanine and cytosine), and natural enzymes can manipulate this information in a highly parallel manner. The power of this approach was first brought to light by 2 Copies of a complementary strand—which computer scientist Leonard M. Adleman in 1994. He encodes the first clause of the equation—are poured showed that a DNA-based computer could solve a type onto the chip. These copies attach themselves to of problem that is particularly difficult for ordinary any strand that represents a valid solution of the clause. computers—the Hamiltonian path problem, which is Any invalid solutions remain a single strand. related to the infamous traveling-salesman problem [see “Computing with DNA,” by Leonard M. Adleman; SCIENTIFIC AMERICAN, August 1998]. ENZYME Adleman started by creating a chemical solution of DNA. The individual DNA molecules encoded every possible pathway between two points. By going through a series of separation and amplification steps, 3 An enzyme removes all the single strands. Adleman weeded out the wrong paths—those, for example, that contained points they were not supposed to contain—until he had isolated the right one. More recently, Lloyd M. Smith’s group at the University of Wisconsin–Madison implemented a similar algorithm using gene chips, which may lend themselves better to practical computing (diagram). 4 Other processes melt away the added complementary Despite the advantages of DNA computing for strands. These steps are repeated with all the clauses otherwise intractable problems, many challenges of the equation. remain, including the high incidence of errors caused by base-pair mismatches and the huge number of DNA nanoelements needed for even a modest computation. DNA computing may ultimately merge with other types of nanoelectronics, taking advantage of the integration and sensing made possible by 5 The DNA strand that survives this whole process nanowires and nanotubes. —C.M.L. represents the solution to the whole equation. BRYAN CHRISTIE

62 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Soon nanodevices may have useful applications—for example, as ultrasensitive detectors of gas molecules and biological compounds.

a large batch of nanotubes may contain only a few working de- wires themselves must be organized into, for example, a two- vices. In April of this year Phaedon Avouris and his colleagues dimensional array. In a report published earlier this year, Duan at the IBM Thomas J. Watson Research Center came up with and another member of my team, Yu Huang, made a very sig- a solution. They started with a mixture of conducting and nificant breakthrough: they assembled nanocircuits by means semiconducting nanotubes and, by applying a current between of fluid flows. Just as sticks and logs can flow down a river, metal electrodes, selectively burned away the conducting ones nanoscale wires can be drawn into parallel lines using fluids. until just semiconducting ones were left. The solution is only In my lab we have used ethanol and other solutions and con- partial, however, because it requires the use of conventional trolled the liquid flow by passing it through channels molded lithography to wire up the random nanotube array and then into polymer blocks, which can be easily placed on the substrate test and modify each of the individual elements, which would where we wish to assemble devices. ultimately number in the billions. The process creates interconnections in the direction of the My group has also been working on a different type of fluid flow: if the flow is along only one channel, then parallel nanoscale wire, which we term the semiconductor nanowire. It nanowires are formed. To add wires in other directions, we is about the same size as a carbon nanotube, but its composi- redirect the flow and repeat the process, building up addition- tion is easier to control precisely. To synthesize these wires, we al layers of nanowires. For instance, to produce a right-angle start with a metal catalyst, which defines the diameter of the grid, we first lay down a series of parallel nanowires, then ro- growing wire and serves as the site where molecules of the de- tate the direction of flow by 90 degrees and lay down another sired material tend to collect. As the nanowires grow, we in- series. By using wires of different compositions for each layer, corporate chemical dopants (impurities that add or remove elec- we can rapidly assemble an array of functional nanodevices us- trons), thereby controlling whether the nanowires are n-type ing equipment not much more sophisticated than a high school (having extra electrons) or p-type (having a shortage of elec- chemistry lab. A grid of diodes, for example, consists of a lay- trons or, equivalently, a surfeit of positively charged “holes”). er of conducting nanotubes above a layer of semiconductor The availability of n- and p-type materials, which are the nanotubes, or a layer of n-type nanowires atop a layer of p- essential ingredients of transistors, diodes and other electron- type nanowires. In both cases, each junction serves as a diode. ic devices, has opened up a new world for us. We have assem- Our approach, which is similar to that being pursued by the bled a wide range of devices, including both major types of team at U.C.L.A. and Hewlett-Packard, is deterministic. We are transistors (field-effect and bipolar); inverters, which transform trying to create arrays with a certain predictable behavior. Form a “0” signal to a “1”; and light-emitting diodes, which pave follows function. An alternative proposed by the group at Rice, the way for optical interconnections. Our bipolar transistors Yale and Penn State is to allow blocks of devices and wires to were the first molecular-scale devices ever to amplify a current. interconnect at random. Later, the ensemble can be analyzed to A recent advance in my lab by Xiangfeng Duan has been the determine how it might be used for storage or computation. In assembly of memory from crisscrossing n- and p-type this case, function follows form. The problem with this proce- nanowires. The memory can store information for 10 minutes dure is that it would take a huge effort to map a complex net- or longer by trapping charge at the interface between the cross- work and figure out what use it could be put to. ing nanowires [see illustration on next page]. Intimately linked to all these efforts is the development of architectures that best exploit the unique features of nanoscale Breaking the Logjam devices and the capabilities of bottom-up assembly. Although BUILDING UP AN ARSENAL of molecular and nanoscale we can make unfathomable numbers of dirt-cheap nanostruc- devices is just the first step. Interconnecting and integrating tures, the devices are much less reliable than their microelec- these devices is perhaps the much greater challenge. First, the tronic counterparts, and our capacity for assembly and orga- nanodevices must be connected to molecular-scale wires. To nization is still quite primitive. date, organic-molecule devices have been hooked up to con- In collaboration with André DeHon of the California In- ventional metal wires created by lithography. It will not be easy stitute of Technology, my group has been working on highly to substitute nanowires, because we do not know how to make simplified architectures that can be generalized for universal a good electrical connection without ruining these tiny wires computing machines. For memory, the architecture starts with in the process. Using nanowires and nanotubes both for the de- a two-dimensional array of crossed nanowires or suspended vices and for the interconnections would solve that problem. electromechanical switches in which one can store information Second, once the components are attached to nanowires, the at each cross point. The same basic architecture is being pur-

www.sciam.com SCIENTIFIC AMERICAN 63 Copyright 2001 Scientific American, Inc. NANOWIRE ARRAY

CRISSCROSSING NANOWIRES “ON” JUNCTION neatly solves a major problem in ELECTRODE molecular-scale electronics: How do you connect wires to components such as transistors or diodes? The wires do double duty, serving both as wires and as components. Each junction is a component, in this case, a miniature relay—an electromechanical SUPPORT switch that is either on (touching) or off (separated). To flip a switch on or “OFF” JUNCTION off, you apply a certain voltage to the two nanowires. The switch will then stay in that position indefinitely. Crisscrossed NANOTUBE semiconductor nanowires have also been used to create switches that are turned on and off electrically, without mechanical motion. And they can form memory and INSULATOR logic arrays—key steps toward the assembly of a nanocomputer.

sued by researchers at U.C.L.A. and Hewlett-Packard, and it range of biological compounds. In our work at Harvard, we resembles the magnetic-core memory that was common in have converted nanowire field-effect transistors into sensors by computers of the 1950s and 1960s. modifying their surfaces with molecular receptors. This tech- nology has the potential of detecting single molecules using only Law of Large Numbers a voltmeter from a hardware store. The small size and sensitiv- TO OVERCOME the unreliability of individual nanodevices, ity of nanowires also make possible the assembly of extremely we may rely on sheer numbers—the gizmos are so cheap that powerful sensors that could, for instance, sequence the entire plenty of spares are always available. Researchers who work human genome on a single chip and serve in minimally invasive on defect tolerance have shown that computing is possible even medical devices. In the nearer term, we could see hybrids of mi- if many of the components fail, although identifying and map- cro and nano: silicon with a nano core—perhaps a high-densi- ping the defects can be slow and time-consuming. Ultimately ty computer memory that retains its contents forever. we hope to partition the enormous arrays into subarrays whose Although substantial work remains before nanoelectronics reliability can be easily monitored. The optimum size of these makes its way into computers, this goal now seems less hazy than subarrays will depend on the defect levels typically present in it was even a year ago. As we gain confidence, we will learn not molecular and nanoscale devices. just to shrink digital microelectronics but also to go where no Another significant hurdle faced by nanoelectronics is digital circuit has gone before. Nanoscale devices that exhibit “bootstrapping.” How do engineers get the circuit to do what quantum phenomena, for example, could be exploited in quan- they want it to? In microelectronics, circuit designers work like tum encryption and quantum computing. The richness of the architects: they prepare a blueprint of a circuit, and a fabrica- nanoworld will change the macroworld. tion plant builds it. In nanoelectronics, designers will have to work like computer programmers. A fabrication plant will cre- MORE TO EXPLORE ate a raw nanocircuit—billions on billions of devices and wires The author’s Web site: cmliris.harvard.edu whose functioning is rather limited. From the outside, it will The Avouris group: www.research.ibm.com/nanoscience look like a lump of material with a handful of wires sticking The Dai group: www-chem.stanford.edu/group/dai out. Using those few wires, engineers will somehow have to The DeHon group: www.cs.caltech.edu/~andre configure those billions of devices. Challenges such as this are The Dekker group: www.mb.tn.tudelft.nl/user/dekker what keeps me tremendously excited about the field as a whole. The McEuen group: www.lassp.cornell.edu/lassp_data/mceuen/homepage/welcome.html Even before we solve these problems, nanodevices may have The Penn State team: stm1.chem.psu.edu useful applications. For example, semiconducting carbon nano- The Rice/Yale team: www.jmtour.com and www.eng.yale.edu/reedlab tubes have been used by Hongjie Dai’s group at Stanford Uni- The Smith group: www.chem.wisc.edu/~smith versity to detect gas molecules, and Yi Cui in my group has used The U.C.L.A./Hewlett-Packard team: www.chem.ucla.edu/~schung/hgrp semiconductor nanowires as ultrasensitive detectors for a wide and www.hpl.hp.com/research/qsr/staff/kuekes.html SLIM FILMS

64 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Copyright 2001 Scientific American, Inc. NANOMEDICINE

Less is

moreinMedicine SOPHISTICATED FORMS OF NANOTECHNOLOGY WILL FIND SOME OF THEIR FIRST REAL-WORLD APPLICATIONS IN BIOMEDICAL RESEARCH, DISEASE DIAGNOSIS AND, POSSIBLY, THERAPY BY A. PAUL ALIVISATOS

The 1966 film Fantastic Voyage treated movie- self-replicating nanometer-scale automatons run amok and de- goers to a bold vision of nanotechnology applied to medicine: stroy the entire biological world? through mysterious means, an intrepid team of doctors and In my view, shared by most investigators, such thoughts be- their high-tech submarine were shrunk to minute size so that long squarely in the realm of science fiction. Still, nanotech- they could travel through the bloodstream of an injured patient nology can potentially enhance biomedical research tools—for and remove a life-threatening blood clot in his brain. In the past example, by providing new kinds of labels for experiments 35 years, great strides have been made in fabricating complex done to discover drugs or to reveal which sets of genes are ac- devices at ever smaller scales, leading some people to believe tive in cells under various conditions. Nanoscale devices could, that such forms of medical intervention are possible and that moreover, play a part in quick diagnostic screens and in genet- tiny robots will soon be coursing through everyone’s veins. In- ic tests, such as those meant to determine a person’s suscepti- deed, in some circles the idea is taken so seriously that worries bility to different disorders or to reveal which specific genes are have emerged about the dark side of such technology: Could mutated in a patient’s cancer. Investigators are also studying them as improved contrast agents for noninvasive imaging and HIGHLY MAGNIFIED VIALS contain solutions of quantum dots—semiconduc- as drug-delivery vehicles. The emerging technologies may not be tor nanocrystals—of specific sizes. The precise size of a quantum dot deter- as photogenic as a platelet-size Raquel Welch blasting away at mines the color it emits after exposure to light. By attaching different sizes a clot with a laser beam, but they are every bit as dramatic be- of dots to different biological molecules, investigators can track the activities cause, in contrast, the benefits they offer to patients and re- of many tagged molecules simultaneously. searchers are real. www.sciam.com SCIENTIFIC AMERICAN 67 Copyright 2001 Scientific American, Inc. The technologies may not be as photogenic as a minute Raquel Welch blasting away at a clot, but they are just as dramatic because their benefits are real.

How exactly can nanotechnology do The answer is that this bacterium has Artificial magnetic crystals of similar all these things? The answer hinges on fixed within it a chain of about 20 mag- dimension might soon serve biomedical one’s definitions. All of biology is ar- netic crystals that are each between 35 research in a novel way. Two groups, guably a form of nanotechnology. After and 120 nanometers in diameter. To- one in Germany and the other at my in- all, even the most complicated creature is gether these crystals constitute a minia- stitution, the University of California at made up of tiny cells, which themselves ture compass. Because the magnetic field Berkeley, are exploring the use of mag- are constructed of nanoscale building of the earth is inclined in most places (it netic nanoparticles to detect particular blocks: proteins, lipids, nucleic acids and points not only north but also down- biological entities, such as microorgan- other complex biological molecules. But ward in the Northern Hemisphere and isms that cause disease. by convention the term “nanotechnol- upward in the Southern), a magnetotac- Their method, like many of the tech- ogy” is usually restricted to artificial con- tic bacterium can follow a magnetic field niques applied today, requires suitable structions made, say, from semiconduc- line up or down to its desired destination. antibodies, which bind to specific targets. tors, metals, plastic or glass. A few inor- This compass is a marvel of natural The magnetic particles are affixed, as la- ganic structures of nanometer scale— nanoscale engineering. For one, it is made bels, to selected antibody molecules, minute crystals, for instance—have al- of the perfect material—either magnetite which are then applied to the sample un- ready been commercialized, notably as or greigite, both highly magnetic iron der study. To detect whether the anti- contrast agents. minerals. The use of multiple crystals is bodies have latched onto their target, an also no accident. At very small scales, the investigator applies a strong magnetic Magnetic Attraction larger a magnetic particle is, the longer field (which temporarily magnetizes the NATURE ITSELF provides a beautiful it stays magnetized. But if the particle be- particles) and then examines the speci- illustration of the usefulness of such in- comes too large, it will spontaneously men with a sensitive instrument capable organic crystals in a biological context: form two separate magnetic domains of detecting the weak magnetic fields em- humble magnetotactic (magnetic-sens- with oppositely directed magnetizations. anating from the probes. Labeled anti- ing) bacteria. Such organisms, which live Such a crystal has little overall magneti- bodies that have not docked to the sam- in bodies of water and their muddy bot- zation and does not make for a very ef- ple tumble about so rapidly in solution toms, thrive only at one depth in the wa- fective compass needle. By building its that they give off no magnetic signal. ter or sediment. Above this position, oxy- compass out of crystals that are of just Bound antibodies, however, are unable gen is too abundant for their liking; be- the right size to exist as stable, single to rotate, and together their magnetic low, too scarce. A bacterium that drifts magnetic domains, the bacterium makes tags generate a readily detectable mag- ) away from the right level must swim the best use of every bit of iron it lays netic field.

back, and so, like many of its cousins, the down. Interestingly, when people design Because the unbound probes produce page 66 ( microbe wields a whiplike tail for pro- media for hard-disk storage, they follow no signal, this approach does away with

pulsion. But how does the buoyant cell exactly the same strategy, using magnet- the time-consuming washing steps usual- M.I.T. tell up from down when gravity has es- ic nanocrystals that are of the proper size ly required of such assays. The sensitivi- sentially no effect on it? to be both stable and strong. ty demonstrated with this experimental technique is already better than with stan- Nanomedicine dard methods, and anticipated improve- Overview/ ments in the apparatus should soon boost Nanometer-scale objects made of Nanoparticles could be used to sensitivity by a factor of several hundred. inorganic materials can serve in deliver drugs just where they are Despite these advantages, the mag- biomedical research, disease needed, avoiding the harmful side netic method probably will not com- diagnosis and even therapy. effects that so often result from pletely replace the widespread practice of Biological tests measuring the potent medicines. labeling probes with a fluorescent tag, presence or activity of selected Artificial nanoscale building blocks typically an organic molecule that glows substances become quicker, more may one day be used to help repair with a characteristic hue when it is ener- sensitive and more flexible when such tissues as skin, cartilage and gized by light of a particular color. Col- certain nanoscale particles are put to bone—and they may even help ors are very useful in various diagnostic work as tags, or labels. patients to regenerate organs. and research procedures, such as when

more than one probe needs to be tracked. IMAGE BY FELICE FRANKEL, WITH TECHNICAL HELP FROM K. F. JENSEN, M. G. BAWENDI, C. MURRAY, KAGAN, B. DABBOUSI, J. RODRIGUEZ-VIEGO

68 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. A GRAND PLAN FOR MEDICINE

he National Nanotechnol- 2 GOAL: New Ways to Treat Disease ogy Initiative includes amongT its goals, or “grand challenges,” a host of TUMOR futuristic improvements in the detection, diagnosis and treatment of disease. Some are depicted here. The NANOSHELL goals, many of which are far from being realized, also feature new aids for vision Nanoparticles would deliver treatments to and hearing, rapid tests specifically targeted sites, including places that for detecting disease standard drugs do not reach easily. For example, susceptibility and responses gold nanoshells (spheres) that were targeted to to drugs, and tiny devices tumors might, when hit by infrared light, heat up able to find problems—such enough to destroy the growths. as incipient tumors, infections or heart problems— 3 GOAL: Superior Implants and to relay the information to an external receiver or fix BONE them on the spot.

1 GOAL: Improved Imaging Improved or new contrast agents COATED would detect problems at earlier, IMPLANT more treatable stages. They might, red for instance, reveal tumors ( ) Nanometer-scale modifications of implant surfaces only a few cells in size. would improve implant durability and biocompat- ibility. For instance, an artificial hip coated with nanoparticles might bond to the surrounding bone more tightly than usual, thus avoiding loosening.

The world of modern electronics is rules that restrict the electrons in atoms emit a photon when the electron falls also full of light-emitting materials. to certain discrete energy levels. An or- back to a lower energy level. This phe- Every CD player, for instance, reads the ganic dye molecule absorbs only photons nomenon is quite different from what disc with light from a solid-state laser of light with just the right energy to lift its happens in bulk semiconductors, which diode, which is made of an inorganic electrons from their quiescent state to allow electrons to occupy two broad semiconductor. Imagine carving out a one of the higher levels available to them. bands of energy. Such materials can ab- vanishingly small piece of that material, That is, the incident light must be exact- sorb photons in a broad range of colors a scoop the size of a protein molecule. ly the right wavelength, or color, to do (all those that have enough energy to The result is a semiconductor nanocrys- the job. The molecule will subsequently bridge the gap between these two bands), tal, or, in the talk of the trade, a “quan- tum dot.” Like nanoscale magnetic crys- A. PAUL ALIVISATOS is a professor in the department of chemistry at the University of Cal- tals, these minuscule dots have much to ifornia, Berkeley, where he received his doctorate in physical chemistry in 1986. A Fellow Hybrid Medical Animation offer biomedical researchers. of both the American Association for the Advancement of Science and the American Physi- As the name suggests, quantum dots cal Society, Alivisatos has received many awards for his research on the physical proper- owe their special properties to the weird ties of nanocrystals. He is a founder of Quantum Dot Corporation, which is working to com- THE AUTHOR

JEFF JOHNSON rules of quantum mechanics, the same mercialize the use of semiconductor nanocrystals as fluorescent labels in biomedical tests.

www.sciam.com SCIENTIFIC AMERICAN 69 Copyright 2001 Scientific American, Inc. 70 emits wavelength oflightaquantumdot the thresholdofbandgap.But they absorbphotonsofallenergiesabove mediate case.Likebulksemiconductors, gap energy.Quantumdotsareaninter- wavelength, correspondingtotheband- but theyemitlightonlyatonespecific help diagnosediseaseordiscovernew tors hadanyideathattheseobjectscould devices. Fewofthepioneeringinvestiga- one dayfashionnewelectronicoroptical in the1970s,thinkingthattheymight tire familyofdistinctlycoloredlabels. semiconducting materialcanyieldanen- on thedot’ssize.Hence,asingletypeof Bar Codes”inthebox ontheoppositepage.) distinct labelsforbiological tests.(Seealso“Nano beads. Theiraimistocreateahugevariety of loaded selectionsofdifferentdots into single the dotsthemselves.Researchershave also colors glowatnearlythesamewavelengths as LATEX BEADS hscssfis tde quantum dots Physicists firststudied SCIENTIFIC AMERICAN — its color filled withquantumdotsofsingle drugs drugs within nanometer-scale packages that control the — depends verystrongly medicines For therapy, one might encapsulate subtle semiconductor nanocrystalsofferisless now beachieved.Butthegreatestbenefit and tissuesforlongerintervalsthancan tigators totrackthegoings-onincells compose. Andthisstabilityallowsinves- ical organicmolecules,whichsoonde- citation andlightemissionthancantyp- withstand significantlymorecyclesofex- molecules. Smallinorganiccrystalscan several advantagesoverconventionaldye for quantumdotsas,well,bright. may bebiased,butIviewtheprospects found thiscompany,somyassessments of MelbourneinAustralia.Ihelpedto National LaboratoryandtheUniversity Indiana University,LawrenceBerkeley Massachusetts InstituteofTechnology, niques developedinmylabandatthe tech- licensed tum DotCorporation,has to commercializethistechnology,Quan- of research,buttheyarenowareality. ological systemsdidindeedrequireyears dots thatwouldfunctionproperlyinbi- ology andmedicine.Makingquantum cations ofquantumdotswouldbeinbi- dreamed thatthefirstreal-worldappli- drugs. Andnoneofthemwouldhave quantum dots. Thebeadcould,forin- latex beadfilled withacombinationof interesting possibilities.Imagine asmall tively, butquantumdotsoffer even more all bemonitoredatonce. gized withasinglelightsource, theycan because allofthesecrystalscanbeener- of adifferentsize(andhencecolor).And biological molecules,eachwithacrystal dots makeitpossibletotagavarietyof ferent wavelengthoflight.Butquantum organic dyemustbeexcitedwithadif- ing isdifficulttoachieve,becauseeach be observedsimultaneously.Suchtrack- and frequentlyseveralcomponentsmust H OPN LEADING COMPANY THE The Rainbow Coalition Copyright 2001 Scientific American, Inc. Semiconductor nanocrystalshave This approachisbeingpursued ac- Biological systemsareverycomplex, — ’ they comeinmorecolors. release in sophisticated wa the push together. Because eachparticlehasmul- es tobothtypes ofparticles,linkingthem complete targetsequencereadily attach- binds totheotherhalf.DNA withthe quence; thesecondsetcarries DNA that that bindstoonehalfofthe target se- gold particles.Thefirstsetcarries DNA ticles studdedwithDNA. scheme employs13-nanometergoldpar- cific geneticsequenceinsolution.Their method totestforthepresenceofaspe- who recentlydevelopedaningenious L. LetsingerofNorthwesternUniversity, the workofChadA.MirkinandRobert mens. Anotherexampleemergesfrom genetic compositionofbiologicalspeci- structures usefulforopticallysensingthe way, buttheyarenottheonlynano- soon servebiomedicalresearchersinthis ic materialinthesample. mediately revealthenatureofgenet- ments theother),resultswouldim- precisely, whenonesequencecomple- netic sequencesmatchclosely(ormore cause bindingtakesplaceonlywhenge- that bindtosequencesinthesample.Be- spectral barcodesofthelibraryDNAs ple tothefullbeadedlibraryandread sues. Theywouldsimplyexposethesam- genes wereactiveincertaincellsortis- investigator wantedtofindoutwhich DNA sequences,asmightbedoneifan al inasampleagainstlibraryofknown could easilycomparethegeneticmateri- quences ofgeneticbuildingblocks. molecules composedofdifferentse- which couldbeattached,say,toDNA labels (billions,potentially),eachof low foranenormousnumberofdistinct tral barcode,ifyoulike.Suchbeadsal- lines withprescribedintensities a prismwillproducefivedistinctspectral give offlight,whichwhenspreadoutby tions. Afterthebeadisilluminated,itwill or fivecolors,inavarietyofconcentra- stance, containfivedifferentsizesofdots, The trickhereistousetwosetsof Semiconductor quantumdotsshould With thesekindsofbeads,technicians ETME 2001 SEPTEMBER — ys. a spec-

SHUMING NIE Indiana University BIO-NANOTECH IN ACTION

he items here could one day enhance the speed and power of biomedical tests, such as those used to screenT small samples of material for the presence of particular genetic sequences. For clarity, the images PROBE DNA have not been drawn to scale. GOLD NANOPARTICLE ANTIBODY

MAGNETIC NANOPARTICLES TARGET DNA DIRECTION OF MAGNETIZATION

ANTIBODY BOUND TO TARGET GOLD PARTICLES MAGNETIC TAGS Gold nanoparticles studded with short segments of DNA could Many tests reveal the presence of a molecule or disease- form the basis of an easy-to-read test for the presence causing organism by detecting the binding of an antibody to of a genetic sequence (black) in a sample under study. DNA that target. When antibodies labeled with magnetic nano- complementary to half of such a sequence (red) is attached particles bind to their target on a surface (foreground), brief to one set of particles in solution, and DNA complementary to the exposure to a magnetic field causes these probes collectively other half (blue) is attached to a second set of particles. to give off a strong magnetic signal. Meanwhile unbound anti- If the sequence of interest is present in the sample, it will bind bodies tumble about in all directions, producing no net signal. to the DNA tentacles on both sets of spheres, trapping the balls This last property makes it possible to read the results without in a dense web. This agglomeration will cause the solution to first washing away any probes that fail to find their target. change color ( from red to blue).

DNA QUANTUM DOTS FROM SAMPLE

PROBE DNA BEAD

BENT CANTILEVER CLEVER CANTILEVERS Biological samples can be screened for the presence of particular genetic sequences using small beams (cantilevers) NANO BAR CODES of the type employed in atomic force microscopes. The surface Latex beads filled with several colors of nanoscale semiconductors of each cantilever is coated with DNA able to bind to one known as quantum dots can potentially serve as unique labels for particular target sequence. A sample is then applied to the any number of different probes. In response to light, the beads Hybrid Medical Animation beams. Binding induces a surface stress, which bends the would identify themselves (and, thus, their linked probes) by affected beams by nanometers—not much, but enough to reveal emitting light that separates into a distinctive spectrum of colors that the bent beams found their specific targets in a sample. and intensities—a kind of spectral bar code. JEFF JOHNSON

www.sciam.com SCIENTIFIC AMERICAN 71 Copyright 2001 Scientific American, Inc. tiple DNA tentacles, bits of genetic ma- these microscopes has proved to be quite strands of DNA to the tops of the can- terial carrying the target sequence can difficult. tilevers. When genetic material carrying glue many particles together. And when The solution appeared in 1996, when a complementary sequence binds to the these gold specks aggregate, their optical workers at Rice University affixed a slen- anchored strands, it induces a surface properties shift markedly, changing the der carbon nanotube to the tip of an stress, which bends the cantilevers sub- test solution from red to blue. Because atomic force microscope, making it pos- tly—by just nanometers—but enough to the outcome of the test is easy to see sible to probe samples just a few nano- be detected. By fabricating devices with without any instrumentation at all, such meters in size. In 1998 Charles M. Lieber many cantilevers and coating each with a system might be particularly useful for and his co-workers at Harvard Univer- a different type of DNA, researchers home DNA testing. sity applied this approach to probing bio- should be able to test a biological sam- molecules, providing a very high resolu- ple rapidly for the presence of specific Feeling the Force tion means to explore complex biologi- genetic sequences (as is now done rou- NO DISCUSSION of bio-nanotechnol- cal molecules and their interactions at the tinely with gene chips) by nanomechan- ogy would be complete without at least most basic level. ical means without the need for labeling. a brief mention of one of the hottest in- But atomic force microscopy may This example, like the others de- struments in science today—the atomic soon be applied to more than just mak- scribed earlier, illustrates that the con- force microscope. Such devices probe ing fundamental scientific measurements. nections between nanotechnology and materials in the same way an old-fash- Last year James K. Gimzewski, then at the practice of medicine are often indi- ioned phonograph reads the grooves in a the IBM Zurich Research Laboratory, rect, in that much of the new work record: by dragging a sharp point over showed with collaborators at IBM and promises only better research tools or the surface and detecting the resulting de- the University of Basel that an array of aids to diagnosis. But in some cases, nano- flections. The tip of an atomic force mi- micron-scale arms, or cantilevers, much objects being developed may themselves croscope is, however, much finer than a like the ones employed in atomic force prove useful for therapy. One might, for phonograph needle, so it can sense far microscopes, could be used to screen instance, encapsulate drugs within nano- smaller structures. Regrettably, fabricat- samples for the presence of certain ge- meter-scale packages that control the ing tips that are both fine and sturdy for netic sequences. They attached short medicines’ release in sophisticated ways. Consider a class of artificial mole- cules called organic dendrimers. Two Petite Plumbing Jobs decades ago Donald A. Tomalia of the Michigan Molecular Institute in Mid- Microfluidics enhances biomedical research land fashioned the first of these intrigu- ing structures. A dendrimer molecule ost of the nanotechnologies now being developed for biomedical use take the branches successively from inside to out- Mform of minute objects immersed in comparatively large quantities of fluid, be it side. Its shape resembles what one would water, blood or a complex experimental concoction. But investigators are also get by taking many sprigs from a tree and building devices to manipulate tiny amounts of such liquids. These so-called poking them into a foam ball so that they microfluidic systems pump solutions through narrow channels, controlling the flow shot out in every direction. Dendrimers with diminutive valves and intense electric fields. are globular molecules about the size of The ability to handle vanishingly small quantities of a solution in this way allows a typical protein, but they do not come biomedical researchers to carry out many different experiments on what might be apart or unfold as easily as proteins do, only a modest amount of sample—and to do so in an efficient manner, with hundreds because they are held together with of tests being performed, say, on the surface of a single glass slide. Microfluidic stronger chemical bonds. devices also offer researchers the means to carry out experiments that could not Like the lush canopies of mature otherwise be done; for example, to deliver test solutions of specific compositions to trees, dendrimers contain voids. That is, different parts of a cell under study. they have an enormous amount of inter- Although many of the components being created for these systems are nal surface area. Interestingly, they can considerably larger than a micron, some experimental devices include nanoscale be tailored to have a range of different dimensions. Notably, Harold G. Craighead’s team at Cornell University has devised cavity sizes—spaces that are just perfect methods for sorting different sizes of DNA fragments in water according to how fast for holding therapeutic agents. Den- the fragments traverse passages measuring 100 nanometers across or travel drimers can also be engineered to trans- through microchannels that repeatedly narrow to a depth of 75 to 100 nanometers. port DNA into cells for gene therapy, These or other nanofluidic devices could potentially increase the speed and reduce and they might work more safely than the costs of separating DNA molecules for sequencing and could in theory be adapted the other leading method: genetically for separating proteins or other molecules. —A.P.A. modified viruses. Other types of nanostructures pos-

72 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. sess high surface area, and these too may prove useful for delivering drugs where they are needed. But dendrimers offer the greatest degree of control and flexibility. It may be possible to design dendrimers that spontaneously swell and liberate their contents only when the appropriate trigger molecules are present. This ability would allow a cus- tom-made dendrimer to release its load of drugs in just the tissues or organs needing treatment. Other drug-delivery vehicles on the horizon include hollow polymer capsules under study by Helmuth Möhwald of the Max Planck Institute of Colloids and In- terfaces in Golm, Germany. In response to certain signals, these capsules swell or compress to release drugs. Also intrigu- ing are so-called nanoshells, recently in- vented by the researchers at Rice. Nanoshells are extremely small beads of glass coated with gold. They can be fashioned to absorb light of almost any wavelength, but nanoshells that capture ORGANIC DENDRIMER, shown in an artist’s conception, could be roughly the size of a protein energy in the near-infrared are of most molecule. Dendrimers harbor many internal cavities and are being eyed as drug-delivery vehicles. interest because these wavelengths easily penetrate several centimeters of tissue. Even more exciting is the prospect that nologists have stated some truly ambi- Nanoshells injected into the body can physicians will make use of nanoscale tious goals. One of the “grand challenges” therefore be heated from the outside us- building blocks to form larger struc- of the National Nanotechnology Initia- ing a strong infrared source. Such a tures, thereby mimicking the natural tive is to find ways to detect cancerous nanoshell could be made to deliver drug processes of biology. Such materials tumors that are a mere few cells in size. molecules at specific times by attaching it might eventually serve to repair dam- Researchers also hope eventually to de- to a capsule made of a heat-sensitive aged tissues. Research on these bold velop ways to regenerate not just bone polymer. The capsule would release its strategies is just beginning, but at least or cartilage or skin but also more com- contents only when gentle heating of the one enterprise already shows that the plex organs, using artificial scaffoldings attached nanoshell caused it to deform. notion has merit: building scaffoldings that can guide the activity of seeded cells A more dramatic application envi- on which to grow bone. Samuel I. Stupp and can even direct the growth of a va- sioned for nanoshells is in cancer thera- of Northwestern is pioneering this ap- riety of cell types. Replacing hearts or py. The idea is to link the gold-plated proach using synthetic molecules that kidneys or livers in this way might not spheres to antibodies that bind specifi- combine into fibers to which bone cells match the fictional technology of Fan- cally to tumor cells. Heating the nano- have a strong tendency to adhere. tastic Voyage, but the thought that such shells sufficiently would in theory destroy What other marvels might the future medical therapies might actually become the cancerous cells, while leaving near- hold? Although the means to achieve available in the not so distant future is by tissue unharmed. them are far from clear, sober nanotech- still fantastically exciting. It is, of course, difficult to know for certain whether nanoshells will ultimate- MORE TO EXPLORE ly fulfill their promise. The same can be Ultrasensitive Magnetic Biosensor for Homogeneous Immunoassay. Y. R. Chemla, H. L. Grossman, said for the myriad other minuscule de- Y. Poon, R. McDermott, R. Stevens, M. D. Alper and J. Clarke in Proceedings of the National Academy of vices being developed for medical use— Sciences USA, Vol. 97, No. 27, pages 14268–14272; December 19, 2000. among them, one-nanometer buckyballs The author’s Web site is at www.cchem.berkeley.edu/~pagrp/ Hybrid Medical Animation made from just a few dozen carbon Information on using gold nanoparticles for DNA testing is available at atoms. Yet it seems likely that some of www.chem.nwu.edu/~mkngrp/dnasubgr.html the objects being investigated today will Information on nanoshells is available at www.ece.rice.edu/~halas/

JEFF JOHNSON be serving doctors in the near future. Information about quantum dots and their use in biomedicine is available at www.qdots.com

www.sciam.com SCIENTIFIC AMERICAN 73 Copyright 2001 Scientific American, Inc. NANOVISIONS

By K. Eric Drexler Machine-Phase Nanotechnology A molecular nanotechnology pioneer predicts that the tiniest robots will revolutionize manufacturing and transform society

IN 1959 PHYSICIST Richard Feynman to put every atom in a selected place molecular repair of the human body. gave an after-dinner talk exploring the (where it would serve as part of some ac- Medical nanorobots are envisioned that limits of miniaturization. He set out from tive or structural component) with no ex- could destroy viruses and cancer cells, known technology (at a time when an tra molecules on the loose to jam the repair damaged structures, remove ac- adding machine could barely fit in your works. Such a system would not be a liq- cumulated wastes from the brain and pocket), surveyed the limits set by phys- uid or gas, as no molecules would move bring the body back to a state of youth- ical law and ended by arguing the possi- randomly, nor would it be a solid, in ful health. bility—even inevitability—of “atom by which molecules are fixed in place. In- Another surprising medical applica- atom” construction. stead this new machine-phase matter tion would be the eventual ability to re- What at the time seemed absurdly would exhibit the molecular movement pair and revive those few pioneers now ambitious, even bizarre, has recently be- seen today only in liquids and gases as in suspended animation (currently re- come a widely shared goal. Decades of well as the mechanical strength typically garded as legally deceased), even those technological progress have shrunk mi- associated with solids. Its volume would who have been preserved using the crude croelectronics to the threshold of the mo- be filled with active machinery. cryogenic storage technology available lecular scale, while scientific progress at The ability to construct objects with since the 1960s. Today’s vitrification the molecular level—especially on the molecular precision will revolutionize techniques—which prevent the forma- In principle, Drexler says, a molecular construction system called an assembler could build almost anything, including copies of itself.

molecular machinery of living systems— manufacturing, permitting materials tion of damaging ice crystals—should has now made clear to many what was properties and device performance to be make repair easier, but even the original envisioned by a sole genius so long ago. greatly improved. In addition, when a process appears to preserve brain struc- Inspired by molecular biology, stud- production process maintains control of ture well enough to enable restoration. ies of advanced nanotechnologies have each atom, there is no reason to dump Those researchers most familiar with focused on bottom-up construction, in toxic leftovers into the air or water. Im- the field of molecular nanotechnology which molecular machines assemble mo- proved manufacturing would also drive see the technology base underpinning lecular building blocks to form products, down the cost of solar cells and energy such capabilities as perhaps one to three including new molecular machines. Biol- storage systems, cutting demand for coal decades off. At the moment, work fo- ogy shows us that molecular machine and petroleum, further reducing pollu- cuses on the earliest stages: finding out systems and their products can be made tion. Such advances raise hope that those how to build larger structures with atom- cheaply and in vast quantities. in the developing world will be able to ic precision, learning to design molecular Stepping beyond the biological anal- reach First World living standards with- machines and identifying intermediate ogy, it would be a natural goal to be able out causing environmental disaster. goals with high payoff. Low-cost, lightweight, extremely To understand the potential of mo- ABOUT THE AUTHOR strong materials would make transporta- lecular manufacturing technology, it tion far more energy efficient and—final- helps to look at the macroscale machine K. ERIC DREXLER is chairman of the Foresight ly—make space transportation economi- systems used now in industry. Picture a Institute, research fellow of the Institute for cal. The old dreams of expanding the robotic arm that reaches over to a con- Molecular Manufacturing, and author of Engines biosphere beyond our one vulnerable veyor belt, picks up a loaded tool, applies of Creation and Nanosystems: Molecular planet suddenly look feasible once more. the tool to a workpiece under construc- Machinery, Manufacturing, and Computation. Perhaps the most exciting goal is the tion, replaces the empty tool on the belt,

74 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. systems engineering. The shortage of mo- lecular systems engineers will probably be a limiting factor in the speed with which nanotechnology can be developed. It is important that critiques of nano- technology are well executed, because vi- tal societal decisions depend on them. If molecular nanotechnology as described here is correct, policy issues can look quite different from what is generally ex- pected. Today most people believe that global warming will be hard to correct— with nanotechnology, excess greenhouse gases could be inexpensively removed from the atmosphere. Current Social Se- curity projections assume increasing numbers of aged citizens in poor health. With advanced medical nanotechnology, tomorrow’s seniors could be more active and healthy than they are now, bringing new meaning to the “golden years.” Likewise, we need to focus now on avoiding accidents and preventing abuse of this powerful technology. Solid work has been done on the problem of heading off major nanotechnology accidents. The Foresight Guidelines, available on the World Wide Web, sketch out proposed picks up the next loaded tool, and so NANOSCRIBE K. Eric Drexler conceived the safety rules [see below]. on—as in today’s automated factories. concept of molecular machine systems But the challenge of preventing Now mentally shrink this entire mech- (a component of one is shown in the background). abuse—the exploitation of this technol- anism, including the conveyor belt, to the ogy by aggressive governments, terrorist molecular level to form an image of a scribing designs and calculations too groups or even individuals for their own nanoscale construction system. Given a bulky to fit in this essay. Fortunately, purposes—still looms large. The closest sufficient variety of tools, this system technical literature providing seemingly analogy to this problem these days is the would be a general-purpose building de- adequate answers has been available difficulty of controlling the proliferation vice, nicknamed an assembler. In princi- since at least 1992, when my book of chemical and biological weapons. The ple, it could build almost anything, in- Nanosystems was published. advance toward molecular nanotechnol- cluding copies of itself. Another well-known chemist objects ogy highlights the urgency in finding ef- Molecular nanotechnology as a field that an assembler would need 10 robot- fective ways to manage emerging tech- does not depend on the feasibility of this ic “fingers” to carry out its operations nologies that are powerful, valuable and particular proposal—a collection of less and that there isn’t room for them all. open to misuse. general building devices could carry out The need for such a large number of ma- MORE TO the functions mentioned above. But be- nipulators, however, has never been es- cause the assembler concept is still con- tablished or even seriously argued. In EXPLORE Engines of Creation: The Coming Era troversial, it’s worth mentioning the ob- contrast, the designs that have received of Nanotechnology. K. E. Drexler. Fourth jections being raised. (and survived) the most peer review use Estate, 1990. One prominent chemist speaking at a one tool at a time and grip their tools Nanosystems: Molecular Machinery, recent event sponsored by the American without using any fingers at all. Manufacturing, and Computation. Association for the Advancement of Sci- These examples point to the difficulty K. E. Drexler. John Wiley & Sons, 1992. ence asked how one could power and di- of finding appropriate critiques of nano- The Foresight Institute and the Foresight rect an assembler and whether it could technology designs. Many researchers Guidelines: www.foresight.org really break and re-form strong molecu- whose work seems relevant are actually Richard Feynman’s lecture “There’s Plenty of Room at the Bottom” can be found at — lar bonds. These are reasonable ques- the wrong experts they are excellent in www.zyvex.com/nanotech/feynman.html

PETER MENZEL tions that can be answered only by de- their discipline but have little expertise in

www.sciam.com SCIENTIFIC AMERICAN 75 Copyright 2001 Scientific American, Inc. NANOFALLACIES

By Richard E. Smalley Of Chemistry,Love and Nanobots How soon will we see the nanometer-scale robots envisaged by K. Eric Drexler and other molecular nanotechologists? The simple answer is never

WHEN A BOY AND A GIRL fall in love, it is cramped region of space measuring no certainly build another copy of itself. It often said that the chemistry between more than a nanometer on each side. could therefore self-replicate, much as them is good. This common use of the In recent years, it has become popu- biological cells do. After a while, we’d word “chemistry” in human relations lar to imagine tiny robots (sometimes have a second nanobot and, after a lit- comes close to the subtlety of what actu- called assemblers) that can manipulate tle more time, four, then eight, then 16 ally happens in the more mundane cou- and build things atom by atom. Imagine and so on. pling of molecules. In a chemical reaction a single assembler: working furiously, For fun, suppose that each nanobot between two “consenting” molecules, this hypothetical nanorobot would make consisted of a billion atoms (109 atoms) bonds form between some of the atoms many new bonds as it went about its as- in some incredibly elaborate structure. If in what is usually a complex dance in- signed task, placing perhaps up to a bil- these nanobots could be assembled at the volving motion in multiple dimensions. lion new atoms in the desired structure full billion-atoms-per-second rate imag- Not just any two molecules will react. every second. But as fast as it is, that rate ined earlier, it would take only one sec- They have to be right for each other. And would be virtually useless in running ond for each nanobot to make a copy of if the chemistry is really, really good, the a nanofactory: generating even a tiny itself. The new nanobot clone would molecules that do react will all produce amount of a product would take a soli- then be “turned on” so that it could start the exact product desired. tary nanobot millions of years. (Making its own reproduction. After 60 seconds Manipulator fingers on the hypothetical self-replicating nanobot are not only too fat; they are also too sticky. Both these problems are fundamental, and neither can be avoided.

Near the center of the typical chem- a mole of something—say, 30 grams, or of this furious cloning, there would be 260 ical reaction, the particular atoms that about one ounce—would require at least nanobots, which is the incredibly large are going to form the new bonds are not 6 × 1023 bonds, one for each atom. At number of 1 × 1018, or a billion billion. the only ones that jiggle around: so do the frenzied rate of 109 per second it This massive army of nanobots would all the atoms they are connected to and would take this nanobot 6 × 1014 sec- produce 30 grams of a product in 0.6 the ones connected to these in turn. All onds—that is, 1013 minutes, which is 6.9 millisecond, or 50 kilograms per second. these atoms must move in a precise way × 109 days, or 19 million years.) Al- Now we’re talking about something very to ensure that the result of the reaction though such a nanobot assembler would big indeed! is the one intended. In an ordinary chem- be very interesting scientifically, it Nanobots in general may not be ter- ical reaction five to 15 atoms near the re- wouldn’t be able to make much on its ribly interesting as a way of making action site engage in an intricate three-di- own in the macroscopic “real” world. prodigious amounts of things, but self- mensional waltz that is carried out in a Yet imagine if this one nanobot were replicating nanobots are really interest- so versatile that it could build anything, ing. If they are feasible, then the notion ABOUT THE as long as it had a supply of the right of a machine that can build anything AUTHOR kinds of atoms, a source of energy and from a CD player to a skyscraper in a RICHARD E. SMALLEY is the Gene and Norman a set of instructions for exactly what to remarkably short time doesn’t seem so Hackerman Professor of Chemistry and build. We could work out these detailed far-fetched. Physics at Rice University. He received the instructions with a computer and then But these self-replicating nanobots 1996 Nobel Prize in Chemistry for the radio them to the nanobot. If the nano- can also be quite scary. Who will control discovery of . bot could really build anything, it could them? How do we know that some sci-

76 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. entist or computer hacker won’t design someone who has worried in print about tional manipulator arm for each one of one that is truly autonomous, carrying a the societal implications of proliferating these atoms. Then it would have com- complete set of instructions for itself? nanobots, the future simply would not plete control of all the goings-on that oc- How do we know that these nanobots need us. cur at the reaction site. won’t mutate and that some of these mu- But how realistic is this notion of a But remember, this region where the tants won’t achieve the ability, like can- self-replicating nanobot? Let’s think chemistry is to be controlled by the nano- cer cells, to disregard any signals that about it. Atoms are tiny and move in a bot is very, very small—about one nano- would otherwise trigger self-destruction? defined and circumscribed way—a chem- meter on a side. That constraint leads to How could we stop them once they ist would say that they move so as to at least two basic difficulties. I call one reached this malignant state? Self-repli- minimize the free energy of their local the fat fingers problem and the other the cating nanobots would be the equivalent surroundings. The electronic “glue” that sticky fingers problem. Because the fin- of a new parasitic life-form, and there sticks them to one another is not local to gers of a manipulator arm must them- might be no way to keep them from ex- each bond but rather is sensitive to the selves be made out of atoms, they have a panding indefinitely until everything on exact position and identity of all the certain irreducible size. There just isn’t earth became an undifferentiated mass of atoms in the near vicinity. So when the enough room in the nanometer-size re- gray goo. nanomanipulator arm of our nanobot action region to accomodate all the fin- Still more frightening, they would by picks up an atom and goes to insert it in gers of all the manipulators necessary to either design or random mutation devel- the desired place, it has a fundamental have complete control of the chemistry. op the ability to communicate with one problem. It also has to somehow control In a famous 1959 talk that has inspired another. Maybe they would form groups, not only this new atom but all the exist- nanotechnologists everywhere, Nobel constituting a primitive nervous system. ing atoms in the region. No problem, physicist Richard Feynman memorably Perhaps they would really become “alive” you say: our nanobot will have an addi- noted, “There’s plenty of room at the bot- by any definition of that term. Then, in tom.” But there’s not that much room. the memorable words of Bill Joy, the NOBELIST Richard E. Smalley dismisses the Manipulator fingers on the hypo- chief scientist at Sun Microsystems and notion of out-of-control nanorobots. thetical self-replicating nanobot are not only too fat; they are also too sticky: the atoms of the manipulator hands will ad- here to the atom that is being moved. So it will often be impossible to release this minuscule building block in precisely the right spot. Both these problems are fundamen- tal, and neither can be avoided. Self- replicating, mechanical nanobots are simply not possible in our world. To put every atom in its place—the vision artic- ulated by some nanotechnologists— would require magic fingers. Such a nanobot will never become more than a futurist’s daydream. Chemistry is subtle indeed. You don’t make a girl and a boy fall in love by push- ing them together (although this is often a step in the right direction). Like the dance of love, chemistry is a waltz with its own step-slide-step in three-quarter time. Wishing that a waltz were a mer- engue—or that we could set down each atom in just the right place—doesn’t make it so. MORE TO EXPLORE The author’s Web site can be accessed at cnst.rice.edu/reshome.html PAM FRANCIS

SCIENTIFIC AMERICAN 77 Copyright 2001 Scientific American, Inc. NANOINSPIRATIONS

By George M. Whitesides The Once and Future Nanomachine Biology outmatches futurists’ most elaborate fantasies for molecular robots

AMONG THE PROMISED FRUITS of nano- ering its ultimate practicality). It would commonly considered to be the products technology, small machines have always be a new type of machine—a universal of human design and intention, why stood out. Their attraction is straightfor- fabricator. It would make any structure, shouldn’t a complex molecular system ward. Large machines—airplanes, sub- including itself, by atomic-scale “pick and that performs a function also be consid- marines, robotic welders, toaster ovens— place”: a set of nanoscale pincers would ered a machine, even if it is the product are unquestionably useful. If one could pick individual atoms from their envi- of evolution rather than of design? take the same ideas used to design these ronment and place them where they Issues of teleology aside, and accept- devices and apply them to machines that should go. The Drexlerian vision imag- ing this broad definition, nanoscale ma- were a tiny fraction of their size, who ines society transformed forever by small chines already do exist, in the form of the knows what they might be able to do? machines that could create a television set functional molecular components of liv- Imagining two types of small ma- or a computer in a few hours at essential- ing cells—such as molecules of protein or chines—one analogous to an existing ly no cost. It also has a dark side. The po- RNA, aggregates of molecules, and or- machine, the second entirely new—has tential for self-replication of the assembler ganelles (“little organs”)—in enormous captured broad attention. The first is a has raised the prospect of what has come variety and sophistication. The broad nanoscale submarine, with dimensions to be called gray goo: myriads of self- question of whether nanoscale machines of only a few billionths of a meter—the replicating nanoassemblers making un- exist is thus one that was answered in the The charm of the assembler is illusory: it is more appealing as metaphor than as reality, and less the solution of a problem than the hope for a miracle.

length of a few tens or hundreds of countable copies of themselves and rav- affirmative by biologists many years ago. atoms. This machine might, so the argu- aging the earth while doing so. The question now is: What are the most ment goes, be useful in medicine by nav- Does the idea of nanoscale machines interesting designs to use for future nano- igating through the blood, seeking out make sense? Could they be made? If so, machines? And what, if any, risks would diseased cells and destroying them. would they be effectively downsized ver- they pose? The second—the so-called assem- sions of their familiar, large-scale cousins, Cells include some molecular ma- bler—is a more radical idea, originally or would they operate by different prin- chines that seem similar to familiar hu- proposed by futurist K. Eric Drexler. ciples? Might they, in fact, ravage the man-scale machines: a rotary motor fixed This machine has no macroscopic ana- earth? in the membrane of a bacterium turns a logue (a fact that is important in consid- We can begin to answer these in- shaft and superficially resembles an elec- triguing questions by asking a more or- tric motor. Others more loosely resemble ABOUT THE AUTHOR dinary one: What is a machine? Of the the familiar: an assembly of RNA and many definitions, I choose to take a ma- protein—the ribosome—makes proteins GEORGE M. WHITESIDES is a professor in the chine to be “a device for performing a by an assembly line–like process. And chemistry department of Harvard University. task.” Going further, a machine has a de- some molecular machines have no obvi- This is his third article for Scientific American. sign; it is constructed following some ous analogy in macroscopic machines: a He is grateful to graduate students Kateri Paul process; it uses power; it operates ac- protein—topoisomerase—unwinds dou- and Abraham Stroock, who made many helpful cording to information built into it when ble-stranded DNA when it becomes too suggestions on this essay. it is fabricated. Although machines are tightly wound. The way in which these

78 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. organelles are fabricated in the cell—an ganic materials derived from them; it WHIPLIKE TAILS, found on many bacteria, are efficient synthesis of long molecules, com- largely removed the carbon dioxide from propelled by nanomotors. The tiny biochemical bined with molecular self-assembly—is a the atmosphere and injected enormous motor turns a rotary shaft that spins the tails, or model for economy and organization, quantities of oxygen. Overall, a radical flagella, and allows the bacteria, such as these and entirely unlike the brute-force meth- change. Cells—self-replicating collections E. coli, to move through liquid. od suggested for the assembler. of molecular nanomachines—complete- And as for ravaging the earth: in a ly transformed the surface and the atmo- that they are—but rather where we sense, collections of biological cells al- sphere of our planet. We do not normal- should look for new ideas for design. ready have ravaged the earth. Before life ly think of this transformation as “rav- Should we be thinking about the Gener- emerged, the planet was very different aging the planet,” because we thrive in al Motors assembly line or the interior of from the way it is today. Its surface was the present conditions, but an outside ob- a cell of E. coli? Let’s begin by compar- made of inorganic minerals; its atmo- server might have thought otherwise. ing biological nanomachines—especially Hybrid Medical Animation sphere was rich in carbon dioxide. Life So the issue is not whether nanoscale the ultimate self-replicating biological rapidly and completely remodeled the machines can exist—they already do—or system, the cell—with nanoscale ma- planet: it contaminated the pristine sur- whether they can be important—we of- chines modeled on the large machines

JEFF JOHNSON face with microorganisms, plants and or- ten consider ourselves as demonstrations that now surround us. How does the bi-

www.sciam.com SCIENTIFIC AMERICAN 79 Copyright 2001 Scientific American, Inc. ological strategy work, and how would straightforward chemical process—poly- being consumed) and other functional it compare with a strategy based on mak- merization—to create large, linear mol- species—sensors, structural elements, ing nanoscale versions of existing ma- ecules. The second is to build molecules pumps, motors. Most of the nanoma- chines, or a new strategy of the type sug- that spontaneously fold themselves into chines in the cell are thus, ultimately, mo- gested by the assembler? functional, three-dimensional structures. lecular catalysts. These catalysts do most This two-part strategy does not require a of the work of the cell: they form the Molecular Copy Machines difficult and sophisticated three-dimen- lipids (fats, for instance) that in turn self- THE CELL is a self-replicating structure. sional pick-and-place fabrication: it sim- assemble into the flexible sheet that en- It takes in molecules from its environ- ply strings beads (for example, amino closes the cell; they make the molecular ment, processes some of them for fuel, acids) together into a necklace (a poly- components necessary for self-replica- and reworks others into the pieces it uses peptide) and lets the necklace self-assem- tion; they produce the power for the cell to make, maintain, move and defend it- ble into a machine (a protein). Thus, the and regulate its power consumption; self. DNA stores the information need- information for the final, functional, they build archival and working infor- ed for fabrication and operation from three-dimensional structure is coded in mation storage; and they maintain the in- one generation to the next. One kind of the sequence of the beads. The three most terior environment within the proper op- RNA (messenger RNA, or mRNA) serves important classes of molecules in the erating parameters. as the temporary transcript of this infor- cell—DNA, RNA and proteins—are all Among the many marvelous molecu- mation, “telling” ribosomes which pro- made by this strategy; the proteins then lar machines employed by the cell, four tein to make. Membranes provide com- make the other molecules in the cell. In are favorites of mine. The ribosome, partments that enclose the working parts, many instances proteins also sponta- made of ribosomal RNA (rRNA) and house portals that control the flux of neously associate with other molecules— protein, is a key: it stands at the junction molecules into and out of the cell, and proteins, nucleic acids, small molecules— between information and action—be- hold molecules that sense the cell’s envi- to form larger functional structures. As a tween nucleic acids and proteins. It is an ronment. Proteins (often cooperating strategy for building complex, three-di- extraordinarily sophisticated machine with other molecules) build everything in mensional structures, this method of lin- that takes the information present in the cell and move its parts when they ear synthesis followed by various levels mRNA and uses it to build proteins. must be moved. of molecular self-assembly is probably The chloroplast, present in plant cells The strategy adopted by the cell to unbeatable for its efficiency. and algae, is a large structure that con- make its parts—and thus to replicate and The cell is, in essence, a collection of tains arrays of molecules that act as maintain itself—is based on two ideas. catalysts (molecules that cause chemical tuned optical antennas, collect photons The first is to use a single, conceptually reactions to occur without themselves from sunlight and employ them to gen- erate chemical fuel that can be stored in ) A TALE OF TWO MOTORS the cell to power its many operations. The chloroplast, incidentally, also con-

verts water to the oxygen that so conta- micrograph minated the atmosphere when life first emerged: the stuff on which our lives de- pend was originally a waste product of cellular light-harvesting! The mitochondrion is the power sta- tion: it carries out controlled combustion of organic molecules present in the cell— typically glucose—and generates power for the system. Instead of pumping elec- trons through wires to run electric mo- tors, it generates molecules of ATP that ROSIER, © 2000 AMERICAN INSTITUTE OF PHYSICS ( move through the cell by diffusion and E that are essential contributors to many biological reactions. The flagellar motor of bacteria is a specialized but particularly interesting nanomachine, because it seems so simi- lar to human-scale motors. The flagellar STANDARD-ISSUE electric motor bears a superficial—albeit striking—resemblance to the motor is a highly structured aggregate of

biochemical rotary motor (top right) that turns the flagella in a bacterium. proteins anchored in the membrane of BALDOR ELECTRIC COMPANY; DAVID D

80 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. many bacterial cells that provides the ro- large ratios of surface to volume, surface The assembler, with its pick-and- tary motion that turns the flagella—the effects—both good and bad—become place pincers, eliminates the many diffi- long whiplike structures that act as the much more important for them than for culties of fabricating nanomachines and propeller for these cells and allow them large devices. Some of these types of of self-replication by ignoring them: to propel themselves through water. It problems will eventually be resolved if it positing a machine that can make any has a shaft, like an electric motor, and a is worthwhile to do so, but they provide composition and any structure by simply structure around the shaft, like the ar- difficult technical challenges now. We placing atoms one at a time dismisses the Considering the many constraints on the construction and operation of nanomachines, it seems that new systems for building them might ultimately look much like the ancient systems of biology.

mature of a motor. The similarity be- will undoubtedly progress toward more most vexing aspects of fabrication. The tween flagellar and electrical motors is, complex micromachines and nanoma- assembler seems, however, from the van- however, largely illusory. The flagellar chines modeled on human-scale ma- tage of a chemist, to be unworkable. motor does not act by using electric cur- chines, but we have a long path to travel Consider just two of the constraints. rent to generate moving magnetic fields; before we can produce nanomechanical First is the pincers, or jaws, of the as- instead it uses the decomposition of ATP devices in quantity for any practical pur- sembler. If they are to pick up atoms with to cause changes in the shape of the mol- pose. Nor is there any reason to assume any dexterity, they should be smaller ecules that, when combined with a so- that nanomachines must resemble hu- than the atoms. But the jaws must be phisticated molecular ratchet, make the man-scale machines. built of atoms and are thus larger than protein shaft revolve. Could these systems self-replicate? At the atom they must pick and place. present, we do not know how to build (Imagine trying to build a fine watch with Nanomachines That Mimic self-replicating machines of any size or your fingers, unaided by tools.) Second is Human-Scale Machines type. We know, from recent biological the nature of atoms. Atoms, especially CAN WE EVER APPROACH the ele- studies, something about the minimum carbon atoms, bond strongly to their gant efficiency of cellular nanomachines level of complexity in a living cell that neighbors. Substantial energy would be by creating tiny cousins of the larger ma- will sustain self-replication: a system of needed to pull an atom from its place (a chines we have invented? Microfabrica- some 300 genes is sufficient for self-repli- problem for the energy supply) and sub- tion has developed as an extraordinarily cation. We have little sense for how to stantial energy released when it is put in successful technology for manufacturing translate this number into mechanical place (a problem of cooling). More im- small, electronically functional devices— machines of the types more familiar to portant, a carbon atom forms bonds transistors and the other components of us, and no sense of how to design a self- with almost everything. It is difficult to chips. Application of these techniques to sustaining, self-replicating system of ma- imagine how the jaws of the assembler simple types of machines with moving chines. We have barely taken the first would be built so that, in pulling the parts—mechanical oscillators and mov- steps toward self-replication in nonbio- atoms away from their starting material, able mirrors—has been technically suc- logical systems [see “Go Forth and Repli- they would not stick. (Imagine trying to cessful. The development of these so- cate,” by Moshe Sipper and James A. build your watch with parts salvaged called microelectromechanical systems Reggia; Scientific American, August]. from another watch in which all the (MEMS) is proceeding rapidly, but the And other problems cast long shad- parts were coated with a particularly functions of the machines are still ele- ows. Where is the power to come from sticky glue: if you could separate the mentary, and they are micro, not nano, for an autonomous nanomachine? There pieces at all, they would stick to your fin- machines. The first true nanoscale MEMS are no electric sockets at the nanoscale. gers.) Problems with the assembler are (NEMS, or nanoelectromechanical sys- The cell uses chemical reactions of specif- also discussed by Richard E. Smalley in tems) have been built only in the past few ic compounds to enable it to go about its his essay on page 76. years and only experimentally [see “Plen- business; a corresponding strategy for Would a nanosubmarine work if it ty of Room, Indeed,” on page 48]. nanoscale machines remains to be devel- could be built? A human-scale subma- Many interesting problems plague oped. How would a self-replicating nano- rine moves easily in water by a combina- the fabrication of nanodevices with mov- machine store and use information? Biol- tion of a rotating propeller—which, in ing parts. A crucial one is friction and ogy has demonstrated a strategy based on spinning, forces the water backward and sticking (sometimes combined in talking DNA, so it can be done, but if one want- the submarine forward—and movable about small devices in the term “stic- ed a different strategy, it is not clear where planes that guide its direction. Bacteria tion”). Because small devices have very to start. that swim actually use structures—fla-

www.sciam.com SCIENTIFIC AMERICAN 81 Copyright 2001 Scientific American, Inc. a nanosubmarine jects 10or100timesmoremassivethan mariner. Cellsinthebloodstream at leastinasensefamiliartosub- would probablybeimpossibletosteer, be lefttochance,becausethetinycraft the destinationofmostvoyageswould proximately 100nanometersinscale, crash againsttheirhulls.Forshipsofap- to theBrownianstormsthatwould nanoscale wouldhavetoaccommodate water molecules.Navigatorsonthe lentless collisionswithrapidlymoving course wouldbefrustratedbythere- tions: anyefforttosteerapurposeful Brownian motion)butinrandomdirec- bounce aboutrapidly(aprocesscalled with themmakeananoscaleobject is rapidonthenanoscale.Collisions much smaller,andtheirthermalmotion smaller thanananosubmarinebutnot dom batteringbywatermolecules. and seriousproblemwouldemerge:ran- one couldfabricateapropeller,new of food.Fornanoscaleobjects,evenif tends inthegeneraldirectionofasource about, withmotionthat,ifallgoeswell, very purposefulpathbutratherdash propeller. Theytypicallydonotsteera or whipsbutserveafunctionsimilarto 82 gella detect diseaseifone for thesophisticatedtasksrequiredto at thenanoscale,theywouldnotwork whether onecouldmakeorsteerdevices a specificdestination.Regardlessof hope toselectageneraldirectionbutnot with it.Atbest,ananosubmarinemight selves init:theysimplytumblealong as dangerous.In manyoftheircharac- the outsideofmost cancercellsflagthem that constituteit.Nosimplemarkers on ing themanybillionsofspecialized cells panoply ofourimmunesystem, includ- plex process or “cancer”isanextraordinarily com- tion ofacellas“normal”or“pathogen” tem nowfunctioninginus.Therecogni- have tomimicaspectsoftheimmunesys- prey. Indoingso,theywouldprobably would havetofocusonfindingtheir cells inthebody,suchascancercells, egy fordetectinganddestroyingdiseased These watermoleculeswouldbe Parts ofthe“littlesubmarine”strat- — SCIENTIFIC AMERICAN that lookmorelikeflexiblespirals — one thatrequiresthefull — do notguidethem- could make them. — ob- the cell isting nanomachines limiting strategies.Thefirstisto takeex- nanomachines, wecomeupagainst two those usedinhuman-scalemachines. these nanometer-scalemachines arefrom ing ishowdifferentthestrategiesusedin ily sophisticatedfunctions.Whatisstrik- performextraordinar- exist, andtheydo ples: inlivingsystems,nanomachinesdo one brilliantlydevelopedsetofexam- remain tobedevised.Biologyprovides them, andthepurposestheywillserve, made, butthestrategyusedtomake would probablyhavetodothesame. ents asdoothercells;alittlesubmarine the immunesystemusesamenutri- would alsorequireenergy.Thecellsof cease tobelittle.Operatingthisdevice bers andanalyticaldevices,itwould devices andreagentsreactioncham- that laboratorywouldrequiresampling tle diagnosticlaboratory,andbecause cells wouldhavetocarryonboardalit- that wastobeahunter-killerforcancer ent fromnormalcells.Alittlesubmarine teristics, theyarenotenormouslydiffer- ML MACHINES SMALL Outdesigning Evolution in anautomotivefactory( acids thatconstituteaprotein( RIBOSOME Copyright 2001 Scientific American, Inc. In thinkingabouthowbestto make — and learnfrom them.Wewill reads alongastrandofRNA( — opposite page will eventuallybe those presentin gold ). Thisassembly-lineprocessbringstomindtheroboticwelders ). purple ) togetinstructionsforstringingtogethertheamino ufacture electronic deviceswillcertainly tromagnets. Techniques devisedtoman- or generatingmagneticfields with elec- moving inastraightlinethrough afluid meter sizes.Nordoprocesses suchas ing donothavecounterparts atnano- cases impossible.Machining and weld- will usuallynotbepracticalandinmany gous tothoseemployedonalargescale versions ofthemusingprocessesanalo- expecting tobeablebuildnanoscale ing atthemachinesthatsurroundusand ers. Butthispathwillbearduous.Look- son tobelievethattherecannotbeoth- tional nanomachines,andthereisnorea- for fabricationandsynthesisoffunc- ogy hasproducedonepracticalmeans mental newtypesofnanosystems.Biol- and independentlytodevelopfunda- and nucleicacids. molecular systemsthatarenotproteins enable ustousebiologicalprinciplesin opment ofnewtypeschemistrymay ceeding downthispath,andthedevel- tions. Geneticengineeringisalreadypro- others thatwillhaveentirelynewfunc- them thatwillserveourpurposes,and that willenableustomakevariantsof these systemsconceptsandprinciples undoubtedly beabletoextractfrom The secondistostartfromscratch SEPTEMBER 2001

JEFF JOHNSON Hybrid Medical Animation fabrication. It must also be able to manu- facture and assemble (or allow to assem- ble) all the pieces needed to make a copy of itself. Biology has solved all these prob- lems, and self-replicating biological sys- tems—from pathogenic bacteria to can- cer cells—are a danger to us. In comput- er systems, self-replicating strings of bits (computer viruses), although not materi- al objects, are also at least a great nui- sance, but only indirectly a danger, to us. If a new system—any system—were able to replicate itself using materials present in the environment, it would be cause for concern. But we now know enough to realize how far we are from re- producing self-replication in a nonbio- logical system. Fabrication based on the assembler is not, in my opinion, a work- able strategy and thus not a concern. For be able to make some simple types of me- ticular reaction catalyzed by this enzyme the foreseeable future, we have nothing chanical nanodevices, but they will be could be better. For most enzymes, and to fear from gray goo. If robust self-repli- limited in what they can do. all structures more complicated than en- cating micro (or perhaps nano) struc- The dream of the assembler holds se- zymes, we have made no effort to dis- tures were ultimately to emerge, they ductive charm in that it appears to cir- cover the alternatives. would probably be chemical systems as cumvent these myriad difficulties. This Biological structures work in water, complex as primitive bacteria. Any such charm is illusory: it is more appealing as and most work only in a narrow range of system would be both an incredible ac- metaphor than as reality, and less the so- temperatures and concentrations of salts. complishment and a cause for careful as- lution of a problem than the hope for a They do not, in general, conduct elec- sessment. Any threat will not be from as- miracle. Considering the many constraints tricity well (although some, such as the semblers gone amok but from currently on the construction and operation of chloroplast and the mitochondrion, unimaginable systems of self-catalyzing nanomachines, it seems that new systems move electrons around with great so- reactions. for building them might ultimately look phistication). They do not carry out bi- So biology and chemistry, not a me- It will be a marvelous challenge to see if we can outdesign evolution. It would be a staggering accomplishment to mimic the simplest living cell.

much like the ancient systems of biology. nary computation and communications. chanical engineering textbook, point in It will be a marvelous challenge to see if They are not particularly robust me- the direction we should look for an- we can outdesign evolution. It would be chanically. Thus, a great many types of swers—and it is also where our fears a staggering accomplishment to mimic function must be invented if nanoma- about organisms or devices that multiply the simplest living cell. chines are to succeed in nonbiological uncontrollably are most justified. In Are biological nanomachines, then, environments. thinking about self-replication, and about the end of the line? Are they the most And what have we learned from all the characteristics of systems that make highly optimized structures that can ex- this about the doomsday scenario of gray them “alive,” one should start with biol- ist, and has evolution sorted through all goo? If a hazard were to arise from nano- ogy, which offers a cornucopia of designs

Corbis possibilities to arrive at the best one? We machines, it would lie in a capability for and strategies that have been successful have no general answer to this question. self-replication. To be self-replicating, a at the highest levels of sophistication. In Jeremy R. Knowles of Harvard Universi- system must contain all the information tackling a difficult subject, it is sensible to ty has established that one enzyme— it needs to make itself and must be able to start by studying at the feet of an accom- triose phosphate isomerase, or TIM—is collect from its environment all the mate- plished master. Even if they are flagella,

MICHAEL S. YAMASHITA “perfect”: that is, no catalyst for the par- rials necessary both for energy and for not feet.

www.sciam.com SCIENTIFIC AMERICAN 83 Copyright 2001 Scientific American, Inc. NANOROBOTICS

By Steven Ashley Nanobot Construction Crews Nanotechnology visionaries find out how difficult it is to develop minuscule robots that can treat diseases or perform pollution-free manufacturing

MOLECULAR MANUFACTURING—the heady notion of assembling almost anything, from computers to caviar, from individ- ual molecules—would change the world, if someone could just find a way to make it work. Imagine nanoassemblers: busy work gangs of “pick and place” robotic manipulator arms, each one tens of nanometers in size. Controlled from on high by a powerful computer, these sim- ple devices would arrange blocks of mol- ecules to make copies of themselves— and these machines would, in turn, build still other nanomachines, which, in turn, would create others, and so on in an ex- ponential expansion. These nanobot construction crews could then be direct- ed to accomplish astonishing tasks such as curing diseases from inside the body and fabricating intricately engineered materials from basic bulk feedstocks at extraordinarily low cost. For years, futurist K. Eric Drexler and his colleague Ralph C. Merkle, the noted cryptographer, nurtured this vision, using computer simulations of nanometer-scale gears, pumps and other molecular ma- chine subsystems. In these images, col- real substance to his computer-concoct- TOP-DOWN FABRICATION is one pathway Zyvex ored spheres indicate the position of each ed concepts. He joined software mogul is taking toward molecular manufacturing. component atom. Several analytical cri- James R. Von Ehr II, an admirer of tiques of these elaborate simulations sug- Drexler’s ideas, who had decided to of its quest. Its experience has already gested that the devices simply wouldn’t spend part of the “low-nine-figure” for- demonstrated the difficulty of moving work as hoped. But perhaps the most bit- tune he garnered from the sale of his from software-wrought molecular ma- ing criticism focused on their status as company, Altsys, to start the first “mo- chines to a real-world nanoassembler. mere digital representations and not real- lecular nanotechnology” company. One of the first things the company had world objects interacting with complex Zyvex, the Richardson, Tex.–based to do was to pull back from the idea of chemical bonding forces in the nanoscale company Von Ehr founded, is aiming building things atom by atom. Today the environment, where macroscale physics high. “We’d love to be the Applied Ma- long-term focus is on fabricating nano- doesn’t always apply and quantum me- terials for the manufacturing base of the scale machine systems from large mole- chanics often prevails. world,” he says, referring to the world’s cules or blocks of molecules. To that end, Then, in 1999, Merkle decided to leading semiconductor equipment man- company scientists are experimenting leave a long-term position at the Xerox ufacturer. But his molecular nanotech- with the bottom-up approach to making

Palo Alto Research Center to try to give nology company is still in the early stages assemblers with molecular nanotechnol- JAMES STEINBERG

84 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. ogy—learning to put together structures constructors on the job, building macro- MEMS program at the Defense Ad- molecule by molecule using atomic force scale objects molecule by molecule vanced Research Projects Agency. microscopes and the like. would take a very long time. Zyvex re- And what about the nanoscale as- In the near term, Zyvex researchers searchers note that a fully self-replicating sembler? “There is significant contro- are developing microelectromechanical machine would not be necessary to man- versy about how long it will take to systems (MEMS), whose structures mea- ufacture an assembler. Still, the task is a achieve ultimate control at the molecu- sure in tens of microns. MEMS devices monumental one: any machine outside lar level,” Merkle notes, adding that “it are manufactured using a top-down ap- the biological sphere that could make could take a decade or two.” proach. They are lithographically pat- any number of copies of itself would So Zyvex has its work cut out for it. terned and etched out from silicon sub- probably lead to a Nobel But it has embarked on strates or other materials. Prize, possibly several. a mission to achieve the Richard Feynman, the ur-guru of the So how does Zyvex get nanoequivalent of a moon field, did say that bigger machines could to an assembler? Recently shot in relative isolation be used to make smaller machines. And Zyvex joined with Stan- with a staff of only 37. In that’s where MEMS comes in. Accord- dard MEMS in Burling- general, the scientific re- ing to Merkle, MEMS fabrication meth- ton, Mass.—a company search community has ods can be used to build robotic assem- that fabricates microsys- distanced itself from this bly arms. These microelectromechanical tems—in a two-year col- project. Several scientists “hands” can assemble submicron hands laborative program to develop litho- working in the field of nanotechnology that can build even tinier hands, and so graphically formed micron-scale manip- derided Zyvex’s scheme but requested forth, step by step until the ultimate in ulator arms and grippers that can as- anonymity to avoid protests from ama- mechanical smallness is achieved: the semble smaller manufacturing devices teur nanotech enthusiasts. Says one re- nanoassembler, a nanometer-scale man- from pallets of precisely positioned “ac- searcher, “We’ve seen no experimental ufacturing device. For Zyvex to accom- tive” parts produced on silicon wafers. “If proof that any portion of their scheme plish its goal, the MEMS robots that it the positional accuracy is good enough, can actually be accomplished. We think creates must be shrunk by a factor of the arm simply has to reach down and it’s a lot of nonsense.” Merkle responds 1,000 or so. Once the requisite down- pick the part up. You don’t need an elab- that “nothing we propose contradicts sizing has been completed, nanobots, as orate sensing system,” Merkle says. the laws of physics.” envisaged, will be used for “atomically According to Von Ehr, Zyvex re- Meanwhile, with Von Ehr’s fortune precise manufacturing” to make virtual- searchers have developed a method by backing up the research effort, Zyvex can ly anything. Universal constructors could which the MEMS manipulator can de- continue to function for many years with- manufacture a Rolex watch, followed tach the parts from the substrate so they out having to market a product, but the by a computer memory and then by are ready for microassembly using self- entrepreneur allows that it would be nice nanomachines that can treat diseases. centering snap-connectors—a useful, if to make some money. “This whole thing Two basic ideas underlie Zyvex’s not exactly new, capability. Manipulator is a lot harder than it first seemed,” he ad- road to nanorobotics, Merkle says. In parts would be created lithographically, mits. Von Ehr has reportedly already positional assembly, mechanical manip- etched out and then disconnected from spent about $20 million on the project, ulators pick up and precisely place ob- the surface substrate so they could be but considering last year’s stock market jects into assemblies. Accurate position- picked up by the MEMS manipulator fall and the reality of the task looming on al control at scales of tens of nanometers and attached to the device by snapping the horizon, he says he is planning to seek means moving large molecules or blocks them into place. Von Ehr hopes these outside investment once financial condi- of molecules where you want them and MEMS manipulators will prove to be an tions improve. “If we’re going to grow,” causing them to bond to an assembly as intermediate technology that could serve he says, “we’ll need more money.” desired. as a moneymaking product, such as a de- No matter how big the envisioned A machine that can build a complex vice to align a fiber-optic cable, as the payoff, however, given the daunting material or device from the bottom up— company pursues its goals. technical difficulties and a 10- to 20-year using an Erector set of molecules or Unfortunately, it’s hard to find any- timeline to possible success, one won- chunks of molecules for components— one in the MEMS field who would be in- ders just who would be willing to put up will require one other critical technolo- terested in this kind of technology or any- significant investment funds. Perhaps gy: the machine must be able to make a one who can contemplate what kind of Zyvex’s trek toward molecular nanotech copy of itself. “If you wish to achieve product it would actually be used to could be financed by small contributions economical production of molecular de- build, says Kaigham J. Gabriel, now pro- from its legions of true believers. vices in large numbers, some form of self- fessor of electrical and computer engi- replication is necessary,” Merkle explains. neering and robotics at Carnegie Mellon Steven Ashley is a staff editor and And unless you had hordes of nano- University and former director of the writer. www.sciam.com SCIENTIFIC AMERICAN 85 Copyright 2001 Scientific American, Inc. NANOFICTION

By Graham P. Collins Shamans of Small Like interstellar travel, time machines and cyberspace, nanotechnology has become one of the core plot devices on which science-fiction writers draw

“READING DREXLER’S Engines of Cre- ed by its proponents to be upon us with- that Escher used rules of geometric sym- ation in 1990 went into the making of in a decade or few, nanotechnology also metry—the rules give form to the im- the world in Queen City Jazz, my first gives science-fiction writers a chance to possible imaginative content. novel, though the book drew from many engage in the art of predicting and warn- other sources: Shakers, ragtime, jazz, ing about possible futures. This role as Antediluvian Nanotech American literature, even Krazy Kat.” So an ad hoc think tank is one that innu- MANY OF DREXLER’S imaginings have writes Kathleen Ann Goonan in the Sum- merable science-fiction writers and fans antecedents in science fiction and feed mer 2001 SFWA Bulletin, the quarterly take on enthusiastically, not only in their into old potent themes of the genre. Sci- of the Science Fiction and Fantasy Writ- fiction but also in endless earnest panel ence fiction has long been fascinated with ers of America, in a brief essay about her discussions at conventions, in online machines in general, such as in the stories award-nominated novel Crescent City newsgroups and discussion boards, and of Jules Verne. The absolute control of Rhapsody, the third book of her musi- in articles labeled (sometimes optimisti- matter promised by nanomachines is a cally structured Nanotech Quartet. cally) nonfiction. It is the culture of the variant of the dream that Homo sapiens Unsurprisingly, Goonan is far from intensely technophilic—even those who can achieve complete mastery over nature the only science-fiction writer to take in- write of techno-dystopias and apoca- and has utter freedom to shape its own spiration from K. Eric Drexler’s vision lypses are enrapt in a love-hate relation- destiny. The dark vision of nanobots run- of molecular nanotechnology, for it is a ship with science and technology. The ning amok is a new wrinkle on the old vision that connects to numerous preex- borders between four dominions—those golem/Frankenstein myth, the dangers of isting themes of science fiction and offers of scientists, writers, readers and science- meddling with godlike powers or bring- writers an extraordinarily broad palette fiction fans—are hopelessly blurred, ing too much hubris to science. The of capabilities, all imbued with the ap- with countless individuals holding joint bright vision of the world, and indeed the pearance of scientific plausibility. Tout- citizenships. nature of humanity, being transformed But it would be a mistake to think into something transcendent and new is that science fiction’s central role is one of another science-fiction standby. NANOTITLES serious prognostication. The question Nanotech burst into the collective DISCUSSED “What if ...?” lies at the heart of science consciousness of technology aficionados QUEEN CITY JAZZ. Kathleen Ann Goonan. fiction, but what comes after the ellipsis at a good time to interface neatly with Tor, 1994. and the answers that stories give are ulti- the mid-1980s wave of cyberpunk sto- CRESCENT CITY RHAPSODY. Kathleen Ann mately not science but literature—that ries, in which characters experience the Goonan. Avon Eos, 2000. strange mix of entertainment and mean- completely programmable virtual reali- FANTASTIC VOYAGE: MICROCOSM. Kevin ingful enrichment of life. The art of any ties of cyberspace. With full-scale mo- J. Anderson. Onyx Books, 2001. fiction writer is the art of the storyteller. lecular nanotech it is not just virtual re- MICROCOSMIC GOD: THE COMPLETE STORIES As Kathryn Cramer (writer and anthol- ality that is programmable. The intelli- OF THEODORE STURGEON, Theodore Vol. 2. ogist and daughter of physicist and fic- gent agents and viruses of cyberspace Sturgeon. Edited by Paul Williams. North tion writer John G. Cramer) writes in become free to roam about in the air that Atlantic Books, 1995. The Ascent of Wonder: we breathe and within our bodies—a cu- BLOOD MUSIC. Greg Bear. Arbor House, 1985. rious inversion of people loading their NANOWARE TIME. Ian Watson. Tor, 1991. The majority of science fiction stories consciousnesses into machines. QUEEN OF ANGELS. Greg Bear. Warner Books, are not plausible extrapolations upon 1990. our current situation, using available in- MICROSCOPIC CRAFT, built using nanotech, / [SLANT]. Greg Bear. Tor, 1997. formation; rather they are Escheresque battle foreign invaders the way an immune THE DIAMOND AGE. Neal Stephenson. impossible objects which use the princi- system does. Even the implant-enhanced eye Bantam, 1995.

ples of science in much the same way sees only a fog sparkling with laser light. FRED GAMBINO

86 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Copyright 2001 Scientific American, Inc. The prospect of absolute control of matter using nanomachines is a variant of the dream that Homo sapiens can achieve complete mastery over nature.

Elements suggestive of now common plore the dormant body of an alien from nanites, and a mutually acceptable peace- nanotech themes appeared in science fic- a shot-down UFO. Lo and behold, the ful outcome is negotiated: they are placed tion well before the advent of the term Lilliputian explorers find themselves con- on a convenient planet where they’ll have “nanotech.” The concept of microscop- fronting alien nanotechnology. more room to live and grow. If only every ic surgery appeared in the 1966 movie Progenitors of nanotech fiction ex- conflict, plague or technological disaster Fantastic Voyage, novelized by Isaac Asi- tend back even further. In the classic 1941 in the real world were solvable with such mov. Of course, instead of using molec- story by Theodore Sturgeon, “Microcos- ease and rationality. ular machinery built of real atoms, a mic God,” a scientist creates a society of The central feature of molecular large scientific-looking contraption mag- miniature creatures (“Neoterics”) that nanotechnology, precise manipulation ically reduces people and machinery to evolve at a rapid pace and produce tech- of atoms, crops up in a classic science microscopic scale by shrinking their very nological wonders. This theme is picked fantasy of 1965: atoms, in violation of numerous physical up in a modern way in Blood Music, by principles. Interestingly, the 2001 novel Greg Bear, first published as a novelette The stuff was dancing particles within Fantastic Voyage: Microcosm, by Kevin in the magazine Analog in 1983 and ex- her[....] She began recognizing famil- J. Anderson, uses the miniaturization panded into a novel in 1985. Despite iar structures, atomic linkages: a car- technology of Fantastic Voyage to ex- predating the popularization of nano- bon atom here, helical wavering ... a tech, Blood Music is frequently cited as glucose molecule. An entire chain of a seminal nanotech story and included molecules confronted her and she rec- STORY EXCERPT: EVERYDAY NANO-NOVELTIES in nanotech anthologies. In the story, a ognized a protein ... a methyl-protein researcher creates intelligent cells, “no- configuration. WE WERE OUT OF OUR MINDS WITH JOY, ocytes,” that escape from confinement [...] she moved into it, shifted an BY DAVID MARUSEK (from Asimov’s Science and spread like an epidemic through hu- oxygen mote, allowed another carbon Fiction, November 1995) manity, destroying it but also seemingly mote to link, reattached a linkage of I held patents for package applications in bringing about a transcendental change oxygen ... hydrogen. many fields, from emergency blankets and to a new form of existence. It’s the end The change spread ... faster and temporary skin, to military camouflage and of the world as we know it, but we’ll all faster as the catalysed reaction opened video paint. But my own favorites, and feel fine afterward. its surface of contact. probably the public’s as well, were my novelty Television has also picked up on Stur- gift wraps. My first was a video wrapping paper geon’s concept. In the 1996 Halloween The novel? Dune, by Frank Herbert, that displayed the faces of loved ones (or episode of The Simpsons, Lisa acciden- in which computers are banned through- celebrities if you had no loved ones) singing tally creates a microscopic society (ingre- out the empire and spaceship pilots nav- “Happy Birthday” to the music of the New York dients: a tooth, Coca-Cola and an electric igate through hyperspace by means of Pops. That dated back to 2025 when I was a shock delivered by Bart) that rapidly ad- drug-induced precognition. While Paul molecular engineering student. vances from the Stone Age through the Atreides is on his way to becoming the My first professional design was the old Renaissance and then far beyond our messianic ruler of the known universe, box-in-a-box routine, only my boxes didn’t own technology. The theme is combined his mother, Lady Jessica, takes part in a get smaller as you opened them, but larger, more soberly with the modern concept of ritual involving “the water of life”—a and in fact could fill the whole room until you nanotech in the episode “Evolution” of deadly poison related to the “melange,” chanced upon one of the secret commands, Star Trek: The Next Generation, which or “spice,” that feeds supernatural pow- which were any variation of “stop” (whoa, aired in 1989, just three years after ers of intuition and prescience. To sur- enough, cut it out, etc.) or “help” (save me, Drexler’s Engines of Creation hit the vive the ritual, Jessica’s consciousness I’m suffocating, get this thing off me, etc.). bookstores. Boy wonder Wesley Crusher dives down into inner space and slows Next came wrapping paper that accidentally releases some “nanites,” tiny time to a crawl to analyze the chemical screamed when you tore or cut it. That led to robots designed to work in living cells, composition of the poison on her tongue paper that resembled human skin. It molded which proceed to evolve into a highly in- and, using a “psychokinesthetic exten- itself perfectly and seamlessly (except for a telligent society that invisibly infests the sion of herself,” transform it into a cata- belly button) around the gift and had a shelf systems of the starship Enterprise and lyst that rapidly detoxifies all the rest of life of fourteen days. You had to cut it to open starts wreaking havoc. Fortunately, in the poison, turning it into a potent but the gift, and of course it bled. We sold classic Next Generation style, at the last not deadly narcotic. mountains of that stuff.

minute, contact is made with the evolved Yet except for the use of psychoki- © 1995 BY BANTAM DOUBLEDAY DELL MAGAZINES; REPRINTED PERMISSION OF THE AUTHOR

88 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. INVISIBLE NANOTECH pervades this scene from Greg Bear’s novel / [Slant]. Mary Choy has transformed her features back to her original Asian look. The pyramidal building is reinforced with flexfuller and contains cryopreserved bodies of the rich and privileged. The swanjet has no conventional ailerons; instead nanodevices on each wing surface control the lift forces by forming thousands of tiny vanes or humps.

nesis in place of a technological frame- miniaturization.” Indeed, devices from lier stories, of what generic Acme nano- work, the entire process sounds like a Dune such as the “hunter-seeker,” a tiny tech can and can’t do. nanotech engineer working at a virtual- poison-tipped flying needle, would be Both of the latter alternatives relate reality station to design a molecule. Or completely at home in nanotech stories. to two approaches to presenting tech- like a scenario from Unbounding the Fu- nology in science fiction. At one ex- ture: The Nanotechnology Revolution, Hot Stuff treme, the text practically contains a re- by Drexler, Chris Peterson and Gayle NANOTECHNOLOGY’S use in science search paper on the author’s hyperspace Pergamit, in which a tourist experiences fiction takes many forms, classifiable by theory and blueprints of the first starship a museum exhibit that simulates the mo- a number of measures. The role of nano- (rather like some of the letters Scientific lecular-scale world, complete with scal- tech ranges from a central part of the American receives). At the other, fa- ing (slowing down) of time. In essence, plot to a relatively incidental part of the mously pioneered by Robert A. Hein- nanotechnology offers to make Dune’s fictional world. The nanotech may be lein, the technology is dropped in with- fantasy of complete human control over developed by humans, or it may be a gift out explanation: “The door dilated.” In the self and the rest of the universe into from aliens, or it may be the aliens. The three words we know we are in a future a reality but in a mass-produced indus- technology may work according to well- with strange new technologies, and we trial fashion rather than through inten- defined rules, or it may be arbitrary mag- are really there because in the future sive individual training and drug-en- ic, scantily clad in trappings of science. commonplace devices such as dilating hanced psychic powers. The rules may be expressly mentioned in doors need no more explanation than Nanotech is also prefigured in Dune the text (perhaps even laboriously de- cellular phones do today. through the Ixians, traders from a rare scribed), or the work may rely on the When you read a large number of high-tech corner of the universe who are reader to tap into a science-fictional con- nanotechnology stories in a short space

JIM BURNS “supreme in machine culture. Noted for sensus reality, acquired from reading ear- of time, some amusing recurring themes

www.sciam.com SCIENTIFIC AMERICAN 89 Copyright 2001 Scientific American, Inc. a sequel to the landmark Queen of An- gels (1990), four people are found dead in an illicit body-modification clinic: they were cooked, literally, when the body-modifying nanotech ran amok. The cause is promptly uncovered by investi- gators when they examine the jars of pastelike “nano” on the shelf:

Mary picks up a bottle, reverses it to read the label.[...] The label confirms her suspicions.[...] “This isn’t medical grade,” Mary says. “It’s for gardens.[...] Any real ex- pert could reprogram it. Apparently they didn’t have a real expert.”

Presumably the victims were broiled be- cause a bug in the badly reprogrammed garden-grade nano made it run wild, gen- erating far too much heat in the process. Later in the book a group of crimi- nals who have infiltrated a huge tetrahe- MINIATURIZED PEOPLE in the 1966 movie Fantastic Voyage dral building make use of some illicitly swim like nanobot surgeons in a patient’s bloodstream. obtained MGN—military-grade nano. Sprayed from a canister like fire-extin- appear. On the one hand, nanotechnol- instance, to shield himself from the vac- guisher foam, the nano deconstructs ob- ogy often becomes a means to accom- uum of space, propel a starship across jects present in the building’s garage and plish anything within the realm of the the galaxy, fire bolts of energy for good rebuilds the atoms into intelligent ro- imagination, while conveniently ignor- or ill, and so on. The nanoware is really botic weaponry. During this process the ing the constraints of physical laws. Cu- just a technological cloak for supernat- garage heats up like an oven, but not too riously, on the other hand, these stories ural magic. In another era the alien de- hot, because “at about four hundred de- reveal some of the actual technical chal- vice would have been some other mind- grees, nano cooks itself.” lenges that molecular nanotechnologists enhancing black box or injectable drug. A spectacular case of spontaneous might confront if they ever were to exe- Yet because this is nanoware—nanobots human nanocombustion occurs in one cute their designs for real-world nano- that rewire the hardware and software of the most surreal sections of Neal bots. For example, it seems that most of your “wetware” (your brain)—one Stephenson’s tour de force The Dia- everyone writing nanotech fiction is does have to worry about the heat gen- mond Age. A secluded cult known as the aware that highly active nanocritters will erated by its functioning. A few of the Drummers is infected with millions of generate heat, a problem of some con- early human volunteers fried their brains nanoprocessors. When two processors when said nanocritters are func- before the right parameters for humans meet in someone’s bloodstream, they tioning inside your body. were worked out: compare notes, perform a computation In one of the early stories, the 1989 and then go on their way: a kind of dis- novella Nanoware Time, by Ian Watson, Heat was a byproduct of all the rapid tributed, Internet-like supercomputer. the nanoware has been brought to hu- molecular activity in the skull while the The computation proceeds mostly at a mankind by seemingly benevolent aliens busy little nanomachines built the nano- steady pace, but occasionally it advances that look like giant golden centipedes. ware. Thus some brains got cooked. in a spurt of activity when myriad par- When the nanoware is injected into a Vance was among the survivors. allel threads of the computation are person, it takes root in the subject’s [His] brain damage was repaired by brought together for synthesis by an brain, supplying him or her with the other nanos; sort of repaired. He’d orgy (exchange of bodily fluids is the key power to ... (can you guess?) harness been rehabilitated, retrained as a waste means of transferring these processors “demons” from a parallel dimension. recycler. and their data between people). The These demons have no will of their own orgy culminates when the nanoproces- but possess extraordinary powers that In one of the many plot threads in sors are loaded into one unlucky woman

the nanowired person can then use, for the 1997 novel / [Slant], by Greg Bear, who is promptly incinerated by the heat © TWENTIETH CENTURY FOX; COURTESY OF EVERETT COLLECTION

90 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. Anyone can obtain free food from public nanotech matter compilers, but they’re not up to the cordon bleu standards of Star Trek’s replicators.

of the nano-orgy that ensues in her Bear’s Queen of Angels are comprehen- generation, Bud gets around on in-line bloodstream. To access the computa- sive depictions of societies completely skates capable of a top speed of more tion’s result, the other Drummers mix changed by nanotechnology. In Queen than 100 kilometers an hour, and his her ashes into a soup—highly reminis- of Angels, a large proportion of the pop- music system is “a phased acoustical ar- cent of the Martian process of “grokking” ulace has been “therapied,” in which in- ray splayed across both eardrums like the dead (in essence, ritual cannibalism jected nanotech devices infiltrate a per- the seeds on a strawberry.” He’s some- to honor and fully appreciate the de- son’s brain to correct psychological im- times a little “hinky” on the skates: im- ceased) in Robert A. Heinlein’s 1961 balances and weaknesses. Many people planted nanosites incessantly twitch his novel Stranger in a Strange Land, but undergo extensive nano-enabled body muscle fibers to maximize their bulk. with the patina of a scientific rationale. modification, ranging from practical en- Together with a testosterone pump in Stephenson’s The Diamond Age and hancements for their occupation to beau- his forearm, “it was like working out at tification and the addition of exotic fea- a gym night and day, except you didn’t tures. A complex tension runs through have to actually do anything and you STORY EXCERPT: DYSTOPIAN NANO-FUTURE the society because of prejudices and at- never got sweaty.” titudes about “transforms,” “high natu- And isn’t that, in the end, what much NO LOVE IN ALL OF DWINGELOO, BY TONY rals,” and “therapied” and “simple un- of nanotech is about? A quasi-scientific Asimov’s Science Fiction, DANIEL (from therapied” individuals. way to get what you want effortlessly November 1995) In Bear’s world, nano comes in jars and at minimal cost. Instant gratification [We lived] on Kokopelli Station. There were like paste. In Stephenson’s The Dia- through the ultimate in high technology. more people living up-cable, thousands and mond Age, the key to nano is “the feed,” Compile your dreams and gather a thousands more. Still, this was just a trickle a type of nanopipeline that runs into bunch of atoms. Just add nano. of refugees compared with the billions who every household, supplying atoms as had died below. Earth was horribly worse, and needed by matter compilers, which are Graham P. Collins is a staff editor it was then that I realized I was seeing the as common as microwave ovens are to- and writer. He is also an occasional future. day. Anyone can obtain free food from science-fiction writer and the The squabbles and wars of the present public matter compilers, but they’re not webmaster of the Science Fiction had played themselves out and we’d done it, up to the cordon bleu standards of Star and Fantasy Writers of America. we’d ruined the planet. The seas were Trek’s replicators; rice they can do, but biohazard cauldrons, seething with an green vegetables come out as a paste. MORE TO ecology of war viruses. The land was haunted Airborne nanotech is ubiquitous, rang- by nanoplasms, the primal form that life had ing from almond-size surveillance mon- EXPLORE The Encyclopedia of Science Fiction. Second taken, been reduced to. Sea and land were at itors to microscopic attack and defense war—over nothing, any longer—just a edition. Edited by John Clute and Peter craft engaged in a constant struggle, like Nicholls. Palgrave, 1993. Also available from meaningless perpetual struggle between an immune system battling invaders. On Grolier on CD-ROM as The Multimedia viral life and nano algorithms caught in a bad days in the poor section of town, Encyclopedia of Science Fiction. perpetual loop. Those who crafted the this ongoing contest looks like a fog shot The Ascent of Wonder: The Evolution of Hard weaponry were dead. through with firefly sparkles of laser Science Fiction. Edited by David G. Hartwell A few million humans survived on the light. A sootlike coating composed of ca- and Kathryn Cramer. Tor Books, 1997. coasts, in the land between the warring The introduction is available online at sualties from this conflict settles on ebbs.english.vt.edu/exper/kcramer/aow.html elements. They were temporarily immune to everything and everyone. Wealthier en- the nano, but none could say for how long, Locus magazine’s online index of science claves, such as that of the Vickys (neo- fiction: www.locusmag.com/index/ since the nano evolved, its sole purpose Victorians), are protected from such Science Fiction and Fantasy Writers of finding ways to beat back the living, zombie troubles by a deep defensive perimeter of America: www.sfwa.org sea—and, incidentally, to remake whatever airborne nanobots. Analog Science Fiction and Fact: people remained into a substance that could Nanofiction is not without humor. www.analogsf.com not wield a gun and could not think to use one. The Diamond Age, particularly early in Asimov’s Science Fiction: www.asimovs.com Yet I found myself completely, the book, is told with abundant wit and Magazine of Fantasy & Science Fiction: unshakably content. Kokopelli was safe. We drollery. The story opens with the ex- www.sfsite.com/fsf/ had severed all but one cablelift, and created ploits of a spectacularly stupid lowlife Sci Fiction: www.scifi.com/scifiction/ defenses that kept the muck below at bay.

© 1995 BY BANTAM DOUBLEDAY DELL MAGAZINES; REPRINTED PERMISSION OF THE AUTHOR named Bud. A parody of the Walkman

www.sciam.com SCIENTIFIC AMERICAN 91 Copyright 2001 Scientific American, Inc. WORKINGKNOWLEDGE

FLEA TREATMENTS Killer Drops

Shampoos, powders, sprays and collars all aim to con- trol fleas on pets, but the most popular treatments to- “SPOT” FLEA TREATMENTS day are the “spot” medications. Squeeze a few drops mix with a pet’s skin oils. A few drops on the skin along a dog’s or cat’s back, and the insec- disperse readily. Tests by maker ticide will control fleas for a month. The products are Merial on midsize dogs given the available only from veterinarians, in doses adjusted recommended dose of the treatment for an animal’s weight. fipronil show that concentrations The formulations spread by mixing with a pet’s quickly spread across the body. And skin oils, which migrate as a result of body movement although concentrations are low and gravity. Some products flow into the sebaceous after 56 days, they are still high glands of hair follicles, where they are stored and se- enough to kill fleas (95 percent creted over time; others remain on the skin’s surface. of fleas die when exposed to Advantage (imidacloprid) from Bayer and Frontline 0.7 microgram per gram of fur). (fipronil) from Merial—the two market leaders—will fan out across the body in less than 12 hours and kill more than 90 percent of fleas by then. Tests by Bayer show that after 28 days, concentrations across a dog’s body decrease to as little as one part per million, but less than one tenth of that amount is needed to kill fleas, according to Bob Arther, Bayer’s manager of parasitology. And because they reside in or on skin, spot compounds do not readily wash off like treat- INSECTICIDE ments that stick to an animal’s hair. (blue) is stored Pets can be given anti-flea pills, sending medica- in a hair’s sebaceous tion into their bloodstream. But a flea must bite the pet gland, which can secrete to be exposed to the insecticide. Some pills do not kill the compound for a adult fleas but make their eggs unviable. The spot month or more. treatments kill virtually all fleas within 18 hours and prevent eggs from being laid. Buyers should be wary of over-the-counter knockoffs; many contain permethrin, which is less effective on dogs and is toxic to cats. Some critics claim that pets can be harmed by in- gesting small amounts of spot treatments as they groom themselves (although the same could be said of a spray, collar or pill). But tests required by the Envi- ronmental Protection Agency show that “animals that had received even gross overdosages had no change in their kidney or liver values,” says Bruce Klink, man- ager of veterinary services at Merial. Manufacturers say the active ingredients do not affect people. No insecticide is 100 percent safe. But be skeptical about “natural” or “chemical-free” alternatives such as vitamin B or garlic. There is little scientific proof that

the potions send fleas fleeing. —Mark Fischetti ILLUSTRATIONS BY GEORGE RETSECK

92 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. ➤ FADING FAD: When a flea bites, it deposits saliva tury. In 1999 microbiologists identified plague in in the skin. Pets can have an allergic reaction that flea-infested squirrels, chipmunks and other wild ro- causes redness and severe itching. As they scratch, dents in 22 counties surrounding Sacramento, Calif.; they develop flea-allergy dermatitis. For years, FAD in 2000 state health officials began issuing regular was the leading skin problem in dogs and cats, but warnings about plague-prone areas there. veterinarians report that “spot” treatments, because of their quick kills, have dramatically reduced the in- ➤ MAN BEFORE BEAST: The French established the cidence of the disorder. world’s first veterinary school, in Lyon in 1762. Vet- erinary science was originally developed not so ➤ BLACK DEATH: Rat fleas carry bacteria that can much to help animals but to improve the under-

DID YOU KNOW... YOU DID cause plague, such as the bubonic plague that wiped standing of zoonotic diseases in order to prevent out one third of Europe’s population in the 14th cen- their transfer to humans.

Fipronil Flow of chloride ions blocked FLEA’S NERVE CELL is the target of the spot medication’s active ingredient, which binds to a specific receptor there. The treatment fipronil (shown at right) blocks the flow of chloride ions that otherwise interrupts nerve signals. This Spot of application hyperexcites a flea’s central nervous system, sending it into a deadly seizure.

Typical medication concentration: 308 micrograms/gram at 24 hours; Nerve cell membrane 7.4 mcg/g at 56 days

Typical concentration: 30.9 mcg/g at 24 hours; 1.8 mcg/g at 56 days

Typical concentration: 91.5 mcg/g at 24 hours; 1.6 mcg/g at 56 days

Copyright 2001 Scientific American, Inc. VOYAGES

Seeing the Earth for Its Faults GEOLOGICAL TOURS AND GUIDES EXPOSE THE SECRETS OF NEW YORK CITY AND BEYOND BY MARGUERITE HOLLOWAY

Continents collided—northwestern Africa crashed and ground against North Amer- ica—and mountains as tall as the Alps rose in New England. Wind and rain beat at the peaks and wore them down, and then the land on which they stood sank below the sea. Waves and tides had their turn, and the towering ranges were largely leveled. The land lifted. The balmy weather turned cold. Ice sheets scraped across bedrock, pushing a wall of rubble before them. The climate warmed again, and glaciers melted, cre- ating a huge lake that lapped behind the wall of glacial till. Ice sheets kept melting; the lake swelled and stretched and, ulti- mately, broke through the wall at its low- est part. Rivers began to flow through valleys that the glaciers had carved. Before politics, culture and high fi- nance, these were the forces that shaped New York City. To see New York thus is to see it with the eyes of a geologist, to see the sweep of millions of years, and to un- derstand why certain things look the way they do and stand where they do. The scar where continents ground together runs down the Bronx River; the remnant Manhattan was cut in two when very NEW YORK SKYLINE owes everything to geology. of the mountains is the city’s bedrock; the high tides swelled the rivers and they met The financial district and Midtown support towering rubble pushed by the glaciers gave rise to each other in the middle of the island, at buildings because bedrock is no more than Long Island; the great lake spilled over at 125th Street, where a fault slices the city. 38 to 80 feet below the surface. In Greenwich Village, however, it is about 260 feet down. the Verrazano Narrows; the Hudson and All this insight and more can come East rivers fill valleys deepened by glaciers. from a three-hour cruise up the Hudson Wall Street and Midtown raised modern River—or the East River or through New Audubon Society’s guide to Familiar Fos-

mountains of glass and granite because York Harbor—with Sidney Horenstein sils of North America and coordinator of The Image Bank there was solid bedrock to build on, but of the American Museum of Natural environmental programs at the museum. in between lay insufficient foundation for History. “Since geology is the basis of “Because if you look at the landscape, skyscrapers—hence low-slung Green- everything, you just cannot help, once you can’t help saying, ‘What is that?’ and wich Village and Soho. Before garbage you get involved in this, to go beyond the ‘How did this happen?’”

and landfill were strewn in its wetlands, geology,” says Horenstein, author of the Horenstein’s Circle Line trips, which GUIDO ALBERTO ROSSI

94 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. for about a decade have taken place in where participants can see the scratches early summer, are as renowned for his carved in schist by roving glaciers and observations about geology as they are some of the boulders left behind by melt- for his historical anecdotes, questionable ing ice. Another lecture series covers geol- puns (“The reason all the lights are lit up ogy and travel: how to recognize geologic here is that we are approaching the Bat- features wherever you find yourself. (For tery”) and sayings. On a recent sunset information about scheduling, check the cruise packed with passengers and their American Museum of Natural History’s picnics, Horenstein explains his answer Web site at www.amnh.org or call the to the controversial question of how the main reservations line at 212-769-5200.) Big Apple got its name: from a bordello Many other paths also wander into owner named Eve and the “apples” who New York’s geologic past. Charles Mer- worked for her. He describes how 300- guerian of Hofstra University, author with million-year-old invertebrate fossils end- John E. Sanders of myriad papers on the ed up in the walls of Riverside Church’s region’s geology, has designed virtual field bell tower: the sandstone was brought in trips of the city and environs, which can from Indiana, site of an ancient tropical be viewed or ordered at www.duke- sea. And a discussion about New York’s labs.com. The Parks Department orga- main rock formations gives Horenstein a nizes some excursions as well. And the chance to quote a favorite refrain: “The South Street Seaport Museum’s off-site Bronx is gneiss, but Manhattan is full of permanent exhibit “New York Un- schist. Inwood has lost its marbles, and earthed” provides a window into the New York is full of faults.” city’s archaeology. For those who want to During other parts of the year, Hor- explore the rocks and faults alone, there enstein gives talks and leads regional ge- is a general site for New York (www.

Corbis ological field trips. “Boats are just one as- albany.net/~go/newyorker/index.html), pect of it,” Horenstein says. “I lead tours with references and materials, as well as all over, just like a gypsy cab.” One, for ex- several good books, including John Kie- ample, is a walking tour of Central Park, ran’s A Natural History of New York City and Wild New York, by Margaret GLACIAL MOVEMENTS left marks on rocks that Mittelbach and Michael Crewdson.

© YANN ARTHUS-BERTRAND can be seen in Central Park. If you prefer instead to conjure the

www.sciam.com SCIENTIFIC AMERICAN 95 Copyright 2001 Scientific American, Inc. VOYAGES

creatures that lumbered through the land- scape—to envision the ground sloths, hairy tapirs, saber-toothed tigers, woolly mammoths and mastodons that were the region’s wildlife during the Pleistocene— you can find fossil hunters to follow. The New York Paleontological Society, which Horenstein founded in 1970, has month- ly meetings most of the year and orga- nizes field trips to look for fossils. One such tour this summer took members to Hamburg, N.Y., to observe Devonian fossils, such as trilobites, in a quarry. An- other tour entailed visiting fossil track- ways in Connecticut’s Dinosaur State Park. (Go to www.nyps.org.) But no matter which city or state or region or country you are in—or whether you have easy access to a natural histo- ry or related museum—you are free to see with a geologist’s eyes. Many states have a paleontological society, and al- though you might have to join to be able to participate in the field trips, mem- bership is typically only about $20 for a family. A number of states also have a mineral society that leads tours—often these societies are members of the Amer- ican Federation of Mineralogical Soci- eties (www.amfed.org)—as well as a sur- vey that details the geology of the region. The Utah Geological Survey, for in- stance, provides local information on di- nosaur fossils. In addition, Horenstein and other ex- perts recommend a helpful series of books. Published by Mountain Press in Missoula, Mont., the Roadside Geology books are written by geologists for laypeople (see www.mountainpresspublish.com). Al- most half the 50 states have a guide, from Roadside Geology of Alaska to Roadside Geology of Wyoming. The press also publishes books on various regions of the U.S., several national parks and a few countries, including Canada and Argenti- na. To see familiar places against the backdrop of aeons is truly a grand view, both humbling and haunting; in the rise and fall of the land in your mind’s eye, you can see the earth breathe.

96 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. REVIEWS

Bloated,Whiny and Self-Important IS THE SCIENTIFIC BUREAUCRACY THE QUINTESSENTIAL SPECIAL-INTEREST GROUP? BY KEAY DAVIDSON

most National Academy of Sciences re- members of the National Academy of SCIENCE, MONEY, AND ports yet reads as briskly as a Dashiell Sciences, as saying. “Every gene that is POLITICS: POLITICAL Hammett detective story. (I devoured it discovered will lead to a cure for cancer. TRIUMPH AND until 4 A.M.) The federal science budget Maybe, but not for a long time. Even the ETHICAL EROSION has gone up, not down, he argues—and Superconducting Super Collider was said by Daniel S. Greenberg presents the statistics to prove it. The to have important implications for im- University of Chicago American people are not anti-science; in proving human health.” Press, 2001 ($35) fact, they are very pro-science, almost un- For anyone naive about the politics of critically so. “By virtually every relevant science, Science, Money, and Politics is Dan Greenberg’s profoundly important measure, the United States leads the the perfect curative. The military names new book depicts American “Big Sci- world in the financing, quality, and vol- subs and aircraft carriers after friendly ence” as a classic self-perpetuating bu- ume of research; it has held this lead since legislators; likewise, federal science bu- reaucracy—bloated, whiny and self-im- the end of World War II, and appears reaucrats affix the names of helpful portant. This bureaucracy defends big bound to maintain, and increase, its su- politicians to their buildings. The Na- (and sometimes indefensible) budgets by premacy far into the new century.” tional Institutes of Health’s “named weaving scare stories about national sci- So what, Greenberg asks, are scien- buildings recognize the good works of entific “illiteracy,” questionable “short- tists griping about? And why do their senators Warren Magnuson, Lister Hill, ages” of scientific personnel, and imag- prime beneficiaries, federal scientific Lowell Weicker, Lawton Chiles, and inary threats from foreign competitors. agencies, increasingly act like Madison Mark Hatfield, and representatives John Greenberg quotes an official of the U.S. Avenue hype factories? Fogarty, William Natcher, Silvio Conte, Office of Management and Budget: “There is too much hype,” he quotes and Claude D. Pepper.” “With the possible exception of veterans, Maxine Singer, one of the few female Politics makes strange bedfellows. In farmers, and college students, there is no group that squeals more loudly over a re- duction of federal subsidies than scien- THE EDITORS RECOMMEND tists. They are the quintessential special EXTINCT BIRDS interest group, and in effect, they make by Errol Fuller. Revised edition. Comstock Publishing the oil industry look like a piker.” Associates, Cornell University Press, Ithaca, N.Y., 2001 Startling words? Isn’t American sci- ($49.95) ence struggling to survive on meager “Never send to know for whom the bell tolls; it tolls for thee.” funds, the result of post–cold war budget John Donne’s line, which appears on the dedicatory page, cuts inspired, in part, by the American sets the tone for this elegantly written book by British painter people’s alleged indifference, even hostil- Errol Fuller. The author of two other books on birds, Fuller de- ity, toward science? scribes here the lives, times and disappearances of more than Poppycock, Greenberg replies in a 80 avian species since 1600. These accounts are illustrated book that is better documented than primarily by beautifully reproduced paintings (from a variety of artists, including John James Audubon, J. G. Keulemans and the author) as well as by early photographs and Keay Davidson is a science writer for the engravings. This revised edition updates the original 1987 publication, reinstating a few San Francisco Chronicle and author of birds found not to have disappeared and, sadly, adding a few more that have.

Carl Sagan: A Life (John Wiley, 1999). BETH PHILLIPS

98 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. TIME TRAVEL IN EINSTEIN’S UNIVERSE: THE PHYSICAL POSSIBILITIES OF TRAVEL THROUGH TIME by J. Richard Gott III. Houghton Mifflin, Boston, 2001 ($25) “One reason that time travel is so fascinating is that we have such a dictions here. The future duration of the human great desire to do it,” Gott writes. And so he explores the possibilities species, for one—more than 5,100 years but less of travel to the past and to the future. Being a professor of astro- than 7.8 million. And because “one of the things we physics at Princeton University, he does not stray from the laws of should understand about time is that we have just a physics in constructing this stimulating odyssey. Being also the man little,” he argues that “the goal of the human space- who has made a number of intriguing predictions based on the Coper- flight program should be to increase our survival nican idea that “your location is not special,” Gott offers several pre- prospects by colonizing space.”

DARK REMEDY: THE IMPACT OF THALIDOMIDE AND ITS REVIVAL AS A VITAL MEDICINE by Trent Stephens and Rock Brynner. Perseus Publishing, Cambridge, Mass., 2001 ($26) Marketed in 1957 as a sedative by Chemie Grünenthal, a German cally. Since then, thalidomide has proved effective in treating more company the authors describe as “notorious for than 130 disorders, many of them autoimmune diseases. But be- rushing drugs to market with inadequate testing,” cause of its menace to reproduction, “everyone involved in the re- thalidomide soon launched its ghastly procession of vival of this dark remedy hopes that it will be a short one”—five deformed babies. It was removed from the market more years, perhaps—to be followed by a safe analog. Stephens, and appeared to be permanently in limbo. But in professor of anatomy and embryology at Idaho State University, and 1964 an Israeli physician treating victims of leprosy historian Brynner tell this haunting story well. tried it on a patient with a severe complication of the All the books reviewed are available for purchase disease. The patient’s condition improved dramati- through www.sciam.com BETH PHILLIPS

www.sciam.com SCIENTIFIC AMERICAN 99 Copyright 2001 Scientific American, Inc. REVIEWS

the 1980s, as the cold war waned, the As Greenberg reminds us, science is a stunts. This admirable book should be U.S. Energy Department, which runs the big, expensive and all-too-human enter- required reading for science policy mak- nuclear weapons complex, feared a ma- prise, as prone to shams and hucksterism ers, science journalists and any American jor slash in its budget. So it began looking as any human endeavor. Shams and huck- who gives a damn whenever science— for a new enterprise and found it: the Hu- sterism are “news,” too; they deserve to be one of the nation’s crown jewels—falls man Genome Project. Naturally, this up- covered alongside NASA’s latest circus into irresponsible hands. set the NIH, whose officials reasoned that they were more suited to map the ge- nome. But the weapons labs are master lobbyists for their own interests. President Gerald Ford’s science advis- er Guy Stever told Greenberg, “I guaran- tee you, if somebody really had complete control of the quality of our system, and could cut out everything at the low level, we could survive beautifully on less mon- ey.” In any case, Greenberg says, there is no consistent evidence of a strong corre- lation between American scientific-tech- nological leadership and the fortunes spent seeking it: “Great Britain has ac- complished a great deal scientifically on relatively modest science budgets; the So- viet Union had relatively little to show for its enormous spending on research.” For four decades, Greenberg has been the conscience of American science writ- ers. He was the first news editor of Science magazine. In 1967 he established himself as the Jessica Mitford of science when he wrote The Politics of Pure Science, an acerbic analysis of science-government re- lations after World War II. He spent 26 years running the acclaimed gadfly news- letter Science & Government Report. We need more Greenbergs. Too many science writers cover science as uncriti- cally as fashion reporters cover fashion. (I am as guilty of this as many of my peers.) Science, Money, and Politics rightly blasts “doting science journalists who faithfully echo the paranoid fears and alarms of the scientific leadership and its publicists.” He scolds medical reporters for their “bait-and-switch” coverage of alleged breakthroughs in the war against cancer and other diseases. Typically, an exciting discovery leads the evening news or is bal- lyhooed on the front page. Follow-up re- search often undermines the claim, but do you see that in the paper? Not likely.

www.sciam.com SCIENTIFIC AMERICAN 101 Copyright 2001 Scientific American, Inc. PUZZLINGADVENTURES

Square Dancing BY DENNIS E. SHASHA

Striding gracefully, her white cape undulating each stage, each dancer can either stand in place or Answer to Last slowly behind her, the director of a famous dance take one step in one of four directions: to the left, Month’s Puzzle company enters my office. “You’ve heard of ghost to the right, forward or backward. There are two The solution to the writers,” she says. “I need a ghost choreographer.” important conditions: two dancers cannot swap first problem of I motion for her to sit, and she outlines the partic- positions during a step, nor can two dancers occu- the unreliable oracle is that you can ulars of her problem. py the same space at the end of a step. Above all, convert your initial The dance company consists of 12 men (colored the dancers want to avoid collisions. stake of $100 to blue in the illustrations below) and 20 women (col- The starting and ending configurations of the at least $9,309 if you ored red). At a key point in their dance, they go dancers are shown below. Can you design their use the optimum from a configuration in which the men surround moves at each step so that they reach the final po- betting strategy, which calls for mak- the women to one in which the women surround sition without violating the conditions? It may help ing a first bet of just the men. The transition has three stages. During to plot the moves on a piece of graph paper. under $81.82. The strategy is fully explained at the Puzzling Adventures page at www.sciam. STARTING CONFIGURATION ENDING CONFIGURATION com. For the second problem, in which you must make all the bets in advance, the best strategy can guarantee you only $1,600. Again, see the Web site for the full explanation.

Web Solution For a peek at the answer to this month’s problem, visit www.sciam.com JOHNNY JOHNSON

102 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc. ANTIGRAVITY

Enter the Dragon Exhibit ONCE UPON A TIME THERE WAS A DRAGON AND A BEAUTIFUL PRINCESS, I MEAN ACTRESS, WHO FORGOT THAT LARGE CARNIVORES HAVE A BASIC INSTINCT BY STEVE MIRSKY

Journalists often cite an old bromide: For another thing, the dragon was no Monitor (which, by the way, makes a ter- “Dog Bites Man” is not news, but, on the doubt well fed, which zoo animals tend rific Father’s Day present), remembered other hand, “Man Bites Dog” is news. to be. As for nearsightedness, Claudio the first dragon he met on Komodo: And on the other foot, which is exactly Ciofi of the Zoological Society of Lon- “God, he stank. Oh my goodness. God, where it happened, “Komodo Dragon don, writing in the March 1999 issue of he was smelly. I called him Stinky. I went Bites Newspaper Editor” is definitely this magazine, revealed that “monitors back and told the family that I met news. Especially when the newspaper ed- [the Komodo is a species of monitor Stinky.” Auffenberg later learned that itor in question is the San Francisco lizard] can see objects as far away as 300 Stinky was a “psychotic animal that had Chronicle’s semi-celebrity executive edi- meters, so vision does play a role in hunt- probably killed two people on Komodo tor Phil Bronstein, who was bitten while ing, especially as their eyes are better at prior to our visit.” Ahhh, good times. getting a special up-close-and-perilous picking up movement than at discerning (The entire lecture can be found at http:// tour of the Los Angeles Zoo’s Komodo stationary objects.” natzoo.si.edu/hilights/lectures.htm.) exhibit in early June, arranged for him by Anyway, my interest in Komodos his full-celebrity wife, the beautiful ac- was born while working with Ciofi on his tress Sharon Stone. The visit was Stone’s article. And even though I learned that idea of an early Father’s Day celebration, the dragons’ saliva harbors numerous which often includes dinner out. species of septic bacteria that can kill Evidently, a zookeeper advised Bron- even if the bite itself doesn’t, I was still stein to remove his white sneakers so they dumb enough to get next to a Komodo at wouldn’t be confused with the dragon’s the National Zoo in Washington, D.C. If customary meal of white rat. This strat- you look very closely at the Ciofi story’s egy clearly backfired. A letter to the photo credits, you’ll see that I snapped Chronicle asked, “Whose bright idea was the dragon. (By the way, my visit to the it to remove the white shoes from the zoo’s dragon enclosure was arranged for white feet of a white man in the hopes of me by my beautiful actress wife, uh, Mor- not confusing a nearsighted, simplemind- gan Fairchild. Yeah, that’s the ticket.) ed, ravenously hungry lizard?” Back in the glamorous world of I take umbrage at this reader’s um- celebrity maulings, Bronstein underwent brage, and I feel I am in a unique position surgery to reattach a few tendons, and to comment on this incident. That posi- both his foot and sense of humor seem in- tion is seated, in front of my computer, For my final rejoinder to the Chroni- tact. He told his own paper, “That’s how far away from any Komodos. For one cle reader, I note that Komodos are more it goes when a movie star and a dragon thing, the Komodo dragon is not merely single-minded than simpleminded—as an are involved.” Meanwhile Stone’s view of a lizard. It is “the world’s largest living ambush-predator, the Komodo must be the foot-and-mouth business was more lizard—a ferocious carnivore—found on cunning and stealthy. And it should Grimm: “Very few women get to see the steep-sloped island of Komodo in the probably stay downwind of its prey, their knight wrestle the dragon. And I Lesser Sunda chain of the Indonesian what with its amazing body funk. In a think he’s kind of fabulous. And kind of archipelago,” according to the hilarious lecture, Walter Auffenberg, the world’s hot.” With the raging infection that usu- and completely accurate bit by the leg- foremost Komodo expert and author of ally follows a Komodo bite, perhaps that

ILLUSTRATION BY MATT COLLINS; PHOTOGRAPH FRANK VERONSKY endary radio comedians Bob and Ray. The Behavioral Ecology of the Komodo heat is just a fever.

www.sciam.com SCIENTIFIC AMERICAN 103 Copyright 2001 Scientific American, Inc. ENDPOINTS

Why do dogs get blue, not red, eyes in flash photos? Q —B. Sloot, Mount Prospect, Ill. Veterinary ophthalmologist J. Phillip Pickett of the Virginia- flection. A buff-colored cocker spaniel will generally show a Maryland Regional College of Veterinary Medicine explains: yellow tapetal reflection. Most young puppies and kittens “Red eye,” the nemesis of amateur photographers, occurs have a blue tapetal reflection until the structures in the back when someone looks directly at the camera while a picture is of the eye fully mature at six to eight months of age. “Color taken. If the flash is on the same axis as the visual axis of the dilute” dogs and cats, such as red Siberian huskies and blue camera, the reflection off the blood vessels in the person’s reti- point Siamese cats, may have no tapetal pigment and may na can produce an eerie, satanic look, the so-called red reflex. therefore exhibit a red reflex just like human beings. Dogs, cats and almost all domestic animals have a special reflective layer in the back of the eye termed the tapetum. In- coming light passes through the animal’s retina and is then If a used needle can transmit HIV, reflected back through the retina a second time from the why can’t a mosquito? tapetal layer. This double stimulation helps these species to Q see better in dim light. The color of this tapetal layer varies —P. Smith, Masclat, France to some extent with an animal’s coat color. A black Labrador Laurence Corash, chief medical officer of Cerus Corporation, retriever, for example, will usually have a green tapetal re- explains: The AIDS virus (HIV) on used needles is infectious when injected into a human because the virus can bind to T cells and start to replicate. The human T cell is a very specific host cell for HIV. When a mosquito feeds on a person with HIV, the HIV enters the insect’s gut, where it cannot find a host. The malaria parasite, in contrast, can survive, multiply and mature in the mosquito’s gut. The parasites then migrate to the insect’s salivary glands. Because mosquitoes inject their saliva when they bite, the parasite is passed along to the next human on whom the insect feeds. The complex interaction between the infectious agent and the mosquito is thus re- quired for malaria transmission. HIV, however, deteriorates in the gut before the mosquito bites again and therefore is not transmitted to the insect’s next victim.

For the complete text of these and many other answers, visit Ask the Experts (www.sciam.com/askexpert).

Scientific American (ISSN 0036-8733), published monthly by Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111. Copyright © 2001 by Scientific American, Inc. All rights reserved. No part of this issue may be reproduced by any mechanical, photographic or electronic process, or in the form of a phono- graphic recording, nor may it be stored in a retrieval system, transmitted or otherwise copied for public or private use without written permission of the publisher. Periodicals postage paid at New York, N.Y., and at additional mailing offices and also Standard A enclosure postage paid (for edition “50000” only). Canada Post International Publications Mail (Canadian Distribution) Sales Agreement No. 242764. Canadian BN No. 127387652RT; QST No. Q1015332537. Subscription rates: one year $34.97, Canada $49, International $55. Postmaster: Send address changes to Scientific American, Box 3187, Harlan, Iowa 51537. Reprints available: write Reprint Department, Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111; (212) 451-8877; fax: (212) 355-0408 or send

e-mail to [email protected] Subscription inquiries: U.S. and Canada (800) 333-1199; other (515) 247-7631. Printed in U.S.A. PATRICK HARDIN

104 SCIENTIFIC AMERICAN SEPTEMBER 2001 Copyright 2001 Scientific American, Inc.