nanoscale materials and devices for future communication

IEEE Communications Magazine • June 2010112 0163-6804/10/$25.00 © 2010 IEEE

NANO-TECHNOLOGY IN COMMUNICATIONS

M. Saif Islam and Logeeswaran VJ, University of California, Davis

Nanoscale Materials and Devices forFuture Communication Networks

INTRODUCTION

During the last two decades, significant progresshas been made in controlling and engineeringnew materials on the nanometer-length scale —at the level of atoms, molecules, and supramolec-ular structures. Many of these nanostructuredmaterials have shown tremendous promise asbuilding blocks for scalable, miniature, and ener-gy-efficient electronics, photonics, magnetics,and electromechanical systems to transformcomputing and communication in the future.Key progress in all these areas will convincinglytransform future communications links to facili-tate faster data transfer rates for connectionswith ubiquitous intelligent ambient systems athome, in the office, and in public places. At pre-sent, these nanodevices for miniaturized proces-sors, memory, circuits, interconnects, andpossibly future self-powered computing systemswith unprecedented intelligence, energy efficien-cy, and scalability are progressively posing anengineering and technological challenge ratherthan remaining as a field of fundamentalresearch. A highly multidisciplinary endeavorand significant successes in the commercializa-tion of nanomaterials and nanodevices will nodoubt lead to strong potential for high economicimpact comparable to the telecommunicationtechnology of the 1990s and the informationtechnology growth of the last decade.

In this article we give an overview of the

anticipated profound impact of these innova-tions in nanomaterials and nanodevices in realiz-ing this vision, and the range of factors that arecritical to the success of the future communica-tion technologies, computing, and intelligentsensor networks.

NANOMATERIALS FORCOMMUNICATION DEVICES

ON-CHIP, CHIP-TO-CHIP, ANDFREE-SPACE INTERCONNECTS

Increased demand for seamless connectivity withintelligent ambient systems and unrestrictedmobility requires faster data transfer rates, whichleads to more memory and computing capabili-ties required in communication devices whileensuring small form factor and lower power con-sumption. For more than a decade, a great dealof work has been focused on improving the con-ventional interconnect technology for on-chip,chip-to-chip, and board-to-board communica-tions by reducing the resistivity of conductorslike copper (Cu) and reducing the dielectric con-stant of interlayer dielectric materials by usinglow-k polymers. As data rates approach 10 bil-lion b/s, the dimension of copper wires approach-es the mean free path of electrons in bulk Cu(40 nm at room temperature), resulting in exces-sive resistivity in the interconnects. This is large-ly due to the contribution of microscopicimperfections in the Cu wire, that signals travel-ing distances even as short as 50 cm are severelydeteriorated, causing higher thermal dissipation.Consequently, interconnects are considered oneof the most difficult challenges gigascale systemintegration faces. Currently, more than 70 per-cent of capacitance on a high-performance chipis associated with interconnects, and the dynamicpower dissipation associated with interconnectsis larger than that of transistors [1].

An additional issue with Cu wire intercon-nects carrying digital information has been thatthey are insufficient to provide the necessaryconnections required by an exponentially grow-ing transistor count, thus impeding the advanceof future ultra-large-scale computing. The para-sitic resistance, capacitance, and inductanceassociated with Cu electrical interconnects arebeginning to limit the circuit performance andhave increasingly become one of the primary

ABSTRACT

New discoveries in materials on the nanome-ter-length scale are expected to play an impor-tant role in addressing ongoing and futurechallenges in the field of communication.Devices and systems for ultra-high-speed short-and long-range communication links, portableand power-efficient computing devices, high-density memory and logics, ultra-fast intercon-nects, and autonomous and robust energyscavenging devices for accessing ambient intelli-gence and needed information will criticallydepend on the success of next-generation emerg-ing nanomaterials and devices. This article pre-sents some exciting recent developments innanomaterials that have the potential to play acritical role in the development and transforma-tion of future intelligent communication net-works.

ISLAM LAYOUT 5/18/10 11:45 AM 12

Authorized licensed use limited to: Qualcomm. Downloaded on July 21,2010 at 20:44:54 UTC from IEEE Xplore. Restrictions apply.

barriers to the progression of deep submicrome-ter ultra-large-scale integration (ULSI) technolo-gy. Primarily, the signal delay in interconnectshas been increasing, becoming a substantial limi-tation to the speed of digital circuits and result-ing in a slowing trend of clock speed inmicroprocessors. All these issues cannot beresolved with current technologies, and innova-tive methods and other novel techniques mustbe pursued to break the interconnect wall. Onepossible interconnect scheme is to use activephotonic interconnects and free-space opticalinterconnects. Figure 1 depicts several differentinterconnect technologies that could benefitfrom the rapid developments in the field ofnanomaterials and devices.

Photonic interconnects such as optical wave-guides or fiber optic cables can carry digital datawith three orders of magnitude capacity morethan electronic interconnects [2]. Unlike elec-tronic data, optical signals can travel tens ofkilometers without distortion or attenuation.Furthermore, several dozens of channels of high-speed data, each with a unique wavelength, canbe packed into a single fiber, in a techniqueknown as wavelength-division multiplexing(WDM). Today, 40 separate signals, each run-ning at 10 Gb/s, can be squeezed onto a hair-thin fiber. Despite all these attractivecharacteristics, the reasons photonic links arenot widely accepted yet are the high cost of thecomponents, large size, and difficulty of integra-tion with complementary metal oxide semicon-ductor (CMOS) circuitry.

Current optical communication devices arespecialized components made from GalliumArsenide (GaAs), Indium Phosphide (InP),Lithium Niobate (LiN), and other exotic materi-als that cannot easily be integrated onto siliconchips. That makes their assembly much more

complex than the assembly of ordinary electron-ics, because the paths the light travels must bemeticulously aligned to micrometer precision. Ina sense, the photonics industry is where the elec-tronics industry was a half century ago beforethe breakthrough of the integrated circuit. Thelogical way for photonics interconnects to moveinto the mass market is to introduce innovativeintegration, high-volume manufacturing, andlow-cost assembly. Fortunately, several nanoma-terials and new nanofabrication methods canplay a very important role in monolithically inte-grating different optical devices on a silicon plat-form, rather than separately assembling eachfrom diverse exotic materials, as outlined in thefollowing sections.

NANOHETEROEPITAXY FOROPTICAL INTERCONNECTS

Direct integration of an assortment of semicon-ductor nanostructures in devices and circuits onsingle crystal surfaces offers attractive opportu-nities in several areas of high-performance com-munication electronics, optoelectronics, sensing,energy conversion, and imaging. Beyond theseapplications, integration of a range of nanostruc-tures on amorphous surfaces offers unlimitedcapabilities for multifunctional materials anddevice integration. In fact, the possibility of low-cost electronics and photonics based on such anapproach would dwarf silicon photonics andother competing technologies. The key con-straints in growing planar epitaxial thin film of asemiconductor on another single crystal sub-strate are lattice and thermal expansion coeffi-cient mismatches, material incompatibilities, anddifferences in crystal structure [3]. These limita-tions can now be circumvented by growing semi-conductor nano-heterostructures that can

IEEE Communications Magazine • June 2010 113

Figure 1. Different interconnect technologies that could benefit from the rapid developments in the field ofnanomaterials and devices. Images courtesy of a) Intel Corporation and Getty Images, e) IBM, and f)Nature Publishing Group.

Fiber opticcommunication

betweenservers

New communication paradigms incomputer networks based on

nanomaterials

a

bc

d

e

f

Communicationbetween devicesand subsystems

Memory/logic

Logic

Optical/plasmonicinterconnects

Direct integration of

an assortment of

semiconductor

nanostructures in

devices and circuits

on single crystal

surfaces offers

attractive opportuni-

ties in several areas

of high performance

communication elec-

tronics, optoelectron-

ics, sensing, energy

conversion and

imaging.

ISLAM LAYOUT 5/18/10 11:45 AM Page 113


IEEE Communications Magazine • June 2010114

accommodate large mismatch due to their smallcrystallite dimensions. Recently, a number ofhighly crystalline III-V nanowires were grown onwafers with lattice mismatch as high as 8.1 per-cent [4]. The ability to grow dissimilar materialswith large lattice mismatches on a single sub-strate removes the integration related constraintsof incorporating photonics with electronics, andtherefore an associated challenge of on-chipelectronic-to-optical (E/O) and optical-to-elec-tronic (O/E) conversions for communication andinformation processing primarily using electronsand the majority of the information transferusing photons. Semiconductor nanostructureswith direct bandgap grown on silicon can offerhigh gain for designing laser diodes as well assuperior absorption characteristics needed forhigh-speed photodetection on a silicon surface.

Yi and coworkers recently reported the het-eroepitaxial growth of highly aligned InPnanowires on silicon substrates [5]. More recent-ly, VJ, Kobayashi, and coworkers introduced anew method of synthesizing III-V nanowires ona non-single crystalline surface that directly relax-es any lattice mismatching conditions anddemonstrated a device for high-speed photode-tection based on InP nanowires grown in theform of nano-bridges between prefabricatedelectrodes made of hydrogenated microcrys-talline silicon as shown in Fig. 2a–d [6]. Thedevice was fabricated on an amorphous glasssurface and was measured to have a bandwidthabove 30 GHz. Grego and coworkers reportedthe integration of bridged nanowires between apair of vertically oriented non-single crystal sur-

faces etched into rib optical waveguides to designnanowire integrated waveguide photodetectorson amorphous surfaces (Fig. 2e and 2f) [7].These capabilities offer opportunities for ultra-fast low-cost free-space optical interconnect forfuture computers and servers. Heteroepitaxiallygrown optically active materials such as III-Vand II-VI for integration with mainstream Sitechnology may enable low-cost, and highly inte-grated ultra-fast devices due to their high carriermobilities and optical absorption coefficients.This opens opportunities for a wide variety ofapplications including intrachip, interchip, andfree-space communications. The capabilities ofgenerating and detecting photons by directbandgap materials on Si substrate, which isknown for its low efficiency in electron-to-pho-ton conversion, will bring about a myriad ofchallenges along with revolutionary opportuni-ties that will impact a large sector of the high-tech industry. Several barriers still remain thatimpede this technology transition from the labo-ratory to real-world applications.

OUT-OF-PLANE OPTICAL INTERCONNECTSOptical pillars and wires along with micropho-tonics technology were employed to demonstrateexternal optical interconnects that can connectdifferent parts of electronic chips via air or fibercables at the expense of a bulky configuration[8]. Unfortunately, optical devices (laser, modu-lators, and detectors) and light guides are~1000× larger in physical dimensions than elec-tronic components, and it is very challenging tocombine these two technologies on the same cir-

Figure 2. InP nanowire photodetector on a glass substrate: a) schematic diagram; b) SEM image; c) schematic close-up; d) 30 GHzhigh-speed data pulse generated by the device [6]; e) bridged axial silicon nanowires on an optical waveguide; f) bridged siliconnanowires between a pair of silicon electrodes.

SEM image of InP Nanowire photodetector

SiO2 waveguide

ActiveNanowire

device

Si NanowirePhotodetector on glass

InP NanowirePhotodetector on glass

b

a

fSEM image of bridged Si Nanowire

c

Ti/Pt

eIntersecting

InPNanowires

Ti/Pt

μc-Si:HG

S

G

SiO2

ActiveNanowire

deviceCoplanar

waveguide

Nanoheteroepitaxy for optical interconnect

Time (ps)High speed data pulse

100

-0.2

Am

plit

ude

(mV

)

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0 200 300 400

DC bias = 1V

18ps

500

d5 μm

IntersectingNanowires




cuit. State-of-the-art electronic devices are fabri-cated with feature sizes in the range of tens ofnanometers, and emerging electronic devicessuch as single electron transistors are designedwith sub-nanometer dimensions. On the otherhand, optical devices can reach a theoretical sizelimit on the order of the wavelengths (~1 μm) ifsophisticated techniques such as photonic crys-tals are used.

PHOTONIC CRYSTALSThe concept of photonic bandgap crystals hasbeen around for more than two decades. A linedefect within a two-dimensional photonicbandgap crystal provides efficient spatial con-finement of light, and works as a building blockin a variety of routing and processing schemes oflight. In contrast, silicon in the form of theCMOS technology platform has been the coredriver in microelectronics. Silicon nanophoton-ics, which allows CMOS platforms to handlelight, thus would offer a wide range of photonicfunctions required for CMOS platforms to pushfurther progress in high-speed data transferrates. Nanophotonic implementations of wave-length multiplexers and demultiplexers byemploying photonic crystals or non-periodicnanophotonic structures offer greatly reducedfootprints and enhanced robustness to fabrica-tion tolerances and temperature variations.While the photonic bandgap crystal and siliconnanophotonics are still subject to the diffractionlimit of light, photonic devices that use surfaceplasmon polaritons (SPPs) and/or energy trans-fer mechanisms relying on optical near-fieldinteractions would pave the road toward ulti-mate photonic integration beyond the diffractionlimit of light.

PLASMONICS FOR INTERCONNECTSAs discussed previously, to overcome the sizelimitation of photonic interconnects, researchersintroduced devices and interconnects based onSPPs, which are electromagnetic waves thatpropagate along the surface of a conductor sup-ported by the fluctuations in the density of elec-trons on a metal surface — which can absorblight, travel along the surface, and re-emit thatenergy. In this manner, plasmonics offers thesynergistic bridge between nanophotonics andnanoelectronics. Initially, plasmonics was mainlyfocused on passive routing of light in waveguideswith dimensions much smaller than the wave-length of the light. However, as the propagationlength in such high-confinement SPP waveguidesis limited to a few tens of micrometers, theyhave not been considered as a better substitutefor high-index dielectric waveguides. It is impor-tant to note that the size of dielectric waveguidesis limited by the fundamental laws of diffraction,which are orders of magnitude larger than theelectronic devices on a chip. Plasmonic devicesand their sub-wavelength dimensions are unique-ly capable of reconciling this mismatch in size,bridging dielectric microphotonics and nanoelec-tronics. Plasmonic devices such as nanoscalelasers, modulators, and detectors offer the poten-tial to generate, modulate, and detect optical sig-nals while transporting them at the speed ofphotons through the same thin metal circuitry.

This offers the ability to combine the superiortechnical advantages of photonics and electron-ics on the same chip. A big challenge to plas-mon-based communication in computer chips isgenerating light from plasmonic lasers that areelectrically biasable and compatible with silicon-based CMOS fabrication techniques. Recentdemonstrations of plasmon-enhanced lightsource and detectors are significant steps towardthis goal of realizing new generation of ultrafastnanoelectronics [9].

FUTURE PROSPECTS OFNANOPHOTONIC COMMUNICATION

Although the recent trend of using multicoreprocessors is providing a temporary respite fromstagnation of microprocessor clock frequencies,it has created daunting challenges to pro-grammability, and ultimately drives today’s sys-tem architectures toward extreme levels ofunbalanced communication-to-computationratios. Interior heat removal capability com-pounded with a huge deficit in chip input/output(I/O) bandwidth due to insufficient I/O intercon-nect density has stalled our high-performancegains. The excessive access time of a micropro-cessor for communication with the off-chip mainmemory is among the major hurdles holdingback ultra-fast computing. Nano-photonics hasthe potential to offer a disruptive technologysolution that will not only enhance the perfor-mance of communications and future computers,but will also contribute to a fundamental trans-formation in the field of computer architecture.Proven ultra-high throughput, minimal accesslatencies, and low power dissipation of nano-photonic devices remain independent of capacityand distance.

Although individual local interconnects arediminutive, their cumulative length is ratherlong, and their share in the delays of criticalpaths and power dissipation is large and grow-ing. For instance, more than 50 percent of inter-connect power is dissipated in localinterconnects.

While for a distance exceeding a few millime-ters energy efficiency for electrical signaling istypically ~10 pJ/b (equivalent to ~10 mW/Gb/s),photonic communication links based onnanophotonics have shown the ability to operateat 100 fJ/b, with potential to reduce the cost tothe level of 10 fJ/b. The unique capability ofphotons for multiwavelength photonic intercon-nects in the DRAM I/O can continue the scala-bility advantage that semiconductor industry hasbeen accustomed to in the last few decades. Theemergence of multicore architectures and theever increasing demand for information process-ing naturally make room for nanophotonics inthe field of intrachip and interchip communica-tions.

Unlike prior generations of optical technolo-gies, nanoscale photonics offers the possibility ofcreating highly integrated platforms with dimen-sions and fabrication processes compatible withnanoelectronic logic and memory devices, and isexpected to have a great influence on how futurecomputing and communication systems willscale. Nanophotonic interconnects can potentially

Although individual

local interconnects

are diminutive, their

cumulative length is

rather long, and

their share in the

delays of critical

paths and power

dissipation is large

and growing.

For instance, more

than 50 percent of

interconnect power

is dissipated in local

interconnects.




revolutionize future computing and communica-tion by bringing parallelism, immense intercon-nection bandwidths, and wavelength routingcapabilities without electrical impedance orcrosstalk limitations.

NANOSENSORS FOR EMBEDDEDUBIQUITOUS INTELLIGENCE

SENSING AMBIENT INTELLIGENCEFuture communication infrastructure withembedded intelligent and autonomous deviceswill be equipped with a large number of robustsensors to ensure seamless connectivity withambient and intelligent networks. In the recentpast, researchers have been actively developingnanoscale sensors, instigating a new era of ubiq-uitous nanosensors. Nanosensors were realizedbased on the amazing progress in the synthesisof nanomaterials with rationally predictable size,shape, and surface properties resulting inextraordinary sensitivity [10].

When reduced to nanoscale dimensions,many materials reveal new properties that canbe desirable for a variety of sensing applications.These attractive characteristics can stem fromhigh surface-to-volume ratios at the nanoscalealong with changes in optical properties (reflec-tivity, absorption, and luminescence), volumetricor surface diffusivity, thermal conductivity, heatcapacity, mechanical strength, and magneticbehavior. Many of these effects can be harnessedto develop new sensitive detection techniques.Examples include utilizing the inherent ultra-high specific surface areas of materials struc-tured (or porous) at the nanoscale to enhancethe reaction rate or overall reactivity with adesired analyte. A single nanosystem will be able

to identify many analytes, in a wide range ofbackgrounds, with great sensitivity. It is antici-pated that in the not-too-distant future, nanosen-sors will become part of our everyday intelligentenvironment, collecting and transmitting enor-mous amounts of information about our healthand wellness, life, and environment, bringing ahost of benefits to our lives. Potential applica-tions include safe driving systems, smart build-ings and home security, smart fabrics ore-textiles, manufacturing systems, transportation,and rescue and recovery operations in hostileenvironments. Indeed, the inundation of a num-ber of nanosensors is predicted to transform theway we interact and communicate with the envi-ronment and our surroundings. The abundanceof inexpensive omnipresent sensors and theirpotential marriage with biotechnology and infor-mation technology will radically change humanlives and communication with the environment.Emerging research in nanosensors continues tobe among the broadest and most dynamic areasof science and technology, bringing togethernumerous researchers in biology, chemistry,physics, biotechnology, medicine, and manyareas of nanoengineering. However, thewidespread applications of nanosensors may alsobring about a myriad of challenges in processingand analyzing the sensor data, and interpretingthe information in an efficient manner.

Integration of various components into sen-sor systems will be a significant technical chal-lenge. The real challenge in nanosensors now isto create the integrated nanoelectronic andmicro-electronic circuits in the form of a univer-sal platform that will allow speedy deploymentas new needs emerge, and new intuitions andunderstanding for rapid transmission and pro-cessing of colossal sums of data. Practical meth-

Figure 3. SEM images of sensors designed with: a) vertically bridged nanowires; b) laterally bridged nanowires; c) ultra-sharp Ga2O3nanowire; d) nanowire FET chem-bio sensor (Image courtesy: Lieber et al. Harvard University); e) carbon nanotube bio-sensor (imagecourtesy: Pacific Northwest National Laboratory); f) molecular functionalization of sensors using dip-pen lithography (image courtesy:NanoInk Inc.); g) Surface Enhanced Resonant Imaging (SERS) of biomolecules (image courtesy: Royal Society of Chemistry); h)schematic of a nanostructure-based gas ionization sensor.

CounterelectrodeAperture

Spac

er

Ohmiccontacts

Silicon nanowire array

Top electrode

Bottom electrode

Nanowire

Nanowire

2μm2μm

Oxide

a

e f g h

b c d

D

S




ods for integrating dissimilar materials anddevices on a substrate need to be developed.Monolithically integrated nanoscale semicon-ducting materials, with diverse bandgap andelectrical and optical properties, could offer theultimate range of capabilities in nanosensors.The realization of a ubiquitous nanosensor withlow-powered and self-powered computing net-works could enable a vast number of novel appli-cations and possibilities of ultra-low-powerwireless sensor networks that have not been pos-sible before. Although many of the excitingdevelopments are only recent, the basic concep-tual issues have now been widely debated andreasonably clarified. At the time of writing thisreview, major theoretical fundamentals havebeen laid and numerous experiments have beenperformed, placing the field of nanosensors on a

sound technological footing. Figure 3 presentssome of the nanomaterials-based sensors demon-strated by the research community in the recentpast.

NANOMATERIALS FOR COMPUTING

MATERIALS FORCOMPUTING LOGIC AND MEMORY

In order to miniaturize devices beyond the ulti-mate limit for conventional CMOS devices, newconcepts such as molecular scale electronics [11],memristors [12], low-dimensional semiconduc-tors, and organic structures and single-electrontransistors are being developed as replacementsto Si-based technologies to address the everincreasing communication speed that requires

Figure 4. Various emerging nanomaterials based communication devices and components: a) DNA can beused as a computing device as well as scaffolds to generate highly organized and complex superstructuresthrough self-assembly; b) CNTs; c) graphene has the highest mechanical strength, enormous intrinsicmobility, zero effective mass, and an electron can travel for micrometers evading scattering at room temper-ature; d) metal catalyzed nanowires can be used to connect two chips by growing them in the shape of verti-cal bridges; e–f) molecular electronic switching junctions and their current-voltage property; g) an array of17 purpose-built oxygen-depleted titanium dioxide memristors [12]; h) nano-antenna based on metamate-rials; i–-j) plasmonic and photonic crystal waveguides capable of confining light to sub-10-nm dimension;k) a nanoscale CNT radio; l–m) molecular scale machines: molecular motors. Image courtesy: A. Zettl,University of California at Berkeley.

Nanoscale materials and

devices for future computing and communication

k

l a

b

c

d

e

f

g

h

i

Plasmonic waveguide

Photoniccrystalj

Carbonnanotube

radioMolecular

motorDNA nanowire

Multi-wall carbonnanotube Graphene

Molecular scale switching junction

Memristor

Metamaterial nanoantenna

Voltage (V)-0.1-0.2

0

???

1 2 3

-1 -2 -3

0.0 0.1 0.2

Off

On

CMOS integrated nanowire interconnects

At the time of

writing this review,

major theoretical

fundamentals have

been laid and

numerous

experiments have

been performed,

placing the field of

nanosensors

on a sound techno-

logical footing.




increasing computational capability with limitedpower. Consequently, nanoelectronics beyondCMOS will take very different approaches inboth active device structures and interconnects.This will require new nanomaterials and inter-connect innovation, fabrication techniques, and,most important, fundamental understanding ofthe interface with devices, low-/high-frequencycarrier transport properties, and long-termdevice reliability.

The rapid advancements in the creation ofnanometric devices will push the limit of minia-turization to unprecedented levels and possiblyexhibit new properties stemming from couplingstandard electronic components with unconven-tional electronic materials such as molecules,redox active proteins, semiconductor nanocrys-tals, nanotubes, grapheme, chalcogenides, andeven organic materials such as DNA [13], toname but a few (Fig. 4). One interestingapproach to device miniaturization is molecularelectronics, which utilizes single molecules orsmall groups of molecules as components inelectronic devices; with feature lengths as smallas 1 nm, it holds the promise of allowing thenext generation of device scaling into the nano-scale range. These new devices must be integrat-ed into complex information processing systemswith billions and eventually trillions of parts, allat low cost. Fortunately, these molecular-scalecomponents lend themselves to relatively large-scale manufacturing processes based on chemicalsynthesis and self-assembly techniques. By takingadvantage of these key tools of nanotechnology,it may be possible to put a cap on the amount oflithographic information required to specify acomplex system, and thus a cap on the exponen-tially rising cost of semiconductor manufacturingtools.

MEMRISTORSA direct byproduct of research in molecular elec-tronics is the demonstration of a new electricalelement called the memristor that not onlybehaves like a resistor, which simply resists theflow of electric current, but also displays thecapability to remember the last current it experi-enced. In the world of integrated circuits, thisability to remember the last current would usual-ly require many different components. Williamsand coworkers, inventors of memristors — thefourth fundamental circuit element along withcapacitor, resistor, and inductor — believe thateach memristor can take the place of 7 to 12transistors and hold its memory without power[12]. In contrast, transistors require power at alltimes, so there is a significant power loss throughthe leakage currents. For decades, attempts tobuild an electronic intelligence that can mimicthe amazing power of a human brain have seenlittle success because of this missing crucial elec-tronic component — the memristor. By combin-ing the properties of memristors with those ofcapacitors and inductors to produce compounddevices called memcapacitors and meminductors,complex processors resembling the human cortexsynapses can be developed with a packing densityof about 1010/cm2 — 10 times more than today’smicroprocessors can offer. It is now speculatedthat it would be possible to implement more than

50Gbytes of low-power memory in mobile devicesin the near future based on memristors.

DNA ELECTRONICSAmong a few other organic nano-materials, DNAis currently regarded as an ideal scaffold fordirected organization of materials at thenanoscale, thanks to its intrinsic physico-chemicalproperties. For instance, its length and sequencecan easily be defined by enzymatic and syntheticmethods with nanometric precision as each basepair contributes about 0.34 nm to a linear DNAstrand. Furthermore, the DNA sequence can beprogrammed to generate highly organized andcomplex superstructures through self-assembly.Lastly, the propensity of two complementaryDNA strands to hybridize into a duplex evenwhen removed from their native environmentadds an important recognition element that canbe exploited for guiding the interaction with sur-rounding objects. Kiehl and coworkers usedDNA, peptides, and proteins as scaffoldings forself-assembling nanoparticles, carbon nanotubes,and molecules into electronic circuitry [14]. Thisapproach could be used to integrate novel devicesat densities far beyond those possible with litho-graphic techniques. Several studies on DNAexhibit a range of electron transport behavior,which is currently dependent on the experimentalconditions. For instance, it has been demonstrat-ed that DNA can act as an insulator, a semicon-ductor, a conductor, and even a superconductor— offering opportunities for future bio-compati-ble devices and circuits.

CARBON-BASED NANOMATERIALSCarbon-based nanomaterials are currently con-sidered among the wonder materials expected tohave a profound impact on future electronics,photonics, biomedical, energy storage, and con-version devices. Carbon nanotubes (CNTs) are10 times as strong as steel with one-sixth theweight. CNTs, which are rolls of one-atom-thickcarbon sheets, show great potential in addressingsome of the key interconnect challenges in futuregenerations of technology, when copper conduc-tivity will degrade substantially because of sizeeffects, and thereby extend the lifetime of electri-cal interconnects. Some of the fascinating prop-erties of carbon nanotubes include very largecurrent conduction capacity, large electron meanfree paths, resistance to electromigration, highmechanical strength, and stability. Graphene,which is a one-atom-thick carbon sheet, is thethinnest material known to date with the highestmechanical strength, enormous intrinsic mobility,zero effective mass, and an electron can travelfor micrometers evading scattering at room tem-perature. This magnificent material can sustainsix orders of magnitude higher current densitiesthan copper, shows record thermal conductivityand stiffness, is impermeable to gases, and recon-ciles such conflicting qualities as brittleness andductility. Relativistic electron transport ingraphene makes it unparalleled in speed forapplications in ultra-fast logic, detectors, sensors,and communication circuits.

Other nanodevices currently being studiedinclude nanomagnetics for spintronics, nanoma-terial-based phase-change memory, and cellular

A direct byproduct

of research in molec-

ular electronics is the

demonstration of a

new electrical

element called the

memristor that not

only behave like a

resistor, which simply

resists the flow of

electric current, but

also display the

capability to remem-

ber the last current it

experienced.




automata realized with extremely low levels ofelectrical currents or charges. Nanomagneticsprovide an opportunity to realize zero-energyswitching logic gates and memory. Phase changememory (PCM) is considered one of the mostpromising candidates for next-generation non-volatile memory, based on its excellent charac-teristics of high speed, large sense margin, goodendurance, and high scalability. Nanoscale quan-tum dots, wires, and wells in resonant nanopho-tonic structures can provide extremely strongmodulation and switching effects with low powerexcitations. The combination of quantum con-finement for electrons with the nanoscale con-finement of photons may lead to very excitingpossibilities in optical switching. Such switchescould operate at very high speeds with energycost per bit much lower than electronicapproaches, enabling greater intelligence in theon-chip communication networks that are crucialfor advanced computing architectures. Conven-tional semiconductor flash memory devicesrequire relatively large program/erase voltages(exceeding 10 V); hence, the amount of energyrequired to store one bit of information exceeds10 fJ, whereas the energy required to store onebit of information in a volatile memory (SRAM)cell (~1 V operation) is less than 1 fJ. Recently,nano-electro-mechanical memory (NEMory) celldesigns have been proposed and demonstrated[15]. These can be very compact and operatewith low voltages (less than 3 V) to achieve lowprogram/erase energy (<1 fJ). The device isanother strong candidate for future embeddednon-volatile memory applications because itpotentially offers a dramatic reduction (morethan 100 times) in program/erase energy.

Computing today is facing a catastrophicpower crisis that has stalled progress across allscales — from handheld devices and personalcomputers to supercomputing systems. Futureadvancement is simply too challenging with con-ventional semiconductor scaling, and breakingthis barrier, the so-called power wall, will clearlyrequire paradigm shifting innovations and trans-formative opportunities that can only be createdby fundamental breakthroughs in nanoscale sci-ence and technologies. At the same time, to dealwith the inaccuracies and instabilities introduced

by fabrication processes and the tiny devices,future nanomaterial-based computer architec-tures have to be able to tolerate an extremelylarge number of defects and faults. This will leadto fault-tolerant architectures for the ultra-largeintegration of highly unreliable nanometerdevices. A significant roadblock to wide-scaleintegration of functional nanomaterial-baseddevices is the difficulty in forming contacts tothem, controlling their surfaces, and purifyingthem to ensure identical physical and electricalproperties. In addition, in all practical devices, itis important to pattern nanoscale electrical con-tacts and interconnects and to fabricate low-resistance Ohmic interfaces to the devices. Allthese challenges are limiting the immediateapplications of these materials and devices incurrent technology.

CONCLUSION:A NEW ERA OF NANOMATERIALS

Nanomaterials — in both unrefined and moreadvanced and engineered forms — are the keyto future technology development and industrialrevolutions, and thus to better and more securesocieties. The ages of ancient civilizations areknown to be categorized by their materials. Dif-ferent eras of history are named after the mate-rials that defined them, for example, the StoneAge, the Bronze Age, and the Iron Age. Truly,the present age will be known as the Nanomate-rials Age, as they should provide our society thenecessary tools to address many pressing techno-logical barriers. Indeed, in the last decade,nanoscience has reached the status of a leadingscience with fundamental and applied implica-tions in all basic physical sciences, life sciences,and earth sciences, as well as engineering andmaterials science.

Several innovations based on nanoscaledevices have the potential to play a unique andimportant role in enhancing the data transmis-sion speed of future communication systems.Consequently, the performance of communica-tion links is experiencing an impact as depictedin Fig. 5. This graph shows the map of energyefficiency and data transfer rate with the dimen-

Figure 5. Performance of communication links with current and emerging nanoscale devices and systems.

Nano-optics

Micro-optics

Nano-optics

Electrical

Dimension of communication devices Length of communication link

Optical links

Dat

a tr

ansf

er r

ate

Ener

gy e

ffic

ienc

y

Electrical




sions of different device technologies. In thepast, devices were relatively slow and bulky. Thesemiconductor industry has performed an incred-ible job in scaling electronic devices to nanoscaledimensions. Unfortunately, interconnect delaytime issues provide significant challenges towardthe realization of purely electronic circuits oper-ating above several gigahertz. In stark contrast,photonic devices possess enormous data-carryingcapacity. Unfortunately, dielectric photonic com-ponents are limited in their size by the laws ofdiffraction, preventing the same scaling as inelectronics. On the other hand, plasmonics andemerging quantum communication techniquesoffer precisely what electronics and photonics donot have at present: the size of electronics andthe speed of photonics. While at this stage it isimpossible to predict the eventual timeline ofdeployment and applications of new nanodevicesin scalable commercial systems, resolute efforton overcoming the roadblocks to wide-scale andcost-effective integration of functional nanoma-terials in communication systems, and under-standing their environmental impact, are likelyto keep the scientific community occupied in thenext decade.

REFERENCE[1] A. Naeemi, R. Sarvari, and J. D. Meindl, “Performance

Comparison between Carbon Nanotube and CopperInterconnects For Gigascale Integration (GSI),” IEEEElectron Device Letters, vol. 26, 2005, pp. 84–86.

[2] S. A. Maier et al., “Experimental Demonstration ofFiber-Accessible Metal Nanoparticle Plasmon Wave-guides for Planar Energy Guiding and Sensing,” AppliedPhysics Letters, vol. 86, 2005, article ID: 071103.

[3] H. Mori et al., “GaAs Heteroepitaxy on an Epitaxial SiSurface with a Low Temperature Process,” AppliedPhysics Letters, vol. 63, 1993.

[4] M. T. Björk et al., “One-Dimensional Steeplechase forElectrons Realized,” Nano Letters, vol. 2, 2002, p. 87.

[5] S. S. Yi et al., “InP Nanobridges Epitaxially Formedbetween Two Vertical Si Surfaces by Metal-CatalyzedChemical Vapor Deposition,” Applied Physics Letters,vol. 89, 2006, article ID: 133121.

[6] Logeeswaran VJ et al., “A 14 ps Full Width at Half Max-imum High-Speed Photoconductor Fabricated withIntersecting InP Nanowires on an Amorphous Surface,”Applied Physics A, vol. 91, 2008, pp. 1–5.

[7] S. Grego et al., “Waveguide-Integrated Nanowire Pho-toconductors on a Non-Single Crystal Surface,” Proc.SPIE, vol. 7406, 2009, article ID: 74060B.

[8] M. Bakir et al., “Mechanically Flexible Chip-to-SubstrateOptical Interconnections using Optical Pillars,” IEEETrans. Advanced Packaging, vol. 31, no. 1, 2008, pp.143–53.

[9] R. J. Walters et al., “A Silicon-Based Electrical Source ofSurface Plasmon Polaritons,” Nature Materials, vol. 9,2010, pp. 21–25.

[10] M. Saif Islam, “The All Pervading Nanosensors,” Int’l. J.Nanotech., Special Issue on Nanosensors, vol. 5, no.4/5, 2008.

[11] C. Joachim, J. K. Gimzewski, and A. Aviram, “Electron-ics using Hybrid-Molecular and Mono-MolecularDevices,” Nature, vol. 408, 2000, pp. 541–48.

[12] D. B. Strukov et al., “The Missing Memristor Found,”Nature, vol. 453, 2008, pp. 80–83.

[13] H. A. Becerril et al., “DNA-Templated Three-BranchedNanostructures for Nanoelectronic Devices,” J. Amer.Chem. Soc., vol. 127, 2005, pp. 2828–29.

[14] Y. Y. Pinto et al., “Sequence-Encoded Self-Assembly ofMultiple-Nanocomponent Arrays by 2D DNA Scaffold-ing,” Nano Letters, vol. 4, 2005, pp. 2399–2402.

[15] W. Y. Choi et al., “Compact Nano-Electro-MechanicalNon-Volatile Memory (NEMory) for 3D Integration,”IEEE Int’l. Electron Devices Mtg. Tech. Digest, 2007, pp.603–6.

BIOGRAPHIESM. SAIF ISLAM ([email protected]) received his Ph.D.degree in electrical engineering from the University ofCalifornia at Los Angeles (UCLA) in 2001 and worked forHewlett-Packard Laboratories, Gazillion Bits Inc., and JDSUniphase Corporation before joining UC Davis in 2004,where he is an associate professor. His work coversultra-fast optoelectronic devices, molecular electronics;and synthesis, device applications and CMOS integrationof semiconductor nanostructures for imaging, sensing,computing, and energy conversion. He has authored/co-authored more than 150 scientific papers, edited 13books and conference proceedings, and holds 29 patentsas an inventor/co-inventor. He has chaired/co-chaired13 conferences, and received NSF Faculty Early CareerAward in 2006, Outstanding Junior Faculty Award in2006, IEEE Professor of the Year in 2005 and 2009, andthe UC Davis Academic Senate Distinguished TeachingAward in 2010.

LOGEESWARAN VJ received his B.Eng. (mechanical) andM.Eng. (mechanical) degrees from the National Universityof Singapore (NUS) in 1997 and 2001 respectively. From1997 to 2004 he was a research engineer in the Depart-ment of Mechanical Engineering at NUS, working on theresearch and development of micromachined inertial sen-sors and resonators. He obtained his Ph.D. degree in elec-trical and computer engineering (ECE) at UC Davis in 2010and is currently a postdoctoral researcher in the ECEDepartment at UC Davis working on nanowire solar cells.His research interests are in photovoltaics, electromagneticmetamaterials, nanowires, MEMS, and NEMS.

Resolute effort on

overcoming the

roadblock to

wide-scale and cost-

effective integration

of functional nano-

materials in

communication

systems, and

understanding their

environmental

impact, are likely to

keep the scientific

community occupied

in the next decade.



nanoscale materials and devices for future communication

Documents