Magazine

September 2012 Vol. 37 No. 9

www.mrs.org/bulletin

Inside: Energy Quarterly

Semiconductor nanowire building blocks

ZnO: Nanogenerators to piezotronics

Fighting cancer with nanoparticle medicines

Smart materials for cell-biomaterial interactions

Self-folding thin-film materials

Extremes of heat conduction

IN THIS ISSUE

High Voltage Engineering

Beam energies from 10 keV up to several 10s of MeV

Beam currents from 100 micro-amps up to several milliamps

Ion species, including H, He, B, P, As and others

Single wafer or batch processing of wafers up to and including 8”

In-air or in-vacuum cassette-to-cassette wafer handling

Electrostatic and/or mechanical wafer clamping

CUSTOMIZED PRODUCTION ION IMPLANTERS

High Voltage Engineering Europa B.V.

P.O. Box 99, 3800 AB Amersfoort, The Netherlands

Tel: 31 33 4619741 [email protected]

www.highvolteng.com

786

OPINION LETTER FROM THE PRESIDENT

MRS BULLETIN VOLUME 37 SEPTEMBER 2012 www.mrs.org/bulletin

September 2012 Volume 37 Number 9 ISSN: 0883-7694 CODEN: MRSBEA

CONTENTSTECHNICAL FEATURES

ON THE COVER

(Counterclockwise from top) 1. Litho-

graphically patterned bi-directionally self-

folding metallic sheets with Cr-Cu bilayer

hinges (Shenoy and Gracias, p. 847).

2. Endothelial cell elongated parallel

to a vertical fibronectin pattern and

membrane extensions along a horizontal

peptide pattern forming a lamella (Kustra

and Bettinger, p. 836). 3. Novel ZnO

nanostructure formed due to the existence

of ±(0001)f polar surfaces (Wang, p. 814).

4. Nanoparticles in a mousea liver scavengedr

by special Kupffer cells marked K (Davis,K

p. 828). 5. Illustration of a time-domain

thermoreflectance (TDTR) measurement

for a transfer-printed structure. Image

courtesy of Jeremyf Miller andr Ryan Durdle,

Imaging Technology Group,y UIUC (Cahill, p. 855). 6. Hierarchical

InGaN nanowire arrays grown on a Si (111) substrate yielding

high surface area photoanodes (Yang, p. 806).

1

2

34

56

806 Semiconductor nanowire building blocks:

From flux line pinning

to artificial photosynthesis

2011 MRS Medal

Peidong Yang

814 From nanogenerators to piezotronics—

A decade-long study of ZnO nanostructures

2011 MRS Medal

Zhong Lin Wang

828 Fighting cancer with nanoparticle medicines—

The nanoscale matters

2011 Fred Kavli Distinguished Lecture

Mark E. Davis

836 Smart polymers and interfaces for dynamic

cell-biomaterials interactions

Stephen Kustra and Christopher J. Bettinger

847 Self-folding thin-film materials:

From nanopolyhedra to graphene origami

Vivek B. Shenoy and David H. Gracias

855 Extremes of heat conduction—Pushing

the boundaries of the thermal conductivity

of materials

2011 MRS Fall Meeting Symposium X

David G. Cahill

Energy Quarterly

797 Editorial

Billboard science

Steve Yalisove

798 Energy Sector Analysis

High-temperature superconductors

change the game

Arthur L. Robinson

FEATURE EDITOR: James Misewich

800 Interview

From materials research to climate change:

David Eaglesham assesses the solar energy

industry

Interviewed by Steve M. Yalisove

and Arthur L. Robinson

802 Regional Initiative

Supercapacitors take charge in Germany

Philip Ball

FEATURE EDITOR: Yury Gogotsi

804 Energy Focus

Tim Palucka

www.mrs.org/energy-quarterly

Blog: www.materialsforenergy.org

787

OPINION LETTER FROM THE PRESIDENT

789 Research/Researchers

Lifetime variation in giant nonblinking QDs

due to switching between neutral and negatively

charged states

Anthony S. Stender

Wireless PV retinal prosthesis shows promise for

restoration of sight

Steven Trohalaki

Optical confinement modifies graphene

transistor characteristics

Tobias Lockwood

Weakly charged cationic nanoparticles unzip DNA

Tim Palucka

792 Science Policy

India to reopen mining for rare-earth elements

Angela Saini

NSF and EC establish collaboration opportunities for

early career scientists

Brazil and China discuss 10-year cooperation plan

Australia’s synchrotron receives renewed funding

NEWS & ANALYSIS

www.mrs.org/bulletin

www.mrs.org/bulletinMRS members—access MRS BulletinS onlinen

www.mrs.org/energy-quarterlyAccess Energy Quarterly online

www.materialsforenergy.orgJoin the conversation in the Materials for Energy blog

www.mrs.org/mymrsMRS Publications Alert—

receive advance Table of Contents by email

http://journals.cambridge.org/mrsbulletin-rssSubscribe TODAY to the MRS BulletinS RSSn Feed

S BULLETIN VOLUME 37 SEPTEMBER 2012 www.mrs.org/bulletin

DEPARTMENTS

ADVERTISERS IN THIS ISSUE

Advanced Research Systems ............................. 835

American Elements .....................Outside back cover

Bruker ............................................ Inside back cover

High Voltage Engineering ............. Inside front cover

JEOL USA, Inc. .................................................... 796

Kurt J. Lesker Company ...................................... 805

MMR Technologies, Inc. ..................................... 827

National Electrostatics Corp. .............................. 863

Park Systems, Inc. .............................................. 785

ULVAC Technologies ............................................813

Page No.

FEATURES

794 Beyond the Lab

Alta Devices moves out of the lab and into the valley

Jessica M. Smith

865 Books

Composite reinforcements for optimum performance

Philippe Boisse

Reviewed by Erik Thostenson

871 Posterminaries

Good reads for the materials researcher

Steve Moss

866 CAREER CENTRAL

864 SOCIETY NEWSYMRS seeks award nominations for 2013

VOLUME ORGANIZERS

2013 Mark T. Lusk, Colorado School of Mines,f USA Eva Olsson, Chalmers University of Technology,f Sweden Birgit Schwenzer, Pacificfi Northwest National Laboratory, USA James W. Stasiak, Hewlett–Packard Co., USA

2012 Lei Jiang, Chinese Academy of Sciences,f ChinaSergei V. Kalinin, Oak Ridge National Laboratory, USAStéphanie P. Lacour, EPFL, SwitzerlandSteven C. Moss, Aerospace Corporation, USA

2011 Kyoung-Shin Choi, Purdue University, USAReuben T. Collins, Colorado School of Mines,f USASean E. Shaheen, University of Denver,f USA

MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin788

EDITORIAL OFFICE 506 Keystone Drive, Warrendale, PA 15086-7573 USA

[email protected] tel 724.779.2747 fax 724.779.8313 www.mrs.org

MRS BulletinS (ISSN:n 0883-7694, print; ISSN 1938-1425, online) is published 12 times a year by the Materials Research Society, 506 Keystone Drive, Warrendale, PA 15086-7573.

Copyright © 2012, Materials Research Society. Permission required to reproduce content. Periodical postage paid at New York, NY, and at additional mailing offices.fi POSTMASTER: Send

address changes to MRS BulletinS inn care of the Journals Department, Cambridge University Press, 100 Brook Hill Drive, West Nyack, NY 10994-2113, USA. Printed in the U.S.A.

Membership in MRS is $115 annually for regular members, $30 for students. Dues include an allocation of $29 ($17 for students) to a subscription to MRS BulletinS . Indi-

vidual member subscriptions are for personal use only. Non-member subscription rates are $363 for one calendar year (12 issues) within North America and $436 elsewhere.

Requests from subscribers for missing journal issues will be honored without charge only if received within six months of the issue’s actual date of publication.

MRS Bulletin isn included in Current Contents®/Engineering, Computing, and Technology; Current Contents®/Physical, Chemical, and Earth Sciences, the SciSearch® online data-

base, Research Alert®, Science Citation Index®, and the Materials Science Citation Index™. Back volumes of MRS BulletinS aren available on microfichefi through University Microfilmsfi Inc.,

300 North Zeeb Road, Ann Arbor, MI 48106, USA.

Editor Gopal R. Rao, [email protected]

Managing Editor Judy Meiksin, [email protected]

Technical EditorLori A. Wilson, [email protected]

Editorial Assistants Ben Moriarty, [email protected] Wilmoth

Associate Technical EditorAniketa Shinde

Art Director Kasia M. Bruniany

Production/Design Andrea Pekelnicky, Christopher R. Roberts,and TNQ

Production EditorCatherine Paduani

Science News EditorTim Palucka

Principal Development EditorElizabeth L. Fleischer

Director of Communications Eileen Kiley Novak

Send Lettersto the Editor [email protected] your name,affi liation,fi and fullcontact information.

2012 MRS BOARD OF DIRECTORS

President Bruce M. Clemens, Stanford University, USAImmediate Past President James J. De Yoreo, Lawrence Berkeley Nationaly

Laboratory, USAVice President and President-Elect Orlando Auciello, Argonne National

Laboratory, USASecretary Sean J. Hearne, Sandia National Laboratories, USATreasurer Michael R. Fitzsimmons, Los Alamos National Laboratory, USAExecutive Director Todd M. Osman, Materials Research Society, USA

Wade Adams, Rice University, USAAna Claudia Arias, University of California–Berkeley,f USAShenda Baker, Synedgen, Inc./Harvey Mudd College, USATia Bensona Tolle, U.S. Air Force Research Laboratory, USADuane B. Dimos, Sandia National Laboratories, USAChang-Beom Eom, University of Wisconsin-Madison,f USAEric Garfunkel, Rutgers University, USAJ. Murray Gibson, Argonne National Laboratory, USAOliver Kraft,r Karlsruhe Institute of Technology,f GermanyHideki Matsumura, Japan Advanced Institute ofe Sciencef ande Technology, JapanStephen K. Streiffer, Argonne National Laboratory, USAJames C. Sturm, Princeton University, USASusan E. Trolier-McKinstry, The Pennsylvania State University, USAPierre Wiltzius, University of California–Santaf Barbara, USA

MRS OPERATING COMMITTEE CHAIRS

Academic Affairs M. Stanley Whittingham, SUNY–Binghamton, USAAwards C. Barry Carter, University of Connecticut,f USAGovernment Affairs Nabil Bassim, U.S. Naval Research Laboratory, USAMeetings Committee David S. Ginley, National Renewable Energy Laboratory, USA Membership YvesYY Chabal, The University of Texasf at Dallas, USAPublic Outreach Aditi Risbud, University of Utahf in Salt Lake City, USAPublications Paul McIntyre, Stanford University, USA

About the Materials Research Society

EDITORIAL BOARD

Paul S. Drzaic (Chair), Apple, Inc., USA

V.S. Arunachalam, Center for Study of Science,f Technology & Policy, India

Marie-Isabelle Baraton, University of Limoges,f France

Robert C. Cammarata, Johns Hopkins University, USA

Laura Fornaro, University of Uruguay,f Uruguay

Hanns-Ulrich Habermeier, Max Planck Institute for Solid State Research, GermanyFiona C. Meldrum, University of Leeds,f UK

Amit Misra, Los Alamos National Laboratory, USA

Julie A. Nucci, Cornell University, USA

Linda J. Olafsen, Baylor University, USA

David N. Seidman, Northwestern University, USA

Carol Trager-Cowan, University of Strathclyde,f UK

Julia R. Weertman, Northwestern University, USA

Eric Werwa, Washington, DC, USA

Steve M. Yalisove, University of Michigan,f USA

MRS OFFICE OF PUBLIC AFFAIRS

Ron Kelley 499 South Capitol St. SW, Suite 600, Washington, DC 20003

The Materials Research Society (MRS), a not-for-profitscientificfi association founded ind 1973, promotes interdiscipli-nary goal-oriented basicd research on materials of technologi-fcal importance. Membership in the Society includes almost16,000 scientists, engineers, and research managers fromindustrial, government, and universityd researchy laboratoriesh in thenUnited Statesd and overd 80r countries.

The Society’s interdisciplinary approach differs from thatof single-disciplinef professional societies because it promotestinformation exchangen across the many technicaly fi eldsfi touchingmaterials development. MRS sponsors three major interna-tional annual meetings encompassing approximately 125y topicalsymposia, and alsod sponsors numerous single-topic scientificfimeetings. The Society recognizes professional and technicaldexcellence and fostersd technical interaction inn localn geographicregions through Sections and Universityd Chapters.

MRS participates in the international arena of materialsfresearch throughh theh International Union ofn Materialsf ResearchSocieties (IUMRS). MRS is a member of thef Alliance forScience & Technology Research in America and isd an affiff liatefiof thef American Institute of Physics.f

MRS publishes The MRS Online Proceedings Library, MRSBulletin, Journal ofl Materialsff Research, MRS CommunicationsS ,and otherd publicationsr related tod current researcht activities.

Special ConsultantsMridula Dixit Bharadwaj

Energy Quarterly Steve M. Yalisove (Chair),V.S. Arunachalam, Anshu Bharadwaj,David Cahen, Russell R. Chianelli,George Crabtree, Abdelilah Slaoui,Guillermo Solórzano,and M. Stanley Whittingham

Advertising/Sponsorship Mary E. Kaufold, [email protected] L. Watterson, [email protected]

Member Subscriptions Michelle Judt, [email protected]

Non-Member [email protected]

789MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin

NEWS & ANALYSIS RESEARCH/RESEARCHERS

Lifetime variation in giant nonblinking QDs due to switching between neutral and negatively charged states

Nano Focus

Quantum dots (QDs) are attractiveas nanoscale light sources,t but thet

fluctuationsfl in emission intensity fromindividual dots, which is known as“blinking,” can limit their application.rIt has recently been shown that CdS/CdSe core–shell nanoparticles with thickhshells do not exhibitt blinking.t However,as reported byd ay Losa Alamos team, Chris-tophe Galland (nowd at thet University ofDelaware), Yagnaseni Ghosh, AndreaSteinbrück, Jennifer A.r Hollingsworth,Han Htoon, and Victord I.r Klimov in the

June 19 issue of Nature Communica-tions (DOI: 10.1038/ncomms1916), gi-ant nonblinkingt QDs (g-QDs) exhibit atpronounced variationd in their emissionrlifetimes due to a switching betweennegatively charged andd neutrald states.

In theirn study,r the researchers investi-gated thed optical behavior ofr individualfCdSe/CdS g-QDs with a 15 monolayerthick shell.k Individual g-QDs producedstable photoluminescence intensitieswith variations fitting a single-peakPoisson distribution.n Such behavior sug-rgests emission occurred from a singlestate. However, lifetime measurementsrevealed the presence of twof equallyweighted lifetimes (19 ns and 39 ns),thereby demonstrating that twot distinctstates contribute to the emission.

This behavior wasr linked tod the pro-cess of nonradiativef Auger recombina-rtion and itsd different effectt ont negativelycharged excitonsd (negative trions) versuspositively charged excitons (positivetrions). Neutral QDs correspond tod thebright optical state. In thin-shell QDs,charged excitons are essentially non-emissive because Auger decay is fast.However, in g-QDs,n Auger processesr arelargely suppressed for negative trions.By adding a single electron to the QD,the number ofr radiativef recombinationpathways is doubled, as is the radiativedecay rate. Charging is thus accompa-nied by a decrease in the photolumi-nescence lifetime without affectingt theintensity (“lifetime blinking”).

The investigators confirmedfi thisd sce-nario by studyingy the effect oft controlledfelectrochemical charge injection on theQD photoluminescence and relaxationrates. At 0t V, the equal distribution be-tween the two lifetimes was again ob-served. When either 0.5 V orV 0.8 Vwas applied, electron injectionn wasn morefavorable, and thed shorter lifetimer (19 ns)predominated without any change inphotoluminescence intensity. The re-search team concluded that observedlifetime fluctuationsfl are connected torandom charging of thef g-QDs withexcess electrons. Moreover, they foundthat chargingt occurs by Auger ionizationrthrough ejection of af hole. This Augerdecay pathway is favored ind g-QDs be-cause of af greatera degreer of conf fi nementfifor holesr than forn electrons.r These resultsshow that nonblinking behavior is notincompatible with random charge fl uc-fltuations in the QD.

Anthony S. Stender

In thin-shell QDs (a) charging is one of thef mechanisms causing luminescence intensityfl uctuationsfl known as blinking. This is because light emission from the negativelycharged exciton (X–X )– is quenched by very fast Auger decayr (negative trion pathway,curved arrows). In contrast, in g-QDs (b), nonblinking emission intensity is observedbecause Auger recombination of thef negative trion is suppressed. The only signatureof chargingf is a doubling of thef radiative decay rate compared to the neutral exciton((( r 2 r; double orange arrow). This is indicated by the plot of time-resolvedfphoton-counting data on the lifetime-intensity distribution map (lower right). Chargefl uctuationsfl are caused by Auger ionization of thef g-QD, that is, by the decay of afbiexciton through the fast positive trion pathway with ejection of thef hole. This pathwayis favored in g-QDs because of af pronounced asymmetry in spatial distributions ofelectron and hole wave functions in these nanostructures.

Wireless PV retinal prosthesis shows promise for restoration of sight

Bio Focus

Retinal degenerative diseases lead todblindnessRR due to the loss of photo-f

receptors even though the inner retinalrneurons remain largely intact. Visualpercepts, also called “phosphenes,”d can

be produced byd electrical activation ofthe inner retinalr neurons. This alternateroute to visual information has the po-tential for restoring sight to the blind.Current retinalt prosthesis designs, withelectrode arrays implanted ind the retinafacing either the ganglion cells or theinner nuclearr layer,r rely on serial telem-etry to deliver stimulationr signals to theelectrodes, requiring bulky receivingy and

processing electronics and ad trans-scleralacable. Surgery isy complex and thed designis diffiff cultfi to scale up to attain highervisual acuity. In addition,n patients cannotuse natural eye movements to scan thevisual scene because retinal stimulationpatterns are transmitted fromd an externalncamera to the retinal implant, indepen-dent oft eyef orientation. These limitationscan be overcome by devices that use

790 MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin


Subretinal photodiode array with triple-diode pixels arranged in a hexagonal pattern.Pixels of 70f μm and 140 μm in size were made. Left inset: Central electrodes are sur-rounded by three diodes connected in series, and by the common return electrode.Right inset: The subretinal implant.

photosensitive pixels but they dependon an external power source.r Recently,however, researchers from the Palankergroup at thet Hansen Experimental Phys-ics Laboratory and the Department ofOphthalmology at Stanford Universitydesigned ad photovoltaic retinal prosthe-sis where video goggles were used tod de-liver bothr power andr visuald informationthrough pulsed NIR illumination,R pre-serving the natural link betweenk imageperception and eye movement withoutcomplex electronics and wiring.d

In an article published ind the June is-sue of Nature Photonics (DOI: 10.1038/nphoton.2012.104; p. 391), Keith Ma-thieson, James Loudin, and co-research-d

ers from Stanford University and theUniversity of California–Santaf Cruz,describe their prosthesisr design in whichnvideo images captured byd ay head-mount-aed camerad are processed byd a portablecomputer. The video goggles use a liq-auid-crystal display (LCD)y illuminated bydpulsed near-infraredd lightd (880–915t nm)to project thet images onto a subretinalphotodiode array (consisting of 70f mpixels, each with ~20 m stimulatingelectrodes), which converts the light totlocal currents that stimulatet the nearbyneurons in the inner nuclear layer ofthe retina.

The researchers fabricated silicond pho-ntodiode arrays consisting of pixelsf with

Optical confi nement modifi es graphene transistor characteristics

Nano Focus

The interaction between light andmatter can be greatly enhanced

within an optical cavity in which thespacing of twof mirrors definesfi a stand-ing electromagnetic wave. Placing asheet oft graphenef in such a cavity cantherefore have profound effects on itsoptoelectronic properties, as shown by

M. Engel of thef Karlsruhe Institute ofTechnology, M. Steiner ofr thef IBM T.J.Watson Research Center inr New York,A. Lombardo of thef University of Cam-fbridge, and theird colleagues.r Their articlerin the June 19 issue of Nature Commu-nications (DOI: 10.1038/ncomms1911)describes how such optical confinementfiof af graphenea transistor allowsr spectrallyselective generation of photocurrentf andteven alters the electrical transport prop-terties of thef material.

The team embedded ad sheet oft gra-f

phene between two optically transpar-ent dielectrict materials, Si3N4 and Ald 2O3,which are in turn enclosed byd silver mir-rrors with a spacing equal to one-half offthe resonant wavelength of thef cavity.At the center ofr thisf optical cavity theanti-node in the optical fi eldfi enhancesthe absorption orn emissionr ofn photonsf bythe graphene at thet resonant wavelength,tand inhibitsd it att othert wavelengths.r Ap-plying a voltagea across the graphene andilluminating the device with a laser gen-rerated 20d times more photocurrent att thet

single diodes as well as those consistingof pixelsf with three diodes connectedin series. These triple-diode pixels canproduce 1.5 V, which triplesh the chargeinjection onn then sputtered iridiumd oxidemfilmfi electrodesm (from 0.5m mC cm 2 for arsingle-diode pixel to 1.5 mC cm 2). Thetriple-diode pixels require light intensitiestthree times higher thanr single-dioden pix-els because the photosensitive pixel areais divided intod three subunits. However,the researchers found thatd theirt single-rand triple-dioded devices had veryd similarythresholds for elicitingr retinal responses.

The researchers tested their designconcept byt stimulating healthy and de-dgenerate rat retinas in vitro with NIRlight intensities at least two orders ofmagnitude below the ocular safetyr limit.They showed thatd thet elicited retinald re-sponses can be modulated byd both lightintensity and pulsed width, although theircurrent optical design allows only forintensity modulation within each videoframe. However, if thef retinal responseis modulated byd varying the pulse width,the researchers said thatd digitalt light pro-tcessing technology can also be used,adding, “Such ah devicea would allowd bothwthe duration and timingd of exposuref tobe precisely controlled on the scale ofindividual pixels. In addition to higherthroughput compared to an LCD, thishigh-speed controld would allowd the se-quential activation of nearbyf pixels tofurther reducer pixel crosstalk—interfer-ence of currentsf from nearby pixels.”

Steven Trohalaki



resonant wavelength,t providing a spec-tral selectivity not observed in uncon-finedfi graphene. Similarly, when a cur-rent ist applied tod unconfi nedfi graphene,d it

heats up and emitsd a featureless thermalspectrum, whereas the graphene in theoptical cavity displays a stronga emissionpeak atk thet resonant wavelength.t

The researchers also found thatd con-tfi nedfi graphened exhibited unusuald electri-cal behavior. At lowt voltages the cavityinhibits the emission of thef low energythermal radiation with wavelengthslonger thanr resonance, and the currenttherefore saturates. As the voltage isincreased, this threshold ford lightr emis-tsion is passed andd thed device resistancedrops accordingly.

Graphene is an ideal material forthis type of devicef because of itsf two-dimensional nature, which allows it toextend ford micrometersr across the centerof thef cavity, and easilyd tunable electri-cal properties. The degree of spectralfselectivity provided byd the optical con-fi nementfi suggests a useful applicationin photodetection, while its influencefl onelectrical transport couldt bed exploited indnanoelectronic devices.

Tobias Lockwood

Weakly charged cationic nanoparticles unzip DNA

Bio Focus

Understanding the interaction be-tween charged nanoparticles and

double-stranded DNA hasA importantimplications for drugr delivery schemesand DNA-templated metallization,in addition to other possible applica-tions. While attempting to packageDNA ontoA nanoparticles as a meansof genef delivery into cells, Anatoli V.Melechko of Northf Carolina State Uni-versity (NCSU), Timothy E. McKnightof Oakf Ridgek National Laboratory, andtheir colleagues discovered somethingcompletely unexpected. Some of thefnanoparticles clumped together,d and indthe process pulled thed double-strandedDNA apart,A at leastt partially.t

“This could bed a newa type of machin-fery that cant be used ford separatingr DNAinto single strands,” Melechko said, ad-mitting that at lot moret work willk needto be done to understand and controlthe phenomenon.

In nature, negatively charged DNAwraps around protein cylinders with apositive charge of +220f to form a com-

plex known as chromatin. A highA posi-tive charge on a protein or a nanopar-ticle causes DNA toA bend andd undergodcompaction. Much work hask been donewith functionalized nanoparticlesd in thisregime. Some research has been donewith weakly positive charged nanopar-dticles, which typically have no effect ontthe conformation of DNA.f In choosinggold nanoparticles (AuNPs) function-alized with thiolated alkane ligandsbearing primary amines with a chargeof +6,f Melechko and his colleaguesexplored the lesser-known transitioncharge region between the weak andk thedstrong regimes.

As reported in the August 16 issueof Advanced Materialsd (DOI: 10.1002/adma.201104891; p. 4261), gel electro-phoresis experiments with the AuNP-DNA showedA ad “mysterya bandy ind an gel—aan extra line [close to the 1000 base pairmarker] that createdt thed main puzzle forus,” Melechko said. Though prior workrby others might signifyt that thist mysteryline was the result oft DNAf compactionAaffecting gel mobility, Melechko and co-dworkers hypothesized thatd somet denatur-ing—or unzipping—ofr thef two-strandedDNA wasA occurring. UV spectroscopyV

showed that at least partial denaturingwas occurring.

To better understand the phenom-enon, co-researcher Yaroslavar Yinglingof NCSUf ran molecular dynamicsr simu-lations involving single AuNPs with sixcharged ligands (AuNP 6NH3

+), andcompared them with other moleculardynamics simulations involving threeof thef AuNP 6NH3

+ units. The single-particle simulations showed bindingd ofthe ligands to both the minor andr majordgrooves of DNA,f but littlet structural al-teration inn then double helix. Basically, theDNA justA stickst to a single nanoparticle.But int the three-particle simulations, thenanoparticles bunch together. “You haveuhydrophobic groups that want to hidebetween themselves, and polar groupsthat grabt on to DNA,” Melechko said.“When they do this clumping and stilldhold ond to DNA, they can rip it apart.”t

The researchers said that “particlesacting in concert cant produce effects notpossible with single particles.” A videoAof thef molecular dynamics simulationcan be viewed at http://youtu.be/9M-58niEOpU.

Tim Palucka

A schematicA and cross section (inset) of thef graphene transistor in an optical cav-ity. The separation L of thef silver mirrors determines the resonant wavelength ofthe cavity, for which the intensity profilefi of thef /2 mode is shown in orange; VgVV ,gate voltage; VdVV , drain voltage. Reproduced with permission from Nat. Commun. 3(2012), DOI: 10.1038/ncomms1911. ©2012 Macmillan Publishers Ltd.

Al2O3

Si3N4

Ag

Ag graphene

V I

L

Vg


SCIENCE POLICYNEWS & ANALYSIS

The battle over rarer earths seems tohave become the most bittert inter-r

national trade dispute this decade. Coun-tries have become increasingly con-cerned thatd Chinat isa gaining a monopolyaover ther production of thesef elements,which are critical to green energy andhigh-tech industries. The United States,dfor example,r once produced alld its rare-earth elements domestically, but hast be-come wholly reliant ont Chinese importsover ther last 15t years. During disputesthrough the World Traded Organization,China said thatd its policies in questionare aimed atd protectingt natural resourcesand achievingd sustainable economic de-velopment. Meanwhile, one of thef fewsolutions for countriesr suffering a supplyashortage is to launch their ownr miningefforts. And thatd ist exactly what China’stneighbor, India, is now planning to do.

India’s Ministry of Minesf has formeda steering committee to investigate re-starting exploration of raref earths, with aparticular eyer on India’sn growing renew-able energy sector. Modern electronictechnologies depend ond rare earths, andwith government targetst for ar total ca-

pacity of 20f GW ofW on-gridf solard powerrby 2022, and 40d GW ofW windf power,d theGeological Survey ofy Indiaf hasa made ex-ploration ofn raref earths a higha priorityh foryits next five-yearfi plan.r

“With limited availability of rarefearths, and thed projected growthd in de-mand, supply chainy vulnerabilityn mayy setyin,” said Rangacharid Krishnan, chief ad-fvisor tor the Center forr Studyr of Science,fTechnology andy Policyd iny Bangalore,n andformer headr ofd thef Metallurgy Divisionat thet Bhabha Atomic Research Centre.

The elements particularly in demandinclude neodymium, which ish used ind thenpermanent magnetst inside the compactmotors of windf turbines, and dyspro-sium, which is used to raise the Curietemperature of thef magnets of electricfvehicles. “The typical weight oft af 1-MWawind generatord isr around 600d kilograms,out oft whichf neodymium is around 25%dand dysprosium is about two to threepercent,” said Krishnan.d Cerium oxideis used ford UVr absorptionV inn solarn panels,rand lanthanumd is required ford catalyticrcracking in then petroleum industry. Indiaalso has a large and increasingd demand

for fluorescentfl bulbstand tubed lights, whichuse terbium, yttrium,and europiumd inm theirnfluorescentfl coatings.t

If Indiaf doesbegin mining rareearths, it will notbe for ther firstfi time.“India was a leadingproducer and sup-plier of raref earths50 years ago,” saidKrishnan, particu-larly of yttrium.fAbout five yearsago, however, thecountry froze min-ing and developmentd

of rare-earthf elements because of af lackof competitionf in the domestic market,which made Chinese imports cheaper.

Now, as before, the Atomic Miner-als Directorate, a unit oft India’sf Depart-ment oft Atomicf Energy, is at thet heartof explorationf since a major sourceof raref earths is monazite sands, fromwhich radioactive thorium and uraniumdare also extracted. “India possesses thelargest depositst of monazitef in then world,mostly in the coastal tracts of Orissaf onthe east coastt andt ind Kerala on the westcoast. Besides the beach sands, mona-zite has been reported ind carbonatites inMeghalaya, Tamil Nadu, and Assam,”said Krishnan.d

According to the Prime Minister’s of-fice,fi these reserves stand atd aboutt 10.7tmillion tons, translating to roughly 5million tons of rare-earthf oxide.

Indian Rare Earths Ltd. (IREL), thecountry’s only rare-earths producer, issetting up a processing plant int the east-ern staten of Orissaf toa produce 11,000 tonsof rare-earthsf chloride, which inh turnn cannbe converted tod rare-earth oxides. Alongwith a smaller plantr alreadyt in operationin Kerala, they are expected tod producearound 2250d tons of rare-earthf oxides inthe last quartert ofr 2012.f

The Japanese firmfi Toyota Tsushoa hasentered intod an agreement witht IREL toLset upt a plant att Visakhapatnamt in East-ern India to produce rare-earth oxides,and the company expects an export ofabout 4000t tons of thef material from this

India to reopen miningfor rare-earth elements

Mineral-rich “black sands” in Kollam, a coastal district in the Indianstate of Kerala.f These sands are a valuable source of monazite.f

Yvo

nM

auric

e

BastnäsiteExtract: Cerium

MonaziteExtract: Thorium and uranium

XenotimeExtract: Dysprosium

Kerala

Tamil Nadu

Orissa

Chhattisgarh

Jharkhand

MeghalayaAssam

gggggg

West Bengal


NEWS & ANALYSIS SCIENCE POLICY

plant. In July 2011, it began construc-tion of a plant that “makes use of thispreviously unexploited mixture as a raw material to produce such rare earths as neodymium, lanthanum, and cerium,”according to the company’s website.

Sizable deposits of xenotime, a phos-phate mineral that contains the heavier rare earth dysprosium, have also beenfound in the states of Chhattisgarh and Jharkhand. Meanwhile, bastnäsite—a source of cerium—has been found in the state of West Bengal. As yet, though, the extent of India’s reserves is not fullyknown. Of the 6500 kilometers of In-dian coastline, the Geological Survey of India has only explored 2200 kilometers,according to Krishnan. He also warns

that the process of starting such miningoperations is likely to be slow: “Fromexploration until setting up an extractionplant may take more than 10 years.”

Despite its latest efforts, however,India is unlikely to challenge China in its dominance over the global supply of rareearths, according to Naresh Pant, associ-ate geology professor at the University of Delhi. “That would require at least anorder of magnitude increase in produc-tion,” he said. India’s rare-earth reserves stand at just more than three milliontons, he said, while China has more than 36 million.

At the moment, the lack of rare-earthdeposits in the European Union means that it imports USD$458 million of rare

earths annually from China. Accordingto the US Department of the Interior,the United States has around 13% of global reserves of rare earths, Russiahas 17%, and Australia has 1.5%, yet all these nations depend on imports, too. Complaints to the World TradeOrganization have focused on the fact that China—which has around 37% of reserves and supplies around 97% of theworld’s rare-earth elements—is threaten-ing businesses by restricting exports. For India, which also relies on Chi-nese rare earths, restarting rare-earths processing, mining, and exploration may at least offer a ray of hope as sup-ply shortages begin to bite.

Angela Saini

NSF and EC establish collaboration opportunities for early career scientists

The US National Science Foundation (NSF) and the European Commis-

sion (EC) signed an Implementing Ar-rangement to provide opportunities for NSF-funded early career scientists and engineers to pursue research collabo-

rations with European colleagues sup-ported through the European Research Council (ERC) awards. The agreement supports collaborations on specifi c proj-fiects while leveraging research fundingand fostering lasting collaborations be-

tween European and US researchers. European Commissioner for Re-search, Innovation and Science Máire Geoghegan-Quinn and NSF Director Subra Suresh signed the arrangement onJuly 13 at the European Science OpenForum in Dublin.

Australia’s synchrotron receives renewed fundingwww.synchrotron.org.au

Monash University, with the support of the Australian government’s

investment of AUD$30 million under Prime Minister Julia Gillard, will man-age the Australian Synchrotron program. Australian universities will also investaround AUD$25 million. Announcingthe funding, Science and Research Min-

ister Senator Chris Evans said with the strong need to undertake research and development to transform industriesand see them through challenges such as climate change, economic change, and skills shortages, there has never been a more vital time to invest in the facility. “As the Australian Synchrotron can

be used to study the most precise natureof any biological and industrial material,it can be used by almost any industry across a wide range of research fields,” fisaid Evans.

The AUD$30 million government in-vestment is being provided by the Aus-tralian Research Council (AUD$25 mil-lion) and National Health and Medical Research Council (AUD$5 million).

Brazil and China discuss 10-year cooperation plan

During the Rio+20 Summit held inJune, Brazil and China signed the

Ten-Year Cooperation Plan 2012–2021. In July, Brazil’s Minister of Science, Technology and Innovation (MCTI),

Marco Antonio Raupp, traveled to China to meet with the Chinese Ministers of Science and Technology, Wan Gang, and Industry and Information Technology, Miao Wei, and the head of the Chinese

National Space Administration (CNSA), Chen Qiufa, to discuss numerous top-ics, including the Bi-National Center for Nanotechnology and memorandums of understanding for biotechnology and meteorology centers.

MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin

BEYOND THE LAB

794

FEATURESFEATURES

Jessica M. Smith

Alta Devices movesout of the lab andinto the valley

In the world ofd technologyf start-ups,especially clean energy, “the val-

ley”—also known as “the valley ofdeath”—refers to the gap in capital be-tween then funding of invention—throughfgovernment grantst and ventured capital—and massd production. Recently, the thin-filmfi solar-cell manufacturer Solyndrafamously failed tod cross this valley, de-claring bankruptcy after acceptingr mil-lions of dollarsf in US government gaptfunding. The controversy around thisparticular case,r and around the fate ofmany solar start-ups generally, meansthat allt eyes are on Alta Devices. As thestart-up begins production on its pilotline, the clean energy community willwatch as Alta Devices attempts to takeits technology out of thef lab and intothe valley.

Alta Devices was founded byd HarryAtwater of thef California Institute ofTechnology and Eli Yablonovic of thefUniversity of California–Berkeleyf incollaboration with venture capitalistAndy Rappaport. The company relieson Yablonovic’s research in the 1980son a technique called epitaxial liftoff.The active layer Altar Devicesa uses in so-lar cells,r gallium arsenide (GaAs), canbe produced inexpensively using thistechnique without degradingt the perfor-mance of thef cell. The team began thecompany in 2007, and hasd subsequentlyworked towardd perfectingd epitaxial lift-off off GaAsf at at laboratory scale. Whenthe pilot productiont line opens later thisryear, this invention will be produced fordthe firstfi timet on large scale.

Epitaxial liftoff isf an effiff cientfi wayto create thin wafers of GaAs.f On topof af high-purity GaAs surface, layers ofGaAs are deposited withd alternating lay-ers of aluminumf arsenide (AlAs). “Wetake advantage of thef serendipitous fact

that the etch selectivity between AlAsand GaAsd is more than 100 million, sothat it’st possible to immerse the structurein an etching solution and remove theAlAs layer completelyr without etchingtthe GaAs layer,” said Atwater.d The freeGaAs wafers are of veryf high purity, andthe original GaAs surface can be usedagain to create more wafers.

Solar cells made from these GaAswafers compare favorably to conven-tional silicon photovoltaic (PV) cells.The highest performing silicon PVcells are made out of high-purityf sili-con, similar tor that found in computerchips, which allows them to reach solarconversion effiff cienciesfi near 30%.r Thesilicon used ind these cells, however, iscostly to produce, and thed cells requirea thick layer in order to reach max-imum effiff ciency.fi

Gallium arsenide PV cells madethrough epitaxial liftoff canf also reachhigh effiff ciency—currentlyfi 23.4% andclimbing. Because of thef properties ofGaAs, effiff cientfi solart cellsr can be madewith a much thinner layerr ofr material.fCombined withd the less expensive pro-cessing method, Alta Devices'a solar cellsrcan be made for ar fraction of thef cost oftsilicon PV cells.V

The conventional wisdom ofm thef pho-tovoltaics community is that any solarcell that ist inexpensive to produce mustalso have a low effiff ciency.fi For instance,rpolymer PV cells, or those made withamorphous cadmium telluride, have ef-ficienciesfi below 10%. As a result, anycost eft fiff ciencyfi is mitigated byd the smallamount of electricityf that will comeout oft onef of thesef solar cells.r Throughepitaxial liftoff, Alta Devices’ galliumarsenide solar cellsr can be both highlyeffiff cientfi andt inexpensive.d

Alta Devices is working toward fur-dther increasing the effiff ciencyfi of theirfsolar cells.r Atwater said,r “We see 30%[effiff ciency]fi as an achievable limit. Thatwas seen as a lunatic idea when we be-gan the company.”

Rappaport agreed, adding, “Fortu-nately, you found ad lunatic.” Rappaport

Alta Devices founder Harry Atwater (right) of thef California Institute of Technologyfwith venture capitalist Andy Rappaport of Augustf Capital describe the early develop-ment of thef company at the Technology Innovation Forum held at the 2012 MaterialsResearch Society Spring Meeting in San Francisco.


FEATURES BEYOND THE LAB

had never invested in a clean energycompany before and saidd hed is unlikelyto invest int one again. “A problemA withsolar start-upsr is that mostt oft thef inno-vation is around productiond methods atscale. It’s expensive to get to scale!”he said. He liked Altad Devices becauseit was the fi rstfi solar company he hadseen that was fundamentally different.“The cost-effiff ciencyfi curve had gottento be conventional wisdom.” That AltatDevices had ad way to deviate from thecost-effiff ciencyfi curve convinced Rap-paport and became the foundation ofthe company.

Alta Devices' leaders believe thatthey can findfi a market for their solarcells soon after theyr optimize the pilotline. They said theird thin,r flexible,fl high-effiff ciencyfi panels will be ideal for aero-rspace and militaryd applications in whichthe right combination of propertiesf ismore important thant cost fort ar PV cell.VThese customers will give the companytime to decrease manufacturing costs,according to Rappaport. “When we arefar downr the learning curve, we shouldbe as cheap, or cheaper,r than anythingelse out there.t Eventually, we should bedthe best solution to utility scale solar,but wet don’t havet to immediately trans-port ourselvest to the end ofd thef learningcurve. We get tot gradually get there.”t

At thet same time, another earlyr ap-plication for ar fl exible,fl high-effiff ciencyfisolar cellr could bed implemented ind de-veloping countries. “Worldwide energy

production and consumption has beenhampered ind a verya general way becauseof thef high infrastructure cost oft distri-fbution. If you’ref able to generate powerlocally, you reduce the dependence onexpensive, diffiff cult-to-constructfi distri-bution systems. We’re building some-thing lightweight, portable, and cheap,dso we’re contributing to a notion of lo-fcal power generation.”r Rappaport saidthat the same properties that allow theAlta Devices’ solar cellsr to be useful tomilitary operations could alsod be usefulin places that lackt energyk infrastructure.

To be used inother countries, par-ticularly in Europe,the solar cells mayneed to be part of afrecycling schemebecause of thef coun-tries’ strict regula-tions with respect totoxic materials. Forinstance, the Restric-tion of HazardousfSubstances Directive,a policy of thef Euro-pean Union, banscadmium in all con-sumer products. Toreceive permission to

sell panels in the European Union, FirstSolar, which makes cadmium telluridesolar cells,r had tod agree to remit allt of thefpanels back tok their factoryr in Arizonaat thet end ofd life.f “People regarded thatdas being a significantfi liability,”t Atwatersaid, referring to the high insuranceh coststo First Solar.t “They’re profitablefi evendoing this recycling. They’re headingoff thef criticism that theyt will leave theworld withd an environmental obligationif theyf go bankrupt.”

In Alta Devices, the toxic element oftconcern is arsenide. Atwater saidr that,increasingly, solar companiesr do recyclepart oft theirf solarr cells.r He said, “I imag-ine Alta will follow suit ast a responsiblecorporate citizen.”

Given then challenges facing solarg start-rups, considering a recyclinga policy beforeylaunching a pilota linet may seemy optimis-mtic. Luckily, the founders of Altaf Devicesaseem tom overflrr owfl withw optimismh thatm theythave discerned thed right combinationt ofntechnology andy businessd strategy. Despitethese assets, to land ond then other sider of“the valley ofy death”f as a solara companyrthat manufacturest a newa materialw will bea signia ficantfi accomplishment.t

Alta Devices’ GaAs high-efficiency,fi ultrathin solar panel.

David Parrillo of Thef Dow Chemical Company (second from left) moderatedthe Technology Innovation Forum with panelists (from left to right) Rappaport;Atwater; Joseph Foster, Alta Devices Vice President of Businessf Development;and Fred Farina, Caltech’s Chief Innovationf Officerfi and Executive Director of thef Officefiof Technologyf Transfer.

View our real-time DRAM analysis videoat www.jeolusa.com/Unrivaled EDS NA

NOAN

ALYS

ISS/

TEM

Unrivaled EDS for TEM

JEM-2800

JEM-ARM200F

0.98sr 100mm2• Exponentially enhances elemental

mapping for nano-area analysis TEM• Automatic retractable design

AnotherExtreme ImagingSolution

Global Solutions Provider for Advanced Technologywww.jeolusa.com • [email protected]

978-535-5900Find us on Facebook and Twitter @jeolusa

MRS BULLETIN VOLUME 37 SEPTEMBER 2012 www.mrs.org/bulletin Energy Quarterly

Energy QuarterlyNews and analysis on materials solutions to energy challenges

Inside:

EDITORIAL

Billboard science

ENERGY SECTOR ANALYSLL IS

High-temperature superconductors

change the game

INTERVIEW

From materials research

to climate change:

David Eaglesham assesses

the solar energy industry

Supplementary: Video selections

from Eaglesham interview online

REGIONAL INITIATIVE

Supercapacitors take charge in Germany

ENERGY FOCUS

ENERGY QUARTERLY ORGANIZERS

CHAIR Steve M. Yalisove,YY University of Michigan, USA

V.S. Arunachalam, Center for Study of Science,

Technology and Policy, India

Anshu Bharadwaj, Center for Study of Science,

Technology and Policy, India

David Cahen, Weizmann Institute, Israel

Russell R. Chianelli, University of Texas, El Paso, USA

George Crabtree, Argonne National Laboratory, USA

Abdelilah Slaoui, InESS, France

Guillermo Solórzano, PUC-Rio, Brazil

M. Stanley Whittingham, State University of New York

at Binghamton, USA

Images incorporated to create the energy puzzle

concept used under license from Shutterstock.com.

Energy Sector Analysis image of a variety of SuperPower 26 HTS

wire types courtesy of SuperPower, Inc.

www.mrs.org/energy-quarterly | www.materialsforenergy.org

ToTT suggest ideas for ENERGY QUARTERLY,

to get involved, or for information

on sponsorship, send email to

[email protected].

Billboard scienceWind Dies,d Sun Sets … You need reliable,d affordable, clean coal electricity.l This iswhat It recently read ond billboard afterd billboardr whiled driving through Pennsylvania.Coal may be cleaner thanr it was,t but itt stillt produces massive amounts of COf 2. WhileI was upset att thet message, I quickly began to admire the ability of thef coal industry

sears its message. This is no accident. Huge amounts of moneyf go into advertisingcampaigns. Claims are even made that solart panelsr emit radiation,t as found ond internetdiscussion forums such as Australia’s “Whirlpool” (http://forums.whirlpool.net.au)and governmentd policyt recommending people should remaind at leastt 4t meters awayto avoid exposured as reported byd the Israeli newspaper Haaretz. The materials com-

the resources to hire advertising agencies to do our workr fork us.r What wet can do is useour talentsr combined withd the internet tot get ourt messager out. While science museumexhibits and NOVAd showsA are great, a largea fraction of thef population never seesr them.Broadcast mediat (e.g., billboards, blogs, radio, and magazinesd like People) are themost commont source of informationf for ther general public. Luckily, these sources ofinformation are losing ground tod the internet. Hence, the materials community has agolden opportunity to reach the general public if wef are smart aboutt it.t Our communityrcan learn the skills of visualf communication. We can learn how to make science simple

major researchr university. The main purpose of thesef departments would bed to trainour scientistsr and engineersd in the elements of visualf communication and armd them

billboards on Pennsylvania highways.Steve Yalisove

MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin • Energy Quarterly

Ene

rgy

Qu

art

erl

y•

Ene

rgy

Sec

tor A

naly

sis

Aquarter centuryr after ther Nobel-prize-winning discoveryin 1986 of thef firstfi “high-temperaturet superconductors”

(HTS), the once heady prospect of transformingf the electri-cal power industry with lossless superconductors operatingat liquidt nitrogend temperature is no longer ar dream. Years ofmaterials research and ad suite of highlyf successful demonstra-tion projects have put HTS not only on the doorstep of thefelectric power grid but of facilitatingf its entry into the 21stcentury, including the increasingly mandatory shift tot green,renewable energy.

The US National Academy of Engineeringf describes the vastnetworks of electrif fi cationfi known as the grid asd “the greatestengineering achievement oft thef 20th century.” But thet futuredemands better: a grid thatd ist not fragmentedt butd trulyt nationalin scope;n where large amounts of powerf canr ben transported overdvast distancest in a fl ashfl by underground cablesd from wherevergenerated tod wherever needed;r where networks are redundantto back up outages; where overloads, short-circuits, losses,and fl uctuationsfl can be instantly compensated; where fl eetsfl ofelectric cars can be plugged intod the grid tod recharge withoutoverloading it; and where the frequency and voltage of thefpower arer reliably maintained (increasinglyd essential for therdigital society).

Many think thek grid is not up to the demands of thef 21stcentury without at seriousa effort tot upgrade. A 2010A report titledt“Science for Energyr Technology” from the US Department oftEnergy’s (DOE’s) Offiff cefi of Basicf Energy Sciences offers anentrée for high-temperaturer superconductor powerr equipment,r[but] “to achieve competitive cost-performance, significantfi im-tprovements over existingr wire performance are still required.”Surveying current majort HTSr challenges, the DOE report says,t“chief amongf these [is] a majora increaser (at leastt at factora ofr two)fin high-temperature superconducting current-carrying capabil-ity under operatingr conditions.”

Superconductors are best knownt forn ther lossless transmissionof dcf electric currents when cooled belowd their transitionr tem-peratures. In service, however, superconductors contain arraysof nanosized,f quantized, “fluxfl tubes” or vorticesr of supercurrentfcirculating around non-superconductingd cores. Vortices are noproblem when pinned atd structuralt defects like dislocations or

impurities in the superconductor, but att at critical current, theybreak free,k dissipating energy asy they move,y thereby introducingya resistance.a In practice,n incorporating pinning defects designedto block vortexk motion raises the critical current tot levels thatare useful, but theret is much room for improvement.r

Determining the maximum critical current thatt couldt bed ob-tained byd introducing the “ideal” distribution of defectsf awaitsthe ability to understand thed behavior ofr largef arrays of vorticesfin a fieldfi ofd pinf sites. In large arrays, the best criticalt currentsare only ~20% of thef theoretical critical current, but whyt is notknown. “There’s some theory but stillt lots of empiricism,”f saidDrew Hazelton of SuperPower,f Inc., a major HTSr wire maker.

The availability ofy HTSf conductors will be an industryn game-ychanger saidr Steve Eckroad of thef Electric Power ResearchrInstitute (EPRI). Up until now, utilities have relied ond a high-voltage, low-current networkt basedk ond copper forr generators,rtransformers, and urband underground cablesd (rural overheadlines are usually aluminum). Superconductors with no dc andonly small ac losses plus high current densityt change the equa-tion to lower voltager and higherd current.r HTS cables have fi vefitimes the capacity in the same cross-sectional area as conven-tional copper cables.r

One way to exploit this capability is by combining HTSwith renewable energy sources for ar truly green grid ind whichremotely generated powerd fromr renewable sources could traveldto distant consumerst over whatr windt andd solard powerr advocatesrcall “green power superhighways.” However, long-distancetransmission is not yett at near-term prospect fort HTS,r owingto the huge capital investment costs and an unproven trackrecord ofd benef fits.fi

Here is where the ongoing materials research could payd off.Eckroad said,d “Our studiesr suggest thatt reducingt the presentcost oft thef superconductor byr a factor ofr twof would bringd thecost of 10-GW,f 1200-mile-long, superconducting cables towithin range of thatf oft conventionalf overhead lines.d Since un-derground dcd cables also offer substantialr environmental, siting,and aestheticd benefitsfi over conventionalr overhead transmissiondlines, they may become an attractive alternative option in somesituations.”

Whether cabler or somethingr else, every HTS power applica-r

High-temperature superconductors change the gameBy Arthur L. RobinsonFeature Editor James Misewich

Are high-temperature superconductors ready toy take on the grid?

James Misewich, Brookhaven National LaboratoryArthur L. Robinson, [email protected]


Ene

rgy

Qu

art

erl

y•

Ene

rgy

Sec

tor A

naly

sis

beginsn with compositeh “wires” in whichn theh superconductoris only one component. Doped bismuth-strontium-calcium-copper-oxygen (BSCCO) compounds emerged as the fi rst-figeneration (1G) HTS material for commercialr use. Kilometer-long wires of BSCCOf in a silver matrixr can now be routinelyproduced, but sincet the matrix is 60–70% of thef volume, thewire is inherently costly. Sumitomo Electric Industries remainsthe principal supplier ofr 1Gf wires.

“Around 1999,”d said SuperPower’sd Hazelton, “we looked atdthe cost projections,t and wed moved tod yttrium-barium-copper-oxygen [YBCO]-based second-generationd [2G] wires.” Othervirtues of 2Gf conductors include higher currentr andt betterd per-rformance in high magnetic fi eldsfi even though the transitiontemperature is somewhat lower. YBCO conductors take theform of af multilayer taper made from superconductor depositedronto a textured metallicd substrate. Two coated-conductor tech-rnologies have caught on,t one researched atd Oakt Ridgek NationalLaboratory and commercializedd byd American SuperconductorCorporation (AMSC) and thed other developedr atd Lost AlamosNational Laboratory and usedd byd SuperPower.

As the demonstration projects suggest, it ist largely govern-ment supportt thatt ist helping to grow an HTS industry aroundthe world. Alan Laudern ofr thef Coalition forn ther Commercial Ap-plication of Superconductorsf estimates that fromt 1993 through2011, DOE offiff cesfi alone have invested aroundd $600d million inthe United Statesd to develop HTS technology spanning materi-als research, wire (or conductor)r fabrication, systems assembly(cables, transformers, rotating machinery), and demonstrationdprojects involving utilities aimed atd establishingt the technicalreadiness of HTSf systems. “The DOE investment hast broughtHTS technology a long way,” says David Knolld of thef South-wire Company, an HTS cable manufacturer.

To take one example, serving congested urband environmentsas demand growsd inexorably is a utility priority. Beginning in

2003, the DOE’s Offiff cefi of Electricityf Delivery and EnergyReliability (DOE-OE) has sponsored demonstrationd projectsat utility substations with costs split roughly 50-50 betweenDOE-OE and industry.d Based ond ac HTS cables with typicallengths of af few hundred meters,d these projects demonstratedthe ability to add capacityd by simply replacing copper withr su-perconducting cables in existing utility conduits, thus avoidingthe expensive and dauntingd prospect oft diggingf up the streetsto install new ones. Though some feel it wast premature, DOE-OE deemed thesed so successful that itt ceasedt supportingd newefforts after 2009.r

Nonetheless, the importance of continuedf governmentd sup-tport at this stage of HTSf development is well illustrated inAsia. The Korean Electric Power Corporationr (KEPCO), 51%percent government-owned,t has a particularly ambitious planfor HTSr in itsn grid andd hasd recently installedy ad distributiona cablenat itst I’cheon substation. Construction of af power substationr inBaiyin, China’s Gansu province, is supported byd the ChineseAcademy of Sciencesf and thed State-owned Assetsd Supervisionand Administrationd Commission of Baiyinf City.

What cant motivate US utilities to take up where DOE leftoff? Unfortunately, said Josephd Minervini of thef MassachusettsInstitute of Technology,f “there are no killer appsr to do the job.”Syed Ahmedd ofd Southernf California Edison says that exploit-ting the benefitsfi of HTSf over ther next decadet is most likelyt tostart witht niche projects to build ad technologya foundation whilegradually expanding markets and growingd the manufacturingcapacity to supply them. Minervini added that “microgrids”for isolatedr militaryd bases and larged urban data centers mayprovide another entréer by avoiding the need ford investmentr bytrisk-averse utilities.

HTS fault-current limiterst (FCLs) fi tfi Ahmed’st scenario nice-ly, as demonstrated ind Europe. The Swedish utility Vattenfallhas installed ind Boxberg,n Germany, an FCLn madeL by the Frenchcompany Nexans. These devices prevent hight currents (faultcurrents) generated byd disturbances in the grid fromd causingoutages, but int the 21st century,t they must bet faster actingr andhigher capacityr to maintain the reliability and securityd of thefnetwork. In essence, HTS do this by rapidly transitioning to thenormal, non-superconducting state when the current exceedstthe material’s critical current.

Off-shore wind energyd may emerge as a renewable-energyapplication for HTS.r Superconducting generators can be morepowerful and much smaller than conventional devices, saidAlex Malozemoff off AMSC,f who foresees that HTSt will per-mit thet relatively compact, high-power windr turbinesd that cantmake off-shore energy affordable. Though HTS-based super-dconducting magnetic energy storage (SMES) is farther fromrcommercial deployment, Qiang Li of Brookhavenf NationalLaboratory said thatd thet fast responset and highd effiff ciencyfi ofHTS-based SMES may make it a contender for the storagetechnology needed to complement intermittent sources likesolar andr windd farms.d

Please visit 10.1557/mrs.2012.205t for ar supplementarya tableon selected high-temperatured superconductor utilityr projects.

The Long Island Power Authority has been operating, since 2008, acable system manufactured by Nexans that utilizes AMSC’s high-temperature superconductor wire and an Air Liquide cooling system.Energized in April 2008, this is the world’s firstfi superconductortransmission voltage cable system, which is capable of transmittingfup to 574 MW of electricityf and powering 300,000 homes. Photocourtesy ofy AMSC.f


Ene

rgy

Qu

art

erl

y•

Inte

rvie

w

David Eaglesham is dedicated to using materials science to address energy problems. After

earning a PhD degree in physics at the University of Bristol (UK), he spent many years at

Bell Laboratories working on semiconductors and later took on management positions at

Lawrence Livermore National Laboratory and Applied Materials. It was at Applied Materials

that he began to connect with the solar industry, just as it was getting hot. When he joined

First Solar in 2006 as Vice President of Technology, it had around 350 employees and about

USD$50 million in revenue. The company now has grown to about USD$4 billion in revenue.

With an extensive portfolio of achievements in scientificfi research and ever eager for new

challenges, Eaglesham left First Solar this summer and has taken a new position with Mg-ion

battery leader Pellion Technologies. We caught up with Eaglesham at a corner brewery in

Ypsilanti, Mich., where we noticed that they were putting up photovoltaic panels on the roof

combined with a solar thermal energy system—a hybrid system. This auspicious beginning

led to an all-encompassing interview spanning the range from materials research to mitigating

global warming.

MRS BULLETINS :NN Several years agowhen you were president of thef Mate-rials Research Society, you called fora “Manhattan Project” for renewablerenergy. Is this still needed?DAVID EAGLESHAM: I think wekneed tod make carefully targeted invest-dments in basic research—that’s actu-ally something that thet governmentdoes very well. And Id think govern-kment investmentst in creating marketscan help to build ad marketplace inwhich all commercial sectors can thencompete without furthert governmentrinvolvement. And, I believe that ast aplanet, we will eventually put at priceon carbon—a carbon tax. The countriesthat aret firstfi tot implement sucht a taxwill be the fi rstfi tot develop low-carbon

technologies and willd ultimately be themost competitive.t

Where are the biggest areas of op-fportunity for materialsr in improvingphotovoltaic (PV) effiff ciencyfi and inreducing cost?With regard tod materials, it’s hard todcome up with a truly new semiconduc-tor becauser photovoltaic (PV) materi-als like cadmium telluride (CdTe) andcopper indiumr selenide (CIS) are goodby virtue of theirf defectr properties.t Inparticular, the recombination velocityat dislocationst and graind boundaries islow. And thatd makest predicting goodnew materials diffiff cult.fi As a result,a lot oft peoplef are working on thewrong problem. The big opportunities

that It see lie in exploiting capabili-ties that thet semiconductor andr LEDd[light-emitting diodes] industries havedeveloped tod achieve higher efr fiff cien-ficies in silicon and otherd materials—rtechnologies like heterojunctions,band engineering,d heteroepitaxy,dopant engineering,t barrier layers,r andcontact engineering.t Figuring out howtto adapt thoset tricks to photovoltaicsin an affordable way while achievinghigh throughputs presents a huge op-portunity. Other majorr PVr challengesVare around metrologies,d control ofthe process, and manufacturability,dand that’sd an area where the materialscommunity excels.

You said a lot of peoplef are workingon the wrong problem. What’s thewrong problem?There’s a big push to try to use earth-abundant elementst in semiconductors,but it’st hard tod findfi somethingd that’smuch rarer thanr tellurium, and indCdTe, only a tiny fraction of thef totalcost oft thef system is the tellurium.Similarly, the affordability of silverf isrnot at problem in crystalline silicon.So we’re trying to fixfi something thatisn’t at problem. What wet really wantis higher efr fiff ciency.fi

Where can materials help promotesustainable, long-term signifi cantfiPV powerV production?r

From materials research to climate change: David Eaglesham assessesthe solar energy industryInterviewed by Steve M. Yalisoveand Arthur L. Robinson

Steve M. Yalisove, University of MichiganArthur L. Robinson, [email protected]


Ene

rgy

Qu

art

erl

y•

Inte

rvie

wI think there’sk going to be sustainable,long-term silicon power productionrwithout goingt beyond thed existingmaterials technology. And ongo-ding improvement oft thef existing thinfilms,fi CdTe and CIS,d will continue tomake them successful. But theret arethings we could dod to move the currenttechnology landscape faster. Cad-mium telluride is the technology that’sfarthest fromt its theoretical capability,so there’s plenty of roomf for raisingr theeffiff ciency.fi In other materials,r produc-ing a low-defect densityt III–V onV glasseither byr heteroepitaxy or byr liftofftechniques would resultd int a very higheffiff ciencyfi device, as would goodd defectdpassivation. Tandem devices are cur-rently under- or unexploitedr andd willdbe a major directionr in the future.Organic devices are problematicbecause there’s still a question whethersomeone can produce a reliable deviceat ant effiff ciencyfi that’s high enough tomake a difference.

What about the vision of paintingf asolar cellr on the side of af wall? Mightorganics be a part of that?fThe notion ofn sprayingf on PVn isVseductive but delusional:t even ifn wefcould figurefi out encapsulation,t there’sstill wiring, mounting, inverters. And Iddon’t knowt howw we’llw test thet device forreliability. Rolling out flexiblefl devicesonto a roofa mayf bey more practical, andthe materials community willy figurefiout howt tow make the installation cost-neffective, which ish the biggest costt oft thefsystem. The fl exiblefl device is not neces-tsarily any organicn device. Whatever itr is,twe’d stilld choose the materials systemthat providest the highest eft fiff ciencyfi inyorder tor lower ther biggest cost.t

What challenges would we face ifwe wanted to get to even 1% of thefglobal energy production with PV?I think thek existing technologies andsome of thef emerging ones can take usmost oft thef way. There’s good reasond tobelieve that at cost oft electricityf around10 cents/kilowatt-hour isr reachable, andthat makest the economic impact oft afcarbon tax affordable. There’s a lot oft

fear thatr at carbon taxwill make the economygrind tod a screaminghalt. But thet differencebetween fossil and PVdelectricity is only about4 cents/kilowatt-hour,and thatd meanst that thettax is affordable.

Another arear wherewe can make headwayis in the “balance-of-systems” cost throughthigher efr fiff ciencyfi[solar cells]r and betterd constructionrpractices. New materials for powerrelectronics are likely to have a bigimpact, but lifetimet is an issue.

Many say that solar, as a signifi cantfienergy provider, is a non-starteruntil storage solutions become eitherbetter orr cheaper.r Do you agree?I believe that renewablest can and willdplay a large and successfuld role atthe grid leveld without storage.t Sincedemand ind any industrial country peaksin the middle of thef day, we can beginby adding solar tor the grid basicallyd asnegative demand duringd peak hours.kSecond, as solar-energy productionforecasting becomes better, we canuse PV asV a predictable component oftour generatingr “mix,” and ultimately,dwe can add wind,d which primarilygenerates energy during the night.Studies show we can achieve very highpenetration of renewablesf into the gridat thet expense of fossil-fuelf sourceswithout storage.t On the other hand,rthere’s a clear pathwayr for hybridsrto decrease the carbon footprint ofttransportation through improved stor-dage solutions, and thisd direction is anextremely interesting path for materialsrinnovation.

The solar industryr is in fl uxfl withIPOs and venture-capital cash infu-sions mixing with production scale-backs and business closings. Whatare the reasons for thisr shakeup?And how does it affect the prospectsof af solution to climate change?This industry exists for regulatoryr

reasons, and recentlyd regulators havechosen to make the PV marketV smaller,tresulting in lower capacity,r lowerprices, a collapse in stock prices,k andbankruptcy for noncompetitiver players.I think thek regulators got prettyt muchthe outcome they wanted: a much morecompetitive and lessd profitablefi solarsector thatr ist serving a significantlyfismaller market.r I think thek underlyingquestion is whether ther shakeup meansthat thet industry is closer tor or fartherrfrom the long-term goal of makingfsolar electricityr competitive with fossilfuels, and clearlyd the industry is muchcloser tor that goalt than it was.t

Regarding climate change, a com-petitive industry is not enough.t Peopleassume that allt we need ared renewabletechnologies that aret competitive withfossil fuels, and thed free market willt doits work. That’s delusional on three dif-ferent levels.t First, no energy marketanywhere in the world isd really com-petitive. Second, where renewables arenow competitive with fossil fuels, asin India, fossils are subsidized. Last, ifPVs were ultimately successful, peoplewould justd uset more electricity, andwe haven’t solvedt anything.d It’s aboutchoosing how we tax. Right now,t mosttax is income tax, so we’re taxinglabor asr opposed tod taxing more evenlyacross fossil-fuel consumption andlabor. If wef want tot import energyt andexport jobs,t we can keep 100% of theftax burden on labor. If wef want tot re-duce the use of fossilf fuels, we have tomake them more expensive. We haveto have a carbon tax. It’s that simple.t


Ene

rgy

Qu

art

erl

y•

Reg

iona

l Ini

tiativ

e

Super-capacitors take charge in GermanyBy Philip BallFeature Editor Yury Gogotsiy

Yury Gogotsi, Drexel UniversityPhilip Ball, [email protected]

The yellow-and-white Stadtbahntrams criss-crossing the streets of

Mannheim in southern Germany lookunremarkable, but somet of themf are lit-erally carrying what couldt bed the key togreener publicr transportation. The elec-trically powered vehicles use 30% lessenergy than their equivalents in mostother cities because they contain on-board systems for capturingr the energythat wouldt otherwised be squandered whendthe trams brake. This energy is convertedinto electricity, which is then stored indevices called supercapacitorsd mountedon the tram roof.

Supercapacitors are power-storagedevices that can supply onboard elec-trical power inr hybrid vehicles. Unlikebatteries, supercapacitors can be chargedand dischargedd ind seconds and cand with-stand many hundreds of thousandsf ofsuch charging cycles. This makes themideal for energy-savingr applications thatcapitalize on transient opportunitiest forrecharging, such as energy capture dur-ing braking, and thatd requiret power tor bedelivered ind short burstst of perhapsf halfa minute or less.r They can, for example,r

help with acceleration or canr restart en-tgines that cutt outt tot reduce fuel use andpollution in stop-and-go traffiff c.fi

Whereas batteries store energy inchemical form—in substances that canreact tot release electrical energy—capac-itors store it byt simply piling up electri-cal charge on two electrodes. The largerthe electrodes and the closer they are,the more energy that cant be stored. Anordinary capacitor consistsr of twof con-ductive plates separated byd an insulating(dielectric) layer. But a supercapacitor(sometimes called an ultracapacitor)holds its charge a little differently. Typi-cally it containst two conductive porouselectrodes—usually made of carbon—fimmersed in a liquid electrolyte andseparated byd a very thin insulating fi lm,fiusually made of af porous polymer. Thecharge is stored by adsorption of ionsfonto the high-surface-area electrodes.When the electrodes are charged, thisproduces a layer ofr oppositelyf chargedions on their surfaces:r a so-calleda electri-dcal double-layer, which is why this typeof supercapacitorf isr often calledn ad double-alayer capacitor.r

The potential of supercapacitorsf to as-sist int powering vehicles was displayedin dramatic fashion in the 24-hour speedrrace at Let Mans this June, when Toyotafieldedfi a hybrid TS030 that used themfor energy-capturer during braking. Thedevices performed perfectly,d but at crashscuppered thed vehicle’s bid ford glory.r

The Mannheim trams are rather morersedate, but witht their ownr onboard pow-der, they can keep running across shortgaps or disruptionsr in the electricity line,for example due to ice or where over-head powerd linesr cannot bet deployed fordaesthetic or technical reasons. The en-ergy source can also be tapped tod driveair conditioning,r automatic windows, orpassenger doors.r

These trams have been reapingn the ben-efi tsfi of supercapacitorsf since 2003 andare now joined byd a host oft otherf public-rtransportation systems in Germany andbeyond. Supercapacitor technology isdeployed, for example,r on Spanish and

French trains and hybridd busesd all overthe world, on construction equipmentsuch ash cranes, and ond garbage-collectionntrucks. On buses, it cant reduce the effec-tive carbon-dioxide emissions by around30%, while the Munich-based heavy-ve-dhicle manufacturer MANr estimated thatdtheir supercapacitor-r fittedfi coachesd eachsave around $4,500d a year onr fuel costs.

Reducing fuel consumption and emis-dsions during the “dead time”d of standingfat stops,t intersections, and trafd fiff cfi lights isa particularlya pressing concern for buses.rSince 2001, MAN has been developinghybrid supercapacitor buses called theLion’s City Hybrid. The current com-mercial model, available since 2010,cuts diesel consumption by up to 30%,and isd now being used ond a small scalein Paris and somed other Europeanr cities.In principle, buses can recharge their su-rpercapacitors not justt duringt braking, butalso at everyt bus stop by making contactwith overhead charging lines for justr afew seconds.

So far ther take-up of thef technologyhas been relatively modest. But itt lookstset tot expand bothd as energy-saving andlow-emission technologiesn become moreimperative and asd the technical capabili-ties of supercapacitorsf improve. Super-capacitors are not byt any means a pana-cea for greenr transportation: they have alower totalr energy density than batteriesby one or twor orders of magnitude,f forexample. But theyt have a higher powerrdensity, delivering much more powerover ar short periodt ofd time.f

“There is no single perfect energy-storage solution, no ‘one size fi tsfi all,’”said materialsd scientist Yuryt Gogotsi atDrexel University in Philadelphia. “A‘battery of thef future’ may be a battery-supercapacitor hybrid having the longlifetime, fast charging, and highd powerof af supercapacitora combinedr withd ah highaenergy density of af battery.”

The main producersn of supercapacitorsfworldwide are Nesscap Energy, based indSouth Korea, and the California-basedcompany Maxwell Technologies. Theseand otherd manufacturersr supply the basic


cells to companies such as Siemens andthe Belgian firmfi 4Esys, which incorpo-rate them into modules that cant be addeddirectly intoy vehicle designs. In Germany,nthe electrical-component manufacturerWIMA inA Mannheim, which special-izes in capacitors and supercapacitors,manufactures both individual cells andmodules. Siemens has developed twoenergy-storage supercapacitor modulesrcalled Sibacd and Sitrac,d which are incor-porated intod the vehicles or ther power-supply lines respectively to capture ener-gy duringy braking. Sitras can alson be usedto maintain the voltage of thef networkduring peaks of highf power demandr ordtemporary outages.

Despite the promise of supercapacitorsfin transportation, their uptaker has beenrather slow and cautious over the pastdecade. “There have been an lota oft projectsfwhere prototypes are tested overd ar con-siderable time,” said Frank Herrmann,an engineer atr WIMA.t The market has,thowever, been picking up speed overd therpast two or threer years, partly becauseof risingf fuel costs. “The economics aredriving it,” said Juergen Auer ofr Ness-fcap’s division in Schondorf, Germany.“It’s very sensitive to what happens todiesel prices.”

And although Germany is noted forits commitment to green energy poli-cies, the kind of supportf from central

government that has been enjoyed by,say, photovoltaic energy (where subsi-dies and guaranteesd of competitivef rateshave boosted production and use) hasnot beent extended tod the supercapacitormarket. “German authorities are prettyconservative,” Auer feels.r There are alsono coherent planst for howr and whered thetechnology might bet introduced. “Eachenergy ory transportr companyt hasy a hybridasolution, whether withr supercapacitorsh orbatteries,” he said, “but thet finalfi custom-ers like cities aren’t readyt to buy it ort im-rplement itt intot their fl eets.fl That makest ita very diffiff cultfi productt tot market.” Auercontrasts this ruefully with China, whereefforts to use supercapacitors in transportn

began only about fourtyears ago but havet re-sulted already in tensof thousandsf of “super-fcap” buses on the road,especially in Shanghai.n“Right nowt the biggestmarket is in China,”Herrmann agreed.

However, Herrmannthinks that Germany’stenergy policy to userenewable rather thannuclear energyr will in-directly help the tech-nology. Offshore windfarms, for example,need regular mainte-nance using ships andhelicopters, and super-dcapacitors should helpd

to reduce the running costs of thesef fl eets.flGermany’s federal system offersm scope

for regionalr projects, and individuald cit-ies such as Mannheim have sometimeslaunched theird ownr initiatives. Siemens’Combino trams, which use the Sibacsystem, have been deployed in severalGerman cities,n including Augsburg, Düs-seldorf, Ulm, and Potsdam, as well asoutside Germany in Amsterdam, Basel,and Budapest for example. The com-pany unveiled ad second-generation tramsystem, called Avenio,d last year.t Mean-while, the Mannheim trams use the Mi-trac supercapacitor systemr produced bydthe Berlin-based company BombardierTransportation. Last year, the German

Ene

rgy

Qu

art

erl

y•

Reg

iona

l Ini

tiativ

etransport operator RNV, which servesthe Rhein-Neckar region and includesMannheim, Heidelberg, and Ludwig-shafen, ordered 11 more Bombardiertrams for itsr 200-km network.

In part, the obstacles to wider user arematters of infrastructuref and design:engineers are more used to thinking interms of hybridf vehicles with batteriesthat require charging cycles of severalfhours, rather than the several secondsneeded byd supercapacitors. But ast theiruse hinges mostly on economics, it ist vi-tal to improve performance while reduc-ing costs. Supercapacitors can have animpressive lifetime—15 to 20 years—butthe initial outlay is still rather high.r

A keyA problem is that the individualcells develop relatively low voltages—typically less than 3 V—because of theflimited electrolyted stability. This meansthat tot obtain the 24 V typicallyV neededfor vehicler systems, several cells must betconnected ind series,n making for relativelyrcostly and bulkyd modules. “We need todincrease voltage per cellr in order tor havehigher energyr density and fewerd cells,”rsaid Herrmann.d

Auer said that some of thef raw ma-terials, such as the porous carbon elec-trodes, are also somewhat costly,t and sodis the manufacturing process. However,Gogotsi insists that “there is no funda-mental reason for supercapacitorsr to beexpensive, because they use just carbon,tpolymer film,fi an inexpensive aluminumfoil, and an organic electrolyte, withno rare or expensive elements.” Costsshould fall, for example,r simply as thescale of productionf increases and manu-dfacturing methods improve. That wouldtnot only boost current uses but enablenew applications to emerge.

This all leaves Auer optimisticr aboutthe prospects. “There’s still a lot oft roomffor futurer growth,” he said. “There arepotential uses everywhere.” He said thatdsupercapacitors are one of thef few elec-tronic components that havet had ad steadi-aly growing market overt recentr years.t

“It takest time,” he said.

Trams in Germany powered by supercapacitors use 30% lessenergy than their equivalents in other regions. Credit: RNV


Ene

rgy

Qu

art

erl

y•

Ene

rgy

Foc

us

ENERGYFOCUS

Tim Palucka

Dielectric core–shell optical antennas can enhancen the trappingof solarf radiationr inn photovoltaicn devices, enabling a 70-nm-athick a-Si:Hk thin filmfi to absorb about ast much radiation asa typicala 400-nm-thick anti-rek fl ectionfl coating thin fi lm.fi Thestructures tested usedd semiconductingd a-Si:H as the core anddielectric materials such ash ZnO and Sid 3N4 as the shell. LinyouCao of Northf Carolinah Statea University andy A.d Paul Alivisatosat thet University ofy California,f Berkeley, and theird colleaguesrreported ind Nano Letters that thet solar radiationr absorptionenhancement comest from multiplicationm ofn contributionsf fromleaky modey resonances in then semiconductor andr anti-red fl ectionfleffects in then dielectric. The size ratio of thef core and shelld is thekey toy optimal absorption. After optimizingr the dielectric shellfor anti-rer flection,fl sizing the semiconductor corer at at 0.5–0.6acore–shell ratio preserves the intrinsic leaky modey resonances.This technology couldy leadd tod thin solarn cellsr with improvedhconversion efn fiff cienciesfi and lowerd costr thant currentlyn availableysolar cells.r

Thinner solar cells use dielectric core–shelloptical antennasNano Lett.o DOI: 10.1021/nl301435r

By studying the morphology and crystald structure of lithium-fsulfur (Li-S)r batteries during operation, researchers at StanfordtUniversity led byd Michael Toney have discovered phenomenadthat largelyt contradicty previoust ex situx studies. Sulfur hasr greatpromise as a cathode in this system because of itsf high energydensity and lowd cost. However, Li-S batteries fail after onlyr afew tens of cycles,f compared tod thousands for Li-ionr batteries.nPrevious ex situ studies have attributed thed short lifetimet tothe dissolved sulfurd formingr electrically insulating crystallineLi2S following discharge and tod sulfur’s failure to recrystallizeat thet cathode following charging. Now, using in operandosynchrotron-based XRDd and transmissiond x-ray microscopy,Toney and colleagues have shown that no crystalline Li2Sforms; they speculatey that int previousn studies, it wast an artifactnof thef ex situ XRD process. Furthermore, recrystallization ofsulfur canr occur dependingr on the cathode morphology. Theyconcluded thatd in operando studies are necessary for furtherrevaluation of sulfurf cathodesr for Li-Sr batteries.

Study clarifi es short Li-S battery lifetimeJ. Am. Chem. Soc. DOI: 10.1021/ja2121926

Transparent nanogenerators use triboelectric effectNano Lett.o DOI: 10.1021/nl300988z

Flexible transparent nanogenerators (FTNGs) based ontriboelectric phenomena could be used as self-poweredsystems for touchscreensr in electronic displays according toresearch done by Zhong Lin Wang of thef Georgia Instituteof Technologyf and co-workers from Xiamen University,China. The triboelectric effect referst to electrificationfi resultingnfrom frictional contact between materials. The FTNGsuse only transparent materials, including a patternedpolydimethylsiloxane (PDMS) thin filmfi sandwiched betweendtwo sheets of polyester,f each capped with an indium tinoxide electrode. The 460-micron-thick devicesk with an area

of 5.4f cm2 produced upd to 18 V ofVelectricity at a current density ofapproximately 0.13 A/cm2 whenfl exed,fl yielding 0.7 A ofA currentf(up to ~13 W ofW power)f at at fl exingflfrequency of 1f Hz. The researchersfabricated PDMS thin filmsfi withlinear, cubic, and pyramidald patternsto increase friction during bending.As reported in Nano Letters, the

triboelectric effect was greatest in devices with pyramidalpatterns, followed byd the cubic, linear, and flatfl PDMSt sheetsin decreasing order.

Artifi cial enzyme could enhancebiofuel production

The Plante Cellt DOI:l 10.1105/tpc.112

Lignin, the tough biopolymer inr plant cell walls that givesthem structural strength, presents significant difficultieswhen trying to convert biomass to biofuels; it interferes

with digestive enzymes thatmust access the sugars insidethe plants to produce thesefuels. Now, using mutagenesisto create an artificial enzymethat inhibits the polymerizationof threef lignin precursors calledmonolignols, researchers atBrookhaven National Laboratoryand thed University of Wisconsin,fMadison, have developed amethod of reducingf the amountof ligninf in Arabidopsis plantsby up to 24 percent withoutcompromising the growth of thef

plants. As reported ind then July 31y edition ofn The Plant Cell,t thisdevelopment couldt signid ficantlyfi reduce the cost oft biofuelsfby removing pretreatment stepst that aret currently needed todreduce lignin content int industrial biofuel processes.

Transgenic plants with an enzyme variant

designed to limit the production of the cell

wall component lignin showed no dramatic

difference in growth (top) compared to the

control plants, but had less lignifiedfi xylem

tissue (stained violet on bottom).

A new high-output, fl exiblefl and trans-

parent trioboelectric nanogenerator

produced from transparent polymer

materials. Credit:tt Zhong Ling Wang.n

806 MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin © 2012 Materials Research Society

Introduction personal journey in nanowire research started withd Profes-

sor Charlesr Lieber atr Harvardt University,d where I investigated

fluxfl line pinning in highn TcTT superconductors using nanowires as

pinning centers. More than 15 years later, I devote most oft myf

efforts in studying these nanostructures for energyr conversion

and storaged purposes.

I would like to acknowledge the pioneering contribution

of R.S.f Wagner atr Bell Laboratories. Much of ourf nanowirer

research today relies on the very powerful method developedd

by Wagner, known as the “vapor-liquid-solid process”d (VLS).1

In 1964, Wagner describedr thed VLS growth of siliconf micro-

and nanoscaled wires or whiskersr and pointedd outd thatt onet of

the catalysts that cant be used isd gold. Gold isd often used todayd

to grow silicon nanowires, as well as copper, nickel, and, more

recently, aluminum. Wagner alsor noted thatd nanowirest can grown

into different crosst directions. Over ther past 10t years, there have

been thousands of papersf on silicon nanowire growth, growth

direction control, and the use of differentf catalysts and sub-

strates, all based ond Wagner’sn original concept.2 Wagner’s work

in the 1960s set thet foundation for muchr of today’sf nanowire

research.

My firstfi nanowire experiments were in the early 1990s,

when manyn ofy usf were still working on high-Tn cTT superconductors

(HTSC). The subject oft myf PhD thesis was to fi ndfi ad way to

introduce linear defectsr within high-temperature superconduc-

tors, in the hope of increasingf the critical current density.t Our

approach was to introduce single-crystalline nanowires into a

high-TcTT cuprate superconductor tor make a composite, to create

stable linear tracks,r and tod increase the critical current densityt

by “pinning” the fluxfl lines.3 This work paralleledk muchd of thef

work conductedk atd thet time using fast iont irradiation to create

linear defectsr within high-temperature superconductors. That

was the start oft myf research in this very exciting area, and itd

Semiconductor nanowire building blocks: From fl ux line pinning to artifi cial photosynthesis

Yang

The following articleg is an edited transcriptd oft thef MRS MedalS Lecturel presented byd Peidong Yangg ong

November 30,r 2011 at thet 2011 Materials Research Society Fall Meetingl ing Boston. The MRS MedalS isl

awarded ford ar specificfi outstanding recentg discoveryt or advancementr thatt hast a major impactr ont the

progress of af materials-related fid eld.fi Yang receivedg thed award ford “outstandingr contributionsg in the

creative synthesis and assemblyd of semiconductorf nanowiresr and theird heterostructures,r and innovationsd

in nanowire-based photonics,d thermoelectrics, solar energyr conversion, and nanofld uidicflfl applications.”

Semiconductor nanowires,r by defiy nition,fi typically havey nanoscale cross-sectional dimensions,

with lengths spanning from hundreds of nanometersf to millimeters. These subwavelength

structures represent a new class of semiconductorf materialsr for investigatingr light generation,

propagation, detection, amplification,fi and modulation. After more than a decade of

research, nanowires can now be synthesized and assembled with specificfi compositions,

heterojunctions, and architectures. This has led to a host of nanowiref photonic and electronic

devices, including photodetectors, chemical and gas sensors, waveguides, LEDs, microcavity

lasers, and nonlinear optical converters. Nanowires also represent an important class of

nanostructure building blocks for photovoltaics as well as direct solar-to-fuel conversion

because of theirf high surface area, tunable bandgap, and efficientfi charge transport and

collection. This article gives a brief historyf of nanowiref research for the past two decades

and highlights several recent examples in our lab using semiconductor nanowires and their

heterostructures for the purpose of solarf energy harvesting and waste heat recovery.

Peidong Yang, Departments of Chemistry,y and Materials Science and Engineering , Lawrence Berkeley National Laboratory; [email protected] DOI: 10.1557/mrs.2012.200

SEMICONDUCTOR NANOWIRE BUILDING BLOCKS: FROM FLUX LINE PINNING TO ARTIFICIAL PHOTOSYNTHESIS


was also the beginning of nanowire research

in Lieber’s group. Shortly after, the landmark

paper by Morales and Lieber (1998) 4 introduced

the laser ablation method for the growth of

silicon nanowires, once again based on the VLS

process. In this case, the vapor was generated

using laser ablation. After that, there was sig-

nifi cant research using VLS processes to grow

semiconductor nanostructures of many different

compositions. Many different vapor deposition

methods were employed, including chemical

vapor deposition (CVD) and several physical

vapor deposition techniques.

After joining the University of California,

Berkeley faculty in 1999, my fi rst research

effort was to investigate the growth mechanism

of semiconductor nanowires. Initially, we tried

to explore in situ growth of Ge nanowires; we

observed the process of VLS growth using an

in situ high temperature transmission electron microscope

(TEM);5 in this case we were investigating the nucleation of a

Ge nanowire, using a gold nanoparticle heated to 800o C as the

catalyst. Based on the Au-Ge phase diagram, at 800 o C, we can

fi rst observe an alloying process that leads to the liquid forma-

tion; with more Ge vapor incorporation, the system eventually

moves into a biphasic region, and we can observe the nucleation

event in real time. The creation of this solid-liquid interface

is the starting point for one-dimensional crystal growth; this

underpins the fundamental nucleation step at the nanometer

scale for all of the VLS processes.

In the 1990s, there were very few papers published in

the semiconductor nanowire fi eld, but there has been a rapid

increase in the past 10 years, and nowadays there are thousands

of papers published every year on this subject, with hundreds of

research groups now active in this exciting area.

Nanophotonics In the last decade, we have seen many important

discoveries in a couple of different directions,

such as nanoelectronics and photonics, and

next I will describe some of the work we did

in these areas. In 2000, we discovered that a

single-crystalline semiconductor nanowire can

serve as a nanoscopic laser cavity ( Figure 1). 6

This is basically a Fabry – Pérot cavity, and

Figure 1b shows the far-fi eld emission patterns

from single nanowires, the power-dependent

laser spectra from these individual nanowires,

and the integrated intensity as a function of the

incident power. For the last couple of years,

we have extended our work on photonics from

nanolasers to subwavelength waveguides, as

well as using a single nanowire to perform

nonlinear optical mixing. 7,8 Due to the high

refractive index of these nanostructures, we can

achieve these functionalities either in air or in a liquid. 9 This is

quite powerful, as we have basically developed a nanoscopic

light source that can be potentially used in a liquid medium.

Recently, we have been trying to utilize these nanoscopic

light sources for endoscopy at the single living cell level. 10

We attached a single nanowire onto an optical fi ber so that

we could deliver light directly to the nanowire waveguide and

then use the waveguided end emission to perform imaging at

the single-cell level. These nanowires are very robust, do not

fracture easily, and can be bent, so they are stable enough to be

used in single-cell endoscopy. As illustrated in Figure 2 , the

nanowire had quantum dots attached to the surface that were

cleaved by UV light. After cleaving, the quantum dots were

delivered to a single cell. As an example, we can deliver the

quantum dots into the cytoplasm and directly to the nucleus. It is

Figure 1. Nanowire nanolasers. (a) Schematic of an optically pumped nanowire laser

cavity. (b) Lasing spectra from an individual nanowire cavity, a far-fi eld optical image of a

lasing 30- μm-long GaN nanowire (left inset), and power dependent curve for the nanowire

cavity (right inset). Reprinted with permission from Reference 23. ©2005, American

Chemical Society.

Figure 2. Nanowire single-cell endoscopy. (a) Schematic of the spatiotemporal delivery

of quantum dots (QDs) into a living cell using a photoactivated nanowire endoscope. Inset

shows QDs were conjugated to the nanowire via photocleavable linkers. (b) Fluorescence

confocal image of a Hela cell after QD delivery showing nanoprobe delivered QDs (red dot)

sitting in the cytoplasm within the cell membrane (green), which was labeled with Alexa

Fluor 488 conjugate of WGA. A 488 nm laser was used to excite both the cell membrane

stain and the QDs. The inset is a dark-fi eld image of the cell during QD delivery with a

nanowire endoscope (bright line). Scale bar is 20 μ m. Reprinted with permission from

Reference 9. ©2012, Nature Publishing Group.



then possible to perform high spatial-resolution imaging. Using

the nanowire as a scanning probe in a liquid environment is a

powerful imaging tool. I believe that in the next few years, we

will continue to see many new developments that integrate all

of these active nanoscopic optical components into integrated

nanophotonic circuits and also use the nanowire probe at

the single-cell level for targeted delivery, imaging, or even

stimulating single living cells.

Thermoelectric energy conversion I want to focus now on how these nanoscale one-dimensional

structures can be used for terawatt energy conversion and stor-

age. We all know that the amount of energy required worldwide

is on the scale of terawatts, and the percentage of renewable

energy in the current energy portfolio is very limited. Many

research groups are now working on ways to save energy or

convert renewable energy. My group approaches this problem

from two different directions. The fi rst relates to saving energy

by developing nanostructured silicon thermoelectrics for waste

heat recovery, and the second, more recent effort, involves

developing nanostructures for solar energy conversion, either

directly into electricity or into liquid fuels through artifi cial

photosynthesis.

If we examine the performance of power plants delivering

15 TW of power through fossil fuels, we fi nd the total energy

expended is 38.6 TW with 40% effi ciency for the heat engine.

This means that about 20 TW is lost as waste heat to the envi-

ronment. The question is whether we can develop a technology

to recover some of this waste heat energy? A thermoelectric

material provides a suitable strategy for this. If we had a ther-

moelectric waste heat recovery module with an effi ciency of

3%, we could recover up to 1 TW of energy; this is a potential

solution to the terawatt energy challenge, provided we can fi nd

the right thermoelectric material to achieve this.

We started to work in this fi eld about 10 years ago, together

with Arun Majumdar in the Mechanical Engineering Depart-

ment at Berkeley. We fi rst examined the ther-

mal transport in a single-crystalline silicon

nanowire. For any thermoelectric material, the

important fi gure of merit is the dimensionless

parameter ZT, which is related to the Seebeck

coeffi cient, the electrical conductivity, and the

thermal conductivity of the material. A high

ZT value means high energy conversion effi -

ciency. We considered using one-dimensional

structures to control thermal transport, essen-

tially to suppress the thermal conduction, while

allowing all the carriers, electrons, or holes to

transport and maintain a reasonable level of the

Seebeck coeffi cient. The design principle, for

Si in particular, was to choose a wire diameter

that is smaller than the phonon mean-free path

(200–300 nm) but larger than the electron mean-

free path (1–2 nm). We measured the thermal

conductivity of individual nanowires using a

suspended microelectromechanical system (MEMS) device

by heating one end and then measuring the thermal transport

to the other end of the MEMS platform. By measuring the

steady-state temperature on the other end, we can calculate

the thermal conductance of the nanowire thermal bridge, nor-

malize it to the nanowire cross-section, calculate the thermal

conductivity, and then examine the size dependence of these

semiconductor structures.

The initial results were very promising. 11 This was the fi rst

study of size-dependent thermal conductivity of Si nanowires.

The thermal conductivity of the Si nanowires was suppressed

by an order of magnitude compared to bulk Si. By simply

decreasing the diameter of the nanowire, we could decrease

the thermal conductivity through phonon-boundary scattering.

However, the thermal conductivity was still quite high, typically

at the level of 10 Wm–1m K–1KK , which is not suffi cient for a good

thermoelectric material. Later, after several years of exploring

various strategies to further decrease the thermal conductivity,

we started investigating a solution-based process for the pro-

duction of silicon nanowires with a rough surface, which was

found to be critical for further thermal conductivity reduction. 12

This is a simple etching process that creates a rough surface on

the silicon nanostructure. It led to the large scale production

of rough silicon nanowires; they can now be made on six-inch

wafers. The nanoscopic surface roughness turns out to be the

key structural feature to further decrease the thermal conduc-

tivity, without signifi cantly decreasing the Seebeck coeffi cient

and electrical conductivity of these nanowires. These nano-

wires now have values of ZT that are two orders of magnitude

greater than that of bulk Si and reach a value of 0.6, as shown

in Figure 3 .

In 2009, together with a Berkeley graduate student, I

founded the Alphabet Energy Company, and we are already

producing modules using these Si nanostructures. The com-

pany currently employs 16 people and is continuing to hire, so

nanowire technology is indeed creating jobs. We will have our

Figure 3. Nanowire thermoelectrics. (a) Cross-sectional scanning electron microscope

image of a rough silicon nanowire array. The inset is a bright-fi eld transmission electron

microscopy image of a segment of a rough Si nanowire. The roughness is clearly seen at

the surface of the wire. The selected area electron diffraction pattern (inset) indicates that

the wire is single crystalline all along its length. Scale bar for the TEM image is 20 nm.

(b) Single nanowire power factor (red squares) and calculated ZT (blue squares). Reprinted

with permission from Reference 11. ©2008, Nature Publishing Group.



fi rst pilot thermoelectric testing in mid-2012. The prospect of

using these nanostructured silicon thermoelectrics for waste

heat recovery is real.

Conversion and storage of renewable energy I will now concentrate on the semiconductor nanostructures

required to convert and store renewable solar energy. Since we

are aiming to operate at the terawatt level, the amount of solar

energy required is enormous, and this must be achieved with

reasonable effi ciency and at a reasonable cost. Before discuss-

ing the use of nanowires for this purpose, I want to emphasize

some of the fundamental materials requirements for designing

solar energy conversion devices. First is the use of the earth’s

abundant elements, which will be absolutely necessary if we are

to work on the terawatt scale; and second, whether we are devel-

oping solar-cell devices or an artifi cial photosynthetic system,

we need to use less-energy-intensive processes to make these

materials and to create the necessary systems. Where possible,

we want to use mild chemistry or mild processing conditions

and minimize the energy used to create these energy conversion

devices; this applies to both photovoltaic (PV) applications and

artifi cial photosynthesis.

We started our research on nanowire-based solar cells about

seven years ago. The fi rst example was to use a dye-sensitized

cell (DSC) as a test platform, with single-crystalline nanowires

as the electrodes, replacing the traditional TiO2 nanoparticle

percolation network. 13 One of the major advantages of using

the single-crystalline nanowire conduction channel in a DSC

cell is that electron diffusivity is enhanced, compared to the

nanoparticle percolation networks. In recent years, there have

been various research efforts utilizing nanowires in composite

solar cells or in a core-shell gemotry. Nanowire structures have

many benefi ts; it is fairly straightforward to design a suitable

interface to facilitate charge separation, and the charge collection

can also be quite effi cient. In addition, the core-shell nanowire

geometry has the advantage of orthogonalized light absorp-

tion and charge separation, a feature quite important for mate-

rial systems with short minority carrier diffusion length. For

nanowire arrays with a suitable periodicity, we can introduce a

light-trapping mechanism so that less material will be needed

for light collection.

With all the benefi ts of using nanowires as the PV active

elements, we should not forget that we are also introducing an

important potential limitation—the large junction area. These

large interfacial areas in nanostructured solar cells naturally

introduce signifi cant trapped states and various channels for

carrier recombination. If we examine the performance of these

nanostructured PVs, we typically fi nd that the fi ll factor is low,

and the open circuit voltage (VocVV ) is low. This is intrinsically

related to the high junction surface area solar cell and is par-

ticularly obvious in the nanowire polymer composite solar cell,

where low fi ll factors and VocVV were typically reported. To miti-

gate this problem, we developed a simple solution procedure

to produce high quality nanowire core-shell solar cells with

excellent fi ll factor and VocVV . 14 We fi rst make CdS nanowires in

solution, then we use a mild cation-exchange chemistry to pro-

duce the core-shell nanostructure, as shown in Figure 4. This

cation exchange chemistry was developed in Paul Alivisatos’s

group at the University of California, Berkeley. All that is

required is to dip a nanowire for a couple of seconds in a

copper solution at 50°C; the core-shell nanowires can then

be fabricated into an asymmetrical junction to measure PV

performance.

It is interesting to note that we can create the inter-

face using mild processing conditions. The resulting

interface is highly coherent, with a minimum amount of

defects or dislocations. The CdS-Cu 2 S interface is illus-

trated in Figure 4b–c . The heterojunction prepared by

this method is atomically well-defi ned with few interface

defects, which enables excellent charge separation and mini-

mal minority carrier recombination. As a result, our nanowire

PV device shows excellent VocVV and fi ll factors. Figure 4e shows

the I–V curve from the individual nanowires, and the current

output is limited by absorption in the visible region, which is

limited in turn by the thickness of the Cu2 S shell ( Figure 4f ).

However, the most important message here is that the fi ll factor

and VocVV of this device are even better than those for traditional

thin fi lms. The insight we gained here is that we can indeed

create highly coherent interfaces using mild chemistry and

less-energy-intensive processes. Such a coherent interface is the

key structural feature that could lead to high fi ll factor and VocVV

for nanostructured solar cells. I believe that with this strategy,

we will be able to push the effi ciency above 10% by adding

more visible light absorbing Cu2 S.

Artifi cial photosynthesis For PVs, we will have to worry about the energy storage

issue because of the intermittency of sunlight irradiation. It

would be much more appealing if somehow we can convert

the solar energy directly into liquid fuels. This is artifi cial

photosynthesis, where we try to mimic what is happening in

nature: to convert solar energy and store it in chemical form,

possibly in the form of hydrogen, methanol, or even gasoline.

In our daily life, we use gasoline in cars, and as a result, CO 2is emitted into the environment. Artifi cial photosynthesis

attempts to convert the CO 2 and water back into liquid fuel,

which will be a truly carbon neutral process. Once again, in

this artifi cial photosynthesis process, we have to emphasize

the need to use earth abundant elements and more energy

effi cient processes.

If we examine the chemical reactions required to generate

solar fuels, either by water splitting or CO 2 reduction (where

there are many possible products), they are all “uphill” reac-

tions, not spontaneous; we will need solar energy as the energy

input and catalysts to drive these reactions. Therefore, in order

to generate solar fuel by artifi cial photosynthesis, we need two

things: the fi rst is quite similar to PVs; we need to have the

semiconductor and the junctions for light harvesting and charge

separation. Then we need to utilize the charge carriers to run the

reactions, instead of extracting them as electricity as in solar



cells. If we consider the materials challenges in developing

this artifi cial photosynthesis system, we fi nd many of them are

relevant to the overall work performed by the materials research

community at large. Two examples are photocathode materials

for hydrogen and CO 2 reduction and photoanode materials

for water oxidation. The photoanode materials are particularly

challenging, and we do not have a solution at this moment. We

also need to develop better catalysts for water reduction, CO 2reduction, and water oxidation. In order to balance the ion fl ux

during this artifi cial photosynthetic process, we also need an ion

conductive membrane that is gas impermeable, mechanically

stable, and optically transparent, and this must sit between the

light absorbing photoanode and photocathode.

Semiconductor nanowires, with their high charge mobility

and high surface area, are good candidates to serve as the

scaffold for the entire artifi cial photosynthetic system. Solar

water-splitting research commenced with Fujishima and Honda

using TiO 2 , 15 which is a UV absorbing material. Although the

basic concept works quite well, this will never result in a high

effi ciency photoelectrochemical (PEC) cell because the fraction

of UV photons in the solar spectrum is limited. Nozik 16 pro-

posed the PEC diode concept based on the concept of a coupled

photocathode and anode, with an ohmic contact in between

so that minority carriers can be used on both sides to run the

water reduction and the water oxidation for complete solar

water splitting. This two-photon process essentially mimics the

photosynthetic process in nature; now we can place two small

bandgap materials back-to-back in a diode confi guration and

make use of the UV and visible spectrum. The major advantage

here is the increased photocurrent from the coupled photocath-

ode and anode system, compared to the single bandgap struc-

ture. The single bandgap structure is limited to the UV region,

while in the two-bandgap system we can make use of the entire

solar spectrum, and current density at a 20 mA/cm 2 level is

possible. Our main goal was to capture the entire solar spectrum

and use it for solar fuel conversion.

We have a couple of good candidates for the photocathodes.

The Si nanowire array is certainly very promising; Figure 5shows black-colored silicon nanowire arrays in the form of

2 cm × 2 cm wafers. We can make these ordered nanowire

arrays shown in Figure 5b , and decorate them with water reduc-

tion catalysts such as Pt or MoS2 nanoclusters. These devices

have reasonable photocurrent output up to 20 mAcm –2m , and

we can implement the “half reaction” of hydrogen generation

Figure 4. Core-shell nanowire solar cell. (a) Schematic of the CdS-Cu 2 S core-shell nanowire solar cell. (b) High-resolution transmission

electron microscopy image of a CdS-Cu2S nanowire at the heterojunction. (c) Constructed inverse fast Fourier transform image along the

growth direction for the area marked in (b); the green area shows the typical lattice fringe distortion at the core-shell interface. (d) Scanning

electron microscopy image of a photovoltaic unit; CdS and Cu 2S are highlighted with yellow and brown false colors, respectively. (e) I–V

characteristics of a core-shell nanowire under 1-sun (AM 1.5 G) illumination. (f ) Wavelength dependence of the photocurrent compared with

simulated nanowire absorption. Photocurrent (red curve) was normalized by the source’s photon fl ux and matches well with the simulated

CdS-Cu 2 S core-shell nanowires absorption spectrum with similar dimensions (blue curve). Reprinted with permission from Reference 13.

©2011, Nature Publishing Group.



this way. 22 More recently, we have also investigated possible

photocathodes for CO2 reduction. For the CO2 reduction, we

chose GaP because CO 2 redox potential is more negative than

the hydrogen redox potential. For this purpose, we developed

a surfactant-free solution process to generate a large quantity

of single-crystalline GaP nanowires ( Figure 6 ).17 We can now

produce these GaP nanowires without an organic coating, and

we can control the doping by adding a zinc precursor as well.

We believe that a combination of Mo3 S 4 cubane molecular

catalysts and high surface area GaP nanowires can lead to a

very promising photocathode.

The anode, however, presents more of a problem, because

we have a very limited choice of meterial. The best candi-

date is Si-doped Fe 2O 3 from Michael Grätzel’s lab at the Ecole

Polytechnique de Lausanne in Switzerland, for which photo-

current is typically in the single digit mA/cm2 range.18 Our

group has been working on Inx Ga 1−x N nanostructures with ax

tunable bandgap from 3.3 eV to 1 eV. 19 We are hopeful that,

by using a tunable bandgap material to cover

the solar spectrum, we will be able to develop

a useful photoanode. The initial testing on the

photocurrent output is quite disappointing, in

the range of 50 μ A/cm 2 , far below that from

what we would expect based on their visible

bandgap.21 We have short carrier lifetime issues

and extraction issues here, and clearly these

InxGa 1−x N nanostructures need to be optimized x

for photoanode applications.

Finally, I will briefl y discuss the way we

plan to implement this two-bandgap confi g-

uration. We have two schemes: the fi rst is an

asymmetric Si-TiO2 core-shell model system;

the second is based on bilayer nanowire fabrics.

In the fi rst case, the Si/TiO 2 asymmetrical junction represents

a fully integrated cathode and anode in a single nanowire

(Figure 7); 20 Si can be decorated once again by the reduction

catalyst, and these arrays can be used directly for water split-

ting. This process is based on CVD; large reactors can be used

to make larger arrays. It is, however, quite energy intensive to

create these asymmetrical junctions. We are also investigating

silicon nanowires decorated by high surface area InGaN nano-

rods ( Figure 8 ),21 and this will be our model system for bal-

ancing the photocurrent from the anode and cathode. However,

contact between these two materials remains to be optimized.

We recently embarked on the concept of bilayer nanowire

fabrics for PEC applications; everything here is processed

in solution, and the fundamental idea is quite similar to the

asymmetric junction. Basically, we want to directly couple the

photoanode and photocathode, both made of nanowire mesh;

one side will perform the water oxidation, the other side will be

used for water reduction, and a proton transport medium will

Figure 5. Silicon nanowire array photocathode. (a) Digital photograph of the silicon

nanowire array on 2 cm × 2 cm wafers. (b) Cross-sectional scanning electron microscopy

images of the silicon nanowire arrays.

Figure 6. Large-scale, solution synthesis of semiconductor nanowires. (a) As-synthesized GaP nanowires in 1 L of hexadecane.

(b) A GaP nanowire membrane fabricated by fi ltration of water-dispersed GaP nanowires through a fi lter paper (white). After drying, the

nanowire membrane became free-standing. (c) Transmission electron microscopy image of the GaP nanowires. Reprinted with permission

from Reference 16. ©2011, American Chemical Society.



be placed in between. We can now make two-inch diameter discs

by laminating two layers of nanowire materials together, for

example, a WO 3 or TiO 2 nanowire mesh on top and a GaP

nanowire mesh underneath, where one is the photocathode

and one is the photoanode. Of course the bandgap and photo-

current output do not match at this point. We are trying to use

this as a model system to demonstrate whether such a bilayer

nanowire fabric idea can work. The initial results are promising.

The main problem right now is the urgent need to discover

better anode materials to replace TiO2 or WO3. Above all, we

believe that these high surface area semiconductor nanowires,

with their high carrier mobilities, will eventually be part of the

artifi cial photosynthesis system as these nanostructures enable the

stacking of the catalysts in the third dimension,

and effectively relax the stringent requirements

for catalysts with high turnover frequency.

Summary Over the past 10 years, we have continued to

develop the science and technology of semi-

conductor nanowires. As of today, nanowires

with different sizes, growth directions, com-

positions, and heterojunctions can be rationally

designed, synthesized, and assembled. We have

already seen major progress in many different

areas of nanowire research (e.g., electronics,

photonics); we can be confi dent that we will

continue to see many more fundamental new

discoveries and science based on this unique

class of nanoscale building blocks. Considering

their unique structural, optical, and electrical

properties, we expect that these nanostructures

will have a signifi cant impact on large scale

clean energy conversion and storage technolo-

gies. In the meantime, the future of nanowire

technology will be largely dependent on how

well we can balance the issues of cost, perfor-

mance, and stability of nanowire-based devices

and systems.

Acknowledgments I wish to thank my research group for their hard

work. I would also like to acknowledge the

generous continued support from the Depart-

ment of Energy, Offi ce of Basic Energy Sciences

over the last 10 years, without which much of

the progress I discussed would be impossible.

References 1. R.S. Wagner , W.C. Ellis , Appl. Phys. Lett. 4 , 89 ( 1964 ). 2. Y.N. Xia , P.D. Yang , Y.G. Sun , Adv. Mater. 15, 353 ( 2003 ). 3. P. Yang , C.M. Lieber , Sciencee 273, 1836 ( 1996 ). 4. A.M. Morales , C.M. Lieber , Sciencee 279 , 208 ( 1998 ).5. Y. Wu , P. Yang , J. Am. Chem. Soc. 123, 3165 ( 2001 ). 6. M. Huang , S. Mao , H. Feick , H. Yan , Y. Wu , H. Kind , E. Weber , R. Russo , P. Yang , Sciencee 292, 1897 ( 2001 ).7. M. Law , D. Sirbuly , J. Johnson , J. Goldberger , R. Saykally ,

P. Yang , Sciencee 305 , 1269 ( 2004 ). 8. Y. Nakayama , P.J. Pauzauskie , A. Radenovic , R.M. Onorato , R.J. Saykally , J. Liphardt , P. Yang , Naturee 447, 1908 ( 2007 ). 9. D.J. Sirbuly , M. Law , P. Pauzauskie , H. Yan , A.V. Maslov , K. Knudsen , R.J. Saykally , P. Yang , PNASS 102, 7800 ( 2005 ). 10. R. Yan , J. Park , Y. Choi , C. Heo , S. Yang , L.P. Lee , P. Yang , Nat. Nanotechnol.7 , 191 ( 2012 ). 11. D. Li , Y. Wu , P. Kim , L. Shi , N. Mingo , Y. Liu , P. Yang , A. Majumdar , Appl. Phys. Lett. 83 , 2934 ( 2003 ). 12. A.I. Hochbaum , R. Chen , R.D. Delgado , W. Liang , E.C. Garnett , M. Najarian , A. Majumdar , P. Yang , Naturee 451, 163 ( 2008 ). 13. M. Law , L. Greene , J.C. Johnson , R.J. Saykally , P. Yang , Nat. Mater. 4 , 455( 2005 ). 14. J. Tang , Z. Huo , S. Brittman , H. Gao , P. Yang , Nat. Nanotechnol. 6, 568 ( 2011 ). 15. A. Fujishima , K. Honda , Naturee 238 , 37 ( 1972 ).16. A.J. Nozik , Appl. Phys. Lett. 30, 567 ( 1977 ). 17. J. Sun , C. Liu , P. Yang , J. Am. Chem. Soc. 133, 19306 ( 2011 ).

Figure 7. Si/TiO 2 asymmetric core-shell nanowire array. (a) Optical image of asymmetric

nanowire array made using soft lithography, showing incandescent color due to array

periodicity. Scanning electron microscopy images of an asymmetric nanowire array (b) and

an individual nanowire (c). Reprinted with permission from Reference 19. ©2011, American

Chemical Society.

Figure 8. High surface area Si/InGaN nanowire photoanodes. Scanning electron

microscopy images of hierarchical Si/Inx Gax 1–x– N nanowire arrays on a Si (111) substratex

with X = 0.08 ∼ 0.1 (a). A fractured wire reveals the cross-section (b) showing that InGaN

nanowires grow vertically from the six Si wire facets. Reprinted with permission from

Reference 20. ©2012, American Chemical Society.



Ensure your voting privileges. Update your member

record/contact information at www.mrs.org today!

COMING SOON—Election of MRS

Officers & Board Members

ballot log-in information from

“Election-America for MRS”

will be posted as available at

www.mrs.org/elections-2012

Need Nano particles or coatings?Call 800-99ULVAC or

email [email protected].

science, fuel cell and automotive applications.

Extremely smooth ultra-thin filmsfi – 0.01– to0.3 nm/sec

Size-controlled nano particles – 1– nm to 30 nm dia.

Dense filmfi formation without process gas

Small target size: 10 mm dia. x 17 mm

Uniformity +/- 10% over 50 mm diametercoated area

Nano Particlesand Coatings

18. A. Kay,y I. Cesar,r M. Grätzel , J. Am. Chem. Soc. 128, 15717 (2006). 19. T. Kuykendall , P. Ulrich , S. Aloni, P. Yang, Nat. Mater. 6, 951 (2007).20. C. Liu, Y. Hwang, H.E. Jeong, P. Yang, Nano Lett.o 11 , 3755 (2011). 21. Y. Hwang, C. Wu , C. Hahn, H. Jeong , P. Yang , Nano. Lett. 12, 1678 ( 2012 ). 22. E. Garnett , P. Yang , Nano Lett.o 10, 1082 ( 2010 ). 23. D.J. Sirbuly,y M. Law, H. Yan, P. Yang, J. Phys. Chem. B 109 , 15190( 2005 ).

Peidong Yang is a professor inr the Department ofChemistry, Materials Science and Engineering,the S.K. and Angela Chan Distinguished Chairin Energy, and a senior faculty scientist at theLawrence Berkeley National Laboratory (LBNL).He received a BS degree in chemistry from theUniversity of Science and Technology of China(1993) and a PhD degree in chemistry fromHarvard University (1997). His postdoctoralresearch was at the University of California, SantaBarbara, after which he joined the Departmentof Chemistry faculty at the University ofCalifornia, Berkeley (1999). He founded thenanoscience subdivision within the American

Chemical Society (ACS) and also co-founded two startups, Nanosys Inc. andAlphabet Energy Inc. In addition, he is an associate editor for the Journal ofl thefAmerican Chemicaln Societyl andy serves on the editorial advisory boards for anumber of journals, including Accounts ofs Chemicalf Researchl andh Nano Letterso .sHis main research interest is in the area of one-dimensional semiconductornanostructures and their applications in nanophotonics and energy conversion.Yang is the recipient of numerous awards and recognitions. He can be reachedat the Department of Chemistry, University of California, Berkeley, MaterialsScience Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720,USA; email [email protected].


Introduction ZnO is one of mostf the important materials in materials

research today. ZnO nanowires have attracted worldwide

attention because of importantf implications in LED, biomed-

ical, solar cell, and electronics applications. Our research

has focused on this material, but we have ventured beyond

building nanostructures to building self-powered nanosys-

tems. Our research on ZnO started in 1999 and continues

today.1 From a structural point of view,f ZnO has a non-

central symmetric wurtzite crystal structure, which naturally

produces a piezoelectric effect once the material is strained.

Zn2+ cations and O2– anions are tetrahedrally coordinated,

and the centers of thef positive and negative ions overlap. If

a stress is applied at an apex of thef tetrahedron, the centers

of thef cations and anions are relatively displaced, resulting

in a dipole moment (Figure 1a). Polarization from all of thef

units results in a macroscopic potential (piezopotential) drop

along the straining direction in the crystal (Figure 1b).2 By

using this potential generated inside the crystal as an acting/

controlling field,fi two new fi eldsfi are created. Piezopotential-

driven transient fl owfl of electronsf in an external load is the

principle of thef nanogenerator (NG);3–5 piezopotential tuned/

controlled charge transport inside the crystal is the principle

behind piezotronics.6,7

From nanogenerators to piezotronics—A decade-long study of ZnO nanostructures Zhong Lin Wang

The following articleg is based ond the MRS MedalS Lecture,l presented byd Zhong Ling Wang on November

30, 2011, at thet 2011 Materials Research Society Fall Meetingl ing Boston. The MRS MedalS isl awarded

for ar specificfi outstanding recentg discoveryt or advancementr thatt hast a major impactr ont the progress

of af materials-related fid eld.fi Wang receivedg thed award ford “seminalr contributionsl in the discovery,

controlled synthesis,d and fundamentald understandingl ofg ZnOf nanowires and nanobelts,d and thed design

and fabricationd of novel,f nanowire-based nanosensors,d piezotronic devices, and nanogenerators.”d

Developing wireless nanodevices and nanosystems is critical for sensing, medical science,

environmental/infrastructure monitoring, defense technology, and even personal electronics.

It is highly desirable for wireless devices to be self-powered without using a battery. We have

developed piezoelectric nanogenerators that can serve as self-sufficientfi power sources for

micro-/nanosystems. For wurtzite structures that have non-central symmetry, such as ZnO,

GaN, and InN, a piezoelectric potential (piezopotential) is created by applying a strain. The

nanogenerator uses the piezopotential as the driving force, responding to dynamic straining

of piezoelectricf nanowires. A gentleA strain can produce an output voltage of upf to 20–40 V

from an integrated nanogenerator. Furthermore, piezopotential in the wurtzite structure

can serve as a “gate” voltage that can effectively tune/control charge transport across an

interface/junction; electronics based on such a mechanism are referred to as piezotronics,

with applications such as electronic devices that are triggered or controlled by force or

pressure, sensors, logic units, and memory. By using the piezotronic effect, we show

that optoelectronic devices fabricated using wurtzite materials can provide superior

performance for solar cells, photon detectors, and light-emitting diodes. Piezotronic devices

are likely to serve as “mediators” for directly interfacing biomechanical action with silicon-

based technology. This article reviews our study of ZnO nanostructures over the last

12 years, with a focus on nanogenerators and piezotronics.

Zhong Lin Wang, School of Materials Science and Engineering , Georgia Institute of Technology ; [email protected] DOI: 10.1557/mrs.2012.186

FROM NANOGENERATORS TO PIEZOTRONICS—A DECADE-LONG STUDY OF ZNO NANOSTRUCTURES


Growth of unique nanostructures The discovery of nanobelts,f ultralong ribbon-like nanostruc-

tures, in 2001 was the start of ourf researchr in oxide nanostruc-

tures.8 Since then, we have focused mainlyd on ZnO, which has

the most splendid nanostructure configurationsfi among all

known materials.9,10 The wurtzite structure of ZnOf is

unique for its non-central symmetry and polar surfaces.

The structure of ZnOf can be described as a number of alter-f

nating planes composed of tetrahedrallyf coordinated O2–

and Zn2+ ions, stacked alternatively along the c-axis. The

oppositely chargedy ionsd produce positively chargedy (0001)-Znd

and negatively charged (0001 )-O polar surfaces, resulting

in a normal dipole moment and spontaneous polarization

along the c -axis as well as a divergence in

surface energy. The electrostatic interaction

energy and distinct chemical activities of thef

polar surfacesr result int the formation of af wide

range of nanostructures,f such as nanosprings,11

nanorings,12 nanobows,13 and nanohelices14

(Figure 2 ).

ZnO nanobelts and the associated unique

nanostructures shown in Figure 2 were grown

using a vapor-solid process without a metal

catalyst.8 For a nanobelt dominated by the

polarized ±(0001) facets, a spontaneous

polarization is induced across the nanobelt

thickness due to the positive and negatived ionic

charges on the zinc- and oxygen-terminated

±(0001) surfaces, respectively. As a result,

a nanospring is formed when a single-crystal

nanobelt rollst up; this phenomenon isn attributed

to a consequencea of minimizingf the total energy

contributed by spontaneous polarization and

elasticity. Alternatively, a nanoringa can be initi-

ated byd circular foldingr of af nanobelta sot that thet

oppositely charged surfacesd meet face-to-face.t

Coaxial and uniradiald loop-by-loop winding of

the nanobelt formst a completea ring. Short-range

chemical bonding among the loops results in

a single-crystal structure. The self-coiling is

likely to be driven by minimizing the energy

contributed byd polar charges,r surface area, and

elastic deformation.8

Vertically aligned ZnO nanowire (NW)

arrays are probably the most importantt struc-t

ture for applications, such as solar cells,15

fi eldfi emission devices,16 UV lasers,17 light-

emitting diodes,g 18 and piezo-nanogenerators.d 2,19,20

Aligned growth of ZnOf nanorods has been

successfully achieved on a solid substrate

using a vapor-liquid-solid or vapor-solid-

solid process using gold nanoparticles as

catalysts,13,21 where the catalyst initiates and

guides growth, and the epitaxial orientation

relationship between the nanorods and the

substrate leads to aligned growth.22 The spatial distribution

of thef catalyst particles determines the pattern of thef grown

NWs. By choosing an optimum match between the substrate

lattice and the desired NWs, epitaxial orientation between

the NW and substrate results in aligned growth of NWsf nor-

mal to the substrate. The distribution of thef catalyst particles

definesfi the locations of thef NWs, and the epitaxial growth

on the substrate results in vertical alignment.

Solution based growthd of ZnOf is one of thef most power-t

ful low-cost, low-temperature, and large-scaled approaches for

aligned NWs.d 23–25 The growth temperatureh is around 80–100°C,d

so the substrate can ben any materialy with anyh shape.y This allows

a broad ranged of applications.f

Figure 1. Piezopotential in a wurtzite crystal. (a) Atomic model of thef wurtzite-structured

ZnO. (b) Scanning electron microscopy image of alignedf ZnO nanowire (NW) arrays

synthesized using a solution-based approach (left image). Numerically calculated

distribution of piezoelectricf potential along a ZnO NW under axial strain: compressive

strain (middle image) and tensile strain (right image). The growth direction of thef NW is

the c-axis. The dimensions of thef NW are L = 600 nm and a = 25 nm; the external force is

fyff =y 80 nN.



Nanogenerators Lateral nanowire based nanogenerators In today’s micro-/nanoscale devices, power consumption is

small. Medical sensing, remote patient monitoring, environ-

mental monitoring, long-range asset tracking, and national

security all require a large number and high density of sen-

sors. In this case, we would like sensors to be self-powered,

and the device could simply generate its own power from the

environment. This is called “self-powered nanotechnology.” 26

For this, we tried to utilize the functionality of ZnO. The goal of

our research is to use our daily activity as small-scale mechan-

ical action. In order to do this, we rely on the piezoelectric

effect. One of the unique advantages of ZnO is that all of the

NWs grow along the c-axis (i.e., in the polar direction) and are

uniaxially aligned. Once the NWs are strained, a macroscopic

piezopotential is created, which can drive a fl ow of electrons,

thus converting mechanical energy into electricity. We started

this work in 2005, using conducting atomic

force microscopy to deform a single NW. 1 We

have developed nanogenerators (NGs) that can

give an output of 50 V at an ideal peak power

density of ∼ 0.5 W/cm3 .27

A single NW laterally bonded on a fl exible

substrate can be used to illustrate the principle

of NG ( Figure 3 a).28 A Schottky contact on at

least one side of the contact is required for a

functioning NG. The entire wire is in an equi-

librium state without strain and power output

( Figure 3b ). When the wire is stretched, the

piezopotential created results in a difference in

the Fermi levels of the two contacts at the ends

of the wire ( Figure 3c ). The electrons in the

external circuit fl ow from the right-hand side to

the left-hand side to compensate for the energy

difference. The electrons cannot fl ow across

the interface due to the presence of a Schottky

barrier SB . The accumulation of the electrons at

the interface partially screen the piezopotential

built in the wire. When the electrons and the

piezoelectric fi eld reach equilibrium, the Fermi

energy on both sides is equal, and there is no

more current fl ow ( Figure 3d ). When the single

wire generator (SWG) is quickly released, the

polarization and piezoelectric fi eld vanish, and

the equilibrium with the accumulated electrons

is broken. The accumulated charge carriers fl ow

back from the left-hand side to the right-hand

side through the external circuit, producing

the second output signal in the opposite direc-

tion ( Figure 3e ). The energy band diagrams

( Figure 3b–e ) show the generation of a pair

of positive and negative voltage/current peaks

with a polarity as assumed for the case. This is

the process of the experimentally observed ac

output ( Figure 3f ). When the wire is cyclically

“fast stretched” (FS) and “fast released” (FR), output electricity

is obtained.

Integration of SWGs is a major step toward practical appli-

cations. We can transfer millions of NWs onto a single substrate

for improving the output. 29 For the structure shown in Figure 4 ,

the output voltage reached 2 V, and the current reached 100 nA.

Vertical nanowire based nanogenerators Using the as-grown, well-aligned NWs, high output NGs have

been fabricated. Figure 5 a shows a new approach we devel-

oped recently for fabricating a high-output, low-cost NG. 30

The entire structure is based on a polystyrene (PS) substrate

of typical thickness 0.5 mm, on which a Cr adhesion layer is

deposited. After sputtering a layer of ZnO seed, densely packed

ZnO NWs are grown on the seed layer as a quasi-continuous

“fi lm” using a solution based growth technique at a tempera-

ture no more than 100°C. Finally, a thin layer of poly(methyl

Figure 2. A collection of novel ZnO nanostructures that were formed due to the existence

of ±(0001) polar surfaces.1



methacrylate) (PMMA) was deposited on top of the fi lm to

serve as an isolation layer, followed by deposition of a thin

gold layer as an electrode. The growth of ZnO NWs has a

unique and distinguishing feature in that they

are aligned along the c -axis, so that the entire

fi lm has a common polar direction. Once the

PS substrate is mechanically bent/deformed

through the substrate, the fi lm at the top surface

is under tensile strain, and the one at the bottom

surface is under compressive strain, resulting

in a piezopotential difference between the

top and bottom electrodes, which drives

the fl ow of electrons in the external load. A

cycling mechanical deformation results in the

back-and-forth fl ow of electrons in response

to the mechanical triggering. By introducing a

strain of 0.12% at a strain rate of 3.56 % S –1 ,

the measured output voltage reached ∼20 V,

and the output current exceeded 8 μ A (corre-

sponding to an ideal maximum power density

of 0.16 W/cm 3 ) ( Figure 5b–c ).30 The total output

can be enhanced by integrating multiple NGs in

series or parallel depending on the application

so that the entire system can be placed, for

example, in shoes, clothes, plastic sheets, and

rotating tires. The advantage of using NWs

is that they can be triggered by tiny physical

motions, and the excitation frequency can be one

Hz to thousands of Hz, which is ideal for harvest-

ing random energy in the environment such as

tiny vibrations, body motion, and gentle air fl ow.

In some cases, alignment of the NWs is not

essential for generating electricity. Using the

conical shape of the ZnO NWs, a ZnO–polymer

composite can also be used to generate electricity. 31

Such an idea can be extended to a general com-

posite between a polymer and NWs, such as

NaNbO3.32 Applying a polarization perpendic-

ular to the composite fi lm can produce a mac-

roscopic piezopotential in the fi lm, which then

generates electricity.

Self-powered nanosystems A nanosystem is an integration of multifunc-

tional nanodevices. The power required to drive

such electronics is in the microwatt to milli-

watt range. It is possible to have self-powered,

maintenance-free biosensors, environmental

sensors, nanorobotics, microelectromechanical

systems (MEMS), and even portable/wearable

electronics.33 An NG was fabricated using

aligned NW arrays (To view supplementary

material for this article, please visit http://dx.doi.

org/10.1557/mrs.2012.186.), which powered a

pH sensor made of a single NW. To demonstrate

the operation of an NG driven self-powered system ( Figure 6a),

we used a single transistor radio frequency (RF) transmitter to

send out a detected electric signal. 34,35 The oscillation frequency

Figure 4. (a) High output nanogenerator (NG) obtained by using the integration of millions of

nanowires transferred to a polymer substrate following the approach presented in Figure 3a .

Scanning electron microscopy images before (far left) and after (right) depositing Au

electrodes and corresponding schematic and real NG device (far right). (b) The packaged

NG gives an output voltage of 2 V (left) and current of 100 nA (right). The insets are enlarged

signals from one cycle of deformation taken from the areas inside the red rectangles. 33

Figure 3. Charge generation and output mechanism of an alternating-current single wire

nanogenerator illustrated using an energy band diagram. (a) Schematic of a ZnO wire

contacted with two metal electrodes, at equilibrium (upper) and stretched (lower).

Energy band diagrams of the device when the wire is at equilibrium (b), tensile

stretched (c), returns to equilibrium (d), and then released (e). In the lower portion of each

is a sketch of the measurement circuit, where a small load resistor RL is introduced, which

is much smaller than the resistance of the wire and/or the contacts. (f) An experimentally

measured output current from the nanogenerator, in which the output regions corresponding

to the processes shown in (b–e) are indicated by the corresponding labels (b–e).

(g) Experimentally measured output voltage and current from the nanogenerator by

repeated deformation. 32 FS, fast stretched; FR, fast released; VB, valence band; CB,

conduction band.



was tuned to be around 90 MHz, and a com-

mercial portable AM/FM radio was used to

receive the transmitted signal. The entire system

is made of the NG, a capacitor for energy storage

(to regulate output power), the sensor signal

modulator, and the wireless rf data transmitter.

For demonstrating the synchronization between

the sensed signal and the signal transmitted,

a phototransistor in a slotted optical switch was

added to the system as the photon detecting

sensor to demonstrate that the self-powered

system can work independently and wire-

lessly. The signal of the photocurrent generated

by the phototransistor as a result of external

light excitation was periodically sent out using

the energy stored in the capacitor. Each time it

was triggered, the signal received by the pho-

totransistor modulated the transmitting signal,

and the information was received by the radio,

and the demodulated signal was recorded from

the headphone jack. Each cycle included an on

(16 ms)/off (5 ms)/on (5 ms)/off (10 ms) status

sequence. Figure 6b is the signal demodulated

by the radio. When the phototransistor and the

transmitter were triggered, there was a pulse

detected beyond the noise background. When

we enlarged this pulse, it contained a segment

of the information that had the same waveform

envelope as the triggering voltage sequence of

the LED, as shown in Figure 6c . This indicates

that the wireless data transmission was achieved

by using this self-powered system over a distance

of ∼ 10 m. 34,35

Hybrid cell for simultaneously harvesting multiple types of energies Our environment has an abundance of energy

forms, including light, thermal, mechanical (such

as vibration, sonic waves, wind), magnetic, chem-

ical, and biological. Harvesting these types of

energies is of critical importance for long-term

energy needs and sustainable development.

Innovative approaches have to be developed

for conjunctional harvesting of multiple types of

energies using an integrated structure/material so

that the energy resources can be effectively and

complimentarily utilized whenever and wherever

one or all of them are available. We initiated an idea

in 2009 for harvesting multiple types of energy

using a single device structure, known as a hybrid

cell (HC) ( Figure 7a).36 The structure is based

on vertical ZnO NW arrays but with the addi-

tion of a solid electrolyte and a metal coating. 37

The solar cell open circuit voltage ( UOC-SCUU ) was

Figure 5. Fabrication of a high output nanogenerator using vertically aligned nanowire

arrays/fi lms grown on two sides of a polymer fi lm. (a) Fabrication process of the nanogenerator.

The lower right part is a photo of a fabricated nanogenerator after packaging. The

bending of the nanogenerator shows good mechanical fl exibility. (b) Output current

and (c) output voltage of a typical nanogenerator.31

Figure 6. (a) Schematic diagram of an integrated self-powered system that can be divided

into fi ve modules: energy harvester, energy storage, sensors, data processor and controller,

and data transmitter and receiver. (b) The voltage sequence used to trigger the sensor,

representing a waveform envelope of the input signal. (c) Recorded signal from the

headphone jack of the radio located 10 m away, showing the technical feasibility of a

wireless self-powered system. 34, 35



0.42 V, and the short circuit current density (J SC-SCJJ ) was

0.25 mA cm –2 . The NG was characterized by introducing

ultrasonic waves through a water medium without sun-

light illumination; the corresponding J–V curve shows that V

UOC-NGUU was ∼0.019 V, and INGII was ∼0.3 pA cm –2. When

only simulated sunlight shines on the HC, the dye-sensitized

solar cell (DSSC) worked ( Figure 7b ), and the optimum

output power density was found to be 32.5 μ Wcm–2 at

JSCJJ = 140 μ Acm–2mm and UOCUU = 0.231 V. When both the DSSC and

NG were simultaneously operating in series, the corresponding

output power density was 34.5 μ Wcm –2mm at J SCJJ = 141 μ Acm –2mm and

UOCUU = 0.243 V. After the ultrasonic wave was turned on, power

density increased (∆PHC ) by 2 μ Wcm –2mm , which represented more

than a 6% enhancement in optimum power output. Therefore,

in addition to enhancing the open circuit voltage, the HC suc-

cessfully added the total optimum power outputs from both the

solar cell and the NG.

With the development of modern medical technology, pow-

ering implantable nanodevices for biosensing using energy

harvesting technology has become a challenge. We developed

a HC for harvesting mechanical and biochemical energies 38

mainly for biomedical applications. The structure is based on

an integrated NW-based NG system and an enzyme-based

biofuel cell (BFC). In this hybridized design, we used piezo-

electric poly(vinylidene fl uoride) (PVDF) nanofi bers (NFs) as

the working component for mechanical energy harvesting. The

working principle of the PVDF NG is based on the piezoelectric

properties of the PVDF NF. As the device is deformed under

alternating compressive and tensile force, the NFs drive a fl ow

of electrons back and forth through the external circuit. 17 The

enzymatic BFC was used to convert the chemical energy of glu-

cose and oxygen in the biofl uid into electricity. The electrodes

were patterned onto a Kapton fi lm and coated with multiwalled

carbon nanotubes, and fi nally immobilized glucose oxidase

(GOx) and laccase form the anode and cathode, respectively,

for the BFC. A HC structure can be fabricated on a single PVDF

NF for energy harvesting. 39

Nanogenerators as active sensors A nanogenerator can also function as an active sensor by using

its electric output as the signal to be detected. Based on such

an idea, we recently demonstrated a self-powered sensor for

detecting low frequency vibrations. 40 Furthermore, we fabri-

cated a self-powered pressure sensor based on the BFC and

NG on a single fi ber,41 as shown in Figure 8 a. ZnO NW fi lms

grown around a carbon fi ber forms a textured fi lm with the

c -axis radially pointing outward. Mechanical straining would

generate a piezopotential across the thickness of the NW fi lm

( Figure 8b ). Thus, the output of the NG is sensitive to the

pressure change. This experiment demonstrates that not only

can we use the HC or NG as an energy harvester, but also as

an active sensor for detecting a mechanical signal from the

environment.

Piezotronics In order to illustrate the basic concept of piezotronics, we fi rst

start from a traditional metal oxide semiconductor fi eld-effect

transistor (MOSFET). For an n-channel MOSFET ( Figure 9a),

the two n-type doped regions are the drain and source; a thin

insulator oxide layer is deposited on the p -type region to serve

as the gate oxide, on which a metal contact is made as the

gate. The current fl owing from the drain to the source under an

applied external voltage VDSVV is controlled by the gate voltage VGVV ,

which controls the channel width for transporting the charge

carriers. A piezotronic transistor is a metal-NW-metal structure,

as schematically shown in Figure 9b once a strain is applied

through the substrate. 41 The fundamental principle of the piezo-

tronic transistor is to control the carrier transport at the metal-

semiconductor interface through tuning at the local contact by

creating a piezopotential at the interface region in the semicon-

ductor by applying a strain. This structure is different from the

complementary metal oxide semiconductor (CMOS) design.

First, the externally applied gate voltage is replaced by an inner

crystal potential generated by the piezoelectric effect, thus, the

“gate” electrode is eliminated. This means that the piezotronic

transistor only has two leads: drain and source. Second, control

over channel width is replaced by control at the interface. Since

Figure 7. Design of a compact hybrid cell (HC) structure

consisting of a dye sensitized solar cell (DSSC) and a

nanogenerator (NG). (a) Schematic illustration of the HC,

illuminated by sunlight from the top and excited by ultrasonic

waves from the bottom. The ITO layer on the DSSC and the

GaN substrate are defi ned as the HC cathode and anode,

respectively; Spiro-MeoTaD is a polymer. (b) A comparison of

power output J–V characteristics of a HC. The top left inset is theVI–V characteristic of the NG when the solar cell is off. The bottomVright inset is a replot of the displayed curves at short circuit. 37



the current transported across an M-S interface

is the exponential of the local barrier height in

the reversely biased case, the ON and OFF ratio

can be rather high due to the non-linear effect.

Finally, a voltage controlled device is replaced by

an external strain/stress controlled device, which

is likely to have applications complementary to

CMOS devices. The initial idea of piezotronics

was fi rst described in 2006 42 and 2007.43,44

When a ZnO NW device is under strain,

there are two typical effects that may affect the

carrier transport process. One is the piezore-

sistance effect 45,46 due to changes in band gap,

charge carrier density and possibly density of

states in the conduction band of the semicon-

ductor crystal under strain. This effect is a sym-

metric effect on the two end contacts and has no

polarity, which will not produce the function of

a transistor. The other is the piezotronic effect

due to the polarization of ions in a crystal that

has non-central symmetry; this has an asymmet-

ric or non-symmetric effect on the local contacts

at the source and drain owing to the polarity

of the piezopotential. In general, the negative

piezopotential side raises the barrier height at

the local contact of a metal/n-type semicon-

ductor, possibly changing an Ohmic contact to

a Schottky contact, or a Schottky contact to an

“insulator” contact; while the positive piezopo-

tential side lowers the local barrier height, changing a Schottky

contact to an Ohmic contact. But the degree of change in the

barrier height depends on the doping type and doping density in

the NW. The piezoelectric charges are located at the ends of the

wire, thus they directly affect the local contacts. The piezotronic

effect is a common phenomenon for the wurtzite family, such

as ZnO, GaN, CdS, and InN. It is important to point out that

the polarity of the piezopotential can be switched by changing

tensile strain to compressive strain ( Figure 9c ). Thus, the device

can be changed from control at source to control at drain simply

by reversing the sign of the strain applied to the device.

Piezotronic eff ect on metal-semiconductor contact When a metal and an n-type semiconductor form a contact, a

Schottky barrier (SB) ( e SB ) is created at the interface if the

work function of the metal is appreciably larger than the elec-

tron affi nity of the semiconductor ( Figure 10 a). If the semi-

conductor also exhibits the piezoelectric effect, a strain in the

structure would produce piezo-charges at the interfacial region.

It is important to note that the polarization charges are distrib-

uted within a small depth from the surface, and they are ionic

charges, which are non-mobile charges located adjacent to the

interface. In such a case, free carriers can only partially screen

the piezo-charges due to the fi nite dielectric permittivity of the

crystal and the limited doping concentration, but they cannot

completely cancel the piezo-charges. The piezo-charges may

Figure 8. Integration of a nanogenerator and a biofuel cell along a single fi ber as a

hybrid cell or a self-powered pressure sensor. (a) Schematic diagram of the design. Inset

scanning electron microscopy images show the ZnO nanowire fi lm. The right-hand side

shows the principle of the fi ber-based nanogenerator (top) along with a photograph of the

as-fabricated device (bottom). (b) Pressure measurement in a bio-liquid using a hybrid

nanogenerator as a self-powered system. The response of the hybrid nanogenerator

system to periodically applied pressure is shown on the left. Plot of the statistically

measured steady current from left as a function of applied pressure, showing a linear

relationship, is on the right. 41

Figure 9. Schematic of (a) an n -channel MOSFET and

(b) a semiconductor nanowire fi eld-effect transistor. Schematic

of a piezotronic transistor with tensile strain (b) and compressive

strain (c), where the gate voltage that controls the channel

width is replaced by a piezopotential that controls the transport

across the metal-semiconductor interface. The colors represent

the distribution of piezoelectric potential inside the nanowire:

red is positive potential, and blue is negative potential. 41



produce mirror charges at the metal side. The positive piezo-

charges may effectively lower the barrier height at the local

Schottky contact, while the negative piezo-charges increase the

barrier height ( Figure 10b–c ). The role played by the piezopo-

tential is to effectively change the local contact characteristics

through an internal fi eld depending on the crystallographic

orientation of the material and the sign of the strain, thus, the

charge carrier transport process is tuned/gated at the M-S

contact. 9,47 Therefore, the charge transport across the interface

can be largely dictated by the created piezopotential, which

is the gating effect. This is the core of the piezotronic effect.

Piezotronic eff ect on p-n junctions When p-type and n-type semiconductors form a junction, the

holes in the p-type side and the electrons in the n -type side tend

to redistribute to balance the local potential; the interdiffusion

and recombination of the electrons and holes in the junction

region form a charge depletion zone. The presence of such a

carrier free zone can signifi cantly enhance the piezoelectric

effect, because the piezo-charges will mostly be preserved

without being screened by local residual free

carriers. As shown in Figure 10d , in the case

that the n-type side is piezoelectric and a strain

is applied, local net negative piezo-charges are

preserved at the junction provided the doping is

relatively low such that the local free carriers do

not fully screen the piezo-charges. The piezopo-

tential tends to raise the local band slightly and

introduce a slow slope to the band structure. 9

Alternatively, if the applied strain is switched

in sign ( Figure 10e ), the positive piezo-charges

at the interface creates a dip in the local band.

A modifi cation in the local band may be effective

for trapping the holes so that the electron-hole

recombination rate may be largely enhanced,

which is very benefi cial for improving the effi -

ciency of an LED. 48 Furthermore, the inclined

band tends to change the mobility of the carriers

moving toward the junction.

Piezotronic transistor A piezotronic transistor is made of a single ZnO

NW with its two ends, the source and drain

electrodes, fi xed by silver paste on a polymer

substrate ( Figure 11 a). Once the substrate is

bent, a tensile/compressive strain is created in

the NW, since the mechanical behavior of the

entire structure is determined by the substrate.

Utilizing the piezopotential created inside the

NW, the gate input for a NW strain-gated tran-

sistor (SGT) is an externally applied strain

rather than an electrical signal. 49 IDSII -VDSVV char-

acteristics for each ZnO-NW SGT are obtained

as a function of the strain created in the SGT

( Figure 11b ). Positive/negative strain is created

when the NW is stretched/compressed. The SGT behaves simi-

larly to an n-channel enhancement-mode MOSFET, apparently

indicating the working principle of the SGT.

The working principle of a piezotronic transistor is illus-

trated by the band structure of the device. 33 A strain free ZnO

NW may have Schottky contacts at the two ends with the source

and drain electrodes ( Figure 11c ). A piezopotential drop from

V+VV to V –VV is created along the NW once it is subjected to mechan-–

ical straining, which reduces and increases the local Schottky

barriers ( Figure 11d ), respectively. This asymmetric effect on

the local contacts is characteristic of the piezotronic effect.

A change in the sign of strain results in a reversal of the piezo-

potential, thus, the polarity of the transport characteristic is

changed ( Figure 11e ).

Based on the piezotronic transistor described previously,

universal logic operations such as inverters, NAND, NOR,

and XOR gates have been demonstrated for performing piezo-

tronic logic calculations, 35 which have the potential to be inte-

grated with MEMS technology for achieving advanced and

complex functional actions. This is an outstanding example

Figure 10. (a–c) Energy band diagrams illustrating the effects of piezoelectric polarization

charges on a Schottky contacted metal-semiconductor interface with and without

the presence of nonmobile, ionic charges at the metal-semiconductor interface. The

piezoelectric charges are indicated at the interface. (d–f) Energy band diagrams illustrating

the effect of piezoelectricity on a p-n junction made of two materials with similar bandgaps.

The distorted band in the presence of piezoelectric charges is indicated by red lines.58 VB,

valence band; CB, conduction band.



of using mechanical straining to generate and

control digital calculation. Furthermore, using

the memristor effect of ZnO, a piezoelectrically

modulated resistive memory cell 50 has also been

fabricated.

We have also developed the theoretical

framework of piezotronics by studying charge

transport across metal-semiconductor contacts

and p-n junctions with the introduction of a

piezopotential.27 In addition to numerical cal-

culations, we derived analytical solutions under

simplifi ed conditions and found that for an M-S

contact, the current to be transported can be

tuned or controlled by not only the magnitude

of the strain, but also by the sign of the strain

(tensile versus compressive).

Traditionally, a transistor is a device that

uses an external voltage to control transport cur-

rent, and it can be utilized to magnify the input

electrical signal. In other words, such a device

uses an electrical signal to tune/control another

electrical signal. In the case of the piezotronic

transistor, we use strain to control electrical

signal rather than amplifying it.

Piezo-phototronics If photon excitation is applied at a M-S contact,

the newly generated electrons in the conduction

band tend to move away from the contact, while

the holes tend to move close to the interface

toward the metal. The accumulated holes at the

interface modify the local potential profi le so

that the effective height of the Schottky barrier

is lowered (Figure 12a), which then increases

the conductance.

On the other hand, the presence of negative

piezo-charges as a result of applying strain can

effectively reduce the conductance, which is

a result of raising the Schottky barrier height

(SBH). Therefore, laser excitation and the

piezoelectric effect can be applied together for

controlling charge transport at the interface.

This coupling between piezoelectricity and

photon excitation 51 is the piezo-phototronic

effect, which uses the piezopotential for effec-

tively controlling carrier generation, separation,

transport, and/or recombination in optoelec-

tronic processes. 9,10 We now use three examples

to illustrate the piezo-phototronic effect in a

photodetector, solar cell, and LED.

Piezo -phototronic effec t in a photodetector The basic principle of a photon detector is

based on the photoelectric effect, in which the

Figure 11. Piezotronic transistor. (a) Schematic of an Ag-ZnO-Ag strain gated

transistor on a fl exible substrate. The deformation in the transistor is indicated through

a change in shape of the substrate. (b) Changes in transport characteristics of a

Ag/ZnO-nanowire/Ag device from symmetric I–V characteristics (black) to asymmetricVrectifying behavior when stretching (red) and compressing (green) the wire. Inset:

equivalent circuit models of the device in corresponding to the observed I–V curves, Vdifferent sizes of diode symbols are used to illustrate the asymmetric Schottky

contacts at the two ends of the nanowire. (c–e) Band diagrams for illustrating the

piezotronic effect on local contact of the device without and with the presence of

piezopotential. The sign of the piezopotential is indicated in the nanowire. 49

Figure 12. Enhancing photon detection sensitivity by the piezo-phototronic effect.

(a) Schematic band diagrams for illustrating the effect of local piezo-charges on the

Schottky contact showing the piezotronic effect on separating the photon induced

charges. (b) Response of a ZnO wire UV detector (in units of A/W) as a function of strain

under different excitation light intensity on a natural logarithmic scale. 55



e-h pairs generated by a photon are separated

by either a p-n junction or a Schottky bar-

rier ( Figure 12a ). If the Schottky barrier is too

high, the holes will be trapped at the semicon-

ductor side so that they cannot be effectively

annihilated by the free electrons in the metal

( Figure 12a ), thus reducing the photocurrent.

If the Schottky barrier is too low, the photon-

generated electrons cannot be effectively driven

away from the interface region, so they can be

easily recombined with the holes, which also

results in low photocurrent. An optimization of

SBH can give the maximum photocurrent. Such

a result was observed in a simple photocell.52

By tuning the SBH in a ZnO wire-based

UV sensor through applying a strain, we can

improve the sensitivity of the UV detector,

even when the illumination intensity is rather

weak.53 The response of the photodetector is

enhanced by 530%, 190%, 9%, and 15% upon

4.1 pW, 120.0 pW, 4.1 nW, and 180.4 nW UV

light illumination, respectively, onto the wire by

introducing a –0.36% compressive strain in the

wire ( Figure 12b ); this effectively tunes the SBH at the contact

by the local piezopotential produced. The sensitivity for weak

light illumination is especially enhanced by introducing strain,

although the strain has little effect on the sensitivity to stronger

light illumination. Our results show that the piezo-phototronic

effect can enhance the detection sensitivity more than fi vefold

Figure 14. Piezo-phototronic in a GaN/ZnO LED. Enhancement of emission light intensity and conversion effi ciency of a ( n( ( -ZnO wire)-

( p( ( -GaN fi lm) LED under applied strain. (a) CCD images recorded from the emitting end of a packaged single wire LED under different

applied strains. (b) Integrated emission light intensities from the data shown in (a), showing a huge increase in the emission intensity with

the increase of the applied compressive strain. The inset is the injection current of the LED at 9 V bias voltage with increase in strain. (c)

Schematic energy band diagram of the p-n junction without (upper) and with (lower, red line) applied compressive strain, where the channel

created at the interface inside ZnO is due to the piezopotential created by strain. The slope of the red line in the lower image at the ZnO

side represents the driving effect of the piezopotential on the movement of the charge carriers.50

Figure 13. Piezo-phototronic effect in a solar cell. (a) Schematic of a fabricated [0001] type

device and top-down optical image of a device. (b) Dependence of short-circuit current Isc

and the open-circuit voltage VocVV on applied strain. (c) The piezopotential distributions in the

stretched device of [0001] type, and the corresponding (d) band diagram of P3HT/ZnO with

the presence of negative piezoelectric charges. The blue line indicates the energy band

diagram modifi ed by the piezoelectric potential in ZnO.55



for pW levels of light detection. This conclusion also holds

for visible light.54

Piezo-phototronic eff ect in a solar cell A solar cell uses photon-generated charges, in which the charge

separation is critical for conversion effi ciency. We now describe

a solar cell based on ZnO- poly(3-hexylthiophene) (P3HT) as an

example to illustrate the piezo-phototronic effect on the output 55

(Figure 13 a). ZnO micro/nanowires have a wurtzite structure

and grow along the [0001] direction. The short-circuit current

IscII and VocVV under different strains are shown in Figure 13 b. VocVV

increases and decreases with increasing the compressive and

tensile strains, respectively. However, IscII shows a relatively

constant value of 0.035 nA under the different strains.

To explain the observed phenomenon, we calculated the

distribution of piezopotential in a single ZnO wire. When

this device is under tensile strain ( Figure 13 c), the negative

piezopotential is in contact with the P3HT. The negative piezo-

electric polarization charges at the interface lift the local con-

duction band level of ZnO, which can result in a decrease of

∆E∆∆ and E VocVV (∆ E∆ ∆ −EE as shown in Figure 13d ).

Piezo-phototronic eff ect in an LED Effective charge recombination is essential for an LED.

The piezo-phototronic effect can be used to enhance LED

output.34 Our experiments were carried using

a n-ZnO/ p/ -GaN device. A normal force was

applied perpendicular to the p-n junction inter-

face, which produces a tensile strain along the

c -axis in the ZnO microwire. At a fi xed applied

bias above the turn-on voltage (3 V), the cur-

rent and light emission intensity increased with

increasing compressive strain ( Figures 14a–b).

The injection current and output light inten-

sity were enhanced by a factor of 4 and 17,

respectively, after applying a 0.093% a- axis

compressive strain, indicating that the conver-

sion effi ciency was improved by a factor of 4.25

compared to applying no strain. This means that

the external true effi ciency of the LED can reach

∼ 7.82% after applying a strain.

The enhanced LED effi ciency is due to

the piezo-phototronic effect. 6,7 Under an

assumption of no-doping or low-doping in

ZnO for simplicity, the numerically calcu-

lated piezopotential distribution in the ZnO

microwire shows ( Figure 14c ) that a nega-

tive potential drop is created along its length

when the ZnO microwire is under a- axis

compressive strain. The fi nite doping in the

wire may partially screen the piezoelectric

charges, but it cannot totally eliminate the

piezoelectric potential if the doping level is

low, thus a dip in the band is possible. The

depletion width and internal fi eld decrease

under this additional component of forward biased voltage.

Subsequently, the injection current and emitting light intensity

under the same externally applied forward voltage increase

when the device is strained.

Theory of piezo-phototronics We have developed the theoretical frame of piezo-phototronics

by studying photon emission at the p-n junction and the photon

detector with the presence of local piezoelectric charges. 56,57

The analytical results under simplifi ed conditions were derived

for understanding the core physics of the piezo-phototronics

devices, and the numerical model was developed for illustrating

the photon emission process and carriers transport characteris-

tics of the piezoelectric LED in a practical case. Furthermore,

the theory for the piezo-phototronic effect on solar cell output

has also been developed. 58

Summary and future perspectives We have devoted over 12 years to studying ZnO nanostructures.

Our systematic studies in the fi eld can be summarized using the

analogy of a tree structure, as shown in Figures 15. The fun-

damental “root” of all the science and technologies developed

is twofold: (1) the fundamental physics is the piezopotential

generated in the inner crystal and the semiconductor properties

of the materials themselves; (2) the basic materials systems

Figure 15. A “tree” representing the fi elds of nanogenerators, hybrid cells for harvesting

multiple types of energies, 36,38,39,60 self-powered systems, piezotronics, piezo-phototronics,

and possibly piezophotonics that have been developed by our group in the last decade.

The fundamental “root” of all these fi elds is piezopotential and semiconductor as the basic

physics, and ZnO as the fundamental materials system. All of the fi elds (branches) are

derived from these roots.10



are wurtzite structures, such as ZnO and GaN. The “branches”

are the fi elds we have developed in the last decade; and the

“fruits” are the important near future applications. Further-

more, fundamental physics can be developed by introducing a

strain-tuned charge at the interface, and it is possible to study

nanoscale piezo-physics and related concepts, which remain

to be investigated.

Developing self-powered nanodevices and nanosystems is

a step toward going beyond Moore’s law. The prototype tech-

nology established by the nanogenerator (NG) sets a platform

for developing self-powered nanotechnology with important

applications in implantable in vivo biosensors, wireless and

remote sensors for environmental monitoring and infrastruc-

ture monitoring, nanorobotics, microelectromechanical systems

(MEMS), and personal electronics. We have devoted seven

years to developing the NG. For one layer of ZnO nanowires

(NWs), the output voltage has now reached 50 V, a current of

120 uA (area size 3 × 3 cm2 ), and an ideal peak power of

∼ 0.5 W/cm 3. The near-term goal for NG is to continuously

improve the power output and integration with other technology

for practical applications.

Energy generation and energy storage are two distinct pro-

cesses that are usually accomplished using two separated units

designed based on different physical principles, such as a piezo-

electric nanogenerator and a Li-ion battery; the former converts

mechanical energy into electricity, and the latter stores electric

energy as chemical energy. Recently, we introduced a method

that directly hybridizes the two processes into one, using the

mechanical energy that is directly converted and simultaneously

stored as chemical energy without going through the intermedi-

ate step of fi rst converting it into electricity. 61 By replacing the

polyethylene separator, as in a conventional Li battery with a pie-

zoelectric PVDF fi lm, the piezoelectric potential from the PVDF

fi lm, as created by mechanical straining, acts as a charge pump to

drive Li ions to migrate from the cathode to the anode accompa-

nied by charging reactions at electrodes. This new approach can

be applied to fabricate a self-charging power cell for sustainable

driving of micro/nano-systems and personal electronics.

Figure 16. Schematic diagram showing the three-way coupling among piezoelectricity, photoexcitation, and a semiconductor, which is the

basis of piezotronics (piezoelectricity-semiconductor coupling), piezophotonics (piezoelectric-photoexcitation coupling), optoelectronics,

and piezo-phototronics (piezoelectricity-semiconductor-photoexcitation). Potential applications of piezotronics and piezo-phototronics and

nanogenerators are projected, which are important future directions for research and applications.10



Piezotronics uses piezopotential as the gate voltage for

tuning the charge carrier transport processes at an M-S contact

or p-n junction. The design of piezotronics may fundamen-

tally change the design of traditional fi eld-effect transistors

by eliminating the gate electrode, replacing the externally

applied gate voltage with an internally created piezopotential,

and controlling the transport of charges through the contact at

the drain (source)-NW interface rather than the channel width.

Piezotronics can be used with silicon-based complementary

metal oxide semiconductor (CMOS) technology, because it

can be integrated on a polymer substrate for fabricating active,

fl exible electronics. Silicon technology provides the speeds and

density of devices, while piezotronics provides the function-

ality required for human-CMOS interfacing. Piezotronics has

the potential to serve as a mechanosensation in physiology, 59

which is about the mechanical stimulations of senses (touch,

hearing and balance, and pain), to convert mechanical stimuli

into neuronal signals; the former is a mechanical actuation, and

the latter is electrical stimulation. Piezotronics has potential

applications in human-Si technology interfacing, smart MEMS,

nanorobotics, and sensors (Figure 16 ).

Piezo-phototronics is a result of a three-way coupling

among piezoelectricity, photonic excitation, and semiconductor

transport, which allows for the tuning and controlling of carrier

generation, separation, transport, and/or a recombination of

charge carriers in optoelectronic processes by strain-generated

piezopotential. The development of this fi eld will have great

impact on LEDs, photodetectors and solar cells fabricated

using wurtzite and other materials ( Figure 16 ).The piezotronics

and piezo-phototronics were invented for such purposes,

and they are considered to be active fl exible electronics or

bio-driven electronics.

Moore’s law has been the roadmap that

directs and drives information technology in

the last few decades (see the vertical axis in

Figure 17 ). Sensor networks and personal

health care have been predicted as major

driving forces for industry in the near-term

future. As we have observed in today’s elec-

tronic products, electronics are moving toward

personal electronics, portable electronics, and

polymer-based fl exible electronics. We are

looking for multifunctionality and diversity

associated with electronics. The drive for tech-

nology in the last half century is miniaturi-

zation and portability/mobility. For example,

having a super-fast computer in a mobile

phone may not be the major driver for future

markets, but consumers are looking for more

functionality, such as healthcare sensors for

blood pressure, body temperature, and blood

sugar level, and interfacing with the environ-

ment using sensors for detecting gases, UV,

and hazardous chemicals. In such a case, IT

is developing along another dimension, as

presented in the horizontal axis in Figure 17 . The near future

development of electronics is moving toward integrating

electronics with multifunctional sensors and self-powered

technology. The goal is to directly interface humans with

the environment in which we live in. A combination of CPU

speed, density of memory and logic, along with functionality,

tends to drive electronics toward smart systems and self-

powered systems, which are believed to be the direction for

near-term electronics.

Acknowledgments This research was supported by DARPA, NSF, BES DOE,

NIH, NASA, US Airforce, MANA, NIMS (Japan), Samsung,

Chinese Academy of Sciences, and Georgia Tech. Thanks

to the contributions made by my group members (not in

any particular order): Yong Ding, Puxian Gao, Jinhui Song,

Xudong Wang, Rusen Yang, Jun Zhou, Yong Qin, Sheng

Xu, Zhengwei Pan, Zurong Dai, Will Hughes, Jin Liu,

Yifan Gao, Jr-Hau He, Ming-Pei Lu, Jung-il Hong, Chen

Xu, Yaguang Wei, Wenzhuo Wu, Youfan Hu, Yan Zhang,

Qing Yang, Weihua Liu, Yifeng Lin, Minbaek Lee, Peng Fei,

Ying Liu, Chi-Te Huang, Tei-Yu Wei, Ben Hansen, Caofeng

Pan, Guang Zhu, Ya Yang, Ying Liu, Sihong Wang, Yusheng

Zhou, Xiaonan Wen, Long Lin, Simiao Niu, Xinyu Xue, Lin

Dong, and more; and my collaborators: Charles M. Lieber,

L.-J. Chen, S.Y. Lu, L.J. Chou, R.L. Snyder, R. Dupuis, J.F. Wu,

Gang Bao, Liming Dai, Jing Zhu, Yue Zhang, Aifang Yu,

Peng Jiang, M. Willander, C. Falconi.

References 1. Z.L. Wang , Mater. Sci. Eng. RR 64 , 33 ( 2009 ). 2. Z.L. Wang , J.H. Song , Sciencee 312, 242 ( 2006 ). 3. Z.L. Wang , Adv. Funct. Mater. 18 , 3553 ( 2008 ).

Figure 17. Future of electronics beyond Moore’s law. The vertical axis represents a

miniaturization and increase of device density, CPU speed, and memory. The horizontal

axis represents the diversity and functionality of personal and portable electronics. 10



www.mrs.orgwww.mmr-tech.com

[email protected] ∙ 650.962.9622

THE WORLD’S RESOURCE FORVARIABLE TEMPERATURESOLID STATE CHARACTERIZATION

TTTAAWWMMEERR

HALL EFFECT MEASUREMENT SYSTEMS

4. Z.L. Wang, R.S. Yang, J. Zhou , Y. Qin , C. Xu , Y.F. Hu , S. Xu , Mater. Sci. Eng. R70, 320 ( 2010 ).5. Z.L. Wang , Nanogenerators fors Self-Poweredr Devicesd ands Systemsd (s GeorgiaInstitute of Technology, SMARTech digital repository,y 2011); http :// hdl/ . handle .net / 1853/ / 39262/ . 6. Z.L. Wang, Nano Todayo 5 , 540 (2010).7. Z.L. Wang, Adv. Mater. ( 2012 ); doi:10.1002/adma.201104365. 8. Z.W. Pan, Z.R. Dai, Z.L. Wang , Science 291 , 1947 (2001). 9. Z.L. Wang, Mater. Today 7 , 26 (2004). 10. Z.L. Wang, X.Y. Kong , Y. Ding , P.X. Gao , W. Hughes , R.S. Yang , Y. Zhang,Adv. Funct. Mater. 14 , 944 (2004). 11. X.Y. Kong, Z.L. Wang , Nano Lett.o 3, 1625 ( 2003 ). 12. X.Y. Kong, Y. Ding, R.S. Yang, Z.L. Wang , Science 303 , 1348 (2004). 13. W.L. Hughes , Z.L. Wang, J. Am. Chem. Soc. 126, 6703 ( 2004 ).14. P.X. Gao , Y. Ding , W.J. Mai, W.L. Hughes, C.S. Lao, Z.L. Wang , Science 309 , 1700 (2005). 15. M. Law ,w L.E. Greene , J.C. Johnson , R. Saykally,y P.D. Yang, Nat. Mater. 4, 455 (2005). 16. X.D. Wang , J. Zhou , C.S. Lao , J.H. Song, N.S. Xu , Z.L. Wang, Adv. Mater.19 , 1627 (2007). 17. M.H. Huang , S. Mao , H. Feick,k H.Q. Yan , Y.Y. Wu, H. Kind , E. Weber ,rr R. Russo ,P.D. Yang, Science 292, 1897 ( 2001 ). 18. J.H. Lim, C.K. Kang, K.K. Kim, I.K. Park , D.K. Hwang, S.J. Park , Adv. Mater.18 , 2720 (2006). 19. X.D. Wang , J.H. Song, J. Liu, Z.L. Wang , Science 316 , 102 (2007). 20. Y. Qin , X.D. Wang , Z.L. Wang, Nature 451, 809 ( 2008 ).21. Q.X. Zhao, M. Willander,r R.R. Morjan , Q.H. Hu, E.E.B. Campbell, Appl. Phys.Lett. 83 , 165 (2003). 22. X.D. Wang , C.J. Summers, Z.L. Wang , Nano Lett.o 3, 423 ( 2004 ). 23. S. Xu , Z.L. Wang, Nano Res.o 4 , 1013 (2011). 24. Y.G. Wei, W.Z. Wu, R. Guo, D.J. Yuan, S. Das , Z.L. Wang, Nano Lett.o10 , 3414 (2010). 25. S. Xu , C.S. Lao , B. Weintraub , Z.L. Wang, J. Mater. Res. 23, 2072 ( 2008 ). 26. Z.L. Wang, Sci. Am. 298, 82 ( 2008 ). 27. Y.F. Hu, L. Lin, Y. Zhang , Z.L. Wang, Adv. Mater. 24, 110 ( 2012 ). 28. R.S. Yang , Y. Qin , L.M. Dai, Z.L. Wang , Nat. Nanotechnol. 4 , 34 (2009). 29. G. Zhu, R.S. Yang, S.H. Wang, Z.L. Wang , Nano Lett.o 10, 3151 ( 2010 ). 30. Y.F. Hu, L. Lin, Y. Zhang , Z.L. Wang, Adv. Mater. 24, 110 ( 2012 ). 31. Y.F. Hu, Y. Zhang , C. Xu , G. Zhu, Z.L. Wang , Nano Lett.o 10, 5025 ( 2010 ). 32. J.H. Jung, M.B. Lee, J.-I. Hong, Y. Ding, C.-Y. Chen, L.-J. Chou , Z.L. Wang ,ACS NanoS 5, 10041 ( 2011 ). 33. S. Xu , Y. Qin , C. Xu , Y.G. Wei, R.S. Yang, Z.L. Wang , Nat. Nanotechnol.5 , 366 (2010). 34. Y.F. Hu, Y. Zhang, C. Xu , L. Lin , R.L. Snyder,r Z.L. Wang , Nano Lett.o 11 , 2572(2011). 35. M.B. Lee , J.H. Bae, J.Y. Lee , C.S. Lee , S.H. Hong , Z.L. Wang, Energy Environ.ySci. 4 , 3359 (2011). 36. C. Xu , X.D. Wang , Z.L. Wang, J. Am. Chem. Soc. 131, 5866 ( 2009 ). 37. C. Xu , Z.L. Wang, Adv. Mater. 23, 873 ( 2011 ). 38. B.J. Hansen, Y. Liu , R.S. Yang , Z.L. Wang, ACS NanoS 4 , 3647 (2010).39. C.F. Pan , Z.T. Li , W.X. Guo , J. Zhu , Z.L. Wang, Angew. Chem. Int. Ed.50, 11192 ( 2011 ).40. A.F. Yu, P. Jiang , Z.L. Wang, Nano Energyo (y 2012); doi:10.1016/j.nanoen.2011.12.006.41. Y. Zhang , Y. Liu, Z.L. Wang , Adv. Mater. 23 , 3004 (2011).42. X.D. Wang, J. Zhou, J.H. Song , J. Liu , N.S. Xu, Z.L. Wang , Nano Lett.o6, 2768 ( 2006 ).43. J.H. He, C.L. Hsin , J. Liu , L.J. Chen , Z.L. Wang, Adv. Mater. 19 , 781 (2007).44. Z.L. Wang, Adv. Mater. 19, 889 ( 2007 ). 45. P.W. Bridgman, Phys. Rev. 42, 858 ( 1932 ). 46. C.S. Smith , Phys. Rev. 94 , 42 (1954).47. J. Zhou , P. Fei, Y.D. Gu , W.J. Mai, Y.F. Gao , R.S. Yang , G. Bao, Z.L. Wang ,Nano Lett.o 8, 3973 ( 2008 ). 48. Q. Yang , W.H. Wang, S. Xu, Z.L. Wang , Nano Lett.o 11, 4012 ( 2011 ). 49. W.Z. Wu , Y.G. Wei, Z.L. Wang , Adv. Mater. 22 , 4711 (2010). 50. W.Z. Wu , Z.L. Wang, Nano Lett.o 11 , 2779 (2011). 51. Y.F. Hu, Y.L. Chang , P. Fei, R.L. Snyder,r Z.L. Wang , ACS NanoS 4 , 1234 (2010). 52. Y.F. Hu , Y. Zhang, Y.L. Chang , R.L. Snyder,r Z.L. Wang , ACS NanoS 4 , 4220 (2010). 53. Q. Yang , X. Guo, W.H. Wang , Y. Zhang, S. Xu, D.H. Lien, Z.L. Wang, ACSNano 4, 6285 ( 2010 ).

54. Y. Liu, Q. Yang, Y. Zhang , Z.Y. Yang , Z.L. Wang, Adv. Mater., in press ( 2012).55. Y. Yang, W.X. Guo, Y. Zhang , Y. Ding , X. Wang, Z.L. Wang , Nano Lett.o 11,4812 ( 2011 ).56. Y. Liu , Q. Yang , Y. Zhang, Z.Y. Yang, Z.L. Wang , Adv. Mater. 24, 1410 ( 2012 ).57. Y. Zhang , Z.L. Wang, Adv. Mater., in press (2012).58. Y. Zhang , Y. Yang , Z.L. Wang, Energy Environ.y Sci. 5, 6850 ( 2012 ).59. http://en/ .wikipedia. org /wiki/ / Mechanosensation/ .60. D. Choi , M.J. Jin, K.Y. Lee , S.-G. Ihn, S. Yun , X. Bulliard, W. Choi, S.Y. Lee ,S.-W. Kim, J.-Y. Choi, J.M. Kim, Z.L. Wang , Energy Environ.y Sci. 4, 4607 ( 2011 ).61. X.Y. Xue , S.H. Wang , W.X. Guo , Y. Zhang, Z.L. Wang , Nano Lett.o ; doi: 10.1021/n13028791 ( 2012 ).

Zhong Lin Wang is the Hightower Chair inmaterials science and engineering, a Regents’Professor, and an Engineering DistinguishedProfessor and director of the Center forNanostructure Characterization at the GeorgiaInstitute of Technology. He received his PhDdegree from Arizona State University iny 1987. Hewas elected as a foreign member ofr thef ChineseAcademy ofy Sciencesf in 2009; a member ofr thefEuropean Academy ofy Sciencesf in 2002; andFellow ofw thef American Physical Society iny 2005,of AAAS in 2006, of the Materials ResearchSociety iny 2008, of thef Microscopy Societyy ofyAmerica in 2010, and of thef World Innovation

Foundation in 2002. He received the 2012 Edward Orton Memorial Lecture Awardfrom the American Ceramic Society, the 2011 MRS Medal from the MaterialsResearch Society, the 1999 Burton Medal from the Microscopy Societyy ofy America,fthe 2001 S.T. Li prize for Outstanding Contribution in Nanoscience andNanotechnology, the 2009 Purdy Awardy from the American Ceramic Society, and theNanoTech Briefs Top 50 award in 2005. Wang has authored and co-authored fivefiscientificfi reference books and textbooks and more than 730 peer-reviewedjournal articles, 50 review papers, and book chapters. He also holds 32 patents.Wang can be reached by email at [email protected].


Introduction would fi rstfi like to thank the Kavli Foundation and the

Materials Research Society for givingr me the opportunity to

present thist talk. I will show the ways in which we are fightingfi

cancer withr nanoparticle medicines (nanomedicines). This is

an area where the nanoscale is very important, and Id will pro-

vide some illustrations to demonstrate it. First, I will provide

some background aboutd cancer,t and thend second, I will discuss

nanoparticles and theird potentialr to create new ways to treat

cancer. Nano-sized particlesd can haven a majora effectr int tryingn to

attack solidk tumors,d and Id will show data from the clinic where

we have been treating patients since 2006.

People have been trying to fi ndfi ad cure for cancerr forr ar long

time; there is evidence of writingsf from ancient Egypt that

describe cancer inr papyrus manuscripts from 1600 BC. Toward

the end of thef last century, cancer became the number one

killer of Americansf (in the United States) under the age of

85, replacing heart disease.t 1 The good newsd is that deathst due

to heart diseaset have declined signifid cantlyfi over ther past fewt

decades, whereas deaths from cancer haver remained relativelyd

constant. As one might expect,t based ond data from the World

Bank,2 this phenomenon is not confit nedfi tod the United States,d

but ist a worldwidea problem:e the totale numberl ofr deathsf from cancerm

is larger thanr then total number ofr deathsf from malaria,m HIV, and

tuberculosis, and thisd number isr predicted tod increase significantlyfi

as the population increases.n The costs of therapiesf to treat cancert

have increased dramaticallyd overy ther last decadet or so,r and thisd

increase is unsustainable in then long term.g The cost oft cancerf tor

society is not onlyt the cost oft thef drugs used, but alsot the loss

of lifef of youngf people,g and thed consequent losst of productivity.f

Healthcare costs for cancerr nowr surpassw that fort allr other diseasesr

or injuries,r including thoseg from roadm accidentsd and heartd disease;t 3

thus, this is a verya largey problem, and oned that willt increase.

Most oft usf know somebody in our neighborhood,r or inr our

family, who is currently undergoing cancer therapy.r Patients

suffer ar significantfi losst in quality of life,f whether itr ist acute or

long-term side effects. We therefore have a number ofr reasonsf

for developingr new therapeutics that cant lower ther death rate

and maintaind a high quality of lifef for patients.r In order tor do

this, we need tod attack atk leastt twot problems: metastatic disease

and drug-resistantd disease.t Cancer isr a metastatica disease; it cant

spread fromd its original site to multiple sites simultaneously.

If wef wish to attack metastatick disease, we require therapies

that actt throughoutt thet body. Additionally, cancer canr “fi ghtfi

back” when being treated byd chemotherapeutics. When cancer

is treated withd conventionalh small-molecule chemotherapeutics,

Fighting cancer with nanoparticle medicines ― The nanoscale matters Mark E.k Davis

The following articleg is an edited transcriptd oft thef Fred Kavlid Distinguished Lectured in Nanoscience

presented byd Marky E.k Davis, on November 27,r 2011, at thet 2011e Materials Research Society Fally Meetingl

in Boston. The Lectureship is supported byd the Kavli Foundation, which supports scientificfi research,

honors scientificfi achievement, and promotesd public understanding ofg scientistsf and theird work.r A video

of thef presentation can be viewed atd www.mrs.org/f11-kavli-video.t

Papyrus writings from 1600–1500 BC describe cancer andr the attempts at treatment. Centuries

later, cancer remains a devastating disease. Given the long history of diffif cultiesfi in developing

cancer therapies, why is there excitement about nanoparticle medicine (nanomedicines) for

fightingfi cancer? This article describes the current understanding of whyf these engineered,

nano-sized medicines, which are highly multifunctionaly chemical systems, have the potential to

provide revolutionary waysy to treat cancer. This point is illustrated by physicaly insights at the

nanoscale that allow for ther development of nanoparticlesf that can function in both animals and

humans. The human data show how we have translated two independent nanoparticle cancer

therapeutics from laboratory curiositiesy to experimental therapeutics in human clinical trials.

Mark E. Davis, California Institute of Technology ; [email protected] DOI: 10.1557/mrs.2012.202

FIGHTING CANCER WITH NANOPARTICLE MEDICINES ― THE NANOSCALE MATTERS


one of the ways it can “fi ght back” is to put proteins on the

surface of the cell that act as chemical pumps. When the mol-

ecules used to treat the patient are exposed to the cancer cells,

the cells pump the drug out, and the treatment is ineffective.

This phenomenon is called “multidrug resistance”; the cell

can pump drugs out even if they have never been used on the

patient before, thus, whole classes of drugs become ineffective.

When this happens to a patient, after numerous treatments for

example, very little can be done. To fi nd a solution to this

problem, (1) we need to develop new therapies that can be

administered systemically and can treat metastatic disease;

and (2) it will not be suffi cient to apply the therapy to the

tumor itself, but we will need to move it through the tumor

and into the cancer cells before it is released to carry out its

cell-killing properties. This problem is much more diffi cult than

just getting drugs to the tumor.

Current cancer treatments Since 1955, the major cancer treatment has been chemotherapy.

The fi rst drug used for metastatic cancer was methotrexate; which

was followed by a series of small molecules (e.g., Adriamycin,

Carboplatin, and Taxol) through the 1970s to 1990s. These

small-molecule treatments are used frequently; for example,

billions of dollars’ worth of Taxol is sold each year. Some of

these drugs are administered orally, but most of them are admin-

istered intravenously (IV); they move through the whole body

by diffusion and convection, both into and throughout cells in

the various areas they can reach. The primary function of these

molecules is to inhibit cell division, but most of the dose quickly

exits via the kidney and into the urine. The small amount of the

dose that remains in the body enters all types of cells, and this

leads to a number of different side effects. If the drug enters the

hair follicles, it kills those cells, and the patient’s hair or eyebrows

fall out; if it enters the cells in the gastrointestinal tract, then it

causes vomiting; and if it enters the bone marrow, it causes loss

of cells that make up the immune system and other blood cells.

Hopefully, some of the drug also kills the cancer

cells, but if the cancer has become multidrug

resistant, this can result in a number of harmful

side effects, with no effect on the cancer.

More recent treatments use “targeted,

molecular medicines.” These can still be small

molecules, such as Gleevec, but now antibody

molecules that are ∼1–5 nm in size are being

used. They still can have signifi cant side effects,

but they are more selective in hitting targets.

Nanoscale treatment of cancer I would like to fi rst briefl y discuss some aspects

of tumor biology. When either primary tumors

or metastatic tumors become larger than ∼1 mm

in size—roughly the thickness of a credit card—

they need to create new blood vessels to be able

to bring oxygen and nutrients to continue to

grow, so they send out signals to the current

blood vessels to grow new ones into the tumor mass. These grow

very quickly and are different from the normal, mature blood

vessels in the body. Figure 1 shows a comparison of normal

blood vessels and those in a tumor. 4 The vasculature for normal

cells is leaky to molecules that are ∼1 nm or smaller in size;

these are the molecules needed for nourishment. In a tumor,

however, these vessels are formed quickly, and they are not

completely closed, as in the case of a mature vessel, so entities

that are tens or even hundreds of nanometers in size can leak out.

The question is whether we can exploit this difference to intro-

duce nanoscale entities into the tumor. Of course, this requires

an injection into the circulatory system of a patient: we want to

move nanoparticles into a tumor, but they also pass through the

circulatory system and enter other organs in the body. When we

started this work in the 1990s, there was very little information

known about the way nanoscale objects interact with these organ

systems. Our aim is to minimize the interaction with these organ

systems and ensure the nanoparticles enter the tumor.

This is illustrated in a schematic from my 1996 patent

application ( Figure 2).5 Our idea was to create a stable colloid

(these were not called nanoparticles at that time) that could be

built with the right properties in order to introduce them into

tumor cells. These particles would contain chemical sensors

that would recognize that they were inside of the cell and allow

them to perform certain functions to release the drug. These

were our objectives, and 15 years later, we have met all of them

in treating patients.

We wanted to try and create new therapies to treat the meta-

static drug-resistant problem in a physician’s offi ce, instead of a

research hospital, and provide a high quality of life for patients.

Of course, to be viable, this treatment needed to be highly effi -

cient. If we can create high quality of life therapeutics, we might

be able to treat a patient for a much longer period of time. Also,

we wanted to have a robust system, while hopefully keeping the

costs under control. At the California Institute of Technology

(Caltech) in the mid-1990s, we started with two approaches:

Figure 1. Comparison of (a) normal vasculature (impermeable to entities larger than ca.

2–4 nm) and (b) tumor vasculature (impermeable to entities larger than ca. 400–700 nm).

Reprinted with permission from Reference 4. ©2006, National Cancer Institute.



one way was to try and build systems that could ultimately

obtain FDA approval to treat patients; the other was to build

model systems, from which we could understand some of the

fundamental principles about the way these nanoparticles inter-

act with various tissues. Very little was known about the way

nanoscale entities interacted with the various organs in the body.

One of the model systems we used was gold nanoparticles func-

tionalized with polyethylene glycol (PEGylated

gold). The research started in 1996, and in 2006,

we were able to treat the fi rst patients, which

illustrates the time and effort required.

These colloidal particles are now referred to

as “nanoparticles,” and they started to become

signifi cant in the early 2000s. I believe that one

of the main reasons for this was the National

Nanotechnology Initiative (NNI), which was

started at that time. I was fortunate to be in the

audience at Caltech when President Clinton

announced the initiation of this program. In

the early 2000s, the National Cancer Institute

started their program in nanotechnology, and

they defi ned nanoparticles to be between 1 nm

and 100 nm in size. In our efforts to attack solid

cancers from systemic administrations, we

found that the range is much more restrictive,

but it is in the middle of this region.

Many of us were very positive about this

approach, but others were concerned about

the negative aspects of injecting nanoparticles

into patients. I was asked to testify in front of

a US Senate subcommittee on the safety of

nanoparticles, and I explained nanoparticles

to the senators by saying that the ratio of a

hundred-nanometer particle to that of a soccer ball is of the same

order of magnitude change as that of the soccer ball to planet

Earth. The senators understood this; they knew a nanoparticle is

supposed to be small. It is also interesting to note that a nanopar-

ticle is very large relative to a molecule, so a molecule that is

smaller than 1 nm, compared to a 100-nm nanoparticle, has the

same ratio as that of the soccer ball to the Goodyear blimp. I do

not think that the senators had an appreciation for this point. We

learned over that last decade or so that the “Goodyear blimp” is

too large for treating solid cancers, and we require something like

the size of a hot air balloon relative to a soccer ball.

We are therefore trying to learn the rules for making nanopar-

ticles on the order of 50 nm, and I will show why I believe this

is approximately the right size. We need to ensure that these

highly multifunctional systems perform correctly in the right

place and at the right time. These are not passive entities; they

must be very dynamic in order to perform the tasks required of

them. We will consider size fi rst. As an engineer, I always think

of bounds, so what is the lower bound? For example, animal

and human kidneys all contain holes, ∼10 nm in diameter, to

allow molecules to pass into urine. A nanoparticle larger than

10 nm in diameter, with the right properties, can circulate in

the blood for a number of hours, whereas a small molecule

would escape directly into the urine. As I mentioned earlier,

most small-molecule drugs are delivered directly into the urine,

because they are smaller than 10 nm. This is a very fi xed bound

if we are dealing with a non-deformable spherical particle, but

for different types of morphologies and aspect ratios, different

results may occur. Figure 3 shows a carbon nanotube passing

Figure 2. Initial schematic for cancer treatment with

nanoparticles. Issues from the initial design that have carried

through to the clinic are cyclodextrin polymer-based colloid

(nanoparticle), sizes below 100 nm in diameter, targeted colloid

(R in diagram), and intracellular delivery and active mechanism

of drug release. Reprinted with permission from Reference 5.

©2009, American Society for Pharmacology and Experimental

Therapeutics.

Figure 3. Excretion of carbon nanotubes by a mouse kidney. (a) Glomerular fi ltration

barrier. Inset: section of renal glomerulus (scale bar is 10 mm); (b) bundled MWNT

agglomerate (circled by dashed black line) in the glomerular capillary. Individual MWNTs

(black arrow) (c) 5 min and (d) 30 min after IV injection crossing the fi ltration membrane.

P, podocyte; BM, basal membrane; FE, fenestrated endothelium; EC, endothelial cell;

RBC, red blood cell. Reprinted with permission from Reference 6. ©2008, Wiley.



through one of the pores in a kidney. 6,7 Joe DeSimone, of the

University of North Carolina at Chapel Hill, has shown that it is

possible to make nanoparticles of any shape, and therefore the

rules will change depending on both the shape and the aspect

ratio. However, the rules we will discuss here are for spherical

nanoparticles that are not very deformable.

What happens if we make particles greater than 10 nm in

size? They will distribute themselves throughout the body

and we need to make sure they disassemble; otherwise they

will be present in the body for a long time. Figure 4 is an

image of 70 nm gold nanoparticles inside a kidney, where

they will remain forever. Fitzpatrick et al. showed that even

two years after an injection, nanoparticles that have neither

disassembled nor dissolved are still present in the kidney. 8

This is unacceptable for treating patients, so we need to make

nanoparticles greater than 10 nm in size that circulate but also

disassemble when required.

Now consider the upper bound. As one would expect, the

smaller the size, the farther these particles can move from the

vessel into the tumor tissue. It is now known that 100 nm is

defi nitely far too large, and our preference is 50 nm +/– 20 nm.

All tumors are different, of course, and they are all hetero-

geneous in nature, but it can be shown that for very permeable

tumors, there is not much difference in the 100- to 30-nm

range for penetration into the tumors. However, for tumors

that are not very permeable, there is great discrimination in

size range. 9

Very small particles have a large relative surface area com-

pared to volume, and there are cells in the body whose function

is to scavenge nanoparticles from nature, such as viruses and

fungi, whose surfaces are all highly electrically negative. It was

known in the 1980s, from microspheres, that if these surfaces

are made to be almost neutral, the scavenging is minimized,

and as they become negative and are more similar to nature’s

particles, the scavenging is increased. We must avoid creating

positive surfaces in the body, because all other

surfaces are negative.

Figure 5 shows a transmission electron

microscopy (TEM) image of a section of a

liver showing the liver cells; the Kupffer cells

marked with a K are the cells that scavenge

particles.10 We made a series of nanoparticles

with a constant zeta potential close to neutral

and with increasing sizes from 25 nm to 160 nm.

These nanoparticles were injected into the tail

vein of a mouse, and the TEM images of single

cells showed that the difference in scavenging

in the liver cells between 100 nm and 70 nm

was dramatic. The smaller the particles, the

fewer remain in the cells; thus, it is clear that

smaller is better, because we do not want to

waste valuable therapeutics by allowing them

to be taken up in these cells. In essence, the

rule that applies at the microscale level is the

same rule that applies at the nanoscale level;

however, the smaller the nanoparticle, the less scavenging

occurs by these cells.

Another point is that these particles need to engage various

molecules on the surface of the cells they enter. If the particles

are too small, they cannot engage enough of these molecules to

cause the membrane to wrap around them; if they are too large, the

membrane cannot wrap around them completely. The optimum

size for this membrane wrapping is about 40–50 nm.11

In summary, over the last decade we have learned that these

nanoparticles must have sizes between 10 nm and 100 nm. If

they are entering cells, they should be as close to electrically

neutral as possible, but it is preferable to err on the side of

negative charge. This gives us a basis for the design rules we

require, independent of the type of drug or payload. Figure 6shows a TEM image of nanoparticles with desirable properties

Figure 4. Gold nanoparticles functionalized with polyethylene glycol (PEGylated gold)

trapped in a kidney. ZnS capped CdSe quantum dots coated with PEG injected into Balb/c

mice were detected two years after injection. The nanoparticles are larger than ca. 10 nm.

Figure courtesy of Chung Hang J. Choi (Caltech).

Figure 5. Cells in liver uptake particles; cells marked with K

are the cells that scavenge particles. Figure courtesy of

S.R. Popielarski (Caltech).



entering the cells and localizing into vesicles. 12 It is interesting

to note that the pH is near neutral in these vesicles and becomes

acidic as they move toward the nucleus of the cell. Therefore,

we create chemical entities on the nanoparticles that can recog-

nize this acidity and trigger a number of changes that allow the

drug molecules to be released. If we are using big molecules

like RNA, we have to actively take them out of vesicles, as

they do not diffuse through the vesicle membranes like small-

molecule drugs. These systems are highly multifunctional;

they stay together during circulation and release the payload

in the right place.

Patient treatment using nanoparticles I will now describe some specifi cs about two nanoparticles

that we have been using to treat patients. The fi rst one involves

a small-molecule drug, and the second one involves RNA.

We started to build new polymers based on a molecule called

“cyclodextrin”; this is a ring of sugar. The reason I chose this

molecule is primarily based on human data. The human dose

of cyclodextrin is 8 grams in the therapeutic called Sporanox;

for comparison, a typical dose of Advil or Motrin is 100 mg to

200 mg—up to 1 gram or higher for arthritic patients. These

cyclodextrin molecules have low toxicity, and my idea was

to use these as molecular building blocks to build polymeric

materials with high functionality. We started creating linear

polymers, where the cyclodextrin was part of the backbone

for a variety of reasons. The fi rst nanoparticle that we created

using cyclodextrin-containing polymers carried a small-

molecule drug called “camptothecin,” that inhibits a par-

ticular protein called “topoisomerase I,” and its mechanism

of action would favor having the drug continually bound

to topoisomerase I. Camptothecin itself is so toxic that it

was never commercialized, although it is a very potent drug

for a wide variety of different cancers. However, there are

two commercial drugs based on the camptothecin core, but

they have other organic functional groups on them to assist

in their function in humans. Sales of these commercially

available molecules have been over a billion dollars a year,

but they have many side effects.

We built a polymer with repeating units of cyclodextrin

and a PEG molecule along the backbone. This can be thought

of as a long, fl exible rope with knots in it; the knots are the

cyclodextrins, and the rope in between is the PEG. We attached

the camptothecin molecules to this polymer chain. When this

polymer is placed in water, the camptothecin molecules hide

in the cyclodextrin; they form what is called an “inclusion

complex.” Some of the molecules on this polymer chain hide

in the cyclodextrins of the same chain, while others also enter

cyclodextrins on other chains When this is done correctly,

it forms a nanoparticle of ∼ 30 nm diameter with a slightly

negative zeta potential that contains about 5–10 polymer

chains. Based on our design rules, this is the right size range

and the right charge. This nanoparticle was originally called

IT-101, but is now denoted CRLX101. The nanoparticle must

also be able to disassemble at the right time, because it is too

large to escape through the kidney. Since the drug-cyclodextrin

interaction holds the nanoparticle together, it disassembles

into single polymer strands when the drug is released; these

polymer strands are made of suffi cient molecular weight or

size so that an individual polymer strand is small enough to

escape the body through the kidney. The nanoparticles will

circulate, enter the tumor, enter cancer cells, release the drug,

and disassemble into single strands that escape from the body

through the urine.

We also used nanotechnology in order to observe these

nanoparticles as they accumulate in tumors. We take a 5 nm

Au nanoparticle and attach PEG molecules with adamantane

at their termini to the gold surface, forming Au-PEG-AD. The

adamantane molecule fi ts snuggly inside the cyclodextrin

so that it can bind to the CRLX101 nanoparticle. Since the

Au-PEG-AD is a fl uorophore, it can be used to locate the

positions of the CRLX101 nanoparticles in a tissue sample.

Figure 7 shows both a cryo TEM image of the therapeutic

particle and a cryo TEM image of the gold particles decorating

the surface of the therapeutic particle. When we take sections

of a tumor, we can use the Au-PEG-AD “stain” to follow the

nanoparticles as they move into the tissue and into the cells.

As already mentioned, we can develop the chemistry such

that the nanoparticle releases the drug in a pre-programmed

way; this particular drug should be released over a long period

of time. This means we can retain a slow release agent over a

period of days in the tumor only.

After years of work, we started treating patients in the

summer of 2006, and the first trial was performed at the

City of Hope. We made a freeze-dried product that was then

reformulated in water in an IV bag. The fl uid in the bag was

then infused into the patient. Several patients had survival times

of a year or more using this treatment.

It is interesting to note that, while these nanoparticles

circulate with a half-life of approximately a day in rats and dogs,

they circulated even longer in human patients—with almost a

two-day half-life. Patient reproducibility was excellent, which

we believe is due to the circulation of nanoparticles not bound

to blood components.

Figure 6. Nanoparticles enter cells as individual particles,

bypass surface pumps, and deliver their payloads. Reprinted

with permission from Reference 12. ©2004, Elsevier.



From the biopsies of patients 14 days after a dose, we could

recognize from that gold stain that there were nanoparticles still

in the tumor and releasing the drug. It was also pleasing to note

that the side effect profi le was extremely low; patients had a

high quality of life, and there were no new side effects due to

the presence of the nanoparticles.

We started to observe encouraging activity over a variety of

different cancers, and the mechanism of long circulation and

long drug release we observed in animals has been extended

to humans. It is encouraging to fi nd that the rules learned from

animals may help us try to learn the rules for application to

humans.

CRLX101, a product of Cerulean Pharma, is currently

undergoing a randomized Phase II trial for a particular type

of lung cancer; the trial involves 150 patients at more than

25 sites, and there are a variety of other Phase II trials that will

start in 2012.

RNA as a therapeutic I now discuss ways of delivering a piece of RNA as a thera-

peutic. The understanding of the molecular biology of cancer

is advancing extremely rapidly. It is now known that cells use

many different pathways as they pass through cell growth, cell

division, and cell death. In cancer, these pathways become

altered, and there is an imbalance such that cells continue to

grow all the time. We would like to develop a way to selectively

attack multiple pathways in a patient-by-patient way. A new

technology is available that may help us do this, called “RNA

interference (RNAi),” in which we use a small piece of RNA

called a “duplex.” An RNA duplex used for RNAi consists of

two small pieces of RNA (around 20–25 base pairs), and when

they enter cells, they can interact with some proteins within the

cell. The proteins that incorporate one of these strands shuttle

it to what is called “messenger RNA” (mRNA),

line it up appropriately, and cut that mRNA at

a very specifi c spot. If the mRNA is cut, then

the protein that is normally synthesized by that

RNA message cannot be created. The under-

standing of this technology won the Nobel Prize

in Medicine and Physiology in 2006 for Craig

Mello and Andrew Fire, and they conducted

their study using worms. It requires a large tran-

sition from worms to human patients, but my

colleagues and I reported the fi rst proof of RNAi

in patients in 2010. Fire said, “If a person has a

tumor, why not take a gene that’s essential for

that tumor and administer double stranded RNA

corresponding to that gene to shut down growth

of that tumor?”13 We did just that.

Most drugs work by binding to proteins,

and since proteins have many different func-

tions, these drugs must be very specifi c for each

protein and each function. But if you attack the

mRNA instead, the molecules you can use are

essentially the same as the mRNA. The only

thing you are doing is changing the orders of the letters on the

duplex RNA. Thus, in principle, we can attack any gene with

any function, whereas in the case of drugs, there are a number

of proteins that are “undruggable” because we do not know how

to drug or attack them. This technology, if it can be developed

as a therapeutic, essentially changes the methodology from

primarily chemistry to informatics. We can examine which

genes have gone wrong and dial in appropriate sequences, but

the chemistry is essentially the same.

This process is analogous to a bathtub with faucets. We

normally turn the water on, but when we have enough water,

we turn it off. This is similar to the way a cell knows how to

turn functions on and off before they have mutated. When they

mutate, the faucet is on all the time, and the cell is constantly

making protein and the cells continue to grow; we are simply

creating water, and the water fl ows out over the tub all of the

time. Traditional drugs work at the protein level, and they are

spending all their time mopping up the water. The proteins

keep being created, and the drugs keep mopping up as much

as they can. Using RNAi, we simply turn the faucet off,

which can be a much more effi cient system to stop the pro-

duction of protein, and this is independent of the protein

being stopped. This method has very high potential to be

broadly applicable.

Consequently, we developed a nanoparticle carrying the

RNA duplexes, again using a polymer, and some other mate-

rials to decorate the surface of the nanoparticle to help it target

cancer cells. These pieces of RNA are large relative to che-

motherapeutics. We can construct a particle carrying about

2000 RNA molecules as a payload. Once again, we can use the

Au-PEG-AD stain because we are using a cyclodextrin particle.

Thus, when we examine the tissues, we can see the particles as

well as the position of the RNA in order to see that it is being

Figure 7. Schematic representation and TEM of the interaction between the 5 nm

Au-PEG-AD and 30–40 nm IT-101. The TEM images show the therapeutic particle (left)

and gold particles decorating the therapeutic particle surface.



delivered throughout the tumor. We can examine individual cells

and again show individual particles circulating from the point

where the animal is dosed to the inside of the cancer cell. We

started treating patients with these nanoparticles in the spring

of 2008, and in 2010, we published the fi rst results showing that

this technology can actually be used in a living human being. 14

Figure 8 illustrates what we believe happens when we

infuse these nanoparticles into the patient: (step 3) they circu-

late, (step 4) enter the tumor, and (step 5) move into the tumor

cells. These nanoparticles contain chemical sensors that can

recognize that they have entered the vesicles, and we have

built in a mechanism to (step 6) bring them out of the vesicles

and release the RNA that then will be (step 7) taken up by the

protein machinery, guide it to the mRNA, and (step 8) cut the

mRNA to (step 9) stop the production of protein. 5 In principle,

if this mechanism was acting in the way we believe it is

acting, we should see a decrease in the mRNA, a decrease in

the protein, and a new fragment of RNA. We were fortunate

to obtain biopsies from patients at three different dose levels:

18, 24, and 30 mg-siRNA/m2. When we examined the tissues,

we saw that for the lowest dose, using the Au-PEG-AD stain,

we saw no nanoparticles. When we examined the intermediate

dose, we started to see some nanoparticles, and then at the

higher dose, we saw many nanoparticles. After a month,

we examined biopsies of one of the patients where we saw

the nanoparticles, and we observed that the nanoparticles had

all disassembled (the disassembled components are suffi ciently

small to escape the body via the kidney). After a repeat dose,

we saw the nanoparticles returning again, in a very reproducible

manner. Thus, for the fi rst time, we have seen a dose-dependent

accumulation of these nanoparticles within tumor cells from a

systemic administration: this is the fi rst example of this using

nanoparticles of any type.

A very encouraging result was to fi nd that we did not see

any nanoparticles in the tissue adjacent to the tumor. We believe

that this is good evidence to suggest that these nanoparticles

accumulate in tumors through the leaky vessels, but cannot

accumulate in the healthy tissue next to the tumor.

When we examined the tissue through staining, we

observed reductions in the protein we were trying to inhibit.

When we looked for either the protein or the level of the

mRNA, we found it is reduced. Additionally, we showed the

presence of RNA fragments after dosing, and by sequencing

these fragments, we revealed that the mRNA was cut at exactly

the right position by the RNAi mechanism. This was the fi rst

example showing that we could perform RNAi in a patient.

Thus, we have demonstrated that RNAi can be successful

in a human patient and gives a high quality of life during

treatment.

The future We are now learning a great deal about design rules and how

to control the properties of these nanoparticles to make them

more biocompatible and more effective. The newer particles

I have just shown have ever-increasing functionality in order

to perform the required function at the right place and at the

right time. There is no doubt that

these nanoparticles will be com-

plex, but hopefully, this will be

worth the effort. We can now

focus on ways to create very

effective therapeutics for solid

tumors and give patients a high

quality of life.

One thing I am very proud

of is that we have been able to

stop the production of an indi-

vidual gene in the tumor of the

patient, thus there is no reason

we cannot stop the production

of multiple genes at the same

time. We could take a biopsy

from a patient, fi nd which genes

are causing their disease, create

RNAs to treat the patient, and be

able to follow how the treatment

of the disease is progressing,

maybe by simply taking a prick

of blood. In the future, one could

envision that there will be an app

on a smartphone that will read

the information from a prick of

blood, call up a physician, and

Figure 8. Schematic of the delivery and function of treating a patient with targeted nanoparticles

containing RNA.



www.materials360online.com

Breaking Research News

MRS News

Hot Topics

Featured Journal Articles

Videos and Podcasts

Education and

Outreach Links

Twitter Feeds

And More

Your Search Stops Here!

Advanced Research Systems

report the results. This will enable the physician to know how

the disease is either regressingr or progressing,r and this infor-

mation couldn bed used dynamicallyd toy decide the next dose,t what

genes it should attack, and when it should be administered to

the patient. This is really a dream, but the basic science and

engineering for every step of thisf process has already been

worked out. It has yet to be integrated, but the basic principles

of suchf a system are all in place, and every step of thef way

required newd nanoscience and nanoscaled engineering. My hope

is that at least some fraction of thisf will be achieved in the

near future.r

In conclusion, I hope I have been able to convincingly show

very high potential for ther use of nanoparticlesf to create new

types of therapiesf for treatmentr oft solidf tumorsd that providet

patients with a high quality of life.f It ist very encouraging for

me to see patients in these early clinical trials having a high

quality of lifef when treated withd these therapies.

I have had thed great privileget of workingf with wonderful

people everywhere—at Caltech,t and withd various companies.

I would very much like to thank all the patients who have

been treated in these trials. It has been a pleasure for mer to

be present int the treatment roomst with them in a number ofr

different trials.t Finally, I would onced again like to thank thek

Kavli Foundation, and MRS,d for givingr me the opportunity to

speak withk you tonight.

References1. A. Jemal , R. Siegel, J. Xu, E. Ward , CA-Cancer J.r Clin. 60, 277 (2010). 2. Data from the World Bank, www.w worldbank . org .3. T. O’Callaghan , Nature 471 ( 7339 ), S2 ( 2011 ).4. M.R. Dreher,r W. Liu , C.R. Michelich, M.W. Dewhirst, F. Yuan, A. Chilkoti,J. Natl. Cancer Inst.r 98 ( 5 ), 335 ( 2006 ).5. M.E. Davis , Mol. Pharmacol. 6 ( 3 ), 659 ( 2009 ).6. L. Lacenda, M.A. Herrero, K. Venner ,r A. Bianco , M. Prato, K. Kostarelos, Small4 ( 8 ), 1130 ( 2008 ).7. A. Ruggiero, C.H. Villa , E. Bander,r D.A. Rey,y M. Bergkvist, C.A. Batt,K. Manova-Todorova, W.M. Deen, D.A. Scheinberg, M.R. McDevitt, PNAS 107( 27 ), 12369 ( 2010 ).8. J.A.J. Fitzpatrick, S.K. Andreko , L.A. Ernst, A.S. Waggoner,r B. Ballou,M.P. Bruchez , Nano. Lett. 9 (7 ), 2736 ( 2009 ).9. H. Cabral , Y. Matsumoto , K. Mizuno, Q. Chen , M. Murakami, M. Kimura ,Y. Terada , M.R. Kano, K. Miyazono , M. Uesaka, N. Nishiyama, K. Kataoka, Nat.Nanotechnol. 6 , 815 (2011).

10. S.R. Popielarski, S. Hu-Lieskovan, S.W. French , T.J. Triche, M.E. Davis,Bioconjugate Chem.e 16 ( 5 ), 1071 ( 2005 ).11. W. Jiang, B.Y.S. Kim , J.T. Rutka , W.C.W. Chan, Nat. Nanotechnol. 3 (3), 145( 2008).12. S. Mishra , P. Webster ,r M.E. Davis , Eur. J. Cell Biol.l 83, 1 ( 2004 ).13. A.Z. Fire , “Gene silencing by doubley stranded RNA” (Nobel Lecture, December 8,r2006), p. 224 ; www.w nobelprize.org / nobel_prizes/ / medicine/ / laureates/ /2006/ /fi// re_filecture . pdf .14. M.E. Davis , J.E. Zuckerman, C.H.J. Choi, D. Seligson , A. Tolcher,r C.A. Alabi,Y. Yen , J.D. Heidel , A. Ribas, Nature 464 , 1067 (2010).

Mark E. Davis is the Warren and KatharineSchlinger Professor of Chemical Engineeringat the California Institute of Technology anda member of the Experimental TherapeuticsProgram of thef Comprehensive Cancer Centerr atrthe City of Hope. His research efforts involvematerials synthesis in two general areas: namely,zeolites and other solids that can be used formolecular recognitionr and catalysis and polymersfor ther delivery ofy af broad range of therapeutics.fHe is the founder ofr Insertf Therapeutics Inc. andCalando Pharmaceuticals Inc. He also has beena member of the scientificfi advisory boards ofSymyx and Alnylam. Davis has more than 375

scientificfi publications, 2 textbooks, and more than 50 patents. He is a foundingeditor ofr CaTTech andh has been an associate editor ofr Chemistry ofy Materialsf ands theAIChE Journal.E He also is the recipient of thef Colburn and Professional ProgressAwards from the American Institute of Chemicalf Engineers (AIChE), the Ipatieff,Langmuir, Murphree, and Gaden Prizes from the American Chemical Society (ACS),yand the National Science Foundation (NSF) Alan T. Waterman Award. He waselected to the National Academy ofy Engineeringf in 1997, the National Academy ofySciences in 2006, and the Institute of Medicine of the National Academies in2011. Davis can be reached by email at [email protected] .


Introductionhost responset to biomaterials forms the foundation for ther

integration of naturalf and syntheticd materials with the human

body.1,2 An understanding of cell-biomaterialsf interactions

within this framework providesk clinically relevant insightst for

the design of newf biomedical innovations ranging from medi-

cal implants to nanoparticle therapies.3 The depth and breadthd

of knowledgef in this domain can be expanded by adopting

new tools and innovative methods. Improved technological

capabilities can be used tod propose increasingly sophisticated

biological questions and testd moret complex hypotheses. Nat-

ural and syntheticd materials along with applied chemistryd and

advanced microfabrication techniques play a central role in

advancing our collectiver knowledge base of biomaterials-tissuef

interactions and biocompatibilityd across many length scales.4,5

The inherent biologicalt cues that aret presented tod cells at inter-t

faces can be precisely controlled byd definingfi the physical prop-

erties, including surface chemistry, mechanical properties, and

microstructure.6–9 Modulating these parameters in then context oft

biomaterials design can provide additional insight intot mecha-

nisms associated withd cell adhesion, migration, spreading, and

differentiation. Consider ther now well-established strategies

of selectivef protein adsorption using soft lithographyt 10,11 and

controlled substrate stiffness using engineered polymer net-r

works.12 These elegant approachest have produced remarkabled

fundamental discoveries.6,13

Static materials with well-definedfi chemical,d physical, and

mechanical properties have contributed significantlyfi to the

identificationfi of mechanismsf that underpin cell-biomaterials

interactions. Assessing cell responses to biomaterials or inter-r

faces with novel compositions, structures, or functionalityr typi-

cally involves the characterization of dynamicf reactions in a

static biomaterial environment.14–17 Interrogating cell popula-

tions through fluorescencefl microscopy, immunohistological

characterization, mRNA quantifiA cation,fi and cytokined profi lingfi

provides information that describest the composite of af cell’s

observable characteristics (phenotype) and insightd intot ongo-

ing active biological responses at discretet time points. While

static materials can provide some element oft insightf intot these

investigations, they are fundamentally limited with respect

to measuring transient responses.t Living cells are inherently

dynamic and continuouslyd adapt to their environmentr in the

relentless pursuit oft homeostasis.f Many of thef kinetic aspects

of cytoskeletonf responses are difficultfi tot interpret whent using

Smart polymers and interfaces for dynamic cell-biomaterials interactions Stephen Kustra and Christopher J.r Bettinger

Fundamental aspects of thef interactions between cells and biomaterials provide a crucial

framework for the design of manyf systems that interact with the human body. A comprehensiveA

understanding of multifactorialf processes may provide an adequate basis for the rational

design of implantablef devices ranging from nanoparticles for drug delivery and scaffolds for

tissue engineering, to artificialfi organs for augmentation and rehabilitation. Recent progress in

the elucidation of cell-biomaterialsf interactions has been predicated on advances in polymer

chemistry, materials engineering, and device microfabrication. The confluencefl of these

developments has stimulated a newfound ability to design soft materials and interfaces with

precise chemical, physical, and mechanical properties. While static surfaces can yield insight

into the interaction of mammalianf cells with medical materials, the study of fundamentalf

cell-biomaterials phenomena will benefitfi significantlyfi from the ability to present biologically

active signals to cells with spatiotemporal control. Specifi cally,fi the use of dynamicf materials

and interfaces can mirror the intrinsic dynamic behavior of livingf cells. This article highlights

recent advances in soft materials design, interfacial engineering, and synthetic polymer

networks in the context of producingf dynamic materials to study cell-biomaterials interactions.

Emerging challenges and future research directions are also discussed.

Stephen Kustra, Carnegie Mellon University,y Pittsburgh, PA; [email protected] Christopher J. Bettinger, Carnegie Mellon University,y Pittsburgh, PA; [email protected] DOI: 10.1557/mrs.2012.185

SMART POLYMERS AND INTERFACES FOR DYNAMIC CELL-BIOMATERIALS INTERACTIONS


static surfaces. Such processes may include adhesion, spread-

ing, and migration.13 Furthermore, seeding cells on engineered

surfaces that present time-invariant bioactive signals (static

cues) leads to a convolution of nascent adhesion and spread-

ing with other types of cellular responses that may be of

interest.

The properties that defi ne the extracellular matrix (ECM)

microenvironment, either natural or synthetic, can be roughly

classified into the following broad categories: chemical

and biochemical (surface energy, protein adsorption, integ-

rin binding), 18,19 mechanical (viscoelasticity, stiffness), 20 and

topographic (geometry of ECM structures). 21–24 The bioactive

signals in the native ECM can be reproduced in synthetic envi-

ronments through a variety of techniques, including protein-

patterning,25 materials engineering,15 and microfabrication. 26–28

For example, the ECM is composed of proteinaceous fi bers with

a characteristic size, geometry, and orientation. These properties

collectively defi ne the topography of the ECM, which can be

mimicked using synthetic substrates.17 Controlling the topo-

graphy of a biomaterial, either natural or synthetic, is a powerful

way to modulate a broad spectrum of cell functions, 24 including

morphology, 7,29 proliferation,30 migration,31,32 and differentia-

tion. 8,33 Engineered cell-topography interactions using synthetic

materials have been explored in numerous contexts, including

controlled differentiation 8,33,34 and regenerative medicine.34,35

Hence, topography is an important aspect of an ECM that can

be engineered to study cell-biomaterial interactions using model

in vitro environments.

There have been many technological advances in real-time

characterization techniques, including sophisticated microscopy

tools and imaging agents. However, parallel advances in the

real-time control of materials are also required. Coordinated

progress in the design and synthesis of smart materials with

downstream characterization techniques can lead to a deeper

understanding of the kinetics of cell-biomaterials interactions.

For example, a variety of cytoskeleton remodeling processes

can be quantifi ed, including microtubule organization, fi lopodia

formation, and other structural changes. 36 Precise measurements

of these processes may serve as biomarkers that yield insight

into the identifi cation and treatment of disease. In addition

to measuring the characteristic time scales of cytoskeleton

remodeling events, dynamic systems can be used as controlled

environments to study complex processes such as stem cell

differentiation or tissue development.

This article focuses on the design of next-generation poly-

meric materials and functional interfaces for applications in

elucidating the fundamentals that drive cell-biomaterials inter-

actions. Specifi c focus is granted to the governing philosophy

that precise spatiotemporal control of cue presentation to cells

using smart materials is benefi cial. Recent advances in pro-

grammable microenvironments that provide new capabilities

in testing advanced hypotheses in biology will be discussed.

The article fi rst outlines the benefi ts of using dynamic systems

to study biology and then provides recent examples from the

literature in the general areas of programmable interfaces and

dynamic polymeric networks. Emerging trends and future chal-

lenges are also highlighted.

Programmable surface chemistry Design considerations in switchable surfaces There are two essential components of a dynamic bioactive

surface: (1) a stimulus that can be arbitrarily applied to the

surface and (2) a material system that is designed to trans-

duce the stimulus into a signal. The mechanism of stimulus

transduction is ideally reversible and very precise. There is an

abundance of materials that satisfy the latter criterion, while

there are relatively few stimuli that are both mild and benign

and therefore appropriate in studying dynamic cell-materials

interactions. Most programmable stimuli for these applica-

tions rely on the selective application of heat, light, mechanical

strain, or electrical signals to the surface. Perhaps the simplest

demonstration of a dynamic surface is reversible stretching,

which has been extensively studied primarily in the context

of measuring the impact of cell mechanics on downstream

function. 37–39 Systems that employ increased complexity by uti-

lizing alternative stimuli-responsive surfaces will be discussed

in this section. Table I highlights representative examples of

responsive materials that are able to alter physical properties

of the surface (left column) via an external stimulus.

Electronic control of self-assembled monolayers Dynamic control of surface chemistry represents a general tool

that is widely applicable, including for self-cleaning surfaces,

the reduction of biofi lm formation, and selective cellular adhe-

sion. There are many possible modes to control the local surface

chemistry, which in turn can be used to modulate the adsorption

of proteins and other functional groups to promote or hinder cell

adhesion. Early examples of this approach have used applied

voltages to alter the confi guration of self-assembled monolayers

(SAM) on gold surfaces. 40 Under small positive voltages, the

terminal anionic carboxylic acid groups of low-density, fl ex-

ible alkane thiols are temporary localized to the surface in a

reversible manner. This transient confi guration leads to the

presentation of hydrophobic domains located in the mid-chain

of the SAM to the free surface. When the voltage is removed,

the carboxylic acid groups become unpinned from the surface

and relax, thereby recovering the anionic surface. This process

is rapid, reversible, and nearly instantaneous. Cell adhesion

can be controlled by manipulating surface chemistry through

the application of external voltages. In one demonstration,

SAM attached to gold surfaces can be patterned into arbitrary

micron-scale geometries through soft lithography and irrevers-

ible removal through the application of a reductive potential.

This technique can be used to study the migration of single

cells that are pre-patterned into select shapes. 41 The guiding

hypothesis for this work is that cells temporarily physically

constrained into a shape that represents an actively migrating

cell will assume a preference for biased migration once the

barriers to spontaneous migration (the SAM in this case) are

removed. This hypothesis has been proven correct: cells



that are pre-positioned with asymmetric geometries showed

enhanced migration in terms of both directionality and velocity

compared to cells that are pre-patterned into isotropic shapes.

Figure 1 summarizes the results of these experiments. The

asymmetric shapes used in these studies are shown in Figure 1a

where the dark area represents the cell adherent area (no SAM).

Confi ning cells within these geometries leads to segregation

of certain cellular components toward the blunt end of the

cell ( Figure 1b ). Figure 1c–d shows time lapse images of

two different cell types as the physical barrier is released

at time t = 0’. This demonstration is a salient example of a

material that can be utilized to test an intriguing hypothesis

that would otherwise be impossible using traditional cell

culture methods.

It should also be noted that the direct modifi cation of sur-

face energy can be achieved by simply altering surface charge

through electrical stimulation. 42 This approach has been utilized

in a number of advanced materials for studying cell-biomaterials

interactions, including conducting polymers. Electronically con-

ductive cell-adhesive substrates composed of naturally occurring

compounds, such as melanins, may also serve as candidate mate-

rials for this strategy. 43 The advantages of this method include

the reversibility of surface properties, robust and predictable

responses, facile fabrication, and micron-scale resolution. This

topic has been reviewed extensively elsewhere. 44

Optical control of surface ligands Spatially selective removal of SAM can also be engineered

through the application of optical stimulus. Laser-scanning

lithography (LSL) is a convenient tool to deliver such a stimulus

with resolutions that are relevant to cell-biomaterial interac-

tions. One such application of this technique utilizes gold-thiol

SAM that are modifi ed with cell adhesive ligands such as

Arg-Gly-Asp residues. 45 The principle of operation is to selectively

pattern successive SAM compositions through cycles of blanket

absorption and subsequent localized desorption ( Figure 2 ). The

process begins by depositing the fi rst SAM layer on a gold-

coated glass. A laser is raster scanned on the surface to induce

selective desorption through localized heating. A second SAM

layer can then be deposited onto available substrate domains.

This process can be repeated in succession to create complete

patterns of complex surface chemistries. The desorption process

is suffi ciently rapid enough to enable spatially targeted removal

of ligands from cell-seeded substrates. The development of this

tool is predicated on the desired ability to study the response of

single cells to the selective removal of substrate cues, including

chemical domains that are known to promote cell adhesion. It

may be possible to selectively detach certain parts of a cell in a

controlled manner. For example, an area of a cell that expresses

a certain protein can be removed, and the downstream alterations

in morphology or regulatory behavior can then be quantifi ed

with high-resolution live-cell imaging. Although the process of

selective ligand removal is irreversible, the ability to manip-

ulate cytoskeleton components with subcellular resolution is

attractive from a technological standpoint. There are other

Table

I.

Sum

mary

of

dyn

am

ic m

ate

rials

and t

heir

ass

oci

ate

d s

tim

uli

.

Sti

muli

Lig

ht

Heat

Ele

ctri

cal

Mech

anic

al

Pro

pert

y R

ever

sible

Irre

vers

ible

Rev

ersi

ble

Irre

vers

ible

Rev

ersi

ble

Irre

vers

ible

Rev

ersi

ble

Irre

vers

ible

Surf

ace

Chem

istr

y A

zoben

zene

iso

mer

izat

ion

46

LS

L o

f S

AM

45

SA

M d

esorp

tion

via

LS

L4

5

SA

M o

rien

tation,

conduct

ing

poly

mer

s 40

SA

M des

orp

tion

41

,11

5

Inte

rface

Topogra

phy

Mic

ro-m

old

edS

MP

80

Contr

olle

d c

rack

form

atio

n a

nd

buck

ling

81,8

3

Nan

ow

rinkl

es7

8,7

9

Mech

anic

al

Modulu

sP

hoto

-dim

eriz

atio

n

of

cinn

amat

e net

work

s10

5,1

16

Photo

lysi

s of

cross

-lin

ked

net

work

s11

3

Die

ls-A

lder

cy

clo-a

dditio

n1

03

,10

4,1

17

This

non-e

xhau

stiv

e lis

t re

pre

sents

a b

rief

surv

ey o

f th

e re

leva

nt

mat

eria

ls t

hat

are

dis

cuss

ed in

this

work

. LS

L, la

ser-

scan

nin

g li

thogra

phy;

SA

M, se

lf-a

ssem

ble

d m

onola

yers

; S

MP,

shap

e-m

emory

poly

mer

.



prominent demonstrations in which the surface energy and

presentation of ligands can be controlled reversibly through

optical methods by employing photo-switchable chemistries

based on azobenzene derivatives. 46

Dynamic topography Substrate topography with features that span the micron- and

submicron length scales serves as a potent cue in directing cell

function.47,48 Reactions to topography are highly conserved

across many phenotypes, including neurons, endothelial cells,

smooth muscle cells, and stem cells. 7,14,24,27,30,31,49 Natural topo-

graphy in ECM proteins infl uences contact guidance responses

and directs in vivo migration. 50–53 Native structures can be repro-

duced in vitro using synthetic substrates with micron-scale

features.54,55 Microfabrication techniques have been employed

to study the role of topographical cues on cell function. 23,56 Syn-

thetic topography uniformly induces altered cellular morphol-

ogies as well as subsequent downstream signaling that affect

proliferation, 29,57,58 migration,17,24,32,35,59–63 differentiation,8,33,34,62

and ECM protein production. 64–66 These responses are likely

to govern decision-making processes regarding interactions

with micro- and nanoparticles such as opsonization and phago-

cytosis, mechanisms by which cells can uptake micron-scale

particles. 67 One theory suggests that fi lopodia, small actin-rich

fragments that extend from the cytoskeleton, engage in an

active substrate exploring function. 68 Although the impact of

topographical microenvironment on gross morphology has

been studied extensively, 69,70 the mechanisms that govern cell-

topography interactions are only starting to be unraveled. 71

Thus, topography is a cue that warrants parallel investigation

along with other substrate properties such as substrate stiffness

and surface chemistry.72

Previous studies have investigated populations of fi xed cells

that are seeded on engineered topographic substrates with well-

defi ned features, 8,14,15,17,73 including bilayer fi lms with controlled

buckling.74 The insight in these studies is limited in part by the

static nature of the substrate. Scant information on the kinetic

responses of these interactions can be extracted using static

surfaces. 16 There are several prominent examples of substrates

with unique structures that can be used as sensing elements.

For example, substrates can be decorated with micron- or

nanometer-scale vertical pillars that extend from the surface.

Adherent cells exert forces on these structures, which distort the

shape of the pillars. 75,76 The degree of tip defl ection is directly

Figure 1. Controlled electrical desorption of self-assembled

monolayers to assess polarized cell migration. (a) Two representative

asymmetric patterns used in biased migration experiments are

shown. The dark area represents the cell adhesive domain and the

approximate shape of the cell prior to release of the migration barrier

(the self-assembled monolayer). The Golgi and the centrosome are

located closer to the half of a cell with the blunt end. The various

components of the cell are identifi ed as follows: actin (red), the Golgi

(green), the nucleus (blue), microtubules (red), and the centrosome

(green), respectively. HUAEC, human umbilical artery endothelial

cells. (b) Asymmetric pattern and corresponding bright fi eld image of

a fi broblast are shown. (c) Time-lapse images (in minutes) show the

motility of a polarized fi broblast after the removal of the migration

barrier. (d) Kidney cells derived from monkeys (COS-7 cells) show

a similar behavior to fi broblasts. This system depicts the ability to

observe real-time changes in a cytoskeletal structure. Reprinted

with permission from Reference 41 . ©2005, National Academy of

Sciences.



related to the force generated by the cell and can be accurately

measured through microscopy (light, fl uorescent, or scanning

electron). 75,76 This technique can measure the forces generated

by a cell that is migrating on a surface with subcellular-scale

precision. 77 There is considerable interest in using structures

on surfaces as a characterization tool to mea-

sure responses. However, there is also merit

in designing dynamic surfaces that can serve

as a stimulus as well. Dynamic substrate

topography enables the quantifi cation of tem-

poral responses of the cytoskeleton, includ-

ing reorganization kinetics. Surfaces that

can actively switch between defi ned surface

topographies may serve as a powerful tool

in cell biology with the potential for broad

impact in a variety of disciplines. This system

could be used to quantify the kinetics asso-

ciated with microtubule organization, fi lo-

podia formation, or other structural changes.

Dynamic topography is an attractive tool to

study cell-biomaterial interactions for several

reasons: (1) interfaces can be designed using

a variety of smart and stimuli-responsive

materials; (2) the fabrication of these sur-

faces is compatible with many standard-

ized techniques such as replica molding

and photolithography; (3) topographic

states are binary, dynamic, and potentially

reversible; (4) there are many available

approaches to fabricate structures across

a broad spectrum of length scales; and

(5) the presence or absence of topography

on a surface can be precisely controlled in

time by employing the strategies described

herein.

Nanowrinkles Next-generation programmable surfaces

will benefi t from the transduction of benign

external stimuli into arbitrary presentation

of topographical features at the material

interface in a reversible manner. There are

many candidate materials that can harness

this principle to ultimately overcome the

limitations of static surface systems. Thermal

loads have been used as stimuli to gener-

ate tunable nanowrinkles. 78 Specifi cally,

temporary fl at substrates are prepared by

depositing thin gold membranes on bulk

polystyrene fi lms. The composites are then

heated, which shrinks the polystyrene fi lms

and subsequently introduces strain-induced

nanowrinkle formation in the gold mem-

brane ( Figure 3 ). Biaxial compression

produces disordered submicron-scale

features through isotropic wrinkling ( Figure 3b ). Uniaxial com-

pression produces nanograting features with short-range order

and wavelengths between 200 nm and 1 μ m ( Figure 3c ). The

advantages of this approach include the ability to functionalize

gold with a variety of thiol-containing compounds and the

Figure 2. Dynamic desorption of self-assembled monolayers (SAMs) using laser-scanning

lithography. (a) Gold-coated glass coverslips are functionalized with alkanethiols. Select

regions are thermally desorbed using laser-scanning lithography (LSL). Bare regions are

exposed to additional alkanethiols. This process is continued to allow for patterning of multiple

SAMs. (b–d) Multifaceted surfaces can be created to present coordinated ligands, including

adhesion-promoting peptides (GRGD) and fi bronectin (HFN). The multifaceted surface displays

independent arrays of GRGD (red horizontal patterns) and HFN (blue vertical patterns). These

dynamic interfaces can be used to study lamella formation and migration in endothelial cells.

(e) Endothelial cells are initially elongated in a direction parallel to the long axis of the HFN

patterns. The cells then produce membrane extensions along GRGD patterns and form lamella

near fi bronectin. Scale bars represent 20 microns. Reprinted with permission from Reference 45 .

©2011, Wiley.



relatively straightforward fabrication process. An alternative

technique utilizes bilayer substrates that consist of a shape-

memory polymer (SMP) with a thin oxide fi lm that is formed

using oxygen plasma. The composite is uniaxially compressed

through thermal activation by heating the fi lms to 150°C. This

technique has been adopted to enhance the assembly of tis-

sue. More specifi cally, substrates with nanoscale features can

organize cells that are precursors to forming functional cardiac

tissue. Using topography to organize these cells produces car-

diac tissue that functions more like the native heart. 79 However,

the formation of dynamic topography using these materials

requires high temperatures that are incompatible with in situ

cell culture and ultimately restrict the ability to signal viable

cell populations in real time.

Shape-memory polymers Thermal stimuli can be directly integrated with cell-seeded

surfaces by adopting phase transitions at or near physiolog-

ical temperatures. In work by Henderson et al., a thermally

activated SMP network is utilized as a material that can pre-

sent dynamic substrate topography. 80 External heating serves

as an irreversible stimulus, which is transduced through a

polyurethane-based SMP. The SMP network (Norland Opti-

cal Adhesive-63) is initially cured on microfabricated negative

molds by photo-cross-linking via thiolene chemistry. Fibro-

blasts are cultured on surfaces with temporary micron-scale

grating features. Thermal activation leads to the recovery of

a fl at substrate. Abolishing the grating features by heating the

material from 30 to 37°C leads to a gradual decay of cell align-

ment and a recovery of isotropic cell spreading. Hence, there

is a gradual restoration to a morphology that is representative

of cells cultured on fl at substrates ( Figure 4 ). The next gener-

ation of materials for this application should be reversible with

minimal hysteresis and exhibit narrower temperature ranges

for thermal transitions.

Reversible topography The central ideal of reversible topography is to apply aspects

of materials engineering to fabricate synthetic substrates with

topographic features that can signal cells in a dynamic manner.

Prospective studies in this area will ultimately lead to a deeper

understanding of the role of cytoskeleton function in response

to the selective presentation of topographic cues. These systems

can be used to elucidate cytoskeleton reorganization kinetics in

a variety of cells, including phenotypes that approximate pro-

genitor or disease states. The design and fabrication of revers-

ible topography for studying cell-biomaterials interactions

requires a cross-cutting interdisciplinary effort that integrates

materials science and engineering, polymer fabrication, solid

mechanics, and cell biology.

Mechanical strain is an effective stimulus that can be

leveraged for use in reversible topography. Takayama et al.

described the reversible mechanical compression of oxidized

polydimethylsiloxane (PDMS) bi-layer substrates for use in

measuring dynamic cell-substrate interactions.81 In this study,

uniaxial compression of a rigid fi lm on an elastomeric substrate

at strains of 11–15% produces grating features with wavelengths

of approximately 7 μ m and amplitudes of 670 nm. Alternat-

ing the presentation and withdrawal of substrate topography

Figure 3. (a) Fabrication of uniaxial and biaxial nanowrinkles

for use as a cell culture platform. (b) Biaxial and (c) uniaxial

scanning electron microscopy images of shrunk polystyrene

sheets with a layer of gold (10 nm). Controlled wrinkling patterns

can be used as a cell culture platform for real-time microscopy

and cytoskeletal dynamic studies. Reprinted with permission

from Reference 78 . ©2009, Wiley.



signals induces aligned and random orientation of C2C12

murine myoblasts, respectively. This process can be repeated

up to nine times without any observable hysteresis. There are

several technical challenges in this approach, including the

limited range of mechanical strains that can induce feature

formation without macroscopic deformation of the substrate.

Mechanical strains over 3.5% can lead to a convolution of

mechanical strain in conjunction with the presentation of

topographic cues. 38 By comparison, a typi-

cal cell stretching experiment would strain a

substrate by approximately 10%.82 Reversible

topography has also been utilized as a method

to control the presentation of adhesion ligands. 83

Briefl y, submicron-scale cracks can be induced

by uniaxially straining bilayers composed of

elastomeric PDMS substrates and thin silica

membranes. First, the unstrained substrate is

coated using cell repulsive detergent (Pluronic

F108). Next, the substrate is strained, thereby

forming micro-cracks that can be chemically

modifi ed with collagen. The resulting surface

consists of a fi eld of cell repulsive pluronic with

cell adhesive collagen that is confi ned within

the micro-cracks. The adhesion domains are

masked under zero strain and exposed during

the application of positive strain induced by uni-

axial tension. The time course and density of

collagen domains on the substrate can therefore

be controlled by simply adjusting the amount of

strain. This approach has been used to validate

the dynamic responses of muscle precursor cells

to pulsed collagen exposure. These cells align

and elongate in response to the exposed colla-

gen domains that are present under 10% strain.

The spreading and alignment can be reversibly

controlled by cycling between strains of zero

and ten percent.

Reconfi gurable polymer networks The intrinsic mechanical properties of a cell can

play a prevalent role in determining cell fate. 13,84

Pathologies in cellular mechanics can lead to

debilitating disease states. 85 Conversely, the

intrinsic mechanical properties of cytoskeleton

components may serve as a valuable biomarker

for diagnostics. 86 For example, the cytoskeleton

of metastatic cancer cells is 70% softer than cor-

responding non-metastatic cancer cells. 87 Cell

mechanics play a critical role in hematologic

diseases as well. 88 Hence, the role of cellular

biomechanics in maintaining homeostasis can-

not be diminished. There are several established

theories of mechanotransduction that have been

supported by subsequent biophysical studies.

The tensegrity model proposed by Ingber posits

that pre-stressed microfi laments and microtubules, structural

components of the cytoskeleton, produce tension in the under-

lying network that ultimately yields a stable structure which,

in part, serves as a physical signal transduction pathway. 89,90

Tensegrity, or tensional integrity, is a concept in which the

mechanical stability of a structure, in this case the cytoskeleton,

is structurally stable through the appropriate combination

of components that are either under tension or compression

Figure 4. (a) Thermal effect of cell culture substrate transition from a grooved surface to

a fl at surface using an SMP network. After cells were cultured for 9.5 h at 30°C, the

substrate was triggered by placing a sample in a 37°C incubator. The traces below the

image represent a cross-section of the contours of the culture surface. (b) The change in

substrate topography controls cell behavior. Confocal images of cells stained with

phalloidin on a temporary grooved structure have microstructures in alignment with

the groove direction (white arrow). After the surface transitions, the cells rearrange, are

randomly oriented, and show stress fi ber formation. Scale bar in (b) is 100 μ m. Reprinted

with permission from Reference 80 . ©2011, Elsevier.



and assembled in a coordinated manner. A complementary

theory posed by Charras and Horton suggested that tensegrity

is maintained by intermediate fi laments, cytoskeleton protein

structures that have a characteristic size that is between micro-

fi laments and microtubules, 91 which is supported by relevant

observations.92,93 Another theory posed by Forgacs suggested

that critical percolation of actin fi laments is essential for mecha-

notransduction. 94,95 These theories have been collectively tested

in the context of constrained cell shape, 6,96 altered substrate

stiffness,97,98 or deformed substrates.99

Cell mechanics is an important design criterion in the

broader effort to transform two-dimensional cell culture models

into more biologically relevant three-dimensional constructs.

This overarching strategy has infl uenced the design of many

bioactive materials, including porous polymer networks and

hydrogels. A parallel strategy may also be devised for the design

of dynamic three-dimensional networks that are able to control

the spatiotemporal presentation of cell signaling proteins and

peptides. Recent efforts have been predicated on repurposing

reversible coupling chemistries that can be applied to control

molecular topology and the extent of cross-linking in the

context of controlling the elasticity of synthetic ECM. Similar

approaches can also be used to control the presentation of

bound ligands for controlled adhesion and integrin activa-

tion. This section intentionally omits the vast literature on

ionic cross-linking of networks and phase change materials,

including stimulus-responsive hydrogels, which have been

extensively reviewed elsewhere. 100,101 Instead, a dedicated

focus is granted toward modular covalent coupling in the

context of precisely controlling the mechanical properties

of the network and the ability to present chemical cues with

spatiotemporal specifi city.

Reversible cross-linking Controlling the cross-linking density in polymeric networks is

a convenient method to modulate the bulk mechanical proper-

ties of synthetic ECM analogues. Furthermore, networks con-

sisting of hydrogels or elastomers are able to achieve elastic

moduli that are similar to a wide range of naturally occurring

tissues.102 The ideal design criteria in engineering reversible

polymer networks are as follows: (1) Hyperbranched network

precursors are functionalized with homobifunctional or com-

plementary bifunctional cross-linking moieties; (2) the chosen

cross-linking chemistries are reversible; and (3) forward and

reverse cross-linking reactions are both inducible through the

application of mild stimuli.

There is a wide range of reversible coupling chemistries that

are suitable for modulating the bulk properties of polymer net-

works. Diels–Alder cycloaddition reactions provide an effi cient

[2+4] cycloaddition mechanism. Diels–Alder coupling using

furan-based dienes and maleimide-based dienophiles has been

used as a strategy for the room temperature cross-linking of

hydrogels.103,104 One potential limitation in applying reversible

Diels–Alder chemistry is the large temperature window that is

required for promoting forward and reverse coupling reactions.

The kinetics of forward Diels–Alder cycloaddition is somewhat

slow at room temperature, while retrocycloaddition reactions

require temperatures of 70°C and higher, depending upon the

composition of the substituents and the solvent environment.

Cinnamates, a class of aromatic molecules based on cin-

namic acid, represent an alternative chemistry that is capa-

ble of reversible coupling via photoexcitation. 105 Cinnamate

derivatives undergo reversible photodimerization upon UV

irradiation. 106 Dimerization forms a cyclobutane ring by [2+2]

photocycloaddition at wavelengths greater than 260 nm,

while photocleavage of the cyclobutane ring is possible by

irradiation at a wavelength less than 260 nm. Hyperbranched

pre-polymers derivatized with pendant cinnamate can pro-

duce reversibly photo-cross-linkable networks. 105 There are

several advantages of synthesizing networks using precursors

that feature cinnamate functionalization. First, cinnamates are

a class of naturally occurring aromatic compounds that are

found in some fruits. Second, photodimerization and cleavage

can be controlled in two dimensions through simple masking

techniques. Third, coupling occurs through homodimerization,

which can simplify the synthesis of network precursors. Despite

these advantages, the high intensities and low wavelengths

used in reversible photodimerization typically preclude the

incorporation of viable cell populations. There are numerous

other reconfi gurable cross-linking chemistries, including those

based on isocyanates/imidazole coupling and free radical poly-

merization of thiolene compounds. This class of cross-linking

chemistry is broadly categorized as covalent adaptive networks

(CAN), which are described in more detail elsewhere. 107 Despite

the potential capabilities in utilizing CAN as programmable

polymer networks and interfaces, there are several practical

processing limitations that can preclude the widespread appli-

cation of these materials for studying dynamic cell-biomaterial

interactions in real time.

Dynamic 3D environments The advantages of reconfi gurability in covalent networks must

be balanced with the benefi ts of in situ manipulation of cell-

seeded networks. Hydrogels offer an attractive platform to cre-

ate covalent networks that are capable of presenting dynamic

three-dimensional cellular microenvironments. Specifi cally,

light-activated networks offer an opportunity to control the

spatiotemporal presentation of cues that cells can recognize

in real time. The core material property that has been utilized

in a number of demonstrations is the ability to modify poly-

mer side chains and backbones using photolysis. Under this

scheme, the single- ( λ = 365 nm) or two-photon ( λ = 740 nm)

absorption of select aromatic species can initiate rearrange-

ment reactions that uncage molecules or modify the properties

of polymer networks. For example, photo-induced rearrange-

ment in polymer backbones can be used for the instantaneous

decomposition of synthetic polymers in response to benign

levels of IR irradiation. 108 In this study, pendant photosen-

sitizers based on 4-bromo-7-hydroxycoumarin protecting

groups are incorporated within linear polyester precursors.



Upon excitation at λ = 740 nm, a cascade of cyclization

and rearrangement reactions proceeds, which leads to sys-

temic degradation of the network. This rapid degradation

reaction yields relatively non-toxic low-molecular weight

by-products, including adipic acid.

Photodeprotection chemistries generally describe the ability

to cleave a portion of a molecule in a selective manner using

light. Photodeprotection can serve as an effi cient method to

selectively alter the compositions of side chains in polymeric

materials. Functional groups containing pendant 2-nitrobenzyl

groups via ester linkages can be cleaved to produce the

nascently deprotected carboxylic acid and the newly formed

2-nitrobenzaldehyde by-product. Variations on this chemistry

have also been used for three-dimensional patterning of recog-

nition molecules in hydrogels via single- or two-photon pattern-

ing of static networks.109,110 This synthetic strategy can also be

used to create dynamic hydrogel biomaterial networks. Anseth

et al. demonstrated the formation of photolyti-

cally cleavable cell-seeded hydrogel networks

based on poly(ethylene glycol) (PEG) func-

tionalized with 2-nitrobenzyl groups. 111 Briefl y,

linear PEG precursors are functionalized with

acrylate-containing 2-nitrobenzyl groups via

ester linkages. These photolabile PEG pre-

cursors form cross-linked hydrogel networks.

The storage modulus of the networks, which is

directly proportional to the cross-link density,

can be reduced irreversibly by over 90% after

exposure to benign UV irradiation of λ = 365 nm

at 10 mW/cm 2 for 10 minutes. Therefore, the

mechanical properties of the network can be

dynamically tuned in space and time through

controlled irradiation. A parallel synthetic

approach can be used to create networks in

which cell adhesion domains can be selectively

removed using photolysis. Human mesenchy-

mal stem cells cultured in PEG hydrogels with

photocleavable cell-binding domains engage

in accelerated chrondrogenesis upon selective

removal of this integrin binding motif. Similar

chemistries can be utilized for real time control

of elasticity in hydrogel substrates as well. 112,113

Practical aspects of these techniques are sum-

marized in Figure 5 . Figure 5a is a schematic

of the formation of a photodegradable polymer

network through free radical photopolymer-

ization. The green groups represent dienes

that can undergo free radical polymerization

to create the cross-linked network. The blue

groups represent photodegradable segments

that can be cleaved to selectively remove parts

of a network. One strategy for selective photo-

cleavage of cross-linked networks uses masking

combined with single-photon excitation to create

two-dimensional structures in a manner that is

similar to photolithography ( Figure 5b ). Another strategy uses

two-photon absorption in which laser scanning microscopy can

be used to create arbitrary three-dimensional features by selective

photocleavage ( Figure 5c ).

Light-activated hydrogel networks are advantageous because

direct manipulation of cell-biomaterial interactions is achiev-

able, although irreversible. Nonetheless, there is a broad set

of biological questions that can be answered using material

systems that are able to administer step changes in biomaterial

response. The emerging challenge is to design materials, sys-

tems, and experiments that are able to eventually provide insight

into disease progression such that complementary treatments

may eventually be discovered.

Conclusions and outlook The continued interest in dynamic cell-biomaterials interac-

tions using smart materials will lead to discoveries that could

Figure 5. Methods for photodegradable hydrogel synthesis and patterning.

(a) Photodegradable hydrogels formed by redox-initiated, free-radical chain polymerization,

which is indicated by R*. The resultant network is comprised of polyacrylate kinetic chains*(green) connected by PEG-based cross-links (black) with photocleavable moieties (blue) in

the backbone. (b) Using a chrome mask, the surface of the photocleavable hydrogel can

be patterned with features of depths that depend upon exposure time. Features can be

generated within minutes of irradiation time. (c) Use of a laser-scanning microscope (LSM)

with a 405 nm laser can also be used to create three-dimensional patterns by precisely

moving the laser in the x–y plane using a scanner. The z-dimension is controlled with ayhigher minimum feature resolution by either controlling the focal plane of the laser or

the location of the stage. Reprinted with permission from Reference 112 . ©2010, Nature

Publishing Group.



resolve key aspects of cytoskeleton function in a variety of

model systems such as neurons, cardiomyocytes, and stem cells.

In addition to fundamental scientifi c discovery, this approach

may be utilized in application-based studies such as in vitro

disease models or assays for drug development. Despite the

proliferation of materials and the rapidly growing set of

fabrication capabilities, the fi eld of dynamic materials for

elucidating cell responses remains in its infancy. The utility

of these systems continues to expand with help from parallel

advances in materials design strategies. Next-generation

dynamic materials will gradually incorporate cues beyond

traditional fi rst-order engineering properties such as surface

energy, feature topography, or elastic modulus. For exam-

ple, a dedicated focus may be the control of dynamic cell-

cell interactions. Preliminary examples exist, but they will

benefi t from continued refi nement.114 Prospective dynamic

systems will require new materials with increased sophisti-

cation in polymer synthesis and macromolecular engineering

approaches.

The prospect for increasing complexity in these systems

must be balanced with the practical utility of such inventions.

There may be reduced scientifi c merit and unrealized down-

stream clinical benefi t by arbitrarily increasing the complexity

of dynamic cell culture materials without regard to the ease of

widespread adoption. One example of a technology that has

successfully navigated this tradeoff is soft lithography. Micro-

fl uidics and micropatterning in biomaterials have been widely

adopted as tools by both biomedical engineers and biologists

alike. This prominent success story can provide a roadmap

for the adoption of future technologies that are predicated on

dynamic materials. Traction of early-stage demonstrations

can be accelerated through two types of research environments:

(1) strategic collaborations between materials engineers and

biomedical researchers who are early adopters of innovative

techniques; and (2) laboratories that are able to comprehen-

sively integrate the design, fabrication, and validation of

dynamic materials in model biological systems under one roof.

Taken together, the prospects for dynamic biomaterials remain

bright and have the potential to enable high impact discoveries

in numerous fi elds ranging from the design of medical implants

to engineering immunotherapies.

Acknowledgments Funding was provided by the following organizations: the

Berkman Foundation; the American Chemical Society Petro-

leum Research Fund (ACS PRF #51980-DNI7); the Proctor &

Gamble Education Grant Program; and the Carnegie Mellon

University School of Engineering.

References 1. N.P. Ziats , K.M. Miller , J.M. Anderson , Biomaterialss 9 , 5 ( 1988 ). 2. J.M. Anderson , A. Rodriguez , D.T. Chang , Semin. Immunol. 20 , 86 ( 2008 ). 3. G. Voskerician , M.S. Shive , R.S. Shawgo , H.v. Recum , J.M. Anderson , M.J. Cima , R. Langer , Biomaterialss 24 , 1959 ( 2003 ). 4. S. Chen , J.A. Jones , Y. Xu , H.-Y. Low , J.M. Anderson , K.W. Leong , Biomaterials31 , 3479 ( 2010 ). 5. J. Anderson , A. McNally , Semin. Immunopathol. 33 , 221 ( 2011 ).

6. C.S. Chen , M. Mrksich , S. Huang , G.M. Whitesides , D.E. Ingber , Sciencee 276, 1425 ( 1997 ). 7. E.K.F. Yim , R.M. Reano , S.W. Pang , A.F. Yee , C.S. Chen , K.W. Leong ,Biomaterialss 26 , 5405 ( 2005 ). 8. E.K.F. Yim , S.W. Pang , K.W. Leong , Exp. Cell Res. 313, 1820 ( 2007 ). 9. J.L. Charest , M.T. Eliason , A.J. García , W.P. King , Biomaterialss 27, 2487 ( 2006 ). 10. Y. Xia , G.M. Whitesides , Angew. Chem. Int. Ed. 37 , 550 ( 1998 ).11. C.J. Lee , M.S. Blumenkranz , H.A. Fishman , S.F. Bent , Langmuirr 20 , 4155 ( 2004 ). 12. D.C. Duffy , J.C. McDonald , J.A. Schueller , G.M. Whitesides , Anal. Chem. 70, 4974 ( 1998 ). 13. A.J. Engler , S. Sen , H.L. Sweeney , D.E. Discher , Cell 126, 677 ( 2006 ). 14. C.J. Bettinger , B. Orrick , A. Misra , R. Langer , J.T. Borenstein , Biomaterials27 , 2558 ( 2006 ). 15. M.J. Dalby , M.O. Riehle , H. Johnstone , S. Affrossman , A.S.G. Curtis ,Biomaterialss 23, 2945 ( 2002 ). 16. C. Oakley , D.M. Brunette , J. Cell Sci. 106 , 343 ( 1993 ).17. G.A. Abrams , S.L. Goodman , P.F. Nealey , M. Franco , C.J. Murphy , Cell Tissue Res. 299 , 39 ( 2000 ).18. A.J. García , D. Boettiger , Biomaterialss 20 , 2427 ( 1999 ).19. T.A. Petrie , J.R. Capadona , C.D. Reyes , A.J. García , Biomaterialss 27 , 5459 ( 2006 ). 20. C.-M. Lo , H.-B. Wang , M. Dembo , Y.-l. Wang , Biophys. J. 79, 144 ( 2000 ). 21. R.G. Harrison , J. Exp. Zool. 9 , 787 ( 1910 ). 22. P. Clark , P. Connolly , A. Curtis , J. Dow , C.D. Wilkinson , J. Cell Sci. 99, 73( 1991 ). 23. R.G. Flemming , C.J. Murphy , G.A. Abrams , S.L. Goodman , P.F. Nealey ,Biomaterialss 20 , 573 ( 1999 ).24. C. Bettinger , R. Langer , J. Borenstein , Angew. Chem. 48, 5406 ( 2009 ). 25. A. Bernard , J.P. Renault , B. Michel , H.R. Bosshard , E. Delamarche , Adv.Mater. 12, 1067 ( 2000 ). 26. C.J. Bettinger , E.J. Weinberg , K.M. Kulig , J.P. Vacanti , Y. Wang , J.T. Borenstein , R. Langer , Adv. Mater. 18 , 165 ( 2006 ). 27. C.J. Bettinger , K.M. Kulig , J.P. Vacanti , R. Langer , J.T. Borenstein , Tissue Eng. 15 , 1321 ( 2009 ). 28. G.C. Engelmayr , M. Cheng , C.J. Bettinger , J.T. Borenstein , R. Langer , L.E. Freed ,Nat. Mater. 7 , 1003 ( 2008 ). 29. C.-H. Choi , S.H. Hagvall , B.M. Wu , J.C.Y. Dunn , R.E. Beygui , C.-J. Kim , Biomaterialss 28, 1672 ( 2007 ). 30. S. Gerecht , C.J. Bettinger , Z. Zhang , J. Borenstein , G. Vunjak-Novakovic , R. Langer , Biomaterialss 28, 4068 ( 2007 ). 31. C.J. Bettinger , Z. Zhang , S. Gerecht , J.T. Borenstein , R. Langer , Adv. Mater.20 , 99 ( 2008 ). 32. K.A. Diehl , J.D. Foley , P.F. Nealey , C.J. Murphy , J. Biomed. Mater. Res. 75A, 603 ( 2005 ). 33. M.J. Dalby , N. Gadegaard , R. Tare , A. Andar , M.O. Riehle , P. Herzyk , C.D.W. Wilkinson , R.O.C. Oreffo , Nat. Mater. 6, 997 ( 2007 ). 34. J.D. Foley , E.W. Grunwald , P.F. Nealey , C.J. Murphy , Biomaterialss 26, 3639 ( 2005 ). 35. S. Patel , K. Kurpinski , R. Quigley , H. Gao , B.S. Hsiao , M.M. Poo , S. Li , Nano Lett. 7 , 2122 ( 2007 ). 36. M.L. Gardel , I.C. Schneider , Y. Aratyn-Schaus , C.M. Waterman , Annu. Rev. Cell Dev. Biol. 26, 315 ( 2010 ). 37. R. Geiger , W. Taylor , M. Glucksberg , D. Dean , Gene Ther. 13 , 725 ( 2006 ). 38. H. Wang , W. Ip , R. Boissy , E.S. Grood , J. Biomechan. 28 , 1543 ( 1995 ). 39. S. Jungbauer , H. Gao , J.P. Spatz , R. Kemkemer , Biophys. J. 95, 3470 ( 2008 ). 40. J. Lahann , S. Mitragotri , T.-N. Tran , H. Kaido , J. Sundaram , I.S. Choi , S. Hoffer , G.A. Somorjai , R. Langer , Sciencee 299, 371 ( 2003 ). 41. X. Jiang , D.A. Bruzewicz , A.P. Wong , M. Piel , G.M. Whitesides , Proc. Natl. Acad. Sci. USA 102, 975 ( 2005 ). 42. A. Gumus , J.P. Califano , A.M.D. Wan , J. Huynh , C.A. Reinhart-King , G.G. Malliaras ,Soft Matterr 6 , 5138 ( 2010 ). 43. C.J. Bettinger , J.P. Bruggeman , A. Misra , J.T. Borenstein , R. Langer ,Biomaterialss 30 , 3050 ( 2009 ). 44. M. Muskovich , C.J. Bettinger , Adv. Healthcare Mater. 1 , 248 ( 2012 ). 45. J.H. Slater , J.S. Miller , S.S. Yu , J.L. West , Adv. Funct. Mater. 21, 2876 ( 2011 ). 46. H.S. Lim , J.T. Han , D. Kwak , M. Jin , K. Cho , J. Am. Chem. Soc. 128, 14458 ( 2006 ). 47. M.J. Dalby , M.O. Riehle , D.S. Sutherland , H. Agheli , A.S.G. Curtis , Eur. J. Cell Biol. 83 , 159 ( 2004 ). 48. M.J. Dalby , M.J.P. Biggs , N. Gadegaard , G. Kalna , C.D.W. Wilkinson , A.S.G. Curtis ,J. Cell. Biochem. 100 , 326 ( 2007 ). 49. L. Peng , M.L. Eltgroth , T.J. LaTempa , C.A. Grimes , T.A. Desai , Biomaterials30 , 1268 ( 2009 ). 50. R.D. Campbell , B.A. Marcum , J. Cell Sci. 41, 33 ( 1980 ). 51. N. Nakatsuji , Amer. Zool. 24, 615 ( 1984 ).



52. D.F. Newgreen , Dev. Biol. 131, 136 (1989). 53. K. Wolf , R. Muller,r S. Borgmann, E.B. Brocker ,r P. Friedl, Blood 102, 3262(2003). 54. P. Weiss , J. Exp. Zool. 100 , 353 (1945).55. G.A. Dunn, J.P. Heath, Exp. Cell Res.l 101 , 1 (1976).56. A.-S. Andersson, P. Olsson, U. Lidberg, D. Sutherland, Exp. Cell Res.l 288,177 ( 2003 ).57. E. Rebollar,r I. Frischauf , M. Olbrich , T. Peterbauer ,r S. Hering , J. Preiner ,rP. Hinterdorfer ,r C. Romanin, J. Heitz, Biomaterials 29 , 1796 (2008).58. J.Y. Lim , H.J. Donahue , Tissue Eng.e 13 , 1879 (2007).59. S. Lenhert , M.-B. Meier ,r U. Meyer,r L. Chi , H.P. Wiesmann, Biomaterials 26, 563 ( 2005 ). 60. J.-P. Kaiser,r A. Reinmann, A. Bruinink , Biomaterials 27 , 5230 ( 2006 ). 61. T. Tzvetkova-Chevolleau , A. Stéphanou , D. Fuard, J. Ohayon , P. Schiavone , P. Tracqui, Biomaterials 29, 1541 (2008). 62. E. Martínez , E. Engel , J.A. Planell , J. Samitier,r Ann. Anat. 191, 126 (2009). 63. E.A. Cavalcanti-Adam , T. Volberg, A. Micoulet , H. Kessler ,rr B. Geiger,r J.P. Spatz ,Biophys. J. 92, 2964 ( 2007 ). 64. L. Chou, J.D. Firth, V.J. Uitto , D.M. Brunette , J. Cell Sci.l 108 , 1563 (1995).65. I. Nagata , A. Kawana , N. Nakatsuji, Development 117, 401 (1993).66. M.D. Rosenberg , Proc. Natl. Acad. Sci. USA 48, 1342 ( 1962 ).67. J.A. Champion , Y.K. Katare, S. Mitragotri , J. Control. Release 121, 3 (2007).68. G. Albrecht-Buehler,r J. Cell Biol.l 69, 275 ( 1976 ). 69. A. Carrel, M.T. Burrows , J. Exp. Med. 13, 571 (1911). 70. P. Weiss, B. Garber,r Proc. Natl. Acad. Sci. USA 38 , 264 ( 1952 ). 71. M.J. Dalby,y A. Hart, S.J. Yarwood, Biomaterials 29 , 282 (2008).72. F. Johansson, P. Carlberg, N. Danielsen, L. Montelius , M. Kanje , Biomaterials27 , 1251 ( 2006 ). 73. C. Oakley,y D.M. Brunette, Cell Motil.l Cytoskel. 31 , 45 ( 1995 ).74. X. Jiang , S. Takayama , X. Q, E. Ostuni , H. Wu , N. Bowden , P. LeDuc , D.E. Ingber,rr G.M. Whitesides , Langmuir 18, 3273 ( 2002 ). 75. J.L. Tan , J. Tien , D.M. Pirone , D.S. Gray,y K. Bhadriraju , C.S. Chen , Proc. Natl.Acad. Sci. USA 100, 1484 ( 2003 ). 76. Z. Li , J. Song, G. Mantini , M.-Y. Lu, H. Fang , C. Falconi , L.-J. Chen , Z.L. Wang,Nano Lett.o 9, 3575 ( 2009 ). 77. O. du Roure , A. Saez, A. Buguin, R.H. Austin, P. Chavrier ,rr P. Siberzan , B. Ladoux ,Proc. Natl. Acad. Sci. USA 102, 2390 ( 2005 ). 78. C.-C. Fu, A. Grimes, M. Long, C.G.L. Ferri , B.D. Rich , S. Ghosh, S. Ghosh,L.P. Lee, A. Gopinathan , M. Khine , Adv. Mater. 21 , 4472 (2009). 79. A. Chen , D.K. Lieu, L. Freschauf , V. Lew , H. Sharma , J. Wang, D. Nguyen ,I. Karakikes , R.J. Hajjar,rr A. Gopinathan , E. Botvinick ,k C.C. Fowlkes, R.A. Li , M. Khine,Adv. Mater. 23 , 5785 (2011). 80. K.A. Davis, K.A. Burke , P.T. Mather ,r J.H. Henderson: Dynamic cell behavioron shape memory polymer substrates. Biomaterials 32, 2285 ( 2011 ). 81. M.T. Lam , W.C. Clem, S. Takayama, Biomaterials 29 , 1705 (2008). 82. J.H.C. Wang , P. Goldschmidt-Clermont , J. Wille, F.C.P. Yin, J. Biomech. 34 , 1563 (2001). 83. X. Zhu , K.L. Mills , P.R. Peters, J.H. Bahng , E.H. Liu, J. Shim, K. Naruse, M.E. Csete,M.D. Thouless, S. Takayama, Nat. Mater. 4 , 403 (2005). 84. R. McBeath, D.M. Pirone, C.M. Nelson , K. Bhadriraju , C.S. Chen, Dev. Cell6 , 483 (2004). 85. S. Suresh, J. Spatz, J.P. Mills, A. Micoulet, M. Dao , C.T. Lim , M. Beil, T. Seufferlein ,Acta Biomatera 1, 15 ( 2005 ). 86. S. Suresh, Acta Materialiaa 55, 3989 ( 2007 ).87. S.E. Cross , Y.-S. Jin , J. Rao, J.K. Gimzewski, Nat. Nanotechnol. 2 , 780( 2007 ).88. M.J. Rosenbluth , W.A. Lam, D.A. Fletcher,r Lab Chipb 8, 1062 ( 2008 ). 89. D.E. Ingber,r J. Cell Sci.l 104 , 613 (1993). 90. D.E. Ingber,r J. Cell Sci.l 116 , 1157 (2003). 91. G.T. Charras, M.A. Horton, Biophys. J. 82, 2970 ( 2002 ). 92. H. Hotani , H. Miyamoto, Adv. Biophys. 26 , 135 (1990).93. F.C. MacKintosh , J. Kas , P.A. Janmey,y Phys. Rev. Lett. 75 , 4425 (1995).94. G. Forgacs, J. Cell Sci.l 108, 2131 ( 1995 ).95. Y. Shafrir ,r G. Forgacs , Am. J. Physiol. Cell Physiol.l 282 , C479 (2002).

96. K.K. Parker ,r A.L. Brock, C. Brangwynne, R.J. Mannix , N. Wang, E. Ostuni ,N.A. Geisse, J.C. Adams, G.M. Whitesides, D.E. Ingber,r FASEB JB 16 , 1195 ( 2002 ). 97. D.E. Discher,r P. Janmey,y Y.-l. Wang , Science 310 , 1139 (2005). 98. T. Yeung, P.C. Georges , L.A. Flanagan , B. Marg, M. Ortiz , M. Funaki , N. Zahir,r W. Ming, V. Weaver,r P.A. Janmey,y Cell Motil.l Cytoskel. 60, 24 ( 2005 ). 99. C.-M. Cheng, R.L. Steward Jr., P.R. LeDuc , J. Biomech. 42 , 187 (2009).100. P. Gupta, K. Vermani , S. Garg , Drug Discov.g Today 7 , 569 (2002).101. N.A. Peppas , W. Leobandung, Journal ofl Biomaterialsf Science,s PolymerEdition 15, 125 ( 2004 ).102. C.J. Bettinger,r Pure Appl.e Chem. 83, 9 ( 2011 ).103. I. Kosif, E.-J. Park, R. Sanyal , A. Sanyal , Macromolecules 43, 4140 ( 2010 ). 104. C.M. Nimmo , S.C. Owen, M.S. Shoichet , Biomacromolecules 12, 824 (2011).105. D. Shi , M. Matsusaki , T. Kaneko , M. Akashi , Macromolecules 41, 8167( 2008 ).106. P. Pinazzi , A. Fernandez , Ultraviolet Lightt Inducedt Reactionsd ins Polymersn( American Chemical Society,y Washington, DC , 1976 ), p. 37.107. C.J. Kloxin, T.F. Scott , B.J. Adzima, C.N. Bowman, Macromolecules 43 ,2643 ( 2010 ).108. N. Fomina , C.L. McFearin, M. Sermsakdi , J.M. Morachis, A. Almutairi ,Macromolecules 44 , 8590 (2011).109. Y. Luo , M.S. Shoichet , Nat. Mater. 3 , 249 (2004).110. R.G. Wylie , M.S. Shoichet , J. Mater. Chem. 18, 2716 ( 2008 ).111. A.M. Kloxin, A.M. Kasko, C.N. Salinas , K.S. Anseth , Science 324 , 59 (2009).112. A.M. Kloxin, M.W. Tibbitt , K.S. Anseth , Nat. Protoc. 5 , 1867 (2010).113. A.M. Kloxin, J.A. Benton, K.S. Anseth, Biomaterials 31 , 1 (2010).114. E.E. Hui , S.N. Bhatia, Proc. Natl. Acad. Sci. USA 104 , 5722 (2007).

Stephen Kustra received his master of sciencedegree in biomedical engineering at CarnegieMellon University and his bachelor of sciencedegree in biomedical engineering from theUniversity of Connecticut. His current researchinterests include biomaterials-based medicaldevices and smart interfaces for understandingfundamentals of cell-biomaterials interactions.Kustra can be reached by email at [email protected].

Christopher J. Bettinger isr an assistant professorat Carnegie Mellon University (CMU) in theDepartments of Materialsf Science and BiomedicalEngineering. He directs the laboratory forBiomaterials-Based Microsystems and Electronicsat CMU, which is broadly interestedy in the designof novel materials and interfaces that promotethe integration of medical devices with thehuman body. Bettinger has received manyhonors, is a co-inventor on several patents, andwasafi nalistfi in theMIT $100KEntrepreneurshipCompetition. Bettinger received an SB degree inchemical engineering, an M.Eng. degree inbiomedical engineering, and a PhD degree in

materials science and engineering as a Charles Stark Draper Fellow, all from theMassachusetts Institute of Technology. He completed his post-doctoralfellowship at Stanford University in the Department of Chemical Engineering asan NIH Ruth Kirschstein Fellow. Bettinger can be reached by email at [email protected].

Frontiers in Thin-Film Epitaxy and Nanostructured Materials JMR Special Focus Issue, July 2013

www.mrs.org/jmr-focus

CALL FOR PAPERSSubmission Deadline November 15, 2012

847MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin© 2012 Materials Research Society

Introduction From the early 1900s, it hast been known that stressest induced

during or postr thin-fit lmfi depositionm cann causen the fi lmsfi to curve.1

Since thin-filmfi stressm can causen delamination, cracking, and pre-d

mature failure of multilayerf devicesr such as integrated circuits,d

a large effort hast been directed atd developingt deposition and

processing methods that minimizet thin-filmfi stress. However,

it has recently become evident that thin-filmfi stresses can

in fact be engineered to shape three-dimensional patterned

micro- and nanostructures with a variety of materialf compo-

sitions, including metals, semiconductors, and polymers.d These

structures that providet new capabilities in electronics, optics,

and medicine can be challenging to fabricate using conven-

tional bottom-up or top-downr techniques. For example,r Prinz

et al.2 reported thed formation of nanotubesf with inner diam-r

eters as small as 2 nm by releasing heteroepitaxially strained

InGaAs/GaAs ultrathin semiconductor films.fi Similarly, so

called roll-upd structures with micro- to nanoscale radii have

been fabricated byd stress engineering of af variety of materials,f

including other semiconductors such as SiGe;3 metals such

as MoCr alloys,r 4 chromium,5,6 and tin;d 7 oxides such as SiOx;8

and polymersd such as chitosan/poly(PEGMA-co-PEGDMA),9

polystyrene/poly(4-vinylpyridine),10 polysuccinimide/

polycaprolactone,11 and differentiallyd cross-linkedy SU8.d 12 Roll-up

structures have enabled newd functionalities for electronics,r 13

optics,14,15 sensing,16,17 microfluidics,fl 12 energy harvesting

and storage,18,19 drug delivery,11,20 tissue engineering,21 and

robotics.22

Stresses can also be engineered within localized regions of

thin fi lmsfi so that they function like hinges to enable out-of-

plane rotation. When these hinges are patterned between rigid

panels, they enable a hands-free origami approach that can be

used tod create three-dimensional micro- and nanostructures.d 23,24

For example,r Syms et al. described the use of surfacef tension

forces in molten solder to perform out-of-plane rotation of

polysilicon fl aps.fl 25 Here, solder wasr lithographically patterned

on a hinge material, such as Au or ar polymer, and liquefiedfi

by heating, causing it to deform to reduce surface energy;

this deformation generated the torque required to rotate the

fl ap.fl Building on the body of literaturef on electromechanical

actuation,26 Smela et al. showed electrochemically controlled

bending and foldingd of microstructures,f such as spirals or cubicr

boxes using bilayer strips or hinges of Auf and polypyrrole

(PPy, doped with sodium dodecyl benzene sulfonate).27,28 In

this case, the Au/PPy strips curved when the PPy thin filmfi

was electrochemically oxidized, causing it to shrink. In addi-

tion, a number ofr activef and passive mechanisms have been

explored, including the use of electrical,f magnetic, electro-

chemical, optical, pneumatic, thermal, and chemical stimuli

to manipulate the stresses in single or multilayerr fir lmsfi so that

they cany ben curved ord foldedr onlyd wheny desired.n 24 Some of thesef

stimuli require that the structures are tethered to substrates,

Self-folding thin-fi lm materials: From nanopolyhedra to graphene origami Vivek B.k Shenoy and David H. Gracias

Self-folding of thinf filmsfi is a more deterministic form of self-assemblyf wherein structures

curve or fold up either spontaneously on release from the substrate or in response to specificfi

stimuli. From an intellectual standpoint, the study of thef self-folding of thinf filmsfi at small size

scales is motivated by the observation that a large number of naturallyf occurring materials

such as leaves and tissues show curved, wrinkled, or folded micro- and nanoscale geometries.

From a technological standpoint, such a self-assembly methodology is important since it

can be used to transform the precision of existing planar patterning methods, such as

electron-beam lithography, to the third dimension. Also, the self-folding of graphenef promises

a means to create a variety of three-dimensionalf carbon-based micro- and nanostructures.

Finally, stimuli responsive self-folding can be used to realize chemomechanical and tether-

free actuation at small size scales. Here, we review theoretical and experimental aspects of

the self-folding of metallic,f semiconducting, and polymeric fi lms.fi

Vivek B. Shenoy, Engineering Department, Brown University; [email protected] David H. Gracias, Departments of Chemical and Biomolecular Engineering , Chemistry and the Institute for Nanobiotechnology,y The Johns Hopkins University ; [email protected] DOI: 10.1557/mrs.2012.184

SELF-FOLDING THIN-FILM MATERIALS: FROM NANOPOLYHEDRA TO GRAPHENE ORIGAMI


while others are tether-free; some mechanisms are reversible,

while others are not.

While the folding of microstructuresf has become well estab-

lished, only very recently was the concept appliedt tod the self-

folding of 100-nm-sizedf (referred to as nanoscale origami)

metallic and dielectricd polyhedra.29–31 These studies have shown

how the precision of e-beamf lithography can be transferred tod

three-dimensional patterning. Simultaneously, origami-inspired

assembly of complexf structures such as truncated octahedra,d 32

self-folding sheets,33 and functionald structures such as silicon

microscanners,34 electrochemical capacitors,35 three-axis

sensors,36 THz metamaterials,37 3D solar devices,r 38 polymeric

capsules,39,40 lab-on-a-chip devices,41,42 and surgical tools43,44

have been realized.

Recently, the idea of origami-inspiredf fabrication with

graphene has been described, and preliminary experimental

results suggest feasibilityt of suchf an approach.45–48 Graphene,

an atomically thin two-dimensional layer of carbonf atoms,

has received attentiond for itsr unique electronic, thermal, and

mechanical properties. While graphene has the highest in-planet

Young’s modulus of anyf materialy measured, it cant ben easily benty

and foldedd justd liket a sheet oft fabricf or paper.r Folded graphened

shows new and distinctived properties. For example,r recent stud-t

ies suggest thatt foldedt graphened exhibits interesting electronic

properties due to the interplay between the gauge fi eldsfi created

by the fold and the external fields.fi 49,50 Carbon atoms at the

folds, warps, and creasesd are generally more reactive, making

folded structuresd attractive for applicationsr in catalysis, energy

storage, and sensing.d

It ist important tot note that self-foldingt of organicf molecular

chains is also an active area of researchf with important focust

areas such as protein folding, molecular foldamers,r 51 and DNAd

origami.52 It ist also noteworthy that theret may be similarities

between the geometric design rules and foldingd pathways of

self-folding molecular chainsr and thosed of patternedf thin-fid lmfi

self-folding structures wherein units interact throught a variety

of primaryf and secondaryd interactions.32 As for practicalr appli-

cations, while the curving and foldingd of molecularf chainsr has

resulted in the creation of complexf 3D structures, including

polyhedra,53 they can be constructed with a somewhat lim-

ited set of organicf molecules. Further, it can be challenging

to incorporate specificfi three-dimensional surface patterns and

integrate different material classes as would be required for

electronic circuits or opticalr modules. A detailedA descriptiond

of organicf molecule folding paradigms is beyond thed scope of

this article, which is focused ond a review of thef mechanisms

and applicationsd of curvingf and foldingd thin filmsfi composed ofd

metals, semiconductors, dielectrics, polymers, and gels.d

The mechanics of curving and folding In order forr ar thin filmfi to curve or foldr outd oft thef plane, there

must bet a bending moment ort torquer that causest it tot rotate.

Patterned thind fi lmsfi will bend ifd theref is an elastic or plasticr

differential stress along the thickness of thef fi lm,fi causing one

side of thef fi lmfi to contract ort extendr relatived to the other. The

stresses used haved originated fromd a variety of mechanisms,f

typically in bilayers, including thermal, electrochemical,

magnetic, heteroepitaxy, and capillarity.d The energy required

to bend ad filmfi is related tod its rigidity. For example,r according

to the theory of elasticity,f the bending energy is F,2

0

1 1,

2F = −k

dAR R

(1)

k isk the bending rigidity, and R and R0 are the radius to which theh

structure curves in the absence of deformation.f 54 This expres-

sion can be used tod derive the curvature due to residual stresses

in multilayer fir lms.fi

Stoney suggested a simple analysis to relate the amount

of stressf in a thin fi lmfi on a thick substratek to the amount oft

bending.1 This analysis has since been generalized tod the case

of layeredf 55–58 and compositionallyd gradedy fid lms.fi 59 For example,r

in order to estimate the curvature and the position of thef

neutral axis for ar multilayer systemr with different thickness

and stress,d one can utilizen models described byd Nikishkovy etv al.t 58

for freestandingr multilayers. Expressions for curvaturer in com-

positionally graded fid lmsfi can be found ind the book byk Freund

and Suresh.d 59 The previous analysis based ond small deforma-

tions has also been generalized tod the non-linear deformationr

regime. As the mismatch strainsh in bilayern fir lmsfi increase, axially

symmetric deformations in bilayern fir lmsfi require an extensionn ofn

the mid-plane in addition to bending. Since the stiffness asso-

ciated withd such deformations is much larger thanr the bending

stiffness, the system will begin a shape transformation toward

cylindrical bending deformation as an alternate shape that ist

energetically favorable,y which hash been observedn ind an numbera ofr

systems.60 In fi nite-sizedfi fid lmsfi undergoing large deformations,

the effect oft edgesf on the direction of bendingf or bifurcationr

has recently been analyzed usingd analytical methods.61 Large

strain can also lead tod plastic deformations in metallic bilayers

leading to complex folding patterns that havet been observed

experimentally and simulatedd usingd finitefi element methods.t 62

Curving patterned fi lms with micro- and nanoscale radii It is relatively straightforward to curve fi lmsfi with milli- to

micrometer scaler radii, and oned only needs to look atk peelingt

paint to observe this phenomenon. Indeed, curved struc-

tures with microscale radii can be formed usingd a variety of

materials,6,8,63 but itt ist much more challenging to curve struc-

tures with nanoscale radii due to the high stresses required tod

achieve such small radii. Curved nanotubesd with diameters as

small as ∼2 nm have been reported upon releasing ultrathin

fi lmsfi (1 ML GaAs:L 1 ML InAs)L of molecularf beamr epitaxially

(MBE) grown strained heterostructuresd (Figure 1a–c).2 How-

ever, MBE is not widelyt accessible,y and heteroepitaxiald curving

of nanostructuresf has been restricted to semiconductors.3,64

Recently, it wast discovered thatd surfacet forces associated withd

the reflowfl of lowf melting point materialst such as tin (Sn) on

nickel (Ni) or aluminar could be utilized to curve filmsfi with

radii as small as 20 nm (Figure 1 d).7 Apart from providing



a relatively facile means to curve structures

with nanoscale radii, this methodology was

combined with e-beam patterned structures

so that nanostructures could be formed not

only with nanoscale radii but also with 20 nm

resolved patterns on their curved surfaces.

The equi-biaxial stress originating from the

thermally driven coalescence of Sn on Ni was

numerically estimated to be very large, on the

order of 5.5 GPa, and is responsible for the

tight nanoscale radii. 62 It is noteworthy that

the roll-up of polymer fi lms has been reported

with radii as small as 100 nm.65 On the micro-

scale, thin fi lms have been curved up with a

variety of constituents typically by engineering

differentially stressed bilayers such as SiO/SiO 2( Figure 1 e), Cr based bilayers, or polymers

( Figure 1 f). In addition to bilayers, differential

stress can also be created within a single poly-

meric thin fi lm by generating a stress gradient

through differential photocross-linking. 12 Of

practical importance, an attractive feature of

curving a structure from planar fi lms is that

patterns with the resolution of planar litho-

graphic techniques can be precisely defi ned

on the curved surfaces of the rolled-up structure

( Figure 1 d, g, h). Hence, electromagnetic

modules, circuits, or microfl uidic channels

( Figure 1 h) can now be defi ned on curved geom-

etries with micro- to nanoscale radii. Another

attractive feature is that multilayered structures

of importance in energy storage 18 or tissue

engineering 21 ( Figure 1 i) can be readily formed.

Self-folding micro- and nanostructures As compared to curved geometries, structures

with discrete folds can be self-assembled by

patterning stressed regions within thin fi lms;

these regions function as “hinges.” A variety

of methods have been utilized, including pneu-

matics, magnetics, polymer shrinkage, elec-

trochemistry ultrasonic actuation, differential

stress, and surface forces, which are compre-

hensively reviewed by Leong et al. 24 These

strategies have been used to self-fold a variety

of functional structures, including inductors 23

( Figure 2a) and solar devices 38 ( Figure 2 b).

One of the concerns with self-folded closed-

form structures such as polyhedra is their lack

of mechanical rigidity, as would be required

for practical applications. The mechanical

rigidity and strength of folded structures can

be increased by utilizing self-aligning hinges

that seal the edges during self-folding. The

Figure 1. Curving patterned fi lms with micro- and nanoscale radii. (a) Schematic

illustration of the strain-induced bending of an InAs/GaAs bilayer after freeing it from

bonding with the substrate. (b–c) High-resolution transmission electron microscopy

images of InGaAs/GaAs nanotubes; the thickness of the self-scrolled bilayer: (b) 4 ML

GaAs + 4ML In x Gax 1− x− As ( x x(( = 0:6); (c) 2 ML GaAs+1 ML InAs. Reprinted with permissionxfrom Reference 2. ©2000, Elsevier. (d) E-beam patterned curved nanostructures formed

due to grain coalescence in thin fi lms during plasma etching of the underlying silicon

substrate. Reprinted with permission from Reference 7. ©2010, Wiley. (e) Scanning

electron microscopy top view (top inset) and cross-section of rolled-up SiO/SiO2

microtubes. Reprinted with permission from Reference 79. ©2012, Royal Society of

Chemistry. (f) Optical micrograph of a 3D hydrogel microstructure self-folded from a

cross-shaped microwell. Scale bar is 100 μ m. Adapted with permission from Reference 9. μ

©2005, American Chemical Society. (g–h) Self-folding of SU8 fi lms featuring

(g) micropatterned cylinders and (h) microfl uidic networks. Scale bar is 250 μ m. Reprinted

with permission from Reference 12. ©2011, Nature Publishing Group. (i) Confocal image of

a 3D reconstruction of different cells within a self-rolled tube; endothelial (red, innermost

layer); smooth muscle cells (green, middle three layers); and fi broblasts (blue, outermost

layer). Reprinted with permission from Reference 21. ©2012, Wiley.



combined use of both folding and sealing hinges has dramat-

ically increased the complexity of sealed polyhedra that can

be formed, including polymeric cubes 66 ( Figure 2 c) or even

fourteen-faced truncated octahedra 32 ( Figure 2 d). As with

curving structures, a very attractive feature of self-folding

with hinges is that patterns defi ned by highly precise planar

lithographic methods can be implemented in

three dimensions to form precisely patterned

nanopolyhedra ( Figure 2 e–h). 29,30

Graphene origami Three important approaches that have been uti-

lized to enable graphene origami are reviewed

( Figure 3).

Graphene folding by growth and transfer from patterned surfaces In recent work, Kim et al. 47 described strategies

to create multiply-folded structures in graphene,

which they termed grafolds ( Figure 3 a−c).

These structures were observed by transmission

electron microscopy (TEM) in both suspended

and supported graphene samples. They dem-

onstrated control over both the direction and

placement of folds by manipulating surface

curvatures during either the graphene synthesis

or transfer process. One of the routes to con-

trolled folding of graphene was to grow gra-

phene on e-beam patterned copper substrates

prior to transfer ( Figure 3 a); copper functions

as the catalyst for chemical vapor deposition

(CVD) growth of graphene. After growing the

graphene and etching the underlying copper,

the graphene was transferred to isotropic sub-

strates, where it exhibited folds more frequently

along the direction of the patterned copper

Figure 2. Self-folding micro- and nanostructures.

(a) Self-assembled inductor with a submillimeter

scale meander layout. Reprinted with permission

from Reference 23. ©2003, IEEE. (b) Millimeter-scale

spherical solar cells self-assembled from fl ower-

shaped fl at Si leafl ets with a thickness of 2 μm.

Optical image of a complete device consisting of

the folded spherical Si shell, inner glass bead, and

printed silver electrodes. Reprinted with permission

from Reference 38. ©2009, National Academy of

Sciences. (c) Bright fi eld image of stained fi broblast

cells encapsulated within a transparent self-folded

cubic container. Reprinted with permission from

Reference 66. ©2011, Springer Science+Business

Media. (d) Scanning electron microscopy (SEM)

image of a self-folded truncated octahedron.

Reprinted with permission from Reference 32.

©2011, National Academy of Sciences. (e–h) SEM

images of e-beam patterned self-folded nanocubes;

(e–f) 50-nm-thick Au double loop split-ring

resonators patterned on 50-nm-thick Al 2 O 3 panels

and the corresponding 500 nm self-folded cube.

(g–h) Hollow squares lithographically patterned

within 13-nm-thick Ni panels and the corresponding

100 nm folded cube. Reprinted with permission

from Reference 30. ©2011, Wiley.



trenches, as can be seen in the TEM images in Figure 3 b−c.

Apart from potential applications of oriented grafolds, the

authors suggest that complex folding of multiple layers such

as pleat folds could enable validation of theoretical predictions

that have suggested a strong infl uence of stacking and interlayer

interactions on the electrical and optical properties of graphene.

Graphene folding induced by capillary forces Using molecular dynamics simulations, Patra et al. 48 ( Figure 3 d)

have shown that water nanodroplets can activate and guide fold-

ing of graphene nanostructures. They suggest that such folding

could be realized by different types of nanodroplet motion, such

as bending, sliding, rolling, or zipping, leading to either stable or

metastable structures such as sandwiches, cap-

sules, knots, and rings. These novel 3D graphene

structures could enable functional nanodevices

with unique mechanical, electrical, or optical

properties, but it is challenging to locally dis-

pense the nanodroplets. Recently, Guo et al. 45

have reported hydration responsive folding

and unfolding of ordered graphene oxide (GO)

liquid crystal phases with long range order

( Figure 3 e–f). They showed that GO sheets of

known lateral dimension spontaneously form

nematic liquid crystal phases that can be

systematically ordered into well-defi ned supra-

molecular patterns by methods such as surface

anchoring, complex fl uid fl ow, and microconfi ne-

ment. Excluded volume entropy and adsorption

enthalpy associated with its partially hydropho-

bic basal planes drive homeotropic interfacial

surface anchoring in GO. Upon drying, some

of the surface-ordered GO phases transform to

GO solids. When rehydrated, the solids undergo

a dramatic anisotropic swelling to recover their

initial size and shape. The unique ability to fold

and unfold distinguishes GO from other molecu-

lar building blocks. The authors note that further

research on the use of other solvents, salt effects,

and the infl uence of drying speed is needed to

control drying patterns. An entire array of GO-

based smart, stimuli-responsive materials can

be achieved using this method, and hydration

responsive folding/unfolding of GO could be

used in controlled release in nano/microelectro-

mechanical devices, or in dynamic space-fi lling

or sealing applications.

Graphene folding by mechanical forcing Zhang et al. 67 folded graphene sheets in solu-

tion using a 200 Watt ultrasound treatment,

which generates random mechanical forces

( Figure 3 g–i). The authors observed that although

graphene is stiff in the planar directions, it is com-

pliant in the out-of-plane direction, which leads to folding along

preferred directions during mechanical stimulation. Through a

statistical investigation of 100 straight folded edges, they found

that free graphene sheets preferentially folded along armchair

(0°, 29 edges) and zigzag (30°, 27 edges) directions during such

intense mechanical stimulation. They rationalized this preference

by calculating the energies of different folding structures using

atomistic simulations and observing global and local energy

minima along the preferred folding directions of 30° and 0°.

Materials with multiple folds Folding is a distinctive feature observed in naturally self-

assembled structures from the folds in intestinal villi, skin,

Figure 3. Graphene origami. (a–c) Directional control of folding formations in graphene

during synthesis. Graphene synthesized using an etch-patterned copper substrate and

transferred to Quantifoil holey carbon transmission electron microscopy (TEM) grids

show (b–c) graphene folds along the direction of the copper patterns (vertical dark lines).

Images courtesy of the Zettl Research Group, Lawrence Berkeley National Laboratory

and University of California at Berkeley. Reprinted with permission from Reference 47.

©2011, American Physical Society. (d) Nanodroplet-assisted folding of a star-shaped

graphene fl ake. Reprinted with permission from Reference 48. ©2009, American Chemical

Society. (e–f) Crumpled graphene nanoparticles fabricated by continuous microdroplet

drying of colloidal graphene oxide suspensions. Scanning electron microscopy images

showing folded sheet structure and extended creases, whose sharp edges suggest plastic

deformation. Reprinted with permission from Reference 46. ©2012, American Chemical

Society. (g–i) TEM image (center) of a multilayer graphene sheet with HRTEM images

(left and right) of the two preferential folding directions formed by intense mechanical

stimulation; the preferred folding directions are 30° apart. The diffraction patterns indicate

that the left image (g) is an armchair folded edge, and the right image (i) is a zigzag folded

edge. Reprinted with permission from Reference 67. ©2010, American Physical Society.



insect wings, and leaves, to those in stratifi ed rock. Hence, it is

important to understand how folded structures can be generated

by self-assembly processes. 68 Materials often form wrinkles to

relieve stress, as is readily observed during thermal evaporation

of a metal on an elastomeric substrate due to a mismatch in

the thermal expansion coeffi cient.69 Using photolithography

to modulate stiffening by ultraviolet (UV) cross-linking, this

wrinkling can be controlled, and semi-regular patterns can be

formed (Figure 4a). 70,71

Recently, a rich phase behavior of wrinkles and folds has

been observed in a UV curable polyurethane bilayer system

under biaxial compressive stress ( Figure 4 b). 72 Materials with

multiple folds can also be precisely engineered by lithographi-

cal patterning of different materials, thereby controlling the

local thin-fi lm mechanical properties such as stress, thickness,

and modulus. This approach has been utilized by Bassik et al. to

create both uni- ( Figure 4 c) 33 and bi-directionally ( Figure 4 d) 73

self-folding metallic sheets with hundreds to

potentially millions of ordered and oriented

folds.

Stimuli responsive folds enable actuation at small size scales The ability to reconfi gure materials so that they

can reversibly transition from a folded to an

unfolded state is important in enabling tether-

free actuation of devices at small size scales.

Moreover, the creation of stimuli responsive,

reversibly folding structures can enable smart

behaviors. For example, Harrington et al. have

described how swellable cellulose layers fi ll-

ing specialized ice plant cells are important

in enabling origami-like unfolding and thus

disperse seeds only in well-hydrated environ-

ments favorable to germination. 74 Randhawa

et al. have described a microgripper-like

device that reversibly opens and closes during

thermal oxidation and reduction of copper thin

fi lms within copper/chromium bilayer hinges

( Figure 5 a). 75 Of relevance to drug delivery,

pH responsive folding and unfolding of poly-

HEMA capsules have been demonstrated

( Figure 5 b). 40,76 In addition to small devices,

reversible folding and unfolding has been

achieved by differential cross-linking of an

epoxy polymer SU-8, to create reconfi gurable

metamaterial sheets that fl atten out and crum-

ple on exposure to acetone and water, respec-

tively ( Figure 5 c−d).12 Half-tone lithographic

patterning of thermal stimuli-responsive poly

( N ( -isopropylacrylamide) gel sheets has been NN

developed with a variety of swelling patterns

( Figure 5 e−f).77 Other strategies have utilized

macroscale electrically wired hinges composed

of stimuli-responsive materials such as shape-

memory alloys (nitinol) to create so called programmable

matter. 78

Concluding remarks The ability to spatially manipulate the mechanical properties

and, more importantly, the stress within thin fi lms, can be uti-

lized to shape them into three-dimensional geometries. This

spatial heterogeneity can be achieved either by macroscopic

deformation, resulting in wrinkles, or by precise lithographic

patterning of nano- or microscale features. As we have observed,

the ability to transform planar structures into 3D structures

enables the creation of precisely patterned 3D materials and

devices on length scales ranging from the tens of nanometers

to centimeters. Additionally, many self-folding mechanisms

are highly parallel and can be implemented in a tether-free

manner to enable both high-throughput manufacturing as well

as parallel actuation at small size scales. Still, a number of

Figure 4. Materials with multiple folds. (a) Example of complex patterns imprinted on the

surface of poly(dimethylsiloxane) and the alignment of buckles around these patterns.

Reprinted with permission from Reference 70. ©2000, American Chemical Society.

(b) Scanning electron microscopy (SEM) image of co-existing wrinkles and a fold in an

ultraviolet-curable polyurethane bilayer. The fold evolves in a region of disordered wrinkles.

As the fold grows, the stress becomes re-oriented perpendicular to the direction of folding

within a region around the fold tip. Adapted with permission from Reference 72. ©2011,

Nature Publishing Group. (c–d) Self-folding sheets. (c) Assembly of 3D metallic structures

from 2D sheets with lithographically patterned mechanical properties. Optical image

snapshot showing the screw-like lift-off and assembly process for a spiral. Reprinted with

permission from Reference 33. ©2008, Wiley. (d) Bi-directionally self-folding metallic sheets

with chromium-copper bilayer hinges. SEM micrograph of a lithographically patterned

cubic core featuring square pores on horizontal faces and triangular pores on vertical

faces. Reprinted with permission from Reference 73. ©2009, American Institute of Physics.



challenges remain on both the theoretical and

experimental front. On the experimental front,

the elucidation of mechanisms to enable high

local stresses in diverse materials, as well as

methods to orient or pattern them, are impor-

tant to create ordered self-folding nanostruc-

tures. Further, heterogeneous integration with

diverse materials is required to enable elec-

tronic, optical, and biomedical functionalities.

In these systems, the spatial stress distribution

would need to be optimized to limit failure after

self-folding due to cracking or delamination.

Multilayer and multiscale simulation methods

need to be developed and applied to investigate

the behavior of these complex materials and to

optimize functionalities.

Acknowledgments D.H.G. acknowledges support from the ARO

(W911NF-09–2-0065) and the NSF grant

CBET-1066898. V.B.S. acknowledges support

from the ARO (W911NF-11–1-0171) and the

NSF. We thank Shivendra Pandey for help with

preparation of fi gures.

References 1. G.G. Stoney , Proc. R. Soc. London, Ser. A 82 , 172 ( 1909 ). 2. V.Y. Prinz , V.A. Seleznev , A.K. Gutakovsky , A.V. Chehovskiy , V.V. Preobrazhenskii , M.A. Putyato , T.A. Gavrilova , Physica E6, 828 ( 2000 ). 3. O.G. Schmidt , K. Eberl , Naturee 410, 168 ( 2001 ). 4. C.L. Chua , D.K. Fork , K. Van Schuylenbergh , J.P. Lu , J. Microelectromech. Syst. 12, 989 ( 2003 ). 5. W.J. Arora , A.J. Nichol , H.I. Smith , G. Barbastathis , Appl.Phys. Lett. 88, 053108 ( 2006 ). 6. P. Tyagi , N. Bassik , T.G. Leong , J.H. Cho , B.R. Benson , D.H. Gracias , J. Microelectromech. Syst. 18 , 784 ( 2009 ).7. J.H. Cho , T. James , D.H. Gracias , Adv. Mater. 22, 2320 ( 2010 ). 8. R. Songmuang , A. Rastelli , S. Mendach , O.G. Schmidt , Appl. Phys. Lett. 90, 091905 ( 2007 ). 9. J.J. Guan , H.Y. He , D.J. Hansford , L.J. Lee , J. Phys. Chem. BB 109, 23134 ( 2005 ). 10. V. Luchnikov , O. Sydorenko , M. Stamm , Adv. Mater. 17 ,1177 ( 2005 ). 11. S. Zakharchenko , E. Sperling , L. Ionov , Biomacromol.12, 2211 ( 2011 ). 12. M. Jamal , A.M. Zarafshar , D.H. Gracias , Nat. Commun.2 , 527 ( 2011 ). 13. S.V. Mutilin , J.S. Vorobyova , A.B. Vorob’ev , M.A. Putyato , V.Y. Prinz , J. Phys. D: Appl. Phys. 44 , 365104 ( 2011 ). 14. L.B. Ma , S. Kiravittaya , V.A.B. Quinones , S.L. Li , Y.F. Mei , O.G. Schmidt , Opt. Lett. 36, 3840 ( 2011 ). 15. E.V. Naumova , V.Y. Prinz , S.V. Golod , V.A. Seleznev , R.A. Soots , V.V. Kubarev , J. Opt. A: Pure Appl. Opt. 11, 074010 ( 2009 ). 16. E.J. Smith , S. Schulze , S. Kiravittaya , Y.F. Mei , S. Sanchez , O.G. Schmidt , Nano Lett. 11 , 4037 ( 2011 ). 17. I. Monch , D. Makarov , R. Koseva , L. Baraban ,D. Karnaushenko , C. Kaiser , K.F. Arndt , O.G. Schmidt , ACS Nanoo 5 , 7436 ( 2011 ).18. C.C.B. Bufon , J.D.C. Gonzalez , D.J. Thurmer , D. Grimm , M. Bauer , O.G. Schmidt , Nano Lett. 10, 2506 ( 2010 ). 19. J.B. Kim , P. Kim , N.C. Pégard , S.J. Oh , C.R. Kagan , J.W. Fleischer , H.A. Stone , Y.-L. Loo , Nat. Photonicss 6 , 327( 2012 ).

Figure 5. Stimuli responsive folds. (a) Optical image of reversible closing and opening

of metallic microgrippers in oxidative and reductive environments. Scale bar is 500 μm.

Reprinted with permission from Reference 75. ©2010, Wiley. (b) Optical time-lapse

images showing reversible folding of poly(2-hydroxyethyl methacrylate-co-acrylic acid)/

poly(2-hydroxyethyl methacrylate) microcapsule bilayers upon a pH increase from

4 to 9 (folding; upper row) and a pH decrease from 9 to 4 (unfolding; lower row). Scale

bar is 400 μm. Reprinted with permission from Reference 40, ©2012, Wiley. (c–d) Bright

fi eld images of a free-fl oating bi-directionally folded sheet with micrometer-scale gold

split-ring resonators (SRRs) patterned on the square faces. The sheet was reversibly

self-assembled such that the SRRs were either (c) in a single plane when the structure

fl attened in acetone, or (d) arranged along the x , x y , and y z axes when the structure was

de-solvated in water. Insets show magnifi ed views of representative regions from the

same fi lm. Scale bars are 3 mm (inset 250 μm). Reprinted with permission from Reference

12. ©2011, Nature Publishing Group. (e–f) Thermal actuation of lithographically patterned

poly ( N -isopropylacrylamide) gel sheets. Reprinted with permission from Reference 77.

©2012, AAAS.



20. H.Y. He , J.J. Guan, J.L. Lee , J. Controlled Released 110 , 339 ( 2006 ).21. B. Yuan, Y. Jin , Y. Sun, D. Wang, J. Sun, Z. Wang , W. Zhang, X. Jiang ,Adv. Mater. 24 , 890 (2012).22. A.A. Solovev,v W. Xi , D.H. Gracias , S.M. Harazim, C. Deneke , S. Sanchez, O.G. Schmidt , ACS NanoS 6, 1751 ( 2012 ). 23. R.R.A. Syms , E.M. Yeatman, V.M. Bright , G.M. Whitesides ,J. Microelectromech. Syst. 12 , 387 (2003).24. T.G. Leong, A.M. Zarafshar,r D.H. Gracias, Small 6, 792 ( 2010 ). 25. R.R.A. Syms , E.M. Yeatman, Electron. Lett. 29 , 662 (1993). 26. Y. Bar-Cohen, Electroactive Polymere (EAP)r Actuators Ass Artifis cialfifi Muscles:lReality, Potential, and Challengesd (s Society of Photo-Optical InstrumentationEngineers, Bellingham, WA, 2004).27. E. Smela , O. Inganas, Q.B. Pei , I. Lundstrom, Adv. Mater. 5, 630 ( 1993 ). 28. E. Smela , O. Inganas, I. Lundstrom , Science 268 , 1735 (1995).29. J.H. Cho , D.H. Gracias, Nano Lett.o 9, 4049 ( 2009 ).30. J.H. Cho , M.D. Keung, N. Verellen, L. Lagae, V.V. Moshchalkov ,v P. Van Dorpe,D.H. Gracias , Small 7, 1943 (2011). 31. J.H. Cho, A. Azam , D.H. Gracias , Langmuir 26, 16534 ( 2010 ). 32. S. Pandey,y M. Ewing , A. Kunas , N. Nguyen , D.H. Gracias , G. Menon, Proc.Natl. Acad. Sci. U.S.A. 108 , 19885 (2011). 33. N. Bassik , G.M. Stern , M. Jamal , D.H. Gracias, Adv. Mater. 20, 4760 ( 2008 ).34. R.R.A. Syms, IEEE PhotonicsE Technol.s Lett. 12, 1519 (2000). 35. H.J. In, S. Kumar ,r Y. Shao-Horn , G. Barbastathis , Appl. Phys. Lett. 88, 083104 (2006). 36. J.H. Cho, S. Hu , D.H. Gracias , Appl. Phys. Lett. 93 , 043505 (2008). 37. J.S. Randhawa, S.S. Gurbani , M.D. Keung , D.P. Demers, M.R. Leahy-Hoppa ,D.H. Gracias , Appl. Phys. Lett. 96, 191108 (2010). 38. X.Y. Guo, H. Li, B.Y. Ahn, E.B. Duoss , K.J. Hsia, J.A. Lewis, R.G. Nuzzo , Proc.Natl. Acad. Sci. U.S.A. 106 , 20149 ( 2009 ). 39. L. Ionov,v Soft Mattert 7 , 6786 ( 2011 ). 40. T.S. Shim , S.H. Kim, C.J. Heo, H.C. Jeon, S.M. Yang, Angew. Chem. Int. Ed.51 , 1420 ( 2012 ). 41. Y.K. Liu , E. Smela, N.M. Nelson , P. Abshire , Proc. Ann. Int. Conf. IEEE EMBSE26, 2534 ( 2004 ). 42. C.L. Randall , Y.V. Kalinin , M. Jamal , T. Manohar,r D.H. Gracias , Lab Chipb11, 127 ( 2011 ).43. Y.W. Lu, C.J. Kim, Appl. Phys. Lett. 89, 164101 ( 2006 ).44. T.G. Leong , C.L. Randall, B.R. Benson , N. Bassik, G.M. Stern, D.H. Gracias ,Proc. Natl. Acad. Sci. U.S.A. 106 , 703 (2009).45. F. Guo, F. Kim , T.H. Han , V.B. Shenoy,y J.X. Huang , R.H. Hurt , ACS NanoS5, 8019 ( 2011 ).46. Y. Chen , F. Guo , A. Jachak , S.-P. Kim , D. Datta , J. Liu , I. Kulaots , C. Vaslet ,H.D. Jang, J. Huang , A. Kane , V.B. Shenoy,y R.H. Hurt, Nano Lett.o 12 , 1996( 2012 ).47. K. Kim, Z. Lee, B.D. Malone , K.T. Chan , B. Aleman , W. Regan , W. Gannett,M.F. Crommie, M.L. Cohen, A. Zettl, Phys. Rev. B 83 ( 2011 ).48. N. Patra, B.Y. Wang, P. Kral, Nano Lett.o 9 , 3766 (2009).49. E. Prada , P. San-Jose , L. Brey,y Phys. Rev. Lett. 105 , 106802 (2010).50. D. Rainis, F. Taddei , M. Polini , G. Leon, F. Guinea, V.I. Fal’ko, Phys. Rev. B83, 8 ( 2011 ).51. S.H. Gellman, Acc. Chem. Res. 31 , 173 (1998).52. P.W.K. Rothemund , Nature 440 , 297 (2006).53. Y. He, T. Ye , M. Su , C. Zhang , A.E. Ribbe, W. Jiang , C.D. Mao , Nature 452 ,198 ( 2008 ).54. L.D. Landau , E.M. Lifshitz, Theory ofy Elasticity,f 3rd ed.d (Pergamon Press,Oxford, 1986).55. S. Timoshenko, J. Opt. Soc. Am. Rev. Sci. Instrum. 11, 233 ( 1925 ).56. C.A. Klein, R.P. Miller ,r J. Appl. Phys. 87 , 2265 (2000).57. C.H. Hsueh, J. Appl. Phys. 91, 9652 ( 2002 ).58. G.P. Nikishkov,v J. Appl. Phys. 94, 5333 ( 2003 ).59. L.B. Freund , S. Suresh, Thin Filmn Materials:m Stress,: Defect Formation,t andSurface Evolution,e 1st paperbackt editionk ( Cambridge University Press , UK,2009).60. A.E. Giannakopoulos , I.A. Blech , S. Suresh, Acta Mater.a 49 , 3671 (2001).

61. S. Alben , B. Balakrisnan, E. Smela, Nano Lett.o 11 , 2280 (2011). 62. J.H. Cho, D. Datta, S.Y. Park, V.B. Shenoy,y D.H. Gracias, Nano Lett.o 10, 5098(2010). 63. E. Moiseeva , Y.M. Senousy,y S. McNamara, C.K. Harnett, J. Micromech.Microeng. 17, N63 ( 2007 ). 64. M.H. Huang , F. Cavallo, F. Liu , M.G. Lagally,y Nanoscale 3, 96 (2011 ).65. K. Kumar ,r V. Luchnikov,v B. Nandan , S. Zakharchenko, L. Ionov ,v Mater. Res.Soc. Symp. Proc. 1272 (2010).66. A. Azam , K.E. Laflinfl , M. Jamal , R. Fernandes , D.H. Gracias , Biomed.Microdevices 13 , 51 (2011).67. J. Zhang, J.L. Xiao, X.H. Meng , C. Monroe, Y.G. Huang , J.M. Zuo , Phys. Rev.Lett. 104, 4 ( 2010 ).68. L. Mahadevan, S. Rica , Science 307 , 1740 (2005).69. N. Bowden, S. Brittain , A.G. Evans, J.W. Hutchinson, G.M. Whitesides,Nature 393, 146 ( 1998 ).70. W.T.S. Huck, N. Bowden, P. Onck, T. Pardoen , J.W. Hutchinson ,G.M. Whitesides, Langmuir 16 , 3497 (2000).71. A. Schweikart, A. Fery,y Microchim. Acta 165 , 249 (2009).72. P. Kim , M. Abkarian, H.A. Stone , Nat. Mater. 10 , 952 (2011).73. N. Bassik , G.M. Stern , D.H. Gracias , Appl. Phys. Lett. 95 , 91901 (2009).74. M.J. Harrington , K. Razghandi , F. Ditsch , L. Guiducci, M. Rueggeberg,J.W.C. Dunlop, P. Fratzl , C. Neinhuis, I. Burgert, Nat. Commun. 2 (2011).75. J.S. Randhawa, M.D. Keung , P. Tyagi , D.H. Gracias , Adv. Mater. 22, 407( 2010 ).76. R. Fernandes, D.H. Gracias, Adv. Drug Deliveryg Rev.y ( 2012 ), doi:10.1016/j.addr.2012.12.012.77. J. Kim, J.A. Hanna , M. Byun , C.D. Santangelo, R.C. Hayward, Science335, 1201 ( 2012 ).78. E. Hawkes, B. An, N.M. Benbernou, H. Tanaka , S. Kim , E.D. Demaine , D. Rus ,R.J. Wood, Proc. Natl. Acad. Sci. U.S.A. 107 , 12441 (2010).79. S.M. Harazim , W. Xi, C.K. Schmidt, S. Sanchez , O.G. Schmidt , J. Mater.Chem. 22, 2878 ( 2012 ).

Vivek B. Shenoy is a professor of engineeringat Brown University. He received his doctoratedegree from the Ohio State University,performed postdoctoral research at Brown,and was a visiting Fellow at the University ofCambridge prior to starting his independentresearch group in 2000. He has published over100 journal articles in the areas of computationalfmaterials science and mechanics.

David H. Gracias is an associate professor atthe Johns Hopkins University. He received hisPhD degree from UC Berkeley, performedpostdoctoral research at Harvard, and workedat Intel Corporation prior to starting hisindependent research group in 2003. He haspublished 80 journal articles and is a co-inventorof 20 patents in the areas of micro- andnanotechnology, self-assembly, and interfacialscience.

NOMINATE A COLLEAGUE TODAY

www.mrs.org/awardsNomination Deadline—October 1, 2012

for one of these prestigious awards from the Materials Research Society

Innovation in Materials Characterization Award Mid-Career Researcher Award

MRS Fellow Outstanding Young Investigator Award

855MRS BULLETIN • VOLUME 37 • SEPTEMBER 2012 • www.mrs.org/bulletin© 2012 Materials Research Society

Introduction The thermal conductivity of materialsf is a wonderful topic

for ar diverse group of materialsf researchers because every type

of materialf carries heat, and issues of thermalf management

and heat transfer appear in almost all of thef sub-disciplines

of materialsf science and engineering. Thermal management

is a well-known limiting factor in the performance of elec-f

tronic materials; oxide thermal barriers extend the operat-

ing temperature of super-alloys;f 1 and in biological systems,

nanoscale thermal effects are thought to be a route toward

new techniques for drug delivery and medical therapies.2

The subject of heatf transport is highly interdisciplinary, not

only across the spectrum of materials,f but it also involves

researchers from many different disciplines in physical

sciences and engineering. The research described here

includes contributions from specialists in inorganic and

physical chemistry; geology; condensed matterd andr materialsd

physics; and mechanical and electrical engineering. I will

begin withn someh introductory commentsy on thermaln conductivity

and a related property, the “interface thermal conductance.”

I will describe examples of extremesf of highlyf thermally

conductive interfaces between two materials3 and interfaces

with low thermal conductance.4 I will then describe examples

of highf thermal conductivity materials, for example, carbon

nanotubes and graphene.d 5 I will describe why it ist diffi cultfi tot

make use of thef high thermal conductivity of carbonf nanotubes

in forming a high performance thermal interface material6

or composite.7

Extremes of heat conduction ― Pushing the boundaries of the thermal conductivity of materials David G. Cahill

The followinge articleg ise an edited transcriptd oft af lecture presentede ind Symposium X: Frontiers of Materialsff

Research, by David G.d Cahill, on November 29, 2011, at the 2011 Materials Research Society Fall

Meeting ing Boston.

Thermal conductivity is a familiar property of materials: silver conducts heat well, and

plastic does not. In recent years, an interdisciplinary group of materialsf scientists, engineers,

physicists, and chemists have succeeded in pushing back long-established limits in the

thermal conductivity of materials.f Carbon nanotubes and graphene are at the high end of thef

thermal conductivity spectrum due to their high sound velocities and relative lack of processesf

that scatter phonons. Unfortunately, the superlative thermal properties of carbonf nanotubes

have not found immediate application in composites or interface materials because of

difficultiesfi in making good thermal contact with the nanotubes. At the low end of thef thermal

conductivity spectrum, solids that combine order and disorder in the random stacking of

two-dimensional crystalline sheets, so-called “disordered layered crystals,” show a thermal

conductivity that is only a factor of 2f larger than air. The cause of thisf low thermal conductivity

may be explained by the large anisotropy in elastic constants that suppresses the density of

phonon modes that propagate along the soft direction. Low-dimensional quantum magnets

demonstrate that electrons and phonons are not the only signifiy cantfi carriers of heat.f Near roomr

temperature, the spin thermal conductivity ofy spin-laddersf is comparable to the electronic

thermal conductivities of metals.f Our measurementsr of nanoscalef thermal transport properties

employ ay variety ofy ultrafastf optical pump-probe metrology toolsy that we have developed over

the past several years. We are currently workingy to extend these techniques to high pressures

(60 GPa), high magnetic fi eldsfi (5 T), and high temperatures (1000 K).

David G. Cahill, University of Illinois at Urbana-Champaign; [email protected] DOI: 10.1557/mrs.2012.201

EXTREMES OF HEAT CONDUCTION ― PUSHING THE BOUNDARIES OF THE THERMAL CONDUCTIVITY OF MATERIALS


I will then discuss the opposite limit, the ultralow thermal

conductivity of “disordered layer crystals.” 8 These materials

have the lowest thermal conductivity of any dense solid, and

I will explain our current thinking about the mechanisms that

create this phenomenon and why it is exciting. I will describe

the utility of high pressures in this fi eld of research, which are

used as an extreme condition that allow us to probe physics in

a way that has not been available before. The example I will

describe is the ice phase VII at pressures up to 20 GPa, which

has a thermal conductivity approximately that of sapphire in

ambient conditions. Ice becomes highly thermally conduc-

tive because pressure increases the strength of the molecular

bonding. 9 Studies of ultralow thermal conductivity and thermal

conductivity at high pressures have been enabled by a new non-

contact measurement technique we developed approximately

10 years ago, time-domain thermorefl ectance (TDTR).17 I will

then briefl y discuss extremes of thermal conductivity due to spin

waves in copper-oxide “spin-ladders.” 10 At room temperature,

spin waves in this material have a reported thermal conductivity

that is close to that of thermal conductivity of Si by phonons.

Thermal conductivity of materials and the thermal conductance of interfaces I will use Λ to denote thermal conductivity. Λ is the linear

coeffi cient that relates a temperature gradient to a heat fl ux;

a material with a high thermal conductivity conducts a large

amount of heat across a small temperature gradient. Another

thermal transport property of interest is the thermal conductance

per unit area, which I will denote by G, which is a linear

transport coeffi cient that relates a temperature drop at an inter-

face to the heat fl ux at that interface. This is a property of the

interface—a high thermal conductance interface produces a

high fl ux of heat for a small temperature difference.

Figure 1 summarizes handbook values of thermal conductiv-

ities of some materials11 with one recent addition, a disordered

form of the layered crystal WSe2 . Diamond has the highest

thermal conductivity of any material at room temperature; this

is because the sound speed in diamond is large, the Debye

temperature is high, and the anharmonicity is weak, so the

mean-free paths of the phonons, the wave-like lattice vibra-

tions that carry heat in dielectrics and semiconductors, are long

and carry heat effi ciently and coherently over long distances.

Graphite has a thermal conductivity that is similar to that of

diamond but only in one direction, in the plane of the carbon

sheets of graphite; graphite has the highest anisotropy of any

material. The thermal conductivity of SiC is somewhat lower.

Cu is also a good thermal conductor, as are Al, Ag, and Au; this

is not due to phonons, but because of electrons, which carry

heat in metals. They have high electrical conductivities, and

the Wiedemann–Franz law states that they should also have

high thermal conductivities. 23 In disordered materials—oxide

glasses, disordered crystals, such as stabilized zirconia, and

amorphous polymers—the thermal conductivity is low. The

thermal conductivity of crystalline polyethylene has recently

been reported to be close to that of Si. 22

The range of observed values of the interface thermal conduc-

tance is approximately 60 at room temperature ( Figure 2 ).4 This

range of 60 is much smaller than the 3 × 10 4 range of thermal

Figure 1. Thermal conductivities and densities of selected

materials. Reprinted with permission from Reference 11.

©2007, American Association for the Advancement of Science.

Figure 2. Thermal conductance of selected interfaces. The

TiN/MgO and Bi/H-terminated-diamond interfaces represent the

extremes of high and low thermal conductance, respectively.

The right axis is the thickness of a layer of oxide glass that has the

same thermal resistance as these individual interfaces. 4 DMM is

the result of a calculation using the diffuse mismatch model. The

term “radiation limit” refers to the upper limit for heat conduction

assuming only elastic channels that couple phonons on one side

of the interface to phonons on the other side of the interface.



conductivities because transport across an interface does not

directly involve the coherence or mean-free path of the heat

carriers. We created the highest thermal conductance interface

by growing an epitaxial TiN layer on an MgO crystal at high

temperatures to create a nearly perfect interface; the thermal

conductance is extremely high, approaching 1 GW/m2 -K.

The power output of a large nuclear power plant, if channeled

through an area of 1 m 2 , would produce a temperature drop

across a TiN/MgO interface of only 1 K.3

To explore the limit of low thermal conductance interfaces,

we prepared samples of Bi on H-terminated diamond. This is

certainly an extreme interface: the stiffest, highest Debye tem-

perature solid (diamond) in contact with the most compliant,

lowest Debye temperature solid (Bi) that can be easily evapo-

rated in a conventional deposition chamber. Heat transport

across the Bi/diamond interface is strongly inhibited, but the

thermal conductance is actually much larger than predicted by

simple models. (The reason for the discrepancy between the

simple models and experiment is still an open question.) The

“equivalent thickness” or the “Kapitza length” (the thermal

conductivity of the adjacent material divided by the interface

thermal conductance) is plotted on the right-hand axis and is

calibrated to the thermal conductivity of a hypothetical oxide

glass with a thermal conductivity of 1 W/m-K. The Bi/diamond

interface has a thermal resistance that is equivalent to approxi-

mately 100 nm of an oxide glass.

A common research activity in materials science is to relate

changes in microstructure with changes in properties, in other

words, to establish the structure–property relationships. This

task is challenging for thermal transport properties because

thermal conductivity and interface thermal conductance are

integrals over all of the thermally excited states that can carry

heat in the material. For example, in a dielectric or a semi-

conductor, where phonons are the dominant heat carriers, the

thermal conductivity is an integral over all modes of the product

of the heat capacity of each mode; the group velocity, the speed

with which the mode propagates; and the mean-free path or the

coherence distance over which the heat carriers can transmit the

thermal energy from one point of the crystal to another.24 If we

make a change in the structure (e.g., we create a superlattice,

introduce defects, or create a nanostructured material), we are

rarely sure how the microstructural changes affect the different

quantities in the integral.

Upper limits of thermal conductivity To give some perspective on the upper bound of the thermal

conductivity of materials, consider diamond, the bulk crystal-

line material with the highest thermal conductivity. It is possible

to make isotopically pure diamond and increase the thermal

conductivity by about 30% by eliminating the mass disorder

related to the natural mixture of C12 and C13 isotopes. 12 Carbon

nanotubes and graphene have even higher thermal conductivity,

as has been demonstrated in a number of experiments. A meas-

urement technique pioneered at the University of California,

Berkeley13 includes two micromachined Si platforms, which

serve as the heaters and thermometers in the experiment. The

platforms are thermally isolated by extremely long silicon-

nitride legs. A nanotube thermally connects one pad to another

so that the conductance of that nanotube can be measured by

applying heat and measuring temperatures. Typically, focused

ion beam deposition of Pt is used to improve the thermal con-

tact between the nanotube and the pads. There is considerable

excitement to try and harness the extremely high thermal con-

ductivity of nanotubes to make a thermal interface material

from a vertical array of carbon nanotubes to connect a silicon

die to a heat sink; that layer would be highly compliant, but it

would also be able to carry large amounts of heat.6

There is also considerable interest in the possibility of

adding nanoparticles or nanowires, such as carbon nanotubes,

to fl uids in order to produce better heat transfer fl uids for heat

engines, or composites for applications in thermal management.

For example, if carbon nanotubes are dispersed in an epoxy at

a concentration beyond what is referred to as the “percolation

limit” (when the entire sample is spanned by a continuous

network of carbon nanotubes), then, ignoring interfacial effects,

the thermal conductivity of the composite material would be

one-third of the product of the volume fraction of the fi bers and

the thermal conductivity of the fi ber. (The factor of one-third

comes from orientational averaging.) If the nanotube thermal

conductivity is 3000 W/m-K, as reported for suspended nano-

tubes, and we add 1% volume fraction of nanotubes into the

epoxy, we should obtain a composite with a thermal conductivity

of approximately 10 W/m-K. 7 That hypothetical composite

would be a remarkable material with a thermal conductivity

higher than the typical thermally conductive adhesives or

polymer composites that can currently be made. This high

conductivity has not been observed however, because the

predicted behavior can only be obtained if the aspect ratio of

the fi bers is large enough. This constraint on the aspect ratio is

where the properties of the interfaces enter into the problem.

The interface conductance between a nanotube and a polymer

matrix is small; therefore, the fi bers have to be extremely long,

with high persistence length to achieve a signifi cant increase

in the thermal conductivity of a composite.

At nanometer length scales, a one-dimensional fi ber geom-

etry may not be the best approach; planar two-dimensional

fi llers might be more advantageous for composites or interface

materials. Balandin et al. 5 applied self-heating of graphene

using a laser, and then Raman scattering was used to measure

the temperature rise; the temperature dependence of the fre-

quency of the Raman active mode is used for thermometry.

There are discrepancies between measurements by different

groups, but the consensus is that the thermal conductivity near

room temperature is on the order of 3000 W/m-K.

The observations of extremely high thermal conductivity

in nanotubes and graphene are for materials that are not in

contact with any surrounding materials. If we want to use the

material to manage heat fl ow in an electronic device or use

these extreme properties to make a composite, then obviously

it must be in contact with another material. Jang et al. 14 studied



the thermal conductivity of graphene layers that were encased

in layers of SiO2; heat was injected by joule heating, and then

the temperatures were measured by resistance thermometry.

The results of this are shown in Figure 3 for different layer

thicknesses of graphene, from a single layer up to 21 layers

of graphene compared to the in-plane thermal conductivity

of graphite. As the number of graphene layers increases, the

conductivity approaches the properties of graphite, as expected,

but the thermal conductivity of the individual layers is strongly

suppressed by interactions with the neighboring materials. This

suppression is not fully understood; computational models of

nanotubes do not show this degree of suppression. Vibrational

modes that involve the rippling of the graphene sheet or

snake-like reptations of nanotubes conduct a signifi cant amount

of heat, and the suppression in conductivity is partly due to

strong damping of those vibrational modes.

To summarize, there are actually two distinct types of inter-

face problems that limit the applications of extremes of thermal

conductivity observed in nanotubes and graphene. The fi rst

problem is the thermally weak interfaces between the materials

and the surrounding environment; the second problem is that

the surrounding materials suppress the ability of a material to

conduct heat.

Thermal conductivity in organic materials The extreme of thermal conductivity of an organic or a molec-

ular structure is also a topic of current interest. We know that

ultrahigh molecular-weight polyethylene, highly drawn, and

highly crystalline, can have extraordinary high thermal conduc-

tivity, reported to be up to 30 W/m-K in commercially available

fi bers. 15 This thermal conductivity is comparable to that of a

sapphire crystal.

It is diffi cult to make large perfect crystals out of polyeth-

ylene. Recently, Shen et al.16 performed an experiment using a

nanofi ber of polyethylene—only a few-hundred nanometers in

diameter—which was synthesized, drawn, and then measured

in a challenging experiment in which the polymer nanofi ber

was attached to an atomic force microscope (AFM) cantilever

that bends when heat fl ows through the cantilever. The mea-

surements need to be carefully calibrated. Shen and co-workers

obtained a thermal conductivity of the polyethylene nanofi ber

on the order of 100 W/m-K, close to that of Si.16

This result raises the question: What is the limit of thermal

conductivity of a molecular or organic structure? There are a

large variety of materials available, but high density polyeth-

ylene is the only one that has been examined in detail. At the

University of Illinois at Urbana-Champaign (UIUC), we are

currently examining a variety of high modulus polymer fi bers

to gain greater insight into the structure/property relationships

for thermal conductivity in this class of materials. We started

with polyethylene fi ber and embedded it in an epoxy, cut it

with a diamond knife, and measured its thermal conductivity.

This is a relatively routine measurement, so I believe we can

explore a large variety of high modulus fi bers relatively quickly.

For the past several years, our routine measurements

of thermal conductivity have employed a method we call

“time-domain thermorefl ectance” (TDTR). 17 This is a pow-

erful technique that has replaced the use of the 3 ω method in

my group. 25 The cost of parts needed to assemble this instru-

ment, approximately $200,000, is large compared to the tools

typically used in thermal sciences but comparable to many

of the scientifi c instruments used in materials research. The

basic approach—a pump optical pulse heats a sample, and a

probe optical pulse measures the temperature of the sample

through a change in refl ectivity that occurs as a function of

temperature—has been known for 25 years. Our contribution

to advancing the fi eld was in improving the optical design and

the methods of data analysis to obtain much more accurate

values of the thermal conductivity than was previously possible.

By analyzing the data as a function of the delay time between

the pump and the pulse and the frequency at which the pump

beam is modulated, a great deal of information about thermal

properties can be extracted, for a number of different geometries

and samples, spanning a wide range of properties.17

Measurements of thermal conductivity I will now concentrate on measurements that are performed

using TDTR. Such measurement systems are rapidly propagating

around the world, and I believe there are now approximately

20 systems operating similar to the one we established at UIUC

several years ago.

The use of thermal waves has a long history in materials

research. In a TDTR measurement, we apply a broad frequency

spectrum of heat, and we examine the temperature response to

that spectrum. Our predecessors showed us the way. Ångström 18

measured the thermal conductivities of metals, iron, and copper

for example, to within a few percent of the accepted values

Figure 3. Thermal conductivity of supported or “encased”

graphene.14 The label on each curve is the number of graphene

layers. The inset shows power laws for thermal conductivity as

a function of temperature.



today. This was achieved withd a temperature boundary condi-

tion oscillatingn at lowt frequency. Fixed temperaturesd (either icer

water atr 0°Ct or steamr at 100°C)t were applied tod one end ofd thef

sample bar, thermometers were mounted alongd the length ofh thef

bar, and thed temperature oscillations of thef thermometers were

measured asd a function of time;f the results were analyzed tod

obtain the thermal conductivity of thef material. The time scales

of thef oscillations were long, because the length scales were

large. These measurements must havet required greatd patience;t

presumably the graduate students or technicians turned the

switch back-and-forth between ice water andr steamd for manyr

hours in order tor collect thet data.

In the modern era, we are working with nanometer lengthr

scales that require short time scales to spatially confinefi the

thermal waves. The thermal diffusivity of anf oxide glass is

approximately 10−6 m2/s, which means that heat diffuses in

an oxide approximately 1 meter per month.24 (The thermal

diffusivity of soilf is comparable to glass, which explains why

pipes in cold climatesd buried 1d meter belowr the surface do not

freeze in the winter.) Writing the diffusivity in sets of unitsf

gives 10–6 m2 /s = 1 mm2 /s = 1 μ m2 / μ/ s = 1 nm2 /ps. If wef are

interested ind samples with thicknesses on the order ofr tensf of

nanometers, then the time scale of heatf diffusiont is on the order

of nanoseconds.f

Lower limits of thermal conductivity Measurements by TDTR were critical to our discovery of

extremely low thermal conductivity in anisotropic materials

that containt high densities of disorderedf interfaces.d For manyr

years, the conventional wisdom for the lowest heat carry-

ing capability of af material has been the “minimum thermal

conductivity,” sometimes referred tod as the “amorphous limit.”

The idea behind the minimum thermal conductivity started

with Einstein; in 1911, he published the fi rstfi theory of heatf

conduction in materials. This paper19 is completely wrong for

crystals; however, for amorphousr materials, Einstein’s model

works quite well. Debye was the firstfi tot explain the thermal

conductivity ofy crystalsf in termsn of phononsf and theird mean-freer

paths. The idea ofa thisf amorphous limit reappearst several times

in the literature over ther past century.t In the high-temperature

limit, the minimum thermal conductivity is given by

2 / 3

min B0.40 ( 2 ),Λ = ν + νl tk

Bn (1)

where kBkk is Boltzmann’s constant, n is the atomic density, and

νl and νt are the sound speeds.d

Einstein’s 1911 paper builtr ont hisn 1907 “Einstein oscillator”

paper, which was the fi rstfi applicationt of quantumf mechanics

to the solid state. The thermal conductivity of thef Einstein

oscillator modelr of 1907f was exactly zero, because Einstein

oscillators are not coupled to each other. In his 1911 paper,

Einstein introduced transportd intot his model by coupling each

Einstein oscillator tor 26 neighbors in a cubic arrangement oft

atoms. That ist a largea number ofr coupledf neighbors,d but het was

trying to make the thermal conductivity as large as possible,

because he realized hisd results were too small. The mistake he

made was to assume that the thermal energy was randomly

transported between vibrations, and he failed to realize that

phonons are the normal modes of af crystal. In other words,r

Einstein assumed thatd thet atoms oscillated withd random phases

rather thanr with the coherent andt collectived lattice motions we

now associate with phonons.h The moral of thisf story thaty ist even

a gianta oft sciencef can sometimes miss ideas that latert seemr like

they should haved been obvious.

Figure 4 shows that thist picture works well for disorderedr

materials that aret homogeneous and isotropic.d 20 The measured

values are in good agreementd witht the model for bothr amor-

phous solids and disorderedd crystals.d The model has no free

parameters, including only the sound speedsd and atomicd den-

sities. Some important materials are included in this fi gure,fi

such as stabilized zirconia,d which provides the thermal barrier

on the turbine blades of everyf airplane; feldspar isr a dominant

mineral in the earth’s crust, so this physical behavior controlsr

heat conductiont in the crust.

To produce materials with a thermal conductivity lower

than the amorphous limit, we can try to make use of reducedf

thermal conduction across interfaces. Many different kindst of

materials have been studied, both in my group and byd other

researchers, but thet most dramatict result int reducing thermal

conductivity below the amorphous limit was obtained using

a disordered form of WSef 2 .8 This material was the result of

David Johnson’s work at the University of Oregonf using a

synthesis technique he calls the “modulated elementald reactants

method.”26 This material is synthesized byd depositing W andW

Se onto a cold substrated in the correct sequencet and withd the

correct quantityt so that thoset materials will form the required

compound whend heated.

Figure 4. Comparison of measuredf and calculated thermal

conductivities of amorphousf solids (filledfi circles) and strongly

disordered crystals (open circles). The calculation is based on

the model of minimumf thermal conductivity.20



We carried out signifi cant analysis of the structure of the disor-

dered fi lms, including x-ray diffraction at a synchrotron source,

but the structure is more directly revealed by cross-sectional

transmission electron microscopy (TEM). The TEM images 21

revealed limited crystalline order in the in-plane direction,

but there is almost complete disorder in the through-thickness

direction. This material is not in equilibrium—annealing creates

some crystallinity in the plane, but the kinetics are not suffi cient

to achieve crystallinity normal to the plane. The structure is

analogous to a deck of playing cards that have been thrown

down on the table and then brought together but have not been

gathered into a deck. A well-stacked deck is analogous to a

crystal, but the structure of this thin fi lm has cards with random

orientation to each other, so each of the interfaces is incoherent.

If we measure the thermal conductivity through this disor-

dered layered structure—using TDTR discussed previously—

the result is a thermal conductivity that is lower than has ever

been observed in a fully dense material, approximately twice

the thermal conductivity of air, or 0.05 W/m-K. The exciting

result is that this value is approximately a factor of 6 less than

the predicted minimum thermal conductivity, based on the

Einstein model, and calculated from the atomic density and

speeds of sound measured using picosecond acoustic tech-

niques. We refer to conductivities far below the expected amor-

phous limit as “ultralow thermal conductivity” and attribute this

behavior, in the case of disordered WSe 2 , to the combination of

disordered stacking of the WSe2 sheets and the strong anisot-

ropy in the bonding. The low speeds of sound normal to the

crystalline sheets relative to the sound velocities within the

plane of the crystalline sheets essentially reduces the density

of states of phonons that carry heat in the cross-plane direction.

Figure 5 shows the thermal conductivity of disordered WSe 2as a function of the measurement temperature. The top plot is

data for the equilibrium single crystal, which has a thermal

conductivity of approximately 1.5 W/m-K at room tempera-

ture. This is a small value for a crystal, but can be understood

based on the low speeds of sound and the large anharmonicity

and anisotropy of the atomic bonding. 27 There is some scatter

in the properties of disordered WSe2 because it is diffi cult to

precisely control the composition in the process of deposition

and annealing.

Heavy ion irradiation The data plotted in Figure 5 are quite remarkable—the thermal

conductivity is the lowest that has ever been observed. Extraor-

dinary results require extraordinary proof, so I suggested to my

student that he fi nd a way to make the low thermal conduc-

tivity disappear. If we believe it is due to the interfaces, and

if we remove the interfaces, the thermal conductivity should

increase. To do this, we irradiated the disordered layer crystal

with 1 MeV Kr + ions. When a high-energy heavy ion passes

through a high atomic number material, the ion creates a track

of molten high-pressure fl uid, and then the material will recrys-

tallize, become amorphous or become a mixture of both as a

result. As the ion dose increased from 10 14 cm −2 to 10 15 cm −2 ,

the thermal conductivity increased toward the value predicted

by the Einstein model ( Figure 6 ). It is not possible to measure

samples exposed to higher ion doses, because the morphology

of the fi lm is disrupted.

A decade ago, before TDTR, we could only dream of

performing this type of experiment using ion beams because of

limitations of the thickness of the layers that can be processed

and measured. The high throughput of the TDTR approach

provides another big advantage: these radiation damage

Figure 5. Thermal conductivity of disordered layered WSe 2 in

the direction normal to the crystalline sheets. Data are labeled

with the thickness of the fi lm. 8 The label “irrad” denotes a layer

that has been irradiated with heavy ions. The dashed-dot line

is the predicted minimum thermal conductivity based on the

measured atomic density and speeds of sound in the cross-

plane direction.



experiments were performed ond one sample and irradiatedd tod

different dosest at differentt spotst on then sample. TDTR wasR used

to measure the thermal conductivity ofy eachf locationh andn accu-d

mulate the data fora thermalr conductivity as a functiona of ionf

dose. We are currently performingy similar experimentsr on UOn 2

and revisitingd some longstanding issues concerning the effects

of radiationf damagen on then thermal conductivity ofy nuclearf fuels.r

Eff ect of extreme pressure Extreme pressure provides a method ford probingr the physics

of heatf conduction. We are using gem anvil cells, SiC and

diamond, to modify the phonon density states, measure the

thermal conductivity byy TDTR,y and testd ourt models.r In ourn case,r

we are studying ice. Water-ice has a wonderfully rich phase

diagram (Figure 7 ).22 Our experimentsr are carried outd att roomt

temperature, starting with the liquid phase,d which crystallizes

at approximatelyt 1 GPa, forming the tetragonal Ice VI phase,

and thend at approximatelyt 2 GPa it formst Ice VII. Ice VII is a

cubic crystal withl interpenetratingh bccg sublattices. If wef reducee thee

temperature, the hydrogen disordern disappears,r forming Iceg VIII.

If wef increase the pressure on Icen VII, the hydrogen bondn even-d

tually becomes symmetric; this is sometimes referred to as

“polymeric ice” because the O−H−O bonds are evenly spaced.

We combine diamond anvild cells and TDTR,d which ish a com-a

pletely optical method, so it cant be easily adapted tod extreme

environments such as high pressure. We pass the pump and

the probe beams through the diamond andd used Al-coated micad

as a substrate. The data analysis is complicated, and thered are

a number ofr differentf parameterst to be determined ind order tor

model the experiment.9 We use the data to test thet applicability

of thef Leibfried–Schlömann equation,29

1/ 3

,γ

DV

T

3

2

ωΛ = f (2)

where V isV the molecular volume,r Dω is the Debye frequency,

and γ is the Grüneisen parameter, which is a measure of thef

anharmonicity. This equation is based ond the idea that acoustict

phonons are the dominant heatt carriers,t and thatd anharmonict

scattering controls the mean-free paths. The thermal conductiv-

ity is inversely proportional to temperature, and ad high Debye

frequency results in high thermal conductivity.

As we increase the pressure at constant temperature, the

main change is an increase in the Debye frequency; in other

words, we make ice with different Debyet frequencies by apply-

ing pressure. The Debye temperature and Grüneisend parameter

can ben obtained fromd them equation ofn statef as measured byd x-rayy

diffraction at at synchrotron source. Data plotted ind Figure 8show good agreement with the prediction of thef Leibfried–

Schlömann equation for ther thermal conductivity as a function

of pressure.f The upper axisr is the relative molar volume,r or ther

volume per moleculer normalized by the volume of Icef VII,

when it is fi rstfi formed at the lowest pressure point. As the

volume decreases to two-thirds of thef original volume,l the thermal

conductivity increases to approximately 30 W/m-K; this is ice

with a thermal conductivity equivalent tot that oft sapphire,f due

to the high Debye temperature created byd the high pressure.

Spin wave conductivity In discussing thermal transport, we are normally dealing with

electrons and phonons.d I have focused ond phonons,n lattice vibra-

tions that aret the dominant heatt carrierst in dielectricn crystals and

semiconductors. In metallic materials, electrons are usually the

dominant carrierst of heat.f Normally, there is no need tod consider

other mechanisms.r But anyt thermal excitation inn then crystal can

conduct heat;t the question is what ist the heat capacityt of thesef

excitations, their group velocity, and the mean-free path or

Figure 6. Effect of heavyf ion irradiation on the cross-plane

thermal conductivity of disorderedf WSe2 .8

Figure 7. Pressure-temperature phase diagram for water.r Liquid

water isr in the upper leftr part of thef diagram; the remaining regions

are labeled by any ice crystalline phase. At room temperature,

tetragonal ice VI is the equilibrium phase between 1 and 2 GPa; ice

VII is the equilibrium phase at pressures greater thanr 2 GPa.



coherence of thef excitations? Ten years ago, it wast discovered

that itt ist possible to have extremely highy thermalh conductivity iny

certain classes of complexf copper-oxides, in which spin waves

are the dominant heat carriers.12 The scattering mechanisms

and thed ultimate limits of thef spin-wave channel of heatf trans-t

port aret not known.t In other words,r we do not yett knowt what

limits the mean-free paths of spin-wavesf or what we might

change in the crystal structure to create greater dispersionr or

longer mean-freer paths. At UIUC,t we are examining the prop-

erties of onef of thef so-called “telephoned number compounds”—r

Ca9La5Cu24 O41. The key structural motif inf these crystals is

the CuO2 ladder, where the spins are strongly coupled, and thed

thermal excitations of thesef coupled spinsd produce a surpris-

ingly high thermal conductivity.

A thermalA conductivity at room temperature as high as

approximately 90 W/m-K has been reported (Figure 9 ).10

The conductivities of thef crystals we are studying are lower

for reasonsr that aret not yett clear.t We are currently examining

how the thermal conductivity measured byd TDTR variesR as a

function of thef modulation frequency of thef pump beam; we

have previously showny thatn thet frequency dependencey provides

information about thet mean-free-paths of heatf carriers.t 30 Our

working hypothesis is that weak magnon-phonon coupling

at low temperatures will suppress the thermal conductivity

measured in a TDTR experimentR when the length scales of

thermal waves are insuffi cientfi to effectively transfer heat

between the lattice and thed spins.31

Conclusions Ultrahigh thermalh conductivity existsy in graphenen and ind carbonn

nanotubes because of theirf uniquer structure. It ist hard tod imag-

ine a material with a higher thermalr conductivity than carbon

atoms bonded tod each other inr a dense network. It ist diffi cult,fi

however, to make use of thisf extreme property. Thermally weaky

interfaces to graphene and nanotubes illustrate that “nano”

is not always a good approach. We often think thatk nanotech-

nology will improve things, but this is not always the case.

There are some real challenges integrating nanotubes as

solutions to thermal management problemst because interfaces

suppress thermal transport, and thed weak couplingk to the matrix

limits thermal performance of nanocomposites.f

We have achieved many advances in measurements over

the past several years. The technique of TDTRf hasR become

“one-stop shopping” for studiesr of thermalf conductivities of

a wide range of materialsf in a wide variety of forms:f bulk

materials, thin films,fi combinatorial samples, composites, and

polymer fir bers.fi We can perform experiments under pressure,r

study radiation effects, and mapd with high spatial resolution.

Conventional wisdom concerningm the lowest possiblet thermal

conductivity iny an densea solid isd not correct.t The search continuesh

for structuresr that cant producen even lowern thermalr conductivities

in densen materials, high thermalh conductivity iny organicn struc-

tures, and highd thermalh conductivity byy spiny waves.n

Acknowledgments Our workr onk thermal transport at UIUC has been supported

by NSF, ONR, DOE-BES, AFOSR, ARO, ARPA-E, and thed

Carnegie-DOE Alliance Center.

References 1. X. Zheng , D.G. Cahill , J.-C. Zhao, Adv. Eng. Mater. 7, 622 (2005).2. P. Keblinski, D.G. Cahill, A. Bodapati, C.R. Sullivan, T.A. Taton , J. Appl.Phys. 100 , 54305 ( 2006 ). 3. R.M. Costescu, M.A. Wall , D.G. Cahill , Phys. Rev. B 67 , 054302 ( 2003 ). 4. H.-K. Lyeo, D.G. Cahill, Phys. Rev. B 73, 144301 ( 2006 ).

Figure 8. Comparison of thef pressure dependent thermal

conductivity of icef at room temperature to prediction based

on the Leibfried–Schlömann equation and the pressure

dependence of thef Debye temperature and Grüneisen

parameter.9

Figure 9. Estimated magnetic thermal conductivity as a

function of temperaturef of thef two-leg spin ladder compounds

Sr14 Cu24O41 and Ca9 La5 Cu24O41 . Reprinted with permission from

Reference 10. ©2007, Springer-Verlag.



5. A.A. Balandin , Nat. Mater. 10 , 569 ( 2011 ). 6. B.A. Cola , J. Xu , C. Cheng, X. Xu, T.S. Fisher ,r H. Hu , J. Appl. Phys.101, 054313 ( 2007 ).7. S. Huxtable , D. Cahill , S. Shenogin , L. Xue, Rahmi Ozisik, P. Barone , M. Usrey,yM.S. Strano , G. Siddons , M. Shim, P. Keblinski , Nat. Mater. 2 , 731 (2003).8. C. Chiritescu , D.G. Cahill , N. Nguyen , D. Johnson, A. Bodapati , P. Keblinski,P. Zschack, Science 315 , 351 (2007).9. B. Chen, W.-P. Hsieh, D.G. Cahill , D.R. Trinkle , J. Li, Phys. Rev. B 83 , 13201(2011). 10. C. Hess, Eur. Phys. J. Spec. Top. 151 , 73 (2007).11. K.E. Goodson , Science 315, 342 (2007).12. J.R. Olson, R.O. Pohl, J.W. Vandersande, A. Zoltan, T.R. Anthony,y W.F. Banholzer ,r Phys. Rev. B 47 , 14850 ( 1993 ). 13. D. Li, Y. Wu , P. Kim, L. Shi, P. Yang, A. Majumdar,r Appl. Phys. Lett. 83 , 2934( 2003 ). 14. W. Jang, Z. Chen, W. Bao, C.N. Lau, C. Dames, Nano Lett.o 10, 3909 (2010).15. H. Fujishiro, M. Ikebe, T. Kashima , A. Yamanaka , Jpn. J. Appl. Phys.36, 5633 ( 1997 ).16. S. Shen, A. Henry,y J. Tong , R.T. Zheng, G. Chen , Nat. Nanotechnol. 5, 251 ( 2010 ).17. D.G. Cahill, Rev. Sci. Instrum. 75 , 5119 (2004).18. A.J. Angstrom , Ann. Phys. 114 , 513 (1861).19. A. Einstein, Ann. Phys. 35, 679 ( 1911 ).20. D.G. Cahill, S.K. Watson , R.O. Pohl , Phys. Rev. B 46 , 6131 (1992).21. S. Kim, J. Zuo, N. Nguyen, D.C. Johnson , D.G. Cahill , J. Mat. Res. 23 , 1064(2008).22. V.F. Petrenko , R.W. Whitworth, The Physicse ofs Icef (e Oxford University Press, UK, 1999). 23. G.S. Kumar ,r G. Prasad , R.O. Pohl , J. Mater. Sci. 28 ( 16 ), 4261 ( 1993 ). 24. D.G. Cahill , R.O. Pohl , Ann. Rev. Phys. Chem. 39, 93 ( 1988 ).25. D.G. Cahill, K.E. Goodson , A. Majumdar,r J. Heat Transfert 124 , 223 (2002).

26. M. Noh , J. Thiel, D.C. Johnson , Science 270 , 1181 ( 1995 ).27. W.-P. Hsieh, B. Chen, J. Li , P. Keblinski, D.G. Cahill, Phys. Rev. B 80, 180302(2009). 28. C. Lobban , J.L. Finney,y W.F. Kuhs, Nature 391, 268 ( 1998 ). 29. G.A. Slack, J. Phys. Chem. Solids 34 ( 2 ), 321 ( 1973 ). 30. Y.K. Koh, D.G. Cahill, Phys. Rev. B 76, 75207 ( 2007 ). 31. D.J. Sanders , D. Walton , Phys. Rev. B 15 , 31 (1977).

David G. Cahill has been a faculty member atthe University of Illinois at Urbana-Champaign(UIUC) since 1991. He earned his PhD degreein condensed matter physics from CornellUniversity in 1989 and then worked as apostdoctoral research associate at the IBMWatson Research Center. In 2005, he wasnamed Willett Professor of Engineering at UIUCand was appointed head of the Department ofMaterials Science and Engineering in 2010. Hisresearch program focuses on developing amicroscopic understanding of thermal transportat the nanoscale; the development of newmethods of materials processing and analysis

using ultrafast optical techniques; and advancing fundamental understanding ofinterfaces between materials and water. Cahill received the Peter Mark MemorialAward from the American Vacuum Society (AVS); is a fellow of the AVS, theAmerican Physical Society (APS), and the Materials Research Society; and ischair-elect of the Division of Materials Physics of the APS. Cahill can be reachedby email at [email protected].

For more information please visit www.mrs.org/mrc or email [email protected].

SOCIETY NEWSY


MRS seeks awardnominations for 2013Deadline: October 1, 2012

Mid-Career Researcher Awardwww.mrs.org/mraThe Materials Research Society is nowaccepting nominations for the Mid-Career Researcherr Awardr tod be presentedat thet 2013 MRS Spring Meeting in SanFrancisco, Calif.

The annual award recognizes ex-ceptional achievements in materials re-search made by mid-career profession-rals. It ist intended tod honor anr individual

who is between the ages of 40fand 52d at thet time of nomina-f

tion. Exceptions may be made for aninterruption in career progression dueto family or militaryr service. The awardrecipient mustt alsot demonstrate notableleadership in the materials area.

The award consistsd of af $5,000 cashprize, a presentation trophy, and ad cer-tifi cate.fi Meeting registration fee, trans-portation, and hoteld expenses to attendthe Materials Research Society SpringMeeting at whicht the award isd presentedwill be reimbursed.

The Mid-Career Researcher Awardis made possible through an endow-ment established by Aldrich MaterialsScience.

Outstanding Young Investigator Award www.mrs.org/oyiThe Materials Research Society is ac-cepting nominations for ther Outstanding

Young Investigator (OYI)r Award tod bepresented atd thet 2013 MRS Spring Meet-ing in San Francisco.

The OYI Award recognizesd outstand-ing interdisciplinary scientifi cfi work inkmaterials research by a scientist ort en-rgineer underr ther age of 36f (as of Janu-fary 1, 2013). The award recipientd musttshow exceptional promise as a develop-ing leader inr the materials area.

The award consistsd of af $5,000a prize,a presentation trophy, and ad citation cer-tificate.fi Reasonable travel expenses toattend the MRS Meeting at which theaward isd presented andd thed meeting reg-istration fee will be reimbursed.

Innovation in MaterialsCharacterization Awardwww.mrs.org/IMCAThe Materials Research Society is ac-cepting nominations for ther Innovationin Materials Characterization Award todbe presented at the 2013 MRS SpringMeeting in San Francisco, where the re-cipient ist invited tod speak atk thet AwardsCeremony.

The award recognizesd an outstandingadvance in materialsn characterization thatnnotably increases the knowledge of thefstructure, composition, in situ behaviorunder outsider stimulus, electronic, me-chanical, or chemicalr behavior, or otherrcharacterization featuren of materials.f It istnot limitedt tod the method ofd characteriza-f

tion or ther class of materialf observed.The annual award consists of af

$5,000 cash prize, a presentationa trophy,and a certifi cate.fi Meeting registrationfee, transportation, and hotel expensesto attend thed MRS Meeting at whicht theaward isd presented willd be reimbursed.

The Innovation in Materials Charac-terization Award hasd been endowed bydToh-Ming Lu and Gwo-Chingd Wang.

Materials ResearchSociety Fellowswww.mrs.org/fellowsThe Materials Research Society seeks torecognize as “MRS Fellow” outstandingmembers who are notable for theirr sus-rtained andd distinguishedd contributionsd tothe advancement oft materialsf research.It ist intended that—representingd excel-lence in science and engineering, anddedication to the advancement of ma-fterials research—the MRS Fellows willcollectively exemplify the highest idealstof accomplishmentf andt serviced embod-ied ind the MRS Mission.

Nomination is open to any MRSmember in good standing, whosemembership has been continuous forat least fivefi years preceding receipt ofthe nomination. MRS Fellow is a life-time appointment.

The deadline for submission of allfnominations is October 1, 2012. Forguidelines and applicationd forms, accessthe MRS website at www.mrs.org/awards or contactr Lorrit Smiley, MaterialsResearch Society, 506 Keystone Drive,Warrendale, PA 15086-7573,A USA;email [email protected].

Registration opens mid-September November 25 – 30, 2012 www.mrs.org/fall2012


FEATURES BOOKS

Composite reinforcements

for optimum performancePhilippe Boisse, Editor

Woodhead Publishing, 2011704 pages, $300.00ISBN 978-1-84569-965-9

The fundamental basis of materialsfscience is in understanding the re-

lationship between the microstructure ofmaterials and theird macroscopicr proper-ties. Unlike traditional materials, whichare usually treated asd homogeneous andisotropic, fiber-reinforcedfi compositescan be highly anisotropic depending onthe localized orientation of fi bers.fi Forfi berfi composites, researchers are veryheavy-handed with the microstructurein that they can control the internalstructure during the composites manu-facturing process through inducing fi berfiorientation duringn lay-up or bendingr andtwisting the fi bersfi using textile tech-niques. The book Composite reinforce-ments for optimumr performance is aimedat addressing this interrelationship be-tween the composite microstructure andthe resulting properties and performance.d

The book isk structured intod four pri-rmary sections: materials for reinforce-rments in composites; structures forreinforcements in composites; proper-ties of compositef reinforcements; and

characterizing and modelingd reinforce-ments in composites. The fi rstfi sectionhighlights the common fiberfi reinforce-rments and their properties,r along witha section addressing carbon nanotubesas an emerging reinforcement. The in-formation contained in this section isfundamental to most textbookst on fiber-fireinforced compositesd and isd presentedin a clear andr straightforwardd way.d

The second sectiond focuses primarilyon textile preform structures for com-posites, addressing woven and braideddstructures in both two-dimensional andthree-dimensional confi gurationsfi andmodeling their geometric properties.This section is comprehensive in thatit coverst the textile forming technologyand thed resulting fabric architectures andcomposite properties.

The third sectiond addresses the prop-erties of thef textiles and compositesd andhighlights both experimentalh and model-ding efforts. The finalfi section examinesthe characterization of andf modelingd ofcomposites. It alsot includes chapters on

modeling the mechanical properties atvarying scales and modelingd of differentfmanufacturing processes, such as of res-fin transfer molding,r injection molding,and fabricd draping/forming processes.

As with many books where multipleauthors contribute individual chapters,there often is some overlap. However,Composite reinforcements for optimumrperformance covers appropriate breadthand depth on a wide variety of topicsfthat aret not oftent found withind a singlereference. The treatment oft textilef com-posites, from a manufacturing, model-ing, and characterization viewpoint isparticularly strong. Two chapters in thebook addressk discontinuous fi berfi com-rposites composed ofd eitherf carbonr nano-tubes or shortr fi bers.fi Both chapters dealprimarily with processing in terms ofexamining fl ow-inducedfl orientation.d Thebook, while already rather long,r wouldbenefitfi fromt a chapter onr the mechanicsof discontinuousf fibersfi and composites.d

At thet University ofy Delaware,f I teachan advanced undergraduated and gradu-date-level course on composite materials,and Id think thatk thist book isk an excellentreference for both scientists and engi-neers working with, or studying, fi berficomposites. I plan to keep this book onkreserve at thet library as a reference forthe course.

Reviewer: Erik Thostensonk of thef De-partment oft Mechanicalf Engineeringland Centerd forr Compositer Materials atthe University of Delaware.f

Abstract Deadline November 1, 2012

Abstract Submission Site

Opens October 1, 2012

www.mrs.org/spring2013


CAREER CENTRAL

Sandia National Laboratories is one of thef country’s largest research facilities employing nearly 8,500 people at

major facilities in Albuquerque, New Mexico and Livermore, California. Please visit our website at www.sandia.gov.

We are seeking applicants for the President Harry S. Truman Fellowship in National Security Science and

Engineering. Candidates for this position are expected to solve a major scientific or engineering problem in their

thesis work or have provided a new approach or insight to a major problem, as evidenced by a recognized impact in

their field.

Sandia’s research focus areas are: bioscience, computing and information science, engineering science, materials

science, nanodevices and microsystems, radiation effects and high energy density science, and geoscience.

Candidates must meet the following requirements: U.S. citizenship, the ability to obtain and maintain a DOE “Q”

clearance, and a Ph.D. (3.5 undergraduate and 3.7 graduate GPA preferred), awarded within the past three years at

the time of application,f or completed Ph.D. requirements by commencement of appointment.f Candidates must be

seeking their first national laboratory appointment (pre postdoc internships included).

The Truman Fellowship is a three-year appointment normally beginning on October 1. The salary is $110,900 plus

benefits and additional funding for the chosen proposal. The deadline to apply is November 1st of eachf year.

For complete application instructions, please visit: http://www.sandia.gov/careers/students_postdocs/fellowships/

truman_fellowship.html

Please submit the complete package to: Yolanda Moreno, Sandia National Laboratories, P.O Box 5800, MS0359,

Albuquerque, NM 87185-0359, or email [email protected]. Please reference Job ID: 640971. All materials must

be received by November 1, 2012.

U.S. Citizenship Required.

Equal Opportunity Employer. M/F/D/V.

Harry S. Truman Fellowshipin National Security Science and Engineering

Operated byd

ASSISTANT PROFESSORMATERIALS CHEMISTRY

The Department of Chemistryf at the University atBuffalo (UB), State University of Newf York invitesapplications for ar tenure-track faculty position inexperimental materials chemistry at the AssistantProfessor level.r All areas of materialsf chemistrywill be considered; candidates with interests inimaging, microscopy, or spectroscopy areparticularly encouraged to apply. The successfulcandidate will contribute, through research andteaching, to a new interdisciplinary graduateprogram in Materials Science and Engineering atUB. An assistant professor isr expected to develop


CAREER CENTRAL

As part oft af major initiative,r the Institute of ChemicalfSciences and Engineering (ISIC) at EPFLt invites appli-cations for severalr faculty appointments in Chemical En-gineering. Exceptional applicants with expertise in en-ergy-related topics, including solar energyr conversion,chemical, electrochemical and biochemical energy con-version, carbon capture and utilization, are especiallyencouraged to apply. Appointments at thet Assistant Pro-tfessor level (tenure track) are envisioned, but seniorfaculty levels (Associate/Full) may also be considered.

A PhD in Chemical Engineering or ar related field and anexcellent trackt record of innovativef research and leader-ship are generally required. The successful candidatewill be expected to establish and direct at vigorous, inde-pendent research program and be committed to excel-lence in teaching at botht the undergraduate and graduatelevels.

Applications including cover letter,r curriculum, vitae,m publi-,cations list, concise, statements of researchf andh teachingd in-terests as well as the names and addressesd (including email)of fivef references should bed submitted ind electronicn formatvia thea website http://sbpositions.epfl.ch/applications/by October 15, 2012.,

The EPFL is consistently evaluated asd one of thef leadinguniversities in sciencen and engineeringd in Europe.n We offerinternationally competitive salaries, start-up resourcesand benefits.d The EPFL aims for ar stronga presence ofwomen amongst itst faculty, and, qualifiedd femaled candi-dates are strongly encouraged tod apply. More informationabout EPFLt and thed Institute of Chemicalf Sciences andEngineering can be found at:d http://www.epfl.ch/ andhttp://isic.epfl.ch. For additionalr information about thistcall for applications, please contact the director of thefinstitute, Prof., Paul Dyson ([email protected]).

Faculty Positions in Chemical Engineeringat the Ecole polytechnique fédérale de Lausanne (EPFL)

Located ind northern New Mexico,w Los, Alamos National Laboratoryl isya multidisciplinary researchy institution engaged ind strategic sciencec onbehalf off nationalf security.l Our. Centerr forr Integratedr Nanotechnologiesdhas the following opening:g

CHEMISTRY POSTDOCFocus on interfacial & colloidal chemistry and photophysicsof carbonf nanomaterials. Colloidal. & interfacial chemistrybackground required to support DOE program generatingsolution-based carbon nanomaterials for photovoltaicrapplications. Requires. experimental background/strongpublication record in colloidal chemistry or nanomaterialrcharacterization and PhD in Chem, Materials, Science, Physics,,or relatedr within past 5 years or soonr to be completed.

Director’s Fellowship or Marier Curie, Richard, P. Feynman,P. J.,Robert Oppenheimer, or, Frederickr Reinesk Fellowships possible.

Apply online at http://bit.ly/CINTPostdoc2 or visitrhttp://careers.lanl.gov and reference vacancy IRC9229.

AA/EOE

MINDS THAT MATTER

Material Sciences Division Director

Lawrence Berkeley National Laboratory (LBNL) is seeking a dynamicscientific leader to serve as the Division Director of thef Materials SciencesDivision (MSD). MSD is a research division of aboutf 700 staff andf guests,involving scientists at LBNL and faculty and students at UC Berkeley,with an annual budget of aboutf $75M. MSD is dedicated to designing,synthesizing, and characterizing the new materials, and discovering andunderstanding the new phenomena that will propel us into the futurehttp://www.lbl.gov/msd/. The successful candidate will have an excellentnational and international reputation and record of accomplishment.f

The Material Sciences Division Director is responsible for providingscientific leadership for the Division’s research programs as well asenhancing existing programs and developing new programs in materialssciences, condensed matter, experimental and theoretical physics,materials chemistry and biomolecular materials. In addition, will overseetwo major DOE user facilities, two research centers and build collaborativeprograms with UC Berkeley and other research institutions.

For a detailed positiondescription and instructionsregarding how to apply,please visit www.lbl.gov,access the careers page andreference job number 74877.

Berkeley Lab is an affirmative action/equal opportunity employer committed to thedevelopment of af diverse workforce.


CAREER CENTRAL

0612

R.K. Mellon Endowed Faculty Positions in Energy

As part oft thef University of Pittsburgh’sf strategic expansionof itsf Center forr Energyr (www.energy.pitt.edu), the SwansonSchool of Engineering invites exceptional applicants forendowed faculty positions at allt ranks in the following keygresearch areas:

Energy deliveryy andy reliability, with an emphasis on elec-tric power transmissionr and distribution systems, advancedpower electronicsr technologies (FACTS and DC systems),power systemr modeling andg analysis, power systemr opera-tion and control, and renewable energy integration.

Materials for energy-relatedr applications, with an em-phasis on experimental and/or computationalr efforts onstructural and functional materials used in harsh service en-vironments, and therefore including corrosiong engineering,catalysts, energy storage, thermo-electrics and sensors.

These key areas also complement ourt existingr andg emergingresearch and education activities in carbon managementn andtutilization, unconventional gasl resourcess , and direct energyt con-yversion andn recoveryd .yy

Established as part oft af recent $22t million gift fromt the Rich-ard King Mellong Foundation, a total of fourf endowedr facultypositions are available: two Professor-level appointments asR.K. Mellon Chairsn ins Energyn andy two Assistant/Associate Pro-fessor appointmentsr as R.K. Mellon Facultyn Fellowsy ins Energyn .yy

The successful candidates will greatly benefit fromt the re-sources fostered by the University of Pittsburgh’sf extensivefacilities, research partnerships, and close proximity to nu-merous energy-related companies and research laboratories.For instance,r the Department oft Energy’sf National EnergyTechnology Laboratory (NETL) recently formed a RegionalUniversity Alliance (RUA) for energyr technology innovationthat ist in partnership with the University of Pittsburghf andfour otherr nationallyr recognized universities.

Interested candidates or candidater teams should apply with asingle pdf filef of thef following: a cover letter;r a full curriculumvita; statements describing teachingg andg research interestsand plans; copies of threef representative publications; andthe names and contact informationt for atr leastt threet refer-ences. Questions and nominations should be addressed toProf. Brian Gleeson, Director ofr thef Center forr Energyrat [email protected].

For ther R.K. Mellon Chair inr Energy position, please apply at:[email protected]

For ther R.K. Mellon Faculty Fellow in Energy positions, pleaseapply at: [email protected]

Screening beginsg immediatelys andy willd continuel untile thel searcheis closed.s The Universitye ofy Pittsburghf ish ans Affirmativen Action,eEqual Opportunityl Employer.y

Phenomenex isx a leading provider of advanced technology solutions for separa-

tion science techniques in the areas of sample preparation, high-performance

liquid chromatography (HPLC), and gas chromatography (GC). The company has

the following job openings—Organic Surface Chemist and Sol Gel Chemist.

Organic Surface Chemist will work with R&D team to develop, enhance, or

investigate new and/or existing separation products and technology. Respon-

sible for researching, developing, and controlling sorbents, in particular their

surface modificationfi in order to convey particular properties useful in separation

science. This may include development of new HPLC, SPE, and other separation

science consumables.

Qualifi cations: PhD degree in Organic Chemistry or Polymer Chemistry re-

quired; industry work experience required; extensive knowledge of HPLC and

SPE production; at least fivefi years of experience in synthesis and derivatization

of silica and polymer.

Sol Gel Chemist will support and/or lead R&D efforts in the design and devel-

opment of high performance, sol gel based medias and technologies for use

in liquid chromatography and solid phase extraction.

Qualifi cations: PhD degree in Analytical Organic/Biochemical areas required

(Masters degree will also be considered); at least three years of experience in

inorganic Sol-Gel Chemistry; at least one year of work experience.

Submission of Application: Interested candidates should submit his/her C.V.

and cover letter to [email protected].

For more information about the position, visit www.phenomenex.com/careers.

CHEMISTS

U.S. Citizenship Required.

Equal Opportunity Employer. M/F/D/V.


CAREER CENTRAL

The Department of Physicsf in the School of Artsf and Sciences an-nounces a tenure-track faculty opening at the assistant professor levelrin Experimental Condensed Matter/Materials Physics. This hire is de-signed to enhance activities within the Physics Department and in a newInstitute for Materialsr Science and Engineering, which will be formallycommissioned in July 2013. The duties of thef position will include, butare not limited to, teaching and advising students, conducting originalresearch and publishing the results, and participating in departmentaland university service. A PhDA degree in a relevant fieldfi is required.Candidates are sought who have highly visible research achievementsand who have a strong aptitude for teachingr and mentoring studentsat the undergraduate and graduate levels. The appointment will beginFall 2013. Information on our departmentr can be found at http://www.physics.wustl.edu.

Applications should consist of thef following: cover letter,r current resumeincluding publication record, statement of researchf interests and plans(up to fivefi pages), statement of teachingf interests and approach (up tothree pages), and names and complete contact information (includingemail addresses) of threef references. Application materials must besubmitted electronically by email as a single filefi in editable (e.g., notpassword protected) PDF format tot [email protected] fullr consideration, applications should be submitted on or beforerNovember 1,r 2012.

Washington University isy ans equal opportunity/equall access/afl fiff rmativefifi action institution.

Women and minoritiesd ares encouraged tod apply.

The Center for Nanoscale Science and Technology (CNST) at

the National Institute of Standardsf and Technology (NIST) in

Gaithersburg, MD anticipates that it will soon have a vacancy

for a Project Leader in the Energy Research Group. The CNST

is a national user facility that supports nanotechnology from

discovery to production by providing industry, academia,

NIST, and other government agencies with access to world-

class nanoscale measurement and fabrication methods

and technology. The expected Project Leader position will

require an exceptional scientist or engineer with a strong

record of creativityf and achievement in the synthesis and

characterization of inorganic-basedf nanomaterials for ap-

plications in batteries, ultracapacitors, solid-state fuel cells,

and related electrochemical energy conversion and stor-

age technologies. The individual should have an extensive

background in chemical and/or materials science or related

disciplines, and a strong interest in developing new instru-

mentation and measurement methods for nanoscale charac-

terization of thef relevant chemical and physical phenomena.

The individual must possess the leadership abilities required

to build a thriving research program; mentor the research

The College of Engineering Sciences, University of Tsukuba, is seeking

a researcher for an appointment at the associate professor level from

1 April 2013 till 31 March 2017. The candidate should have experi-

ence in condensed matter and/or materials research in a broad sense

including soft matter research, either theoretical or experimental. As

part of the university’s drive to further global integration, the successful

candidate is also expected to contribute to English education and science

& engineering education in English at the College of Engineering Sci-

ences, the Graduate School of Pure and Applied Sciences, and through

campus-wide programs. The candidate must have a doctorate degree,

and English must be the candidate’s fi rstfi language.

Applications should be sent by REGISTERED MAIL to: Nobuyuki Sano,

Dean of the College of Engineering Sciences, University of Tsukuba,

Tsukuba, Ibaraki 305-8573, Japan, to arrive no later than October 19,

2012. See application details at http://www.tsukuba.ac.jp/update/jobs/

pdf/h24boshu_matse_en.pdf.

Inquiries should be directed to T.Takemori at [email protected].

ASSOCIATE PROFESSORCondensed Matter/Materials Research + English Education

FACULTY POSITIONYExperimental Condensed Matter/Materials Physics

PROJECTLEADER | ENERGY RESEARCH GROUP

Center for Nanoscale Science and Technology

National Institute of Standardsf and Technology

activities of postdocs;f have a successful record of interact-f

ing with multiple disciplines; be interested in contributing

to ongoing projects within the CNST and NIST related to

measurements of energy-relatedf processes, materials, and

devices; and be able to eff ectivelyffff communicate with a wide

variety of audiences.f

For additional information about the Center for Nanoscale

Science and Technology, please visit www.nist.gov/cnst.

Positions may be filledfi at any appropriate level (NIST pay

band III-V, current salary $60,989 to $153,200). NIST offT ersffff a

comprehensive benefitsfi package that includes, in part, paid

vacation, sick leave,k holidays, life insurance, health benefi ts,fi

on-site child care, and participation in the Federal Employees

Retirement System.

Candidates must have a degree in chemical science, materi-

als science, engineering, or physical science, or equivalent

experience combined with education. A PhD degree is desir-

able. Send expressions of interestf to [email protected].


CAREER CENTRAL

The Department of Materialsf Science and Engineering at Johns Hopkins University invites applications for ar junior-level, tenure-track faculty positionpreferably in computational biomaterials. Major areasr of interestf include cell mechanics, self-assembly, transport phenomena in biological systems,bio-inspired engineering, and tissue engineering. We will also consider outstandingr candidates in experimental biomaterials. Preference will be givento applicants at the assistant professor level,r but consideration will also be given to exceptionally qualifi edfi candidates at higher ranks.r

Johns Hopkins University offers world-class research and teaching environment in biological and medical sciences with extensive opportunities forcollaboration with the Johns Hopkins School of Medicine,f the School of Publicf Health and the Krieger Schoolr of Artsf and Sciences. Collaborativeopportunities also exist with the Institute of Nanobiotechnology,f the Whitaker Biomedicalr Engineering Institute, the Translational Tissue EngineeringCenter, the Johns Hopkins Engineering in Oncology Center, and the Center ofr Cancerf Nanotechnologyr Excellence.

The successful candidate will be expected to establish an independent, internationally recognized research program and to contribute fully to theundergraduate and graduate educational missions of thef department. Applicants should have a PhD degree or equivalentr in materials science andengineering or ar related fi eld;fi postdoctoral experience is desirable. Candidates must have demonstrated ability to undertake independent, interdisci-plinary, and collaborative research. Additional information about the department may be found at http://materials.jhu.edu.

All applications should be submitted electronically as a single PDF document to [email protected]. Applications should include a cover letterrdescribing the principal expertise and accomplishments of thef applicant, a complete resumé, statements of researchf and teaching interest, and thenames and contact information for atr least three references. For fullr consideration, applications should be received by November 1, 2012.

FACULTY POSITION Department of Materials Science and Engineering and Sheridan Libraries

The Department ist committed tod building a diverse educational environment;l women and underrepresentedd minoritiesd are strongly encouragedy tod apply. The Johns Hopkins University isy an EEO/AA Employer.A

FACULTY POSITION IN ADVANCED MATERIALSSchool of Materials Science and Engineering | College of Engineering, Architecture, and Technology

Oklahoma State University

The College of Engineering,f Architecture and Technology (CEAT) at Oklahoma State University (OSU) seeks applicants and

nominations for a tenure-track position at the assistant or associate professor level. The successful candidate will join an

existing group of facultyf in the Advanced Materials Program housed in the 123,000 square foot Helmerich Advanced Tech-

nology Research Center (HATRC) on the OSU campus in Tulsa. The vision for the HATRC is to be internationally recognized

for advanced materials research, graduate education, and new enterprise development.

Applicants should have research interests which complement thrusts in advanced/nanomaterials useful for energy systems,

biological/medical systems, and information technologies. There is a particular interest in candidates with background in

materials for energy systems such as batteries, fuel cells, and solar energy conversion. Applicants should have an earned

PhD degree in materials science and engineering or ar related field.fi Research experience beyond doctoral studies is desirable.

The successful candidate will be expected to develop an externally funded, internationally recognized research program in

advanced materials; to excel in teaching at both the undergraduate and graduate levels; and to work collaboratively across

the university and State.

Applications should include a letter ofr application;f curriculum vitae; descriptions of twof research projects with plans to secure

external funding; a statement of teachingf interests and philosophy; and the names and contact information of fif vefi references.

Applications should be submitted electronically to [email protected]. Review of applications will begin

immediately and continue until the position is fi lled.fi The target starting date is January 2013 if thef successful candidate is

available. More detailed information about the position and the HATRC may be obtained by visiting the College web site

(http://www.ceat.okstate.edu/).

Oklahoma State University isy an Equal Opportunity/Affil rmativefifi Action/E-Verify Employer.y


FEATURES POSTERMINARIES

Good reads for the materials researcher

Iread constantly. It is part of who and what I am. Reading, whether fiction or nonfi fiction, is a pleasure and helps me infi

life and in my work. I hereby share with you some recent reads, recommending them to your attention. I understand that some of you will not like any of these books, but I hope that most of you find one or more interesting and useful.fi The fi rst book on my list is fi The New Science of Strong Materials (or Why You Don’t Fall through the Floor) by J.E.

Gordon. This is a classic, first fipublished in 1968, which hasbeen reprinted several times.The edition that I read is in pa-perback published by PrincetonUniversity Press in 2006, with an introduction by science writ-er Philip Ball. The book delvesdeeply into strength, cohesion, stress, and strain. It covers cracks, crack stopping, and dislocations. Chapters addresscomposite materials, wood, ce-ramic, and metals. The book left me with a sense of wonder and

a deep appreciation for research in this area that has improved all of our lives. Next, I recommend the autobiography of Eric R. Kandel,who received the Nobel Prize in Physiology or Medicine in2000 for his work on memory storage in the brain. His book isIn Search of Memory (The Emergence of a New Science of the Mind). This technical autobiography covers details of his lifeas well as research during his lifetime into the phenomenology associated with memory storage and retrieval. These are areas in which I am profoundly ignorant, having overlooked themthroughout my education. Nonetheless, I found the book read-able and enjoyable. Kandel does an excellent job of juxtaposing his work with that of others in the fi eld. The edition that I read fiwas in paperback published by Norton in 2006. Mark P. Silverman has written several highly interesting and readable books. I recommend A Universe of Atoms, An Atom in the Universe. The version that I read was in hardcover published in 2002 by Springer-Verlag. It is a revised version of And Yet It Moves published by Cambridge University Press in 1993.Although the book is a smorgasbord of strange and interest-ing physics, I was most highly interested in his discussions of interference effects using electrons. Joe Jackson has written a marvelous book discussing his-torical events around the race to discover oxygen. The book,entitled A World on Fire (A Heretic, an Aristocrat, and the Raceto Discover Oxygen), follows the lives of Joseph Priestley (the

heretic), Antoine Lavoisier (the aristocrat), and others duringthe time before, during, and after the French revolution. This era saw the rise and fall of the phlogiston theory and its replace-ment with one of the foundations of modern chemistry. Thebook highlights the tragic lives of both of the main protagonists in light of the revolutionary era in which they lived. I read ahardback version of the book published in 2005 by Viking. Giancarlo Ghirardi has written an excellent book about thefoundations of quantum theory. The book, entitled Sneaking aLook at God’s Cards (Unravel-ing the Mysteries of Quantum Mechanics), covers the is-sues raised by superposition,interference, and entangle-ment. Although he discusses the usual topics in vogue inquantum information, includ-ing quantum cryptography,quantum communication, and quantum computers, the meat of the book is the discussion of quantum theory and nonlocal-ity. He highlights the disputes between Einstein and Bohr, particularly related to the famous Einstein–Podolski–Rosen(EPR) paper and Bohr’s response; and discusses hidden vari-ables approaches to quantum theory, philosophical notions such as contextuality, many-worlds theories, and quantum histories.He concludes the book with a study of dynamic reduction ef-forts, including an approach developed by himself, Alberto Rimini, and Tullio Weber. The book is deep in content about thefoundations of quantum theory, but should be accessible to thosewho have had a standard course in quantum mechanics and have knowledge of the Dirac notation of quantum states. The ver-sion of the book I read was in hardback published in Eng-lish by Princeton University Press in 2004. The book was originally published in Italianunder the title Un’occhiataalle carte de Dio (il Saggia-tore, Milano, 1997). I recommend any of the books written by Emil Wolf. I have several editions of Principles of Optics, theclassic work by Born and Wolf, and have used it to my advantage throughout my


FEATURES POSTERMINARIES

career. I recently read Professor Wolf’s book Introduction tothe Theory of Coherence and Polarization of Light. The version that I read is a hardback published by Cambridge UniversityPress in 2007. He treats both spatial and temporal coherence, second-order and higher order coherence effects, phenomena produced by multiple sources with differing states of coherence, coherence effects on polarization, scattering, and a unified treat-fiment of polarization and coherence. The book will be useful to those with a general interest in optics, to astronomers, and to materials scientists who deal with scattering, coherence, and polarization of electromagnetic fields.fi All of my career, I have been fascinated by elementaryparticle physics and the giant machines used by high-energy researchers to understand elementary particles. Sometime agoMichael Riordan wrote a history of the work to establish the quark theory, entitled The Hunting of the Quark (A True Storyof Modern Physics). The work sets the research in place with the personalities involved, from Murray Gell-Mann, Richard Feynman, Geoffrey Chew, Shoichi Sakata, George Zweig,Sheldon Glashow, Abdus Salam, Steven Weinberg, and a host of others. If you have ever wondered about the eightfold way, bootstrapping, S-matrix theory, partons, and quarks, this is ahighly readable history. The version that I read was a paperback published by Touchstone/Simon and Schuster in 1987. Nancy Thorndike Greenspan has written an excellent biog-raphy of Max Born entitled The End of the Certain World (The Life and Science of Max Born). The biography covers both per-

sonal and intellectual aspects of Born’s life and work. The book covers much of Born’s work, but highlights his work on quantum mechanics. Hisrelationships with Albert Einstein, Neils Bohr, Fritz Haber, Werner Heisenberg, Wolfgang Pauli, P.A.M. Di-rac, Max Planck, Pascual Jor-dan, and others are discussed in detail. The backdrop of the rise to power of Nazism and its impact on Born and oth-ers is addressed in detail. The book is highly readable. The

version that I read was a hardback published by Basic Books in 2005. Dietrich Stoltzenberg has written an excellent biographyof Fritz Haber entitled Fritz Haber (Chemist, Nobel Laureate,German, Jew). This book covers both personal and intellectualaspects of Haber’s life and work, highlighting significant areas fisuch as the Haber–Bosch process (synthesis of ammonia, nitro-gen fixation) and the Born–Haber cycle (calculation of lattice fienergies, the energy required to form a crystal from its ions).Haber’s relations with many of the chemists and physicists of that era are discussed in detail, including Albert Einstein, Max Born, Rudolf Stern, Richard Willstatter, Carl Bosch, Carl

Engler, James Franck, Walther Nernst, and Max Planck. Thebook discusses Haber’s industrial affiff liations, his academic fiwork and affiff liations, his Nobel Prize received in 1920, hisfiparticipation in the fi rst World War, and his exile from Germany fiafter the ascension of the Nazi Party. Haber’s life story is filled fiwith intellectual triumph and personal tragedy. The version of the book that I read was a hardback published in 2004 by theChemical Heritage Press.

For an introduction to special relativity, I highly recom-mend N. David Mermin’s book It’s About Time (Understand-ing Einstein’s Relativity). The version that I read is a hardback published in 2005 by Princeton University Press. It covers theusual paradoxes, moving clocks, synchronization of clocks,space-time geometry, and E = mc2. It is highly readable. Michael F. Barnsley has written an excellent book entitled SuperFractals (Patterns of Nature). The version that I read was in hardback published by Cambridge University Press in 2006. The book follows on previous work in iterated function systems to intro-duce new ideas such as su-per iterated function systems and fractal tops. The book re-quires knowledge of at least some calculus and may betruly accessible only to thosewith more mathematicalbackground. Nonetheless,the graphics are spectacular and alone worth the price of the book. The connectionsto real fi gures found in naturefiis amazing. So many of my colleagues have acquired an interest in the programming language called Scientifi c Python that I recently fidetermined to learn to program in this language. Consequently,I am working my way through several books to help: Begin-ning Python (From Novice to Professional) by Magnus LieHetland (Apress, 2008), Python Algorithms (Mastering BasicAlgorithms in the Python Language) by Magnus Lie Hetland (Apress, 2010), and A Primer on Scientifiii c Programming with fiPython by Hans Petter Langtangen (Springer, 2009). Althoughthe books are available in paper, I am reading them on my iPad using the Kindle app. Lest you believe that I am invested only in technical books, I also have some recommendations in fiction. For my friends fiat MRS Headquarters in Pittsburgh, I recommend the fantasy series by Wen Spencer including Tinker, Wolf Who Rules, Elf-home, and others in which a signifi cant part of Pittsburgh hasfibeen transported to be a part of a planet fi lled with elves. If you fiare interested in adventure and mystery, I recommend the Jack Reacher novels by Lee Childs, the novels of Dick Francis, and the Spenser novels of Robert B. Parker.

Steve Moss

XRDInnovation with Integrity

D8 DISCOVER with DAVINCI ULTRA GID:The ultimate in surface sensitivity for ULTRA thin films.

Contact us for more details and a system demonstration! www.bruker.com/d8discover

Push-button, motorized switching between in-plane and out-of-plane geometry

In-plane grazing incidence diffraction (IP-GID) using full line-focus

100 times higher intensity than traditional methods

Easy access to the full range of thin film XRD applications

Skimming the surface has never been this focused!

Callfor LIVEDemo

op

toelectro

nics

macro

mo

l e

sing

le crystal

diam

on

d m

icrop

ow

d

ium

Now

iuiu Inven

tu

mw

u.

diu

m sp

on

ge

semico

nd

u

Nd

:YAG

metam

aterials

euro

piu

m p

ho

green

techn

olo

gy

LED lig

htin

g

hafn

ium

tub

ing

carbo

n n

ano

tub

es titan

ium

rob

otic p

arts

99

.99

9%

ruth

eniu

m sp

heres

rare earth m

etalssp

intro

nics

go

ld n

ano

particles

ultra h

igh

pu

rity me

germ

aniu

m w

ind

ow

serb

ium

do

ped

fiber o

ptics

yttrium

foil

silicon

carbid

e

osm

ium

nickel fo

amq

uan

tum

do

ts p

latinu

m in

k

laser crystals

bio

synth

etics

shap

e mem

ory allo

ys

alternative en

ergy

galliu

m arsen

ide

galliu

m lu

mp

battery lith

ium

catho

de

tantalu

m

aerosp

ace ultra-lig

ht allo

ysirid

ium

crucib

les

cerium

po

lishin

g p

ow

der

dysp

rosiu

m p

ellets

tantalu

m allo

ys

atom

ic layer dep

ositio

n

thin

film

stron

tium

do

ped

lanth

anu

mIII-IV

nitrid

e materials

scand

ium

-alum

inu

m

org

ano

-metallics

crystal gro

wth

anti-b

allistic ceramics

CIG

S

sup

er alloys

dielectrics

ph

oto

voltaics

ovskite

con

es

pallad

ium

sho

tsu

rface fun

ction

alized n

ano

particles

solar en

ergy

iron

liqu

ids

ano

de

cob

alt

fuel cell m

aterials

erbiu

multra

vanad

ium

bu

ckey ba

sup

ercon

du

c

misch

metal

targets

© 2

00

1-2

01

1. A

me

rican

Ele

me

nts is a

U.S

. Re

giste

red

Tra

de

ma

rk.

AA

NM

ER

IC

ET

SL

EM

EN

®T

he

Wo

rld’s

Ma

nu

fac

ture

r of

En

gin

ee

red

& A

dv

an

ce

d M

ate

rials

catalog: americanelem

ents.com

TM

ion

ic

Co

Ni

Cu

Zn

Ga

Ge

As

SeB

rK

rK

Ca

ScT

iV

Cr

Mn

Fe

Al

SiP

SC

lA

rN

aM

g

BC

NO

FN

eLi

Be

He

H

Rh

PdA

gC

dIn

SnSb

TeI

Xe

Rb

SrY

Zr

Nb

Mo

TcR

u

IrP

tA

uH

gT

lPb

Bi

PoA

tR

nC

sB

aLa

Hf

TaW

Re

Os

Gd

Tb

Dy

Ho

ErTm

Yb

LuC

ePr

Nd

PmSm

Eu

CmB

kC

fEs

FmM

dN

oLr

ThPa

UN

pPu

Am

Mt

Ds

Rg

Cn

Uut

Uuq

Uup

Uuh

Uus

Uuo

FrR

aA

cR

fD

bS

gB

hH

s

Documents

Magazine