Copyright 2001 Scientific American, Inc.40 SCIENTIFIC AMERICAN SEPTEMBER 2001 IMAGE BY FELICE...

Copyright 2001 Scientific American, Inc.

w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 39

artNANOFABRICATION

The

of

RESEARCHERS ARE DISCOVERING CHEAP,

EFFICIENT WAYS TO MAKE STRUCTURES

ONLY A FEW BILLIONTHS OF A METER ACROSS

BY GEORGE M. WHITESIDES AND J. CHRISTOPHER LOVE

BuildingSmall

INTRICATE DIFFRACTION PATTERNS are 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.


“Make it small!”is a tech-nological edict that has changed theworld. The development of microelec-tronics—first the transistor and then theaggregation of transistors into micro-processors, memory chips and con-trollers—has brought forth a cornucopiaof machines that manipulate informa-tion by streaming electrons through sil-icon. Microelectronics rests on tech-niques that routinely fabricate structuresalmost as small as 100 nanometersacross (that is, 100 billionths of a meter).This size is tiny by the standards ofeveryday experience—about one thou-sandth the width of a human hair—butit is large on the scale of atoms and mol-ecules. The diameter of a 100-nanome-ter-wide wire would span about 500atoms of silicon.

The idea of making “nanostruc-tures” that comprise just one or a fewatoms has great appeal, both as a scien-tific challenge and for practical reasons.A structure the size of an atom repre-sents a fundamental limit: to make any-thing smaller would require manipulat-ing atomic nuclei—essentially, transmut-ing one chemical element into another. Inrecent years, scientists have learned var-ious techniques for building nanostruc-tures, but they have only just begun to

investigate their properties and potentialapplications. The age of nanofabricationis here, and the age of nanoscience hasdawned, but the age of nanotechnol-ogy—finding practical uses for nano-structures—has not really started yet.

The Conventional ApproachRESEARCHERS may well develop nano-structures as electronic components, butthe most important applications couldbe quite different: for example, biolo-gists might use nanometer-scale particlesas minuscule sensors to investigate cells.Because scientists do not know whatkinds of nanostructures they will ulti-mately want to build, they have not yetdetermined the best ways to constructthem. Photolithography, the technologyused to manufacture computer chipsand virtually all other microelectronicsystems, can be refined to make struc-tures smaller than 100 nanometers, butdoing so is very difficult, expensive andinconvenient. In a search to find betteralternatives, nanofabrication researchershave adopted the philosophy “Let a thou-sand flowers bloom.”

First, consider the advantages anddisadvantages of photolithography. Man-ufacturers use this phenomenally pro-ductive technology to churn out three bil-

lion transistors per second in the U.S.alone. Photolithography is basically anextension of photography. One firstmakes the equivalent of a photographicnegative containing the pattern requiredfor some part of a microchip’s circuitry.This negative, which is called the mask ormaster, is then used to copy the patterninto the metals and semiconductors of amicrochip. As is the case with photogra-phy, the negative may be hard to make,but creating multiple copies is easy, be-cause the mask can be used many times.The process thus separates into twostages: the preparation of the mask (aone-time event, which can be slow andexpensive) and the use of the mask tomanufacture replicas (which must berapid and inexpensive).

To make a mask for a part of a com-puter chip, a manufacturer first designsthe circuitry pattern on a convenientlylarge scale and converts it into a patternof opaque metallic film (usually chromi-um) on a transparent plate (usually glassor silica). Photolithography then reducesthe size of the pattern in a process anal-ogous to that used in a photographicdarkroom [see illustration on oppositepage]. A beam of light (typically ultravi-olet light from a mercury arc lamp)shines through the chromium mask, thenpasses through a lens that focuses the im-age onto a photosensitive coating of or-ganic polymer (called the photoresist) onthe surface of a silicon wafer. The partsof the photoresist struck by the light canbe selectively removed, exposing parts ofthe silicon wafer in a way that replicatesthe original pattern.

Why not use photolithography tomake nanostructures? The technologyfaces two limitations. The first is that theshortest wavelength of ultraviolet lightcurrently used in production processes isabout 250 nanometers. Trying to makestructures much smaller than half of thatspacing is like trying to read print that is

40 S C I E N T I F I C A M E R I C A N S E P T E M B E R 2 0 0 1

IMAG

E B

Y FE

LIC

E F

RAN

KE

L, W

ITH

TE

CH

NIC

AL H

ELP

FR

OM

KAT

ER

I E

. P

AUL;

CO

UR

TESY

OF

GE

OR

GE

M.

WH

ITE

SID

ES

Ha

rva

rd U

niv

ersi

ty (

pa

ge

38

)

� The development of nanotechnology will depend on the ability of researchers toefficiently manufacture structures smaller than 100 nanometers (100 billionthsof a meter) across.

� Photolithography, the technology now used to fabricate circuits on microchips,can be modified to produce nanometer-scale structures, but the modificationswould be technically difficult and hugely expensive.

� Nanofabrication methods can be divided into two categories: top-down methods,which carve out or add aggregates of molecules to a surface, and bottom-upmethods, which assemble atoms or molecules into nanostructures.

� Two examples of promising top-down methods are soft lithography and dip-penlithography. Researchers are using bottom-up methods to produce quantum dotsthat can serve as biological dyes.

Overview/Nanofabrication


too tiny: diffraction causes the features toblur and meld together. Various techni-cal improvements have made it possibleto push the limits of photolithography.The smallest structures created in massproduction are somewhat larger than100 nanometers, and complex micro-electronic structures have been madewith features that are only 70 nanome-ters across. But these structures are stillnot small enough to explore some of themost interesting aspects of nanoscience.

The second limitation follows fromthe first: because it is technically difficultto make such small structures using light,it is also very expensive to do so. The pho-tolithographic tools that will be used tomake chips with features well below 100nanometers will each cost tens to hun-dreds of millions of dollars. This expensemay or may not be acceptable to manu-facturers, but it is prohibitive for the bi-ologists, materials scientists, chemists andphysicists who wish to explore nanosci-ence using structures of their own design.

Future NanochipsTHE ELECTRONICS industry is deeplyinterested in developing new methods fornanofabrication so that it can continue itslong-term trend of building ever smaller,faster and less expensive devices. It wouldbe a natural evolution of microelectron-ics to become nanoelectronics. But be-cause conventional photolithography be-comes more difficult as the dimensions ofthe structures become smaller, manufac-turers are exploring alternative technolo-gies for making future nanochips.

One leading contender is electron-beam lithography. In this method, thecircuitry pattern is written on a thin poly-mer film with a beam of electrons. Anelectron beam does not diffract at atom-ic scales, so it does not cause blurring ofthe edges of features. Researchers haveused the technique to write lines withwidths of only a few nanometers in a lay-er of photoresist on a silicon substrate.The electron-beam instruments current-ly available, however, are very expensiveand impractical for large-scale manufac-turing. Because the beam of electrons isneeded to fabricate each structure, theprocess is similar to the copying of a

manuscript by hand, one line at a time.If electrons are not the answer, what

is? Another contender is lithography us-ing x-rays with wavelengths between 0.1and 10 nanometers or extreme ultravio-let light with wavelengths between 10and 70 nanometers. Because these formsof radiation have much shorter wave-lengths than the ultraviolet light current-ly used in photolithography, they mini-mize the blurring caused by diffraction.These technologies face their own set ofproblems, however: conventional lensesare not transparent to extreme ultravio-let light and do not focus x-rays. Fur-thermore, the energetic radiation rapid-

ly damages many of the materials used inmasks and lenses. But the microelectron-ics industry clearly would prefer to makeadvanced chips using extensions of fa-miliar technology, so these methods arebeing actively developed. Some of thetechniques (for example, advanced ul-traviolet lithography for chip produc-tion) will probably become commercialrealities. They will not, though, make in-expensive nanostructures and thus willdo nothing to open nanotechnology to abroader group of scientists and engineers.

The need for simpler and less expen-sive methods of fabricating nanostruc-tures has stimulated the search for un-

BR

YAN

CH

RIS

TIE

GEORGE M. WHITESIDES and J. CHRISTOPHER LOVE work together on unconventional meth-ods of nanofabrication in the department of chemistry at Harvard University. Whitesides,a professor of chemistry, received his Ph.D. from the California Institute of Technology in1964 and joined the Harvard faculty in 1982. Love is a graduate student and a member ofWhitesides’s research group. He received his bachelor’s degree in chemistry from the Uni-versity of Virginia in 1999 and his master’s degree from Harvard in 2001.TH

E AU

THO

RS

A laser beam writes the

circuit pattern for a

microchip on a layer of

light-sensitive polymer

that rests atop a layer

of chromium and a

glass substrate. The

sections of polymer

struck by the beam can

be selectively removed.

1

2

3

CONVENTIONAL PHOTOLITHOGRAPHY

The exposed sections of

chromium are also removed,

and the rest of the polymer

is dissolved. The result is a

mask—the equivalent of a

photographic negative.

When a beam of ultraviolet light is directed at

the mask, the light passes through the gaps

in the chromium. A lens shrinks the pattern by

focusing the light onto a layer of photoresist

on a silicon wafer.

The exposed parts of the photoresist are

removed, allowing the replication of the pattern

in miniature on the silicon chips.

4

LASER BEAM1

23

4

GLASS SUBSTRATE

CHROMIUM LAYER

ULTRAVIOLET LIGHT

MASK

SILICON WAFER WITH LAYER OF PHOTORESIST

SILICON CHIPS

LENS

w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 41Copyright 2001 Scientific American, Inc.


BR

YAN

CH

RIS

TIE

SOFT LITHOGRAPHY

MAKING AN ELASTIC STAMP

MICROMOLDING IN CAPILLARIES

1 A liquid precursor to

polydimethylsiloxane (PDMS) is

poured over a bas-relief master

produced by photolithography or

electron-beam lithography.

2 The liquid is cured into a rubbery solid that

matches the original pattern. 3 The PDMS stamp is peeled off the master.

The PDMS stamp is placed on a hard surface,

and a liquid polymer flows into the recesses

between the surface and the stamp.

The polymer solidifies into the

desired pattern, which may contain

features smaller than 10 nanometers.

21

LIQUID PRECURSOR TO PDMS

SOLIDIFIED POLYMER

2 The thiols form a self-assembled monolayer on the gold

surface that reproduces the stamp’s pattern; features in the

pattern are as small as 50 nanometers.

MICROCONTACT PRINTING

PDMS STAMP

PHOTORESISTMASTER

SELF-ASSEMBLEDMONOLAYER

Printing, molding and other mechanical processescarried out using an elastic stamp can producepatterns with nanoscale features.

Such techniques can fabricate devices that might be used in optical communications orbiochemical research.

LIQUID POLYMER

GOLD SURFACE

THIOL INK

The PDMS stamp is inked with a solution consisting of

organic molecules called thiols and then pressed against

a thin film of gold on a silicon plate.

1


conventional approaches that have notbeen explored by the electronics indus-try. We first became interested in the top-ic in the 1990s when we were engaged inmaking the simple structures required inmicrofluidic systems—chips with chan-nels and chambers for holding liquids.This lab-on-a-chip has myriad potentialuses in biochemistry, ranging from drugscreening to genetic analysis. The chan-nels in microfluidic chips are enormousby the standards of microelectronics: 50microns (or 50,000 nanometers) wide,rather than 100 nanometers. But thetechniques for producing those channelsare quite versatile. Microfluidic chips canbe made quickly and inexpensively, andmany are composed of organic polymersand gels—materials not found in theworld of electronics. We discovered that

we could use similar techniques to cre-ate nanostructures.

The methods represented, in a sense,a step backward in technology. Insteadof using the tools of physics—light andelectrons—we employed mechanical pro-cesses that are familiar in everyday life:printing, stamping, molding and em-bossing. The techniques are called softlithography because the tool they have incommon is a block of polydimethyl-siloxane (PDMS)—the rubbery polymerused to caulk the leaks around bathtubs.(Physicists often refer to such organicchemicals as “soft matter.”)

To carry out reproduction using softlithography, one first makes a mold or astamp. The most prevalent procedure isto use photolithography or electron-beam lithography to produce a pattern ina layer of photoresist on the surface of asilicon wafer. This process generates abas-relief master in which islands of pho-toresist stand out from the silicon [seetop illustration on opposite page]. Thena chemical precursor to PDMS—a free-flowing liquid—is poured over the bas-relief master and cured into the rubberysolid. The result is a PDMS stamp that

matches the original pattern with aston-ishing fidelity: the stamp reproduces fea-tures from the master as small as a fewnanometers. Although the creation of afinely detailed bas-relief master is expen-sive because it requires electron-beamlithography or other advanced techniques,copying the pattern on PDMS stamps ischeap and easy. And once a stamp is inhand, it can be used in various inexpen-sive ways to make nanostructures.

The first method—originally devel-oped by Amit Kumar, a postdoctoral stu-dent in our group at Harvard Universi-ty—is called microcontact printing. ThePDMS stamp is “inked” with a reagentsolution consisting of organic moleculescalled thiols [see middle illustration onopposite page]. The stamp is then broughtinto contact with an appropriate sheet of

“paper”—a thin film of gold on a glass,silicon or polymer plate. The thiols reactwith the gold surface, forming a highlyordered film (called a self-assembledmonolayer, or SAM) that replicates thestamp’s pattern. Because the thiol inkspreads a bit after it contacts the surface,the resolution of the monolayer cannotbe quite as high as that of the PDMSstamp. But when used correctly, micro-contact printing can produce patternswith features as small as 50 nanometers.

Another method of soft lithography,called micromolding in capillaries, in-volves using the PDMS stamp to moldpatterns. The stamp is placed on a hardsurface, and a liquid polymer flows bycapillary action into the recesses betweenthe surface and the stamp [see bottom il-lustration on opposite page]. The poly-mer then solidifies into the desired pat-tern. This technique can replicate struc-tures smaller than 10 nanometers. It isparticularly well suited for producingsubwavelength optical devices, wave-guides and optical polarizers, all ofwhich could be used in optical fiber net-works and eventually perhaps in opticalcomputers. Other possible applications

are in the field of nanofluidics, an exten-sion of microfluidics that would involveproducing chips for biochemical researchwith channels only a few nanometerswide. At that scale, fluid dynamics mayallow new ways to separate materialssuch as fragments of DNA.

These methods require no specialequipment and in fact can be carried outby hand in an ordinary laboratory. Con-ventional photolithography must takeplace in a clean-room facility devoid ofdust and dirt; if a piece of dust lands on themask, it will create an unwanted spot onthe pattern. As a result, the device beingfabricated (and sometimes neighboringdevices) may fail. Soft lithography is gen-erally more forgiving because the PDMSstamp is elastic. If a piece of dust getstrapped between the stamp and the sur-

face, the stamp will compress over thetop of the particle but maintain contactwith the rest of the surface. Thus, the pat-tern will be reproduced correctly exceptfor where the contaminant is trapped.

Moreover, soft lithography can pro-duce nanostructures in a wide range ofmaterials, including the complex organ-ic molecules needed for biological stud-ies. And the technique can print or moldpatterns on curved as well as planar sur-faces. But the technology is not ideal formaking the structures required for com-plex nanoelectronics. Currently all inte-grated circuits consist of stacked layers ofdifferent materials. Deformations anddistortions of the soft PDMS stamp canproduce small errors in the replicatedpattern and a misalignment of the pat-tern with any underlying patterns previ-ously fabricated. Even the tiniest distor-tions or misalignments can destroy amultilayered nanoelectronic device.Therefore, soft lithography is not wellsuited for fabricating structures withmultiple layers that must stack preciselyon top of one another.

Researchers have found ways, how-ever, to correct this shortcoming—at


These methods require no special equipment and in factcan be carried out by hand in an ordinary lab.



BR

YAN

CH

RIS

TIE

least in part—by employing a rigid stampinstead of an elastic one. In a techniquecalled step-and-flash imprint lithogra-phy, developed by C. Grant Willson ofthe University of Texas, photolithogra-phy is used to etch a pattern into a quartzplate, yielding a rigid bas-relief master.Willson eliminated the step of making aPDMS stamp from the master; insteadthe master itself is pressed against a thinfilm of liquid polymer, which fills themaster’s recesses. Then the master is ex-posed to ultraviolet light, which solidifiesthe polymer to create the desired replica.A related technique called nanoimprintlithography, developed by Stephen Y.Chou of Princeton University, also em-ploys a rigid master but uses a film ofpolymer that has been heated to a tem-perature near its melting point to facilitatethe embossing process. Both methods canproduce two-dimensional structures withgood fidelity, but it remains to be seenwhether the techniques are suitable formanufacturing electronic devices.

Pushing Atoms AroundTHE CURRENT REVOLUTION innanoscience started in 1981 with the in-vention of the scanning tunneling micro-scope (STM), for which Heinrich Rohrerand Gerd K. Binnig of the IBM Zurich

Research Laboratory received the NobelPrize in Physics in 1986. This remark-able device detects small currents thatpass between the microscope’s tip andthe sample being observed, allowing re-searchers to “see” substances at the scaleof individual atoms. The success of theSTM led to the development of otherscanning probe devices, including theatomic force microscope (AFM). Theoperating principle of the AFM is simi-lar to that of an old-fashioned phono-graph. A tiny probe—a fiber or a pyra-mid-shaped tip that is typically betweentwo and 30 nanometers wide—is broughtinto direct contact with the sample. Theprobe is attached to the end of a can-tilever, which bends as the tip movesacross the sample’s surface. The deflec-tion is measured by reflecting a beam oflaser light off the top of the cantilever.The AFM can detect variations in verti-cal surface topography that are smallerthan the dimensions of the probe.

But scanning probe devices can domore than simply allow scientists to ob-serve the atomic world—they can also beused to create nanostructures. The tip onthe AFM can be used to physically movenanoparticles around on surfaces and toarrange them in patterns. It can also beused to make scratches in a surface (or

more commonly, in monolayer films ofatoms or molecules that coat the surface).Similarly, if researchers increase the cur-rents flowing from the tip of the STM, themicroscope becomes a very small sourcefor an electron beam, which can be usedto write nanometer-scale patterns. TheSTM tip can also push individual atomsaround on a surface to build rings andwires that are only one atom wide.

An intriguing new scanning probefabrication method is called dip-pen lith-ography. Developed by Chad A. Mirkinof Northwestern University, this tech-nique works much like a goose-featherpen [see illustration at left]. The tip of theAFM is coated with a thin film of thiolmolecules that are insoluble in water butreact with a gold surface (the same chem-istry used in microcontact printing).When the device is placed in an atmo-sphere containing a high concentrationof water vapor, a minute drop of watercondenses between the gold surface andthe microscope’s tip. Surface tensionpulls the tip to a fixed distance from thegold, and this distance does not changeas the tip moves across the surface. Thedrop of water acts as a bridge over whichthe thiol molecules migrate from the tipto the gold surface, where they are fixed.Researchers have used this procedure towrite lines a few nanometers across.

Although dip-pen lithography is rel-atively slow, it can use many differenttypes of molecules as “inks” and thusbrings great chemical flexibility to nano-meter-scale writing. Researchers havenot yet determined the best applicationsfor the technique, but one idea is to usethe dip-pen method for precise modifica-tions of circuit designs. Mirkin has re-cently demonstrated that a variant of theink used in dip-pen lithography can writedirectly on silicon.

An interesting cousin to these tech-niques involves another kind of nano-structure, called a break junction. If youbreak a thin, ductile metal wire into twoparts by pulling sharply, the processseems abrupt to a human observer, but itactually follows a complex sequence.When the force used in breaking the wireis first applied, the metal begins to yieldand flow, and the diameter of the wire

DIP-PEN LITHOGRAPHY

PYRAMIDAL TIP

of an atomic force micro-

scope (AFM) is coated with

a thin film of thiol molecules.

A minute drop of water

condenses between the

microscope’s tip and a gold surface.

The thiols migrate from the tip to the surface,

where they form a self-assembled monolayer.

AFM CANTILEVER

AFM TIP

GOLD SURFACE

THIOL MOLECULES

SELF-ASSEMBLEDMONOLAYER

DROP OF WATER


decreases. As the two ends move apart,the wire gets thinner and thinner until, inthe instant just before breaking, it is a sin-gle atom in diameter at its narrowestpoint. This process of thinning a wire toa break junction can be detected easily bymeasuring the current that flows throughthe wire. When the wire is slender enough,current can flow only in discrete quanti-ties (that is, current flow is quantized).

The break junction is analogous totwo STM tips facing each other, and sim-ilar physical rules govern the current thatflows through it. Mark A. Reed of YaleUniversity has pioneered a particularly in-ventive use of the break junction. He builta device that enabled a thin junction to bebroken under carefully controlled condi-tions and then allowed the broken tips tobe brought back together or to be heldapart at any distance with an accuracy ofa few thousandths of a nanometer. By ad-justing the distance between the tips in thepresence of an organic molecule thatbridged them, Reed was able to measurea current flowing across the organicbridge. This experiment was an impor-tant step in the development of technolo-

gies for using single organic molecules aselectronic devices such as diodes and tran-sistors [see “Computing with Mole-cules,” by Mark A. Reed and James M.Tour; Scientific American, June 2000].

Top-Down and Bottom-UpALL THE FORMS of lithography wehave discussed so far are called top-downmethods—that is, they begin with a pat-tern generated on a larger scale and re-duce its lateral dimensions (often by afactor of 10) before carving out nano-structures. This strategy is required infabricating electronic devices such as mi-crochips, whose functions depend moreon their patterns than on their dimen-sions. But no top-down method is ideal;none can conveniently, cheaply andquickly make nanostructures of any ma-terial. So researchers have shown grow-ing interest in bottom-up methods,which start with atoms or molecules andbuild up to nanostructures. These meth-ods can easily make the smallest nano-structures—with dimensions betweentwo and 10 nanometers—and do so in-expensively. But these structures are usu-

ally generated as simple particles in sus-pension or on surfaces, rather than as de-signed, interconnected patterns.

Two of the most prominent bottom-up methods are those used to makenanotubes and quantum dots. Scientistshave made long, cylindrical tubes of car-bon by a catalytic growth process thatemploys a nanometer-scale drop ofmolten metal (usually iron) as a catalyst[see “Nanotubes for Electronics,” byPhilip G. Collins and Phaedon Avouris;Scientific American, December 2000].The most active area of research inquantum dots originated in the labora-tory of Louis E. Brus (then at Bell Labo-ratories) and has been developed by A. Paul Alivisatos of the University ofCalifornia at Berkeley, Moungi G.Bawendi of the Massachusetts Instituteof Technology, and others. Quantumdots are crystals containing only a fewhundred atoms. Because the electrons ina quantum dot are confined to widelyseparated energy levels, the dot emitsonly one wavelength of light when it isexcited. This property makes the quan-tum dot useful as a biological marker [see


PhotolithographyAdvantages: The electronics industry is already familiar withthis technology because it is currently used to fabricatemicrochips. Manufacturers can modify the technique to producenanometer-scale structures by employing electron beams,x-rays or extreme ultraviolet light.Disadvantages: The necessary modifications will be expensiveand technically difficult. Using electron beams to fashionstructures is costly and slow. X-rays and extreme ultravioletlight can damage the equipment used in the process.

Scanning Probe MethodsAdvantages: The scanning tunneling microscope and the atomicforce microscope can be used to move individual nanoparticlesand arrange them in patterns. The instruments can build ringsand wires that are only one atom wide.Disadvantages: The methods are too slow for mass production.Applications of the microscopes will probably be limited to thefabrication of specialized devices.

Soft LithographyAdvantages: This method allows researchers to inexpensivelyreproduce patterns created by electron-beam lithography orother related techniques. Soft lithography requires no specialequipment and can be carried out by hand in an ordinarylaboratory.Disadvantages: The technique is not ideal for manufacturing themultilayered structures of electronic devices. Researchers aretrying to overcome this drawback, but it remains to be seenwhether these efforts will be successful.

Bottom-Up MethodsAdvantages: By setting up carefully controlled chemicalreactions, researchers can cheaply and easily assemble atomsand molecules into the smallest nanostructures, withdimensions between two and 10 nanometers.Disadvantages: Because these methods cannot producedesigned, interconnected patterns, they are not well suited forbuilding electronic devices such as microchips.

Nanofabrication: Comparing the MethodsResearchers 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.


“Less Is More in Medicine,” on page 66].One procedure used to make quan-

tum dots involves a chemical reaction be-tween a metal ion (for example, cadmi-um) and a molecule that is able to donatea selenium ion. This reaction generatescrystals of cadmium selenide. The trick isto prevent the small crystals from stick-ing together as they grow to the desiredsize. To insulate the growing particlesfrom one another, researchers carry outthe reaction in the presence of organicmolecules that act as surfactants, coatingthe surface of each cadmium selenideparticle 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. Thereaction can generate particles with a va-riety of shapes, including spheres, rodsand tetrapods (four-armed particles sim-ilar to toy jacks).

It is important to synthesize the quan-tum dots with uniform size and composi-tion, because the size of the dot deter-mines its electronic, magnetic and opticalproperties. Researchers can select the sizeof the particles by varying the length oftime for the reaction. The organic coatingalso helps to set the size of the particles.When the nanoparticle is small (on thescale of molecules), the organic coating isloose and allows further growth; as theparticle enlarges, the organic moleculesbecome crowded. There is an optimumsize for the particles that allows the moststable packing of the organic moleculesand thus provides the greatest stabiliza-tion for the surfaces of the crystals.

These cadmium selenide nanoparti-cles promise some of the first commercialproducts of nanoscience: Quantum DotCorporation has been developing thecrystals for use as biological labels. Re-searchers can tag proteins and nucleicacids with quantum dots; when the sam-ple is illuminated with ultraviolet light,the crystals will fluoresce at a specificwavelength and thus show the locationsof the attached proteins. Many organicmolecules also fluoresce, but quantumdots have several advantages that makethem better markers. First, the color of


QUANTUM DOT ASSEMBLY

When the crystal reaches its

optimum size, the organic

molecules coat its surface in

a stable packing.

1

2

3

A chemical reaction brings

together cadmium ions

(purple), selenium ions

(green) and organic

molecules (red spheres

with blue tails).

The organic molecules act

as surfactants, binding to

the surface of the cadmium

selenide crystal as it grows.

BR

YAN

CH

RIS

TIE

Crystals called quantum dots contain only a few hundred atoms andemit different wavelengths of light, depending on their size. They maybecome useful as biological markers of cellular activity.


a quantum dot’s fluorescence can be tai-lored by changing the dot’s size: the larg-er the particle, the more the emitted lightis shifted toward the red end of the spec-trum. Second, if all the dots are the samesize, their fluorescence spectrum is nar-row—that is, they emit a very pure color.This property is important because it al-lows particles of different sizes to beused as distinguishable labels. Third, thefluorescence of quantum dots does notfade on exposure to ultraviolet light, asdoes that of organic molecules. Whenused as dyes in biological research, thedots can be observed for convenientlylong periods.

Scientists are also investigating thepossibility of making structures from col-loids—nanoparticles in suspension. Chris-topher B. Murray and a team at the IBMThomas J. Watson Research Center areexploring the use of such colloids to cre-ate a medium for ultrahigh-density datastorage. The IBM team’s colloids con-tain magnetic nanoparticles as small as

three nanometers across, each composedof about 1,000 iron and platinum atoms.When the colloid is spread on a surfaceand the solvent allowed to evaporate,the nanoparticles crystallize in two- orthree-dimensional arrays. Initial studiesindicate that these arrays can potential-ly store trillions of bits of data per squareinch, giving them a capacity 10 to 100times greater than that of present mem-ory devices.

The Future ofNanofabricationTHE INTEREST in nanostructures is sogreat that every plausible fabricationtechnique is being examined. Althoughphysicists and chemists are now doingmost of the work, biologists may alsomake important contributions. The cell(whether mammalian or bacterial) is rel-atively large on the scale of nanostruc-tures: the typical bacterium is approxi-mately 1,000 nanometers long, and

mammalian cells are larger. Cells are,however, filled with much smaller struc-tures, many of which are astonishinglysophisticated. The ribosome, for exam-ple, carries out one of the most importantcellular functions: the synthesis of pro-teins from amino acids, using messengerRNA as the template. The complexity ofthis molecular construction project farsurpasses that of man-made techniques.Or consider the rotary motors of the bac-terial flagella, which efficiently propel theone-celled organisms [see “The Once andFuture Nanomachine,” on page 78].

It is unclear if “nanomachines” tak-en from cells will be useful. They willprobably have very limited application inelectronics, but they may provide valu-able tools for chemical synthesis andsensing devices. Recent work by Carlo D.Montemagno of Cornell University hasshown that it is possible to engineer aprimitive nanomachine with a biologicalengine. Montemagno extracted a rotarymotor protein from a bacterial cell and

connected it to a metallic nanorod—acylinder 750 nanometers long and 150nanometers wide that had been fabricat-ed by lithography. The rotary motor,which was only 11 nanometers tall, waspowered by adenosine triphosphate(ATP), the source of chemical energy incells. Montemagno showed that the mo-tor could rotate the nanorod at eight rev-olutions per minute. At the very least,such research stimulates efforts to fabri-cate functional nanostructures by demon-strating that such structures can exist.

The development of nanotechnologywill depend on the availability of nano-structures. The invention of the STMand AFM has provided new tools forviewing, characterizing and manipulat-ing these structures; the issue now is howto build them to order and how to de-sign them to have new and useful func-tions. The importance of electronics ap-plications has tended to focus attentionon nanodevices that might be incorpo-rated into future integrated circuits. Andfor good technological reasons, the elec-tronics industry has emphasized fabri-cation methods that are extensions ofthose currently used to make micro-chips. But the explosion of interest innanoscience has created a demand for abroad range of fabrication methods,with an emphasis on low-cost, conve-nient techniques.

The new approaches to nanofabrica-tion are unconventional only becausethey are not derived from the microtech-nology developed for electronic devices.

Chemists, physicists and biologists arerapidly accepting these techniques as themost appropriate ways to build variouskinds of nanostructures for research.And the methods may even supplementthe conventional approaches—photo-lithography, electron-beam lithographyand related techniques—for applicationsin electronics as well. The microelec-tronics mold is now broken. Ideas fornanofabrication are coming from manydirections in a wonderful free-for-all ofdiscovery.


More information about nanofabrication can be found at the following Web sites:

International SEMATECH: www.sematech.org/public/index.htm

The Whitesides group at Harvard University: gmwgroup.harvard.edu

The Mirkin group at Northwestern University: www.chem.northwestern.edu/~mkngrp/

The Willson group at the University of Texas at Austin:willson.cm.utexas.edu/Research/research.htm

The Alivisatos group at the University of California at Berkeley: www.cchem.berkeley.edu/~pagrp/

The Bawendi group at M.I.T.: web.mit.edu/chemistry/nanocluster/

The Montemagno group at Cornell University: falcon.aben.cornell.edu/

M O R E T O E X P L O R E

Bottom-up methods start from atomsor molecules and build up to nanostructures.


Date post:	17-Mar-2021
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Copyright 2001 Scientific American, Inc.40 SCIENTIFIC AMERICAN SEPTEMBER 2001 IMAGE BY FELICE...

Documents