119 Nanowires 4. Nanowires - UFAMhome.ufam.edu.br/berti/nanomateriais/Nanowires.pdf · 119...

119

Nanowires4. Nanowires

Mildred S. Dresselhaus, Yu-Ming Lin, Oded Rabin, Marcie R. Black, Jing Kong, Gene Dresselhaus

This chapter provides an overview of recentresearch on inorganic nanowires, particu-larly metallic and semiconducting nanowires.Nanowires are one-dimensional, anisotropicstructures, small in diameter, and large in surface-to-volume ratio. Thus, their physical properties aredifferent than those of structures of different scaleand dimensionality. While the study of nanowiresis particularly challenging, scientists have madeimmense progress in both developing syntheticmethodologies for the fabrication of nanowires,and developing instrumentation for their charac-terization. The chapter is divided into three mainsections: Sect. 4.1 the synthesis, Sect. 4.2 the char-acterization and physical properties, and Sect. 4.3the applications of nanowires. Yet, the reader willdiscover many links that make these aspects ofnanoscience intimately interdepent.

4.1 Synthesis ............................................. 1214.1.1 Template-Assisted Synthesis .......... 121

4.1.2 VLS Methodfor Nanowire Synthesis ................. 124

4.1.3 Other Synthesis Methods ............... 1264.1.4 Hierarchical Arrangement

and Superstructures of Nanowires .. 128

4.2 Characterization and Physical Propertiesof Nanowires ....................................... 1304.2.1 Structural Characterization ............ 1304.2.2 Mechanical Properties................... 1354.2.3 Transport Properties ..................... 1364.2.4 Optical Properties ......................... 147

4.3 Applications ......................................... 1524.3.1 Electrical Applications ................... 1524.3.2 Thermoelectric Applications ........... 1544.3.3 Optical Applications ...................... 1544.3.4 Chemical and Biochemical

Sensing Devices ........................... 1574.3.5 Magnetic Applications................... 158

4.4 Concluding Remarks ............................. 159

References .................................................. 159

Nanowires are attracting much interest from those seek-ing to apply nanotechnology and (especially) thoseinvestigating nanoscience. Nanowires, unlike otherlow-dimensional systems, have two quantum-confineddirections but one unconfined direction available forelectrical conduction. This allows nanowires to be usedin applications where electrical conduction, rather thantunneling transport, is required. Because of their uniquedensity of electronic states, in the limit of small diam-eters nanowires are expected to exhibit significantlydifferent optical, electrical and magnetic properties totheir bulk 3-D crystalline counterparts. Increased sur-face area, very high density of electronic states andjoint density of states near the energies of their vanHove singularities, enhanced exciton binding energy,diameter-dependent bandgap, and increased surfacescattering for electrons and phonons are just some of

the ways in which nanowires differ from their corre-sponding bulk materials. Yet the sizes of nanowires aretypically large enough (> 1 nm in the quantum-confineddirection) to result in local crystal structures that areclosely related to their parent materials, allowing the-oretical predictions about their properties to be madebased on knowledge of their bulk properties.

Not only do nanowires exhibit many properties thatare similar to, and others that are distinctly differentfrom, those of their bulk counterparts, nanowires alsohave the advantage from an applications standpoint inthat some of the materials parameters critical for certainproperties can be independently controlled in nanowiresbut not in their bulk counterparts. Certain propertiescan also be enhanced nonlinearly in small-diameternanowires, by exploiting the singular aspects of the 1-Delectronic density of states.

PartA

4

120 Part A Nanostructures, Micro-/Nanofabrication and Materials

Table 4.1 Selected syntheses of nanowires by material

Material Growth Technique Reference

ABO4-type Templatea [4.2]

Ag DNA-template, redox [4.3]

Template, pulsed ECDb [4.4]

Au Template, ECDb [4.5, 6]

Bi Stress-induced [4.7]

Template, vapor-phase [4.8]

Template, ECDb [4.9–11]

Template,pressure injection

[4.12–14]

BiSb Pulsed ECDb [4.15]

Bi2Te3 Template, dc ECDb [4.16]

CdS Liquid-phase (surfactant),recrystallization

[4.17]

Template, ac ECDb [4.18, 19]

CdSe Liquid-phase (surfactant),redox

[4.20]

Template, ac ECDb [4.21, 22]

Cu Vapor deposition [4.23]

Template, ECDb [4.24]

Fe Template, ECDc [4.25, 26]

Shadow deposition [4.27]

GaN Template, CVDc [4.28, 29]

VLSd [4.30, 31]

GaAs Template,liquid/vapor OMCVDe

[4.32]

a Template synthesisb Electrochemical deposition (ECD)c Chemical vapor deposition (CVD)d Vapor–liquid–solid (VLS) growthe Organometallic chemical vapor deposition (OMCVD)f Liquid phase synthesisg Self assembly of nanocrystals (in liquid phase)

Material Growth Technique Reference

Ge High-T,high-P liquid-phase, redox

[4.33]

VLSd [4.34]

Oxide-assisted [4.35]

InAs VLSd [4.36]

MgO VLSd [4.37]

Mo Step decoration,ECDb+ redox

[4.38]

Ni Template, ECDb [4.11, 39, 40]

Pb Liquid-phasef [4.41]

PbSe Liquid phase [4.42]

Self assemblyof nanocrystalsg

[4.43]

Pd Step decoration, ECDb [4.44]

Se Liquid-phase,recrystallization

[4.45]

Template,pressure injection

[4.46]

Si VLSd [4.47]

Laser-ablation VLSd [4.48]

Oxide-assisted [4.49]

Low-T VLSd [4.50]

W Vapor transport [4.51]

Zn Template, vapor-phase [4.52]

Template, ECDb [4.53]

ZnO VLSd [4.54]

Template, ECDb [4.53, 55]

Furthermore, nanowires have been shown to pro-vide a promising framework for applying the bottom-upapproach [4.1] to the design of nanostructures fornanoscience investigations and for potential nanotech-nology applications.

Driven by (1) these new research and developmentopportunities, (2) the smaller and smaller length scalesnow being used in the semiconductor, optoelectronicsand magnetics industries, and (3) the dramatic develop-ment of the biotechnology industry where the action isalso at the nanoscale, the nanowire research field hasdeveloped with exceptional speed in the last few years.Therefore, a review of the current status of nanowire re-

search is of significant broad interest at the present time.It is the aim of this review to focus on nanowire proper-ties that differ from those of their parent crystalline bulkmaterials, with an eye toward possible applications thatmight emerge from the unique properties of nanowiresand from future discoveries in this field.

For quick reference, examples of typical nanowiresthat have been synthesized and studied are listed in Ta-ble 4.1. Also of use to the reader are review articles thatfocus on a comparison between nanowire and nanotubeproperties [4.56] and the many reviews that have beenwritten about carbon nanotubes [4.57–59], which can beconsidered as a model one-dimensional system.

PartA

4

Nanowires 4.1 Synthesis 121

4.1 Synthesis

In this section we survey the most common syntheticapproaches that have successfully afforded high-qualitynanowires of a large variety of materials (Table 4.1).In Sect. 4.1.1, we discuss methods which make useof various templates with nanochannels to confine thenanowire growth in two dimensions. In Sect. 4.1.2, wepresent the synthesis of nanowires by the vapor–liquid–solid mechanism and its many variations. In Sect. 4.1.3,examples of other synthetic methods of general ap-plicability are presented. The last part of this section(Sect. 4.1.4) features several approaches that have beendeveloped to organize nanowires into simple architec-tures.

4.1.1 Template-Assisted Synthesis

The template-assisted synthesis of nanowires is a con-ceptually simple and intuitive way to fabricate nano-structures [4.62–64]. These templates contain verysmall cylindrical pores or voids within the host ma-terial, and the empty spaces are filled with the chosenmaterial, which adopts the pore morphology, to formnanowires. In this section, we describe the templatesfirst, and then describe strategies for filling the tem-plates to make nanowires.

Template SynthesisIn template-assisted synthesis of nanostructures, thechemical stability and mechanical properties of the tem-plate, as well as the diameter, uniformity and densityof the pores are important characteristics to consider.Templates frequently used for nanowire synthesis in-clude anodic alumina (Al2O3), nanochannel glass, iontrack-etched polymers and mica films.

Porous anodic alumina templates are produced byanodizing pure Al films in selected acids [4.65–67].Under carefully chosen anodization conditions, the re-sulting oxide film possesses a regular hexagonal array

1 µm100 nm

a) b)

Fig. 4.1 (a) SEM images of the topsurfaces of porous anodic aluminatemplates anodized with an averagepore diameter of 44 nm (after [4.60]).(b) SEM image of the particle track-etched polycarbonate membrane, witha pore diameter of 1 μm (after [4.61])

of parallel and nearly cylindrical channels, as shownin Fig. 4.1a. The self-organization of the pore structurein an anodic alumina template involves two coupledprocesses: pore formation with uniform diameters andpore ordering. The pores form with uniform diam-eters because of a delicate balance between electricfield-enhanced diffusion which determines the growthrate of the alumina, and dissolution of the aluminainto the acidic electrolyte [4.68]. The pores are be-lieved to self-order because of mechanical stress atthe aluminum–alumina interface due to expansion dur-ing the anodization. This stress produces a repulsiveforce between the pores, causing them to arrange ina hexagonal lattice [4.69]. Depending on the anodiza-tion conditions, the pore diameter can be systematicallyvaried from ≤ 10 up to 200 nm with a pore density inthe range of 109 –1011 pores/cm2 [4.13, 25, 65, 66]. Ithas been shown by many groups that the pore size dis-tribution and the pore ordering of the anodic aluminatemplates can be significantly improved by a two-step anodization technique [4.60, 70, 71], where thealuminum oxide layer is dissolved after the first an-odization in an acidic solution followed by a secondanodization under the same conditions.

Another type of porous template commonly usedfor nanowire synthesis is the template type fabricatedby chemically etching particle tracks originating fromion bombardment [4.72], such as track-etched polycar-bonate membranes (Fig. 4.1b) [4.73, 74], and also micafilms [4.39].

Other porous materials can be used as host tem-plates for nanowire growth, as discussed by Ozin [4.62].Nanochannel glass (NCG), for example, contains a reg-ular hexagonal array of capillaries similar to the porestructure in anodic alumina with a packing densityas high as 3 × 1010 pores/cm2 [4.63]. Porous Vycorglass that contains an interconnected network of poresless than 10 nm was also employed for the early

PartA

4.1


study of nanostructures [4.75]. Mesoporous molecularsieves [4.76], termed MCM-41, possess hexagonally-packed pores with very small channel diameters whichcan be varied between 2 and 10 nm. Conducting organicfilaments have been fabricated in the nanochannels ofMCM-41 [4.77]. Recently, the DNA molecule has alsobeen used as a template for growing nanometer-sizedwires [4.3].

Diblock copolymers, polymers that consist of twochain segments different properties, have also been uti-lized as templates for nanowire growth. When twocomponents are immiscible in each other, phase seg-regation occurs, and depending on their volume ratio,spheres, cylinders and lamellae may self-assemble. Toform self-assembled arrays of nanopores, copolymerscomposed of polystyrene and polymethylmethacrylate[P(S-b-MMA)] [4.79] were used. By applying an elec-tric field while the copolymer was heated above theglass transition temperature of the two constituent poly-mers, the self-assembled cylinders of PMMA could be

20 30 40 50 602θ (deg)

Intensity (arb. units)

(012)

(202)

(024)

a)

b)

c) (110)

Fig. 4.2a–c XRD patterns of bismuth/anodic aluminananocomposites with average bismuth wire diameters of(a) 40 nm, (b) 52 nm, and (c) 95 nm [4.78]. The Miller in-dices corresponding to the lattice planes of bulk Bi areindicated above the individual peaks. The majority of theBi nanowires are oriented along the [101̄1] and [011̄2]directions for dW ≥ 60 nm and dW ≤ 50 nm, respectively(after [4.13,78]). The existence of more than one dominantorientation in the 52 nm Bi nanowires is attributed to thetransitional behavior of intermediate-diameter nanowiresas the preferential growth orientation is shifted from [101̄1]to [011̄2] with decreasing dW

aligned with their main axis perpendicular to the film.Selective removal of the PMMA component affordedthe preparation of 14 nm diameter ordered pore arrayswith a packing density of 1.9 × 1011 cm−3.

Nanowire Template-Assisted Growthby Pressure Injection

The pressure injection technique is often employed forfabricating highly crystalline nanowires from a low-melting point material and when using porous templateswith robust mechanical strength. In the high-pressureinjection method, the nanowires are formed by pressure-injecting the desired material in liquid form into theevacuated pores of the template. Due to the heatingand pressurization processes, the templates used forthe pressure injection method must be chemically sta-ble and be able to maintain their structural integrityat high temperatures and at high pressures. Anodicaluminum oxide films and nanochannel glass are twotypical materials used as templates in conjunction withthe pressure injection filling technique. Metal nanowires(Bi, In, Sn, and Al) and semiconductor nanowires (Se,Te, GaSb, and Bi2Te3) have been fabricated in anodicaluminum oxide templates using this method [4.12, 46,78].

The pressure P required to overcome the surfacetension for the liquid material to fill the pores witha diameter dW is determined by the Washburn equa-tion [4.80]

dW = −4γ cos θ/P , (4.1)

where γ is the surface tension of the liquid, and θ

is the contact angle between the liquid and the tem-plate. To reduce the required pressure and to maximizethe filling factor, some surfactants are used to de-crease the surface tension and the contact angle. Forexample, the introduction of Cu into the Bi melt canfacilitate filling the pores in the anodic alumina tem-plate with liquid Bi and can increase the number ofnanowires that are formed [4.13]. However, some ofthe surfactants might cause contamination problemsand should therefore be avoided. Nanowires producedby the pressure injection technique usually possesshigh crystallinity and a preferred crystal orientationalong the wire axis. For example, Fig. 4.2 shows thex-ray diffraction (XRD) patterns of Bi nanowire ar-rays of three different wire diameters with an injectionpressure of ≈ 5000 psi [4.78], showing that the major(> 80%) crystal orientation of the wire axes in the 95and 40 nm diameter Bi nanowire arrays are, respec-tively, normal to the (202) and (012) lattice planes,

PartA

4.1


which are denoted by [101̄1] and [011̄2] when usinga hexagonal unit cell, suggesting a wire diameter-dependent crystal growth direction. On the other hand,30 nm Bi nanowires produced using a much higherpressure of > 20 000 psi show a different crystal ori-entation of (001) along the wire axis [4.14], indicatingthat the preferred crystal orientation may also dependon the applied pressure, with the most dense packingdirection along the wire axis for the highest appliedpressure.

Electrochemical DepositionThe electrochemical deposition technique has attractedincreasing attention as a versatile method for fabricatingnanowires in templates. Traditionally, electrochemistryhas been used to grow thin films on conducting surfaces.Since electrochemical growth is usually controllablein the direction normal to the substrate surface, thismethod can be readily extended to fabricate 1-D or0-D nanostructures, if the deposition is confined withinthe pores of an appropriate template. In the electro-chemical methods, a thin conducting metal film isfirst coated on one side of the porous membrane toserve as the cathode for electroplating. The length ofthe deposited nanowires can be controlled by vary-ing the duration of the electroplating process. Thismethod has been used to synthesize a wide variety ofnanowires, such as metals (Bi [4.9, 74]; Co [4.81, 82];Fe [4.25, 83]; Cu [4.73, 84]; Ni [4.39, 81]; Ag [4.85];Au [4.5]); conducting polymers [4.9, 61]; superconduc-tors (Pb [4.86]); semiconductors (CdS [4.19]); and evensuperlattice nanowires with A/B constituents (such asCu/Co [4.73, 84]) have been synthesized electrochemi-cally (Table 4.1).

In the electrochemical deposition process, the cho-sen template has to be chemically stable in theelectrolyte during the electrolysis process. Cracksand defects in the templates are detrimental to thenanowire growth, since the deposition processes pri-

a) b)

1 µm100 nm

Fig. 4.3 (a) SEM image of a Bi2Te3

nanowire array in cross section show-ing a relatively high pore filling factor.(b) SEM image of a Bi2Te3 nanowirearray composite along the wire axis(after [4.16])

marily occur in the more accessible cracks, leavingmost of the nanopores unfilled. Particle track-etchedmica films or polymer membranes are typical tem-plates used in simple DC electrolysis. To use anodicaluminum oxide films in the DC electrochemical de-position, the insulating barrier layer which separatesthe pores from the bottom aluminum substrate hasto be removed, and a metal film is then evaporatedonto the back of the template membrane [4.87]. Com-pound nanowire arrays, such as Bi2Te3, have beenfabricated in alumina templates with a high filling fac-tor using the DC electrochemical deposition [4.16].Figure 4.3a,b, respectively, shows the top view andthe axial cross-sectional SEM images of a Bi2Te3nanowire array [4.16]. The light areas are associ-ated with Bi2Te3 nanowires, the dark regions denoteempty pores, and the surrounding gray matrix isalumina.

Surfactants are also used with electrochemical de-position when necessary. For example, when usingtemplates derived from PMMA/PS diblock copolymers,a methanol surfactant is used to facilitate pore fill-ing [4.79], thereby achieving a ≈ 100% filling factor.

It is also possible to employ an ac electrodeposi-tion method in anodic alumina templates without theremoval of the barrier layer, by utilizing the rectifyingproperties of the oxide barrier. In ac electrochemicaldeposition, although the applied voltage is sinusoidaland symmetric, the current is greater during the ca-thodic half-cycles, making deposition dominant overthe stripping, which occurs in the subsequent anodichalf-cycles. Since no rectification occurs at defect sites,the deposition and stripping rates are equal, and nomaterial is deposited. Hence, the difficulties associatedwith cracks are avoided. In this fashion, metals, such asCo [4.82] and Fe [4.25, 83], and semiconductors, suchas CdS [4.19], have been deposited into the pores of an-odic aluminum oxide templates without removing thebarrier layer.

PartA

4.1


a) b)

0.1µm 15 nm

Fig. 4.4 (a) TEM image of a single Co(10 nm)/Cu(10 nm)multilayered nanowire. (b) A selected region of the sampleat high magnification (after [4.84])

In contrast to nanowires synthesized by the pressureinjection method, nanowires fabricated by the elec-trochemical process are usually polycrystalline, withno preferred crystal orientations, as observed by XRDstudies. However, some exceptions exist. For exam-ple, polycrystalline CdS nanowires, fabricated by anac electrodeposition method in anodic alumina tem-plates [4.19], possibly have a preferred wire growthorientation along the c-axis. In addition, Xu et al. haveprepared a number of single-crystal II–VI semiconduc-tor nanowires, including CdS, CdSe and CdTe, by DCelectrochemical deposition in anodic alumina templateswith a nonaqueous electrolyte [4.18, 22]. Furthermore,single-crystal Pb nanowires were formed by pulse elec-trodeposition under overpotential conditions, but nospecific crystal orientation along the wire axis was ob-served [4.86]. The use of pulse currents is believed tobe advantageous for the growth of crystalline wires be-cause the metal ions in the solution can be regeneratedbetween the electrical pulses and therefore uniform de-position conditions can be produced for each depositionpulse. Similarly, single-crystal Ag nanowires were fab-ricated by pulsed electrodeposition [4.4].

One advantage of the electrochemical depositiontechnique is the possibility of fabricating multilayeredstructures within nanowires. By varying the cathodicpotentials in the electrolyte, which contains two dif-ferent kinds of ions, different metal layers can becontrollably deposited. Co/Cu multilayered nanowireshave been synthesized in this way [4.73,84]. Figure 4.4shows TEM images of a single Co/Cu nanowire whichis about 40 nm in diameter [4.84]. The light bands rep-resent Co-rich regions and the dark bands representCu-rich layers. This electrodeposition method provides

a low-cost approach to preparing multilayered 1-Dnanostructures.

Vapor DepositionVapor deposition of nanowires includes physical va-por deposition (PVD) [4.8], chemical vapor deposition(CVD) [4.29], and metallo-organic chemical vapor de-position (MOCVD) [4.32]. Like electrochemical depo-sition, vapor deposition is usually capable of preparingsmaller-diameter (≤ 20 nm) nanowires than pressure in-jection methods, since it does not rely on the highpressure and the surface tension involved to insert thematerial into the pores.

In the physical vapor deposition technique, the ma-terial to be filled is first heated to produce a vapor, whichis then introduced through the pores of the template andcooled to solidify. Using a specially designed experi-mental setup [4.8], nearly single-crystal Bi nanowiresin anodic aluminum templates with pore diameters assmall as 7 nm have been synthesized, and these Binanowires were found to possess a preferred crystalgrowth orientation along the wire axis, similar to theBi nanowires prepared by pressure injection [4.8, 13].

Compound materials that result from two reactinggases have also be prepared by the chemical vapor de-position (CVD) technique. For example, single-crystalGaN nanowires have been synthesized in anodic alu-mina templates through a gas reaction of Ga2O vaporwith a flowing ammonia atmosphere [4.28, 29]. A dif-ferent liquid/gas phase approach has been used toprepare polycrystalline GaAs and InAs nanowires ina nanochannel glass array [4.32]. In this method, thenanochannels are filled with one liquid precursor (suchas Me3Ga or Et3In) via a capillary effect and thenanowires are formed within the template by reactionsbetween the liquid precursor and the other gas reactant(such as AsH3).

4.1.2 VLS Method for Nanowire Synthesis

Some of the recent successful syntheses of semiconduc-tor nanowires are based on the so-called vapor–liquid–solid (VLS) mechanism of anisotropic crystal growth.This mechanism was first proposed for the growth ofsingle crystal silicon whiskers 100 nm to hundreds ofmicrometer in diameter [4.88]. The proposed growthmechanism (Fig. 4.5) involves the absorption of sourcematerial from the gas phase into a liquid droplet of cat-alyst (a molten particle of gold on a silicon substratein the original work [4.88]). Upon supersaturation ofthe liquid alloy, a nucleation event generates a solid

PartA

4.1


precipitate of the source material. This seed serves asa preferred site for further deposition of material atthe interface of the liquid droplet, promoting the elon-gation of the seed into a nanowire or a whisker, andsuppressing further nucleation events on the same cata-lyst. Since the liquid droplet catalyzes the incorporationof material from the gas source to the growing crystal,the deposit grows anisotropically as a whisker whosediameter is dictated by the diameter of the liquid alloydroplet. The nanowires thus obtained are of high purity,except for the end containing the solidified catalyst asan alloy particle (Figs. 4.5 and 4.6a). Real-time obser-vations of the alloying, nucleation, and elongation stepsin the growth of germanium nanowires from gold nan-oclusters by the VLS method were recorded by in situTEM [4.89].

Reduction of the average wire diameter to thenanometer scale requires the generation of nanosizedcatalyst droplets. However, due to the balance betweenthe liquid-vapor surface free energy and the free energyof condensation, the size of a liquid droplet, in equilib-rium with its vapor, is usually limited to the micrometerrange. This obstacle has been overcome in recent yearsby several new methodologies:

1. Advances in the synthesis of metal nanoclustershave made monodispersed nanoparticles commer-cially available. These can be dispersed on a solidsubstrate in high dilution so that when the temper-ature is raised above the melting point, the liquidclusters do not aggregate [4.47].

2. Alternatively, metal islands of nanoscale sizes canself-form when a strained thin layer is grown orheat-treated on a nonepitaxial substrate [4.34].

10 nm

Si

SiOx

[111]

a) b) c)

100 nm

Fig. 4.6 (a) TEM images of Si nanowires produced after laser-ablating a Si0.9Fe0.1 target. The dark spheres witha slightly larger diameter than the wires are solidified catalyst clusters (after [4.48]). (b) Diffraction contrast TEM im-age of a Si nanowire. The crystalline Si core appears darker than the amorphous oxide surface layer. The inset showsthe convergent beam electron diffraction pattern recorded perpendicular to the wire axis, confirming the nanowire crys-tallinity (after [4.48]). (c) STEM image of Si/Si1−xGex superlattice nanowires in the bright field mode. The scale bar is500 nm (after [4.90])

Si/Metalcatalyst (liquid)

Si vapor

NanowireGrowth

Si vaporSi/Metalcatalyst (liquid)

Si (solid)

Fig. 4.5 Schematic diagram illustrating the growth of siliconnanowires by the VLS mechanism

3. Laser-assisted catalytic VLS growth is a methodused to generate nanowires under nonequilibriumconditions. Using laser ablation of a target con-taining both the catalyst and the source materials,a plasma is generated from which catalyst nan-oclusters nucleate as the plasma cools down. Singlecrystal nanowires grow as long as the particle re-mains liquid [4.48].

4. Interestingly, by optimizing the material propertiesof the catalyst-nanowire system, conditions can beachieved for which nanocrystals nucleate in a li-quid catalyst pool supersaturated with the nanowirematerial, migrate to the surface due to a large sur-face tension, and continue growing as nanowiresperpendicular to the liquid surface [4.50]. In thiscase, supersaturated nanodroplets are sustained onthe outer end of the nanowire due to the low solubil-ity of the nanowire material in the liquid [4.91].

A wide variety of elemental, binary and com-pound semiconductor nanowires has been synthesized

PartA

4.1


via the VLS method, and relatively good controlover the nanowire diameter and diameter distributionhas been achieved. Researchers are currently focus-ing their attention on the controlled variation of thematerials properties along the nanowire axis. In thiscontext, researchers have modified the VLS synthe-sis apparatus to generate compositionally-modulatednanowires. GaAs/GaP-modulated nanowires have beensynthesized by alternately ablating targets of the corre-sponding materials in the presence of gold nanoparti-cles [4.92]. p-Si/n-Si nanowires were grown by chem-ical vapor deposition from alternating gaseous mixturescontaining the appropriate dopant [4.92]. Si/Si1−xGexnanowires were grown by combining silicon froma gaseous source with germanium from a periodicallyablated target (Fig. 4.6c) [4.90]. NiSi-Si nanowires havebeen successfully synthesized which directly incorpo-rate a nanowire metal contact into active nanowiredevices [4.93]. Finally, using an ultrahigh vacuumchamber and molecular beams, InAs/InP nanowireswith atomically sharp interfaces were obtained [4.94].These compositionally-modulated nanowires are ex-pected to exhibit exciting electronic, photonic, andthermoelectric properties.

Interestingly, silicon and germanium nanowiresgrown by the VLS method consist of a crystalline corecoated with a relatively thick amorphous oxide layer(2–3 nm) (Fig. 4.6b). These layers are too thick to bethe result of ambient oxidation, and it has been shownthat these oxides play an important role in the nanowiregrowth process [4.49, 95]. Silicon oxides were foundto serve as a special and highly selective catalyst thatsignificantly enhances the yield of Si nanowires with-out the need for metal catalyst particles [4.49, 95, 96].A similar yield enhancement was also found in thesynthesis of Ge nanowires from the laser ablation ofGe powder mixed with GeO2 [4.35]. The Si and Genanowires produced from these metal-free targets gen-erally grow along the [112] crystal direction [4.97],and have the benefit that no catalyst clusters are foundon either ends of the nanowires. Based on these ob-servations and other TEM studies [4.35, 95, 97], anoxide-enhanced nanowire growth mechanism differ-ent from the classical VLS mechanism was proposed,where no metal catalyst is required during the laserablation-assisted synthesis [4.95]. It is postulated thatthe nanowire growth is dependent on the presence ofSiO (or GeO) vapor, which decomposes in the nanowiretip region into both Si (or Ge), which is incorporatedinto the crystalline phase, and SiO2 (or GeO2), whichcontributes to the outer coating. The initial nucleation

100 nm

Fig. 4.7 TEM image showing the two major morpholo-gies of Si nanowires prepared by the oxide-assisted growthmethod (after [4.95]). Notice the absence of metal par-ticles when compared to Fig. 4.6a. The arrow points at anoxide-linked chain of Si nanoparticles

events generate oxide-coated spherical nanocrystals.The [112] crystal faces have the fastest growth rate,and therefore the nanocrystals soon begin elongatingalong this direction to form one-dimensional structures.The SimO or GemO (m > 1) layer on the nanowiretips may be in or at temperatures near their moltenstates, catalyzing the incorporation of gas moleculesin a directional fashion [4.97]. Besides nanowires withsmooth walls, a second morphology of chains of unori-ented nanocrystals linked by oxide necks is frequentlyobserved (indicated by an arrow in Fig. 4.7). In addi-tion, it was found by STM studies that about 1% ofthe wires consist of a regular array of two alternatingsegments, 10 and 5 nm in length, respectively [4.98].The segments, whose junctions form an angle of 30◦,are probably a result of alternating growth along differ-ent crystallographic orientations [4.98]. Branched andhyperbranched Si nanowire structures have also beensynthesized by Whang et al. [4.99].

4.1.3 Other Synthesis Methods

In this section we review several other general pro-cedures available for the synthesis of a variety ofnanowires. We focus on bottom-up approaches, whichafford many kinds of nanowires in large numbers, and

PartA

4.1


do not require highly sophisticated equipment (suchas scanning microscopy or lithography-based methods),and exclude cases for which the nanowires are not self-sustained (such as in the case of atomic rows on thesurface of crystals).

A solution-phase synthesis of nanowires with con-trollable diameters has been demonstrated [4.45, 101],without the use of templates, catalysts, or surfac-tants. Instead, Gates et al. make use of the anisotropyof the crystal structure of trigonal selenium and tel-lurium, which can be viewed as rows of 1-D helicalatomic chains. Their approach is based on the masstransfer of atoms during an aging step from a highfree-energy solid phase (e.g., amorphous selenium)to a seed (e.g., trigonal selenium nanocrystal) whichgrows preferentially along one crystallographic axis.The lateral dimension of the seed, which dictates thediameter of the nanowire, can be controlled by the tem-perature of the nucleation step. Furthermore, Se/Tealloy nanowires were synthesized by this method, andAg2Se compound nanowires were obtained by treatingselenium nanowires with AgNO3 [4.102–104]. In a sep-arate work, tellurium nanowires were transformed intoBi2Te3 nanowires by their reaction with BiPh3 [4.105].

More often, however, the use of surfactants isnecessary to promote the anisotropic 1-D growth ofnanocrystals. Solution phase synthetic routes have beenoptimized to produce monodispersed quantum dots,(zero-dimensional isotropic nanocrystals) [4.106]. Sur-factants are necessary in this case to stabilize theinterfaces of the nanoparticles and to retard oxida-tion and aggregation processes. Detailed studies onthe effect of growth conditions revealed that they canbe manipulated to induce a directional growth of thenanocrystals, usually generating nanorods (aspect ratioof ≈ 10), and in favorable cases, nanowires with highaspect ratios. Heath and LeGoues synthesized germa-nium nanowires by reducing a mixture of GeCl4 andphenyl-GeCl3 at high temperature and high pressure.The phenyl ligand was essential for the formation ofhigh aspect ratio nanowires [4.33]. In growing CdSenanorods [4.20], Alivisatos et al. used a mixture of twosurfactants, whose concentration ratio influenced thestructure of the nanocrystal. It is believed that differentsurfactants have different affinities, and different ab-sorption rates, for the different crystal faces of CdSe,thereby regulating the growth rates of these faces. Inthe liquid phase synthesis of Bi nanowires, the ad-ditive NaN(SiMe3)2 induces the growth of nanowiresoriented along the [110] crystal direction from smallbismuth seed clusters, while water solely retarded the

growth along the [001] direction, inducing the growthof hexagonal-plate particles [4.105]. A coordinatingalkyl-diamine solvent was used to grow polycrystallinePbSe nanowires at low temperatures [4.42]. Here,the surfactant-induced directional growth is believedto occur through to the formation of organometalliccomplexes in which the bidentate ligand assumes theequatorial positions, thus hindering the ions from ap-proaching each other in this plane. Additionally, thealkyl-diamine molecules coat the external surface ofthe wire, preventing lateral growth. The aspect ratioof the wires increased as the temperature was low-ered in the range 10 ◦C < T < 117 ◦C. Ethylenediaminewas used to grow CdS nanowires and tetrapods bya solvothermal recrystallization process starting withCdS nanocrystals or amorphous particles [4.17]. Whilethe coordinating solvent was crucial for the nanowiregrowth, its role in the shape and phase control was notclarified.

Graphite

Elektrodepositionof MoO2 nanowires

Reduction to Mo0 in H2

at 500 °C for ≈ 1h

Cast poly(styrene) film

Lift-off of embeddedMo0 nanowires

Poly(styrene)

Fig. 4.8 Schematic of the electrodeposition step edge dec-oration of HOPG (highly oriented pyrolytic graphite) forthe synthesis of molybdenum nanowires (after [4.38, 100])

PartA

4.1


Stress-induced crystalline bismuth nanowires havebeen grown from sputtered films of layers of Bi andCrN. The nanowires presumably grow from defects andcleavage fractures in the film, and are up to several mil-limeters in length with diameters ranging from 30 to200 nm [4.7]. While the exploration of this techniquehas only begun, stress-induced unidirectional growthshould be applicable to a variety of composite films.

Selective electrodeposition along the step edges inhighly oriented pyrolytic graphite (HOPG) was usedto obtain MoO2 nanowires as shown in Fig. 4.8. Thesite-selectivity was achieved by applying a low over-potential to the electrochemical cell in which the HOPGserved as cathode, thus minimizing the nucleationevents on less favorable sites (plateaux). While thesenanowires cannot be removed from the substrate, theycan be reduced to metallic molybdenum nanowires,which can then be released as free-standing nanowires.Other metallic nanowires were also obtained by thismethod [4.38, 100]. In contrast to the template syn-thesis approaches described above, in this method thesubstrate only defines the position and orientation ofthe nanowire, not its diameter. In this context, othersurface morphologies, such as self-assembled groovesin etched crystal planes, have been used to generatenanowire arrays via gas-phase shadow deposition (forexample: Fe nanowires on (110)NaCl [4.27]). The crosssection of artificially prepared superlattice structureshas also been used for site-selective deposition of par-allel and closely spaced nanowires [4.109]. Nanowiresprepared on the above-mentioned substrates wouldhave semicircular, rectangular, or other unconventionalcross-sectional shapes.

4.1.4 Hierarchical Arrangementand Superstructures of Nanowires

Ordering nanowires into useful structures is anotherchallenge that needs to be addressed in order to har-

a) b) c) d)

Fig. 4.9a–d SEM images of (a) sixfold- (b) fourfold- and (c) twofold-symmetry nanobrushes made of an In2O3 core andZnO nanowire brushes (after [4.107]), and of (d) ZnO nanonails (after [4.108]). The scale bars are (a) 1 μm, (b) 500 nm,(c) 500 nm, and (d) 200 nm

ness the full potential of nanowires for applications. Wewill first review examples of nanowires with nontrivialstructures, and then proceed to describe methods usedto create assemblies of nanowires of a predeterminedstructure.

We mentioned in Sect. 4.1.2 that the preparation ofnanowires with a graded composition or with a super-lattice structure along their main axis was demonstratedby controlling the gas phase chemistry as a function oftime during the growth of the nanowires by the VLSmethod. Control of the composition along the axial di-mension was also demonstrated by a template-assistedmethod, for example by the consecutive electrochem-ical deposition of different metals in the pores of analumina template [4.110]. Alternatively, the composi-tion can be varied along the radial dimension of thenanowire, for example by first growing a nanowire bythe VLS method and then switching the synthesis con-ditions to grow a different material on the surface ofthe nanowire by CVD. This technique was demon-strated for the synthesis of Si/Ge and Ge/Si coaxial(or core–shell) nanowires [4.111], and it was shownthat the outer shell can be formed epitaxially on theinner core by a thermal annealing process. Han et al.demonstrated the versatility of MgO nanowire arraysgrown by the VLS method as templates for the PLDdeposition of oxide coatings to yield MgO/YBCO,MgO/LCMO, MgO/PZT and MgO/Fe3O4 core/shellnanowires, all exhibiting epitaxial growth of the shellon the MgO core [4.37]. A different approach wasadopted by Wang et al. who generated a mixture ofcoaxial and biaxial SiC-SiOx nanowires by the catalyst-free high-temperature reaction of amorphous silica anda carbon/graphite mixture [4.112].

A different category of nontrivial nanowires is thatof nanowires with a nonlinear structure, resulting frommultiple one-dimensional growth steps. Members ofthis category are tetrapods, which were mentioned inthe context of the liquid phase synthesis (Sect. 4.1.3).

PartA

4.1


Subphase

Hydrophobicnanorods

100 nm

Fig. 4.10 A TEM image of a smectic phase of a BaCrO4

nanorod film (left inset) achieved by the Langmuir–Blodgett technique, as depicted by the illustration (af-ter [4.113])

In this process, a tetrahedral quantum dot core is firstgrown, and then the conditions are modified to in-duce one-dimensional growth of a nanowire from eachone of the facets of the tetrahedron. A similar pro-cess produced high-symmetry In2O3/ZnO hierarchicalnanostructures. From a mixture of heat-treated In2O3,ZnO, and graphite powders, faceted In2O3 nanowireswere first obtained, on which oriented shorter ZnOnanowires were crystallized [4.107]. Brushlike struc-tures were obtained as a mixture of 11 structures ofdifferent symmetries. For example, two, four, or sixrows of ZnO nanorods could be found on different corenanowires, depending on the crystallographic orienta-tion of the main axis of the core nanowire, as shownin Fig. 4.9. Comblike structures made entirely of ZnOwere also reported [4.54].

Controlling the position of a nanowire in the growthprocess is important for preparing devices or test struc-tures containing nanowires, especially when it involvesa large array of nanowires. Post-synthesis methods toalign and position nanowires include microfluidic chan-nels [4.114], Langmuir–Blodgett assemblies [4.113],and electric field-assisted assembly [4.115]. The firstmethod involves the orientation of the nanowires bythe liquid flow direction when a nanowire solution isinjected into a microfluidic channel assembly and bythe interaction of the nanowires with the side wallsof the channel. The second method involves the align-

ment of nanowires at a liquid–gas or liquid–liquidinterface by the application of compressive forces onthe interface (Fig. 4.10). The aligned nanowire filmscan then be transferred onto a substrate and lithog-raphy methods can be used to define interconnects.This allows the nanowires to be organized with a con-trolled alignment and spacing over large areas. Usingthis method, centimeter-scale arrays containing thou-sands of single silicon nanowire field-effect transistorswith high performance could be assembled to makelarge-scale nanowire circuits and devices [4.99, 116].The third technique is based on dielectrophoretic forcesthat pull polarizable nanowires toward regions of highfield strength. The nanowires align between two iso-lated electrodes which are capacitatively coupled toa pair of buried electrodes biased with an AC volt-age. Once a nanowire shorts the electrodes, the electricfield is eliminated, preventing more nanowires fromdepositing. The above techniques have been success-fully used to prepare electronic circuitry and opticaldevices out of nanowires (Sects. 4.3.1 and 4.3.3). Al-ternatively, alignment and positioning of the nanowirescan be specified and controlled during their growth bythe proper design of the synthesis method. For exam-ple, ZnO nanowires prepared by the VLS method weregrown into an array in which both their position onthe substrate and their growth direction and orienta-tion were controlled [4.54]. The nanowire growth regionwas defined by patterning the gold film, which servesas a catalyst for the ZnO nanowire growth, employingsoft-lithography, e-beam lithography, or photolithogra-phy. The orientation of the nanowires was achieved byselecting a substrate with a lattice structure matchingthat of the nanowire material to facilitate the epitaxialgrowth. These conditions result in an array of nanowireposts at predetermined positions, all vertically alignedwith the same crystal growth orientation (Fig. 4.11).Similar rational GaN nanowire arrays have been syn-thesized epitaxially on (100)LiAlO2 and (111)MgOsingle-crystal substrates. In addition, control over thecrystallographic growth directions of nanowires wasachieved by lattice-matching to different substrates.For example, GaN nanowires on (100)LiAlO2 sub-strates grow oriented along the [110] direction, whereas(111)MgO substrates result in the growth of GaNnanowires with an [001] orientation, due to the dif-ferent lattice-matching constraints [4.117]. A similarstructure could be obtained by the template-mediatedelectrochemical synthesis of nanowires (Sect. 4.1.1),particularly if anodic alumina with its parallel and or-dered channels is used. The control over the location

PartA

4.1


a) b)

100 nm1 µm1 µm

c)

Fig. 4.11a–c SEM images of ZnO nanowire arrays grown on a sapphire substrate, where (a) shows patterned growth,(b) shows a higher resolution image of the parallel alignment of the nanowires, and (c) shows the faceted side-walls andthe hexagonal cross section of the nanowires. For nanowire growth, the sapphire substrates were coated with a 1.0–3.5 nmthick patterned layer of Au as the catalyst, using a TEM grid as the shadow mask. These nanowires have been used fornanowire laser applications (after [4.122])

of the nucleation of nanowires in the electrochemi-cal deposition is determined by the pore positions andthe back-electrode geometry. The pore positions canbe precisely controlled by imprint lithography [4.118].

By growing the template on a patterned conductivesubstrate that serves as a back-electrode [4.119–121]different materials can be deposited in the pores at dif-ferent regions of the template.

4.2 Characterization and Physical Properties of Nanowires

In this section we review the structure and prop-erties of nanowires and their interrelationship. Thediscovery and investigation of nanostructures werespurred on by advances in various characterization andmicroscopy techniques that enabled material charac-terization to take place at smaller and smaller lengthscales, reaching length scales down to individual atoms.For applications, characterizing the structural propertiesof nanowires is especially important, so that a repro-ducible relationship between their desired functionalityand their geometrical and structural characteristics canbe established. Due to the enhanced surface-to-volumeratio in nanowires, their properties may depend sen-sitively on their surface conditions and geometricalconfigurations. Even nanowires made of the samematerial may possess dissimilar properties due to dif-ferences in their crystal phase, crystalline size, surfaceconditions, and aspect ratios, which depend on the syn-thesis methods and conditions used in their preparation.

4.2.1 Structural Characterization

Structural and geometric factors play an importantrole in determining the various attributes of nanowires,such as their electrical, optical and magnetic proper-ties. Therefore, various novel tools have been developedand employed to obtain this important structural in-

formation at the nanoscale. At the micrometer scale,optical techniques are extensively used for imagingstructural features. Since the sizes of nanowires are usu-ally comparable to or, in most cases, much smallerthan the wavelength of visible light, traditional opti-cal microscopy techniques are usually limited whencharacterizing the morphology and surface features ofnanowires. Therefore, electron microscopy techniquesplay a more dominant role at the nanoscale. Sinceelectrons interact more strongly than photons, electronmicroscopy is particularly sensitive relative to x-rays forthe analysis of tiny samples.

In this section we review and give examples of howscanning electron microscopy, transmission electronmicroscopy, scanning probe spectroscopies, and diffrac-tion techniques are used to characterize the structuresof nanowires. To provide the necessary basis for de-veloping reliable structure–property relations, multiplecharacterization tools are applied to the same samples.

Scanning Electron MicroscopySEM usually produces images down to length scalesof ≈ 10 nm and provides valuable information regard-ing the structural arrangement, spatial distribution, wiredensity, and geometrical features of the nanowires. Theexamples of SEM micrographs shown in Figs. 4.1 and4.3 indicate that structural features at the 10 nm to

PartA

4.2

Nanowires 4.2 Characterization and Physical Properties of Nanowires 131

10 μm length scales can be probed, providing infor-mation on the size, size distribution, shapes, spatialdistributions, density, nanowire alignment, filling fac-tors, granularity, etc.. As another example, Fig. 4.11ashows an SEM image of ZnO nanowire arrays grownon a sapphire substrate [4.122], which provides evi-dence for the nonuniform spatial distribution of thenanowires on the substrate, which was attained by pat-terning the catalyst film to define high-density growthregions and nanowire-free regions. Figure 4.11b, show-ing a higher magnification of the same system, indicatesthat these ZnO nanowires grow perpendicular to thesubstrate, are well-aligned with approximately equalwire lengths, and have wire diameters in the range20 ≤ dW ≤ 150 nm. The SEM micrograph in Fig. 4.11cprovides further information about the surface ofthe nanowires, showing it to be well-faceted, form-ing a hexagonal cross section, indicative of nanowiregrowth along the 〈0001〉 direction. Both the uniformityof the nanowire size, their alignment perpendicular tothe substrate, and their uniform growth direction, assuggested by the SEM data, are linked to the good epi-taxial interface between the (0001) plane of the ZnOnanowire and the (110) plane of the sapphire substrate.(The crystal structures of ZnO and sapphire are es-sentially incommensurate, with the exception that thea-axis of ZnO and the c-axis of sapphire are related al-most exactly by a factor of 4, with a mismatch of lessthan 0.08% at room temperature [4.122].) The well-faceted nature of these nanowires has important impli-cations for their lasing action (Sect. 4.3.2). Figure 4.12shows an SEM image of GaN nanowires synthesized bya laser-assisted catalytic growth method [4.30], indicat-ing a random spatial orientation of the nanowire axesand a wide diameter distribution for these nanowires, incontrast to the ZnO wires in Fig. 4.11 and to arrays ofwell-aligned nanowires prepared by template-assistedgrowth (Fig. 4.3).

Transmission Electron MicroscopyTEM and high-resolution transmission electron mi-croscopy (HRTEM) are powerful imaging tools forstudying nanowires at the atomic scale, and they usu-ally provide more detailed geometrical features than areseen in SEM images. TEM studies also yield informa-tion regarding the crystal structure, crystal quality, grainsize, and crystal orientation of the nanowire axis. Whenoperating in the diffraction mode, selected area electrondiffraction (SAED) patterns can be made to determinethe crystal structures of nanowires. As an example, theTEM images in Fig. 4.13 show four different morpholo-

gies for Si nanowires prepared by the laser ablationof a Si target [4.123]: (a) spring-shaped; (b) fishbone-shaped (indicated by solid arrow) and frogs egg-shaped(indicated by the hollow arrow), (c) pearl-shaped, while(d) shows the poly-sites of nanowire nucleation. Thecrystal quality of nanowires is revealed from high-resolution TEM images with atomic resolution, alongwith selected area electron diffraction (SAED) pat-terns. For example, Fig. 4.14 shows a TEM image ofone of the GaN nanowires from Fig. 4.12, indicatingsingle crystallinity and showing (100) lattice planes,thus indicating the growth direction of the nanowire.This information is supplemented by the correspondingelectron diffraction pattern in the upper right. A morecomprehensive review of the application of TEM forgrowth orientation indexing and crystal defect charac-terization in nanowires is available elsewhere [4.124].

The high resolution of the TEM also permits thesurface structures of the nanowires to be studied. Inmany cases, the nanowires are sheathed with a nativeoxide layer, or an amorphous oxide layer that forms dur-ing the growth process. This can be seen in Fig. 4.6bfor silicon nanowires and in Fig. 4.15 for germaniumnanowires [4.35], showing a mass–thickness contrastTEM image and a selected-area electron diffraction pat-tern of a Ge nanowire. The main TEM image shows thatthese Ge nanowires possess an amorphous GeO2 sheathwith a crystalline Ge core that is oriented in the [211]direction.

2 µmµm

Fig. 4.12 SEM image of GaN nanowires in a mat arrange-ment synthesized by laser-assisted catalytic growth. Thenanowires have diameters and lengths on the order of10 nm and 10 μm, respectively (after [4.30])

PartA

4.2


a) b)

100 nm 100 nm

50 nm 300 nmc) d)

Fig. 4.13a–d TEM morpholo-gies of four special forms of Sinanowires synthesized by the laserablation of a Si powder target.(a) A spring-shaped Si nanowire;(b) fishbone-shaped (indicated bya solid arrow) and frogs egg-shaped(indicated by a hollow arrow) Sinanowires; and (c) pearl-shapednanowires, while (d) shows polysitesfor the nucleation of silicon nanowires(indicated by arrows) (after [4.123])

Fig. 4.14 Lattice-resolved high-resolution TEM image ofone GaN nanowire (left) showing that (100) lattice planesare visible perpendicular to the wire axis. The electrondiffraction pattern (top right) was recorded along the [001]zone axis. A lattice-resolved TEM image (lower right)highlights the continuity of the lattice up to the nanowireedge, where a thin native oxide layer is found. The direc-tions of various crystallographic planes are indicated in thelower right figure (after [4.30]) �

Dynamical processes of the surface layer ofnanowires can be studied in-situ using an environmen-tal TEM chamber, which allows TEM observations tobe made while different gases are introduced or as thesample is heat-treated at various temperatures, as il-lustrated in Fig. 4.16. The figure shows high-resolutionTEM images of a Bi nanowire with an oxide coat-ing and the effect of a dynamic oxide removal processcarried out within the environmental chamber of theTEM [4.125]. The amorphous bismuth-oxide layercoating the nanowire (Fig. 4.16a) is removed by expo-sure to hydrogen gas within the environmental chamberof the TEM, as indicated in Fig. 4.16b.

55 nmnm

100100

010010

100100010010

11–0

11–0

01–0

01–0

1–0000

1–0000

1–1010

1–1010

By coupling the powerful imaging capabilities ofTEM with other characterization tools, such as anelectron energy loss spectrometer (EELS) or an en-ergy dispersive x-ray spectrometer (EDS) within the

PartA

4.2


10100 nmnm

111111022022

[211

][2

11]

(100)(100)

(111)(111)

GeGe

GeO

GeO

2

[0[01–

1]1]

1–1111

(01(011 –

)

(01(011 –

)

Fig. 4.15 A mass–thickness contrast TEM image of a Genanowire taken along the [01̄1] zone axis and a selected-area electron diffraction pattern (upper left inset) (af-ter [4.35]). The Ge nanowires were synthesized by laserablation of a mixture of Ge and GeO2 powder. The core ofthe Ge nanowire is crystalline, while the surface GeO2 isamorphous

TEM instrument, additional properties of the nanowirescan be probed with high spatial resolution. Withthe EELS technique, the energy and momentum ofthe incident and scattered electrons are measuredin an inelastic electron scattering process to pro-vide information on the energy and momentum ofthe excitations in the nanowire sample. Figure 4.17shows the dependence on nanowire diameter of theelectron energy loss spectra of Bi nanowires. Thespectra were taken from the center of the nanowire,and the shift in the energy of the peak position(Fig. 4.17) indicates the effect of the nanowire diam-eter on the plasmon frequency in the nanowires. Theresults show that there are changes in the electronicstructure of the Bi nanowires as the wire diameter de-creases [4.126]. Such changes in electronic structureas a function of nanowire diameter are also observedin their transport (Sect. 4.2.2) and optical (Sect. 4.2.3)properties, and are related to quantum confinementeffects.

EDS measures the energy and intensity distributionof x-rays generated by the impact of the electron beamon the surface of the sample. The elemental composi-tion within the probed area can be determined to a highdegree of precision. The technique was particularly use-ful for the compositional characterization of superlattice

Oxidelayer

Before After H2 annealing at 130 °C for 6 h

Fig. 4.16 High-resolution transmission electron microscope(HRTEM) image of a Bi nanowire (left) before and (right) afterannealing in hydrogen gas at 130 ◦C for 6 h within the environmen-tal chamber of the HRTEM instrument to remove the oxide surfacelayer (after [4.125])

nanowires [4.90] and core–shell nanowires [4.111](Sect. 4.1.2).

Scanning Tunneling ProbesSeveral scanning probe techniques, such as scan-ning tunneling microscopy (STM) [4.127], electric

Energy loss (eV)

8 10 12 14 16 18 20 22

35 nm

60 nm

90 nm


Fig. 4.17 Electron energy loss spectra (EELS) taken fromthe centers of bismuth nanowires with diameters of 35, 60and 90 nm. The shift in the volume plasmon peaks is dueto the effect of wire diameter on the electronic structure(after [4.126])

PartA

4.2


field gradient microscopy (EFM) [4.13], magnetic fieldmicroscopy (MFM) [4.40], and scanning thermal mi-croscopy (SThM) [4.128], combined with atomic forcemicroscopy (AFM), have been employed to study thestructural, electronic, magnetic, and thermal properties

a)

b)

c) d)

Fig. 4.18a–d STM height images, obtained in the constantcurrent mode, of MoSe chains deposited on an Au(111)substrate. (a) A single chain image, and (b) a MoSe wirebundle. (c,d) Images of MoSe wire fragments containingfive and three unit cells, respectively (after [4.127]). Thescale bars are all 1 nm

0 μm 1.25 μm 2.5 μm

0 μm

1.25 μm

2.5 μm0 μm 1.25 μm 2.5 μm

a) b)

Fig. 4.19 (a) Topographic image ofa highly-ordered porous aluminatemplate with a period of 100 nmfilled with 35 nm diameter nickelnanowires. (b) The correspondingMFM (magnetic force microscope)image of the nanomagnet array,showing that the pillars are magne-tized alternately up (white) and down(black) (after [4.40])

of nanowires. A scanning tunneling microscope can beemployed to reveal both topographical structural infor-mation, such as that illustrated in Fig. 4.18, as well asinformation on the local electronic density of states ofa nanowire, when used in the STS (scanning tunnelingspectroscopy) mode. Figure 4.18 shows STM height im-ages (taken in the constant current STM mode) of MoSemolecular wires deposited from a methanol or acetoni-trile solution of Li2Mo6Se6 onto Au substrates. TheSTM image of a single MoSe wire (Fig. 4.18a) exhibitsa 0.45 nm lattice repeat distance in a MoSe molecu-lar wire. When both STM and STS measurements aremade on the same sample, the electronic and structuralproperties can be correlated, as for example in the jointSTM/STS studies on Si nanowires [4.98], showing al-ternating segments of a single nanowire identified withgrowth along the [110] and [112] directions, and differ-ent I–V characteristics measured for the [110] segmentsas compared with the [112] segments.

Magnetic field microscopy (MFM) has been em-ployed to study magnetic polarization of magneticnanowires embedded in an insulating template, such asan anodic alumina template. For example, Fig. 4.19ashows the topographic image of an anodic alumina tem-plate filled with Ni nanowires, and Fig. 4.19b demon-strates the corresponding magnetic polarization of eachnanowire in the template. This micrograph shows thata magnetic field microscopy probe can distinguish be-tween spin-up and spin-down nanowires in the nanowirearray, thereby providing a method for measuring inter-wire magnetic dipolar interactions [4.40].

X-ray AnalysisOther characterization techniques that are commonlyused to study the crystal structures and chemical com-positions of nanowires include x-ray diffraction andx-ray energy dispersion analysis (EDAX). The peak po-

PartA

4.2


sitions in the x-ray diffraction pattern can be used todetermined the chemical composition and the crystalphase structure of the nanowires. For example, Fig. 4.2shows that Bi nanowires have the same crystal structureand lattice constants as bulk bismuth. Both the x-raydiffraction pattern (XRD) for an array of aligned Binanowires (Fig. 4.2) and the SAED pattern for individ-ual Bi nanowires [4.13] suggest that the nanowires havea common axis of nanowire alignment.

As another example of an XRD pattern for an ar-ray of aligned nanowires, Fig. 4.20 shows the x-raydiffraction pattern of the ZnO nanowires that are dis-played in Fig. 4.11. Only (00�) diffraction peaks areobserved for these aligned ZnO nanowires, indicat-ing that their preferred growth direction is (001) alongthe wire axis. Similarly, XRD was used to confirmthe different growth directions of GaN nanowire ar-ray grown epitaxially on (100)LiAlO2 and (111)MgOsubstartes [4.117].

EDAX has been used to determine the chem-ical compositions and stoichiometries of compoundnanowires or impurity contents in nanowires. However,the results from EDAX analysis should be interpretedcarefully to avoid systematic errors.

4.2.2 Mechanical Properties

Thermal StabilityDue to the large surface area-to-volume ratio innanowires and other nanoparticles, the thermal stabil-ity of nanowires is anticipated to differ significantlyfrom that of the bulk material. Theoretical studies ofmaterials in confined geometries show that the melt-ing point of the material is reduced in nanostructures,as is the latent heat of fusion, and that large hysteresiscan be observed in melting–freezing cycles. These phe-nomena have been studied experimentally in three typesof nanowire systems: porous matrices impregnated witha plurality of nanowires, individual nanowires sheathedby a thin coating, and individual nanowires.

The melting freezing of matrix-supported nanowirescan be studied by differential scanning calorimetry(DSC), since large volumes of samples can thus be pro-duced. Huber et al. investigated the melting of indiumin porous silica glasses with mean pore diameters rang-ing from 6 to 141 nm [4.129]. The melting point ofthe pore-confined indium shows a linear dependence oninverse pore diameter, with a maximum melting pointdepression of 50 K. They also recorded a 6 K differencein the melting temperature and the freezing tempera-ture of 12.8 nm diameter indium. The melting profile

30 40 50 602θ (deg)

70 80 90


(002)

(004)

Al2O3 (110) Al2O3 (220)

Fig. 4.20 X-ray diffraction pattern of aligned ZnO nano-wires (Fig. 4.11) grown on a sapphire substrate. Only [00�]diffraction peaks are observed for the nanowires, owing totheir well-oriented growth orientation. Strong diffractionpeaks for the sapphire substrate are found (after [4.122])

of the pore-confined indium in these samples is broaderin temperature than for bulk indium, as expected for theheterogeneity in the pore diameter and in the indiumcrystal size aspect ratio within the samples.

Sheathed nanowires provide an opportunity tostudy the melting and recrystallization of individualnanowires. The shell layer surrounding the nanowireprovides confinement to keep the liquid phase withinthe inner cylindrical volume. However, the shell–nanowire surface interaction should be taken intoaccount when analyzing the phase transition thermody-namics and kinetics. Yang et al. produced germaniumnanowires coated with a thin (1–5 nm) graphite sheath,by pyrolysis of organic molecules over VLS-grownnanowires, and followed the melting and recrystalliza-tion of the germanium by variable temperature TEMimaging [4.130]. The melting of the nanowires was fol-lowed by the disappearance of the electronic diffractionpattern. It was found that the nanowires began melt-ing from their ends, with the melting front advancingtowards the center of the nanowire as the tempera-ture was increased. During the cool-down part of thecycle, the recrystallization of the nanowire occurred in-stantaneously following significant supercooling. Theauthors report both the largest melting point suppres-sion recorded thus far for germanium (≈ 300 ◦C), anda large melting–recrystallization hysteresis of up to≈ 300 ◦C. Similarly, carbon nanotubes have been filledwith various low-temperature metals [4.131]. A nanoth-

PartA

4.2


ermometer has been demonstrated using a 10 nm liquidgallium filled-carbon nanotube, showing an expansioncoefficient that is linear in temperature and identical tothe bulk value [4.132].

A different behavior was observed in free-standingcopper nanowires [4.134]. In this system, there is lit-tle interaction between the nanowire surface and thesurroundings, and the nanowire is not confined in itsdiameter, as in the case of the sheathed nanowires. Ther-mal treatment of the free-standing nanowires leads totheir fragmentation into a linear array of metal spheres.Thinner nanowires were more vulnerable than thickernanowires to the thermal treatment, showing constric-tions and segmentation at lower temperatures. Analysisof the temperature response of the nanowires indi-cates that the nanowire segmentation is a result ofthe Rayleigh instability, starting with oscillatory per-turbations of the nanowire diameter, leading to longcylindrical segments, that become more separated andmore spherical at higher temperatures. These observa-tions indicate that annealing and melting are dominatedby the surface diffusion of atoms on the entire surfaceof the nanowire (versus tip-initiated melting).

4.2.3 Transport Properties

The study of electrical transport properties of nanowiresis important for nanowire characterization, electronicdevice applications, and the investigation of unusualtransport phenomena arising from one-dimensionalquantum effects. Important factors that determine thetransport properties of nanowires include the wire diam-

1

Counts

8×105

6×105

4×105

2×105

02 3 4

Conductance (2e2/h)

Au –AuGoldwires

Nano-contact

a) b)

Fig. 4.21 (a) Schematic representation of the last stages of the con-tact breakage process (after [4.133]). (b) Histogram of conductancevalues built with 18 000 gold contact breakage experiments in air atroom temperature, showing conductance peaks at integral values ofG0. In this experiment the gold electrodes approach and separate at89 000 Å/s (after [4.133])

eter, (important for both classical and quantum sizeeffects), material composition, surface conditions, crys-tal quality, and the crystallographic orientation alongthe wire axis for materials with anisotropic material pa-rameters, such as the effective mass tensor, the Fermisurface, or the carrier mobility.

Electronic transport phenomena in low-dimensionalsystems can be roughly divided into two categories:ballistic transport and diffusive transport. Ballistictransport phenomena occur when the electrons cantravel across the nanowire without any scattering. Inthis case, the conduction is mainly determined by thecontacts between the nanowire and the external circuit,and the conductance is quantized into an integral num-ber of universal conductance units G0 = 2e2/h [4.135,136]. Ballistic transport phenomena are usually ob-served in very short quantum wires, such as thoseproduced using mechanically controlled break junctions(MCBJ) [4.137, 138] where the electron mean free pathis much longer than the wire length and the conductionis a pure quantum phenomenon. To observe ballistictransport, the thermal energy must also obey the relationkBT � ε j −ε j−1, where ε j −ε j−1 is the energy separa-tion between subband levels j and j −1. On the otherhand, for nanowires with lengths much larger than thecarrier mean free path, the electrons (or holes) undergonumerous scattering events when they travel along thewire. In this case, the transport is in the diffusive regime,and the conduction is dominated by carrier scatteringwithin the wires, due to phonons (lattice vibrations),boundary scattering, lattice and other structural defects,and impurity atoms.

Conductance Quantizationin Metallic Nanowires

The ballistic transport of 1-D systems has been ex-tensively studied since the discovery of quantizedconductance in 1-D systems in 1988 [4.135, 136]. Thephenomena of conductance quantization occur when thediameter of the nanowire is comparable to the electronFermi wavelength, which is on the order of 0.5 nm formost metals [4.139]. Most conductance quantization ex-periments up to the present were performed by bringingtogether and separating two metal electrodes. As thetwo metal electrodes are slowly separated, a nanocon-tact is formed before it breaks completely (Fig. 4.21a),and conductance in integral multiple values of G0is observed through these nanocontacts. Figure 4.21bshows the conductance histogram built with 18 000contact breakage curves between two gold electrodesat room temperature [4.133], with the electrode sep-

PartA

4.2


aration up to ≈ 1.8 nm. The conductance quantizationbehavior is found to be independent of the contact ma-terial, and has been observed in various metals, suchas Au [4.133], Ag, Na, Cu [4.140], and Hg [4.141].For semimetals such as Bi, conductance quantizationhas also been observed for electrode separations aslong as 100 nm at 4 K because of the long Fermiwavelength (≈ 26 nm) [4.139], indicating that the con-ductance quantization may be due to the existence ofwell-defined quantum states localized at a constrictioninstead of resulting from the atom rearrangement as theelectrodes separate. Since conductance quantization isonly observed in breaking contacts, or for very narrowand very short nanowires, most nanowires of practi-cal interest (possessing lengths of several micrometer)lie in the diffusive transport regime, where the carrierscattering is significant and should be considered.

I–V Characterizationof Semiconducting Nanowires

The electronic transport behavior of nanowires may becategorized based on the relative magnitudes of threelength scales: carrier mean free path �W, the de Brogliewavelength of electrons λe, and the wire diameter dW.For wire diameters much larger than the carrier meanfree path (dW �W), the nanowires exhibit transportproperties similar to bulk materials, which are indepen-dent of the wire diameter, since the scattering due tothe wire boundary is negligible compared to other scat-tering mechanisms. For wire diameters comparable toor smaller than the carrier mean free path (dW ≈ �Wor dW < �W), but still much larger than the de Brogliewavelength of the electrons (dW λe), the transport innanowires is in the classical finite size regime, wherethe band structure of the nanowire is still similar to thatof bulk, while the scattering events at the wire bound-ary alter their transport behavior. For wire diameterscomparable to the electronic wavelength dW ≈ λe, theelectronic density of states is altered dramatically andquantum subbands are formed due to the quantum con-finement effect at the wire boundary. In this regime, thetransport properties are further influenced by the changein the band structure. Therefore, transport properties fornanowires in the classical finite size and quantum sizeregimes are highly diameter-dependent.

Researchers have investigated the transport prop-erties of various semiconducting nanowires and havedemonstrated their potential for diverse electronic de-vices, such as for p-n diodes [4.142, 143], field effecttransistors [4.142], memory cells, and switches [4.144](Sect. 4.3.1). So far, the nanowires studied in this

context have usually been made from conventionalsemiconducting materials, such as group IV and III–Vcompound semiconductors, via the VLS growth method(Sect. 4.1.2), and their nanowire properties have beencompared to their well-established bulk properties. In-terestingly, the physical principles for describing bulksemiconductor devices also hold for devices based onthese semiconducting nanowires with wire diameters oftens of nanometers. For example, Fig. 4.22 shows thecurrent–voltage (I–V ) behavior of a 4-by-1 crossed p-Si/n-GaN junction array at room temperature [4.142].The long horizontal wire in the figure is a p-Si nanowire(10–25 nm in diameter) and the four short vertical wiresare n-GaN nanowires (10–30 nm in diameter). Each ofthe four nanoscale cross points independently formsa p-n junction with current rectification behavior, asshown by the I–V curves in Fig. 4.22, and the junc-tion behavior (for example the turn-on voltage) canbe controlled by varying the oxide coating on thesenanowires [4.142].

Huang et al. have demonstrated nanowire junc-tion diodes with a high turn-on voltage (≈ 5 V) byincreasing the oxide thickness at the junctions. Thehigh turn-on voltage enables the use of the junction in

Current (nA)

– 4Bias (V)

2000

1500

1000

500

0– 2 0 2 4

Fig. 4.22 I–V behavior for a 4(p) by 1(n) crossed p-Si/n-GaN junction array shown in the inset. The four curvesrepresent the I–V response for each of the four junc-tions, showing similar current rectifying characteristics ineach case. The length scale bar between the two mid-dle junctions is 2 μm (after [4.142]). The p-Si and n-GaNnanowires are 10–25 and 10–30 nm in diameter, respec-tively

PartA

4.2


Current (nA)

Bias (V)

400

200

0

– 200

– 400–1 – 0.5 0 0.5 1

Vg (V):0

1

2

3

S D

G

102

100

10–2

0 1 2 3 4 5

Current (nA)

Gate (V)

Fig. 4.23 Gate-dependent I–V characteristics of a crossednanowire field-effect transistor (FET). The n-GaN nanowireis used as the nanogate, with the gate voltage indicated(0, 1, 2, and 3 V). The inset shows the current versusVgate for a nanowire gate (lower curve) and for a globalback-gate (top curve) when the bias voltage is set to 1 V(after [4.142])

a nanoscale FET, as shown in Fig. 4.23 [4.142] whereI–V data for a p-Si nanowire are presented, for whichthe n-GaN nanowire with a thick oxide coating is usedas a nanogate. By varying the nanogate voltage, theconductance of the p-Si nanowire can be changed bymore than a factor of 105 (lower curve in the inset),whereas the conductance changes by only a factor of 10when a global back-gate is used (top curve in the insetof Fig. 4.23). This behavior may be due to the thin gatedielectric between the crossed nanowires and the bettercontrol of the local carrier density through a nanogate.Based on the gate-dependent I–V data from these p-Sinanowires, it is found that the mobility of the holes inthe p-Si nanowires may be higher than that for bulk p-Si,although further investigation is required for completeunderstanding.

Because of the enhanced surface-to-volume ra-tios of nanowires, their transport behavior may bemodified by changing their surface conditions. Forexample, researchers have found that by coating n-InP nanowires with a layer of redox molecules, suchas cobalt phthalocyanine, the conductance of the InPnanowires may change by orders of magnitude upon al-tering the charge state of the redox molecules to providebistable nanoscale switches [4.144]. The resistance (or

conductance) of some nanowires (such as Pd nanowires)is also very sensitive to the presence of certain gases(e.g., H2) [4.145,146], and this property may be utilizedfor sensor applications to provide improved sensitivitycompared to conventional sensors based on bulk mater-ial (Sect. 4.3.4).

Although it remains unclear how the size effectmay influence the transport properties and device per-formance of semiconducting nanowires, many of thelarger diameter semiconducting nanowires are expectedto be described by classical physics, since their quan-tization energies �

2/(2med2W) are usually smaller than

the thermal energy kBT . By comparing the quantizationenergy with the thermal energy, the critical wire diam-eter below which quantum confinement effects becomesignificant is estimated to be 1 nm for Si nanowires atroom temperature, which is much smaller than the sizesof many of the semiconducting nanowires that havebeen investigated so far. By using material systems withmuch smaller effective carrier masses me (such as bis-muth), the critical diameter for which such quantumeffects can be observed is increased, thereby facilitat-ing the study of quantum confinement effects. It is forthis reason that the bismuth nanowire system has beenstudied so extensively. Furthermore, since the crystalstructure and lattice constants of bismuth nanowires arethe same as for 3-D crystalline bismuth, it is possi-ble to carry out detailed model calculations to guideand to interpret transport and optical experiments onbismuth nanowires. For these reasons, bismuth can beconsidered a model system for studying 1-D effects innanowires.

Temperature-Dependent ResistanceMeasurements

Although nanowires with electronic properties similarto their bulk counterparts are promising for construct-ing nanodevices based on well-established knowledgeof their bulk counterparts, it is expected that quan-tum size effects in nanowires will likely be utilizedto generate new phenomena absent in bulk materials,and thus provide enhanced performance and novel func-tionality for certain applications. In this context, thetransport properties of bismuth (Bi) nanowires havebeen extensively studied, both theoretically [4.147]and experimentally [4.8, 10, 78, 148–150] because oftheir promise for enhanced thermoelectric performance.Transport studies of ferromagnetic nanowire arrays,such as Ni or Fe, have also received much attentionbecause of their potential for high-density magneticstorage applications [4.151].

PartA

4.2


The very small electron effective mass componentsand the long carrier mean free paths in Bi facilitate thestudy of quantum size effects in the transport propertiesof nanowires. Quantum size effects are expected to be-come significant in bismuth nanowires with diameterssmaller than 50 nm [4.147], and the fabrication of crys-talline nanowires with this diameter range is relativelyeasy.

Figure 4.24a shows the T dependence of the re-sistance R(T ) for Bi nanowires (7 ≤ dW < 200 nm)synthesized by vapor deposition and pressure injec-tion [4.8], illustrating the quantum effects in theirtemperature-dependent resistance. In Fig. 4.24a, theR(T ) behavior of Bi nanowires is dramatically dif-ferent from that of bulk Bi, and is highly sensi-tive to the wire diameter. Interestingly, the R(T )

1 1

T (K)

2

1.5

1

0.5

00 100 10 100

0 50 100 150 200 250 300

4

3

2

1

0

2.5

2

1.5

1

T (K)

T (K) T (K)

a) R (T )/R (300 K) b) R (T )/R (290 K)

c) R (T )/R (300 K)

48 nm

36 nm

28 nm

7 nm

70 nm

200 nm

Bulk Bi

200 nm

400 nm

1 μm

2 μm

36 nm

70 nm (polycrystalline)

70 nm

Fig. 4.24 (a) Measured temperature dependence of the resistance R(T ) normalized to the room temperature (300 K)resistance for bismuth nanowire arrays of various wire diameters dW (after [4.8]). (b) R(T )/R(290 K) for bismuth wiresof larger dW and lower mobility (after [4.10]). (c) Calculated R(T )/R(300 K) of 36 and 70 nm bismuth nanowires. Thedashed curve refers to a 70 nm polycrystalline wire with increased boundary scattering (after [4.78])

curves in Fig. 4.24a show a nonmonotonic trend forlarge-diameter (70 and 200 nm) nanowires, althoughR(T ) becomes monotonic with T for small-diameter(≤ 48 nm) nanowires. This dramatic change in the be-havior of R(T ) as a function of dW is attributed toa unique semimetal–semiconductor transition phenom-ena in Bi [4.78], induced by quantum size effects. Biis a semimetal in bulk form, in which the T -pointvalence band overlaps with the L-point conductionband by 38 meV at 77 K. As the wire diameter de-creases, the lowest conduction subband increases inenergy and the highest valence subband decreases inenergy. Model calculations predict that the band over-lap should vanish in Bi nanowires (with their wireaxes along the trigonal direction) at a wire diameter≈ 50 nm [4.147].

PartA

4.2


The resistance of Bi nanowires is determined bytwo competing factors: the carrier density that in-creases with T , and the carrier mobility that decreaseswith T . The nonmonotonic R(T ) for large-diameterBi nanowires is due to a smaller carrier concentrationvariation at low temperature (≤ 100 K) in semimet-als, so that the electrical resistance is dominated bythe mobility factor in this temperature range. Basedon the semi-classical transport model and the estab-lished band structure of Bi nanowires, the calculatedR(T )/R(300 K) for 36 and 70 nm Bi nanowires isshown by the solid curves in Fig. 4.24c to illus-trate different R(T ) trends for semiconducting andsemimetallic nanowires, respectively [4.78]. The curvesin Fig. 4.24c exhibit trends consistent with experimentalresults. The condition for the semimetal–semiconductortransition in Bi nanowires can be experimentally de-termined, as shown by the measured resistance ratioR(10 K)/R(100 K) of Bi nanowires as a function ofwire diameter [4.152] in Fig. 4.25. The maximum inthe resistance ratio R(10 K)/R(100 K) at dW ≈ 48 nmindicates the wire diameter for the transition of Binanowires from a semimetallic phase to a semiconduct-ing phase. The semimetal–semiconductor transition andthe semiconducting phase in Bi nanowires are examplesof new transport phenomena resulting from low dimen-sionality that are absent in the bulk 3-D phase, andthese phenomena further increase the possible benefitsfrom the properties of nanowires for desired applica-tions (Sect. 4.3.2).

0 50 100 150 200

1.6

1.2

0.8

0.4

0

R (10 K)/R (100 K)

Wire diameter (nm)

≈ 48 nm

Fig. 4.25 Measured resistance ratio R(10 K)/R(100 K) ofBi nanowire array as a function of diameter. The peakindicates the transition from a semimetallic phase toa semiconducting phase as the wire diameter decreases(after [4.153])

10 100

1

0.1

T (K)1

R (T )/R (300 K)

Zn(4 nm)/Vycor glass

Zn(9 nm)/Al2O3

Zn(15 nm)/SiO2

T 1

Fig. 4.26 Temperature dependence of the resistance ofZn nanowires synthesized by vapor deposition in vari-ous porous templates (after [4.52]). The data are given aspoints, the full lines are fits to a T 1 law for 15 nm diameterZn nanowires in an SiO2 template, denoted by Zn/SiO2.Fits to a combined T 1 and T−1/2 law were made for thesmaller nanowire diameter composite samples denoted byZn (9 nm)/Al2O3 and Zn 4 nm/Vycor glass

It should be noted that good crystal quality isessential for observing the quantum size effect innanowires, as shown by the R(T ) plots in Fig. 4.24a.For example, Fig. 4.24b shows the normalized R(T )measurements of Bi nanowires with larger diameters(200 nm–2 μm) prepared by electrochemical deposi-tion [4.10], and these nanowires possess monotonicR(T ) behaviors, quite different from those of the corre-sponding nanowire diameters shown in Fig. 4.24a. Theabsence of the resistance maximum in Fig. 4.24b is dueto the lower crystalline quality for nanowires preparedby electrochemical deposition, which tends to producepolycrystalline nanowires with a much lower carriermobility. This monotonic R(T ) for semimetallic Binanowires with a higher defect level is also confirmedby theoretical calculations, as shown by the dashedcurve in Fig. 4.24c for 70 nm wires with increased grainboundary scattering [4.154].

The theoretical model developed for Bi nanowiresnot only provides good agreement with experimental re-sults, but it also plays an essential role in understandingthe influence of the quantum size effect, the bound-ary scattering, and the crystal quality on their electricalproperties. While the electronic density of states maybe significantly altered due to quantum confinement

PartA

4.2


effects, various scattering mechanisms related to thetransport properties of nanowires can be accountedfor by Matthiessen’s rule. Furthermore, the transportmodel has also been generalized to predict the trans-port properties of Te-doped Bi nanowires [4.78], Sbnanowires [4.155], and BiSb alloy nanowires [4.156],and good agreement between experiment and theory hasalso been obtained for these cases.

For nanowires with diameters comparable to thephase-breaking length, their transport properties maybe further influenced by localization effects. It hasbeen predicted that in disordered systems, the extendedelectronic wavefunctions become localized near defectsites, resulting in the trapping of carriers and giving riseto different transport behavior. Localization effects arealso expected to be more pronounced as the dimension-ality and sample size are reduced. Localization effectson the transport properties of nanowire systems havebeen studied on Bi nanowires [4.158] and, more re-cently, on Zn nanowires [4.52]. Figure 4.26 shows themeasured R(T )/R(300 K) of Zn nanowires fabricatedby vapor deposition in porous silica or alumina [4.52].While 15 nm Zn nanowires exhibit an R(T ) behav-ior with a T 1 dependence as expected for a metallicwire, the R(T ) of 9 and 4 nm Zn nanowires exhibits

0.16

0.14

0.12

0.1

0.08

0.06

0.04

0.02

0

– 0.02

– 0.040

0

3

2

1

0

5

4

3

2

1

ΔR(B)/R(0)

T (K)20 40 60 80 100

0 0.01 0.021/d w (nm–1)

1 2 3 4 5B (T)

Bm (T)

Bm (T)

a) b)

c)

65 nm

109 nm

Fig. 4.27 (a) Longitudinal magnetoresistance, ΔR(B)/R(0), at 2 K as a function of B for Bi nanowire arrays with diam-eters of 65 and 109 nm before thermal annealing. (b) The peak position Bm as a function of temperature for the 109 nmdiameter Bi nanowire array after thermal annealing. (c) The peak position Bm of the longitudinal MR (after thermalannealing) at 2 K as a function of 1/dW , the reciprocal of the nanowire diameter (after [4.157])

a temperature dependence of T−1/2 at low tempera-tures, consistent with 1-D localization theory. Thus, dueto this localization effect, the use of nanowires withvery small diameters for transport applications may belimited.

MagnetoresistanceMagnetoresistance (MR) measurements provide aninformative technique for characterizing nanowires,because these measurements yield a great deal ofinformation about the electron scattering with wireboundaries, the effects of doping and annealing on scat-tering, and localization effects in the nanowires [4.150].For example, at low fields the MR data show a quadraticdependence on the B field from which carrier mobilityestimates can be made (Fig. 4.27 at low B field).

Figure 4.27 shows the longitudinal magnetoresis-tance (B parallel to the wire axis) for 65 and 109 nmBi nanowire samples (before thermal annealing) at 2 K.The MR maxima in Fig. 4.27a are due to the classicalsize effect, where the wire boundary scattering is re-duced as the cyclotron radius becomes smaller than thewire radius in the high field limit, resulting in a decreasein the resistivity. This behavior is typical for the longi-tudinal MR of Bi nanowires in the diameter range of 45

PartA

4.2


to 200 nm [4.8,149,150,157], and the peak position Bmmoves to lower B field values as the wire diameter in-creases, as shown in Fig. 4.27c [4.157], where Bm varieslinearly with 1/dW. The condition for the occurrence ofBm is approximately given by Bm ≈ 2c�kF/edW wherekF is the wave vector at the Fermi energy. The peakposition Bm is found to increase linearly with increas-ing temperature in the range of 2–100 K, as shownin Fig. 4.27b [4.157]. As T is increased, phonon scat-tering becomes increasingly important, and thereforea higher magnetic field is required to reduce the resis-tivity associated with boundary scattering sufficiently tochange the sign of the MR. Likewise, increasing thegrain boundary scattering is also expected to increasethe value of Bm at a given T and wire diameter.

The presence of the peak in the longitudinal MR ofnanowires requires a high crystal quality with long car-rier mean free paths along the nanowire axis, so thatmost scattering events occur at the wire boundary in-stead of at a grain boundary, at impurity sites, or atdefect sites within the nanowire. Liu et al. have inves-tigated the MR of 400 nm Bi nanowires synthesized byelectrochemical deposition [4.74], and no peak in thelongitudinal MR is observed. The absence of a magne-toresistance peak may be attributed to a higher defectlevel in the nanowires produced electrochemically andto a large wire diameter, much longer than the carriermean free path. The negative MR observed for the Binanowire arrays above Bm (Fig. 4.27) shows that wireboundary scattering is a dominant scattering processfor the longitudinal magnetoresistance, thereby estab-lishing that the mean free path is larger than the wirediameter and that a ballistic transport behavior is indeedobserved in the high field regime.

In addition to the longitudinal magnetoresistancemeasurements, transverse magnetoresistance measure-ments (B perpendicular to the wire axis) have also beenperformed on Bi nanowire array samples [4.8,150,157],where a monotonically increasing B2 dependence overthe entire range 0 ≤ B ≤ 5.5 T is found for all Binanowires studied thus far. This is as expected, since thewire boundary scattering cannot be reduced by a mag-netic field perpendicular to the wire axis. The transversemagnetoresistance is also found to be always larger thanthe longitudinal magnetoresistance in nanowire arrays.

By applying a magnetic field to nanowires at verylow temperatures (≤ 5 K), one can induce a transitionfrom a 1-D confined system at low magnetic fieldsto a 3-D confined system as the field strength in-creases, as shown in Fig. 4.28 for the longitudinal MRof Bi nanowire arrays of various nanowire diameters

(28–70 nm) for T < 5 K [4.150]. In these curves, a sub-tle steplike feature is seen at low magnetic fields, whichis found to depend only on the wire diameter, and isindependent of temperature, the orientation of the mag-netic field, and even on the nanowire material (see forexample Sb nanowires [4.155]). The lack of a depen-dence of the magnetic field at which the step appearson temperature, field orientation, and material type in-dicates that the phenomenon is related to the magneticfield length LH = (�/eB)1/2. The characteristic lengthLH is the spatial extent of the wave function of electronsin the lowest Landau level, and LH is independent ofthe carrier effective masses. Setting LH(Bc) equal to thediameter dW of the nanowire defines a critical magneticfield strength Bc below which the wavefunction is con-

1.012

1.006

1.004

1

0 1 2 3 4 5

R (B)/(B = 0 T)

B (T)

1.008

1.004

1

1.02

1.01

1

1.1

1.05

1

0.13 T1.85 K

3 K

Bi J1A70 nm

4 K

Bi J4A48 nm0.3 T

1.35 K2 K4 K

Bi J2B36 nm

0.54 T1.39 K

1.97 K

4.32 K

Bi J5A28 nm

0.87 T1.39 K

1.97 K

4.33 K

Fig. 4.28 Longitudinal magnetoresistance as a function ofmagnetic field for Bi nanowires of the diameters indicated.The vertical bars indicate the critical magnetic field Bc atwhich the magnetic length equals the nanowire diameter(after [4.150])

PartA

4.2


fined by the nanowire boundary (the 1-D regime), andabove which the wavefunction is confined by the mag-netic field (the 3-D regime). The physical basis for thisphenomenon is associated with confinement of a singlemagnetic flux quantum within the nanowire cross sec-tion [4.150]. This phenomenon, though independent oftemperature, is observed for T ≤ 5 K, since the phasebreaking length has to be larger than the wire diameter.This calculated field strength Bc indicated in Fig. 4.28by vertical lines for the appropriate nanowire diameters,provides a good fit to the steplike features in these MRcurves.

The Shubnikov–de Haas (SdH) quantum oscillatoryeffect, which results from the passage of the quantizedLandau levels through the Fermi energy as the fieldstrength varies, should, in principle, provide the mostdirect measurement of the Fermi energy and carrier den-sity. For example, Heremans et al. have demonstratedthat SdH oscillations can be observed in Bi nanowiresamples with diameters down to 200 nm [4.159], andthey have demonstrated that Te doping can be used toraise the Fermi energy in Bi nanowires. Such infor-mation on the Fermi energy is important because, forcertain applications based on nanowires, it is neces-sary to place the Fermi energy near a subband edgewhere the density of states has a sharp feature. How-ever, due to the unusual 1-D geometry of nanowires,other characterization techniques that are commonlyused in bulk materials to determine the Fermi energyand the carrier concentration (such as Hall measure-ment) cannot be applied to nanowire systems. Theobservation of the SdH oscillatory effect requires crys-tal samples of very high quality which allow carriers toexecute a complete cyclotron orbit in the nanowire be-fore they are scattered. For small nanowire diameters,large magnetic fields are required to produce cyclotronradii smaller than the wire radius. For some nanowiresystems, all Landau levels may have passed throughthe Fermi level at such a high field strength, and insuch a case, no oscillations can be observed. The lo-calization effect may also prevent the observation ofSdH oscillations for very small diameter (≤ 10 nm)nanowires. Observing SdH oscillations in highly dopedsamples (as may be required for certain applications)may be difficult because impurity scattering reducesthe mean free path, requiring high B fields to satisfythe requirement that carriers complete a cyclotron orbitprior to scattering. Therefore, although SdH oscilla-tions provide the most direct method of measuring theFermi energy and carrier density of nanowire samples,this technique may, however, not work for small-

diameter nanowires, nor for nanowires that are heavilydoped.

Thermoelectric PropertiesNanowires are predicted to hold great promise for ther-moelectric applications [4.147, 160], due to their novelband structure compared to their bulk counterparts andthe expected reduction in thermal conductivity associ-ated with enhanced boundary scattering (see below).Due to the sharp density of states at the 1-D sub-band edges (where the van Hove singularities occur),nanowires are expected to exhibit enhanced Seebeckcoefficients compared to their bulk counterparts. Sincethe Seebeck coefficient measurement is intrinsically in-dependent of the number of nanowires contributing tothe signal, the measurements on nanowire arrays of uni-form wire diameter are, in principle, as informative assingle-wire measurements. The major challenge withmeasuring the Seebeck coefficients of nanowires lies inthe design of tiny temperature probes to accurately de-termine the temperature difference across the nanowire.Figure 4.29a shows the schematic experimental setupfor the Seebeck coefficient measurement of nanowirearrays [4.161], where two thermocouples are placed onboth faces of a nanowire array and a heater is attachedto one face of the array to generate a temperature gra-dient along the nanowire axis. Ideally, the size of thethermocouples should be much smaller than the thick-ness of the nanowire array template (i. e. the nanowirelength) to minimize error. However, due to the thin-ness of most templates (≤ 50 μm) and the large size ofcommercially-available thermocouples (≈ 12 μm), themeasured Seebeck coefficient values are usually under-estimated.

The thermoelectric properties of Bi nanowire sys-tems have been investigated extensively because oftheir potential as good thermoelectric materials. Fig-ure 4.29b shows the measured Seebeck coefficientsS(T ) as a function of temperature for nanowire ar-rays with diameters of 40 and 65 nm and differentisoelectronic Sb alloy concentrations [4.154], andS(T ) results for bulk Bi are shown (solid curve) forcomparison. Thermopower enhancement is observedin Fig. 4.29b as the wire diameter decreases and asthe Sb content increases, which is attributed to thesemimetal–semiconductor transition induced by quan-tum confinement and to Sb alloying effects in Bi1−xSbxnanowires. Heremans et al. have observed a substan-tial increase in the thermopower of Bi nanowires as thewire diameter decreases further, as shown in Fig. 4.30afor Bi(15 nm)/silica and Bi(9 nm)/alumina nanocom-

PartA

4.2


Fig. 4.29 (a) Experimental setup for the measurement ofthe Seebeck coefficient in nanowire arrays (after [4.161]).(b) Measured Seebeck coefficient as a function of temper-ature for Bi ( , ) and Bi0.95Sb0.05 ( , ) nanowires withdifferent diameters. The solid curve denotes the Seebeckcoefficient for bulk Bi (after [4.154]) �

posites [4.52]. The enhancement is due to the sharpdensity of states near the Fermi energy in a 1-D system.Although the samples in Fig. 4.30a also possess veryhigh electrical resistance (∼ GΩ), the results for theBi(9 nm)/alumina samples show that the Seebeck co-efficient can be enhanced by almost 1000 times relativeto bulk material. However, for Bi nanowires with verysmall diameters (≈ 4 nm), the localization effect be-comes dominant, which compromises the thermopowerenhancement. Therefore, for Bi nanowires, the optimalwire diameter range for the largest thermopower en-hancement is found to be between 4 and 15 nm [4.52].

The effect of the nanowire diameter on the ther-mopower of nanowires has also been observed in Znnanowires [4.52]. Figure 4.30b shows the Seebeckcoefficient of Zn(9 nm)/alumina and Zn(4 nm)/Vycorglass nanocomposites, also exhibiting enhanced ther-mopower as the wire diameter decreases. It is found thatwhile 9 nm Zn nanowires still exhibit metallic behavior,

1×106

1×105

1×104

1×103

1×102

1×101

1×100

0 100 200 300

0

– 50

– 100

– 150

|S| (μV/K)

T (K)T (K)0 100 200 300

S (μV/K)a) b)

Bulk Zn

9 nm, Al2O3 sample 1

9 nm, Al2O3sample 2

15 nm, SiO2sample 1

15 nm, SiO2sample 2

Bulk Bi

Bi 200 nm diameter wires

Zn 9 nm/Al2O3

Zn 4 nm/Vycor

T 1

– π2kB

6e

Fig. 4.30 (a) Absolute value of the Seebeck coefficient of two Bi(15 nm)/silica and two Bi(9 nm)/alumina nanocompositesamples, in comparison to bulk Bi and 200 nm Bi nanowires in the pores of alumina templates (after [4.52]). The full lineon top part of the figure is a fit to a T−1 law. The Seebeck coefficient of the Bi(9 nm)/alumina composite is positive; therest are negative. (b) The Seebeck coefficient of Zn(9 nm)/Al2O3 and Zn(4 nm)/Vycor glass nanocomposite samples incomparison to bulk Zn (after [4.52])

S (μV/K)

0

– 20

– 40

– 60

– 80

Temperature (K)0 100 200 300

a)

b)

65nm Bi40nm Bi65nm Bi0.95Sb0.05

45nm Bi0.95Sb0.05

Bulk Bi

Heater

Thermocouple

To voltmeterHeat Sink

Nanowiresample

To voltmeter

PartA

4.2


the thermopower of 4 nm Zn nanowires shows a differ-ent temperature dependence, which may be due to the1-D localization effect, although further investigation isrequired for definitive identification of the conductionmechanism in such small nanowires.

Quantum Wire SuperlatticesThe studies on superlattice nanowires, which possessa periodic modulation in their materials compositionalong the wire axis, have attracted much attentionrecently because of their promise in various appli-cations, such as thermoelectrics (Sect. 4.3.2) [4.90,162], nanobarcodes (Sect. 4.3.3) [4.110], nanolasers(Sect. 4.3.3) [4.92], one-dimensional waveguides, andresonant tunneling diodes [4.94, 163]. Figure 4.31ashows a schematic structure of a superlattice nanowireconsisting of interlaced quantum dots of two differentmaterials, as denoted by A and B. Various tech-niques have been developed to synthesize superlatticenanowire structures with different interface conditions,as mentioned in Sects. 4.1.1 and 4.1.2.

In this superlattice (SL) nanowire structure, theelectronic transport along the wire axis is made possi-ble by the tunneling between adjacent quantum dots,while the uniqueness of each quantum dot and its 0-D characteristic behavior is maintained by the energydifference of the conduction or valence bands be-tween quantum dots of different materials (Fig. 4.31b),which provides some amount of quantum confine-ment. Recently, Björk et al. have observed interestingnonlinear I–V characteristics with a negative dif-ferential resistance in one-dimensional heterogeneousstructures made of InAs and InP, where InP serves asthe potential barrier [4.94, 163]. The nonlinear I–Vbehavior is associated with the double barrier reso-

a)

b)

A BLA LB

mA mBEC

A +εnmA

ECB +εnm

B

D

Fig. 4.31 (a) Schematic diagram of superlattice (seg-mented) nanowires consisting of interlaced nanodots A andB of the indicated length and wire diameter. (b) Schematicpotential profile of the subbands in the superlatticenanowire (after [4.162])

0 5 10 15 20

2

1.5

1

0.5

0

ZT

Segment length (nm)

PbSe/PbS SL nanowire

PbSe0.5S0.5 alloy

PbSe

PbS

Fig. 4.32 Optimal ZT calculated as a function of segmentlength for 10 nm diameter PbSe/PbS nanowires at 77 K,where optimal refers to the placement of the Fermi levelto optimize ZT . The optimal ZT for 10 nm diameter PbSe,PbS, and PbSe0.5S0.5 nanowires are 0.33, 0.22, and 0.48,respectively (after [4.153])

nant tunneling process in one-dimensional structures,demonstrating that transport phenomena occur in su-perlattice nanowires via tunneling and the possibilityof controlling the electronic band structure of the SLnanowires by carefully selecting the constituent mater-ials. This new kind of structure is especially attractivefor thermoelectric applications, because the interfacesbetween the nanodots can reduce the lattice ther-mal conductivity by blocking the phonon conductionalong the wire axis, while electrical conduction maybe sustained and even benefit from the unusual elec-tronic band structures due to the periodic potentialperturbation. For example, Fig. 4.32 shows the cal-culated dimensionless thermoelectric figure of meritZT = S2σT/κ (Sect. 4.3.2) where κ is the total thermalconductivity (including both the lattice and electroniccontributions) of 10 nm diameter PbS/PbSe superlatticenanowires as a function of the segment length. A higherthermoelectric performance than for PbSe0.5S0.5 al-loy nanowires can be achieved for a 10 nm diametersuperlattice nanowire with segment lengths ≤ 7 nm.However, the localization effect, which may becomeimportant for very short segment lengths, may jeop-ardize this enhancement in the ZT of superlatticenanowires [4.153].

Thermal Conductivity of NanowiresExperimental measurements of the temperature depen-dence of the thermal conductivity κ(T ) of individualsuspended nanowires have been carried out on study thedependence of κ(T ) on wire diameter. In this context,

PartA

4.2


0 100 200 300

100

80

60

40

20

0

κ (W/(m K))

T (K)

D = 115 nm

56

37

22

Fig. 4.33 Predicted thermal conductivities of Si nanowiresof various diameters (after [4.168])

measurements have been made on nanowires down toonly 22 nm in diameter [4.164]. Such measurements arevery challenging and are now possible due to techno-logical development in the micro- and nanofabricationof miniature thermal sensors, and the use of nanometer-size thermal scanning probes [4.128, 165, 166]. Theexperiments show that the thermal conductivity of smallhomogeneous nanowires may be more than one orderof magnitude smaller than in the bulk, due mainly tostrong boundary scattering effects [4.167]. Phonon con-finement effects may eventually become important innanowires with even smaller diameters. Measurementson mats of nanowires (Fig. 4.12) do not generally give

100

10

1

0.1

Gth (T )/16g0

6000100060010060

b)

Temperature T (mK)

a)

Fig. 4.34 (a) Suspended mesoscopic phonon device used to measure ballistic phonon transport. The device consists ofan 4 × 4 μm2 phonon cavity (center) connected to four Si3N4 membranes, 60 nm thick and less than 200 nm wide. Thetwo bright C-shaped objects on the phonon cavity are thin film heating and sensing Cr/Au resistors, whereas the darkregions are empty space. (b) Log–log plot of the temperature dependence of the thermal conductance G0 of the structurein (a) normalized to 16g0 (see text) (after [4.169])

reliable results because the contact thermal resistancebetween adjacent nanowires tends to be high, which isin part due to the thin surface oxide coating which mostnanowires have. This surface oxide coating may alsobe important for thermal conductivity measurements onindividual suspended nanowires because of the relativeimportance of phonon scattering at the lateral walls ofthe nanowire.

The most extensive experimental thermal con-ductivity measurements have been done on Sinanowires [4.164], where κ(T ) measurements havebeen made on nanowires in the diameter range 22 ≤dW ≤ 115 nm. The results show a large decrease in thepeak of κ(T ), associated with Umklapp processes as dWdecreases, indicating a growing importance of boundaryscattering and a corresponding decreasing importanceof phonon–phonon scattering. At the smallest wirediameter of 22 nm, a linear κ(T ) dependence is foundexperimentally, consistent with a linear T dependenceof the specific heat for a 1-D system, and a temperature-independent mean free path and velocity of sound. Fur-ther insights are obtained through studies of the thermalconductivity of Si/SiGe superlattice nanowires [4.170].

Model calculations for κ(T ) based on a radia-tive heat transfer model have been carried out forSi nanowires [4.168]. These results show that thepredicted κ(T ) behavior for Si nanowires is simi-lar to that observed experimentally in the range of37 ≤ dW ≤ 115 nm regarding both the functional formof κ(T ) and the magnitude of the relative decreasein the maximum thermal conductivity κmax as a func-tion of dW. However, the model calculations predict

PartA

4.2


a substantially larger magnitude for κ(T ) (by 50% ormore) than is observed experimentally. Furthermore,the model calculations (Fig. 4.34) do not reproduce theexperimentally observed linear T dependence for the22 nm nanowires, but rather predict a 3-D behavior forboth the density of states and the specific heat in 22 nmnanowires [4.168, 171, 172].

Thermal conductance measurements on GaAsnanowires below 6 K show a power law dependence,but the T dependence becomes somewhat less pro-nounced below ≈ 2.5 K [4.165]. This deviation fromthe power law temperature dependence led to a moredetailed study of the quantum limit for the thermalconductance. To carry out these more detailed experi-ments, a mesoscopic phonon resonator and waveguidedevice were constructed that included four ≈ 200 nmwide and 85 nm thick silicon nitride nanowirelikenanoconstrictions (Fig. 4.33a), and this was used toestablish the quantized thermal conductance limitof g0 = π2k2

BT/(3h) (Fig. 4.33b) for ballistic phonontransport [4.169, 173]. For temperatures above 0.8 K,the thermal conductance in Fig. 4.33b follows a T 3 law,but as T is further reduced, a transition to a linearT dependence is observed, consistent with a phononmean free path of ≈ 1 μm, and a thermal conductancevalue approaching 16g0, corresponding to four mass-less phonon modes per channel and four channels intheir phonon waveguide structure (Fig. 4.33a). Ballis-tic phonon transport occurs when the thermal phononwavelength (380 nm for the experimental structure) issomewhat greater than the width of the phonon waveg-uide at the waveguide constriction.

4.2.4 Optical Properties

Optical methods provide an easy and sensitive tool formeasuring the electronic structures of nanowires, sinceoptical measurements require minimal sample prepara-tion (for example, contacts are not required) and themeasurements are sensitive to quantum effects. Opti-cal spectra of 1-D systems, such as carbon nanotubes,often show intense features at specific energies near sin-gularities in the joint density of states that are formedunder strong quantum confinement conditions. A vari-ety of optical techniques have shown that the propertiesof nanowires are different to those of their bulk coun-terparts, and this section of the review focuses on thesedifferences in the optical properties of nanowires.

Although optical properties have been shown toprovide an extremely important tool for characteriz-ing nanowires, the interpretation of these measurements

is not always straightforward. The wavelength of lightused to probe the sample is usually smaller than the wirelength, but larger than the wire diameter. Hence, theprobe light used in an optical measurement cannot befocused solely onto the wire, and the wire and the sub-strate on which the wire rests (or host material, if thewires are embedded in a template) are probed simulta-neously. For measurements, such as photoluminescence(PL), if the substrate does not luminescence or absorbin the frequency range of the measurements, PL mea-sures the luminescence of the nanowires directly andthe substrate can be ignored. However, in reflection andtransmission measurements, even a nonabsorbing sub-strate can modify the measured spectra of nanowires.

In this section we discuss the determination of thedielectric function for nanowires in the context of effec-tive medium theories. We then discuss various opticaltechniques with appropriate examples that sensitivelydifferentiate nanowire properties from those also foundin the parent bulk material, placing particular empha-sis on electronic quantum confinement effects. Finally,phonon confinement effects are reviewed.

The Dielectric FunctionIn this subsection, we review the use of effectivemedium theory as a method to handle the opticalproperties of nanowires whose diameters are typicallysmaller than the wavelength of light, noting that ob-servable optical properties of materials can be related tothe complex dielectric function [4.174, 175]. Effectivemedium theories [4.176, 177] can be applied to modelthe nanowire and substrate as one continuous compos-ite with a single complex dielectric function (ε1 + iε2),where the real and imaginary parts of the dielectricfunction ε1 and ε2 are related to the index of refraction(n) and the absorption coefficient (K ) by the relationε1 + iε2 = (n + iK )2. Since photons at visible or infraredwavelengths see a dielectric function for the compos-ite nanowire array/substrate system that is differentfrom that of the nanowire itself, the optical transmis-sion and reflection are different from what they wouldbe if the light were focused only on the nanowire. Onecommonly observed consequence of effective mediumtheory is the shift in the plasma frequency in accordancewith the percentage of nanowire material that is con-tained in the composite [4.178]. The plasma resonanceoccurs when ε1(ω) becomes zero, and the plasma fre-quency of the nanowire composite will shift to lower(higher) energies when the magnitude of the dielectricfunction of the host materials is larger (smaller) thanthat of the nanowire.

PartA

4.2


Although reflection and transmission measurementsprobe both the nanowire and the substrate, the opticalproperties of the nanowires can be determined indepen-dently. One technique for separating out the dielectricfunction of the nanowires from the host is to use aneffective medium theory in reverse. Since the dielec-tric function of the host material is often known, andthe dielectric function of the composite material can bemeasured by the standard method of using reflectionand transmission measurements in combination witheither the Kramer–Kronig relations or Maxwell’s equa-tions, the complex dielectric function of the nanowirescan be deduced. An example where this approach hasbeen used successfully is for the determination of thefrequency dependence of the real and imaginary partsof the dielectric function ε1(ω) and ε2(ω) for a paral-lel array of bismuth nanowires filling the pores of analumina template [4.179].

Characteristic Optical Properties of NanowiresA wide range of optical techniques are available for thecharacterization of nanowires, to distinguish their prop-erties from those of their parent bulk materials. Somedifferences in properties relate to geometric differences,such as the small diameter size and the large length-to-diameter ratio (also called the aspect ratio), while othersfocus on quantum confinement issues.

Probably the most basic optical technique isto measure the reflection and/or transmission ofa nanowire to determine the frequency- dependent realand imaginary parts of the dielectric function. This tech-nique has been used, for example, to study the bandgap and its temperature dependence in gallium nitridenanowires in the 10–50 nm range in comparison to bulkvalues [4.180]. The plasma frequency, free carrier den-sity, and donor impurity concentration as a functionof temperature were also determined from the infraredspectra, which is especially useful for nanowire re-search, since Hall effect measurements cannot be madeon nanowires.

Another common method used to study nanowires isphotoluminescence (PL) or fluorescence spectroscopy.Emission techniques probe the nanowires directly andthe effect of the host material does not have to beconsidered. This characterization method has beenused to study many properties of nanowires, suchas the optical gap behavior, oxygen vacancies inZnO nanowires [4.55], strain in Si nanowires [4.181],and quantum confinement effects in InP nanowires[4.182]. Figure 4.35 shows the photoluminescence ofInP nanowires as a function of wire diameter, thereby

providing direct information on the effective bandgap.As the wire diameter of an InP nanowire is decreasedso that it becomes smaller than the bulk exciton diam-eter of 19 nm, quantum confinement effects set in,and the band gap is increased. This results in anincrease in the PL peak energy. The smaller the effec-tive mass, the larger the quantum confinement effects.When the shift in the peak energy as a function ofnanowire diameter Fig. 4.35 is analyzed using an ef-fective mass model, the reduced effective mass of theexciton is deduced to be 0.052 m0, which agrees quitewell with the literature value of 0.065 m0 for bulkInP. Although the linewidths of the PL peak for thesmall-diameter nanowires (10 nm) are smaller at lowtemperature (7 K), the observation of strong quantumconfinement and bandgap tunability effects at roomtemperature are significant for photonics applications ofnanowires (Sect. 4.3.3).

The resolution of photoluminescence (PL) opticalimaging of a nanowire is, in general, limited by thewavelength of light. However, when a sample is placedvery close to the detector, the light is not given a chanceto diffract, and so samples much smaller than the wave-length of light can be resolved. This technique is knownas near-field scanning optical microscopy (NSOM) andhas been used to successfully image nanowires [4.183].For example, Fig. 4.36 shows the topographical (a) and(b) NSOM PL images of a single ZnO nanowire.

Magnetooptics can be used to measure the elec-tronic band structure of nanowires. For example,magnetooptics in conjunction with photoconductancehas been proposed as a tool to determine band parame-ters for nanowires, such as the Fermi energy, electroneffective masses, and the number of subbands to beconsidered [4.184]. Since different nanowire subbandshave different electrical transmission properties, theelectrical conductivity changes when light is used toexcite electrons to higher subbands, thereby provid-ing a method for studying the electronic structure ofnanowires optically. Magnetooptics can also be used tostudy the magnetic properties of nanowires in relationto bulk properties [4.27, 185]. For example, the surfacemagnetooptical Kerr effect has been used to measure thedependence of the magnetic ordering temperature of Fe-Co alloy nanowires on the relative concentration of Feand Co [4.185], and it was used to find that, unlike inthe case of bulk Fe-Co alloys, cobalt in nanowires in-hibits magnetic ordering. Nickel nanowires were foundto have a strong increase in their magnetooptical activitywith respect to bulk nickel. This increase is attributed tothe plasmon resonance in the wires [4.186].

PartA

4.2


1.3 1.5 1.7 1.4 1.5 1.6

10 30 50

Energy (eV)

1.6

1.5

1.4

1.55

1.5

1.45

Energy (eV)

Diameter (nm)10 30 50

Diameter (nm)

10 nm

15 nm

20 nm

50 nm

10 nm

15 nm

20 nm

50 nm

RT 7 K

a) Intensity (arb. units) b)

c) PL max (eV) d)

RT 7 K

Fig. 4.35a–d Photoluminescence ofInP nanowires of varying diameters at7 K (b,d) and room temperature (a,c)showing quantum confinement effectsof the exciton for wire diameters ofless than 20 nm (after [4.182])

a) b)Fig. 4.36 (a) Topographical and(b) photoluminescence (PL) near-field scanning optical microscopy(NSOM) images of a single ZnOnanowire waveguide (after [4.183])

Nonlinear optical properties of nanowires havereceived particular attention since the nonlinear be-havior is often enhanced compared to bulk materialsand the nonlinear effects can be utilized for manyapplications. One such study measured the second har-monic generation (SHG) and third harmonic generation

(THG) in a single nanowire using near-field opticalmicroscopy [4.187]. ZnO nanowires were shown tohave strong SHG and THG effects that are highlypolarization-sensitive, and this polarization sensitivitycan be explained on the basis of optical and geomet-rical considerations. Some components of the second

PartA

4.2


harmonic polarization tensor are found to be enhancedin nanowires while others are suppressed as the wirediameter is decreased, and such effects could be of inter-est for device applications. The authors also showed thatthe second-order nonlinearities are mostly wavelength-independent for λ < 400 nm, which is in the transparentregime for ZnO, below the onset of band gap absorp-tion, and this observation is also of interest for deviceapplications.

Reflectivity and transmission measurements havealso been used to study the effects of quantum con-finement and surface effects on the low-energy indirecttransition in bismuth nanowires [4.189]. Black et al.investigated an intense and sharp absorption peak inbismuth nanowires, which is not observed in bulk bis-muth. The energy position Ep of this strong absorptionpeak increases with decreasing diameter. However, therate of increase in energy with decreasing diameter|∂Ep/∂dW| is an order of magnitude less than that pre-dicted for either a direct interband transition or forintersubband transitions in bismuth nanowires. On theother hand, the magnitude of |∂Ep/∂dW| agrees wellwith that predicted for an indirect L-point valence to T -

0.05

01000 2000 3000 4000

1

0.8

0.6

0.4

0.2

0

1/λ (cm–1) 1/λ (cm–1)1000 2000 3000 4000

0.1a) Experimentally measured b) Simulation of the indirect L–T transition

L-point T-point

Trigonal (z)

T

AB

Binary (x)

B

Bisectrix (y)

[012]Γ

Fig. 4.37 (a) The measured optical transmission spectra as a function of wavenumber (1/λ) of a ≈ 45 nm diameterbismuth nanowire array. (b) The simulated optical transmission spectrum resulting from an indirect transition of an L-point electron to a T -point valence subband state. The insert in (a) shows the bismuth Brillouin zone, and the locationsof the T -point hole and the three L-point electron pockets, including the nondegenerate A, and the doubly-degenerate Bpockets. The insert in (b) shows the indirect L–T point electronic transition induced by a photon with an energy equalto the energy difference between the initial and final states minus the phonon energy (about 100 cm−1) needed to satisfyconservation of energy in a Stokes process (after [4.188])

point valence band transition (Fig. 4.37). Since both theinitial and final states for the indirect L–T point valenceband transition downshift in energy as the wire diam-eter dW is decreased, the shift in the absorption peakresults from a difference between the effective massesand not from the actual value of either of the masses.Hence the diameter dependence of the absorption peakenergy is an order of magnitude less for a valence to va-lence band indirect transition than for a direct interbandL-point transition. Furthermore, the band-tracking ef-fect for the indirect transition gives rise to a large valuefor the joint density of states, thus accounting for thehigh intensity of this feature. The enhancement in theabsorption resulting from this indirect transition mayarise from a gradient in the dielectric function, whichis large at the bismuth–air or bismuth–alumina inter-faces, or from the relaxation of momentum conservationrules in nanosystems. It should be noted that, in con-trast to the surface effect for bulk samples, the wholenanowire contributes to the optical absorption due to thespatial variation in the dielectric function, since the pen-etration depth is larger than or comparable to the wirediameter. In addition, the intensity can be quite signif-

PartA

4.2


icant because there are abundant initial state electrons,final state holes, and appropriate phonons for makingan indirect L–T point valence band transition at roomtemperature. Interestingly, the polarization dependenceof this absorption peak is such that the strong absorp-tion is present when the electric field is perpendicularto the wire axis, but is absent when the electric field isparallel to the wire axis, contrary to a traditional polar-izer, such as a carbon nanotube where the optical E fieldis polarized by the nanotube itself and is aligned alongthe carbon nanotube axis. The observed polarizationdependence for bismuth nanowires is consistent witha surface-induced effect that increases the coupling be-tween the L-point and T -point bands throughout the fullvolume of the nanowire. Figure 4.37 shows the exper-imentally observed transmission spectrum in bismuthnanowires of ≈ 45 nm diameter (a), and the simulatedoptical transmission from an indirect transition in bis-muth nanowires of ≈ 45 nm diameter is also shown forcomparison in (b). The indirect L–T point valence bandtransition mechanism [4.188] is also consistent withobservations of the effect on the optical spectra of a de-crease in the nanowire diameter and of n-type doping ofbismuth nanowires with Te.

Phonon Confinement EffectsPhonons in nanowires are spatially confined by thenanowire cross-sectional area, crystalline boundariesand surface disorder. These finite size effects give riseto phonon confinement, causing an uncertainty in thephonon wavevector which typically gives rise to a fre-quency shift and lineshape broadening. Since zonecenter phonons tend to correspond to maxima in thephonon dispersion curves, the inclusion of contribu-tions from a broader range of phonon wave vectorsresults in both a downshift in frequency and an asym-metric broadening of the Raman line, which developsa low frequency tail. These phonon confinement ef-fects have been theoretically predicted [4.191, 192] andexperimentally observed in GaN [4.190], as shownin Fig. 4.38 for GaN nanowires with diameters in therange 10–50 nm. The application of these theoreti-cal models indicates that broadening effects should benoticeable as the wire diameter in GaN nanowires de-creases to ≈ 20 nm. When the wire diameter decreasesfurther to ≈ 10 nm, the frequency downshift and asym-metric Raman line broadening effects should become

2000

1500

1000

500

00 200 400

Raman shift (cm–1)600 800 1000


T = 300 K

λi = 514.5 nm

E2

E2

E1(TO)

A1(TO) E1(LO)

A1(LO)

GaNnanowires

GaNfilm

Sapphire

Fig. 4.38 Room-temperature Raman scattering spectra ofGaN nanowires and of a 5 μm thick GaN epilayer film withgreen (514.5 nm) laser excitation. The Raman scattering re-sponse was obtained by dividing the measured spectra bythe Bose–Einstein thermal factor [4.190]

observable in the Raman spectra for the GaN nanowiresbut are not found in the corresponding spectra forbulk GaN.

The experimental spectra in Fig. 4.38 show thefour A1 + E1 +2E2 modes expected from symme-try considerations for bulk GaN crystals. Two typesof quantum confinement effects are observed. Thefirst type is the observation of the downshift and theasymmetric broadening effects discussed above. Obser-vations of such downshifts and asymmetric broadeninghave also been recently reported in 7 nm diameter Sinanowires [4.193]. A second type of confinement effectfound in Fig. 4.38 for GaN nanowires is the appearanceof additional Raman features not found in the corre-sponding bulk spectra and associated with combinationmodes, and a zone boundary mode. Resonant enhance-ment effects were also observed for the A1(LO) phononat 728 cm−1 (Fig. 4.38) at higher laser excitation ener-gies [4.190].

PartA

4.2


4.3 Applications

In the preceding sections we have reviewed many ofthe central characteristics that make nanowires in somecases similar to and in some cases very different fromtheir parent materials. We have also shown that someproperties are diameter-dependent, and these proper-ties are therefore tunable during synthesis. Thus, it isof great interest to find applications that could bene-fit in unprecedented ways from both the unique andtunable properties of nanowires and the small sizes ofthese nanostructures, especially in the miniaturizationof conventional devices. As the synthetic methods forthe production of nanowires are maturing (Sect. 4.1)and nanowires can be made in reproducible and cost-effective ways, it is only a matter of time beforeapplications will be seriously explored. This is a timelydevelopment, as the semiconductor industry will soonbe reaching what seems to be its limit in feature sizereduction, and approaching a classical-to-quantum sizetransition. At the same time, the field of biotechnol-ogy is expanding through the availability of tremendousgenome information and innovative screening assays.Since nanowires are similar in size to the shrinkingelectronic components and to cellular biomolecules, itis only natural for nanowires to be good candidatesfor applications in these fields. Commercialization ofnanowire devices, however, will require reliable massproduction, effective assembly techniques and qualitycontrol methods.

In this section, applications of nanowires to elec-tronics (Sect. 4.3.1), thermoelectrics (Sect. 4.3.2), op-tics (Sect. 4.3.3), chemical and biochemical sensing(Sect. 4.3.4), and magnetic media (Sect. 4.3.5) are dis-cussed.

4.3.1 Electrical Applications

The microelectronics industry continues to face tech-nological (in lithography for example) and economicchallenges as the device feature size is decreased, es-pecially below 100 nm. The self-assembly of nanowiresmight present a way to construct unconventional de-vices that do not rely on improvements in photolithogra-phy and, therefore, do not necessarily imply increasingfabrication costs. Devices made from nanowires haveseveral advantages over those made by photolithog-raphy. A variety of approaches have been devised toorganize nanowires via self-assembly (Sect. 4.1.4), thuseliminating the need for the expensive lithographic tech-niques normally required to produce devices the size

of typical nanowires that are discussed in this review.In addition, unlike traditional silicon processing, dif-ferent semiconductors can be used simultaneously innanowire devices to produce diverse functionalities. Notonly can wires of different materials be combined, buta single wire can be made of different materials. Forexample, junctions of GaAs and GaP show rectifyingbehavior [4.92], thus demonstrating that good electronicinterfaces between two different semiconductors can beachieved in the synthesis of multicomponent nanowires.Transistors made from nanowires could also hold ad-vantages due to their unique morphology. For example,in bulk field effect transistors (FETs), the depletionlayer formed below the source and drain region resultsin a source–drain capacitance which limits the oper-ation speed. However, in nanowires, the conductor issurrounded by an oxide and thus the depletion layercannot be formed. Thus, depending on the device de-sign, the source–drain capacitance in nanowires couldbe greatly minimized and possibly eliminated.

Device functionalities common in conventionalsemiconductor technologies, such as p-n junctiondiodes [4.142], field-effect transistors [4.144], logicgates [4.142], and light-emitting diodes [4.92, 194],have been recently demonstrated in nanowires, show-ing their promise as building blocks that could be usedto construct complex integrated circuits by employ-ing the bottom-up paradigm. Several approaches havebeen investigated to form nanowire diodes (Sect. 4.2.2).For example, Schottky diodes can be formed by con-tacting a GaN nanowire with Al electrodes [4.143].Furthermore, p-n junction diodes can be formed atthe crossing of two nanowires, such as the crossingof n- and p-type InP nanowires doped by Te and Zn,respectively [4.194], or Si nanowires doped by phos-phorus (n-type) and boron (p-type) [4.195]. In additionto the crossing of two distinctive nanowires, hetero-geneous junctions have also been constructed insidea single wire, either along the wire axis in the formof a nanowire superlattice [4.92], or perpendicular tothe wire axis by forming a core–shell structure of sil-icon and germanium [4.111]. These various nanowirejunctions not only possess the current rectifying proper-ties (Fig. 4.22) expected of bulk semiconductor devices,but they also exhibit electroluminescence (EL) that maybe interesting for optoelectronic applications, as shownin Fig. 4.39 for the electroluminescence of a crossedjunction of n- and p-type InP nanowires [4.194](Sect. 4.3.3).

PartA

4.3

Nanowires 4.3 Applications 153

In addition to the two-terminal nanowire devices,such as the p-n junctions described above, it is foundthat the conductance of a semiconductor nanowire canbe significantly modified by applying voltage at a thirdgate terminal, implying the utilization of nanowiresin field effect transistors (FETs). This gate terminalcan either be the substrate [4.30, 196–199], a separatemetal contact located close to the nanowire [4.200],or another nanowire with a thick oxide coating in thecrossed nanowire junction configuration [4.142]. Theoperating principles of these nanowire-based FETs arediscussed in Sect. 4.2.2. Various logic devices perform-ing basic logic functions have been demonstrated usingnanowire junctions [4.142], as shown in Fig. 4.40 for theOR and AND logic gates constructed from 2-by-1 and1-by-3 nanowire p-n junctions, respectively. By func-tionalizing nanowires with redox-active molecules tostore charge, nanowire FETs were demonstrated withtwo-level [4.144] and with eight-level [4.201] mem-ory effects, which may be used for nonvolatile memoryor as switches. In another advance, In2O3 nanowireFETs with high-k dielectric material were demon-strated, and substantially enhanced performance wasobtained due to the highly efficient coupling of thegate [4.202]. A vertical FET with a surrounding gategeometry has also been demonstrated, which has thepotential for high-density nanoscale memory and logicdevices [4.203].

V0(V)

5

4

3

2

1

0OR address level

5

4

3

2

1

V0(V)

AND address level

543210

54321

V0(V)

V0 (V)

Vi (V)

Vi(V)0 1 2 3 4 5

0 1 2 3 4 5

a) b)

11

100100

00

01 10 11

Vi1 Vi2

V0

np

Vi1 Vi2

V0

Vi1 Vi2

V0

Silicon oxide

Silicon oxide

np Vi1 Vi2

V0

R

Vc1

V0 Vi2Vi1 Vc2

Vc1

V Vi2Vi1 Vc2

Vc1

OR

AND

c) d)

Fig. 4.40a–d Nanowire logic gates:(a) Schematic of logic OR gate con-structed from a 2 (p-Si) by 1 (n-GaN)crossed nanowire junction. The in-set shows the SEM image (scalebar: 1 μm) of an assembled OR gateand the symbolic electronic circuit.(b) The output voltage of the cir-cuit in (a) versus the four possiblelogic address level inputs: (0,0); (0,1);(1,0); (1,1), where logic 0 input is 0 Vand logic 1 is 5 V (same for below).(c) Schematic of logic AND gate con-structed from a 1 (p-Si) by 3 (n-GaN)crossed nanowire junction. The in-set shows the SEM image (scale bar:1 μm) of an assembled AND gateand the symbolic electronic circuit.(d) The output voltage of the circuitin (c) versus the four possible logicaddress level inputs (after [4.142])

Intensity (counts)

1.6

1.2

0.8

0.4

01.6 2 2.4 2.8 3.2

Forward bias (V)

a) b)

5 μm

5 μm

4

3

2

1

0– 2 0 2

Current (μA)

Voltage (V)

Fig. 4.39a,b Optoelectrical characterization of a crossed nanowirejunction formed between 65 nm n-type and 68 nm p-type InPnanowires. (a) Electroluminescence (EL) image of the light emit-ted from a forward-biased nanowire p-n junction at 2.5 V. Inset,photoluminescence (PL) image of the junction. (b) EL intensity asa function of operation voltage. Inset, the SEM image and the I–Vcharacteristics of the junction (after [4.194]). The scale bar in theinset is 5 μm

Nanowires have also been proposed for applicationsassociated with electron field emission [4.204], such asflat panel displays, because of their small diameter andlarge curvature at the nanowire tip, which may reducethe threshold voltage for electron emission [4.205]. In

PartA

4.3


this regard, the demonstration of very high field emis-sion currents from the sharp tip (≈ 10 nm radius) of a Sicone [4.204], from carbon nanotubes [4.206], from Sinanowires inside a carbon nanotube [4.207], and fromCo nanowires [4.208], has stimulated interest in thispotential area of application for nanowires.

The concept of constructing electronic devicesbased on nanowires has already been demonstrated,and the next step for electronic applications would beto devise a feasible method for integration and massproduction. We expect that, in order to maintain thegrowing rate of device density and functionality in theexisting electronic industry, new kinds of complemen-tary electronic devices will emerge from this bottom-upscheme for nanowire electronics, different from whathas been produced by the traditional top-down approachpursued by conventional electronics.

4.3.2 Thermoelectric Applications

One proposed application for nanowires is for thermo-electric cooling and for the conversion between thermaland electrical energy [4.171, 209]. The efficiency ofa thermoelectric device is measured in terms of a di-mensionless figure of merit ZT , where Z is defined as

Z = σ S2

κ, (4.2)

where σ is the electrical conductivity, S is the See-beck coefficient, κ is the thermal conductivity, and Tis the temperature. In order to achieve a high ZT andtherefore efficient thermoelectric performance, a highelectrical conductivity, a hugh Seebeck coefficient anda low thermal conductivity are required. In 3-D systems,the electronic contribution to κ is proportional to σ inaccordance with the Wiedemann–Franz law, and nor-mally materials with high S have a low σ . Hence anincrease in the electrical conductivity (for example byelectron donor doping) results in an adverse variation inboth the Seebeck coefficient (decreasing) and the ther-mal conductivity (increasing). These two trade-offs setthe upper limit for increasing ZT in bulk materials, withthe maximum ZT remaining ≈ 1 at room temperaturefor the 1960–1995 time frame.

The high electronic density of states in quantum-confined structures is proposed as a promising possibil-ity to bypass the Seebeck/electrical conductivity trade-off and to control each thermoelectric-related variableindependently, thereby allowing for increased electri-cal conductivity, relatively low thermal conductivity,and a large Seebeck coefficient simultaneously [4.210].

For example, Figs. 4.29 and 4.30a in Sect. 4.2.3 showan enhanced in S for bismuth and bismuth-antimonynanowires as the wire diameter decreases. In addition toalleviating the undesired connections between σ , S andthe electronic contribution to the thermal conductivity,nanowires also have the advantage that the phonon con-tribution to the thermal conductivity is greatly reducedbecause of boundary scattering (Sect. 4.2), therebyachieving a high ZT . Figure 4.41a shows the theoret-ical values for ZT versus sample size for both bismuththin films (2-D) and nanowires (1-D) in the quantum-confined regime, exhibiting a rapidly increasing ZT asthe quantum size effect becomes more and more im-portant [4.210]. In addition, the quantum size effect innanowires can be combined with other parameters totailor the band structure and electronic transport behav-ior (for instance, Sb alloying in Bi) to further optimizeZT . For example, Fig. 4.41b shows the predicted ZTfor p-type Bi1−xSbx alloy nanowires as a function ofwire diameter and Sb content x [4.211]. The occurrenceof a local ZT maxima in the vicinity of x ≈ 0.13 anddW ≈ 45 nm is due to the coalescence of ten valencebands in the nanowire and the resulting unusual highdensity of states for holes, which is a phenomenon ab-sent in bulk Bi1−xSbx alloys. For nanowires with verysmall diameters, it is speculated that localization effectswill eventually limit the enhancement of ZT . However,in bismuth nanowires, localization effects are not signif-icant for wires with diameters larger than 9 nm [4.52].In addition to 1-D nanowires, ZT values as high as≈ 2 have also been experimentally demonstrated inmacroscopic samples containing PbSe quantum dots (0-D) [4.212] and stacked 2-D films [4.167].

Although the application of nanowires to thermo-electrics appears very promising, these materials arestill in the research phase of the development cycle andare far from being commercialized. One challenge forthermoelectric devices based on nanowires lies in find-ing a suitable host material that will not reduce ZTtoo much due to the unwanted heat conduction throughthe host material. Therefore, the host material shouldhave a low thermal conductivity and occupy a volumepercentage in the composite material that is as low aspossible, while still providing the quantum confinementand the support for the nanowires.

4.3.3 Optical Applications

Nanowires also hold promise for optical applications.One-dimensional systems exhibit a singularity in theirjoint density of states, allowing quantum effects in

PartA

4.3


3

2

1

0

100

90

80

70

60

50

40

30

20

10

Antimony content (at %)0 10 20 30 40 50 60

dW (nm)

a) ZT b) Wire diameter (nm)

0.25 0.5 0.75

1

0.25

1.2

1

0.75 0.5

1-D

2-D

Bi at 77 KTrigonal direction

0 5 10 15 20 25 30

Fig. 4.41 (a) Calculated ZT of 1-D (nanowire) and 2-D (quantum well) bismuth systems at 77 K as a function of dW,denoting the wire diameter or film thickness. The thermoelectric performance (ZT ) is expected to improve greatly whenthe wire diameter is small enough for the nanowire to become a one-dimensional system. (b) Contour plot of optimal ZTvalues for p-type Bi1−xSbx nanowires versus wire diameter and antimony concentration calculated at 77 K (after [4.211])

nanowires to be optically observable, sometimes evenat room temperature. Since the density of states ofa nanowire in the quantum limit (small wire diameter)is highly localized in energy, the available states quicklyfill up with electrons as the intensity of the incident lightis increased. This filling up of the subbands, as wellas other effects that are unique to low-dimensional ma-terials, lead to strong optical nonlinearities in quantumwires. Quantum wires may thus yield optical switcheswith a lower switching energy and increased switchingspeed compared to currently available optical switches.

Light emission from nanowires can be achieved byphotoluminescence (PL) or electroluminescence (EL),distinguished by whether the electronic excitation isachieved by optical illumination or by electrical stim-ulation across a p-n junction, respectively. PL is oftenused for optical property characterization, as describedin Sect. 4.2.4, but from an applications point of view, ELis a more convenient excitation method. Light-emittingdiodes (LEDs) have been achieved in junctions betweena p-type and an n-type nanowire (Fig. 4.39) [4.194]and in superlattice nanowires with p-type and n-typesegments [4.92]. The light emission was localizedto the junction area, and was polarized in the su-perlattice nanowire. An electrically driven laser wasfabricated from CdS nanowires. The wires were as-sembled by evaporating a metal contact onto an n-typeCdS nanowire which resided on a p+ silicon wafer. Thecleaved ends of the wire formed the laser cavity, so

that in forward bias, light characteristic of lasing wasobserved at the end of the wire [4.213]. LEDs havealso been achieved with core–shell structured nanowiresmade of n-GaN/InGaN/p-GaN [4.214].

Light emission from quantum wire p-n junctionsis especially interesting for laser applications, becausequantum wires can form lasers with lower excitationthresholds than their bulk counterparts and they alsoexhibit decreased sensitivity of performance to tempera-ture [4.215]. Furthermore, the emission wavelength canbe tuned for a given material composition by simplyaltering the geometry of the wire.

Lasing action has been reported in ZnO nanowireswith wire diameters that are much smaller than thewavelength of the light emitted (λ = 385 nm) [4.122](Fig. 4.42). Since the edges and lateral surfaces of ZnOnanowires are faceted (Sect. 4.2.1), they form opticalcavities that sustain desired cavity modes. Comparedto conventional semiconductor lasers, the exciton laseraction employed in zinc oxide nanowire lasers ex-hibits a lower lasing threshold (≈ 40 kW/cm2) thantheir 3-D counterparts (≈ 300 kW/cm2). In order to uti-lize exciton confinement effects in the lasing action,the exciton binding energy (≈ 60 meV in ZnO) mustbe greater than the thermal energy (≈ 26 meV at 300 K).Decreasing the wire diameter increases the excitationbinding energy and lowers the threshold for lasing. PLNSOM imaging confirmed the waveguiding propertiesof the anisotropic and the well-faceted structure of ZnO

PartA

4.3


Wavelength (nm)370 380 390 400


b

a

Excitation

UV laser output

Wavelength (nm)


380 390 400

Fig. 4.42 A schematic of lasing in ZnO nanowires and thePL spectra of ZnO nanowires at two excitation intensities.One PL spectrum is taken below the lasing threshold, andthe other above it (after [4.122])

nanowires, limiting the emission to the tips of the ZnOnanowires [4.183]. Time-resolved studies have illumi-nated the dynamics of the emission process [4.216].

Lasing was also observed in ZnS nanowires in an-odic aluminum oxide templates [4.217] and in GaNnanowires [4.218]. Unlike ZnO, GaN has a small exci-ton binding energy, only ≈ 25 meV. Furthermore, sincethe wire radii used in this study (15–75 nm) [4.218] arelarger than the Bohr radius of excitons in GaN (11 nm),the exciton binding energy is not expected to increase inthese GaN wires and quantum confinement effects suchas those shown in Fig. 4.35 for InP are not expected.However, some tunability of the center of the spectralintensity was achieved by increasing the intensity of thepump power, causing a redshift in the laser emission,which is explained as a bandgap renormalization as a re-

sult of the formation of an electron–hole plasma. Heat-ing effects were excluded as the source of the spectralshift. GaN quantum wire UV lasers with a low thresholdfor lasing action have been achieved using a self-organized GaN(core)/AlGaN(shell) structure [4.219].

Nanowires have also been demonstrated to havegood waveguiding properties. Quantitative studies ofcadmium sulfide (CdS) nanowire structures show thatlight propagation takes place with only moderate lossesthrough sharp and even acute angle bends. In addition,active devices made with nanowires have shown that ef-ficient injection into and modulation of light throughnanowire waveguides can be achieved [4.220]. By link-ing ZnO nanowire light sources to SnO2 waveguides,the possibility of optical integrated circuitry is intro-duced [4.221].

Nanowire photodetectors were also proposed. ZnOnanowires were found to display a strong photocurrentresponse to UV light irradiation [4.222]. The conduc-tivity of the nanowire increased by four orders of mag-nitude compared to the dark state. The response of thenanowire was reversible, and selective to photon ener-gies above the bandgap, suggesting that ZnO nanowirescould be a good candidate for optoelectronic switches.

Nanowires have been also proposed for another typeof optical switching. Light with its electric field nor-mal to the wire axis excites a transverse free carrierresonance inside the wire, while light with its elec-tric field parallel to the wire axis excites a longitudinalfree carrier resonance inside the wire. Since nanowiresare highly anisotropic, these two resonances occur attwo different wavelengths and thus result in absorp-tion peaks at two different energies. Gold nanowiresdispersed in an aqueous solution align along the elec-tric field when a DC voltage is applied. The energy ofthe absorption peak can be toggled between the trans-verse and longitudinal resonance energies by changingthe alignment of the nanowires under polarized lightillumination using an electric field [4.223, 224]. Thus,electro-optical modulation is achieved.

Nanowires may also be used as barcode tags for op-tical read-out. Nanowires containing gold, silver, nickel,palladium, and platinum were fabricated [4.110] byelectrochemical filling of porous anodic alumina, sothat each nanowire consisted of segments of variousmetal constituents. Thus many types of nanowires canbe made from a handful of materials, and identifiedby the order of the metal segments along their mainaxis, and the length of each segment. Barcode read-out is possible by reflectance optical microscopy. Thesegment length is limited by the Rayleigh diffraction

PartA

4.3


12

9

6

0

Parcticle length

a) b)

1 µm

Fig. 4.43 (a) An optical image ofmany short bar-coded Au-Ag-Au-Auwires and (b) an FE-SEM image of anAu/Ag barcoded wire with multiplestrips of varying length. The insert in(a) shows a histogram of the particlelengths for 106 particles in this image(after [4.110])

limit, and not by synthesis limitations, and thus can beas small as 145 nm. Figure 4.43a shows an optical im-age of many Au-Ag-Au-Ag barcoded wires, where thesilver segments show higher reflectivity. Figure 4.43bis a backscattering mode FE-SEM image of a singlenanowire, highlighting the composition and segmentlength variations along the nanowire.

Both the large surface area and the high conductiv-ity along the length of a nanowire are favorable for itsuse in inorganic–organic solar cells [4.225], which offerpromise from a manufacturing and cost-effectivenessstandpoint. In a hybrid nanocrystal–organic solar cell,the incident light forms bound electron–hole pairs (ex-citons) in both the inorganic nanocrystal and in thesurrounding organic medium. These excitons diffuse tothe inorganic–organic interface and disassociate to forman electron and a hole. Since conjugated polymers usu-ally have poor electron mobilities, the inorganic phaseis chosen to have a higher electron affinity than the or-ganic phase so that the organic phase carries the holesand the semiconductor carries the electrons. The sepa-rated electrons and holes drift to the external electrodesthrough the inorganic and organic materials, respec-tively. However, only those excitons formed within anexciton diffusion length from an interface can disas-sociate before recombining, and therefore the distancebetween the dissociation sites limits the efficiency ofa solar cell. A solar cell prepared from a compositeof CdSe nanorods inside poly(3-ethylthiophene) [4.225]yielded monochromatic power efficiencies of 6.9% andpower conversion efficiencies of 1.7% under A.M. 1.5illumination (equal to solar irradiance through 1.5 timesthe air mass of the Earth at direct normal incidence).The nanorods provide a large surface area with goodchemical bonding to the polymer for efficient chargetransfer and exciton dissociation. Furthermore, theyprovide a good conduction path for the electrons toreach the electrode. Their enhanced absorption coeffi-

cient and their tunable bandgap are also characteristicsthat can be used to enhance the energy conversion effi-ciency of solar cells.

4.3.4 Chemical and BiochemicalSensing Devices

Sensors for chemical and biochemical substances withnanowires as the sensing probe are a very attrac-tive application area. Nanowire sensors will potentiallybe smaller, more sensitive, demand less power, andreact faster than their macroscopic counterparts. Ar-rays of nanowire sensors could, in principle, achievenanometer-scale spatial resolution and therefore pro-vide accurate real-time information regarding not onlythe concentration of a specific analyte but also its spa-tial distribution. Such arrays could be very useful, forexample, for dynamic studies on the effects of chemicalgradients on biological cells. The operation of sen-sors made with nanowires, nanotubes, or nanocontactsis based mostly on the reversible change in the con-ductance of the nanostructure upon absorption of theagent to be detected, but other detection methods, suchas mechanical and optical detection, are conceptuallyplausible. The increased sensitivity and faster responsetime of nanowires are a result of the large surface-to-volume ratio and the small cross section available forconduction channels. In the bulk, on the other hand,the abundance of charges can effectively shield externalfields, and the abundance of material can afford manyalternative conduction channels. Therefore, a strongerchemical stimulus and longer response time are neces-sary to observe changes in the physical properties ofa 3-D sensor in comparison to a nanowire.

It is often necessary to modify the surface of thenanowires to achieve a strong interaction with the an-alytes that need to be detected. Surface modificationsutilize the self-assembly, chemisorption or chemical re-

PartA

4.3


activity of selected organic molecules and polymerstowards metal and oxide surfaces. Examples include:thiols on gold, isocyanides on platinum, and siloxaneson silica. These surface coatings regulate the bindingand chemical reactivity of other molecules towards thenanowire in a predictable manner [4.226].

Cui et al. placed silicon nanowires made by the VLSmethod (Sect. 4.1.2) between two metal electrodes andmodified the silicon oxide coating of the wire throughthe addition of molecules that are sensitive to the ana-lyte to be detected [4.227]. For example, a pH sensorwas made by covalently linking an amine-containingsilane to the surface of the nanowire. Variations in thepH of the solution into which the nanowire was im-mersed caused protonation and deprotonation of the−NH2 and the −SiOH groups on the surface of thenanowire. The variation in surface charge density regu-lates the conductance of the nanowire; due to the p-typecharacteristics of a silicon wire, the conductance in-creases with the addition of negative surface charge.The combined acid and base behavior of the surfacegroups results in an approximately linear dependenceof the conductance on pH in the pH range 2 to 9, thusleading to a direct readout pH meter. This same typeof approach was used for the detection of the bindingof biomolecules, such as streptavidin using biotin-modified nanowires (Fig. 4.44). This nanowire-baseddevice has high sensitivity and could detect streptavidinbinding down to a concentration of 10 pM (10−12 mol).Subsequent results demonstrated the capabilities ofthese functionalized Si nanowire sensors as DNA sen-sors down to the femtomolar range [4.228]. The chem-

Time (s)

1600

1500

1400

1300

1200

Conductance (nS)

0 200 400

b)a)

SiNW

SiNW

1

2

3

Fig. 4.44 (a) Streptavidin molecules bind to a silicon nanowirefunctionalized with biotin. The binding of streptavidin to biotincauses the nanowire to change its resistance. (b) The conductanceof a biotin-modified silicon nanowire exposed to streptavidin ina buffer solution (regions 1 and 3) and with the introduction of a so-lution of antibiotin monoclonal antibody (region 2) (after [4.227])

ical detection devices were made in a field effect tran-sistor geometry, so that the back-gate potential could beused to regulate the conductance in conjugation withthe chemical detection and to provide a real-time di-rect read-out [4.227]. The extension of this device todetect multiple analytes using multiple nanowires, eachsensitized to a different analyte, could provide for fast,sensitive, and in situ screening procedures.

A similar approach was used by Favier et al., whomade a nanosensor for the detection of hydrogen fromof an array of palladium nanowires between two metalcontacts [4.44]. They demonstrated that nanogaps werepresent in their nanowire structure, and upon absorp-tion of H2 and formation of Pd hydride, the nanogapstructure would close and improve the electrical con-tact, thereby increasing the conductance of the nanowirearray. The response time of these sensors was 75 ms,and they could operate in the range 0.5–5% H2 beforesaturation occurred.

4.3.5 Magnetic Applications

It has been demonstrated that arrays of single-domainmagnetic nanowires can be prepared with controllednanowire diameter and length, aligned along a commondirection and arranged in a close-packed ordered array(Sect. 4.1), and that the magnetic properties (coercivity,remanence and dipolar magnetic interwire interaction)can be controlled to achieve a variety of magnetic ap-plications [4.40, 79].

The most interesting of these applications is formagnetic storage, where the large nanowire aspect ra-tio (length/diameter) is advantageous for preventingthe onset of the superparamagnetic limit at which themagnetization direction in the magnetic grains can bereversed by the thermal energy kBT , thereby resulting inloss of recorded data in the magnetic recording medium.The magnetic energy in a grain can be increased byincreasing either the volume or the anisotropy of thegrain. If the volume is increased, the particle size in-creases, so the resolution is decreased. For sphericalmagnetized grains, the superparamagnetic limit at roomtemperature is reached at 70 Gbit/in2. In nanowires,the anisotropy is very large and yet the wire diam-eters are small, so that the magnetostatic switchingenergy can easily be above the thermal energy whilethe spatial resolution is large. For magnetic data stor-age applications, a large aspect ratio is needed for thenanowires in order to maintain a high coercivity, anda sufficient separation between nanowires is neededto suppress interwire magnetic dipolar coupling. Thus

PartA

4.3

Nanowires References 159

nanowires can form stable and highly dense magneticmemory arrays with packing densities in excess of1011 wires/cm2.

The onset of superparamagnetism can be pre-vented in the single-domain magnetic nanowire ar-rays that have already been fabricated using eitherporous alumina templates to make Ni nanowires with

35 nm diameters [4.40] or diblock copolymer tem-plates [4.79] to make Co nanowires, with meandiameters of 14 nm and 100% filling of the templatepores (Sect. 4.1.1). The ordered magnetic nanowire ar-rays that have already been demonstrated offer theexciting promise of systems permitting 1012 bits/in2

data storage.

4.4 Concluding Remarks

In this chapter, we reviewed the synthesis, character-ization and physical properties of nanowires, placingparticular emphasis on nanowire properties that differfrom those of the bulk counterparts and potential appli-cations that might result from the special structures andproperties of nanowires.

We have shown that the newly emerging field ofnanowire research has developed very rapidly over thepast few years, driven by the development of a vari-ety of complementary nanowire synthesis methods andeffective tools for measuring nanowire structure andproperties (Sects. 4.1 and 4.2). At present, much of theprogress is at the demonstration-of-concept level, withmany gaps in knowledge remaining to be elucidated,theoretical models to be developed, and new nanowiresystems to be explored. Having demonstrated that manyof the most interesting discoveries to date relate to

nanowire properties not present in their bulk materialcounterparts, we can expect future research empha-sis to be increasingly focused on smaller diameternanowires, where new unexplored physical phenom-ena related to quantum confinement effects are morelikely to be found. We can also expect the develop-ment of applications to follow, some coming sooner andothers later. Many promising applications are now atthe early demonstration stage (Sect. 4.3), but are mov-ing ahead rapidly because of their promise of newfunctionality, not previously available, in the fields ofelectronics, optoelectronics, biotechnology, magnetics,and energy conversion and generation, among others.Many exciting challenges remain in advancing both thenanoscience and the nanotechnological promise alreadydemonstrated by the nanowire research described in thisreview.

References

4.1 R.P. Feynman: There’s plenty of room at the bottom,Caltech Eng. Sci. 23, 22 (1960)

4.2 Y. Mao, S.S. Wong: General, room-temperaturemethod for the synthesis of isolated as well as arraysof single-crystalline ABO4-type nanorods, J. Am.Chem. Soc. 126, 15245–15252 (2004)

4.3 E. Braun, Y. Eichen, U. Sivan, G. Ben-Yoseph: DNA-templated assembly and electrode attachment ofa conducting silver wire, Nature 391, 775–778 (1998)

4.4 G. Sauer, G. Brehm, S. Schneider, K. Nielsch,R.B. Wehrspohn, J. Choi, H. Hofmeister, U. Gösele:Highly ordered monocrystalline silver nanowire ar-rays, J. Appl. Phys. 91, 3243–3247 (2002)

4.5 G.L. Hornyak, C.J. Patrissi, C.M. Martin: Fabrica-tion, characterization and optical properties of goldnanoparticle/porous alumina composites: The non-scattering Maxwell–Garnett limit, J. Phys. Chem. B101, 1548–1555 (1997)

4.6 X.Y. Zhang, L.D. Zhang, Y. Lei, L.X. Zhao, Y.Q. Mao:Fabrication and characterization of highly ordered

Au nanowire arrays, J. Mater. Chem. 11, 1732–1734(2001)

4.7 Y.-T. Cheng, A.M. Weiner, C.A. Wong, M.P. Balogh,M.J. Lukitsch: Stress-induced growth of bis-muth nanowires, Appl. Phys. Lett. 81, 3248–3250(2002)

4.8 J. Heremans, C.M. Thrush, Y.-M. Lin, S. Cronin,Z. Zhang, M.S. Dresselhaus, J.F. Mansfield: Bismuthnanowire arrays: Synthesis, galvanomagnetic prop-erties, Phys. Rev. B 61, 2921–2930 (2000)

4.9 L. Piraux, S. Dubois, J.L. Duvail, A. Radulescu,S. Demoustier-Champagne, E. Ferain, R. Legras:Fabrication and properties of organic, metalnanocylinders in nanoporous membranes, J. Mater.Res. 14, 3042–3050 (1999)

4.10 K. Hong, F.Y. Yang, K. Liu, D.H. Reich, P.C. Sear-son, C.L. Chien, F.F. Balakirev, G.S. Boebinger: Giantpositive magnetoresistance of Bi nanowire arrays inhigh magnetic fields, J. Appl. Phys. 85, 6184–6186(1999)

PartA

4


4.11 A.J. Yin, J. Li, W. Jian, A.J. Bennett, J.M. Xu: Fab-rication of highly ordered metallic nanowire arraysby electrodeposition, Appl. Phys. Lett. 79, 1039–1041(2001)

4.12 Z. Zhang, J.Y. Ying, M.S. Dresselhaus: Bismuthquantum-wire arrays fabricated by a vacuum melt-ing and pressure injection process, J. Mater. Res. 13,1745–1748 (1998)

4.13 Z. Zhang, D. Gekhtman, M.S. Dresselhaus, J.Y. Ying:Processing and characterization of single-crystallineultrafine bismuth nanowires, Chem. Mater. 11, 1659–1665 (1999)

4.14 T.E. Huber, M.J. Graf, P. Constant: Processing andcharacterization of high-conductance bismuth wirearray composites, J. Mater. Res. 15, 1816–1821 (2000)

4.15 L. Li, G. Li, Y. Zhang, Y. Yang, L. Zhang: Pulsedelectrodeposition of large-area, ordered Bi1−xSbx

nanowire arrays from aqueous solutions, J. Phys.Chem. B 108, 19380–19383 (2004)

4.16 M.S. Sander, A.L. Prieto, R. Gronsky, T. Sands,A.M. Stacy: Fabrication of high-density, high aspectratio, large-area bismuth telluride nanowire arraysby electrodeposition into porous anodic aluminatemplates, Adv. Mater. 14, 665–667 (2002)

4.17 M. Chen, Y. Xie, J. Lu, Y.J. Xiong, S.Y. Zhang, Y.T. Qian,X.M. Liu: Synthesis of rod-, twinrod-, and tetrapod-shaped CdS nanocrystals using a highly orientedsolvothermal recrystallization technique, J. Mater.Chem. 12, 748–753 (2002)

4.18 D. Xu, Y. Xu, D. Chen, G. Guo, L. Gui, Y. Tang:Preparation of CdS single-crystal nanowires by elec-trochemically induced deposition, Adv. Mater. 12,520–522 (2000)

4.19 D. Routkevitch, T. Bigioni, M. Moskovits, J.M. Xu:Electrochemical fabrication of CdS nanowire ar-rays in porous anodic aluminum oxide templates,J. Phys. Chem. 100, 14037–14047 (1996)

4.20 L. Manna, E.C. Scher, A.P. Alivisatos: Synthesis of sol-uble and processable rod-, arrow-, teardrop-, andtetrapod-shaped CdSe nanocrystals, J. Am. Chem.Soc. 122, 12700–12706 (2000)

4.21 D. Routkevitch, A.A. Tager, J. Haruyama, D. Al-Mawlawi, M. Moskovits, J.M. Xu: Nonlithographicnano-wire arrays: fabrication, physics, and deviceapplications, IEEE Trans. Electron. Dev. 43, 1646–1658 (1996)

4.22 D.S. Xu, D.P. Chen, Y.J. Xu, X.S. Shi, G.L. Guo, L.L. Gui,Y.Q. Tang: Preparation of II–VI group semiconductornanowire arrays by dc electrochemical depositionin porous aluminum oxide templates, Pure Appl.Chem. 72, 127–135 (2000)

4.23 R. Adelung, F. Ernst, A. Scott, M. Tabib-Azar,L. Kipp, M. Skibowski, S. Hollensteiner, E. Spiecker,W. Jäger, S. Gunst, A. Klein, W. Jägermann,V. Zaporojtchenko, F. Faupel: Self-assemblednanowire networks by deposition of copper ontolayered-crystal surfaces, Adv. Mater. 14, 1056–1061(2002)

4.24 T. Gao, G.W. Meng, J. Zhang, Y.W. Wang, C.H. Liang,J.C. Fan, L.D. Zhang: Template synthesis of single-crystal Cu nanowire arrays by electrodeposition,Appl. Phys. A 73, 251–254 (2001)

4.25 D. Al-Mawlawi, N. Coombs, M. Moskovits: Magnetic-properties of Fe deposited into anodic aluminum-oxide pores as a function of particle-size, J. Appl.Phys. 70, 4421–4425 (1991)

4.26 F. Li, R.M. Metzger: Activation volume of α-Fe par-ticles in alumite films, J. Appl. Phys. 81, 3806–3808(1997)

4.27 A. Sugawara, T. Coyle, G.G. Hembree, M.R. Schein-fein: Self-organized Fe nanowire arrays preparedby shadow deposition on NaCl(110) templates, Appl.Phys. Lett. 70, 1043–1045 (1997)

4.28 G.S. Cheng, L.D. Zhang, Y. Zhu, G.T. Fei, L. Li, C.M. Mo,Y.Q. Mao: Large-scale synthesis of single crystallinegallium nitride nanowires, Appl. Phys. Lett. 75,2455–2457 (1999)

4.29 G.S. Cheng, L.D. Zhang, S.H. Chen, Y. Li, L. Li,X.G. Zhu, Y. Zhu, G.T. Fei, Y.Q. Mao: Ordered nano-structure of single-crystalline GaN nanowires ina honeycomb structure of anodic alumina, J. Mater.Res. 15, 347–350 (2000)

4.30 Y. Huang, X. Duan, Y. Cui, C.M. Lieber: Gallium nitridenanowire nanodevices, Nano Lett. 2, 101–104 (2002)

4.31 X. Duan, C.M. Lieber: Laser-assisted catalytic growthof single crystal GaN nanowires, J. Am. Chem. Soc.122, 188–189 (2000)

4.32 A.D. Berry, R.J. Tonucci, M. Fatemi: Fabrication ofGaAs, InAs wires in nanochannel glass, Appl. Phys.Lett. 69, 2846–2848 (1996)

4.33 J.R. Heath, F.K. LeGoues: A liquid solution synthesisof single-crystal germanium quantum wires, Chem.Phys. Lett. 208, 263–268 (1993)

4.34 Y. Wu, P. Yang: Germanium nanowire growth viasimple vapor transport, Chem. Mater. 12, 605–607(2000)

4.35 Y.F. Zhang, Y.H. Tang, N. Wang, C.S. Lee, I. Bello,S.T. Lee: Germanium nanowires sheathed with anoxide layer, Phys. Rev. B 61, 4518–4521 (2000)

4.36 S.J. May, J.-G. Zheng, B.W. Wessels, L.J. Lauhon:Dendritic nanowire growth mediated by a self-assembled catalyst, Adv. Mater. 17, 598–602(2005)

4.37 S. Han, C. Li, Z. Liu, B. Lei, D. Zhang, W. Jin, X. Liu,T. Tang, C. Zhou: Transition metal oxide core-shellnanowires: Generic synthesis and transport studies,Nano Lett. 4, 1241–1246 (2004)

4.38 M.P. Zach, K.H. Ng, R.M. Penner: Molybdenumnanowires by electrodeposition, Science 290, 2120–2123 (2000)

4.39 L. Sun, P.C. Searson, L. Chien: Electrochemical de-position of nickel nanowire arrays in single-crystalmica films, Appl. Phys. Lett. 74, 2803–2805 (1999)

4.40 K. Nielsch, R. Wehrspohn, S. Fischer, H. Kronmüller,J. Barthel, J. Kirschner, U. Gösele: Magnetic proper-ties of 100 nm nickel nanowire arrays obtained from

PartA

4


ordered porous alumina templates, MRS Symp. Proc.636, D1.9–1–D1.9–6 (2001)

4.41 Y. Wang, X. Jiang, T. Herricks, Y. Xia: Single crystallinenanowires of lead: Large-scale synthesis, mecha-nistic studies, and transport measurements, J. Phys.Chem. B 108, 8631–8640 (2004)

4.42 E. Lifshitz, M. Bashouti, V. Kloper, A. Kigel,M.S. Eisen, S. Berger: Synthesis and characteriza-tion of PbSe quantum wires, multipods, quantumrods, cubes, Nano Lett. 3, 857–862 (2003)

4.43 W. Lu, P. Gao, W.B. Jian, Z.L. Wang, J. Fang: Perfectorientation ordered in-situ one-dimensional self-assembly of Mn-doped PbSe nanocrystals, J. Am.Chem. Soc. 126, 14816–14821 (2004)

4.44 F. Favier, E.C. Walter, M.P. Zach, T. Benter,R.M. Penner: Hydrogen sensors and switches fromelectrodeposited palladium mesowire arrays, Sci-ence 293, 2227–2231 (2001)

4.45 B. Gates, B. Mayers, B. Cattle, Y. Xia: Synthesis,characterization of uniform nanowires of trigonalselenium, Adv. Funct. Mater. 12, 219–227 (2002)

4.46 C.A. Huber, T.E. Huber, M. Sadoqi, J.A. Lubin, S. Man-alis, C.B. Prater: Nanowire array composites, Science263, 800–802 (1994)

4.47 Y. Cui, L.J. Lauhon, M.S. Gudiksen, J. Wang,C.M. Lieber: Diameter-controlled synthesis of sin-gle crystal silicon nanowires, Appl. Phys. Lett. 78,2214–2216 (2001)

4.48 A.M. Morales, C.M. Lieber: A laser ablation methodfor the synthesis of crystalline semiconductornanowires, Science 279, 208–211 (1998)

4.49 N. Wang, Y.F. Zhang, Y.H. Tang, C.S. Lee, S.T. Lee:SiO2-enhanced synthesis of Si nanowires by laserablation, Appl. Phys. Lett. 73, 3902–3904 (1998)

4.50 M.K. Sunkara, S. Sharma, R. Miranda, G. Lian,E.C. Dickey: Bulk synthesis of silicon nanowires us-ing a low-temperature vapor-liquid-solid method,Appl. Phys. Lett. 79, 1546–1548 (2001)

4.51 S. Vaddiraju, H. Chandrasekaran, M.K. Sunkara: Va-por phase synthesis of tungsten nanowires, J. Am.Chem. Soc. 125, 10792–10793 (2003)

4.52 J.P. Heremans, C.M. Thrush, D.T. Morelli, M.-C. Wu:Thermoelectric power of bismuth nanocomposites,Phys. Rev. Lett. 88, 216801–1–216801–4 (2002)

4.53 Y. Li, G.S. Cheng, L.D. Zhang: Fabrication of highlyordered ZnO nanowire arrays in anodic aluminamembranes, J. Mater. Res. 15, 2305–2308 (2000)

4.54 P. Yang, H. Yan, S. Mao, R. Russo, J. Johnson,R. Saykally, N. Morris, J. Pham, R. He, H.-J. Choi:Controlled growth of ZnO nanowires and their opticalproperties, Adv. Funct. Mater. 12, 323–331 (2002)

4.55 M.J. Zheng, L.D. Zhang, G.H. Li, W.Z. Shen: Fabrica-tion and optical properties of large-scale uniformzinc oxide nanowire arrays by one-step electro-chemical deposition technique, Chem. Phys. Lett.363, 123–128 (2002)

4.56 M.S. Dresselhaus, Y.-M. Lin, O. Rabin, A. Jorio,A.G. Souza Filho, M.A. Pimenta, R. Saito, G.G. Sam-

sonidze, G. Dresselhaus: Nanowires and nanotubes,Mater. Sci. Eng. C 23, 129–140 (2003), also in: Currenttrends in nanotechnologies: From materials to sys-tems, Eur. Mater. Res. Soc. Symp. Proc., Vol. 140, ed.by W. Jantsch, H. Grimmeiss, G. Marietta (Elsevier,Amsterdam 2002)

4.57 R. Saito, G. Dresselhaus, M.S. Dresselhaus: Physi-cal Properties of Carbon Nanotubes (Imperial CollegePress, London 1998)

4.58 M.S. Dresselhaus, G. Dresselhaus, P. Avouris: Carbonnanotubes: Synthesis, structure, properties and ap-plications, Springer Ser. Top. Appl. Phys. 80, 1–447(2001)

4.59 R.C. Haddon: Special issue on carbon nanotubes,Acc. Chem. Res. 35, 997–1113 (2002)

4.60 Y.-M. Lin, X. Sun, S. Cronin, Z. Zhang, J.Y. Ying,M.S. Dresselhaus: Fabrication, transport propertiesof Te-doped bismuth nanowire arrays. In: MolecularElectronics: MRS Symposium Proceedings, Vol. 582,ed. by S.T. Pantelides, M.A. Reed, J. Murday, A. Avi-ran (Materials Research Society Press, Pittsburgh2000) pp. 1–6, Chap. H10.3

4.61 C.R. Martin: Nanomaterials: A membrane-basedsynthetic approach, Science 266, 1961–1966 (1994)

4.62 G.A. Ozin: Nanochemistry: synthesis in diminishingdimensions, Adv. Mater. 4, 612–649 (1992)

4.63 R.J. Tonucci, B.L. Justus, A.J. Campillo, C.E. Ford:Nanochannel array glass, Science 258, 783–785(1992)

4.64 J.Y. Ying: Nanoporous systems and templates, Sci.Spec. 18, 56–63 (1999)

4.65 J.W. Diggle, T.C. Downie, C.W. Goulding: Anodic ox-ide films on aluminum, Chem. Rev. 69, 365–405(1969)

4.66 J.P. O’Sullivan, G.C. Wood: The morphology andmechanism of formation of porous anodic filmson aluminum, Proc. R. Soc. Lond. A 317, 511–543(1970)

4.67 A.P. Li, F. Müller, A. Birner, K. Neilsch, U. Gösele:Hexagonal pore arrays with a 50-420nm interporedistance formed by self-organization in anodic alu-mina, J. Appl. Phys. 84, 6023–6026 (1998)

4.68 J.P. Sullivan, G.C. Wood: The morphology, mech-anism of formation of porous anodic films onaluminum, Proc. R. Soc. Lond. A 317, 511–543 (1970)

4.69 O. Jessensky, F. Müller, U. Gösele: Self-organizedformation of hexagonal pore arrays in anodic alu-mina, Appl. Phys. Lett. 72, 1173–1175 (1998)

4.70 F. Li, L. Zhang, R.M. Metzger: On the growth of highlyordered pores in anodized aluminum oxide, Chem.Mater. 10, 2470–2480 (1998)

4.71 H. Masuda, M. Satoh: Fabrication of gold nanodotarray using anodic porous alumina as an evap-oration mask, Jpn. J. Appl. Phys. 35, L126–L129(1996)

4.72 E. Ferain, R. Legras: Track-etched membrane – dy-namics of pore formation, Nucl. Instrum. Methods B84, 331–336 (1993)

PartA

4


4.73 A. Blondel, J.P. Meier, B. Doudin, J.-P. Ansermet: Gi-ant magnetoresistance of nanowires of multilayers,Appl. Phys. Lett. 65, 3019–3021 (1994)

4.74 K. Liu, C.L. Chien, P.C. Searson, Y.Z. Kui: Structuraland magneto-transport properties of electrode-posited bismuth nanowires, Appl. Phys. Lett. 73,1436–1438 (1998)

4.75 C.A. Huber, T.E. Huber: A novel microstructure:semiconductor-impregnated porous Vycor glass,J. Appl. Phys. 64, 6588–6590 (1988)

4.76 J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz,C.T. Kresge, K.D. Schmitt, C.T.-W. Chu, D.H. Ol-son, E.W. Sheppard, S.B. McCullen, J.B. Higgins,J.L. Schlenker: A new family of mesoporous mo-lecular sieves prepared with liquid crystal templates,J. Am. Chem. Soc. 114, 10834–10843 (1992)

4.77 C.-G. Wu, T. Bein: Conducting polyaniline filamentsin a mesoporous channel host, Science 264, 1757–1759 (1994)

4.78 Y.-M. Lin, S.B. Cronin, J.Y. Ying, M.S. Dresselhaus,J.P. Heremans: Transport properties of Bi nanowirearrays, Appl. Phys. Lett. 76, 3944–3946 (2000)

4.79 T. Thurn-Albrecht, J. Schotter, G.A. Kästle, N. Em-ley, T. Shibauchi, L. Krusin-Elbaum, K. Guarini,C.T. Black, M.T. Tuominen, T.P. Russell: Ultrahigh-density nanowire arrays grown in self-assembleddiblock copolymer templates, Science 290, 2126–2129(2000)

4.80 A.W. Adamson: Physical Chemistry of Surfaces (Wiley,New York 1982) p. 338

4.81 R. Ferré, K. Ounadjela, J.M. George, L. Piraux,S. Dubois: Magnetization processes in nickel andcobalt electrodeposited nanowires, Phys. Rev. B 56,14066–14075 (1997)

4.82 H. Zeng, M. Zheng, R. Skomski, D.J. Sellmyer, Y. Liu,L. Menon, S. Bandyopadhyay: Magnetic propertiesof self-assembled Co nanowires of varying lengthand diameter, J. Appl. Phys. 87, 4718–4720 (2000)

4.83 Y. Peng, H.L. Zhang, S.-L. Pan, H.-L. Li: Mag-netic properties and magnetization reversal of α-Fenanowires deposited in alumina film, J. Appl. Phys.87, 7405–7408 (2000)

4.84 L. Piraux, J.M. George, J.F. Despres, C. Leroy,E. Ferain, R. Legras, K. Ounadjela, A. Fert: Gi-ant magnetoresistance in magnetic multilayerednanowires, Appl. Phys. Lett. 65, 2484–2486 (1994)

4.85 S. Bhattacharrya, S.K. Saha, D. Chakravorty:Nanowire formation in a polymeric film, Appl. Phys.Lett. 76, 3896–3898 (2000)

4.86 G. Yi, W. Schwarzacher: Single crystal superconduc-tor nanowires by electrodeposition, Appl. Phys. Lett.74, 1746–1748 (1999)

4.87 D. Al-Mawlawi, C.Z. Liu, M. Moskovits: Nanowiresformed in anodic oxide nanotemplates, J. Mater.Res. 9, 1014–1018 (1994)

4.88 R.S. Wagner, W.C. Ellis: Vapor-liquid-solid mecha-nism of single crystal growth, Appl. Phys. Lett. 4,89–90 (1964)

4.89 Y. Wu, P. Yang: Direct observation of vapor-liquid-solid nanowire growth, J. Am. Chem. Soc. 123, 3165–3166 (2001)

4.90 Y. Wu, R. Fan, P. Yang: Block-by-block growthof single-crystalline Si/SiGe superlattice nanowires,Nano Lett. 2, 83–86 (2002)

4.91 S. Sharma, M.K. Sunkara, R. Miranda, G. Lian,E.C. Dickey: A novel low temperature synthesismethod for semiconductor nanowires. In: Syn-thesis, Functional Properties and Applications ofNanostructures: Mat. Res. Soc. Symp. Proc., SanFrancisco, Vol. 676, ed. by H.W. Hahn, D.L. Feldheim,C.P. Kubiak, R. Tannenbaum, R.W. Siegel (MaterialsResearch Society Press, Pittsburgh 2001) p. Y1.6

4.92 M.S. Gudiksen, L.J. Lauhon, J. Wang, D.C. Smith,C.M. Lieber: Growth of nanowire superlattice struc-tures for nanoscale photonics and electronics,Nature 415, 617–620 (2002)

4.93 Y. Wu, J. Xiang, C. Yang, W. Lu, C.M. Lieber: Single-crystal metallic nanowires and metal/semiconductornanowire heterostructures, Nature 430, 61–65(2004)

4.94 M.T. Björk, B.J. Ohlsson, T. Sass, A.I. Persson, C. The-lander, M.H. Magnusson, K. Deppert, L.R. Wallen-berg, L. Samuelson: One-dimensional steeplechasefor electrons realized, Nano Lett. 2, 87–89 (2002)

4.95 N. Wang, Y.H. Tang, Y.F. Zhang, C.S. Lee, S.T. Lee:Nucleation and growth of Si nanowires from siliconoxide, Phys. Rev. B 58, R16024–R16026 (1998)

4.96 Y.F. Zhang, Y.H. Tang, N. Wang, C.S. Lee, I. Bello,S.T. Lee: One-dimensional growth mechanism ofcrystalline silicon nanowires, J. Cryst. Growth 197,136–140 (1999)

4.97 S.T. Lee, Y.F. Zhang, N. Wang, Y.H. Tang, I. Bello,C.S. Lee, Y.W. Chung: Semiconductor nanowires fromoxides, J. Mater. Res. 14, 4503–4507 (1999)

4.98 D.D.D. Ma, C.S. Lee, Y. Lifshitz, S.T. Lee: Periodic ar-ray of intramolecular junctions of silicon nanowires,Appl. Phys. Lett. 81, 3233–3235 (2002)

4.99 D. Whang, S. Jin, C.M. Lieber: Large-scale hier-archical organization of nanowires for functionalnanosystems, Jpn. J. Appl. Phys. 43, 4465–4470(2004)

4.100 M.P. Zach, K. Inazu, K.H. Ng, J.C. Hemminger,R.M. Penner: Synthesis of molybdenum nanowireswith millimeter-scale lengths using electrochemicalstep edge decoration, Chem. Mater. 14, 3206–3216(2002)

4.101 B. Gates, Y. Yin, Y. Xia: A solution-phase approachto the synthesis of uniform nanowires of crystallineselenium with lateral dimensions in the range of10-30 nm, J. Am. Chem. Soc. 122, 12582–12583 (2000)

4.102 B. Mayers, B. Gates, Y. Yin, Y. Xia: Large-scale syn-thesis of monodisperse nanorods of Se/Te alloysthrough a homogeneous nucleation and solutiongrowth process, Adv. Mater. 13, 1380–1384 (2001)

4.103 B. Gates, Y. Wu, Y. Yin, P. Yang, Y. Xia: Single-crystalline nanowires of Ag2Se can be synthesized by

PartA

4


templating against nanowires of trigonal Se, J. Am.Chem. Soc. 123, 11500–11501 (2001)

4.104 B. Gates, B. Mayers, Y. Wu, Y. Sun, B. Cattle, P. Yang,Y. Xia: Synthesis and characterization of crystallineAg2Se nanowires through a template-engaged re-action at room temperature, Adv. Funct. Mater. 12,679–686 (2002)

4.105 H. Yu, P.C. Gibbons, W.E. Buhro: Bismuth, telluriumand bismuth telluride nanowires, J. Mater. Chem.14, 595–602 (2004)

4.106 X. Peng, J. Wickham, A.P. Alivisatos: Kinetics of II–VI,III–V colloidal semiconductor nanocrystal growth:’Focusing’ of size distributions, J. Am. Chem. Soc.120, 5343–5344 (1998)

4.107 J.Y. Lao, J.G. Wen, Z.F. Ren: Hierarchical ZnO nano-structures, Nano Lett. 2, 1287–1291 (2002)

4.108 J.Y. Lao, J.Y. Huang, D.Z. Wang, Z.F. Ren: ZnOnanobridges and nanonails, Nano Lett. 3, 235–238(2003)

4.109 N.A. Melosh, A. Boukai, F. Diana, B. Gerardot,A. Badolto, P.M. Petroff, J.R. Heath: Ultrahigh-density nanowire lattices and circuits, Science 300,112–115 (2003)

4.110 S.R. Nicewarner-Peña, R.G. Freeman, B.D. Reiss,L. He, D.J. Peña, I.D. Walton, R. Cromer, C.D. Keat-ing, M.J. Natan: Submicrometer metallic barcodes,Science 294, 137–141 (2001)

4.111 L.J. Lauhon, M.S. Gudiksen, D. Wang, C.M. Lieber:Epitaxial core-shell and core-multishell nanowireheterostructures, Nature 420, 57–61 (2002)

4.112 Z.L. Wang, Z.R. Dai, R.P. Gao, Z.G. Bai, J.L. Gole: Side-by-side silicon carbide-silica biaxial nanowires:Synthesis, structure and mechanical properties,Appl. Phys. Lett. 77, 3349–3351 (2000)

4.113 P. Yang, F. Kim: Langmuir–Blodgett assembly ofone-dimensional nanostructures, ChemPhysChem 3,503–506 (2002)

4.114 B. Messer, J.H. Song, P. Yang: Microchannel net-works for nanowire patterning, J. Am. Chem. Soc.122, 10232–10233 (2000)

4.115 P.A. Smith, C.D. Nordquist, T.N. Jackson, T.S. Mayer,B.R. Martin, J. Mbindyo, T.E. Mallouk: Electric-field assisted assembly and alignment of metallicnanowires, Appl. Phys. Lett. 77, 1399–1401 (2000)

4.116 S. Jin, D.M. Whang, M.C. McAlpine, R.S. Friedman,Y. Wu, C.M. Lieber: Scalable interconnection and in-tegration of nanowire devices without registration,Nano Lett. 4, 915–919 (2004)

4.117 T. Kuykendall, P.J. Pauzauskie, Y.F. Zhang, J. Gold-berger, D. Sirbuly, J. Denlinger, P.D. Yang: Crystal-lographic alignment of high-density gallium nitridenanowire arrays, Nat. Mater. 3, 524–528 (2004)

4.118 H. Masuda, H. Yamada, M. Satoh, H. Asoh, M. Nakao,T. Tamamura: Highly ordered nanochannel-arrayarchitecture in anodic alumina, Appl. Phys. Lett. 71,2770–2772 (1997)

4.119 O. Rabin, P.R. Herz, S.B. Cronin, Y.-M. Lin, A.I. Ak-inwande, M.S. Dresselhaus: Nanofabrication using

self-assembled alumina templates. In: Nonlitho-graphic and Lithographic Methods for Nanofabrica-tion: MRS Symposium Proceedings, Boston, Vol. 636,ed. by J.A. Rogers, A. Karim, L. Merhari, D. Norris,Y. Xia (Materials Research Society Press, Pittsburgh2001) pp. D4.7–1–D4.7–6

4.120 O. Rabin, P.R. Herz, Y.-M. Lin, A.I. Akinwande,S.B. Cronin, M.S. Dresselhaus: Formation of thickporous anodic alumina films and nanowire arrayson silicon wafers and glass, Adv. Funct. Mater. 13,631–638 (2003)

4.121 O. Rabin, P.R. Herz, Y.-M. Lin, S.B. Cronin, A.I. Akin-wande, M.S. Dresselhaus: Arrays of nanowires onsilicon wafers, 21st Int. Conf. Thermoelectr. Proc.ICT ’02 Long Beach (IEEE, Piscataway 2002) pp. 276–279

4.122 M.H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind,E. Weber, R. Russo, P. Yang: Room-temperatureultraviolet nanowire nanolasers, Science 292, 1897–1899 (2001)

4.123 Y.H. Tang, Y.F. Zhang, N. Wang, C.S. Lee, X.D. Han,I. Bello, S.T. Lee: Morphology of Si nanowiressynthesized by high-temperature laser ablation,J. Appl. Phys. 85, 7981–7983 (1999)

4.124 Y. Ding, Z.L. Wang: Structure analysis of nanowiresand nanobelts by transmission electron microscopy,J. Phys. Chem. B 108, 12280–12291 (2004)

4.125 S.B. Cronin, Y.-M. Lin, O. Rabin, M.R. Black,G. Dresselhaus, M.S. Dresselhaus, P.L. Gai: Bismuthnanowires for potential applications in nanoscaleelectronics technology, Microsc. Microanal. 8, 58–63(2002)

4.126 M.S. Sander, R. Gronsky, Y.-M. Lin, M.S. Dressel-haus: Plasmon excitation modes in nanowire arrays,J. Appl. Phys. 89, 2733–2736 (2001)

4.127 L. Venkataraman, C.M. Lieber: Molybdenum selenidemolecular wires as one-dimensional conductors,Phys. Rev. Lett. 83, 5334–5337 (1999)

4.128 A. Majumdar: Scanning thermal microscopy, Annu.Rev. Mater. Sci. 29, 505–585 (1999)

4.129 K.M. Unruh, T.E. Huber, C.A. Huber: Melting andfreezing behavior of indium metal in porous glasses,Phys. Rev. B 48, 9021–9027 (1993)

4.130 Y.Y. Wu, P.D. Yang: Melting and welding semi-conductor nanowires in nanotubes, Adv. Mater. 13,520–523 (2001)

4.131 P.M. Ajayan, S. Iijima: Capillarity-induced fill-ing of carbon nanotubes, Nature 361, 333–334(1993)

4.132 Y. Gao, Y. Bando: Carbon nanothermometer con-taining gallium, Nature 415, 599 (2002)

4.133 J.L. Costa-Krämer, N. Garcia, H. Olin: Conductancequantization histograms of gold nanowires at 4 K,Phys. Rev. B 55, 12910–12913 (1997)

4.134 M.E.T. Morales, A.G. Balogh, T.W. Cornelius, R. Neu-mann, C. Trautmann: Fragmentation of nanowiresdriven by Rayleigh instability, Appl. Phys. Lett. 84,5337–5339 (2004)

PartA

4


4.135 D.A. Wharam, T.J. Thornton, R. Newbury, M. Pep-per, H. Ahmed, J.E.F. Frost, D.G. Hasko, D.C. Peacock,D.A. Ritchie, G.A.C. Jones: One-dimensional trans-port and the quantization of the ballistic resistance,J. Phys. C 21, L209–L214 (1988)

4.136 B.J. van Wees, H. van Houten, C.W.J. Beenakker,J.G. Williamson, L.P. Kouvenhoven, D. van der Marel,C.T. Foxon: Quantized conductance of point contactsin a two-dimensional electron gas, Phys. Rev. Lett.60, 848–850 (1988)

4.137 C.J. Muller, J.M. van Ruitenbeek, L.J. de-Jongh: Conductance and supercurrent disconti-nuities in atomic-scale metallic constrictions ofvariable width, Phys. Rev. Lett. 69, 140–143(1992)

4.138 C.J. Muller, J.M. Krans, T.N. Todorov, M.A. Reed:Quantization effects in the conductance of metal-lic contacts at room temperature, Phys. Rev. B 53,1022–1025 (1996)

4.139 J.L. Costa-Krämer, N. Garcia, H. Olin: Conductancequantization in bismuth nanowires at 4 K, Phys. Rev.Lett. 78, 4990–4993 (1997)

4.140 C.Z. Li, H.X. He, A. Bogozi, J.S. Bunch, N.J. Tao:Molecular detection based on conductance quanti-zation of nanowires, Appl. Phys. Lett. 76, 1333–1335(2000)

4.141 J.L. Costa-Krämer, N. Garcia, P. Garcia-Mochales,P.A. Serena, M.I. Marques, A. Correia: Conductancequantization in nanowires formed between microand macroscopic metallic electrodes, Phys. Rev. B55, 5416–5424 (1997)

4.142 Y. Huang, X. Duan, Y. Cui, L.J. Lauhon, K.-H. Kim,C.M. Lieber: Logic gates and computation from as-sembled nanowire building blocks, Science 294,1313–1317 (2001)

4.143 J.-R. Kim, H. Oh, H.M. So, J.-J. Kim, J. Kim, C.J. Lee,S.C. Lyu: Schottky diodes based on a single GaNnanowire, Nanotechnology 13, 701–704 (2002)

4.144 X. Duan, Y. Huang, C.M. Lieber: Nonvolatile mem-ory and programmable logic from molecule-gatednanowires, Nano Lett. 2, 487–490 (2002)

4.145 E.C. Walter, R.M. Penner, H. Liu, K.H. Ng, M.P. Zach,F. Favier: Sensors from electrodeposited metalnanowires, Surf. Interface Anal. 34, 409–412(2002)

4.146 E.C. Walter, K.H. Ng, M.P. Zach, R.M. Penner,F. Favier: Electronic devices from electrodepositedmetal nanowires, Microelectron. Eng. 61/62, 555–561 (2002)

4.147 Y.-M. Lin, X. Sun, M.S. Dresselhaus: Theoretical in-vestigation of thermoelectric transport properties ofcylindrical Bi nanowires, Phys. Rev. B 62, 4610–4623(2000)

4.148 K. Liu, C.L. Chien, P.C. Searson: Finite-size effects inbismuth nanowires, Phys. Rev. B 58, R14681–R14684(1998)

4.149 Z. Zhang, X. Sun, M.S. Dresselhaus, J.Y. Ying, J. Here-mans: Magnetotransport investigations of ultrafine

single-crystalline bismuth nanowire arrays, Appl.Phys. Lett. 73, 1589–1591 (1998)

4.150 J. Heremans, C.M. Thrush, Z. Zhang, X. Sun,M.S. Dresselhaus, J.Y. Ying, D.T. Morelli: Magne-toresistance of bismuth nanowire arrays: A pos-sible transition from one-dimensional to three-dimensional localization, Phys. Rev. B 58, R10091–R10095 (1998)

4.151 L. Sun, P.C. Searson, C.L. Chien: Finite-size effectsin nickel nanowire arrays, Phys. Rev. B 61, R6463–R6466 (2000)

4.152 Y.-M. Lin, S.B. Cronin, O. Rabin, J.Y. Ying,M.S. Dresselhaus: Transport properties and ob-servation of semimetal-semiconductor transitionin Bi-based nanowires. In: Quantum ConfinedSemiconductor Nanostructures: MRS SymposiumProceedings, Boston, Vol. 737-C, ed. by J.M. Buriak,D.D.M. Wayner, F. Priolo, B. White, V. Klimov,L. Tsybeskov (Materials Research Society Press, Pitts-burgh 2003) p. F3.14

4.153 Y.-M. Lin, M.S. Dresselhaus: Transport propertiesof superlattice nanowires and their potential forthermoelectric applications. In: Quantum ConfinedSemiconductor Nanostructures: MRS SymposiumProceedings, Boston, Vol. 737-C, ed. by J.M. Buriak,D.D.M. Wayner, F. Priolo, B. White, V. Klimov,L. Tsybeskov (Materials Research Society Press, Pitts-burgh 2003) p. F8.18

4.154 Y.-M. Lin, O. Rabin, S.B. Cronin, J.Y. Ying, M.S. Dres-selhaus: Semimetal-semiconductor transition inBi1−xSbx alloy nanowires and their thermoelectricproperties, Appl. Phys. Lett. 81, 2403–2405 (2002)

4.155 J. Heremans, C.M. Thrush, Y.-M. Lin, S.B. Cronin,M.S. Dresselhaus: Transport properties of antimonynanowires, Phys. Rev. B 63, 085406–1–085406–8(2001)

4.156 Y.-M. Lin, S.B. Cronin, O. Rabin, J.Y. Ying, M.S. Dres-selhaus: Transport properties of Bi1−xSbx alloynanowires synthesized by pressure injection, Appl.Phys. Lett. 79, 677–679 (2001)

4.157 Z. Zhang, X. Sun, M.S. Dresselhaus, J.Y. Ying,J. Heremans: Electronic transport properties of sin-gle crystal bismuth nanowire arrays, Phys. Rev. B 61,4850–4861 (2000)

4.158 D.E. Beutler, N. Giordano: Localization and electron-electron interaction effects in thin Bi wires andfilms, Phys. Rev. B 38, 8–19 (1988)

4.159 J. Heremans, C.M. Thrush: Thermoelectric power ofbismuth nanowires, Phys. Rev. B 59, 12579–12583(1999)

4.160 L.D. Hicks, M.S. Dresselhaus: Thermoelectric figure ofmerit of a one-dimensional conductor, Phys. Rev. B47, 16631–16634 (1993)

4.161 Y.-M. Lin, S.B. Cronin, O. Rabin, J. Heremans,M.S. Dresselhaus, J.Y. Ying: Transport propertiesof Bi-related nanowire systems. In: AnisotropicNanoparticles: Synthesis, Characterization and Ap-plications: MRS Symposium Proceedings, Boston,

PartA

4


Vol. 635, ed. by S. Stranick, P.C. Searson, L.A. Lyon,C. Keating (Materials Research Society Press, Pitts-burgh 2001) pp. C4301–C4306

4.162 Y.-M. Lin, M.S. Dresselhaus: Thermoelectric prop-erties of superlattice nanowires, Phys. Rev. B 68,075304 (2003)

4.163 M.T. Björk, B.J. Ohlsson, C. Thelander, A.I. Persson,K. Deppert, L.R. Wallenberg, L. Samuelson: Nanowireresonant tunneling diodes, Appl. Phys. Lett. 81,4458–4460 (2002)

4.164 D. Li, Y. Wu, P. Kim, L. Shi, P. Yang, A. Ma-jumdar: Thermal conductivity of individual siliconnanowires, Appl. Phys. Lett. 83, 2934–2936 (2003)

4.165 T.S. Tighe, J.M. Worlock, M.L. Roukes: Direct ther-mal conductance measurements on suspendedmonocrystalline nanostructures, Appl. Phys. Lett. 70,2687–2689 (1997)

4.166 S.T. Huxtable, A.R. Abramson, C.-L. Tien, A. Ma-jumdar, C. LaBounty, X. Fan, G. Zeng, J.E. Bowers,A. Shakouri, E.T. Croke: Thermal conductivity ofSi/SiGe and SiGe/SiGe superlattices, Appl. Phys. Lett.80, 1737–1739 (2002)

4.167 R. Venkatasubramanian, E. Siivola, T. Colpitts,B. O’Quinn: Thin-film thermoelectric devices withhigh room-temperature figures of merit, Nature 413,597–602 (2001)

4.168 C. Dames, G. Chen: Modeling the thermal conduc-tivity of a SiGe segmented nanowire, 21st Int. Conf.Thermoelectr. Proc. ICT ’02, Long Beach (IEEE, Piscat-away 2002) pp. 317–320

4.169 K. Schwab, J.L. Arlett, J.M. Worlock, M.L. Roukes:Thermal conductance through discrete quantumchannels, Physica E 9, 60–68 (2001)

4.170 D. Li, Y. Wu, R. Fan, P. Yang, A. Majumdar: Thermalconductivity of Si/SiGe superlattice nanowires, Appl.Phys. Lett. 83, 3186–3188 (2003)

4.171 G. Chen, M.S. Dresselhaus, G. Dresselhaus, J.-P. Fleurial, T. Caillat: Recent developments inthermoelectric materials, Int. Mater. Rev. 48, 45–66(2003)

4.172 C. Dames, G. Chen: Theoretical phonon thermal con-ductivity of Si-Ge superlattice nanowires, J. Appl.Phys. 95, 682–693 (2004)

4.173 K. Schwab, E.A. Henriksen, J.M. Worlock, M.L. Roukes:Measurement of the quantum of thermal conduc-tance, Nature 404, 974–977 (2000)

4.174 M. Cardona: Light Scattering in Solids (Springer,Berlin Heidelberg 1982)

4.175 P.Y. Yu, M. Cardona: Fundamentals of Semiconduc-tors (Springer, Berlin Heidelberg 1995), Chap. 7

4.176 J.C.M. Garnett: Colours in metal glasses, in metallicfilms, and in metallic solutions, Philos. Trans. R. Soc.Lond. A 205, 237–288 (1906)

4.177 D.E. Aspnes: Optical properties of thin films, ThinSolid Films 89, 249–262 (1982)

4.178 U. Kreibig, L. Genzel: Optical absorption of smallmetallic particles, Surf. Sci. 156, 678–700 (1985)

4.179 M.R. Black, Y.-M. Lin, S.B. Cronin, O. Rabin,M.S. Dresselhaus: Infrared absorption in bismuthnanowires resulting from quantum confinement,Phys. Rev. B 65, 195417–1–195417–9 (2002)

4.180 M.W. Lee, H.Z. Twu, C.-C. Chen, C.H. Chen: Op-tical characterization of wurtzite gallium nitridenanowires, Appl. Phys. Lett. 79, 3693–3695 (2001)

4.181 D.M. Lyons, K.M. Ryan, M.A. Morris, J.D. Holmes:Tailoring the optical properties of silicon nanowirearrays through strain, Nano Lett. 2, 811–816 (2002)

4.182 M.S. Gudiksen, J. Wang, C.M. Lieber: Size-depentphotoluminescence from single indium phosphidenanowires, J. Phys. Chem. B 106, 4036–4039 (2002)

4.183 J.C. Johnson, H. Yan, R.D. Schaller, L.H. Haber,R.J. Saykally, P. Yang: Single nanowire lasers,J. Phys. Chem. B 105, 11387–11390 (2001)

4.184 S. Blom, L.Y. Gorelik, M. Jonson, R.I. Shekhter,A.G. Scherbakov, E.N. Bogachek, U. Landman:Magneto-optics of electronic transport in nanowires,Phys. Rev. B 58, 16305–16314 (1998)

4.185 J.P. Pierce, E.W. Plummer, J. Shen: Ferromagnetismin cobalt-iron alloy nanowire arrays on w(110), Appl.Phys. Lett. 81, 1890–1892 (2002)

4.186 S. Melle, J.L. Menendez, G. Armelles, D. Navas,M. Vazquez, K. Nielsch, R.B. Wehrsphon, U. Gösele:Magneto-optical properties of nickel nanowire ar-rays, Appl. Phys. Lett. 83, 4547–4549 (2003)

4.187 J.C. Johnson, H. Yan, R.D. Schaller, P.B. Petersen,P. Yang, R.J. Saykally: Near-field imaging of non-linear optical mixing in single zinc oxide nanowires,Nano Lett. 2, 279–283 (2002)

4.188 M.R. Black, Y.-M. Lin, S.B. Cronin, M.S. Dresselhaus:Using optical measurements to improve electronicmodels of bismuth nanowires, 21st Int. Conf. Ther-moelectr. Proc. ICT ’02, Long Beach, ed. by T. Caillat,J. Snyder (IEEE, Piscataway 2002) pp. 253–256

4.189 M.R. Black, P.L. Hagelstein, S.B. Cronin, Y.-M. Lin,M.S. Dresselhaus: Optical absorption from an indi-rect transition in bismuth nanowires, Phys. Rev. B68, 235417 (2003)

4.190 H.-L. Liu, C.-C. Chen, C.-T. Chia, C.-C. Yeh, C.-H. Chen, M.-Y. Yu, S. Keller, S.P. DenBaars: Infraredand Raman-scattering studies in single-crystallineGaN nanowires, Chem. Phys. Lett. 345, 245–251(2001)

4.191 H. Richter, Z.P. Wang, L. Ley: The one phononRaman-spectrum in microcrystalline silicon, SolidState Commun. 39, 625–629 (1981)

4.192 I.H. Campbell, P.M. Fauchet: The effects of mi-crocrystal size and shape on the one phononRaman-spectra of crystalline semiconductors, SolidState Commun. 58, 739–741 (1986)

4.193 R. Gupta, Q. Xiong, C.K. Adu, U.J. Kim, P.C. Eklund:Laser-induced Fano resonance scattering in siliconnanowires, Nano Lett. 3, 627–631 (2003)

4.194 X. Duan, Y. Huang, Y. Cui, J. Wang, C.M. Lieber:Indium phosphide nanowires as building blocks

PartA

4


for nanoscale electronic and optoelectronic devices,Nature 409, 66–69 (2001)

4.195 Y. Cui, C.M. Lieber: Functional nanoscale electronicdevices assembled using silicon nanowire buildingblocks, Science 291, 851–853 (2001)

4.196 Y. Cui, X. Duan, J. Hu, C.M. Lieber: Doping and elec-trical transport in silicon nanowires, J. Phys. Chem.B 104, 101–104 (2000)

4.197 G.F. Zheng, W. Lu, S. Jin, C.M. Lieber: Synthesisand fabrication of high-performance n-type sili-con nanowire transistors, Adv. Mater. 16, 1890–1891(2004)

4.198 J. Goldberger, D.J. Sirbuly, M. Law, P. Yang: ZnOnanowire transistors, J. Phys. Chem. B 109, 9–14(2005)

4.199 D.H. Kang, J.H. Ko, E. Bae, J. Hyun, W.J. Park,B.K. Kim, J.J. Kim, C.J. Lee: Ambient air effects onelectrical characteristics of gap nanowire transistors,J. Appl. Phys. 96, 7574–7577 (2004)

4.200 S.-W. Chung, J.-Y. Yu, J.R. Heath: Silicon nanowiredevices, Appl. Phys. Lett. 76, 2068–2070 (2000)

4.201 C. Li, W. Fan, B. Lei, D. Zhang, S. Han, T. Tang, X. Liu,Z. Liu, S. Asano, M. Meyyappan, J. Han, C. Zhou: Mul-tilevel memory based on molecular devices, Appl.Phys. Lett. 84, 1949–1951 (2004)

4.202 B. Lei, C. Li, D.Q. Zhang, Q.F. Zhou, K. Shung,C.W. Zhou: Nanowire transistors with ferroelec-tric gate dielectrics: Enhanced performance andmemory effects, Appl. Phys. Lett. 84, 4553–4555(2004)

4.203 H.T. Ng, J. Han, T. Yamada, P. Nguyen, Y.P. Chen,M. Meyyappan: Single crystal nanowire verticalsurround-gate field-effect transistor, Nano Lett. 4,1247–1252 (2004)

4.204 M. Ding, H. Kim, A.I. Akinwande: Observation of va-lence band electron emission from n-type siliconfield emitter arrays, Appl. Phys. Lett. 75, 823–825(1999)

4.205 F.C.K. Au, K.W. Wong, Y.H. Tang, Y.F. Zhang,I. Bello, S.T. Lee: Electron field emission from sil-icon nanowires, Appl. Phys. Lett. 75, 1700–1702(1999)

4.206 P.M. Ajayan, O.Z. Zhou: Applications of carbonnanotubes. In: Carbon Nanotubes: Synthesis, Struc-ture, Properties and Applications, Springer Ser.Top. Appl. Phys., Vol. 80, ed. by M.S. Dresselhaus,G. Dresselhaus, P. Avouris (Springer, Berlin Heidel-berg 2001) pp. 391–425

4.207 M. Lu, M.K. Li, Z.J. Zhang, H.L. Li: Synthesis of car-bon nanotubes/si nanowires core-sheath structurearrays and their field emission properties, Appl. Surf.Sci. 218, 196–202 (2003)

4.208 L. Vila, P. Vincent, L. Dauginet-DePra, G. Pirio,E. Minoux, L. Gangloff, S. Demoustier-Champagne,N. Sarazin, E. Ferain, R. Legras, L. Piraux, P. Legag-neux: Growth and field-emission properties ofvertically aligned cobalt nanowire arrays, Nano Lett.4, 521–524 (2004)

4.209 G. Dresselhaus, M.S. Dresselhaus, Z. Zhang, X. Sun,J. Ying, G. Chen: Modeling thermoelectric behaviorin Bi nano-wires, 17th Int. Conf. Thermoelectr. Proc.ICT’98, Nagoya, ed. by K. Koumoto (IEEE, Piscataway1998) pp. 43–46

4.210 L.D. Hicks, M.S. Dresselhaus: The effect of quantumwell structures on the thermoelectric figure of merit,Phys. Rev. B 47, 12727–12731 (1993)

4.211 O. Rabin, Y.-M. Lin, M.S. Dresselhaus: Anoma-lously high thermoelectric figure of merit in Bi1−xSbx

nanowires by carrier pocket alignment, Appl. Phys.Lett. 79, 81–83 (2001)

4.212 T.C. Harman, P.J. Taylor, M.P. Walsh, B.E. LaForge:Quantum dot superlattice thermoelectric materialsand devices, Science 297, 2229–2232 (2002)

4.213 X. Duan, Y. Huang, R. Agarwal, C.M. Lieber: Single-nanowire electrically driven lasers, Nature 421, 241(2003)

4.214 F. Qian, Y. Li, S. Gradecak, D.L. Wang, C.J. Barrelet,C.M. Lieber: Gallium nitride-based nanowire radialheterostructures for nanophotonics, Nano Lett. 4,1975–1979 (2004)

4.215 V. Dneprovskii, E. Zhukov, V. Karavanskii, V. Poborchii,I. Salamatini: Nonlinear optical properties of semi-conductor quantum wires, Superlattice. Microst.23(6), 1217–1221 (1998)

4.216 J.C. Johnson, K.P. Knutsen, H. Yan, M. Law, Y. Zhang,P. Yang, R.J. Saykally: Ultrafast carrier dynamics insingle ZnO nanowire and nanoribbon lasers, NanoLett. 4, 197–204 (2004)

4.217 J.X. Ding, J.A. Zapien, W.W. Chen, Y. Lifshitz, S.T. Lee,X.M. Meng: Lasing in ZnS nanowires grown on anodicaluminum oxide templates, Appl. Phys. Lett. 85, 2361(2004)

4.218 J.C. Johnson, H.-J. Choi, K.P. Knutsen, R.D. Schaller,P. Yang, R.J. Saykally: Single gallium ni-tride nanowire lasers, Nat. Mater. 1, 106–110(2002)

4.219 H.J. Choi, J.C. Johnson, R. He, S.K. Lee, F. Kim,P. Pauzauskie, J. Goldberger, R.J. Saykally, P. Yang:Self-organized GaN quantum wire UV lasers, J. Phys.Chem. B 107, 8721–8725 (2003)

4.220 C.J. Barrelet, A.B. Greytak, C.M. Lieber: Nanowirephotonic circuit elements, Nano Lett. 4, 1981–1985(2004)

4.221 M. Law, D.J. Sirbuly, J.C. Johnson, J. Goldberger,R.J. Saykally, P. Yang: Ultralong nanoribbon wave-guides for sub-wavelength photonics integration,Science 305, 1269–1273 (2004)

4.222 H. Kind, H. Yan, B. Messer, M. Law, P. Yang:Nanowire ultraviolet photodetectors and opticalswitches, Adv. Mater. 14, 158–160 (2002)

4.223 B.M.I. van der Zande, M.R. Böhmer, L.G.J. Fokkink,C. Schöneberger: Colloidal dispersions of gold rods:synthesis and optical properties, Langmuir 16, 451–458 (2000)

4.224 B.M.I. van der Zande, G.J.M. Koper, H.N.W. Lekker-kerker: Alignment of rod-shaped gold particles by

PartA

4


electric fields, J. Phys. Chem. B 103, 5754–5760(1999)

4.225 W.U. Huynh, J.J. Dittmer, A.P. Alivisatos: Hybridnanorod-polymer solar cells, Science 295, 2425–2427(2002)

4.226 L.A. Bauer, N.S. Birenbaum, G.J. Meyer: Biologi-cal applications of high aspect ratio nanoparticles,J. Mater. Chem. 14, 517–526 (2004)

4.227 Y. Cui, Q. Wei, H. Park, C. Lieber: Nanowire nanosen-sors for highly sensitive and selective detection ofbiological and chemical species, Science 293, 1289–1292 (2001)

4.228 J. Hahm, C. Lieber: Direct ultra-sensitive electri-cal detection of DNA and DNA sequence variationsusing nanowire nanosensors, Nano Lett. 4, 51–54(2004)

PartA

4

Date post:	01-Apr-2018
Category:	Documents
Upload:	doantuyen
View:	227 times
Download:	1 times

119 Nanowires 4. Nanowires - UFAMhome.ufam.edu.br/berti/nanomateriais/Nanowires.pdf · 119...

Documents